II III
ARMY CONSTRUCTION
ENGINEERING RESEARCH LAB
U.S. ENVIRONMENTAL
PROTECTION AGENCY
University of Pittsburgh
PROCEEDINGS:
FIRST NATIONAL
SYMPOSIUM/WORKSHOP
ON ROTATING BIOLOGICAL
CONTACTOR TECHNOLOGY
HELD AT CHAMPION, PENNSYLVANIA
FEDRUARY 4-6, 1980
VOLUME I
SPONSORED BY UNIVERSITY OF PITTSBURG
IN COOPERATION WITH
US ARMY CONSTRUCTION ENGINEERING RESEARCH LABORATORY
AND USEPA MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
EDITED BY ED D. SMITH, R. D. MILLER, AND Y. C. WU
JUNE. 1980
-------
SYMPOSIUM ORGANIZING COMMITTEE
Yeun C. Wu (Chairman)
Department of Civil Engineering
University of Pittsburgh
Ed D. Smith
Environmental Division
U.S. Army Construction Engineering
Research Laboratory
E. J. Opatken
Wastewater Research Division
U.S. Environmental Protection Agency
R. D. Miller
Environmental Health Engineering Branch
U.S. Army Environmental Hygiene Agency
J. A. Borchardt
Department of Civil Engineering
University of Michigan
-------
DISCLAIMER
These proceedings have been reviewed by the US Army Construction
Engineering Research Laboratory, the University of Pittsburgh and the
US Environmental Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the views and policies
of the US Army Construction Engineering Research Laboratory, the University
of Pittsburgh or the US Environmental Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for
use.
-------
FOREWORD
As our population increases, and our costs for treating wastewater
increase, greater emphasis is being placed upon the development of simple to
operate, effective, reliable, energy conservative, and efficient pollution
abatement technology. Rotating Biological Contactor (RBC) technology meets
these requirements. Too often, however, RBC technology has not even been con-
sidered as a viable treatment technology wherever wastewater pollution abate-
ment problems exist.
The objective of this Symposium/Workshop was to provide a forum and focal
point in which an interdisciplinary group of scientists, engineers, planners,
academicians, researchers, consultants, and sewage treatment plant operators
could exchange ideas, present technical information, define the problems,
assess the state of knowledge and identify the research needs regarding rotat-
ing biological contactors.
More than 345 participants representing a wide variety of experiences and
viewpoints attended the symposium. Many of these attendees expressed satis-
faction that for the first time, they had had the opportunity for extended
face-to-face communication with persons with widely different perspectives
about what the RBC problems are and how they might be resolved.
This event provided a unique national platform for the presentation of
new knowledge and the most advanced thinking on all aspects of RBC technology.
The fact that the Symposium/Workshop included 68 papers and that participants
traveled from Japan, Canada, Switzerland, Italy, Belgium, Sweden, Denmark,
France, and Norway, is testimony that significant global interest exists in
RBC technology; that RBC technology is being applied to a broad range of func-
tions; that application of RBC technology is increasing; and that RBC applica-
tions are important solutions to wastewater pollution abatement efforts.
The symposium focused mainly upon municipal and industrial RBC applica-
tion with speakers presenting papers ranging from the theoretical to the
highly practical. In particular, the intent was to:
1. Provide a history, overview, and perspective,
2. Present research results,
3. Present practical experiences,
4. Provide a forum for discussing the problems of compliance with pollu-
tion abatement through use of the RBC process,
5. Encourage information exchange and the transfer of technology, and
6. Identify research needs.
ill
-------
To accomplish these goals, the symposium was designed to provide informa-
tion concerning various aspects of the theory, design, operation, and evalua-
tion of the RBC treatment system. This information should significantly
improve the state-of-the-art understanding of the RBC process, thus optimizing
treatment performance. Moreover, it is hoped that through the definition of
specific research needs, a large portion of interested research talents in the
environmental engineering profession will be diverted to RBC scrutiny. More
importantly, it is hoped that RBC technology will be considered as a waste
treatment option whenever a wastewater pollution problem exists. It was cal-
culated that these 68 papers represent a doubling of the state of knowledge
for RBC technology.
No attempt has been made to edit, reformat or alter the material provided
except for printing production requirements or where obvious errors or
discrepancies have been detected. Any statements or views here presented are
totally those of the speakers and are neither condoned nor rejected by the
Symposium/Workshop co-sponsors.
Ed D. Smith, Ph.D.
Yeun C. Wu, Ph.D.
Roy D. Miller, Ph.D.
Co-editors
±v
-------
ABSTRACT
This document is a compilation of 68 papers presented at the First
National Symposium/Workshop on Rotating Biological Contactor (RBC) Technology
sponsored by the University of Pittsburg in cooperation with the US Army Con-
National Symposium/Workshop on Rotating Biological Contactor (RBC) Techr
sponsored by the University of Pittsburg in cooperation with the US Arm^
struction Engineering Research Laboratory (Champaign, IL) and the USEPA
of Reseach and Development's Municipal Engineering Research Laboratory (
Army
'leering Kesearcn Lctuoratory ^oriciinpaigri, IL; ana tne uSEPA Office
Development's Municipal Engineering Research Laboratory (Cin-
cinnati, Ohio). The Symposium/Wokshop was held 4-6 Feb 80 at Champion,
Pennsylvania.
The papers presented in the three-day Symposium and the findings of the
research needs Workshop comprise the major portion of these proceedings.
Question and answer sessions preceeding each paper and a list of participants
are provided.
The Symposium/Workshop proceedings will document pesent knowledge regard-
ing RBC technology. The papers are divided into 11 major topic areas:
1. Perspective, Overview, History
2. Process Variables and Biofilm Properties
3. Municipal Wastewater Treatment
4. Biokinetc Studies
5. Air Drive and Supplemental Aeration
6. Industrial Wastewater Treatment
7. Concepts and Models
8. Upgrading Primary and Secondary Waste Treatment Systems With RBC's
9. Design and Operation
10. Nitrification and Denitrification
11. Selections and Economics
The Research Needs Workshop discussions were taped and are presented as
an appendix. These proceedings document present knowledge regarding RBC's and
are disseminated as a definition and establishment of priorities for research.
-------
ACKNOWLEDGEMENTS
Because of the space limitation, the organizing committee cannot list
all of the persons who had contributed to this Symposium/Workshop, but
we wish to mention that Dr. R. K. Jain, Chief of the Environmental Division,
USACERL, Dr. M. L. Williams, Dean of the School of Engineering, the
University of Pittsburgh, were among them. Their support and advice are
gratefully appreciated. In addition, the organizing committee deeply
appreciates keynote speaker, Dr. R. L. Bunch of the U.S. Environmental
Protection Agency for his effort and time, and also all session chairmen
who successfully monitored the paper presentations. Special thanks is
given to Ed Opatken of the Wastewater Research Division (Municipal
Environmental Research Laboratory, Office of Research and Development,
USEPA, Cincinnati, Ohio 45268) for his help in selecting a Symposium
site and for looking after numerous logistic details before, during, ,
and after the symposium. His cooperation was invaluable to the succes§
of this Symposium/Workshop.
Finally, the organizing committee would like to thank all of the
participants for their interest and enthusiasm that assured success of
the Symposium. The Symposium assistants from the Seven Springs Mountain
Resort and Department of Civil Engineering of the University of Pittsburgh
did most of the typing and clerical work. Their pleasant and efficient
assistants were instrumental in getting us through the entire period of
this Symposium. To all these people we express our greatest appreciation.
The proprietors who provided RBC exhibits are also thanked. Many others
contributed ideas, time and effort; though not all can be named, they
are warmly and sincerely thanked.
VI
-------
SYMPOSIUM CONCEPT AND DESIGN
The symposium was intended to provide a mechanism for:
1) Getting to know other people, including meeting and talking infor-
mally with other people working on similar projects. This enabled
the formation of a resource base which can increase and improve the
information available.
2) Sharing experiences, which included learning who is doing what, how
they're doing it, what problems they've encountered, and how these
problems have been overcome or are being approached.
3) Examining differences in perspective in such areas as mathematical
modeling of the RBC process, and the economics associated with the
process. It is believed that some progress toward resolving some of
these differences occurred.
4) Providing input to State and Federal agencies relative to action
deemed desirable to accelerate utilization of RBC technology.
The symposium was designed to facilitate the exchange of information and
ideas. To this end, there were "keynote" speakers for each of several topics.
The symposium featured one general session, several concurent sessions and one
workshop session.
The workshop session began with opening remarks by one person, followed
by a brief sharing of experiences by the panel members. At that point, the
floor was opened up for general discussion with anybody free to make a contri-
bution.
It is believed that the Symposium will have, as a final result, the
effect of accelerating the rate at which RBC technology is utilized as an
economically viable treatment technology in the United States.
VII
-------
Page Intentionally Blank
-------
Table of Contents
SYMPOSIUM ORGANIZING COMMITTEE ---- .... ............................... i
DISCLAIMER ................ . ---- . ..................................... i i
FOREWORD [[[ iii
ABSTRACT. ... ....................... . ................................. v
ACKNOWLEDGEMENTS ...................................... . .............. vi
SYMPOSIUM CONCEPT AND DESIGN ....................... . ......... . ....... vii
PART I: GENERAL SESSION
Keynote Address
Rotating Biological Contactors - Are All Systems Go,... .............. 1
R. L. Bunch
Kenyote Address
Technology and Public Policy ................... . ...... . ........ . ..... 5
R. K. Jain
A History of the RBC Process ......................................... 11
Ed D. Smith and J. T. Bandy
EPA Research Program for RBC ......................................... 27
E. J. Opatken
ASCE Water Pollution Management Task Committee Report on
"Rotating Biological Contactor for Secondary
Treatment" ..................................... . ......... . ......... . . 31
Shankna K. Banerji
Current Status of Municipal Wastewater Treatment With RBC
Technology in the U.S ................................................ . 53
W. H. Chesner and J. J. lannone
PART II: PROCESS VARIABLES AND BIOFILM PROPERTIES
Hydraulic Characteristics of the RBC .............. . .................. . 71
H. 01 em and R. F. Unz
Physical Factors in RBC Oxygen Transfer ............................ . . 87
B. J. Kima nd A. H. Molof
Effect of Carbon, Ammonia Nitrogen and Hydraulic Loading
Rates, RPM, and Exposed Surface Area Variations on
RBC Performance ..... . ......... . ..................................... . 103
G. Hoag and W. Hovey
Hydraulic and Organic Forcing of a Pilot-Scale RBC Unit .............. 119
L. W. Orwin and C. D. Siebenthal
Effect of Organic Loading on RBC Process Efficiency and
Fixed-Film Thickness ....................... . ......................... 137
-------
Microfauna and RBC Performance: Laboratory and Full-Scale
Systems 167
G. Hoag, W. Widmer and W. Hovey
The Characteristics of Rotating Biological Contactor Sludge 189
C. F. Ouyang
PART III: MUNICIPAL WASTEWATER TREATMENT
Data Evaluation of a Municipal RBC Installation, Kirksville,
Mi ssouri 205
R. R. Dupont and R. E. McKinney
Rotating Biological Contactor for the Treatment of Wastewater
i n India 235
A. N. Khan and V. Raman
High Salinity Wastewater Treatment Using Rotating Biological
Contactors.., 259
N. E. Kinner and P. L. Bishop
Full-Scale Rotating Biological Contactor for Secondary
Treatment and Nitrification 269
J. A. Hitdlebaugh and R. D. Miller
Nitrogen and Phosphorus Removal With Rotating Biological
Contactors 295
Robert J. Hynek and Hiroshi lemura
Operational Advantages Obtained by Incorporating a Bio-Drum
in An Activated Sludge Process 325
George R. Fisette
Evaluation of Rotating Biological Disc in a Sewage Treatment
Process in Package Plant Applications 349
Bob Joost and Mike Vesio
PART IV: BIOKINETIC STUDIES
Dynamics of Microbial Film Processes.... 365
W. G. Characklis and M. G. Trulear
Effects of Organic Loading and Mean Solids Retention Time
on Nitrification in RBC Systems. 409
F. M. Saunders, R. L. Pope, and M. A. Cruz
Role of Suspended Solids in the Kinetics of RBC Systems 433
D. F. Kincannon and S. Groves
The Kinetics of a Rotating Biological Contactor Treating
Domesti c Sewage 449
A. Pano, J. H. Reynolds, and E. J. Middlebrooks
-------
Fixed Film Nitrification Surface Reaction Kinetics and Its
Application in RBC Systems 469
C. S. Huang
A Kinetic Model for Treatment of Cheese Processing Wastewater
With a Rotating Biological Contactor 491
W. J. Mikula, J. H. Reynolds, D. 'B. George, D. B. Porcella,
and E. J. Middlebrooks
PART V: AIR DRIVE AND SUPPLEMENTAL AERATION
Aerated RBC's - What Are the Benefits 515
Kevin J. McCann and Richard A. Sullivan
Performance Evaluation of Air Driven RBC Process for Municipal
Waste Treatment 525
R. Srinivasaraghaven, C. W. Ren, and S. Liljegren
Surf act: Current Developments and Process Applications 553
J. D. Cowee and R. A. Sullivan
Plant Scale Investigation of RBC Process Supplemental Aeration....... 575
R. Srinivasaraghavan, C. W. Reh, and James Canaday
Effect of Supplemental Air on RBC Process Domestic Waste 599
J. T. Madden and R. B. Friedman
Use of Supplemental Air to Correct An Oxygen Limitation
Condition of an Operating RBC System 611
J. F. Lagnese, Jr.
Operational Experience of Oxygen-Enriched Rotating Biological
Contactors 637
J. C. Huang
PART VI: INDUSTRIAL WASTEWATER TREATMENT
Wastewater Treatability Studies for a Grassroots Chemical
Complex Using Bench Scale Rotating Biological Contactors 661
Joe C. Watt and C. J. Cahill
The Treatment of Saline Wastewaters Using a Rotating Biological
Contactor. 691
Mark E. Lang and S. L. Klemetson
RBC for Munitions Wastewater Treatment..... 711
P. 6. Chesler and 6. R. Eskelund
Removal of Waste Petroleum Derived Polynuclear Aromatic
Hydrocarbons by Rotating Biological Discs 725
John T. Tanacredi
xi
-------
Treatment of Phenol-Formaldehyde Resin Wastewater Using
Rotating Biological Contactors 733
L. W. Bracewell, D. Jenkins, and W. Cameron
Energy Recovery From Anaerobic Rotating Biological Contactor
(AnRBC) Treating High Strength Carbonaceous Wastewaters 759
A. A. Friedman and S. J. Tait, Sr.
PART VII: CONCEPTS AND MODELS
The Biological Approach to the Rotating Disc Process 791
C. G. Steiner
Factors Affecting Attachment and Development of Biological
Films on Solid Media 803
E. J. La Motta and R. F. Hickey
A Conceptual Model of RBC Performance 829
C. P. L. Grady, Jr., and H. C. Lim
Recarbonation of Wastewater Using the RBC 861
C. I. Noss, Roy D. Miller, and Ed D. Smith
Prediction of RBC Plant Performance for Municipal Wastewater
Treatment 887
Y. C. Wu, Ed D. Smith, and John Gratz
RBC Design Considerations for Industrial Wastewaters 909
A. A. Friedman
Analysis of Design of Rotating Biological Contactors 921
J. A. Mueller, J. Famularo, and J. Fitzpatrick
PART VIII: UPGRADING PRIMARY AND SECONDARY WASTE TREATMENT SYSTEMS WITH RBC
Upgrading Primary Tanks With Rotating Biological Contactors 961
0. K. Scheible and J. J. Novak
Upgrading Trickling Filter Effluents With a RBC System 997
Calvin P. C. Poon, Howard Chin, Ed D. Smith, and
W. J. Mikucki
Upgrading- Existing Waste Treatment Facilities Utilizing the
Bio-Surf Process 1015
Richard A. Sullivan and Robert J. Hynek
Rotating Biological Contactor Process for Secondary Treatment
and Nitrification Following a Trickling FiTter 1035
R. D. Miller, C. I. Noss, A. Ostrofsky, and R. S. Ryczak
xii
-------
PART IX: DESIGN AND OPERATION
A Design Approach for Rotating Biological Contactors Treating....,
Industrial Wastewaters., 1065
W. W. Eckenfelder, Jr. and L. Vandevenne
Empirical Oxygen Transfer Relation in the RBC Process 1077
B. F. Severin, H. Brociner, J. E. Dumanowski,
J. T. Su, and M. M. Garvitch
Comparison of Full Scale RBC Performance With Design Criteria 1101
C. C. Chou, R. J. Hynek, and R. A. Sullivan
First USA Air Drive RBC Units Operational Experience and
Performance Indian Creek Wastewater Treatment Plant,
Cincinnati, Ohio 1127
M. M. Sehirtzinger
Design and Operation of Two Rotating Biological Contactor
Plants at Fundy National Park, New Brunswick, Canada 1137
T. Viraraghavan, R. C. Landine, E. Winchester,
and G. H. Jenkins
The Gladstone, Michigan Experience Performance of a 1.0 MGD
RBC Plant in a Northern Michigan Community. 1147
T. C. Williams and R. J. Berner
PART X: NITRIFICATION AND DENITRIFICATION
The Effect of Organic Loading on Nitrification in RBC
Wastewater Treatment Processes 1165
K. Ito and T. Matsuo
Nitrification Enhancement Through pH Control With RBC..... 1177
J. M. Stratta and David Long
Nitrification of Municipal Wastewater Using Rotating Biological
Contactors 1193
J. C. O'Shaughnessy, F. C. Blanc, Peter Brooks,
Alan Silbovitz and Richard Stanton
Pilot Scale Studies on the Nitrification of Primary and
Secondary Effluents Using Rotating Biological Discs at the
Metropolitan Sanitary District of Greater Chicago 1221
D. R. Zenz, E. Bogusch, M. Krup, T. B. S. Prakasam,
and C. L. Hing
Use of Rotating Biological Contactors for Nitrification at
the City of Guelph Water Pollution Control Plant, Guelph,
Ontar i o, Canada 1247
P. M. Crawford
xiii
-------
Denitrification of Nitrified Municipal Wastewater Using
Rotating Biological Contactors 1275
F. C. Blanc, J. C. O'Shaughnessy, D. J. Connick,
and Donald Wood
Nitrogen Removal in Rotating Biological Contactors Without
the Use of External Carbon Source 1301
Hallvard Odegaard and B. Rusten
Nitrification/Denitrification Studies With Rotating Biological
Contactors 1319
A. 6. Smith and R. K. Khettry
PART XI: SELECTIONS AND ECONOMICS
Design Considerations for a 16 MGD RBC Treatment Facility...... 1343
W. F. Barry and J. W. Heine
An Evaluation of the Cost-Effectiveness of the Rotating
Biological Contactor Process in Combined Carbon Oxidation and
Ni tri f i cat ion Appl i cat i ons 1357
J. L. Pierce and L. A. Lundberg
Computerized Cost Effective Analysis of Fixed Film Nitrification
Systems 1383
Paul T. Sun, Steve R. Struss, and M. J. Cull inane, Jr.
Comparative Cost-Effectiveness Analysis of Rotating Biological
Contactor and Activated Sludge Processes for Carbon Oxidation 1413
L. A. Lundberg and J. L. Pierce
WORKSHOP ON RBC RESEARCH NEEDS 1429
APPENDIX A: LIST OF PARTICIPANTS 1463
APPENDIX B: FLOOR DISCUSSIONS AFTER EACH SESSION
Session 1. General Discussion. 1481
Session 2. Process Variables and Biofilm Properties 1483
Session 3. Municipal Wastewater Treatment 1489
Session 4. Biokinetic Studies 1497
Session 5. Air Drive and Supplemental Air 1501
Session 6. Industrial Wastewater Treatment..... 1515
Session 7. Concepts and Models 1519
Session 8. Upgrading Primary and Secondary Waste
Treatment Systems With RBC 1525
Session 9. Design and Operation 1533
Session 10. Nitrification and Dentrification...... 1537
Session 11. Selections and Economics 1543
-------
PART I: GENERAL SESSION
Keynote Address
ROTATING BIOLOGICAL CONTACTORS - ARE ALL SYSTEMS GO
By
Robert L. Bunch
Chief, Treatment Process Development Branch
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio ^45268
It is a real pleasure for me to be here this morning to "kick off" the
first symposium ever devoted to the rotating biological contacting process
(RBC). I am happy to see the excellent turnout for this symposium. Most of
the experience and competency in this field are represented here today. The
agenda indicates you will be quite busy in the next three days, but I am
confident it will be a pleasant and productive experience for you. I am
certain the proceedings, which will be published from this meeting, will
provide very good technical guidance to those who could not attend.
The discovery of the wheel about 5000 years ago was one of the most
important steps in man's development. Wheeled carts and wagons were much
easier and faster to pull than sledges. The wheel soon found use in mechanics
in controlling,the flow of power. The three power sources used in the Middle
Ages, animals, water and wind, were all exploited by means of the wheel. For
example, waterwheels, windmills and beasts of burden were all used to drive
millstones for grinding grain. Today's civilization would not be possible
without the wheel. In the next three days, we will be discussing the wheel
and how it can further be used to benefit mankind. It is a strange coinci-
dence that the first wheels for carts and RBC process were made of wood. In
the latter case, sheets of plywood served as discs. In fact, Dr. Buswell
back in 1929 referred to his process as the biological wheel.
One of the few pleasures of getting old is that you are able to reminisce.
You never realize how far you have progressed until you look back. It was in
-------
the middle 60*s that I visited the University of Stuttgart and saw Dr. Popel's
pilot plant and discussed the features of the RBC. Much progress has been
made in our understanding of the process. We have come a long way since I
was Project Officer for the first full-scale application of the RBC process
in the United States on municipal wastewater at Pewaukee, Wisconsin. At first
thought, it seems strange that with all the interest displayed in the RBC
systems in the early thirties that the process should lay dormant until the
middle sixties. Factors that probably prevented the early adoption were the
deep depression in the thirties, World War II in the forties, and the popular-
ity of the trickling filter process in the fifties. Lastly, most of man's
technical progress has been achieved in the last 62 years. If we divide
man's last 50,000 years of existence in lifetimes of 62 years, we would have
about 800 lifetimes. Fully 650 of these were spent in caves. Only during
the last two lifetimes was the electric motor used. The majority of all the
material goods we use today were developed in the present lifetime or the
last 62 years.
Enough of the reminiscing, let's look at the future. According to the
19f8 EPA needs survey of treatment facilities, there were 59 RBC facilities
in operation, 68 under construction and 305 required but not funded. The
growth rate is increasing geometrically. We did not come here today to praise
the RBC system nor to criticize it, but to decide where the voids are in our
knowledge on design criteria and operating conditions. If we are to take
full advantage of all the good features of this process, we need to know
these voids. Let's briefly look at some of the design considerations.
SIZING AN RBC PLANT
How do we size an RBC plant? Should the design of an RBC plant be
based on hydraulic flow or organic loading? Should contact time be a con-
sideration? The early developments of wastewater treatment processes in Eng-
land and the USA were based on domestic sewage. Since all domestic sewage,
in those days, contained about the same organic strength and the K-factors
were similar, it was easier to report data in terms of hydraulic flow and
percentage reduction. This parameter is convenient because for a given physi-
cal system, hydraulic loading will be inversely proportional to detention
time. If the first order reaction kinetics are applicable, percentage removal
will be independant of organic concentrations.
As our country became more industrialized, the domestic wastes became
mixed with industrial wastes. Today we refer to this mixture as municipal
wastewater. To account for this change, the sanitary engineers introduced a
design factor called population equivalent. This was supposed to compensate
for the differences in organic strength of wastes. This factor, however, did
not take into consideration the treatability of the wastewater nor the K-factor.
Unfortunately today we still refer to the efficiency of wastewater treatment
systems in terms of percentage reduction. Percentage reduction has little
meaning unless related to the strength and type of wastes. If a plant is
heavily loaded organically, it is possible to have a 90% reduction and still
produce a very poor effluent as opposed to one normally loaded achieving a very
low BOD with the same percentage reduction. Percentage reduction can be mis-
leading. For example, it is much easier to achieve a 90? reduction of phos-
phorus on a raw wastewater that has 10 mg P/l than one that has 5 mg P/l. An
-------
effluent standard of 1 mg P/l can be met with mineral addition but to achieve
0.5 rag P/l would probably require filtration. In the future we will be more
concerned with residuals in the effluent rather than about percentage reduc-
tion.
The measure of unit processes in the future will be cost vs residual.
In other words, how much does it cost to obtain a certain residual. For
example, to obtain an effluent containing less than 1 rag P/l costs about $50
per million gallons (3785 m3). TO achieve a residue of less than 0.5 mg P/l
increases the cost twofold and for a limit of 0.05 mg P/l the cost soars
twentyfold over the cost for 1 mg P/l.•
Since the early development work on the RBC process was done on domestic
sewage, it was natural that the hydraulic flow was first used to size the
plants. In that the basis for the RBC system is biological degradation,
should not the controlling factors be based on microbiological principles?
If so, then organic loading on the discs and oxygen mass transfer efficiencies
of the system will be the controlling factors. Many investigators have con-
cluded that the RBC system follows first order kinetics, but with varying
reaction rates with various stages.
DISC CONFIGURATION
Most RBC plants designed today have equal disc surface areas for each
successive stage of treatment. Dissolved oxygen (DO) profiles follow a pattern
of rapid initial decline and slow recovery in successive stages. Under heavy
organic loadings, the liquor from the first stage can be distinctly anaerobic.
Thus, increasing the hydraulic loading and/or organic concentration can stress
the system. Increasing the number of stages will increase the total treatment
potential or the system, reducing the stress. The excess organic material
left untreated by the first stage can be treated by the second and third
stage, etc. Having several stages can dampen out hydraulic surge and organic
slugs.
For a given disc area and disc speed, the amount of oxygen transferred'
is fixed. An overstressed system will reduce the DO below the critical con-
centration and the efficiency of the system will be drastically reduced. It
is not the relative disc size or oxygen transfer efficiency, but the absolute
oxygen transfer capability of the system which determines whether the DO will
be reduced below the critical concentration.
How do we alleviate the stress on the first stage? Should the first
stage contain more disc area than the successive stages? Is the addition of
liquid aeration a better design? Would step feeding of the system with part
of the load added to the second stage be more cost effective? The addition
of aeration would give the plant more flexibility in handling different types
of wastes. Aeration may be the answer in situations where the waste charac-
teristics are changed significantly by the addition of new industrial wastes
after the plant has been constructed. Most certainly during the next three
days disc configuration should be high on the list of topics to be discussed.
-------
DISC SIZE AND ROTATIONAL VELOCITY
Much of the data upon which mathematical models and full-scale designs
are based were obtained from small pilot plants. The assumption was made
that a unit area on a small disc is equivalent in organic removal capacity to
that of a larger diameter disc. Because peripheral velocity (tip speed) is
directly proportional to the disc diameter, both rotational velocity (RPM)
and peripheral velocity cannot be simultaneously scaled. Tip speed has been
used in most cases as the scale-up parameter.
Evidence has been accumulating in the literature that larger discs have
poorer oxygen transfer characteristics than small discs at the same tip speed.
These studies would indicate that the possibility exists of introducing vari-
able size discs in plants with smaller discs in the first stage where there
is a high oxygen demand. Are further considerations of the diameter of disc
and rotational velocity in order?
COSTS
The cost of constructing an RBC plant for a small treatment plant is
almost as great as that of an equivalent activated sludge plant. The RBC
process, being a unit modular process, does not have the scale-up advantages
that other systems do. Ways need to be found to lower construction costs
without increasing the service rate. Can cheaper material for the discs be
found? Bacteria can grow on practically any material. All that needs to be
done is rotate it. In the milder climatic portions of the USA a more open
system could be designed. Lessons can be learned from the petrochemical
industry for they have reduced cost by eliminating expensive structures.
Less costly protection of the discs from elements can be designed.
In closing, I make the plea that each one of you try to make this sympo-
sium a workshop where new ideas and unexplored needs of the rotating biological
contact process are discussed informally. The conference will be considered
a success if we can clearly set forth the present knowledge on design criteria
and define the research and development needs to fill the voids in our know-
ledge. The adaptation of wheels as gears was a conceptual leap. Engaging
wheel rims to transmit or modify motion was not obvious. Quantum improvements
in the RBC system are not obvious, but I am confident that there are among us
today many who will continue to improve the biological wheel.
-------
Keynote Address
TECHNOLOGY AND PUBLIC POLICY
by
Ravi Jain*
Mr. Chairman, ladies and gentlemen; it is indeed an honor
for me to be here. I hope to listen to your presentations this morning
and learn from you about the RBC technology. After looking at the roster
of attendees, it is clear that you are indeed a distinguished group of
participants.
I should tell you that for this keynote address I did receive
advice from a number of sources. Dr. Ed Smith, who is one of the
organizers of this conference, sent me a note and he said, "I know your
keynote address will be thought-provoking, clever, and dynamic," and
then he proceeded to attach an example keynote address from another
conference, ostensibly to assist me with preparing my remarks. The
example keynote address would have taken about 37 minutes to deliver.
I asked Dr. Wu, Chairman of the Symposium Organizing Committee, as
*Chlef, Environmental Division, USA-CERL, Champaign, IL. For the
1979-80 academic year, Dr. Jain was on leave from CERL to study Public
Administration and Policy at Harvard-.
-------
to how much time I have for my presentation. His response was: "not
much."
So, with the charge of making my presentation thought-
provoking and dynamic, and with the requirement of not using much time,
I have decided to share a few ideas with you on the characteristics of a
professional scientific community like the one you represent here and
ideas on technology and public policy.
As you know, a group of scientists, researchers, and other
professionals, like yourselves, form a unique community. This com-
munity, as Daniel Bell has stated,1 is such where the sovereignty is
not coercive and the conscience is individualistic and sharing. As an
imago (or an image), it comes closest to the ideal of the Greek polls.
Robert Merton (a philosopher and author of Sociology of Science), has
9
stated that ethos of a scientific community has four elements. Two of
these elements,that are particularly relevant to this conference, deal
with sharing of knowledge and participating in organized skepticism.
This conference is an example of a scientific community where many of
you are willing to share your knowledge and scientific discoveries and
at the same time participate in an organized skepticism in.an effort to
scrutinize,and learn from the discoveries of others. So this sets the
1. Bell, Daniel, The Coming Of Post Industrial Society. Basic Books,
New York, 1973 (P 380).
2. Merton, Robert K., The Sociology Of Science, The University of
Chicago Press, Chicago, IL, 1973 (P 270).
-------
stage and the environment for this symposium.
Next, let me briefly comment on technology and public
policy. In the context of this symposium, two aspects of technology:
technology development and technology transfer might be of interest.
Technology Development: It is clear that as scientists and researchers
you are involved in developing new technology. Technology developed
has to accommodate many conflicting requirements on our resources such
as energy and other material and human resources. The RBC technology
may well provide answers to some of these issues.
Technology Transfer: The process of technology transfer, however, is
difficult to understand if one were to look at science and technology in
the context of the 19th century when science dealt primarily with machines
and physical tools, i.e., hardware. Today technology consists increas-
ingly of "software". This software deals with procedures, methodologies,
and systematization of ways of doing things as opposed to merely speci-
fications for things. For the research organizations, technology transfer
would have two distinct components. The scientific information which the
organization obtains from the outside scientific community, this could be
referred to as an "input" and the "output" , would be the scientific con-
cepts, procedures and methodologies, developed by the organization for
the potential user. You would agree with me that any research organiza-
tion represents only a very small portion of the total scientific community;
-------
therefore, interaction with the wider scientific community is essential.
This conference can serve as a vehicle for technology transfer both on
the input side and the output side. For instance, on the input side by
helping us relate to the wider scientific community as represented by this
group, and also on the output side by documenting research results for
the use of practitioners who are also represented here.
As you know, tied to technology development and transfer
is national productivity. It is interesting to note that the United States
is far ahead of other industrialized countries in major technological
innovations. However, if one were to look at trends in productivity
represented by output per person hour, it is quite a different story. The
U.S. productivity gain between 1960 - 76 was the smallest of the other
o
major industrialized countries. Some figures for the U.S. for 1979 show
a decline in productivity which is quite alarming. It is possible that a
lack of sufficient investment for transferring and implementing new tech-
nology accounts for, to some degree, this decline in productivity.
I would suggest that your effort here would help immensely
towards bridging this gap between technology development and tech-
nology transfer.
The last item I would like to discuss with you is public
alicy as it relates to the environmental issues. While most of us
3. Science Indicators - 1976, National Science Board, National
Science Foundation, 1977 (P 35).
-------
understand that public policy affects the extent to which resources are
available for clean water, clean air and environmental protection and
also resources for developing technology necessary for achieving these
goals, let me suggest, we often ignore the effect technology has on
policy. If clean water, clean air and other environmental amentities
are important, then, it is essential that when the trade-offs between
conflicting demands on resources are made and public policy decisions
reached that people like yourselves — who are knowledgeable not only
in the environmental issues but also in economic and social issues — be
Involved in these deliberations. This is essential if your knowledge
of not only existing technology but also emerging technology is to posi-
tively affect national policy. I believe more can be done in this area.
I would simply like to close my address by commending the
many sponsors and organizers of this symposium. Organizing a symposium
like this requires considerable effort. Many of you who have organized
similar activities know exactly what I am talking about. I hope to listen
to your papers and get an opportunity to exchange ideas with you. As
Dr. Wu mentioned, many of our participants have come here from other
countries; all of you have come here leaving behind other important com-
mitments. Your willingness to share your ideas and participate in this
symposium are commendable.
-------
Page Intentionally Blank
-------
A HISTORY OF THE ROTATING BIOLOGICAL
CONTACTOR PROCESS
By
E. D. Smith
Environmental Engineer, U.S. Army Construction Engineering
Research Laboratory, Champaign, IL 61820
J. T. Bandy
Environmental Engineer, U.S. Army Construction Engineering
Research Laboratory, Champaign, IL 61820
Introduction
In comparison to many other sewage treatment technologies (e.g.,
activated sludge and trickling filters) relatively few dollar and manpower
research resources have been spent studying Rotating Biological Contactor
(RBC) technology. The resultant lack of knowledge is due, in part, to the
fact that the RBC process is relatively new in the United States. In fact,
only a few RBC plants have been operational for more than a few years. Most
of these are utilized for secondary sewage treatment with a few used for
upgrading existing sewage treatment plants (nitrification and denitrification)
or industrial waste treatment applications. Although millions of dollars have
been spent by American industries and municipalities for RBC process equip-
ment, the latest wastewater treatment guidance documents reveal a conspicuous
lack of information regarding the RBC unit process. For instance, many excel-
lent documents which provide design and operation and maintenance
criteria/guidelines are readily available for traditional technologies such as
the activated sludge and trickling filter processes. An example of such a
11
-------
publ*cation is the excellent EPA report - Process^ Control Manual for Aerobic
Wastewater Treatment Facilities (1). The purpose of the publication is to
provide guidance to optimize the performance and to help establish process
control techniques for trickling filter and activated sludge systems. There
is no comparable manual for 1BC technology. Other examples which demonstrate
the novel nature of KBC technology in the United States are two excellent EPA
documents - (1) Upgrading^ TricklingFilters (2) and (2) Process Design Manual
forUpgrading Existing Wastewater Treatment Plants (3). They do not mention
KBC technology. In addition, commonly used "state-of-the-knowledge" documents
which are designed as guidance for the selection of wastewater treatment sys-
tems based upon economic considerations either do no*" have KBC cost curves
(capital, O&M, energy, etc.) or the curves are dated. Because of the relative
newness of KBC technology in the U.S., guidance is scarce with regard to KBC
applicability, design, O&M and economic considerations.
The lack of empirical data and guidelines is complicated by the fact that
there is no well-defined theory of design and operation accepted by all RBC
manufacturers. Activated sludge, trickling filter and most other wastewater
treatment processes may be designed and constructed without significant depen-
dence upon equipment proprietors. This is not the case with KBC technology.
Design engineers who have selected KBC technology are extremely dependent upon
proprietors' design curves. This situation is compounded by the fact that the
various BBC proprietors/manufacturers have differing philosophies of design
and varying media densities, structure, etc.
Status of RBC TechnologyToday
RBC technology has been very popular in Europe for many years, and
recently, in the U.S., has become increasingly popular for both municipal and
industrial utilization. The extent and magnitude of interest regarding RBC
technology becomes immediately evident when one contemplates the recent number
of publications reporting KBC related research and operations experience.
This symposium is further evidence of the interest the various sectors
(private, academic, research, government agency, regulatory, A/E, professional
organization, design engineer, industrial, and plant operators) have regarding
RBC technology. All of the above and other professionals involved in wastewa-
ter treatment and management are represented at this symposium. Other man-
ifestations of interest with KBC technology include the followingj
1. The American Society of Civil Engineers (ASCE) has formed a "Rotating
Biological Contactor Task Committee." (A report from this committee is
scheduled in this symposium).
2. The U.S. Army Construction Engineering Research Laboratory and the
U.S. Army Medical Bioengineering Research and Development Laboratory is inves-
tigating RBC technology applicability for upgrading existing Army sewage
treatment plants.
12
-------
3. The federal Highway Administration, U.S. Department of Transportation
has cooperated in research associated with the RBC process as a treatment
method for wastewaters of roadside packs (4).
4* An ad hoc committee has been formed to evaluate the applicability of
KBC technology for the People's Republic of China.
5. Proprietors/manufacturers of RBC equipment have increased dramati-
cally during the last few years. Many of them are represented at this sympo-
sium.
Legislative Requirements
In 1972, Congress initiated a comprehensive program to restore and main-
tain the quality of the nation's rivers and lakes by passing amendments to the
Federal Water Pollution Control Act (P.L. 92-500). The 1977 Clean Water Act
(P.L. 95-217) reaffirmed this commitment through additional amendments which
strengthened a number of the provisions of P.L. 92-500. These two laws
require that industrial and municipal waste treatment operations constrain
their point source wastewater effluents within prescribed limits of quality.
In fact, certain mandatory penalties are stipulated and are enforced by the
EPA.
Several wastewater technologies are available as candidate mechanisms for
meeting these secondary or even more stringent NPDES permit stipulations.
Each of these technologies exhibits its own inherent technical and
economic attributes (advantages/disadvantages, etc.). There has been a recent
tendency among consulting firms to choose the more capital- and energy-
intensive, and the more complex technology should be used when it is applica-
ble to particular wastewater problems. However, it is more sensible to choose
simple to operate, economical and reliable technology whenever possible. RBC
technology is conducive to meeting these requirements. In particular, if one
evaluates (and compares to other processes) RBC energy scenarios,
operational/maintenance requirements, efficiency, reliability under various
environmental and loading conditions, it becomes evident that RBC technology
should be considered as an option whenever municipal and industrial pollution
abatement is required.
Personnel interested in considering the RBC process as an option are
faced with finding answers to the following questions:
1. How can I insure that 1BC technology is right for my particular
situation?
2. How much does it cost?
3. Are the RBC units easy to install and start up? What about site
preparation?
13
-------
4« Can we obtain the process and install it into a. tight compliance
schedule.
5» What are the RBC's operational and maintenance problems/costs?
6« How does RBC technology compare with other technologies?
7» Is the process reliable and effective under a variety of climatic
conditions and under hydraulic, organic, and ammonia loadings?
8. What are the appropriate design criteria?
9» What are the system's land requirements?
10. What are its skill and manpower requirements?
11. What are the process advantages/disadvantages?
12. Can the process be retrofitted to existing secondary equipment to
meet biochemical oxygen demand (BOD), suspended solids (SS), and ammonia
requirements?
13. What industrial pollutants will RBC technology successfully treat?
14. What about nuisances (odors, filter flies)?
15. How does energy consumption compare to other processes?
16. What are the sludge characteristics?
17. What is the need for clarification prior to disinfection and
discharge, and what design criteria are appropriate for clarification?
18. What are the life expectancies of major control components?
19. What new developments are anticipated for RBC technology?
20. What is the effect of extremely low temperatures?
21. What are the safety considerations?
22. What information is available?
23. What are the opinions of RBC plant operators? What kind of problems
can I expect?
14
-------
Need for an RBC State-of-the-Knowled%e Definition
It is anticipated that this symposium will answer many of these questions
by providing a state-of-the-knowledge definition of RBC technology. At the
very least, the symposium will provide a forum for identifying problems. This
problem definition will be interpreted into a prioritized list of research
needs. As the research is performed, the design and 0/M problems may be
solved with a resultant increased popularity of RBC technology.
History of the Rotating Biological Contactor
According to a recent EPA report (5):
The RBC concept of treating waste streams biologically
has been known for many years, but it was not until
strong, lightweight plastics become available that signi-
ficant interest in the technique began to develop. The
treatment technique is to grow biologically active masses
. on a series of discs that slowly rotate, alternately
exposing the biomass to the air and to the wastewater.
In early models, the discs were made of metal and were heavy,
cumbersome, and subject to corrosion. Recent models have
discs fabricated of polyethylene or polystyrene. Many inves-
tigators have found advantages for the RBC over activated
sludge or other conventional treatment systems based on spe-
cialized circumstances...
Historical information is also provided in a Civil Engineering article
(6) titled "Behind the Rapid Rise of the Rotating Biological Contactor":
The rotating biological contactor goes back to the 1920s.
Investigators in both the U.S. and Germany experimented
with using rotating wood surfaces. But wood surfaces
were impractical to manufacture and deterioriated, and in
those days, few communities were putting in secondary
treatment.
Not much more happened until the 1950s. In that decade,
investigators at Stuttgart University, West Germany,.
attempting to improve the secondary treatment process,
experimented with wooden and plastic flat disks rotating
in wastewater.
In 1959, J. Conrad Stengelin began to manufacture 2 and 3
meter diameter expanded polystyrene disks in West Ger-
many. The first commercial installation went on stream
there in 1960. But the rotating disk process was not
cost competitive with the activated sludge process ; ini-
tial capital costs were considerably more than for
15
-------
sludge plants. Nonetheless, many small plants
were installed in Germany in the 1960s — most serving
less than 1000 people. These small municipalities were
willing to pay more in initial cost to get a plant
requiring little maintenance and low energy consumption.
After 1960, further development of the rotating biologi-
cal contactor stopped in Europe. But between 1960 and
1965 in the U.S., Allis-Chalmers did much development of
rotating disks.
In 1970, Allis-Chalmers sold its rotating biological
contactor technology to the Autotrol Corp. (Milwaukee,
Wise). At that time, the polystyrene disks were still
not competitive with the activated sludge process. Even
as late as 1972, Autotrol had sold only a few RBC instal-
lations for sewage treatment. The capital cost of the
polystyrene disks was simply too high.
Breakthrough Sparks Growth of RBCs
Then, in 1972, came an important breakthrough: the
development of a more compact disk, one with much more
surface area for a given volume. Until then, the RBC
unit consisted of a series of parallel, flat 0.5-in.-
thick expanded polystyrene sheets, each separated by a
0.75-in. space. Now, Autotrol came out with an arrange-
ment of l/16-in.-thick polyethylene sheets with a 1.2-in.
space separating them filled with a honeycombed
polyethylene configuration. Whereas the standard poly-
styrene RBC unit was 10 ft in diameter and 17 ft long
with 21,000 ft of surface area, the new polyethylene RBC
unit was 12 ft in diam., 25 ft long, with 100,000 ft of
surface area. In recent years, Autotrol has developed a
still more compact arrangement for nitrification applica-
tions — the distance between adjacent polyethylene
sheets being only 0.6 in., with total surface area of a
standard RBC being 150,000 ft ...
The use of RBC technology in Europe (particularly in Germany) has been
quite extensive, and over 700 installations (some with more than 25 years of
experience) are presently in operation. Most of them are small, but the muni-
cipal plant at Ponavischigen, West Germany, serves about 100,000 persons (7).
Since 1972, the number of wastewater treatment facilities in the United
States utilizing rotating biological contactors has increased more than 300,
with another 300 now in the planning stages (8). An excellent historical
review of the RBC process can be found in Ph.D. thesis of C. G. Grieves sub-
mitted to Clemson University.
16
-------
RBC Literature Review
The following literature review provides information concerning various
aspects of theory, design, and operating experience associated with RBC sys-
tems.
Historically, rotating biological contactors have been used to remove
organic carbon from wastewater. This process was later expanded to include
nitrification and denitrification of wastewater. One of the earliest reports
of RBC application in the United States is by Welch (9), who successfully
treated highly concentrated wastes using an RBC system installed at Allis-
Chalmers, West Allis, WI. In terms of chemical oxygen demand (COD), as much
as 800 lb/1000 cu ft/day (1.28 kg/m /day) removal was recorded. Torpey, et
al. (10) reported a 10-stage RBC with aluminum disks which decreased BOD from
124 mg/1 in the influent to 9 mg/1 in the effluent after 5 months. Nitrifica-
tion also occurred, which reduced the ammonia nitrogen content of the effluent
(NH -N) from 14.2 mg/1 to 5.7 mg/1 and correspondingly increased the nitrate
from zero to 10.4 mg/1 in the effluent. Antonie (11), in his study of the RBC
process response to fluctuating flow, reported significant chemical oxygen
demand (COD) removal when the hydraulic residence time of wastewater was
approximately 60 minutes. Hydraulic surge, which reduced the residence time
to 30 minutes or less, resulted in low COD reductions. In a later report,
Antonie (12) noted successful applications of the RBC process for treating
various food and nonfood processing wastes. In an EPA demonstration project
using the RBC system as a full-scale secondary treatment plant, Antonie (13)
reported good BOD removal and some nitrification. In the winter, the system
was placed in an enclosure to protect the biomass from freezing temperatures.
In a pilot study conducted by LaBella, et al. (14) it was reported that the
RBC process at a hydraulic loading of 1 gal/sq ft/day (0.04 m /m /day) could
remove BOD from winery wastes at an efficiency comparable to that of an
activated sludge process. However, the yearly operating cost of the RBC pro-
cess was found to be $6000 per year less than the activated sludge process for
a flow of 0.34 to 0.44 MGD (1290 to 1665 m /day). Chittenden, et al. (15)
also used the RBC system to treat anaerobic lagoon effluents. At a hydraulic
loading of 4.0 gpd/sq ft/day (0.16 m /m /day), increasing the rotating speed
of the first stage to 6 rpm produced a 79.5 percent BOD reduction and an
overall BOD reduction of 83*2 percent from an influent having an average of
225 mg/1 BOD. Higher hydraulic loading and lower rotational speeds resulted
in poor efficiency of BOD removal and little or no dissolved oxygen in the
system. Using a synthetic wastewater for an RBC process study, Stover, et al.
(16) reported that more than 90 percent COD removal was possible as long as
the organic loading was kept below approximately 400 lb/1000 cu ft/day (0.64
kg/m /day). Using the same RBC system for slaughterhouse waste treatment,
only 70 percent COD removal was achieved, even though the organic loading was
low at 100 lb/1000 cu ft/day (0.64 kg/m /day). Increasing the loading to 400
lb/1000 cu ft/day (0.64 kg/m /day) reduced removal efficiency to 15 percent.
Expressed in Ib COD/day/1000 sq ft of disk surface area, the maximum COD remo-
val for slaughterhouse waste was approximately 4.0 Ib COD/day/1000 sq ft (19.5
g/m /day) at loadings of 8 Ib COD/day/1000 sq ft (39 g/m /day) or higher. An
investigation (17) to determine the efficiency of the RBC process on raw
wastewater from a liquid detergent manufacturing plant was performed.
17
-------
Selected parameters were chosen for measurement, including (COD), COD, BOD,
MBAS and DO.
Applications of the RBC process for nitrification of wastewater or sludge
supernatant have been reported. Weng, et al. (18) evaluated various parame-
ters affecting the process performance and showed that among influent loading,
flow rate, rotational disk speed, detention time, effective disk surface area,
and submerged disk depth, only influent loading, flow rate, and effective disk
surface area were important in determining nitrification efficiency (tempera-
ture steady at 20°C and disk rotating speed at 10.5 or more rpm). In effect,
NH»-N loading was the only controlling factor.
Antonie (19) reported that at various treatment plants using the Bio-surf
RBC process, as much as 0.8 Ib NH -N/day/1000 sq ft (3.9 g/m /day) could be
removed. Generally, 90 to 95 percent nitrification was obtainable. A pilot
plant study conducted by Hao, et al. (20) showed excellent NH -N removal at
the Columbus, Indiana, sewage treatment plant. In January and February, when
cold temperatures prevailed, 50 to 60 percent NH--N removal was obtained at a
hydraulic loading of 2.5 gpd/sq ft (0.013 m/m /day) and 90 to 95 percent
NH,-N removal at 1.5 gpd/sq ft (0.06 m /m /day). When high strength ammonia
wastewater (780 mg/1 NH -N on the average) was applied to a four-stage RBC
system, Lue-Hing, et al. (21) found that at an overall NH -N loading of 15.6
Ib of NH.-N/day/1000 cu ft (25 g/m /day) and a wastewater temperature of 10°C,
99.4 percent of the NH.-N was removed; at an overall loading of 43.5 Ib of
NH2-N/day/1000 cu ft (70 g/m /day) and a wastewater temperature of 20°C, 99.8
percent of the NH.-N was removed. The maximum removal rates in the first
stage ranged from 95 Ib of NH -N/day/1000 cu ft (152 g/m /day) at 9°C to 170
Ib of NH -N/day/1000 cu ft (272 g/m /day). Recirculation of effluent in the
RBC process showed insignificant improvement of nitrification.
Temperature sensitivities of the RBC system have been evaluated by
Murphy, et al. (22) over a range of 5 to 25 C. For both nitrification and
denitrification, RBC temperature sensitivities were reported to be similar to
those of suspended growth systems having long sludge retention times.
With more than 4 months of RBC nitrification study at the Belmont Waste-
water Treatment Plant at Indianapolis, Indiana; Reid, Quebe, Allison, Wilcox
and Associates, Inc. (23) reported that although the RBC process appeared a
feasible alternative nitrification process for waste containing relatively
consistent NH -N loadings, the process was unable to consistently maintain low
(less than 1.0 mg/1) NH -N levels in the effluent when the influent NH--N load
varied. In the same study, it was found that the RBC system could reduce the
total BOD_ (carbonaceous portion only) in the clarified activated sludge
effluent from 8 to 18 mg/1 to 6 to 13 mg/1. The percentage of BOD^ removal
was low (0 to 57 percent) compared to the secondary treatment process. How-
ever, soluble carbonaceous BOD removal was more successful (1 to 10 mg/1 to 1
to 3 mg/1, or 0 to 80 percent removal). By removing a portion of the treated
effluent total suspended solids (TSS), an effluent with BOD less than 10 mg/1
can be obtained with no difficulty.
18
-------
Other studies also indicate the inability of the RBC system to remove
total BOD_. leh, et al. (24), Lagnese (25), and Sullivan, et al. (26) col-
lected ana analyzed operational data from various full-scale RBC plants and
concluded that design of RBC systems should be based on soluble BOD. loading,
rather than on total BOD- loading. In using the RBC system for upgrading
existing secondary treatment plants and for tertiary treatment, it is impor-
tant to recognize the inability to remove particulate BOD_, particularly when
the particulate portion of the total BOD,, is high. When the RBC unit is
operated in series and following secondary treatment, a less efficient perfor-
mance can be expected, since the wastewater contains a higher fraction of
refractory organics. Finally, nitrified effluent from the RBC unit contains
nitrogeneous oxygen demand (NOD) which can be a significant portion of the
effluent BOD,.. Lagnese (27) suggested that a nitrification inhibitor be used
in the BOD,, analysis to eliminate NOD from the analysis. However, this
approach may require some revision or clarification of the NPDES permit.
Important RBC design considerations include the characteristics of wastewater
to be treated and the degree of treatment desired. These considerations dic-
tate such system parameters as number of stages, speed of RBC rotation, reac-
tion tank volume, media density, and pretreatment.
According to a literature search performed by Griffith, et al. (28), sys-
tems treating municipal wastewater usually provide for two to four stages for
secondary treatment and up to 10 stages if further treatment is required.
Disk rotation velocities of 1 fps (peripheral velocity) are common for initial
stages, with lower velocities (0.5 fps) used in later stages as the oxygen
demand in the wastewater is reduced. Disk reaction tank volumes which provide
0.12 gal/sq ft (4.89 1/m ) of disk (including disk volume), or 1-hour deten-
tion time, at a hydraulic loading rate of 0.06 m /day/m (1.5 gpd/sq ft) of
disk area, are common. A wide range of hydraulic and organic loading rates
has been reported for systems treating domestic wastewater. Hydraulic loading
rates ranging from 0.004 to 0.17 m /day/m (0.09 to 44.1 gpd/sq ft) of disk
surface area and organic loading rates of 0.20 to 6.0 Ib BOD per day/1000 sq
ft (0.98 to 2.93 g/m ) of disk surface area are documented. Systems having
disks aligned parallel to the direction of flow and perpendicular to the
direction of flow have been described. The disk reaction tank is generally
contoured to the shape of the disks, which improves mixing of the wastewater
within each stage. Documented disk materials include aluminum, polystyrene,
polyethylene, and plexiglas. Desirable properties in a disk material are low
density and rigid shape. Disk diameters range from 6 in. to 12 ft (15.2 cm to
365.8 cm), with spacing between disks ranging from 3/8 to 3/4 in. (0.96 cm to
1.9 cm). The disk is generally immersed in the wastewater to between 40 and
50 percent of its diameter, with the only criterion being that its entire sur-
face becomes wet.
Final solids removal facilities are generally incorporated into the total
treatment scheme. A biomass generation of approximately 0.4 Ib (.16 kg) of
dry solids per pound of BOD removal has been reported. Systems used to tran-
sport the settled biological solids to storage and treatment facilities
include screw conveyors, scraper/bucket schemes, and pumps.
19
-------
Description of Modern Process (29)
In. its present form, the rotating biological contactor process consists
of a series of closely spaced discs (10-12 feet in diameter) mounted on a hor-
izontal shaft and rotated while about one half their surface area is immersed
in wastewater. The media commonly used in Europe and originally introduced
into the U.S. consists of a series of parallel, closely spaced flat discs.
Now many U.S. manufacturers offer media with a lattice structure. This more
complex structure offers more surface area per unit volume.
When the process is placed in operation, the microbes in the wastewater
begin to adhere to the rotating surfaces and grow there until the entire sur-
face area of the discs is covered with a 1/16 to 1/8 inch layer of biological
slimes. As the discs rotate, they carry a film of wastewater into the air,
where it trickles down the surface of the discs, absorbing oxygen. As the
dises complete their rotation, this film mixes with the reservoir of wastewa-
ter, adding to the oxygen in the reservoir and mixing the treated and par-
tially treated wastewater. As the attached microbes pass through the reser-
voir, they absorb other organics for breakdown. The excess growth of microbes
is sheared from the discs as they move through the reservoir. These dislodged
organisms are kept in suspension by the moving discs. Thus, the discs serve
several purposes. They provide media for the buildup of attached microbial
growth, bring the growth into contact with the wastewater, and aerate the
wastewater and suspended microbial growth in the wastewater reservoir. The
speed of rotation is adjustable.
The foregoing description was excerpted from a recent EPA technology
transfer publication. It describes the operation of RBCs for organics removal
and, with minor modifications, for nitrification. When RBCs are used for den-
itrification, the entire disc is submerged and rotation provides mixing but
not oxygen exchange.
The rotating biological contactor process, like any other treatment tech-
nology, has inherent advantages and disadvantages of which prospective users
should be aware.
Advantages
Rotating biological contactors have a number of characteristics which
commend them to the design engineer. They can provide a very high degree of
treatment. They require less area than most other comparable processes. They
can be retrofit easily to existing plants.
RBCs show high efficiency in oxygen transfer. They handle organic over-
loading well due to the large biomass on the discs. Since they involve
attached growth, they are much less likely to fail through washout when condi-
tions adverse to the biological growth occur. There is no bulking, foaming,
or floating of sludge to interfere with a plant's overall efficiency. Short
circuiting in the biological reactor cannot occur.
20
-------
In laying out a plant, IBCs offer advantages beyond their relatively low
area requirements. Because most RBC units operate with a net increase in
hydraulic head, pumping which would otherwise have been needed may be obvi-
ated. Less excavation is required for RBCs than for activated sludge aeration
tanks, a characteristic of the process which is especially helpful in high
water table areas. Finally, RBCs are versatile both in the functions which
they perform and in the flexibility with which they can be configured. There
is even a choice in the methods for rotating the discs. Mechanical drive
units can be employed or an air drive mechanism can be used which has fewer
moving parts and which uses less energy.
However the discs are rotated, RBC technology uses up to 50 percent less
energy than activated sludge units. Over the lifetime of a plant, this can be
a very important advantage. The low speed of the mechanical drive units
reduces their maintenance requirements and prolongs their lives.
Rotating biological contactors are simple to operate. There are no
sludge or effluent recycle streams, although recycle has been shown to be
advantageous in some applications. The sloughed biomass settles well and can
be more reliably removed than the solids from an activated sludge tank. Cla-
rifier design and operation, which frequently limits the performance of plants
relying on other processes, is far less critical in 1BC installations.
Because Rotating Biological Contactor treatment is simple and stable with
respect to most potentially upsetting fluctuations in influent flow and qual-
ity, it requires fewer process decisions by the operator than do activated
sludge processes. Thus, satisfactory operation can be achieved with less
highly skilled personnel than are needed for activated sludge. This factor
could mean considerable savings in operating a treatment plant as well as mak-
ing the operation of a plant more predictable. Since less than optimal
operating procedures have been cited as a leading cause of plants failing to
achieve the results for which they were designed, the advantages of the simple
RBC process may be greater in practice than a comparison of design perfor-
mances would indicate. The rotating disc process lends itself well to upgrad-
ing existing treatment facilities. Because of its modular construction, low
head loss and shallow excavation, it can be installed to follow existing pri-
mary treatment plants, including Imhoff tanks and septic tanks.
Disadvantages
RBC technology is not without its share of problems. The oldest U.S.
plants have been in operation for only 7 years. The structural integrity of
RBC units is untested by time. In one instance, plastic media tore loose from
its drive shaft. It has been a common experience for tie rods to loosen and
cause uneven rotation and need for realignment. Oil leaks from drive units
are common.
Although low maintenance costs are often cited as an RBC advantage, these
costs are strictly proportional to plant capacity, exhibiting none of the
economies of scale observed with other non-modular technologies. Similarly,
21
-------
area requirements for RBC installations are proportional to plant capacity so
that an RBC advantage for small- and medium-scale plants becomes a liability
in very large capacity applications. The use of air drives reduces the rela-
tionship between plant size and maintenance costs because each shaft does not
require a separate electromechanical drive. Air drives avoid another disad-
vantage which has been cited for RBCs ; rotational speed can be continuously
adjusted by turning a few valves. Altering the rotational speed of elec-
tromechanical drives requires modifying each drive unit. A large plant may
have dozens.
Enclosures are necessary where very low air and wastewater temperatures
occur in order to achieve acceptable performance. Installations in warmer
areas may also require enclosures for protection against wind, precipitation,
and vandalism. Provision of enclosures increases an RBC installation's ini-
tial cost and is thus a disadvantage, although protected RBCs probably operate
more stably, especially in winter.
When grit and primary solids removal is inadequate, suspended solids may
accumulate in RBC reactors. Foul odors and falling process efficiency ensue.
This is a potential RBC disadvantage which can be avoided by ensuring that the
RBC's influent has had good primary treatment. When excess solids do pass
through to an RBC unit, they can be periodically pumped out of the reactor.
While the RBC process is a relatively stable one, RBC operation can be
disrupted by many of the same influent fluctuations which upset other
processes. Organic and hydraulic shock loadings are handled comparatively
well by RBCs, but some loss of process efficiency will occur. Toxic sub-
stances in the influent may cause a sometimes catastrophic loss of biomass
from the discs. Process efficiency will fall. Recovery, however, is usually
more rapid than that of trickling filters which have been similarly insulted.
Extremes of wastewater pH have an adverse effect upon RBC system performance.
This, of course, is a disadvantage common to all biological treatment
processes.
It is common for organisms to develop on RBC media which are whitish in
color. This white biomass, which is probably composed of Thiotrix or Beggia-
toa, is of little concern when it appears in small patches. As these patches
expand to cover a significant proportion of the discs, however, process effi-
ciency falls. The white biomass phenomenon is associated with septic
influents containing high concentrations of hydrogen sulfide. It can be
prevented or cured by preaeration of the wastewater or by the addition of oxi-
dizing materials such as hydrogen peroxide to the water.
Overloading of the first stage of an RBC installation can c'ause odors to
develop and less than adequate removals to occur. Where this problem is
observed or anticipated, extra surface area can be provided in the first
stage, alleviating the overload conditions. When the overloads are episodic,
equalization upstream from the RBC reactor can be as useful with this technol-
ogy as with others.
22
-------
Some disadvantages have been charged to the RBC process which will prob-
ably disappear as the technology matures. Extensive and intensive controversy
exists regarding design criteria, optimum rotational speeds, matrix design,
media configuration, recirculation requirements, surface-to-volume ratio for
the reaction chambers, and appropriate scale-up procedures. Compared to many
other modeling efforts, RBC modeling is in its infancy. Further operational
experience, additional research, and symposia such as this one can be expected
to remedy these shortcomings.
23
-------
REFERENCE LIST
(I) USEPA, "Process Control Manual for Aerobic Wastewater Treatment Facili-
ties," EPA-403-9-77-006 (March 1977).
(2) USEPA, Office of Water Program Operations (WH-547), Washington, DC,
20460, EPA-430/9-78-004, MCD-42 (June 1978).
(3) USEPA Technology Transfer Publication, "Process Design Manual for Upgrad-
ing Existing Wastewater Treatment Plants" (October 1974).
(4) Manos, G. P., University of Akron, Akron, OH, Institute for Technical
Assistance. Final Report, Field and Laboratory Evaluation of Roadside
Park Wastewater Treatment Plants (March 1977).
(5) F. T. Lense, S. E. Mileski, and C. W. Ellis, Effects of Liquid Detergent
PlantEffluent on the Rotating Biological Contactor, EPA-600/2-78-129
(Industrial Environmental Research Laboratory, Office of R and D, USEPA
(June 1978).
(6) "Behind the Rapid Rise of the Rotating Biological Contactor," Civil
Engineering, Gene Ballaire, Series on Water Pollution Control: No. 11
(American Society of Civil Engineers (January 1979), pp 72-73 (copyright,
1979).
(7) Borchardt, J. A., J. K. Shin, and T. H. Chung, EPA-600/2-78-061, Nitrifi-
cation of Secondary Municipal Waste Effluents by Rotating Bio-Discs (June
1978).
(8) Gene Dallaire, Civil Engineering, Series on Water Pollution Control No.
11, "Behind the Rapid Use of the Rotating Biological Contactor" (January
1979), p 72 (copyright, 1979).
(9) F. M. Welch, "Preliminary Results of a New Approach in the Aerobic Bio-
logical Treatment of Highly Concentrated Waste," Proceedings, 23rd Purdue
Industrial Waste Conference (May 1968).
(10) W. N. Torpey, H. Heukelekian, A. J. Kaplovsky, and R. Epstein, "Rotating
Disks with Biological Growths Prepared Wastewater for Disposal or
Reuse," Journal of the Water Pollution Control Federation, 43 (November
1971).
(11) R. L. Antonie, "Response of the Bio-Disc Process to Fluctuating Wastewa-
ter Flows," presented at the 25th Purdue Industrial Waste Conference
(May 1970).
24
-------
(12) R. L. Antonie and R. J. Hynek, "Operating ikperience with Bio-Surf Pro-
cess Treatment of Food-Processing Wastes," paper presented at the 28th,
Purdue Industrial Waste Conference (May 1973).
(13) R. L. Antonie, D. L. Kluge, and J. H. Mielke, "Evaluation of a Rotating
Disk Wastewater Treatment Plant," Journal of Water Pollution Control
Federation, 46 (March 1974), Part 1.
(14) S. A. LaBella, I. H. fhaker, and J. E. Tehan, "Treatment of Winery
Wastes by Aerated Lagoon, Activated Sludge, and Rotating Biological Con-
tactor," paper presented at the 27th Purdue Industrial Waste Conference
(May 1972).
(15) J. A. Chittenden and W. J. Wells, "Rotating Biological Contactors Fol-
lowing Anaerobic Lagoons," Journal of theWater Pollution Control
Federation, 43 (May 1971).
(16) E. L. Stover and D. F. Kincannon, "Evaluating Rotating Biological Con-
tactor Performance," Water and Sewage Works, 123 (March 1976).
(17) Lense, F. T., S. E. Mileski, and C. W. Ellis, Effects of Liquid. Deter-
gentPlant Effluent onthe Rotating BiologicalConttactor, Industrial
Environmental Research Laboratory, Cincinnatic, OH, June 1978.
(18) C. N. Weng and A. H. Molog, "Nitrification in the Biological Fixed-Film
Rotating Disk System," Journal of the Water Pollution Control Federa-
tion, 46 (July 1974).
(19) R. L. Antonie, "Nitrification of Activated Sludge Effluent: Bio-Surf
Process," Parts I and II, Water and Sewage Works, 121 (November-December
1974).
(20) 0. Hao and G. F. Hendricks, "Rotating Biological Reactors Remove
Nutrients," Parts I and II, Water and SewageWorks, 122 (October-
November 1975).
(21) C. Lue-Hing, et al., "Biological Nitrification of Sludge Supernatant by
Rotating Disks," Journal of the Water Pollution Control Federation, 48
(January 1976).
(22) K. L. Murphey, et al., "Nitrogen Controls Design Considerations for
Supported Growth Systems," paper presented at the 48th Annual Confer-
ence, Water Pollution Control Federation, Miami Beach, FL (October
1975).
(23) Nitrification Systems for Ammonia-Nitrogen Removal (Reid, Quebe, Alli-
son, Wilcox & Associates, Inc., 19 )•
25
-------
(24) C. W. Reh, et al., "An Approach to Design of RBCs for Treatment of Muni-
cipal Wastewater," paper presented at the ASCE National Environmental
Engineering Conference, Nashville, TN (July 1977).
(25) J. F. Lagnese, "Evaluation of RBC Used to Upgrade Municipal Plant to
Secondary Standards," paper presented at the Technical Conference, WPCA,
Pennsylvania (April 1978).
(26) R. A. Sullivan, et al., "Upgrading Existing Waste Treatment Facilities
Utilizing the BIO-SURF Process," paper presented for Autotrol Corpora-
tion (May 1978).
(27) J. F. Lagnese, Ibid.
(28) G. T. Griffith, R. H. F. Young, and M. J. Chun, Rotating Disc Sewage
Treatment System for Suburban Developments and High Density Resorts of
Hawaii, Water Resources Research Center, University of Hawaii, Honolulu,
Hawaii, Technical Memorandum Report #56 (January 1978).
(29) Environmental PollutionControl Alternatives — Municipal Wastewater, pp
21-24, EPA 635/5-76-012, Environmental Pollution Control EPA Technology
Transfer Publication (USEPA).
26
-------
EPA RESEARCH PROGRAM FOR RBC
By
Edward J, Opatken
Chemical Engineer
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 1(5268
The EPA municipal wastewater research effort on Rotating Biological Con-
tactors (RBC's) was initiated in 1967- A contract was awarded to Allis
Chalmers to test a bench-scale unit in their laboratory. This contract was
extended to conduct an evaluation using the laboratory RBC unit on domestic
wastewater in Milwaukee, Wisconsin. This was then followed by a contract to
pilot test the RBC's at Pewaukee, Wisconsin, in 1969, and it was during this
period that Allis Chalmers sold their RBC interests to Autotrol Corporation.
EPA supported these series of studies with approximately $360)000 in contract
awards between 1967 and 1969- The Pewaukee pilot plant contract amounted to
$33,000 and Autotrol continued the pilot study with corporate resources for
an additional six months beyond the EPA project completion date.
The early research effort progressed from the laboratory scale in Mil-
waukee, through the pilot scale in Pewaukee to a full-scale demonstration in
Pewaukee, between 1967 and 1971- These concentrated research and development
initiatives accelerated the introduction of RBC's as an alternative secondary
treatment process.
While EPA was supporting the research and development work in Wisconsin,
a concurrent grant was awarded to Rutgers University to conduct a literature
search. This was then followed with a research grant to study a bench-scale
RBC unit.
27
-------
These initiatives into laboratory and pilot plant scale evaluations were
instrumental in establishing the feasibility of RBC's, and the next step was
a demonstration with full-scale equipment at Pewaukee, Wisconsin in 1971.
The Village of Pewaukee, along with Autotrol Corporation, conducted a compara-
tive evaluation of the RBC process with an existing trickling filter plant
and concluded that RBC's were an effective treatment process. Additional
work was done on phosphorus removal by mineral addition. The EPA share of
the project cost was approximately $400,000.
In 1972, EPA granted the West Virginia University an award of $16,000 to
evaluate an RBC unit located at a summer camp, Camp Horseshoe in West Virginia.
Following these projects which were directed towards speeding the intro-
duction of new technology, the EPA research program was decelerated. The
private sector came to the forefront and took up the research slack and intro-
duced novel equipment and process innovations to improve the cost effective-
ness of the process. The number of manufacturers supplying RBC's grew and
with this increase came further improvements spurred by competition for sales.
The EPA research program for RBC's became selective. In the mid 1970's
grants were awarded for specific or unique situations. The University of
Michigan studied nitrification with RBC's in a pilot facility at Saline,
Michigan. The City of Edgewater, New Jersey, was awarded a grant to evaluate
a novel application in which a primary clarifier was converted to a secondary
system by installing RBC's above a false floor.
The entire RBC research effort was conducted via extramural grants, con-
tracts, and cooperative agreements with universities and municipalities. This
approach enabled EPA to handle a diverse program with a minimum of personnel
and to perform these investigations within a minimal time frame.
The rapid growth in the number of treatment plants employing RBC's by the
late 1970's has caused a re-evaluation of the research program. The present
technology used for designing RBC facilities is being questioned. Peripheral
speed as a scale-up factor is being questioned. DO sags in the initial stage
require corrective action. The EPA research program had earlier taken the
position that development efforts that lead to equipment modifications and
result in performance advantages should remain in the domain of the suppliers.
However, the technical questions that are being raised effect performance and
capital costs. Both of these items impact the EPA Construction Grants Program
and answers are required to improve the cost effectiveness of the process.
The RBC research program has again turned around to address these questions
and a new program was developed to provide answers to these questions concern-
ing RBC's.
The symposium on RBC's at Seven Springs is geared towards producing a
state-of-the-art on the latest technology. There are 70 plus papers that will
be presented during this symposium that will cover'practically all aspects on
RBC technology. In addition, there is a Research Needs Workshop on Tuesday
night that is aimed at defining technical gaps in the present process so that
solutions can be prescribed and evaluated for bridging these gaps and improve
the overall effectiveness of the process.
28
-------
The second phase of the EPA research program was an award of a contract to
Roy F, Weston Consulting Engineers. They will assess the present design prac-
tices and evaluate the applicability of the various parameters to adequately
predict the performance of RBC's. Secondly, they will study several RBC oper-
ating plants to determine if performance is within the design specifications.
This phase of the program is well underway and Warren Chesner will bring us
up-to-date on its progress later on this morning.
The third phase of the EPA research program is a cooperative agreement
under evaluation with the City of Columbus, Indiana, to study the questions
that are being raised on RBC's. The City of Columbus has 10 lines of RBC's.
Each line consists of 8 shafts. At the present time only seven lines are used.
It is our intention to modify two of these lines to evaluate various RBC design
parameters with full-scale equipment. This would allow a direct comparison
between the two lines. One of the major issues confronting the designers of
RBC facilities concerns the application of hydraulic or organic loadings as
the preferred basis for specifying the surface area requirements on RBC's.
To evaluate this function, the following ground rules were established for
conducting the test at Columbus.
1, A comparative evaluation between organic and hydraulic loadings would
be performed.
2. The flows to both systems would be controlled.
3- The influent to both systems would be identical.
U. Diurnal variation would be incorporated into the flow control system
with a maximum to average ratio of 1.5 and a minimum to average ratio
of 0.7. The system will also be capable of operating between 50 and
200$ of design flow to stress the RBC treatment trains.
5. The RBC's are plant scale facilities. That is, the use of pilot scale,
or more specifically, less than 10 foot diameter disks were forbidden
to avoid controversy over the peripheral speed scale-up parameter.
6. DO's would be continuously recorded at critical locations on each
system.
7. Chemical characterization would be conducted on the influent, effluent
and at the various stages.
Following this evaluation, the preferred mode of operation, organic or
hydraulic, should be established. In addition, a secondary objective is to
identify and improve the limiting factor governing RBC performance. If the
data on DO identified this parameter as a contributing factor for limiting
RBC performance, then provisions will be made to modify the process to improve
DO levels. Several methods will be tried. Among them are:
1. Increase rotating speed at critical locations.
2. Force feed air at critical locations.
29
-------
3. Evaluate the effect of lower iramergence which increases the air
contact time.
This project is scheduled to start during May, 1980, and its estimated
completion date is October, 1982. The approach should provide definitive
answers to many of today's questions and should advance the technology and
performance of RBC's.
30
-------
ASCE WATER POLLUTION MANAGEMENT TASK COMMITTEE REPORT ON
"ROTATION BIOLOGICAL CONTACTOR FOR SECONDARY TREATMENT"
By
Shankha K. Banerji
Chairman of Task Committee on RBC
Professor of Civil Engineering
University of Missouri
Columbia MO
Introduction
ASCE Environmental Engineering Division, Water Pollution Management
Committee (WPMC) established a task committee in October 1977 to write a State-
of-the-Art report on Rotating Biological Contactor for Secondary Wastewater
Treatment. The task committee has completed a second draft of the report
which is under review by the committee members. It is expected that after
final reviews, this report will be published later this year. Selected portions
of the second draft of the report are presented here.
Currently, the most common ^.secondary biological treatment methods include
the trickling filter process and its modifications, air or pure oxygen activated
sludge process and its modifications, and rotating biological contactor (RBC)
process. This paper will briefly summarize the present knowledge on the design,
application and selection of the RBC process for municipal wastewater treatment.
Process Description
The (RBC) process is an aerobic, continuous flow, wastewater treatment
system designed for municipal and many industrial wastewaters. The RBC
process converts the influent soluble biodegradable organic wastewater con-
stituents into biomass and off-gases. Biomass generated by RBC units is
separated from the wastewater carrier stream in a sequential secondary
clarifier. Settled wastewater is first introduced into a tank containing a
series of high density polyethylene discs (media) attached to a horizontal
shaft (Figure 1). In U.S. practice, the discs are mounted in the tank so that
31
-------
w
to
Liquid
Layer
Biofilm
,. Liquid
Boundary
Layer
02
CO
Shaft
02
Biological
Growth
'/"//rrrr
Tank
Sloughed
Biomass
Figure 1. Rotating Biological Contactor Disc
-------
about forty percent of the media area is submerged in the wastewater. The film
of biomass growing on the media is responsible for the removal and conversion
of the biodegradable organic wastewater constituents. The media is rotated
continuously by mechanical or air drive systems so that the biomass film is
alternately exposed to fresh wastewater in the tank and the air above the tank.
For some specially designed RBC systems, additional treatment results from
the development of unattached, suspended biomass culture in the mixed liquor
of the tank.
To avoid short circuiting in the tank, groups of discs are segregated by
baffling into stages. For small installations, the flow path is usually per-
pendicular to the disc faces. Larger installations are usually designed with
a single shaft or a series of shafts constituting each stage with wastewater
flow parallel to the disc faces. The microbial population in each stage can
vary significantly depending on wastewater loading conditions. Heavy growth
and substrate removal usually occur in the first stage followed by both de-
creasing growth and carbonaceous removal in succeeding stages. Where nitri-
fication of wastewater is desired, the latter stages can be constructed with
more media surface area per shaft length since biomass production is reduced
and bridging between adjacent discs is less likely to occur.
As a result of continuous rotation, the media carries a film of wastewater
into the air where oxygen is transferred through the liquid film surface. Both
oxygen and organic substrate materials diffuse through the liquid film into the
growing biomass film where they are consumed for growth and respiration purposes.
Excess dissolved oxygen in the wastewater film is mixed with the contents of
the bulk liquor in the tank and results in aeration of the wastewater carrier
stream.
Shearing forces exerted on the growing biomass film result in excess bio-
mass being periodically sloughed from the media into the wastewater carrier
stream. This sloughing action prevents bridging and clogging between adjacent
discs. The disc mixing action keeps sloughed biomass solids in suspension
until they are removed from the RBC tank and separation occurs in the final
clarifier. In essence, the rotating media is used to both provide a support
surface for microorganism growth and to assure an opportunity for contact
between the microorganisms, the substrate and oxygen.
Figure 2 shows a flow diagram for a typical RBC treatment plant. Float-
able and settleable solids in the wastewater are first removed by primary
treatment. The primary effluent then flows to the multi-stage RBC unit where
biological removal of organic material occurs. Each RBC stage tends to
operate as a completely mixed, fixed film, biological reactor. Treated waste-
water and sloughed biomass flows from stage to stage with progressively increasing
substrate removal occuring. Sloughed biomass is separated from the carrier
stream in the final clarifier and the underflow solids are disposed of by
conventional means. Operation of the process is on a once through basis with
no need for effluent recycling.
The RBC process differs from the trickling filter by having a significantly
longer retention time (8 to 10 times that of a trickling filter) and a dynamic,
rather than stationary media; and from the activated sludge process, by having
an attached rather than a suspended biomass and not dependent on suspended
33
-------
culture separation and recycle. In the former case, higher levels of treatment
are achieved by the RBC process and in the latter, the RBC process has less
susceptibility to upset from changes in hydraulic or organic loading in either
the reactor or clarifier.
Process Applicability
The RBC process may be used to remove a major protion of the biochemical
oxygen demand (BOD), and ammonia-nitrogen (NH3-N) from any biodegradable
wastewater. The process is applied to treat domestic sewage in plants ranging
from small package facilities to large municipal sewage treatment plants. Also,
wastewater from dairies, bakeries, meat and poultry processors, pulp and
paper mills, animal feed lots, distilleries, canneries, refineries and other
biodegradable industrial wastewaters can be treated by the process.
Process Hardware
The process hardware consists of closely packed circular plastic media
mounted on a shaft. The shaft is supported on bearings and connected through
a gear box to an electric motor. The plastic media consists of corrugated
polyethylene material. In one instance, the media consists of a drum filled
with 38 mm plastic balls (Bio-Drum Process). The shaft rotates at 2-4 rpm
inside a concrete or steel tank. The shaft length varies from 5 to 20 ft.
depending upon the size of the unit. The diameter of the packed media on the
shaft varies from 4 ft to 12 ft. depending on the capacity of the unit. For
higher degrees of treatment and larger flow capacities several modular units
may be placed in parallel or in series depending upon the configuration desired
as shown in Figure 2. The RBC media is about 40% submerged in a trapezoidal,
semi-circular or rectangular tank, with intermediate partitions in some situa-
tions. To maintain performance under cold weather conditions, the modular RBC
units are provided with fiberglas enclosure with access doors & ventilation.
Alternatively, the RBC units can be housed in a conventional insulated struc-
ture that covers a whole battery of units.
In a recent development, air is introduced to aid in rotating the media in
the tank. Figure 3 shows an air drive RBC system. In this process, plastic
cups are welded onto the periphery of the media over the entire length of the
contactor. A small air header placed in the tank underneath the media allows
air to be released along the tank length. The released air is captured in
the plastic cups causing buoyant forces to rotate the shaft. Radial passages
in the media periphery cause a portion of the released air to flow upwards into
the corrugated media sections. The supplementary aeration and increased
turbulence achieved from this is sufficient to allow a reduction in rotational
velocity of the media while still achieving the same degree of treatment. The
air drive process requires about 25% less units for a given application com-
pared to mechanical drive systems.
Hi s to ri ca1 Ba c kg rou nd
The RBC systems as presently used evolved from the research work of
Pope! and Hartman in West Germany in 1955 (1). However, earlier researchers
in the USA had developed similar devices. Buswell, in 1929, developed a unit
called "Biological Wheel", which was similar to the present RBC units and whose
purification capacity was thought to be based on biological principles. Later
in 1931, pursuing this line of thought, Maltby patented a process that was based
34
-------
Radial Passages
Media
End View — Air Drive Schematic
Figure 3: Air Drive RBC Unit
35
-------
Primary Clan"tier
Raw __
Wastewater
w
en
Multf - Stage Rotating Biological Contactors
Final
Clarifier
Sludge Treatment
Figure 2. Flow Diagram for Rotating Biological Contactor Process
-------
on the biological wheel principles (2). Hartmann (!) credits Travis
for the idea behind the RBC process. Travis in 1901 installed wooden strips
in "Hydrolytic Tanks", (settling tanks) that were to catch cloudy non-settling
solids from the wastewater by adsorption. These strips accumulated solids
on their surfaces & eventually these solids would slough off the strips to the
settling tank hopper. The development of the contact aeration by Hays in USA
(3) and others in Europe (4), was a logical improvement of Travis's idea.
The application of air below the wooden slots was to retain sludge floes in
the aeration tank and improve effluent quality. However, these ideas were not
integrated to produce the present RBC process until 1960 when Pope! &
Hartmann developed their immersion drip-filter (trauchtropfkorpern) (1). The
first eommerical RBC was installed in 1960 in West Germany & soon after it was
widely applied throughout Europe (5). In U.S. Allis-Chalmers Company began
development work in mid 1960's and presently there are-several companies
offering these systems for commerical applications.
PROCESS DEVELOPMENT
Operational Characteristics
The RBC systems employed for secondary waste treatment study usually
involve 2-10 stages in small-scale laboratory units or 2-6 stage in full-
scale pilot plants. Due to the change in physical and chemical properties of
wastewater to be treated in each stage, the biochemical nature of micro-
floral populations as well as metabolic end products varies significantly.
Figure 4 shows the distribution of biochemical oxygen demand (BOD), chemical
oxygen demand (COD), ammonia nitrogen (NH3~N), nitrite (N02) and nitrate (N03)
dissolved oxygen, suspended solids, dry weight of biomass, and pH in multi-
stage rotating biological contactor systems (6,7,8 ).
Apparently, the organic carbon in both high-strength industrial waste
and normal-strength domestic waste can be effectively removed by the RBC
system, but the degree of BOD and COD removal is highly dependent upon the
rate of hydraulic loading applied to the system and the number of stages
employed. Observations from Figure 4 reveal that the majority of the removal
of biodegradable organic matter is achieved within first six successive stages.
It was also found, however, that the rapid uptake of carbonaceous matter
occurred in the first three stages. It is believed that improved removal
achieved by successive stages is due to improved residence time distribution
obtained by staging and the development of a biomass population in each stage
that has adapted to treat the specific waste characteristics found in that stage.
Study of nitrogen transformation as seen in Figure 4 indicates that the
conversion of NH3-N to nitrate and nitrate does not take place until stage 5
where the BOD concentration has been reduced to about 20 mg/K Before Stage
5, the concentration of NH3-N increases slightly over influent because of the
hydrolysis of organic nitrogen in the biological growth. Figure 4 also shows
that the minimum level of dissolved oxygen occurs at the first and second stages
and, thereafter, the content of dissolved oxygen in wastewater increases in
each successive stage. This result explains why the microbial activities as
expressed by oxygen utilization, change in each stage of the RBC unit. Rapid
oxygen consumption results in a low concentration of dissolved oxygen in
Stages 1 and 2. This suggests that the most active cell growth apparently
takes place within these stages, where the wastewater is initially contacted
37
-------
P
E
-------
with biomass already developed on the disc surfaces. Studies of steady-state
disc biofilm thickness at various stages by Pretorius (8), Pescod, et.al (9),
and Sack, et.al (10) have confirmed the above statement.
The development of the types of microorganisms on the media is certainly
dependent upon the nutrients in the wastewater entering each individual stage.
Torpey, et. al (7,11), Pescod, et. al (9), Sack, et. al (10), and Antonie, et.
al (12) conducted examinations to determine biological solids characteristics
on the media under various operational conditions in treating domestic and
industrial wastes. Sack, et. al (10) found that in a four-stage RBC system,
the overall appearance of sludge organisms ranged from a black stringly growth
with white gelatinous patches on first and second stages, to a greenish-brown
slime on third and fourth stages. However, the general findings based upon
these studies reported by the above investigators are:
"The predominant organisims including Sphaerotilus and zoogleal
bacteria are present on all discs. Besides these two important
kinds, the diversity and abundance of free-swimming protozoa
(Paramecium, Cyclidium, Ocomonas, Oxytrichia, and Euglena) are present
in the first few stages. The growth of rotifers (Epiphanes and Proales),
stalked ciliates (Vorticella), nematodes (Ethmolaimus), and a loop
forming fungus (Athrobotrys) together with algae (Coelastrum, Chlorella,
Fragilaria and Pinnularia) take place in the last few stages only when
organic loading is low but high enough to support microbial growth.
The quickly developed biofilm at the earlier stages of the RBC system
is much thicker than bacterial slime produced on the later discs.
"The mechanisms of attached growth in a RBC treatment system is
described as the filamentous organisms (Sphaerotilus, Geotrichum,
Bacillus) actually serving a sort of skeltal system on which other
microorganisms are able to attach. The thickness of biofilm is sub-
stantially reduced in each stage as a result of significant reduction
in filamentous populations, and that is caused by the marked change
of carbonenergy level in wastewater after passing it through each stage.
Both Pseudomonas dem'trificans and Beggiatoa alba are also present in
the RBC system indicating that there are the involvements of nitrogen
and sulphur transfers inherent in the systems.
Operational Parameters
The major factors controlling the RBC system operation and performance are
known to be:
1. Influent wastewater substrate concentration
2. Residence time of wastewater (or surface hydraulic loading)
3. Wastewater temperature
4. Media rotational speed.
In addition to the above control parameters, the effects of disc immersion
depth and disc surface area configuration and density on the treatment effic-
iency of the RBC system may also be significant. However these two parameters
have been standardized for the purpose of optimizing the process design and
operation. The current practice with regard to the immersion depth requirement
is to ensure that 40% of the total disc surface area is submerged in the waste-
39
-------
water in the biological reactor. The total effective disc surface area for a
full scale treatment plant is determined for disc diameters commonly in the
range of 10-12 ft., although the treatability study is often carried out by a
relatively small pilot plant.
Effect of InfluentWastewater Concentration
The influence of the initial wastewater concentration on the removal of
BOD and Hti^-H at various hydraulic loading rates is illustrated in Figures 5
and 6. It is apparent from Figure 5 that a linear relationship between % BOD
removal and hydraulic loading is found when treating both industrial and
domestic wastes. However, the rate of BOD removal is entirely dependent upon
the initial concentration of BOD in the wastewater. At a specific hydraulic
loading, the BOD removal for a domestic waste increases as the initial con-
centration of BOD increases. On the other hand, a decrease in the BOD removal
is observed with increasing-initial BOD concentration when treating low-
biodegradable industrial wastes.
The removal of Ammonia nitrogen under different applied hydraulic
loading rates is also affected by the initial wastewater characteristics.
Figure 6 shows that in the range of 9.5-36.0 mg/1 NH3-N, the rate of
ammonia nitrogen oxidation decreases as the initial concentration of ammonia
nitrogen increases. In addition, under the high initial NH3-N concen-
tration, the fraction of ammonia nitrogen remaining increases significantly
with an increase in hydraulic loading.
Effect of Residence Time of Wastewater (or SurfaceHydraulicLoading)
The influence of varying residence time of wastewater on the efficiency
of BOD and ammonia nitrogen removal is shown in Figure 7. It is evident that
the system performance is closely associated with the liquid process retention
time or the residence time of wastewater. At residence times less than 100
minutes, the removal of BOD and NHs-N always decreases as the flow rate in-
creases or the residence time decreases. The substrate removal does not
increase significantly much beyond a residence time of 100 minutes.
As indicated earlier, the stability as well as the efficiency of a RBC
waste treatment system is highly dependent upon the surface hydraulic loading
rate. The parameter is normally expressed as flow per unit time per unit
surface area (gpd/ft2) covered by biological growth and is inversely related
to residence time. Many researchers have reported that Increasing disc surface
area or decreasing surface hydraulic loading increases substrate utilization
(24,25,27). The increase in substrate utilization is mainly attributed to the
longer wastewater residence time and the fact that the amount of active biomass
on the disc surface relative to the substrate loading has increased (lower
F/M ratio).
The effect of hydraulic loading on the removal of BOD and ammonia nitrogen
is shown in Figures 8 and 9. The studies of Antonie, et. al (15) and Tucker
(21) indicate that the RBC process is approximately first order with respect
to BOD and ammonia nitrogen removal, that is the rate of bio-oxidation reactior
is proportional to the amount of oxizidable organic matter or inorganic
nitrogen remaining. However, it is generally observed as in Figures 8 and 9,
that the efficiency of substrate assimilation is reduced as the surface
40
-------
100
90
2 80
o
r*
11
K
O 70
a
100
90
§ 80
o
o
es
Q
O 70
a
60
© Bleached Kraft, Gillespie (13)
x Ins;Board(240 mg/l),Gillespie {13}
* Un. BI. Kraft, Gilles pie (13)
A Ins. Board (500 mg/l), Gillespie (13
<>3I. Sulfite, Cillespic (13)
O Dairy Waste, Antonie (12)
All 4 stages SB treatment plants
\0.5
\mg/i)
r-4 \\ N
f ft4,000 (900 mg/l)
\ \ , '- x
/S
(H6 mg/l)
(240 mg/1)
INDUSTRIAL WASTES
IM-
(225 mg/l)
S55
"~—
mg/l)
Sack, et al (10)
Clark.et al (14)
O 4 stages,
A 4 stages,
9 2 stages, Antonie,ctal( 15)
T 4 stages,
® 3 stages,
9 4 stages,
^ 4 stages,
)jC 10 stages,
Chittendpa.etal
(16)
Borchardt (17)
Malhotrs.gt al
Hao.et al (19)
Allis-Chalmer;
(20)
•78
DOMESTIC SEWAGE
_L
10
0123456 789
Hydraulic Loading, gpd/ft^
Figure 5. - Effect of Initial BOD Concentration on BOD Reduction at Various Hydarulic Loadings
(Initial BOD Concentration in Parentheis)
-------
100
90
a-*
80
OS
c
-------
100
50
40
30
ZO
10
80
I
I
a
a
S.
o
40
I
o
90
I
a
«70
O
n
60
50
0.06
.
6
® 4 stages, Clark, et al (14)
4 stages, Tucker (21)
6 stages, Malhotra, et al (18)
2 stages, Antonie.etal (15)
O 10 stages, Torpey, ot al (11)
DOMESTIC SEWAGE
0.10
\
®4<
20 40 60 80 100 120
Residence Time of Waste-water, Minutes
140 160
Figure 7. Effect of Flow Rate & Residence Time on BOD and Ammonia-N
Removal
43
-------
100
80
** 60
o
O)
Q
§ 40
20
0
9-3
2,4 stages
12.2-19.4°C
9 3.2-4.6 rpm
stages
19.9°C
rpm, 15-30 rpm
stages
3.8-10.00C
2.0-2,9 rpm
O 2 stages, Antonie, et.al (15)
0 4 stages, Sack, et.al (10)
© 4 stages, Antonie (12)
O 10 stages, Allis-C halraers (20)
d 4 stages, Clark, et.al (14)
i
0
3456
Hydraulic Loading, gpd/ft2
Figure 8. Effect of Hydraulic Loading on BOD Removal
-------
Ln
100
80
rt
I 60
d
o
Cfl
o
£
rt
• i-4
O
g
g
40
20
6 stages,
stages,
stages,
es.
es.
Malhotra, et al (fS)
Clark, et al (14.)
Antonie, et al (|*£"}
Antonie (1*3^
Tucker (2 |)
stages
19.9 C
2, 9 rpm
2 stages
V 14.4-19. 9 C, 3.2-4.6 rpm
5678
6
,2
2345
Hydraulic Loading, gpd/£t*
Figure 9. Effect of Hydraulic Loading on Ammonia Nitrogen Removal
-------
hydraulic loading rate increases. Additionally, it is found that the influence
of hydraulic loading on substrate removal is also dependent upon other par-
ameters such as temperature, disc rotating speed, and number of stages.
Effect of Wastewater Temperature
The results of eight earlier investigations as shown in Figure 10 and
11 demonstrated that increasing temperature increases the rate of substrate
utilization. However, the magnitude of the temperature effect on the removal
of carbonaceous and nitrogen compounds is actually determined by the applied
hydraulic loading rate as indicated earlier.
Figure 10 shows that at the wastewater temperatures of 55°F or above, the
percentage BOD removal increases with temperature is not significant. However,
at temperatures greater than 55°F, the hydraulic loading changes cause signif-
icant % BOD removal changes. As the hydraulic loading rate increases, the
inhibition caused by the low wastewater temperature becomes more appreciable.
A similar effect on ammonia removal efficiency by nitrification with temp-
erature is shown in Figure 11. It seems that biological nitrification
exhibits a greater sensitivity to temperature and hydraulic loading rate. In
general, as the hydraulic loading rate is controlled at less than 1.0 gallon
per day per square foot, the percentage of ammonia nitrogen removal is not
greatly influenced by the wastewater temperature unless it reaches below 55°F.
However, a significant decrease in the % ammonia nitrogen removal results
from a temperature drop after the applied hydraulic loading rate is in excess
of 1.0 gallon per day per square foot.
The temperature correction factor which is used for the conversion of
treatment efficiency at any temperature to standard temperature of 20°C or
68°C has been studied by Antonie (22) and Weng (24) in both pilot-scale and
full-scale operations. The results indicate that the temperature correction
factor varies in each stage of a RBC system and also is related to the degree
of temperature fluctuation. It is generally found that the correction factor e
(in the equation KT = Kgs 6 T-68} where T is the temperature in °F) is approx-
imately equal to 1.017 in the first few stages having a temperature higher
than 55°F. The temperature correction factor is reduced from 1.017 to 0.645
as a result of temperature drop from 55°F to 40°F. According to Weng (24), a
lower correction factor (<1.017) is commonly observed in the last few stages
even at wastewater temperature higher than 55°F.
Effect of Media Rotational Speed
Disc rotation affects wastewater treatment in several ways. It provides
contact between the biomass and the wastewater, it shears excess biomass, it
aerates the wastewater, and it provides the necessary mixing velocity in each
stage. Increasing the rotational velocity increases the effect of these
factors. However, there is an optimum rotational velocity, above which treat-
ment levels are no longer increased. This optimum velocity will vary with
wastewater conditions, i.e., the optimum velocity is higher for more concen-
trated industrial wastes and lower for more dilute domestic wastes.
The effect of rotating disc speed on the performance of the RBC system is
shown in Figure 12. It is apparent from Figure 12 that the effect of rotating
disc speed on BOD or COD removal cannot be well defined unless the process
operation is classified in accordance with disc size. For a 2-4 stage treat-
46
-------
Temperature, °F
100
32
.104
o o
e
80
fO
>
O
01
a:
a
O
CO
60
0.7 gpd/m
1.5 gpd/ft2
3.0 gpd/ft2
3.5 gpd/ft2
© 2
4 stages.
4 stages,
stages.
Sack, et.al (10
Antonie, Kluge
Mielke (23)
Antonie, et.al (15)
© 4 stages, Clark, et.al, (14)
010 stages, Allis-Chalmers (20)
9 4 stages, Antonie (22)
40
10 20 30
Temperature, °C
Figure 10. Effect of Temperature on BOO Removal
40
-------
*»
CD
100
90
to
1 80
O)
on
c
03
O)
O
4J
70
C
O
•f 60
50
40
0.5 gpd/ff:
1.0gpd/£t2
1.5 gpd/ft2
2.0 gpd/ft2
2.5 gpd/ft:
2. 6 gpd/ft2
stages, Malhotra, et al
stages, Hao , et al (}g)
stages, Antonie.etal
10
20
Temperature, C
30
40
Figure 11. Effect of Temperature on Ammonia Nitrogen Removal
-------
100
80
70
• 60
o
O)
oe
Q 90
o
CD
80
70
.Domestic Sewage
Stage Number* 4
Disc Size > 6.5 ft
Ref.
Clark, et al (14)
Sack, et al (iQ)
Antonie, Kluge, & Mielke (23)
Domestic Sewage
Stage Numbers 2
Disc Size> 6.5 ft
Ref,
Antonie, et al (15)
_L
Dome81ie Sewage Stage Number* 4-10
Disc Size <&3 ft
Ref.
Torpey.etal {7){11)
Allis-Chalmers (20)
Poon.et al (26) •
Soft Drink Haste stage Number = 5
Disc Size = 0.62 ft
Initial GOD -1, 000 mg/l
Ref.
Pescod, et al (9)
100
90 ^
30 ~
Q
O
70 g
O
60 Q
Cr
50 U
40
23450 123
Hydraulic Loading, gpd/ft^
Figure 12. Effect of Rotating Disc Speed on BOD Removal
-------
ment plant having the disc size greater than 6,5 ft., the optimum speed of
disc rotation is approximately 4.6 rpm. Whereas in a small 4-10 stage pilot
plant having a disc size of less than 3 ft., the most effective speed of
rotation is somewhere between 6-10 rpm. The disc rotating speed in excess
of the optimum velocity does not improve performance and wastes a significant
amount of energy. Thus, rotational velocity of RBC media is an important
design criterion. Testing of various diameter media has indicated that a
fixed peripheral velocity can be used to determine the required RPM for any
media diameter as long as there is no oxygen limitation on substrate removal.
Under organic loading conditions where an oxygen supply limit exists, the
angular velocity of RPM is the proper scale factor for various diameter
media. The principle factor determining oxygen supply is the surface renewal
rate per unit of wastewater flow and a proper scale-up for this factor is
angular velocity. When an oxygen limit does not exist, the essential factors
are hydraulic shear and mixing energy, both of which scale-up directly with
peripheral velocity. Power requirements increase expenentially with increases
in media velocity. For example, doubling the rotational velocity will
typically increase the power consumption five-fold. Typicaly, full-scale
RBC units consume from two to three KW per 100,000 sq. ft. of the media
surface area when rotated at 1.6 RPM or a peripheral velocity of 60 ft.
per minute. To significantly increase this rotational speed is not econ-
omically justifiable, particularly when considering that.a 20-year present
worth analysis of energy costs indicate that each horse-power of energy
is worth $2,500 to $5,000. Therefore, a rotational speed of 1.6 RPM is
considered a practical upper velocity limit to use, even when treating highly
concentrated wastes (28).
Summary
The paper describes some of the factors that affect the performance of
RBC process for wastewater treatment. The major factors controlling the per-
formance of the RBC process are: Influent Wastewater Substrate Concentration,
Residence Time of Wastewater (or surface hydraulic loading), Wastewater
temperature and media rotational speed.
Acknowledgement
The paper presented herein was taken from a draft report prepared by
WPMC Task Committee on Rotating Biological Contactor. The members of the
committee who contributed to the draft report ara: Donald F. Kincannon,
Alfred A. Friedman, Ronald L. Antonie, R. Srinivasaraghavan, Ed Smith,
Yeun C. Wu, A.F. Gaudy, Jr., and S.K. Banerji. The completed report after
review will be published by ASCE. Acknowledgement is made to ASCE Environ-
mental Engineering Division Executive Committee for allowing this report to
be presented here.
50
-------
References
1. Hartmann, H., "Investigation of the Biological Clarification of Waste-
water Using Immersion Drip Filters." 9^, Stuttgarter Berichte zur
Seidlungswassenirtschaft, R. Oldenbourg, Munich (1960).
2. Maltby, A.T., "Process & Apparatus for Treating Sewage & Other Organic
Matters", Patented June 23, 1931 U.S. Pat. No. 1,811,181.
3. Hardenbergh, W.A. and Rodie, E.A., "Water Supply & Waste Disposal",
International Textbook Co., Scranton, PA (1961).
4. Imhoff, K. and Fair, G.M., "Sewage Treatment" John Wiley & Sons, New
York (1956).
5. Antonie, R., "Fixed Biological Surfaces-Wastewater Treatment", CRC Press,
Cleveland, OH (1976).-
6. Labella, S.A., et. al. "Treatment of Winery Wastes by Aerated Lagoon,
Activated Sludge, and Rotating Biological Contactor", Proceedings 27th
Purdue Industrial Waste Conference, 803,(1972).
7. Torpey, W.N., et al. "Rotating Discs with Biological Growths Prepare
Wastewater for Disposal and Reuse", Water Pollution Control Federation
Journal, 43, 2181,(1971).
8. Pretorius, W.A., "Some Operational Characteristics of A Bacterial Disc
Unit", Water Research, 5, 1141, (1971).
i
9. Pescod, M.B\, et al., "Biological Disc Filtration for Tropical Waste
Treatment", Water Research, 6, 1509, (1972).
10. Sack, W.A., et al., "Evaluation of the Bio-Disc Treatment Process for
Summer Camp Application", EPA Project No. S-800707, National Environ-
mental Research Center, Cincinnati, OH.(1973).
11. Torpey, et al., "Rotating Biological Disc Wastewater Treatment Process-
Pilot Plant Evaluation", EPA Project No. 17010 EBM, Environmental Pro-
tection Agency, Washington DC, (T974).
12. Antonie, et al., "Preliminary Results of a Noval Biological Process for
Treating Dairy Wastes", Proceedings of 24th Annual Purdue Industrial
Waste Conference, Purdue University, (May 1969).
13. Gillespie, W.J., et al., "A Pilot Scale Evaluation of Rotating Biological
Surface Treatment of Pulp and Paper Mill Wastes", Proceedings 29th
Purchase Industrial Waste Conference, 1026,(1974).
14. Clark, et al., "Performance of a Rotating Biological Contactor Under
Varying Wastewater Flow", Journal Water Pollution Control Federation,
50, 896,(1978).
15. Antonie, R.L., et al., "Application of Rotating Disc Process to Municipal
Wastewater Treatment11, EPA Project No. 17050 DAM, Autotrol Corporation,
Milwaukee, WI,(November, 1971).
51
-------
16. Chittenden, J.A., et al., "Rotating Biological Contactors Following
Anaerobic Lagoons", Journal Water Pollution Control Federation, 43/746,
U9Z1).
17. Borchardt, J.A., "Biological Waste Treatment Using Rotating Discs",
Biotechnol & Bioeng, Symp. No. 2, 131, John Wiley & Sons, Inc., New
York,(1971).
18. Malhotra, S.K., et al., "Performance of a Bio-Disc Plant in a Northern
Michigan Community", Presented at the 1975 WPCF Conference at Miami
Beach, Florida.
19. Hao, 0., et al., "Rotating Biological Reactors Removal Nutrients - Part
I & II, "Water and Sewage Works", 122, 10, 70, and 11, 48,(1975).
20. Allis-Chalmers, "Municipal Sewage Treatment with a Rotating Biological
Contactor", EPA Contract No. 12-12-24, Federal Water Pollution Control
Administration, Department of the Interior,(May 1969).
21. Tucker, A.L., "A Case Study of Ammonia Nitrogen Oxidation Rates in a
Biological Rotating Disc System" Master Thesis, Department of Civil
Engineering, University of Pittsburgh,(September, 1976).
22. Antonie, R.L., "Rotating Biological Contactor for Secondary Wastewater
Treatment" Presented at Culp/Western/Culp WWT Seminar South Lake Tahoe,
State!ine, Nevada, (1976).
23. Antonie, R.L., Kluge, D.L., and Mielke, J.H., "Evaluation of a Rotating
Disc Wastewater Treatment Plant", Journal Water Pollution Control Fed-
eration, 46, 498,(1974).
24. Weng, C., "Biological Fixed-Film Rotating Discs for Wastewater Treatment",
PhD Thesis, School of Engineering and Science, New York University,(1972).
, _ _._. t«^^—,. i . .. , ...
25. Hartmann, H., "Investigations of the Biological Clarification of Waste-
water Using Immersion Drip Filters", Stuttgart Report of City Water
Economy, Published by Oldenbourg, Munich, Vol. 9,(1960).
26. Poon, C.P., et al., "Rotating Biological Contactors Treat Island's Saline
Sewage", Water and Sewage Works, 125. 62,(February 1978).
27. Popel, F., "Estimating Construction and Output of Immabrow Drip Filter
Plants1,' Eidg. Technische Hochschule, Stuttgart, Germany,(1963).
28. Antonie, R.L., Personal Communication (1979).
52
-------
CURRENT STATUS OF MUNICIPAL WASTEWATER TREATMENT
WITH RBC TECHNOLOGY IN THE U.S.
By
Warren H. Chesner, Ph.D., P.E,
Manager, New York Office
Roy F. Weston
John J, lannone
Associate Project Engineer
Roy F, Weston
1.0 OVERVIEW
The U.S. Environmental Protection Agency's Municipal Environmental
Research Laboratory (MERL) and Roy F. Weston, Inc. are currently engaged in
the evaluation of the Current Status of Municipal Wastewater Treatment With
Rotating Biological Contactor (RBC) Technology in the United States.
The study is outlined in six major categories:
1. Review of Available Process Equipment
2. Identification of Existing Facilities
3. Review of Existing Design Procedures
4. Evaluation of Field Process Performance Relative
To Design
-------
5. Evaluation of Field Operating Difficulties
6. Evaluation of O&M and Power Requirements
This paper summarizes the data collected and evaluated to date,
briefly outlining equipment, existing facilities, design guides and power
requirements, focusing on operating difficulties at surveyed facilities.
A more detailed report is envisioned by early summer,
2.0 AVAILABLE PROCESS EQUIPMENT
KBC equipment can be placed into three basic categories: the media;
the mechanical or drive components; and the tank or reactor. Each
equipment manufacturer offers their own variation of media and drive com-
ponents .
Table 2.1 highlights the nominal parameters associated with the media,
mechanical components and tanks.
Some previous difficulties at RBC facilities have resulted from
equipment problems, which affect both the mechanical and the process
performance of the system. The media material, support, shaft strength, tank
shape, baffling arrangement and clearance are some of the items which have
adversely affected previous RBC performance. Further discussion of these
problems are addressed in Section 6.0,
3.0 IDENTIFICATION OF EXISTING FACILITIES
There are approximately two hundred-sixty three RBC installations
currently treating municipal wastewater, and fifty-eight installations
treating industrial wastewater in the United States. Table 3.1 lists the
distribution of municipal treatment facilities by flow range. Approximately
twenty-five percent of the existing facilities are package plants provided
as complete systems by RBC manufacturers. The largest operating RBC
facility is an eighty snaft, fifty-four MGD facility at Alexandria, Virginia.
Figure 3.1 displays the regional distribution of RBC municipal
facilities in the United States. The vast majority are located in the north,
eastern and midwestern portion of the country. New York State "has the
largest number of facilities, but approximately eighty percent of the
thirty-nine identified plants are package plants.
4.0 REVIEW OF EXISTING DESIGN PROCEDURES
Carbonaceous RBC removal rates per unit surface area are related to
wastewater concentration, flow rate and temperature.
54
-------
TABLE 2.1
BACIS RBC EQUIPMENT DIMENSIONS
MEDIA;
Disc
Shape
Material
Diameter
Surface Are*
Spacing
Construction
Circular
High Density Polyethylene
Standard: 12.0 Feet
Range: 6.0 - 21.0 Feet
Per 25-Foot Single Stage Shaft -
Standard Media; 100,000 - 104,000 Square Feet
High Density Media: 150,000 - 156,000 Square Feet
Standard Media; 1.25 Inches
High Density Media; 0.75 Inches
Segmented (8 Pieces): Steel Supported
Unltlzed: Heat Welded Self-Supported
MECHANICAL:
Shaft
Shape
Material
Thickness
Length
Cross-Section: Octagonal, Round Square
Steel
Municipal; 0.75 Inches
Industrial: 1.50 Inches
Standard: 25 Feet
Range: 6.0 - 25 Feet
Motors
Horsepower Ratings
3.5, 5.0, 7.5
Drive Units
Multl-V-Belts
Chain and Sprocket
Enclosed Cartridge
TANK:
Shape
Material
Contoured, Filleted, Flat Bottom
Steel (0.25-Inch Thickness)
Concrete (1.0-Foot Thickness)
Baffles
Overflow; Underflow
-------
TABLE 3.1
TOTM.JjUHiEB OF OPERATING BBC IHSmtATIOHS (1979)
Type
Municipal
cn
cr>
lndustri«1
Flow Range (HOD)
Total
0
0.1
0.5
1.0
5.0
10.0
0
0.1
0.5
1.0
5.0
- 0.1
- O.S
- 1.0
- 5.0
- 10.0
- 20.0
> 20.0
Sub-ToUt, Municipal:
- 0.1
- 0.5
- 1.0
- 5.0
- 10.0
(1)
74
57
41
T>
8
5
... k
263
20
13
2
5
0
18
Sub-Tot«1, Industrial: 58
(1) Size distribution not »v»!1»bl«
-------
©
-
/i "/
o\©
FIGURE 3.1
REGIONAL DISTRIBUTION OF
OPERATING RBC PLANTS
-------
In recent years, EEC systems have been mathematically simulated,
attempts have been made to establish scale-up factors, and operating data
has been fitted to design curves.
Current field design procedures are based upon manufacturers*
established curves which define effluent concentrations (soluble or total
BOD) in terms of applied hydraulic loading and influent concentration, as
illustrated in Figure 4.1. Starting with a desired efficiency or effluent
concentration and a given influent concentration, the designer simply
selects the defined hydraulic loading which establishes the surface area
requirement. These curves (Figure 4.1) are based upon field operation and
have been established by developing relationships such as that presented in
Figure 4.2. The shaded area shown in Figure 4.2 represents a range of disc
performance as observed by the manufacturers.
It is generally agreed that direct scale-up of RBC's, based upon unit
area removal rates developed in pilot studies on small discs, cannot be .
extrapolated to full scale twelve-foot diameter installations. As a result,
facility design is currently based upon previous operating experiences
(design curves) or full scale pilot studies.
Staging RBC units should theoretically optimize the process by
establishing desirable cultures in each stage and maximize removal rates.
Current design guides recommend staging to achieve high removal rates,
but have not definitively established the relationship between staging and
existing design curves.
Surface corrections are recommended for wastewater temperatures below
55°F. Figure 4.3 illustrates three different existing temperature correction
curves currently recommended. The lower curve illustrates temperature
correction based upon a decreasing reaction defined by K2/K1 = 1.042(T~20).
The upper curves simulate greater reductions in reaction rates at these lower
temperatures.
5.0 POWER REQUIREMENTS
Power drawn by operating RBC units is a function of disc surface
area, rotational speed and the organic removal rate of the individual RBC
shaft. Energy requirements increase as each of the above parameters
increase. Variations in the media configuration of each manufacturer can
also be expected to affect the drag and hence, the energy utilization of
the system.
58
-------
Q
O
m
o
z
ui
100
75
50
25
5O
4O
o
S 30
20
u.
u.
UI
10
0
HYDRAUUC LOADING (6PD/FT.2)
BASIS' TOTAL BOD
INFLUENT SBOD(mg/l)
i
150
INFLUENT BOD
CONCENTRATIONS (mg/l)
J
K>
IO
HYDRAULIC LOADING (GPD/FT.2}
BASIS'SOLUBLE BOD
FIGURE lf.1
EFFLUENT DESIGN CURVES
59
-------
o
U.
*
o
w
O
O
o
5
Q
X
to 3
CO
-J
«-•
o
UJ
o
Z
UJ
a
o
o
03
22§mg/I
'
1
I
1
345
HYDRAULIC LOADING (6PD/FT.2)
FIGURE 1}.2
DESIGN LOADING CURVES
6
8
-------
5.0
DC
O
4.0
3.0
I-
O
bJ
er
cc
o
o
a: 2.0
Ct
kl
O.
bJ
1.0
40
45
5O
55
WASTEWATER TEMPERATURE
FIGURE 4.3
TEMPERATURE CORRECTION CURVE
61
-------
Results of a literature survey of RBC power requirements with
respect to trickling filters and conventional activated sludge is presented
in Table 5.1, which categorizes RBC energy requirements per million gallons
between trickling filters and activated sludge systems. Such an analysis
is helpful in preliminary assessments of energy requirements, but must be
tempered with organic removal rates and efficiences to fully assess the
actual effective use of power supplied for treatment.
Preliminary results of field power studies are reported in Table 5.2,
which depicts metered power drawn at five operating facilities. Power
drawn per one-hundred-thousand square feet or shaft average 2.9 kw varying
from 1.9 to 4.1 lew. Power tests (manufacturers) conducted on clean media
at two locations, average approximately 1.4 kw per one-hundred-thousand
square feet. Energy utilization per million gallons varied from
five hundred-fifty five to three hundred-forty two kw-h^lG, averaging
four hundred-ten kw-h/MG. This is consistent with the range observed in the
literature survey (Table 5.1).
Energy utilization per million gallons is related to the hydraulic
loading and the rotational speed. Higher hydraulic loadings result in
more favorable energy/flow ratios; however, these loadings must be
commensurate with effective treatment. Lower rotational speeds will also
reduce metered kw, but must insure adequate mixing and aeration.
6.0 FIELD OPERATING DIFFICULTIES
Seventeen RBC facilities have been surveyed to date to aid in the
review of operating and performance problems. These seventeen facilities
were selected because they represent facilities which have equaled or
exceeded eighty percent of their design flow and, as a result, were
considered more likely to display process operating difficulties if they exist.
Table 6.1 outlines all reported difficulties. Some of these are
minor and some are major. Some are directly attributed to RBC equipment and
operation and some the result of design omissions and/or questionable
operation practices. The analysis is designed to identify and review those
reoccurring problems which limit RBC performance, and to
identify problems which can be minimized in future facilities.
Table 6.1 tabulates the actual and design hydraulic loadings of each
plant, highlighting structural difficulties reported along with design
loading problems and their impacts.
6.1 STRUCTURAL PROBLEMS
Structural problems are divided into three categories: media; shaft;
and bearing. Of the seventeen plants surveyed, four experienced some media
difficulties; three shaft problems; and two bearing problems.
62
-------
TABU 5.1
LITERATURESURVgYj,. JEHjEKCT REQUIREMENTS OF SECONDARYTREATMENT PROCESSES
Protest..
Kumber Of Plant*
Average KVH/HC,
Range KVm/HG
Source
Trickling Filter
633
(132)
-------
*»
16
Flex (HGD)
0,2
0.5
1.3
2,8
TABLE 5.2
FIELD POWER TEST RESULTS
Hettred KW Oraw/100.000 ft.
2.6
1.9
2.7
3,3
Average: 2.9
KU-Hr./HCC
555
357
410
3W,
3M
Averiga: 410
Hydrjulle lotdlnq
H.L. RPH
1.18 1.5
2.02 1.75
1.13 1.0
1.89 1.8
2.30 1.6
Cl««n
(1) Stamford Htdla
(2) Hanufacturers Av«r»ge Power Tests «t frlnceton, Illinois »nd Platnvllle, Connecticut
-------
TABLE 6.1
REVIEW OF OPERATING HBC FACILITIES
on
in
Plant
No.
1
2
3
4
5
6
7
B
9
10
It
12
13
H)
15
16
17
Design Flow
(MCD)
1.9
1.13
0.65
0.20
0.216
0.5
1.3
0.39
1.0
0.5
0.6
0.65
1.75
1.0
2.0
2.8
0.35
Percent
Of Design
108
83
80
85
91
100
100
136
80
100
80
1.00
80
80
100 (2)
103 (3)
95 '
Design Hydraul Ic
Loading GPD/Ft.2
1.72
2.17
3.70
1.18
1.03
2.02
1.13
2.85
2.5
1.56
1.0
2.6
5.72(1)
1.89
2.5
2.3
1.75
(Totals:
S true tufa! Problems
Media
X
X
X
X
i.
Shaft
X
X
X
3
Bearing
X
X
2
Design Load In
Raw Uastewater
X
X
X
X
X
X
X
X
X
X
X
X
X
X
111
q Problems
S 1 destreams
X
X
X
X
It
Impacts
D.O.
X
X
X
X
X
X
X
7
(5)
Beggtotoa
X
X
X
X
X
X
6
SS Accumulations
X
X
X
X
X
X
6
Hydraulic Overload:
Washout and Diluted Waste
X
X
X
X
X
X
X
7
Effluent Requirr-
ment Problems
BOO
X
X
w
2
SS
X
X
X
X
(«*)
i»
NHj-H
X
X
2
(1) Followf trickling filter
(2) 1/2 of plant RBC's In use
(3) 2/J of plant RBC'J tn use
(It) UC;t followed by polishing lagoon
{5} Nuisance bacterial growth
-------
Table 6.2 details the specific problems associated with the structural
difficulties. Reported media problems include: movement of segmented,
steel supported discs, which result in plastic shear, failure of the hub
which links the media to the shaft in unitized systems, shifting media on
support rods and ultraviolet degradation of the media. Reported shaft and
bearing problems include a total of six shaft failures in three plants and
two bearing problems; one associated with flooding, and the other with
improper lubrication.
Corrective actions for structural failures in almost all cases
require equipment replacements. Shaft failures and/or media movement or
failure are major equipment problems, which must be corrected if RBC
systems are to be considered viable treatment options. Media degradation
from exposure to ultraviolet radiation, results in brifctleness and can lead
to ultimate structural failure.
Equipment manufacturers are aware of the importance associated with
structural problems. Recent actions by major manufacturers have been aimed
at improving the structural integrity of the media, their support and shaft
strength. The addition of supplementary carbon to the media and insuring
that the media is continually covered, should help reduce the affects of
ultraviolet degradation.
6.2 DESIGN LOADINGS
Of the seventeen plants surveyed, Table 6.1 lists fourteen reported
variations in raw wastewater loading and four reported variations resulting
from sidestreams. Impacts associated with loading difficulties include:
seven reporting low DO concentrations; six reporting undesirable bacterial
growth; four reporting solids accumulation in undesirable locations; and
seven reporting washouts or diluted wastewater resulting from hydraulic
overloads.
Table 6.3 further defines the cause of the individual loading
problems at each facility. Difficulties include excessive hydraulic loads
resulting from inflow and infiltration (I/I); excessive flow resulting from
water running in the winter to avoid freeze-up, and peaking. Unaccounted
industrial contribution, sidestreams, and septage all increase organic loads
above those anticipated, Sidestreams include aerobic digestor, anaerobic
digestor, and sludge lagoon supernatant return. Poorly operated digesters
can substantially increase the return solids as well as organic load to the
RBC unit. Long collection system detention times were often associated
with incoming septic waste. The absence of primary clarifiers increased
the susceptibility of the RBC systems to potential solids and organic
loading problems. Excessive solids detention time in the primary clarifier
helped create septic conditions and poor settling. Solids deposition in the
channel and the tank were deemed significant in creating septic environments.
Solids deposition in tanks have been attributed to insufficient tank mixing
and dead spots, resulting from poor designs and overflow baffles. Excessive
equalization basin aeration in one facility increased the soluble organic
loading to the RBC unit by hydrolizing some of the suspended solids.
66
-------
TA8L€ 6.5
STRUCTURAL DIFFICULTIES
Descrlpt Ton
Movement Of Segmented Discs; Plastic Damage
Plastic Hub Failure In Unlttzed Construction
Exposure Of Media To Sunlight; Ultraviolet Degradation
Shifting Media On Supports
Corrective Action
Support Tightened
Replaced
Replace Damaged Cover (Carbon Black)
Eliminate Play Between Media and Supports
Description
Shaft Failure; Two Shafts
Shaft Failure: Threa Shafts
Shaft Failure; One Shaft
Corrective Action
Replacement
Replacement
Replacement
Description
Bearing Seizure; Pumps Flooded
Uneven Bearing Wear; Lack Of Proper Lubrication
Corrective Action
Replacement
Repair Or Replace
-------
Corrective actions for the itemized difficulties in Table 6.3 vary
widely. Improper design loadings resulting from poor facility planning is
difficult to correct after installation without increasing the plant
capacity or reducing the loading. Sidestreams are loadings which are often
omitted in the design of wastewater treatment facilities. Designers must
account for these loads in the design of RBC systems. Since in many cases,
Sidestreams are dependent upon sludge thickening, dewatering, and
treatment methods, current design curves (typical domestic wastewater) do
not fully account for these loads. Given existing ~si3estream~ conditions,
operators should avoid shock return loads and timewise distribute the
recycled flow to the head of the facility. Channel, tank aeration, and the
elimination of dead spots can help reduce solids accumulation in existing
facilities. Where this is not feasible, solids must be withdrawn directly.
6.3 NUISAMCE BACTERIAL GROWTHS
Microbes of many kinds are present in RBC films. Conditions which
favor the growth of nuisance organisms that result in the colonization
of initial stages of RBC media, have been reported. Resultant impacts
include reduced organic removal rates and a shifting of loadings from the
initial to the latter stages of a multistage system.
A sulfide oxidizing filamentous aerobe and microaerophile, from the
Beggiotoa family commonly referred to as Beggiotoa, has in some cases been
Identified as the nuisance organism and has become associated with this
particular problem in RBC systems. For purposes of this discussion, the
nuisance organism will be referred to as Beggiotoa, This slow growing
population has a milky white appearance when predominating RBC surfaces.
The odor from the film in this condition is somewhat septic and unpleasant
and markedly different from a healthy disc culture. The biomass is also
thinner than would be expected for a first stage.
Beggiotoa problems have been associated with two reoccurring
environmental conditions:
1, Low DO
2. Reducing environment capable of sulfide production
Since Beggiotoa are microaerophiles, they can exist at low DO
concentrations. Low DO, coupled with sulfide availability, weuld appear to
present optimum condition for Beggiotoa propogation at the expense of other
aerobic microbes.
Six of the seventeen plants surveyed reported nuisance organism or
Beggiotoa problems. Almost all of these plants reported low DO and side-
streams or suspended solids accumulation creating conditions for sulfide
production. Two facilities surveyed did not report low RBC tank dissolved
oxygen concentration, but did report suspended solids accumulation and
localized septic conditions making sulfide production possible.
68
-------
TABLE 6,3
LOADIHC DIFFICULTIES
en
Number Of Plants
7
1
Description
Predominant Effect
I/I
Winter Tap Flow {Avoid Freezing)
Penklng/Hydraullc Shock toads
Industrial Contributions
Sldest reams
Septage Contribution
Long Collection System Detention Time
Ho Primary Clarification
Solids Deposition In Channels
Solids Deposition In Reactor
Septic Primary Clsrlfler Sludge
Excessive Equalization Basin Aeration
Hydraulic Overload; Diluted Waste
Hydraulic Overload; Diluted Waste
Sol Ids Washout
Organic Overload; DO Deficiency; Sulflde Production
Organic Overload; 00 Deficiency; Sulflde Production
Organic Overload; BO Deficiency; Sulflde Production
DO Deficiency, Sulflde Production
High Organic and Solids Load; DO Deficiency; Sulflde Production
DO Deficiency; Sulflde Production
Sulflde Production
DO Deficiency; Sulflde Production; Increased Soluble Organic Loading
Increased Soluble Organic Loading
-------
It is currently felt that the condition which leads to sulfide
production (incoming septic waste, suspended solids accumulation, long
primary clarifier detention times) is the predominant factor associated with
Beggiotoa problems. Beggiotoa are capable of surviving in highly aerobic
or microaerobic conditions. Under microaerobic conditions, their
competitive advantage is probably enhanced. Microaerobic conditions will
exist xcithin the RBC film at depths where oxygen penetration becomes limiting.
If sulfides are available, this environment is probably capable of
catalyzing a Beggiotoa take over.
Potential corrective actions include; the elimination of conditions
which cause the sulfide production; oxidation of sulfide prior to the RBC
systems.
7.0 CLOSING
Preliminary investigation of BBC systems have identified operating
difficulties which are currently being addressed by manufacturers, design
engineers and researchers. Some of these difficulties are common to
biological systems, while some are unique to RBC systems. Major items
needing resolution include:
- Solution to equipment problems
- Continued design procedure improvements
- Solutions to nuisance organism problems
Continued RBC research and data generation will continue to enhance
our ability to utilize the treatment capabilities of these systems.
70
-------
PART II. PROCESS VARIABLES AND BIOFILM PROPERTIES
HYDRAULIC CHARACTERISTICS OF THE RBC
By
Harvey Olem
Environmental Engineer
Division of Water Resources
Tennessee Valley Authority
Chattanooga, Tennessee
Richard F. Unz
Associate Professor
Department of Civil Engineering
The Pennsylvania State University
University Park, Pennsylvania
INTRODUCTION
In the course of pilot scale investigations with the rotating
biological contactor (RBC) in the treatment of acid mine drainage (1),
divergent results were obtained with different size treatment units when
equivalent hydraulic loading and disc peripheral velocities were applied.
It was the objective of this study to learn if the observation of lower
ferrous iron removal efficiencies with a larger size RBC was related to
hydraulic characteristics of the unit.
BACKGROUND
Full scale RBC systems, scaled according to pilot plant data, may
not always meet design expectations. Godlove et al. (2) found that
soluble COD removal from petroleum refinery wastewaters was about 14
percent less efficient in full scale RBC systems than in pilot plant
units. Murphy and Wilson (3) reported about 16 percent lower COD
removal efficiencies for a 2.0-m diameter RBC unit treating wastewater
71
-------
at a peripheral velocity equivalent to a parallel 0.5-m unit. Chesner
and Molof (4) found that smaller discs operated at equivalent peripheral
velocities to larger discs resulted in increased oxygen transfer with
correspondingly higher COD removal efficiencies. Recently, Friedman,
et al. (5) described experiments with a pilot scale RBC unit operating
at different rotational speeds and it was concluded that heavily loaded
or older plants approaching design load may not be able to meet design
performance because differences in oxygen transfer exist between pilot
and full scale systems operated at equivalent peripheral velocites.
In previous work performed by Olem and Unz (1) involving the treat-
ment of acid mine drainage by the RBC, ferrous iron oxidation efficiency
of the 2.0-m RBC was about 10 percent lower than that of the 0.5-m unit
when operated in parallel and under equivalent conditions of peripheral
velocity and hydraulic loading rate (Figures 1 and 2). The discrepancy
was not believed due to an oxygen deficiency, however, since high reactor
dissolved oxygen concentrations (greater than 10 mg/1) were observed for
both units, presumably owing to the low temperatures (typically 10 C)
and oxygen demand of the mine water relative to wastewaters rich) in
organic matter. '
In this study, chemical tracers were employed to determine if signi-
ficant hydraulic differences exist between the different size RBC units
when operated at equivalent hydraulic loading rates and peripheral disc
velocites.
METHODS
Rotating Biological Contactors
Field studies were conducted employing commercial RBC pilot units
(Autotrol Corp. Milwaukee, Wise.) equipped with 0.5- and 2.0-m diameter
discs (Figure 3). Each unit consisted of four stages of closely spaced
and corrugated high-density polyethylene discs suspended on a shaft with
approximately 40 percent of the surface area immersed in the volume of a
corrosion-proof trough. The individual stages were separated by baffles.
Flow through the 0.5-m unit was facilitated by a 2.5-cm diameter hole
in each baffle (Figure 4, left). A serpentine flow pattern was formed
in the 2.0-m unit by an arrangement of piping on the outside of stage
compartments (Figure 4, right). A rotating bucket mechanism controlled
mine water feed from the influent chamber of each unit to stage
compartments.
RBC units were enclosed in a building for protection from the
natural environment. Mine waters were delivered from the source with
the aid of two 0.25-kW centrifugal pumps in connection with a 7.6-cm
diameter foot valve and appropriate lengths of flexible polyethylene
piping. Mine water was pumped to the feed chambers of pilot units at
slightly greater than the desired flow rate of 119 m3/d. Excess flow
from the feed chamber was discharged through an overflow pipe in order
to maintain a constant liquid level of fresh mine water in the feed
chamber. The static level of mine water in the feed chamber and the
72
-------
100
90
u
fc_
-------
100
90
•*-
c
«
u
t-
0>
^80
x 70
60
0
i
Y= 0.636 X + 70.6
R = 0.668
468
TEMPERATURE,°C
10
12
Figure 2. Ferrous iron oxidation efficiencies obtained at different
acid mine drainage temperatures with the 2.0-m RBC.
74
-------
FEED BUCKET
DRIVE SYSTEM
FEED
CHAMBER
INFLUENT
STAGES OF DISCS
EFFLUENT
Figure 3. Schematic of rotating biological contactor pilot units.
75
-------
INFLUENT
FEEDWELL
Si
S2
S3
*
S4
LO
EFFLUENT
INFLUENT
TT
FEED WELL
Q
s +
©
"*"
S2 -*-
(D
S3 +>
O
S4 •+-
EFFLUENT
Figure 4. Plan and side view showing flow pattern and stage sampling
locations for trough of 0.5-ra RBC (left) and 2.0-m RBC
(right). Drawing is not to scale.
76
-------
number of feed buckets employed was in accordance with the desired flow
rate. Flow rates were checked by collecting effluent samples for one
minute in 3-1 and 100-1 containers, respectively, for the 0.5- and 2.0-m
units. Both units were operated at an equivalent peripheral disc velocity
of 19 m/min and a hydraulic loading of 0.16 m3/d-m2. Attainment of a
peripheral velocity of 19 m/min required 13 and 2.9 rpm for the 0.5- and
2.0-m unitsj, respectively.
Tracer Additions
During equilibrium operation of the RBC units, a concentrated solu-
tion of lithium chloride (1.0 g/1) was added instantaneously to one feed
bucket of each RBC unit just prior to discharge into stage one. The
volume of tracer solution added to the 2.0-m unit was proportionately
larger on the basis of trough volume. Samples collected from each
stage and effluent at specified time intervals were analyzed for lithium
content by atomic absorption spectrophotometry. In addition, as a check
on the validity of the lithium technique and for convenience of field
analyses, sodium chloride was employed as an independent tracer and
monitored by specific conductance. A concentrated solution of sodium
chloride (230 g/1) was added simultaneously along with the lithium
chloride solution.
RESULTS AND DISCUSSION
Recovery of lithium chloride and sodium chloride tracers in each
stage of the RBC units is presented in Figures 5 and 6. There existed
a background specific conductance in the mine water which ranged from
1020 to 1100 ymhos/cm. There was no detectable background level of
lithium. The final set of samples for lithium analysis was collected
20 min following the attainment of baseline specific conductance
(160 min).
Samples collected from stage one of both units revealed peak
lithium content within 30 sec of tracer addition. Higher lithium con-
centrations were recovered from stage one of the 0.5-m RBC than from
the same stage of the 2.0-m unit, which indicated more short-circuiting
of flow in the larger size RBC to subsequent stages.
Pintenich and Bell (6) described a procedure for evaluation of
tracer data to quantify the hydraulic characteristics of continuous
flow treatment basins. The method is based on comparison of the shapes
of standardized tracer recovery curves to those for an ideal completely
mixed basin. Relative amounts of complete mix, plug flow, and dead space
volumes may be obtained. Overall lithium recovery curves for the 0.5-
and 2.0-m RBC units are presented in Figure 7. Percent recovery curves
were developed by integration of the areas under a standardized recovery
curve (C/C vs. t/T).
Both units were found to be approximately 80 percent completely
mixed. Dead space in the trough of the 2.0-m RBC, although relatively
low for continuous flow reactors, was 10 percent as compared to 2 percent
77
-------
A STAGE I
o STAGE 2
° STAGE 3
EFFLUENT
0 20 40 60 80 100 120 140 160
TIME, min
0 20 40 60 80 100 120 140 160
TIME, min
Figure 5. Recovery of lithium chloride tracer in each stage of 0.5-m
(top) and 2.0-m (bottom) RBC units.
78
-------
4500 -
A STAGE 1
O STAGE 2
D STAGE 3
A EFFLUENT
o
o
lij
85 2000
1000
-10
3000
1000
-10
30
50 70 90
TIME, min
no
130
150
10
30
5O 7O 9O
TIME, min
130
150
Figure 6t Recovery of sodium chloride tracer in each stage of 0.5-m
(top) and 2.0-m (bottom) RBC units.
79
-------
o
o
o
2.0
1.5
1.0
0.5
0
* 78 %
P =20%
D = 2 %
0 0.5 1.0 1.5 2.0
t/T
2.5 3.0
-------
for 0.5-m unit. The larger RBC displayed only 10 percent plug flow
characteristics while the 0.5-m unit had 20 percent plug flow volume.
Pintenich and Bell (6) observed good performance for clarifiers with
15 to 20 percent plug flow characteristics. Baffled basins such as in
the RBC should show similar plug flow characteristics. It was antici-
pated that the serpentine flow pattern of the 2.0-m RBC unit (Figure 4)
would allow improved plug flow characteristics over the 0.5-m unit.
Villemonte and Rohlich (7) described certain dimensionless ratios
which may be employed to evaluate the hydraulic efficiency of continuous
flow reactors. These indices were applied to the results of lithium addi-
tion to RBC units (Tables 1 and 2). The ratio, t /T» is a measure of
short-circuiting, dead spaces, and effective tank volume. It should be
near zero for the first stage of the RBC (ideal mixing) and near unity
for the final stage (ideal settling or plug flow). Comparison of this
index for the two RBC units revealed more short-circuiting, dead spaces,
and a lower effective tank volume in the 2.0-m unit. Another ratio,
tgo/tiQ, measures dispersion of reactor contents; a mixing function.
The 2.0-m unit displayed near ideal dispersion of reactor contents in
the first stage. The dispersion ratio decreased considerably for the
latter stages of both units, presumably due to the baffled, multi-stage
configuration.
Although RBC units were operated at equivalent hydraulic loading,
different ratios of surface area-to-trough volume for the units resulted
in different theoretical retention times (Table 1). The lower theoreti-
cal retention time for the 2.0-m 1BC may also have been a factor in the
observation of divergent treatment performance between the two units.
In addition, there exists a sizeable difference in available surface
area for the two RBC units (Table 3). Hydraulic loadings applied in
this study were based solely on disc surface area. However, a greater
proportion of the total surface area in the 0.5- than in the 2.0-m unit
was attributed to the trough. Olem and Unz (8) observed similar viable
iron-oxidizing bacterial densities in comparison of trough and disc
surfaces of the same stage. Thus, calculation of the effective surface
area of smaller diameter RBC units should include trough surfaces. It
is likely that if the units had been sized on the basis of total available
surface area, closer agreement would have been obtained in tracer evalua-
tions and comparative treatment performance.
Friedman et al. (5) concluded that new, lightly loaded RBC systems
should be able to meet design expectations when sized on the basis of
pilot plant data. This would only be applicable to lightly loaded
systems because differences in oxygen transfer between pilot and full
scale would be less important. Similarly, differences in oxygen transfer
for the two different size RBC units in treatment of acid mine drainage
were likely very low owing to the high reactor dissolved oxygen levels
present. The combination of a longer residence time for mine water in
the trough of the smaller RBC and relative differences in total available
disc surface areas probably accounted for much of the difference in
performance observed between the two units. The exact contribution of
observed hydraulic differences to the divergent treatment results is
not known.
81
-------
Table 1. Cumulative flow through times for lithium tracer
in stages of 0.5- and 2.0-m BBC units
Time, min
0.5-m RBC Stages 2.0-m RBC Stages
Parameter 1234 1234
t a 3 15 36 47 3 6 27 37
P
t1Qb 1.9 8.1 17.3 28.9 1.8 5.1 15.0 20.1
t-.C 33.9 54.3 77.0 98.7 36.1 53.4 76.3 96.5
-------
Table 2. Comparison of dimensionless time ratios for
0.5- and 2.0-m RBC units.
Parameter
0.5-m RBC Stages
1234
2.0-m RBC Stages
1234
t /T
P
0.22 0,55 0.88 0.86
17.8 6.7 4.5 3.4
0.25 0.25 0.75 0.77
20.1 10.5 5.1 4.8
Measures average short-circuiting, dead spaces, and effective tank
, volume. It is 1.0 for ideal settling and zero for ideal mixing.
Measures dispersion. It is 1.0 for ideal settling and 21.9 for ideal
mixing.
83
-------
Table 3. Comparison of available surface area for
different size RBC units.
RBC Unit
Bench Scale
Pilot Scale
Prototype
Full Scale
Disc
Diameter
15 cm
0.5 m
2.0 m
3.6 m
Surface
Disc
0.438
21.8
738.1
9,290
2
Area, m
Trough3
0.10
2.14
18.8
50
Portion of Total
Surface Area Due to
Trough, Percent
19.1
8.9
2.5
0.5
Available surface area was calculated from dimensions of experimental
RBC units and estimated for full-scale RBC by use of one 7.6 m shaft
placed in a contoured basin.
84
-------
AND CONCLUSIONS
Two conservative tracers were employed to liydraulically characterize
two parallel RBC units with different disc diameters under equivalent
conditions in order to determine possible explanations for the observa-
tion of lower ferrous iron removal efficiencies in the treatment of acid
mine drainage. The observed decrease in treatment performance for scale-
up to the 2.0-m RBC was due, in part, to the following factors:
1. The 0,5-m RBC more closely simulated the ideal hydraulic
characteristics of a series of completely mixed treatment
basins than did the 2.0-m unit.
2. The residence time of mine water in the 0.5-tn RBC trough
was longer than the larger unit when operated at equivalent
hydraulic loading rates (54.4 min vs. 48.0 min) due to
differences in surface area-to-trough volume ratios for the
two units.
3. The proportion of total RBC surface area assumed by the trough
floor was greater in the 0.5-m RBC than in the 2.0-m RBC. In
the calculation of hydraulic loading, only the disc surface
area is considered. Since the trough area, which is colonized
by active iron oxidizing bacteria, does not enter into the
determination of effective surface area, it is apparent
that the 0.5-m RBC would demonstrate the greater treatment
capacity of the two units.
ACKNOWLEDGEMENTS
This work was performed for the Institute for Research on Land
and Water Resources at The Pennsylvania State University under Grant
R-805132 from the U.S. EPA. At the time this work was performed,
Harvey Olem was a doctoral candidate in the Department of Civil
Engineering, The Pennsylvania State University.
REFERENCES
1. Olem, H. and R. F. Unz, "Rotating-Disc Biological Treatment of Acid
Mine Drainage." Jour Water Poll. Control. Fed. (In press).
2. Godlove, J. W., W. C. McCarthy, H. H. Comstock, and R. 0. Dunn, "Kansas
City Refinery's Wastewater Management Program Using Rotating Disc
Technology." Paper presented at 50th Annual Conf, Water Poll.
Control Fed., Philadelphia, Pa. (Oct. 1977).
3. Wilson, R. W., K. L. Murphy, and J. P. Stephenson, "Effect of Scale-
Up in Establishing Design Loadings for RBC's." Paper presented
at 51st Annual Conf. Water Poll. Control Fed., Anaheim, Calif.
(Oct. 1978).
85
-------
4. Chesner, W. H. and A. H. Molof, "Biological Rotating Disk Scale-
Up Design: Dissolved Oxygen Effects." Progr. Water Technol.
9.:811-819 (1977).
5. Friedman, A. A., L. E. Robbins, and R. C. Woods, "Effect of Disk
Rotational Speed on Biological Contactor Efficiency." Jour.
Water Poll. Control Fed. 5±:2678-2690 (1979).
6. Pintenich, J. L. and R. A. Bell, "Hydraulic Characterization of
Wastewater Treatment Basins." Proc. 33rd Ind. Waste Conf.
Purdue University, W. Lafayette, Ind. Ann Arbor Science Publ.
449-456 (1979).
7. Villemonte, J. R. and G. A. Rohlich, "Hydraulic Characteristics
of Circular Sedimentation Basins." Proc. 17th Ind. Waste
Conf., Purdue University, W. Lafayette, Ind. 682-702 (1962).
8. Olem, H. and R. F. Unz, "Acid Mine Drainage Treatment with Rotating
Biological Contactors." Biotechnol. Bioeng. 19_: 1475-1491 (1977).
86
-------
PHYSICAL FACTORS IN RBC OXYGEN TRANSFER
By
Byung Joon Kim
Graduate Student, now with the
New York State Department of Environmental Conservation
Alan H. Molof
Department of Civil and Environmental Engineering
Polytechnic Institute of New York
Introduction
Although the Rotating Biological Contactors (RBC) process is
classified as an aerobic system, oxygen transfer rates have not been
sufficiently considered in present design. Rather, the emphasis has
been centered on an increase in the RBC surface area.
When Hartman (1960) introduced the practical use of RBC, he
claimed that the dissolved oxygen (DO) in the reactor did not have
significance in treatment efficiency because adequate amount of oxy-
gen could be supplied during the air exposed cycle. The importance
of the DO in the mixed liquor has been cited by workers in defining a
minimum oxygen balance. Welch (1968) presented data showing con-
siderable decline of treatment efficiency when operational DO concen-
tration dropped below 1.5 mg/1 for his operating conditions of 500
mg/1 COD and 30 minutes retention time. Weng and Molof (1974)
found that nitrification took place only in the stages where DO was
greater than about 2 mg/1 in a six stage laboratory reactor for oper-
ating condition of 178 mg/1 COD and 48 minutes per stage retention
time. However there has not been a corresponding emphasis on the
RBC as an aeration device to deliver the oxygen to maintain this DO
level. Furthermore, treatment of high BOD industrial wastewater
requires more study of the factors involved in RBC oxygen transfer.
The RBC biological film is fixed on the disk surface and is
rotated alternately between the air and the liquid. One effect of
rotation is to provide a means of better aeration by carrying a liquid
film into the air after the disk completes its liquid immersion cycle.
87
-------
Another effect is to provide turbulence in the mixed liquor surface
and subsurface volume increasing mass transfer. Oxygen transfer
takes place at the interfaces between the air-liquid film, liquid film-
microbial fixed film, air-mixed liquor, microbial fixed film-mixed liquor
and air-microbial fixed film.
This work involves a study of some of the physical factors
affecting RBC oxygen transfer into the mixed liquor. Three different
sized laboratory scale RBC units were used in non-steady state clean
water tests. The volumetric oxygen transfer coefficient (K,a)2Q was
calculated from the laboratory data. The physical parameters studied
included space between the disks, size of the disks, rotational velo-
city, peripheral velocity and number of disks per stage. It was the
purpose of the study to correlate the volumetric oxygen transfer
coefficient (KLa)20 with significant physical factors. As a result, the
volume renewal number (Nv) is developed from the theory and data
as a more practical efficient tool to predict physical oxygen transfer
in the RBC process.
RBC Oxygen Transfer
Bintanja et. al (1975) used the Yamane and Yoshida solution to
solve Pick's second law. The boundary conditions were:
t = 0, 0< x < 6, c = CL
t > 0, x = 6, c = Cc
S3
t > 0, x = 0, 3c/3x = 0
where 6 was the liquid film thickness on the disk. They con-
cluded that:
K •
.L M ^ j — OkV0'5 -
2o 6 ~ 6 when 5 < 0.8 (2)
Experimental K, value were 49% to 87% of the theoretical KT value.
lj Lt
Chesner and Molof (1976, 1977) found that the smaller RBC with
the higher RPM had better efficiency than the larger RBC with lower
RPM. The peripheral velocity was set at the same level that is used
in present plant design. They also reported that rotational velocity
was the better DO scale-up function than the peripheral velocity.
Friedman et. al (1979) confirmed the significance of rotational
velocity and presented an equation using Bintanja et. al (1975) data;
88
-------
In KL = 1.31 In u> + 14.78
where the unit of KL was 10" m/s and that of m was RPM.
Zeevalkink et. al (1978) solved the Navier-Stokes equation for
determining liquid film thickness on the RBC and also verified the
equation by experiment. They concluded that:
6 = 1.2 • (vc)°*5 * (10"4m) (3)
where v is the vertical component of the peripheral velocity at
c
the point where the disk emerges from the water.
Ouano (1978) correlated the overall liquid phase mass transfer
coefficient (K,.) and Reynolds number by dimensional analysis. The
result was:
V-/A,.
K
Dm
2
where D u>)3 was Reynolds number.
Zeevalkink et. al (1979) explained that the deviation of the
Bintanja et. al (1975) model was due to incomplete mixing in the
reactor and derived an experimental equation for
as KT = 2 I (1 - 4.21 exp n = )
' -
-------
3. The liquid film thickness is determined by the Zeevalkink
et, al (1978) model.
The volume renewal number can be defined as the ratio between
the liquid film flow rate and the effective reactor volume, or
Nv =Qf/VE
where, NV: volume renewal number (T
-1,
(4)
Qf: liquid film flow rate, which is the total film flow
3-1
volume per unit time (L T )
3
V£: effective reactor volume (L )
Under the above assumptions, correlation of the volumetric
oxygen transfer coefficient (K,.a) and the volume renewal number
(Nv) is expected as
KLa = a (Nv)b (5)
It should be noted that the use of a half circle submerged reac-
tor is an ideal case in order to simplify the calculations. In case of
any other physical conditions, the constants a and b will be changed
accordingly. The conceptual scheme is shown in Fig. 1.
The liquid film flpw rate can be derived by;
Q =
(ui
GO
c
(6)
-S ,
•SJ
ZQ.
RGURE I. CONCEPTUAL SCHEME
90
-------
where n/2 • (iu • D ) is a single disk surface area per unit time
carrying the liquid film into the air and (w
thickness (6) function from equation (3).
D)°'5 is the liquid film
The effective volume is:
= K,
S.
(7)
where the liquid volume underneath the disks,
n • e • (R + e/2) * (2S + t™), is neglected.
Therefore, the volume renewal number (Ny) is obtained by combining
equation (6) and (7) without including constants K, and K2.
v
D
0.5
,-1
Substituting equation (8) into equation (5) results in;
KLa = a(Nv)b = a(u)1'5 D0*5 S~lf
Experimental
(8)
(9)
Three different size laboratory scale geometrically similar RBC
units were designed to measure the dissolved oxygen concentrations
with time under various physical conditions. The schematic drawing
of the units and dimension data are shown in Figure 2 and Table 1
respectively. The disks were 6, 12, and 24 inch diameter flat circles
made from 0.25 inch thickness (t^) clear plastic.
23
00 £
meter
movable vail
FIGURE 2. LABORATORY SCALE RBC UNITS.
91
-------
Table 1. Reactor dimensions (unit: inch)
D
H
Ha
es
R==H +H +e
L
6.0
0.47
2.53
0.125
3.125
12.0
12.0
0.94
5.06
0.25
6.25
24.0
24.0
1.88
10.12
0.50
12.50
48.0
Movable walls and spacers were provided to vary the effective
volume of liquid CV~E). Effective liquid volume in the case of a single
disk at 85% submerged depth is shown in Table 2.
Table 2. Effective volume of reactor
S \ D
(1/4) (1/4)
(1/2) (1/2)
(3/4) (3/4)
(1) (D
(3/2) (3/2)
6 inch
106
208
310
412
12 inch
420
830
1240
1650
, unit-mi
24 inch
1700
3330
4960
6590
9840
The number of disks and the space between the disks were
varied. The space between disks was twice the space between the
disk and wall. Rotating velocity was varied from 5.2 to 164 RPM by
a gear and sprocket combination connected to a 1/6 HP A.C. motor.
The standard design peripheral velocity is about 1 ft/sec. The
peripheral velocity in this study ranged from 0.5 to 4 ft/sec.
The physical variables are summerized in Table 3.
Table 3. Physical variables.
D (inch)
N (No. of Disks)
io (RPM)
VD (ft/sec.)
S (inch)
6.0
1-3
84-21
4.0-0.5
0.25-2.0
12.0
1-4
84-10.5
4.0-0.5
0.5-2.0
24.0
1-3
42-5.2
4.0-0.5
0.5-3.0
The oxygen transfer test was basically the same as the non-
steady state clean water test procedure in Standard Methods (1975).
New York City (Brooklyn, N.Y.) tap water was aerated about twelve
hours to reach oxygen saturation. The water and room temperature
was maintained at 20 ± 0.5°C. The DO meter (Beckman 777 oxygen
analyzer) was calibrated with the oxygen saturated water. Winkler
tests were performed to confirm the oxygen saturation. Reagent
grade sodium sulfite and cobalt chloride were used to deoxygenate the
water. The test run was over when the reaerated sample become
about 90% to 95% saturated. Since accurate saturation values were
known, the log deficit least square method was used to analyze the
data. The temperature effect was compensated for by using
92
-------
(KLa)2Q = (KLa)T • (1.024) (20'T)
Results
The volumetric oxygen transfer coefficients (1C. a) were cal-
culated from the experimental data under the various physical con-
ditions .
The results of measured K.,a are presented graphically on full
log paper in Figure 3,4 and 5. Figure 3 depicts the disk size (D)
effect on K^a. A linear relationship between the disk diameter (D)
and K,a was shown. The averaged slope of 0.39 indicates that KTa
0 39
is proportional to D " . Figure 4 shows the effect of space (S)
between disks on KTa. The average slope was - 0.86. KTa is pro-
/\ oc Lt L>
portional to S ° .
Fig. 5 depicts the rotational velocity (tu) effect on It, a. The
slope of the lines was not constantly linear in contrast to the results
for D and S. The reason appeared to be a difference in hydraulic
regime. A discontinuity in the liquid film flow occured somewhere on
the dotted line in Figure 5 and ICa was drastically decreased under
this hydraulic condition. The average slope of 1.07 indicates that
1 07
K..a is proportional to ui . The number of disks each having the
same volume/area relationship was varied to detect the wall effect and
turbulence characteristics. No appreciable effect on KTa was noticed.
JLi
The combination of effects of the rotational velocity, the disk
size and the disk space resulted in;
Kra = K • u,1'07 • D0-39 - S-°'86 <10)
JLi
The equation coefficients strongly suggest that the experimental K-a
correlates well with the proposed NV. Figure 6 depicts the relation-
ship between the calcualted NV (w1*5 D0*5 S"1) and KLa on a full log
paper. The relationship can be expressed as
log KLa = 0.732 log (Nv) - 2.96
or KLa = 0.0011 (NV)'4 (11)
with the correlation coefficient (r) of 0.991. This equation is valid
for the clean flat disks, e/R = 0.042 and H /R™ = 0.15
a i
Discussion
One of Bintanjal et. al (1975) solutions was equation (2);
93
-------
0.08
0.04 —
D (inch ) l2
FIGURE 3. DISK 3Zp EFFECT
24
94
-------
/4
05
S (mch) i / 2
FIGURE 4- SPACE EFFECT
i. o
95
-------
105
21 RpM 42
FIGURE 5. ROTATIONAL VELOCITY EFFECT
96
-------
I.Or
08
06
Ugend
D
6 *
12 4
24 «
Crt
03
Q2
X
(minf
0.1
0.08
0.06
OW
o.o%r
^o
logKLa *
R • O.Sl
Q732 log
91
- 2.96
100
_^_ ^
N
1000 6030 800D IODQO
FIGURE 6. REUT10NSHIP BETWEEN ^ "& \ '.
-------
K = __ When - — ___ < 0.8 (2)
By using Zeevalkink et. al (1978) data for ui and 6, it can be calcula-
ted that at 6/(D tR) * =0.8 and 1.7, the corresponding peripheral
velocity was about 1 ft/sec, and 2.2 ft/sec, respectively. As the
disk size increases, tn increases. Therefore, 6/(D tn) ' will be
it In xv
less than 0.8 in most cases at plant scale . Thus, it appears that
equation (2) would better represent full scale plant performance.
A comparison of the Bintanja et. al equation (2) with the volume
renewal number (Nxr) would be of interest. Since 6 is proportional to
05
(u£» * and tR is inversely proportional to u», K, in equation (2) can
be expressed as;
KL = K u)' E' (12)
The common functions are noticed in equation (8) and (12). This
shows a possible application of the volume renewal number (Nv)
concept to full scale plant operation .
Ouano (1978) obtained a straight line by plotting
on full log paper. Since he did not vary the disk size, the
plotting was simply the relationship of 1C, and u>. The Ny. concept
shows that the u» term is more important than the D term. This is in
contrast to the Ouano equation where D is more important than tu.
Conclusion
The volumetric oxygen transfer coefficient (K,a) in the RBC
process was correlated with the volume renewal number (Nv) or
K, a=a(Nv) . The volume renewal number represented the liquid film
flow rate per unit effective reactor volume. The important physical
factors in calculation of the volume renewal number included the
rotational velocity (w), the disk size (D) and the space between the
disks (S). The volume renewal number (Ny) was defined as
Nv = u,1"5 D0-5 S'1.
98
-------
The relationship of KLa and NV under the conditions of e/R
0.042, H /R™ = 0.15 and a clean flat disk is:
a JL
log KLa = 0.732 log (Nv) - 2.96, or
O TV)
KLa = 0.0011 (Nvr''**
99
-------
References
APHA, AWWA and WPCF (1975). "Article 207 Oxygen Transfer"
Standard Methods for Examination of Water and Wastewater, 14th ed.,
82.
Bintanja H.H.J., Van Der Erve J.J.V.M., and Boelhouwer C. (1975).
"Oxygen Transfer in Rotating Disc Treatment Plant" Water Research,
Vol. 9, 1147.
Chesner W.H. and Molof A.H. (1976). "Relative Performance of Dif-
ferent Sized Biological Rotating Disks" presented at 49th Annual WPCF
Conference, Minneapolis, Minnesota.
Chesner W.H. and Molof A.H. (1977). "Biological Rotating Disk
Scale-up Design: Dissolved Oxygen Effects" Prog. Wat. Tech., Vol.
9, 811.
Friedman A.A., Robbins L.E. and Woods, R.C. (1979). "Effect of Disk
Rotational Speed on Biological Contractor Efficiency" Jour. Wat. Pollut.
Control Fed., Vol. 51, 2678.
Hartman H. (1960). "Untersuchung uber die Biologische Reinigung
von Abwassar mit Hilfe von Tauchtropfkorperanlagen" Stuttgarter
Berichte-zur Siedlungswas-serwirtschaft Kommissionsverlag, Band 9, R,
Oldenbourg, Munich, Germany (Translation).
Ouano E.A.R. (1978). "Oxygen Mass Transfer Scale-up in Rotating
Biological Filter" Water Research, Vol. 12, 1005.
Welch F.M. (1968)." "Preliminary Results of a New Approach in the
Aerobic Biological Treatment of Highly Concentrated Wastes" Pro-
ceedings of 23rd Purdue Industrial Waste Conference. Purdue Uni-
versity, Lafayette, Indiana.
Weng C.N. and Molof A.H. (1974). "Nitrification in the Biological
Fixed-film Rotating Disk System" Jour. Wat. Pollut. Control Fed.,
Vol. 46, 1674.
Zeevalkink J.A., Kelderman P. and Boelhouwer C. (1978).
"Liquid Film Thickness in a Rotating Disc Gas-Li o^iid Contactor" Water
Research, Vol. 12, 577.
Zeevalkink J.A., Kelderman P., Visser D.C. and Boelhouwer C.
(1979). "Physical Mass Transfer in a Rotating Disc Gas-Liquid
Contactor" Water Research, Vol. 13, 913.
100
-------
Nomenclature
2
A : Area where the mass transfer occurs (L )
2
A : Projected area of disk on water surface (L )
2
At : Surface area of reactor (L )
a,b,K : Constant.
D : Diameter of disk (D)
*? —I
D : Molecular diffusion coefficient (L T )
e : Distance from the disk rim to inner lining of reactor (L)
H : Distance from the disk center to the liquid free surface
a
v_
P
8
M
P
U)
(L)
H : Submerged disk depth (L)
S
KL : Overall liquid phase mass transfer coefficient (LT~ )
KTa : Volumetric Oxygen transfer coefficient (T~ )
L*
L : Reactor length (L)
N : Number of disks
NV : Volume renewal number (T )
Qf : Film flow rate (L3T"1)
R : Radius of disks (L)
R™ : Radius of reactor (L)
S : Half space between disks (L)
tR : Contact time per rotation (T)
v : Vertical component of the peripheral velocity at the ,
point where the disk emerges from the water (LT~ )
3
Effective reactor volume (L )
Peripheral velocity of disk (LT~ )
Liquid film thickness (L)
Absolute viscosity of liquid (MLT~ )
Density of liquid (ML" )
Rotational velocity (T"1)
101
-------
Page Intentionally Blank
-------
EFFECT OF CARBON, AMMONIA NITROGEN AND
HYDRAULIC LOADING RATES, RPM, AND EXPOSED SURFACE
AREA VARIATIONS ON RBC PERFORMANCE
By
George E. Hoag,
Research Assistant
and
Wendell H. 'Hovey
Assistant Professor
Department of Civil Engineering
University of Connecticut
Storrs, Connecticut
U.S.A.
INTRODUCTION
Sometimes a. system being studied is well understood, and it is possible
to postulate a plausible model from considerations of the governing physical
laws. This type of model would be a theoretical or mechanistic model because
it is based directly on the theory governing the process. However, the case
often arises where the system is too complicated, or the important physical
laws are not known, and a theoretical model cannot be postulated. In this
case an emperical model might be of value. The responses of the system to
changes in conditions often helps determine which physical laws are governing.
Then an attempt can be made to postulate theoretical models. Even before an
emperical model can be constructed, it may be necessary to determine which are
the controlling factors, and what their effects are on the system. The experi-
mental program reported on in this paper is such a screening study, designed
to determine which of a number of variables are important ones, and what their
effects are on process performance.
103
-------
In such an experimental program it might appear necessary to study every
factor over its entire range. This experimental design would examine many
combinations of all factors at many levels. This is an inefficient and expen-
sive way to experiment. In many cases a more realistic approach would be to
design & sequence of more modest experiments so that variables can be dropped
or added as information about the system is gained. A two level factorial de-
sign is often used, and this was the approach chosen by the authors.
Rotating Biological Contactor Study
Basic and independent factors which possibly effect RBC effluent quality
and performance are hydraulic flow rate, carbon mass flow rate, ammonia nitrogen
mass flow rate, percent of disc area exposed to the atmosphere at any given time,
and disc revolutions per minute (rpm). These factors were studied in this ex-
perimental program. Influent concentrations were not chosen as variables for
study because they are not basic factors, but are derived factors, equal to
mass flow rates divided by hydraulic flow rates.
If experiments are run with each of the five variables set at only two
levels, there are 2 =*32 possible combinations of factors. From these 32 experi-
ments the effects of changing variables one, two, three, four, and five at a
time can be determined using standard experimental analyses. An excellent ref-
erence is Statistics for Experimenters, by Box, Hunter and Hunter (Wiley, 1978).
It is generally accepted that single factor and two factor interaction ef-
fects are greater than the higher order interaction effects. If this assumption
is made, then sixteen or one-half 2 experiments can be run, using one half of
all the possible combinations of five factors at two levels. From such a. study,
single and two factor interaction effects can be determined. However, the ex-
perimenter is gambling that the higher order effects are not significant. Be-
fore discussing levels of factors, it is necessary to discuss the experimental
RBC because many of the currently accepted values for factor levels are based
on RBC disc diameter and surface area.
Experimental RBC
It was felt that the results of the experimental program would be more
applicable to full scale RBCs the larger the discs on the experimental RBC were.
Three foot diameter discs were chosen after consideration of the logistics of
operating various size systems. A three foot diameter disc has 1/16 the area
of a full scale 12 ft. diameter disc, and 9 times the surface area of a typical
laboratory unit.
The RBC used in this study has four stages, each stage with five discs,
three feet in diameter. The total disc surface area is 284 square feet, or 71
square feet per stage. The discs were made of 1/8 inch thick plexiglass, and
were spaced 1/2 inch apart. The tanks in which the discs rotate are 37 inches
in diameter, and 4.5 inches wide per stage. Constant hydraulic flow rates were
maintained through the use of a constant head tank. Constant feed rates were
achieved through the use of multi-channel perfusion pumps. Flat plexiglass
discs were used because the hydraulics of the flow on the discs would be easier
to model later on, if necessary, and there is a much better definition of disc
surface area. It was also felt that flat discs would give a baseline to which
other types of disc surfaces could be compared.
104
-------
Factor Levels
RBCs are often characterized on the basis of peripheral velocity (length-
time ~1), hydraulic loading rates (volume-area ~^~ - time "I) and organic load-
ing rates (mass-area - time ~^). Previous studies have suggested that per-
ipheral velocity of 60 ft./min. is desirable. The experimental RBC was opera-
ted at peripheral velocities of 56 and 76 feet per minute, corresponding to 5.94
and 8.10 rpm. These rpms were dictated by gear ratios available.
Common hydraulic loading rates are in the order of 1 or 2 gallons per day
per square foot of disc area. The experimental RBC had hydraulic loading rates
of 0.96 and 1.63 gallons per square foot per day. These loading rates were ob-
tained by maintaining flows at 1032 and 1750 liters per day.
Organic loading rates range from 0.25 to 10 pounds of BOD5 per one thousand
square feet per day. The loading rates used in this experiment were 258 and 500
grams of glucose per day, which are theoretically equivalent to 3.6 and 7 pounds
of BOD5 per thousand square feet per day.
Ammonia nitrogen mass loading rates are normally in the 0.1 to 0.2 pounds
NH3~N per thousand square feet per day range. By using NH^-N mass flow rates
of 41 and 65 grams per day, loading rates of 0.3 and 0.5 pounds NI^-N per thou-
sand square feet per day were achieved.
RBCs are usually constructed so that 50 to 60 percent of the disc surface
area is exposed to the atmosphere at any time. The experimental RBC was built
so that 60 or 74 percent of the disc area could be exposed by varying the water
level in the tanks.
These five factors were varied systematically. The factor levels, condi-
tion numbers with corresponding coded factor levels and the order of experimen-
tation are shown in Table 1.
EXPERIMENTAL PROCEDURES
RBC discs were rotated by a constant speed gear motor and a chain drive.
RPMs were changed by changing sprockets. Hydraulic loading rates were varied
by a valve in the water line from an aerated constant head tank. Tap water,
which originated from shallow wells tapping an aquifer under a river, and re-
ceiving only chlorination was used as the feed water. Both the constant head
tank and the water bath surrounding the RBC were heated to maintain a constant
20°C temperature. Percent of disc exposed was controlled by an adjustable
overflow weir on the effluent side of the last stage. Carbon and nitrogen mass
loading rates were controlled by varying concentrations of feed solutions which
were pumped to the first stage by multi-channel perfusion pumps. The feed com-
ponents were not all mixed together to prevent precipitation and to reduce the
possibility of bacterial degradation. Table 2 shows feed solution compositions
for both high and low level factor loading rates.
In addition to the carbon and nitrogen feed solutions, sodium bicarbonate
and sodium hydroxide were added to the second stage to minimize pH and Alkalin-
ity on nitrification. The pH of the second stage was maintained at pH 8.5 and
effluent alkalinity was kept above 100 mg/1 as CaC03-
105
-------
TABLE 1
ONE-HALF 25 EXPERIMENTAL DESIGN
Factor
Number
1
2
3
4
Level
Factor
Revolutions per minute (RPM)
Hydraulic Flow Rate (liters/day)
Disc Surface Area Exposed (percent)
Nitrogen Mass Loading Rate
(grams NH--N per day)
Carbon Mass Loading Rate
(grains glucose per day)
5.9
1032
60
41.5
258
8.1
1750
74
65.0
500
EXPERIMENTAL
CONDITION
NUMBER
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
ORDER
RUN
7
1
12
8
13
2
5
16
4
6
11
10
15
3
9
14
FACTOR LEVELS
Factor 12345
•f
106
-------
TABLE 2
FEED COMPOSITIONS
Carbon Feed Solutions
Solution 1 C,HnoO,'H_0
D L2. D 2
Solution 2 KH_PO.
Solution 3 FeCl3'6H20
Solution A
MgS04'7H20
Nitrogen Feed Solution
Solution A (NH,)2SO,
Concentration (grams/Liter)
(+) Level (-) Level
283.81
119.325
2A2.520
0.1935
2.064
3.339
33.387A
194.112
146.47
61.577
125.15
0.0999
1.065
1.66A
16.6A2
100.162
107
-------
The RBC was started up by pumping 18 liters of trickling filter effluent
into the first stage over a 24 hour period. Constant carbon, ammonia and hy-
draulic flow rates were established and maintained. Effluent quality was mon-
itored until steady state conditions were achieved.
Experimental runs all proceeded in the same manner. Factor levels were
set as indicated by the experimental design. Because changes in nitrification
will cause changes in pH and alkalinity, fourth stage pH and alkalinity, as
well as nitrates, were monitored. When these values stabilized, it was assumed
that the system was at steady state. During the initial phase of the experimen-
tal program a week was allowed to go by before this monitoring began. After
three days of sampling we consistently found that the system had already reached
steady state. During the later portion of the experimental program monitoring
began immediately. We found that system steady state conditions were achieved
in about four days.
Once steady state conditions had been reached the three to five day sampling
program began. Samples were taken from the influent and first, second, third,
and fourth stages. All samples were analyzed for filtrate COD (0.45 micron
membrane filter), nitrates (electrode method), ammonia (electrode and distilla-
tion methods), total Kj eLdahl nitrogen, pH, alkalinity, and total volatile sol-
ids. Influent and stage one, two, three, and four temperatures and dissolved
oxygen concentrations were also determined. All analyses were done according
to Standard Methods (14th Edition).
RESULTS
Experimental results which pertain to this report are presented in Table 3.
A one—half 2~* experimental design allows the experimenter to determine the effect
of varying five different factors either one or two at a time. The effect repor-
ted is the average result of raising the factor or factors from the low (-) level
to the high (+) level used in the experimental program while all remaining fac-
tors were held constant. In addition to analyzing the concentration data presen-
ted in Table 3, selected mass flow and mass removal (or generation) data were
developed by multiplying concentrations or differences in concentrations by hy-
draulic flow rate. These generated data were also analyzed. All analyses are
presented in Tables 4 through 7. "Factor" refers to the operational variable
which, when raised from its (-) level to its (+) level caused, on the average,
the effect listed in the table. "12", "13", etc., refers to raising 1 and 2
together or 1 and 3 together.
Caveats
The effects of changing the level of a factor, either by itself, or in
combination with another factor, on a response whose value is dependent on the
level of that factor should not be considered. For instance, the effects of
changing flow rate on concentration should not be considered; since concentra-
tion is a function of the flow rate. Or, the effect of increasing carbon mass
flow rate on COD mass removal or concentration should not be discussed.
If data is to be manipulated, the manipulations have to be done before the
factorial analysis is performed. For example, efficiencies of removal must be
calculated for each of the sixteen experimental runs and then analyzed to deter-
mine the effects of changing factors on efficiencies.
108
-------
TABLE 3
EXPERIMENTAL RESULTS
COD (mg/1)
D.O. (mg/1)
CONDITION
NUMBER
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Influent
307.0
266.4
171.5
345.3
278.5
524.6
307.0
102.35
256.9
489.2
327.2
158.7
511.2
263.3
148.1
326.1
Stage 1
104.2
24.3
32.4
72.0
34.0
47.4
75.6
33.5
55.2
51.7
59.8
46.9
60.6
29.8
39.2
39.5
Stage 4
35.8
12.1
19.2
31.1
23.5
14.5
47.4
5.1
39.7
37.1
27.3
37.3
28.7
19.4
31.7
17.8
Influent
27.4
38.5
44.1
22.1
35.8
37.9
24.9
22.7
83.6
79.9
44.0
31.0
69.2
74.3
42.7
34.4
Stage 1
22.5
23.1
30.8
11.0
26.5
20.4
11.6
20.4
67.6
61.3
31.9
16.9
49.9
59.6
32.8
33.9
Stage 4
0.88
0.57
0.00
0.14
0.00
0.43
0.72
0.00
15.80
29.30
7.69
1.73
14.12
12.10
2.76
3.50
Influent
0.92
1.12
0.78
0.93
0.84
0.97
0.92
1.04
1.09
0.80
0.85
1.19
1.23
0.82
1.10
1.25
Stage 2
2.82
5.02
8.05
0.93
10.90
1.89
2.82
2.12
12.70
5.27
4.62
13.93
1.92
8.09
1.83
9.49
Stage 4
11.57
27.30
21.10
16.73
35.90
18.98
11.56
25.23
61.83
27.50
26.70
41.17
44.23
51.00
39.47
32.50
Stage 1
0.20
5.34
6.28
1.90
4.98
4.04
3.10
6.83
4.38
2.03
4.38
7.55
1.37
5.20
6.28
4.78
Stage 4
3.63
4.86
8.54
9.42
5.91
5.75
7.59
9.86
2.55
4.28
6.34
5.92
3.51
3.35
5.31
7.44
-------
TABLE 4
FACTOR EFFECTS ON CHEMICAL OXYGEN DEMAND
EFFECT
FACTOR
AVERAGE
1
2
3
4
5
12
13
14
15
23
24
25
34
35
45
COD CONCENTRATION (mg/1)
Stage 1
50.0
-14.1
-1.0
-10.9
-5.1
27.3
10.4
-0.7
3.2
-7.9
5.1
-1.9
-3.6
-0.3
-5.7
-16.4
FACTOR
NUMBER
1
2
3
4
5
Stage 4
27.4
-8.6
-0.5
-5.2
7.6
5.4
0.3
-7.5
7.2
-1.1
1.9
-4.8
2.4
-3.2
-0.3
-11.8
FACTOR
RPM
FLOW RATE
COD MASS REMOVAL RATE
Stage 1
319
38
-6
25
33
221
-36
-30
-20
60
-53
-11
55
-1
43
40
(grams/day)
Overall
351
39
7
24
17
247
-29
-21
-21
47
-61
-12
56
6
29
34
EXPOSED
NITROGEN MASS LOADING RATE
CARBON MASS
LOADING RATE
110
-------
TABLE 5
FACTOR EFFECTS ON AMMONIA NITROGEN
EFFECT
NH«-N Concentration
J (mg/1 NH3-N)
NH«-N Mass Removal Rate
(grams/day)
FACTOR
Stage 1
Stage 4
Stage 1
Overall
AVERAGE
32.51
5.61
14.1
51
1
2
3
4
5
12
13
14
15
23
24
25
34
35
45
-3.36
-17.71
-1.26
23.43
-4.40
-2.83
6.75
0.75
6.05
3.29
-13.03
1.28
0.87
-1.49
4.40
0.73
-7.08
-2.81
10.53
2.98
-2.18
-1.12
0.84
1.77
2.17
-6.83
1.09
-2.70
-2.00
2.58
4.6
3.5
-6.3
4.1
-2.1
-4.1
-7
8
0.7
2.8
-3.6
-5.8
1.5
-0.1
1.6
-1.0
9
3
—2
16
8
—7
7
0
5
-2
7
-1
6
3
FACTOR
NUMBER
1
2
3
4
5
FACTOR
RPM
FLOW RATE
SURFACE AREA EXPOSED
NITROGEN MASS LOADING RATE
CARBON MASS LOADING RATE
111
-------
TABLE 6
FACTOR EFFECTS ON NITRIFICATION
EFFECT
NQ.V-N Concentration
(tng/1)
NO_-N Production Rate
(grams/day)
FACTOR
Stage 2
Stage 4
Stage 2
Overall
AVERAGE
5.78
30.86
6.02
39.3
1
2
3
4
5
0.14
0.60
-1.79
2.88
-4.11
-1.37
-7.86
3.00
19.63
14.28
0.68
3.95
1.55
3.76
-2.53
0.1
8.9
3.9
25.6
17.4
12
13
14
15
2.15
0.90
3.79
1.22
5.82
0.51
-3.40
1.78
3.09
0.10
6.59
1.21
7.4
0.4
3.4
2.6
23
24
25
-1.
1.
2.
02
08
09
-2.48
,07
,16
-2.32
3.22
0.31
-1.4
1.9
1.9
34
35
2.01
2.41
-0.75
3.20
2.18
3.29
0.1
2.9
45
0.30
-1.61
1.33
1.9
FACTOR
NUMBER
FACTOR
1
2
3
4
5
RPM
FLOW RATE
SURFACE AREA EXPOSED
NITROGEN MASS LOADING RATE
CARBON MASS LOADING RATE
112
-------
TABLE 7
FACTOR EFFECTS ON STAGE 1 DISSOLVED OXYGEN
FACTOR
EFFECT (mg/1 D.O.)
AVERAGE
1
2
3
4
5
12
13
14
15
23
24
25 -
34
35
45
1.29
0.84
1.70
0.57
0.41
-3.13
-0.58
0.44
-0.05
0.09
-0.35
0.81
-0.63
-0.74
0.63
0.42
FACTOR
NUMBER
1
2
3
4
5
FACTOR
RPM
FLOW RATE
SURFACE AREA EXPOSED
NITROGEN MASS LOADING RATE
CARBON MASS LOADING RATE
113
-------
The effects presented in a one-half 2 factorial analysis are the effects
of changing the factors listed aliased with the effects of changing all the re-
maining factors not listed. For example, what appears to be the effect of chan-
ging factors 2 and 3 together could be the effect of changing 2 and 3, or the
effect of changing 1, 4 and 5 together, or a combination of both effects. Re-
member, the experimenter is gambling that the single and two factor effects are
greater than the three, four, and five factor interaction effects.
These experiments were not replicated, so there is no direct estimate of
the variance. This means that there is no immediate way to determine which
effects are statistically significant and which effects are merely due to ran-
dom variations of system performance. This problem can be partially solved
through the use of normal probability paper.
If the calculated effects were only due to random variations about a mean
value, the effects would have been distributed approximately normal with mean
zero. The mean would be zero because the effects are calculated from differ-
ences between pairs of responses, and, on the average, the difference between
random numbers would be zero. Thus, the ordered effects, if not significant,
would plot as a straight line on normal probability paper.
If the ordered effects do not plot as a straight line on normal probability
paper, they are not distributed normal. If the removal of some effects allows
the rest of the effects to plot as a straight line, the removed effects are
not easily explained as chance occurrences, while the remaining effects are.
The effects presented in Tables 4 through 6 were plotted on normal probability
paper and analyzed. The results of these analyses are presented in Table 8,
and are also discussed in the following section along with the results shown
in Tables 4 through 7.
ANALYSES
When considering factor effects it is necessary to overlook the obvious
effects and discuss that which may appear to be less important. One could
guess that if flow rate was increased while mass loading rates remained the
same, concentrations would decrease. Or, it could be anticipated that concen-
trations would inoease if mass loading rates increase , flow being held con-
stant. Similarly, mass removal rates would probably increase if mass loading
rates were increased. Once the effects are calculated, it is very difficult,
if not impossible, to rationalize apparent contradictions by multiplying effects
of one factor by effects of another or subtracting interaction effects of sev-
eral factors from effects of another. So, when considering the information in
Tables 4 through 7 and in Table 8, a lot of the obvious will be ignored, In
addition, even though analyses summarized in Table 8 indicated that only cer-
tain effects were significant, other effects revealed by analyses summarized
in Tables 4 through 7 will be discussed. The authors think that the analyses
summarized in Table 8, while helpful, are not absolute, and should be "taken
with a grain of salt."
Factor Effects on Chemical Oxygen Demand
If the results of analyses summarized in Table 8 are believed, and the
obvious effects ignored, there are no significant effects.
114
-------
TABLE 8
RESULTS OF PROBABILITY ANALYSES
RESPONSE
Stage 1 COD Concentration
Stage 4 COD Concentration
Stage 1 Removal Rate
Overall COD Removal Rate
SIGNIFICANT FACTOR OR
FACTORINTERACTION
EFFECTS
INCREASE
5
none
5
5
'DECREASE
none
none
none
none
Stage 1 NHL-N Concentration
Stage 4 NHv-N Concentration
Stage 1 NH--N Removal Rate
Overall NBL-N Removal Rate
Stage 2 N03-N Concentration
Stage 4 NQ«~N Concentration
Stage 2 N03~N Production Rate
Overall N03~N Production Rate
4
4
none
4, 5, 2, 12
none
4, 5
14
4
24, 2
24, 2
none
none
23, 2, 5
none
23, 5
none
FACTOR
NUMBER
1
2
3
4
5
FACTOR
RPM
FLOW RATE
SURFACE AREA EXPOSED
NITROGEN MAS LOADING RATE
CARBON MASS LOADING RATE
115
-------
One can choose to ignore Table 8 and the analyses for significance, and
examine Table 4. It will be found that the factors which effect COD concen-
tration the most, either decreasing or increasing it, are all trivial factors;
i.e., flow rate, or flow rate in combination with an other factor, or carbon
mass loading, singly or in combination. The same can be said for COD mass
removal rates.
Factor Effects on Ammonia Nitrogen
Considering the results of the probability analyses, and ignoring the ob-
vioup effects, there are three significant effects. Overall ammonia removal
rates were increased when carbon mass loading rate (5), flow rate (2), RPM
and flow rate (12) were increased. The following non-trivial major effects
were observed, but may or may not be significant, depending upon one's belief
in the validity of the probability analyses.
Ammonia Concentration.
An increase in RPM and carbon mass loading rate (15) caused an increase
in ammonia concentration. However, an increase in carbon mass loading (5)
alone caused a decrease in ammonia concentration. It is easy to eliminate
the contradiction by saying that neither effect is statistically significant.
Ammonia Concentration.
An increase in carbon mass loading rate (5) increased stage 4 ammonia con-
centration the most. An increase in surface area exposed (3) caused the great-
est decrease.
Overall Ammonia Removal.
Increasing RPM (1) decreased removal the most, Increasing carbon mass
loading rate (5) also increased removal. Increasing RRM and flow rate together
(12) caused the greatest decrease in removal.
Factor Effects on Nitrification
There were four statistically significant, yet non-trivial effects. In-
creasing carbon mass loading rate (5) caused three of them. Stage 2 N03~N con-
centration decreased, stage 2 NO,, production rate decreased, and stage 4 NO~-N
concentration increased. If one examines Table 6, ignoring Table 8, one can
find no major non-trivial, but perhaps statistically non-significant effects.
CONCLUSIONS
The results of the analyses done to date are ambiguous. There were extreme
variations in the performance of the RBC during the experimental program. One
only needs to scan Table 3 to confirm that. Something caused those variations,
and these experimenters feel it was variations in the factors being studied.
Further analyses of the data are anticipated, and hopefully these analyses
will give more insight into the system.
For the time being, some information has come out of the analyses which
adds to the knowledge of the RBC system.
116
-------
Consider first only information in Table 8.
RPM was a statistically significant factor, in combination with a non-
trivial factors, once.
Flow rate was a statistically significant factor, in combination with
other non-trivial factors, twice,
Surface Area Exposed was a statistically significant factor, in combination
with a non-trivial factor, once.
Nitrogen Mass Loading Rate was never a non-trivial, statistically signifi-
cant factor.
Carbon Mass Loading Rate was a non-trivial, statistically significant
factor, by itself, four times.
If one includes the more subjective analyses presented in the "Analyses"
section, Table 9 can be developed.
TABLE 9
FREQUENCY MENTIONED AS
IMPORTANT FACTOR
FACTOR
RPM
FLOW RATE
SURFACE AREA EXPOSED
NITROGEN MASS LOADING RATE
CARBON MASS LOADING RATE
FREQUENCY MENTIONED
ALONE
1
1
1
0
7
IN COMBINATION
3
2
0
0
1
One must remember that the factors were identified as important or significant
more times than indicated above. However, they were termed "trivial" because
the factor was part of the response being mentioned. The data are still being
analyzed, and attempts are being made to "normalize" the derived data so these
factors become "non-trivial".
117
-------
Page Intentionally Blank
-------
HYDRAULIC AND ORGANIC FORCING OF A PILOT SCALE RBC UNIT
By
Leonard W. Orwin
Project Manager
Charles D. Siebenthal
Associate
Metcalf & Eddy, Inc.
Palo Alto, California
The San Francisco Master Plan for water pollution control recommends
that a new Southwest water pollution control plant (WPCP) be
constructed. The Southwest WPCP would include secondary treatment for
about 22 million gallons per day (Mgal/d) of dry-weather sewage to
effluent quality levels required for ocean discharge. Since San
Francisco has combined sewers, wet-weather sewage would also be treated
in the secondary plant. A separate wet-weather plant would be
constructed to handle wet-weather flows that exceed the hydraulic
capacity of the secondary plant. Wet-weather flows as high as 500
Mgal/d could be possible depending on the design of the
storage/transport system.
In support of the Southwest facilities planning effort, Metcalf & Eddy
conducted a pilot plant program to study the serious technical problems
associated with treating both wet- and dry—weather flows. During wet
weather, the composition of the feed to the secondary plant would
continuously change over an extremely wide range, and the effect of
these changes on the quality of the secondary effluent was unknown.
Thus, pilot testing of the secondary processes was necessary to
determine (1) what percentage of the wet-weather treatment capacity
119
-------
could be achieved by hydraulically forcing the secondary plant, and
(2) what design modifications would be necessary to achieve this maximum
hydraulic capacity. Another objective of the pilot testing was to
determine the stability of the secondary processes under extreme diurnal
and wet-weather treatment load transitions.
This paper presents the results of the pilot plant tests designed to
study the response of a rotating biological contactor (RBC) to rapid
changes in hydraulic and organic loadings. These results should be
valuable to others considering this process for locations experiencing
similar wide variations in wastewater composition (for example,
communities with combined sewers or unusually high infiltration/inflow
rates).
CHARACTERISTICS OF THE TREATMENT PROBLEM
The proposed dry-weather treatment plant will be required to handle wide
variations in both flow and pollutant loadings. The typical diurnal
flow pattern includes a low of about 8 Mgal/d during the early morning
hours and peak flows of about 35 to 40 Mgal/d during the late morning
and early afternoon. The BOD concentrations vary in a similar fashion
with concentrations as low as 50 milligrams per litre (mg/L) during the
periods of low flow and peak concentrations as high as 250 mg/L
coinciding with the peak flows. Typically, the time period between the
early morning low and the later morning peak is 3 to 4 hours and,
consequently, the BOD loading rate can increase by as much as 2,500%
during this short time span.
During wet weather, the short-term variations in loading conditions
could vary even more drastically depending on the peak hydraulic design
capacity selected for the dry-weather plant. During a storm, the BOD
and TSS concentrations may vary from as low as 15% of the dry-weather
average to as high as 200 or 300%. Furthermore, the alkalinity and
conductivity of the wastewater will vary more or less in direct
proportion to the amount of stormwater dilution. These variations in
influent characteristics may significantly affect the performance and
stability of biological secondary processes and, consequently, must be
carefully considered in the selection of the recommended process.
Since the RBC system is a fixed culture process, several advantages were
anticipated over the competitive suspended biomass processes (air and
pure oxygen activated sludge) for the proposed Southwest WPCP.
• • Primary clarifiers could be operated at higher overflow rates
during wet weather without seriously affecting the RBC process
since the inert solids carried over in the primary effluent
would tend to pass through the RBC units with little effect on
BOD removal efficiency. In a suspended biomass system, these
inert solids accumulate in the aeration basins and reduce the
fraction of the total biomass, which is biologically active.
120
-------
• Since the RBC biomass is attached to the rotating disks, high
hydraulic loadings will not cause biomass "washout." Large
increases in hydraulic loadings to a suspended biomass system
will result in a temporary transfer (washout) of biomass from
the aeration basins to the secondary clarifiers leading to
lower BOD removal efficiency.
* Significant variations in loading conditions to an activated
sludge process require careful operator attention to make
proper adjustments to air or oxygen supplies, sludge recycle
rates, and secondary clarifier operating conditions in order
to maintain a viable biomass. Very little operator attention
would be required during hydraulic forcing of an RBC system
since the biomass is retained on the disks.
Pilot scale tests were performed to demonstrate the extent to which an
RBC unit can be hydraulically forced relative to normal design loadings
and to determine the response of an RBC system to rapid increases in BOD
loadings over short-time periods. Additional tests were conducted to
evaluate the effects of high feed solids from an .overloaded primary
clarifier; however, those tests will be described in a subsequent paper.
TEST FACILITY DESCRIPTION
The pilot scale RBC unit was a part of an extensive pilot plant facility
designed to test a number of physical/chemical processes for wet—weather
treatment in addition to the tests on biological secondary treatment
processes. Photographs of the pilot plant facilities, located primarily
In the courtyard of the Richmond—Sunset WPCP in Golden Gate Park, are
presented in Figure 1. These photographs indicate the basic scale of
the pilot units: e.g., the RBC unit was equipped with 2.0 metre
diameter disks and the UNOX system was in the standard 8 ft by 40 ft
trailer. All biological treatment units were elevated to provide
gravity transfer to the secondary clarifiers.
The RBC pilot unit was an Autotrol 2.0 metre unit with four disks
providing about 7,900 square feet (ft ) of surface area. The nominal
hydraulic capacity of the unit was 6 gallons per minute (gal/min) based
on a hydraulic loading of 1.1 gallons per square foot per day
(gal/ft *d) . The pilot unit was operated at a rotational speed of about
3 revolutions per minute (rpm) during all tests. Feed to the RBC unit
was typically primary effluent from a pilot—scale clarifier operated at
a surface overflow rate of about 1,000 gal/ft 'd. Effluent from the RBC
unit flowed by gravity to a 5 foot diameter secondary clarifier. During
the hydraulic forcing tests, a large portion of the RBC effluent was
bypassed around the secondary clarifier to avoid excessive overflow
rates.
A schematic process flow diagram of the RBC pilot facility is presented
in Figure 2. Composite samples, taken at 8 hour intervals, were
collected by the use of small metering pumps or timer—controlled
solenoid valves at the flow points indicated in the figure.
121
-------
NOTE
(a) RBC UNIT
(b) CHEMICAL MIX AND LAMELLA SETTLER
(c) HPO ACTIVATED SLUDGE (IN TRAILER)
(d) GENERAL VIEW
(e) AERIAL OF RICHMOND SUNSET WPCP SHOWING
PILOT PLANT IN COURTYARD
Figure 1. Southwest WPCP pilot plant.
122
-------
DEQRITTED
RAW SEWAGE
SLUDGE OVERFLOW
ABBREVJ.AT IONS
HCV = HAND CONTROL VALVE
M = MOTOR
SP = SAMPLING POINT
V - VESSEL
EXCESS TO
DRAIN
^EFFLUENT
>• SLUDGE
Figure 2. Process flow diagram, RBC unit.
-------
HYDRAULIC FORCING TESTS
The experimental plan for the hydraulic forcing tests Involved operating
the RBC unit at progressively higher feedrates by applying the step
increases shown in Figure 3. The unit received diurnal variations in
wastewater composition but was operated at constant flow conditions
after each step increase. Samples were collected for analysis according
to the schedule shown in Table 1. The sampling points are shown
schematically on the process flow diagram. Figure 2.
70 r
BO
SO
40
»- 30
20
10
STEADY-STATE DESIGN OPERATING CONDITIONS
RBC (V-105)
FEED RATE 6 gal/Bin .
HYDRAULIC LOADING RATE t.l gii/ft -d
RBC SECONDARY CLARIFIER (V-109)
OVERFLOW RATE
700 fal/ft -d
0 I 2 3 4 5 6 7 8 8 101112131415
DAY OF TEST RUN
Figure 3. RBC flow increases, hydraulic forcing tests.
124
-------
Table 1. SAMPLING PROGRAM, RBC HYDRAULIC
FORGING TESTS
Effluent primary Effluent secondary
clarifies: V-102 RBC unit clarifier V-109
Sampling point No.
Sample type
pH
Temperature
Dissolved oxygen
Conductivity
Total suspended solids
Volatile suspended solids
Alkalinity
104
Composite
3
3
3
3
3
3
1
108
Grab
3
3
3
—
3
3
—
112
Composite
3
3
3
3
3
3
1
Settleable solids — 3
Total organic carbon (TOC) 3 — 3
Soluble TOC 3 12a 3
Biochemical oxygen demand (BOD) 3 — 3
Soluble BOD 3 — 3
Chemical oxygen demand (COD) 3 — 3
Soluble COD 3 3
Ammonia nitrogen lb — it
Freon extractable material I*3 — lfc
a. One sample from each stage per shift.
b. 24 hour composite.
Results of the hydraulic forcing tests are summarized in Table 2 along
with comparable data collected during another test run on this unit at
essentially nominal design conditions. The RBC unit was operated at
hydraulic loadings greater than 1,000% of design and organic loadings of
up to 370% of design (baaed on a nominal design loading of 1.4 pounds of
soluble BOD per 1,000 ft /d). The last line of Table 2 indicates
whether the feed was all dry-weather flow or partially wet-weather flow.
Due to the rainfall that occurred during many of the test periods, the
influent BOD concentrations varied widely from day to day. Thus, the
average influent and effluent concentrations listed in Table 2 and the
percent removals calculated from them must be interpreted only as trends
with increasing hydraulic load.
125
-------
Table 2. RBC HYDRAULIC FORCING TEST RESULTS
97l
Hydraulic loading rate, % of design rate3
225 275 325 375 470 475 590 775
1,040
Test period
Duration, hours
Feedrate, gal/min
Hydraulic loading,
gal/ft2 -d
Hydraulic 'detention, min.
Organic loading ,
Ib BODsol/1 , 000 f t^ * d
Organic removal,
Ib BODsoJ/lrOOQ ft2 -d
Influent BOD, mg/L
Total
Soluble
Effluent BOD, mg/L
Total
Soluble
Effluent total suspended
solids, ng/L
BOD removal, %
Total
Soluble
COD removal, %
Total
Soluble
TOG removal, %
Total
Soluble
Feed type
__
216
5.8
1.06
172
0.79
0.75
160
90
24
4.2
21
85.2
95.3
79.9
70.2
74.2
52.9
Dry
D
96
13.0
2.38
77
1.48
1.24
133
75
29
12
22
78.2
84.0
76.9
71.5
64.5
56.5
Wet
A
104
15.
8
2.88
63
1.25
1.13
102
52
23
5
12
77.
90.
67.
47.
53.
39.
Wet
5
4
2
4
5
4
E
40
19.0
3.45
53
1.85
1.21
123
64
43
22
58
65.0
65.6
39.6
36.3
41.0
44.4
Dry
F
80
21.7
3.96
46
2.31
1.88
108
70
44
13
30
59.2
81.4
61.4
46.6
59.3
49.1
Dry
B
72
27.2
4.96
37
3.30
2.72
145
80
46
14
34
70.0
82.5
62.3
63.5
54.7
38.0
Dry
G
88
27.6
5.02
36
3.48
2.60
140
83
49
22
33
65.0
74.7
57.5
56.5
57.6
49.2
Dry
C
40
34.2
6.23
29
1.82
1.09
78
35
52
14
32
55.1
60
52.9
59.9
30.7
20.7
Wet
H
168
45.0
8.20
22
5.19
3.69
115
76
55
22
46
52.2
71.1
44.4
32.2
—
—
Dry
I
256
60.5
11.03
16.5
4.87
2.48
101
53
55
26
43
45.5
50.9
40.9
36.0
—
—
Wet
a. Based on nominal design rating of 6 gal/min.
b. This column represents the performance at nominal design loading.
Scanning the data In Table 2 from left to right, it is clear that in
general as hydraulic load increases, the effluent soluble BOD tends to
rise from 4 mg/L up to 26 mg/L. Likewise, the soluble BOD removals tend
to decrease from 95% down to 51%. However, to compensate for stormwater
dilution of the feed, the data are best interpreted as a function of
applied organic loading.
The organic removal achieved as a function of applied organic loading,
both measured in units of pounds of soluble BOD per day per 1,000 ft of
disk area (Ib BOD/1,000 ft -d), is shown in Figure 4. In Figure 4, the
open circles represent dry-weather tests; the dark circles represent
tests with some rain. Note that as the loading increases up to about 3
lb/1,000 ft "d, the removal increases linearly with a slope of 0.82,
indicating a removal of about 82% on a mass basis.
126
-------
5 r
LEGEND
• WET TESTS
© DRY TESTS
1234
SOLUBLE ORGANIC LOADING.Ib BOD/1,000 ft2- d
Figure 4. RBC performance, organic removal,
The data point at a loading of 5.19 lb/1,000 ft 'd represents operation
at 45 gal/min under wet-wnather conditions. The organic removal was
only 51%, a reduction due to a combination of short hydraulic detention
time (16.5 minutes) and low influent soluble BOD concentrations (only 53
mg/L due to stormwater dilution, versus 70 to 90 mg/L for dry weather).
Another measure of performance is the effluent soluble BOD
concentration. The effluent soluble BOD concentration as a function of
organic loading is shown in Figure 5. Again, effluent BOD tends to
increase more or less predictably as the organic loading increases.
Only one data point is more than 5 mg/L from the correlating line, which
is an excellent margin of error in BOD determinations at low values
between 0 and 25 mg/L.
Similar conclusions hold for chemical oxygen demand (GOD) removals,
although the data are not presented in this paper.
127
-------
30 t~
20
10
LEGEND
• IET TESTS
G DRY TESTS
0
O
O
G
SOLUBLE ORGANIC LOADING, Ib BOD/1,000 ft d
Figure 5. RBC performance, effluent soluble BOD.
The total organic carbon (TOG) reductions tend to decrease more rapidly
than either BOD or COD because a portion of the TOG represents
nonbiodegradable organics; thus, the maximum possible TOG reduction via
biodegradation is less than 100%. Also, the effluent soluble TOG tends
to increase more quickly as the organic load increases, indicating that
some refractory metabolic byproducts may be produced under highly
loaded, short residence time operations.
The overall performance data from Table 2 are superimposed on the
vendor's RBC design curves in Figure 6. Only the data for test periods
at feedrates up to 45 gal/min are shown. The hydraulic loading for the
final test period at 60 gal/min is off scale.
Each operating point is represented in Figure 6 as two black dots
connected by an arrow. In each case, the arrow runs from the measured
performance to the theoretical performance. Thus, arrows pointing
upward indicate that at the measured hydraulic and organic loading, the
test resulted in a lower effluent BOD concentration than predicted by
the design curves.
Note that in only one test period (E) does the arrow point downward,
indicating that the actual effluent BOD was higher than that predicted
by the design curves. Test E was a short test period that was
terminated by very heavy rains. Thus, the data from Test E can probably
be ignored in this analysis.
128
-------
40
30
_ 20
10
SO
40
30
25
20
10
INFLUENT SOLUBLE BOO, mg/L
ISO 120 100 80 70 60 50 45 40
•ASTEiATER TEHPER*TURi>58'f
0L
HYDRAULIC LOADING RATE,(al/ft*•rt
Figure 6. 1BC design curves.
To consider Test B, it was necessary to arbitrarily extrapolate the
design curves to much higher organic loads. The result is a predicted
effluent BOD of 42 mg/L versus an actual value of 12 mg/L. The design
curves could not be extrapolated in this manner to accommodate Tests H
and I. In both cases, the predicted effluent concentrations would have
been greater than the influent concentrations.
In general, it appears that the design curves are conservative and that
the RBC can be much more heavily loaded for short periods of time than
would be expected based solely on the design curves.
SIMULTANEOUS HYDRAULIC AND ORGANIC FORCING TESTS
A second series of tests was performed to evaluate the effects of more
extreme forcing conditions over short periods of time. In these tests,
the step increases in feedrate were made during the morning hours when
the Influent soluble TOC value changes from a low value of about 30 mg/L
at 6:00-7:00 a.m. to a peak value of about 60 mg/L at 10:00-11:00 a.m.
The tests were performed on 4 consecutive days with step increases in
the RBC unit feedrate planned to follow the schedule shown in Table 3.
129
-------
Table 3. PLANNED INCREASES IN
RBC UNIT FEEDRATE
Time
5:00
6:00
7:00
7:30
8:00
8:30
9:00
9:30
10:00
11:00
Noon
1:00
RBC
Day 1
10
10
10
20
20
20
20
20
20
20
20
10
feedrate
Day 2
10
10
10
20
30
30
30
30
30
30
30
10
, gal/min
Day 3 Day 4
10
10
10
20
30
40
40
40
40
40
40
10
10
10
10
20
30
40
50
50
50
50
50
10
Grab samples of primary effluent (RBC feed), RBC 4th stage, and
elarifier effluent were collected according to the following schedule:
Primary effluent
RBC 4th stage
RBC elarifier effluent
Sample frequency
5:00-7:00 7:00-10:00 10:00-1:00
30 minutes 15 minutes 30 minutes
30 minutes 15 minutes 30 minutes
30 minutes 30 minutes 30 minutes
The RBC unit feedrates during the 4 days of testing are shown in Figure
7. The flowrate was adjusted according to the planned 'schedule except
on April 18 when the operator had difficulty increasing the RBC feedrate
to the desired rate of 20 gal/min because a bypass valve had been left
open- Consequently, the RBC feedrate did not reach 20 gal/min until
about 9:00 a.m. rather than at 7:30 a.m. as planned. The corresponding
increases in soluble organic loading are shown in Figure 8. The
response of the RBC unit in terms of soluble TOC in the feed, 4th stage,
and secondary elarifier effluent for each day is shown graphically in
Figure 9.
On April 18, 1978, the peak soluble organic loading was 2.1 Ib TOC/1,000
ft -d at about 10:30 a.m. The peak soluble TOC concentration in the RBC
4th stage was about 30 mg/L, and the elarifier effluent peak soluble TOC
value was also 30 mg/L if the questionable data point at 10:30 a.m. is
130
-------
60
50
40
30
20
10
4 5 6 7 8 9 10 11
APRIL 18,1978
12 456789 101!g
O
APRIL 19,1978 *
oU
10
n
<
1r*
/
)f)f
„
;
!
>
i
r
1
^
f
>ff2
ii
-Yi
r
J
r
j
%
»^
4 5 8 7 8 9 10 11
APRIL 20,1978
12 456789 1011*12
APRIL 21,1978
Figure 7. RBC feedrates.
131
-------
o
3.5
3.0
o cs
"C *t.f
3 - 2.5
o o
z ° 2.0
«c
£ ~ 1.5
T
iasai.
je«F
456789 10 11 g12 4 5 6 7 8 9 10 11* 1 2
o g
APRIL 18, 1978 APRIL 19, 1978
4.5
4.0
3.5
3.0
4567891011*12 4567111011*12
o o
Z X
APRIL 20, 1978 APRIL 21, 1978
Figure 8. RBC organic loading rates.
132
-------
4 5 6 7 • S 10 lie
e
a
APRIL 20. 1978
1 2
4 S 8 7 8 9 10 11 <= 1 2
CD
z
APRIL 21, 1978
LEiEMP
0 8IC UM1T FEED
e 4th SIA8E
• SEGWDA1V CLAMIFIEM EFFLUENT
Figure 9. RBC soluble TOG response,
133
-------
ignored. On the basis of hydraulic residence time alone, the response
of the 4th stage should lag an increase in feed concentration by about
45 minutes during this test. Since the clarifier flowrate was
maintained constant at about 15 gal/min for all tests, the response in
the clarifier effluent should lag the 4th stage response by about 2.5
hours in each test.
On April 19, 1978, the deterioration in clarifier effluent quality was
not significantly greater than in the first test even though the peak
organic loading of about 2.8 lb/1,000 ft -d was about 35% greater. The
4th stage soluble TOG did reach a peak of 34 mg/L, which was about 13%
greater than the peak value in the first test.
On the third test day, April 20, 1978, the peak organic loading was
about 3.6 Ib TOC/1,000 ft "d, which was about 70% greater than in the
first test. The clarifier effluent reached a peak value of 40 mg/L
soluble TOG, which was about 33% higher than in the first test. The
rapid increase in 4th stage TOG, which occurred between 8:30 and 9:00
a.m., was the initial response to the sudden increase in loading.
Although a peak value of about 44 mg/L was apparently reached at 9:00
a.m., the subsequent TOG values leveled off at about 34 mg/L.
On the final day of the test, April 21, 1978, the RBC feedrate was
increased to 50 gal/min, resulting in a peak organic loading of about
4.4 Ib TOC/1,000 ft -d. This represents an increase of about 210% over
that of the first test. The 4th stage TOG values during this test show
a steady increase in response to the increased hydraulic and organic
loading, peaking at about 44 mg/L. This is an increase of about 50%
over the corresponding peak value in the first test- The clarifier
effluent TOG showed a similar response.
Although the effluent TOG quality deteriorated as the organic loading
increased during the series of tests, it is interesting to view the data
in terms of the organic removal achieved as a function of applied
organic loading. These data, along with corresponding data collected
during the hydraulic forcing tests, are presented in Figure 10. It
should be noted that the data points presented for the simultaneous
forcing tests are average values corresponding to the final 2 to 3 hours
of operation in each daily test, while the data from the hydraulic
forcing tests represent relatively long-term, steady-state values.
Considering only the data collected during the simultaneous forcing
tests (Curve I), for increases in soluble TOG loading of up to about 3.5
lb/1,000 ft *d, there is a net benefit in terms of total mass of TOC
removed. At higher organic loadings, the organic removal apparently
decreases, indicating that the process is stressed beyond its maximum
capacity for soluble TOC removal. These results are consistent with the
data from the hydraulic forcing tests, which are shown in Figure 10 for
comparison.
134
-------
H
W
U1
i.o t-
r 1.5
i.o
0.5
1E6END
A HYDRAULIC'AND ORBAHIC FORCING TESTS
D HYDRAULIC FORCING TESTS
0.5
1.0
1.5
2.0
2.5
3.0
3.5
SOLUBLE ORGANIC LOADING, Ib TOC/1,000 ft -d
B
4.0
4.5
Figure 10. RBC performance, organic removal.
-------
There were no apparent operational difficulties associated with these
When the loadine rates were returned to
-------
EFFECT OF ORGANIC LOADING ON RBC PROCESS
EFFICIENCY AND FIXED-FILM THICKNESS
By
Cheng-Nan Weng
Associate
Buck, Seifert and Jost, Inc., Consulting Engineers
Englewood Cliffs, New Jersey
Alan H. Molof
Associate Professor of Civil Engineering
Polytechnic Institute of New York
Brooklyn, New York
Secondary or biological treatment of wastewater has been provided
principally by the activated sludge and trickling filter processes. The former can
be described as a slurry or suspension process while the latter is a fixed film
process. One substantial modification of the trickling filter process involves the
use of biological films attached to rotating disks.
The rotating biological contactor (RBC) system, as now practiced, consists
of a series of reactors each of which contains a number of closely spaced rotating
vertical disks partially submerged in the wastewater. A biomass similar to
trickling filter slime is established on the surface of the disks. As the disks rotate,
they carry the biomass film saturated with wastewater into the air where it is
aerated to provide the dissolved oxygen (DO) required for aerobic biological
activity. In summary, the rotating disk is utilized as a supporting media for
biological growth, as a mechanism for aeration, and as a means of contacting the
microorganisms with the wastewater.
-------
For better design and operation of the RBC process, it is important to
understand the interaction of the loading and operating variabies such as organic
loading, flow rate, rotational disk speed, detention time, disk surface area,
submerged disk depth, and wastewater temperature. It is also important to
understand the effect of biological film age and thickness on the organic removal
efficiency.
This work describes the effects of detention time, influent concentration,
flow rate and biological film age and thickness on organic removal of RBC.
EXPERIMENTAL PROCEDURE
To study the above effects, a six-stage RBC system was constructed with
each stage 6 in. (15.24 cm) deep, 4 in. (10.16 cm) wide, and 8 in. (20.32 cm) long.
The system was partially submerged in a constant temperature water bath.
Each stage contained one or two 6 in. (15.24 cm) diam., 0.25 in. (0.64 cm)
thick plastic disks. The available surface area from each disk was 0.442 sq. ft. or
63.6 sq. in. (410 sq. cm). When there were two disks in each stage, the two disks
were spaced on 1.75 in. (4.45cm) centers. These were mounted vertically and
parallel to the flow on a 0.5 in. (1.27 cm) diam. horizontal stainless steel shaft.
The shaft rotated at controlled speeds by a roller chain connected to a 1/6 hp
(0.124 kw) ratiomotor.
In this type of laboratory reactor, the wall area effect could mask the effect
of the rotating disk area. Therefore, it was necessary to prevent biological growth
on the walls and bottoms of the reactor and restrict the biological film
accumulation to the rotating disks. This was accomplished by using two sets of
0.25 in. (0.64 cm) thick reactor linings for each stage and changing the linings three
times each day.
The substrate selected is shown in Table I. The protein was present in the
nutrient broth representing 65 percent of the chemical oxygen demand (COD), the
carbohydrate was present as glucose representing 25 percent of the COD, and the
fatty acid was present as sodium oleate representing 10 percent of the COD.
The effect of film age on organic removal rate was studied under three
different test conditions. These three conditions differed by flow, detention time,
or organic loading. The first study was made at a flow rate of 21 I/day, a loading
of 7.56 g COD/day, or 0.63 Ib. COD/100 sq. ft. apparent disk area/day (30.8 g/sq.
m/day), and a detention time of 96 min. per stage while the second study was done
at a flow rate of 42 I/day, a loading of 15.12 g COD/day, or 1.26 Ib.
COD/100 sq. ft. apparent disk area/day (61.6 g/sq. m/day), and a detention time of
24 min. per stage. The third study was made at the same flow rate and organic
loading as the second study, but at a detention time of 48 min. per stage. All three
studies were made using a single disk per stage, an influent COD concentration of
360 mg/1, a rotational disk speed of 30 rpm, and a liquid temperature of 20 C. In
the third study, the effect of film thickness on organic removal rate was also
investigated.
138
-------
The biological growths on the first three disks (stages) were removed
successively starting from the third disk with the first disk being the last. This
reverse procedure was used so that the feed to each stage would not be changed as
would be the case if the disk cleaning went from stage one to stage three. At least
four detention times of continuous slime removing was done before samples were
withdrawn for COD testing in order to insure the purging of the original treated
liquid in the reactor. Two successive runs were made on each disk.
The thickness of the biological film was determined with a Bausch and Lomb
phase contrast microscope. The microscope was focused on the top, and then on
the bottom of the film. The film thickness was determined from the number of
divisions turned on the fine adjustment between the two focusings. The microscope
was calibrated by focusing the top and bottom of one to five pieces of Corning
cover glass which ranged in thickness from 174.78 microns to 873.92 microns. The
actual thickness of the Corning cover glass was determined with a precision
micrometer. By plotting the thickness against the number of divisions turned on
the fine adjustment between the two focusings, it was found that each division on
the fine adjustment of the phase contrast microscope used in this investigation
represented 1.8863 microns of thickness.
Six and eight random film thickness measurements were made each time,
respectively, on the four pieces of cover glass and the eight pieces of cloth tape
which were attached symmetrically on the two sides of the disk.
All analytical determinations were made according to the recommendations
of Standard Methods.
RESULTS
Effect of Detention Time at
Constant Organic Loading
The effect of detention time was studied by two different methods which
are summarized in Table II. The detention time was first varied by keeping both
reactor volume and organic loading constant, and varying both flow and influent
concentratrion. This was studied at three different levels - 24, 48 and 96 min. per
stage. Varying detention time was also achieved by keeping flow rate, influent
concentration and organic loading constant, and varying only the reactor volume.
This was tested at three levels - 17, 24 and 48 min. per stage.
Varying Both Flow and Influent Concentration
Data on the effect of detention time by varying both flow and influent
concentration are summarized in Figure 1. It was shown that COD reduction for
the first stage reactor increased as detention time was increased from 24 min. to
96 min. However, increasing detention time had little effect on the overall organic
removal efficiency which was about 90% COD reduction.
139
-------
Varying Reactor Volume
The results on the effect of detention time by varying reactor volume are
shown in Figure 2 as COD reduction in each stage versus stage number. The COD
reduction in the first stage increased only very slightly as detention time per stage
was increased from 17 min. to 48 min. An increase in detention time had little
effect on the overall organic removal efficiency which was about 91% COD
reduction. The results show less fluctuation from decreasing detention time by
varying reactor volume than from varying both flow and influent concentration.
From both Figure 1 and Figure 2 it is readily seen that stage 4 removed
more substrate than stage 3 under all detention time conditions studied. This was
contradictory to the first order reaction theory if all conditions except organic
concentration were the same in both stages. It was found that mixed liquor
dissolved oxygen was absent in the first three stages, but was present in the last
three under all detention time conditions investigated. Therefore, it appears that
the absence of the mixed liquor dissolved oxygen was the cause for this reverse
condition.
Further insight into controlling mechanisms can be gained by assuming the
reaction was first order. If the biochemical reactions taking place in the RBC were
"first order" in character, or the rate of the reaction was proportional to the
amount of oxidizable organic matter remaining at any time, a straight line should
be obtained when the logarithms of the concentrations remaining were plotted
against the linear scale of time. Since the percentage COD remaining is
proportional to the concentration remaining and the detention time of the liquid in
the RBC is proportional to the number of stages, data were plotted as the
logarithm of the percentage COD remaining versus stage number.
The results showed that two or three straight lines in series instead of one
were obtained depending on the organic loading. Typical results exhibited three
straight lines and are shown in Figure 3. The first straight line extended over the
first three stages where mixed liquor dissolved oxygen was absent. The second line
covered the next two stages where mixed liquor dissolved oxygen was present. The
last line represented the last stage where the influent organic concentration was
low. The slope of the second straight line is greater than those of the other two.
Therefore it can be stated that organic utilization in the first three stages was
dissolved oxygen limited, the fourth and fifth stages was diffusion limited, and the
last stage was organic concentration limited.
Effect of Influent Concentration at
Constant Flow
The effect of influent concentration was studied under two different slime
area conditions - single disk and double disks per stage.
Single Disk Reactor
The effect of influent concentration was first tested using a single disk per
stage at four feed levels - 180, 360, 540 and 720 mg/1 COD at a flow rate of 42 1
140
-------
per day and a detention time of 48 min. per stage. The results are shown as stage
number versus COD reduction rate in Figure 4 and as percentage COD reduction in
Figure 5.
As expected, organic utilization in each stage increased as influent COD
concentration was increased to 730 mg/1 except in the fourth stage where more
organics were removed at an influent COD of 347.5 mg/1 than at influent COD
concentrations of 533 and 730 mg/1. This was a result of the mixed liquor dissolved
oxygen being absent at influent COD concentrations of 533 and 730 mg/1 and being
present at an influent COD of 347.5 mg/1.
For a six-stage reactor, the total COD reduction rate was increased from
6.70 grams per day to 20.70 grams per day as the influent COD concentration was
increased from 178 mg/1 to 730 mg/1. The two highest feed levels yielded almost
equal COD reductions in the last two stages even though the COD concentrations in
the reactors were different. At an influent COD of 533 mg/1, the effluent COD's
from the last two stages were 151.7 and 98.7 mg/1. At an influent COD of 730
mg/1, the effluent COD's were 288 and 237 mg/1. This indicates that the RBC
system was overloaded at these two high feed levels.
At low influent concentration, the last few stages are at low COD levels and
do not need to utilize their organic removal capacity. As a result, nitrification
takes place. As can be seen from Figure 4, the last three stages at an influent
COD of 178 mg/1 and the last two stages at an influent COD of 347.5 mg/1 did not
need to utilize their (Organic removal capacity and as expected, nitrification was
found in these stages.
Since the mixed liquor dissolved oxygen was absent in the first three stages
at an influent COD of 347.5 mg/1 and the first four stages at influent COD
concentrations of 533 and 730 mg/1, stage 4 removed more organics than stage 3
when the influent COD was 347.5 mg/1, and stage 5 removed more organics than
stage 4 when the influent COD concentrations were 533 and 730 mg/1. Stage 4 also
removed more organics at an influent COD of 347.5 mg/1 than at influent COD
concentrations of 533 and 730 mg/1. Again, the depletion of mixed liquor dissolved
oxygen was found to have a negative effect on the organic removal efficiency of
the RBC.
Figure 5 shows that the percentage COD reduction increased markedly as
influent COD concentration decreased from 730 mg/1 to 178 mg/1 for the RBC up
to three stages. For systems of more than three stages, this was also true as the
influent COD concentration decreased down to about 350 mg/1, and then leveled
out and became independent of influent concentration at low feed concentrations.
Double Disk Reactor
The effect of influent concentration was also studied using double disks per
stage at three levels - 360, 540 and 720 mg/1 COD at a flow rate of 42 liters per
day and a detention time of 48 min. per stage. The results are summarized in
Figures 6 and 7.
It is of interest to note the similarity in COD removal rates in stage 2 even
though the effluent concentrations from stage 2 were different, namely, 103.1,
141
-------
272.5, and *32.1 mg/1 COD for feeds of 3*5.*, 531.5, and 717,7 mg/1 COD
respectively. As the number of stages increased, the COD utilization decreased
more rapidly with decreasing influent concentration. As can be seen from Figure 6,
the last three stages at an influent COD of 3*5 mg/1, the last two stages at an
influent COD of 532 mg/1, and the last stage at an influent COD of 718 mg/1 did
not need to utilize their organic removal capacity and as expected, NH--N removal
took place in these stages.
Again, the absence of the mixed liquor dissolved oxygen was found to have a
negative effect on the organic removal efficiency. The mixed liquor dissolved
oxygen was absent in the first stage at an influent COD of 3*5 mg/1, the first two
stages at an influent COD of 532 mg/1, and the first three stages at an influent
COD of 718 mg/1. Stage 3 was found to remove more organics than stage 2 when
the influent COD was 532 mg/1 and stage * removed more COD than stage 3 when
the influent COD was 718 mg/1. Stage 3 also removed more organics at an influent
COD of 532 mg/1 than at an influent COD of 718 mg/1. Stage 2 utilized more
organics at an influent COD of 3*5 mg/1 than at an influent COD of 532 mg/1.
Figure 7 shows that the percentage COD reduction decreased markedly as
influent COD was increased from 3*5 mg/1 to 718 mg/1 for the RBC up to four
stages. For systems of more than four stages, the percentage COD reduction
became independent of influent concentration.
Effect of Flow Rate at Constant
Influent Concentration
Flow rate (hydraulic loading) determines the rate of organic addition and the
detention time of the liquid in the RBC system. The effect of flow rate was
studied at four levels - 1*, 21, *2 and 63 1 per day at an influent COD concentration
of 360 mg/1.
Figure 8 shows the effect of flow rate on the COD reduction rate. The COD
reductions in the first two stages increased only very slightly wHen the flow rate
was increased from *2 I/day to 63 I/day. This indicates that the first two stages
were overloaded at a flow rate of 63 I/day.
Since the mixed liquor dissolved oxygen was absent in the first one, three,
and four stages at flow rates of 21, *2 and 63 I/day respectively, and present in all
stages at a flow rate of 1* I/day, stage 1 removed more COD at a flow rate of
1* I/day than at a flow rate of 21 I/day, and stage * had a higher COD reduction at
a flow rate of *2 I/day than at a flow rate of 63 I/day. Stage 2 also removed more
COD than stage 1 when the flow rate was 21 I/day, and stage * had a higher COD
reduction than stage 3 when the flow rate was *2 I/day. Stage 5 also removed
more COD than stage * at a flow rate of 63 I/day. However, in general, increasing
flow rate increases the COD reduction rate.
Figure 9 shows that the percentage COD reduction increased markedly as
flow rate was decreased from 63 I/day to 1* I/day for RBC up to three stages. For
reactors of more than three stages, this was also true as flow rate was decreased
down to about *2 I/day and then leveled out and became independent of flow rate
at low feed rates.
142
-------
Effect of Biological Film Age
The results of the biological film study were calculated in terms of COD
removal rate (mg/hr) and correlated with film age as shown in Figures 10, 11 and 12
for the first, second and third tests respectively. There was a rapid organic
utilization by films in the early periods of their growth. Utilization rates were
then increasing more slowly and relative stability was eventually achieved. The
slight COD reduction at time zero was most likely achieved by the suspended
microorganisms. Figure 10 shows that it took the first, second and third disks 25,
34.5 and 38 hours respectively to recover the efficiencies to their former levels.
Figure 11 indicates that 33, 50 and 42 hours were required, respectively, by the
first, second and third stage disks to reestablish the normal efficiencies of
treatment. It was also found in a third test that 34, 44 and 33 hours were required,
respectively, by the first, second and third disks to recover their normal
efficiencies as shown in Figure 12. In these three figures, the two straight line
portions of each curve were drawn by the method of least squares.
While Figures 10, 11 and 12 delineate average rates, they do not demon-
strate trends that might have occured during the individual runs. However, the
scatter of data shown in these figures suggests that there may be significant
trends. Therefore, data are presented separately for runs 1 and 2 of the second
stage disk of the first study in Figure 13. The trends indicated by the data were
thought to be a reflection of the dynamic nature of the mixed-culture microbial
films. Sharp decreases were observed in the COD removal rates between 40
and 60 hours. However, a relative stability in substrate utilization rates would be
apparent when films were older than approximately 70 hours as shown in Eigure 13,
and as evidenced by the results obtained from the equilibrium period . These
results of an increase, a decrease and an increase in substrate utilization support
only those of Hoehn » although he used film thickness instead of film age as the
independent variable.
Effect of Biological Film Thickness
Figure 14 shows the effect of film thickness on the COD removal rate. The
two straight line portions of each curve were drawn by the method of lease squares.
It is evident that there was a rapid uptake of organics by films in the early
periods of their growth. Utilization rates were then increasing more slowly and
relative stability was finally achieved. Similar trends of rate changes have been
reported by other investigators * * . Figure 14 shows that reestablishments of the
normal efficiencies of treatment were achieved when the films on the first, second
and third disks were, respectively, 135, 220 and 265 microns thick.
While Figure 14 depicts average rates, it does not manifest trends that
might have occurred during the individual runs. Therefore, data are presented
separately for runs 1 and 2 of the third stage disk in Figure 15. Sharp decreases
were observed in the COD removal rates between 350 and 600 microns. A relative
stability in organic utilization rates was apparent when films were thicker than
approximately 600 microns or more than 70 hours old as shown in Figures 15 and 13
respectively, and as evidenced by the constancy of the data f rom studies during the
equilibrium period . These results support only those of Hoehn .
143
-------
DISCUSSION
Effect of Detention Time at
Constant Organic Loading
A decrease in detention time from 96 min. per stage to 2^ min. per stage by
varying flow and influent concentration at a constant organic loading was found to
have no appreciable effect on the organic removal efficiency of the RBC, It was
also found that decreasing detention time from ^8 min. per stage to 17 min, per
stage by varying reactor volume at a constant organic loading had no significant
effect on the organic removal. Due to the restrictions in the physical conditons,
lower detention time levels were not obtained. The results indicate that the
critical detention time was equal to or less than 17 min. per stage, and the system
could be operated at a detention time of 17 min. per stage without losing
efficiency.
Effect of Influent Concentration at
Constant Flow
It was found that increasing influent concentration at a constant flow
resulted in an increase of the COD removal rate while the percentage COD
reduction decreased. These results are not unexpected. Organic removal is also
mass transfer limited even if the mixed liquor dissolved oxygen content is a
controlling factor. Mass transfer is directly proportional to the concentration
gradient. Increasing influent concentration increases the concentration gradient
which, in turn, increases mass transfer and thereby increases the organic removal
rate. However, above a certain loading the biomass and the resulting organic
removal rate is not sufficient to prevent the overall percentage reduction from
dropping to a lower level.
Effect of Flow Rate at Constant
Influent Concentration
Increasing flow rate at a constant influent concentration resulted in an
increase in both effluent COD concentration and rate of COD reduction with a
decrease in the percentage and mg/1 COD reduction. The most plausible
explanation for these results is that organic removal is mass transfer limited
because organics in the layer of liquid immediately adjacent to the slime layer is
depleted rapidly. Mass transfer is directly proportional to the concentration
gradient. At low flow rates the concentration gradient penetrates into the bulk
liquid film thus reducing the magnitude of the concentration gradient. Increasing
feed rate reduces penetration of the concentration gradient into the bulk of the
film, until at high feed rates depletion of organics is limited to the area
immediately adjacent to the slimej as a result, the concentration gradient extends
over a shorter length and is numerically larger . At high flow rates, the
concentration gradient is limited to a thin liquid film adjacent to the slime layer,
and further increases in liquid feed rate have no effect. Therefore, organic
removal becomes independent of flow rate at high feed rates, as seen in Figure 8.
144
-------
Effect of Film Age and Thickness
Each run made in this investigation to assess the organic removal by
biological films indicated clearly that at film ages between 40 and 60 hours or
thicknesses between 350 and 600 microns the rates of organic removal changed
markedly. The composite data of Figures 10, 11, 12 and 14 would suggest that the
removal rates stabilized at some constant values when films were between 25 and
50 hours old or between 135 and 265 microns thick.
Torpey et al. found that only 18 hours after cleaning were required to
restore a biological growth on the disk surfaces of the first stage with the
reestablishment of normal efficiency of treatment. They did not report the
procedures used for this evaluation. However, it seems that no continuous slime
removing was made by them before samples were withdrawn in order to purge the
original treated liquid in the reactor. This is a possible explanation for the
difference between the limiting film ages obtained by Torpey et al. and this
investigation. Another possible explanation for this difference is that different
cultural and operational conditions existed in the two studies.
Perhaps the most significant trait of the composite removal rate curves
(Figures 10, 11, 12 and 14) is that they tend to support the theory that the
assimilation capacity of the biological films remains constant beyond some limiting
film age or thickness. It should be emphasized, however, that this is based on
composite data. It is useful, therefore, only in the determination of the average
results.
The limiting film ages found for disks one and two in the first study were
much less than those found in the second and third studies. This is most likely due
to the fact that the organic loading in the first study was only one-half of that in
the other two studies. Therefore, the organic removal rate in the first study was
less than those in the other two studies. This, in turn, enabled the films in the first
study to reestablish their normal efficiencies in shorter times. However, the
limiting film age for the third disk in the first study, 38 hours, was more than that
in the third study. This event along with the fact that each disk in each study
required different periods of time to recover their former efficiency may possibly
be explained on the basis of different cultural conditions on each disk.
Figure 14 shows that reestablishments of the normal efficiencies of
treatment were obtained when the films on the first, second and third disks were,
respectively, 135, 220 and 265 microns thick. Greater film thicknesses resulted in
no increase in the organic removal rates. These values were taken as the
thicknesses of the active microbial films. The above three active film thicknesses
differ from each other. This again may possibly be explained on the basis of
different cultural conditions on each disk.
Data presented by Kornegay and Andrews and Tomlinson and Snaddon
show that the organic remova.1 stabilized at the maximum rates observed when the
limiting thicknesses were reached. However, both groups of data are of composite
data.
145
-------
Even in studies conducted with strictly controlled laboratory systems, the
investigator cannot duplicate exactly the cultural conditions each time.
Differences in microbial metabolism are brought about by slight shifts in
predominance of organism-type. Therefore, if the results of metabolic studies of
several film cultures are considered together, the fluctuations in data from the
individual studies will most likely be masked. An average result will therefore be
defined.
Cultures were not reproducible in every detail in this investigation. The
individual organic removal curves reflect differences that were masked by the
composite curves.
The individual utilization rate curves (Figures 13 and 15) were interpreted as
indicating that some factor or factors caused organic uptake by the microbial films
to decrease sharply when they were between ^0 and 60 hours old or 350 and
600 microns thick. Relative stability of organic removal rates and recovery to
their former levels were achieved by the time ages of approximately 70 hours or
thicknesses of about 600 microns were reached. Thereafter, a quasi-steady state
with respect to organic removal was apparent.
o
Sanders showed that organic removal rates for microbial films decreased
after the limiting film thickness for oxygen diffusion to the lower microbial layer
had been exceeded. However, there is evidence, based on Sanders' own data, that
the rates were starting to increase at the time the experiments were terminated.
Had the growth not dropped off, recovery of the films might have been manifested
by further increases in the organic removal and a reestablishment of the former
rates.
Maier did not study films less than 480 microns thick. It is possible that
changes in organic removal rate would occur before this thickness was reached.
The observed decreases in the individual organic removal rates might be
caused by some film instability brought about by the limitation of some necessary
growth factor. This growth factor may have been oxygen, as proposed by Sanders ,
Kornegay and Andrews „, and Maier , or it may have been nutrient as proposed by
Tomlinson and Snaddon . The limiting of either oxygen or nutrient would result in
population changes within the film, and a period of readjustment would be required.
It is postulated that during this period the organic removal rate would decrease
particularly if endogenous respiration were increasing. The latter would result in
utilization of nutrient either stored in the film or supplied by the microorganisms
themselves and, in effect, reduce the removal of organics from the incoming
wastes. Organisms within the film would be dying during this period which would
provide an additional source of nutrient to the living microorganisms.
The findings of Hoehn and Ray for the reversal effects of film thickness on
nutrient utilization rates are in good agreement with the findings of this study,
although the absolute magnitude of the maximum removal rates and the film
thickness at which they occurred differ from those of this investigation due to
different cultural conditions in the two studies. Their finding that composite
results mask the reversal effect was also noted in this work.
146
-------
SUMMARY AND CONCLUSIONS
1. Organic removal in the RBC can be limited by the dissolved oxygen content,
diffusion or organic concentration.
2. Increasing influent concentration at a constant flow resulted in a decrease
of both the percentage COD reduction and the dissolved oxygen content of
the mixed liquor, while the rate of COD removal increased.
3. Increasing flow rate at a constant influent concentration resulted in an
increase in both effluent COD concentration and rate of COD reduction,
with a decrease in the percentage and mg/1 COD reduction and dissolved
oxygen content of the mixed liquor.
4. At low organic loadings, most of the organics is removed in the first few
stages of the RBC with very little being utilized in the following stages.
This allows NH--N removal to take place in these later stages when the
COD and BOD are reduced to 50 mg/1 and 14 mg/1 respectively.
5. Organic removal by mixed-culture biological films initially decreases when
some limiting film age or thickness is reached. However, the organic
removal will increase again to recover its former level as the films get older
or grow thicker. A quasi-steady state (with regard to organic removal) will
then be established.
ACKNOWLEDGEMENTS
This research was supported in part by a Research Fellowship from the
Office of Water Programs of the Environmental Protection Agency, No. 1-F1-WP-
26, 708-01, and by a Federal Water Quality Administration Traineeship, Grant
No. 5T2-WP-186.
147
-------
REFERENCES
1. Standard Methods for the Examination of Water and Wastewater, 12th
Edition, American Public Health Association, New York (1965).
2. Weng, C.N., "Biological Fixed-Film Rotating Disks for Wastewater
Treatment." Ph.D. thesis, New York Univ., Bronx, N.Y. (1972).
3. Hoehn, R.C., and Ray, A.D., "Effects of Thickness on Bacterial Film."
Journal Water Pollution Control Fed., 45, 2302 (1973).
4, Maier, W.3., Behn, V.C., and Gates, C.D., "Simulation of the Trickling Filter
Process." Journal of the San. Eng. Div., ASCE, 93, SA4, 91-112 (1967).
5. Tomlinson, T.G., and Snaddon, Dorothy, H.M., "Biological Oxidation of
Sewage by Films of Microorganisms," Air and Water Pollution International
Journal, 10, 865 (1966).
6. Torpey, W.N., Heukelekian, H., Kaplovsky, A.3., and Epstein, R., "Rotating
Disks with Biological Growths Prepare Wastewater for Disposal or Reuse,"
Jour. Water Poll. Control Fed., 43, 2181 (1971).
7. Kornegay, B.H., and Andrews, 3.F., "Kinetics of Fixed-Film Biological
Reactors," JWPCF, 40, Research Supplement, R460-R468 (1968).
8. Sanders, W.M., III, "The Relationship Between the Oxygen Utilization of
Heterotrophic Slime Organisms and the Wetted Perimeter." Ph.D. thesis,
The Johns Hopkins Univ., Baltimore, Md. (1964).
9. Weng, C.N., and Molof, A.H., "Nitrification in the Biological Fixed-Film
Rotating Disk System." Journal Water Pollution Control Fed., 46, 1674
(1974).
148
-------
TABLE I
Synthetic Wastewater Substrate
Material Amount*
Trace salt solution,. **ml 16.66
(NH^)2SO^,g 0.264
NH^Cl,g 2.537
CaCl2.2H2O,g 0.29*
MgCl2.6H2O,g 0.407
CaSOr2H2O,g 1.350
MgSOr7H20,g 3.650
Na3POr12H2O,g 3.650
KH2P04,g 6.434
K2HPOrg 16.MO
Na2HPOr7H2O,g 1.500
Nutrient broth, ***g 9.000
Glucose, g 3.798
Sodium oleate,g 0.603
COD,mg/l 360
*Diluted to 45 1 with deionized tap water.
»*Dilute 5.0 g of FeCl3.6H2O, 0.672g of A1C13.6H2O, 0.342 g of
CoCl2.6H2O, 0.15g of MnSO^.H2O, 0.06 g of (NH^JgMo^^.Wj
and 0.01 g of ZnCl2 to 1 1 with distilled water.
***Bio Cert™ Nutrient Broth, Dehydrated, a-1089-C, Fisher
Scientific Company, Pittsburgh, Pa.
149
-------
H
Ul
o
Table II
Detention Time Study Conditions
Organic Loading
Study
Varying Flow
and Influent
Concentration
Varying Only
Reactor
Vo lume
g/day
15.3?
14.59
15.25
14.64
15.02
14.59
lb/ .
1000 ft3/
day
114.2
108.4
113,3
108.8
111.6
108.4
as COD
1W 0
100 ft2/
day
1.28
1.22
1.27
1.22
1.25
1.22
Flow
Rate
(I/ day)
84
42
21
42
42
42
Influent
COD
Cone.
-------
Detention Time
(min/stage)
96
12345
Stage Number
Figure l Effect of Detention Time by Varying Flow and
Influent Concentration at Constant Organic
Loading
151
-------
Detention Time
(min/stage)
Figure 2
2 34 5 6
Stage Number
Effect of Detention Time by Varying Reactor
Volume at Constant Organic Loading
152
-------
Dissolved Oxygen Content Limited
Diffusion (Mass
Transfer) Limited
Substrate
Limited
STAGE NUMBER
Figure J "^Effect of Dissolved Oxygen and Substrate
Concentration
153
-------
Influent COD
(me/I)
730
3 4
Stage Number
Figure * Effect of Influent Concentration on COD
Reduction at Constant Flow with Single Disk
154
-------
100
90
80
70
5 60
o
•e
§
e*
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800
Influent COD Concentration, mg/1
Figure 5 Effect of Influent Concentration on % COD
Reduction at Constant Flow with Single Disk
'155
-------
8f-
Influent
COD, mg/1
Figure 6
23456
Stage Number
Effect of Influent Concentration on COD
Reduction at Constant Flow with Double Disk
156
-------
100
90
80
70
jj 60
•u
o
50
§
30
20
10
6 Stages
1 Stage
! I
I
I I I I I
0 100 200 300 400 500 600 700 800
Influent COD Concentration, mg/1
Figure 7 Effect of Influent Concentration on % COD
Reduction at Constant Flow with Double Disk
157
-------
4.0i—
63 I/day
3 4
Stage Number
42 I/day
21 I/day
14 I/day
Figure 8 Effect of Flow Rate on COD Reduction at Constant
Influent Concentration
158
-------
o
100
90
80
70
60
1 50
o
o
« 40
»<
30
20
10
6 Stages
I I
L t I I I I
10 15 20 25 30 35 40 45 50
Flow Rate, Liters/Day
I I I
2.4 1.6 0.8
Detention Time/Stage, Hours
55 60 65
0.533
Ffgure 9 Effect of Flow Rate on % COD Reduction at Constant
Influent Concentration
159
-------
110
100
90
^ 80
2 ,„
UJ
< 60
< 50
O
Sko
tu
* 30
Q
> U 20
10
—n— STAGE 1
mm**m STAGE 2
•«O»i STAGE 3
J ......... i ....... i, ....... ! ........ i
i., ..... ,,i,.,..j
o 10 20 30 ho 50 60 ?o 80 90 100 no 120 130
BIOLOGICAL FILM AGE, HOURS
Figure 10 Effect of Slime Age on COD Removal Rate - First Study
-------
~ O- •" Stage 2
——O—— Stage 3
10 20 30 40 50 60 70 80 90 100
Biological Film Age, Hours
Figure II Effect of Slime Age on COD Removal Rate -
Second Study
161
-------
-,-,-Q—._ Stage 3
0 10
20 30 40 50 60 70 80
Biological Film Age, Hours
100
Figure 12 Effect of Slime Age on COD Removal Rate -
Third Study
162
-------
Oi
1
**
Ul
§
D
O
u
UJ
<
Ul
I
UJ
Q£
Q
O
U
loo
90
80
70
60
50
fco
30
20
10 t
RUN 1
-> RUN 2
I
1
0 10 20 30 ^0 50 60 70 80 90
BIOLOGICAL FILM AGE> HOURS
Figure 13 COD Removal Rate at Second Stage
163
-------
140
130
120
110
100
90
« 80
4J
&
* 70
n)
1 60
pi
o
8 50
40
30
20
10
0
o
o
o
-—6-
n
Stage 1
— O
—O
•~ Stage 2
-—Q—. Stage 3
200
Mean Film Thickness, Microns
250 300 350 400 450 500 550 600 650 700 750 800 850
I I I I I I I I I I I I I
Figure 1*,,. Effect of Slime Thickness on COD Removal Rate
-------
H
CTi
(Jt
•a
110
100
, 90
i
° 80
it
i
o
! 70
60
50
40
o
§ 30
8 20
10
0
i
50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850
Mean Film Thickness, Microns
Figure 15 COD Eemoval Rate at Third Stage
-------
Page Intentionally Blank
-------
MICROFAUNA AMD RBC PERFORMANCE:
"LABORATORY AND FULL-SCALE SYSTEMS
By
George E. Hoag
Research Assistant
Wilbur J. Widmer
Professor
Wendell H. Hovey
Associate Professor
Department of Civil Engineering
University of Connecticut
Storrs, Connecticut, U. S. A.
Introduction
The transformation of soluble substrate to cells, metabolic products and
excreta is the basis of the rotating biological contactor process. Metabolic
activities associated with the rotating biological contactor occur in fixed-
film, or as they are more commonly termed, biofilm cultures. Bacterial
populations exist as a response to combinations of various environmental
factors. The sum of all these factors constitute, in an ecological sense,
a niche, or a specialized environment suitable for microbial occupation.
These saprophytic bacteria are in turn predated by holozoic organisms such
as the Phyla: Protozoa, Rotifera and Nematomorpha. Some members of these
Phyla may prey upon either bacteria, algae or other members of the above-
mentioned Phyla. Thus, the niches for these organisms are defined by the
presence of other organisms as well as environmental factors.
Factors that affect the growth of microbial populations are generally:
I. Physical, II. Chemical, and III. Biological. These factors individually
and synergistically alter suitability and as a consequence, abundance of
particular organisms. The abundance of an organism in an environment can be
167
-------
used as a measure of its Importance in that ecological structure.
Biological treatment processes select for particular organisms by
creating very defined niches. However, while these processes are fundamen-
tally dependent on the activity of living organisms, their design does not
account for the specific characteristics of the organisms.
Recently, problems associated with activated-sludge bulking have
stimulated morphological studies of sludge-floe particles.!•2- Control of
activated sludge filamentous bulking by increased hydraulic mixing was
examined by Chudoba, et al. (1973).3- They found that increased staging
caused more distinct biological cultures, and a decrease of filamentous
organisms in pre-clarified effluent.
Hawkes (1963) integrated fundamental relationships of biology and
ecology to trickling filter and activated sludge operations.^- He recog-
nized that there was a succession of microfaunal organisms affecting the
efficiency of biological treatment. An informative representation of
microfaunal food chains was developed for both processes.
McKinney (1962) briefly discusses predominance of microfauna in
wastewater treatment processes and presents an interpretation of micro-
faunal succession with respect to degree of treatment.5- He indicates
that protozoa are particularly valuable indicators of the performance and
stability of treatment processes.
Various interactions between aquatic organisms have been developed by
Bungay and Bungay (1968). A list and simple definitions are found in
Table 1.
TABLE1
MICROBIAL INTERACTION
Interaction Definition
Neutralism No interaction
Commensalism One benefits, other unaffected
Mutualism Each member benefits from the other
Competition A race for nutrients and space
Amensalism One adversely changes the environ-
ment for the other
Parasitism One organism steals from the other
Predation One organism ingests the other
Although Table 1 contains a convenient list of interactions, and all are
possible in the rotating biological contactor. However, clear cut boundaries
cannot always be defined.
Recently, Commensalism, Mutualism, Amensalism and Predation have been
regarded as important factors mediating activities of aquatic organisms.7.
168
-------
Competition and predaticm, in these authors* view, are of great impor-
tance in the biofilms associated with rotating biological contactors.
Changes in process parameters such as hydraulic retention time, cellular
retention time, and specific growth rates of bacteria and protozoa may have
an effect on the predator-prey relationships in the biofilm.
The importance of predator—prey interactions was examined in continuous
flow systems by Pirt and Bazin (1972).8. The necessity of maintaining
predator and prey growth rates either greater or less than the dilution rate
was discussed. They claimed that more efficient substrate conversion and/or
microfaunal predation can be attained if these dilution rates are properly
controlled.
A reduction in effluent biomass due to protozoa predation of bacteria
is reported to be significant (Sudo and Aiba, 1972).9« They state that the
average yield of protozoan mass from bacterial mass is 0.5.
Most investigations of microfaunal interactions as a response to changes
in operating parameters are conducted in either pure or mixed suspended
growth systems. The list of environmental factors affecting organisms in
the RBC process may be different from the list of factors that affect
organisms in other processes. Where the lists have factors in common, the
biological responses to the common factors may be different.
Full-scale activated sludge and trickling filter culture studies have
revealed that particular species and taxa occur with greater frequencies in
these cultures. Curds and Cockburn (1970a, 1970b) examined a variety of
trickling filters and activated sludge units. They determined the frequency
that individual species were present in these processes.10•> H- They found
that ciliated protozoa were generally the most abundant microfauna in
effluents from both processes. Also, the microfauna of both processes were
similar in that the organisms were from the same orders. Individual species
within these orders did vary. Activated sludge units that did not contain
ciliated protozoa contained significant populations of flagellates. Effluent
from these plants was turbid and of low quality.
To date, there is a lack of published investigations addressing the
types, abundance and ecology of microfauna in rotating biological contactors.
In view of the importance of microfauna in biological wastewater treatment
processes, we decided to examine the organisms associated with these bio-
films. The following were objectives of this study:
I, To examine what types of bacteria are responsible for soluble
substrate conversion.
II. To identify predatory microfauna that ingest substrate removing
bacteria.
III. To investigate successional patterns exhibited by the micro-
fauna in a rotating biological contactor.
IV. To evaluate the possibility of using microfauna as indicators
of effluent quality and stability of rotating biological con-
tactor biofilms.
169
-------
Methods and Materials
Results from the examination of biofilms from two rotating biological
contactors are presented in this paper. A plexiglass laboratory scale unit
was constructed and an experimental program initiated to analyze various
components of the system. Results from a if-25 factorial design experiment
have been reported elsewhere. A full-scale Clow-Envirex rotating
biological contactor used to treat an overloaded trickling filter effluent
was examined during the course of study. It was hoped that during the
course of study, the full-scale RBC would replace the trickling filter.
However, this did not occur. Consequently, the sampling program on the
full-scale unit was limited. These results will be discussed in this paper.
The laboratory unit was a continuous flow, four stage reactor.
Specific design of the unit and the experimental program has been described
by Hoag and Hovey (1980). Some basic characteristics at the laboratory and
full-scale units are found in Tables 2 and 3.
TABLE 2
CHARACTERISTICS OF ROTATING BIOLOGICAL
CONTACTOR UNITS STUDIED
Laboratory Full-Scale
Unit Unit
Number of Stages . 4 6
Discs per Stage 3
Disc Diameter (ft) 5 12
Surface Area per Stage (ft2) 71 100,000
Total Surface Area (ft2) 284 2.4 x 106
BOD Mass Loading Rate
(lbs/1000 ft2 - day) 3.6 or 7 0.12
NHg-N Mass Loading Rate
(lbs/1000 ft2 - day) 0.3 or 0.5 0.17
Peripheral Velocity (ft/min) 56 or 76 57
170
-------
TABLE 3
ONE-HALF 25 EXPERIMENTAL DESIGN
Factor
Number
1
2
3
4
Level
Factor (-)
Revolutions per minute (RPM) 5.9
Hydraulic Flow Rate (liters/day)
Disc Surface Area Exposed (percent)
Nitrogen Mass Loading Rate
(grams NH~-N per day)
Carbon Mass'Loading Rate
(grams glucose per day)
.1
032
60
41.5
258
1750
74
65
500
.0
EXPERIMENTAL
CONDITION
NUMBER
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
ORDER
RUN
7
1
12
8
13
2
5
16
4
6
11
10
15
3
9
14
FACTOR LEVELS
Factor 12345
.1
4- +
171
-------
In brief, the full—scale treatment plant was an overloaded trickling
filter plant upgraded through the use of rotating biological contactors.
At the time of the experiment, the rotating biological contactor was treating
trickling filter effluent Characteristics of the full-scale RBC given in
Table 2. This treatment plant did not control pH or alkalinity. Low
alkalinities, low pH and low nitrogen mass loading rates resulted in a
significant underloading of the process,
EXPERIMENTAL PROCEDURE
During each of the 16 laboratory experimental runs, biofilm samples were
removed from the discs by means of a glass slide. The culture was micro-
scopically examined immediately after it was removed from the discs. Micro-
fauna present in the cultures were identified using taxonomic keys by Kudo
(1966)13.} Martin (1968)14- and Gibbons (1974).15- Due to the difficulty of
removing a fixed mass and volume of biofilm and then making representative
serial dilutions, a quantitative analysis of population number was not
possible. Relative numbers of individuals of a particular species were .„
determined in a manner similar to a method used by Curds and Cockburn (1970).
Population levels were recorded by estimating relative numbers of a species
using the following levels: (+) few, (4+) many, (x) very many, and (xx)
extremely high level. While the numbers of organisms are not numerically
estimated, the levels presented indicate proportionate sizes of the popula-
tions .
Biofilm was removed from the laboratory discs approximately six inches
from the outer edge of the discs. Biofilm samples from the full scale unit
were taken at a distance of 2.0 feet (0.6 m) from the outer edge of the
discs. Samples used for the determination of wastewater chemical charac-
teristic were taken from the bulk liquid in all cases.
In addition to the microscopic examination of the biofilms, many visual
changes of the texture, color and density were observed during the experi-
mental period.
Results
The characteristics of the biofilm cultures varied from stage to stage
and often shifted between experimental runs. Biofilm thickness, in the first
stage, while not measured, was observed to become very rough textured and
thick with increases in carbon mass loading. Similarly, the latter stages
increased thickness and roughness with increases in nitrogen mass loadings.
Microfauna of the biofilm samples also varied between stages and runs.
Four classes, represented by 16 orders of protozoa, were present at some
time during the laboratory study. Frequencies of occurrence of the various
microfauna taxa, separated by stages, are found in Table 4. While Protozoa
were classified to species, the Phyla Rotifera and Nematomorpha were
classified to order.
As indicated in Table 4, members of the class Sarcodina occurred 69%
of the time in the 3rd and 4th stage laboratory biofilm cultures. Amoeba
172
-------
TABLE 4
CLASSIFICATION 01 MICROFAIMA FROM
LABORATORY RBC BIOFILMS
TAXONOMIC CLASSIFICATION
FREQUENCY
Percentage of Biofilms
containing Species
Phylum
Class
Subclass
Order
Genus
Species
Colpoda inflata
Order Suctorida
Podophyphra elongata
Podophyphra fixa
Order Hymenostomatida
Saprophilus muscarum
Uronema marina
Colpodium colpada
Parameeium caudatum
Parameeium bursaria
Glaucoma scintillans
T A G E
1234
Protozoa
Class Ciliata
Order Gymnostomatida
Chilodonella cucculus
Traecophylum pus ilium
Dldinium nasutum
Orthodonella guttula
Coleps hirtus
Spathidium spathula
Prorodon griseus
Prorodon teres
Litonotus lamellae
Litonotus fasciola
Order Trichostomatida
6
6
_
6 6
_
6
_
_ _
19 50
13 44
6
6
-
_
-
-
6
6
63
44
6
6
6
6
6
-
-
56
44
6
13 6
13
31 19 25 38
6
6
31 13 38 13
6
44 56 50 38
173
-------
TABLE 4 (cont.)
123
Order Peritrichida
Vortieella campanula
Vorticella mlcrostoma
Vortieella acquilata
Vorticella nebuifera
Opercularia coarctata
Epistyllis plicatis
Telotrochidum henneguyi
Order Hypotrichida
Euplotes patella
Euplotes harpa
Stylonchia mytilus
Stylonchia pulsata
Osytricha
Class Sarcodina
Order Aaoebina
Amoeba guttula
Amoeba proteus
Amoeba verucosa
Amoeba gorgonia
Amoeba limax
Amoeba striata
Order Testacea
Diffuglia oblongata
Order Proteomyxida
Nuclearia simplex
Order Heliozoida
Heterophrys myriopoda
Class Mastigophosa
Order Chrysomoadida
Monas amoebina
Monas obliqua
Monas vulgaris
19 31 44
6 6 13
38 38 19
19
25 13 13
6 6
6 19 13
6
6
6
13
6
6
6
19 25
19 69 56
13 -
6 38 38
- - 6
13 13
25 69
6 44 31
6 6 25
19 6 6
6
6 13 19
174
-------
TABLE 4 (cont)
1234
Order Phytomonadida
Chlamydomonas sp.
Carteria globosa
Order Euglenida
Astasla dangeardi
Peranema trichophorum
Class—Zoomastigophosea
Order - Rhizomastigina
Mastigamoeba longifilium
Mastigamoeba reptans
OrderProtomastigida
Bodo caudatus
Bodo globosus
Bodo lens
Bodo mutabilis
Cercobodo radiatus
Pleuromanas jaculans
Order Polynias tigida
Tetramitus pyrofonais
Phylum Rotifera
Class Digonata
Order Bdelloidea
Philodina
Rotatia sp.
Class Monogononta
Subordis Plolme
Encentrum sp.
Phylum Nematomorpha
6
13
13
6
6 31
13 19
13
6
6
6
6
31 63 38
6
25 88 88 38
175
-------
guttula, A. proteus, A. gorgonia, Diffuglia oblongata, Huclearia simplex and
Heterophrys myriopoda were the species most commonly found of this Class. Class
Ciliata was represented by 31 species and occurred in 44%, 56%, 63%, and 56% of
the cultures in Stages 1 through 4, respectively. Both Free-Swimming Ciliates
and Attached Ciliates were numerous during the course of experimentation. Class
Mastigophora species (Flagellates) occurred infrequently. The only flagellates
which were present in 20% or more of the biofilms per stage was Peranema trich-
ophorum (31% of 4th Stage Cultures). Phylum Rotifera was absent in all 1st
stage cultures but occured in 31%, 63%, and 38% of the 2nd, 3rd, and 4th stage
cultures, respectively. Phylum Nematomorpha occurred with the greatest fre-
quency of any group (88% in both 2nd and 3rd stages).
Filamentous Bacteria were always dominant in the first stage cultures.
A very feathery like series of filaments were quite evenly distributed over the
entire discs. Although some of the experimental runs had lower depths of submer-
gence, the cultures tended to cover the entire disc. At times, the bacterial
growth near the center of the disc was not as thick and/or rough as was found
on the submerged (wetted) area. The filiamentous bacteria were not confined to
the first stage cultures in all cases. Two types of filamentous bacteria were
identified during the study. A Sphaerotilus sp. (probably S. natans) and a
Norcardia sp. were the major filamentous species. A Zooglea sp. (probably Z.
ramigera) was present most of the time in the 2nd, 3rd, and 4th stages cultures.
These bacteria formed a biological matrix in which the other microfauna lived.
The amount of surface area these bacteria matrices provided to the predatory
microfaunal was quite astounding to the microscope viewer.
The roughness and thickness of these bacterial matrices varied from species
to species. The roughest cultures were Sphauotilus and Norcardia. Cultures
associated with the Zoogleal colonies were the smoothest. Roughness and texture
of the individual species did change with changes the process factors. Table 5
indicates which bacteria were dominant in the four stages for the 16 experimental
conditions.
Analysis
Reduction of the number; of individual taxa was conducted for the following
reasons: First, while individual species may not be present in all RBC biofilms,
members of the same Order may perform a function similar to that of the absent
species. Second, examination of all species may not be necessary because of
similarities in types food ingested and in means of motility. Third, this meth-
odology would have a much greater application for treatment plant operators.
Fourth, the necessity of tedious taxanomic identification to the level of species
is alleviated. Accordingly, species were grouped together by motility. The
groups, and orders contained in the groups are shown in Table 6.
Another method of reducing the burden of species identification is counting
the number of species per stage without identifying the species. A list of the
number of microfauna species per stage (excluding bacteria) for the 16 experi-
mental runs is shown in Table 7. The average number of microfauna species per
stage is also shown in Table 7. Although the number of species per stage in-
creases from stage to stage, the rate of increase is much lower going from
Stage 3 to Stage 4.
176
-------
TABLE 5
BACTERIA IDENTIFIED IN BIOFILMS
Expr
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
*
A
S
X
4
XX
4+
X
X
+
XX
XX
4-
4-
X
4
1
N Z
X X
XX
XX
X
X
XX
44
XX
XX
X
XX
XX
XX
2
S N
+
X
X XX
X
44- 44-
X
4+ x
4+ 44-
X
XX
4 xx
4 4
4- x
XX
x x
S
Z
XX
X
4+
X
XX
X
X
X
4+
44-
XX
X
X
T
3
S
44-
4
4
x
4
44-
+
+
X
X
X
A
N Z
XX
44- xx
4- xx
XX
x x
4- xx
44- x
XX
X X
XX X
XX X
XX
XX
XX
X X
G E
4
S N Z
XX
XX
XXX
4- xx
44- xx
XX
4+ 44- xx
XX
XXX
XX
4 4 xx
XX
XX
4 4 xx
4- = few *S = Sphaer.oti.lus sp.
44- = many N = Norcardia sp.
x = very many Z = Zooglea sp.
xx = extremely high level
j *
177
-------
Another response was that, in some of the runs, the number of microfaunal
species decreased from stage 3 to 4. This condition was observed in experi-
ments 1, 4, 6, 10, 13 and 16.
Table 6
Grouping of Microfauna
Group Orders or Phyla included in groups
Sarcodina Amoebina
Testacea
Proteomyxida
Heliozoida
Flagellates Chrysomoadidg
Phytomonadida
Euglenida
Zoomastigophorea
Prototaastigida
Polymastigida
Free swimming Ciliates Crymnostomuatida
Trichostomatida
Hymenos tomat ida
Hypotrichida
Attached Ciliates Suctorida
Peritrichida
Rotifers Bdelloidea
Monogononta
Nematodes Phylum Nenatomorpha
These 6 runs represented 6 of the 8 runs with high glucose mass loading rates
(500 g/day).
Table 7
Numbers of Microfaunal Species per stage
Exp.
Run
1
2
3
4
5
6
7
8
1
1
—
4
1
5
1
3
5
2
6
_
11
7
10
3
6
6
STAGE
3
11
—
10
7
10
4
11
9
4
9
-
11
5
13
4
12
12
178
-------
Table 7 (cont.)
Numbers of Microfaunal Species per Stage
Exp
Run
1
STAGE
2 3
4
9 4256
10 1487
11 3689
12 3 7 15 17
13 3 4 15 12
14 -
15 4 12 22 15
x 2.9 6.3 9.9 10.3
Microfauna in the 4 stage in all of the laboratory experiments runs were
classified according to the groups listed in Table 6. Relative populations of
each group were determined. An organism was identified as the dominant (most
numerous) organism in each of the biofilm cultures. The group that this organ-
ism was a representative of was then identified as the dominant (most numerous)
group. The dominant group in each stage was recorded. In some instances more
than one organism was dominant. To investigate the possibility that a particu-
lar group was repeatedly dominant in a stage, Figure 1 was developed. The num-
ber of times a group was dominant was recorded for each stage, then plotted in
Figure 1.
While only two groups were ever dominant in Stage 1 (Free-swimming Ciliates
and Flagellates), all groups exhibited dominance in the latter stages in at
least one of the experimental runs. This indicates that the niche provided in
stage 1 is more specialized and/or restrictive than niches found in later stages.
The second stage was most frequently dominated by populations of attached_
ciliates. In 12 of the 14 runs examined they were either the dominant organ-
isms or shared dominance with another organism while the attached ciliates were
a dominant organism in stage 3 four times and in stage 4 three times their dom-
inance in the laboratory rotating biological contactors was most pronounced in
stage 2. Nematodes were a dominant group in stage 2 in 7 runs.
Rotifers were the group most often predominant in the 3rd stage (8 of 14
runs). They showed large increase in the number of times they were a predom-
inant organism from stage 2 to 3 (from 1 to 8 times). A decrease in dominance
from 8 to 1 times was also observed from stage 3 to 4 cultures.
Attached Ciliates exhibited a similar large decline in dominance from stage
2 to 3. Because they exhibit a large increase, followed by a rapid decrease in
dominance, these two groups of organisms could be used as indicators of RBC pro-
cess kinetics. The environment in stage 3 varied more than that of any other
stage. All groups were dominant in at least 4 of the 16 runs.
Dominance in the 4th stage biofilms was most frequently shown by the Sar-
codinians. This group was the only one that steadily increased the number of
times it was a dominant group going from stage 1 to stage 4. Free-swimming
179
-------
0)
o
e
to
s
o
0>
U
•H
1
P.
3
O
M
60
CO
>$
U
4J
0)
12
11
10
Free-Swimming
Ciliates
^Attached
Ciliates
Sarcodinians
Free-Swimming
Ciliates
Nematodes
Attached
Ciliates
lagellates
Rotifers
RBC STAGE
Figure 1
Dominance ofMicrofauna in
Laboratory RBC Cultures
180
-------
Ciliates and Nematodes were also predominant in some of the 4th stage cultures,
After the dominant groups in a biofilm were determined, COD and NH_—N con-
centrations in the bulk liquid were noted for that particular stage and run.
Population of a group in the biofilm was plotted against concentration of
either COD or NH -N (Figures 2 and 3).
Concentrations of various chemical species are descriptive of the environ-
mental condition the biofilms are exposed to. The development of a relationship
between concentration of a chemical species and mierofauna population would pro-
vide increased understanding of the species succession found in RBC biofilms.
Figures 2 and 3 are graphs of concentration versus population levels of Free-
swimming Ciliates, attached Ciliates, Rotifers and Sarcodinians.
It can be seen from examination of Figure 2, that these are two basic dif-
ferences in the types of areas enclosed for each of the four groups.
First, consider the range in COD values that the groups occurs in. Free
swimming ciliates occur over the broadest range of COD concentrations, followed
by attached ciliates and then Sarcodinians. Rotifers exist over the narrowest
range. Secondly, it appears that free-swimming ciliates occurred in high COD
concentrations than attached ciliates. Rotifers occurred in still a lower COD
concentration. Sarcodinians occurred at the lowest concentration of COD while
there is overlap of the areas in Figure 2, there are also areas unique to some
groups.
Ammonia Nitrogen (NH.,-N) concentration and population levels are related
for the four groups in Figure 3. Similar effects in the concentration ranges
of NH--N and COD were observed. Both free-swimming ciliates and the attached
ciliates occur in wide ranges of NHL-N concentrations. Rotifers and Sarcodin-
ians were restricted to more narrow conditions, especially at higher population
levels.
A similar methodology to the one used in this study, relating COD and NH«-N
concentrations to group population levels, may also be used for individual
species. While this has not been done by these authors, there are indications
that the area enclosed for an individaul organism would be less than that for
a group.
Results From Full Scale Analyses
Microscopic examinations of the Plainville, Connecticut, RBC treatment
plant indicate that the number of species per stage decreased from 17 in the
1st stage to 12 in the 6th stage. Chemical and biological characteristics of
the full scale operations are found in Table 8. The pH and alkalinity concen-
trations are noticeably less than those maintained in the laboratory study.
66% of the nitrification occurred in the 1st stage.
The succession of dominant groups from stage to stage is shown in Figure
4. The peak of attached ciliates followed by a large decline in dominance
from stage 1 to 2 is very similar to that observed in the laboratory unit.
A peak in the dominance of flagellates, similar to that observed in the
laboratory unit was more pronounced in the Full-scale process. The flagellate
181
-------
H
S)
0)
co e
NJ Q|
cd
iH
XX
Sarcodinians Rotifers
x
Free-Swimming
Ciliates
10
20
30
40
50
60
70
80
COD (mg/1)
CONCENTRATION
Figure 2
Effects of GOD concentration on Microfaunal Populations
-------
Sarcodinians
Rotifers
03
W
XX
0)
OJ
c
o
•H
4-1
ca
H
ft
o
CM
Attached Ciliates
AAAA/\AAAA'
|ree Swimming
Ciliates
I
10.0
1 1 1 1 1
20.0 30.0 40.0 50.0 60.0
NH3-N (rag/1)
CONCENTRATION
I |
70.0 80.0
Figure 3
Effectsof NH -N Concentration on Micrpfaunal Populations
-------
Table 8
Chemical and Biological Characteristics
of the Full Scale RBC
STAGE
INFLUENT
1st
2nd
3rd
4th
5th
6th
CONCENTRATION
NKU-N NO-N
J
36.7
9.4
3.86
2.26
1.30
1.67
1.38
0.86
24.4
28.8
30.4
32.6
32.0
36.6
(fflg/D
pH Alkalinity*
7.27
6.83
6.63
6.56
6.51
6.57
6.63
232.0
112.0
70.5
62.0
54.2
57.0
57.0
NUMBER OF
MICROFAUNAL
SPECIES
17
17
13
13 '•
12
12
* Alkalinity expressed as
mg/1 as CACO,,
184
-------
XX _
00
(Jl
fi
o
•H
4J
a
o
PM
Attached
Ciliates
Flagellates
Free
swimming
Ciliates
4-
Rotifers
Sarcodinians
ttached Ciliates
Free-swimming
Ciliates
Flagellates
RBC STAGE
Figure 4
POPULATIONS OF MICROFAUNA IN FULL SCALE RBC
-------
species responsible for this peak was Monas obliqua and Mastigamoeba reptans.
Dominance of Rotifers and Sarcodinians in the full scale biofilms was similar
to that of the laboratory study. However, the slow increases of the Sarcodin-
ians and the rapid increases and decreases of Rotifers was not observed in the
laboratory.
Although there are similarities between the laboratory and Full-scale stu-
dies, no major comparisons should be made. The lack of pH and Alkalinity con-
trol creates a very different niche for organisms and their abundance is very
likely an indication of this. The biofilm cultures nearly had identical zoo-
glea matrices when examined with a microscope. The full scale biofilms were
much thinner and in some instances did not uniformly cover the entire discs.
Conclusions
There is a succession of microfauna from stage to stage in both laboratory
and full scale biofilms. In the first stage filamentous bacteria dominate bio-
films and are responsible for the removal of BOD. Free-swimming ciliates were
the primary predators of the filiamentors bacteria in the first stage of the
laboratory unit. Two genera of filamentous bacteria were either singly or in
combination identified in the first stages of all 16 laboratory runs. A^Zoo-
glea sp. (probably Zooglea ramigera) played an increasely important role in
latter stage biofilms associated with nitrification.
Attached Ciliates were the most frequently dominant microfaunal group (ex-
cluding bacteria) in the 2nd stage-laboratory and 1st stage-full scale units.
Rotifers and Sarcodinians were most frequently the dominant groups in the 3rd
and 4th stage laboratory biofilms and in the 3rd through 6th stages of the full
scale unit.
While there were similarities in the full-scale and laboratory scale re-
sults, nitrogen-mass loading rates, pH and alkalinity regimes were quite dif-
ferent. The decrease of total microfaunal species per stage (excluding bacteria)
in the full-scale unit may have been the result of lower Nitrification reaction
rates.
Microfauna was classified to the level of species whenever possible, but
results in this paper indicate that this may not be necessary in order to assess
the overall microbiological condition of the biofilm. Until there is an ade-
quate data base, classification of microfauna to the level of species should be
attempted whenever possible.
The average number of microfauna per stage (excluding bacteria) increase
with increased degree of treatment in the laboratory study. Some individual
runs did exhibit decreases in the 4th stage cultures.
Population levels of four microfaunal groups were related to concentration
of COD and NHo-N. There were also shifts of microfaunal groups related to chan-
ges in COD and NH -N concentrations. Ranges for many of the microfaunal groups
were quite large. Future investigators should address the relationships between
microfaunal groups or individual species and chemical concentrations in effluents
from RBC reactors. Rotating Biological Contactor researchers should incorporate
whenever possible, microbiological examinations of biofilm cultures, to expand a
minute data base. Hopefully this will provide further insights into the biologi-
cal activities of the organisms responsible for the efficient operations of the
process.
186
-------
References
1. Merkel, G.J., "Observations on the Attachment of Thiotrix to Biological
Surfaces in Activated Sludge" Water Research 9, 10. (1975).
2. Eikelbloom, D.H., "Filamentous Organisms observed in Activated Sludge",
Water Research, 9, 4. (1975).
3. Chudoba, J., Ottova, ?., and Madera, V., "Control of Activated Sludge Fil-
amentous Bulking - I. Effect of the Hydraulic Regime or Degree of Mixing in
an Aeration Tank" Water Research, 7, 8. (1973)
4. Hawkes, H.A., The Ecology of Waste Water Treatment, The MacMillan Company,
New York, (1963).
5. McKinney, R.E., Microbiology For Sanitary Engineers, McGraw-Hill Book
Company, New York (1962) .
6. Bungay, H.R., and Bungay, M.L., "Microbial Interactions in Continuous
Culture" Advances in Applied Microbiology, 10, 269-290 (1968).
7. Curds, C.R., "Interactions Involving Protozoa", In Aquatic Microbiology,
Eds. Skinner, F.A. and Shewam, J.M. Academic Press, New York, (1977)
8. Pirt, S.J. and Bazin, M.J., "Possible Adverse Effect of Protozoa on Ef-
fluent Purification Systems", Nature, 239, 290 (1972).
9. Sudo, R., and Aiba, S., "Massand Monoxenic Culture of Voricella Micro-
stoma Isolated from Activated Sludge", Water Research, 7, 4 (1973).
10. Curds, C.R., and Gockburn, A., "Protozoa in Biological Sewage Treatment
Processes: I. A Survey of the Protozoan Fauna of British Percolating
Filters and Activated Sludge Plants" Water Research 4, 3 (1970).
11. Curds, C.R., and Cockburn, A., "Protozoa in Biological Sewage Treatment
Processes: II Protozoa as Indicators in the Activated Sludge Process"
Water Research 4, 3 (1970).
12. Hoag, G.E., and Hovey, W.H. "Effect of Carbon, Ammonia Nitrogen, Hydraulic
Loading Rates, RPM and Exposed Surface Variations on RBC Performance" Pro-
ceeding of First National Symposium on Rotating Biological Contactor Tech-
nology, February, 1980.
13. Kudo, R.R., PROTOZOOLOGY, 5th edition, Charles C. Thomas, Springfield,
Illinois, (1966).
14. Martin, D., Microfauna Of Biological Filters, Oriel Press Limited, Newcastle
upon Tyne (1968).
15. Bergey's Manual Of Determinate Bacteriology, 8th edition eds. Buchanan,
R. E. and Gibbons, N.E., the Williams and Wilkins Company, Baltimore (1974).
187
-------
Page Intentionally Blank
-------
THE CHARACTERISTICS OF ROTATING BIOLOGICAL CONTACTOR (RBC) SLUDGE
By
C. F. Ouyang
Deputy Chief Engineer, Sewerage Engineering Department
Taipei City Government, Taiwan, China
Introdyc tion
The concept of using natural aerobic microbe in municipal sewage treat-
ment has become an important design factor. Activated sludge process has
been popularly adopted since it useless space, treats large quantity of
sewage quickly and obtained better quality effluent. However, requirement
of high degree management skill and high power consumption are disadvantage.
The RBC process which applies aerobic microbe treatment is become more
important and popular. This process has a buffer character against varia-
tion of load. It generates less sludge and eliminates the bulking and foam-
ing problems which usually cause great trouble in activated sludge process,
and also has no odor and clogging problem which trickling filter process
has. It also requires less power and operation as well.(l)
The article discloses data from a pilot experiment of sludge character-
istics using RBC process as reference in design RBC sludge treatment
facilities.
FACILITIES AND
Experiments were performed with 2 RBC pilot plants, one with 2 stages
and the other 4 stages. The inlet water was municipal sewage which has
been primarily treated and then introduced by pump to the RBC for treat-
ment, or had been pretreated with coarse screen, grit removal and fine
screen and then introduced to RBC for direct treatment. Pilot plants, are
described as follows:
(1) 2-stages Pilot Plant: (Unit: mm)
189
-------
Overall Size: 1550L x 440W x 210H
Inlet: SOL x 440W x 75H
Reaction Tank: 950L x 440W
Gross Capacity: 68.9 liters. Net Capacity: 61.3 liters.
Liquid Volume/Disc Surface Area: 3.64 l/m^
RBC Body: 405 0 x 440W x 2 stages, acrylic board: 2 mm thick.
(30 pieces in 1st stage at 13 mm interval, 23 pieces in
2nd stage at 15.3 mm interval, total surface area is
16.86 m".)
Settling Tank: 500L x 440W x 200H (Net Capacity: 0.044 tn)
Settling Tank Outlet: SOL x 440W x 60H
1BC Rotation Speed: 2-20 rpm
Driving Motor: 0.25 KW
(2) 4-stages Pilot Plant: (Unit: mm)
Overall Size: 2000L x 720W x 560H
Reaction Tank: 1550L x 616W x 265H
RBC Body: 473 0 x 300L x 4 stages (foamed plastic board, 9 pieces per
stage, 36 pieces in total.)
Total RBC Surface Area: 23 m2
Reaction Tank Capacity: 151 liters
Rotation Speed: 13 rpm
Driving Motor: 0.4 KW
BOD REMOVAL
RBC process is a multi-stage continuous treatment, therefore, the
number of stage is an important factor for BOD removal. Fig 1 gives the
experimental results of the rate of BOD removal by a 2-stages and a 4-
stages RBC process respectively under the same load. It shows that the
more the stages are, the better BOD removal efficiency is. Fig 2 indicates
the process of the 1st stage can remove about 651 of BOD and the sequential
stages have lower efficiency. The BOD removal curve demonstrates that
while the number of stages is up to 3 or more, the removal efficiency is
negligible. However, an appropriate number of stages will have buffer
effect on the shock loading of flow and water quality change. Generally
2—4 stages may be recommended.
BOD removal speed has a close relation with the property of raw
wastewater. Fig 3 shows its reaction. It is proceeded in two stages and
its reaction is a first order reaction.
Ct
In — = K11
u0 i
Where C0: Influent BOD (%)
Ce: Remaining BOD (%) after t time
—1 •
K-,: BOD removal speed (hr )
t: reaction time (hrs)
The 1st stage-ireaction has a BOD reduction ranging from 100% to 35%
and K, is 5.0 (hr ). During the 2nd stage reaction, the BOD reduction
is from 35% to 10%, K-[ is 1.75 (hr"^). These are much greater than those
190
-------
IOO
O
Fig.
0.4 O.8 1.2 1.6
Contact tlmt ( hrs)
I . No. of R BC sfo««* and
BOD removal rat*
2. BOO removal rate un*«r
•octi stag«
100
0.5 I.O 1.5 2.0
Contact ttow (hr«)
Flf. 3 BOO r«nm»ol *p*«d
2.5
IO
L2O £
p
3O
40 |
50 |
-60 Q
7O O
8O OD
h 9O
3.O
191
-------
in aetived sludge process. The time of reaction treatment will be complet<-
ed within 2 hours.
According to our long term experiment, the following two assumptions
are made:
(1) Organic load and flow load of SBC are factors which affect BOD removal
rate.
(2) Rotation speed of RBC and G value (Liquid Volume/Disc surface area)
should be of the optimum data selected according to experience.
With reference to the data from continuous operation, relations of
influent wastewater concentration and removal rate under different load
are shown in Fig 4. The result can be written as:
y = 100 - (BL x a ) (1)
where y = BOD removal rate (%)
Ot = the slope of BOD loading and BOD removal rate
Then, common slope of all influent wastewater concentrations are
analyzed:
- . 126-1 (2)
a Gin + 10.07 v '
where Gin = Influent wastewater BOD concentration (mg/l)
Institute (2) into (1), then,
Gin + 10.07
B*L-
'- r
126.1 Gin
Remaining BOD
- 126.1 B.L.
Cin+ 10.07
126.1 Gin H.L. __
10.07) x 1,000
2
z
where B.L. = BOD loading (g BOD/m day)
2
H.L. *= Flow rate (1/m day)
192
-------
CHARACTERISTICS OF BIO-FILM
KBC treatment exposes the disc to air under an appropriate rotation
speed and provides oxygen for reproduction requirement for various microbes,
then, bio-film is grown. Microbes, by means of Exo-enzymes, absorb basic
elements in sewage and perform metabolism. Part of it becomes thermal
energy and the other becomes the film,
Under normal operation, compositions of bio-film in different stages
are as follows: (SS composition in Table 1 and VS composition in Table 2.)
Table 1: Composition of Bio-film
Stage
1
2
3
4
Qty.
(g/m2)
32.6
28.7
28.8
22.4
Cone.
(%)
4.3
4.4
5.8
5.9
Chemical Composition (%)
C
35.2
39.1
37.1
37.2
H
5.4
6.7
5.7
6.0
N
5.8
6.9
6.8
6.8
0
21.4
23.4
25.3
25.5
Ash
32.2
23.7
24.5
24.5
Thermal Value
(Cal/g.vss)
5,790
4,230
4,000
3,010
VS/TS
(*)
67.8
76.3
75.5
75.5
Table 2: Chemical Composition of fS of Bio-film
Stage
1
2
3
4
Organic Composition (%)
C
51.9
51.2
49.1
49.3
H
8.0
8.8
7.5
7.9
N
8.6
9.0
9.0
9.0
0
31.5
31.0
33.4
33.8
Thermal Value
(Cal/g.vss)
8,540
5,540
5,300
3,990
The molecular formula of this composition is C, 0 H_ N-. , 0 .
4.z o.l U.D 2.1
!
Content of volatile substance is 67.8% in the 1st stage and 77-78% in
the 2nd and sequential stages.(2) Ash content in the 1st stage is higher
than other stages. It might be due to the absorption of inorganic matter
in the influent wastewater. Sludge concentration is 4.3% in the 1st or
2nd stage and 5.8% in the 3rd or 4th stage. The latter film is harder due
to the bio-film formed from the growth of nitrobacteria. No significant
difference is occurred in the CHNO ratio among these stages.
i
Regarding the appearance of bio-film, during the 1st stage, because
of the high load, the sludge is in light dark color and the micro-organism
appears are merely Sphaerotilus, Zooglea and other filamentous bacteria.
During the 2nd stage, the film is in tan or brown color with slurry; the
protozoa therein is Rotaria, Diplogaster, Zoothonmium etc. During the 3rd
stage, the film is harder then previous one and it is in tan color with
Rotaria etc. Microbes therein are less than that in the 2nd stage. The
film in the 4th stage is similar to that in the 3rd stage. There are only
a few microbes, such as Podophrya etc. Since microbes appear at the RBC
film is much more than that in activated sludge process, the ecosystem is
more stable, and is more resistant to the variation of influent wastewater
193
-------
too
90--
Influtnf
4>
O
O
O
m
80
70
5 IO 15 2O 25
BOD Surface area loading (g/mldoy)
Fig. 4. relation of BOD removal rate with influent
BOD and surface area loading
7O
*••»
**
\
O»
I 6°
V
•o
j= 50
^
I
ro
40
• TO 20 3O 4O 5O 6O 7O
BOD Surface area loading (9/ml-day)
Fig. 5. relation of BOD surface area loading
and bio-film dry weight ( 4 stages).
194
-------
quality. This is one of the reason why this method is easier to operate.
In the experiment of pilot plant, the thickness and the dry weight of
the bio-film due to the BOD surface loading (see Fig 5,) vary at each stage,
due to the rotating speed. Fig 6 shows the relationship between the thick-
ness of the bio-film and rotating speed.
BBC is relying on the bio-film to treat wastewater. The MLSS in the
activated sludge aeration tank is generally considered to be equivalent to
the sum of the microorganism and SS in the RBC divided by the net capacity
of the contact tank. This is defined as equivalent SS (ESS), which is a
function of removing the organic matter in the wastewater. Thus it can be
written as
F/M = B.L./DS
2
Where B.L.: BOD surface area loading (g.BOD/m .day)
2
DS: Dry sludge weight (g.SS/m )
F : BOD (mg/1)
M : MLSS (mg/1)
In 4-stages pilet plant experiment, the first stage ESS is 10,000 -
13,500 mg/1, and decreases stage by stage. The 4th stage ESS is about
4,000 - 8,500 mg/1. The average ESS is 7,000 - 11,500 mg/1 for the whole
process. The average ESS for the two stage pilot plant under different
rotating speed is shown in table 3. It is found that the average ESS for
the 2-stages experiment is approximately the same as that for 4 stages
experiment, specially under the same speed where the average ESS is 9,000
mg/1 for 2-stages experiment.
Table 3: Average SS and ESS of 2-stages experiment
Rotating
speed
(KPM)
4
7
10
15
Peripheral
velocity
(m/min)
6.5
9.9
14.1
21.2
Average SS in
tank (mg/1)
Stage 1
466
370
327
294
Stage 2
263
100
78
76
Average bio— film
dry weight (g/m^)
Stage 1
66.3
60.7
54.8
28.4
Stage 2
30.9
23.9
22.6
21.6
ESS
(mg/1)
13,940
12,230
11,170
6,960
ESS and BOD relation under the aerobic conditions vary, depending on
the loading. From the experiment, no definite relation between F/M loading
and removed rate at each stage can be evaluated. The BOD removal in the
experiment can be up to 50% at the 1st stage and it differs substantial at
each stage. This may be due to the reason that very high adhesion of the
SS on the disc at the 1st stage experiment result on the reduction of non-
organic SS in the wastei Fig 7 show in the relationship between average
F/M and BOD removal rate.
Table 4 described the relation between the F/M loading and BOD removal
for the 4-stages experiment. Because the aeration time for the activated
sludge process is about 4 times of the RBC process, the F/M ratio of 0.2 -
195
-------
1.5
E
E
~ I.O
I 0.5
^-
i
_o
m
Thickness
Weight
I st stage
7
rotating
IO
speed (rpm)
15
SO
SO
4O
JO
CD
2O
6.5 9.9 14.1 21.2
Peripheral velocity ( m / min )
1.0
E
E
u
JO
CD
2 nd stage
Thickness
7 K>
rotating speed (rpm)
IS
60 E
40
> 6.5 a9 14.1
Peripheral velocity (m/min)
Fig. €. relation of peripheral velocity with Bio-film
thicliness ojid dry weigfit
196
-------
IOO
o
£
§ 9O
o
a] so
70
O,l Q2 O.3 O.4 as O£
F/M loading (gBOD/gEss-day)
Fig. 7 average F/M loading and BODremool rate
10
TO
S3 3O 4O 5O 6O
Setting Mm* (min)
Fig. 8 Settling «eed of the RBC effluent water
SO 9O IOO IIO 12O
197
-------
0.4 g.BOD/g.ESS.day for the RBC process can be considered as the same as
that for the activated sludge process.
Table 4: Aerobic digestion of RBC sludge
\.
days ^s.
raw sludge
5
10
15
T.S.
(mg/D
12,630
11,000
6,300
5,800
V.S.
(mg/1)
9,230
7,500
4,010
3,330
V.S. reduce
rate (%)
18.6
56.4
63.8
BOD
(mg/1)
11,400
6,000
2,900
1,200
SV30
(%)
98.0
95.0
73.0
40.0
Supernatant
BOD (mg/1)
255
140
110
SETTLING AND THICKENING CHARACTERISTICS OF SLUDGE
Suspended solid in RBC reaction tank is merely down-scaled bio-film
and a few SS of influent wastewater. Concentration of SS in the tank
varies upon the concentration of influent wastewater and rotating speed.
It increases while the load is high and thickness of bio-film is increasing.
When the load is lowering, the sloughing of bio-film will be also increased.
Removed sludge radiates outwards with the Zoogloea sludge as a center
a great amount of sponge-like bacteria filament. Under microscope they are
branch-like suspension with lower specific gravity but larger diameter. The
diameter found is about 0.1 mm -1.2 mm.
Generally, SS in the final reaction tank is less than 200 mg/1, the
solid loading in final settling tank is about one-tenth of that in activated
sludge process. In the measurement at the overflow weir of reaction tank,
SS is about 100-150 mg/1, the SV30 is about 1-1.5%, hence, SVI is about 100.
The settling character is excellent.
Settling speed varies upon the difference of raw wastewater concentration
However, large SS sinks quickly and freely within the first 10 minutes. The
curve in Fig 8 shows the initial sinking speed. If the outlet port has an
average SS of 150 mg/1 and SV3Q is 1.5%, fihen, after a settling for 30
minutes, the sludge concentration is about 10,000 mg/1.
The lowest settling speed of slough off bio-film as mentioned is about
6-15 cm/min, i.e. 80-216 m/day. If the surface loading of a clarifier is
under 30 m^/m2.day, it is enough to settle the flow and separate it from
water.
Sludge from final sedimentation tank is withdrawn to a 1,000 ml
cylinder. The interface is measured at definite time intervals. The
interface'change with various sludge concentrations are shown in Fig 9.
Thickening speed lowers when concentration increases. After keeping still
for 6 hours, the thickening nearly terminated.
Since SS is high, the interface of the liquid is obvious. In the
beginning, the zone falls with a constant speed. However, the lower
concentrated layer is thickened and the sinking speed at the middle is
lowered. Thereafter, by the accumulated weight, the inter water being
198
-------
SS(mQ/«)
925
10,070
11,910
9,030
8,000
(9,770)
8,460
4,540
5,540
4,970
0.51 246 12
Thickening time (hrs)
Fig 9. Volumt occupied by tht thickend sludge with thickening time
-------
pressed out. Settling speed becomes very slow (compression zone). The
thickening character is identical to that of activated sludge process and
in a layer deposit.
PRODUCTION OF SLUDGE
Treatment of sludge is the most difficult problem in future sewage
treatment. We have researched the sludge production by directly treat the
sewage which is merely through grit removal and fine screen and the sludge
production from presettled sewage. The actual sludge production from
direct treatment and pretreatment is respectively shown in Fig 10. In
direct treatment has a wider range of 0.37 - 1.15 Kg.SS/Kg.BOD.removal.day,
average is 0.67 KgSS/KgBOD removed.day. Thus indicates that the conversion
rate between BOD removal and sludge is very low. For secondary treatment,
depending on the flow rate, the sludge production rate is 0.34 - 0.55 Kg.SS
/Kg.BOD removal-day is about one third of that from activated sludge
process, which is considered as the major advantage of this process.
SLUDGE DIGESTION AND DEWATERING
The sludge generated from the direct treatment by the RBC process, has
a water content of about 98.8% with the average quality of SS about 9,200
mg/1. The VSS has been 18% for the duration of 5 days and 56% for 10 days
respectively, if the aerobic digestion is applier under the room temperature
of 15 - 20°C. This result indicates that the digestion due to the RBC
process has been more efficient than that of the primary plus activated
sludges.(3) pig 11 shows the relationship between the VSS reduction rate
and the aerobic digestion duration. The digestion characteristics due to
the RBC sludge with the aerobic digestion are also given in Table 4.
Further, the raw sludge obtained from the RBC process went through the
thickening process. It was found that the sludge has a water content of
97.5%. The sludge then was proceeded by anaerobic digestion with a temper-
ature of about 36 C and under the room temperature of 18 - 20 C. Table 5
shows the characteristics of anaerobic digestion. After 30 day the VS
reduction rate is approximately 38% and the sludge alkality is about 1,500-
2,400 mg/1.
Table 5: Anaerobic digestion of RBC sludge
Sludge ^^^-^^^
Raw sludge
36°C
Digestion
sludge
19°C
Digestion
sludge
15 days
30 days
15 days
30days
S.S.
(mg/1)
23,720
18,460
16,600
19,270
18,230
BOD
(mg/1)
25,000
9,200
4,800
11,600
5,400
T.S.
(mg/1)
25 , 300
17,750
18,500
20,610
17,460
V.S.
(mg/D
14,500
11,200
8,900
13,150
9,290
Alk.
(mg/1)
1,580
2,430
1,195
1,440
Water
cont.(%)
97.40
98.20
98.26
97.93
98.35
The raw sludge produceed by the RBC paeaft,€.i^S',then, is dewatered
after the anaerobic digestion for the duration of 15 and 30 days respec-
tively. The dewatering characteristics of the obtained sludge cake can
200
-------
z i.o
o
a 0.9
x OB
v>
06
05
direct treatment
Secondary
treatment
6O 7O 8O 9O IOO
BOO removal rate (%)
Fig. IO. BOO removal rate and sludge production
7O
60
o 4O
m
10
activate sludge
5 10 15
Reaction time (doy»)
I I relationship of VSS removal rate and
aerobic digestion reacton time
201
-------
be compared with other of the sludge- as shown in Table 6.(2) It appears
that the specific resistance of the dewatered sludge is lower than those
of the primary plus activated sludge as well as those of the activated
sludge. The water content of the obtained sludge cake is also lower.
Table 6: The characteristics of
compared with others
RBC sludge dewatering and
Sludge
RBC direct
treatment
sludge
Primary plus
activated
sludge
activated
sludge
raw sludge
15 days
digestion
30 days
digestion
Feeding
FeCl3(%)
5
10
5
10
5
10
5
10
5
10
Water cont.
of Sludge cake
84.9
78.8
83.1
68.3
80.5
65.0
79.8
83.1
82.3
86.1
Specific resistance
R x 107(sec2/g)
18.4
12.8
19.6
12.6
25.8
10.8
27.0
27.0 >'
16.8
15.2
CONCLUSIONS
1. The speed of BOD removal in the RBC treatment is a 1st order reaction.
The process can be proceeded into two stages. The BOD remaining rate
rangjd from 100% to 35%, Kj = 5.0 (hr"1) and from 35% to 10% KJ = 1.75
(hr ) in the 1st and 2nd stage respectively.
2. The thickness of the Bio-film was grown depending on the sewage loading
and the rotating speed. Under the two stages process and the peripheral
velocity under 14.1 m/min, the dry weight of bio-film are 54.79 g/m2 and
22.56 g/m2 for the 1st and 2nd stages respectively. When the peripheral
velocity is high than 21.2 m/min, the dry weight of the bio-film is only
28.41 g/m2 for the first stage and 21.56 g/m2 for the 2nd stage.
3. The ESS in the contact tank varies depending on the load. Under the 4-
stages process of the RBC treatment, the ESS is 6.970 - 11,460 mg/1 with
the average of 9,000 mg/1, if the BOD loading is under 10 - 641 g.BOD/m2.
day. For the 2-stages process, the ESS has an average of 11,000 mg/1 and
has a concentration about 3-4 times of that of MLSS from the convential
activated sludge process. Hence, the contact time by the RBC process in
only below % of that of the activated sludge process. This may me the
reason where the RBC is more efficient.
4. Generally, the larger the F/M loading is, the lower the BOD removal. The
F/M loading of the RBC process is about 0.2 - 0.4 g.BOD/g.ESS.day, which
is approximately equal to that of the conventional activated-sludge
process.
5. The bio-film from the RBC process has an average VSS of about 74% and a
water content of 95%. Its chemical composition is C, _H0 N~ ,0». The
4-./ o U.o 2.
202
-------
thermal value is about 8,500 cal/g.SS in the 1st stage, which is
comparetively higher than that in the 2nd-4th stage (about 5,200 cal/g.
SS).
6. The sludge has settled freely and relatively fast. After 10 minutes,
the settling almost terminated, because, only limit amount of the settled
sludge was found. The early settling of the sludge has a velocity be-
tween 6-15 cm/min, which is faster than that due to the activated
sludge process. A surface loading of 30 m-Ym^.day for the clarifier
design is recommended of a rate of less than 20 mg/1 for the effluent SS
is required.
7. The sludge production under the BBC secondary treatment process varies
depending on the influent loading and its BOD, The average sludge
production rate is approximately equal to 0.54 Kg.SS/Kg.BOD.removal.
The sludge production due to the direct treatment process has a wider
range of 0.37 - 1.15 Kg.SS/Kg.BOD.removal (average is 0.67 Kg.SS/Kg.BOD.
removal).
8. The VSS reduction rate can reach up to 56% for a duration of 10 days if
the RBC direct treatment sludge using aerobic digestion applied.
Consequently, it seems to have better digestion than conventional
activated sludge.
9. After the thickening of the sludge from the RBC direct treatment, the
raw sludge and anaerobic digested sludge is dewatered by feeding FeClo.
Its specific resistance is less than those of primary plus activated
sludge, and activated sludge.
ACKNOWLEDGE
The author wishes to thank Professor W. F. Yaa&« . Environment
Engineering Institute of National Taiwan University, for all the help and
guidance in his capacity as supervisor of this research.
Dr. C. S. Chou should also be thanked for his recommendation for the
establishment of the pilot plant. The author is also indebted to Dr. T.
Matsuo, Tokyo University and Dr. Okuno, for their valuable comments.
Thanks are also given to Mr. W. K. Liu, Director of the Sewerage
Engineering Department, Taipei City Government, for his support and
encouragement.
Thank is also given to Executive Yan National Science Council for
financing this research.
REFERENCE
(1) John W. Clark; Warren Viessman, Jr; Mark J. Hammer; Water Supply and
Pollution Contron. P.558, Third Edition. Harper & Row, Publishers.
(2) Malhotra, S. K., T. C, Williams and W. L. Morley: Performance of a
Bio-Disk Plant in a Nothern Michigan Community Prepared For: Present-
ation at the 1975 WPCF Conference at Miami Beaach Florida, October
5-10 (1975).
203
-------
(3) R. E. Mckinney, J. E. Symons: Advances in Water Pollution Research, 2,
440. Pergamon Press (1964).
(4) Ronald L. Antoine: Fixed Biological Surface Wastewater Treatment.
(5) U. S. EPA, Water Pollution Research Ser. 17050 DAM 11/71. (1971):
Application of Rotating Disc Process to Municipal Waste Water
Treatment.
204
-------
PART III: MUNICIPAL WASTEWATER TREATMENT
DATA EVALUATION OF A MUNICIPAL RBC
INSTALLATION, KIRKSV1LLE, MISSOURI
1 By
R. Ryan Dupont
Graduate Assistant, Department of Civil Engineering
Ross E. McKinney
N.T. Veatch Professor of Environmental Engineering
University of Kansas
Lawrence, Kansas
INTRODUCTION
The Rotating Biological Contactor System, RBC>is becoming more and more
competitive with the conventional fixed film and suspended growth systems
because of its simplicity, ease of operation, and low operation and main-
tenance costs. Design of RBC systems has been based primarily on surface
hydraulic loading, HLA, as gal/d/sq ft of media surface, and only recently
have the manufacturers suggested BOD influent concentration as having some
effect on system performance. Independant researchers have considered both
organic and hydraulic loading in analyzing significant parameters and results
have been somewhat contradictory. Some researchers have suggested system
performance was strictly a function of HLA and some investigators have said
surface organic loading, OLA, as Ib BOD/d/1000 sq ft of media surface was
the controlling parameter. These Initial analyses were conducted using pilot
plant results and suffered from a severe lack of long-term operational data
from full scale RBC installations. The research'conducted at the University
of Kansas and summarized in this paper was aimed at clarifing some of the
inconsistencies reported In the literature. Full scale operational data
from a 5 MGD RBC plant at Kirksville, Missouri was analyzed to determine
the parameters that had a significant effect on system performance, to check
field results against design theory, and to make recommendations on parameters
that should be used for future design and analysis of the RBC system.
205
-------
EFINJTION OF TERMS
Many different terms have been used In the analysis of RBC process
fficlency to describe basically the same parameters. Shown in Figure 1
urface hydraulic loading, HLA, as gal/d/sq ft is determined simply from
he flowrate in MGD divided by total media surface area in million sq ft.
he HLA was found to have common units by all authors reporting hydraulic
oadlng data. The second parameter of interest is organic surface loading.
LA, as Ib BOD or COD/d/1000 sq ft media surface area. As seen in Figure 1,
rganic concentration and flowrate or hydraulic loading have an effect ton
LA. The significant characteristic of OLA is that it not only reflects
he driving force for the diffusion of substrate Into the biofilm through
he Influent BOD concentration term but also the effect of reaction time
i the system as indicated by the hydraulic loading term.
FIGURE 1 VARIABLES USED IN DATA ANALYSIS
fDRAULIC SURFACE LOADING— HLA
HLA = GAL/DAY/SQ FT OF MEDIA SURFACE AREA.
« Q in MGD
SA In mill ion sq ft
16ANIC SURFACE LOADING— OLA
OLA = LB BOD or COD/DAY/1000 SQ FT OF MEDIA SURFACE AREA
~ (Q- ' n MGD)* (BOD I nf1uent concent ra t Ion I nmg/1) *8.
SA in 1000 sq ft
Correspondingly:
OLA = (BOD i nf 1 u en t con centrati on In mg /1_) *(H LA) * 8.3 k
1000
Many investigators have presented OLA as Ib BOD/d/1000 cu ft media
ilume. The RBC system is a fixed film process and as such, microbial
tlvity is concentrated at the media surfaces. It seems more theoreti-
lly sound to be concerned with loading on a surface basis rather than on
volumetric basis and all data in this analysis Is presented In this manner.
tal organic loading, OL, and BOD removal both as Ib BOD/d, were used for
oss comparlsions between specific periods within the overall study period.
rcent BOD and SS removal values were also used in the analysis for com-
rlson of Kirksvllle data with data in the literature.
EVIOUS STUDIES
Initial use of rotating media for treatment of wastewater was by Doman
) In 1925. His experimental "contact filter" consisted of concentrically
unted 20 gage steel discs, 1 inch on centers, that was rotated i rpm in
e effluent from a septic tank. He used the discs to accumulate "active
robic material" for removal of colloids from the wastewater. The biological
ture of the process was not appreciated by Doman and the system was doomed
failure. Anaerobic conditions plagued the system's operation and poor
206
-------
results caused abandonment of the concept for almost 30 years. The next use
of rotating media for wastewater treatment came In the late 1950's with the
investigations of Hartman and Popel (2) at the University of Stuttgart. One
meter diameter plastic discs were used as support media and tests were run
to gather information for process development of what they called "Immersion
Drip Filter," These investigations, along with the development of expanded
polystrene as an inexpensive construction material, led to commercial use
of the "Immersion Drip Filter" in Europe as early as 1960.
Al1is-Chalmers developed the RBC in the United States independently
from European efforts in the late 1960's and upon learning of European
activities, came to a licensing agreement with J. Conrad Steneglin Company
of Tuttlinger, West Germany for sales and manufacturing rights for the RBC
process in the United States under the trade name BIO-DISC (2), The process
was subsequently bought by Autotrol Corporation in 1970 and has been marketed
since then under the trade name BIO-SURF.
Most early process performance analyses were carried out by manufacturers
and it has been only since about 1970 that results of independent investigation
have been readily available. Table 1 is a summary of important RBC performance
analysis results and design recommendations found in the literature. Hartman's
study in 1965 (3) was the first presentation of modern RBC design criteria. He
recognized the effect of influent BOD concentration on BOD removal thropgh
"decomposition curves" he presented. These "decomposition curves" represented
percent BOD removal versus 1/HLA for varying wastewater strengths ranging from
less than 100/mg/1 to over 600 mg/1 influent BOD. He also recommended a
maximum OLA of 20.5 1b BOD/d/1000 sq ft to the first stage of the RBC for
efficient process performance.
Welch and Antonie presented manufacturer's pilot plant results in their
studies of synthetic waste from 1968 to 1970. Welch (4) investigated both
hydraulic and organic loading variables in his early studies but design and
operational analyses, based on organic loading, were abandoned by manufact-
urers in the later studies in lieu of HLA or detention time in the reactor.
All of the studies showed an initial rapid increase in percent removal with
an increase in detention time, leveling off to a constant percent removal
between 30 and 90 minutes depending on the waste type used in the study.
Welch found a constant slope for BOD removal versus OLA, both as lb/d/1000
cu ft, to an OLA of 650 lb/d/1000 cu ft, beyond which a constant BOD removal
as lb/d/1000 cu ft was achieved.
After 1971 studies began to look into the relationship between loading
and removal on a pound basis. Pescod and Nair (5). Cochran, Burn, and Dostal
(6), and Stover and KIncannon (7) all showed constant rates of removal of
organics on a pound basis for loadings up to 250 Ib COD/d/1000 cu ft for
bottling waste, 10.6 Ib BOD/d/1000 sq ft for cannery waste, and 3.48
Ib COD/d/1000 sq ft for synthetic sucrose waste, respectively.
The study by Malhorta, Williams, and Morley (13) in 1975 is the only
presentation of full scale, long-term RBC plant data found In the literature.
A large amount of average monthly data was provided, but no attempt was made
to quantify performance in terms of graphs or correlations between significant
parameters. In analyzing the data it was observed that once the plant estab-
lished its microbial population, organic removal was more a function of
207
-------
RELATIONSHIPS
SHOD rcnovil vs 1/HLA
800 reoovd vs OLA
TAmr i Rfvirn or unc rrncirncY AHAIYSIS simiirs
WASTE RESULTS RECOMMENDATIONS
SCOO reoowl vs t^
JCOD removal vs tj,
JBOD removal vs HU
HLA vs XBOQ rtnoval
SCOO removal vs OLA
l/d/KCF
S80D reooval vi HLA
800 rceoval vs OLA
f/d/KSF
BOD effluent vs OLA
J/d/KSF
variable
synthetic
synthetic
synthetic
500 mg/1
primary eff
170 rng/1
anaerobic
packinghse
waste
bottling
1000 s>g/l
puniciple
400 ag/1
winery
1003 mg/1
cannery
varying Q,
BOO, SS
primary eff
varying BOD,
Q, SS
decomposition curves to yltld
SA. requirements for different
Ml with different Influent BQO's
u§e of decomposition curves
for design, max OLA to first
stage of 20.51/d/KSF
constant slope to 650 Ib/d/KCF
then pounds removed/d/KCF is constant
decreasing COD influent yield
higher constant rate of removal.
all level off at 30 nin
logrithnlcally approach SOS
removal at 60 mln.tn
approach 9Qi removal at approx.
90 rain
decreasing BOD 1nf 206-145 mg/1
through the study. 652 removal
HLA=8ga1/d/SF. 83* removal HLA=
4 gal/d/SF. 95* of total removal
by first stage
951 removal 8250 */d/KSF
85 X removal P3Q!H/d/KSF
95! removaieHLA'0,75 gal/d/SF
(OLA*6.31 Ib/d/KSF)
constant rate of removal for
QLA=3-10.6 f/d/KSF
increased BOO effluent for
increased OLA
t_ 60 mln for 90*3. removs!
tn"50 min for 20 rog/1 SS and
0 800 eff, 80S NH3-N reiwaval
slnnle staje unit »t an HLA
of 4gal/d/KSF
OLA=250 l/d/KSF for bottling
waste
HLA=0.75 gal/d/SF design
OLA-1 l/d/KSF for BOD eff
quality of 15 mg/1
1COD roitov*! vs COD inf
Full scale plant data
synthetic constant HLA=0.5 gal/d/SF, constant
sucrose rate of COD removal5 90*P 1.9x and
236 ng/1 3.6x initial COO 'inf. peak OLA of
3.48 l/d/KSF
immlciple OLA-1.11-2.41 IBOD /d/KSF
J HSO HLA-0.97-2.14 gal/d/SF
removal a function of temperature
T- 45-60 F 86-93S removal KSO F
94-971 removal T>SO F
XBOO renoval vs no stages synthetic 95S removal through the unit «ith same
carbohydrate OLA from 3 HLA and BOD inf combinations
BOO 1nf HLA OLA
550 0.?S 3.13
8SO 0.50 3.54
1870 0.25 3.48
COO removal vs OLA
l/d/KSF
ICOD renova! vs OLA
slaughterhse 701 COO removal for OLA=3-7 */d/KSF
4.7 l/d/KSF removed at OLA 7l/d/KSF'
carbohydrate 92S COO removal for OLA=6.7-26.7l/d/KSF
logrithmic decrease of % removal to 40X
at 47 f/d/KSF
AUTKOR(S)
Hartnan
(1965)
Welch
(1968)
Welch
(1968)
Antonte
(1S70)
Antonie
(1970)
Chlttenden
Wells
0971)
Pescod S
Sair
(1971)
Labella et al
(1972)
Cathran,Burn
Postal
(1973)
AM berg i
Kwong
(1974)
Stover &
Kincannon
(1975)
Halhorta,
Uilliams,
Herley
(1975)
Stover &
Kincannon
(1976)
Stover £
Kincannon
(1976)
Stover I
Kincannon
'(1976)
208
-------
temperature than organic or hydraulic loading. Resultant removal efficiencies
were 86 to 93 percent BOD removal for temperatures below 50°F and Bk to 97
percent BOD removal for temperatures above 50°F. The overall temperature
range was k$ to 68°F while OLA varied from 1.11 to 2,41 Ib BOD/d/1000 sq ft
and HLA varied from 0.97 to 2.\k gal/d/sq ft.
Stover and Kincannon's 1976 study (14) of carbohydrate and slaughterhouse
waste treatment with RBC's Is significant in that It showed definitively that
percent COD removal is not strictly a function of HLA or influent COD concen-
tration, but is dependent on the combination of the two factors, namely the
total organic loading to the system. Not only was the driving force for
diffusion into the mlcrobial film, as given by COD concentration, an important
variable affecting the performance of the RBC unit, but also the ability for
the btomass to metabolize the waste in a given detention time, as reflected
by HLA, was important. Influent COD concentrations ranged from 550 to 1670
mg/1 and HLA values ranged from 0.75 to 0.25 gal/d/ sq ft to produce OLA
values ranging from 3.4 to 3,5 Ib COD/1000 sq ft. Resultant percent COD
removals varied only between 94 and 96 percent and indicated a definite
relation between OLA and COD removal efficiency as shown in Figure 2. De-
creased COD removal efficiency resulting from organic overloading was also
shown in their results. In Figure 3 their plot of COD removal versus COD
applied as lb/d/1000 sq ft shows a constant rate of removal of COD up to an
OLA of 7 lb/d/1000 sq ft, beyond this a constant mass of COD of 4.7 lb/d/1000
sq ft was removed, causing a decrease in removal efficiency expressed as a
percent as shown in Figure 4. The effect of varying wastewater character-
istics is also evident from Figure 4 from the relative removal efficiencies
for the carbohydrate and the slaughterhouse wastes used in the study. The
soluble, readily degradable synthetic carbohydrate waste showed 90 percent
COD removal up to 26.7 lb/d/1000 sq ft before the constant COD removal
situation developed, resulting in decreased percent COD removal. The
slaughterhouse waste, on the other hand, showed a maximum 70 percent removal
with decreasing efficiency beginning at only 7 lb/d/1000 sq ft because of
the large amount of fats, oils, and slowly degradable organic materials in
this wastewater.
Several other studies were important in developing a conceptual back-
ground for the understanding of the microbiological relationships within the
RBC biofilm. The discussion of substrate utilization in relation to oxygen
and substrate diffusion in the RBC biofilm by Hoehn and Ray (15) in 1973 was
based on work first presented by Sanders (16) in 1964 and by Kornegay and
Andrews (17) in 1968 and with RBC pilot plant results described above serve
as the basis for Figure 5- Three theoretical limitation regions and the
limiting parameter in each are-identified for BOD removal as percent and
lb/d/1000 sq ft versus influent BOD or OLA. Initially substrate is limiting
because of diffusion limitation into the biofilm. Once a minimum load or
concentration is reached the driving force for diffusion into the biofilm
is such that metabolism and corresponding percent substrate removal becomes
independent of substrate concentration and is limited by the mass of sub-
strate reaching the organisms on the RBC discs. More substrate would be
removed as more is applied as shown in Figure 5b. In this region the mass
of organisms on the discs increases proportionally to the substrate applied
as neither oxygen nor substrate are diffusion limited within the biofilm.
209
-------
a COOSMWWW sa ft
Figure 2. Relationship of A COO applied (!bs./day/1OOO
sq.ft.) with A COO removed (Ibs./day/lOOO sq.ft.) for
slaughterhouse wastewater at various applied organic
loadings.
i .
Figure 3. Comparison of percent » COD removal versus
applied organic loading to the RBC for carbohydrate and
slaughterhouse wastewatera.
Figure 4. Percent A COO remaining
per stage with carbohydrate waste-
water for -various organic concentra-
tions and various flow rates (all re-
sulting in the same total applied or-
ganic loading).
(From Stover and Kincannon, 19?6 (14), Water and Sewage WbHcs)
210
-------
% BOD
REMOVAL
A=SUBSTRATE DIFFUSION LIMITATION
B=SUBSTRATE MASS LIMITATION
C=OXYGEN LIMITATION
5a. BOD INFLUENT or OLA
BOD REMOVAL
LB/D/1000 SQ. FT
A=SUBSTRATE DIFFUSION LIMITATION
B=SUBSTRATE MASS LIMITATION
C=OXYGEN LIMITATION
5b. BOD INFLUENT or OLA
FIGURE 5 THEORETICAL LIMITING REGIONS WITHIN THE RBC BIOFILM
211
-------
Higher organic loadings cause the biofilm to grow to a depth where oxygen
diffusion within it is limited due to oxygen depletion from mlcrobial meta-
bolism of substrate within the biofilm. An active aerobic layer at the
biofllm surface is established having a constant depth independent of total
biofilm depth, and removes a constant mass of substrate independent of
substrate loading. The removal of a constant mass of substrate produces.
the logrlthmic decrease in percent removal of BOD in this oxygen limiting
region as shown In Figure 5a.
From this brief review of the literature it becomes obvious that analyses
and results from past RBC work have been all but uniform and results have not
been generally comparable nor compatible. They do reflect, however, the
importance of influent BOD concentration and HLA, used together as OLA, to
describe RBC performance.
PUNT DESCRIPTION
The Kirksvl1le,Missouri wastewater treatment plant was the first full-
scale municipal RBC plant to be built in Missouri and serves a population of
roughly 20,000. It went JntX) operation in June, 1976, as part of a $7
million sewerage system improvement program. Two trickling filter plants
and 12 single-celled lagoons were replaced by the RBC system. A schematic
flow diagram of the Kirksville plant is shown In Figure 6.
Flow entered the plant through a 36 inch influent line and was detained
in a wet well where two, 66 inch diameter, screw pumps lifted the sewage to
a level where it could flow by gravity through the rest of the plant. An
emergency holding basin was provided with a capacity of 2.1 KG to prevent
bypassing of excess flows. From the screw pumps the flow passed through a
grit chamber and comminutor/bar- rack structure before being split and
entering the two primary clarifiers. Effluent from the primary tanks was
split again before it entered the four flow paths of the RBC media. Once
through the RBC tanks, the flow was combined and split to flow Into two
secondary clarifiers. Clarified wastewater flowed through the chlorine
contact chamber, through a Parshall flume for measurement, and into an
adjacent stream.
Sludge from the secondary clarifier was recycled to the front of the
plant via a 10 inch return line to the wet well. It was settled in the
primary tank along with primary solids and the combined sludge was pumped
from the primary clarifier to a two stage digester. Digested sludge was
then disposed of primarily as liquid for application to farm lands via a
tank truck, A storage basin for digested sludge was also provided for
winter months when land application was not possible. Vacuum filters were
provided as well for dewatering of raw or digested sludge for final dis-
posal in the city sanitary landfill.
The rotating biological media and accompanying equipment was provided by
the Autotrol Corporation. Four flowpaths of five shafts each were used for
this 5 MGD plant and each shaft of media consisted of 11 foot diameter rings
of corrugated polyethylene media, concentrically mounted on a square steel
shaft 24 inches on a side. Shafts were 25 feet in overall length and were
212
-------
WHUTHUCTUHtNOt
BUMHM.I. Tuff.
to
GA3 MIKIMC
JL
FIGURE 6 KIRKSVILLE WASTEWATER TREATMENT PLANT SCHEMATIC FLOW DIAGRAM - SECTION
-------
supported at their ends and rotated with approximately 40 percent of their
area submerged in wastewater. Flow entered the trapezoidal concrete reactor
chambers perpendicular to the shafts and an overflow weir separated each
compartment so short circuiting was prevented and each isolated tank approxi-
mated a completely mixed reactor. Each of the 20 shafts was rotated by its .
own drive unit rated at a maximum 7i HP. Table 2 shows pertinent equipment
and design specifications for the Kirksville RBC plant. No information was
readily available for design values for influent BOD and SS so average values
of 200 mg/1 were assumed for each in the table.
TABLE 2 RBC SYSTEM DESIGN SPECIFICATIONS
CLARIFIERS:
MFC
Type
SWD
Feed
Overflow
Sludge Withdrawl
RBC UNITS:
INFLUENT:
EFFLUENT:
MFC
Type
Configuration
Rotator Motor HP
Surface Area
Shaft
Total
Tank Construction
Cover
Mode of Operation
0_, MGD
8005. mg/1
SS. mg/1 Q
Min Temp. F
R.T., Hrs.
HLA, gal/d/sq ft
Equiv MLVSS, mg/1
Break HP/Shaft
Break HP Total
Clairifier SOR
gal/d/sq ft
BOD5, mg/1
SS, mg/1
Smith £ Loveless
2 primary, 2 secondary, 70' Diameter
10'
Center
Peripheral Weir
Mechanical Scraper, Full
Skimming
Surface
AUTOTROL, Inc.
B10SURF
20 Shafts, *t Paths § 5
20 & 7.5
Shafts
95,400 sq ft
1,900,000 sq ft
Trapezoidal, One Interstage Baffle/Shaft
Bui Id ing
Flow Perpendicular to Shafts
Design
5.0
(200)
(200)
45
1.25
2.6
15,000
3
60
650
30
30
1600
30
30
214
-------
DATA PRESENTATION
Data used in this evaluation were from the monthly reports of the
Kirksville, Missouri WWTP for the period extending from January 1977 to
November 1978. Two monthly reports were not available, June and July 1977,
and were omitted from the data analysis. The data included flow, influent
and effluent BOD, influent and effluent SS, and water temperature. Flow
and temperature data were generally measured on a daily basis, while BOD and
SS were initially determined on a daily basis until mid-March 1978 when they
were measured on an average of three times a week.
Derived parameters from the available data included HLA as gal/d/sq ft,
OLA as Ib BOD/d/1000 sq ft, BOD removal as Ib BOD/d/1000 sq ft, percent BOD
removal and percent SS removal. Unfortunately no primary, secondary or in-
dividual stage measurements of these parameters were made and analysis of
the treatment efficiency of the RBC in isolation was not possible. System
evaluation then was based on overall plant performance in the removal of SS
and total BOD as loading conditions changed. Average weekly values were
used to indicate overall system performance while extreme daily values were
used to indicate performance under extreme loading conditions.
In looking at the data several periods of data within the study period
become important. August to September 1977 and December 1977 to May 1978
were periods of reduced surface area in the system because of structural
failure of the first shaft of two separate flows paths. The surface area
reduction from 1,9 to l.k million sq ft was taken into account in the data
and allows analysis of the RBC operations under high surface loading
conditions. Spring 1977 was also an important time at the plant because a
discharge of highly concentrated metal plating wastes from an industry in
the city stopped all biological activity In the digesters. Subsequently
solids were allowed to build up in the primary clarifiers before being
vacuum filtered for disposal rather then being digested. This period
allows for comparison of RBC removal efficiency under adverse high solids
conditions to operations under more normal conditions assumed to occur
during the balance of the study period.
Three other specific periods become important in the analysis. First,
September to December 1977 because It was a time of highly variable flow
with relatively constant influent BOD and SS. Second, August to December 1978
because it was a time of relatively constant flow with highly variable in-
fluent BOD and SS. These two periods are of interest in determining if changes
in flow and concentration effect removal efficiency for the same organic load-
ing. The third period of interest was February 1978 because of the extremely
high BOD arid relatively low SS received at the plant during that time.
No soluble BOD data was available so the ratio of BOD/SS was used to
give a relative value for the soluble nature of the waste. It was felt that
with this ratio being maintained throughout the study period at a relatively
constant value a relatively constant waste type would be entering the plant.
Municiple waste characteristics would not normally be expected to change
drastically and BOD/SS was used to indicate possible discharges of other types
of wastes into the system. The highly soluble waste received at the plant in
February 1978 was thought to come from a creamery in the city. Several other
Isolated incidents of high BOD/SS ratios occurred throughout the study period,
215
-------
but February 1978 was isolated because the BOD/SS ratio remained high for a
full four week period, providing a sufficient period for results to show any
effect on removal efficiency that might occur from the highly soluble organic
loading.
Figure 7 shows the variation of weekly flow and HLA values. Flow
increased gradually from 1 MGD in January 1977 to over 2.5 MGD at the end
of the study period, still at half design flow. This gradual increase was
related to growth of the area and to more of the collection system coming on
line to the new plant. Peak flows that occurred during the study period were
generally related to seasonal variations in rainfall and subsequent infiltra-
tion. Average flow for September to December 1977 was 2.5 MGD and varied
from 1.6 to k.2. MGD during that period. Average flow for August to December
1978 was 2.5 MGD and varied only between 2.1 and 3.4 MGD. Flow during
February 1978 was relatively constant at 1.7 MGD ranging from 1.6 to 1.8 MGD.
HLA exhibited similar trends in the data. Values of HLA ranged from 0.75 to
3-5 gal/d/sq ft and were quite variable at times as was flow. An average
value of HLA for September to December 1977 was 1.3 gal/d/sq ft, for August
to December 1978 was 1.3 gal/d/sq ft, and for February 1978 was 1.2 gal/d/sq
ft. Variations of HLA during these same periods were 0.8 to 2.2 gal/d/sq ft,
1.1 to 1.8 gal/d/sq ft, and 1.1 to 1.3 gal/d/sq ft respectively.
Figure 8 shows influent BOD and SS weekly average variations with time.
Again no soluble BOD data was available from Kirksville but the ratio of
influent BOD/SS was used as an indicator of the soluble nature of the waste.
BOD/SS averaged 1.24 for the study period less February 1978. During February
1978 the ratio of BOD/SS averaged 2.03. This four week period of high BOD and
low SS was seen to represent the lowest BOD removal efficiency of the whole
study period. Influent BOD varied from 113 to 441 mg/1 during the study
period and averaged 237 mg/1. Influent BOD averaged 192 mg/1 for the period
September to December 1977, 262 mg/1 for August to December 1978, and 383 mg/1
for February 1978. Influent BOD ranged between 127 and 257 mg/1, 165 and
343 mg/1> and 299 and 441 mg/1 during the three periods respectively.
Suspended solids did not show the variation that BOD did and ranged
from 103 to 312 mg/1, averaging 191 mg/1. Influent SS generally ranged
between 100 and 200 mg/1 except for the period of May 1977 where SS steadily
rose from 242 to 312 mg/1. This was the period when digester problems
occurred and the solids build up in the system was thought to be the reason
for the rise in influent SS.
Figure 9 shows the variation of weekly OLA and BOD removal for the
study period. OLA reflects both hydraulic and influent BOD variations and was
seen to vary widely throughout the study period. OLA varied from 1.2 to 4-9
lb/d/1000 sq ft and averaged 2.5 lb/d/1000 sq ft for the whole study period.
OLA averaged 2.0 lb/d/1000 sq ft and varied from 1.2 to 3.6 lb/d/1000 sq ft
during the period September to December 1977- OLA for August to December 1978
varied between 1.9 and 3.9 lb/d/1000 sq ft and averaged 2.8 lb/d/1000 sq ft.
February 1978 OLA values averaged 3.7 lb/d/1000 sq ft, and ranged from 3-2 to
4.2 lb/d/1000 sq ft. BOD removal followed OLA closely for nearly all of the
study period except during February 1978 when the highly soluble organic
loading was applied to the plant.
216
-------
to
z:
03
ta
in
cc
—i
a:
10 --
g ..
B --
7 --
B --
S --
H --
3 --
2 •-
I
0
H
1
B
S
H
3
2
I
KIRKSViLLE RBC PERFDRMRNCE &RTR
U
13T7
D
d
isia
D
KIRK5V1LLE RBC PERFDRMRNCE BRTR
a
13*7"3
isie
FIGURE "7 WEEKLV VRRIBT1DNS OF HVDRRULI C PRRRMETERS
217
-------
U.
25
in
in
1000 •-
300 -•
B00 -•
700 -•
B00 -•
K00 -•
"400 •-
300 -•
200 • •
100 ••
0 --
KIRKSVILLE RBC PERFDRHRNCE BHTR
KIRK5VILLE RBC PERFDRMRNCE DR7H
-*-
J
1377
d
I37B
FIGURE 1 WEEKLY VRRIflTIDNS OF WHSTEWRTER INFLUD4T PRRHME1ER5
218
-------
£3
in
E3
in
a
ra
El
CD
OS.
en
IB
3
B
•7
E
S
H
3
•5 J.
E)
121
D •!*
s • •
•7
E
£
H
3 --
2 -•
J ••
El
KIRK5VILLE RBC PERFDRMRNCE
R
J
1311
D
R
KIRK5VILLE RBC PERFDRMRNCE DRTR
d
IBT7
J
13*79
FIGURE 9 WEEKLV VHRIRTIDNS DF SURFRiE LORDING RN5 REHDVRL
PHRRME7ERS
-------
Figures 10 and 11 represent removal efficiency as percent BOD and SS
removal and resulting effluent quality In mg/1 of BOD and SS. Percent BOD
removal generally ranged between 80 and 95 percent during the study and averaged
88 percent for the whole period less February 1978. February 1978 weekly
average BOD removal values ranged between 78 and 8k percent and averaged only
81 percent, the lowest of any four week period during the study. Low BOD
removal during February 5-11, 1977 was due to a reported influent BOD
concentration of only 23 mg/1 which yielded 22 percent BOD removal. Average
removal without this value would be 90 percent for the period. February 25
to March 2, 1978 was also a low BOD removal efficiency period averaging little
better than 81 percent. Again low removal efficiency was due not to high
loading on the system, but to low influent BOD concentrations reported. In-
fluent BOD's as low as 83 mg/1 resulted in very low removal efficiencies and
thus lowered the average removal value for that period. BOD removal for
Septmeber to December 1977 was fairly constant ranging from 80 to 95 percent
and averaged 90 percent, August to December 1978 yielded BOD removals ranging
from 87 to 98 percent, averaging 3k percent for the period.
SS removal remained over 90 percent for most of the study period but
was seen to deteriorate after solids built up in the system as a result of
digester problems in the spring of 1977- Despite the digester problems
however, removal remained at an average of 91 percent during the latter
part of 1977- Surge hydraulic loadings were also shown to affect SS removal
but percent SS removal during the variable flow period of September to December
1977 still averaged 90 percent and ranged from SO to 95 percent. August to
December 1978, the period of constant flow and varying organic loading, pro-
duced a range of SS removal to 87 to 98 percent with an average removal of
3k percent. SS removal ranged from 86 to 90 percent and averaged 88 percent
for February 1978.
Effluent quality as shown in Figure 11 provided valuable information
as to the trouble areas of the system. Effluent BOD averaged Zk mg/1 for
the study period less February 1978. Effluent quality in February 1978
deteriorated to an average 70 mg/1 with peaks reaching 75 mg/1 during the
*"»*-«4- *•*§»*•»! ?§<** A * A «".<•%*• **#* **•*?> 1 s\ *• r\ 1 »* ir**-**-*"**""**"
-------
LU
K
Ct
a
m
*%
I UU
30 *
B0 -
10 •
E0 -
SB -
H0 -
30 -
20 •
10 •
a .
W^ H IWxA M/^^
. ][ vi V^
KJRKSVILLE RBC PERFDRHRHCE 5RTR
E
an
in
in
e%
unun. unuoiJ
100 -
H0 -
B0 •
70 -
E0 -
50 •
H0 •
30 •
20 -
10 -
0 -
r.
1ST? me
^V^ ^1^1^
V
KJRK5VILLE RBC PERFDRKRNCE &RTR
IfiilDUfltlDi.
IS7T I SIS
FIEURE 10 WEEKUM VFIRIRtiDNS DF REMDVHL EFFICIENCY PHRRME7ER5
-221
-------
100
KIRK5VJLLE RBC PERFDRMRJKE &RTH
IBB
90 •-
B0 ••
70 -•
3 E0 -•
~ STB •-
UJ
in
in
30 -•
20 -•
4
10 ••
KIRKSVILLE RBC PERFnRMRNCEl bHTR
R
U
I SIT
J
I31B
FIGURE I! HEEKLM VRRIRTIDM5 DF EFFLUENT QUHL
222
D
PRRRMETER5
-------
identified In the discussions above. All values are average weekly values
from data found in the Klrksville treatment plant monthly reports.
TABLE 3 RBC LOADING CHARACTERISTICS
PERIOD
PARAMETER S-D 1977
Q (MGD)
Average 2.5
Range 1.6-4.2
HLA (gal/d/sq ft)
Average 1.3
Range 0.83-2.2
BOD/SS 1.24*
BOD Influent (mg/1)
Average 192
Range 12J-257
SS Influent (mg/1) ^
Average 191
OLA (lbBOD/d/1000 sq ft)
Average 2.0
Range 1.2-3.6
%BOD Removal
Average 84
Range 78-90
%SS Removal
Average 90
Range ' 80-95
Effluent BOD (mg/1)
Average 29
Range 18-43
Effluent SS (mg/1)
Average 16
Range 8-30
February 1978
1.7
1.6-1.8
1.2
1.1-1.3
2.03
383
229-441
191*
3.7
3-2-4.2
81
78-84
88
86-90
70
63-75
23
20-30
A-D 1978
2.5
2.1-3-4
1.3
1.1-1.8
1.24*
262
165-343
191*
2.8
1.9-3-9
93
87-97
94
87-98
17
11-25
11
5-18
sr
Average for the entire study period
The relationships between BOD removal and loading parameters were then
investigated. Figure 12 shows the relationship between percent BOD removal
and influent BOD. Without daily extremes a general increase In percent
removal resulted from an increase in influent BOD. The correlation co-
efficient was extremely low however, and the relation did not prove to be
an adequate estimator of BOD removal efficiency. With daily extreme values
added the curve approached the theoretical curve shown earlier in Figure 5,
with an initial increase in removal to a limiting value of influent BOD of
approximately 100 mg/1, beyond which an approximately constant removal of
90 percent resulted, independent of influent BOD concentration to a value of
700 mg/1.
223
-------
3C
tu
EC
ta
«
x:
LU
e
a
ta
c*
1 tlW
90 •
H0 •
10 •
B0 •
£0 •
H0 •
30 •
20 •
12 •
B.
•_^M^- — ""^^
• *«•>*<• *
CDRRELRTIDN COEFFICIENT" 0.2H33
5LOPE= 0.013
^-INTERCEPT" B3.3
STHMDRRD CEVIflTIDN= H.7
KIRKSVILLE RBC PERFORMRNCE DHTR
.=HEEKLM RVERRGES
0 1 00 200 300 H00 £00 B00 100 B00 300! 1 00i
am INF (ME/LJ
100 n
30 •
B0 •
70 •
B0 -
£0 -
H0 -
30 •
20 -
10 -
d -
V
..'.*-'.• • ^ x x X
w " '*•* :** *•"""• X
XJBk "» 1* • * W W
*<*«••• A IP
I J---"T.- . •**
' ^' *'••"• "
X
X KIRKSYILLE RBC PERFDRHRNCE DHTR
X
x
x
x
.«HEEKLM RVERHGES
X=EXTREME 0R1LH VRLUES
JI3 ^ 106 INF ^ HHI
.X
1— * 1 .1 i t i i >
0 100 200 300 M00 £00 E0B 700 000 300 1000
BOD INF (HG/LJ
FIGURE \1 PERCENT 1DB REHDVHL VERSUS INFLUENT BOB
CDNCENTRRtlDM/ HITH RNB NITHDUT BHILM EXTREME
VRLUES
224
-------
Similar results were observed for the plot of percent BOD removal versus
OLA shown in Figure 13- A slight increase in BOD removal with increased OLA
was seen in the plot of weekly averages; but again correlation was very poor
between variables, making the relation impractical for predictive purposes.
With the addition of extreme daily values a rapid increase in percent removal
occurred to approximately 90 percent at 1 lb BOD/d/1000 sq ft, beyond which
percent removal remained constant for values of OLA to 9.6 lb BOD/d/1000 sq ft.
Figure \k shows percent BOD removal as a function of HLA. Generally,
increased HLA produced decreasing percent BOD removal, but, again, correlation
between variables was extremely low. When daily extreme values were added,
correlation of the variables increased; but, correlation was still very low
and no definitive relation was evident from the plot. HLA was found to be an
imprecise indicator of RBC performance on a percent removal basis and other
relationships were investigated to see if they would be more descriptive of
RBC removal efficiency.
BOD removal on a pound/d/1000 sq ft basis was plotted against HLA for
specific influent BOD concentration ranges as shown in Figure 15- These
ranges were chosen to produce the lines of best fit and were shown to
describe the data extremely well. Daily extreme values were added to yield
information on whether peak loadings would cause deviations from weekly
average plots. Table 4 gives the least squares statistical results for the
data plotted for Figure 15. It can be seen that the data correlated well
for all ranges of influent BOD with and without daily extreme values.
Figure 16 is another form of Figure 15, being BOD removal as lb/d/1000
sq ft versus influent BOD for a given HLA range. Ranges of HLA were again
chosen to yield maximum correlation coefficients and daily extreme values
showed no significant deviation from monthly average trends. Table 5 shows
the statistical data for the set of curves in Figure 16. Again the data
correlated extremely well and could be used as an indicator of RBC per-
formance for RBC analysis and design.
Figure 17 shows BOD removal versus OLA as lb/d/1000 sq ft and exhibits
the same relationship reported in the literature. A constant rate of removal
resulted for general weekly average values from 1.2 to 5-0 lb BOD/d/1000 sq ft
and was maintained with the addition of daily extreme values to a loading of
9-6 lb BOD/d/1000 sq ft. No leveling off of the relationship was observed at
the high loading conditions, suggesting that the Kirksville plant was mass
limiting even at these extreme loadings. With the high correlation coeffi-
cient of 0.9869 for weekly average values, this relationship can be used as
an estimator of RBC BOD removal on a lb/d/1000 sq ft basis for a given organic
loading. The most important variable in analysis of RBC performance, the one
most indicative of BOD removal efficiency, was OLA. For this reason OLA
should be used for design and analysis of RBC's rather than HLA or detention
time In the system.
CONCLUSIONS
The results of the described research and review of the literature enabled
the following conclusions to be made:
225
-------
LU
o:
LU
ce
a
m
1 UKI i
H0 -
B0 *
70 -
E0 -
£0 •
M0 -
30 •
20 -
10 -
0 -
0
100 •
30 -
B0 •
70 •
60 •
£0 •
H0 •
3i -
20 •
m •
0 •
0
... ••" . ""* " * "
..-.'•.*.• ^ -
,
•
KJRK5Y1LLE RBC PERFDRMRNCE BRTfl
. -WEEKLY
HVERFIBES
CBRRELRTIDM CDEFFICIENT= 0.0B33
SLDPE« 0.3SE
Y- INTERCEPT" BB.S
STHN0HRD DEViHTJDN«= M.Q
1 2 3 H
S E "7 B 3 IE
DLH CLB/D/iB12B SB FTJ
*.'*••*'.."*. •
• ^r:r"."VjJ" -•
- f^''"'-" '': "'
x KJRKSVILLE
X
K
X
! 2 3 H
i XK x
x xx
y v
A ^y
RBC PERFORHRNCE DRTR
.«HEEKLH HVERREES
X-EXTREME DRILY VRLUES
1 .2 > DLH } H.3
£ E 1 B § IE
DLR (LH/&/I000 BB FT/
FIGURE 13 PERCENT BOB REHOVBL VERSUS DLH/ HItH RND
WITHOUT DRILM EXTREME VRLUES
226
-------
5
tc
s
§
1 UU
90 •
aa •
7B •
50 •
SB •
M0 •
32 •
20 •
10 •
0 •
H
100 -
30 -
BB -
70 •
E0 •
SB •
MB •
3B •
20 •
IB -
71 *
,** * *" *.
— H^^Li^,
***'f '••'*" * * ~ ~~~ " ~~ — —
* » * * * • t ' —
KIRKSVJLLE RBC PEHFDRHRNCE 5HTH
."MEEKLY RVERHGE5
CDRRELHTIDN CDEFFJ Cf ENT«-0. ISSI
SLOPE** -i .EKB
" V _ t lf*T r*d /"CT3T** QQ £2
1 *" i N I fcJTv V. t." 1 ** Oil * ZI
STflWDHRD BEVIHTIOH* M.B
U 1 1 . I I 1 I 1 it | i , I I • - 1 1 I 1 1 HI,
i~ - - - -_J>Jiuji™«i.«..«j« •*._. _.i_j -..i™.-i|z ..... ...j | ..,_-_j_. p- ,.i.i.ii.wniig»iu.i>uiinn.« ( . ^
1 2 3 M £ S
MLR (ERL/D/5B FTJ
-4^:. "j K
" XK* ^* !^r^*^^^JL^^^
— ~^_
X
.-HEEKLM RVERRBE5
• X**EXTREME BR1LV VRLUES x
B.7 > HLH > 2.S
CORRELHTIDH COEFFI C! EWT—1.3S23
5LDPE« -3.H75
1-IMTERCEPT" 32.3
5THNDHRD &EVJRT1DN« B.S
KIRKSVILLE RBC PEHFORKHNCE BRTH
.1. . 1 T. — . >• till » 1 .1 ' t 1 HI . 1 1 II II
E5
•2 3
HLH tBRL/5/Sa
FIGURE IH PETRCCNT HDD REMDVHL VERSUS HLHr HltH RN0
WITHOUT amur EXTREME VBLUES
227
-------
c
R
c
ffi
BDD 1NFLUDJ1
CME/LJ
HB0- 3JB-
MIS! 333
BOD INFLUENT
CME/L;
X-EXTREHC
DRILH
VHLUE5
I El-220
IBi-liB
I 2 3 H
HLR tEHL/D/S3 FT;
IK BOD REHDVBL tLB/D/IHaa SQ FT; VERSUS HLR FDR R
El YEN INFLUENT BOD O3NCENTRFITIDN RHNGE/ WITH
RND WITHOUT DRILX EXTREME VflLUES
228
-------
TABLE 4 BOD REMOVAL VS HLA CURVE STATISTICAL DATA
BOD INFLUENT RANGE WITHOUT
(mg/1) R
0-60
61-105
106-160
161-220
221-260
261-315
316-399
400-475
476-700
0.
0.
0.
0.
0.
9594
9572
9637
9341
9717
DAILY EXTREMES
M Y
0.
1.
1.
2.
2.
98
26
53
10
97
-0.
0.
0.
0.
-0.
02
14
32
00
03
WITH DAILY EXTREMES
R M Y
0.9937 0.30 -0.20
0.
0.
0.
0.
0.
0.
0.
0.
9927
9786
9690
9707
9842
9533
9875
9072
0.
0.
1.
1.
2.
2.
"1
4.
33
87
32
76
09
78
52
34
0.26
0.04
0.03
0.05
0.08
-0.13
-0.25
0.04
WHERE: R=CORRELATION COEFFICIENT
M=SLQPE OF LINE OF BEST FIT
Y=Y-INTERCEPT OF LINE OF BEST FIT
TABLE 5 BOD REMOVAL VS BOD INFLUENT CURVE STATISTICAL DATA
HLA RANGE
(GAL/DAY/SQ FT)
0.
1.
1.
2.
50-1.
01
51
51
-1.
-2.
-4.
00
50
50
50
WHERE:
WITHOUT
R
0.
0.
0.
8778
9162
731
1
DAILY
M
0.
0.
0.
006
009
013
R=CORRELAT10N
EXTREMES
Y
0.
0.
0.
COEFF1
06
1
1
C
2
0
1
WITH DAILY EXTREMES
R M Y
0.9489 0.006 -0.04
0.9591 0.009 -0.06
0.9649 0.016 -0.51
0.9582 0.023 -0.07
M=SLOPE OF LINE OF BEST FIT
Y=Y-INTERCEPT OF LINE OF BEST FIT
229
-------
RRNEE (GHL/D/SH
0 100 20B 300 H00 £00 E00 702 B00 H00 1000
BD5 INF (HB/LJ
HLR RRNBE
CGRL/D/SH FTJ
RVEPHEES
X-EXTREME DRILf TRLUES
B IBB 200 300 400 S00 B00 *7BB B00 900 IB00
BOD INF (HB/LJ
FIGURE IB BQD REMDYRL (LB/D/I000 50 FTJ VERSUS BOO INFLUENT
FOR R EIVCN HLR RRNBE/ NITH HMD WITHOUT DRILY
EXTREME VflLUES
230
-------
IB
9
B
7
s
N
3
2
I
B
KIRKSY1LLE RBC PERFORMHNCE DRTR
.-HEEKLX RYtRBEES
CDRRELHT1DN CDEFFKICNI- B.iSES
SLOPE- 0.BS3
Y-JN7ERCCPt« -B.B
5TRNDRRD DEYiflTIDN- B.I
3 M S B 1
OLR tLH/D/iaaa so FTJ
B
IB
IB
S
B
7
3 * "
2 ••
I ••
B
CIRK5VILLC RSC PEFFDRMflNCE &R7H
.-HEEKLM RVERHEE5
X-CX7REME DHILM VHLUES
1."2 1 OLH ) S.B
CDRRELRTIDN CBEFFl CJEJTI
SLOPE- 0.S2B
t-INTERCEPT- -B.I
ETRNDRRD &EVinTiQN« B.2
0.S3M1
1 I ' t •"••' "I I > i
2 3 M S 1 1
DLR CLB/D/IBBB SB FTJ
II BOD RErtOVHL {LB/&/1BB0 SB FTJ VERSUS QLH/
H1TH RN0 NITHOUT DRlL^f EXTREME VHLUES
231
-------
1. The Klrksville RBC units performed efficiently under normal organic
and hydraulic loading conditions, giving an average BOD removal efficiency
of 88 percent with an average BOD of 2k mg/1 over OLA values from 1.2 to
4.9 Ibs BOD/day/1000 sq ft and HLA values from 0.5 to 3.2 gals/day/sq ft.
2. High soluble organic loadings adversely affected RBC performance, giving
an average BOD removal of 81 percent with an average BOD of 70 tng/1 at
an OLA of 3-7 Ibs BOD/day/1000 sq ft and an HLA of 1.2 gals/day/ sq ft.
3- Highly variable hydraulic loadings reduced the treatment efficiency of
the RBC units as a result of reduced contact time and surges on the
final clarlfiers.
k. By itself, HLA was not a satisfactory indicator of RBC performance;
but could be combined! with organic concentration to yield the OLA
which correlated read'ily with RBC operations.
5- The best indicator for RBC operations appeared to be the soluble,
BOD/day/1000 sq ft loading rate.
r
i
6. The lack of adequate data prevented the development of more precise
design criteria for RBC units but the results of this study helped
to indicate the data required for proper evaluation.
ACKNOWLEDGEMENTS
We would like to thank the City of Klrksville for their help in providing
the data necessary for the study. We would also like to thank Bob Lamberton
of Larkin and Associates for his help in providing background information on
the design and operation of the Kirksville plant. Also Jane Boatwright,
Secretary of the Environmental Health Engineering Department, deserves a great
deal of thanks for all her help and patience in typing this paper.
232
-------
BIBLIOGRAPHY
1. Doman, J., "Results of Operation of Experimental Contact Filter with
Partially Submerged Rotating Plates," Sewage Works Journal, J_, 555,
(October 1929).
2. Antonie, Ronald L., Fixed Biologleal Surface Wastewater Treatment, The
BiologicalContactor,CRC PressInc., Cleveland, Ohio, (1976).
3- Hartman, Ing H, , "The BIO-DISC Filter," Reprint from Oesterrelchische
Wasserwirtschaft, No 11/12, (1965).
4. Welch, Fred M., "Preliminary Results of a Hew Approach in the Aerobic
Biological Treatment of Highly Concentrated Wastes," Proceedings of
the 23rd Purdue Industrial Waste Conference, 428, (1968).
5. Pescod, M.B. and Nair, J.W., "Biological Disc Filtration for Tropical
Waste Treatment," Water Research, j>, 1509, (1972).
6. Cochrane, Max W,, Burn, Robert J., and Dostal, Kenneth A., "Cannery
Wastewater Treatment With RBC and Extended Aeration," EPA Technology
Series EPA-R2-73~024, (April 1973).
7- Stover, E.L., and Kincannon, D.F., "One Step Nitrification and Carbon
Removal," Water and Sewage Works, 122, 66, 6, (June 1975).
8. Antonie, Ronald L., "Application of the BIO-DISC Process to Treatment
of Domestic Wastewater," Presented to 43rd Annual Conference of the
Water Pollution Control Federation, Boston, Massachusetts, (October
4-9, 1970).
9- Antonie, Ronald L., "Response of the BIO-DISC Process to Fluctuating
Wastewater Flows," Proceedings, 25th Purdue Industrial WasteConference,
425, (1970).
10. Chittenden, Jimmie A., and Wells, James W., Jr., "BOD Removal and
Stabilization of Anaerobic Lagoon Effluent Using a Rotating Biological
Contactor," Presented to the 43rd Annual Conference of the Water Pollution
Control Federation, Boston, Massachusetts, (October 4-9, 1970).
11. Label la, Salvatore A, Indravandon, H.T., "Treatment of Winery Wastes by
Aerated Lagoon, Activated Sludge Process, and Rotating Biological Contactor
or RBC," Proceedings of the 27th AnnualPurdue Industrial Waste Conference,
(1972).
12. Ahlberg, N.R. and Kwong, T.S., "Process Evaluation of an RBC for Municipal
Wastewater Treatment," Research Paper #W2041 , Wastewater Treatment Section,
Pollution Control Planning Branch, Ministry of Environment, Toronto, Ontario
(1974).
13. Molhorta, S.K., Williams, T.C., and Morley, W.L., "Performance of a
BIO-DISC Plant !n a Northern Michigan Community," Presented to the 1975
WPCF Conference Miami Beach, Florida, (October 5~9» 1975).
233
-------
BIBLIOGRAPHY (Cont)
Stover, E.L., and Kincannon, D.F., "Evaluating RBC Performance," Water
and Sewage Works, 3, 88, (March 1976).
15. Hoehn, R.C., and Ray, A.D., "Effects of Thickness on Bacterial Film,"
Journal Water Pollution Control Federation, k$_, 11, 2303 (November 1973)
16. Sanders, W.M. , "Oxygen Utilization by Slime Organisms in Continuous
Culture," Proceedings of the 19th Purdue Industrial Waste Conference,
986 (May 5-7, 1964)!
17, Kornegay, B.H., and Andrews, J.F., "Kinetics of Fixed Film Biological
Reactors," Journal Water Pollution Control Federation, ^0, 11, R1»60
('November 1968)
234
-------
ROTATING BIOLOGICAL CONTACTOR FOR THE TREATMENT OP
WASTEWATER IN INDIA
By
A.N. Khan
Scientist* Sewage Treatment Division
V. Raman
Scientist & Head
Environmental Engineering Consultancy & Sewage Treatment Divn.
National Environmental Engineering Research Institute, Nagpur,
India.
INTRODUCTION
There is a long standing need in India of providing sa-
tisfactory, low cost and simple sewage treatment facilities
for isolated houses, institutions, hotels, small and medium
sized eommuaities. Such facilities should require occasional
inspection and maintenance, flexibility in construction and
operation, relatively unskilled supervision, maintenance and
operation, and should occupy limited area. They should be
capable of treating wide variation of flow and organic load
producing effluent of desirable standards of purity with res-
pect to BOD and suspended solids* The Rotating Biological
Contactor (RBC) or Bio-disc system, a secondary biological
treatment process, is claimed to have these advantages which
appear to be suitable to Indian conditions.
The RBC or Bix>-disc unit consists of a series of
closely spaced circular discs mounted on a horizontal rota-
ting shaft. The shaft along with the discs are fixed in
a semi—circular cylindrical tank through which the waste-
water flows* with the water level just below the shaft. While
rotating at a iow speed, the disc surface is alternately
exposed to the atmosphere and wastewater. The disc serves
as media for growth and adhesion of biological slime, device
for bringing the slime and film of water in contact with air,
and creating mildly turbulent mixing conditions within the
tank contents. As the disc rotates, the biological slime on
any sector of the disc is alternately dipped into the waste-
water, where the slime metabolises non—settleable and dis-
235
-------
solved organic matter and aerated with each revolution of
the disc. The settleable organic matter and the sloughed film
due to excess growth from the disc, passes along with the
effluent as suspensions which are removed in the subsequent
stage in the secondary settling tank.
REVIEW OP LITERATURE
The idea of RBC originated in USA in 1928, and it was
referred to as 'Biological Wheel' and in Germany as
•Immersion Drip Filter'(l). Further developments in the pro-
cess took place during the last two decades mainly in •
Germany and USA. Large number of plants are now in vogue
in Europe and USA as reported by Anthonie(2). Laboratory
and pilot plant studies in understanding the kinetics and
evolving design criteria are reported by Antonie(3,4).
Torpey et al<5) Khan and Siddiqi^6', Raman and Khan *7'8;,
Pescod and Nair^j PretoriusUOf steels *H), Bruce and
Merkens(12), Clark et al U3) and Kluge and Kipp v!4j.
The first municipal wastewater treatment plant in Pewaukee,
USA using RBC process constructed by Autotrol Corporation
on full scale was in operation since 1971 to treat a flow
of 1179 n\3/day ( 0.47 mgd); and the first large scale RBC
treatment unit with a capacity of .20 mgd for upgrading the
existing treatment unit is in operation in Philadelphia.
Based on the studies reported in USA, Europe and
Thailand, the general requirements for design and opera-
tion are as follows: The rotational speed of the disc is li-
mited to 1 to 10 rpm and the clear spacing between the discs
is kept at 2 to 5 cms with a gap of about 5 cms at the
bottom. A Detention time of 10 to 90 minutes, organic disc
loading rate of 3 to 16 gms/m2/day, hydrauloc loading rate
of 0.04 to 0.11 m3/day/m3 of disc area, power consumption
(depending on capacity) of 0.3 to 2.6 kwh/kg BOD removed and
power requirements of 0.4 to 1 watt/m2 of disc area are the
other relevant parameters, while the BOD removal efficiency
ranges from 80 to 95% at wastewater temperatures of 10°C to
20°C.
OBJECTIVE OF STUDY
The RBC as a treatment device is not yet in use in
India, eventhough it has great potential due to its compact-
ness and simplicity of operation and favourable tempera-
ture conditions. As such, it is felt that the feasibility
of its use in India should be investigated under laboratory
and field conditions. The studies aim to achieve the objec-
tives of formulating relationship amongst the treatment effi-
ciency, organic loading, hydraulic loading, rotational speed,
power consumption, and evolve compact units for various treat-
ment capacities.
236
-------
MATERIALS AND MEMOES
Initially, laboratory model studies were carried out
with synthetic sewage and settled sewage, followed by studies
on pilot plant treating raw municipal sewage.
A. Laboratory Model Using SyntheticSewage andDomestic Sewage
The laboratory model set up of KBC as shown in
Fig. 1 consisted of 20 on diameter, 4 mm thick, as-bestos
cement discs (with perforations) with clear spacing of
2 cms. They were centrally mounted to a horizontal shaft
rotated by a fractional horse power motor ( % HP ) fitted
with reduction gear and adjustable belt drive for varying
the rotational speed ( 3,5 and 8 rpm). The discs along with
shaft assembly were mounted to a perspex tank ( of cross
section similar to Imhoff tank), About 40 per cent of the
disc area was submerged in wastewater. The bottom portion of
the tank acted as a settling tank cum sludge storage tank,
The reaction—cum-settling chamber was partitioned into small
compartments to contain varying number of discs.
It was possible to feed separately each of the compart-
ment and work independently of the others with separate
inlet and outlet arrangements. Continuous feeding was accom-
plished by electrolytic pump connected to feed bottle con-
taining synthetic or raw settled sewage. Pig. 2 shows the
set up of the laboratory model experiment. The sloughed
material and the settled organic material were periodically
removed from bottom storage chamber. The feed was varied for
various hydraulic and organic loading rates, The influent
and effluent samples were collected at particular time in-
tervals regularly and analysed for the usual parameters like
temperature, pH, BOD, COD, S.S, TDS, NHa-N. The units were
continuously dosed at loading rates varying from 6.2 to 42
gras of BOD per square meter of disc area per day. The synthe-
tic sewage was prepared by dissolving a high protein cereal
and milk powder in water to get a BOD§ of aoout 250 rag/1.
Later, the experiments were repeated with settled
domestic municipal sewage, and the unit was operated at or-
ganic load rates ranging from 6.2 g/day/m2 to 31,0 g/day/ra2
of disc area, at rotational speeds of 3,5 and 8 rpm.
further studies were carried out using cylindrical
semicircular chamber fitted with shaft and some discs with
separate settling compartment, the total capacity remai-
ning the same as that of Imhoff type tank.
237
-------
PULLEY;
SHAFT-
MOTOR
-DRIVE BELT
I 2 3
U
5
LJ
6
LJ
9
60 Cm
4
Ul~
II
5 6
UF
2
1J*^4
23
26 Cm
OUTLET
PLAN OF THE UNIT
8 Cm
4 mm. THICK CEMENT
ASBESTOS DISCS
SECTION
FIG. I : LABORATORY UNIT OF ROTATING BIOLOGICAL CONTACTOR.
238
-------
w
VO
/ A j—
HH MI-
ELECTROLYTIC
PUMP '
INFLUENT
FEED -H
BOTTLE
DRIVING BELT
EFFLUENT k-P ----
R. a C. UNIT-
FIG, 2 : ROTATING BIOLOGICAL CONTACTOR SHOWING FEEDING ARRANGEMENT
-------
B. PILOT PLANT.
(i) Large size composite RBC cum Settling Tank with
A.C, Sheets.
The studies were carried out initially on a large size
pilot plant of Irahoff tank type with the settling chamber below
the reactor. The discs were made from 1.22m x 1.22m square pie-
ces of asbestos cement sheets ( 136 numbers) by chopping off
the corners to give octagonal shape with surface area of 1.226
square metres. They were mounted on a mild steel shaft rotated
at 5 rpm by 5 H.P. motor fitted with reduction gear and chain
drive. The settled domestic sewage was fed at the rate of
47.67 m3/day for a continuous period of 10 to 12 hours in a day.
The studies had to be discontinued after 4 months operation
due to some structural failure of the discs.
(ii) Small size Plant with .PVC circular sheets.
Fig, 3* shows the details of the set up of the RBC where
the secondary settling chamber is a separate unit. The tank
located in NEERI campus, Nagpur, consisted of semicircular mild
steel tank of 1,22 metre diameter and 1.52 m long. 40 plane
discs made of PVC of 1 meter diameter were fixed centrally to a
mild steel shaft of 3.7 cm diameter and they were spaced longtu—
dinally at clear interval of 2.5 cms. The clearance from the
bottom of the tank to the bottom edge of disc was kept at
7»5 cms* The shaft was rotated at 5 rpm by a 1.5 kw ( 2 H»P.)
electric motor fitted with reduction gear and belt drive» The
unit was fed by raw sewage tapped from a distribution chamber
where raw municipal screened sewage was pumped. The effluent
from the disc unit passed on to the rectangular settling tank
( of mild steel) of size 1,22m x 0.9lm x 0.508 m depth with
a liquid capacity of 0,5 m3. The settleable organic material
and the 'humus1 or sloughings from the film attached to the discs
were removed in the settling tank. The feed to the system
was for 8 to 12 hours continuously in a day, and the remain-
ing hours of the day, the RBC was working without any feed. The
flow measurements were carried out by V-notch attached to a
separate chamber after the settling tank. Occasionally, volu—
metrically also the flow was computed. The performance of RBC
was observed under two conditions viz. open type and enclosed
type covered by a mild steel perforated cover C fig. 4 ). Later,
some modifications were made in the outlet for the settling
tank ( vide Fig. 4) by providing serrated weirs instead of
5 cms. diameter circular opening initially.
The studies were carried out first with open reactor and
secondary settling tank with a single pipe outlet and later the
same units were modified with the reactor closed by a ventila-
ted semi-circular lid and the settling tank provided with
serrated weir outlets,
240
-------
INFLUENT
INLET
wmm
, , I, I ! i I .. •> •
MOTOR WITH
REDUCTION
GEAR
FINAL EFFLUENT
X"
RBC TANK SETTLING TANK
No. OF DISC=40 SIZE * 1-22 m.X 0-914 m.XO-508 m.
DISC DIAMETER, Im. LIQUID VOLUME = 0-508 m3
THICKNESS, 3m.m.
CLEAR SPACING
BETWEEN THE DISCS = 2-5 Cm.
TANK SIZE* I-22m.did x 1-52 m LENGTH
LIQUID VOLUME * 0- 726 ii|3
FIG. 3 : ROTATING BIOLOGICAL CONTACTOR PILOT PLANT WITH PVC DISCS
INFLUENT
INLET
VENTILATED COVER
EFFLUENT OVERFLOW
WEIR 30Cm.X3OCm.X30a
MOTOR WITH
REDUCTION
GEAR
RBC TANK
SETTLING TANK
FIG. 4 : RBC PILOT PLANT WITH COVER AND WITH MODIFIED SETTLING TANK
241
-------
t-%
r'v--- |pvf:
Photograph i 1 Rotating Biological Contactor showing
Biological Slime on Discs.
242
-------
Photograph!2 RBC Pilot Plant with modified settling tank
and ventilated cover.
• . _i --. ,~y .T • - <<5sr - .">•' v.jJ'- ••"•%• ' * ••'V '. • '•!*"" ,-X»~, . t- •*'. V-,«4i' .V.'1"-"',
.••£?•;• •-'-S^^:;---^-- •• ^-^a^^v^^o
^;::'--;v' •ft:;v-^;>:5^:i^<.s^^t'is!i^?ss;s.
Photograph t 3 RBC Pilot Plant with cover removed.
243
-------
OBSERVATIONS AND DISCUSSIONS.
A* Laboratory SBC Model Composite (Imhoff -bank type)
Reactor.
i) Using Synthetic Sewage
In the Imhoff tank type reactor with settling chamber below
the disc chamber , it took about 7 to 10 days to reach for
the optimum growth of biological film on the disc surfaces
and steady state conditions. Table 1 shows the performance
data of the reactor with reference to hydraulic load, tempera-
ture, COD, BOD and S.S. of the influent and effluent, and the
biomass accumulation on the discs. The organic loading rates
{ in terms of BODs) ranged from 6.2 to 42 gms/day/m2 of disc area
( 1.2 to 8.7 lbs/day/1000 square feet of disc area). The effi-
ciencies varied from 55 to 88 percent removal with reference
to BODs in the decreasing order of organic loadings. The BOD
of the influent varied trom 233 to 276 mg/1 while the S.S. of
the influent was 196 mg/1. The S.S. in the effluent ranged bet-
ween 18 to 25 mg/1.
ii) Using Settled Domestic Sewage.
The same composite unit was later charged with settled
domestic sewage for different loading rates working at three
different rotating speeds namely 3 rpm, 5 rpm and 8 rpm. Dif-
ferent organic loadings were achieved by utilising different
chambers having different numbers of discs and adjusting if
necessary the hydraulic flow rate. It took about 5 to 7 days
for the biological slime to grow and attain a steady state.
Table 2 shows the average performance characteristics of the
unit. The average wastewater temperature was 28°C. It is seen
from Table 2 that 90 per cent overall BODs removal could be ob-
tained with organic loading rate of 10 g/m2/day, when the
unit was operated at 5 rpm. The BODs of the final effluent
after settling was always less than 23 mg/1 ( 8 to 23 mg/1)
during the period of study, while the BODs of the influent to
the reactor varied from 102 to 130 mg/1, for the three dif-
ferent speeds considered. The suspended solids present in the
settled effluent varied from 8 to 15 mg/1.
Fig. 5 & Table 2 & 3 show the average overall performance
of the disc system at the three speeds studied viz. 3,5 and
8 rpm. There was marginal increase in the efficiency of
removal of BOD with increase in speed for the same organic loa-
ding rates.
iii) Laboratory RBC Unit with Semi-Circular reactor
and separate settling
A comparative study for a brief period was made regar-
ding the performance of semicircular tank reactor with separa-
te settling tank and composite Imhoff tank reactor using
settled domestic sewage. It is seen from Table 4 that the
efficiency of BODs removal of 83% could be obtained at an orga
nic loading rate of 14.9 g/m2/day for the semi-circular unit,
244
-------
LABORATORY" STUDY DATA ON THE PERFORMANCE OP
RBC UNIT ( IMHOFF TANK TYPE, COMPOSITE ) FOR
THE TREATMENT OF SYNTHETIC SEWAGE ( 1970-1972)
PARAMETERS
No. of discs
Surface area
EXPERIMENT NO.
(m2)
Hydraulic loading, n»3/m2/day
Detention time (hrs. )
Organic loading, g/m^/day
BOD (mg/1)
1
COD (mg/1)
, Influent
; Effluent
E % reduction
Influent
Effluent
% reduction
Effluent suspended solids,
mg/1.
2
Bioraass, VSS g/m of disc
surface
1
6
0.39
0.0256
8.4
6.0
243
28
88.4
438
72
84.0
12.0
18
2
4
0.26
0.0385
6.0
10.6
276
33
87.8
466
60
86.3
12.0
25
&
3
I
0.06
0.166
1.9
42.4
276
124
54.6
466
195
58.0
23.0
25
* For each experiment No. at least 10 observations
were recorded and the average given.
* Disc speed - 3 RPM
* Hydraulic flow for 24 hours in all the experiments
was 10 litres.
* Liquid temperature varied between 21-22°C.
* Influent Suspended solids 196 mg/1.
245
-------
TANK TYPE) TREATING SETTLED SEWAGE
DISC RPM
** EXPT. No.
Hydraulic
load.
Organic load
g/m2/day
BODs mg/1
Influent
Effluent
% reduction
COD ma/1
Influent
Effluent
% reduction
EH
Influent
Effluent
Suspended
Sol ids. ma/1
Influent
Effluent
% reduction
Biomass 0
VSS, g/m*
1
0.051
6.64
130
10
92
290
48
83
7.3
7.7
91
11
88
19.02
3
2
0.099
11.76
118
15
87
261
72
72
7.3
7.6
80
9
89
28.16
3
0.287
29.29
102
29
71
218
96
55
7.2
7.4
45
18
60
26.68
1
0.047
6.20
130
9
93
278
47
83
7.3
7.6
74
8
89
20.80
5
2
0.066
10.00
150
14
90
309
66
78
7.3
7.6
63
5
92
23.36
3
0.115
14.88
129
23
83
265
80
70
7.3
7.5
80
12
85
17.92
( 1975 )
1
0.051
6.64
131
8
94
304
59
80
7.3
7.6
86
10
88
16.00
8
2
0.099
12.20
122
8
93
269
64
76
7.1
7.5
76
8
89
23.00
3
0.288
31.00
108
22
79
234
90
61
7.2
7.5
51
15
70
18.00
* Average Liquid temperature was 28°C.
4- Municipal Sewage
** Number of Discs 6,4,1, -fov 4&Li*xS 1,2, a.™*. 3 m
246
-------
TABLE „ 3 • WEIGHT OP BIOMASS DEVELOPED ON THE DISCS
AT DIFFERENT DISC SPEEDS AND LOADS (USING
SETTLED DOMESTIC SEWAGE ) - L&B MODEL.
Disc Speed
RPM
.
3 |
I
5 j
8 1
I
Organic Loading
Rate * g/m^/day
6.7
11.7
6.3
10.0
6.7
12.2
Biomass
VSS,g/m2 of
disc area
19.20
28.16
20.80
23.36
16.00
23.00
* A month after start of operation.
247
-------
2 SO
o
1—
8 80
Q
I-LJ
tr
g 70
CD
^
60
50
*X? V
'k
•v
"S.
-x.
X
8— .5 r
r
*»
pm
pm
5 15 25 35
Organic loading (g/rrr/day)
FIG. 5 ORGANIC LOADING VS. PERCENTAGE BOD
REDUCTION AT 3,5 AND 8 RPM
248
-------
TABLE • 4: COMPARATIVE STUDIES USING SETTLED SEWAGE WITH
IMHOFF TYPE AWD SEMI CIRCULAR CHAMBERS WITH
SEPARATE SETTLING TANK ( OCT. 1975 TO MARCH
1976) AVERAGE VALUES*.
PARAMETERS
No. of Discs
2
Surface area, m
Chamber (s) capacity (litres)
Hydraulic loading rate.
Detentime time, (hrs. )
Organic Loading rate,
g/m2
B.O.D. raq/1
Influent
Effluent
% reduction
C.O.D. mq/1
Influent
Effluent
% reduction
Effluent Suspended Solids
mg/1
2
Biomass, g/m
EXPERIMENT NO.
*
1
2
0.13
2.05
0.115
3.3
14.9
130
31
76
307
87
71
15
62.8
**
2
2
0.13
2.18
0.115
3.4
14.9
130
22
83
307
70
76
10
76.8
* Imhoff type of Tank with Discs.
** Semicircular tank with Discs followed by a semicircular
settling tank. ( Vol. 0.950 •»• 1.230 = 2.180 litres)
•*• Average value of 21 observations.
- Disc speed,5 RPM
- Average liquid temperature, 24°C.
- Total flow to the unit for 24 hours, 15 litres.
249
-------
while for the composite Inihoff tank type, the efficiency of
BQD$ removal was 76%, under identical conditions. The removal
of the S.S. in the former type was 83% while for the compo-
site type, was 76%,
B. Pilot Plant Studies
1) Pilot Plant with Asbestos Cement Sheets Using Settled
Domestic Sewage,
The pilot plant composite RBC reactor ( with settling
tank below) was operated at a rotational speed of 5 rpm for
feed rate of 47,67 m3/day of settled domestic sewage. While
the RBC reactor was operating thoughtout the day* the sewage
was fed only for 10 to 12 hours continuously during the day,
An overall BODs reduction of 77 percent was observed* while
there was 50 percent removal of ammonia nitrogen and 3096 removal
of Phosphate ( vide Table 5 }. The studies had to be dis-
continued after three months of continuous operation due to
structural defects noticed in the disc assembly.
ii) Pilot Plant with PVC circular Pise XJsincr Raw
Domestic Sewage,
The raw domestic municipal sewage pumped from a sump re-
ceiving sewage from a municipal manhole was fed by gravity
to pilot RBC plant with PVC discs ( vide Fig. 3 and 4 ) through
a distribution chamber. The SBC reactor disc were operated
continuously throughout the day* while the flow of sewage was
restricted to 8 to 10 hours during the day time* due to opera-
tional difficulties. For a period of 2 years* the RBC reactor
was operated by keeping it open to the atmosphereand at a constant
speed of 5 rpm. The speed of 5 rpm was selected based on
laboratory studies and with reference to reduced power consump-
tion.
The PVC discs were initially found to be smooth and the
biological film adhering to the disc surface was not apparently
'thick'. After 3 months of operation with smooth discs* the
disc surface was showered and coated with fine sand which was
fixed to the discs by 'Wavin PVC Cement** Table 6 give the
performance data of the pilot plant studies for varying condi-
tions viz. open reactor tank with smooth PVC disc and rough-
ened with coated sand* closed reactor tank with modifications
to the outlet of settling tank. The hydraulic flow rate ran-
ged from 4.54 to 5.77 m3/day corresponding to an organic loading
rate of 15.2 to 27.7 g/m2 of disc are* per day. The BODs of
influent was almost consistent during the period of the .day
and as such samples were collected at a tine 2 to 3 hours after
the pumping of sewage started. The effective detention time
for the stated ranges of flow in the RBC reactor tank was 1.6
to 1.26 hours* while for the secondary settling tank it was 1.1
to, 0.9 hours. The influent BOB during the period of the stu-
dies varied from 218 mg/1 to 308 rag/1*
250
-------
TABLE -5: SHOWING THE DATA OBTAINED FROM RBC PILOT PLANT
STUDIES BDR THE TREATMENT OF SETTLED DOMESTIC
SEWAGE USING DISCS MADE UP OF ASBESTOS CEMENT
SHEETS.
PARAMETERS
AVERAGE VALUES AND RANGE
Temperature °c.
Total Plow (10-12 hrs/day),m3
Hydraulic Load, n*3/m2/day
Organic Load, g/m2/day
B.O.B. ma/1 t
Influent
Effluent
(20-30)
47.67
0.143
16.1
114
(86-035)
26
(15-45)
% reduction
C.O.P. ma/1 n
Influent
Effluent
% reduction
Influent
Effluent
% reduction
Influent
Effluent
% reduction
77
280
(210-360)
54
(30-90)
80
17.0
(13.6,19.0)
8.5
(6.5-10.4)
50
11.5
(8.5-14.0)
8.0
(6.2-11.0)
30
* The table gives the average values of 25 observations during
the operation of RBC from January, 1974 to April, 1974.
* Mo. of Discs mounted en two shafts t 136.
2
*Surface Area of each disc (on both sides) t 2.45 m
2
* Total surface area of 136 discs s 333.2 ra .
* Disc speed* 5
251
-------
TABLE-6: DATA ON THE PERFORMANCE OF RBC PILOT PLANT WITH PVC DISCS
FOR THE TREATMENT OF RAW DOMESTIC SEWAGE.
PARAMETERS
RBC tank Temp.°C.
pH of liquid in
RBC tank
Jan. '76
to
Apr. '76
1*
24-27
7.4-7.5
Oct. '76
to
Feb. '77
2 **
20-28
7.5-7.8
PERIOD OF
Feb. '77
to
June '77
3 **
26-31
7.6-7.7
OPERATION
Aug. '78
to
Mav'79
4 **
29-35
7.5-7.9
June '79
to
Aucr. '79
5** *
29-34
7.6-7.9
Total flow/day
(8-12 hrs/day)(m3)
5.77
5.58
4.49
4.54
4.99
rate , m J/m2/day
Organic loading
rate, g/m2/day
BOD, mg/1
COD, mg/1
Inf.
E2
Inf.
El
E2
v Inf.
Suspended I „
solids mg/1 I 1
Ilnf.
El
E2
(
No3 -N,mgA:
Inf.
El
* E2
Bio-mass on
Disc, cr/m2
0.090
27.7
308
98(68%)
70(77%)
573
217(58%)
132(73%)
109
54(50%)
36 (67%)
-
-
-
-
-
-
27.7
0.087
25.6
295
32(89%)
24(91%)
588
105(82%)
87 (85%)
405
44 (89%)
24 (94%)
37
22(41%)
15(60%)
Nil
1.0
1.3
34.7
0.070
22.3
319
75(77%)
55(83%)
551
127(77%)
90(84%)
355
38(89%)
26(93%)
38
30
26
Nil
2.0
2.3
50.0
0.071
15.2
216
48(78%) 86
40(82%) 45
600
134(78%) 199
98(84%) 75
314
68(78%) 152
22(90%) 44
40
23 (43%) 16
18(55%)
Nil
1.2
1.3
78.3
0.078
20.7
255
(66%)
(82%)
662
(70%)
(88%)
466
(59%)
(90%)
32
(50%)
16
Nil
2.1
4.2
63.0
No. of Discs of 1m. dia x 40; Total surface Area : 64.138 m^
Size of semi-circular contact tank: l«22m dia, 1.52m length.
Depth : 0.72 m3 3
Size of Settling tank: 1.22m x 0.91m x 0.508m; volume: 0.50m
Inf: Raw Domestic Sewage Influent
EI : Effluent from RBC Contact tank
: Final Effluent after settling tank : (%) Percentage Efficiency
The RBC plant was operated with smooth PVC discs,without cover.
The RBC plant was operated with ventilated cover and modified
settling tank*using sand coated discs
-------
There was gradual reduction of efficiencies of removal
of BOD§ and S.S as the organic feed loading rate was increased
from 15g/m2 of disc area/day to 28g/m2 of disc area/day. The
overall BODs removal varied from 77 to 91 per cent and S.S.
removal varied from 67 to 94 percent in the decreasing order of
organic loading rate. The removal of 8005 in the reactor
alone ranged from 68 to 89 per cent, Ammonia nitrogen to an
extent of 50 to 60 percent was removed in the system.
During the period of studies at various seasons of the
year the temperature of sewage water varied from 2QOC to 35°C.
The whole plant was worked under natural field conditions
without any control over temperature* organic strength etc.
Only the hydraulic flow rate was regulated.
As can be seen from Table 5, the efficiency of removal
of BODs in the reactor with smooth PVC discs was lower by about
15 to 18 percent compared to the reactor with sand coated
PVC discs. Under similar conditions with sand coated PVC
discs* the removal of BOD in RBC reactor which was open was higher
by about 10 to 15 percent to that of reactor which was
closed by ventilated lid. It was to some extent improved
with respect to BOD and S3 removal after modification of the
outlet arrangements using serrated weirs. Settling tank with
modified outlet arrangements was being used for the closed
reactort while earlier the same reactor without closed lid was
operated along with settling tank with single pipe outlet.
It may be noted that the sewage that was fed to the
system, had only preliminary treatment of screening and grit
removal and no primary settling. Perhaps, the performance
efficiency of the system could be bettered by charging
with primary settling sewage. Otherwise, by providing a second
stage smaller reactor unit in series with the existing one,
the efficiency could possibly be improved.
ORGANISMS IN THE SLIME
After the start up of the plant, about 7 to 10 days
of continuous operation was required to build up an optimum
growth of biological slime adhering to the disc and to reach
a steady state condition producing an effluent of desired
quality.
The microscopic examination of slime revealed mixed
culture of Protozoans, Rotifers, Nematodes, Filamentous
bacteria. Fungi and Algae, while particular species dominated
under different conditions of working of the reactor. Iden-
tified organisms are shown below
253
-------
PROTOZOA
P^aramaecium
ALGAE
Vorticella SB.
EpjLstyl.is ,^.spA
Ameoba pro teus
(SLaucoma sp.
Aspidisca costata
Carchesium sp.
Opercul aria sp.
Iiionotus sp.
Podophyra sp.
uneinata
Colpoda SP.
Suplotus sp.
ROTIFERS
Rotaria rotatoria
Lecane sp.
PhjLlodina sp.
NEMATODSS
Rhabditis Larvae
Doriolamus sp.
O s ci 1 1 at or ia sp.
Spirulina sp.
Phormidrum sp.
Chi amvdomonas sp.
Selenastrum sp..
Chlorella SQ.
Actlnosphaerimn SP
Anacystis SP.
Synedra sp.
i SP.
Diatoma sp.
Tabellaria sp.
Nj'fczschi.a sp.
Navicula SP.
FOiMGI
Fusariuin sp.
FILAMENTOUS BACTERIA
Sphaero-tilus na-tans
Laboratory studies with synthetic sewage showed abundance
of filamentous bacteria, SphaerQtilus natans. When the pilot
plant reactor was operated with raw sewage under ventilated cover*
the biological organisms present in the slime adhering to the
disc showed abundance of filamentous fungus, Fusariuenimsp.
and a small proportion of algae. Under conditions of pilot plant
reactor completely exposed to the atmosphere ( without the
ventilated cover) and operating with raw sewage, algae was also
found in abundance in the slime.
The average concentration of bio mass present on the
discs of the pilot plant varied from 35 to 78 g/m2 of disc area.
On an average, the total weight of the biomass adhering
to the surfaces of 40 discs in the pilot plant, was estimated
to be about 4 kgs; while the total suspended solids present
in the mixed liquor in the reactor was found to be aoout 0.25 kg,
which works out to about 6 percent of the total biomass pre-
sent in the discs. The total quantity of sludge accumulated
in the settling tank for a fixed period was measured and worked
out to about 0.4 gm per gm of BOD applied.
LOADING AND POWER REQUIREMENTS
From the results of the performance of the pilot plant*
it is possible to obtain efficiencies of purification ranging
from 82 to 90 percent for loading rates of 16 to 20 g/m2 of
disc area per day when screened and degritted raw domestic
Sewage ( pumped from municipal manhole) is applied to the system.
254
-------
Power consumption as measured by Wattmeter connected to the
drive motor of 1.5 kw capacity* worked out to 1 to 1.25 kw/hr
for one kg. of BOD removed. It is to be noted that the power
rating of the motor was much higher than that required, and as
such, the power consumption seemed to be relatively higher than
that reported elsewhere, even-though the consumption will only
be slightly less than that for conventional activated sludge
process. The plant could very well operate with a 0. 5 kw motor
instead of 1.5 kw motor.
COST ASPECTS
The break up of the cost of different components of
the pilot RBC fabricated at NEERI are as follows*
1. Geared Motor ... Rs. 5,890.00
2. 40 Nos. circular PVC
disc of 1m dia ... Rs. 4,680,00
3. Settling tank piping
& fabrication .». Rs» 4,500.00
Rs. 15,070.00
Contingencies Rs. 930.00
Total Rs. 16,000.00 < as in the year
1976). '
The cost of the system is equivalent to about 2000 US
dollars, for treating a flow of 0.5 m3/hour ( 110 gallons per
hour) of raw sewage with BOD of 250 to 300 mg/1 and working at
an efficiency range of 82 to 90%.
SUMMARY" AND CONCLUSION
1. The Rotating Biological Contactor (RBC) or Bio-disc
due to its compact construction, simplicity of operation and
favourable climatic conditions has great potential for use in
the treatment of wastewaters in India. Accordingly, studies
were conducted under lacxsratory and field conditions with the
objective of its feasibility for use in India, and formulating
relationship amongst the treatment efficiency, organic and
hydraulic loadings, rotational speed and power consumption.
2. Performance studies were carried out in a laboratory
Rotating Biological Contactor using synthetic and domestic
sewage for nearly two years period, for two types namely, com-
posite Imhoff tank type reactor with settling tank below the
reactor, and RBC reactor with separate settling tank.
255
-------
3. Later* studies were carried out using raw degritted and
screened municipal sewage, on the performance of RBC pilot
plant with 40 PSC circular discs of one meter diameter followed
by a reetangular settling tank. The performance characte-
ristics were studied under two conditions viz. when the reac-
tor was open* and when the reactor was enclosed by ventilated
cover.
4. The operation of the pilot plant was carried out under
field conditions, and there was no control as regards the strength
of wastewater and temperature. The RBC was operating conti-
nuously, while the sewage flow was limited to 8 to 10 hours
during the day time, feeding the reactor at a constant rate
of 0.5 m^/hour ( variation ± 10 percent) under ambient con-
ditions and temperature of wastewaters varying from 20°C to
35°C.
5. Under the tropical climatic conditions prevailing at
Nagpur* India, the RBC pilot plant achieves overall efficien-
cies of removal of 82 to 90 percent, for BOIfcj a^ 20°C, and
90 to 93 percent for Suspended solids at an organic loading
rates of 16 to 20 g/m^ of disc area/day and hydraulic loading
rate of 0,07 to 0.08 m3/n>2/day when the applied feed of raw
screened and degritted domestic sewage has BOD concentration
of 250 to 300 mg/1, suspended solids concentration of about
400 mg/1 and the temperature of sewage in the reactor varied
from 200 to 33°C for different seasons of the year. Other
design parameters include the submergence of 45 percent for the
discs, rotational speed or 3 to 5 rpm, effective detention
time in the reactor of 1.5 hours.
6. Such a compact unit occupying an overall area of 1%
Meter by 4% Meter can as well serve a population of 50 to 100
persons depending on the water consumption. A 0.5 K.W.
motor with reduction gear and belt drive would suffice to rota-
te the discs.
7» The desludging from the settling tank need to be car-
ried out only once in a week or two weeks. The quantity of
sludge produced is estimated at 0.4 kg per kg of BOD removed
in the system.
8* There were no major mechanical troubles in the
RBC Pilot Plant during its continuous operation for two years.
The bearings have to be occasionally lubricated or replaced.
9» The cost of the PVC discs ( in India) form nearly 35
to 40 percent of the total cost of the RBC system. The cost
can be reduced by using alternative cheap materials like split
bamboo* aluminium sheet etc. for discs. By using FVC discs,
the per capita capital cost works out to Rs. 160 for the unit
serving 100 persons.
256
-------
10. The power consumption varied from 1 to 1.2 KWH/kg BOD
removed, which could be brought down by using a lower horse
power rating motor drive.
11. Further field studies on pilot plant scale should be
continued with alternative materials for disc, feasibility
of use of wind mills as drive mechanism for rotation of discs,
increasing loading rates, additional surfaces, methods to
decrease power consumption, and using the RBC as based on
extended aeration principle for reduction of sludge volume and
its easy disposal.
12. Open type of RBC gives higher efficiency for BOD
removal than that for the enclosed type.
13. The RBC can be successfully operated with screened,
degritted sewage, avoiding primary settling tank and the pro-
blem of disposal of primary sludge. The efficiency can be
improved by working the RBC in stages.
ACKNOWLEDGEMENTS
The studies could be carried out due to the enthusiasm
and encouragements shown by Prof. N. Majumder, Ex-Director,
NEERI and Dr. B.B, Sundaresan, Director, NEERI, NagpurCIndia).
REFERENCES
1. Borchardt, J.A., 'Biological Waste Treatment Using Rota-
ting Discs', Biological Waste Treatment edited by R.P.
Canale, Interscience Publishers, N.Y ( 1971 ).
2. Antonie, R.L., Kluge, D.L. & Mielke, J.H., 'Evaluation of
a rotating disc waste water treatment plant1 JWPCF, 46
3, March 1974.
3. Antonie, R.L, Kochler, F, J. * Application of Rotating disc
process to Municipal Waste Water Treatment* U.S. EPA,
Project No. 17050 DAM/11/1971.
4» Antonie, R.L., 'Response of the Bio—disc process to fluc-
tuating wastewater flows1 25th Purdue Univ. Conf. May 1970.
5. Torpey, W.N. et al 'Rotating disc with biological growth
prepare waste water for disposal or reuse*, JWPCF,
Nov. 1971.
6. Khan A.N., Siddiqi R.H., 'Treatment of Waste water by
Biological Discs* Indian J. Environmental Health* _14,
4, 289-296, 1972.
257
-------
7. Raman V.? KJ*Mn A.N., 'Sewage Treatment for small communi—
ties: Indian Association for Water Pollution Control,
Technical Annual No. 4, 1977*
8. Technical Digest, National Environmental Engineering
Research Institute, Nagpur, India No. 60, April 1978.
9» Pescod M.B., Nair J.V., 'Biological disc filtration for Tro-
pical Waste Water Treatment', Water Research, g. 12, 1509-
1523, Dec. 1972.
10. Pretorius, W.W., 'Rotating Disc unit - A waste Treatment
system for small communities'. Water Pollution Control,
72(b), 1973.
11. Steels, I.H., 'Design basis for the rotating disc process
Eff. Wat. Treat. J. 14* 431, 1974.
12» Bruce A.M.j Merkens J.C.< 'Developments in Sewage Treat-
ment for small communities1, Proc. Eight P.H.E. Conf.
Loughborough Univ. of Technology, January 1975. (Publi-
shed in Aspects of Sewage Treatment pjoceedings, eidted
by John Pidcford),
13. Clark, J.H., Moseng I«M.» Asano T.,' 'Performance of a
Rotating Biological Contactor under varying waste water
flows'. JWPCF, 5_, 50, 896, May 1978.
14. KLuge D.L., Kipp R.J., 'Evaluation of the RBC process for
Municipal Wastewater Treatment U.S. EPA-600/2-78-028,
March 1978.
258
-------
HIGH SALINITY
USING
ROTATING BIOLOGICAL CONTACTORS
By
Nancy E. Kinner
Paul L. Bishop
Department of Civil Engineering
University of New Hampshire
Durham, New Hampshire, U.S.A.
Introduction
Design of wastewater treatment facilities for small, offshore com-
munities presents unique engineering problems associated with conventional
biological treatment. Islands are often inhabited by seasonal populations
Isolated from the basic services available on the mainland. Limited fresh-
water supplies are frequently conserved by using seawater as a carriage
medium for sanitary wastes. The salinity of the domestic wastewater can vary
from 0% to 3.5$ depending on the source of the waste at a given time. To
further complicate design problems, systems must be able to adapt to varying
flows, little maintenance and intermittent nature of island generated elec-
trical energy.
The inherent flexibility in the operation of the rotating biological
contactor (RBC) process makes it an ideal candidate for this set of con-
ditions. As part of an evaluation of the ability of an RBC unit to treat
saline domestic wastewater, mierobial communities attached to disk surfaces
were studied. Disk populations growing in freshwater and seawater based
sewage were compared as a function of their distance from the influent end
of the unit and the hydraulic loading rate.
Most previous research on saline domestic wastewater treatment has
been conducted on the activated sludge process. Using a bench scale unit
259
-------
Ludzaek and Noran (l) found that free swimming ciliates survived in a con-
tinuous high chloride artificial waste made from NaCl (3.55? salinity). When
the chloride content alternated between high and low concentrations, hypo-
trichs and stalled ciliates occurred, but the latter were small and in-
active. The activated sludge took from one to five weeks to adapt to steady
state conditions during saline treatment. Kincannon and Gaudy (2) fed arti-
ficial waste of 3.0$ salinity to a batch activated sludge system. They found
a 30$ decrease in substrate removal efficiency. When the salinity increased
to 4.5%, operation was impaired. They hypothesized that a change in species
composition might occur in saline treatment. In subsequent research, Kincannon
and Gaudy (3) used a continuous flow activated sludge unit to assess the
treatment capabilities of organisms exposed to an artificial waste containing
NaCl (salinity 3.0$ and 4.0$). After a period of acclimation of up to two
days, the system achieved excellent removal capacity. In another bench scale
study of an activated sludge treatment plant, Burnett (4) conducted research
using domestic waste mixed with seawater. Acclimation occurred within thir-
teen days after changing from freshwater to seawater based sewage (salinity
3-2% to 3.8$). The system operated well under these conditions. He noted
a rapid die-off of rotifers and stalked ciliates and motile ciliates with
increasing salinity. After the first few days of acclimation, motile ciliates
reappeared, but the stalked ciliates and rotifers remained absent. Kessick
and Manchen (5) experimented with another bench scale activated sludge
system and found that freshwater seed bacteria could adequately treat the
soluble fraction of the domestic sewage containing artificial sea salts
(salinity 3.6%), Recently Tokuz and Eckenfelder (6) observed a similar treat-
ment capacity between a bench scale activated sludge unit fed artificial
waste at 0% salinity and one at 5.0$ salinity.
Research on the effect of salt concentration on the treatment capacity
of trickling filters has been more limited. Stowell (7) reported that an
intermediate type high rate filter with a 1:1 reeireulation ratio achieved
BOD removals up to 90$ when treating waste from San Quentin Prison (salinity
1.1/f to 1.5$). Lawton and Iggert (8) found that a bench scale trickling fil-
ter treating an artificial waste was able to recover in one day after the
salinity was increased to 2.0$ with NaCl. Mills and Wheatland (9) found
no change in the removal efficiency of a bench scale trickling filter after
adding NaCl (final salinity 1.2$) to an artificial waste. In this study
an intermittent application of waste of salinity 1.2$ to 3.6$, did cause
a decreased efficiency.
Rotating biological contactors are a relatively new waste treatment system
compared with the activated sludge and trickling filter processes. Therefore
research into the various applications of the unit has been rapidly expan-
ding. The major work done on BBC saline waste treatment is reported by Mikucki
and Poon (10), Poon and Mikucki (ll) and Poon, Chao and Mikucki (12). A
pilot plant study was conducted using a mixture of domestic waste and an
artificial supplement and seawater with a final salinity of 2.1$. This
system worked well in removing BOD after a few days of acclimation. Poon,
Chao and Mitaicki (12) reported that the growth on the disks contained
eatuarine forms of filamentous fungi and algae.
A review of the literature to date reveals that little information is
260
-------
available on the aicrobial populations inhabiting saline waste treatment
systems. Populations for our evaluations were obtained from an BBC pilot
plant fed settled domestic sewage mixed with artificial sea salts to a
salinity of 1.0$ to 2.0%. A qualitative analysis was conducted using photo-
microscopy to assess the difference in community structure between fresh-
water and salt water based samples. To avoid lowering the BOD and to closely
simulate seawater based sewage, a small volume of highly concentrated artifi-
cial seawater was added directly to the waste.
The qualitative differences between the populations occurring on the
disk surfaces in the freshwater and salt water based sewage appeared minimal.
Under both conditions the community structure consisted of a primary sub-
strate composed of filamentous organisms and a zoogleal mass, which sup-
ported active populations of rhizopods and stalked ciliates. At an hy-
draulic loading rate of one gallon per day per square foot (gpd/ft ) stalked
ciliates occurred in all four compartments of the units. At a rate of 2
gpd/ft2 these organisms did not appear until- the second compartment.
Materials and Methods
A pilot plant consisting of four separate rotating biological contactors
was assembled in the Durham, New Hampshire sewage pumping station. Each RBC
consisted of a tank made from a plexiglass half cylinder four feet long and
eight inches in diameter. Each was divided by plexiglass plates into four
separate compartments. Wastewater flowed from one compartment to the next
over notched weirs in the dividing plates. A horizontal stainless steel
shaft supported 64 disks, each with a seven inch diameter, for a total sur-
face area of 34 square feet. Two of the units had disks made of polyethylene,
while the other two units had polyurethane sealed masonite disks. The disks
in all units were equally spaced, with 16 disks per compartment: All units
were rototated at 12 revolutions per minute yielding a peripheral speed of
0.3*7 feet per second. The temperature in the pump station decreased during
£he course of the experiments and the wastewater temperature ranged from 21
C to 13.5 C. The units were exposed to a constant low level incandescent
light of less than 10 footcandles. 2Tne hydraulic loading rate for one unit
with each type of disk was 1 gpd/ft j total retention time of wastewater in
these RBC's was 3.6 hours. The other two units had an hydraulic loading rate
of 2 gpd/ft^ and a retention time of 1.& hours. All disks had approximately
40% of their surface area submerged in the wastewater at a given time .
Raw sewage was taken from the pump station channel just prior to the bar
rack and comminutor. The liquid was filtered through a i inch wire mesh and
a 1/8 inch wire mesh to remove large particulates. It was then pumped in 4
inch plastic hose to a 30 gallon plastic primary clarifier ( retention time
3.5 hours). Sludge from the clarifier was removed every two to three days.
The flow of the settled sewage was pumped into four separate lines of i inch
polyethylene tubing, each of which supplied an individual constant head feed
tank. Influent was delivered directly to the first compartment of each RBC
unit by gravity flow through 1/6 inch polyethylene tubing from the over-
head feed tank. Flow rate was controlled by stopcocks inserted in the influent
lines .
261
-------
The system described above was modified slightly during salt water oper-
ation. Concentrated artificial seawater (salinity 13.3/0 was made daily by
preparing a saturated solution of Utility Marine Mix^ and tap water. The salt
water flowed by gravity from two 30 gallon reservoirs (plastic) into the lines
carrying settled sewage to the constant head feed tanks. Flow rate was reg-
ulated by an in-line valve. The salinity of the salt water based sewage
ranged from 1.0$ to 2.0%.
During the startup period the units received freshwater settled sewage.
After approximately two weeks the units reached steady state operation as
assessed by BOD sampling. The disk populations were removed and observed as
outlined below. The system was then switched to the salt water based waste.
After one week steady state was achieved and samples were removed and observed.
The system was returned to the freshwater based sewage and the entire pro-
cedure repeated.
BOD tests were done on the soluble fraction of the waste to determine
influent and effluent quality. Samples were filtered through Whatman 40
filter paper or the equivalent. For freshwater based sewage the procedure
outlined in Standard Methods (13) was followed. For salt water based sew-
age Martin's (.14) procedure was used with a dilution water of Q% salinity.
Dissolved oxygen was measured with a YSI 51A Oxygen Probe and Meter^.
Samples for microscopic examination were scraped from the surface of
one disk in each compartment of the RBG's. The scrapings weighed 2 to 3
grams. We assumed that each compartment was completely mixed and therefore
the microbial population would be fairly uniform regardless of position with-
in the compartment. Samples were immediately transferred to sterilized
plastic bags. A few milliliters of wastewater from the* -.sampled compartment
was added to the bag. The samples were sealed and stored in a refrigerator
in the laboratory at 4.4°C until examined. Observations usually occurred
within 24 hours, however some samples were held up to one week with no visible
change in composition upon microscopic examination. Random samples taken from
each bag were placed on a glass slide and covered with a glass cover slip.
In some cases one or two drops of 10$ MgCl2 were, added to slow the protozoans.
All fields on a slide were examined with an Olympus BHA microscope. Ob-
servations were recorded in the form of written notes and photomicrographs.
Kodak Ektachrome ASA 64 daylight film was used with an LED and two LD 45
filters for all photomicroscopy. Resources used in identifying organisms
included Kudo (15), 'Bergey's jfaiual (16) and Barnes (17). Identifications
were subsequently confirmed ^""specialists in protozoology and micro-
biology.
Results
The soluble influent BOD averaged 150 mg/1. It showed a typical diurnal
pattern for a small town; lower concentration in the daytime and higher at
night . During steady state operation on freshwater and salt water based
sewage the soluble effluent BOD from the RBC's was' consistently below 30
2 Utility Chemical Company, Paterson, New Jersey
YSI Company, Yellow Springs, Colorado
262
-------
mg/1.
Tables 1 and 2 list the general types of organisms found on the disks
during steady state conditions when the feed was freshwater and salt water
"based sewage, respectively.
The populations present during freshwater and salt water operations
were very similar. A filamentous organism grew on all the disks, forming
a dense mat varying in color from white to dark brown. Growth was patchy
in the last two compartments of the RBC's. The other predominate population
was a zoogleal mass of bacteria. Both the filaments and the zoogleal mass
were embedded in a thick mucilage. Zooflagelates and rhizopods (amoebae)
were also observed in all compartments, under both conditions, as were
fungal fruiting bodies and unidentified cysts. The fruiting bodies and cysts
were found sporadically. Fewer rotifers and nematodes were observed during
the saline conditions.
r\
At an hydraulic loading rate of 1 gpd/ft'' motile and nonmotile peritrichs
were present throughout the entire unit. The nonmotile peritrichs included
solitary and colonial stalked forms. At the higher loading rate (2 gpd/ft*-)
peritrichs did not appear in the first compartment, but were present in all
the remaining compartments.
There was no difference observed between the populations grown on the
polyethylene and masonite disks.
Discussion
"There has been, and still is, too much engineering and too little micro-
biology in this field of environmental sanitation" (IS). Fifteen years later
this generalization is still applicable. One reason for the lack of micro-
biological research is the difficulty encountered in isolating and identifying
specific organisms occurring in wastewater treatment systems. Though
Sphaerotilus (19, 20), Beggiatoa (20), and fungi (12) have all been reported
as major components of RBC microbial communities, there have been no reported
instances of successful culturing of these organisms from the disks. Iden-
tification of genera by microscopic examination is not a foolproof micro-
biological method. Therefore we have not attempted to specify the type of
filament growing on the disks in our experiments. Research is currently under-
way in our laboratory to isolate and culture the filaments and make a positive
identification of the species.
The identification of zoogleal bacteria and specifically Zooglea ramigera
is difficult, though it has been reported growing in RBC's (19~ECrabtree
and McCoy (21) conclude that Z. ramigera may consist of a heterogenous popula-
tion of bacteria. This view is supported by Una and Dondero (22), who report
that two forms of zoogleal mass exist. The most predominant and "biochemically
active" form is composed of nonzoogleal bacteria (22). In our research we
have not attempted to identify the species of bacteria present in the zoo-
gleal mass. We can only confirm that a zoogleal type of growth occurs on the
disks.
We know relatively little about the types of organisms growing on the
263
-------
Table 1. Organisms Present in RBC's Treating Freshwater Based Sewage
Hydraulic
Loading
Rate ,,
gpd/ft^
1,0
2.0
Compartment
1
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Metazoans
Nematodes
Compartment
2
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Plants
Oomycetes
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Compartment
3
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Unidentified cysts
Compartment
4
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zoof lagellate s
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Plants
Phycomycetes
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Unidentified cysts
-------
Table 2. Organisms Present in RBC's Treating Salt Water Based Sewage
10
a*
ui
Hydraulic
Loading
Rate ~
gpa/f-r
1.0
2.0
Compartment
1
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Ehizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Compartment
2
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Unidentified cysts
Compartment
3
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Plants
Phycomycetes
Unidentified cysts
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Unidentified cysts
Compartment
4
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Metazoans
Nematodes
Rotifers
Unidentified cysts
Monerans
Filaments
Zoogleal mass
Animals
Protozoans
Zooflagellates
Rhizopods
Holotrichs
Peritrichs
Unidentified cysts
-------
BBC disks, and their community interactions. The disks provide a surface to
which filamentous organisms adhere. It appears from our observations that
the filaments provide a primary substrate for all the other members of the
community including nonfilamentous bacteria, protozoans, rotifers and nema-
todes. The primary substrate serves two functions. 1. As a habitat, the
filaments provide a refuge for bacteria from large celled predators (24) as
well as a substrate to protect the other microorganisms from the shearing
force of the fluid as the disks are rotated. 2. The filaments also are a food
source for some of the protozoans (25). The relationships between filament-
ous forms and their associated inhabitants are noted by several authors (25,
26, 27).
The protozoans may play several roles in the community structure. We
have adapted the theories developed by Reid (28) to explain protozoan activity
on the RBC disks. Protozoans are mainly carnivorous. As primary carnivores
they prey on free bacteria, helping to maintain bacterial activity by control-
ling the size of the bacterial population. As secondary carnivores, some
protozoans eat the primary carnivores preventing excess predation. Proto-
zoans also assimilate bacterial metabolic by-products, contributing to remov-
al efficiency. This simple model of the community structure of the disk
growth provides a basis for future research on BBC microbial ecology.
In our observations we found little difference in the organisms present
during freshwater and salt water based sewage treatment. Only the rotifers
and nematodes appeared to be present in reduced numbers under saline conditions.
In two previous bench scale activated sludge experiments on saline sewage,
rotifers were absent (4), and stalked ciliates were absent (-4) or reduced in
size and inactive (l). We are unable to explain the reasons for the inhibitions
of nematodes and rotifers and the ability of stalked ciliates to survive under
saline conditions in our RBC experiments. It is possible that salinity toler-
ance is species specific (25) and/or that some inherent characteristic of the
RBC process may mitigate the inhibition of the stalked ciliates. Further re-
search should be conducted in these areas.
The bactericidal action of the seawater on non-marine species is well
known (29, 30, 31). In previous experiments on saline biological waste treat-
ment it has been assumed that marine and estuarine organisms would predomin-
ate due to the die-off of freshwater forms (4,12, 32). Many factors contrib-
ute to the die-off of non-marine microorganisms in seawater. The two major
factors, Jones (33) concludes, are the low nutrient concentrations in sea-
water, and heavy metal toxicity. In continuous culture in seawater, E. coli
can outcompete marine bacteria if substrate concentrations are similar to
those in sewage (34). In saline sewage the negative effects of seawater are
mitigated, as the nutrient concentration is high and the heavy metals are
complexed by organics (33). We observed no change in species composition under
saline conditions. We think that the freshwater sewage bacteria are able to
survive because the environment they "perceive" is not inhibitory. We are
presently conducting further research on heavy metal complexation in sa-
line sewage.
The existence of active peritrich populations on our disks indicates that
good treatment was occurring (23, 36). This is supported by the low soluble
BOD's we measured (less than 30 mg/l). The organisms we observed were similar
266
-------
to those found by other researchers in activated sludge (36, 37), trickling
filters (38, 39) and RBC's (19, 20) which also exhibited good BOD removals.
Our microbial populations did not show a distinct succession along the length
of the unit, which had been observed by others (19, 20). There may be a change
in the bacterial species which we could not determine in these experiments.
Identification and quantification of organisms and their position along the
unit should be examined as it affected by RBC operating parameters.
Our findings confirm those of Poon and Mikueki (11) that RBC's can achieve
BOD removal of greater than 30 mg/1 when treating saline domestic wastewater.
RBC's do offer a possible treatment alternative for small offshore communities
which use seawater as a carriage medium for sanitary wastes. We found that
microbial populations on the disks during the treatment of salt water based
sewage were similar to their freshwater counterparts. Under both conditions
peritrichs were present, indicating a healthy and active microbial growth.
We developed acsimple explanation of the interrelationships among disk
microorganisms. Many questions remain unanswered and we hope that continued
research on the microbial populations and their ecology inEBC's will help
provide the solutions.
References
1. Ludzack,F.J. and D.K. Noran. " Tolerance of high salinities by conventional
wastewater treatment processes", J1PCF 37(10):1404 (1965).
2. Kincannon,D.F. and A.F. Gaudy, Jr.. " Some effects of high salt concentra-
tions on activated sludge", JWPCF 38(7):1148 (1966).
3. Kincannon,D.F. and A.F. Gaudy, Jr.. " Response of biological waste treat-
ment systems to changes in salt concentrations", Biotech. Bioengg. 10:483
(1968).
4. Burnett,I.E.. " The effect of salinity variations on the activated sludge
process", Water and Sewage Works 121:37 (1974).
5. Kessick,M.A. and K.L. Manchen. " Salt water domestic waste treatment", JWPCF
48(9):2131. (1976).
6. Tokuz,R.H. and W.W. Eckenfelder, Jr.. " The Effect of F/M Ratio on an Activa-
ted Sludge System Treating High Salinity Wastewater", Proceedings of the
33rd Industrial Waste Conference, Purdue University, West Lafayette,
Indiana, May 1978.
7. Stowell,E.R.. " Sewage Treatment in California Institutions", Sewage and
Industrial Wastes 26(11):1347 (1954).
8. Lawton,G.W. and C.Y. Eggert. " Effect of High Sodium Chloride Concentration
on Trickling Filter Slimes", Sewage and Industrial Wastes 29(11):1228 (1957),
9. Mills,E.V. and A.B. Wheatland. " Effect of Saline Sewage on the Performance
of Percolating Filters", Waste and Waste Treatment Journal 9:170 (1962).
10. Mikucki,W.J. .and C.P.C. Poon. " An Evaluation of the Bio Surf Process for the
Treatment of Saline Domestic Wastewaters", Proceedings of the 8th Annual
Offshore Technology Conference, Houston, Texas, May 1976.
11. Poon,C.P.C. and W.J. Mikucki. " Rotating Biological Contactors Treat Island's
Saline Sewage", Water and Sewage Works :62 February (1978).
12. Poon,C.P.C., Y. Chao and W.J. MikucM. " Factors Controlling Rotating Biologi-
cal Contactor Performance", JWPCF 51(3):601 (1979).
13. American Public Health Association. Standard Methods for the Examination of
Water and Wastewater, 14th editIon. Washington, D.C. (1976T.
14. Martin, D'.F..Marine Chemistry, Volume 1, Marcel Dekker Inc.,.New York (1968).
267
-------
15. Kudo,R.R.. Protozoology-, 5th edition. Charles C. Thomas, Springfield (1966).
16. Holt,J»G. (ed.). The Shorter Sergey's Manual of Determinative Bacteriology,
8th edition. Williams and Wilkins Company, Baltimore (1977).
17. Barnes,R.D.. Invertebrate Zoology, 3rd edition. W.B. Saunders Company, Phila-
delphia (1974T
18. Wuhrmann,K.. " Microbial Aspects of Water Pollution Control", Adv. in Appl.
MicroMol. 6:119 (1964).
19. Torpey,W.N., H. HeukellMan, A.J. Kaplovslcy and R. Epstein. " Rotating Disks
with. Biological Growths Prepare Wastewater for Disposal or Reuse", JWPCF
43(U):2181 (1971).
20. Pretorius,W.A.. " Some Operational Characteristics of a Bacterial Disc Unit",
Water Research 5:1141 (1971).
21. Crat>tree,K. and E. McCoy. " Zooglea rajaigera Itzigsohn, Identification and
Description", Int. J. Syst. Bacteriol. 17(1):1 (1967).
22. Un3,R.F. and N.C. Dondero. " Nonzoogleal Bacteria in Wastewater Zoogleas",
Water Research 4:575 (1970).
23. Jones,A.R.. The Ciliates. St. Martin's Press, New York (1974).
24. Alexander,M.. MicrobiaT~E.cology. John Wiley and Sons, Inc., New York (1971).
25. Sleigh, M.. The Biology of Protozoa. Edward Arnold Ltd., London (1973).
26. Dondero,N.C.. " The Sphaerotilus-Leptothrix Group", Ann. Rev. Microbiol. 29:
407(1975).
27. Curtis,E.J.C.. " Sewage fungus: Its nature and effects", Water Res. 3:289
(1969).
28. Reid,R.. "Fluctuations in Populations of 3 Vorticella Species from an Acti-
vated-Sludge Sewage Plant", Protozool. 16U):103 U969).
29. Zobell,C.E.. " Bactericidal Action of Sea-Water", Proc. Soc. Exp. Biol. Med.
34:113 (1936).
30. Carlucci,A.F. and D. Pramer. " An Evaluation of Factors Affecting the Survival
of Escherichia coli in Sea Water I. Experimental Procedures", Appl. Microbiol.
7:388 (1960).
31. Qrlob,G.T.. " Viability of Sewage Bacteria in Sea Water", Sewage and Industrial
Wastes 28(9):1147 (1956).
32. Petros,J.K. and E.M. Davis. " Organic degradation by biological treatment of
hypersaline wastewaters", Proceedings of the 31st Industrial Waste Con-
ference, Purdue University, West Lafayette, Indiana, May 1976.
33. Jones,G.E.. " The fate of freshwater bacteria in the sea", Dev. Ind. Microbiol.
12:141 (1971).
34. Jannasch,H.W.. " Competitive elimination of Enterobacteriaceae from seawater",
Appl. Microbiol. 16(10):1616 (1968).
35. Curds,C.R., Cockburn,A. and J.M. Vandyke. " An experimental study of the role
of the ciliated protozoa in the activated sludge process", Wat. Pollut,
Control 67:312 (1968).
36. 5ydenham,D.H.J.. "A re-assessment of the relative importance of ciliates,
rhizopods and rotatorians in the ecology of activated sludge", Hydrobiol.
38:553 (1971).
37. Varma,M.M., Findley,H.E. and G.H. Bennett. " Population dynamics of protozoa
in wastewaters", JWPCF 47(1):85 (1975).
38. Lackey,J.B.. " A •'study of soae ecological factors affecting the distribution
of protozoa", Ecol. Monogr. 8(4):501 (1938).
39. Curds,C.R. and A. Cockburn. " Protozoa in biological sewage treatment processes.
I. A survey of the protozoan fauna of British percolating filters and activated
sludge plants", Water Res. 4:225 (1970).
268
-------
FULL-SCALE ROTATING BIOLOGICAL CONTACTOR
FOR SECONDARY TREATMENT AND NITRIFICATION
BY
JOHN A. HITDLEBAUGH, P.E.
ROY D. MILLER, Ph.D., P.E.
US ARMY ENVIRONMENTAL HYGIENE AGENCY
ABERDEEN PROVING GROUND, MARYLAND
INTRODUCTION
A full-scale RBC system, designed to achieve effluent limitations of
10 mg/1 BOD_ and 2 mg/1 NH»-N for domestic wastewater, performed at less than
design expectations under both summer and winter conditions. This RBC system
constitutes the biological treatment portion of a 6 MGD wastewater treatment
plant serving a major US Army installation (effective population of 40,000)
and consists of 36 stages arranged with 6 treatment trains of 6 stages each.
The design hydraulic loading at 6 MGD is 1.33 gpd/sq ft of -media surface.
During the summer (wastewater temperature of 26 C and average plant flow of
3.7 MGD), the treatment plant effluent BOD;, and NH,-N levels were higher than
NPDES permit limitations and design expectations. High levels of effluent
BOD,, resulted primarily from oxygen demand of suspended solids and nitrifica-
tion in the BOD bottle. In fact, effluent soluble-BOD- was consistently
measured at less than 5 mg/1. High levels of effluent NlL-N resulted from
DO limiting conditions (less than 1 mg/1) in several RBC stages and-from rela-
tively low pH (less than pH 7.0) in latter RBC stages. During the winter
(wastewater temperature of 13 C and average plant flow of 4.6 MGD), RBC per-
formance actually improved; DO limiting conditions did not exist. The RBC
system effectively removed soluble organic material during the winter and was
more effective in removing NH —N than during the summer. NH_—N levels of
design expectation were, however, still not met. Approximately 370 Ibs/day of
269
-------
NKL-N were removed during the summer evaluation period while approximately
500 Ibs/day of NHL—N were removed during the winter — a 36 percent improve-
ment. This improved performance occurred during the winter even though
hydraulic and organic loads were higher and biological activity was considered
relatively low due to lower wastewater temperature.
Conclusions from these studies were that low DO levels in initial RBC
stages and low pH levels in latter RBC stages adversely affected biological
activity. The importance of evaluating performance in each stage of an RBC
system was shovm necessary to effectively judge design criteria and pinpoint
operational problems. In addition, both soluble (filtered) and carbonaceous
(nitrification suppressed) BOD- should be used during evaluations, particu-
larly where systems are designed for nitrification. Finally, the dependence
of nitrification on prior soluble-BOD- removal was highly evident. Recommen-
dations for future RBC system designs include use of supplemental aeration to
overcome limiting DO levels and chemical feed to maintain optimal pH levels.
BACKGROUND
The 6 MGD, domestic wastewater treatment plant had been upgraded from a
trickling filter system for secondary treatment to a rotating biological con-
tactor (RBC) system for secondary treatment and nitrification. Effluent permit
parameters and limitations for the plant discharge are listed in Table 1. The
upgraded treatment plant flow diagram and RBC system are shown in Figures 1 and
2, respectively. As indicated, the upgraded plant consists of a bar screen,
Parshall flume, comminutor, aerated grit chamber, primary clarifiers, RBC system
w/pump station, secondary clarifiers, chlorine contact chambers and a step
aerator. Two anaerobic digesters (high-rate and secondary), together with a
vacuum filter and sludge drying beds, are used for sludge handling and disposal.
Secondary clarifier sludge (recirculated flow), digester supernatant and vacuum
filter filtrate and washings are all returned to the head of the plant. The
plant has a maximum hydraulic capacity of 18 MGD. The RBC system consists of
36 stages arranged in a configuration of 6 treatment trains with 6 stages each;
3 of regular density and 3 of high density media (see Figure 2). Although each
unit operation of the treatment plant was evaluated during both summer and
winter conditions, this paper addresses only the RBC system.
LITERATURE REVIEW
RBC Treatment Process
The RBC process consists of a series of plastic disks of which 40 percent
of the surface area is immersed in wastewater (see Figure 3). As the disks
rotate, the entire media surface develops a culture of microbiological organ-
isms. The organisms adhere and multiply to form a uniform growth referred to
as a fixed-film. The biomass supported by the plastic media picks up a thin
layer of nutrient laden water as it rotates through the wastewater. The film
of water trickles over the microorganisms which remove dissolved organics and
oxygen. The rotation of the media through the wastewater not only allows for
aeration and mixed liquor, but also provides shear forces which cause sloughing
of excess growth.
RBC units commonly operate in series with the number of units depending on
the organic and/or hydraulic load to be treated. The function of the first
stages is to remove organic material, with subsequent stages removing ammonia
270
-------
In cases where nitrification is used to meet effluent NEL-N standards. Nitri-
fication usually does not begin until the soluble—BOD,, level and corresponding
large populations of heterotrophic organisms have been adequately lowered. The
actual reason that heterotrophs and autotrophic nitrifiers do not co-exist in
equal quantities throughout successive RBC stages is not clearly understood,
but it is reported by1sQme that the activities of. the two populations do not
occur simultaneously. * * However, others report that optimal pH levels for
nitrification (e.g., pH 7.0 -8.5) appear to favor initiation of nitrification
simultaneously with low-level soluble-BCD,, removal; whereas, nitrification at
suboptimal pH levels,(e.g., pH 6.6-7.0) is not initiated until soluble—BOD_
removal is complete. * The amount of nitrification achieved has been corre-
lated to the hydraulic loading of the system, usually expressed as the volume
of wastewater applied to a square measure of surface area per day. One to j-
4 gpd/sq ft have often been used as standard loading rates for pilot plants
and full scale wastewater treatment facilities. * *
The change in hydraulic load also changes the organic load as more food is
introduced to the active component of the waste treatment system. It has
recently been suggested that shortcomings observed in the quality of treatment
by RBC units was due to excessive organic loading, while operating at less than
hydraulic design capacity. The question of which parameter, hydraulic loading
or organic loading, to use for proper design and operation of an RBC process
has not been resolved.
As with other biological processes, sufficient dissolved oxygen (DO) must
be available in the wastewater within the RBC system to insure adequate treat-
ment for BOD removal and nitrification. Wastewater DO'levels of 1 to 2 mg/1
are generally considered to be the minimum requirement to avoid DO limiting
conditions. Frequently, RBC systems have been designed to provide oxygen mass
transfer via disk rotation through the wastewater and air. However, in some
cases, this has been considered a shortcoming of the process since supplemental
oxygen must sometimes be provided to prevent DO limiting conditions.
Oxygen Demand of Wastewater
A major criterion used to determine the extent of pollution of receiving
waters is the measurement of oxygen required for the stabilization of organic
matter present in the system. The total amount of oxygen necessary to stabi-
lize a waste is referred to as the oxygen demand. The ultimate oxygen demand
includes not only the amount of oxygen required to stabilize oxidizable carbon-
aceous materials, but also that which is required to microbially transform
ammonia-nitrogen to nitrate-nitrogen. For untreated domestic sewage there is
little oxygen demand by nitrifier populations for the first 8 days of stabili-
zation. Therefore, the BOD,, test is normally considered as representing the
oxygen damand of carbonaceous material. However, total BOD,, is a poor
indication of treatment where a significant population of nitrifying bacteria
are present. For sewage that has received secondary treatment and nitrifica-
tion, conversion of ammonia to nitrate in the BOD bottle may significantly
increase the BOD_ measurement and erroneously indicate a lesser degree of
treatment than that actually received.
In the RBC system, as other biological treatment processes, nitrifying
organisms may be interspersed among the heterotropic population which utilize
carbonaceous materials. The relative concentrations of both populations, at
any specific point in the treatment train, are a consequence of the nutrient
271
-------
supply, environmental conditions (pH and temperature), and the degree of treat-
ment received. Therefore, to adequately assess treatment performance, it is
necessary to know how much of the observed oxygen demand in the BOD,, test was
required to stabilize the carbonaceous materials. The presence of autotrophic
bacteria complicates the BOD_ measurement at the end of the treatment train
and gives "false positives" when testing for regulatory compliance.
The purpose of biological treatment relative to carbonaceous material is
the conversion of soluble organics to particulate bacteria. However, the un-
filtered BOD,, test represents a measure of soluble as well as insoluble organic
matter and NH,-N oxidation. Biological treatment need not be applied to removal
of colloidal and suspended organics. Suspended solids that contribute to oxygen
demand can be removed by physical—chemical processes such as gravity settling
and filtration. The practical consequence is that optimal treatment for removal
of oxygen demand may be removal of suspended solids and not biological treat-
ment. The use of filtered and unfiltered BOD tests should indicate relative
fractions of oxygen demand as originating from soluble or particulate material.
In addition, the filtered BOD test should not undergo nitrification, because
initial nitrifying populations in the BOD bottle would be reduced to Insignif-
icant levels and the BOD bottle is subsequently seeded with raw sewage (i.e.,
heterotrophic bacteria). Thus, the unfiltered BOD,, test is an unreliable param-
eter from which to judge biological treatment performance.
Nitrogen Control
The principle of biologically induced nitrogen removal in wastewater treat-
ment facilities is wholly based on the activity of populations of autotrophic
nitrifying and denitrifying bacteria and their capability to sequentially
oxidize and reduce nitrogen from ammonia to nitrate to nitrogen gas. Nitrifi-
cation is the oxidation of NH_-N to nitrate, and denitrification is the reduc-
tion of nitrate to nitrogen gas. Different types of microorganisms are required
for each action. The extent of their use in wastewater treatment depends upon
the end objective. Nitrification is used to control wastewater effluent levels
of ammonia, but both nitrification and denitrification must be used to control
total nitrogen levels in wastewater effluents. Although process technology
for ammonia-nitrogen removal includes breakpoint chlorination, ammonia strip-
ping, ion exchange, and nitrification/denitrification, this paper deals only
with nitrification.
In addition to nitrification/denitrification, microorganisms other than the
nitrifiers and denitrifiers require nitrogen for growth. The amount of nitro-
gen assimilated during oxidation of carbonaceous material has been generally
placed at 5 percent of the oxygen demand (i.e., BOD to N = 20 to 1). The
consequence is two fold: (1) nitrogen must be present for biological oxidation
of carbonaceous material, and (2) removal of ammonia-nitrogen during biological
treatment of wastewaters may be due to assimilation, not necessarily due to
nitrification.
The importance of nitrogen control in wastewater effluents is its impact
on receiving waters. As ammonia becomes oxidized to nitrate, the dissolved
oxygen level of water is decreased. Ammonia-nitrogen at concentrations of
0.25 to 0.30 mg/1 are lethal to fish within 14 to 21 days'. Nitrate is readily
available for assimilation by plant life, causing algal blooms when present in
too large a quantity. Also, nitrate can cause methemoglobinemia in infants
when contaminated water is used as a drinking water supply.
272
-------
Nitrification
The two microbial genera usually associated with nitrification are
Nitrosomonas and Nitrobacter. Both genera of organisms are "autotrophic nitri-
fying bacteria indicating that energy for growth is derived from the oxidation
of inorganic nitrogen. The oxidation of ammonia to nitrate is a two step pro-
cess requiring both organisms for the conversion. Nitrosomonas transforms
ammonia to nitrite while Nitrobacter further oxidizes nitrite to nitrate. The
overall oxidation of ammonia by these organisms is given by the following
equation:
202 + 2 HCO-3 ^ + 2 H2C°3 + H2°
As ammonia is oxidized, carbonate is utilized, As nitrate formation
occurs, carbonic acid is produced. This microbiologically induced change in
the carbonate buffering system results in the destruction of alkalinity at a
rate of 7.1 mg (as CaCO „) per mg of ammonia oxidized. As the nitrification
process reduces the alkalinity and increases the carbonic acid concentration,
the pH of the wastewater may drop as low as pH 6,0, and adversely impact the
rate of nitrification. This decrease in pH can be minimized by aeration to
strip CO^ from the wastewater, or by insuring the presence of excess alkalin-
ity.
Primary environmental conditions for optimal rates of nitrification are pH
and temperature. The reported pH optima cover a wide range, but the consensus
is that^as the pH decreases, the.. rate of nitrification declines. Sawyer,
et all, and Engel and Alexander have reported pH optima for nitrification
between 8.0 and 9.0, and 7.0 and 9.0, respectively. Painter has stated that
nitrification processes cease at or below pH 6.3 to 6.7. Poduska and
Andrews have shown that abrupt changes in pH from 7.2 to 5.8 markedly re-
duced the ammonia oxidation by nitrifiers while the reversal in pH restored
the original nitrification rate.
Temperature optima for nitrification are generally reported by various
authors at about 30°C with a range of 28-35°C. ' ' ' ' ' Temperature
influences hetero trophic and autotrophic microorganisms, thereby affecting
secondary treatment and nitrification efficiencies. The nitrification rate is
more temperature sensitive than the rates for organic removal. Nitrification
rates decrease about 50 percent for each 10 C drop in wastewater temperature
below about 30 G. For example, the nitrification rate at 10 C would be about
half that of 20 C. Secondary treatment efficiency is less likely to be af-
fected by temperature changes, probably due to microbial population diversity
and other system constraints. Organic removal rates for fixed-film processes
should decrease about 25 percent for each 10 C drop in wastewater temperature
below about 30 C. For example, the rate of biological activity in a trickling
filter process, at 10 C would be about 75 percent of that of 20 G. However,
the actual temperature effect on a biological process is probably characteris-
tic only of that system.
MATERIALS AND METHODS
RBC Process
The RBC system evaluated was designed to remove BOD_ (secondary treatment)
and NH3~N (nitrification) to 10 mg/1 and 2 mg/1, respectively, at a 6
273
-------
design flow during both summer and winter conditions. These design parameters
were chosen on the basis of NPDES permit limitations in effect during the
design phase of the wastewater treatment plant upgrade program. (The seasonal
variance of the NPDES permit, as shown in Table 1, became effective during
construction of the upgraded facility and, essentially, relaxed the require-
ments for winter operation.) The system, shown in Figure 2, consists of six
trains of EEC's with six stages per train (i.e., a total of 36 RBC units).
The first three stages were designed primarily for BOD,, removal and consist of
18 standard shafts of 100,000 sq ft of surface area each. The last three
stages were primarily designed for NBL-N removal and consist of 18 high-
density media shafts of 150,000 sq ft of surface area each. At a 6 MGD design
flow, the overall hydraulic loading of the system is 1.33 gpd/sq ft. The RBC
units were manufactured by Autotrol Corporation (Bio-Surf process) . The BOD,.
removal portion of the RBC system was designed based on hydraulic loading
(gpd/sq ft) versus BOD,, removal (percent) curves. A BOD_ influent concentra-
tion (based on total or unfiltered BODj.) to the RBC system of 140 mg/1 was
used for the design. The NH_~N removal portion of the system was designed
using specific removal rates of NH -N concentration and was based on an influ-
ent NHL—N concentration to the nitrification phase of the RBC system of
15.8 mg/1 NH_-N. Based on a dye study, the hydraulic detention time of the
RBC system was 2 hours and 30 minutes at a flow rate of 5.5 MGD (influent plus
recirculated and sidestream flows).
Sampling and Analyses
Treatment train No. 4 (see Figure 2) was used as the primary train to
evaluate performance of the overall RBC process. During both summer and
winter studies, grab samples of the RBC influent and wastewater in each of the
6 stages of train No. 4 were collected at various times during the studies to
determine changes in wastewater characteristics through the system. Tempera-
ture and DO data were taken at each sample period using a YSI, Model 57 DO
meter. Twenty-four hour, flow proportioned composite samples were also col-
lected at the RBC system influent, effluent and the treatment plant effluent.
Sample point locations are shown in Figures 1 and 2.
Twenty-day BOD versus time curves for the STP effluent were developed from
24-hour flow-proportioned composite samples. BOD values were measured every
day for the first 10 days and every other day, thereafter. Nitrification was
suppressed by the addition of ammonium chloride in order to determine the BOD
exerted by carbonaceous and nitrogenous substances.
All analytical chemistry procedures were conducted by the Environmental
Chemistry Division of the US Army Environmental Hygiene Agency. A mobile
laboratory was set up at the installation for performance of the requisite
laboratory work. Sampling and analyses were conducted in accordance with .,
Standard Methods for the Examination of Water and Wastewater, 14th Edition
or Methods for Chemical Analysis of Water and Wastes. Tests for soluble-
BODj. and TOG were conducted on the filtrate of samples filtered through a
0.45 urn filter.
RESULTS AND DISCUSSION
The RBC system influent and effluent characteristics and critical waste-
water treatment plant effluent values are summarized in Tables 2 and 3 for
both summer and winter studies.
274
-------
During the summer study, the mean treatment plant effluent BOD,, level was
11 mg/1, one rag/1 higher than the monthly permit limitation. To reduce the
BOD level below 10 mg/1 did not necessarily mean that additional RBC surface
area was required. In fact, there was more than sufficient carbonaceous sur-
face area available for BOD,, removal. Detailed analysis of the RBC system
wastewater characteristics and an evaluation of the BOD_ test procedure con-
firmed this. As shown in Figure 4, soluble BOD,, and TOG values were reduced
to essentially constant values within the initial three RBC stages. Hence,
the other three stages were available as extra capacity for soluble—BOD_ re-
moval and nitrification. The RBC system effluent soluble-BCD- averaged less
than 5 mg/1. Detailed evaluation of soluble-BCD,, removal through the RBC
system can be obtained from Table 5. For example, 9 mg/1 (24-15 mg/1) of
soluble BOD,, were removed from stage 1 at an average flow rate of 4.5 MGD.
As shown in Figure 4, wastewater DO concentrations were about 1 mg/1 or
less in stages 1 through 4. Hence, the oxidation rate of organics was
limited by DO level. (The DO levels observed were that of the bulk liquid.
Since most oxidation of organics takes place at the media interface, the DO
level at the interface is believed to be much less than observed values;
therefore, the media interface DO level would limit oxidation rates.)
Another factor indicating that DO limiting conditions existed was the nature
of the biomass growing on the first two stages of the RBC trains. A white
biomass, indicative of the autotrophic, sulfur bacteria, Beggiatoa, was pre-
dominant. Beggiatoa utilize hydrogen sulfide and sulfur as energy sources
in the presence of oxygen, according to the following equations:
2H2S + °2 = S2 + 2H2°
S2 + 302 + 2H2° " 2H2S04
Although Beggiatoa exist under aerobic conditions, anaerobic conditions
must be present for the formation of hydrogen sulfide and sulfur. Because
of the low DO conditions of the bulk liquid, oxygen transfer to the fixed
film was severely limited. Therefore, the bacteria within the film inter-
face were most likely anaerobic leaving only the bacteria on the outside sur-
face of the media interface aerobic. As such, the anaerobic bacteria pro-
duced the hydrogen sulfide and provided an energy source for the Beggiatoa
to thrive at the media interface. When plenty of hydrogen sulfide is present,
Beggiatoa store the sulfur in the cells, giving the organism the distinctive
milky appearance as was seen on the first two RBC stages (see Plate 1)J In
later stages, where the oxygen level of the bulk liquid increased, the
Beggiatoa became visually non-existent (see Plate 2)* The predominance of
Beggiatoa noted, not only aggravates soluble-BOD- removal; but, because of
sulfuric acid formation, the pH dependent nitrifying bacteria (optimum pH
8.0 - 8.5) are aggravated as well.
The RBC system analysis showed that very little soluble BOD,, remained after
treatment. This obviously indicates that other factors contributed to the
treatment plant effluent BOD_. In fact, both suspended solids and ammonia
were found to exert a five-day oxygen demand in this case (see Table 4). More
biological treatment may not be needed to overcome this, because the purpose
of biological treatment is conversion of soluble organics to C0?, H»0, and
particulate matter (i.e., suspended solids). Removal of suspended solids may
be called for to reduce the effluent BOD_. Carbonaceous BOD,, (nitrification
*Not reproducible in these proceedings.
275
-------
suppressed) versus total BOD,, indicated that ammonia nitrogen was also a factor
in exerting an oxygen demand in the treatment plant effluent. Since part of
the RBC process involves nitrification, relatively large numbers of nitrifying
bacteria were present in the plant effluent and in the BOD bottles. Hence,
significant nitrification occurred within the first five days. Data in Table 4
show that nitrification accounted for about 7 mg/1 (11-4 mg/1) of the BOD^,
Figure 5, an example of the BOD versus time data collected, also shows signifi-
cant nitrification. Consequently, suspended solids (2 mg/1 BOD ) and nitrifi-
cation (7 mg/1 BOD,.) accounted for most of the BOD,. (11 mg/1) in the wastewater
treatment plant effluent. While the effluent did not strictly meet the BOD,.
limitations, the reasons for this were oxygen demand of suspended solids ana
NH_-N oxidation. The RBC system was, therefore, performing adequately for
soIuble-BOD removal at the existing conditions (26°C and 3.7 MGD). It is
important to note, however, that at the 6 MGD design flow rate, adequate BOD
removal could not be assumed. In fact, although adequate soluble-BCD- removal
occurred, it must be understood that the system can be considered to nave been
operating at suboptimal levels for soluble-BOD- removal because of DO limiting
conditions. If supplemental oxygen were provided to the system at the same
existing conditions, effluent soluble-BOD- levels would still be about the
same except that less surface area would Be required to remove the same amount
of soluble-BOD t-. Then, more surface area would be available for nitrification.
During the summer study, the treatment plant effluent NH_-N level was sig-
nificantly higher than the monthly effluent limitation (Tables 1 and 3). The
observed level of ammonia-nitrogen in the effluent was 6.2 mg/1 while the
limitation (and design expectation) is 2.0 mg/1. As with BOD,., the short-
comings in meeting effluent limitations for ammonia-nitrogen are not obvious.
Nitrification can be affected primarily by pH and temperature, but also by the
sequence of organic removal (i.e., nitrification does not begin to any sig-,.
nificant degree at neutral pH levels until soluble-BCD- has been oxidized).
Hence, to evaluate 1BC performance for NHL-N removal, progression of treatment
within the RBC stages as well as environmental conditions must be assessed and
discussed. NHL—N data are presented in Table 5 and shown in Figure 6.
Nitrification began in RBC stage 4 and continued through stage 6, Nitrifica-
tion data is also supported by TKN and N0_/N0» (Figure 6) and alkalinity
changes (Figure 7). NH_-N oxidation rates were also believed to be limited by
low DO levels in some or the RBC stages, as discussed earlier for BOD,, removal.
In addition, NH,-N oxidation rates were limited by low pH levels shown in
Figure 6. Optimum pH for nitrification is about pH 8.0-8.5. ' whereas
observed values were pH 6.5-6.7 in latter RBC stages.. (The summer wastewater
temperature of 26 C was believed to be near optimum for nitrification.)
During the winter study, the RBC system was operating under markedly dif-
ferent conditions than during the summer. Wastewater temperature was, of
course, much lower, averaging 13 C, and the hydraulic and organic loadings
were significantly higher (see Tables 2 and 3). These factors would expectedly
result in poorer RBC performance than noted during the summer. To the con-
trary, as shown in Table 3, the average wastewater effluent BOD. level during
the winter was about the same as the summer BOD,, level (10 mg/1 during the win-
ter and 11 mg/1 during the summer). Also, average N1L-N levels were actually
lower (5.1 mg/1 during the winter and 6.2 mg/1 during the summer), albeit
NH_-N design expectations were still not met. The only positive effect during
the winter study that could be attributed to this noticeable improvement in
RBC system performance was the wastewater DO levels in the RBC system. Because
of the low wastewater temperatures, adequate oxygen mass transfer occurred
276
-------
as shown in Figure 8t DO limiting conditions were non-existent. Also, the
Beggiatoa bacteria had disappeared.
As shown in Figures 8 and 9, concentration profiles representing soluble
organic removal and ammonia oxidation through the RBC system are not signifi-
cantly different than those from the summer study. Although soluble organic
removal may have been slightly more sluggish, adequate removal still occurred
such that nitrification began in the fourth stage. Wastewater pH was, again,
too low to insure optimum nitrification (see Figure 10).
The effect of adequate wastewater DO levels in the RBC system cannot, in
this case, be fully appreciated by merely-comparing summer vs winter treatment
plant effluent and RBC svstem character"!sMr=:. Th^co i-r.mr\^T-->"o^no /-.nitr ^u~.~
-------
7. Low pH levels can be easily corrected by chemical feed to maintain a
pH level of 7.0 or higher.
8. BODj. analyses should include soluble-BOD from each RBC stage when
I
5
J _ CLxLd-L-j' ij^-ij o * AW \_I-UI-L ,i.nv^.L.v_m.t_ ^ u JL. u i-r .JL. *_» j_* v/j_* _
evaluating RBC system performance; likewise, BOD analyses should include
carbonaceous and nitrogeneous oxygen demands.
ACKNOWLEDGMENTS
The authors wish to extend special thanks to Majors John C. Mcllrath and
Jeremiah J. McCarthy, Water Quality Engineering Division, and Mr. David Rosak
and his staff, Environmental Chemistry Division, all of the US Army Environ-
mental Hygiene Agency. Their technical and administrative assistance has
been instrumental in the conduct of this work.
278
-------
LITERATURE CITED
1. Wild, H.E., Jr., C.N. Sawyer, and T.C. McMahon, "Factors Affecting Nitrifi-
cation Kinetics," _J. Water Pollut. Control Fed., 43:1845 (1971).
2. Steiner, C.G., "The New Rotating Disk Process," Advance Publ. Copy (1978).
3. Hockenbury, M.R,, G.T. Daigger, and C.P. Grady, Jr., "Factors Affecting
Nitrification," J_. of. _the _Env. Div. :9 (1977).
4. Process Design Manual for NitrogenControl, U.S.EPA Tech. Transfer (1975).
5. Miller, R.D., _e_t _al, "Rotating Biological Contactor Process for Secondary
Treatment and Nitrification Following a Trickling Filter," Technical Report
7905, US Army Medical Bioengineering R & D Laboratory, June 1979.
6. Antonie, R.L., D.L. Kluge, and J.H. Mielke, "Evaluation of a Rotating Disk
Wastewater Treatment Plant," _J. Water Pollut. Control Fed. 46:498 (1974).
7. Antonie, R.L., "Rotating Biological Contactor for Secondary Wastewater
Treatment," Presented at Culp/Wesner/Culp WWT Seminar, South Lake Tahoe, NV
(1976).
8. Sawyer, G.N. and P.L. McCarty, "Chemistry for Sanitary Engineers,"
McGraw-Hill Book Company, New York, NY (1967).
9. Busch, A.W., Aerobic Biological Treatment of Waste Waters, Oligodynamics
Press, Houston, TX (1971).
10. Clark, J.W. and Viessman, Jr., "Water Supply and PollutionControl,"
International Textbook Company, Scranton, PA (1965).
11. Smart, G., "The Effect of Ammonia on Gill Structures of Rainbow Trout,"
i- Fish Biol., 8:471-475 (1976).
12. Gruener, N, and H.I. Shuval, "Toxicology of Nitrites," Environmental
Quality and Safety, 2:219-229 (1973).
13. Sawyer, C.N., H.E. Wild, and T.C. McMahon, "Nitrification and Denitri-
fication Facilities Wastewater Treatment," U.S.EPA, Tech. Transfer (1973).
14. Engel, M.S. and M. Alexander, "Growth and Metabolism of _N. europeae,"
J_. Bacteriol., 76:217 (1958).
15. Painter, H.A., "A Review of the Literature on Inorganic Nitrogen Metabo-
lism in Microorganism," Water Res.. 4:393 (1970).
16. Poduska, R.A. and J.F. Andrews, "Dynamics of Nitrification in the
Activated Sludge Process," 29th Industrial Wastes Conf., Purdue Univ., IN
(1974).
279
-------
17. Buswell, A.M., T. Shiota, N. Lawrence, and I.V. Meter, "Laboratory
Studies on the Kinetics of the Growth of Nitrosomonas with Relation to the
Nitrification Phase of the BOD Test," Appl. Microbiol., 2:21 (1954).
18. Deppe, K. and H. Engel, "Untersuchungen uber die Temperaturabhangigkeit
der Nitratbildung durch Nitrobacter winogradskii Buch. bei ungehemmtem und
gehenmtem Wachstum," Zentbl. Bakt. Parasitk de II. 113, 561-568 (1960).
19. Laudelout, H. and L. vanTichelen, "Kinetics of the Nitrite Oxidation by
Nitrobacter winogradskii," J. Bacteriol., 79:392-42 (1960).
20. Balakrishnan, S. and W.W. Eckenfelder, "Nitrogen Relationship in
Biological Waste Treatment Processes - II, Nitrification in Trickling Filters,"
Water Res., 3:167 (1969).
21. Haug, R.T. and P.L. McCarty, "Nitrification with the Submerged Filter,"
U.S.EPA Grant //17010EPM (1971).
22. Haung, C.S. and N.E. Hopson, "Temperature and pH Effect on the Biological
Nitrification Process," Presented at the New York WPCA, New York, NY (1974).
23. Siddigi, R.H., et al, "Elimination of Nitrification in the BOD Determin-
ation with 0.1M Ammonia Nitrogen," JWPCF, 39:579 (1967).
24. Standard Methodsforthe Examination of Water and Wastewater, 14th Edition,
American Public Health Association. American Water Works Association, Water
Pollution Control Federation (1976).
25. Methods for Chemical Analysis of Water and Wastes, USEPA, Doc. No.
EPA-625-16-74-Q03 (1974).
26. Frobisher, M., Fundamentals of Microbiology, Charles E. Tuttle Company,
Tokyo, Japan, 6th Edition (1957).
280
-------
ABBREVIATIONS
BOD
BOD
BOD -S
CaC03
CBOD
co2 ^
DO i
Eff
FC
gpd/sq ft
HC03
H2C°3
H2S
Inf
1
Ibs
mg
MGD
ml
ym
pmho/em
N
Biochemical oxygen demand is the amount of oxygen required
by microorganisms to oxidize dissolved organics in a waste-
water .
BOD measured after 5 days
BOD of a filtered sample (BOD,, of soluble organic material)
BOD measured after 20 days
Calcium carbonate
Carbonaceous BOD_. Oxidation of any NH,,--N present in the
sample is chemically inhibited.
Carbon dioxide
Dissolved oxygen
Effluent
Fecal coliform
Gallons per day per square foot
Bicarbonate ion
Carbonic acid
Hydrogen sulfide
Sulfuric acid
Influent
liter
pounds
milligram
Million gallons per day
milliliter
Micrometer
Micromhos per centimeter
Nitrogen
281
-------
NH-K Ammonium ion
NH--N Ammonia expressed as nitrogen
N02/N(X, Nitrite plus nitrate expressed as nitrogen
NO-3 Nitrate ion
0 Oxygen
pH Negative logarithm of hydrogen ion concentration
EBC Rotating Biological Contactor
S,, Elemental sulfur
SS Suspended solids
T Alk Total alkalinity
Temp C Temperature in degrees Celsius
TKN Total Kjeldahl nitrogen
TOC-S Total organic carbon of a filtered sample (soluble TOG)
YSI Yellow Springs Instruments
282
-------
TABLE 1. NPDES PERMIT PARAMETERS AND LIMITATIONS*
pH
Chlorine Residual
Fecal Colifora (FC)
Suspended Solids (SS)
Five-day Biochemical Oxygen
Demand (BOD )
Ammonia Nitrogen (NH,-N)
Dissolved Oxygen (DOJ
Monthly Average
Summer
6.0 - 9.0
Min cone to comply
w/FC limit
200/100 ml
30 mg/1
10 mg/1
2.0 mg/1
Greater than
6.0 mg/1
Monthly Average
Winter
6.0 - 9.0
Min cone to comply
w/FC limit
200/100 ml
30 mg/1
20 mg/1*
5.0 mg/1*
Greater than
8.5 mg/1
* Limitations at the time of design did not allow for winter variance,
TABLE 2. RBC INFLUENT AND EFFLUENT CHARACTERISTICS
15 - 21 AUGUST 1978
Avg. Wastewater Flow =4.5 MGD*
Influent Effluent
Conductivity
(ymho/cm)
T Alk
SS
BOD
CBOlL
BOD -S
TOC - S
TKN
NH -N
NO/NO-N
960
158
69
72
48
21
23
21
16.0
0.05
930
90
63
61
28
4
11
8.9
6.2
8.9
23-29 JANUARY 1979
Avg. Wastewater Flow =5.2 MGD*
Influent Effluent
922
174
110
126
33
24
25,
16,
0.89
873
97
89
96
9
11.1
4.8
9.29
* STP Influent flow + recirculated flow.
All units are mg/1 unless otherwise noted.
283
-------
TABLE 3. SEWAGE TREATMENT PLANT EFFLUENT VALUES
15 - 21 AUGUST 1978
pH
SS
Flow
6.7 (median)
11 mg/1
6.2 mg/1
9 mg/1
3.7 MGD
23 - 29 JANUARY 1979
7.0 (median)
10 mg/1
5.1 mg/1
13 mg/1
4,6 MGD
TABLE 4. TREATMENT PLANT EFFLUENT BOD LEVELS
BOD,.
Filtered (soluble) BOD
Unfiltered (total) BOD
Suppressed Nitrification
(carbonaceous only) BOD
Analysis:
BOD
20
2
11
4
No data
45
8
2 mg/1 were due to soluble organics; 4 mg/1 were due to
carbonaceous material (2 mg/1 soluble + 2 mg/1 suspended);
7 mg/1 (11—4 mg/1) were due to nitrification.
BOD2Q: At 20 days, 37 mg/1 (45-8 mg/1) were due to nitrification.
284
-------
TABLE 5. RBC SYSTEM DATA SUMMARY*
AUGUST (WASTEWATER TEMP = 26°C)
Avg Flow: 3.7 MGD (plant flow) +0.8 MGD (recirculated flow)
£l T-Alk DO BpD--S TOC-S TKN NH -N
RBC Influent 6.8 159 3.4 24 24 23 16
,04
Stage
1
2
3
4
5
6
6
6
6
6
6
6
.8
.8
.75
.6
.55
.5
159
158
149
126
104
93
1.4
0.7
0.8
1.3
1.9
2.2
15
9
6
4
3
3
18
14
11
10
9
9
22
20
19
14
11
9
16
16
15
12
8.3
6.6
<0.
0.
1.
4.
7.
9.
04
2
4
4
7
7
JANUARY (WASTEWATER TEMP = 13°C)
Avg Flow: 4.6 MGD (plant flow) +0.6 MGD (recirculated flow)
£l T Alk DO BOD -S TOC-S TKN NH.-N NO /NO -N
RBC Influent
7.1
163
5.7
32
23
24 16
0.9
Stage 1
2
3
4
5
6
7.1
7.1
7.1
6.95
6.8
6.75
163
161
152
124
106
87
3.3
2.7
2.9
3.5
4.2
4.8
24
12
10
7
6
5
17
13
11
10
9
8
23
21
21
16
14
12
15
14
14
9.1
6.7
3.9
0.6
0.8
1.8
5.3
7.7
10.2
* Mean of 5 sets of grab samples collected at various times for each study
period. pH values represent-median values. Units expressed as mg/1 unless
otherwise noted.
285
-------
FIGURE 2 ROTATING BIOLOGICAL CONTACTOR SYSTEM
( NOT TO SCALE )
30°V-NOTCH
WEIRS -Z-*
to
00
SLUICE
GATES '
EFFLUENT
TROLJRH "»
^>*"~----^^
•2X" I "*
1
RBC BANK |
NO 1 NO 2 NO 3 '^
FLOW
FROM
PUMP
CT*TI/MLI
1
1
t
_»t -X
ir
i i
i i
jf :•£€'_
""""*«^- ^--•'""'""^"""^"•^^ ^^*^*^^^^*-»^ ^^^"^
*•• |
1
1
1
i
! NO 4 NO 5 NO 6
STAGE 6
STAGE 5
STAGE 4
STAGE 3
STAGE 2
STAGE 1
Jkr- 90° V- NOTCH
WEIRS} 39 WEIRS PER BANK
^SLUICE GATES
-------
INFLUENT
RBC MEDIA a SHAFT
r
EFFLUENT
REGULAR DENSITY MEDIA
100,000 sqft/sliaft
HIGH DENSITY MEDIA
150,000 sq ft/shaft
MEDIA
ROTATION
WASTE WATER
FLOW
FIGURE 3 RBC PROCESS DIAGRAM
287
-------
Xmg/i)
DC-S
Tlfl/l)
/l)
5,0
4.0
3.0
i
2.0
30
20
10
30
10
0
RBC
Inf
STAGE
RBC
Inf
STAGE i
RBC
Inf
STAGE I
UNIT LENGTH .STAGES
FIGURE 4. SOLUBLE BOD5, SOLUBLE TOC AND DO LEVELS THROUGH THE RiC
SYSTEM,SUMMER STUDY
288
-------
FIGURE 5 BIOCHEMICAL OXYGEN DEMAND VERSUS TIME, SAMPLE POINT 6- STP EFFLUENT
G - STANDARD BOD PROCEDURE - SEEDED W/RAW INFLUENT
A- NH,-N SUPPRESSION (C BOD) USING O.IM NH.CL- SEEDED W/RAW INFLUENT
j 4
50
M
CD
*£>
X
D>
E
5
LU
Q
Z
o
x.
o
30
20
o
o
m
10
8 10
TIME (DAYS)
12
14
16
18
20
-------
25 -
20 -
NH3-N ,5
TKN
N0/N03
(mg/I)
10 -
5 -
© -NH3-N
A -TKN
Q -N02/N03
RBC STAGE I
Inf
234
UNIT LENGTH, STAGES
FIGURE 6 NH3-N,TKN AND N02/N03 LEVELS THROUGH THE RBC
SYSTEM, SUMMER STUDY
-------
PH
7.O
6.8
6.6
6.4
6.2
RBC
Inf
STAGE
200
I5O
T Alk
(mg/l)
IOO
50
RBC
Inf
STAGE! 2 3
UNIT LENGTH, STAGES
FIGURE? CHANGES IN ALKALINITY AND pH THROUGH RBC SYSTEM,
SUMMER STUDY
291
-------
D0(mg/l)
TOC-S
(mg/1)
BODS-S
(mg/I)
5.0
4.0
5.0
2.0
l.O
30
IO
30
20
10
RBC STAGE I
Inf
RBC
Inf
STAGE 1
RBC
Inf
STAGE
UNIT LENGTB.STAGES
FIGURE 8 SOLUBLE BOD5, SOLUBLE TOC AND DO LEVELS THROUGH THE
RBC SYSTEM , WINTER STUDY
292
-------
25
20
NH3-N
TKN
N02/N03
(mg/l)
15
10
0L
O-NH3-N
A-TKN
Q-N02/N03
RBC
Inf
STAGE I 2 3
UNIT LENGTH, STAGES
FIGURE 9 NH3-N,TKN AND NO/N03 LEVELS THROUGH THE RBC
SYSTEM, WINTER STUDY
293
-------
7.0
PH
6.4
6,2
200
TAIk
(mg/l)
50
RBC
Inf
RBC
Inf
STAGE I
STAGE I
UNIT LENGTH, STAGES
FIGURE 10 CHANGES IN ALKALINITY AND pH THROUGH THE RBC
SYSTEM- WINTER STUDY
294
-------
NITROGEN AND PHOSPHORUS REMOVAL WITH
ROTATING BIOLOGICAL CONTACTORS
By
Robert J. Hynek, Manager
Process Verification and Pilot Plant Program
Autotrol Corporation
Milwaukee, Wisconsin
and
Hiroshi lemura
Chief Process Engineer
Nippon Autotrol K.K.
Tokyo, Japan
Introduction
As of September, 1979, over 200 municipalities in the United States and
Canada had chosen Autotrol's Bio—Surf Process for wastewater treatment. Of
these, 25 percent were installed to remove BOD and to nitrify the sewage.
Another 5 percent were installed to nitrify effluent from existing secondary
treatment plants to meet new discharge limitations. About 10 percent were
installed to include phosphorus removal and one of the largest is designed
for denitrification.
Those installations concerned with nutrient removal range in size from 0.1 to
37 MGD. Average flow for the 70 plants was 3 MGD, with 25 plants within 2
MGD of this agerage. This wide and relatively uniform distribution of plant
sizes attests well to the application of the Bio-Surf Process to communities
295
-------
of all sizes, while at the same time offering a non-complex system using a
minimum amount of energy.
The first portion of this paper will discuss two case histories of nitrifi-
cation and phosphorus removal with the Bio-Surf Process in the United States.
The latter portion of this paper will be presented by my colleague from
Nippon Autotrol, Mr. Hiroshi lemura. His discussion will include information
on several installations designed for nitrification and denitrification.
Full-scale performance since 1976 and a new design rationale for methanol addi-
tion resulting from this experience should be of keen interest to this
audience.
Case No. 1 — Gladstone, Michigan
The first case history illustrates application of the Bio-Surf Process design-
ed for BOD removal from primary clarifier effluent. Nitrification was not
required, but is achieved at no extra cost because of the temperature
corrections made for BOD removal at low winter temperatures. Phosphorus
removal was required and is achieved by addition of alum and anionic polymer
to the Bio-Surf effluent prior to secondary clarification.
This plant is located in Gladstone, Michigan on the southern exposure of the
Upper Pennisula to Lake Michigan. The existing real estate was narrow and
little space was available for treatment expansion between the existing
primaries and the lake. This situation and the desire to maintain process
simplicity led to the installation of six rotating biological contactors in
two parallel trains. This arrangement allowed placement immediately adjacent
to the rectangular primaries under an expanded common building, while at the
same time allowing for construction of two new secondaries next to the lake.
Startup of this facility designed for treatment of 1 MGD began in 1974 and has
provided a most significant contribution to Autotrol's knowledge of the Bio-
Surf Process and to current design rationales based on full-scale, rather than
pilot-scale, performance.
Table A summarizes monthly average Process performance for the year of 1977.
This performance is very typical of performance since startup to the present
day, showing effluent BOD and suspended solids well within discharge require-
ments of 35 mg/1 and a minimum of 80 percent phosphorus removal.
The data in Table I were divided into the summer and winter periods (June
through October and November through May) to illustrate performance at the
different wastewater temperatures of 63°F (17.2°C) and 49°F (9.4°C). It is
quite obvious that only nitrification is affected, as would be expected by
the transition from warm to cold conditions, whereas BOD, phosphorus and
suspended solids removals are very consistent at 94, 87 and 81 percent,
respectively.
Table B summarizes monthly operational costs with respect to chemicals, elect-
rical power and miscellaneous utilities for the same seasonal and annual
periods. Little seasonal differences are seen for alum and polymer costs per
million gallons treated per day, whereas relatively large differences are seen
for chlorine and utilities. Higher chlorine consumption in summer is under-
standable from decreased solubility and greater reactivity considerations.
296
-------
The increase in utilities costs in winter reflect the purchase of larger
amounts of natural gas to maintain anaerobic digester temperatures at proper
levels.
Seasonal treatment costs per million gallons are quite uniform in either case,
however, varying only about 10 percent from the annual average.
Cost analysis for nutrient removals can be looked at in several ways. How-
ever, since this plant was designed and constructed for secondary treatment
and phosphorus removal, all costs should be so analyzed. In that event, the
alum and polymer costs are those principally related to phosphorus removal
and, on an annual basis, the City of Gladstone is expending seventy-six cents
per pound of phosphorus removed per million gallons treated. The balance of
the costs for chlorine, power and utilities are largely associated with BOD
and suspended solids removal. For 1977, these costs were 4.1 cents per pound
removed, for a. total of 80.1 cents for removal of the three pollutants.
Nitrification could be regarded as being obtained at no cost. However, the
records are complete enough to allow differentiation between the removal of
each "nutrient" category. Field evaluations by Autotrol reveal that soluble
BOD oxidation is essentially complete with 50 percent of the equipment in
summer and 67 percent in winter. Nitrification occurs on the remainder in
each case. In addition, electrical energy consumption has been measured at
Gladstone with polyphase wattmeters in each season. With the above informa-
tion one can compute daily energy costs for nitrification with rotating bio-
logical contactors for installations similar to Gladstone. Calculations
summarized in Table C show that less than five cents of electrical energy are
required for nitrification of one pound of ammonia nitrogen per million
gallons treated per day with the Bio—Surf Process at Gladstone.
297
-------
Table A
Monthly Bio-Surf Process Performance - Gladstone, Michigan
- 1977 -
Summer Winter _ _ Annual
Q, MGD 0.748 0.707 0.724
°F 63 49 55
mg/1 TOg/1 mg/1
Raw FE %R Raw FE' %R Raw FE %R
BOD5 127 7 94.5 155 8 94.8 143 8 94.4
NH3-N 15.8 1.1 93.0 15.0 4.9 67.3 15.4 3.5 77.3
6.5 1.2 81.5 6.1 1.2 80.3 6.3 1.2 81.0
TSS 131 15 88.5 155 17 85.2 122 16 86.9
298
-------
Table B
Monthly Operational Costs - Gladstone, Michigan
- 1977 -
Dollars Per Million Gallons Treated*
Alum
Chlorine
Polymer
Power
Utilities
Total
Summer
28.60
6.00
4.31
50.76
3.79
Winter
28.15
3.45
4.14
59.25
16.92
Annual
28.33
4.51
4.21
55.71
9.98
$ 93.46
$ 111.91
$ 102.74
*Alum, chlorine and polymer costs were $0.041, $0.170 and
$1.90 per pound, respectively.
299
-------
Table C
Estimated Nitrification Energy Costs - Gladstone, Michigan
- Primary Clarifier Effluent -
Per Million Gallons Treated
LBS/Day
KWHR/LB
KOTR $/LB*
Stammer Winter Annual
117 81.0 95.5
1.13 1.08 1.11
0.045 0.045 0.044
Calculated using $0.040/KWHR
300
-------
Case No. 2 - Cadillac^ Mich_i§an.
The second most popular application of the Bio-Surf Process for nutrient
removal is nitrification of secondary effluent following pretreatment with
alternative processes. At Cadillac, Michigan, an existing activated sludge
plant required upgrading for nitrification and phosphorus removal. The final
design incorporated ferric chloride addition for phosphorus removal ahead
of the nitrification section.,
Eight rotating biological contactors began operation in 1976 to meet discharge
requirements of 1.5 mg/1 NHo-N during the summer months of June through
October. Performance did not meet requirements and thorough investigation
culminating in tracer studies' revealed hydraulic shortcircuiting because of
excessive underflow openings in the baffles separating the reactors. Per-
formance improved measureably after corrections were made to improve staging
and effluent requirements have since been met consistently.
Current annual operation consists of shutting down the RBC shafts on November
1st and re-starting on April 15th. This was done to take advantage of the
annual opportunity to reduce laboratory and maintenance schedules and also to
reduce overall energy consumption. Performance in this mode of operation
for 1979 is summarized in Table D. As seen in Table D, both the Bio-Surf and
final effluent averages are below the 1.5 mg/1 requirement for the 5-month
period. It is also evident that the secondary clarifier effluent NHo-N
declined considerably toward mid-year, then abruptly increased in October by
approximately 90 percent, from 5.9 to 11.0 mg/1. This pattern reflects
eventual nitrification in the activated sludge basin. This results from re-
circulation of Bio-Surf nitrifiers removed by the final sand filters back to
the head end of the plant for separation by the primary clarifier prior to
eventual digestion. The abrupt decrease in this pre-nitrification cannot be
explained from plant records, but both the Bio-Surf and the final effluent
quality remained within the 1.5 mg/1 specification.
An interesting sidelight to these 1979 data was made possible this past year
by monitoring ni.trif.ier growth on the biological contactors from startup on
April 15th through shutdown on November 1st. This data was developed from
hydraulic load cells under the idle end bearing of the first and last shaft
in one of the tanks. Hydraulic pressure readings were taken during rotation
on a frequent basis in conjunction with normal influent and effluent sampling
routines. These pressure readings were converted to dynamic biofilm thickness
equivalents with Autotrol formulae and Table E summarizes biofilm response to
influent NH-j-N following flow startup. Growth was evident within two weeks
and was maximized to biological equilibrium by mid-June. A decline began
shortly afterward in response to mid-month minimums in daily influent con-
centrations and a relatively flat, but gradually declining profile was
maintained until mid-September. At this time, the partial nitrification
occurring in the activated sludge unit abruptly failed, and by the latter part
of October a slight uptrend in biofilm weight was evident.
Several factors no doubt contributed to this pattern of nitrifier growth.
Although influent NHj-N must be by far the major factor, it is normal for
predator growth to respond to bacterial growth and to reduce the population
until equilibrium is established between influent strength, bacterial kinetics
301
-------
and predatipn efficiency. This phenomenon has been observed many other times
in our BOD removal studies, but these data represent the first documentation
in a low BOD, secondary nitrification situation. It will be most interesting
to see if a similar pattern develops in 1980.
302
-------
Table D
Nitrification of Secondary Effluent - Cadillac, Michigan
- 1979 -
NH3-N. mg/1
Month Q, MjGD T°F ~S.C.E.' B.S.E. F.E.
June 1.62 63 13.1 1.40 1.00
July 1.61 65 8.9 0.67 0.43
August 1.61 67 8.6 0.76 0.57
September 1.50 66 5.9 0.60 0.42
October 1.51 63 11.0 1.52 1.22
1.57 65 9.5 0.99 0.73
303
-------
Table E
Bio-Surf Biofilm Response to NH
Cadillac, Michigan - 1979
30-day NH0-N, mg/1
Biofilm Thickness
Inches
Date
April
May
June
July
Aug.
Sept.
Oct.
15*
30
15
30
15
30
15
30
15
30
15
30
15
30
Influent
16.5
13.1
13.1
8.9
8.6
5.9
11.0
"Effluent
16.3
1.5
1.4
0.7
0.8
0.6
1.5
S-l
0.000
0.008
0.014
0.022
0.026
0.014
0.014
0.014
0.016
0.014
0.013
0.007
0.004
0.007
S-4
0.000
0.006
0.010
0.016
0.016
0.010
0.013
0.013
0.015
0.013
0.013
0.005
0.003
0.007
*Flow began.
304
-------
Table 1
SUMMARY OF RBC INSTALLATIONS IN JAPAN
FLOW
WASTEWATER
DOMESTIC
FOOD
INDUSTRIAL
GARBAGE
ANIMAL
SITES
309
188
196
66
23
%
40
24
25
8
3
SHAFTS
495
377
384
110
48
%
44
11
25
6
1
1000 mj/d
98.7
25.5
82.3
12.8
3.2
MGD
26.1
6.7
21.7
3.4
0.8
782 100 1,414 100 222 58.7
305
-------
Table 2
BIO-SURF NITROGEN REMOVAL INTALLATIONS IN JAPAN
NAME
WASTE
SOURCE
FLOW
nr/d MGD
NH?-Ny MG/L
INFL. EFFL.
START
MIYASAKI
CITY
NIPPON
CHEMICAL
IWAKI
CITY
CHIBA
PREF.
YOKKAICHI
CITY
HAMMATSU
CITY
Garbage 350
Dump
0.092
Brine 2,000 0.528
Garbage 200
Dump
Domestic 65 0,017
Garbage 500 0.132
Dump
Garbage 400 0.106
Dump
200
(TN)
200
0.053 200
20
250
.55
50
(75%)
20
(90%)
3
(85%)
10
1976
1977
1978
1979
1979
1980
306
-------
Table 3
NITRIFICATION AND DENITRIFICATION DESIGN CRITERIA
CASE NO. 1 - MIYAZAKI CITY
LOADING RATES
A. NITRIFICATION SECTION
BOD5: 0.9 g/m2-D (0.18 #/l,000 ft2-D)
Hydraulics: 18.7 l/m2-D (0.46 gpd/ft2-D)
NH3-N: 3.7 g/m2-D (0.76 #/1,000 ft2-D)
Time: 6.9 Hours
B. DENITRIFICATION SECTION
Hydraulics: 42.2 l/m2-D (1.04 gpd/ft2-D)
Methanol: 1.9 times N03—N Concentration
N02-N + N03-N: 7.4 g/m2-D (1.52 ///1,000 ft2-D)
Time: 9.9 Hours
C. RE-AREATION SECTION
BOD5: 20.0 g/tn2-D (4.09 #/l,000 ft2-D)
Hydraulics: 269 l/m2-D (6.62 gpd/ft2-D)
Time: 0.57 Hours
307
-------
Table 4
BIO-SURF PERFORMANCE - MIYAZAKI CITY
- 1977 -
DESIGN Q NH-r-N, MG/L NOy-N, MG/L
MONTH
Jan.
Feb.
Apr.
May
Jun.
Jul,
Nov.
May /Nov.
%
21
19
34
47
53
43
71
54
°C/F
9/48
8/46
22/72
24/75
30/86
30/86
23/73
27/81
IN
101
104
128
132
119
138
128
129
NITR.
Tr
Tr
28
4
4
3
4
4
DENITR.
Tr
Tr
21
4
3
3
3
3
IN
1.9
1.6
22
1
1
1
1
1
NITR.
102
108
26
70
68
67
77
70
DENITR.
15
Tr
11
13
5
14
2
8
PEAK FLOW OCCURRENCE
Nov. 153 20/68 97 0 0 1 59
308
-------
Table 5
BIO-SURF PERFORMANCE - MIYAZAKI CITY
- 1977 -
BOD, MG/L TSS, MG/L CaC03, MG/L
MONTH
Jan.
Apr.
May
Jun.
Jul.
Aug.
Nov.
Avg.
IN
8
18
23
30
19
-
24
20
NITR.
5
5
6
3
6
9
2
5
DENITR.
3
8
14
10
4
4
-
7
PEAK FLOW
IN
7
11
48
72
37
38
101
45
NITR.
4
27
27
240
198
-
92
98
DENITR. IN NITR. DENITR.
1 1,020 324 503
2 1,068 169 192
4 1,082 345 384
4 1,104 354 475
9
3
13
5 1,067 298 389
OCCURRENCE
Nov. 16 8 5 5 264 53 10 783 337 484
309
-------
Bio-Surf Process in Japan
Historical
The RBC process was introduced to Japan approximately 10 years ago in 1970.
Initial emphasis was directed to secondary treatment of food processing
wastewater, as there were many small companies supplying Japan and foreign
countries with canned fruit, mandarin oranges being only one example.
Tertiary treatment started six years later in 1976.
The growth of the RBC process in Japan has been very rapid. Table 1 sum-
marizes the number of locations, reactors and the quantity of flow for the five
major wastewater categories identified as domestic, food processing, indus-
trial (such as pulp and paper), garbage dumps and animal breeding. The
number of firms actively pursuing this market is approximately 20, with only
two or three of significant corporate size. Nippon-Autotrol is the major
active company with a manufacturing, R & D and sales organization, accounting
for 20 percent of the installations, 26 percent of the rotating contactors
and 25 percent of the quantity treated.
There are three basic reasons for the rapid growth; (1) low energy and small
space, "very appealing to industrial firms; (2) maintenance and process
simplicity; and (3) impetus provided by the government in the form of the
Japan. Sewage Works Agency recognizing applicability to municipal wastewater.
At the present time there are no regulations restricting discharge of nitrogen
or phosphorus in wastewaters. However, problems associated with nutrient
discharge intensify with each summer season and many Japanese are becoming
very concerned with the implications for the immediate and long range future.
One of these problems is called the 'Red Tide*, which occurs annually in many
of the inland seas of Japan. The microorganism responsible for this
phenomenon, resulting in significant fish and shellfish kills, has stimulated
much research. The identity and life mechanisms have not yet been fully
defined, but nitrogen and phosphorus removal will no doubt play a major role
in control measures.
Another problem is referred to as "Flower of Water", or water bloom, in Lake
Sagami and Lake Biwa. These lakes provide drinking water to the metropolitan
areas of Tokyo and Osaka and water flavor is noticeably affected.
A third problem has been associated x^ith the rice growing industry. Poor
rice yields have been traced to high nitrogen and phosphorus levels, with the
result being excessive stalk growth versus the desired kernel growth.
The fourth problem is linked to the newly-defined limits of available drinking
water supplies. Water re-use is. now being promoted and is gaining wider
acceptance each year. Wastewater from hand washing, kitchens and cooling
towers are being collected sepaxately, treated biologically with RBC units
and sand filters, and then chlorinated in order to reuse for sanitary flushing,
car washing and lawn irrigation, etc.
310
-------
Table 2 summarizes information relative to six Bio-Surf nitrification and/or
removal installations, five of which are in operation at the present time.
Flow ranges from 65 to 1000 m3/D (17,000 to 528,000 gpd), with NH3-N concen-
trations in the range of 20 to 200 mg/1. Runoff from garbage dumps obviously
is the major problem area for the present, as 4 of the 6 plants shown are
directed to this problem.
Of the six plants, Miyazaki City, Iwaki City and Chiba Prefecture use
methanol as the source of carbon for denitrification. Nippon Chemical is
only a nitrification application fbr the present. Yokkaichi City utilizes
BOD from the incoming wastewater as a source of carbon, and this coupled
with 3-1 effluent recycle rates effectively eliminates methanol costs for
denitrification.
Hammatsu City is now under construction and will incorporate the new Aero-Surf
Process for both nitrification and denitrification.
The following discussion will concentrate on the Miyazaki City application,
since it is the first installation and has been in operation since October of
1976.
Case No. 1 - Miyazaki City
This Bio-Surf plant is installed adjacent to a land area created by garbage
dumping. Leachate from this landfill flows to a river and at a point down-
stream is withdrawn as river water for rice field irrigation. The nitrogen
content of this river water was found to be the principal cause of the
previously discussed poor rice yield and several studies were made to select
a process to solve the problem. Unit processes evaluated were activated
sludge, trickling filters, lagoons, stripping and the RBC process. Simplicity
of construction, operation, maintenance and low energy proved to be the
decisive factors and the Bio-Surf Process was selected.
The leachate is characterized as a low BOD (10 - 20 mg/1), low suspended
solids (2 - 30 mg/1), normal pH (7.6 - 7.9), low phosphorus (non-detected)
and high ammonia (100 - 120 mg/1). Alkalinity was sufficient for nitrifica-
tion (800 - 1,100 mg/1), and temperature was expected to be at least 15°C
(59°F) year round.
Design loading criteria for 6005, hydraulic flow, NHg-N, equivalent detention
time, methanol dosage, N02*-N and NO^-N are summarized for the total installa-
tion in Table 3. Influent BOD5, total nitrogen, and suspended solids values
were defined as 50, 200 and 100 mg/1. Respective final effluent values were
defined as 20, 50 and 25 mg/1.
Figure 1 illustrates the process flow and key sampling points for the
Miyazaki installation. Two 2-stage Bio-Surf shafts in series were provided
for BOD removal and nitrification. Each shaft was 7.5-m (25.6 ft.) in length
and media diameter was 3.6-m (11.8 ft.), for a total surface area of
18,680 m2 (200,810 ft2).
311
-------
A single 4-stage Bio-Surf shaft of special design for completely submerged
operation with 8,490 m2 (91,265 ft2) was provided for denitrification. For
reaeration, and residual methanol oxidation, a small 4-stage Bio-Surf shaft
was provided. Media diameter is 2.0 m (6.6 ft) on a 4.5 m (14.8 ft) shaft.
o o
The single and final clarifier provided an overflow rate of 15.8 nrvm -D
(388 gal/ft2-D) and a detention time of 4.7 hours. Underflow solids are
transported by truck for disposal elsewhere.
Plant startup occurred on October 15, 1976 at a flow of 157 m3/D (0.041 MGD) ,
or 45 percent of design. Temperature was 20°C (68°F) , BOD was about 10 mg/1,
NHg-N was 118 mg/1, and pH was 7.9 units. After two weeks small amounts of
N02~N were detected and startup seemed to be progressing very well. Tempera-
ture dropped within another two weeks to 12°C (54°F) and, disappointedly, no
~N was detected.
A review of the design data pointed out that lack of phosphorus may be a
factor. On November 29th, phosphoric acid was added to a 3.2 mg/1 as P con-
centration. Three days later NO^-N rapidly increased to a maximum value of
39 mg/1. Six days later the N02~N began to decrease and by December 20th,
nitrification was essentially complete.
Denitrification followed and was complete in 30 days. Figure 2 illustrates
the startup NHg-N, NO£-N and NO^-N concentration profiles during the 4-month
October 1976 through January 1977 period. It is clearly evident from this
graph that satisfaction of the phosphorus deficiency was the key factor to
proper nitrification, as all other factors were in the usual order (tempera-
ture, dissolved oxygen, etc.).
Table 4 summarizes nitrification performance for the January- November , 1977
period. Plant flow gradually increased to 71 percent of design but the
NH3-N concentration was at or slightly in excess of the 100-120 mg/1 design
for most of the year. Nitrification was greater than 96 percent, particularly
in the May-November period, with final effluent Nt^-N concentrations averaging
3 mg/1.
Denitrification was also very successful and the total nitrogen effluent
requirement of 50 mg/1 was easily met. Although data is sparse for nitrogen
forms other than ammonia or the oxides, evidence as early as January, February
and May revealed less than 20 mg/1 was being discharged routinely.
Figure 3 illustrates NH3-N removal as a function of NH3-N loading. Dots
represent full-scale performance at Miyazaki. Squares denote pilot plant
data developed elsewhere on municipal wastewater in Japan. The correlation
is extremely good and signifies no difference despite the great difference in
wastewater source.
Figure 4 illustrates ammonia nitrification as a function of influent BOD5
concentration. It is quite evident that removal decreases rapidly when BODs
concentration exceeds 35 mg/1. This represents a confirmation of information
held during the design phase in 1975' that the converse was true (that high
degrees of nitrification would occur in the region of 30 mg/1
312
-------
Figure 5 illustrates NC^-N and NC^-N removal in the denitrification unit
process as a function of loading to the submerged Bio-Surf shaft. The data
is again contrasted with pilot data on municipal waste (Dots are full-scale,
squares and triangles are pilot scale.) From'Figure 5 it is clear that con-
centration is not a factor, as 90 percent removal is attained at comparable
loading rates despite concentration differences of two or four-fold.
Figures 6 and 7 provide an illustration of supplementary data with respect
to nitrification and denitrificationl Samples were taken from the individual
four stages of treatment in each case and various analytical parameters are
plotted to show the fate of each during progressive degrees of treatment.
These data are somewhat limited in that samples were taken on only one day in
each case. However, the data is believed quite representative.
It is evident from Figure 6 that nitrification is nearly complete after only
two of the four stages' of treatment. That some denitrification was achieved
is evident from the total nitrogen line which shows a significant decrease
occurred during nitrification. This is not surprising, as other Bio-Surf
studies have shown that simultaneous nitrification—denitrification reactions
occur in the early stages of treatment when soluble BOD is available as a
carbon source.
Similarly, Figure 7 shows denitrification is essentially complete by the
second stage of treatment. Dissolved oxygen concentrations historically are
zero in Stage One and thus support the small amount of denitrification that
is shown, to occur there. These and other stage data confirm that residual
dissolved oxygen entering the first stage with the NQ^-N is preferentially
consumed by the heterotrophic bacteria and that not until Stage Two is
NOj-N the sole source of oxygen.
The Miyazaki plant demonstrated its ability to perform under high hydraulic
overload on November 16, 1977. Heavy rains resulted in a flow of 534 m^/D
(0.141 MGD), or 152 percent of the design maximum flow of 350 m^/D. Despite
this hydraulic overload, effluent 8005, N^S-N, NOg- N and TSS values were 5,
0, 0 and 10 mg/1, respectively.
CaseMo.2 - Yokkaichi City
This is a similar application for nitrification and denitrification at a
garbage dump, although both hydraulic and nitrogen loads are significantly
higher (See Table 2).
Experience gained at Miyazaki City and elsewhere resulted in a new design
rationale to minimize or eliminate entirely the methanol requirement as a
carbon source for denitrification. Bio-Surf nitrified effluent at this
installation will be recirculated and combined with raw sewage in a sub-
merged Bio-Surf reactor. It is predicted that a recycle ratio of 3 parts
of nitrified effluent to one part of raw sewage will essentially eliminate
and significantly reduce the need for methanol.
313
-------
The installation consists of two systems in series, with the first half
employing two fully-submerged, anoxic reactors followed by five standard
submergence aerobic reactors for the nitrified effluent recirculation scheme.
The latter half of the installation has a fully-submerged anoxic shaft follow-
ed by a standard submergence aerobic reactor for reaeration and BOD polishing.
Table 6 summarizes design loadings and Figure 8 illustrates the process
schematic for this plant designed to produce a final effluent with zero con-
centrations of BOD, NH3-N and N03-N.
314
-------
Design
FIGURE 1
PROCESS SCHEMATIC AND SAMPLING POINTS
-MIYAZAKI CITY-
^
^
i — i
i
No. I
BIO-SURF FOR NITRIFICATION
3.6fflp * 7,5ml i 2 slogs
Areo : 9340 m?
Volume : 50 m3
Oetanlioii Tlm« : 3,43Hr
Design
Flo«: 346.9 m3 0
PH: 7.0 - a.6
BOO'. 2O mg/l
SS: 25 mg/l
T-N: SO rng/l
BIO-SURF FOR BOO REMOVAL
Z.OmP • 4.5ml « 4
Area '. I3OO
Volume : 6,3 m3
Tim«: 34min
8OD tood ". 20 g/ m2-D
No, 2
810-SURF FOR NITRIFICATION
7.5f*l * 2 slog*
Aieo ". 9340
Volume : 50
Detention Time
810-SURF FOR DEVITRIFICATION
. 7,5ml * 4
Ar«a : 849O m2
Volume: 144 m3
Detention Time : 9.87 Mr
NO3-N Load: 7.4 g/mz-0
-------
START-UP PROFILES-MIYAZAKl CITY
FIGURE 2
INF. NH3-N (mfl/ I )
EFF. NH3-N (m<>/ I )
EFF. N03-N (m«/ I )
EFF,, N02-N (/ I )
OCT
-------
i
o
5
UJ
Z
UJ
o
tt
o
5
5
iOOi-
90
FIGURE 3
BIO-SURF NITRIFICATION
MIYAZAKI CITY VS. MUNICIPAL PILOT DATA
MARK •
MARK
INF, NH -N I
TEMP
INF. NH3-N
TEMP
i 7 ~ 33° C
: 8~ 2Om«,
,' 17~ 22°C
80
NH:, -N LOADING ( Gr / M2 D )
-------
FIGURE 4
AMMONIA REMOVAL VS. TOTAL BOD
H
00
100
z
U
0
O J
£ 5
51
3 £
z
o
5
90
80
OO
O
O
J 1 I
O
INF. NH3-N I
TEMP. I 7 ~ 33° C
I I I
10
20 30
4O
50
BOD CONCENTRATION (m9/ I )
-------
FIGURE 5
BIO-SURF DENITRIFICATION
MIYAZAKJ CITY VS. MUNICIPAL PILOT DATA
100
o
5
w
K
Ul
0
o
o:
90
80
70
UJ
K
H
z 60
UJ
t"
K
t 50
MARK »
MARK
MARK
INF. NOa -N : 0~ 2 mg / I
INF. N03 -N : 35~ 110 mg / I
TEMP. I 7~33° C
INF. NO3- N ; I I~I8 mg / I
TEMP, : I7~22°C
INF, N03- N i 4~I8 mg/ I
TEMP. ; I8~2I°C
-------
NITRIFICATION SECTION
•s.
o>
E_^
Z
o
H
cc
u
o
z
o
o
160
140
120
100
8O
60
40
20
FIGURE 6
(OCT-26-77)
IOOO
8 ~
7 -
6 -
o
o
INF.
I
2
I
3
I
4
900
800
700
600
500 '
4OO
300
2OO
STAGE NUMBER
320
-------
FIGURE 7
STAGE DATA-M1YAZAKI CITY
DENITRiFlCATlON SECTION
(NOV-9-77)
50
r 40
e»
£
•z.
o
ec
30
2O
g 10
u
•z
o
o 0
- 5OO
PH
INF.
STAGE NUMBER
321
-------
FIGURE 8
FLOW DIAGRAM OF YOKKAICHI CITY
GARBAGE WASTEWATER TREATMENT PLANT
ACTIVATED CARBON
U)
to
to
EQUALIZATION
TANK
HjPOa UiOH ft Cl3 ACID ALKALI POLYMER
CHLORtNATION
is
NO.l
DENITRIFICATION
NO, 2 MIXING TANK
OENITRIFICATION
SLUDGE
STORAGE
GRIT REMOVAL
NO.l BOD REMOVAL AND RE-AERATION FINAL CLARIFIER
NITRIFICATION
SAND FILTER
STORAGE
-------
1 1 *
PRIMARY
SEDIMENTATION
i
i
STATION
SP3
/.
ROTAT.N9
CONTACTORS
spy
FINAL
SEDIMENTATION
I
1
i
t
CHLORINE
CHAMBERS
AERATION
LADDER
SP6-T
/2MILE DRAINAGE
ICOMMINU
U)
to
w
AERATED
SHIT
CHAMBER
/ FLUME\
S*l
BAR
SCREEN
SUPERNATANT
SLUDOE RETURN
FILTRATE RETURN
WASHWATER RETURN
LAND
DISPOSAL
PIGURE I.
SEWAGE TREATMENT PLANT FLOW DIAGRAM AND SAMPLE POINT(SP) LOCATIONS
-------
Table 6
DESIGN PARAMETERS - YOKKAICHI CITY
LOADING RATES
A. NO. 1 DENITRIFICATION
BOD5: 17.4 g/m2-D (3.57 #/1,000 ft2-D)
Hydraulics: 140 l/m2-D (3.45 gpd/ft2)
Methanol: None
N03-N + N02-N: 6.3 g/m2-D (1.29 #/1,000 ft2-D)
Time: 0.9 Hours
B. NO. 1 NITRIFICATION
BOD
: 1.7 g/m2-D (0.35 #/1,000 ft2-D)
Hydraulics: 50 l/m2-D (1.23 gpd/ft2-D)
NH3-N: 3.0 g/m2-D (0.62 #/1,000 ft2-D)
Time: 2.6 Hours
C. NO. 2 DENITRIFICATION
BOD5: 13.6 g/m2-D (2.79 #/1,000 ft2-D)
Hydraulics: 76 l/m2-D (1.87 gpd/ft2-D)
Methanol: 108 kg/D (238 #/D)
N02-N + N03-N: 4.5 g/m2-D (0.92 #/1,000 ft2-D)
324
-------
OPERATIONAL ADVANTAGES OBTAINED
BY INCORPORATING A BIO-DRUM™
IN AN ACTIVATED SLUDGE PROCESS
By
GEORGE R. F1SETTE, P.E.
Product Manager
Ralph 8. Carter Company
192 Atlantic Street
Hackensack, N. J.
Introduction
In the continual search for better methods of wastewater treatment,
several processes have been developed In recent years that are vast im-
provements over the basic trickling filter and activated sludge process.
These new methods have included pure oxygen activated sludge, static tube
aerators, rotating biological contactors and various types of improved
media for trickling filters. The most recent addition is a process that
obtains a synergistic benefit from combining the activated sludge pro-
cess with a rotating biological contactor (RBC).
This process concept was developed in Denmark during 1972. There are
a number of Bio-Drum^ systems operating in Denmark, Southeast Asia, Saudi
Arabia, Canada, and the United States.
The Bio-Drum device consists of a steel cylindrical cage 8 feet in dia-
meter by 8 feet long, a media within fhe cage consisting of polyethylene
balls, a drive system, and mounting arms with bearings to enable multiple
drums to be connected together on a single drive system. The cage con-
sists of a structural metal framework around which is attached a perforated
screening -used to retain the balls. The end plates of the cage are design-
ed to move inward to compress the balls in order to hold them in a rigid,
locked position. The balls are molded of high density polyethylene to which
carbon black has been added for ultraviolet protection. The drive system
consists of a two speed motor attached to a gear reducer which transmits
325
-------
power to a chain and sprocket assembly. Multiple drums can be connected together
by means of couplings and solid-lube, pillow block bearings. The bearings
selected never need to be lubricated and are designed to operate in a water
environment. The drums and drive assembly are designed to go together in
one, two or four drum assemblies. In this manner, a single drive motor can
rotate up to four drums on multiple axes. FIGURE 1 is a photograph of a Bio-
Drum.
In addition to the polyethylene balls within the drum, there are in-
stalled around the perifery of the drum polyethylene containers known as
"waterlifts". These water!ifts are part of the pumping and diffuser capa-
bility of the Bio-Drum.
While the utilization of fixed film media with the activated sludge
process is not a new concept' ' ', the Carter Activated Biofilm Method™
(CABM™) is the first process that effectively combines them in a single
basin. FIGURE 2 is a schematic of the CABM process, see also TABLE 2, a
preliminary design for a 2.5 MGD Bio-Drum CABM system. This rotating bio-
logical contactor process makes use of sludge recycle and a high strength
suspended solids culture. The basic process concept is to obtai-n a very
high rate biological removal process utilizing an extremely large bacteria
inventory in order to keep detention time and basin size to a minimum. The
typical hydraulic detention time is on the order of one to two hours. This
results in a basin 1A to 1/10 the size of that for conventional activated
sludge systems. It is the large inventory of bacteria obtained from both
the suspended culture (MLSS) and the biofilm on the polyethylene balls that
results in an overall design utilizing a minimum of equipment, land area,
and basin volume. Typical MLSS concentrations in the Bio-Drum basin are on
the order of 5,000-10,000 mg/1. Combined with the bacteria slime on the Bio-
Drum balls, a total effective suspended solids concentration of 15,000-25,000
mg/1 is obtained. Since the energy input of the Bio-Drum is into a very
small basin, one obtains a process with a high degree of efficiency since the
drive motor of a dual Bio-Drum is only 5 horsepower.
The Bio-Drum, utilizing the CABM process, is a high rate, short detention
type process that is extremely stable to shocks, peaks, toxic compounds, and
is extremely efficient'^' ' ' -) due to the combination of fixed film and
suspended growth media. The low horsepower, typically 10-15 horsepower per
MGD, provides for a low cost operating system. In addition, the simple design
of the Bio-Drum further minimizes routine maintenance and operating cost. The
Bio-Drura due to its floating nature, greatly simplifies both construction cost
and installation cost. Indeed, it can be easily retrofitted to existing basins
and lagoons with minimal changes to existing structures. A further and highly
significant advantage of a Bio-Drum system is its ability to provide for both
BOD and ammonia removals in a single basin in a single stage process. This
ability to obtain ammonia removals in a single stage system further reduces
the amount of capital equipment, operating processes, and land required to
meet present and future effluent standards.
MECHANICAL DETAILS
As previously described, the Bio-Drum consists of three major components--
the cage, the polyethylene balls, and the drive assembly. The drum (cage)
consists of an angle-iron sub-assembly that is designed to support the weight
326
-------
of the completed drum. To this sub-assembly the covering of expanded metal
(screening) Is attached. The two circular end plates for the drum are designed
on a tracking system to move Jn or out approximately 6 inches each. These
movable end plates are advanced by means of bolts at numerous pressure points
around the cross-sectional area. The structural metal is first wire brushed
then the completed drum and covering is given a finished coat of a rust in-
hibftive paint. The ping-pong sized (38 mm) balls are manufactured out of
high density polyethylene to which carbon black has been added for UV (ultra-
violet) protection. These balls are produced by an injection technique known
as blow-molding. This results in a ball with minimum seam and no perforations.
During manufacture, each individual ball is tested to be surethereare no holes
or leaks in any individual ball. Approximately 1/4 of a million are included
in each 8 foot diameter by 8 foot long cage. In addition to the balls, poly-
ethylene containers of approximately one liter each are attached to all of
the periferial structural members around the outside edge of the cage. These
waterlifts are attached to the structural angle iron by means of plastic ties.
FIGURE 3 shows a waterllft in use. In order to show detail in the photo-
graph the lift is inside a small box to keep the balls back. The box does
not exist in a working drum. The center axis of the cage consists of a thick
walled 8 inch diameter pipe. Flanges are then welded to each end of the pipe
for attaching the drum to the drive component or companion drum assemblies.
The same basic drive assembly is used to power one, two, or four connected
drums. The drive assembly consists of a two-speed motor, a gear reducer, and
chain and sprocket sub-assembly. The 2-speed motor is either 3,5, or 7 1/2
Hp as necessitated by the number of drums, and runs at 900/1800 RPM. The motor
is norma 11 y operated in 900 RPMcondition with the higher speed being saved for
temporary process conditions. The gear reducer is a Koellmann gear reducer
with a 43 to 1 reduction ratio. This gear reducer operates in a flooded hy-
draulic condition using standard greases and lubricating oils. It comes with
a sight tube for continual monitoring of the oil level within the gear reducer.
The chain and sprocket sub-assembly consists of a 5" drive sprocket mounted on
the gear reducer and a 39" master sprocket mounted on the companion .flange of
one of the drums. The chain is a heavy duty roller style' chain with a master
link. The supporting structure for the Bio-Drum consists of a.torsion-tube yoke
which attaches to the basin walls or other permanent structure. The drum axis
Is attached to the support arms by the means of solid-lube pillow blocks. These
bearings are designed to operate in a flooded and wet environment. The bearings
do not need any Iubrication, and as they wear the bearings inner surfaces can be
rotated to increase the life of an individual bearing. Multiple drums are
connected together by a coup 1 ing which allows for some independent movement of
the drums relative to each other. All drums are connected to the common drive
system. FIGURE 4 shows a single drum with its mounting yoke and drive system.
When the Bio-Drum assembly is installed andoperating, the manufacturer
supplies an initial startup service that consists of verifying the proper
wiring and rotational direction of each drum assembly and provides instructions
on the simple adjustment of the end plates to maintain the balls in the state
of proper compression. It is to be expected during the first several months
that the end plates would have to be adjusted several times as the balls work
Into their final permanent locations. After this initial shakedown period, no
further service or attention need be given to the balls or drum structure itself.
327
-------
The only routine maintenance that is required will be for checking and lubri-
cation of the gear reducer. It is advisable to have on hand as spare parts
several master links and several feet of chain in the eventuality that a
chain breaks or is damaged. Routine maintenance to the chain and sprocket
assembly would consist of every 6 to 12 months inspecting the chain and sprock-
ets for wear, it is recommended that at least every two or three years the
drive sprocket and chain be automatically replaced as a preventive main-
tenance procedure.
Under normal and anticipated operating conditions and environment. It
will not be necessary to provide a building or covers for the Bio-Drum assem-
bly. However, It may be advantageous for process and operator convenience to
have covers over the Bio-Drum assembly area. Due to the packed nature of the
Bio-Drum and the fact that the major volume and bacteria film inventory is
within the drum itself, sun, rain, and wind have a very minor effect on the
growth of the biofilm. It must be remembered" that the waterllfts are put—
posely designed to pour water over the balls and biofilm when they are ex-
posed to the atmosphere. Therefore, rain will have less effect than the fall-
Ing water built into the design of the drum. In cold environments, because
of the short detention time within the bio-basin, there is a very small tem-
perature change experienced in this area. As long as the incoming waste is
sufficiently above the freezing point, the small amounts of ice that may form
around the edge of the basin are not detrimental to the mechanical operation
of the Bio-Drum. In addition, operating experience and past history show that
since there is no splash, spray or -foam with a Bio-Drum system there is very
small tendency for ice to accumulate on any of the super-structure. However,
It must be remembered that cold water temperatures are detrimental to the
biological process and, in fact, result in a significant slowing (retarda-
tion) of the biological process. It may be for this purpose then, advantageous
to provide covers to retain heat within the biological system.
PROCESS DETAILS
The -2-speed drive system is incorporated into the Bio-Drum for when a
biological process is subjected to upsets. Whether these upsets be'brought
about by the normal durinal cyclic loading or by BOD peaks from industrial
wastes or by infiltration, the higher second speed Is available to increase
the overall performance of the biological system. Ity increasing the rota-
tional speed of the drum, it is possible to obtain greater oxygen transfer,
up to 4 times more, greater mixing, and more rapid intermixing of the in-
coming BOD with the fixed film bacteria. FIGURE 5 shows the effect of re-
cycle upon the operating performance and BOD removal of a rotating biologi-
cal contactor, specifically the Bio-Drum. To date imost rotating biological
contactors have operated in a plug flow mode without the benefit of solids
recycle. As a matter of fact, previous testing and! research done by both
manufacturers and independent consultants had determined/that a RBC system
received no benefit from the recycle of suspended solids''' '. Those RBC
systems presently using sludge recycle all require air diffusers and ac-
cessory blowers. However, due to the mechanical aerator capability of the
Bio-Drum, the Bio-Drum does not need supplemental diffusers and does obtain
a benefit from the recycle of suspended solids. Imvdeed, the benefit seems
to be as a result of a synergistic combination of the two types of biologi-
cal processes, fixed film and suspended growth. FHGURE 5 shows that with
recycle considerably more pounds of BOD may be appllied to a given cubic foot
of media or conversely, for a fixed BOD waste considerably less capital equip-
328
-------
merit Is required to obtain the same degree of removal.
While the Bio-Drum Is capable of sustaining a suspended culture (MLSS) of
2,000 to 10,000 mg/1, economics dictate that the concentration of MLSS be
about 7*000 mg/1. However, if there are external restraints; such as exist"
Ing tank sizes clarifier efficiency, etc; then the CABM process can be de-
signed to operate at lower MLSS levels.
The quantity of bacteria on the Bio-Drum media can be calculated by
noting the change in the flotation level of the drum as the film accumu-
lates, its bouyancy, and measurements of the film thickness and moisture
content. If all of the biofilm were removed from the B!o-0rum and placed
in suspension in the basin an effective MLSS of 15,000 to 25,000 mg/1 would
be obtained.
A further benefit of the combination of fixed film and suspended growth
is the ability of the Bio-Drum with the CABM process to obtain removals of
both BOD and ammonia in the same single step basin. It is theorized that
this ability, not obtainable with other conventional RBC equipment, results
from the potentially older sludge age of the bacteria in the biofilm at-
tached to the drum itself. The recycle suspended solids are considered
to consist mostly of carbonaceous bacteria used in the removal of BOD. The
ammonia removing bacteria (NITROBACTER and NITROSOMUS) are .considered to be
attached to the bio-ball media. Because of the slower growth rate of these
nitrifying bacteria, they have a requirement of a considerably older sludge
age than the more rapid growing carbonaceous bacteria. A logical explana-
tion then, for the ability to remove ammonia at the same time as BOD re-
quires that there be a separate inventory of nitrifying bacteria, indepen-
dent of the recycle stream which would typically have a relatively young
sludge age.
It has been alluded that the Bio-Drum operates as a mechanical aerator.
Several unusual features have been incorporated, or are a natural part of the
design of the Bio-Drum. It is these features and design benefits that allow
the Bio-Drum to operate as a mechanical aerator. The most obvious feature is
the waterlifts which are attached to the external periphery of the cage. There
are approximately ]kk of these lifts in each drum. These lifts are capable
of providing a direct water pumping rate of 100 gallons per minute per drum
at'the normal rotational speed (high speed is 200 gpm). This means that a
single Bio-Drum is capable of providing a complete hydraulic turnover of the
water within its domain in less than 1 1/2 hours by direct pumpage only. This
does not take into consideration that movement of water brought about by rota-
tion of the drum structure itself; nor intrained flow. Thus, a single drum is
pumping by direct water movement approximately 150,000 gpd. As this pumped
water cascades down over the media during the atmospheric portion of each
cycle, oxygen is transferred from the atmosphere into this falling, cascading
sheet of water. Of course, the normal oxygen transfer as experienced by rotat-
ing biological contactors takes place during the atmospheric portion of the
rotation from the air directly to the bacterial slime adhered to the media.
The oxygen transfer, therefore, into the wastewater is above and beyond
that needed for the bacterial slime attached to the media. Secondly, as these
lifts, now empty of water, are resubmerged into the wastewater, the air that
is trapped within bubbles out of the lifts up through the media. In this
manner, the lifts are acting as a coarse bubble diffuser. TABLE 3 details the
oxygen transfer of various RBC's and diffusers. The direct air pumpage by the
329
-------
lifts is 13 CFM, or 19,000 cubic feet of air per day per drum. As this air
flow passes up through the media, oxygen is transferred Into the wastewater
as the bubbles are continually broken and re-agglomerated during their travel
through the rotating media. In addition to the oxygen transfer and contacting
provided by the waterlifts, a considerable portion of oxygen transfer and energy
Input Into the basin takes place as water and air flow into and out of the drum
during each rotation.
The Bio-Drum is designed to rotate at 2,6 rpm, giving a peripheral speed
of 1.1 ft. per second at the normal condition. The high speed condition used
for special process conditions Is at 5'.2 rpm, giving a peripheral speed of 2.2
ft. per second.
As the drum rotates the water is continually falling out of the back side
of the drum, from within the media back onto the water surface, giving a wake
and a small wave. This additional agitation of the water and splashing action
within the drum again results in oxygen transfer to the wastewater. Because
of the many sources of oxygen transfer and the mechanical design of the Bio-
Drum, the system is able to support a high concentration of mixed liquor sus-
pended solids. The oxygen requirement for endogenous respiration and for BOD
oxidation is provided by the drum for the portion of the BOD removal that takes
place by the suspended culture.
Due to the extremely high concentration of bacterial solids, approaching
15,000-25,000 mg/1, a very rapid assimilation of the BOD takes place. Indeed,
the biological processes are known to be completed within approximately 15 to
20 minutes. The major requirement for the longer hydraulic detention time as
used within the Bio-Drum CASH design of 1 to 2 hours is brought about by the
necessity to protect the design for shock loadings, for hydraulic p'eaks, and
for the nonideality of the intermixing of the bacteria with the incoming BOD.
The basin size Is primarily a function of hydraulic flow rate and such
associated factors as peaking, hydraulic cycles, BOD, and temperature fluct-
uations. The basin conventionally Is designed with a depth only slightly
greater than 1/2 the diameter of the drum, (i.e., k.5~5 feet water depth).
The clearance between the bottom of the drum and the basin floor is kept as
small as possible within the restraints of the design. The basin is norm-
ally designed in a fairly square to slightly rectangular pattern in order to
optimize the completely mixed nature of the process. Conventional wisdom
dictates that for most waste treatment plants, a parallel scheme be incor-
porated whereby two parallel biological basins are provided, each with a
capacity slightly greater than half the hydraulic design.
An important consideration to the proper operation of the CABM system is
the ability of the Bio-Drum to keep the mixed liquor suspended solids in uniform
suspension. At the concentration levels normally encountered of 5,000-8,000
mg/1, It is necessary that sufficient energy be put into the wastfewater to
prevent separation and deposit ion of the solids. Even though the total oper-
ating horsepower of a Bio-Drum system is quite low, the applied horsepower,
I.e., the energy imparted directly into wastewater within the basin, is quite
high. Under normal design considerations, the energy applied directly to the
wastewater by the rotating drum system Is in excess of 2.5 hp./k ft?» In
addition, the higher speed capability of the drive system allows this value
330
-------
to be raised !n those circumstances where a faster settling solids or a higher
density solids is obtained. Furthermore, the clearance between the bottom of
the drum and the basin floor is kept to a minimum in order to maximize the
scouring velocity along the bottom of the basin to resuspend any solids that
have a tendency to settle towards the bottom of the basin. The high direct
pumping capacity of the drum itself of approximately 150,000 GPD, greatly con-
tributes to the ability of the total Bio-Drum CABM system to maintain uniform
solid suspension. In a basin with an average hydraulic detention time of 1
hour, the capability of the Bio-Drum by direct pumping alone to turn over or
recirculate the entire basin volume in approximately 1.5 hours provides add-
itional evidence of the level of agitation within the Bio-Drum basin.
A second and equally important constraint is that the Bio-Drum exhibits
no tendency for internal clogging by excessive bacterial growth, sloughing
off of excess growth, or developing anaerobic areas within the internal volume
of the drum itself. Two features of the Bfo-Drum mechanical design were inten-
tionally incorporated to alleviate this problem. However, the proof of the
drum's ability to perform without clogging and without going anaerobic is best
obtained from evidence of operating units. The units that have been installed
and operating for 3 to k years have shown no tendency to clog as evidenced by
the ^lack of foul odors, or a decrease in the performance rating of the drum
system with time. If the internal area of a drum were to clog or go anaerobic,
or otherwise decrease the performance capability of the Bio-Drum, then a detei—
ioration should be noted in the quality of the product produced by the treatment
plant. To date there has been no evidence of a decrease in performance, rather
the system appears to increase in removal ability with time. The rotational
speed of the drum, combined with the small interstitual spaces, means that the
water veloci ty around the balls and through the spaces must be very high. The
size of the balls was purposely selected in order to arrive at the minimum
practical void space of 35? • In this manner, the Bio-Drum design maximizes
the surface area available for bacterial film to attach and maximizes the vel-
ocity of the water through the open spaces within the packing. Combined with
these high internal water velocities, is the scrubbing effect obtained from
the rising air bubbles released by the waterlifts. These rising air bubbles
provide an expanding action or scrubbing action that, in conjunction with
the high water velocity, maintains a uniform thickness of biofilm on each ball.
These re leased/removed bacteria particles are removed as discreet and small
particles within the flowing water stream as the drum rotates. In this manner,
large chunks of matter as typically slough off a trickling filter have been
prevented. Furthermore, during the atmospheric portion of each rotational
cycle, the bacterial film is further rinsed with a gentle irrigation of water
released from the waterlifts. This gentle flow further tends to remove any
loosely attached and light particles as it flows down over the balls during
the rotational journey. It should also be realized that any plugging tend-
ency which Is exhibited becomes self-defeating. In that as a particular
area would tend to plug, the interstitual velocity in that area would increase
and the excess materials woufd be rapidly removed due to the higher velocity
of water within that section of the drum.
Since it is conceivable to use Bio-Drum in a treatment plant that does
not have a primary clarifier, consideration must be given to preventing
clogging or plugging of the drum screen by large foreign objects. -Such
things as rags and plastic bags, etc-, must be removed from the drum or
331
-------
prevented from reaching the drum in order to prevent an aesthetic problem,
a potential odor problem, and some interferences with the hydraulic flows
within a drum. Therefore, it is recommended that in those applications with-
out a primary clarifier, that a bar screen of some design and a grinder
should be utilized in order to reduce the size of the extraneous objects and
their potential for adhering to the drum screening. In this manner, since
the drum screen itself is very smooth and presents a small profile for string-
like objects to hang on, the tendency of the drum to have string or cloth
attached to its outside surface should be minimized.
PROCESS DESIGN
As with all process equipment, a method of calculation or determining
the size and quantity of equipment necessary to process a given load is
necessary. While the theoretical equations have been derived for both the
activated sludge system and for biofilm devices, at present there are no
known equations describing the combined system. Therefore, the Bio-Drum
and CABM system are designed on a semi-empirical basis as follows.
Based upon both pilot and full-scale testing and operating plants, a
relationship has been determined which related the applied BOD, the cubic
foot of drum, and the desired quality of effluent. The design parameter
used to determine the quantity of drums required is a volumetric loading
factor, F/V (ibs. BOD/day/ft.^ ). This loading factor is used in the manner
similar to the design parameter used for sizing trickling filters and plastic
media-type filters. In conjuction with the drum loading parameter the F/M
ratio (Ibs. BOD/day/lb. MLSS or MLVSS as preferred), is used in an iterative
(trial and error) scheme to arrive at the final design. A balance must be
arrived at between the F/M ratio, the F/V ratio, and the basin size.
FIGURE 6 shows the interrelationships between BOD removals, F/V, and F/M.
At a given F/V loading, the F/M must be decreased in order to improve perfor-
mance. For a given BOD waste, F/M can be decreased by raising the MLSS con-
centration or increasing the basin volume.
The objective is to maximize the total solids inventory while maintaining
the F/M concentration within a practical range that can be maintained by the
Bio-Drum. This means that the typical F/M value obtained is directly related
to the basin size and the concentration allowed, which ranges from 5,000 to
18,000 mg/1. The basin size is primarily governed by flow considerations.
Some allowance in basin size must be considered for unusual BOD or ammonia
requirements. However, the normal basin has a capacity of 1 hour hydraulic
detention time. The upper range of the basin size is approximately 2 hours
hydraulic detention time. Once the basin size has been established, then
the MLSS concentration allowed can be determined and the F/M ratio calculated
at this point. The F/M ratio so calculated must then be related to the F/V .
ratio to verify that the F/M is within the operating range at the F/V loading
selected. TABLE 2 is a preliminary design calculation for a 2.5 MGD municipal
waste.
Of prime consideration in the operation of the CABM process is the design
and operation conditions of the clarifier and recycle flows. It is extremely
332
-------
advantageous that the recycle flow be kept to as small a value as possible in
order to minimize the clarifier surface area required for proper sedimentation
and solids separation. Therefore, the clarifier design selected must have a
highly efficient separation capability and a high degree of thickening activity.
It is imperative that the solids underflow be at the maximum percent solids
concentration obtainable. With proper design of the clarifier and thickenings
action, the recycle flow would be in the range of 10-100% of the inlet flow.
Obviously, it is preferred to minimize the recycle flow in the range of 10-301.
This high concentration is obtainable even though the MLSS concentration is
quite high, because of the improved settling characteristics of.the sludge
that results from the combined biofilm activated sludge system/ '
In addition to these major factors affecting the design and process, there
are several minor parameters which must also be considered in arriving at a
polished design. In municipal designs, and even in some industrial applications
considerations must be given to hydraulic peaks. These may be the conventional
daily cyclic flows as evidenced by most municipal plants, or they may be
process related flow peaks dependent upon the manufacturing processes of a
specific plant, wherein washwater of batch-type processes are dumped at dis-
creet time intervals. Furthermore, there are really two types of hydraulic
peaks. In one case, the BOD concentration remains constant; in other'words,
the total load of BOD incoming to the plant increases in direct proportion
to the hydraulic flow. In the other case, the pounds loading of BOD incoming
to the plant remains essentially constant and we basically have dilution of
the normal flow by increased hydraulic Input. In evaluating the effect of
these hydraulic flows on a Bio-Drum design, if the duration of the peak is
of the order 1/2 to 1 1/2 hours, little or no changes are necessary in the
basic design for a Bio-Drum system. On the other hand, if we are looking
at hydraulic surges that may run 2 hours or more, then consideration must
be given in either case to the effect of these flows on the Bio-Drum pro-
cess. If the increased BOD loading during hydraulic peaks represents a sign-
ificant portion of the day, then the plant will have to be oversized to allow
for proper removal of these increased loadings. In the case of dilution or
infiltration problems, the major constraint will be on the design and operation
of the clarifier and recycle streams, to prevent the washout of biological
solids. However, in either case the fact that both a fixed film culture
and a suspended culture are available for treatment means that the effects
of these hydraulic peaks may be rapidly neutralized.
If the plant Is experiencing an upsurge in BCD, this can be offset and
performance maintained by increasing the amount of recycle and changing the
drum to its higher speed of operation. The system is now capable of providing
additional oxygen and suspending additional solids to treat this surge In
BOD. In the condition where one obtains a hydraulic surge without corres-
ponding BOO Increase, the optimal alternative is to increase sludge recycle
only slightly and maintain a close watch on the clarifier performance such
that excess solids do not wash over the top and to prevent an excessive
buildup of sludge and a sludge blanket. It is also possible to obtain
surges in BOD without corresponding changes in the hydraulic flow. Uner this
condition, It is only necessary to Increase the solids recycle and increase
the drum speed to provide the additional activity necessary for removal of
the increase in BOD. Again, consideration must be given in the normal
design to the magnitude of this Increase relative to the average design
333
-------
for which the plan was originally designed.
A somewhat different type of peaking problem is that experienced in small
municipalities and industries which operate on a 5 day or 6 day cycle, namely
the lack of flow on weekends or for other extended periods. With a convent-
ional rotating biological contactor there is an extreme tendency over these
low periods for the rotating contactor to have such a high degree of evaporation
that it essentially dehydrates the entire basin over the course of one or two
days. The prevention of this occurring, therefore, requires these type of plants
to be designed with a hydraulic recycle strictly to maintain sufficient hydraulic
flow to keep the bacterial cultures wet and active. In the Bio-Drum and CABM
process, however, the recycle flow is already incorporated within the design
and due to the high suspended solids inventory there is always sufficient
bacterial available to act as the food source for the fixed film culture.
In other words, the process can go into an endogenous type of respiration
or cannibalistic activity during these periods of starvation.
A more severe problem associated with peaks and spasmodic flows is the
unfortunate occurance of a toxic dump or poison entering the system. Under
these conditions, the advantages of a combined fixed film and suspened growth
system shows its superiority to the operation of either type of singular sys-
tem. If the system were a pure fixed film culture system, the toxic load
would quite probably wipe out the entire plant and severely hamper removal
and process efficiency. On the other hand, with a strictly activated sludge
type culture it is possible that the sludge would never acclimate to the
constituants of the toxic input and, therefore, each and every time a toxic
cyclic occured the activated sludge plant would also suffer in performance
capability. However, in the combined Bio-Drum CABM system the biofilm can
adjust over time to the toxic material and provide a reserve bacterial in-
ventory capable of treating the toxic component.
As mentioned previously, temperature, per se, does not affect the mech-
anical reliability of the Bio-Drum design. However, the lowering of temp-
erature during the winter does severely hamper the biological processes.
Therefore, the Bio-Drum system must be designed with consideration for this
effect. If one wishes to maintain the same degree of removal in the winter
as in the summer without the use of heat or covers to retain heat, then the
Bio-Drum system must be oversized in order to produce proper removals during
cold weather conditions. Because of the packed type nature of the Bio-Drum
and the fact that the water surface and film surface on the media are not
directly exposed to the atmosphere; the evaporation and heat loss from a
Bio-Drum system is minimized in comparison to other types of biological
processes. This means that in temperate climates where only occasional
cold weather is experienced the Bio-Drum basin will not tend to drop its
temperature below the freezing point as Is common with other biological
contactor devices.
COMPARISON TO OTHER PROCESSES
Inspection of TABLE k shows that the CABM process is quite similar in
design perimeters to both the high rate activated sludge process and the
contact stabilization sludge process. Comparison of the Bio-Drum to fixed
334
-------
film type processes, listed in TABLE 5, shows that on an organic loading per unit
volume of media the Bio-Drum is, in fact, considerably higher than conventional
trickling filter and RBC type devices. However, if one looks at the hydraulic
load in gallons per day per square foot of media, the Bio-Drum device is loaded
considerably less than conventional trickling filter designs and well below the
hydraulic loading for conventional or improved processes. Just as loading peri-
meters and loading values are specific for individual processes and different
from all others;the Bio-Drum and CASH have specific design and loading perimeters
that relate only to the Bio-Drum design. However, if one puts the design peri-
meter of all processes on a comparable basis wherein the loading is measured in
pounds of act ive bacteria, which ?s nearly impossible with fixed film devices or
per pound of suspended solids (MLSS) then it can be seen that the Bio-Drum is
really no different than any other biological process.
Another measurement or comparison of the various processes that can be eval-
uated is the energy requirement, i.e. horsepower, utilized in the removal of BOD
and in the operation of the various types of processes. With a few conservative
assumptions, the horsepower required to treat one million gallons per day of raw
waste can be estimated for the various generic types of equipment; mechanical
aerators, diffused aerators, submerged turbine aerators, rotating biological con-
tactors, and Bio-Drum. TABLE 6, lists some of the values of horsepower required
to treat one MGD of waste. The 15 horsepower shown for the Bio-Drum is consider-
ably lower than the more conventional and expected numbers of ^0 to 50 horsepower
for various mechanical systems. However, these high numbers are really an indic-
ation of the inefficiency of the processes in that the greater the horsepower
required to treat an equivalent amount of BOD the more expensive the annual oper-
ating cost is for an individual plant. Furthermore, a comparison of the applied
horsepower, defined as that horsepower transmitted into the volume of water under
treatment within a given basin shows that the Bio-Drum is within the range of
"conventional" treatment processes. TABLE 6 shows these values of horsepower per
thousand cubic feet of basin water under treatment. Part of the reasoning for
the inefficiencies of the more conventional mechanical processes is brought about
by the size of the basins involved under treatment which require large quantities
of horsepower to maintain proper biological mixing and contacting.
SUMMARY
The difference between the Carter Activated Biofilm Methoa and other
biological processes can be summarized in a few words; the same, but with IMPROVE-
MENTS. However, the Bio-Drum1*^ differs from other biological treatment equip-
ment in being a more efficient device for removing BOD. The features and bene-
fits of the Bio-Drum and CABM process are summarized in TABLE 1.
335
-------
TABLE 1
Feature
floats
B10-DRUM
MECHANICAL
compact 8' x 81 size
low stress on bearings
low stress on drive
simplified design,
few moving parts
Benefit
reduced installation costs,
no supports required
easy to instal1
long 11fe
long life
low maintenance costs
CARTER ACTIVATED BIOFILM METHOD
PROCESS
high organic loading
high BOD removal/unit
short detention time
low Hp
BOD 6 NH-j in single basin
high bacteria solids inventory
built in compensation for BOD, flow,
and toxic variations
less capital.equipment
low capital costs
small basin & land required
low operating costs
less unit processes,
lowest capital costs
very stable process-
resist upsets
easy to operate
336
-------
TABLE 2
PRELIMINARY DESIGN BRIEF
CARTER BIO-DRUM™ and ACTIVATED BIOFILM METHOD
Project Location: Anytown, USA Date: 12/12/79
Consulting Engineer: RBC Consultants, Inc.
Someplace, USA
Attn: Joe Consultant, PE
Prepared by: G. R. Fisette, PE
BA<
(1)
(2)
(3)
CO
(5)
(6)
(8)
(9)
(10)
(11)
(12)
51 C DATA
a) Waste
b) Process
Design, Flow
BOD Influent to Bio-Drum Basin
BOO Effluent
Remova 1
NH-j Influent to Bio-Drum basin
WMPffliionf-
NH, Removal
j
SS Influent
pH range
Dissolved oxygen level to be maintained in the
Bio-Drum basin
Temperature of Waste in Dio-Drum basin a) Winter
b) Summer
Muni
Acti
2.5
A170
625
85
500
n/a
n/a
i»170
6-8
1-5
15
25
c i pal
vated Bi
MGD
Ibs
Ibs
°/
'0
Ibs
Ibs
%
Ibs
PH
mg/1
°c
°C
of i 1m
BOD/day
BOD/day
N-NH,/day
N-NH3/day
SS/day
337
-------
BIO-DRUHS
(13) Media loading rate (F/V) required 0.67 Ibs BOD/d/ft3
(14) Hydraulic detention time required 1.0 hrs.
(15) Vojume of Bio-Drums required™ 6224 ft3
/ t tern 13 \ ,
\ item #13J - (4170
(16) Number of Bio-Drums required= 16 uni ts
/item #15\ =1 6224\
\ 390 I \ 330 J
(17) Basin volume occupied by Bio-Drums=
-(Item #16) (127) - (16) (127)= 2000 ft3
(18) Total Basin volume =
a) f item #2 x 106 1 (5temll4\ + (item #17}
V TBD J \ I •
(2.5 x lO6^ f
• _g_ i ^>0^ +(2000) = 16,000 ft3
") (volume, ft3) (7.^8) - (16,00) (7.48) - 120,000 gal
Minimum area required at 5 feet water depth = ~
item ffloa \ == f 16,000
(20) Minimum area required for Bio-Drum units =
(item #16) (iMi) = ( 16 ) (144) = 2300 ft2
(21) Basin layout for dual assemblies*
a) Width » / item #19 or 20V = -^ ft
'-J
b) Length = (width, ft) (2) = 64 ft
(22) Power required = (item #16) (2.5) = (16) (2.5) - 4p BHp
338
-------
MLSS
(23) Solids loading rate (F/M) required
(2k) Pounds of MLSS In basin =
(item 13 ) - (^170)
Wss)
(item #23)
(25) Concentration of MLSS in basin =
(|tem_#2fj_x_10_6) - (7580 x 106 \
( (8.31*) (Item #l8b)) \ (8.3*0 (120,000) I
(26) Effective mass of MLSS in system =
(item #16) (6001 solids/drum) + (item #2*0 =
(16) (600) + (7580) =
(27) Effective concentration of MLSS in system =
(item #26)(TO6)
(8.3*0 (item #186)
(17.200) ( IP6 )
(8.34)(120,000)=
(28) Estimated Sludge Age of recycle solids =
0.55
7575
Ibs BOO/d/#MLS
» 7580 Ibs MLSS
mg/1
17,200 Ibs MLSS
17,000 mg/1
1 bs MLSS i n sys tern \
Ibs MLSS wasted/day/
/ 7580 -t- 10,600\
4170 * 0.75
item #2^ + clarifier solids
item #3
# MLSS/#BOO/d
5.8
days
CLARIFIER
(29) Approximate clarifier surface area
item #2 V/2.5 - IP6
overflow rate, GPD/ft2)\ 500
5000
ft2
339
-------
TABLE 3
OXYGEN TRANSFER CAPABILITIES OF
BIO-DRUMT/f AND OTHER AERATION DEVICES' '
20
Apparatus ICa SOTR No
Single 8' $ x 81 Bio-Drum
Normal speed § 1.1 fps 2.88 1.4 0.70
Dual 8' ^ x 8' Bio-Drum
Normal speed § 1.1 fps 3-33 2.8 0.85
Peak speed @ 2.2 fps 12.? 10.8 1.5
Pilot unit, 24" x 18'* Bio-Drum^
Low speed @ 0.5 fps 4.27 0.014 O.l6
w Normal speed @ 1.1 fps 8.40 0.028 0.25
o Mid speed § 1.6 fps 14.2 0.04? 0.33
Peak speed § 2.1 fps 22.7 0.075 0,48
High speed @ 2.6 fps 37-0 0.12 0.71
Rotating Biological Contactor (biodisc; '
12' 4> standard media-niechanical drive § 1.0 fps 1.18 0.17
12' 4> standard media-air drive @ 1.0 fps 4.10 0.58
12' 4* high density media-mechanical drive 8 1.0 fps 1.13 0.16
12' high density media-air drive @ 1.0 fps 3-55 0.50
Coarse-bubble diffusers^6' 18) — 0.8-2.0 0.8-1.6
Kj_a • Mass transfer coefficient, hr~'
SOTR * Standard mass transfer rate, Ibs 02/hr
No » Standard mass transfer efficiency, Ibs 02/Hp-hr
-------
TAttif
1VPICAL ACTIVATED StUDOl OESIGH IOAOIHII
Process Q2 Uptake OT
mg 02/hr
g MLSS Nrs
Conventional 7-15 '1-8
lll<|h Rate 15-25 2-
-------
TABLE 5
TYPICAL FIXED FILM PROCESS DESIGN LOADINGS
Process
Convention
w
to Intermediate
High Rate
Plastic media
Roughing filter
ABF«1, 13, 14)
RB(/S, 6, 11, 12)
CABM^W
Depth
ft
6-8
6-8
3-8
<_40
3-20
10-40
6
6
Remova 1
I
80-85
50-70
65-80
65-85
40-65
80-95
80-95
80-95
Organic
Loading
Ibs BOD/d per
kft3 acre ft
5-25
15-30
25-300
1300
MOO
100-300
70-150
500-800
200-1000
700-1400
1000-1300
?
?
-
-
-
Hydraul
MGD
1-4
4-10
10-20
15-90
60-180
-
-
_
ic Loading DT
GPO
IF" hrs
25-90
90-230
230-900
350-2100
1400-4200
2-4
1-4 1-3
8-20 1/2-2
-------
TABLE 6
COMPARISON OF POWER REQUIREMENTS
low speed surface
Turbine-sprayer
Diffused air, medium bubbles
Pure oxygen, cryogenic
RBC
Bio-Drum
ABF'9' 14>
Basis: 1 MGD of
Transfer Rate
Ibs 02/Hp-hr^
1.7
1.4
1.4
2.7
—
0.85
7
180 mg/1 BOD
Loading
Ibs BOD/d/kft3
»;
--
--
--
--
110
700
200
Hp/MGD
40
48
48
25
38
15
21
Hp/k#BOD/d
27
32
32
17
25
10
17
Mixing
Hp/kft3
1.2
1 jt
1 ^4
3.4
4.7
2.7
7
-------
FIGURE 1
2.5 MGD BIO-DRUM™ INSTALLATION
USING THE ACTIVATED B1OF1LM METHOD™
Clean
Effluent
to Sludge Processing
by 8FPrM ^
Sludge Recycle
1
12'
FIGURE 2
344
-------
FIGURE 3
i mi itii u^^t- «! rr i
1111111 n A i ;Tf ^Sm ^ >• .*!••'..
1 i t ( 1 I 111 *r« [ W *** ^fc —
lllHUIlMM W T
v-jitf—-. -.*-"* '"^v
FIGURE
345
-------
FIGURE 5
w
**
er>
4)
a.
2.
o
o
oo
100
90
80
EfTect of Sludcje Recycle upon Performnnce
of Rotating Bioloqioil Contactors
\
\
70
60
\ Rotating \
' Biological \
I Contactor Domain
\
\ (no sludge recycle) .
\
\
Bfo-0rumw Domain
th sludge recycle)
X
50
N
0.1
0.2
0.3
O.'l
0.5
0.6
0.7
0.8
Media Loading Rate, F/V, Ibs BOD/d/ft3
12/12/79
-------
T*M
FIGURE 6
Performance of the Carter Bio-Drum1" f, CABM1
For BOD Removal
100
c
0)
CO U
m
>
it)
on
o
o
to
80
70
50
0.1
0.2 0.3 O.*i 0.5 0.6
Bio-Drum Loading Rate, F/V, Ibs BOD/d/fl3
-------
References
1. R. S. Stenquist, et al, "Long-Term Performance of a Coupled Trickling Filter-
Activated Sludge Plant", JWPCF, %9 (Nov., 77), p. 2265-228%,
2. R. F. Roskopf, et al, "Trickling Filter-Activated Sludge Combinations",
JEED/ASCE, Oct., 76, p. 1005-1018.
3. WPCF/ASCE, 1977, "V/astewater Treatment Plant Design".
k. G. Shell 5 S. Spence, "The Bio-Drum™ and Activated Bfofilm Method™ ",
paper presented at WWEHA Conference, Atlanta, GA, V19/77.
5. R. L. Antonie, "Fixed Biological Surfaces-Wastewater Treatment", CRC Press,
1976.
6. W. N. Torpey, et al , "Rotating Biological Disc Wastewater Treatment Proce'ss
Pilot Plant Evaluation", E.P.A. publication PB232133, 197%.
7. Euro Matic, Danish report.
8. G. A. Gagnon and C. J. Crandall, JWPCF, "Review and Evaluation of Aeration
Tank Design Parameters", pp. 832-841, May 1977, Vol. kS, No. 5.
9. Neptune Microfldc brochure.
10. Envirodisc brochure.
11. 8enjes, "Evaluation of Biological Wastewater Treatment Processes", paper pre-
sented at Wastewater Treatment and Reuse Seminar, South Lake Tahoe, CA, Oct.
27-28, 1976.
12. R. L. Antonie, "Rotating Biological Contactor for Secondary Wastewater Treat-
ment", paper presented at WWT Seminar, South Lake Tahoe, CA, Oct. 27-28, 1976.
13- A. Slechta, "Activated Bio-Filter Process for Biological Wascewater Treatment",
paper presented at WWT Seminar, South Lake Tahoe, CA, Oct. 27-28, 1976.
1%. Metcalf 5 Eddy, Inc., "Wastewater Engineering: Treatment, Disposal, Reuse",
2nd edition, 1979, Mcuraw Hill.
15. W. F. Owen & A. F. Slechta, "Organic Removal or Nitrification With a Combined
Fixed/Suspended Growth Biological Treatment System", paper presented at WPCF
Conference, Miami Beach, FL, Oct. 5-10, 1975.
16. ASCE, EPA report, Proceedings: Workshog Toward an Oxygen Transfer Standard,
600/9-78-021, April 1979.
17. Robert J. Hynek and Charles Chi-Su Chou of Autotrol Corporation, "Development
and Performance of Air-Driven Rotating Biological Contactor", paper presented
at Pollution Control Association of Ontario Conference, April 1979.
18, Gerry Shell, private conversation, Oct. 1979.
348
-------
EVALUATION OF A ROTATING BIOLOGICAL DISC
IN A SEWAGE TREATMENT PROCESS
IN PACKAGE PLANT APPLICATIONS
BY
Bob Joost
Tait/Bio-Shafts, Inc.
Mike Veslo
Purestream Industries, Inc.
The Rotating Biological Disc Process, sometimes referred
to as the RBC, RED, RBS, Bio-Shaft, Bio-surf, Aero-surf, Enviro-
surf, etc. hereafter will be referred to, as the RBC Process.
This process caused a very significant impact in the wastewater
treatment market. This statement is obvious when one considers
the excellent program we are participating in at the First
National Symposium on Rotating Biological Contactor Technology.
However, the simplicity of process flow control, minimal
maintenance and operator attention, economical installed costs,
low energy consumption and according to many, the panacea for
all process problems, has given the RBC Process^ usage for many
applications. The Federal Construction Grant Program contributing
needed construction monies and a tremendous marketing effort by
disc manufacturers also spirited the advance and the industry
acceptance of the RBC Process.
Conservatively, there are over 5,000 RBC installations in
the United States and abroad. In the next ten years, this figure
will easily double. What caused this phenomenal growth of a
biological waste treatment process that dates back over 50 years?
What resurrected this almost forgotten biological wastewater
treatment process? First and foremost, immediately after World
War II Germany, with high population densities located in small
towns and villages surrounded by hilly and mountainous regions,
needed an inexpensive, dependable, low energy consumption and
minimal operator attention wastewater treatment system to produce a
high quality wastewater effluent. Regulatory agencies were
faced as \ve are today, with pollution of good water short rivers
and streams. The University of Stuttgart, Germany, evaluated
many processes to meet these water pollution control objectives.
Among the wastewater treatment processes investigated was the
349
-------
Rotating Biological Discs, Or, commonly then referred to in
Germany as the Immersion Drip Filter. The use of the relatively
new material, plastics, substituting for the wood and metal
biological support surfaces, described in the Weigand Patent (1)
and the Maltby "Biological Wheel" Patent (2), accomplished the
needed economies of the RBC Process. A European disc manufacturer
continued the usage of the RBC Process for relatively small sized
installations. The virtues of an inexpensive, dependable, low
energy consumption and minimal operator attention wastewater
treatment system to produce high water quality effluents made the
RBC Process ideal for low flow applications. The conservativeness
of the European disc manufacturer in process application and
structural design contributed heavily to the credence and
reliabilities of the RBC Process.
Independently, in the mid 60's, I headed a group at Allis
Chalmers Co, in Milwaukee that developed on its own a RBC Process.
The European work at that time was unknown to us. Later, after
our own developments and eventual communications with the European
disc manufacturer, we were convinced the RBC Process had
commercial application in the United States. The wastewater
treatment industry at this time, was not the high technological,
several billion dollar a year sales, regulated industry as we
know today. Generally, the sanitary engineer used a particular
wastewater treatment process he had experienced, continued to
use that same process over and over for every application.
Rule of thumb designs, personal experiences and time proven
wastewater treatment processes were the order of the day.
Some educators and a few pioneers in the industry tried to
advance the technology of wastewater treatment. Unfortunately,
since a waste treatment project carried the individual sanitary
engineer's P. E. stamp and the engineer's reputation was on the
line, very few new technologies were tried.
The Federal Water Pollution Control Administration (FWPCA)
program of complete funding for research of new wastewater
treatment technologies and upwards of 90% for demonstration
projects of these new technologies was the breakthrough the
RBC Process required in the United States. This enabled basic
RBC work and several full-scale demonstration projects. Some
of these projects proved an immediate success, demonstrating
the RBC values of high performance, simple operation, low
maintenance and low power consumption. In the mid 70's the
Federal General Accounting Office (GAO) instructed all applicants
for federal construction grant funds to submit cost evaluations
for all applicable wastewater treatment processes including the
RBC for their projects. This made consulting engineers take
a realistic appraisal of the RBC Process.
The many advantages of the RBC Process found its way into
numerous applications. These applications can be categorized
as follows:
350
-------
A. Small, medium and large municipal sewage treatment
plants.
B. Commercial applications such as subdivisions,
resorts, shopping centers, trailer courts, and
campgrounds.
C. Complete treatment of industrial process wastewater.
D. Pre-treatment of industrial process wastewater.
This latter application for pre-treatment of industrial
process wastewaters will become more prevalent as the State and
municipal regulatory agencies enforce the reporting waste
discharges to municipal sewage systems. The reporting of waste
material with pollution concentrations larger than domestic
wastewater requires additional municipal treatment capacity. The
municipality will charge an extra sewage discharge fee to these
industrial dischargers. Reduction of the industrial waters
pollution concentration to the levels of domestic wastewater,
and the resulting savings of these extra municipality discharge
fees, can in most instances, amortize the cost of wastewater
treatment equipment very rapidly.
The Federal Construction Grant Program and associated cost
effective evaluations gave the RBC Process many opportunities for
different effluent discharge standards in the municipal market.
The RBC Process is used for carbonaceous BOD reduction,
nitrification, following overloaded existing processes, parallel-
ing overloaded existing processes, and more recently, the RBC
Process was used in a de-nitrification application. Unfortunately,
this very large municipal market with federal funds and federal
regulations with so-called "non-restrictive specifications,"
enable some disc manufacturers to take short cuts in design and
applications. Insufficient in-house testing and the
Inabilities to use the required process disc design parameters
resulted in many problems. Horror stories of complete failures
of installations are commonplace. One of the largest installation
failures included 96 RBC shafts. There are several RBC
installations that have been and are still being completely
replaced for the third time. New installation failures are
being reported daily. The disc competitive situation in the
municipal marketplace will still result in future horror stories.
The original purpose of the European RBC objectives are being
ignored. The RBC originally developed as an inexpensive,
dependable, operator insensitive waste treatment process, are
causing consulting engineers to take another look before
selecting the RBC Process.
The Bio-Shaft Co. which I founded, has rigidly kept the
European RBC objectives in mind. Excellent process performance
without structural disc or shaft problems have been our trademark.
351
-------
However, you can't put bread on the table in the federally funded
municipal market with this philosophy. The competitive
situation does not give Tait/Bio-Shafts credit for stating "We
are the only disc manufacturers not to have a shaft or disc
failure."
Looking for a market area where Tait/Bio-Shafts can market
our dependable RBC product, we went back to the basics. The RBC
Process was initially developed for the small size installations.
Many successful small size installations are in operation.
Pretorius (3) said it quite well:
"Treatment works for small communities should be
relatively cheap, reliable, and easy to operate
and be maintained by unskilled labor, but should
produce the same high standard of effluent as that
from a larger well-controlled plant. The rotating
disc unit (RDU) seems to fulfill these prerequisites."
The RBC equipment is only a portion of the complete wastewater
treatmentsystem. Primary treatment, secondary clarifier,
disinfection and sometimes, tertiary treatment will complete a
system. While we can have a simple dependable RBC biological
reactor, all would be for naught if the rest of the wastewater
treatment system equipment does not support the simple and
dependable philosophy. We looked for a market that needed a
wastewater treatment system where dependability, simplicity and
high performance are the first and foremost objectives. Where
is there a market that can support the development and expense
of a disc manufacturer?
By coincidence, a very large market in need of these features
exists. The pre-engineered complete wastewater treatment package
system, more commonly referred to as the commercial market
requires the RBC Process features. Dependability, simplicity and
high performance. The required higher effluent quality standards
and the enforced reporting of all point discharge sources make
the RBC Process features desirable. In the past, the extended
air and contact stabilization processes were supplied for these
commercial applications. With proper, consistent and knowledgable
operations, the past effluent water quality operators for small
installations, enforcement of reporting standards and higher water
quality would appear to see a decreasing use of these processes.
The need will still exist for the pre-engineered complete
wastewater treatment package system because of many small low
applications.
Tait/Bio-Shafts, Inc. and its sister division, Purestream
Industries, Inc., combined its expertise to offer the RBC Process
in a complete pre-engineered wastewater treatment system. Figure
No. 1 shows the flow scheme which is used for a complete wastewater
treatment application. As can be seen, the flow scheme is surge
or flow equalization, primary clarification, RBC reactor, final
settling and disinfection. Sludge from final settling and
352
-------
primary settling is directed to the aerobic digester. Figure
No. 2 is an elevation view with cutaway of the system. Please
note the surge and aerobic digestor is aerated. Both primary and
secondary sludge is air lifted to the aerobic digestor.
Figure 2A is an artist conception of the assembled system.
Each equipment component of the system is a modular design,
completely pre-engineered. All of the components selected have
the same dependable history as the RBC. These components are
not only dependable, but are operator insensitive, low installed
and operating costs, with the ability to consistently produce
high effluent water quality. Figure No. 3 shows the selection
tables for the 6 ft. 8 inch diameter RBC system. Figure No. 4
shows selection tables for the 11 ft. diameter RBC system.
Please note the design flow selections encompass the range of
7,000 to 250,000 GPD. Other selection tables for various
system BOD removals are included in Purestream Industries
brochure TBRI7500. These brochures are available in the
Exhibitor Hall. We also have available for your review a
scaled-down model of this complete pre-packaged RBC system.
The RBC reactor has been conservative sized for the
commercial market. In the past, commercial applications equip-
ment has been sized with a minimal amount of process information.
Therefore, anticipating the same lack of complete information,
the RBC has been sized to insure obtaining the required high
effluent water qualities. Tait/Bio-Shaft used the staged
wastewater treatment process design for BOD removal. Each stage
is sized to remove 50% of the BOD level that the individual
stage sees. A hydraulic loading of 8 GPD per sq. ft. per
stage can easily accomplish this performance. The use of the
staging concept will give the consulting engineer another
valuable tool in which to select the required number of stages
for a specific design output.
The use of modular system components can enable the engineer
or regulatory agencies to substitute or interchange specific system
components design parameters to meet the individual installation
requirements. Figure No. 5 shows a basic modular component for
the primary modular. This includes the surge compartment,
aerobic digestor, and primary clarification. By enlarging or
decreasing the length dimension, incremental capacities of each
individual component can be obtained.
Figure No. 6 is the RBC reactor tank in a stage configuration.
Removal of internal baffles and manipulation of the length
dimension will again produce incremental system capacity
changes. Figure No. 7 is the final clarification tank with
chlorine contact compartment. Again, length changes can
•accommodate various component sizing.
The use of modular system component techniques enables the
engineer to use the specific modular component for other process
353
-------
applications. Inserting the RBC reactor component ahead of an
existing biological process will rough treat an overloaded
secondary treatment system. In these applications, the bio-
reactor modular alone can sufficiently reduce the BOD directed
to the existing biological process to the point where the overall
system will be more effective. Inserting the RBC reactor
component behind an existing secondary treatment system can be
used to polish or upgrade the overall system treatment
efficiency. Once again, the modular concept of a bio-reactor
makes an existing wastewater treatment plant upgrade or overload
a simple solution.
Using the biological reactor with a new clarifier can be used
very advantageously for an industrial pre-treatment installation.
The treated water will be discharged to a municipal sewer
system and as mentioned before, substantial savings to the
industrial user can be experienced.
Applications for the small community or complete pre-
engineered complete wastewater packaged system in effect will
give this market area all the advantages of a larger built-in-
place municipal system. It also has an inherent advantage that
all components will be supplied by one company. Purestream
Industries using the Tait/Bio-Shaft Rotating Biological Disc
is the only company to supply a pre-ehgineered, complete
wastewater packaged RBC system in this country.
354
-------
REFERENCES
(1) Weigand, Philipp: "Method of Biological Cleaning of
Sewage," German Reich Imperial Patent Office,
Patent Specification No. 135755 Class 85C, published
October 21, 1902.
(2) Maltby, Arthur T.: "Process and Apparatus for Treating
Sewage or Other Organic Matter," United States Patent
Office, Patent No. 1.811,181, dated June 23, 1931.
(3) Pretorius, W. A.: "The Rotating-Disc Unit: A Waste
Treatment System for Small Communities," Water
Pollution Control 1973, pages 721-724.
355
-------
w
Ul
inc.
PUDT PLAN RDR BIO-RBCTOfi
SYSTEM WITH GW1TY
CLABIRER i RBCTQR
GENERAL DETAILS
Pig. 1
BR-OOI-PP
P8R-041
-------
-------
358
-------
FIGURE 3
Selection tables for secondary treatment
95.6% Removal
FLOW
7.000
8,000
9,000
10,000
12,500
15,000
17,500
20,000
22,500
as.ooo
30,000
35,000
40,000
45.000
50.000
60,000
70,000
^80,000
_ MODEL
PBR6D5
PBR 606
PBR6O7
PBR6D8
PBR6D9
PBR6O11
PBR6D13
PBR 6015
PBR6D17
PBR6D19
PBR 6023
PBH 6D27
PBR6D31
PBR6O17
PBR6D19
PBH 6D23
PBR 8027
PBR 6D31
SEACTORS
1
1
1
t
1
1
1
1
1
1
1
t
1
2
2
2
2
2
VOL.
700
800
900
1.000
1,250
1.500
1.750
2.000
2,250
2.500
3,000
3,500
4,000
4,500
5,000
6.000
7,000
8.000
VOL.
670
670
750
835
1,160
1,250
1.500
1.670
t,8T5
2,170
2.500
3,000
3,340
3,750
4,170
5.000
5,830
6,670
»OL,(F>>
140
160
180
200
250
300
350
400
450
500
600
700
800
900
1,000
1,200
1,400
1.600
6 Ft,-8 In. Dia. Reactor I
SECONDARY
CLAD, VOL.
670
670
750
835
1,160
1,250
1,500
1,670
1,875
2,1?0
2.500
3,000
3,340
3,750
4,170
5,000
5,830
6,670
CCT
VOL.
146
177
188
208
260
312
365
416
470
521
625
729
835
940
1,050
1,250
1,480
1,670
TOTAL
HP
3V«
3«
3»
3 "A
3V.
4
4
-------
FIGURE 4
Selection tables for secondary treatment
95.6% Removal 11 Ft. Dia. Reactor
FLOW
25.000
30.000
35.000
40.000
45.000
50,000
60,000
70,000
80.000
90.000
100,000
125,000
150,000
175,000
200,000
225.000
V 250.000
MODEL
PBR11D7
PBR 11 Dfl
PBR 11010
PBR 11011
PBR 11012
PBR11D14
PBR 11017
PBR 11020
PBR 11022
PBR 11 025
PBR11D28
PBR11D36
PBR 11 022
PBR 11 024
PBR 11028
PBR 11032
PBR 11 036
NO. OF
REACTORS
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
SURGE
VOL
2.5OO
3.000
3.500
4.000
4,500
5,000
6,000
7,000
8,000
9,000
10,000
12.500
15,000
17,500
20,000
22,500
25,000
PRIMARY
VOL.
2.170
2,500
3.000
3,340
3.750
4,170
5.000
5,830
6.670
7,500
8,340
10,845
12,500
15.000
16,700
38,333
41,666
DIGESTER
VOUF1)
600
600
700
800
900
1,000
1.200
1,400
1,600
1,800
2,000
2,500
3,000
3.500
4.000
4.500
5.000
SECONDARY
CLAR.VOL.
2.170
2,500
3,000
3,340
3,750
4.170
5,000
5,830
6.670
7.500
8.340
10,845
12,500
15.000
16.700
38.333
41,666
CCT
VOL.
521
625
729
835
940
1,050
1,250
1,460
1.670
1,880
2,090
2.605
3,125
3.650
4,170
4,690
5,210
TOTAL
HP
3*4
3V.
5U
6
6»
6Vi
6Vi
6»
9
9
9
9
10V4
13
13
18
18 J
91 % Removal 1 1 Ft. Dla. Reactor
FLOW
25.000
30,000
35,000
40,000
45,000
50,000
60.000
70.000
80.000
90,000
100,000
125.000
150.000
175.000
200,000
225.000
V 250.000
MODEL
PBR11C7
P8R11C8
PBR11C10
PBR11C11
PBR11C12
PBR11C14
PBR11C17
PBR11C20
PBR11C22
PBR11C25
PBR11C28
PBR11C36
PBR11C43
PBR11C50
PBR11C28
PBR 11C32
PBR11C36
NO. OF
REACTORS
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
SURGE
VOL.
2,500 .
3,000
3,500
4,000
4,500
5,000
6.000
7,000
8,000
9,000
10.000
12,500
15.000
17,500
20,000
22,500
25.000
PRIMARY
VOL.
2,170
2,500
3,000
3,340
3.750
4.170
5.000
5,830
6,670
7.500
8.340
10.845
12.500
15,000
16,700
38,333
41,666
DIGESTER
VOL.(F1
500
600
700
800
900
1,000
1.200
1,400
1,600
1,800
2,000
2,500
3,000
3,500
4,000
4,500
5.000
SECONDARY
CLAP- VOL.
2,170
2.500
3,000
3,340
3,750
4,170
5,000
5,830
6,670
7,500
8,340
10,845
12,500
15,000
16.700
38,333
41,666
CCT
VOL.
521
625
729
835
940
1,050
1,250
1.460
1.670
1.880
2,090
2,605
3,125
3,850
4,170
4.690
5,210
TOTAL
HP
3V.
3*4
5U
5V.
5V<
8
BVi
6»
9
9
9
9
9
12
13
18
18 J
-------
U)
en
—FOUNOJTION P*0 THIONEtt
KTEflhVCO ffr ENQiNCCfl
k |l^%v--:^-^WT?
\ *'-^ k^ ^-2"Oft>lMS~ . . . . -J ^
\ h -'.-—
Ef CTION 'C'-'C*
D tr—^
3ESCN FLOW OK
industries inc.
PRIMAFN MODULE FOR
BIO-REACTOR SYSTEM
GENERAL DETAILS
BR-45<50-T-I
Pig. 5
-------
"W
!S
3
i I
to
•H
Jx,
362
-------
PLANT DMEJ6IOHS
A • C E
mduBtnes inc.
THIRD MODULE FOR
BO-REACTOR SYSTEM
GENERAL DETAILS
°BR-4§?50-T-3
Pig. 7
-------
-------
PART IV: BIOKINETIC STUDIES
DYNAMICS OF HICROBIAL FILM PROCESSES
by
W. G. Characklis
Professor of Civil Engineering
M. G. Trulear
Department of Civil Engineering
School of Engineering
Montana State University
BOzeman, Montana, U.S.A...
Abstract
Microbial film formation at a surface is the net result of several
physical, chemical and microbial processes including the following:
1. Transport of dissolved and particulate matter from the
bulk fluid to the surface.
2. Firm microbial cell attachment to the surface.
3. Microbial transformations (growth, reproduction, etc.) at
the surface resulting in production of organic matter.
4. Partial detachment of the deposit due primarily to
fluid shear stress.
The properties and structure of the resulting biofilm accumulation
reflect the relative rates of the above processes and influence the
observed rate coefficients (e.g., coefficients for attachment rate and
microbial substrate removal rate) and transport coefficients (e.g.,
diffusion coefficient, thermal conductivity and viscosity of biofilms).
The transport coefficients are useful in analyzing the effects of biofilm
development on fluid frictional resistance and heat transfer resistance.
This paper presents a framework for analyzing the interrelated proc
esses contributing to biofouling. Available rate and composition data a
presented so that the relative process rates can be compared.
are
365
-------
INTRODUCTION
Bacteria stick firmly, and often with specificity, to almost any
surface submerged in an aqueous environment. The bacteria attach by
means of a matrix of polymers, primarily polysaccharides, that extend
from the cell surface and form a mass of tangled fibers, termed a
glycocalyx. The adhesion mediated by the glycocalyx determines parti-
cular locations of bacteria in many aquatic environments. The cells grow
and reporduce at the surface increasing the mass of cells and their asso-
ciated material, i.e., the biofilm.
Biofilm processes may be beneficial as exemplified by fixed-film
wastewater treatment processes (e.g., trickling filters and rotating bio-
logical contactors). In addition, biofilms frequently play a major role
in stream purification processes. In fact, microbial activity in natural
waters is concentrated at the interfaces. ' ' However, biofilms
can be quite troublesome in certain engineering systems. For example,
biofilms in water conduits can cause energy losses resulting from in-
creased fluid frictional resistance and increased heat transfer resis-
tance. Table 1 lists the effects and relevance of biofilm processes
to various sectors of our society.
Prior research from our group has been directed to problems of
fouling (primarily sponsored by NSF and Electric Power Research Insti-
tute) . Fouling refers to the formation of inorganic and/or organic de-
posits on surfaces. In cooling systems, these deposits form on conden-
ser tube walls increasing fluid frictional resistance and heat trans-
fer resistance. Four types of fouling, alone or in combination, may
occur:
1. crystalline fouling caused by precipitation of CaCO ,
CaSOy or silicates
366
-------
2. corrosion fouling resulting from formation of insula-
ting layers of metal oxides on the tubes
3. fouling due to adherence of particulate matter on
tube surfaces
4. biofouling resulting from attachment and growth of
microorganisms.
Our work has been concerned with biofouling.
The most common method for controlling fouling biofilm development
and maintaining heat exchange performance is periodic chlorination.
Chlorine, added to cooling water, serves either to kill the microor-
ganisms or to hydrolyze the extracellular polymers which hold the bio-
film together. The chlorine dosage and application schedule is typi-
cally determined by (1) observation of condenser performance as indica-
ted by plant steam back-pressure, or (2) operator experience.
Recently, concern over residual toxicity from hypochlorous acid
or its reaction products has resulted in federal regulations which limit
the allowable concentrations of free available chlorine in cooling water
discharges. At the present time, there is no sound basis for assessing
the impact of the regulations. Our previous investigations stemmed
from the apparent need for a better basic understanding of biofilm de-
velopment and biofilm destruction so that the impact of these new regu-
lations on power generation could be evaluated.
The nature of the problem and the widespread concern regarding
biofilm processes (evidenced in Table 1} led us to more fundamental
studies in the dynamics of biofilm processes and its relevance in areas
besides biofouling. Results of these and other studies are described
in succeeding sections of this document. We propose to extend the work
by further study of the fundamental processes which contribute to bio-
film formation and destruction including the following:
367
-------
Table 1, Effect and relevance of bi.ofi.lms on various rate processes.
Effects Specific Process
Heat transfer reduction
Increase in fluid fric-
tional resistance
Mass transfer and
chemical transformations
to
(71
03
Biofilm formation on condenser tubes
and cooling tower fill material.
Energy losses.
Biofilm formation in water and sewage
conduits as well as condenser and heat
exchange tubes. Causes increased power
consumption for pumped systems or
reduced capacity in gravity systems.
Biofilm formation on ship hulls.
increase in fuel consumption.
Causes
Accelerated corrosion due to processes
in the lower layers of the biofilm.
Material deterioration in metal conden-
ser tubes, sewage conduits, and cooling
tower fill material.
Biofilm formation on remote sensors, sub-
marine periscopes, sight glasses, etc.
Reduces effectiveness.
Concerns
Power industry
Chemical process industry
U.S. Navy
Solar energy systems
Municipal water supply and sewage
collection
Power industry
Chemical process industry
Solar energy systems
U.S. Navy
Shipping industry
Power industry
U.S. Navy
Municipal water supply and sewage
collection
Chemical process industry
U.S. Navy
Water quality data collection
Detachment of microorganisms from bio-
films in cooling towers. ReleaseSL patho-
genic organisms (e.g., Legipnella) in
aerosols.
Public health
Biofilm formation and detachment in
drinking water distribution systems.
Changes water quality in distribution
system.
Biofilm formation on teeth.
plaque and cavities.
Dental
Attachment of bacteria to animal cells.
Diseases of lungs, intestines, and
urinary tract.
Water supply industry
Public health
Dental health
Human health
-------
Table 1. (continued)
Effects
Specific Process
Extraction of and oxidation of organic
and inorganic compounds from water and
wastewater. Reduction in "pollutant"
load. For example, rotating biological
contacters, biologically-aided carbon
adsorption, and benthai stream activity.
Biofilnt formation in industrial produc-
tion processes. Reduced product quality.
Immobilized microorganisms or community
of microorganisms for conducting specific
chemical transformations.
Concerns
Wastewater treatment
Water treatment
Stream analysis
Pulp and paper industry
Chemical process industry
-------
1. transport rate of microbial cells from the bulk liquid to
the surface.
2. interfacial phenomena resulting in adsorption/attachment
of microbial cells to the surface.
3. reactions within the biofilm which increase the mass of
the accumulation (e.g., microbial growth, exopolysaccha-
ride production).
4. strength of the biofilm deposit as related to partial bio-
film removal in a fluid shear field.
PROCESS ANALYSIS
Process analysis refers to the application of scientific methods to
the recognition and definition of problems and the development of procedures
for their solution. This generally requires 1) mathematical specification
of the problem for the given physical situation, 2) development of a mathe-.
matical model, and 3) synthesis and systematic presentation of results to
ensure full understanding. The process denotes an actual series of opera-
tions or treatment of materials as contrasted with the model, which is a
mathematical description of the process (1).
Fundamentals (after Churchill (2))
The fundamental relationships which underlie process analysis are the
equations for conservation of mass, momentum and energy which are a result
of the laws of thermodynamics and Newton's laws of motion. The conserva-
tion equations are generally expressed in terms of intensive factors
(independent of system mass) such as composition, velocity and temperature.
The conservation equations also introduce physical properties such as
thermodynamic properties, stoichiometric coefficients, transport rates
and chemical reaction rates.
The thermodynamic properties, such as density, heat capacity and
chemical equilibrium constants can be estimated with reasonable confidence
370
-------
from mechanistic models and can be measured with reasonable accuracy. The
transport coefficients, such as viscosity, thermal conductivity and diffu-
sivity, and the chemical reaction rate coefficients can rarely be predicted
and are difficult to measure. Even the definition of these latter quan-
tities is somewhat arbitrary.
The physical, chemical and biological transformations of interest in
biofouling are completed in a certain period of time. With respect to
biofouling, a specified change may signal the shutdown of operations and
the beginning of cleaning operations. The time required for this specified
change is inversely proportional to the rate at which the process occurs,
Thus, rate is the most important quantity in process analysis.
The equations for conservation of mass, energy, momentum and
chemical species equate the rate of accumulation to the net rate of input
by flow and the net rate of input by various rate processes such as chemi-
cal reaction, diffusion, radiation, convection and viscous dissipation.
The process rates are fundamental quantities in that they can be generalized
and correlated simply with factors such as temperature, pressure, com-
position, velocity and diameter which describe the environment. The
rate of accumulation and the net rateof input by flow are herein called
rates of change. These rates of change are observed or measured quantities
which may be the result of several process rates. They cannot be correlated
simply or generalized. It is essential that rates of change not be con-
fused with process rates (Fig. 1}.
The procedures involved in process analysis are: 1) the mathematical
description of the rate of change, i.e., the rate of accumulation in batch
operations and the net rate of input by flow in continuous operations,
2) the experimental measurement of the rate of change and the determination
371
-------
RATE PROCESS ANALYSIS
Volumetric
Row rate Q fc
Inlet
Concentration CQ
C
1 V,
Q
3
C j
Reaction
Rate
C
k,C
MATERIAL BALANCE ON COMPONENT C
d(vc)
dt
Rate of
Accumulation
\
\
RATES OF
Q (C0-C)
Net Rate of
Input by Flow
CHANGE
k.C
Output by
Reaction
PROCESS RATE
Figure 1
PROCESS RATES can be correlated with temperature,
composition and geometry
372
-------
of process rate from the experimental measurements, 3) the correlation and
generalization of the process data, and 4) the use of the rate data in process
calculations, including conditions outside the range of the experimental
conditions.
Introduction to Microbial Fouling Processes
Microbial fouling is the net result of several physical, chemical
and biological processes including the following:
o organic adsorption at the wetted surface
o transport of the microorganisms to the wetted surface
o microorganism attachment to the surface
o metabolism and growth of attached microorganisms
o detachment or reentrainment of biofilm by fluid shear stress
Organic adsorption. Microorganisms select their habitats on the
basis of many factors, including the nature of the wetted surface (material
of construction and surface roughness).
Figure 2 illustrates an initially "clean" surface exposed to turbulent
flow of a fluid containing dispersed microorganisms, nutrients, 'and organic
macromolecules. Adsorption of an organic monolayer occurs within minutes
of exposure as shown in Fig. 3. Investigations have shown that materials
with diverse surface properties (e.g., wettability, surface tension, elec-
trophoretic mobility) are rapidly conditioned by adsorbing organics when
exposed to natural waters with low organic concentrations.
Transport of microbial particles to the surface. Figure 4 indicates
the physical transport of bacterial particles from the bulk fluid to the
surface covered by an organic film. Within a turbulent flow regime, par-
ticles suspended within the fluid are transported to the solid surface
by two mechanisms: molecular diffusion and turbulent eddy transport.
Theory indicates that the flux of particles to the surface increases with
373
-------
FLOW
o
o
O
c
T-2-^-
5 y
i.
o °^ '
CELL ^o
Unfouled Surface
Fig. 2. Initially clean surface exposed to a turbulent flow of fluid
containing microorganisms and associated material.
FLOW
Organic Adsorption
Fig. 3., ' Adsorption of organic material from the bulk fluid.
Flux and Attachment
Fig. 4. Flux and attachment of microbial cells to
the surface from the bulk fluid.
374
-------
increasing fluid velocity and particle concentration. However, particle
flux is also strongly dependent on the physical properties of the particles
(e.g., size, shape, density) and may be influenced by other hydrodynamic
processes near the attachment surface.
Microorganism attachment to the surface. Research suggests the
existence of a two-stage attachment process: reversible adhesion followed
by an irreversible attachment. Reversible attachment refers to an initially
weak adhesion of a bacterium to a surface. Organisms still exhibit Brownian
motion and are readily removed by mild rinsing. Conversely, irreversible
attachment is usually aided by the production of extracellular polymers
and attached cells can not be removed easily.
Many microbial cell attachment studies have been conducted at rela-
tively low fluid shear rates or under quiescent conditions. Rates of
accumulation determined from these studies are potentially mass transfer-
limited and may not be applicable to condenser biofouling where fluid
shear rates are quite high (equivalent to average fluid velocities of
6 ft/sec in a 3/4' I.D. tube).
Metabolism and growth of attached microorganisms. Attached and dispersed
microorganisms assimilate nutrients, synthesize new biomass and produce
extracellular polymers. Biomass production on the surface is the net
result of cell division and extracellular product formation, as shown in
Fig. 5.
Biofilm growth has been described by a wide variety of rate expressions
whose rate constants are functions of pH, temperature, limiting nutrient
concentration, nutrient type, terminal electron acceptor, and organism
concentration.
Postulated rate expressions for nutrient depletion by a fixed biofilm
375
-------
FLOW
o o o
-o-
, T ^ r ^ u
^g^*^3^^M^jg
Flux and Growth
Fig. 5. Continued flux of microbial cells to the surface
with simultaneous growth processes occurring.
FLOW
TJ
fcP
TJ
o <# *
•*
O fff
,_ O ^/rfT'yf o
•P ~ ^ *^~r*eS. vj
Ofe** o. oi< »-
.^ «fe- o
O O ^Q^
^^CfeO^a^ ^vr
rvp/v-r -fc? v^
^U^-/i-
-------
are numerous, but all agree that nutrient depletion rates are first order
in biofilm mass for thin films and that diffusion rates in the biofilm can
often control the overall removal rate of nutrients.
Reentrainment of biofilm by fluid shear. At any point in the development
of a biofilm, portions of biofilm peel away from the inert surface and are
reentrained in the fluid flow (Fig. 6). Reentrainment is a continuous
process of biofilm removal and is highly dependent on hydrodynamic conditions.
Sloughing, on the other hand, appears to be a random, massive removal of
biofilm attributed to oxygen/nutrient depletion deep within biofilms.
Sloughing is more frequently witnessed with thicker, dense films especially
in laminar flow. More work is needed to quantify either effect.
In summary biofouling is the net result of all these processes occur-
ring simultaneously (Pig. 7). However, at specific times during biofilm
development, certain processes may contribute more than others.
PROPERTIES AND COMPOSITION OF BIOFILMS
Microorganisms, primarily bacteria, adhere to surfaces ranging from
the human tooth and intestine to the metal surface of condenser tubes ex-
posed to turbulent flow of water.' The microorganisms "stick" by means of
extracellular polymer fibers, fabricated and oriented by the cell, that
extend from the cell surface to form a tangled matrix termed a "glycocalyx"
by Costerton _et al_. , (3) . The fibers may conserve and concentrate extra-
cellular enzymes necessary for preparing substrate molecules for ingestion,
especially high molecular weight or particulate substrate which is available
in natural waters.
The biofilm surface is highly adsorptive, partially due to its poly-
electrolyte nature, and can collect significant quantities of silt, clay
and other detritus in natural waters.
377
-------
Physical, chemical and biological properties of biofilms are dependent
on the environment to which the attachment surface is exposed. The physical
and chemical microenvironment combine to select the prevalent microorganisms
which, in turn, modify the environment of the surface. As colonization
proceeds and a biofilm develops, the microenvironment is altered and bulk
biofilm properties change. Biofilm properties and changes that occur during
biofilm development are critical to determining the effect of biofilms
on fluid and heat transport in turbulent flow systems but have been largely
ignored in past studies.
Physical Properties
The most fundamental biofilm properties are volume (thickness) and
mass. In turbulent flow systems, wet biofilm thickness (Th) seldom exceeds
1000 JJm (4). The biofilm mass can be determined from the wet biofilm thick-
ness if the biofilm dry mass density (P™,) is known, p , reflects the
attached dry mass per unit wet biofilm volume and measured values in turbu-
lent flow systems range from 10 - 50 mg/cm . p increases with increasing
turbulence (4) and increasing substrate loading (5,4) as indicated in Figs.
8 and 9. The increase in p with increasing turbulence may be caused by
one of the following phenomena;
1. selective attachment of only certain microbial species
from the available population
2. microorganism response to environmental stress
3. fluid pressure forces "squeeze" loosely bound water from
the biofilm.
The relatively low biofilm mass densities compare well with observed
water content of biofilm (6,7,8).
The transport properties of biofilm are of critical importance in
quantifying effects of biofilms on mass, heat and momentum transfer.
378
-------
40r
35 -
30
25
0»
J=*>
><
"U5
§ 15
E
= .0
o
CD
Fig. 8. Effect of fluid velocity on
biofilm density for glucose
loading rate < 25 mg m~2 min~l
in a 1.27 cm I.D. tube [4].
a 4 e 3 10
Average Fluid Velocity (ft/sec)
Fig. 9. Effect of glucose
loading rate on bio-
film density.
so
40
O
O>
"
tn
|20
O
a 10
379
o Trulear [48'
• Zelver [4]
50
100
I5O
20C
Glucose Loading Rate (mg/rrr-min)
-------
Table 2. Experimental diffusion coefficient measurements from the literature
[9].
Reactant
Oxygen
Oxygen
Glucose
Glucose
Oxygen
Ammonia
Nitrate
Oxygen*
Glucose*
* Tests
Diffusivity
10~5cm2s-l
1.5
0.21
0.048
0.06-0.6
2.2
1.3
1.4
0.4-2.0
0.06-0.21
conducted under
Dfloc/DH20
xlOO%
70
8
8
10-100
90
80
90
20-100
10-30
a variety of
Biomass
Type
Bacterial
Slime
Fungal
Slime
Zoogl ea
ramigera
Zoogl ea
ramigera
Mixed
Culture
Nitrifier
Culture
Mixed
Culture
experimental
Growth
System
Rotating
Tube
Fluidized
Reactor
Fluidized
Reactor
Fluidized
Reactor
Fluidized
Reactor
Fluidized
Reactor
Procedure
Reaction
Products
Analysis
Nonlinear
Curve Fit
Nonlinear
Curve Fit
Two
Chamber
Two
Chamber
Two
Chamber
Refer-
ence
[12]
[13]
[14]
[15]
[16]
[17]
conditions.
Table 3. Viscoelastic properties of biofilm developed at 40°C at a fluid
shear stress of 3.3Nm~2. Glucose was growth-limiting and was
applied at 6.2 mg m~2 min"^- [ 7 ] •
Elastic (storage) Modulus
Viscous (loss) Modulus
59.5 N m-2
118 N m~2
380
-------
Table 4. Thermal conductivity of biofilm and other selected materials -relevant
to biofouling of heat exchangers.
Material
Biofilm
Water
Carbon Steel
Steel
Stainless Steel (type 316)
Aluminum 5052
Cupronickel 10% 706
Copper
Titanium
Thermal
Conductivity
(W nT1^"1)
0.68 ±0.27
0.71 ±0.39
0.57 ±0.10
0.61
0.62
51.92
46.86
16.30
138.46
205.85
44.71
384.
16.44
Temperature
(°C)
28.3±0.3
26.7±0.3
28.3±0.3
26.7
28.3
* 0-100
18
0-100
20
100
0-100
18
0-100
Reference
[6]
[18]
[19]
[20]
[19]
[19]
[20]
[19]
[20]
[19]
(commercial pure)
Glass
0.6-0.9
[18]
381
-------
Diffusion coefficients for various compounds through microbial aggregates
have been reported in the literature (9), mostly for floe particles (Table 2).
Matson and Characklis (9) report variation in the diffusion coefficient for glu-
cose and oxygen with growth rate and carbon-to-nitrogen ratio. In biofilms,
the diffusion coefficient is most probably related to biofilm density. In
situ rheological measurements indicate that the biofilm is viscoelastic
with a relatively high viscous modulus as indicated in Table 3 (7). Re-
ported biofilm thermal conductivities are presented in Table 4. As expected
from reported water content, biofilm thermal conductivity is not sig-
nificantly different from water.
Chemical Properties
Inorganic composition of biofilms undoubtedly varies with the chemical
composition of the bulk water and probably affects the physical and biolo-
gical structure of the film. Calcium, magnesium and iron affect intermole-
cular bonding of biofilm polymers which are partially responsible for the
structural integrity of the deposit. For example, EDTA is effective,
although costly, in detaching biofilm (7). In heat exchangers, corrosion
products and inert suspended solids can adsorb to the biofilm matrix and
influence its chemical composition. Table 5 reports the range of inorganic
composition observed in selected biofilms.
The organic composition of the biofilm is strongly related to the
energy and carbon sources available for metabolism. Classical papers (10,11)
have demonstrated the effect of environment and microbial growth rate on
the composition of the cells and their extracellular products. For example,
nitrogen limitation can result in production of copious quantities of micro-
bial extracellular polysaccharides. Table 6 presents data on the composition
of biofilms developed in the field and in the laboratory. In terms of ma-
382
-------
Table 5. Chemical properties of biofilms obtained from fouled surfaces
experiencing excessive fractional losses (after Characklis [8])
REFERENCE
[21] [22] [22] [23]
Water
Volatile Fraction
Fixed Fraction
Si (as percent fixed fraction)
Fe
Al
Ca
Mg
Mn
87 85.6 90
2.5 2.7 1.9
10.5 11.7 8.1
7.0 11.8
18.5 7.9
7.5
1.0 5.6
2,5
59.5 56.3
95
2.4
2.6
12.5
1.4
3.9
3.2
4.9
[7]
96
3.2
0.8
Table 6, Chemical composition of biofilms obtained in the field and laboratory
emphasizing the primary constituents (C,N,P).
Source
Biofilm —
power plant
condenser
Biofilm -
laboratory
reactor
Biofilm -
laboratory
reactor
E, call
% DRY WEIGHT
Fixed
C N P Solids C/N C/P
6.4-13.8 0.5-3.0 — — 2-27
42.8 10.0 — — 4.3
19.0 9.2 1.8 20 2.1 10.5
50.0 14.0 3.0 — 3.6 16.7
Refer-
ence
[24]
[25]
[7]
[26]
383
-------
cromolecular composition, Bryers (27) has measured protein-to-polysaccharide
mass ratios ranging from 0 to 10 (polysaccharide concentration in terms of
glucose and protein concentration based on casein) with increasing biofilm
accumulation. Other chemical analyses of biofilm have been reported by
Bryers and Characklis (28).
Biological Properties
The organisms which colonize the attachment surface will strongly
influence biofilm development rate and biofilm chemical and physical pro-
perties. However, organism-organism and organism-environment interactions
undoubtedly shift population distributions during biofilm accumulation.
Several investigators have observed succession during biofouling (29,30).
The first visible sign of microbial activity on a surface is usually
small "colonies" of cells distributed randomly on the surface. As biofilm
development continues, the colonies grow together forming a relatively uni-
form biofilm. The viable cell numbers are relatively low in relation to
4 8-3
the biofilm volume (10 - 10 cm biofilm) occupying only from 1-10% of
the biofilm (7). Costerton (31) and Jones et_ al_., (32), present photomicro-
graphs which corroborate these data in natural and laboratory systems.
In many cases, filamentous forms emerge as the biofilm develops further.
Hyphomicrobium, Sphaerotilus and Beggiatoa are frequently identified. The
filamentous forms may gain an ecological advantage as the biofilm develops
since their cells can extend into the flow to obtain needed nutrients or
oxygen which may be depleted in the deeper portions.
384
-------
RATE PROCESSES CONTRIBUTING TO BIOFILM DEVELOPMENT
Transport from the Bulk Fluid to the Wall
The transport and deposition of entrained particles from a turbulent
stream onto the walls of a pipe have been investigated by Friedlander and
Johnstons (33) for large particles (>lym in diameter) and by Lin (34) for
small particles (soluble components). Seal (35) has described a unified
approach to cover a size range including these extremes which agrees reasonably
well with published data. In microbial fouling, the particles of interest
range in size from approximately 0.5-20 ym and are "in between" the size
ranges that have occupied many researchers' interests. Therefore, Beal's
model will be used for this discussion.
Seal begins with an equation describing particle flux, N, as follows:
/"If"™*
N = (D + D ) — ...
e dy (1)
-2 -1
where N = particle flux (L t )
2 —1
D = diffusion coefficient {L t )
2 -1
D = eddy diffusion coefficient (L t )
e
C = particle number concentration (L )
y = distance from the wall (L)
The diffusion coefficients for various particles of interest are pre-
sented in Table 7. Eddy diffusion coefficients are dependent on turbulent
intensity and therefore vary across the pipe section for a given particle.
Eddy diffusion coefficients are characteristically orders of magnitude
greater than diffusion coefficients (i.e., D » D). Beal integrated Eq.(1)
to obtain N , the flux at the wall:
w
N = KG (2)
w D avg
38.5
-------
Table 7 . Calculated diffusion coefficients for particles of interest.
Molecular
Weight
(g mol" )
Tobacco mosaic 31,400,000
virus
Myoglobin 17,000
Glucose 180
Particle3
Particle
Diameter
(urn)
-
-
_
10
5
3
3
1
0.1
0.01
Diffusion
Coefficient x 1010
9 —1
(cm^ sec )
58.8 l
5320 *
97300 1
67300 2
4.3
8.6
14.2
18.5
42.8
430
4300
Temperature
(°c)
20
20
25
20
20
20
30
20
20
20
1 Calculated from Wilke-Chang equation.
2 Measured Value [8 ]
3 Calculated from Stokes-Einstein equation.
386
-------
where
Kv P
_-i
K = deposition coefficient (Lt )
v = radial velocity of the particle (Lt )
K = transfer coefficient (Lt )
p = "sticking" efficiency of the particles (dimensionless)
The radial velocity of the particle, v , can be considered the result
of the following: (1) motility generated by internal energy (e.g., chemo-
tactic response by a bacterial cell) , v , (2) sedimentation, v , (3) Brownian
M s
motion, v , and (4) fluid motion, v_. An analysis of the relative contribu-
B I
tion of the components indicates that v is dominant for particles up to
B
about 1 pm in turbulent flow. Above 1 pm, fluid motion (in the turbulent
flow regime) is most important.
Two simplified forms of Eq. (3) are useful. If K»pv , the only par-
ticle concentration gradient is very close to the wall and
N = pv C (K»pv ) (4)
w r avg r
If pv »K, transport through the bulk fluid in the pipe controls wall flux.
and
N = KO
w avg (pv »K) (5)
For microbial fouling, pv »K, based on calculations assuming the fol-
lowing (Table. 6):
particle density = 1.07 g/cm
particle diameter = 0.01 - 10 pm
avg. fluid flow
velocity = 50-350 cm/sec (in a 1.27 cm ID tube)
sticking efficiency =1.0
387
-------
s
SE
o
I «cr2
El
u.
Ill
o
o
u
o)
<
o:
|0"3
10-4
I
150 250 350
V, FLUID VELOCITY (CM S"1)
Fig. 10. The influence of fluid velocity (1.27 cml.D. tube) and particle diam-
eter (particle density = 1.07 gcm~3) on the transfer coefficient, K,
calculated from Beal [35].
388
-------
Figure 10 illustrates the effect of fluid velocity and particle diameter
on K. Bryers (27) results suggest that sticking efficiency is considerably
less than 1.0 (approximately 0.01) and calculations reported in Pig. 10
must be used cautiously.
Accumulation of organic material at the wall in the early stages of bio-
fouling (assuming growth is negligible) is a result of two fluxes: (1) micro-
bial cells and (2) organic macromolecules. So
CV total * (VM +
-------
Table 8. A comparison of particle flux and mass flux for a hypothetical water
source flowing at 150 cm -1 in a circular tube (1.27 I.D.). Macro-
molecular particle flux is much greater than microbial particle flux.
However, the reverse is true for mass flux.
Carbon Content (%)
Particle Mass Concentration in
Bulk Fluid (yg cm~3)
Molecular Weight (g mol~l)
Particle Number Concentration in
Bulk Fluid (cm-3)
Macromolecule
40
2
50,000
6 x 10!3
Microbial Cell
50
7.6 x
—
1 x 106
Transfer Coefficient (cm s"1) 6 x 10~4 1 x 10~4
(from Table 7)
Particle Flux (cm~2 s'1) 3.6 x 1010 100
Mass Flux (g cm"2 s'1) 5 x 10~23 1.5 x 10~9
1Assuming cell diameter = 3 ym and cell density = 1.07 g/cm3
390
-------
Molecular Fouling
Figure 2 illustrates an initially "clean" surface exposed to turbulent
flow of a fluid containing dispersed microorganisms, nutrients, and organic
macromolecules. Within a short period (minutes) of exposure, adsorption
of measurable quantities of the organic molecules occurs (Fig. 3). These
molecules are usually polysaccharides of glycoproteins. Loeb and Neihof
(37) and DePalma et_ al_. (38) have measured rates of adsorption in sea-
water, and Bryers (27) has observed adsorption in a laboratory system.
Rates and extent of adsorption of these investigations are presented in
Table 9, Maximum accumulation from molecular fouling is less than 0.1 urn.
The rate of molecular fouling is much greater than rates of microbial fouling
which is generally reported in terms of days or weeks. Consequently, mole-
cular fouling can be considered instantaneous. However, based on "thickness"
measurements, molecular fouling can have no significant effect on fluid
flow or heat transfer. Nevertheless, the surface properties resulting
from adsorption of an organic film may affect the sequence of microbial
events which follow.
Costerton et al. (3) have discussed the pronounced specificity of
some bacteria that attack only a particular animal host tissue and suggest
that specificity may be explained by the specificity of the host-tissue
glycocalyx. It remains to be seen whether a surface, wetted by the adsorp-
tion of organic molecules indigenous to that environment, will be initially
colonized by a specific microbial cell.
Brash and Samak (39) present experimental evidence that significant
turnover occurs in molecular fouling films (proteinaceous) on polyethy-
lene. This may be occurring in microbial systems even when relatively
thick films have developed.
391
-------
Table 9 . Maximum rate and extent of molecular fouling.
Maximum Rate
(nm/min)
0.15-0.45
0.004
0.044
0.01 s
0.22 s
Maximum
Accumulation
(nm)
30-80
7.1
77.3
13.5 5
22.5 s
Maximum
Accumulation
(Ug COD /cm2)
1.5
2.5
Surface
Pt1
Ge2
Ti2
glass3
glass **
Reference
[37]
[38]
[27]
ilmmersed in quiescent Chesapeake Bay water (3-4° C) containing 2.3 mg carbon/1,
salinity between 9-16 °/oo and pH between 7.9-8.2
2Gulf of Mexico water (22° C) flowing past the surface as a fluid sheat stress
of 7.1 N/m2. Salinity was 34 °/oo. Carbon concentration not reported.
9Medium consisted of a sterile 1:1 w/w of trypticase soy broth - glucose mix-
ture (34°C; pH 8). The glass surfaces were immersed in tubes placed in a
mechanical shaker. Carbon concentration was approximately 80 mg carbon/1.
"Medium was effluent (30°C; pH 8) from a chemostat (10-20 mg/1 COD, 3 mg/1 poly-
saccharide) with no primary substrate remaining. Microorganisms were present
(approximately 10^ cells/ml) but no cells attached during the period of in-
terest. Fluid shear stress was 3.8 N/m2.
SEstimated from measurements of chemical oxygen demand (COD) adsorbed per unit
area. Assumed COD of protein is 0.855 mg COD/mg protein and protein density
is 1.3g protein/cm^.
392
-------
Attachment of Microorganisms
Previous research (40,41) suggests the existence of a two-stage
attachment process: (1) reversible adhesion followed by, (2) an irrever-
sible attachment. Reversible attachment refers to an initially weak ad-
hesion of a cell which can still exhibit Brownian motion but is readily
removed by mild rinsing. Conversely, irreversible attachment is a perma-
nent bonding to the surface, usually aided by the production of extra-
cellular polymers. Cells attached in this way can only be removed by rather
severe mechanical stress. Marshall (29) and Corpe (42) have implicated
polysaccharides and glycoproteins in irreversible attachment.
Most of the research on cell attachment has been conducted at very
low fluid shear stress or in quiescent conditions. There is yet to be a
demonstration of reversible adhesion in turbulent flow.
Fletcher (43) presents an excellent analysis of the rate of attach-
ment of a marine pseudomonad. The model and results are very similar to
those observed in molecular adsorption from solutions onto surfaces. In
both cases, the process rate is controlled by the concentration in the
bulk solution. The Langmuir adsorption isotherm model was used by Fletcher
and assumes the following:
1. adsorption is limited to a monocellular layer
2. adsorbed components are restricted to specific adsorption
sites
3. the heat of adsorption is independent of surface coverage.
Fletcher observed no cell desorption and modified the Langmuir model
accordingly.
According to Fletcher's development, the rate of attachment is pro-
portional to the cell concentration in the bulk fluid and the probability
393
-------
of a cell contacting a vacant adsorption site;
R = k X_ (1-0) (11)
£>
—2 —1
where R = rate of cell attachment {L t )
_3
X = cell number concentration in the bulk fluid (L )
B
6 = fraction of surface covered by cells (dimensionless)
--I
k = rate constant (Lt )
The number of cells adsorbed is proportional to the fraction of surface
covered so:
X = k'8 (12)
JC4
where X = number of cells adsorbed per unit area {L )
£\
-2
k1 = number of cells required for total coverage (L )
Then, by combining Eqs. (11) and (12):
R = k XB (1 - ^ ,
k' can be obtained from "adsorption isotherms" and is reported to be from
72 5
2-3 x 10 cells/cm . k is estimated at 1-8 x 10 cm/min from Fletcher's
data.
Fletcher indicates more rapid attachment with log phase cells and
slower rates with cells from the stationary phase and death phase, respec-
tively. More cells attach at higher temperatures (ranging from 3 - 20°C).
The nature of the attachment surface is an important factor affecting
attachment in heat exchangers. Wettability or critical surface tension,
is the property used most frequently to describe surface characteristics
in microbial attachment studies (44,45). In seawater, cell attachment
increases with increasing critical surface tension of the surface (in-
cluding glass, copper, polyethylene, teflon) with the exception of the
copper surface on which fewer cells attached (44). The copper may inhibit
394
-------
cell attachment by inhibiting a metabolic process necessary for attachment.
Even so, there are many examples of biofouling on cupronickel condenser
surfaces (46).
Microbial Reaction Processes
Table 10 is a matrix representation of the fundamental microbial
reaction processes. The rows of the matrix define the stoichiometry of the
process. For example, the growth process (Row 1) can be represented by
the following stoichiometric equation where the V's are stoichiometric
coefficients:
-V s - V z - V e + V x + V p + v a = 0 (14)
-L & ,3 ~t O O
The columns of the matrix define the "observed" rate as opposed to the process
rate. For example, observed substrate removal rate (Col. 1) can be repre-
sented by the sum of process rates as follows:
-qs = o^y + a2m + a3k - a4kL (15)
The a's are "observed" stoichiometric ratios. Table 10 only begins to
describe the complexity of microbial metabolism. Nevertheless, this amount
of "structure" is useful in modelling the rate processes affecting biofilm
development.
Trulear and Characklis (5) have observed substrate removal rate, q »
and net biomass production rate, y , in an experimental biofilm reactor.
It was convenient to express the substrate removal rate as:
qsMA = %jpMlh
—I
where q = specific substrate removal rate (t 5
M = total biofilm mass (M)
/\
p = biofilm density (ML )
Th = biofilm thickness (L)
A = wetted surface area (L )
395
-------
fable 10. A matrix representation for the fundamental microbial rate processes.
*
PROCESS RATE
FUNDAMENTAL PROCESS
Process Rate
Growth 11
Maintenance
exogenous m
endogenous k
Product k
Formation ^
Death
loss of k
viability
lysis k_ '
OBSERVED HATE
i uiu til uric, i KT
REACTANTS
Substrate
s
_
- -
-
C+)
qs
Nutrient
z
_
+
-
qz
Electron
Acceptor
e
-
-
—
_
qe
PRODUCTS
Biomass
XT xd
-t-
(+)
4-
(+)
Vn
Product
Pe Pi
+ (+)
+
+
+ 4-
+
S
Metabolite
a
+
+
+
+
qa
Ad
Pe
Pi
s
z
specific production or removal rate (t~^)
net specific growth rate or specific biomass production rate
total biomass concentration (ML~^)
inert solids concentration (ML~3)
extracellular microbial product concentration (ML )
intracellular microbial product concentration
substrate concentration (ML )
nutxient concentration (ML~ )
electron acceptor concentration (ML )
396
-------
The substrate removal rate, defined in this way, increases in propor-
tion to biofilm thickenss up to a critical thickness beyond which removal
rate remains constant (Fig. 11). The critical thickness is observed to in-
crease with influent substrate concentration (S.) or, more fundamentally,
surface loading rate. This behavior is confirmed by other investigators
(47,25,4) and is attributed to nutrient diffusional limitations within
the biofilm. Once the biofilm thickness exceeds the depth of substrate or
oxygen penetration into the biofilm (Fig. 12), the removal rate is unaffected
by further biofilm accumulation.
Observed substrate removal rate cannot be used to distinguish between
growth, maintenace, product formation, and death. It seems clear from other
data (27) that product formation (primarily polysaccharide) is significant
in the early stages of biofilm formation. Maintenance requirements become
important as the film gets thicker and substrate does not entirely pene-
trate the biofilm. These other process rates have not been measured and
are critical for determining stoichiometric coefficients and predicitng
biofilm development rates.
The substrate removal rate is also dependent on fluid velocity (Fig. 13).
At low fluid velocities, a relatively thick mass transfer boundary layer (6)
can cause a liquid phase diffusional resistance which decreases substrate
concentration at the liquid-biofilm interface and thereby decreases sub-
strate removal rate (Fig. 14).
Detachment of Biofilm
As the biofilm grows thicker, the fluid shear stress at the biofilm
interface generally increases. Also, as biofilms grow thicker, the poten-
tial for substrate, oxygen or nutrient limitation in the deeper portions
is great. These limitations may weaken the biofilm matrix and cause de-
397
-------
120 T
q M
^
,.
A
(mg/hr)
Substrate
Removal
Rate, 8Q
60 -•
40 -
20 .
100
20 40 60 80
Biofilm Thickness, Th
Fig. 11. The influence of biofilm thickness and substrate loading on
substrate removal rate [5].
TH
S=St
s=o
TH
.-Z 1_^ $=0
Fig. 12. Diffusional resistances in biofilms result in a. constant substrate
removal rate after the biofilm reaches a critical thickness beyond
which substrate cannot penetrate.
398
-------
SUBSTRATE
REMOVAL
RATE
q M.
(mg/hr)
90-r
80- •
70 •
^ tf
20 40 60 80 100
FLUID VELOCITY (cm/sec)
120
Fig. 13. Influence of fluid velocity on substrate removal rate [5],
I 1
A
TH
!
.i
u
TH
S=Ss
Fig. 14. Increasing fluid velocity past the biofilin results in a
smaller viscous sublayer (6) which increases mass trans-
fer rate through the liquid phase.
399
-------
tachment. Trulear and Characklis (5) report that the biofilm detachment
rate increases with increasing substrate removal rate, probably because
thicker biofilras result (Fig. 15).
Overall Rate of Biofilm Development
The development of a biofilnf is adequately described by a sigmoidal-
/'
shaped curve (Fig. 16). The slope of this curve at a particular time is
the net biofilm "development" rate, R , and is plotted vs time in Fig. 17a
for two experiments at different substrate loadings (5), The rate increases
to a maximum value corresponding to the sigmoidal inflection point and
then decreases to zero. Trulear (48) has measured maximum biofilm develop-
—5
ment rates during an experiment which range from 8.3 - 66 x 10 mg biofilm
-2 —1 -1
cm min for glucose concentrations of 6 and 130 mg 1 , respectively.
Since the biofilm detachment rate, IL. is proportional to biofilm thick-
ness, a higher detachment rate is observed in the high substrate loading
experiment (Fig, 17b). Thus,
*D + Rp * yn (17)
where y is the net biomass production rate. At steady state p = R_
n n ~&
since thickness remains constant.
The effect of fluid velocity on the plateau (or maximum) biofilm
thickness is illustrated in Fig. 18 for various substrate loadings. An
increase in fluid velocity increases biofilm detachment rate which minimizes
the plateau biofilm thickness. However, at low substrate loadings, fluid
velocity seems to have little effect on the plateau thickness.
SUMMARY
Microbial film formation has been discussed in terms of the more
fundamental physical, chemical and biological processes which contribute
to the biomass accumulation at a surface. The discussion suggests that
400.
-------
. Bio film
Detachment
Rate,
10 "
(mg/hr)
Biofilm
Detachment
Rate,
(mg/hr)
15 -
10 -.
5 -.
v = 124 cm s~~
/• v = 96 cm s
-1
20 40 60
Time (hrs)
v » 124 cm s
-1
v = 96 cm s
-1
f/
20 40 60 30 100
Biofilm Thickness, Th (ym)
Fig. 15. The influence of biofilm thickness and fluid velocity on biofilm
detachment race [5].
401
-------
I 60
50-.
S 40
CJ
1 30
^20-1-
E
fa 10
•
10 20'" 30
Time (hrs)
40
50
Fig. 16. The progression of biofilm development [5]
17a
Net Biofilm
Development
Rate
30 r
20 -•
(mg/hr-> 10
Biofilm
Detachment
Rate
(mg/hr)
17b
10 •
S. = 35 mg/1
= 7 mg/1
20 40 60
Time (hrs)
Fig. 17.
Time (hrs)
Changes in biofilm development rate, IL,, and biofilm detachment rate,
Rp, at different substrate loading rates. The sum, RD + Rpis the net
biomass production rate, Mn, for the system [5],
402
-------
IOQO r
800
in
en
05 600
C
.X
E
ffi400
|
I
*x
o
200
\S| = (OOmg/l
\
\
\
\
\
\
\
\
A. --- s.= 5 mg/,
3.0 4.O
3.0
4.O
5.0 8,0 7.0
0.875" 00 (8 BWG
3.O
5.0
0.300" ID
6.0
7.0
8.0
Average Fluid Velocity (ft/sec
Fig. 18. The influence of fluid velocity and substrate loading on the maximum
(or plateau) thickness attained in a 1.27 cm I.D. tube [7].
403
-------
more attention must be directed at the following topics:
1. More information is needed on the physical, chemical and
biological properties and structure of biofilms as a func-
tion of the interfacial environment.
2. Mathematical models for process rates as a function of bulk
concentrations, surface characteristics, and biofilm composi-
tion are needed to ascertain the rate-controlling process
in a given environment.
3. The models must be tested by experiments under controlled
conditions.
The results of such programs will lend insight necessary for scientists
and engineers to design appropriate systems which utilize reactive fixed-
film surfaces.
ACKNOWLEDGMENTS
Much of the reported work was accomplished at Rice University with financial
support from the National Science Foundation (ENG77-26934). The manuscript
was prepared by Ms. S. Vehnekamp.
404
-------
REFERENCES
1. Himmelblau, D.M.» and Bischoff, K.B. 1968. Process Analysis and Simula-
tion. John Wiley and Sons, Hew York.
2. Churchill, S.W. 1974. The Interpretation and Use of Rate Data: The
Rate Concept. McGraw-Hill Book Co., New York.
3. Costerton, J.W., Geesey, G.G., and Cheng, K.J. 1978. How Bacteria
Stick. Sci. Amer. 238:86-95.
4. Zelver, N. 1979. Biofilm Development and Associated Energy Losses in
Water Conduits. M.S. Thesis, Rice University, Houston, Texas,
5. Trulear, M.G., and Characklis, W.G. 1979. Dynamics of Biofilm Processes.
Presented at 34th Annual Purdue Industrial Waste Conference, W. Lafayette,
Indiana, May 8-10. 58 pp.
6, Nimmons, M.J. 1979. Heat Transfer Effects in Turbulent Flow Due to
Biofilm Development. M.S. Thesis, Rice University, Houston, Texas.
100 pp.
7. Characklis, W.G. 1979. Biofilm Development and Destruction. Final
Report, Electric Power Research Inst., RP 902-1, Palo Alto, CA., 160 pp.
8. Characklis, W.G. 1973. Attached Microbial Growths - II, Frictional
Resistance Due to Microbial Slimes. Wat. Res. 7:1249-1258.
9. Matson, J.V., and Characklis, W.G. 1976. Diffusion into Microbial
Aggregates. Wat Res. 10:877-885.
10. Herbert, D. 1961. The Chemical Composition of Microorganisms as a
Function of Their Environment. Symp. Soc. Gen. Microbiol. 11:391-416.
11, Schaechter, M., Maal^e, O. and Kjeldgaard, N.O. 1958. Dependency
on Medium and Temperature of Cell Size and Chemical Composition
During Balanced Growth of Salmonella typhurium. J. Gen. Microbiol.
19(3):592-606.
12. Tomlinson, T.G. and Snaddon, D.M. 1966. Biological Oxidation of
Sewage by Films of Microorganisms. Int. J. Air. Wat. Pollut. 10:865-881.
13. Mueller, J.A., et_ al_. 1966. Oxygen Diffusion through a Pure Culture
°^ Zooglea ramigera. Proc. 21st Ann. Purdue Industrial Waste Conf.
962-995.
14. Baillod, R.D., and Boyle, W.C. 1970. Mass Transfer Limitations in
Substrate Removal. J. San. Eng. Div., A.S.C.E. 96(SA4):525-545.
15. Pipes, D.M. 1974. Variation in Glucose Diffusion Coefficient through
Biological Floe. M.S. Thesis, Rice University, Houston, Texas. 50 pp.
405
-------
16. Williamson, K.J., and McCarty, P.L. 1973, Verification Studies of
the Biofilm Model for Bacterial Substrate Utilization. Presented at
Wat. Pol. Conf. Fed. Meeting, Cincinnati, Ohio.
17. Matson, J.V. 1975. Diffusion through Microbial Aggregates. Ph.D.
Thesis, Rice University, Houston, Texas. 100 pp.
18. Weast, R.D. (ed.). 1973. Handbook of Chemistry and Physics. CRC
Press, Cleveland, Ohio.
19. Perry, R.H. and Chilton, C.H. (eds.). 1973. Chemical Engineers Hand-
book. McGraw-Hill Book Co., New York.
20. Atomic Energy Commission. 1955. Reactor Handbook. Vol. 2, AECD-3646,
U.S. G.P.O., Washington, D.C.
21. Pollard, A.L. and House H.E. 1959. An Unusual Deposit in a Hydraulic
Tunnel. J. Pwr. Div., A.S.C.E. 85(PO6):163-171.
22. Minkus, A.J. 1954. Determination of the Hydraulic Capacity of Pipelines.
J. New Engl. Wat. Wks. Assoc. 68:1-10.
23. Arnold, G.E. 1936. Crenothrix Chokes Conduits. Engrg. News Rec.
116:774-775.
24. Anderson, M.R., Vacarro, R.F. and Toner, R.C. 1976. A New Concept
in Power Plant Operation to Control Slime Bacteria in Steam Electric
Condenser Cooling Systems. Presented at the Symposium on Biological
Fouling in Power Plants, Johns Hopkins University, Baltimore, MD.
25. Kornegay, B.H., and Andrews, J.F. 1967. Characteristics and Kinetics
of Biological Film Reactors. Final Report, FWPCA, Research Grant No.
WP-01181, U.S. GPO, Washington, D.C.
26. Gunsalus, I.e. and Stanier, R.Y. 1960. The Bacteria. Vol. 1, Academic
Press, New York.
27. Bryers, J.D. 1979. Rates of Initial Biofouling in Turbulent Flow
Systems. Ph.D. Thesis, Rice University, Houston, Texas.
28. Bryers, J.D. and Characklis, W.G. 1979. Measurement of Primary Biofilm
Formation. Presented at Condenser Biofouling Control Symposium, Atlanta,
GA. March 25-29.
29. Marshall, K.C. 1976. Interfaces in Microbial Ecology. Harvard Uni-
versity Press, Cambridge, Mass.
30. Corpe, W.A. 1978. Ecology of Microbial Attachment and Growth on Solid
Surfaces. Proc. Microbiology of Power Plant Thermal Effluents. (R.M.
Gerhold, ed.), University of Iowa, Iowa City, IA, pp. 57-66.
31. Costerton, J.W. 1979. The Mechanism of Primary Fouling of Submerged
Surfaces by Bacteria. Presented at Condenser Biofouling Control
Symposium, Atlanta, GA. March 25-29.
406
-------
32. Jones, H.C., Roth, I.L. and Sanders, W.M. 1969. Electron Microscopic
Study of a Slime Layer. J. Bact., 99:316-325.
33. Friedlander, S.K. and Johnstons, H.F, 1957. Deposition of Suspended
Particles from Turbulent Gas Streams. Ind. Eng. Chem., 49(2):1151-1156.
34. Lin, C.S., Moulton, R.W. and Putnam, G.L. 1953. Mass Transfer Between
Solid Wall and Fluid Streams. Ind. Eng. Chem. 45:636-640.
35. Beal, S.K. 1970. Deposition of Particles in Turbulent Flow on Channel
or Pipe Walls. Nucl. Sci. Engrg. 40:1-11.
36. Brock, T.D. 1979, Biology of Microorganisms. 3rd ed. Prentice Hall,
Englewood Cliffs, N.J.
37. Loeb, G.I. and Neihof, R.A. 1975. Marine Conditioning Films. In
Applied Chemistry at Protein Interfaces. Adv. Chem. Ser. 145, ACS,
Washington, D.C. pp. 319-335.
38. DePalma, V.A., Goupil, D.W. and Akers, C.K. 1979. Field Demonstration
of Rapid Microfouling in Model Heat Exchangers: Gulf of Mexico,
November 1978. Presented at 6th OTEC Conf., Washington, D.C. June 19-22.
39. Brash, J.L. and Samak, Q.M. 1979. Dynamics of Interactions Between
Human Albumin and Polyethylene Surface. J. Colloid, Interface Sci.,
in press.
40. Marshall, K.C., Stout, R., and Mitchell, R. 1971. Mechanism of the
Initial Events in the Sorption of Marine Bacteria to Surfaces. J.
Gen. Microbiol. 68:337-348.
41. Zobell, C.E. 1943. The Effect of Solid Surfaces Upon Bacterial Activity.
J. Bact. 46:39-59.
42. Corpe, W.A. 1970. An Acid Polysaccharide Produced by a Primary Film-
Forming Marine Bacterium. Dev. Ind. Microbiol. 11:402-412.
43. Flethcer, M. 1977. The Effects of Culture Concentration and Age,
Time, and Temperature on Bacterial Attachment to Polystyrene. Can.
J. Microbiol. 23:1-6.
44. Dexter, S.C. 1976. Influence of Substrate Wettability on the Formation
of Bacterial Slime Films on Solid Surfaces Immersed in Natural Sea
Water. Proc. 4th Int'l. Cong, on Marine Corr. and Fouling, Juan-Les-Pins,
Antibes, France.
45. Fletcher, M. and Loeb, G.I. 1979. Influence of Substratum Characteristics
on the Attachment of a Marine Pseudomonad to Solid Surfaces. Appl. Env.
Microbiol. 37:67-72.
46. Garey, J. 1979. Personal communication.
47. LaMotta, E.J. 1974. Evaluation of Diffusional Resistances in Substrate
Utilization by Biological Films. Ph.D. Thesis, Univ. of North Carolina,
Chapel Hill, N.C.
407
-------
48. Trulear, M.G. 1979. Rate of Biofilm Development in a Turbulent Flow
Field. M.S. Thesis, Rice University, Houston, Texas.
408
-------
EFFECTS OF ORGANIC LOADING AND MEAN SOLIDS
RETENTION TIME ON NITRIFICATION IN RBC SYSTEMS
By
F. Michael Saunders
Assistant Professor of Civil Engineering
Georgia Institute of Technology
Atlanta, Georgia
Rodney L. Pope
Engineering Science Inc.
Atlanta, Georgia
Miguel A. Cruz
Department of General Engineering
University of Puerto Rico
Mayaguez, P.R.
Oxidation of ammonia-nitrogen in rotating biological contactor
systems can be economically achieved with domestic and industrial
wastewaters using biofilm growths containing nitrifying bacteria.
The design of such wastewater nitrification systems is controlled by
the kinetics of growth of the autotrophic bacteria, Nitrosomonas and
Nitrobacter, which grow at rates lower than those of heterotrophic
bacterial populations routinely used in removing carbonaceous organic
matter. In addition, autotrophic nitrifying populations are sensitive
to such wastewater characteristics as temperature, pH and dissolved
oxygen concentration. Therefore, in systems developed to concurrently
remove carbonaceous organic matter and oxidize ammonia nitrogen, the
growth characteristics of nitrifying bacteria establish the minimum
specific growth rate for use in the design of the process system.
Considerable data and experimental relationships1"9 are available
which quantitatively express the effects of wastewater characteristics
409
-------
and process operational parameters on the growth of pure and enriched
culture suspensions of Nitrospmonas and Nitrobacter. The objective of
the research herein was therefore to experimentally determine if these
fundamental relationships for nitrifying bacteria could be applied
directly to attached films in RBC systems. It was, therefore, necessary
to develop procedures for evaluating growth rates of nitrifying bacteria
in mixed heterogenous cultures. Mean solids retention time, a parameter
routinely utilized in the evaluation of net growth rates in suspended
growth systems, was utilized to relate RBC reactor performance to the
growth characteristics of attached biofilms. The primary objective was
to determine if growth and ammonia removal rates for attached nitrifying
bacteria could be predicted with relationships established for growth
rates of nitrifying bacteria in pure and mixed culture suspensions. The
ultimate objective of the research initiated with this project is to
determine if these growth rate relationships can be used in design,
operation and evaluation of RBC systems.
EXPERIMENTAL PROCEDURES
Experimental Reactor. A laboratory-scale RBC system was used to
evaluate the effect of organic loading and mean solids retention time,
SRT, on nitrification. The continuous-flow RBC reactor used in the study
is presented schematically in Figure 1. As indicated in Figure 2, mixed
liquor from the RBC reactor was continuously pumped through a glass-walled
heat exchanger and discharged to the surface of a filtration sieve
(U. S. Sieve No. 80; sieve opening = 0.177 mm). Mixed liquor passed
through the sieve by gravity into the RBC reactor leaving only sloughed
biomass on the sieve surface. The mixed liquor recycle flow rate was
maintained at 1 2,/min, a rate sufficient to circulate the total volume
of the RBC reactor approximately one time every 6 rain. The use of the
heat exchanger, in conjunction with a reactor cover which minimized heat
and evaporative losses, was sufficient to maintain mixed liquor tempera-
tures between 21 and 25°C. The gravity-flow sieve effectively retained
sloughed biomass and mixed liquor suspended solids concentrations were.
maintained within limits typically experienced in full scale systems.
The physical characteristics of the RBC reactor are indicated in
Table 1.
TABLE 1
Physical Characteristics of Laboratory-scale RBC Reactor
Characteristic Parameter Value
Number of discs 12
Disc diameter 25.4 cm
Disc thickness 0.3 cm
Disc spacing 2.5 cm
Disc submergence 7.6 cm
Total wetted area 1.0m2
Rotational velocity 0.5 rps
Tip velocity 40 cm/s
410
-------
INFLUENT
EFFLUENT
END VIEW
o
o
o
o
o
o
FRONT VIEW
Figure 1. Schematic Diagram of End and Front View of Laboratory-
scale RBC Reactor
411
-------
PUMP *» HEAT EXCHANGER
SIEVES
DILUTION WATER
to
ORGANIC SUBSTRATE
INORGANIC SUBSTRATE
SAMPLE
Figure 2. Schematic Diagram of RBC Reactor System
-------
These parameters were held at constant values throughout all experimental
runs. Disc rotational velocity was set at 0.5 rps to provide an adequate
supply of dissolved oxygen throughout the experimental study. An examina-
tion of gas transfer characteristics of the reactor system indicated that
the overall oxygen mass transfer coefficient was approximately 1.1 x 10~3
cm/s (20°C), which resulted in a maximum oxygen transfer rate of 8.7 g 02/d.
Wastewater Composition. A synthetic wastewater was utilized to
simulate the soluble portion of a domestic wastewater. The synthetic
wastewater was introduced into the reactor in three separate flows, i.e.
an inorganic substrate flow, an organic substrate flow and a dilution
water flow. The inorganic and organic substrates were each applied at a
rate of 5 ml/min, while dilution water was applied at 40 ml/min for a
total flow of 50 ml/min and a hydraulic retention time of 2.0 hr. The
composition of the inorganic and organic substrates is presented in
Table 2. The COD concentration of the composite influent wastewater was
TABLE 2
Composition of Synthetic Wastewater
Inorganic Substrate
Compound
(NH,)2SO^
Mass Fraction
12.15
0.22
Organic Substrate
Compound
Acetic Acid
Benzoic Acid
Butyric Acid
Mass Fraction
1.7
1.4
1.7
KH2P04
Na2HPO
7H20
1.12
2.87
4.40
Citric Acid
Formic Acid
Lactic Acid
Propionic Acid
4.6
3.2
3.2
1.8
CaCl2
MgSO,
NaHCO.
0.66
0.66
22.15
27.95
26.37
Valeric Acid
Arabinose
Galactose
Sucrose
Xylose
1.7
10.3
10.3
26.0
10.3
MnSO,
CuSC
CoCl
H20
5H20
6H20
10H20
0.07
0.13
0.04
0.05
1.16
Phenol
Humics (Tea)
(FeCl3 • 6H20)
0.2
22.6
(1.0)
varied from 49 to 190 mg/1 while influent ammonia concentrations ranged
from 15.6 to 20.5 mg NH4+-N/£,. The COD/N ratio of the influent wastewater,
therefore, ranged from 2.5 to 10.5. Phosphorus and other essential
nutrients were provided in sufficient quantities so as to not limit the
growth of attached microbial populations.
413
-------
Operation of REG Reactor. A primary objective of the study was to
determine if classical data1"*^ for nitrification growth and substrate
removal rates could be applied to attached films in RBC reactor systems.
To examine net microbial growth rates in the RBC system, mean solids
retention time, SRT, values were controlled during a series of eight
experimental runs operated over a range of organic loading rates. Since
only organic loading rates were varied and nitrification growth rates
were to be examined, an experimental technique was developed to simul-
taneously control and monitor attached film growth rates. Several pro-
cedures were used to routinely monitor the rate of accumulation of (1)
sloughed biomass and (2) attached biomass.
Biomass which sloughed periodically from disc surfaces was collected
and monitored daily as mixed liquor suspended solids and effluent sus-
pended solids. Mixed liquor suspended solids were those collected on a
sieve in the recycle system and those actually in suspension within the
RBC reactor. Effluent suspended solids were collected with an effluent
composite sampling system,
In addition to natural sloughing of attached biomass, a portion of
the biomass attached to disc surfaces was mechanically removed on a
regular basis to more effectively control SRT values for the attached
biofilm. The twelve discs in the experimental reactor were subdivided
into 4 groups of 3 discs each, with disc sides numbered from 1 to 6.
The biomass attached to similarly numbered disc sides (i.e., one side
of each of four discs) was mechanically removed at intervals of 6, 12 or
18 days. All disc surfaces were sequentially scraped one time every
6, 12 or 18 days. For example, at a scraping interval of 6 days, all bio-
mass on disc sides numbered 1 was removed on day 1, all biomass on disc
sides numbered 2 was removed on day 2 and so forth until all sides had
been scraped once in a 6 day period. The mass quantity of biomass removed
from each disc was monitored every 1, 2 or 3 days, i.e. for scraping
intervals of 6, 12 or 18 days, respectively. The scraping of biomass
was continued until a steady state response was achieved for each reactor.
The controlled scraping and recovery of attached biomass, and sub-
sequent analysis of biomass contained in the mixed liquor and effluent,
provided sufficient data for determination of the net rate at which
biomass accumulated in the RBC system. To calculate SRT values for the
RBC system, a measure of the total quantity of attached biomass was
required. Periodically, the total quantity of attached biomass was
estimated through removal of a fixed portion of biomass from each of the
6 disc sides in a single set of discs. Experimental data collected in
this manner indicated that attached biomass was linearly distributed as
a function of time of biomass accumulation. Therefore, controlled
scraping of disc surfaces and recovery of all biomass naturally sloughed
within the RBC reactor allowed for the examination of biomass growth
rates as measured with SRT values. Further details of the scraping
procedures and biomass monitoring techniques are presented by Cruz*1
and Pooe12.
414
-------
Analytical Procedures. Analytical methods and procedures presented
in Standard Methods13 were followed in the analysis of wastewater and
biomass properties. A micro-COD**""13 procedure was used to monitor
all effluent COD concentrations. The coefficient of variation for
replicate effluent samples containing 10 to 50 mg COD/£ was 2 to 5%,
indicating excellent analytical precision. Ammonia nitrogen was
determined using an ammonia specific-ion electrode (Orion Research,
Cambridge, MA) and a standard addition technique. ' Nitrite-
and nitrate-nitrogen concentrations were determined with diazotization
and chromo trophic acid procedures, respectively. Total kjeldahl
nitrogen of biomass solids and filtered effluent samples was determined
with an automated method (Industrial Method 28-69A, Technicon Corp.,
New York). Suspended solids measurements for mixed liquor, effluent
and attached biofilm samples were determined with Gooch crucibles
containing glass fiber filter mats.
RESULTS
A total of nine experimental runs were performed sequentially in
three phases over a period of 11 months. Phase A included four runs
in which the effect of attached biofilm SRT on nitrification efficiency
was examined. Organic and nitrogen loading rates were maintained at
relatively constant values of 4-4.6 g COD/m • d (0.82-0.94 Ib COD/
1000 ft2 • d) and 1.11-1.45 g N/ra2 • d (0.23-0.3 Ib N/1000 ft2 • d) ,
respectively.
Phase B included one experimental run in which the effect of
hydrolysis of organically-bound nitrogen was examined. Glycine was
the sole source of nitrogen during this run and was supplied at a rate
of 1.62 g N/m2 • d (0.33 Ib N/1000 ft2 • d) . Organic loading for this
run was increased to 5.9 g COD/m2 • d (1.21 Ib COD/1000 ft2 • d) as a
result of the use of glycine as a nitrogen source.
Four experimental runs were included in Phase C to examine the
effect of increased organic loading on nitrification efficiency.
Organic loading was sequentially increased from 6.4 to 13.7 g COD/m • d)
(1.31 to 2.8 Ib COD/1000 ft2 • d) while nitrogen loading rates remained
at 1.3 to 1.44 g N/m2 • d (0.26 to 0.29 Ib N/1000 ft2 • d) . Experi-
mental conditions for the nine experimental runs are summarized in
Table 3.
SRT Values. The calculation of SRT was accomplished using values
for total attached and suspended biomass in the RBC system, M_, and the
average rate of accumulation (i.e. wastage) of biomass, r , in the
reactor system. SRT was then calculated using Equation 1.
SRT - — (1)
w
The rate of wastage of biomass, r , was equal to the summation of the
rate of accumulation of biomass in the mixed liquor and rate of discharge
415
-------
TABLE 3
Experimental Operating Conditions for Phases A, B and C
Parameter
Experimental Run
Scraping
Influent
Interval, d
DO, mg/1
Al
None
3.1
A2
6
3.4
A3
12
2.4
A4
18
2.6
Bl
18
2.2
Cl
12
2.2
C2
12
1.5
C3
12
3.1
C4
12
3
.3
Influent COD, mg/1
*
Influent Nitrogen,* mg/1
Hydraulic Retention Time, h
2
Hydraulic Loading, gpd/ft
Organic Loading, Ib COD/1000 ft2«d
Nitrogen Loading, Ib N/1000 ft2-d
49.0 54.0 59.0 62.0 84
19.6 17.5 19.8 15.6 23.1 19.2 20.5 20.2 18
2.1 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0
1.67 1.74 1.79 1.74 1.72 1.74 1.72 1.72 1.77
0.84 0.82 0.94 0.90 1.21 1.31 1.95 2.46 2.8
0.27 0.25 0.30 0.23 0.33 0.28 0.29 0.29 0.26
*Nitrogen as NH, -N for all runs except Bl which was as Organic-N.
-------
of biomass in the reactor effluent. The rate of accumulation of mixed
liquor biomass was equal to that retained on the recirculation sieve
and that removed directly from the mixed liquor in the reactor. Both
mixed liquor fractions were removed for analysis on a 1, 2, or 3 day
cycle, depending on the scraping interval used, and expressed herein as
average concentrations of mixed liquor suspended solids.
Measurement of total attached biomass could not routinely be achieved
without significant disruption of process performance. However, the net
accumulation of total biomass attached to disc surfaces over 6, 12 and
18 day scraping intervals was monitored throughout each experimental run
and these data were used to calculate total system biomass. By periodically
scraping a portion of the attached biomass from the six disc surfaces
in a set of four discs, the distribution of attached biomass on disc
surfaces was established for each run.
Data presented in Figure 3 indicate the distribution of attached
biomass solids during experimental runs A4 and Bl. The organic
loading rates during run Bl were higher than those for A4 and, as
expected, total attached biomass was higher for run Bl. Further-
more, the data in Figure 3, and those for other runs, indicate that the
quantity of attached biomass on disc surfaces was a linear function of
the time of accumulation, i.e. the time since biomass was last scraped
from a disc surface. The rate of growth of attached biomass was there-
fore uniform throughout the reactor system and total quantities of
attached biomass on disc surfaces varied primarily as a result of disc
scraping techniques. Furthermore, the quantity of biomass scraped from
similarly numbered disc surfaces, in accord with the scraping intervals
used, was indicative of the total quantity of attached biomass on other
disc surfaces. Total attached biomass, ML, was then estimated with solids
data for biomass scraped from disc surfaces and the associated elapsed
time intervals for which biomass had accumulated on other disc surfaces
in the reactor system.
Average values of the individual suspended and attached solids
fractions for the eight experimental runs during which controlled
wastage of attached biomass was practiced are presented in Table 4.
Mixed liquor suspended solids concentrations (calculated as an average
value using data for biomass solids collected on the recirculation
sieve) ranged from 33 to 487 mg/1 and were generally consistent with
values reported by Antonie (i.e. 49 to 275 mg/1). Effluent suspended
solids concentrations were maintained at low concentrations, i.e. 3.7
to 9,8 mg/1 and reflected the high efficiency with which mixed liquor
solids were removed from suspension in the recirculation system. This
furthermore indicated that the majority of the biological uptake of
influent carbonaceous and nitrogenous oxygen demand was achieved with
attached, and not suspended, biomass.
SRT values for the eight runs varied from 1.3 to 3.6d reflecting
the rapid growth of the attached biomass. These rapid growth rates
were dictated by intentional biomass wastage, however, natural
sloughing of biomass accounted for a significant portion of overall biomass
wastage rates, as reflected in the disparity between actual SRT values and
the scraping intervals of 6, 12 and 18 days.
417
-------
oo
0.8
w 0.7
ui
o
OQ
I
o
o
0.6
li-
CC
CO
o
CO
5 0.5
z
O
CO
g 0.4
0.3
0.2
0.1
Bl
A4
100
200
300
400
500
600
TIME OF BIOMASS ACCUMULATION, hr
Figure 3. Distribution of Attached Biomass in the RBC System for Experimental
Runs A4 and Bl
-------
TABLE 4
Suspended and Attached Solids and Mean SRT Data
for the RBC System
Experimental
Run
A2
A3
A4
Bl
Cl
C2
03*
C4*
Mixed
Liquor
Suspended
Solids
(mg/1)
33
212
232
290
198
363
330
487
Effluent
Suspended
Solids
(mg/1)
5.2
4.3
5.0
9.8
5.1
7.3
3.7
7.7
"r
Total RBC
Biomass
(g)
1.63
5,64
4.77
7.23
3.04
6.25
5.27
7.96
r
w
Rate of
Wastage of
Biomass
(g/d)
1.26
2.4
1.43
2.03
1.66
3.18
2.66
4.04
Mean
SRT
^
r
w
(d)
1.3
2.4
3.3
3.6
1.8
2.0
2.0
2.0
*Calculations are for highly transitory periods
419
-------
Nitrification Efficiency. During experimental phase A, organic and
nitrogen loading rates were held at constant levels. The average COD/N ratio
of the influent wastewater was 3.1, indicating a low organic loading rate
(i.e. influent COD = 49-62 mg/£). SRT was varied from 1.3 to 3.3d for the
three runs (A2-A4) in which controlled wastage of attached biomass was
performed. During the first experimental run (Al), an attached biofilm
was allowed to develop without controlled wastage of attached biomass. While
it was not measured, the SRT value for this run was much higher than the
highest value (SRT = 3.3d for A4) achieved during runs with controlled bio-
mass wastage.
Effluent nitrogen data for phase A were presented elsewhere11* and are
summarized in Table 5. Effluent ammonia and nitrate concentrations for runs
Al, A3 and A4 indicated that nitrification was achieved at the 91 to 99%
level, i.e. virtually complete nitrification. Nitrification was not achieved
during run A2, the run with the lowest SRT value (i.e. SRT = 13d), due to the
washout of nitrifying bacteria from the attached film.
TABLE 5
Steady State Ammonia - Nitrogen Concentrations in Influent and
Effluent Wastewaters During Phase A
Run
Al
A2
A3
A4
Influent
NHi^-N
(mg/JO
19.6
17.5
19.8
15.6
NHi^-N
(mg/£)
0.1
14.4
1.0
0.1
Effluent
N03--N
(mg/jQ
15.8
0.2
18.7
14.0
N02~-N
(mg/JO
<0.01
<0.01
0.80
<0.01
Percent
Nitrification
99
1
91
99
* Percent of effluent soluble nitrogen attributable to NC>3 -N
Since effective nitrification was achieved at low SRT values using
ammonia as the sole source of nitrogen, an experimental run was performed
during phase B to examine the effect of hydrolysis of organically-bound
nitrogen on nitrification efficiency. The RBC system was operated during run
Bl at the same scraping interval as the previous experimental run (A4) to
minimize the time to achieve steady state conditions. The low SRT value
associated with this run was used to determine if the rate of hydrolysis of
organic-nitrogen was a rate-limiting step with respect to nitrification. As
indicated in Table 6, influent organic-N, i.e. glycine, was hydrolysed and
liberated ammonia was oxidized at a 93% efficiency level. Hydrolysis of
organic nitrogen was, therefore, not a rate limiting step, even at a high
growth rate, i.e. a SRT value of 3.6d.
Four experimental runs were conducted during phase C to examine the
effects of increased influent organic matter concentrations on nitrification.
Influent COD was increased in step intervals, as indicated in Figures 4 and 5,
from an average COD concentration of 90 mg/£ to 190 mg/£ and approached that
420
-------
Q
O
o
20 -
206
214
222
230
DATE
238
244
figure 4.
Influent and Effluent COD Concentrations for
Experimental Run Cl
-------
244
252
260
268
276 282
DATE
Figure 5. Influent and Effluent COD Concentrations for
Experimental Runs C2-C4
-------
Run
TABLE 6
Steady State Ammonia- and Organic-Nitrogen Concentrations in
Influent and Effluent Wastewaters for Phases B and C
Influent Effluent
_j_ Soluble _ Percent
Org-N NHij -N Org-N NOs -N* Nitrification"
(mg/£) (mg/£)
(mg/£)
(mg/£)
Bl
Cl
C2
C3
04
23 . 1
19.2
20.5
20.2
18
0.3
0.5
0.1
0.2
0.8
0.5 14.8
16.7
16.0
15.9
1 ry •%'ytyt'ft
93
98
99
98
95
* N02 —N concentration was less than detectable limit (<0.01 mg N/£)
** Percent of effluent soluble nitrogen attributable to N03~-N
*** Effluent nitrate-N indicated was not a steady-state value
equivalent to a low strength domestic wastewater during run C4. Effluent
ammonia-nitrogen concentrations remained low throughout the four runs as
indicated in Figures 6 and 7.
High levels of nitrification were achieved during all runs ranging from
95 to 98%, as indicated in Table 6. Therefore, organic loading up to 13.7 gCOD/
m «d (2.8 lb/1000 ft2*d) had no negative impact on nitrifying bacteria in
attached biofilms. Factors contributing to this favorable response included
mixed liquor dissolved oxygen concentrations which averaged 5.0mg/£. Mixed
liquor pH values, in addition, were stable at 7.3 and temperature averaged 23°C.
Rirther study is required to examine attached film nitrification, especially at
low mixed liquor dissolved oxygen concentrations.
Nitrogen M ass Balance. Formation of nitrate-nitrogen by nitrifying
bacteria may result in denitrification in anoxic portions of an attached bio-
film. To determine if denitrification occurred in the RBC system and to
examine the extent to which nitrogen was removed by inclusion into microbial
cell mass, a detailed nitrogen balance was conducted for the nine experimental
runs.
The mass flows of nitrogen in influent and effluent wastewaters and in
the biomass removed from mixed liquor and disc surfaces were examined for
each experimental run and are presented in Table 7. Quantitative data were
obtained for all nitrogen fractions except the wasted biomass nitrogen for
run Al. Within the limits of cumulative analytical capabilities, the majority
of influent nitrogen was detected in the effluent wastewater and wasted biomass.
For runs A2 through C4, the overall nitrogen balance averaged 103% of influent
nitrogen. Therefore, no measurable denitrification occurred within the RBC
system. In addition, wasted biomass nitrogen data indicated that biomass
nitrogen content varied from 8.7 to 13.5% nitrogen and increased with biomass
SRT value.
423
-------
•&
E
*.
g
i-
Z
UJ
o
I
z
HI
C3
O
cc
0 SDOOOOO
206
214
222
230
244
DATE
Figure 6. Influent and Effluent Soluble Nitrogen
Concentrations during Experimental Run Cl
-------
z
g
<
DC
H
Z
LU
CJ
Z
o
o
z
UJ
o
o
DC
I-
O
O EFFNH4
A EFFNO3
242
282/
Figure 7, Influent and Effluent Soluble Nitrogen Concen-
trations during Experimental Runs C2, C3 and C4
-------
10
cr>
INFLUENT NITROGEN
TABLE 7
Nitrogen Balance for RBC System
Al A2 A3 A4 Bl Cl C2 C3 C4
,*
NHij+-N
Organic-N
EFFLUENT NITROGEN*
NHi^-N
N02I-N
N03 -N
Organic-N
*
WASTED BIOMASS NITROGEN
Organic-N
1.33 1.24
— —
0.01 0.97
0 0
1.08 0.01
— —
0.12
1.45
—
0.07
0.06
1.35
—
0.20
1.11
—
0.01
0
1.02
—
0.19
—
1.62
0.02
0
1.05
0.06
0.30
1.36
—
0.03
0
1.25
—
0.19
1.44
—
0.01
0
1.12
—
0.39
1.41
—
0.02
0
1.11
—
0.30
1.30
—
0.05
0
0.88
—
0.48
NITROGEN BALANCE (% of
influent-N) 82 89 116 110 88 108 106 101 108
* Nitrogen expressed as (g-N/m2*d)
-------
Organic Removal Efficiency. The removal of influent COD was excellent
for all experimental runs as indicated in Figure 8. Effluent organic quality
improved slightly with increased SRT values, as expected15. At organic loading
rates as high as 13.6 gCOD/m2«d (2.8 lbCOD/1000 ft2-d) COD removal efficiency
was 92% and compared very favorably with pilot-scale systems loaded at similar
levels10.
DISCUSSION
Examination of SRT values in Table 4 and nitrification data in Tables 5
and 6 indicated that nitrification efficiency was related to SRT. As presented
in Figure 9, nitrification did not occur below a SRT value of 1.8 days. However,
at and above SRT values of 1.8 days, nitrification was virtually complete and
the data indicated a response typical of the growth of nitrifying bacteria.
When examining growth relationships for these bacteria, a Monod-like hyper-
bolic relationship16 is used, i.e.
^ (2)
Ks + S
where y = net specific growth rate constant, y = maximum net specific growth
rate, S = concentration of limiting substrate and Ks = 'half -velocity constant.
Values of Kg for nitrifying bacteria range from 0.18 to 1.0 mg N/2,^ indicating
that ammonia and nitrite oxidation reactions proceed at maximum rates as near-
zero order reactions, at substrate nitrogen concentrations of 1.5 *1**
Nitrifying bacteria then typically grow very rapidly at or near critical wash-
out growth rates (i.e. y) while continuing to remove ammonia- and nitrite-
nitrogen to sub-mg/£ levels1"5.
The response of the attached biofilms in the experimental RBC system
was consistent with that of nitrifying bacteria. Washout of nitrifying
populations from the RBC system occurred abruptly between SRT values of 1.3
(Run A2) and 1.8 (Run Cl) days. This furthermore indicated that the maximum
net specific growth rate constant, p, for the nitrifying population was
between 0.56 and 0.83 d"1 , in accord with the relationship SRT = (1/y)15.
Of the nitrifying bacteria, Nitrosomonas is the slowest growing
bacterium1 '^'^ and, therefore, the oxidation of ammonia to nitrite is the
rate-controlling reaction. Examination of reported values of p for^
Nitrosomonas , as presented in Table 8, indicated that the range of y values
obtained in this experimental study were consistent with pure and mixed
culture data for Nitrosomonas, i.e. y = '0.17 - 1.08 d *.
While determining the precise p value for the nitrifying population in
the experimental system was not possible, the response of the attached bio-
film population was virtually identical with that predicted with data for
nitrifying bacteria. Therefore, the use of classical nitrification data in
predicting and modelling the response of RBC systems is justified. This
conclusion is further substantiated in Figure 10, in which experimental
nitrification data from numerous wastewater treatment studies19'21"26 are
427
-------
60
50
40
•5
O
O
" 30
IU
UJ
20
10
1234
SRT,d
Figure 8. Effluent Soluble COD for RBC
Reactor System
428
-------
100
I 73
2
CC
01
o
cc
UJ
0.
50
25
SRT.d
Figure 9. Nitrification Efficiencey
for Attached Biofilms as, a
Function of Biofilm SRT
z
o
o
il
z
1U
o
cc
UJ
0.
100
80
60
40
20
O
O
O
This Study
5 10
SRT.d
15
Figure 10. Nitrification Efficiency as a
Function of SRT from Numerous
Wastewater Studies19'21~26()
-------
presented with those from this study. The response of the RBC system in
this study was then consistent with that for numerous wastewater treatment
studies, providing further justification for the use of classical growth
and substrate removal relationships1"9 in the design and evaluation of RBC
systems.
TABLE 8
A
Values of y for Nitrosomonas in Pure and Mixed Cultures
Temperature
°C
21
25
25
21
23
25
23
20
P
d-1
0.85
0.88
0,55
0.85
0.37
0.17
1.08
0.94
Reference
8
17
15
18
19
20
3
6
With regard to modelling RBC systems for concurrent carbonaceous BOD
removal and nitrification, Mueller et a.1,27 presented a comprehensive mathe-
matical model of an RBC system. The model was calibrated and verified with
actual operational data from full-scale systems. A critical component of
the RBC model was the use of the classical Monod-like growth relationship16
to predict growth and substrate removal rates for nitrifying^bacteria.
Although the predicted values of the growth rate constants, y and Ks, for
nitrifying bacteria were significantly lower than reported values27, results
of the study reported herein strongly indicate that such modelling approaches
should be vigorously pursued.
SUMMARY AND CONCLUSIONS
Nitrification can be concurrently achieved with removal of carbonaceous
organic matter in single-stage RBC systems at high biofilm growth rates. In
addition, mean solids retention time, SRT, is a critical variable with respect
to the retention of nitrifying bacteria in attached biofilms. Reported values
for the growth constants, p and Kg, c'an be used to predict critical SRT values
at which washout of attached nitrifying populations will occur, as well as
establish effluent ammonia levels from RBC systems.
The hydrolysis of organically-bound nitrogen, when provided as simple
amino acids, i.e. glycine, is not a rate limiting reaction and does not
impede nitrification, even at high biofilm growth rates. Organic loading
rates up to 13.6 gCOD/m2-d did not impede nitrification at attached biofilm
SRT values slightly higher than those resulting in washout. The continuous
availability of dissolved oxygen within the RBC system, however, provided
condtions for near optimal growth of nitrifying bacteria. Further studies
must be pursued under oxygen limiting conditions to more effectively evaluate
the use of Monod kinetic relationships15 in the design of RBC systems. The
430
-------
close agreement of the results of this experimental study with kinetic
data and relationships for Nitrosomonas and Nitrobacter cultures confirmed
the validity of the use of these relationships in modelling, designing and
operating RBC systems.
REFERENCES
1. Painter, M. A., "A Review of Literature on Inorganic Nitrogen Metabo-
lism in Microorganisms", Water Res., 4_, 393 (1970).
2. Downing, A. L., et al., "Nitrification in the Activated Sludge Process",
J. Proc. Inst. Sew. Purif., 130, 10 (1964).
3. Poduska, R. A., "A Dynamic Model of Nitrification for the Activated
Sludge Process", Ph.D. Thesis, Clemson Univ., Clemson, SC (1973).
4. Sharraa, B., and Ahlert, R. C., "Nitrification and Nitrogen Removal",
Water Res., 11, 897 (1977).
5. Williamson, K., and McCarty, P. L., "Verification Studies of the
Biofilm Model for Bacterial Substrate Utilization", Jour. Water Poll.
Control Fed., 48_, 281 (1976).
6. Buswell, A. M., et al., "Laboratory Studies on the Kinetics of Growth
of Nitrosomonas with Relation to the Nitrification Phase of the BOD
Test", Appl. Microbiol., 2^, 21 (1954).
7. Lees, H., and Simpson, J. R., "The Biochemistry of the Nitrifying
Organisms. 5. The Nitrite Oxidation by Nitrobacter", Biochem Jour.,
65, 297 (1957).
8. Knowles, G., et al., "Determination of Kinetic Constants for Nitrifying
Bacteria in Mixed Culture with the Aid of Electronic Computer", J. Gen.
Microbiol., 38, 263 (1965).
9. Engel, M. S., and Alexander, M., "Growth and Autotrophic Metabolism of
Nitrosomonas europea", J. Bacteriol., 76, 217 (1958).
10. Antonie, R. L., Fixed Biological Surfaces-Wastewater Treatment, CRC
Press, Inc., Cleveland, OH (1976).
11. Cruz, M. A., "Effect of Cell Residence Time on the Oxidation of Carbon-
aceous and Nitrogenous Material with a Rotating Biological Contactor
System", SCEGIT-77-2016, School of Civil Engineering, Georgia Institute
of Technology, Atlanta, GA (1977).
12. Pope, R. L., "Effect of Cell Residence Time on Nitrification with a RBC
System", Thesis presented in partial fulfillment of requirements for Master
of Science in Sanitary Engineering, Georgia Institute of Technology (1978).
13. American Public Health Assoc., Standard Methods for the Examination of
Water and Wastewater-14th Edition, APHA, Washington, D.C. (1976).
14. Saunders, F. M., and Pope, R. L., "Effect of SRT on Nitrification in RBC
Systems", Proc. ASCE Environ. Engr. Division Spec. Conf., San Francisco,
CA, 361 (1979).
15. Lawrence, A. W. and McCarty, P. L., "Unified Basis for Biological Treat-
ment, Design and Optimization", Jour. San. Eng. Div., ASCE, 96, SA3, 757
(1970).
16. Monod, J., "The Growth of Bacterial Cultures" Ann. Rev. Microbiol. , _3_,
371 (1949).
17. Loveless, J. E., and Painter, M. A., "The Influence of Metal Ion Concentra-
tion and pH Value on the Growth of a Nitrosomonas Strain Isolated from
Activated Sludge", J. Gen. Microbiol., 52,1 (1968).
431
-------
18. Gujer, W., and Jenkins, D., "The Contact Stabilization Process-Oxygen
and Nitrogen Mass Balances, SERL Report 74-2, Sanitary Engineering
Research Lab., University of California, Berkley (1974).
19. Balakrishnan, S., and Eckenfelder, W. W., "Nitrogen Relationships in
Biological Treatment Processes I. Nitrification in the Activated Sludge
Process", Water Res., 3, 73 (1969).
20. Melamed, A., et al., "BOD Removal and Nitrification of Anaerobic Effluent
by Activated Sludge", Adv. Water Poll. Res., !_, 1 (1970).
21. Jenkins, D., and Garrison, W. E., "Control of Activated Sludge by Mean
Cell Residence Time", Jour. Water Poll. Control Fed., 40, 1905 (1968).
22. Prakasam, T. B. S., and Loehr, R. C., "Microbial Nitrification and
Denitrification in Concentrated Wastes", Water Res., _6, 859 (1972).
23. Stover, E. L., and Kincannon, D. F., "Evaluating Rotating Biological
Contactor Performance", Water and Sew. Wks., 123, 3, 88 (1976).
24. Wuhrmann, K., "Objectives, Technology, and Results of Nitrogen and
Phosphorus Removal Processes", In Adv. Water Qual. Improv., E. Gloyna
and W. W. Eckenfelder, Editors, pp 21-48, Univ. of Texas Press, Austin,
TX (1965).
25. Johnson, W. K., and Schroefer, G. J., Nitrogen Removal by Nitrification
and Denitrification", Jour. Water Poll. Control Fed., 36, 1015 (1964).
26. Lawrence, A. W., and Brown, C. G., "Design and Control of Nitrifying
Activated Sludge Systems", Jour. Water Poll. Control Fed., 48, 1834 (1976)
27. Mueller, J. A., et al., "Nitrification in Rotating Biological Contactors",
pres. at 51st Annual Water Poll. Control Fed. Conf., Anaheim, CA (1978).
432
-------
ROLE OF SUSPENDED SOLIDS IN THE
KINETICS OF RBC SYSTEMS
BY
Don F. Kincannon
Professor and Chairman
Bioenvironmental Group
School of Civil Engineering
Oklahoma State University
Stillwater, Oklahoma
Steve Groves
Environmental Engineer
Envirodyne
St. Louis, Missouri
INTRODUCTION
The rotating biological contactor is receiving increased interest as a
biological wastewater treatment process. As a wastewater treatment process,
the kinetics of the biological population are very important in the design
and/or operation of the process. While some researchers have speculated
that the RBC process may possess some of the behavioral traits of both the
fixed film and suspended culture processes, the majority, if not all, of the
design methods developed have discounted the effects of the mixed liquor
suspended solids present in the system. The biofilm concept has been com-
monly used for kinetic description of the RBC process.
Kornegay and Andrews (1) developed a kinetic model to describe fixed
reactors. In their mass balance they included the suspended solids. How-
ever, they made the assumption that RBC would be operating under "wash out"
conditions and the suspended solids would play no part in substrate removal.
Therefore, their actual model does not include suspended solids. They also
found that there was an active film thickness. The substrate utilization
433
-------
reached a steady state value after the biological film reached a thickness
of 70 y. Hoehn and Ray (2) fully supported the active film thickness theory,
however, they found that the active film thickness was 200 p instead of 70 p.
In general, the kinetic models do not take into account the actual bio-
mass. They consider the area that the biofilm covers and base their models
on this area. This study was undertaken to study whether or not the suspend-
ed solids play a role in the RBC process. It was also of interest to deter-
mine whether or not the consideration of actual biomass would provide a means
of comparing the RBC process with the activated sludge process.
MATERIALS AND METHODS
The model RBC unit used in this study consisted of a plexiglass tank
divided into four stages with four polyethylene discs in each stage. Each
disc was approximately 1/8 inch thick and 6 inches in diameter. This result-
ed in a total disc surface area of 6.28 square feet or 1.57 square feet per
stage. The volume of the liquid in the reactor was 5.1 liters. This pro-
vided a forty percent submergence of the discs. The hydraulic flow rates to
the RBC were maintained through the use of a constant head tank which re-
ceived a continuous flow of tap water. The flow from the constant head tank
was regulated by a valve combined with a flow meter on the tank outlet line.
Water from the constant head tank fed by gravity into a wet well, where it
was mixed with the concentrated synthetic waste to achieve the desired organ-
ic concentration. The synthetic waste was pumped to the wet well using a
Cole-Farmer Masterflex pump. From the wet well, mixture flowed by gravity
into the first stage of the RBC. The rotational speed of the discs was
maintained at 10 rpm.
The synthetic waste used in this study contained glucose as the sole
carbon source. All required nutrients were added in excess so that carbon
was the limiting growth factor. The COD of the influent wastewater was
maintained at 300 mg/1.
The RBC was initially seeded with effluent from the primary clarifier
of the Stillwater, Oklahoma, Wastewater Treatment Plant. The RBC was allowed
to operate as a batch unit for three days and then operated as a continuous
flow reactor. The RBC was operated in this manner for two weeks to allow
the development of a biological growth. Analyses were initiated after two
weeks. COD and suspended solids were run daily until a steady state condi-
tion was established. After steady state had been achieved, samples were
collected on two consecutive days. These were averaged and recorded as the
results of that phase of the study. In addition to COD and suspended solids
measurements, samples were also collected for determining the residual COD
and for conducting a batch growth study.
The residual COD determination consisted of taking one liter of the
RBC effluent and aerating it as a batch reactor for one week. COD analysis
was conducted at various times to ascertain the residual COD. Batch growth
studies using a shaker were conducted to determine the maximum specific
growth rate.
434
-------
Studies were conducted at five different flow rates. These flow rates
were 36 Si/day, 71 Si/day, 143 H/day, 178 H/day and 250 SL/day. These resulted
in hydraulic loadings of 1.5 gpd/ft2, 3.0 gpd/ft2, 6.0 gpd/ft2, 7.5 gpd/ft2
and 10.5 gpd/ft2.
RESULTS
The ACOD remaining at each stage for the five flow rates studied is
shown in Figure 1. ACOD represents the total amount of organic matter avail-
able as substrate to the microorganisms. It is determined by substrating the
residual COD from the observed COD. Figure 1 shows that the ACOD removed
by stages follows zero order kinetics for flow rates of 36 SL/day
(1.5 gpd/ft2) and 71 £/day (3.0 gpd/ft2). All other flow rates showed
kinetics approximating first order.
Figure 2 shows the mixed liquor suspended solids concentrations ob-
tained at each stage for the flow rates studied. The suspended solids in-
creased with the first stages and then decreased in the latter stages. Also,
the suspended solids concentrations decreased as the flow rates increased.
Suspended solids concentrations of 2000mg/l were reached for flow rates of
36 and 71 &/day, whereas, a suspended solids concentrations of 820 mg/1 was
the maximum achieved at a flow rate of 250 SL/day.
Figure 3 shows a comparison of the dilution rate and the maximum spe-
cific growth rate that was obtained by batch studies. The dilution rate was
greater than the maximum specific growth rate for all growth rates studied.
However, the dilution rate and growth rate were close for the 36 £/day and
71 fL/day flow rates. It must be recognized that the batch procedure used
for obtaining growth rates may not give the true growth rates that are occur-
ing in the continuous flow reactor. Therefore, it is possible that the true
growth rate in the RBC reactor was greater than the dilution rate. If this
is true, then the suspended solids would be effective in removing substrate
at these flow rates. Also, it appears that suspended solids are retained in
each stage rather than being held for only the detention time. This would
provide a smaller dilution rate for the solids. The retention of solids has
been observed at low hydraulic loadings for all RBC studies conducted in this
laboratory. The maximum specific growth rates for the 143 SL/day, 178 SL/day
and 250 SL/day flow rates were much lower than the dilution rate and the possi-
bility that the suspended solids would be effective in removing substrate is
very small. Therefore, it appears that the suspended solids may remove sub-
strate at low dilution rates (high detention times) but not be responsible
for substrate removal, at higher dilution rates (low detention times).
Two parameters were used to evaluate the role of the suspended solids.
These were specific substrate utilization and specific substrate utilization
rate. The specific substrate utilization is given as
S. - S
i e
— (suspended solids not included)
A.
m
435
-------
or
S_ C
» o
e
(suspended solids included)
v .r
X i A V
m s
where
S. - influent COD, mg/£
S = effluent ACOD, mg/£
X^ = mass of mierorganisms on discs, mg
X » suspended solids in reactor, mg/£
S
V = liquid volume of RBC reactor, £
The specific substrate utilization rate is given as
(S± - Se) F
X
m
or
(S± - Se) F
~X 4~X F~
m s
where
F = flow rate, £/day
The mass of microorganisms on the rotating discs were calculated by using the
active film theory. It is evident that an accurate estimate of the amount of
biological solids present on the discs and actively participating in the sub-
strate removal is not easy to make. The mass of microorganisms was calculat-
ed by multiplying the/ disc surface area times the density of the microorgan-
isms, times the active film thickness. The density of the microorganisms
was taken as 95 mg/em^ and the active film thickness was taken as 200 u« As
mentioned earlier, Kornegay and Andrews (1) reported an active film thickness
of 70 ]i and Hoehn and Ray (2) reported an active film thickness of 200 u«
Famularo, Mueller, and Mulligan (3) used their model to calculate an active
film thickness of 120 y. This provides a rather large tange of active film
thicknesses to choose from. In the study being reported, an active film
thickness of 200 u was selected because this value gave results that appeared
to be more reasonable than those obtained from other film thickness. The
biomass per stage would be
2 22 3
1.57 ft /stage x 929 cm /ft x 95 mg/cm x 0.02 cm = 2771 mg/stage
and the substrate utilization may be calculated for various stages and flow
rates .
Specific substrate utilization calculated by using only the active
film biomass is shown as a function of the effluent ACOD in Figure 4. It is
clearly seen that two different relationships exists. One for the 36 £/day
and 71 £/day flow rates and another relationship for the 148 £/day, 178
£/day and 250 It/day flow rates.
Specific substrate utilization can also be calculated by taking into
account the suspended solids for the 36 £/day and 71 £/day flow rates. The
436
-------
350
o 36 I/day 1.5 gpd/ft2
D 71 I/day 3.0 gpd/ft21
A 143 I/day 6.0 gpd/ft2_
0178 I/day 7.5 gpd/ft2
0250 I/day 10.5 gpd/ft2-
Figure 1. A COD Remaining vs.
RBC Stage.
437
-------
o>2000
£
CO
13
"5
CO
1600
<
'^
800
400
0
36 I/day
-143 I/day
71 I/day
250 I/day
1234
Stage
Figure 2. Suspended Solids vs.
RBC Stage.
438
-------
50
1
CO
"D
o 40
•*-»
CO
DC
.2 30
CD
+*
cd
DC
O
20
10
0
1 I 1 I
Dilution Rate
0
^Maximum
/^Growth Rate
VA^
320
80 160 240
Flow Rate, I/day
Figure 3. Maximum Growth Rate
and Dilution Rate vs.
Flow Rate.
439
-------
suspended solids were not included for the higher flow rates. This relation-
ship is shown in Figure 5. It is seen that all data fit one curve. This is
in contrast to Figure 4 in which there are two different curves.
Another parameter that is often used to describe activated sludge is
the specific substrate utilization rate. This parameter includes the hy-
draulic flow rate. Figure 6 shows the relationship between the specific
substrate utilization rate (active biofilm solids only) and the effluent
ACOD. The data fits one curve with the relationship
(S. - S ) F 39 S
i e e
X 232 + S
m e
When the suspended solids are included with the 36 a/day and 71 H/day flow
rates, a straight line relationship is obtained (Figure 7). The mathematical
relationship is
(S - S ) F
X +X V - °'13 Se
m s
Regression analyses were conducted with the data presented in Figures 6 and 7
and these relationships gave the best fit. The correlation coefficient for
the data presented in Figure 6 was 0.93 and the correlation coefficient for
the data presented in Figure 7 was 0.95.
The food to microorganism ratio has been used to evaluate activated
sludge processes for a number of years with much success. Bentley and
Kincannon (4) have also used the food to microorganism ratio to evaluate
biological towers. This ratio can also be used to evaluate the performance
of an RBC. The active film biomass and suspended solids at flow rates of
36 ifday and 71 Si/day were used to calculate the food to microorganism ratio.
Figure 8 shows the treatment efficiency obtained at various ratios. This
Figure also suggests that the suspended solids should be considered as part
of the active biomass at higher detention times. It is also seen that re-
moval rates greater than 95 percent were achieved at food to microorganism
ratios below 1.0.
Figure 9 shows the treatment efficiencies obtained for various hydrau-
lic detention times. The only flow rates that produced detention times
capable of achieving treatment efficiencies of 95 percent or greater were
36 A/day and 71 Si/day. These flow rates gave a hydraulic loading of 1.5
gpd/ft2 and 3.0 gpd/ft2.
DISCUSSION
The results of this study show that there are conditions where the
suspended solids are active in removing substrate in an RBC. The general
concept is that the suspended solids are washed out of the reactor and are
not active in substrate removal. This certainly is the case when the dilu-
tion rate exceeds the maximum specific growth rate. However at low hydraulic
loadings, the dilution rate may be less than the maximum specific growth rate.
440
-------
Ill
0 40 BO 120 160
Se» mg/l
Figure 4. Specific Substrate
Utilization Without
Suspended Solids vs.
Effluent ACOD.
441
-------
.08
.06
CD
CO
1
mmmm
CO
CO
<
•» •
E
X
.04
.02
0
- o
u O
0
D
160
40 80 120
Se, mg/l
Figure 5. Specific Substration
Utilization Including
Suspended Solids vs.
Effluent ACOD.
442
-------
16
12
u.
-------
16
12
LL.
0
CO
CO
CO
X
0
i r
i I I 1 r I
0 40 80 120 160
Se, mg/I
Figure 7. Specific Substrate Utilization
Rate Using Suspended Solids
vs. Effluent ACOD.
444
-------
o
CM
o
CM
CD
CO
CM
o
o
o
00
CO
o
CO
O
CM
CO
X
E
X
03
DC
E
CO
"c
co
o>
o
o
o
i
T3
o
o
LL
o
O
I
I
O
O
LL
CO
o
c
0 .
"5.2
LUOC
"HE
CD CO
EC
CD 5-
> O
CO
0
L_
1}
D5
LI
445
-------
100
8 so
0
Q.
o 60
0
'o
40
0
i
20
0
•o
0123
Hydraulic Detention Time, hours
Figure 9. Treatment Efficiency vs.
Hydraulic Detention Time.
446
-------
Under these conditions, the suspended biological solids would be active in
utilizing substrate. It appears that this is what happened in this study.
At hydraulic loadings of 1.5 and 3.0 gpd/ft » the batch growth studies gave
maximum specific growth rates close to the dilution rates. Higher hydraulic
loadings resulted in the dilution rate being much greater than the maximum
specific growth rate.
After making the assumption that the biological suspended solids were
only active in substrate removal at hydraulic loadings of 1.5 and 3.0 gpd/ft ,
comparisons were made between parameters using suspended solids and those not
using suspended solids. A single relationship between specific substrate
utilization and effluent ACOD was obtained only when the suspended solids
were included in calculating the specific substrate utilization.
A Monod like relationship between the specific substrate utilization
rate and effluent ACOD was observed when the suspended solids were not in-
cluded. However, a first order relationship was observed when the suspended
solids were included.
These two comparisons show that the suspended solids do exert an effect
on the kinetics of the RBC process. This is especially true at low effluent
substrate requirements. The normal way to achieve a low effluent substrate
is by using a low hydraulic loading. This produces a detention that allows
the suspended solids to not "wash out". This then allows the suspended
solids to be active in removing substrate. Since present design models do
not consider suspended solids, most designs have a built-in safety factor.
This may be an advantage of not considering suspended solids when designing
an RBC process. However, suspended solids consideration could provide a
more economical design.
It was also an interest of this study to determine whether or not
parameters using biotnass could be used to describe the RBC process. Such
parameters would allow a more direct comparison between the RBC process and
the activated sludge process. It was found that food to microorganism ratios
could be calculated using an active biomass and suspended solids that compare
well with food to microorganism ratios calculated for activated sludge. RBC
food to microorganism ratios of one or less gave treatment efficiencies of
95 percent or better.
A reaction rate constant that compares with Eckenfelder's activated
sludge reaction rate constant was also determined in this study. Figure 7
shows this relationship. The slope of the straight lines is 0.13 l/mg/£-day.
This reaction rate compares well with Eckenfelder's constant that has been
reported for easily biodegraded organic wastewaters.
This study has shown that suspended solids can play an important role
in the kinetics of RBCs and that they should not be completely disregarded.
The actual importance of the suspended solids depends upon a particular
situation.
447
-------
REFERENCES
1. Kornegay, G. H. and Andrews, J. F., "Kinetice of Fixed Film Biological
Reactors," Jour. Water Poll. Control Federation, 40, R460, 1968.
2. Hoehn, S. C. and Bay, A. D., "Effects of Thickness on Bacterial Films,"
Jour. Water Poll. Control Federation, 45, 2302, 1973.
3- Famularo, James A., Muller, A. and Mulligan, Thomas, "Application of
Mass Transfer to Rotating Biological Contactors," Jour. Water Poll.
Control Federation, 50, 653, 1978.
4. Bentley, Terry L and Kincannon, Don F., "Application and Comparison of
Activated Sludge Design and Operational Parameters to Biological Tow-
ers," Water and Sewage Works, Reference Number, 1976.
448
-------
THE KINETICS OF A ROTATING BIOLOGICAL CONTACTOR
TREATING DOMESTIC SEWAGE
Abraham Pano
Research Assistant
Utah State University, Logan, Utah, U. S. A.
Oames H, Reynolds
Assistant Professor, Civil and Environmental Engineering
Utah State University, Logan, Utah, U. S. A.
E. Joe Middlebrooks
Dean, College of Engineering
Utah State University, Logan, Utah, U. S, A.
INTRODUCTION
Background
The Rotating Biological Contactor (RBC) has been employed for
biological treatment of municipal and industrial wastewater for several
years. The process has been shown to be efficient and cost effective
in various types of applications. Traditionally, the design of RBC
systems has been based primarily on empirical relationships and design
curves developed from pilot plant studies. This empirical design
approach has ignored the basic concepts of biological substrate
removal kinetics. Although several recent investigations have
developed various models to describe' the performance of the RBC process,
there is limited data available describing the kinetic constants
associated with substrate removal. Also, there is very little
available information concerning the effect of temperature on the
kinetic constants associated with RBC substrate removal.
449
-------
Objectives
The general objective of this study is to develop the kinetic
constants describing carbonaceous substrate removal in the RBC process
treating domestic sewage and to determine the effect of temperature
on these constants. The kinetic constants developed will then be
employed to develop a rational approach to RBC systems design for
treatment of municipal wastewater.
To accomplish the above general objective, the following specific
objectives will be achieved.
1. Develop a Monod type [Monod 1942], steady state model which
describes carbonaceous substrate removal in the RBC process.
2. Determine the values of the kinetic constants (maximum specific
growth rate,l3; half saturation constant, xs; decay coefficient,
kg; and the yield coefficient,!) for the above model as a
function of temperature.
3. Develop, using laboratory scale RBC units treating domestic
sewage, the data base required to achieve the above specific
objectives.
4. Employ the kinetic constants produced to develop a rational RBC
design procedure for domestic wastewater.
Scope
This paper details preliminary results obtained from the initial
phase of the study. The results presented in this paper are limited to
only one temperature (15°C) and the analysis is preliminary in nature.
The study is currently underway at Utah State University, Logan, Utah,
and will be completed in the near future. The results of the entire
study will be presented at a later date.
PREVIOUS INVESTIGATIONS
Initial attempts to model the performance of rotating biological
contactors (RBC) were empirical in nature and mainly employed regression
analysis [Hartman 1965, Jost 1969, Anton-Le and ffelah 1969, and Weng
and Molof 1974]. Their efforts generally ignored temperature effects
and were not directly related to microbial substrate removal. Substrate
kinetic removal equations have been developed by several investigators
[Criewes 1972, Hans ford et at. 1976, Benjes 1978, Kornegay and Andrews
1968, Komegay 1972, Kovnegay 1975, and Clark et al. 1978]. In general
these models employed either saturation kinetics [Monod 1942] or first
order kinetics to describe substrate removal. Usually these equations
were limited to a single stage system or treated a multi-stage system
as a single unit.
Kovnegay and Andrews [1968] investigated the kinetics of fixed
film biological reactors at 25°C using a rotating drum and glucose as
the substrate. Under controlled flowrates, glucose concentration, and
450
-------
attached film thickness, they found active biomass thickness to be 70p,
the half saturation constant (K8) 121 mg/£ (glucose), and the maximum
specific growth rate (p.) 0.28 hr'1 (6.7 day"1). Grieves [1972]
developed a theoretical dynamic and steady state model for the RBC
using kinetic constants from the literature. He verified his model
by conducting dynamic tests at 20°C using glucose as the substrate.
Clark et al. [1978] investigated the kinetics of BOD removal under
varying wastewater flows and concentrations, using primary effluent
with soluble BOD5 of 32-88 mg/i. Working with a four-stage RBC unit,
they obtained a yield coefficient (J) of 0.96 (based on soluble BOD5),
a half saturation constant (K8) from 431 mg/£ (first stage) to 18 mg/S,
(fourth stage), and maximum growth rates (y) of 4.4 day"1 (first stage)
to 0.3 day'1 (fourth stage).
Recent investigations have employed either mass transfer models
or have combined mass transfer concepts with substrate removal kinetics
to describe RBC performance \_Sohroeder 1976, Friedman et al. 1976, and
Famularo et al. 1978]. These equations have generally been applied in
oxygen limited substrate removal situations.
MATERIALS AND METHODS
Four, four stage, 38 cm diameter, laboratory scale rotating
biological contactor (RBC) units furnished by the Environmental Systems
Division, George A. Hormel and Company, Coon Rapids, Minnesota, will
be employed to develop the data necessary to determine the values of
the kinetic constants. These units will receive settled domestic
sewage from the Hyrum City Wastewater Treatment Plant, Hyrum, Utah.
The sewage is collected during a 30 minute period three times per week
and is transported to the Utah State University campus and stored at
approximately 2°C until fed to the RBC units. The sewage is collected
at the same time of day to ensure uniform composition and strength.
The units will be operated at four different organic loading
rates at four different temperatures. This paper details the
preliminary results of the 15°C experimental design. The hydraulic
loading rate will be held constant and is the same for each unit. The
experimental apparatus is shown schematically in Figure 1.
Tables 1 and 2 indicate the hydraulic loading rate, the organic
loading rate, and the liquid temperature of each unit during the 15°C
experimental phase. The organic loading rates ranged from 5.9 g COD/
m2/day in Unit A to 22.0 g COD/m2/day in Unit D. The various organic
loading rates are achieved by diluting the raw sewage with dechlorinated
tap water. The liquid temperature in the RBC units ranged from 14.1
to 16.6 °C. This temperature variation was a result of heat loss
through the stages rather than temperature differences between similar
stages.
The influent to the system and the effluent from each stage was
monitored using 24-hour composite samples (20 minute intervals) and
occasionally grab samples every two days during steady state conditions
451
-------
Tap water
Drainage
Figure 1. Schematic of experimental apparatus employed to develop
data base for developing kinetic constants.
Table 1. Average hydraulic and organic loading rates in the four
laboratory scale rotating biological contactor units.
Parameter
A
Unit
Average
hydraulic loading
rate (a/day)
Average organic
loading rate
(grams of COD/
m /day)
274±12 285±6
5.9
11.5
283±14 29U9
15.5
22,0
452
-------
Table 2. Average liquid temperature (°C) in each stage of each
laboratory scale rotating biological contactor unit during
data collection phase.
Stage
Number
1
2
3
4
Unit
A
16.6+0.4
15.4±0.5
14.8±0.3
14.8±0.3
B
16.0±0.3
15.4+0.4
14.4±0.5
14.1±0.5
C
16.0+0.4
15.0±0.4
14.4±0.5
14.1±0.4
D
16.4±0.3
15.7±0.2
15.U0.2
14.7+0.3
Average
emperature
16.3
15.4
14.7
14.4
for chemical oxygen demand (COD), suspended solids, volatile suspended
solids, ammonia-nitrogen, nitrite-nitrogen, nitrate-nitrogen and
total Kjeldahl nitrogen. In addition, in-situ measurements of flow,
temperature, pH and dissolved oxygen were conducted. The ampule
technique [Oceanographies 1978] was employed for COD analysis while
all other analyses were conducted according to Standard Methods [APHA
1975].
At the conclusion of the data collection phase, the entire solids
from each disc in each stage was removed from the disc and analyzed for
total and volatile solids. Thus, the total biomass of the system was
determined.
Kinetic Constants
The model employed in the study is based on Monod [1942] substrate
removal kinetics and assumes steady state conditions.
The substrate removal in a single stage Rotating Biological
Contactor (RBC) can be equated to the growth of microorganisms as
shown in equation 1.
a
J dt
(D
where
s
c
J
= flowrate
= influent substrate concentration
= effluent substrate concentration
= yield coefficient
mass of biomass produced
mass of substrate removed
453
-------
Jy
-3T- = change in biomass per unit time
Monod [1942] indicated that the change in biomass under substrate
limiting conditions can be represented by a saturation function
where
&L = pY S X \
dt M \Kg + Sj
y = maximum specific growth rate
X = biomass concentration
K = half saturation constant (-i.e., substrate concentration at
s a growth rate equal to half the maximum specific growth
rate)
S = limiting substrate concentration
Equations 1 and 2 can be combined and expressed in linear form as
shown below.
IK \ ,
(3)
Q(S0-S )Y
e>
If the value of the yield coefficient, Y, is known, experimental
data can be fitted to equation 3 and the value of the kinetic constants,
0 and K , can be determined.
S
The value of the yield coefficient, Y, can be obtained by writing
a mass balance biomass in a single RBC stage as shown in equation 4.
This equation assumes no change in the mass of biomass occurring once
steady state conditions are achieved.
Biomass Produced - "»«» H- Decay (4)
Equation 4 may be expressed mathematically as
where
X = effluent biomass concentration (generally measured as
e volatile suspended solids)
kj = decay coefficient
X = total biomass (generally measured as volatile suspended solids)
Equation 5 does not distinguish between "active" and "nonactive"
biomass. Combining equation 5 with equation 1 and rearranging into
linear form results in equation 6.
454
-------
(6)
Experimental data fitted to equation 6 will result in values for
the yield coefficient, Y, and the decay coefficient, k^. Thus, all of
the desired kinetic constants can be determine from equations 3 and 6.
Solving these equations with experimental data collected at various
temperatures will yield kinetic constant values at these temperatures.
Thus, the effect of temperature on the kinetic constants can be
determined. A previous study conducted by the authors indicates that
this effect may be described by an Arrhenius type relationship
iMikula 1979, Mikula et al, 1980].
RESULTS AND DISCUSSION
Process Performance
The laboratory scale rotating biological contactor (RBC) units
were operated for approximately one month before steady state conditions
arhipvpri in parh <;t.aop. Thp at.t.anhprl hinmas^ in thp first
-------
Table 3. Dissolved oxygen and pH of the laboratory s€ale rotating
biological contactor (RBC) units.
Unit
A
B
C
D
Stage
Number
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
December 1 ,
pH
7.80
7.85
7.90
8.00
7.70
7.80
7.85
7.90
7.70
7.80
7.85
7.85
7.65
7.80
7.85
7.85
1979
Dissolved
Oxygen
(Mg/1)
5.0
6.7
7.4
7.8
3.9
5.3
6.8
7.5
3.1
4.6
5.8
7.6
2.5
3.6
4.5
10.2*
December 24,
pl-l
7.70
7.65
7.60
7.60
7.80
7.75
7.65
7.50
1979
Dissolved
Oxygen
(Mg/1)
4.0
5.1
4.3
5.2
2.8
3.2
3.4
3.9
* Suspected analytical error
Table 4. Total volatile solids biomass present in each stage of the
laboratory scale rotating biological contactors.
1
2
3
4
Volatile Solids in Each Stage (grams/m2)
ouayc
Number
Unit A
Unit B
Unit C
Unit D
27.8
6.6
4.1
2.1
36.4
29.6
10.9
6.2
47.5
36.0
21.5
13.8
46.7
38.8
23.2
16.8
456
-------
20
'5-
O)
e
-------
- C-NH4
D-NH4
• C-N03+N02
• D-N03+N02
1 2 3
Stage Number
Figure 3. Nitrogen response of laboratory scale rotating biological
contactors, Units C and D.
458
-------
\
o>
E
1
Q
O
O
275 -i
i
250 -
225 -
200 -
t
175 -
1 50 -
125 -
100 -
75 J
50 -
25 -
£_ \*J
n ..
!
\
\
\
\
\
\
\ + — UNIT-A
' s* 1 IKI T T O
\ -X" UNI ~o
\ \
\ \ '•* — UNIT-C
\ \
\ \ »— UNIT-D
\ I
\ i
\ \
'. \ \
\ \ \
\ » *
\ \ \
\ \ \
\ \ \
\ \ \
\ \\
\ \ \
V '^v\
\ \ \ \
\ • \\
\ \\» *---^.
\ \1 "—« " B
V II^-r.7.rr.rr-A
^ _ • — gzi: ^g
0
1 2 3
Stage Number
Figure 4. Carbonaceous substrate removal in the laboratory scale
rotating biological contactor units. Note: (1)0= Total
Chemical Oxygen Demand (COD) of the influent to the system.
(2) Effluent values for stages 1, 2, 3, and 4 are in terms
of soluble COD.
459
-------
Kinetic Constant Determination
The steady state data employed with equations 3 and 6 to determine
the carbonaceous substrate removal kinetic constants are reported in
Table 5. The values for the influent substrate concentration, S0 , are
measured in terms of total chemical oxygen demand (COD). While the
effluent substrate concentrations are measured in terms of soluble
COD. This approach assumes that all the influent particulate organic
material is solubilized within the reactor and is available for uptake
by the microorganisms.
The values 'for the total biomass, x, present in each reactor are
measured in terms of volatile solids and the entire depth of biomass
is assumed to be active. This assumption was made because measurements
for the fraction of active biomass were not conducted.
The data obtained from Unit A were not included in the development
of the kinetic constants due to the high degree of nitrification
occurring within this unit. Initial attempts to include the Unit A
data in the analysis resulted in unrealistic and inconsistent results.
Future data analysis at other temperatures will attempt to account
for the nitrification phenomenon.
The results of the linear regression for equation 6 to determine
the values for the yield coefficient, J, and the decay coefficient, k^,
are reported in Table 6. The yield coefficients varied from 0.81 in
the fourth stage to 1,44 in the second stage. These values are
based on COD removal. These values are similar to those found by
Clark et al. [1978] who reported yield coefficients of 0.68 to 1.11
based on soluble biochemical oxygen demand (BODS).
The values for the decay coefficient, k$, range from 0 days"1 in
the third stage to 0.44 days'1 in the second stage. These values appear
to be relatively high and are probably highly influenced by the incor-
poration of the total volatile biomass in the analysis rather than
only the "active" biomass.
The values for the other kinetic constants developed from
equation 3 are reported in Table 7. The regression plots of equation
3 for each stage are shown in Figures 5 to 8. The values for the
maximum specific growth rate, fl, appear to decrease with stage, except
for the second stage. The half saturation values, Ks, also decrease
significantly in the fourth stage. The values of y range from 1.47
day'1 to 2.92 day"1. Kornegay and Andrews [1968] reported a y of 6.7
days"1 at 25°C and Clark et al, [1978] reported ay of 4.4 days'1 in
the first stage of an RBC unit and 0.3 days"1 in the fourth stage of
an RBC unit. The relatively small values of y reported in this
present study are probably due to the low temperature (15°C). Reynolds
[1975] reported that the value of 0 for algae is temperature related
and is smaller at lower temperatures.
The values for the half saturation constant, JCS, range from 6.0
mg COD/£ is the fourth stage to 67.4 mg COD/£ in the second stage.
460
-------
Table 5. Steady state data employed to develop kinetic constants for carbonaceous substrate removal at 15°C,
t
Influent Total COD
t
Effluent Total COD'
Effluent Soluble COD
Unit Stage
u"11 Number No. of
Effluent Volatile
Suspended Solids
Average No. of Average No. of Average No. of Average
Samples Concentration Samples Concentration Samples Concentration Samples Concentration
(rag/A) (mg/£) (mg/£) (mg/£)
A 1 8
2
3
4
B 1 9
2
3
. 4
C 1 11
2
3
4
D 1 12
2
3
. 4
76.1 2
4
2
141.4 2
3
2
2
192.6 5
4
5
4 /
265.2 5
5
5
. 4
41.9
53.3
35.0
80.0
47.4
55.9
41.5
123.6
101.9
96.7
77.1
168.5
131.7
136.2
107.7
4
3
2
2
2
3
2
3
6
5
5
4
6
7
6
6
20.7
23.4
20.6
21.3
28.3
23.1
24.1
22.5
38.3
38.2
33.5
24.6
52.1
42.4
53.6
40.2
4
5
5
5
5
5
5
5
7
7
7
7
7
7
7
7
24.0
27.7
25.3
20.7
52.3
37.8
30.2
33.7
85.4
58.5
72.0
54.8
141.3
122.5
99.1
73.5
- Influent to first stage
Influent to next stage
-------
1.75 -I
n 1.50 -
to
TO
1.25 -
"o 1.00 -
CO
co .75 -
.50 -
x .25 H
0
0
2 3
102*Se~1 (mg/I)"1
Figure 5. Regression analysis for determination of 'the maximum specific
growth rate, 0, and half saturation constant, Ks, in the
first stage.
1.75 i
n 1.50 -
to
TO
1.25 -
"« 1.00 -
co
co .75 -
.50 -
x .25 H
0
B2
0
2 3
102*Se~1 (mg/l)'1
Figure 6. Regression analysis for determination of the maximum
specific growth rate, 0, and the half saturation constant,
K , in the second stage.
s
462
-------
1.75 -i
co 1.50 -
1.25 -
ID
-o
C3
"« 1.00 -
CO
I
co .75 -
a .50 -
x .25 H
R2=0.93
0
0
234
102*Se~1 (mg/l)'1
Figure 7. Regression analysis for determination of the maximum
specific growth rate, y, and the half saturation constant,
Ks, in the third stage.
1.75 -|
«, 1-50 -
1.25 -
(O
"O
i
~v i.oo H
CO
I
o
co .75 -
o .50 -
.25 -
0
R2=0.89
102»Se~1 (mg/l)'1
Figure 8. Regression analysis for determination of the maximum
specific growth rate, ft, and the half saturation constant,
K , in the fourth stage.
o
463
-------
Table 6. Results of analysis for determination of the yield coefficient,
7, and the decay coefficient, fc,.
Stage Number
Parameter
Yield coefficient, Y 1.00 1.44 1.04 0.81
Decay coefficient, 0.42 0.44 0 0.07
k, (days"1)
Regression coefficient, 0.991 0.990 0.880
* The values for the fourth stage are based on Units C and D only and,
therefore Rz is meaningless.
Table 7. Values for the maximum specific growth rate, 0, and half
saturation constant, K , at 15°C.
s
Parameter Stage Number
Maximum specific growth 2.43 2.92 2.26 1.47
rate, fl, (days'1)
Half saturation constant, 56.6 67.4 61.5 6.0
K, (mg CODA)
Regression coefficient, R2 0.87 0.89 0.93 0.89
464
-------
The Ks values for the final three stages are approximately the same.
The fourth stage Ks value is substantially less than the other three
values. This could be due to the occurrence of nitrification in the
fourth stage. The low fourth stage Ks value also indicates that the
liquid entering the fourth stage is less biodegradable. This is
expected as the easily biodegradable organic compounds are assimilated
in the earlier stages of treatment. These Ks values are slightly
lower than those reported by Clark et al. [1978] who reported Ks values
ranging from 8 mg BOD5/£ to 431 mg BOD5/£. However, the KS values in
this study are similar to those reported by Mikula [1979], who reported
Ks values ranging from 10 to 186 mg CODA.
SUMMARY AND CONCLUSIONS
Little information is available concerning the kinetics of
carbonaceous substrate removal in the rotating biological contactor
(RBC). The general objective of this study was to develop the kinetic
constants describing carbonaceous substrate removal in the RBC process
treating domestic sewage and to determine the effect of temperature on
these constants. This paper details preliminary results of only the
15°C phase of the study. The study is currently underway and the
completed study results will be reported at a later date.
The values for the kinetic constants in a Monod steady state
carbonaceous substrate removal equation have been developed for 15°C
using data collected from four laboratory scale four stage RBC units
treating domestic sewage. The values for the yield coefficient, Y,
ranged from 0.81 to 1.44, while the values for the decay coefficient,
kg, ranged from 0 to 0.44 days"1. The maximum specific growth rate, 0,
varied froml.47to 2.92 days"1 while the half saturation constant, Ks,
ranged from 6.0 to 67.4 mg COD/&.
Additional study will verify the values of these kinetic constants
and also indicate the effect of temperature on substrate removal.
These results will be presented at a later date.
ACKNOWLEDGMENTS
Equipment and financial assistance for this study were provided
by the Environmental Systems Division, George A. Hormel and Company,
Coon Rapids, Minnesota.
REFERENCES
Monod, J. 1942. Recherches SUP la croissance des cultures bacteriennes.
Hermann & Cie. Paris. 211 pp. (original not seen; cited in
Herbert, D., R. Ellsworth, and R. C. Telling. 1956. The
continuous culture of bacteria; a theoretical and experimental
study. Jour. Gen. Microbiol. 14:601-622).
Hartmann, H. 1960. The dipping contact, filter. Oesterreichische
Wasserwirtschaft 11/12:264 (original not seen; cited in Cruz
[1977]).
465
-------
Joost, R. H. 1969. Systematic^ in using the rotating biological
surface (RBS) waste treatment process. Proa. 24th Purdue Ind.
Waste Conf., Purdue Univ.., Lafayette, Indiana, pp. 365-373.
Antonie, R. I., and F. M. Welch. 1969. Preliminary results of a novel
biological process for treating dairy wastes. Proa. 24th Ind.
Waste Conf.3 Purdue Univ. pp. 115-126.
Weng, C.-N., and A. H. Molof. 1974. Nitrification in the biological
fixed-film rotating disk system. Jour. Water Poll. Contr. Fed.
46(7):1674-1685.
GrieveSj C. G. 1972. Dynamic and steady state models for the rotating
biological disk reactor. Ph.D. Dissertation, Clemson University,
252 pp.
Hansford, G. S., J. F. Andrews, C. G. Grieves, and A. D. Carr. 1976.
A steady state model for the rotating biological disc reactor.
Water Res. (G.B.) in press. 55 pp.
Benjes, H. H., Jr. 1978. Small communities wastewater treatment
facilities - biological treatment systems. In: Design Seminar
Handout Small Wastewater Treatment Facilities. USEPA. Technology
Transfer. January 1978. 94 pp.
Kornegay, B. H., and J. F. Andrews. 1968. Kinetics of fixed film
biological reactors. Jour. Water Poll. Contr. Fed. 40(11):R460-
468.
Kornegay, B. H. 1972. Modeling and simulation of fixed film biologi-
cal reactors, p. 257-298. In T. M. Keinath and M. P. Wanielista
(eds.) Mathematical Modeling in Environmental Engineering. Eighth
Annual Workshop, Association of Environmental Engineering Professors,
Nassau, Bahamas, December 18-22.
Kornegay, B. H. 1975. Modeling and simulation of fixed film biologi-
cal reactors for carbonaceous waste treatment, p. 271-318. In
T. M. Keinath and M. P. Wanielista (eds.) Mathematical modeling
for water pollution control processes. Ann Arbor Science
Publishers Inc., Ann Arbor, Michigan.
Clark, J. H., E. M. Moseng, and T. Asano. 1978. Performance of a
rotating biological contactor under varying wastewater flow.
Jour. Water Poll. Contr. Fed. 50(5):896-911.
Schroeder, E. D. 1976. Water and Wastewater Treatment. McGraw-Hill
Inc. New York. pp. 304-309.
Friedman, A. A., R. C. Woods, and R. C. Wilkey. 1976. Kinetic
response of rotating biological contactors. Proa. 31st Ind.
Waste Conf.j Purdue Univ. Ann Arbor Science Publishers, Inc.,
Ann Arbor, Michigan, p. 420-433.
466
-------
Famularo, J., J. A. Mueller, and T. Mulligan. 1978. Application of
mass transfer to rotating biological contactors. Jour. Water
Poll. Contr. Fed. 50(4):653-671.
Oceanography. 1978. Chemical oxygen demand (Standard Ampule Method).
Oceanography International Corporation, College Station, Texas.
APHA. 1975. Standard Methods for the Examination of Water and
Wastewater. 14th Ed. Amer. Pub. Health Assn. New York.
Mlkula, William J. 1979. Performance characteristics and kinetics
of substrate removal in the treatment of a cheese processing
wastewater with a rotating biological contactor (RBC). M. S.
Thesis, Utah State University, Logan, Utah.
Mikula, William J., J. H. Reynolds, D. B. George, D. B. Porcella,
E. J. Middlebrooks. 1980. A kinetic model for treatment of
cheese processing wastewater with a rotating biological contactor.
Presented at First National Symposium on Rotating Biological
Contactor Technology. Feb. 4-6, University of Pittsburgh,
Pittsburgh, Pa.
Reynolds, J. H., E. J. Middlebrooks, D. B. Porcella, and W. J. Grenney.
1975. Effects of temperature on growth constants of Selenastrwn
capricornutum. Jour. Water Poll. Contr. Fed. 47(10):2674-2693.
APPENDIX
Nomenclature
a = Particulate COD to MS
BOD = Biochemical oxygen demand
COD = Chemical oxygen demand
DO = Dissolved oxygen
k^ = Endogeneous respiration
#Q = Half saturation constant
£3
Q = Flow rate
S = Substrate concentration
S0 = Influent substrate concentration
5_ = Effluent substrate concentration
&
t =. Time
V = Stage volume
VS = Volatile solids
X = Biomass amount
Y = Yield coefficient
0 = Maximum specific growth
467
-------
Page Intentionally Blank
-------
FIXED FILM NITRIFICATION SURFACE REACTION KINETICS
AND ITS APPLICATION IN RBC SYSTEMS
By
Ching-San Huang, Ph.D., P.E.
Consultant
USAF Occupational and Environmental Health Laboratory
Brooks AFB, Texas
INTRODUCTION
The waste removal process in a biological fixed film, as in the
trickling filter, or in the rotating biological contactor (RBC), depends
on mass transfer from the wastewater to the slime layer, followed by
metabolism of the waste by microorganisms. The carbonaceous substrate
removal in the fixed film process has been found to be a substrate
diffusion-!imited, heterogeneous model by numerous studies (Ames, et al.»
1962, Atkinson, et al., 1967, Gulerich, et a!., 1968, Maier, 1969, and
Ross, 1970). The nitrification process, however, may differ from the
carbonaceous substrate removal process because the autotrophic nitrifiers
have a much lower growth rate than the heterotrophs and also the ammonia
diffusivity in water is several times higher than the carbonaceous
substrate diffusivity such as glucose (Perry, 1963). Therefore, the
surface reaction model in nitrification process should be investigated,
so that the correct model can be applied in the fixed film nitrification
process design.
This paper includes two stuides: STUDY I used a stationary inclined
pla.te on which a nitrifying biological film was developed for nitrifi-
cation surface reaction model study; STUDY II used a bench scale rotating
biological contactor to check if the findings from the stationary
biological film process are also applicable to the rotating biological
film process.
-------
SURFACE REACTION MODELS IN BIOLOGICAL FIXED FILM PROCESS
Two possibilities exist for the geometry of the interface between the
microorganisms and the liquid phase:
1. The Pseudo-Homogeneous Model: This model assumes the biological
oxidation process takes place throughout the liquid film as if the microbial
population were suspended in the liquid film, and there is no diffusional
resistance to retard the rate of reaction, as shown in Figure 1(A);
2. The Heterogeneous Model: This model assumes that the biochemical
reaction occurs at the interface of the liquid and microbial mass as depicted
in Figure 2(A).
STUDY I: STATIONARY BIOLOGICAL FILM STUDY
1 . Model Development
In order to -describe the surface reaction models mentioned above in a
more comprehensible form, the mathematical models should be developed. To
describe the problem in mathematical terms, some assumptions were made:
a. A steady-state, in the biological sense, exists in that the
thickness, composition and mass of the biological film remains invariant with
respect to time;
b. The laminar flow regime is fully developed; and
c. The liquid film thickness on the inclined plate is defined by
Nusselt's equation:
sine
'
in which s = thickness of liquid film, y = viscosity of liquid,
Q = hydraluic loading, p = liquid density, g = acceleration due to gravity,
8 = angle of the plane with the horizontal .
(1) The Pseudo-Homogeneous Model —
If the liquid phase diffusional resistance is neglected, a
mass balance applied to a length, Ay, of unit width, as in Figure 1(B), yields:
Q dS = rsdy ........... . .................................. (2)
in which Q = hydraulic loading, S = substrate concentration, r - reaction
rate, and 6 - liquid film thickness.
The reaction rate in the nitrification process on a fixed
film reactor was . found to be zero-order by Huang, et al . (1974a), That is,
r = -k ..; ................................................ (3)
470
-------
SLIME LAYER
GROWTH
SURFACE
LIQUID FILM
(A) , POSSIBLE GEOMETRY FOR THE INTERFACE
BETWEEN MICROORGANISMS AND LIQUID
(B), SCHEMATIC DIAGRAM
FIGURE 1 - PSEim-HOmGENEOUS MODEL OF THE BIOLOGICAL FILM
471
-------
SLIME LAYER
GROWTH H
SURFACE
LIQUID FILM
o
T
(A) , POSSIBLE GEOMETRY FOR .-ME INTERFACE
BETWEEN MICROORGANISMS AND LIQUID
(B), SCHEMATIC DIAGRAM
FIGURE 2 - OTEL.OF THE FILM
472
-------
in which k = reaction rate constant, and Equation (2) becomes:
Q dS = -k6dy ............................... ............... (4)
Integration Equation (4) and substituting Equation (1) gives:
S0 - Se= 0/Re)2/3- (Rv- Pa)1/3 ................. ......(5)
in which S_ and Se = influent and effluent substrate concentration,
respectively;
Reaction Number, Rv = g^sin1^ ; Packing Number, Pa = l* P* 9 S1'n8 .
Reynolds Number, Re = ^ (Fulford, 1964); and L = length of biologicaWeaetor.
For the pseudo-homogeneous film flow reactors, it has been
found that recirculation has a beneficial effect on the removal of total
soluble substrate (Kehrberger, et a!., 1969, and Atkinson, et a!., 1963).
(2) The Heterogeneous Model —
Under steady-state conditions an equilibrium situation must
exist such that the rate of transfer of substrate from the main body of the
liquid film to the bacterial phase will be balanced by the rate of removal
and degradation by the organisms. Thus the overall rate will be determined
by the relative resistances due to diffusion and removal .
The heterogeneous model can be broken into other sub! eve! s;
(a) A reaction controlled situation- if the substrate
removal rate at the reaction site is much slower than the substrate diffusional
rate to the reaction site, the reaction rate will control the surface reaction
model ;
(b) A diffusion controlled situation- if the diffusion rate
is so low that the substrate diffused into the reaction site is practically
consumed immediately, the diffusional rate will control the surface reaction
model ;
(c) A diffusion limited situation- this is the condition
which exists in between the reaction controlled situation and the diffusion
controlled situation and so both the reaction rate and the diffusional rate
are controlling.
The nitrification rate on a fixed film reactor was
found to be a zero-order reaction (Huang, et a!., 1974a). That is, there is
always an excess amount of substrate at the reaction surface to allow for
maximum reaction rate. The condition in which there is always an excess
amount of substrate at the reaction site indicates that neither a substrate
473
-------
diffusion controlled situation nor a diffusion limited situation can exist in
the fixed film nitrification process. For this reason, only the reaction
controlled situation warrants a further investigation.
Under the situation that the substrate diffusional. rate
will not control the reaction rate, a differential mass balance applied to a
length &y in a unit width, as in Figure 1(B), can be described as follows:
QdS=rdy .............. ... ..................... ..-(6)
Substituting Equation (3) into Equation (6) and integrating and substituting
previously defined numbers, Equation (6) becomes:
S0 - Se = (l/Re) • (Rv. Pa)1/3 ............... . ..... (7)
Equation (5) and Equation (7) can be placed in the
'
S0 - Se = K- (l/Re)M» or
general form:
CRe2/Rel)M ....... .......... ..- ......... (8)
in which K = (Ry • Pa) » M = exponential constant, S0 = influent substrate
concentration, Se-| and Se2 = effluent substrate concentration in the first
hydraulic loading and the second hydraulic loading, respectively; Rei and
Re2 = Reynolds Number in the first hydraulic loading and the second hydraulic
loading, respectively.
From Equation (5), the M value for a pseudo-homogeneous
model is 2/3, and from Equation (7), the M value for a reaction controlled
heterogeneous model is 1 .
2. Method and Experimental Results
A stationary inclined plate surface, 4-in. wide and 36-in. long,
provided a support surface for the slime growth. The surface of the plate
was covered with a fiberglass screen which served as a structural framework
for slime growth. This reactor was supported by angle frames with an inclined
slope of 10° with the horizontal. Feed was introduced into a stilling basin,
where it passed over a precise mechanical overflow weir. A schematic view of
the physical model is shown in Figure 3. This kind of biological film reactor
has been used by many researchers for the study of biological behavior in
wastewater treatment processes.
A synthetic wastewater was used for this study. A summary of the
constituents of the synthetic wastewater is as follows:
474
-------
.G
V—«
PG
u_
475
-------
varied
NaHC03 300 mg/1
FeCl3 3.5 mg/1
CaClg 1 ml/1*
MgS04 1 ml/I*
Phosphate Buffer Solution 1 ml/1*
Na2Mo04- 2H20 0.5 mg/1
Glucose 10 mg/1
Tap water 100 ml/I
Distilled water . To make-up 1 liter
*Note: Standard BOD dilution water nutrient (Standard Methods, 1975)
A boric acid-sodium hydroxide buffer solution was used to maintain a pH of
8.5 ± 0.1, for the optium pH for nitrifiers is in this range (Huang, et a!.,
1974E).
The analytical methods in this study followed Standard Methods (1975).
Four forms of nitrogen constituents, i.e., NHg-N, NO£-N NQo-N, and organic-N
were analyzed. The organic nitrogen content of the synthetic water was nil,
and the recovery of all other nitrogen forms was good. Since the
ammonia-nitrogen oxidation controls the nitrification rate, the
ammonia-nitrogen analysis was the primary measurement of nitrification
efficiency.
Surface Reation Model Study—
According to Equation (8), the M values can be obtained by varying
the Reynolds Numbers, i.e., by varying the hydraulic loadings. By using two
sets of inclined fixed film reactors and reproducing the experiment at
different times, the M values were calculated for the data obtained and
plotted in Figure 4. Figure 4 reveals that the surface reaction model of
nitrification in a biological film is pseudo-homogeneous.
• As indicated previously, recirculation of the effluent has a
beneficial effect on removal of total soluble substrate in a pseudo-homogeneous
film flow reactor; therefore, a side study for recirculation effect was also
performed. The results for recirculation ratios of 0.5, 1.0, 1.5, and 2.0 all
showed that recirculation improved the nitrification in the fixed film reactor
as shown in Figrue 5. Improving the efficiency of the system by recirculation
also supports the conclusion that the correct surface reaction model for the
nitrification in the fixed film process is of the pseudo-homogeneous type.
The design equation for a fixed film nitrification process can be
derived by putting Equation (5) in another form:
S0 - Se = ak/(Q)2/3 (9)
in which a = constant reflecting the characteristics of the packing media.
Equation (9) indicates that the nitrification efficiency in a fixed film
476
-------
0)
CO
eg
01
CO
1,0
0,8
0,6
O.if
0,2
0,0
* FLOW RATE: 19 TO 110 ML/MIN
0 flow RATE: 17 TO 97 ML/MIN
A FLOW RATE: 52 TO 220 ML/MIN
0,0 0,1 0,2 0,3 '0,4 0,5 0,6 0,7
(Re)2/(Re)i
0,8 0,9 1,0
FIGUE 4 - REYNOLDS NUMBER RESPONSE STUDIES
-------
00
INFLUENT RATE(!) = 30 ML/MIN
TEMPERATURE = 22 C
pH = 8,4
0
RECIRCULATION RATioCR/I)
FIGUE 5 - ECIRCULATId EFFECT ON THE NITRIFICATION RATE
-------
process depends on the characteristics of the packing media used and the
reciprocal of the two-third power of the hydraulic loading. Therefore, the
hydraulic loading and the packing media should be properly selected during
the treatment plant design,
STUDY II: ROTATING BIOLOGICAL CONTACTOR STUDY
Conceptually, RBC units are similar to other kinds of fixed film
biological treatment systems such as trickling filters. However, rotation
of the RBC media provides a more positive supply of oxygen and nutrients to
the bacteria than trickling filters.
In organic substrate removal studies of RBC systems, an oxygen-
substrate diffusion limited situation are often assumed for organic substrate
removal kinetics derivation (Famularo, et al., 1978, Hansford, et a!., 1978,
and Schroeder, 1976), In RBC nitrification process kinetics, however, much
less work has been done. The pseudo-homogeneous model found from the
stationary fixed film process, as described in STUDY I, could be applied to
the RBC system.
1. Mass Balance in RBC Unit
If the pseudo-homogeneous surface reaction model is applied to an/
RBC system (see Figure 6(A)), the mass nitrification rate per disc face is
then
Mz = rsN AT [[[ (10)
In which Mz = mass nitrification rate per unit time; r = reaction rate;
Sfij = liquid film thickness on the slime layer at rotating speed N rpm; and
A-J = .contact surface area per disc face.
Because the discs are closely spaced, a continuous nitrification rate
function can be made which is analogous to a plug-flow reaction through a
whole shaft of the RBC unit. Therefore,
in which V-j = liquid volume per disc face; S = NH3-N concentration; and
t - contact time.
The RBC nitrification reaction order found to follow a zero-order reaction
by several studies (Murphy, et al. , 1973, Torpey, et al . , 1973, and Weng, et
al., 1974) and by this study, which will be described later. That is,
r « -k ...... . ...................... .................. . ....... .....(3)
The liquid film thickness, SM, can be defined as follows (Bintanja, et al . ,
1975):
-------
SHAFT
N
, &
£0
J.IQUID FILM(<£ )
^BACTERIA SLIME
-Disc
H RPM
(A) SCHEMATIC CROSS-SECTIONAL VIEW OF AN RBC Disc
VARI-SPEED
MOTOR
INFLUENT
RBC UNIT
SETTLING
TANK
EFFLUENT
IWSTE SLUDGE
(B) FLOW DIAGRAM OF THE RBC TEST UNIT
6 - FLOW OF AN RBC
480
-------
in which 5|\j = liquid film thickness at rotating speed N rpm (pm);
K] = constant; y = viscosity of wastewater (kg mass)/(m)(sec); N = rotating
speed (rpm); R = radius of disc at the average tangential velocity point (m);
p = density of wastewater (kg/m3); and g = acceleration due to
gravity (m/sec2).
The constant K-| calculated by Bintanja, et al. (1975) is 0.93.
At a certain wastewater temperature, and a certain disc size,
Equation (12) can be expressed as
5N = K2- N1/2 .........(13)
where K2 is a constant.
Substituting Equation (3), (10), and (13) into Equation (11), and
integrating over the entire RBC unit, yields
V (S0 - Se) = K- -N1/2 A t (14)
in which V = liquid volume in RBC unit; S0, Se = influent and effluent
NH3-N concentration, respectively; K = k K2 = reaction rate constant;
A = total contact surface area; and t = contact time.
Since the contact time is close to the hydraulic retention time,
t = V/Q (15)
where Q is the flow rate. Substituting Equation (15) into Equation (14),
yields
- 9^ 119
j—sfil- K N17^ .,
Equation (16) indicates that NH3-N mass removal rate per unit area is a
function of the square root of the rotating speed.
2. Method and Experimental Results
A bench scale RBC unit which consists of 3-3/4 in. diameter discs
was used for this study. The flow direction was parallel to the disc shaft.
The schematic flow diagram is shown in Figure 6(8). Recycling was employed
only during the Recirculation Study mode. The synthetic wastewater used was
similar to the one used in STUDY I. The analytical procedures also followed
Standard Methods (1975).
a. Reaction Rate Study—
The reaction rate study was performed by feeding the matured RBC
unit with different initial NF^-N concentrations at a constant flow rate and
a constant rotating speed. The results for pH = 8.5 and pH = 7.2 are
plotted in Figure 7. The test for pH = 7.2 was done in few days because the
nitrifying bacteria will acclimate to a lower pH as shown in Figure p
(Huang, et al., 1974b).
481
-------
Q
o °o°
16
PH8.5
Q
112
^10
8 L - FH7.2
0
16
1 12
10
0 1 23
RECIRCUUTION RATIO
FIGUE 8 -'^CIRCULATION RATIO VS',' N%-N
0 20 i{0 60 80 100 120
INITIAL NHj-N CONCENTRATION CHG/L)
7- INITIAL MHj-N VS, NHj-N RBWAL
482
-------
8 &
CD
LO
(%) S'8
CD
i iy aivy
oo
CxJ
CM
CD
CXI
or:
^
a
fe
5
j—«
ts
€/)
- CXJ
OO
td
UD
oun
!=£
CD
I
en
U_
483
-------
Figure 7 indicates that the NH3-N removal process follows a zero-order
reaction.
b. Recirculation Study—
A recirculation study was performed by recycling the underflow
from the settling tank. The underflow MLSS concentration was approximately
600 mg/1 . The results are shown in Figure 8. It can be seen from Figure 8
that the NHo-N removal rate increased up to 16 percent at a recirculation ratio
of 0.9.
c. Rotating Speed Study—
During this mode of study, the influent
hydraulic loadings were kept in four groups:
Sroup Flow Rate (ml/min)
I 12 +_2
II 20+2
III 30+2
IV 35+2
-N concentrations and
Initial NHg-N Cone, (mg/1)
33 +_2
43+2 and 52+3
47 + 3
54 + 4
Rotating speeds were varied from 6.5 rpm to 128 rpm. Temperature was room
temperature and wastewater pH was 8.5 +0.2. The results are plotted in
Figure 10.
Figure 10 reveals that the NI^-N mass removal rate per unit area
increases proportional to the square root of the rotating speed up to a
point (Point B in Figure 10), and then levels off. The rotating speed at
Point B is approximately 75 rpm, which corresponds to a peripheral velocity
of 1.23 ft/sec. This indicates that the pseudo-homogeneous model is also
valid for the RBC unit used in this study with the rotating speeds from
6.5 rpm to 75 rpm.
The liquid film thickness at 75 rpm is 72 ym as calculated from
Equation (12). According to the conceptive geometry of the
pseudo-homogeneous model (See Figure 1(B)), a liquid film thickness of 72 pm
may be close to the "effective" slime thickness. And this is probably why
the NHq-N mass removal rate levelled-off at a rotating speed of 75 rpm in
this study.
DISCUSSION
According to STUDY I and STUDY II, the pseudo-homogeneous surface
reaction model derived from a stationary inclined nitrifying fixed film also
applies to an RBC unit. That is, the NH3-N mass removal rate per unit area
in proportional to the 0.5 power of the rotating speed up to 75 rpm, or
equivalent to a peripheral velocity of 1.23 ft/sec in the bench scale RBC
unit used. The NH3-N mass removal rate per unit area then levelled off at
a higher rotating speed.
Weng, et al. (1974) used a 6-in. diameter RBC unit for their
484
-------
1 1
1 I
CM
oo
CD
to
CD
LT»
cn
oo
to
ir\
,OJ
CD
W "
_m
C/3
CD
fe
C3
oo
I
CD
485
-------
nitrification study. The rotating speeds tested were from 10.5 rpm to 42 rpm,
or a peripheral velocity of from 0.27 ft/sec to 1.10 ft/sec. Their results
indicated that the Nt^-N removal is a function of the 0.53 power of the rotating
speed.
The liquid film thickness on an RBC disc is a function of the square root
of the rotating speed as shown in Equation (12). That is, the liquid film
thickness increases if the EEC rotating speed increases. According to the
conceptive geometry of a pseudo-homogeneous model as shown in Figure 1(B),
increasing the liquid film thickness will also increase the NH3-N removal until
the liquid film thickness approaches to the "effective" slime thickness.
Therefore, the NH3-N removal level-off point on Figure 10 may be used to
estimate the "effective" slime thickness on a nitrifying RBC unit. The RBC
'unit used in this study has an "effective" slime thickness of 72 yra estimated
from this method. The thickness of the effective microbial film in carbonaceous
substrate removal system has been reported with the range from 70 ym to 200 \im
(Kornegay & Andrews, 1969, and Tomlinson & Snaddon, 1966).
According to one RBC manufacturer (Autotrol Corp., 1978), the nitrifica-
tion RBC unit design should be based on hydraulic loading and the desired
effluent KHj-N concentration. The peripheral velocity is recommended at 1.0
ft/sec. This peripheral velocity corresponds to 1.6 rpm in a 12-ft diameter
RBC unit. According to this study, the maximum NHj-N removal will be achieved
at a peripheral velocity of 1.23 ft/sec, or equivalent to 2.0 rpm in a 12-ft
diameter RBC unit. Of course, the differences in RBC configurations and in
wastewater characteristics may show a different optimum rotating speed for
NH3~N removal. Also, both the peripheral velocity and the rotating speed
should be used for pilot study scaleup factors as reported by Friedman, et al.
(1979). Therefore, a pilot plant test may be required for a nitrification RBC
plant design. The NHj—N mass removal rate and power consumption data at
different rotating speeds should be obtained to evaluate the economic tradeoffs
so that an optimum system can be designed.
CONCLUSIONS
Based on the results of these two studies, some conclusions can be drawn
as follows:
1. The nitrification process follows a zero-order reaction both in
stationary fixed film system and in rotating fixed film system.
2. The pseudo-homogeneous surface reaction model can be applied to both
trickling filter systems and RBC systems. That is, the nitrification efficiency
in a trickling filter process depends on the characteristics of the packing
media used and the reciprocal of the two—third power of the hydraulic loading;
in RBC nitrification system, NE^-N removal is a function of the square root of
the rotating speed up to a certain speed.
3. The active microbial film in a nitrifying RBC unit may be estimated by
locating the optimum NH3~N removal rotating speed and the corresponding liquid
film thickness at that rotating speed. The RBC unit used in this study has an
active (or effective) slime thickness of approximately 72 ym.
486
-------
ACKNOWLEDGEMENT
The author would like to express sincere appreciation to:
Maj Gary A. Fishburn for his continuous encouragement and advice to this
study;
Mr Mark A. Willis for his gathering valuable literature and information;
SSgt Gene D. Jenkins for his laboratory technical assistance;
Capt John H. Rentier and Capt Elliot K. Ng for their fruitful comments and
suggestions; and
USAF Occupational and Environmental Health Laboratory for the support of .
this study.
487
-------
REFERENCES
Ames, W.F., Collins, W.Z., and Behn, V.C. (1962), "Diffusion and Reaction in
the Biota of the Trickling filter," Technical Report No. 19, Department of
Mechanical Engineering, University of Delaware, Newark.
Atkinson, B., Bush, A.W., and Dankins, G.S. (1963), "Recirculation, Reaction
Kinetics, and Effluent Quality in a Trickling Filter Flow Model," Jour. Wat.
Pol. Cont. Fed., 35:10:1307.
Atkinson, B., et al. (1967), "Kinetics, Mass Transfer, and Organism Growth in a
Biological Film Reactor," Trans. Instn. Chem, Engrs., 45:6:T257.
Autotrol Corp. (1978), Autotrol Wastewater Treatment System DESIGN MANUAL,
Milwaukee, Wisconsin. ,
Bintanja, H.H., van der Erve, J.J.V.M., and Boelhouwer, C. (1975), "Oxygen
Transfer in a Biological Disc Treatment Plant," Wat. Res., 9:12:1147.
Famularo, J., Mueller, J.A., and Mulligan, T. (1978), "Application of Mass
transfer to Rotating Biological Contactors," Jour. Wat. Pol. Cont. Fed.,
50:4:653.
Friedman, A.A., Robbins, I.E., and Woods, R.C. (1979), "Effect of Disk
Rotational Speed on Biological Contactor Efficiency," Jour. Wat. Pol. Cont.
Fed., 51:11:2678.
Fulford, G.D. (1964), "The Flow of Liquids in Thin Films," in Advances in
Chemical Engineering, Vol. 5, Edited by Drew, T.B., Hoopes, J.W., Jr., and
Vermenlen, T., Academic Press, New York, p. 151.
Gulerick, W., Renn, C.E., and Liebman, J.C. (1968), "Role of Diffusion in
Biological Waste Treatment," Environ. Sci. & Tech., 2:3:113.
Hansford, G.S., et al.'(1978), "A Steady-State Model for the Rotating Biological
Disc Reactor," Wat. Res., 12:10:855.
Huang, C.S., and Hopson, N.E. (1974a), "Nitrification Rate in Biological
Processes," Jour. Environ. Engg. Div., ASCE, 100:EE2:409.
Huang, C.S., and Hopson, N.E. (1974b), "Temperature and pH Effect on the
Biological Nitrification," Presented at the Annual Winter Meeting, New York
Wat. Pol. Cont. Assoc., New York City.
Kehrberger, G.J., and Bush, A.W. (1969), "The Effects of Recirculation on the
Performance of Trickling Filter Models," Proc. of 24th Purdue Indus. Waste
Conf., 37-52.
Kornegay, B.H., and Andrews, J.F. (1969), "Application of Continuous Culture
Theory to the Trickling Filter Process," Pore, of 24th Purdue Indus. Waste
Conf., 1398-1425.
488
-------
Maier, W.J. (1969), "Mass Transfer and Growth Kinetics on a Slime Layer, A
Simulation of the Trickling Filter," Ph.D. Thesis, Cornell Univ.
Murphy, K.L., et al. (1977), "Nitrogen Control:Design Consideration for Sup-
ported Growth Systems," Journ. Wat. Pol. Cont. Fed., 49:4:549.
Perry, J.H. (1963), Chemical Engineers Handbook, 4th-ed, McGraw-Hill Book Co.,
New York.
Ross, L.W. (1970), "Contribution to the Theory of Trickling Filter Performance,"
Wat. Res., 4:7:517.
Schroeder, E.D. (1976), Water and Wastewater Treatment, McGraw-Hill Book Co.,
New York.
Standard Methods for the Examination of Water and Wastewater (1975), APHA,
AWWA, WPCF, American Public Health Assoc, 14th-ed.
Tomlinson, T.G.S and Snaddon, H.M. (1966), "Biological Oxidation of Sewage by
Film of Micro-organisms," Int. Jour. Air Wat. Pol., 10:11-12:865.
Torpey, W., et al. (1973), "Effects of Exposing Slims on Rotating Discs to an
Atmospheres Enriched with Oxygen," in Adv. in Wat. Pol. Res., Proc. of 6th Int.
Conf. held in Jerusalem, June 18-23, 1972, Pergamon Press, London.
Weng, C.N., and Molof, A.L. (1974), "Nitrification in the Biological Fixed-Film
Rotating Disk System," Jour. Wat. Pol. Cont. Fed., 46:7:1674.
489
-------
Page Intentionally Blank
-------
A KINETIC MODEL FOR TREATMENT OF
CHEESE PROCESSING WASTEWATER WITH A ROTATING
BIOLOGICAL CONTACTOR
by
William J. Mikula
Engineer, O'Brien & Gere Engineers, Inc.
707 Westchester Ave., Suite 211
White Plains, New York, U. S. A.
James H. Reynolds
Assistant Professor, Civil and Environmental Engineering
Utah State University, Logan, Utah, U. S. A.
Dennis B. George
Assistant Professor, Civil and Environmental Engineering
Utah State University, Logan, Utah, U. S. A.
Donald B. Porcella
Environmental Biologist
Tetra Tech Inc., 3700 Mt. Diablo Blvd.
Lafayette, California, U. S. A.
E. Joe Middlebrooks
Dean, College of Engineering
Utah State University, Logan, Utah, U. S. A.
INTRODUCTION
Rotating Biological Contactors (RBC) have become established as
competitive, cost-effective systems for the treatment of biodegradable
wastewaters. These units are being used to treat many different types
of wastewaters in a multitude of. treatment process configurations.
Although RBC's are used in many industrial and domestic wastewater
treatment schemes, there is incomplete information on the performance
491
-------
capabilities of the RBC. Specifically, the mechanisms of substrate
removal have not been fully postulated and verified.
The general objective of this study was to develop a kinetic model
which describes the performance of a rotating biological contactor
treating cheese processing plant wastewater. The model employs
Michaelis-Menton-Monod kinetics and steady state conditions to describe
the system performance as a function of temperature and organic loading
rate.
To accomplish the above general objectives the following specific
objectives were achieved using a four stage, four foot diameter, pilot
plant rotating biological contactor (RBC) treating cheese processing
plant wastewater.
1. Determine the pilot plant performance at three separate organic
loading rates (hydraulic loading rates) under steady state condi-
tions.
2. Develop a Michaelis-Menton-Monod equation which describes the
performance of the pilot plant on a stage by stage basis.
3. Determine the kinetic constants associated with the model by
stage and with temperature, using the pilot plant performance data.
PREVIOUS INVESTIGATIONS
Initial attempts to model the performance of rotating biological
contactors (RBC) were empirical in nature and mainly employed regression
analysis \Hartman 1965, Jost 1969, Antonie and Welch 1969, and Weng
and Molof 1974]. Their efforts generally ignored temperature effects
and were not directly related to microbial substrate removal. Substrate
kinetic removal equations have been developed by several investigators
[Grieves 1972, Hans ford et al. 1976, Benjes 1978, Komegay and Andrews
1968, Komegay 1972, Komegay 1975, and Clark et al. 1978]. In general
these models employed either saturation kinetics [Monod 1942] or first
order kinetics to describe substrate removal. Usually these equations
were limited to a single stage system or treated a multi-stage system
as a single unit.
Recent investigations have employed either mass transfer models
or have combined mass transfer concepts with substrate removal kinetics
to describe RBC performance \_Schroeder 1976, Friedman et al. 1976, and
Famularo et al. 1978]. These equations have generally been applied in
oxygen limited substrate removal situations.
MATERIALS AND METHODS
Wastewater Source
The Cache Valley Dairy Association Plant in Amalga, Utah, served
as the site for the research project. Monterey jack, swiss, and cheddar
cheese are the principal products of the plant. Approximately 430
farms contribute a total of 327,000 kg/day (720,000 Ib/day) of raw
492
-------
milk for the cheese making process.
All of the process wastes from the plant are discharged to an
aerated primary lagoon followed by several facultative lagoons. A
portion of the final lagoon effluent is recycled back to the plant
for use as non-contact cooling water which, after use, is returned to
the wastewater treatment facility.
Experimental Apparatus
Wastewater was pumped from the headworks of the treatment facility
to a 15.2 cm (6 inch) wide, pilot-scale Hydrasieve screen (Bauer
Brothers Division, Combustion Engineering). The Hydrasieve was
equipped with a 1.0 mm (.04 inch) screen size opening during the two
lower hydraulic loading rates (HLR) in the experimental procedure. At
the highest loading rate, a 1.5 mm (.06 inch) screen size opening was
utilized. This system removed most of the suspended solids from the
wastewater (see Figure 1).
The rotating biological contactor (RBC) (Environmental Systems
Division, Geo. A. Hormel & Company) was a four stage, 1.19 m (47 inch)
diameter, pilot-scale system. The disc media was type SC (sinusoidal
circular) extruded polyethylene. Each stage contained 12 discs, which
yielded a total system area of 145 m2 (1680 ft2). Flow through the
stages was perpendicular to the shaft, or parallel to the disc faces.
Flow to successive stages was allowed via externally mounted 5.1 cm
(2 inch) pipe connections. The discs were kept at approximately 37
percent submergence by adjusting the fourth stage outlet piping.
Figure 1. Schematic drawing of the experimental apparatus.
493
-------
Because of the placement and hydraulics of the RBC unit, the first
stage had a liquid depth of 44.4 cm (17.5 inches) while stages two
through four had a liquid depth of 45.7 cm (18 inches). The dimensions
of the RBC are summarized in Table 1.
The influent and effluent from each stage was monitored on a
weekly basis using 24-hour composite samples with analyses being
performed for suspended solids, volatile suspended solids, total
chemical oxygen demand, soluble chemical oxygen demand, and nutrients.
All analyses were conducted according to Standard Methods \_APHA 1975].
In addition, in-situ measurements of flow, temperature, pH, and
dissolved oxygen were conducted.
Table 1. Summary of the pilot-scale RBC dimensions.
Parameter
Value
Units
Number of stages
Number of discs/stage
Area/disc
Area/stage
Total area
Net first stage volume
Remaining stage volumes (net)
Specific surface area
Rotation speed
Linear velocity
4
12
3.25 (35)
39 (420)
156
.113 (3.99)
.123 (4.34)
119.4 (36.4)
6
22,5 (73.9)
m2(ft2)
m2(ft2)
m2 ft2)
m3(ft3)
m3(ft3)
m2/m3(ft2/ft3)
rpm
m/min (ft/min)
Experimental Procedure
Full-scale testing began in January, 1978, and was completed at
the end of June, 1978. The system was operated at three different
hydraulic loading rates (HLR) as shown in Table 2. Rotation speed
was kept constant at 6 rpm throughout the experiment.
Table 2. Summary of the experimental hydraulic loading rates.
Loading
Number
1
2
3
Mean daily
flow rate
m£/min (gpm)
850 (.22)
1250 (.33)
3300 (.87)
Mean hydraulic
loading rate
m3/m2day (gpd/ft2)
0.008 (.19)
0.011 (.28)
0.030 (.75)
Hydraulic
Mean daily retention
flow H (gal) time (hrs)
1225 (320) 9.5
1880 (475) 6.5
4750 (1250) 2.41
1 Actual measured retention time. Retention times for the first two
loading rates were interpolated.
494
-------
Each HLR was maintained until steady-state conditions were
reached. The criterion defining steady-state was a relatively constant
effluent soluble COD concentration, i.e., three consecutive weeks
within ±10 percent of a mean concentration. Once steady-state was
reached, the system was switched to the next HLR and operated until
steady-state was again reached. In this fashion, data were collected
for three complete and distinct loading rates.
MODEL DEVELOPMENT
Derivation
The model development is similar to that employed by Komegay
[1972] and employs the following assumptions:
1. The liquid portion of each stage is completely mixed.
2. The growth of attached microorganisms is limited by the concentra-
tion of a particular nutrient.
3. Organism decay is negligible as compared with microorganism
growth.
4. Substrate removal is a saturation function with respect to
substrate applied, i.e., Monod kinetics apply.
5. Substrate removal occurs both within the biofilm of the attached
microorganisms and in the liquid portion of the RBC.
6. Substrate removal for microorganism maintenance is negligible
compared with substrate removal for growth.
7. The attached microorganism growth on the entire disc area is
included.
8. Substrate removal in the biofilm occurs through the entire depth
of the biofilm,
9. Suspended microorganism growth contributes to substrate removal.
10. Suspended biomass consists entirely of sloughed attached biomass,
and there is no growth of suspended microorganisms.
11. The stage total suspended solids concentration consists entirely
of suspended biomass.
The model development in this study diverges from Komegay [1972]
due to differing assumptions. The first difference is in the use of
the entire surface area of the discs in computing the mass of attached
microorganisms. In this study, the biofilm developed uniformly across
the entire surface area of the discs; therefore, the entire disc area
served as a support medium for the microorganisms. Komegay assumed
that growth was confined to the wetted submerged segment of the discs.
The second difference is in the use of the entire depth of biofilm
in computing the active mass of attached microorganisms. Here it was
assumed that the entire depth of biofilm contributed to substrate
removal, and thus the total mass of attached growth (dry weight mass)
consisted of active microorganisms. Komegay differentiated between
total and active biofilm depth, and assumed that substrate removal
occurred only in an active portion of the total depth. The remainder
of the biofilm was assumed to be diffusion limited.
495
-------
Thirdly, it is assumed that the suspended organisms contribute to
substrate removal, whereas Koimegay assumed that substrate removal
was confined solely to the active layer of the biofilm.
It is also assumed that the suspended biomass consists entirely
of sloughed attached biomass, and there is no growth of suspended micro-
organisms. Furthermore, it is assumed that the stage total suspended
solids (TSS) concentrations consist entirely of suspended biomass.
The total active mass of microorganisms, then, in each stage (x\)
equals the total attached biomass per stage plus the suspended biomass
(TSS) in each stage. That is,
(1)
where
A = total area of disc (L2)
w
X~ = total attached biomass concentration (M/L2)
V = liquid volume of the reactor (L3)
X = stage TSS concentration (M/L3)
e»
The model was derived in a fashion analogous to suspended growth
model derivations. Performing a substrate mass balance on a single
stage yields:
/ net mass rate of \_ /mass input] [mass removal\ /mass output\
^change of substrate/ \ rate / \ rate /" \ rate J
which may be expressed mathematically as:
pl$l (3)
)1
J
growth
where
V = liquid volume of the reactor (L3)
F0 = influent flow rate (L3/T)
F! = effluent flow rate (L3/T)
So = influent concentration of the growth- limiting substrate
(M/L3)
Si = effluent concentration of the growth-limiting substrate
(M/L3)
t - time (r)
The substrate mass removal rate is represented by microbial growth.
Since Vt the reactor liquid volume, is a constant, it can be removed
from the differentiation. Therefore, equation (3) becomes:
496
-------
net
VfdS;
growth
(4)
Employing the concepts for microorganism growth and substrate
removal developed by Monod [1942], assuming steady state conditions
and utilizing Equation 1, Equation 4 can be reduced to:
V
\dt
X + VX
(5)
net
I
where
= maximum specific growth rate of the microorganisms (T~l)
Y
K
half saturation constant (Af/L3); substrate concentration
where y = V20
Here K and y represent the kinetic constants for the total biomass
population, i.e., the attached and the suspended biomass. Since the
suspended biomass is assumed to represent only sloughed attached
biomass, the suspended biomass specific growth rate y is assumed to
equal y for the attached biomass. Bintanja et at. [1976] made the
same assumption in developing a steady-state model to predict carbon-
aceous removals in an RBC, and achieved excellent fit of the data.
The growth yield lisa measure of the mass of microorganisms
produced per mass of substrate removed. In this study it is defined
as:
TSS
n
y =
n
(Asoluble COD)
(6)
n
where
n
= stage number.
Although many authors use the volatile suspended solids (VSS) concen-
tration to indicate the viable fraction of the mass of microorganisms
produced, in this study TSS concentration was used.
show a definite temperature dependency.
higher temperatures cause more energetic
between molecules, and thus increase the
This temperature dependency is usually
Reaction rate constants
This is due to the fact that
and more frequent collisions
chance of chemical reaction.
represented by an Arrhenius-type relationship [Bailey arid Oil-Is 1977],
Similarly, growth rates of bacterial cultures are proportional to
temperature [Stanier et al. 1975]. Thus the temperature dependency
of the maximum specific growth rate can be incorporated into an
Arrhenius-type relationship:
497
-------
-E /RT
0 = Ae a (7)
where
A = frequency factor (T~l)
E - activation energy (cal/mol)
R = universal gas constant (1.987 cal/mol°K)
T = absolute temperature (°K)
Substituting Equation 7 into Equation 5 results in Equation 8
which describes substrate removal as a function of temperature.
- *) =
-E /RT
Ae a (A X, +
M..../
I
Si
K
s
(8)
In this study it is assumed that the system parameters in Equation (8)
all change with stage. Wu and Kao [1976] and Clark et at. [1978] found
that the kinetic constants for a two stage suspended growth system, and
a four-stage RBC, respectively, all changed with stage. Komegay [1972]
assumed that the kinetic constants did not vary with stage; however,
Kornegay did assume that substrate removal was additive with stage.
The concept of summing the substrate removal contributions of each
stage can be utilized to describe RBC stage performance [Komegay 1972].
It follows that Equation (8) can be generalized for a multi-stage
reactor as:
-E /RT.
A.e ai * (A X. + V.X )
n i w fa ^ si / 5
t.
% + si
(9)
where
n = total number of stages
i = number of specific stage
Since F is constant through the stages, it can be removed from the
summation. Equation (9) represents steady-state substrate removal
in a multi-stage reactor, and is the general model used in this study.
Application
The primary intent of the modeling effort in this study is to
obtain the kinetic parameters - A, the frequency factor, EQ, the
activation energy, and Ks, the half-saturation constant - for each of
498
-------
the four stages in the reactor. One solution technique is to input
stage data into a linearized version of Equation (9), and determine
the stage kinetic parameters from the resultant plots. However, since
there are three unknowns, one of the parameters would have to be
assumed in order to use the linearization method.
A more rigorous approach is to use a nonlinear curve-fitting
solution technique to determine the optimum values for the kinetic
parameters. The "NONLIN" FORTRAN program developed by Grenney [1973]
is clearly suited for this purpose. Following the User's Manual
written by Cleave [1978], the stage data from this study is entered
into the NONLIN program along with initial estimates from A, Ea, and
Ks. The NONLIN program uses an iterative finite-difference technique
to converge to the optimum correlation. Thus the "best-fit" values
of A, Ea, and K8 for each stage are determined.
Nonlinear equations used in the NONLIN program must be of the
form y = f(x,z, ...) where y is the dependent variable, and x, 2,
etc., are the independent variables. Rearranging Equation (9) so
that F becomes the single dependent variable, Equation (9) becomes:
-E IRT .
A.e °"i ^ (A Xf + V.X J
n t. w j • i- S-?* / ~
Y. \K + S.
^ \ S£ i.
Utilizing Equation (10) is not as cumbersome a task as it may appear.
This is because the previous stage best-fit solution becomes a constant
in solving for the successive stage best-fit. For example, for the
first stage, Equation (10) is:
(stage 1 term) (11)
Of) "01
where F has been subscripted to indicate the stage in question. For
simplicity, the entire term in the summation is called the "stage 1
term". For the second stage, Equation (10) becomes:
Fi,2 = q 1 a (stage 1 term) + } (stage 2 term) (12)
' 1 — <-> 2 DO'" *3 2
From Equation (11 ),
(stage 1 term) = Fl(s0 - Si) (13)
Therefore, Equation (12) becomes:
499
-------
Here FI is the flow rate predicted from the best-fit of the first
stage data. Similarly, for the third stage, Equation (10) becomes:
23 = F0""!1 + F2°" + ,„ v (stage 3 term) (15)
- - -
,
' '
where Fa is the predicted flow rate contribution of the second stage,
and is found from equation (14):
^ - r, {f^f} = pj^y (stage 2 term) 06)
Therefore ,
F2 = F1>2 - Fi |g;:g;j (17)
For the fourth stage, Equation (10) becomes:
(Sn— Si ^ f"?n — 9o ^ (Sn— S?} 1
E' •—* IP \ 0 1 / i PI \ ^ Q *^ Z / _L TJJ \ u j / _i_ * jc*l"U(^Q A *f*o yen I
M . A « . «« jj» . . i 1- ^* ^ 7 s* f-i \ ' ^ Q I t~t ' f-i \ * * \ ^ l*dMK ^ ww I til /
(18)
where F3 is found in the manner analogous to the above solution for Fa.
In general, for n a 2, Equation (10) becomes:
(stage n term)
1u> I
Equation (19) serves as the general form of the model incorporated
into the NONLIN program for stages 2 through 4. Equation (11) serves
as the simplified form for the first stage.
RESULTS AND DISCUSSION
Steady-State Data
In this study, steady-state was defined as non-time-varying
effluent soluble COD concentration. The criterion defining steady-state
was three consecutive weeks of effluent soluble COD concentrations
with values between ±10 percent of their mean concentration. The
weeks chosen to represent steady-state for each of the three loading
conditions are summarized in Table 3.
The three-week periods chosen represented the best steady-state
approximations for their respective loading conditions. However, the
500
-------
Table 3. Summary of the steady-state parameter values entered into the
summation model (Equation 9).
Hydraulic
loading
rate,
mfc/mi n
850
850
850
1250
1250
1250
3300
3300
3300
850
850
850
1250
1250
1250
3300
3300
3300
850
850
850
1250
1250
1250
3300
3300
3300
850
850
850
1250
1250
1250
3300
3300
3300
Inf 1 uent
Staap soluble
Stage COD, mgA
(So)
1 357.2
389.8
506.8
269.7
502.2
469.0
1263.8
464.6
855.0
2 357 . 2
389.8
506.8
269.7
502.2
469.0
1263.8
464.6
855.0
3 357.2
389.8
506.8
269.7
502.2
469.0
1263.8
464.6
855.0
4 357.2
389.8
506.8
269.7
502.2
469.0
1263.8
464.6
855.0
Stage
effluent
soluble
COD, mgA
(*„)
98.8
101.8
76.8
81.3
81.8
76.9
496.6
330.9
461.7
76.0
78.7
65.3
63.2
63.2
66.0
453.9
264.0
313.5
64.6
69.1
48.0
79.6
184.3
106.0
368.3
337.9
285.9
57.0
61.4
55.6
55.8
39.0
46.7
181.6
126.7
167.2
Stage
temperature
°K
(T)
295.3
295.3
295.3
300.6
300.6
300.6
305.8
305.8
305.8
294.6
294.6
294.6
299.1
299.1
299.1
304.4
304.4
304.4
293.9
293.9
293.9
297.4
297.4
297.4
302.9
302.9
302.9
293.2
293.2
293.2
295.9
295.9
295.9
301.5
301 . 5
301.5
Attached
biomass
g/m2
*f
74.0
74.0
74.0
66.5
66.5
66.5
88.2
88.2
88.2
31.1
31.1
31.1
36.6
36.6
36.6
69.2
69.2
69.2
16.9
16.9
16.9
24.0
24.0
24.0
53.5
53.5
53.5
14.7
14.7
14.7
22.0
22.0
22.0
44.0
44.0
44.0
Suspended
biomass
TSS, mg/Jl
Xs
115.7
365.0
165.3
179.7
636.3
291.1
245.1
206.8
183.8
154.6
219.9
96.3
132.3
339.7
302.0
447.6
437.5
384.8
196.0
137.5
63.5
79.6
184.3
106.0
368.3
337.9
285.9
101.0
131.5
482.7
259.6
302.0
238.7
615.0
684.0
299.7
Growth
yield
1
.45
1.27
.38
.95
1.51
.74
.32
1.55
.47
.51
.61
.21
.56
.71
.73
.53
1.66
.58
.64
.37
.14
.35
.40
.22
.36
.96
.47
.30
.35
1.05
1.21
.54
.53
.34
.76
.16
Flow
rate
mVday
F
.22
.22
.22
.80
.80
.80
4.75
4.75
4.75
.22
.22
.22
.80
.80
1.80
4.75
4.75
4.75
.22
.22
.22
.80
.80
.80
4.75
4.75
4.75
1.22
1.22
1.22
1.80
1.80
1.80
4.75
4.75
4.75
501
-------
criterion defining steady-state was not met by the 1250 and 3300 m£/
min loading rates. For the 1250 m£/min HLR, the upper and lower
boundaries of the three data points deviated 18.2 and -13.5 percent
from the mean, respectively. For the 3300 m£/min HLR, the upper and
lower boundaries deviated 14.6 and -20.0 percent from the mean,
respectively.
The other parameters required to solve the model equation (Equa-
tion 9) were obtained from the data of these nine weeks. The actual
data values of RBC influent soluble COD (S0), stage effluent soluble
COD (sn], stage temperature (r), attached biomass (%f], suspended
biomass (xs), growth yield (I), and flow rate (F) were used in the
model (Table 3). The value entered into the model for the total disc
area per stage (Aw] was 39.0 m2. The stage volumes (v^} used were
.113 m3 for stage one, and .123 m3 for stages 2, 3, and 4.
The equation for growth yield was modified, assuming that the
stages cannot be treated as independent reactors in terms of growth
yield. Thus the growth yield was assumed to be represented by a
summation model, with the growth yield in any particular stage equal
to the total system growth yield minus the particular stage growth
yield. For the first stage, then, the equation was the same, i.e.,
growth yield equalled TSSra/( soluble COD)n. For stages 2 through 4,
J was defined as:
n n-l
I TSSi I TSS^
^= M izi (20)
y (Asoluble COD). I (Asoluble COD).
i=l * i=l ^
where n equals stage number. The resulting values of y predicted by
Equation (20) varied from 0.16 to 1.66.
Model Results
The best-fit solutions for the stage kinetic parameters, subject
to the constraints listed, are shown in Table 4. The results in
Table 4 show that the frequency factor A is essentially the same for
stages 1 and 2, but decreases through stages 3 and 4. The frequency
factor is related to the number of chemical reactions in the biomass,
and thus shows that the number of reactions is biomass concentration
dependent.
The activation energy Ea is a quantitative measure of the energy
input needed to cause a reaction to occur. In terms of microorganism
growth, it is the heat energy required for substrate removal by the
microorganisms. The results of this study show that Ea is independent
of stage, since the values for all four stages are essentially constant,
502
-------
Table 4. Summary of the stage kinetic parameters found using the summation model (Equation 9) with the
NONLIN program.
in
o
w
Constraint
None
Fourth Stage
K fixed at
s
0.1 mg/A
Stages 3 and
4 lumped
together as
one stage
Stage
1
2
3
4
1
2
3
4
1
2
3-4
Frequency
Factor, ,4
day-1
1.5 x 1012
1.4 x 1012
.46 x 1012
,26 x 1012
1.5 x 1012
1.4 x 1012
.46 x 1012
1.6 x 1012
1.5 x 1012
1.4 x 1012
2.5 x TO12
Activation
Energy, E
cal/mola
17780
17850
17810
17790
17780
17850
17810
18510
17780
17850
17800
Maximum
Specific Growth
Half-Saturation Rate, ft, day"1
C°n(mq/£) KS Flow Rate> mjl/m"in R
32
186
59
-36
32
186
59
0.1
32
186
10
850
.109
.084
.027
.015
.109
.084
.027
.027
.109
.084
.145
1250
.185
.132
.039
.020
.185
.132
.039
.036
.185
.132
.202
3300
.307
.223
.068
.034
.307
.223
.068
.064
.307
.223
.349
.8283
.8559
.8404
.7636
.8283
.8559
.8404
.7404
.8283
.8559
.7825
-------
The mean value of Ea for all four stages is 17807.5 cal/mol (17.81
kcal/mol), and the standard deviation is 30.95.
Murphy et al. [1977] found A, the frequency factor, to be 9.45 *
10 x hr"1 (2.27 x 1Q12 day-1), and Ea to be 13900 cal/mol in an RBC
pilot-scale nitrification scheme. Wong-Chong and Loehr [1975], in a
chemostat study of nitrification, found Ea for the ammonia oxidation
step to be between 16.0 and 21.6 kcal/mol, depending on the pH in the
chemostat. The values of Ea for nitrite oxidation were similarly pH
dependent, and ranged from 14.0 to 39.6 kcal/mol.
The values of Ea found in this study are in the range reported by
Wong-Chong and Loehr, and the values of A are of the same order of
magnitude as those reported by Murphy et al. This may indicate that
the activation energy Ea and the frequency factor A are independent of
the type of substrate - carbonaceous versus nitrogenous - used,by the
microorganisms. Corresponding values of Ea and A for carbonaceous
removal systems could not be found in the literature. '
The values for KQ in Table 4 are listed as whole numbers because
of the tremendous variability of these constants in mixed cultures.
The first stage value of Ks indicates an abundance of usable organic
carbon in the substrate. Thus one-half the maximum growth occurs at
the low concentration of 32 rngA. The low Ks value also implies that
the microbial population is well adapted to the carbonaceous substrate,
i.e., assimilation of the compounds in the substrate occurs quite
readily.
The Ks value of 186 mgA in the second stage indicates that the
second stage microbial mass has a reduced concentration of usable
carbonaceous substrate in its environment, and thus maximum growth
requires a much higher "second stage carbonaceous substrate" concentra-
tion. The higher Ks concentration in the second stage may also be
indicative of greater species diversity, with resulting varying
nutritional needs.
The value of 59 mg/& for the third stage Ks indicates that the
third stage biomass is more adapted to its available substrate than
the second stage biomass, and thus a lower concentration of "third
stage substrate" will achieve maximum growth.
The negative K~ in the fourth stage is indicative of model
inadequacy beyond the third stage. This appears to be an artifact
of the summation model and is not to be considered a true representa-
tion of the fourth stage kinetics. The summation model adds a
decreasing increment to Equation (9) with each successive stage. Thus
a point can be reached where the stage data is best-fit by subtracting
it from the previous stage terms. This is what happened at the fourth
stage in this study. The values of Ks found in this study show an
increase from stage 1 to stage 2, followed by a decrease through stages
3 and 4. A similar trend in Ks values was found by Clark et al.
[1978] in a four stage pilot scale RBC treating domestic wastewater.
The KS values found by Clark and coworkers for the four stages were
504
-------
431, 546, 32, and 8 mg/fc, respectively. The Ks values found in the
first two stages were much higher than the corresponding Ks values
found in this study. This was probably due to a number of reasons,
e.g., different substrate, higher loading conditions, and lower
temperatures.
Because Equation (9) yielded a negative Ks in the fourth stage,
it was decided to verify whether or not zero-order kinetics actually
governed carbonaceous substrate removal in the fourth stage. Therefore,
the fourth stage data were entered into the NONLIN program with KS
arbitrarily fixed at 0.1 mg/£. The resulting predictions (Table 4)
for A, Ea, and 0, the maximum specific growth rate, were inconsistent
with the previous stage data. Therefore, it appears that zero-order
kinetics did not govern substrate removal in the fourth stage.
Also, since the fourth stage data yielded a negative K8, the data
from stages 3 and 4 were lumped together as one stage and entered in
the NONLIN program to determine if these two stages could be treated
as one stage. The resulting values for 0 (Table 4) were greater than
for any previous stage, indicating the inappropriateness of the
concept, i.e., the two stages have different kinetic characteristics.
The values for 0 in Table 4 show a decrease with stage, which is
expected. The highest maximum specific growth rate in the first
stage is about nine times the highest maximum specific growth rate in
the fourth stage. The 0 values have been calculated for each flow
rate, since these flow rates occurred at different average temperatures.
The high values for RZ , the correlation coefficient, indicate
that the summation model fits the data, and that carbon was in fact
the limiting nutrient. The decreased Rz for the fourth stage is
attributed to the inadequacy of the model at that stage.
In comparing the values of 0 obtained in this study with litera-
ture values, a tremendous discrepancy is noted. The values of 0
obtained in this study are an order of magnitude less than published
literature values. For example, Gaudy et al. [1967] found 0 to be
5.10 day"1 (BOD5 basis) for a suspended growth reactor treating skim
milk. Wu and Kao [1976] found a y of .91 day'1 (BOD5 basis) in a
suspended growth reactor treating yeast waste. Clark et al. [1978],
using a four-stage pilot scale RBC, found a 0 of 4.4 day"1 (BOD5 basis)
in the first stage.
The discrepancy appears to be based on two reasons. First of all,
although several authors have used COD as the basis for kinetic
equations [Heukelekian et al. 1951, Benedek and Horvath 1967, Gaudy
and Gaudy 1971, and Bin-ban j a et al. 1976], the trend has been to apply
BOD5 as the basis. In this study, the COD removed per stage was
probably much greater than the BOD5 removed per stage because of the
high COD:BOD5 ratio. Therefore, since COD removal is in the denomina-
tor of our model equation, this could have severely reduced 0.
Secondly, the discrepancy in 0 is probably more directly related
to the low hydraulic loading rates employed in this study. For
505
-------
example, the hydraulic loading rates in this study varied from .008
to .030 m3/m2-day (.19 to .75 gpd/ft2). Clark et al. [1978] varied
HLR's from .044 to .196 m3/m2-day (1.09 to 4.81 gpd/ft2). This
difference alone almost negates the order of magnitude difference
in p.
In this study, the hydraulics of the RBC prevented any higher
increase in. HLR. The .028 m3/m2-day HLR (.75 gpd/ft2) was the
maximum safe loading rate for the chosen rotation speed that would
allow each stage to operate without carryover of wastewater into the
next stage.
Simulation
For the data simulation, the stage kinetic parameters obtained
from the NONLIN solution technique (Table 4) were substituted back
into the model. The average influent soluble COD concentration,of
the three data points used for each HLR was then entered into the
model as So, along with the average values of the remaining steady
state data (Table 3). Stage effluent concentrations were then solved
for, and the predicted stage effluent concentration became the
influent for the next stage. Thus the model was forced to fit the
same data used in generating the kinetic parameters to predict the
stage effluent concentrations. This procedure is not a verification
of the model, but an exercise to show the model "fit."
The simulation results, using the value of 0.10 mg/£ for the
fourth stage tfs, are shown in Figures 2-4. As it is seen, the simula-
tions become less accurate with increasing loading rate. For the
850 m£/min and the 1250 m£/min flow rates (Figures 2 and 3), the
assumption of zero-order kinetics in the fourth stage is inaccurate.
This is especially noted in the 850 m£/min flow rate, where the fourth
stage simulation deviates from the actual data. The 3300 m£/min HLR
(Figure 4) shows a different pattern of stage soluble COD concentration.
The steady-state data show a greater range in concentrations than the
other two flow rates, and the fourth stage has significant effects on
substrate removal. The system reached its maximum biomass production
at the 3300 mA/min HLR, i.e., saturation was reached and zero-order
kinetics (with respect to substrate) governed. Therefore, the
simulation shown in Figure 4 appears to accurately represent the data.
The downward inflection after stage 3 does correspond to the decrease
in fourth stage soluble COD concentration.
The simulations of the steady-state data, lumping stages 3 and 4
together, are shown in Figures 5-7. The resulting curves do not
simulate the third and fourth stages, and this is most apparent at
the lower concentrations, i.e., the 850 m£/min flow rate (Figure 5).
The two stages act differently, and lumping them together creates
greater inaccuracy in the simulation.
Several possible reasons exist for the poor model simulation at
the higher loading rates. They are:
506
-------
Ul
O
-4
600-
_ 500^
•».
o>
60CH
O DATA OF WEEK 4
D 5
A 6
O DATA OF WEEK 15
O 16
A 17
INF
STAGE NUMBER
STAGE NUMBER
Figure 2. Simulation of the 850 mfc/min HLR
steady-state data, using 0.1 mg/£ as
the fourth stage value of K .
Figure 3. Simulation of the 1250 m£/min HLR steady-
state data, using 0.1 mg/£ as the fourth
stage value of K .
-------
o
oo
1400
1200-
o»
1000-
UJ
o
o
o
tu
m
o
V)
800-
600-
400-
200-
O DATA OF WEEK 22
D 23
A 24
INF I 2 3
STAGE NUMBER
60CH
500^
o>
g
i
UJ
o
8
g
o
CO
.- 4004
300H
200-
100-
INF
O DATA OF WEEK 4
D 5
A 6
I 2
STAGE NUMBER
4
Figure 4. Simulation of the 3300 m£/min HLR
steady-state data, using 0.1 mg/£ as
the fourth stage value of K .
S
Figure 5. Simulation of the 850 m£/min HLR steady-
state data, lumping stages 3 and 4.
-------
60Ch
5009
o>
Q 400-
300-
in
O
100-
O DATA OF WEEK
O
15
16
17
WOOi
O DATA OF WEEK 22
D 23
A 24
INF
2 3
STAGE NUMBER
2 3
STAGE NUMBER
Figure 6. Simulation of the 1250 mJl/min HLR
steady-state data, lumping stages
3 and 4.
Figure 7. Simulation of the 3300 m£/min HLR
steady-state data, lumping stages
3 and 4.
-------
1. Soluble COD was chosen to represent the carbonaceous substrate.
The kinetic parameters obtained are lower than those that would
be obtained using soluble BOD5 or soluble organic carbon as the
carbonaceous substrate. Better correlations may have been achieved
using these other parameters.
2. Inaccurate values for the growth yield, y, may have been used.
The steady-state values of growth yield used in the model varied
from .16 to 1.66. Although high growth yields were expected
due to maximum substrate utilization, the computed values greater
than 1.0 probably should not have been used; Growth yields of
greater than 1.0 probably occur instantaneously, or for short
periods of time, but not at steady-state. /The dynamics of sludge
production in the RBC are quite complicated, and the attempts in
this study may not have successfully established an empirical
basis for Y.
3. The entire depth of attached biomass plus the suspended biomass
was assumed to affect substrate oxidation. This assumed that
diffusion of oxygen or substrate into the biofilm was not limited
by temperature, film thickness, dissolved oxygen concentration,
etc. In this study, however, low DO concentrations and high
wastewater temperatures at the highest HLR could have limited
substrate oxidation in the attached biomass. Other authors have
used a partial depth of biofilm in modeling substrate removal
\Kornegay and Andrews 1968 and Williamson and MoCarty 1976].
The suspended biomass comprised only a small percent of the total
mass of the system; therefore, its inclusion in the model probably had
negligible effects.
ENGINEERING SIGNIFICANCE
The steady-state model (Equation 9) developed in this study can
be utilized as a tool in designing RBC systems for similar applications.
This is done by constructing a nomograph of flow rate versus required
disc area for desired percent soluble COD removals. The isoconcentra-
tion lines (percent soluble COD removal lines) are found by solving
Equation (9) for Aw, the disc area, as a function of the flow rate F.
The resulting nomograph is shown in Figure 8. Equation (9) is solved
in this manner using an average value for each of the kinetic constants.
In constructing this nomograph, the four stage RBC is treated as
one stage, so that the disc area needed to obtain a specified percent
COD removal at a specified flow rate is the total disc area for the
entire multi-stage system. Nomographs could similarly be constructed
for each stage.
Figure 8 was constructed based on the average influent soluble
COD concentration of 575 mg/£. The influent soluble COD concentration
determines the position of the isoconcentration lines, since the
required disc area is dependent on ACOD removed. Thus a different
influent soluble COD concentration would shift the position of the
510
-------
u
OC
5.0-
4.0-
3.0
o 2.0
u_
1.0-
80 85 90
95
0 50 100 150 200 250
TOTAL DISK AREA, m2
Figure 8. Design nomograph relating flow rate with total disc area for
various percent soluble COD removals.
isoconcentration lines. Note that the isoconcentration .lines are
nonlinear because Xft the attached biomass concentration, varies with
the flow rate F, IT x~ did not vary with F, the isoconcentration lines
would be linear. ^
SUMMARY AND CONCLUSIONS
A four stage 1.2 m (47 inch) diameter Rotating Biological Contactor
treatment system was operated continuously for six months in treating
a cheese processing plant wastewater. The system was operated under
flow rates of 850, 1250, and 3300 m£/min, and steady-state effluent
carbonaceous substrate concentrations were achieved for each of the
three flow rates. Treatment process performance was measured via
weekly chemical analyses of 24 hour composite samples of the influent,
stages, and effluent from the system. System physical parameters were
monitored regularly at the field site.
Carbon was determined to be the growth limiting nutrient, and
a steady-state Monod kinetics model was developed to simulate carbona-
ceous removal in the RBC. In the model, the temperature dependency of
the maximum specific growth rate 0 was represented by an Arrhenius-type
relationship. A nonlinear FORTRAN curve-fitting solution technique
was used to solve for the stage kinetic parameters E , the activation
511
-------
energy, A, the frequency factor, and Ks, the half-saturation constant.
The kinetic parameters found from the curve-fitting technique are
listed in Table 4.
Leaving the kinetic parameters unbounded and free to "float"
during the solution technique, Ea ranged from 17.78 to 17.85 kcal/mol
in the four stages. Likewise, A varied from 1.5 x 1012 to .26 x lo12
day"1, and Ks ranged from 186 to -36 mg/A. The corresponding values
of $ ranged from .307 to .015 day"1. The high correlations obtained
for each stage indicate that the assumptions used in deriving and
applying the model are valid.
From the results of this study, the following conclusions can be
made:
1. In the lower concentration ranges encountered, soluble COD removal
follows first-order kinetics, while at the higher concentration
ranges, soluble COD removal is zero-order.
2. Carbon is the growth limiting nutrient in the wastewater, and the
average C:N ratio for the steady-state data is 44.6:1.
3. Carbonaceous substrate removal can be best described by the Monod
kinetics model of Equation (9).
4. The maximum specific growth rate, {3, of the microorganisms
decreases with available substrate and temperature, among other
factors, from .307 day1 in the first stage to .015 day'1 in the
fourth stage.
5. The maximum specific growth rate temperature dependency can be
described by an Arrhenius-type equation.
6. The activation energy, Ea, is independent of stage, and has a mean
value of 17807.5 cal/mol.
7. The half-saturation constant, Ks, changes with stage, and the
values for the four stages are 32, 186, 59, and -36 mg/£,
respectively.
8. The summation model (Equation 9) becomes inaccurate at the fourth
stage, as shown by the negative Ks value in this stage.
9. Using an empirical equation for Y, the growth yield, the steady-
state values for J varied from ,16 to 1.66.
10. The percent soluble COD removal was 88.0, 88.0, and 71.1 percent
for the 850, 1250, and 3300 m£/min flow rates, respectively.
ACKNOWLEDGMENTS
The equipment for the project was provided by the Environmental
Systems Division of George A. Hormel and Company, Coon Rapids, Minnesota.
Their assistance and encouragement are greatly appreciated. In
512
-------
addition, financial support was provided by the U. S. Environmental
Protection Agency Graduate Training Program Grant Number T-900861,
Manpower Planning and Training Branch, Washington, D.C.
REFERENCES
Hartmann, H, 1960, The dipping contact filter. Oesterreiehische
Wasserwirtschaft 11/12:264 (original not seen; cited in Cruz
[1977]).
Joost, R. H. 1969. Systemation in using the rotating biological
surface (RBS) waste treatment process. Proa. 24th Purdue Ind.
Waste Conf., Purdue Univ., Lafayette, Indiana, pp. 365-373.
Antonie, R. L., and F. M. Welch. 1969. Preliminary results of a novel
biological process for treating dairy wastes. Proc. 24th Ind.
Waste Conf., Purdue Univ. pp. 115-126.
Weng, C.-N., and A. H. Molof. 1974. Nitrification in the biological
fixed-film rotating disk system. Jour. Water Poll. Contr. Fed.
46(7):1674-1685.
Grieves, C. G. 1972. Dynamic and steady state models for the rotating
biological disk reactor. Ph.D. Dissertation, Clemson University.
252 pp.
Hansford, G. S., J. F. Andrews, C. 6. Grieves, and A. D. Carr. 1976.
A steady state model for the rotating biological disc reactor.
Water Res. (G.B.) in press. 55 pp.
Benjes, H. H., Jr. 1978. Small communities wastewater treatment
facilities - biological treatment systems. In: Design Seminar
Handout Small Wastewater Treatment Facilities. USEPA. Technology
Transfer. January 1978. 94 pp.
Kornegay, B. H., and J. F. Andrews. 1968. Kinetics of fixed film
biological reactors. Jour. Water Poll. Contr. Fed. 40(11):R46Q-
468.
Kornegay, B. H. 1972. Modeling and simulation of fixed film biologi-
cal reactors, p. 257-298. In T. M. Keinath and M. P. Wanielista
(eds.) Mathematical Modeling in Environmental Engineering.
Eighth Annual Workshop, Association of Environmental Engineering
Professors, Nassau, Bahamas, December 18-22.
Kornegay, B. H. 1975. Modeling and simulation of fixed film biologi-
cal reactors for carbonaceous waste treatment, p. 271-318. In
T. M. Keinath and M. P. Wanielista (eds.) Mathematical modeling
for water pollution control processes. Ann Arbor Science
Publishers Inc., Ann Arbor, Michigan.
Clark, J. H., E. M. Moseng, and T, Asano. 1978. Performance of a
rotating biological contactor under varying wastewater flow.
Jour. Water 'Poll. Contr. Fed. 50(5):896-911.
513
-------
seen; cited in Lawrence, A. W., and P. L. McCarty. 1970, Unified
basis for biological treatment design and operation. Jour.
San. Eng. Div.3 ASCE 96(SA3):757-778,*
Benedek, P., and I. Horvath. 1967. A practical approach to activated
sludge kinetics. Mater Res. 1(10):663-682. (original not seen;
cited in Lawrence and MoCarty [1970]}.
Gaudy, A. F., Jr., and E. T. Gaudy. 1971. Biological concepts for
design and operation of -the activated sludge process. USEPA
17090 FQO 09/71. 154 pp.
Bintanja, H. H., J. J. Brunsmann, and C. Boelhouwer. 1976. The use
of oxygen in a rotating disk process. Water Res. (G.B.) 10(6):
561-565.
Williamson, K., and P. L. McCarty. 1976. A model of substrate
utilization by bacterial films. Jour. Water Poll. Contr. Fed.
49(l):9-25.
* Reference list incomplete at time of publication. Please contact
author for complete references.
514
-------
PART V: AIR DRIVE AND SUPPLEMENTAL AERATION
AERATED RBC'S - WHAT ARE THE BENEFITS
By
Kevin J. McCann
Applications Engineer
Autotrol Corporation
and
Richard A. Sullivan
Manager, Process & Application Engineering
Autotrol Corporation
INTRODUCTION
Since 1973 Autotrol Corporation has conducted extensive testing, evaluating
the impact of the aeration process on the operation of RBC units,. 2ar.li.er
studies were conducted solely under the auspices of the Company at its Corp-
orate laboratories and at test facilities located in the Milwaukee Metropoli-
tan Sewerage Commission's Wastewater Treatment Facility South Shore plant.
Other pilot studies have been performed in cooperation with private engineers
such as Greeley and Hansen (Portsmouth, VA), Metcalf & Eddy Engineers (South-
west Water Pollution Control Plant, San Francisco, and Jenks and Harrison Inc.
(Union Sanitary District, CA) . The conclusions derived from these studies
have been confirmed in more recent full-scale evaluations comparing the
515
-------
effluents of aerated biological contactors (ABC) and standard mechanically
driven rotating biological contactors (RBC) at identical loadings. These
tests have been conducted by Greeley & Hansen at the Alexandria, VA Waste-
water Treatment Plant and Duncan, Lagnese & Associates at an industrial
installation. The results of these studies will be more fully discussed in
papers to be presented later in this conference. This presentation summarizes
the benefits of subjacent aeration in the RBC process.
RBC~ Unaerated Mechanical Drive units
To appreciate the benefits of aerating an RBC unit one should have some
familiarity with typical process limitations and problems encountered by
standard mechanically driven units. Overloading, anaerobic growth, beggiatoa
and lack of operator control comprise the majority of problems and criticisms
encountered at RBC facilities experiencing difficulties. Most of these
criticisms are interrelated and can be discussed under a few general areas.
a) Overloading
For every RBC system, treating either municipal or industrial waste, there
is a. limit to the rate at which the waste is applied beyond which first stage
RBC units will not be able to maintain aerobic conditions. Once pushed into
this anaerobic region there are a number of problems which can develop.
Undesirable organisms will proliferate and retard oxidation of the BOD
substrate. The two most common bacterial forms are anaerobic and sulfur
oxidizing micro-organisms.
The anaerobic bacteria normally co—exist with aerobic micro-organisms form-
ing the underlayer of the biofilm. Mathematical model studies have shown
that the aerobic layer is extremely thin, its depth controlled in early
stages by oxygen diffusion and in downstream stages by substrate diffusion.
These same models indicate that aerobic layers average only .004 to .008
inches thickness with any greater depth attributed to anaerobic growth.
When overloading occurs, the tremendous growths which tend to develop are
primarily inactive anaerobic bacteria. This extra "dead load" potentially
reduces the life expectancy of the equipment.
However, the anaerobic layer is not inactive. Any sulfate present in the
waste is reduced to sulfide under anaerobic conditions. The sulfide becomes
a food source for beggiatoa, or sulfur oxidizing bacteria. Although an aerobic
bacterial form, this chemotroph is extremely successful in competing with
carbon bacteria, but provides no removal of BQD. Further, the growth is
gelatinous, effectively blocking oxygen transfer and further inhibiting carbon
bacteria.
Even with a. correctly loaded system, beggiatoa presents a potential problem if
sulfide already exists in the system due to either septic conditions in the
collection system or the primary clarifier, an industrial waste discharge
having a. high sulfide content o£ any other uncontrollable source of sulfide.
In addition to causing reduced BOD removal through the system, beggiatoa growth
has poor settling characteristics and will normally create a pinpoint white
floe which easily crests the effluent weirs in the secondary clarifier.
516
-------
b) Lack ofOperating Control
This historically has been and continues to be one of the most serious
arguments by engineers and operating personnel against the use of RBCs. It
is especially strong in small applications where the operator frequently
does not have the luxury of removable baffles for some measure of control.
It is the unforeseen situation causing a short term problem, for which there
is no easily applied correction, which makes this criticism valid. With a
strictly mechanically driven RBC. The only potential controls available to
the operator are rotational speed adjustment and removable baffles. However,
the rotation of the RBC unit provides both aeration and shearing force for
controlling biomass thickness.
Rotational speeds are fairly uniform in the industry between 1.4 and 1.7 rpms.
These numbers cannot be lowered because of reduced oxygen transfer efficiency.
This loss of oxygen transfer efficiency would further complicate an already
overloaded first stage, while, reduced shearing force would cause further
proliferation of anaerobic biomass. Increasing rotational speed is not a
viable alternative since an increase to ven 2.25 rpm would cause the KBC process
to be non-competitive with other biological treatment systems. This seemingly
small adjustmant would double the power consumption, doubling the operating costs.
Baffle adjustment for staging alternative does not easily lend itself to short
term problems. This option entails draining tankage to remove baffles and would
most likely be accomplished after the problem had passed.
ABC — Aerated Biological Contactors
Understanding these major limitations of the traditional RBC unit should give
greater appreciation for the following improvements which have been evaluated
for ABC systems.
a) Thinner Biomass
From the earliest research, the major concern of investigators attempting to
optimize any biological fixed film treatment system has been control of
biomass. Few fixed film processes,' if any, offer an operator an "activf,"
method for controlling film thickness.
As we have discussed, RBC systems effectively control the film thickness
through a "passive" system of shearing forces created by the rotation of the
unit.
Figure 1 presents data from six consecutive test periods of the South Shore
test facility comparing biofilm thickness for stages 1 and 3 for KBC and ABC
systems each loaded at identical rates. For only the first two test periods
for Stage 3 of the ABC system and the first test period for Stage 3 of the
RBC system, is the biofilm thin enough that a completely aerobic growth is
indicated,
At every point, however, the ABC system is carrying less anaerobic growth
than the RBC train. A dramatic disparity in growth thickness is especially
obvious in the first stage. The two systems start to approach the same
thickness in stage 3, with stage 4 showing little or no difference in biomass
thickness.
517
-------
Figure 1
South Shore Test Facility
0.12
V)
G>
JC.
o
•E 0.09
«^«»
V)
V)
o
c
~ 0.06
f-
E
H—
I
£ 0.03
m
0
RBC - stage 1
ABC - stage 1
RBC - stage 3
ABC - stage 3
I I
3 4
Test Period
518
-------
In these later stages the low organic loadings act as a control on the bio-
film thickness without any need for further "outside" measures, or operator
control.
The major implication of this benefit is that high density media, which
provides more surface area per cubic foot, because of the closer spacing of
the plastic media layers, can be used earlier in the treatment train without
fear of bridging caused by excessive bacterial growth. The thinner biofilm
can be attributed to the stripping action of the bubbles rising across the
surface of the media. In addition to keeping the biomass to a thickness
which can be kept primarily aerobic, the decreased weight carried by the
shaft can only increase the life cycle of the equipment.
b) Higher P.O. Concentrations
Since the aeration arc of the RBC unit is augmented by an additional source
of oxygen in an ABC system, comparisons of identically loaded side by side
evaluations have always shown improved D.O. profiles for the ABC units.
Figure 2 indicates such a comparison for the first stages at South Shore.
Plotting the change in dissolved oxygen versus organic loading the ABC system,
Figure 2 shows that the ABC units have excess oxygen transfer capability above
loadings of 5 Ibs of soluble BOD. per day per 1000 ft^ surface area. RBC
systems' oxygen transfer capability, however, is exceeded above loadings of
3 Ibs of soluble BOD per day per 1000 ft^ of surface area.
This advantage permits an ABC system to be loaded at higher levels and
remain aerobic. This coupled with the higher ORP prevents septic conditions
from prevailing and avoids the undesirable effects of an overloaded system.
Further, ABC units can more easily control situations where sulfides already
exist in the waste prior to entering biological treatment. This has been
demonstrated in full-scale studies at Alexandria, VA and in pilot plant
studies at San Fransicso. In both cases sulfide content of the influent
stream is such that an RBC system during summer operation cannot meet effluent
criteria without drastically lowering the organic loading to prevent beggiatoa
poliferation.
Metcalf & Eddy recommended, in their Summary Report for the Southwest Water
Pollution Control Plant Project, that "... RBC units be used following
primary sedimentation and that they be mechanically driven with supplemental
air provided or .that they be air driven." This was based upon the Engineer's
experience with a beggiatoa problem and his evaluation that the "... correction
of the problem by supplemental air addition was demonstrated dramatically...
where the addition of air to the wastewater flows under the discs eliminated
the beggiatoa growth within a matter of hours. The air not only corrects
the low dissolved oxygen condition but, through the shearing action of the
air bubbles results in a thinner layer of biomass on the disc, allowing the
entire biomass to remain aerobic."
c) Increased BOD'Removal
Primarily due to the higher dissolved oxygen concentrations, denoting im-
proved oxygen transfer, coupled with the control of undesireable bacterial
forms, increased BOD removal rates have been noted for ABC systems. Figures
3 and 4 show side by side comparisons of an RBC and ABC systems at Alexandria,
519
-------
D)
E
%—»
c
o
D)
>»
X
O
TJ
0
_>
O
w
03
Figure 2
South Shore Test Facility
First Stages
a
a
0246
Applied Organic Loading
(Ibs. sol. BOD/day/1000 sq.ft.)
520
-------
Virginia. The stage analysis for soluble BOD concentrations shows lower
concentrations for each stage of the ABC units than the RBC stages, although
both systems were loaded at similar rates.
This improved performance has been eveluated to a much greater extent at the
South Shore test facilities and has resulted in Autotrol Corporation
publishing a separate design curve for ABC systems. The major improvements
have been noted in early stages, the more highly loaded units, demonstrating
further that this advantage is primarily an oxygen transfer improvement
phenomena.
This reduced surface area requirement, coupled with the more efficient use of
higher density media, has resulted in the elimination of 10% to 20% of the
number of shafts required, depending on the application.
d) Operating Flexibility
Because of the improved oxygenation and additional shearing forces for bio-
mass control, the operator has available one more level of control with an
ABC system - air flow rate. These can be adjusted as required during seasonal
peaks or unanticipated discharges from heavy industrial users. Further, the
system can be earily "turned down" during seasonal low periods.
While this would permit the operator to take advantage during low periods of
additional power savings, the major thrust is the ABC's ability to safeguard
against the unforeseen heavy load.
CONCLUSION
In summary by applying aeration equipment to an RBC unit, an engineer can
offset the traditional criticisms of earlier applications. Further, by
efficient application, an ABC system can offer a necessary safety factor to
the total installation without increase in total life cycle cost.
521
-------
Figure 3
OJ
£
%..*
o
c
o
O
CO
Oi
a
*•»
CO
Q
O
CD
O
CO
50 *,
40
30
20
10
0
infl.
Alexandria, VA
9/21/78
row 3 - RBC
row 4 - ABC
1234
Stage Number
522
-------
Figure 4
CD
E
x^
o
c
o
O
-------
Page Intentionally Blank
-------
PERFORMANCE EVALUATION OF AIR DRIVEN
RBC PROCESS FOR MUNICIPAL WASTE TREATMENT
By
R. Srinivasaraghavan
Associate, Greeley and Hansen
Carl W. Reh
Partner, Greeley and Hansen
Sven Liljegren
Director, Pinners Point STP
INTRODUCTION
The existing Pinners Point Wastewater Treatment Plant is
a primary plant treating an average flow of 12 mgd. The
plant is to be upgraded to provide secondary treatment at a
design average flow of 16.4 mgd. The effluent criteria for
the secondary plant are 30 mg/1 biochemical oxygen demand and
30 mg/1 suspended solids.
Based on preliminary comparative process evaluation
studies, the Rotating Biological Contactor (RBC) process has
been chosen as the secondary treatment process at Pinners
Point. A pilot plant study was conducted to establish bases
of design for the secondary treatment facilities.
EXPERIEMNTAL SET-UP
Pilot Unit
An "aerosurf" RBC pilot unit was used in the study. The
pilot unit was a four-stage air-driven RBC, 10 feet long and
a 10.4 feet diame.ter. The total surface area of the unit
was 23,000 ft^. The disc was rotated by air provided through
the diffusers located at the bottom of the tank.
The primary effluent was the feed to the pilot plant.
The waste flow to the RBC pilot unit was controlled by using
a throttle valve to obtain the desired rate. The wet well
was aerated to reduce the sulfide level and provide a positive
525
-------
ORP in the sewage to avoid the growth of undesirable organisms.
The RBC effluent was settled for thirty minutes in a six-inch
diameter experimental column to simulate the operation of the
secondary settling tanks. A diagram of the pilot plant is
shown on Figure 1.
Operating Conditions
The start-up period lasted for about three weeks in
September, 1977. The test program lasted approximately six
months and consisted of seven phases. Overall hydraulic load-
ings were adjusted to cover a range of 1.0 to 3.0 gpd/ft2 during
Phases I through V. The shaft speed was maintained, through
the adjustment of air flow rate, at about 1.60 rpm through all
phases except Phase V when the speed was 1.04 rpm.
Phases VI and VII studied the effect of instantaneous and
diurnal variations in flow on RBC performance. Plant operating
reports indicate an instantaneous flow fluctuation cycle to be
20 to 25 minutes long, with a flow variation of 5 to 30 mgd
during the day and 3 to 15 mgd during the night. A cyclical
flow pattern as indicated on Figure 2 was maintained to approx-
imately simulate this condition in the pilot study. The flow
fluctuation was accomplished by using three constant head tanks
with outlet solenoid valves, each controlled by a timer.
A summary of the program schedule and the operating con-
ditions are shown in Table 1. Settling and thickening tests
were performed on the RBC effluent using an eight-foot settling
column and one liter glass cylinder, respectively.
Data Collection
As indicated on Figure 1, samples were collected at five
sampling points for the RBC influent, effluent, and wastewater
in three intermediate stages. All samples from the RBC unit
were composited. Grab samples were taken for settling tests.
The weekly analytical schedule is shown in Table 2.
RESULTS AND DISCUSSION
Influent Characteristics
The characteristics of the influent to the RBC unit are
shown in Table 3. The values shown are the means of the daily
526
-------
determinations. The percent coefficient of variation for
each parameter is also shown to indicate the extent of in-
fluent quality fluctuation.
The overall average influent BOD5 and SBODs were
104 + 34 mg/1 and 42 + 22 mg/1, respectively, and the SS and
VSS concentrations were 79 +_ 44 mg/1 and 67 j^ 40 mg/1, re-
spectively. It can be seen that there were considerable fluc-
tuations in BOD5 and SBOD5 in each phase as reflected by the
high coefficient of variation values. Even greater amounts of
variation were observed in SS and VSS concentrations in the
influent.
Table 4 shows the ratio of BOD§ to COD on both soluble and
total bases. The ratio ranged from 36 percent to 59 percent,
indicating a significant amount of not readily biodegradable
fraction in the influent. The average COD to BODs ratio did not
change significantly between phases, averaging about 2.3.
The average DO concentration in the RBC influent after
aeration in the wet well ranged from 2.3 mg/1 in Phase I to
7.3 mg/1 in Phase VI. This increase is believed to be due to
the decrease in water temperature from September to March.
RBC Performance
Effect of Loading:
Loading to RBC has been expressed in three different ways:
1. Hydraulic, gallons/ft2/day
2. Soluble Organic, pounds SBOD5/1,OOQ ft2/day
3. Total Organic, pounds 8005/1,000 ft2/day
Each has been plotted against the performance parameters, SBODj
and BOD§ pounds removed, effluent quality, and removal effi-
ciency, and summarized in Table 5. All plots are based on
linear regression calculations.
The total BODs are 30-minute settled effluent values and
the SBODs are RBC effluent values. Total 8005 values are plot-
ted only to obtain the order of magnitude of BODs concentrations
to be expected in the effluent. The thirty-minute settling
test only grossly simulated the performance of a full-scale
527
-------
secondary clarifier. It is anticipated that the effluent
quality would be better with a full-scale secondary clarifier.
o Hydraulic Loading
The average hydraulic loading was varied from
1.0 to 3.0 gpd/ft2 during the test program.
The pounds SBOD5 and BODs removed, effluent SBODs
and BODs , and SBODs and BOD5 removal efficiencies
are plotted as functions of hydraulic loading for
all phases on Figures 3 through 8 .
It can be seen from Figures 7 and 8 that, in spite
of high fluctuations, the treatment efficiency
generally decrease with an increase in hydraulic
loading. Figures 5 and 6 show the same result, but
in terms of effluent quality. Overall the correla-
tion is not satisfactory for performance prediction.
o Soluble Organic Loading
The average soluble organic loading ranged from 0 . 5
to 1.2 pounds SBOD5/1,OOQ ft2/day. Since the in-
fluent SBODs varied widely, individual data points
were plotted in addition to the average phase data.
An average phase value plot of pounds SBODs removed
versus pounds SBODs applied is shown on Figure 9.
Figure 10 is the same graph using daily data points.
It can be seen that excellent correlation exists
between the pounds of soluble organics applied and
removed. The correlation coefficient is 0.97 which
is extremely good for a biological system. The slope
of the regression line which is representative of the
percent removal efficiency indicates about 80 percent
removal .
Figure 11 is a plot of pounds BOD§ removed versus
SBODs loading. The correlation coefficient for this
graph, 0.92, is also relatively high. Figures 9, 10
and 11 show that the pounds of organics removed in-
creased linearly with increases in soluble organic
loading, indicating nonoxygen limiting conditions.
On Figures 12 and 13, the average percent SBODs re-
moval and BODs removal are plotted against SBODs
528
-------
loading for each phase. The regression lines in-
dicate that the performance of the RBC decreases
with an increase in soluble organic loading.
Figures 14 and 15 are plots of effluent SBOD^ and
BODs against SBODs loading. These plots indicate
that effluent quality decreases as the soluble
organic loading decreases.
o Total Organic Loading
The average total organic loading ranged from 1.1
to 3.5 pounds BOD5/1,000 ft2/day. As with the
soluble BODs, the influent total BODs varied widely,
and individual data points were plotted in addition
to the average phase data.
Figures 16 and 17 are average phase value plots
of SBODs removed and BODs removed versus BODs load-
ing. The same correlation is shown on Figure 18
using daily data points. The plots based on BODs
loading have high correlation coefficients: 0.89
for SBODs removal and 0.95 for BODs removal. The
regression lines indicate nonoxygen limiting con-
ditions .
The SBODs and BODs effluent concentrations and re-
moval efficiencies are plotted as functions of 6005
loading for each phase on Figures 19 through 22.
These graphs show that the treatment efficiency and
effluent quality decrease as the total BODs load-
ing increases.
Effluent Characteristics
The characteristics of the RBC pilot plant effluent are
shown in Table 6. Both BODs and SS are the properties of the
30-minute settled effluent.
The effluent ammonia nitrogen (NH3~N) concentrations were
compared to the influent NH3-N concentrations. The average
NH3-N concentration for all seven phases decreased from 30 +_
10 mg/1 in the influent to 24 + 9 mg/1 in the effluent. This
decrease is due to bacterial assimilation. The largest de-
crease in NHg-N occurred during Phase II when the average NH3~
concentration .decreased from 36 to 19 mg/1 across the RBC, in-
dicating some nitrification.
529
-------
In order to estimate the BODs in the effluent, an at-
tempt was made to correlate the particulate BODs (total
BODs - soluble BODs) aru^ SS removals during the settling pro-
cess. Figure 23 is a plot of particulate BOD5 removal ver-
sus SS removal, mg/1, during the 30-minute settling test.
The particulate BODs removal was determined by the differ-
ence between the total BODs concentrations in the RBC effluent
and 30-minute settled effluent, assuming there was no change
in the SBOD§ concentrations during the settling process. The
slope of the regression line is 0.37, indicating 0.37 mg par-
ticulate BODs Per each mg SS.
Settling Characteristics
Settling tests were performed during each phase to deter-
mine the required surface overflow rate, gpd/ft^, and surface
area of settling tank. The tests indicate better settling
characteristics during the lower soluble organic loading phases.
Figure 24 is a plot of settled effluent suspended solids
as a function of soluble organic loading. The purpose of the
plot is not to determine effluent SS at various loading but to
observe the effect of organic loadings on general settling
characteristics reflected by 30-minute settling test. The re-
gression line indicates that the effluent quality decreases as
the SBODs loading increases.
Figure 25 shows a plot of percent SS removal versus surface
overflow rate calculated from the settling tests conducted for
all of the phases. A band is shown to illustrate the range of
values observed.
Sludge Production
The RBC influent solids to effluent solids ratio was plot-
ted to determine if there is any net sludge increase or de-
crease due to synthesis or oxidation. This is shown on Figure
26. No definite trend is apparent. Therefore, the mean value
is chosen for the estimation of sludge production. The sludge
production is approximately equal to 0.83 times the amount of
suspended solids entering the RBC.
Effect of Rotational Speed
A. shaft speed of about 1.60 rpm was maintained through
all of the phases except Phase V when the speed was 1.04 rpm.
530
-------
Comparing Phase V with Phase II and III which had similar sol-
uble organic loadings (about 0.6 pounds SBOD5/1,000 ft^/day),
there is no significant difference in effluent quality. The
SBOD5 removal was lower when the rotating speed was reduced,
but this is probably due to a lower influent SBODs concentra-
tion in Phase V.
Comparing Phase V to all of the phases, no significant
difference is observed in the performance of the RBC. This is
probably because oxygen limiting conditions did not occur during
any of the phases.
Effect of Diurnal Flow Variation
During Phase VI and VII, the RBC influent flow was varied
in a four-step cyclical flow pattern to simulate day and night
flow variations. There is no significant difference between
the results of these phases and the first five phases. This in-
dicates that the diurnal flow variation of the magnitude and
range experimented has no significant effect on the performance
of the RBC unit.
SUMMARY
The pilot plant data generally seemed to correlate bet-
ter with loading when loading is expressed as organic rather
than hydraulic. The performance characteristics showed a direct
dependence on the soluble BOD5 loading.
The bases of design for the full scale RBC process is shown
in Table 7. Based on these design conditions and the RBC pilot
plant study, the performance of the RBC can be estimated as
follows:
o Effluent SBOD5
From the regression equation developed based on
data shown on Figure 10, it is estimated that
the effluent SBOD5 at the design loading rate of
1.4 pounds SBOD5/1,000 ft2/day would be 17 mg/1.
o Effluent SS
Better than 80 percent of the RBC effluent sus-
pended solids can be removed in the secondary clar-
ifier at the design surface overflow rate of
531
-------
740 gpd/ft^ and a hydraulic retention time of
2.97 hours. Assuming an RBC effluent SS con-
centration of 88 mg/1, the effluent SS concen-
tration would be 18 mg/1.
o Effluent BOD5
The ratio of particulate BODs to SS in the efflu-
ent was found to be 0.37. An effluent with 18 mg/1
SS will have a particulate BODs of about 7 mg/1.
Therefore, the total 8005, soluble and particulate,
in the effluent is estimated to be about 24 mg/1.
o Sludge Production
The secondary sludge production under the assumed
design conditions is indicated to be approximately
0.83 times the influent suspended solids.
o Effect of Rotating Speed
Comparing Phase V, when the rotating speed was
changed from about 1.6 rpm to 1.04 rpm, with the
other phases, no significant difference is observed
in the performance of the RBC.
o Diurnal Variation
The diurnal flow variation of the magnitude and
range experimented during Phases VI and VII has
no significant effect on the performance of the
RBC unit.
532
-------
TABLE 1
Study Phases and Hydraulic Loading Characteristics
Phase
I
II
III
IV
V
VI
VII
Period
9/18/77- 9/30/77
10/ 9/77-10/30/77
ll/ 6/77-12/ 2/77
12/11/77- I/ 4/78
I/ 8/78- 1/31/78
2/22/78- 3/ 7/78
3/ 8/78- 3/22/78
Number
of Days
at
Steady State
13
22
27
25
24
14
15
RBC
Influent
Average
Temperature
OF
81
73
69
61
57
56
55
Shaft
Speed
rpm
1.61
1.71
1.62
1.66
1.04
1.55
1.58
Mean
Hydraulic
Loading
gpd/ft2
3.00
1.00
1.79
3.03
2.16
2.74
1.41
533
-------
TABLE 2
Sampling and Testing
pH Temp. D.O. SBOD*
Influent 555 5
Stage 1 - - 5 5
Stage 2 - - 5 5
Stage 3 - - 5 5
Stage 4555 5
30 Min. -
w Settled
w Effluent
Settling Test - Once a week during steady state
OP ANALYSES PER
BOD SS VSS COD SCOD* NHs-N S- ALK.
55533 12 1
555"""" — — *»
555-- --
555-- --
55533 1* 2 1*
5K _*..•«_
•3 _ — — — — —
operation in each phase.
* .Analyses on Filtered Samples
-------
Ln
oo
TABLE 3
RBC Influent Characteristics, mg/1
BOD5
Phase Mean
I 144
II 146
III 93
IV 100
V 84
VI 90
VII 96
Overall
Average 104
(1) % COV =
covd)
25
20
30
27
30
22
25
33
Percent
SBOD5
Mean
49
71
41
40
29
34
40
42
COV
62
35
51
40
52
36
36
53
Coefficient
SBODs
34
49
44
40
35
38
42
COD
Mean
328
262
202
226
181
211
237
40 226
of Variation =
SCOD
COV Mean
25
32
23
21
16
22
35
30
Standard
126
124
126
134
105
133
119
COV
21
29
45
32
22
31
16
123 32
Deviation x
SS
Mean COV
92 29
108 29
70 21
84 40
84 102
64 33
58 36
79 56
100
VSS DO
Mean COV Mean COV
75 34 2.3 16
97 34 3.1 16
54 28 3.8 14
75 43 4.4 19
68 106 5.9 12
54 43 7.3 6
48 40 7.0 10
67 59 4.7 37
Mean
-------
TABLE 4
Influent COD/BODs Ratios
Phase
I
II
III
IV
V
VI
VII
COD
BODS
2.34
1.80
2.26
2.40
2.32
2.32
2.80
BOD5(100)
COD
43
59
44
42
43
43
36
SCOD
SBODs
3.01
1.79
3.28
4.00
4.11
5.06
2.74
SBODS(IOO)
SCOD
33
56
31
25
24
20
36
536
-------
U)
TABLE 5
Performance of RBC Pilot Plant at Various Loadings
Loadings
Phase
I
II
III
IV
V
VI
VII
SBODs
Lbs SBODs Lbs BOD5 gpd Removed
1000 ft2/day 1000 ft2/day ft2" lbs/1000 ft2/day
1.22
0.57
0.65
1.07
0.60
0.78
0.47
3.53
1.14
1.42
2.62
1.60
2.05
1.13
3.00
1.00
1.79
3.03
2.16
2.74
1.41
0.98
0.54
0.51
0.60
0.40
0.59
0.38
BOD5
Effluent
Removed SBODs
lbs/1000 ft2/day mg/1
2.73
1.08
1.11
1.67
1.10
1.72
0.97
9
8
7
12
9
9
8
Effluent
BOD5
mg/1
35
17
19
34
23
18
13
% SBODs
Removal
81.6
88.7
82.9
70.0
69.0
75.1
75.4
% BOD5
Removal
75.7
88.4
79.6
66.0
70.2
80.5
85.2
(1)
concentrations are 30-minute settled effluent values; SBODs concentration are RBC effluent values.
-------
TABLE 6
Effluent Characteristics, mg/1
Phase
I
II
III
IV
V
CJ1
00 t/T
OD
VII
Overall
Average
BODs CD
Mean COV
35 26
17 40
19 47
34 38
23 40
18 27
13. 26
23 5
SBODs '
Mean
9
8
7
12
9
9
8
9
P5
COV
77
92
43
33
49
47
32
65
SBOD5
26
47
37
35
39
50
62
39
COD
Mean
220
141
155
171
140
159
141
156
(2)
COV
25
24
21
24
24
18
29
27
SCOD(2)
Mean
89
65
83
95
77
89
85
83
COV
17
30
40
27
11
34
17
33
SS
Mean
32
24
20
24
18
11
9
20
U)
COV
42
74
75
46
63
61
56
71
VSS(2) DOW
Mean
73
88
49
51
44
36
48
55
COV
25
35
31
37
32
36
59
47
Mean
1.7
4.2
4.3
5.1
5.8
7.3
7.5
5.1
COV
35
7
6
27
17
5
8
36
(1) 30-minute settled effluent values
(2) RBC effluent values
For effluent before settling, BOD5 = 50 +_ 19 (Range: 20-135) mg/1
SS = 71 + 31 (Range: 8-224) mg/1
-------
TABLE 7
RBC Bases of Design(1)
(16.4 mgd)
1. RBC dimensions
Diameter, f t. 12
Surface area per unit, ft2 100,000
2. Maximum shaft speed, rpm 1.6
3. Number of shafts per tank 6
4. Number of tanks 12
5. Number of shafts 72
6. Annual average loading
Total BODs, lbs/1,000 ft2/day 2.7
Soluble BODs, lbs/1,000 ft2/day 1.4
Hydraulic, gpd/ft2 2.3
7. Annual average influent concentration
Total BODs, mg/1 148
Soluble BODs, mg/1 75
Suspended Solids, mg/1 88
(1) Design year: 2010
539
-------
FICVIE I
TO 39 111.
smugs CHUM
SMUIRt HIHT1—
o
IKFtVENT
FROK PRIHMT
SE1UIIII TANKS
RBC PILOT PLANT
z
o
«€
a
K
a
4.4
3.3
2.1
>• 1.1
UJ
in
-x.
3:
runic 2
2.3
1.7
i.l
•o
o.
IX
o
««
o
en
a
0.58 =>
UJ
in
15 20 25 30 35
ELAPSED TIME, MINUTES
PHASE VI AND PHASE VII CYCLICAL FLOH PATTERN
SIMULATING DIURNAL VARIATION
-------
FIGURE 3
i.o
0.5
at
a
o
HYDRAULIC LOADING, gpd/ft2
SBOOs REMOVED vs HYDRAULIC LOADING
I-.B
i.o
UJ
o:
2
l-« 2.0 3,8
HYDRAULIC LOADING.'gpd/ft*
80D5 REMOVED vs HYDRAULIC LOADING
4.9
FIGURE 4
541
-------
a
a
CO
v>
z
UJ
o o
t 2 3
HYDRAULIC LOADING, gpd/ft2
EFFLUENT SBOOg *s HYDRAULIC LOADING
FIGURE 5
10
B
IS
10
J.B
2.0
1.0
HYDRAULIC LOADING, gpfl/ft2
EFFLUENT BOD5 vs HYDRAULIC LOADING
FIGURE 6
4,0
542
-------
FIGURE 7
ion
o
3
(X
If)
o
o
00
I*.
»?
o
t i.e . 2.0 a.a 4.9
HYDRAULIC UJAOING,gpo/ft2
EFFICIENCY (* SBOD5 REMOVAL) »s HYDRAULIC LOADING
IBB
at
§ '
?B
UJ
U
FIGURE 8
1.0 2.0 3.0
HYDRAULIC LOADING,gpd/ft2
EFFICIENCY^* BODg REMOVAL) HYDRAULIC LOADING
4.0
543
-------
FIGURE 9
i.o
e>
a
O
a
LU
O
CQ
C/3
0.5
0 0.5 t.O 1.9
ORGANIC LOADING, SBQ05 iD/IOOOft2day
SBOOs REMOVED « SBOD5 LOADING (MEAN VALUES]
FIGURE 10
a
tr
•v.
»*
•«-*
*" 10
CO
o
a
S'
a"
UJ
0
2 8.5
ec
in
a
0
m
to
0
*
O
a
A
A
4-
X*
'MUSE 1
PHJSE 11
nun in
PH»SE i»
PHtSE <
PH»SE (1
PH1SE T!
• £
:
oo *
F'
,
a
a
..
o ./r
*
D
-
t .X^
0
If - 0.802 I - 0
COEFFICIENT = 0
"
02
968
0.5 1.0 1.5 2.8
ORGANIC LOADING. SBOOg Ib/1000 ft^/tiay
SBOOj REMOVED vs SBOD5 LOADING (DAILY VALUES)
544
-------
3.11
FIGURE 11
ta
•a
~ J.O
t.g
in
o
o
z.
0.5 1.0
ORGANIC LOADING, SBQOg ID/1000
1.9
80D5 REMOVED vs SBOOj LOADING
FIGURE 12
100
in
a
o
X
LU
O
10
I 0-5 1.0
ORGANIC LOADINS, SBODg ID/1000 ft2/day
1.8
EFFICIENCY (55 SBODg REMOVAL! vs SBOD5 LOADING
545
-------
FIGURE 13
100
"
LU
g
u
u.
"
98
o o.s i.s ).:
ORSANIC LOADING, SBOD5 Id/1000 ft2/day
EFFICIENCY (1 BQD5 REMOVAL) *s SBODg LOADING
CD
V>
FIGURE 14
10
t O.J 1.0
ORGANIC LOADING, SBOOg tb/1000 ftVday
EFFLUENT SBOOs vs SBOD5 LOADING
546
-------
41
30
O
to 20
10
FIGURE 15
0.5 1.0 t.S
ORGANIC LOADING. SBOD5 lb/1000 ftVday
EFFLUENT BOD5 vs SBOD5 LOADING
FIGURE 16
1.0
0.9
O
3=
UJ
ac
O
CD
1.0 2.0 3.0
ORGANIC LOADING, 60DS lb/1000 ftVday
SBOD5 REMOVED vs BOD5 LOADING
4.0
547
-------
REMOVED, 1 ft/ 1000 ftVday
— • ** **
KM «a ca
in
C3
CD
B
oX
y
o
O S^
X
^
o
/
X
3 1.0 2.0 3.0 4.
ORGANIC LOADING, BQ05, lb/1000 ftVday
FIGURE 17
0
BOD5 REMOVED vs BOOs LOADING ' M* AN VALUES »
4.B
«•• 3.0
— 2.0
in 1.0
o
» MtS
• PIUS'
-A PHIS
• a PHI:
• II
• I
: ft
MH4-
T « D.B8
CtEFFICI
- 0.16
:MT ' 0.95S
FIGURE 18
t.O 2.0 3.0 4.0 S.O
ORGANIC LOADING, B005 Ib/tOOO ft2/day
BOD5 REMOVED is 80D5 LOADING (DAILY VALUES)
6.0
548
-------
GO
V)
10
FIGURE 19
1.0 2.0 3.0
ORGANIC LOADING, BQ05 Ib/IOOO ftVday
EFFLUENT SBOD5 vs BOD5 LOADING
4.0
FIGURE 20
4ft
1
0
g 20
z
u.
u. ..
0
^
o^-9'
O
O ^*
^*
s^
^^
o
0
^
^
0
U ID ID
ORGANIC LOADING. B005 lb/1000 ftVday
EFFLUENT BOD5 vs BOD5 LOADING
4jJ
549
-------
O
z
LLJ
u
9C
70
10
FIGURE 21
1.0 2.0 3.0
ORGANIC LOADING, BOD5 • ID/1000 ftVday
EFFICIENCY (% SB005 REMOVAL) vs BOD? LOADING
4.1
100
90
tO
'e 70
z
UJ
u
tZ
fc 60
FIGURE 22
1.0 2.0 3.0
ORGANIC LOADING! BOD5 Ib/IOOO ftVday
EFFICIENCY (S B005 REMOVAL) vs B005 LOADING
4.0
550
-------
too
80
_ 80
^
ex
40
20
-B ir
okL
0
on
e
• 0
COBREUTIOK
COEFFICIEKtT- 0.838
20 40 80
80
SS, Rlg/l
tOO 120 140
BOD5 REMOVAL vs. SS REMOVAL
DURING SETTLING PROCESS
FIGURE 23
40
30
10
0 0.} t.O 1.5
ORGANIC LOADING; SBOO: ID/IOOO ftVday
SETTLED EFFLUENT SS vs SBODs LOADING
FIGURE 24
551
-------
FIGURE 25
in
in
to
HI
II
cc
C/l
a
tu
D-
f>
=>
to
Tl
It
SI
41
II
>\
\t\
\*
V *
\>
\
\ o
Y-i
\
•
:\
a • v
* *
I.'*
*
a
V
\
\ *
S
V
\
«x
o
. *
O
4
0
\
\
\
•
0
•
0
A
&
4-
^
X
2 6
•
•
o
^
4
a
o
\
\
\
mm i
mm ii
mm in
ruAsi n
fHA5E 1
mm vt
misE vii
Nv
%
O
• o *-
+
2000 4000 6000 tOOO
OVERFLOW RATE gpd/ftj
10000
4 SS REMOVAL vs SURFACE OVERFLOW RATE
3
FIGURE 26
O o
0.5 1.0 1.5
ORGANIC LOADING, SBOD, ibs/1000 ft'/day
SLUDGE PRODUCTION RATIO »* SBODs LOADING
-------
•SBRFACT: CURRENT DEVELOPMENTS
AND
PROCESS APPLICATIONS
By
Jeffrey B. Cowee
Application Engineer
Autotrol Corporation
Milwaukee, Wisconsin
and
Richard A. Sullivan
Manager, Process & Applications Engineering
Autotrol Corporation
Milwaukee,
INTRODUCTION
A new secondary wastewater treatment concept called the SURFACT Process has
been developed for upgrading existing activated sludge plants to higher cap-
acities, with no additional land usage and little or no increase in energy
consumption. The Surfact process is a biological system resulting froa the
combination of a rotating fixed film contactor, called Bio-Surf, and- an act-
ivated sludge system. The process can be applied to augment any overloaded
activated sludge system or for new treatment facility construction. Revamp-
ing of existing systems is enhanced by minimal capital and operating expendi-
tures while allowing for unusually rapid construction scheduling.
553
-------
In the Surfact process, rotating biological contactor (RBC) media assemblies
are installed in the aeration basin of an activated sludge system, as shown
in Figure 1. The media for the rotating contactor is. fabricated from poly-
ethylene plastic. Air cups are then attached to the periphery of the media
to capture the diffused air, which is the preferred means of driving the
shaft. The air being diffused at the bottom of the aeration tank provides
rotation of the shafts primarily- through capture in the air cups, although
some assistance is derived from the hydraulic roll of the tank fluids. A
supplementary air header, using air from the main blower air supply, can be
installed for additional control of rotational speed.
The combination of fixed film growth and suspended growth within a single
tank, provides additional biological solids In the system, making higher
treatment capabilities possible. The result can be a higher treatment level
at the same flow rate or the same level of treatment at an increased flow
rate, or a'combination of increased flow rate at increased treatment levels.
By merging the two systems, the Surfact process increases the efficiency and
stability of the existing activated sludge system,, inherent process benefits
include:
- Higher Treatment Efficiency - Higher levels of treatment in excess
of 90% result from increased "sludge age" by adding fixed bio-*
logical culture to the existing suspended culture.
— Process Stability — Because of the high biological solids inventory
on the media, the Surfact system is less susceptible to process
upset from hydraulic or organic shock loads.
- Flexibility - The Surfact process lends itself to upgrading treatment
facilities because of its modular construction and low hydraulic head
loss.
- Maintenance and Power Consumption - Minimal maintenance is required
on the RBC units. Utilization of the existing aeration system
achieves higher treatment efficiencies with minimal additional power.
- Low Initial Cost - The Surfact concept permits upgrading facilities
with minimal capital expenditures. Little or no plant modification
is required to install the Aero-Surf units in existing plants.
- Ease of Nitrification - By providing the proper sludge age, the
Surfact process allows for, nitrification to proceed without, costly
separate aeration, settling and sludge recycle systems. Under
proper loading conditions, the Surfact process provides both 8005
removal and nitrification in the tankage.
- Improved Sludge Characteristics - Resulting sludge has improved
settling characteristics, permitting secondary clarifiers to be
designed for'relatively higher overflow-rates. When upgrading plants
for higher degrees of treatment, additional secondary clarifier
surface area will not be requir/ed.
554
-------
SURFACT PROCESS
Rotating
Biological Contactor
Media Rotation
Air Cups
Supplemental Air Header
V :;avv: :;v:^
Air Diffuser
FIGURE 1
END VIEW SKETCH BIO-SURF MEDIA SUPERIMPOSED
ON AN ACTIVATED SLUDGE AERATION TANK.
HYDRAULIC ROLL AND AIR CAPTURE WITHIN THE TANK
CAUSES ROTATION OF THE MEDIA.
AUTOTROL CORP.
555
-------
PILOT PLANT OPERATIONS AND HISTORY
The Surfact process is currently undergoing pilot and full-scale testing. To
date, both the full-scale demonstration plant testing and the pilot scale
testing is proving the concept to be a very effective and economical means
of upgrading existing activated sludge plants. The process has demonstrated
capability to improve flow characteristics, improve loading rates and success-
fully provide single stage nitrification at loading rates in excess of those
considered practical for single stage activated sludge nitrification systems.
Due partially to the success of Philadelphia Phase I studies, several other
pilot facilities have been designed and placed in operation throughout the
world. A detailed discussion of these investigations would be both cumber-
some and incomplete at this point. However, data generated from some of
these studies is included in tables and figures as noted.
A brief summary of the ongoing pilot studies appears in Table 1., Unfortunate-
ly, of the six pilot facilities, data is available fron only three. The data
available from two of these units has only recently become available in a
preliminary and incomplete fashion.
Philadelphia, Pennsylvania
Based on the predicted process advantages, the first pilot work was initiated
by the City of Philadelphia, Pennsylvania in 1974. All testing was performed
at the Northeast Water Pollution Control Plant in one of the existing aera-
tion basins. The primary objective of inital studies was to determine
whether the RBC units would rotate under the conditions of captured power
from the aeration system. Other objectives included observation of the ten-
dencies to support uniform biomasses and structural observations related to
clogging of the media. The studies also included observation of•the con-
tactors rotational speeds during periods in which air was supplied by means
of the special supplemental diffuser pipe directly under the Aero—Surf unit.
The unit was tested at two tank locations and observations of the biological
growth were made at each location. Preliminary results were favorable to the
point that additional testing was conducted.
Since favorable indications were observed during the initial evaluation,
additional testing of the Surfact system continued with the installation of
one full-scale air driven shaft in the existing aeration tank. The air
driven shaft incorporated a plurality of circumferential cups placed on the
media periphery such that a major portion of shaft rotational power could be
accomplished with existing aeration tank air supply and hydraulic motion.
Due to the capture of aeration tank energy, minimal additional power is
required to obtain shaft rotation.
After favorable operating results were obtained with the full-scale shaft,
a prototype installation of 22 shafts incorporated in two passes of the
existing tank was undertaken. This installation necessitated the isolation
of an existing secondary clarifier to avoid .the intermixing of the existing
suspended activated sludge culture and the Surfact mixed liquor.
556
-------
(J1
(Jl
Location
Domestic
Philadelphia, PA
Milwaukee, WI
(Autotrol)
Reno-Sparks, NV
Foreign
Japan
Japan
Sweden
Table 1
SUMMARY OF ONGOING SURFACT PILOT STUDIES
Unit
type Size Start-Up
Carbon Only ,„ Full 9/77
Single Stage
Nitrification Bench 9/79
Single Stage Bench 10/79
Nitrification
1
Carbon Only Intermediate Fall '79
Single Stage Bench 12/79
Nitrification
Carbon Only Intermediate 12/79
Expected Data Included
Completion In Paper Comments
. Not Known Yes Heavy
Industrial
Influent
Winter '80 . Yes Influent
Solids
Control
Difficult
Summer '80 Yes Phostrip
process
& Return
Sludge
Plugging
Not Known No No Data
Available
Not Known No. No Data
Available
Not Known No No Data
Available
-------
A settled sewage pumping station provided a known controlled flow to the
isolated reactor. A separate return sludge pumping station was also provided.
Because of the simplicity of gurfact retrofit, construction was completed in
only sixty days.
Operation of the full-scale demonstration plant was initiated in mid-October
of 1977. Stabilization of operation occurred by the end of December if 1977.
The most significant start-up problems were related to return and waste
sludge pumping facilities.. Virtually no structural clogging problems were
observed in relation to the internal structure of the disc units. By the
end of eleven months of testing, several significant combinations of conditions
had been observed and a generous amount of data had been recorded.
The conditions of testing and results generated appear as Tables 2 and 3,
respectively. An in-depth discussion of the data accumulated during Phase I
was presented by Michael D. Nelson, et al, at the 51st Annual WPCF Con-
ference.
Following a brief down-time, a second testing phase was initiated. The
aerated tank was dewatered and the existing coarse bubble diffuser system
was removed from under the Aero-Surf units and replaced, with a. relatively
high efficiency fine bubble, ceramic type diffuser system.
Data included in Tables 4 and 5 demonstrate the performance with fine bubble
diffusers.
Milwaukee, Wisconsin
The pilot plant in Milwaukee, Wisconsin is operating at the Autotrol Corporate
Testing Facility located at the South Shore Wastewater Treatment Facility. The
unit is receiving primary clarifier effluent from the channel feeding the
existing activated sludge reactors. The Surfact unit is operating in parallel
with the activated sludge plant. Operating data summaries appear as Tables
6 and 7.
Two primary objectives were established for the Surfact Study in Milwaukee.
These were to verify operating results from Philadelphia and to establish
operational parameters for nitrification. In essence, verification of Philal-
elphia parameters would also establish a suitable environment for at least
partial nitrification. This is due to the fact that under many of the test
conditions and environmental conditions were adequate to allow the establish-
ment of nitrifying cultures. It is believed that .nitrification was not
achieved at Philadelphia due to a significant degree due to toxic or inhibi-
tory industrial influents.
The Period I study at South Shore was aimed at verifying the validity of a
single stage nitrifying concept. The conclusions from this period are pre-
sented in Table 7. The overall conclusion is that single stage nitrification
is viable and economically feasible. Definitive analysis of parameters and
environmental conditions will be undertaken in Period III of the study.
558
-------
Ul
U1
TABLE 2
PHILADELPHIA, PA.
CHRONOLOGY OF OPERATION
I
PERIOD NO. START DATE STOP DATE FLOW SRT DESCRIPTION AND
._
_.
I
II
III
IV
V
VI
9/14/77
10/1/77
12/23/77
1/31/78
2/27/78
4/4 /78
5/26/78
7/17/78
10/1/77
12/4/77
1/9/78
2/14/78
4/2/78
5/12/78
7/16/78
8/3/78
12
12
12
12
10
15
10
10
MGD
MGD
MGD
MGD
MGD
MGD
MGD
MGD
LOW
3.
3.
4.
5.
4.
7.
1.
3
3
4
5
4
0
7
Preliminary start up and debugging.
Start up problem resolution period and debugging
Low tank suspended solids SRT and normal combined SRT. (1)
Normal tank suspended solids SRT and high combined SRT.
Normal suspended solids SRT.
Low tank suspended SRT and medium combined SRT..
High tank suspended SRT and high combined SRT,
Operation without benefit of return sludge -
non-equilibrium and non-acclimated culture.
VII 8/4/78 8/13/78 10 MGD 1-7 Operation without benefit of return sludge -
pseudo-equilibrium and acclimated culture.
(1) Tank suspended solids refers to those solids contributed by the activated sludge process.
Combined solids refers to those contributed by both the activated sludge process and air driven
rotating biological contactor media surface slimes.
Source: Nelson, et al.
-------
TABLE 3
PHILADELPHIA, PA
OPERATIONAL DATA
I
TEST PERIOD:
START DATE
END DATE
REM3VAL:
SBODs
TBODs
SS
INFLUENT:
SBOD5
TBODc
SS
EFFLUENT:
SBODs
TBODs
SS
TOTAL REMOVAL:
TBODs
SS
SECONDARY CLARIFIER
SOR f
I
12/23/77
1/9/78
(*)
W
(*)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(rag/1)
(mg/D
(*)
(%)
FIER
/ft2)
ME(hrs)
83
82
72
82
181
115
10
31
29
85
86
720
2.8
II
1/31
2/14
93
87
84
97
198
163
7
25
26
88
87,
720
2.8
III
2/27
4/2
93
87
84
101
183
153
7
23
23
89
89
600
3.3
IV
4/4
5/12
86
76
67
86
190
136
. 12
44
42
78
81
900
2.2
V
5/26
7/16
95
91
82
92
158
100
5
14
18
92
93
600
3.3
VI
7/17
8/3
87
85
79
91
159
97
12
24
20
89
92
N/A
N/A
VII
8/4
8/13
88
86
83
65
135
107
7
. 18
18
91
93
600
3.3
-------
TABLE 3 (Cant.)
PHILADELPHIA, PA
OPERATIONAL DATA
PHASE I
in
H1
TEST PERIOD: I
START DATE i?m/77
END DATE 1/9/78
PF APT AT* •
lU-lrtUlUJv.
LOADING
(iTBODgA/lOOOft2)
(#SBODsA/1000ft2)
(#TBODsA/1000ft3j
(#SBODsA/1000ft3J
F/M
WITH RBC
WITHOUT RBC
SRT
WITH RBC
WITHOUT RBC
DETENTION TIME (hrs)
AIR RATES: (NOTE BELOW)
AIR/VOLUME (ft3/gal)
AIR/LOAD
(100Qft3/#TBOD5R)
. (1000ft3/#SBOD5R)
OXYGEN/LOAD
f & n-"*A tfrFPnT^rDl
(^ it U £/\" 1 £SUU sKJ
(#02A#SBOD5R)
POWER CONSUMPTION:
POWER/LOAD
(KWH/'TBQDgR)
8.40
4.03
67.80
32.50
0.50
0.99
3.28
1.77
4.0
0.9
0.65
1.35
0.49
1.02
0.61
1.27
II
1/31
2/14
8.99
4.68
73.94
38.48
0.34
0.56
4.42
2.82
4.0
1.2
0.70
1.35
0.50
0.96
0.59
1.13
III
2/27
4/2
6.94
4.08
57.09
33.54
0.30
0.75
5.54
2.70
4.8
1.2
0.87
1.48
0.62
1.06
0.76
1.29
IV
4/4
5/12
10.83
5.49
89.09
45.16
0.55
1.74
4.44
1.92
3.2
0.8
0.51
1.01
0.42
0.83
0.50
1.00
V
5/26
7/16
6.01
3.63
49.43
29.86
0.27
0.43
6.97
4.35
3.3
1.1
0.90
1.49
0.61
1.01
0.71
1.18
VI
7/17
8/3
6.69
3.83
55.05
31.48
0.79
10.50
1.07
0.08
3.3
0.6
0.49
0.45
0.26
0.52
0.46
0.79
VII
8/4
8/13
5.12
2.54
42.14
20.89
0.55
8.28
1.72
0.13
3.3
0.4
0.40
0.81
0.29
0.58
0.58
1.17
(KWH/ffSBODsR)
NOTE; All air rates quoted at a diffuser efficiency of 3%
-------
Ul
-------
in
tn
LO
TEST PERIOD:
START DATE
END DATE
REMOVAL
SBODs
TBOD5
SS
INFLUENT:
SBOD5
TBODS
SS
EFFLUENT:
SBODs
TBODs
SS
TOTAL REMOVALS:
TBODs
SS
SOR
TABLE 5
PHILADELPHIA, PA
OPERATIONAL DATA
PHASE II
I LA IB II
11/14/78 11/14/78 11/30/78 12/10/78
12/7/78 11/29/78 12/7/78 3/7/79
III IIIA IIIB IV V
3/1/79 3/1/79 3/13/79 5/26/79 6/9/79
4/25/79 4/12/79 4/25/79 6/7/79 6/19/79
(V
CD
(*)
Cmg/l)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
LS:
(*)
(1)
^RIFIER:
/ft2)
riME (hrs
91
87
80
95
183
148
8
24
28
89
90
900
,) 2.2
91
86
79
93
181
148
8
25
29
88
90
900
2.2
92
88
82
101
187
149
8
21
26
91
92
900
2.2
80
64
53
89
168
158
17
57
68
75
73
891
2.5
92
88
85
92
176
220
7
20
25
91
90
360
5.5
91
87
85
94
181
236
7
22
26
90
90
360
5.5
94
90
85
89
165
168
5
16
22
93
91
360
5.5
93
87
83
69
129
107
5
17
18
92
93
420
4.7
97
91
84
79
144
116
2
13
17
94
93
480
4.1
-------
(Jl
en
TEST PERIOD:
START DATE
END DATE
REACTOR:
LOADING
(WBODrA/lOOOft2,)
(fSBODsA/lOOOft2)
(fTBODsA/lOOOft3)
(#SBODsA/1000ft3)
F/M
WITH RBC
WITHOUT RBS
SRT
WITH RBC
WITHOUT RBC
DETENTION TIME (hrs.)
AIR RATES:
AIR/VOLUME(ft3/gal)
AIR/LOAD
(1000ffV//TBOD5R)
(1000£t3/#SBOD5R)
OXYGEN/LOAD
(002A///TBOD5R)
TABLE 5 (cont)
PHILADELPHIA, PA
OPERATIONAL DATA
II
I IA IB II III IIIA IIIB IV V
11/14/78 11/14/78 11/30/78 12/10/78 3/1/79 3/1/79 3/13/79 5/26/79 6/9/79
12/7/78 11/29/78 12/7/78 2/7/79 4/25/79 4/12/79 4/25/79 6/7/79 6/19/79
10,41
5.42
85,64
44.59
0.47
0.75
4.58
2.80
3.2
10.30
5.28
84.74
43.44
0.76
2.69
3.2
10.63
5.72
87.43
47.05
0.47
0.71
4.54
3.01
3.2
11.47
6.05
94,34
49.76
0.57
1.17
.4.66
2.56
2.7
8.33
4.35
66.00
34.47
0.27
0.35
5.74
4.56
4.0
8.40
4.38
67.78
35.34
0.27
0.36
6,20
4.79
4.0
8.24
4.43
61.64
33.14
0.28
0.33
5.85
4.98
4.0
7,53
4.01
56.30
29.98
0.35
0.44
4.20
3.39
3.4
9.62
5.26
71.95
39.34
0.46
0.58
8.00
6.36
3.0
POWER CONSUMPTION:
POWER/LOAD
(KWH/#TBOD5R)
(KWH/#SBOD5R)
An error in the reporting and computation procedure has
prevented inclusion of this data at time of this printing.
NOTE: All air rates quoted at a diffuser efficiency of %.
-------
TABLE 6
MILWAUKEE, WISCONSIN
OF
PERIOD NO. START DATE STOP DATE FLOW
5/16/79 7/16/79 Variable
I 7/26/79 8/20/79 950 gpd
II 9/4/79 Summer80 Variable
III Fall- 80 SummerSl Variable
DESCRIPTION AND COMMENTS
Preliminary Start-up and debugging
Verification of single stage nitrification capability
with Surfact. Process successfully completed.
Verify Philadelphia Phase I results and determine
approximate design parameters.
Determine design parameters for single stage nitrifica-
tion facilities.
U1
in
Average daily flow
Operational Mode
Maximum flow
Minimum flow
Average RAS flow
Average WAS flow
Removal Efficiencies:
TBODs -—=-944
SBOD5 = 941
NOTE: Due to operational difficulties with
TABLE 7
MILWAUKEE, WISCONSIN
OPERATIONAL DATA
FOR
TEST PERIOD I
SINGLE STAGE NITRIFICATION VARIFICATION
950 gpd Av. Influent Characteristics:
diurnal flow reariation TBODs = 231> mg/1
1.200 gpd SBOD5 - 93 mg/1
290 gpd TSS = 236 mg/1
570 gpd SCOD = 135 mg/1
50 gpd NH3-N = 26 mg/1
HC03 = 250 mg/1
PH =7.7
Temp. = 19° C
D.O. = 0.9 mg/1
secondary clarifier samples are decanted 24 hour
composits of Surf act Tankage Effluent. Biomass WAS and MS
samples were grab.
Av. Effluent Characterist.;
TBOD5 " 15 mg/1
SBOD5 - 6 mg/1
NH3-N = 5 mg/1
Plant Operating Conditions:
F/M overall =0.29
Reactor;
MLSS » 3,051 mg/1
MLSS = 2,197 mg/1
SHAFT:
MLSS - 1,100 mg/1
MLSS = 891 mg/1
-------
Test Period II is currently ongoing. Results appear to fall within the
ranges predicted by Philadelphia. However, several conclusions and recommenda-
tions are becoming clear. Specifically, the process is seeing a break in
SBODc in the front end of the reactor basin, indicating that step feed or
step aeration would allow for a significant increase in flow and load
capability. Concurrent with these increases, effluent quality would be
improved.
Reno-Sparks, Nevada
The pilot plant in Reno, Nevada is operating at the Reno-Sparks Wastewater
Treatment Facility. The unit is operating in a similar fashion to the unit
at South Shore. One major exception is noted in that phosphorus stripping
process called PHOSTRIP is in use. The return method from this process
appears to be affecting the activated sludge significantly but does not
appear to have a major or adverse impact on the fixed film solids inventory.
The effect on the reactor basin solids inventory is a significant reduction
in concentration.
The activated sludge system appears to have entered a recovery mode shortly
after initiation of the PHOSTRIP process. Further investigation 'into the
performance reduction indicates possible plugging of solids flow lines within
the return sludge system. Proper maintenance of these lines may significantly
increase performance and effluent quality.
The data generated, although not presented here, indicates a significant
stabilizing effect of the fixed film. In conjunction with the data from
South Shore and Philadelphia, this data indicates an inherent capability of
Surfact systems to provide stable treatment under conditions which would
otherwise shut down conventional activated sludge systems.
CONTACTOR INSTALLATION
Several methods of contactor installation have been proposed. The two
primary categories are shafts mounted in a "longitudinal" axis and those
mounted in a "transverse" axis. Several advantages are common to both
systems.
Longitudinal placement may be defined as locating an RBC unit parallel to the
longest dimension of the tankage as shown in Figure 1. Thus, in a standard,
spiral roll activated sludge tank, the shafts would be mounted end to end in
such a fashion as to take advantage of hydraulic roll concurrently with air
capture. This physical placement may be accomplished by means of either
supporting the shafts off the bottom of the tank, or by hanging a cantilever
across the existing wall structure.
Several advantages are seen for this type of installation. The shafts and
mounting structure could be assembled exterior to the existing tankage. This
allows for installation of a complete shaft and support structure in one
crane lift. This reduces the amount of. time required for installation and the
length of time an individual basin would be out of service.
566
-------
Transverse installation could be defined as mounting shafts 90° from the
longest dimension of the tank. In esscr\.cs,. the simplest fashion of install-
ing shafts is to secure the bearings to .existing • wall atructures-' This
installation allows for the advantages of completely preparing the tank for
installation without dewatering, shaft installation without a supplemental
structure and a requirement for one single crane lift. Figure 2 is an
example of this installation method.
Length of time out of service is of major importance to most wastewater treat-
ment facilities requiring upgrading. In many cases the process of removing
tanks from service seriously hinders effluent quality and must be avoided.
GENERAL DISCUSSION
Indepth study of the data generated indicates that Surfact systems function in
a similar fashion to activated sludge processes. Similarities can be seen in
SRT relationships, effectiveness of RBC biomass versus suspended biomass,
oxygen uptake rates, F/M relationships, retention time and sludge characteris-
tics.
Percent Removal Versus SRT
Figure III shows the relationship between SRT and percent removal of 8005,
soluble BOD5 and suspended solids. The relationship depicted .is similar to
that expected for activated sludge with percentage removal increasing with an
increasing SRT. Significant increases in percentage removal occur below an
SRT of 5 to 6 days. Above six days the expected leveling off occurs. Percent-
age removal of soluble BOD ranges from 83 - 97 percent with the majority of
data in excess of 90 percent.
The consistent performance depicted is indicative of a strong biological
process.
Iffectiyeness of Fixed Growth
One of the questions of primary importance is a determination of biological
growth effectiveness. In specific, a determination of relative effectiveness
of the suspended growth versus that maintained on the RBC surfaces is requir-
ed. Table 7 contains data comparing operation during a period of no sludge
return to periods high suspended mass-concentrations. The indication is that
fixed film masses are at least as effective as suspended masses.
Uptake RateStudies - Biofilm VersusSuspended Culture
Several studies have been performed on the oxygen uptake rates of fixed film
versus suspended film reactors. Table 8 is a compilation of the k rate
determinations conducted at Philadelphia. As shown, the specific oxygen uptake
rates are roughly equivalent. The work done by the disc and suspended cultures
is thus shown to be roughly equivalent on a unit mass basis.
567
-------
FIGUrtE 2
RBC UNIT
WATER LEVEL
ACTIVATED SLUDGE TANK
SURFACT PROCESS WITH RBC UNIT
INSTALLED IN TRANSVERSE POSITION
568
-------
UI
en
UJ
cc
H
z
UJ
o
a:
UJ
a.
100
90
80
70J.
FIGURE 3
SURFACT PERCENT REMOVAL VS SRT
SBOD
8
SRT (days) AUTOTROL CORP 1/80
-------
Table 8
Philadelphia, Pennsylvania
Biological Mass Effectiveness Comparison
PHASE II II I
PERIOD JEI VI VI
SBOD5 Applied (//) 13,361 12,460 6,755
SBODS Effluent 17 88
Xfflg/1)
% SBOD5 Removal 80 90 87
MLVSS - Total (#) 45,786 36,812 19,566
MLVSS - RBC (5) 44 35 93
tf SBOD5 REM/#MLVSS 0.24 0.31 0.30
F/M with RBC 0.57 0.66 0.69
570
-------
Ul
-J
Table 9
Philadelphia, Pennsylvania
Uptake Rate Studies
Oxygen Uptake Ratio
PERIOD
.III
IV
V
FLOW
10
15
10
SRT
5.5
4.4
7.0
LBS BOD/
1000 cf
57
89
49
rag 0?/mg
Suspended
.022
.063
.039
VSS/hr
Biofilm
.022
.048
.035
K rate
tng?0 /1/hr K 10
Suspended
35
76
74
Biofilm Suspended
33
78
39 .160
Biofilm
,174
-------
Data generated from several RBC installations confirms the uptake rates on
fixed film reactors presented here. Sources of data are found in the
history and bibliography prepared for this seminar by Ron Antonie and Dannette Lank.
Volatile Mass Comparison
In comparing the concentrations of volatile masses on the RBC shafts no the
mixed liquor there is an average of 7.9 percent more volatile mass per
pound on the disc surface area than appea-ring in the mixed liquor. Consider-
ing all cases, there is a range of 0 - 18 percent improvement in the volatile
solids per pound of total solids maintained on the RBC as opposed to the
suspended mass maintained within the aeration tank. Therefore, from a volatile
solids concentration standpoint, the disc splids appear to be more viable.
Based on Table 10, treatment capacity appears to be independent of the MLVSS
location and primarily dependent on total biological solids inventory within
the reactor system.
Thickening Test Results
A series of pilot-scale thickening tests were conducted on the waste activated
sludge generated from the Philadelphia Surfact Facility. The investigations
included both gravity and dissolved air floatation thickening. Several runs
were performed on both secondary sludge and combined secondary and primary
sludge.
The gravity thickening results fell within the normal expected ranges for
activated sludge plants. Average results indicated concentration in execess
of 2% with a 3 hour detention time.
Results from the floatation thickening study were impressive. Secondary
sludge, introduced at 1.0% concentration, thickened to 4.5% without addition
of supplemental chemicals. The resulting capture rate was 99% at an air to
solids ratio of 0.008. These ranges are within extremely economical operating
ranges while performing at well above expected concentration increase and
capture rates.
CONCLUSION
Conclusions drawn from the data presented show the Surfact Process to be
superior to a standard activated sludge process based on treatment reliability,
adjustment to varying flow and load conditions and reduced dependence upon
secondary clarifier operations. The process can readily be applied to over-
loaded activated sludge systems or to the construction of new treatment
facilities. Upgrading is enhanced through the minimal construction time
required, ease of operation and low power consumption.
The combination of fixed and suspended solid growth within a single reactor
tank provides a significant increase in biological solids without increasing
solids loads on the secondary clarifier. Thus, the capability for treatment
is not dependent totally upon the capture, settling and return of solids to
572
-------
Table 10
Philadelphia, Pennsylvania
Comparison of Treatment Based on MLV.S.S.
PHASE I I II
PERIOD II III I?
F/M with RBC 0.34 0.30 0.35
SBOD5 Removal (%) 93 93 93
MLVSS-Stispended 2163 1577 2184
(mg/1)
MLVSS-Shaft (mg/1) 1373 1512 560
MLVSS-Shaft (% of 39 49 29
Total)
SBODS Applied (#) 9708 8423 8056
Effluent SBOD5 7 75
(mg/1)
573
-------
the activated sludge process. The result can be a higher treatment level .
at the same rate of flow or the same level of treatment at a significantly
higher rate of flow or a combination of increased flow and treatment level.
The Surfact Process, by merging two systems, increases the efficiency and
stability of the existing activated sludge system. Inherent process benefits
include:
- Higher Treatment Efficiency
- Increased Process Stability
- Increased Flexibility
- Reduced Maintenance and Power Consumption
- Low Initial Cost
— Ease of Nitrification
- Improved Sludge Characteristics
- Increased Mass Without Increased Oxygen Provision
Ease of nitrification has been demonstrated at the South Shore Wastewater
Treatment Facility. The use of single stage reactors enhances the upgrading
of existing activated sludge systems. This concept is based upon carbon
removal within the existing tank structure and nitrification primarily attain-
ed by the fixed film.
The data related to single stage nitrification and high load periods indicates
that a Surfact system is significantly more stable than a standard activated
sludge facility. Based on loading parameters, the Surfact system can be
operated well beyond the bounds of standard activated sludge operation,.
574
-------
PLANT SCALE INVESTIGATION OF SBC
PROCESS SUPPLEMENTAL AERATION
By
R. Srinivasaraghavan
Associate, Greeley and Hansen
Carl W. Reh
Partner, Greeley and Hansen
James Canaday
Deputy Engineer - Director
Alexandria Sanitation Authority
1. BACKGROUND
The advanced wastewater treatment plant at Alexandria, Virginia in-
cludes preliminary treatment, primary settling, Rotating Biological Con-
tactor (RBC) secondary treatment, carbon adsorption, phosphorus removal,
filtration, ion exchange for nitrogen removal and chlorination. A process
flow diagram of the plant is shown on Figure 1. This is the largest plant
in operation in the U.S.A. employing RBC process at a design capacity of
54 mgd and was designed based on pilot study data. The secondary treatment
consists of 56 motor driven RBC shafts in 14 tanks of 4 stages each. The
design criteria is shown in Table 1.
The RBC process was placed on line in 1977. The operating results
indicated less than expected performance. The RBC discs developed a very
thick biological growth white in color. This was identified as beggiatoa.
The beggiatoa predominance was evident in all stages of the RBC tanks. On
the lead stages, where the discs were covered with beggiatoa, a black
under layer was observed. It appeared that this under layer was not aerobic
due to the thickness of the growth.
575
-------
SCREEN
»-
GRIT
REMOVAL
»-
PRIMARY
SETTLING
RBC
SECONDARY
SETTLING
in
~J
CTl
DISINFECTION
ION
EXCHANGE
FILTRATION
TERTIARY
SETTLING
CARBON
ADSORPTION
FIGURE I
ALEXANDRIA ADVANCED WASTEHATER TREATMENT PLANT
PROCESS FLOW DIAGRAM
-------
TABLE 1
ASA Plant Design Bases
(mg/1)
Flow 54 mgd
Inf. BODs 220 Eff. BOD5 3
Inf. SBOD5 30 Eff. SBOD5 1
Inf. N 42 Eff. N 1
Inf. P 15 Eff. P 0.2
-------
The start-up data are compared with the pilot study data and design
bases in Table 2. The average influent SBODg during start-up period was
significantly higher than during the pilot study period. It is noted that
the SBOD5 during start—up averaged 13 mg/1 higher than pilot study data.
Any increase in RBC effluent SBOD5 would be removed by carbon adsorption
process which would, therefore, require more than anticipated level of per-
formance by the carbon columns and more frequent regeneration of the carbon
bed.
An experimental program was developed to investigate the apparent
differences in the RBC performance observed during the start-up and pilot
study periods. This paper presents the information on tests conducted and
the findings and conclusions of this dtudy.
2. EXPERIMENTAL PROGRAM
The experimental program was designed to eliminate unlikely hypo-
theses and converge on an approach to improve treatment efficiency. The
first part of the experimental program comprised the following:
o SBODg tests on composite samples before and after treatment
with coagulant chemicals. The effects of 100 mg/1 Fed3
and 200 mg/1 alum were studied.
o Batch carbon adsorption tests on 24-hour composite samples
of the RBC effluent. The initial and final SCOD concentra-
tions were measured after two hours of contact with Calgon
"Filtrasorb 400" activated carbon, 10 grams per liter.
The second part of the experimental program included on-line testing
as follows:
The north RBC tanks were used as control units while the south
RBC tanks were used as test units to study the operational
modifications shown on Figure 2. The modifications were:
Tank
No. Description of Experiment
2 Addition of hydrogen perioxide to first stage (5 mg/1)
4 Supplemental aeration in each stage (180 cfm)
8 Operation in 1-1-1-1 mode
10 Operation in 3-1 mode
12 Operation in single-stage mode
The supplemental aeration equipment could not be installed in time
to collect data in parallel with other on-line experiments conducted. The
effect of supplemental aeration was studied at a different time on a long
term basis subsequent to the completion of other on-line tests.
578
-------
TABLE 2
RBC Pilot, Design and Start-Up Data
Parameter
Pilot
Design
Start-Up
Flow, mgd
Loading
GPD/1,000 Ft2
PPD SBODs/1,000 Ft:
Influent SBOD5, mg/1
Effluent SBOD5, mg/1
Chemicals Addition
54
30
11.5
3.6
38
12
Alum
9.6
2.5
30
12
Alum
5.4
2.7
60
25
None
579
-------
R8C IHFLUENT COMPOSITE
SAMPLING POINT—^
• R8C TANK NUMBER
DESIGNATION
N
PR MARY
EFFLUENT
t-u
ae
UJ
ul V
UJ
V.
©
"(15)
.(35)
_@
•H
nTf^
2-1-1 MODE ~~t
2-1-1 J
2-1-1 S
2-i-l >
2-1-1 !
2-1-1
2-1-1
(ODE ~~~C
100E — -G
-------
During the test period, 24-hour composite samples were collected by
automatic samplers at each sampling point shown on Figure 3. The samples
were analyzed for SBOD5, BOD5, COD and suspended solids concentrations.
3. RESULTS AND DISCUSSIONS
a. Coagulation Tests
The SBOD5 concentrations of coagulated and uncoagulated samples are
shown in Table 3. There are significant differences between the values.
The SBOD^ concentrations in the samples analyzed were reduced by coagulation
with both ferric and alum. Data are summarized as follows:
Ferric Alum
Number of sets 5 9
Reduction of SBOD5, mg/1:
Mean 16 11
Maximum 21 28
Minimum 12 4
Based on the mean of all samples, the reduction in SBOD^ by coagulation
with ferric was greater by 5 mg/1 than by coagulation with alum. The ex-
pected SBODcj reduction by ferric coagulation is on the order of 15 mg/1.
b. Carbon Adsorption Tests
The SCOD concentrations were used to indicate the performance of
carbon adsorption. The SBOD5 concentrations are not considered reliable
indicators in the low ranges expected. The SCOD values observed before and
after contact with carbon in four tests are shown in Table 4. In all of
the tests, the SCOD concentration was significantly reduced by carbon
adsorption; the mean reduction was 77 percent. Based on a BOD/COD ratio of
0.33 (as was observed during the pilot work), an initial COD concentration
of 40 mg/1 and a 75 percent reduction, the estimated SBOD^ in the effluent
would be as follows:
Initial SCOD, mg/1 40
Percent reduction 75
Final SCOD, mg/1 10
BOD5/COD Ratio 0.33
Estimated Final SBOD^, mg/1 3
c. On-Line Testing
In this section, all on-line test data except the supplemental aeration
test are discussed. The supplemental aeration test data are discussed in the
next section.
An overall summary of the SBODj- results on composite effluent samples
from the RBCs in each stage configuration, along with their control, is
presented in Table 5. Hydraulic and organic loadings to all RBC tanks for
this test period (Jan. 17 - Feb. 17, 1978) are summarized as follows:
581
-------
TABLE 3
Coagulation Test Data Summary
Date
FeCl3, 100
2/4/78
2/5/78
Alum, 200
2/13/78
2/14/78
2/15/78
2/16/78
Sample Point
mg/1
RBC Inf.
RBC Eff . - ID
RBC Eff. - 2D
RBC Inf.
RBC Eff. - 2D
Mean of
all samples
mg/1
RBC Inf.
RBC Eff. - 2D
RBC Inf.
RBC Eff. - ID
RBC Eff. - 2D
RBC Inf.
RBC Eff. - ID
RBC Eff. - 2D
RBC Eff. - 9D
Mean of
all samples
Before
Coagulation
58
41
30
44
28
40
73
37
39
32
26
63
42
28
11
40
SBOD5/ mg/1
After
Coagulation
40
29
16
23
12
24
45
24
35
14
18
53
34
24
12
29
Reduction
18
12
14
21
16_
16
28
13
4
18
8
10
8
4
9
11
582
-------
TABLE 4
Activated Carbon Test Data Summary
Sample
Date Point
2/13/78 2D
2/14/78 2D
2/15/78 2D
2/16/78 9D
Mean
Median
Initial
44
30
43
39
39
43
SCOD^11 , mg/1
Final
3
14
14
6
9
9
Reduction
41
16
29
33
30
34
Percent
Reduction
93
53
67
85
77
79
(1) Coagulated with. alum.
583
-------
TABLE 5
Summary of On-Line Data'-1-'
SBOD5 mg/1
Sample
Point
inf.
ID
2D
7D
in
CO
*> 8D
9D
10D
11D
12D
Stage
Configuration
NA
2-1-1
2-1-1
2-1-1
1-1-1-1
2-1-1
3-1
2-1-1
Single
No, of
Samples
20
20
20
20
20
17
1
20
20
Max.
93
71
92
80
74
60
57
55
55
Min.
41
24
12
19
19
21
19
27
19
Range
52
49
80
61
55
39
38
28
36
Mid-
range
67
48
52
50
47
41
38
41
37
Mean
61
47
41
37
35
36
37
37
36
Median
55
44
36
33
32
31
35
35
34
Remarks
Control : 2D
Peroxide
Control : 8D
4-Stage Mode
Control: 10D
2-Stage Mode
Control : 12D
Single-stage Mode
(1) No coagulant chemicals were added to the plant or to the samples before analysis.
-------
Max. Min. Mean
Sewage quantity, mgd 54 36 44
Hydraulic loading, gpd/sf 9.6 6.4 7.9
Influent SBOD5, mg/1 93 41 61
Organic loading, ppd SBOD^/
1,000 sf 6.5 2.6 4.0
The suspended solids concentrations were widely scattered. This is attri-
buted to nonrepresentative solids samples obtained by the sampling equip-
ment. For this reason, it is considered that the SBODc; concentrations are
the primary performance indicators in this study.
The average effluent SBOD5 concentration from the RBCs on the
1-1-1-1, 3-1 and single stage configurations and the respective control
units which were operated on the 2-1-1 mode did not differ significantly,
ranging from 35 to 37 mg/1. These results indicate that in terms of SBODc
removal, the performance of the RBCs is not affected by stage configuration.
Tank No. 2, where peroxide was added to the first stage, showed a
mean effluent SBODc, of 41 mg/1 compared to 47 mg/1 for its control, Tank
No. 1. The peroxide unit performed marginally better than its control unit
with about 10 percent difference in SBOD^ removal.
The poorer performance of Tanks Nos. 1 and 2 (average effluent SBODc
of 41 and 47 mg/1) compared to Tanks Nos. 7-12 (average effluent SBODc of
35 to 37 mg/1) may be due to additional aeration of sewage along the RBC
influent channel. It has been observed all along that the biology at the
west end tanks included greater numbers of the higher order growths and less
beggiatoa growth than the east end tanks. It apperas, therefore, that
aeration may help in improving the biological growth and enhancing SBOD,-
removal.
Peroxide treatment did affect the biology observed on the first stage
of RBC Tank No. 2 compared to its control tank. Before the test program
began, the growth on the RBC media in the first pass of Tank No. 2 appeared
similar to that observed at Tank No. 1, with patches of white biomass,
beggiatoa, mixed with brown growth on top of a black underlayer. After two
days of peroxide treatment, the white patches disappeared from the peripheral
layers of the media where very thin areas and bare spots were observed.
After ten days the white patches had entirely disappeared from the leading
passes of Tank No. 2 and the bare spots were replaced with anew thin layer
of brown growth.
The peroxide addition appears to have .had a positive effect on bio-
logy as well as SBOD^ removal in the tank where the beggiatoa growth was
noted. While the improvement in treatment efficiency due to peroxide could
not be quantified based on these results, it is anticipated that peroxide
addition would control the beggiatoa growths especially during summer con-
ditions.
585
-------
d. Supplemental Aeration
The supplemental aeration test data analysis has been divided into
two phases, related to the addition of FeCl3 to the primary settling tanks,
as follows:
o No chemical addition
o Daily chemical addition
(1) No Chemical Addition
During this phase, the effect of single stage aeration versus four
stage aeration was studied. The average RBC effluent SBODc, values for the
control and single stage supplemental aeration tanks were 19 and 16 mg/1,
respectively. The data shows some marginal improvement in SBODc, removal
with single-stage supplemental aeration. When the aeration was extended to
all four stages of the RBC, the average effluent SBODc decreased to 11 mg/1,
while the control unit produced an effluent with SBOD5 of 18 mg/1 for the
same period. The daily influent and effluent SBODc, values for this p"hase
are shown on Figure 3. It is noted that fluctuations in the effluent SBODc
concentration in the supplemental aeration tank had been dampened considerably
compared to the control unit which follows influent SBODc closely.
The daily SBOD,. removals are correlated to the applied load for the
control and supplemental aeration units on Figures 4 and 5. It is evident
from these plots that supplemental aeration unit performance was superior
to the control unit. The improvement in performance is higher when aeration
is provided to all four stages than single stage aeration. Therefore, only
four stage aeration was studied in the subsequent phases.
(2) Daily Chemical Addition
During this phase, 20 mg/1 of ferric chloride as Fe and 0.5 mg/1
polymer were added to the primary settling tanks. The influent SBOD5 to
RBC's averaged 32 mg/1, over 150 days of data collected. The average ef-
fluent SBODc for the control and supplemental aeration units for the same
period were 14 and 6 mg/1, respectively. Some statistical data on these
values are shown in Table 6. The daily influent and effluent SBODc, values
for the two units are plotted on Figure 6.
The SBODc removal versus applied load relationship for the two units
is shown on Figure 7. It is to be noted that the specific removal rate ex-
pressed in pounds of SBODc, removed per unit surface RBC area per unit time
is clearly higher for the supplemental aeration unit than for the control
unit. It is also observed that the difference in removal becomes more
pronounced as the loading increases as indicated by the diverging regression
lines shown on Figure 7. The correlation coefficients as noted on Figure 7
for the two lines are 0.83 and 0.93 testifying to the excellent correlation
between the two parameters plotted.
586
-------
70
SIHGIE PASS
AERATION
.FOUR PASS
AERATION
iG
50
Qfl
E
O
r- 40
O
o
o
m
o
o
CD
C/3
20
MFLUEMT
-CONTROL TAHK
EFFIUEMT
•AERATED TAMK
EFFLUENT
30 40 SO
TIME, (DAYS)
80
70
FIGURE 3
INFLUENT AND EFFLUENT SBOD5 CONCENTRATION
KITH NO CHEMICAL FEED
80
587
-------
Ul
00
oo
SBOD5 LOADING, PPD/1000 SF
FIGURE H
SBOD5 REMOVED VERSUS SBOD5 LOADING
FOR SINGLE STAGE AERATION
-------
3.0
Ul
03
1.0 2.0
SBOD5 LOADING, PPD/1000 SF
FIGURE 5
SBOD5 REMOVED VERSUS SBOD5 LOADING
FOR FOUR STAGE AERATION
-------
TABLE 6
On-Line Supplemental Aeration Test Data with Daily
Chemicals Addition
Parameter Minimum
Flow 18
Influent 88005 Cone., mg/1 H
Control Unit Effluent
SBOD5 Concentration, mg/1 4
Supplemental Aeration Unit
SBODs Concentration 1
Ul
o Control Unit SBODs
% Removal 1°
Supplemental Aeration Unit
% Rfimnva 1 44
Maximum Mean
51 27
59 32
24 14
15 6
88 54
99 81
Standard
Deviation
5
9
4
2
14
10
Mode
28
32
14
5
50
83
Median
27
32
14
5
56
83
-------
60
AERATED TANK EFFLUENT
60 70 60
TIME, (DAYS)
FIGURE 6
INFLUENT AND EFFLUENT
SBOD5 CONCENTRATION
DAILY CHEMICAL FEED
MO
-------
OO
o
o
a
a,
o_
CD
«S
O
to
a
o
CD
00
t-
h
t-
C
CD _
Sc
a »-
•
CD _
CD i-
00 «
CO
a
LU
O -
LU
O
CO
oo
fc-^ —
dS OOl/Qdd 'Q3AOW3H SQ09S
592
-------
The effluent SBOD5 concentrations for the supplemental aeration and
control units are plotted against the loading on Figure 8. While the data
are -scattered, as anticipated, it is clear that the effluent quality from
the supplemental aeration tank is superior to the control tank effluent in
terms of SBOD concentration. It is also seen from Figure 8 that the effect
of loading on effluent SBODc is more severe for the control unit than on
supplemental aeration unit.
_(3)f Summary
The supplemental aeration test data are summarized in Table 7. It
is noted that the aerated unit produced effluent of better quality in every
case. The average values of SBODc removal for single stage and four stage
aeration with and without chemical addition were 60, 69 and 81 percent,
respectively, while average percent removal in the control unit was approxi-
mately 50 percent. The average SBOD- concentrations for these three modes
of operations were 16, 11 and 6 mg/1.
4. COST COMPARISON FOR INSTALLATION OF THE RBC SUPPLEMENTAL AERATION
The treatment criteria require plant effluent BODc and SBODj- con-
centrations of 3 and 1 mg/1, respectively. Removal of the SBODg is achieved
both in the RBCs and the activated carbon columns. Therefore, lower
efficiency of the RBCs requires higher efficiency of activated carbon column
and vice versa. The tests described above have shown that SBOD^ removal
efficiency of the RBCs can be improved by providing supplemental aeration to
the existing motor driven RBC units. An analysis was performed to compare
the costs of removing the SBODc in RBC process versus its removal by carbon
adsorption. The following assumptions were used in costs development:
RBC Performance
The mean effluent SBODg concentrations of 6 mg/1 from aerated
tank and 13 mg/1 from unaerated tank are assumed.
Carbon Regeneration
A uniform carbon exhaust rate of 8.2 Ibs. per Ib. SBODg removed
is assumed for regeneration requirements. A 6 percent loss of
carbon during regeneration is assumed for make-up carbon re-
quirements .
Operating Costs
The operating costs used for the cost estimate are as follows:
Labor, dollars per man-year 15,000
Make-up carbon, dollars per Ib. 0.50
No. 2 fuel oil, dollars per gallon 0.525
Electricity, dollars per kilowatt-hr. 0.0317
Ferric chloride (28 percent), dollars
per dry ton 128
Polymer, dollars per Ib. 1.20
593
-------
U1
vO
*»
1.0 2.0
SBOD5 LOADING, PPD/1000 SF
FIGURE 8
EFFLUENT SBOD5 CONCENTRATION VERSUS SBODs LOADING
FOR FOUR STAGE AERATION WITH DAILY CHEMICAL ADDITION
-------
TABLE 7
Summary of Supplemental Aeration
Test Data.
I/I
ID
(Jl
Mode of Operation
Without Chemical Addition:
Single Stage Aeration
Four Stage Aeration
Daily Chemical Addition:
Pour Stage Aeration
Average
Plant SBOD5, mg/1
Flow Influent Effluent
.mgd Control Supp. Air
30 40 19 16
30 35 18 11
30 32 14 6
SBODs Removal, %
Control Supp. Air
53 60
49 69
56 81
-------
Maintenance Costs
The annual maintenance costs for equipment were estimated at 5
percent of the capital cost.
Supplemental AerationEquipment Costs
The supplemental aeration equipment capital and installation costs
were quoted by the manufacturer of the RBC units (Autotrol Corpora-
tion) at 53,125 per unit. The initial capital cost was amortized
at a 7 percent interest rate over the assumed design life.
The alternatives investigated for cost analysis are as follows:
Alternative
No. 1 No RBC supplemental aeration
No. 2 Supplemental aeration installed at
four stages of each RBC tank (for
total of 56 units installed).
Five-year assumed design life.
Both alternatives include chemical addition to primary tanks with
carbon adsorption following RBCs to produce same quality effluent.
A comparative cost summary of the two alternates is shown in
Table 8.
The chemical cost is common to both alternates at $492,000 per
year. The carbon regeneration and make-up cost is highest for
Alternate 1 at $540,000 per year. This is due to higher exhaustion
rate because of higher effluent SBOD5 from RBCs without supple-
mental aeration. Carbon regeneration and make-up cost for
Alternate 2 is $194,000 per year.
The initial cost of providing supplemental aeration to RBC tanks
is estimated at $175,000 for 4 stage aeration. The annual costs
including the amortized and O&M costs for supplemental aeration
for Alternate 2 is $86,500.
The total annual costs for the alternates are $1,032,000 and
$774,000, respectively. The corresponding unit costs per pound
of SBODg removed are 93* and 70£, respectively.
The cost evaluation, under the assumptions used show the instal-
lation of RBC supplemental aeration equipment is cost effective.
596
-------
TABLE 8
Cost Summary for SBODs Removal
Basis: Plow 36.5 MGD
in
1.
2,
Chemical
Item
feed
to
primary
Carbon regeneration and
tanks ,
make-up,
$1000 per year
$1000 per year
No. 1
492
540
Cost of Alternate (1)
No.
492
194
2
3. RBC supplemental aeration equipment:
Capital cost, $1000
Amortized cost, $1000 per year
Operating cost, $1000 per year
Total annual cost, $1000 per year
4. Total cost for 88005 removal:
$1000 per year
Dollars per Ib, SBODs removed by overall process
Dollars per 1,000 gallons sewage
1,032
0.93
0.077
175
43
46
89
774
0.70
0.058
(1) These costs do not include the cost of existing facilities such as RBC's, carbon columns,
and regeneration equipment.
-------
5. CONCLUSIONS
The following conclusions can be drawn from the tests conducted.
o A portion of the organic material observed in the filtrate of
primary effluent and RBC effluent samples, and measured as
SBODs is due to a colloidal fraction. Based on the overall
mean SBOD5 reduction of primary effluent and RBC effluent
samples treated with ferric chloride, this fraction appears
to be about 15 mg/1.
o The three RBC operational modes tested did not indicate signi-
ficant difference in performance.
o The peroxide treatment controlled beggiatoa predominance and
established a healthy culture on the RBC discs. Increases in
SBODg removal on the order of 10 percent were experienced
using perioxide.
o The supplemental aeration of the RBCs essentially eliminated
beggiatoa population on the discs. The thickness of biomass
growth on the aerated discs was much thinner compared to the
unaerated units.
o The average SBODc removal was increased by about 27 percent
across the RBCs. The average effluent SBOD^ concentration
was reduced by 57 percent. The fluctuations in SBOD^ in the
effluent was attenuated considerably.
o The comparative cost analysis indicate that the removal of
SBOD5 would be cheaper by supplemental aeration of RBCs than
by carbon adsorption for Alexandria Treatment Plant.
598
-------
" EFFECT OF SUPPLEMENTAL AIR ON ROTATING BIOLOGICAL CONTACTOR
PROCESS DOMESTIC WASTE"
By
J.T. Madden & R.B. Friedman
Clow Corporation
Envirodisc Systems
Beacon, New York
Introduction
The general acceptance of Rotating Biological contactors for secondary
and advanced wastewater treatment has led to the divulging of many different
opinions concerning the use of various techniques to improve the wastewater
treatment process with this equipment. One such modification that has
received widespread interest among consulting engineers and designers has
been the use of air to aid the biological oxidation that occurs on the media
surface area. Much of the efficiency of the Rotating Biological Contactor
can be credited to the relatively high concentration of biological growth
that is exposed alternately to the food source in the wastewater and oxygen
in the atmosphere. This vigorous contact in a short detention time is not
matched in any other treatment process. We are going to look at the work
that has been done thus far in the use of supplemental air provided for
mechanically driven Rotating Biological Contactors.
DISCUSSION
Before we look at the direct application of supplemental air in varying
quantities.to wastewater of varying strengths, we are first going to review
the work done by Warren Chesner, Roy F. Weston Company , concerning the use
of air with the Rotating Biological Contactor process. In their paper pre-
sented at the New York Water Pollution Control Federation in January, 1977.
it was stated that increased amounts of dissolved oxygen in the wastewater
stream of the Rotating Biological Contactor process above 3 to 4 mg/1 did
599
-------
not materially effect the performance of the Rotating Biological Contactor
process. The data also showed that thero was virtually no effect on the
suspended solids nor was there any appreciable effect in the reduction in
soluble or total BOD as a result of maintaining a high dissolved oxygen.
O
Chesner's conclusion supports earlier work by Welch who found that a
mixed liquor dissolved oxygen above 1.5 my/1 did not significantly affect
the Rotating Biological Contactor process.
On the other hand, information developed from Alexandria, VA as shown
on Table 1, indicates that copious amounts ol supplemental air added to a
mechanical drive Rotating Biological Contactor System, did affect the
effluent quality. At this installation, approximately 200 cubic feet per
minute of air was added for each Rotating rcioloqical Contactor unit with all
of the air available to add oxygen rather than a substantial portion used
for rotational force. The addition of air did reduce the soluble BOD, but
had little or no effect on the total BOD. Refer again to Table 1. This
installation is reported to be a side by side comparison of wastewater having
similar characteristics of temperature, BOD, suspended solids and nutrients.
In both of the cited examples, it should be noted that the treatment process
for the Rotating Biological Contactor is being performed on what is commonly
called domestic wastewater. By definition, domestic wastewater normally has
influent strengths of 150 to 300 mg/1 BOD and suspended solids, 15 to 35 mg/1
ammonia nitrogen and a temperature range from 40°F ho 75°l-'.
Now that we have examined both ends of a determined spectrum for this
presentation, let us examine what is happening in between the extremes and
see if a conclusion can be drawn for further study, investigation and use.
At the Purdue Industrial Wastewater Conference in 1979, Chou & Hynek
presented information on Autotrol*s South Shore Pilot Testing Program in
Milwaukee, WI. This program compared their mechanical drive and a prototype
air drive treating domestic wastewater at relatively low soluble BOD loadings
and when using their relatively closed media.4 The media is termed as being
relatively closed as the flow of wastewater and air throughout the interior
surfaces of the media are dependent on a limited number of radial flow
passages. The data presented shows that side by side runs for comparison did
have equal influent concentrations, but different hydraulic and absolute
organic loadings. Chou & Hynek concluded that the addition of air via the
prototype air drive system did increase the soluble BOD removal.
STUDY AT COLD SPRING, NY
A study was started in June, 1979 at the Village of Cold Spring, NY
Wastowciter Treatment Facility, using a Clow Envirodisc Rotating Biological
Contactor having a total nominal surface area of 11,000 SF. The tankage pro-
vided was designed to enable testing of each of the four stages independently
of the other stages. The system was also designed to allow varying quantities
of air to be discharged at the bottom of the tank in any stage and its effect
measured independently of any of the other stages involved in the study.
During the testing period from June through November, 1979, two parallel two
stage flow paths were created, equal in all ways, except that air was added
to one of the flow paths. The amount of air added was equivalent to 150% of
the non-motive or biologically active air that is available in an air driven
600
-------
Rotating Hiolocjical Conl:actor. ((.0 ei-'M/100,OOO »!••) .
Figure 1 shows the testing locations, sampling locations and a general
outline of the flow diagram showing how independent and separate testing
could be achieved on wastewaters that were identical in all significant
characteristics. During the first phase of the study, relatively high
loadings of BOD were applied to determine whether or not 3 Ibs. soluble
BOD/1000 SF/day is the relatively low maximum removal rate as stated by the
Autotrol Corporation.
Figure 2 shows loading rates of up to 14 3bs/1000 SF with resulting
removal rates of up to G lbs/1000 SI-'. Upon completion of this first phase,
successive phases at increasingly lower loadings were run.
Table 2 summarizes all of the data from June through November, 1979 and
removals as a result of adding supplemental air to an open media at
rates equal to 150% of the non-motive air in an air drive system.
It is interesting to note and quite germane to the title of this
presentation that by introducing supplemental air to wastewater of the
strength that these wastewaters had, there was virtually no difference
in removals of soluble BOD between the Clow path that had supplemental
air and the flow path that did not have any supplemental air, again
as shown on Table 2. It is also noted, that the substrate loadings and
removals in the initial phases were significantly above.those which are
normally encountered in waste treatment plants that have been designed
for treatment of domestic wastewater, while latter phases of the testing
program were run at loadings normally found in such facilities.
Figure 3 shows the percent of removals at various "normal" loadings with
a 6ne stage process, both for the Cold Spring Study using Clow's open media
and a study using closed media. Again by looking at this data, it can readily
be seen that no significant reduction in soluble BOD occurred as a result of
adding supplemental air. Similarly, Figure 4 shows the percent of removals at
various loadings with a two stage process. The points for the closed media
are taken directly from Autotrol*s South Shore, Milwaukee Test program
published data, while the points for open media are taken directly from Clow's
Cold Spring Study.
Conclusions of the South Shore program show that closed media with an air
drive did reduce the soluble BOD loadings as compared with a closed media
mechanical driven Rotating Biological contactor. At Alexandria, VA as
mentioned earlier, the addition of larqo amounts; of stipplemontal air to the
same type of closed modia also reduood the solublo HOD, althouqh not. tho
total BOD. One could conclude then i'roro those two studies that the addition
of supplemental air to a relatively closed type media can and does improve
the removal of soluble BOD.
On the other hand our studies at Cold Spring, using a relatively open
media with supplemental air showed that there is no increase in the soluble
BOD removal or advantage to providing supplemental aeration at the loading
rates studied. If we then take the published data of pounds applied versus
removal rates for the closed media and compare them at the same loading rates
with Clow's open media as in Figure 3 and 4, we see that regardless of
601
-------
whether its mechanical drive or supplemental air to a mechanical drive or
air drive, that the type of media that is open and that doer? not depend upon
radial passages, appears to consistantly remove a greater percentage of the
applied soluble BOD.
CONCLUSION
The data presented here clearly shows that with the use of a relatively
open Rotating Biological Contactor media, the addition of supplemental air
to enhance wastewater treatment with mechanically driven Rotating
Biological Contactors does not offer any significant improvement over units
that do not have supplemental air provided. Secondly, by comparing
published data on soluble BOD removals usinq a closed media with our
Studies using an open media, the open type media Rotating Biological
Contactor removed a greater percentage of soluble BOD regardless of whether
additional or supplemental air was provided, and finally, the open type
media as used in the Cold Spring Study does not have an upper removal rate
limit of 3 lbs/1000 SF/day for soluble BOD as reported for the closed type
media; removal rates of up to 6 lbs/1000 SF/day were observed.
602
-------
REFERENCES
1. Chesner, W., Roy Western Engineers, "Effect of the Dissolved
Oxygen on the Rotating Biological Contactor Process.*1
presented at N.Y.W.P.C.P.A. 1977 Annual Conf., New York,
M.V.
2. Welch, P.M. (1968) "Preliminary Results of a Hew Approach in the
Aerobic Biological Treatment of Highly Concentrated Wastes,"
Proceedingsof the 23rdPurdue Industrial Waste Conference,
Purdue University, Lafayette, Indiana.
3. Strinivasaraghavan, R,, Greeley and Hanson Engineers, Personal
Commun i cat ions, November, 1979.
4. Hynek, R.J. and Chou, C.C.S., Autot.rol Corporation (1979)
"Development and Performance of Air-Driven Rotating Biological
Contactors.1* Presented at 1979 Purdue Annual Industrial
Waste Conf,
5. Friedman, R., and Roeber, J., "Energy Reduction Consideration for
the RBC Process." Presented at. Energy Optimation of Waste and
Wastewater Management for Municipal and Industrial Application
Conf., December, 1979, New Orleans, LA.
6. Autotrol Corporation, "Wastewator Treatment Systems (1979) "Design
Manual." Pages B-ll and D-3.
603
-------
TABLE 1
SUMMARY OF DATA FROM ALEXANDRIA, VA
MODE No. OF BODr SBODq SS
STAGES mg/T mg/K rag/1
Avg of 5 month's M ^3 15 48
testing MS M 8 50
Without chemical M 1 19
addi tion MS 1 16
M 4 28
MS k 11
With chemical M 4 13
addition MS k 6
¥i = Autozro! Keen an leal Drive R.5.C.
MS= Autotrol Mechanical Drive R.B.C. with 200 CFM supplemental
Air added per shaft
604
-------
TABLE 2
COLD SPRING, NY TEST PROGRAM (1)
AVERAGE OF ALL TESTING
June 28, 1979 ~ November 31, 1979
157.1
kj.k
86.3
^0.9
A. 67
*». 19
113-2
119-*
Total BOD5 Soluble Dissolved Total (2)
BOOr Oxygen Suspended
Solids
mg/1 mg/1 mg/1 mg/1
In f1uent
Mechanical Drive
1 st Stage
Mechanical Drive k6.3 39-8 4.51 118.5
with supplemental
air - 1st Stage
Mechanical Drive 30.1 28.6 5-19 101.1
2nd Stage
Mechanical Drive 30.2 27-6 5.21 106.1
with supplemental
air - 2nd Stage
(1) Clow Envirodisc Model A-10
(2) TSS - Prior to clarification
605
-------
Tc
PLAHT
.
i
i ' i
L_; _j
RBC
s,
T
k/
FLO*/ wren (sc
COLD SPRING, NY
P/LCT
'££ A-/O
aooc FT 2
606
-------
is 14.0
5
I3.t>
§
V)
§
IQ.O
^ 9.t>
^
8.0
7.0
A5^
A A
A A
A A
AA
A A
AA
70
FIG. 2
S BOD
ay CLOW /?&c AT HIGH
607
-------
7.0
6.0
n
fc
-------
Z.S
J
1.0
I
5
8:
O
CQ
1,5
0.5
RPM~ 1.5-
T * 55*'
J$ l/Tfo T"A
O H
A CtOtV
A *
7O
t//>f>'L
A A
O
A A
A A
«
O A A
2 BOO
FIQ, 4- Ft/zsr
-------
-------
USE OF SUPPLEMENTAL AIR TO CORRECT AN OXYGEN LIMITATION
CONDITION OF AN OPERATING RBC SYSTEM
By
Joseph F. Lagnese, Jr.
Duncan, Lagnese and Associates, Inc.
INTRODUCTION
Howes Leather Company is a vegetable sole leather tanner with two operat-
ing plants, one in Curwensville, Pennsylvania and one in Durbin, West Virginia.
The original treatment system for each plant relied upon a combination of
primary lagoons for solids removal and storage, and secondary aerated lagoons
for some biological reduction of organics. In 1975, a six month waste charac-
terization and bench scale treatability evaluation of the Durbin plant wastes
was undertaken in preparation for enhanced treatment capability at both plants.
This effort revealed the wastes to be of high lime content and high
organic strength. Alhough amenable to biological treatment after significant
pretreatment, the persistant tannin content of the wastes demonstrated adverse
effect in the operation of the dispersed growth experiments, particularly with
solids separation. Thereafter, in the late winter and summer of 1976, pilot
plant studies of activated sludge and rotating biological contactors (RBC)
were conducted at the Durbin tannery. This evaluation favored the use of the
rotating biological contactors as the secondary biological process for the
application required at the Curwensville, Pennsylvania tannery, which had the
earliest compliance schedule of the two tanneries.
For the Curwensville plant, the State of Pennsylvania established ef-
fluent standards limiting total biochemical oxygen demand during warm weather
conditions to 900 pounds per day, calculated as the total of 1.5 times the
effluent five day biochemical oxygen demand (BODj.) and 4.56 times the effluent
ammonia nitrogen. Based on the estimated flow of 200,000 gallons per day and
the predicted influent ammonia nitrogen level of 75 mg/1, not anticipated to
be signicantly reduced through the treatment process, the effluent BOD,, would
have to be controlled to a level of about 131 mg/1. For the period of Novem-
ber through April, the State permits the BOD,, to increase to 228/mg/l.
611
-------
The two tanneries were considered to be similar in the tanning procedures
utilized and in the resulting wastes. The only difference in the two plants
was in the production level, with Curwensville processing about 1400 hides per
day and Durbin, about 1900 hides per day.
PILOT PLANT EVALUATION
The pilot plant work at Durbin provided the design basis for the Cur-
wensville treatment plant.
Pretreatment for the pilot plant was comprised of primary settling,
equalization, pH adjustment and phosphorus enrichment. An 8200 square foot,
four stage pilot plant RBC system, leased from the Autotrol Company, was used
for the test work. It was operated at a rotary speed of 2.9 rpm and a
peripheral speed of 60 fpm, the same peripheral speed as is standard for this
manufacturer's full scale units. After preliminary experimentation with general
flow rate ranges, a more detailed evaluation, upon which the design criteria
were formulated, was carried out at flow rates in the general range of 0.25
gpd/sf, to 0.75 gpd/sf.
The characteristics of the influent to the pilot plant RBC are summarized
in Table 1. The soluble portion (SBOD) of both the influent and settled
effluent BOD_ averaged 85 percent.
Graphical presentations of the pilot plant data are given in Figures 1
and 2. It is evident that as the SBOD loading increased, the specific removal
rate increased, but the overall efficiency of removal decreased. At all test
loadings, the SBOD removal efficiency remained above 88 percent.
Within the hydraulic and SBOD loadings tested, the major influence on
performance appeared to be SBOD loading. For the limited amount of per-
formance data obtained at comparable SBOD loading, a small improvement in SBOD
removal was observed for decreased hydraulic loading, but there was too little
data to determine with much certainty the quantitative significance of this
indicated effect.
The level of performance for the first stage reactor at the lowest and
highest hydraulic loading is shown graphically in Figure 3. Although showing
a slightly less consistency of performance, the low specific hydraulic loading
appears to show no dramatic sign of performance limitations due to either
oxygen or biomass limitation. However, at the highest hydraulic and SBOD
loading, the system appears to have reached a maximum SBOD removal rate of
about 9 pounds per day per 1000 square feet. At this level of removal, it is
apparent that the system was oxygen and/or biomass restricted. Having only
limited and somewhat conflicting dissolved oxygen data for the first stage, it
cannot now be determined with certainty the specific cause of this indicated
removal limitation.
However, it is now apparent that the oxygenation capability of the pilot
plant's first stage was disproportionally higher than for a full scale system
at the same peripheral speed. This was unfortunately not realized at the time
the pilot plant's overall performance was being evaluated for development of
612
-------
TABLE 1
RBC INFLUENT CHARACTERISTICS
PILOT PLANT STUDY
Average Range
TBOD1' 939 mg/1 585 - 1290 rag/1
SBOD 798 mg/1 500 - 1080 rag/1
COD 1650 mg/1 1400 - 1870 mg/1
TSS 250 mg/1 140 - 790 mg/1
TKN 65 mg/1 37 - 88 mg/1
NH3 - N 58 mg/1 35 - 81 mg/1
pH 7.4 (median) 6.4 - 8.2
Temperature 68°F 55 - 75°F
Total 5-day Biochemical Oxygen Demand
613
-------
4.0
U
k mo
"
i
1,0
0.5 +
LEO END
Q 0.23 OPO/ SF
0,4? GPO/ SF
o O.M ere/
PILOT PLANT PERFORMANCE
1.9 2,0 IJ S« 3.9
S»OO APPLICATION HATi (UIS/OAY/ IOOOSF )
614
-------
4.0
3.5
- 10
LINE OF UNIMUM PERFORMANCE
I «
Q 0.23 6*>0/SF
• OAT 0*0/ «F
« 0 «i wo/ SF
U
LO
03*
0
.FiCURg J2^
PLANT PERFORMANCE
IO
10 40
so «o TO w *o OB 10 no tao 140 iao
MOO trPUJENT CONCCNTIttTON (IMA.)
ITO
615
-------
II JO-
CO-
i*o-
? luD-
•*«• .
r
SJijo-
! ic-1-
ID-
FIRST STAGE PERFORMANCE
- PILOT PLANT TESTING -
ID-
U>
SB 40 SO 6.0
70 ao ao 100 uo izo lio 1*0
SBOO APPLICATION RATE I LBS/DAY/IOOO SF )
&0
ITO
616
-------
design criteria for the full scale installation. The problems relating to
scale-up limitations due to the higher oxygenation capability of the pilot
plant's smaller disc sizing (regardless of comparable peripheral speed control)
did not come to our attention until the full scale problem was upon us.
Although BOD was the most closely monitored parameter for performance
measurement of the RBC, some performance data for chemical oxygen demand
(COD), total suspended solids (TSS) and nitrogen was obtained. Effluent COD
varied between 227 and 436 mg/1. Effluent TSS averaged 38 mg/1 with a range
of 20 to 96 mg/1. No nitrification was observed for any of the testing,
although conversion of organic nitrogen to ammonia nitrogen through the RBC
was consistently observed.
DESIGM BASES OF RBC FACILITY
The design bases for the Curwensville RBC facility is summarized in
Table 2.
Allowing for no change in the predicted 75 mg/1 influent ammonia con-
centration through the RBC, a plant discharge TBOD of 131 mg/1 (219 pounds per
day) would thus be required for the total oxygen demand to remain within the
900 pounds per day summertime limit of the DER and EPA permits. On a SBOD
basis, the effluent concentration to be achieved by the RBC was set to be 112
mg/1. Based on the estimated influent SBOD concentration of 800 mg/1, the 1BC
would be required to remove 1149 pounds of SBOD per day.
Referring to Figure 2, a removal rate of 3.25 pounds of SBOD per day per
1000 square feet was indicated for this required effluent concentration. A
total surface area of 353,000 square feet was indicated.
To provide a margin of safety and to work within the manufacturers'
standard sizing of four stage systems, a unit comprised of four shafts, each
25 feet long, and each having a rated surface area of 100,000 .square feet, was
selected and installed. The total surface area of 400,000 square feet, re-
lated to the design hydraulic loading and required SBOD removal, provides a
specific hydraulic loading of 0.5 gpd/sf and a specific SBOD removal rate of
2.9 pounds per day per 1000 square feet of total shaft area.
A flow diagram of the treatment system, as it was initially provided, is
given in Figure 4.
INITIAL OPERATING EXPERIENCES
The construction of the plant was completed in early 1978, with all
treatment units in operation by April, 1978. It became evident soon there-
after that the RBC facility was seriously stressed under the conditions which
then prevailed. The symptoms were ominous - significant presence of hydrogen
sulfide in and around the RBC facility, unfavorable growth conditions on the
surface, and depleted dissolved oxygen levels in the first stage.
Concurrently, it was discovered that the waste load to the RBC was much
greater than the design anticipated. A summary of this initial waste load
characterization is given in Table 3.
617
-------
TABLE 2
DESIGN BASES FOR
SIZING RBC FACILITY
CURWENSVILLE TANNERY
Flow (Equalized) 200,000 gals/day
Soluble BOD
Loading 800 mg/1 1336 Ib/d
Required Effluent 112 mg/1 187 Ib/d
Required Removal 688 mg/1 1149 Ib/d
Minimum Sizing Criteria, Based
On Pilot Plant Findings (Figure 2)
4-Stage Removal Rate 3.25 Ibs/day/lOOOsf
Total Area Required 353,000 sf
Selected RBC System
4 - Stage, 25 ft. Shaft Per Stage
Total Area Provided 400,000 sf
618
-------
POLYMER-
CHLORINE
rn
RAW WASTE
PUMPING
ITZS ! 1 1
; PHMARY ' EQUALIZATION •
JCLARIFIER 1 TANK j
s1
ii
* "f-4 "»StKKE TO UBOOK
-"tuf STEAM FOR
TEMP CONTROL 1
\
1
RECEIVING
STBEAW
/^''^CHIOBINE!^
i CONTACT j I
I TANK I '
AERATION
mt.rtj*tfjn t*MtK*^t *~r7_ •*- |
f •>- __ _x
*».Wr ^^ ' ^^ ' ^^ p __ .^_ __ ^pm «_. ,
*..^ », , r^/*r v—^
SLUDGE
NOTE AU. TREATMENT UNITS FROM PRIMARY
CLARiFIER TO FINAL CLARIFIER ME
UNDER ROOF
FIGURE 4
FLOW DIAGRAM
TREATMiNT PLANT
CURWENSV1LLE T3M4NERY
619
-------
TABLE 3
RBC INFLUENT WASTE CHARACTERISTICS
DURING PERIOD OF PLANT START-UP
Flow 210,000 gallons/day
TBOD 1490 mg/1
SBOD 980 mg/1
TSS 900 mg/1
Sulfide 27 mg/1
TKN 360 mg/1
NH3-N 180 mg/1
pH 8
620
-------
As indicated, the flow was higher, as was the SBOD load. Further, the
suspended solids were greater, as is probably the reason for the higher pro-
portion of non-soluble BOD than found in the pilot plant wastes. The presence
of high sulfide concentration was also unexpected from our pilot plant work.
Because of the total oxygen demand limitations of the permit, the influent
presence of ammonia nitrogen higher than the effluent design expectations was
another serious difficulty revealed by this evaluation of initial operating
conditions.
Thereupon, a multi-faceted program was initiated to reduce waste load and
increase treatment capability in order that permit requirements could be met.
This was eventually accomplished with the full cooperation and support of the
Meadville Office of the Pennsylvania DER. This discussion is mostly limited
to the efforts made to enhance RBC performance.
The first relief attempt was to remove the baffle between the first and
second shaft to reduce the first stage specific BOD loading rate. Also, flow
reduction to the RBC unit was made. The major concern at this point was to
relieve the hydrogen sulfide generation problem. The results of these efforts
are tabulated in Table 4. Although the specific hydraulic and SBOD loads were
reduced to or below design levels, the effluent SBOD concentration remained
high as did the hydrogen sulfide levels in and around the RBC basins. A
dissolved oxygen level could not be maintained in the mixed liquor of any of
the stages.
The problem appeared to be more than oxygen limitations related to BOD
conversion. Both the high sulfide and suspended solids content of waste to
the RBC were significant contributors to the pervasivenss of the reduced con-
ditions within the RBC. Influent sulfide levels varied between 30 - 60 mg/1,
and lesser values persisted through the stages. Though each shaft assembly
was equipped with external devices to increase turbulence and minimize solids
deposition, substantial deposition did occur midway between the shafts pro-
viding a source for anaerobic conditions within the tank. High mixed liquor
solids concentrations in the 1500 - 2000 mg/1 range, with a 50 percent vola-
tile content were observed, as were high oxygen uptake rates in the 2-3 mg/1
per minute range.
The attached growth on the surfaces was gray to black and quite thick.
Under the bioraass, there was a thin crystalline inorganic layer on the media
surface, assumed to be a calcium carbonate accumulation.
USE OF HYDROGEN PEROXIDE
As an expedient measure to temporarily relieve the sulfide problem,
hydrogen peroxide was added to the RBC influent. At a reduced waste flow of
about one-half of the design basis, hydrogen peroxide was used initially at
the rate equivalent to approximately 35 mg/1 of influent flow. It was in-
creased later to higher levels sufficient to totally oxidize all sulfides. At
even the lower dose of hydrogen peroxide which did not totally remove all
sulfides, significant improved performance was observed. A summary of two
test runs when hydrogen peroxide was used is given in Table 5.
621
-------
TABLE 4
INITIAL PERFORMANCE OF RBC
Specific
Hydraulic
Loading
(gpd/sf)
0.45
0.45
0.45
0.25
Specific
SBOD
Loading
(ppd/lOOOsf)
3.0
2.1
1.9
1.6
Specific
SBOD
Removal
(ppd/lOOOsf)
1.67
1.00
0.90
1.18
Percent
SBOD
Removal
56
48
47
74
Effluent
SBOD
Cone.
(mg/1)
353
300
267
221
622
-------
TABLE 5
USE OF HYDROGEN PEROXIDE
TO IMPROVE RBC PERFORMANCE
Specific Hydrogen Specific Specific Percent Effluent
Hydraulic Peroxide SBOD SBOD SBOD SBOD
Loading Cone. Loading Removal Removal Cone.
(gpd/sf) (mg/1) (ppd/lOOOsf) (ppd/lOOOsf) (%) (mg/1)
0.283 35 2.85 2.35 82 213
0.17 125 1.17 1.11 95 47
623
-------
In the first test of higher flow and lower hydrogen peroxide dose, the
sulfide in the RBC influent was reduced from an average level of 46 mg/1 to 14
mg/1. For the most part, the SBC effluent showed only occasional signs of
sulfide. The dissolved oxygen level remained consistently low at 0.1 mg/1 in
all stages. The heavy biomass accumulation in the RBC surfaces was, however,
significantly reduced. At the higher hydrogen peroxide dose and lower flow,
total sulfide control was attained. With this control and with the reduced
SBOD loading of 1.17 pounds per day per 1000 square feet, good performance of
the 1BC was realized for the first time.
However, with the high cost ($100/day) of the hydrogen peroxide and the
limited hydraulic loading, other methods of obtaining improved treatment
performance were considered. These efforts were in two directions, one in-
volving treatment facility alterations and the other involving waste pro-
duction alternatives.
PRODUCTION CHANGES TO IMPROVE WASTE CHARACTERISTICS
Although detailed discussion of the production changes to improve waste
characteristics is beyond the intent of this presentation, it should be noted
that an interested and technically oriented plant manager worked diligently
and effectively to reduce both the constituent and volumetric level of the
wastes from the time of plant start-up to the present. This program is still
continuing. It is pertinent to note that as changes were made in production
practices to reduce flow and critical waste constituents, problems were often
encountered in product quality. This was particularly problemsome and costly
due to the lengthy time sequence involved in processing a raw hide. There-
fore, any modification in processing which proved to impair product quality
could be damaging to 15 - 20 days of hide processing, before the effect of
re-adjustment could be realized.
Somewhat offsetting this risk for the tannery was the potential savings
in chemicals and water such a conservation effort could provide. At this
time, substantial reduction in lime, sulfide, ammonium chloride and water has
been realized.
A summary of the reductions in raw waste quantity and quality which were
effected from June, 1978 to September, 1979, at approximately the same hide
processing level, is given in Table 6. The reduction is significant for all
parameters, being highest for TBOD and TSS. A comparison of the settled TBOD
and TKN for the two periods tends to somewhat decrease the significance of the
constituent reduction, as relates to RBC feed. Nevertheless, the reduction is
still significant and overall helpful to the intended improved performance of
the RBC.
FACILITY ALTERATIONS FOR IMPROVED PLANT PERFORMANCE
From the beginning of the full scale operation, it was realized that the
amount of suspended material entering the RBC and becoming deposited in the
basins was contributing to the oxygen limitation problem of the RBC. The
amount of suspended matter in the pre-settled and equalized wastes was also
being increased by partial precipitation of soluble constituents, as a result
of the acid adjustment of the pH to 7 - 9 just prior to the RBC, Laboratory
experiments also revealed further pH adjustment to a level of 4 - 5 signifi-
cantly reduced the BOD by coagulation of the protein. With the reduction of
624
-------
TABLE 6
REDUCTION OF RAW WASTE CONSTITUENTS
BY PRODUCTION CHANGES
Flow
COD
TBOD
TSS
TIN
Sulfide
June 1978
210,000 gal/day
15,000 Ibs/day
8,200 Ibs/day
(2,600 Ibs/day)
10,600 Ibs/day
1,160 Ibs/day
( 630 Ibs/day)
460 Ibs/day
170 Ibs/day
September 1979
130,000 gal/day
10,000 Ibs/day
3,300 Ibs/day
(1,900 Ibs/day)
3,600 Ibs/day
620 Ibs/day
(485 Ibs/day)
315 Ibs/day
110 Ibs/day
Note: Data in parenthesis is for settled wastes.
625
-------
this substrate portion, the resulting waste could be expected to be more
amenable to biological degradation.
It was also evident that improved mixing and oxygenation in the equali-
zation tank would be beneficial to the control of the sulfide generation
problem in the wastes prior to the RBC and, more particularly, prior to the
desired addition of the acidification and intermediate clarification pro-
cesses. Because of the high pH of the primary settled wastes, the aeration of
the equalization tank provided a potential further benefit of reducing the
ammonia content of the wastes by air stripping, and concurrently reducing the
alkalinity and subsequent acid requirements for protein coagulation.
Accordingly, an acid feed/mix system and an intermediate clarifier were
installed between the equalization tank and the RBC unit, to permit the acid
coagulation of the equalized wastes and the reduction of suspended matter in
the feed to the RBC. Concurrent with placing these alterations into operation,
the use of hydrogen peroxide was discontinued.
The use of supplemental air with the RBC also was considered a beneficial
alteration, in that such use would decrease oxygen limitation conditions and
create increased agitation to prevent the thick biomass accumulation and the
sub-surface anaerobic conditions in which sulfur reduction occurred.
This facility alteration program was undertaken in stages and, in fact,
was not totally implemented at the time this presentation was being prepared.
The acidification system and intermediate clarifier were installed first, with
operation commencing in November 1978. Supplemental air was added to the
first shaft in January 1979, to the second shaft in April 1979 and to the
third and fourth shafts in August 1979. The turbine aerators for the equali-
zation tank are just now being installed. Initially, in late 1978, two
surface aerators were taken from the existing aerated lagoons as an interim
effort to aerate the equalization tank. Although reasonably successful in
terms of maintaining an aerobic environment and stripping ammonia, this old
equipment's operation was unreliable and had to be totally discontinued half
way through the testing period.
All of the facility changes were innovatively detailed, fabricated and
installed by plant engineering and maintenance staff.
The desired intent of developing comparative data for the variable
conditions of production and treatment was made difficult by (1) the EPA
pressure to expedite compliance of effluent standards, (2) the trial and error
nature of the production changes being attempted for improved wastewater
conditions and (3) the problems involved in retrofitting and altering existing
facilities within a constrained space and time situation. Adding further to
the burden of this evaluation was the required experimentation being carried
on at this plant with a second stage biological system to achieve nitrifi-
cation capabilities within a very expedited compliance schedule to meet EPA
proposed BAT standards.
The test effort on the first stage RBC facility reached final frustration
on October 31, 1979 when an RBC shaft broke. As of January, 1979, the RBC
facility was still out of operation, with all four shafts being replaced by
the manufacturer.
626
-------
A flow diagram of the altered treatment system is given in Figure 5. A
sketch of the supplemental air system installed in the RBC is given in Figure
6.
PERFORMANCE OF ALTERED SYSTEM, WITH SUPPLEMENTAL AERATION OF RBC
Prior to the completion of the installation of the diffused air system in
the first two stages, but following the completion of the intermediate
clarifier and aeration of the equalization tank contents, there was a short
period of time when some limited observations were made of this change in RBC
performance. Although the dissolved oxygen level in the first two shaft mixed
liquor volumes remained absent, or at very low fractional level, the sulfide
generation problem appeared to be greatly improved. At reduced hydraulic
loadings of approximately .28 gpd per square foot of total shaft area,
effluent SBOD levels of 100 - 120 mg/1 were being achieved at SBOD removal
rates of near 1.5 pounds per day per 1000 square feet of total shaft area.
The major impact on RBC performance, however, occurred with the use of
supplemental air. A graphical representation of the testing performed with
supplemental aeration of the RBC, is shown in Figure 7.
As shown, the best performance was achieved when the wastes were
additionally pretreated by acid coagulation and clarification. For this
series of tests, supplemental aeration was limited to the two shafts of the
first stage, at a total air flow of approximately 600 cfm. This is equivalent
to about 12 cfm per lineal foot of shaft.
The data for partially aerated RBC receiving acid treated wastes do not
indicate for the SBOD range tested that the maximum SBOD removal rate had yet
been attained. Without acid pretreatraent, the maximum SBOD removal rates is
indicated to be approximately 2.5 pounds per day per 1000 square feet of total
shaft area.
Although there is some indication for the non-acid treated conditions
that supplemental aeration of all shafts at an air flow of 1100 cfm results in
better performance than when only the first stage is aerated, more data is
required before a final judgement can be made on this point.
For the acid treated conditions, some experimentation was made with
effluent recycling. Two data points shown in Figure 7 for where non-settled
effluent from the RBC was recycled at a rate of 50% of the influent rate
suggest somewhat reduced performance when compared to the non-recycle con-
ditions of operation. It should be noted, however, that the influent flow
rate (172,000 gpd) during the recycle testing was somewhat higher than the
influent flow rate (138,000 gpd) during the non-recycle testing periods. On a
specific hydraulic rate basis, the loading, including recycle flow, for the
recycle operation was 0.68 gpd/sf as compared to 0.36 gpd/sf for the non-
recycle operation.
A comparison of aerated and non-aerated operation of the RBC treating
these tannery wastes is given in Figure 8. Interestingly, the best performance
with supplemental aeration of the RBC with acid pretreatment was similar to
the pilot plant performance of the nonaerated RBC unit, under similar hy-
draulic conditions.
627
-------
.j-U,. EXHAUST
IS3SB VENTILATION
r~n
AOUACUARD
SOteEMNO
CMLORWE
PRIMARY
--«r ! I i
! EQUALIZATION !
i TANK ;
•«
"V
INTE«M
CUWFIER
MIXER
wicss
SLUDGE
'SLUDGE
LANDFILL
STEAM
COMPRESSED
Aft
LEGEND
—~ EXISTING
ALTERATIONS
FIGURE S
ALTERATIONS TO
TREATMENT SYSTEM
CURWENSVIU.E TANNERY
628
-------
I
4" STAINLESS STEEL MESDER-
SECTION THRU BIO-DISC
SECTION UJ-Qj
FIGURE 6
INSTALLATION DETAILS
RBC DIFFUSES ASSEMBLY
629
-------
JVAL BATE ( tBS/DAr/000 SF)
. * N in v
* p o> o u«
*•"
w
DC
I
u<
1.0
o>
o-
c
j
\
\
t
-
A
,/
/
*
ft /
. '
\
\
D
X
D
®
_L
Q ACID
600
B SAM
«n%
Q68
O NON
600
• NON
IOO
***^'
D O
—
O
EGEND
PRETREATED Q36 GPD/SF
CFM TO SHAFTS 1 AND 2
E AS ABOVE EJtCEPT FOR
. EFFLUENT RECYCLE AND
GPD/SF
-ACID TREATED, 0.36 GPD/
CFM TO SHAFTS 1 AND 2
-ACID TREATED, Q36 GPtV
CFM TO ALL SHAFTS
B
e
e " ' !
' i •
O
SF,
SF,
RBC PI
SUPPLE
,
FIGURE 7
IRFORMANC
MENTAL AE
!
i
1
E WITH
RATION '
) 25 SO 75 CO I2S ISO ITS 200 22S 2!
SBOD EFFLUENT CONCENTRATION (MG/Lt
630
-------
3.5-
3.0-
,2.9-
I
i E.O-
i.S-
1.0-
0.5"
WITH ACID TREATMENT,
SETTLING AND
SUPPLEMENTAL AERATION
OF RBC
IMUM PERFORMfif
»ILO
T RBC
i
j
•Pf
~i
i
i
i
'
.WITH NTERMEOIATE SETTLING AND
SUPPLEMENTAL AERATION OF RBC
WITH ACID TREATMENT
AND SETTLING, BUT NO
AERATION OF RBC
MiTiM. PERFORMANCE
ORIGINAL SYSTEM-
'X
,
1
FIGURE 8
,
1
rOMPARATIV
c
E PERFORMANCE :
>F RBC I
i i
!
! 1
SO
7S
100 IZ5 ISO
S BOO EFFLUENT CONCENTRATION (MO/L)
175
ZOO
zzs
631
-------
Examining performance of the RBC by stages, it appears that for the acid
treated wastes the aerated first stage operation is not oxygen or biomass
limited up to the maximum observed specific SBOD removal rate of approximately
6 pounds per day per 1000 square feet of first stage area. However, for the
wastes not acid coagulated and settled, the aerated first stage operation is
apparently biomass limited at a specific SBOD removal rate of about 4.25
pounds per day per 1000 square feet of first stage area. This is shown
graphically in Figure 9. Based on the relatively significant level of dis-
solved oxygen in the first stage mixed liquor for this period of testing, it
does not appear to be a problem of oxygen limitation. For the most part, the
dissolved oxygen level in the aerated first stage averaged over 1 mg/1, vary-
ing between 0.6 and 2.1 mg/1. Surprisingly, the third shaft (second stage)
and fourth shaft (third stage), during this testing of the RBC with only first
stage aeration, maintained lower dissolved oxygen levels at the 0,5 mg/1
level. It had been expected, based on the low specific SBOD removal rates in
these last two stages as related to the normal oxygenation capability of the
rotating discs, that the dissolved oxygen levels would increase. However,
when aeration was added to the last two shafts, the dissolved oxygen level
gradient changed to the more expected increasing type, with dissolved oxygen
concentrations reaching levels of 3 to 4 under the third shaft and 4 to 5
under the fourth shaft.
It is evident that the major removal of SBOD occurs in the first stage,
with only minor removal occurring in the last two stages. A summary of the
stage removal rates for the acid treated and non-acid treated waste testing of
the aerated RBC is given in Table 7. It is noted that both the first stage
and the combined stages for wastes of reduced protein content have a greater
specific removal rate of SBOD than do the wastes which did not have benefit of
acid coagulation and settling for reduction of protein. Interestingly, the
reverse is true for the last two stages, probably because some of the protein
content in the lesser pretreated waste is degraded in the latter stages.
In more closely evaluating shaft performance during the period of protein
reduction and first stage aeration, the data indicates that as total SBOD
loading and removal increases through the RBC unit, the percent of the total
removed by the first stage reduces somewhat as the removal percentages by the
last two stages increase. This is shown graphically in Figure 10.
For this same period of pretreatment to reduce the protein content, some
analyses were made to determine the change in the BOD rate constant as the
flow passed through the stages. The "k" value decreased from 0.26 for stage 1
influent to 0.15 and 0.08 for stages 2 and 3 influents, and 0.06 for stage 3
effluent.
Also shown in Figure 10 are two data points for the same pretreatment
conditions and supplemental aeration except that 50% recycle of effluent has
been provided. It is apparent that some of the SBOD removal has been trans-
ferred from the first stage to the second stage, by the effect of recycle.
Presumably, as the recycle rate would be increased further, the specific
removal rate would be more uniform throughout all the stages. This limited
testing indicates, however, that the overall removal rate would be reduced
somewhat by such effluent recycling possibly due to the increased hydraulic
rate.
632
-------
7.0-
C.O
& 9.0- •
I
4jO--
S
3.0
a:
D
«n 2.0 T
ACD PHETREATEO, 0.7Z
NON-ACID TREATED.
FIRST STAGE PERFORMANCE
WITH SUPPLEMENTAL AERATION
-OF 600 CFM-i 1
25 90 75 100 125 ISO 175 200 22S 250 275 SCO 325 35O 375 400 425 450
SBOD EFFLUENT CONCENTRATION (MG/L)
633
-------
40 •
^50--
e
u20
0
a
%
s
tf «-
:
) NT RECYO.
3 J.D%RECYt
NOTE- RBC
*F»
T*C
NCI
AND
i;0.36GPD/!
LE;0.6TOPO
PROVIDED SI
AT ION OF 6OC
SHAFTS. PR
UDED ACID a
SETTLING
**
JPPLEMENTAi
> CFM TO FIRS
'TREATMEr\*'y
iAQULATK*
T
Q
EACH OF 1ST
(STASi
'
— ©
TWO SHAFTS
. I)
.
FIGURE 10
COMPARISON Of S SOD
REMOVAL KATES BY STAGES.
2,O
2,2
2.4
2.6 Z.t 3.0 32 3,4
S BOD REMOVAL RATE, ALL SHAFTS (LBS/DAY/ 1000 SF)
634
-------
TABLE 7
AVERAGE SPECIFIC SBOD REMOVAL RATES
FOR RBC STAGES
Supplemental
Aeration With
Acid/Settling
Pretreatment
Supplemental
Aeration Without
Acid/Settling
Pretreatment
1st Stage
2nd Stage
3rd Stage
All Stages
5.1 ppd/lOOOsf
0.7 ppd/lOOOsf
0.4 ppd/lOOOsf
2.83 ppd/lOOOsf
3.7 ppd/lOOOsf
1.4 ppd/lOOOsf
0.7 ppd/lOOOsf
2.38 ppd/lOOOsf
635
-------
Similar to the pilot plant, the full scale RBC with its best performance
of SBOD removal still did not effect nitrification. The effluent ammonia
nitrogen concentration remained within a range of 180 to 220 mg/1. Although
COD removal was substantial through the entire treatment system, effluent
values from the RBC remained in the 350 - 450 mg/1 range.
Although the use of supplemental aeration with the RBC involves addition-
al electrical energy to provide the compressed air, the SBOD removal rate per
horsepower for the RBC in this application still appears to be reasonable.
Based on the high performance curve of Figure 7 and a total of 20 draw horse-
power for the blower and the four rotary drives, the SBOD removal rate per
horsepower to meet summertime permit requirements of 112 mg/1 is approximately
2.7 pounds of SBOD per hours per horsepower. Obviously, if diffused aeration
is used for both supplemental oxygenation and shaft rotation, the energy
efficiency should be enhanced.
SUMMARY AND CONCLUSION
The use of supplemental aeration with the RBC unit at the Curwensville
Tannery significantly improved the performance of this process. Supplemental
aeration, through limited diffusion to the basin contents, provides an
additional source of oxygen for biomass respiration and for prevention/control
of sulfate reduction. The diffused air flow below the rotating assemblies also
assists to maintain a thinner biofilm on the plastic surfaces, which reduces
the possibility of developing sub-surface anaerobic conditions which can
promote sulfide generation with these wastes of high sulfate content. In
combination with improved pretreatment and product processing modifications,
the diffused air supplemented RBC facility was able to achieve design based
pilot plant performance and to exceed present effluent permit requirements for
BOD.
The serious problems caused by the generation of sulfide within the tank
contents of the RBC and within the attached biomass should be viewed as a
limitation to the use of non-aerated RBC for wastes of high sulfur or sulfate
content.
Upon repair to the fullscale RBC facility and installation of the turbine
aerators in the equalization tank, additional testing will be undertaken.
This will be to determine the effect of equalization tank aeration in reducing
the ammonia content of the waste by air stripping, and agglomerating and
precipitating BOD constituents for removal in the intermediate clarifier prior
to the RBC. This approach for reducing the BOD content to the RBC is prefer-
red to the acid coagulation procedure, because of the difficulties involved in
concentrating and dewatering the acid sludges.
It is also intended that a more quantitative evaluation be made of the
effect of calcium accumulation on the rotary surfaces both in terms of
structural stress and biological effect.
The problems in applying pilot plant experience to full scale design as
reported herein is further evidence of the difficulty which can be encountered
in scaling-up RBC pilot plant results. Although it is evident that the
differences in waste characteristics were in part responsible for the perfor-
mance problems of the fullscale system, significant contribution to the
problem was also due to the much higher oxygen supported SBOD removal rate of
636
-------
OPERATIONAL EXPERIENCE OF
OXYGEN-ENRICHED ROTATING BIOLOGICAL CONTACTORS
By
Ju-Chang Huang
Professor of Civil Engineering and Director of
Environmental Research Center
University of Missouri-Rolla
•Rolla, Missouri, U.S.A.
INTRODUCTION
The use of rotating biological contactors (RBC) for treating both
municipal and industrial wasters has gained considerable popularity in
recent years. This is partly because of its low energy need and ease
of operation and maintenance. It is also partly because the attached
microbial growth on the RBC unit can allow different groups of micro-
organisms to exist at different discs within a single treatment unit.
This provides a valuable feature for achieving biological nitrification.
One of the major factors limiting the performance of an RBC unit is
the availability of oxygen in the treatment system. In the literature
it has been well documented that the rate of organic stabilization in such
a system is generally limited by the oxygen penetration rather than by
the substrate diffusion into the biological film.l*2 it is only in a
multi-stage system that the substrate diffusion may become a rate-limiting
factor in the last stages of the system.3>4 jn fact, numerous field
installations have experienced variable degrees of septic problems in
their treatment units. Many researchers3>5,6,7 have also found that
the organic stabilization rate increases with the disc rotating speed.
This is because a higher rotating speed would cause a greater oxygen
transfer efficiency. Unfortunately, the rotating speed cannot be increased
indefinitely without causing some major drawbacks. First, power requirement
increases exponentially with the disc rotating speed.3,8 Secondly, an
excessively high rotating speed creates a high hydraulic shearing force
637
-------
which may interfere with satisfactory development of biomass on the disc
surface. Antonie8,9 has suggested that the optimal peripheral speed for
treating domestic wastewater is about 18.2 m/min (60 ft/min). Therefore,
in order to increase the oxygen penetration in the RBC system, some appro-
priate means other than unlimitedly increasing the rotating speed is neces-
sary. There are at least two possible methods to accomplish this. One
is to use an .enclosed RBC system and replace air with pure oxygen as the
feed gas. The other is to pressurize the enclosed RBC system using either
air or pure oxygen. In either case, the partial pressure of oxygen in the
gaseous phase is increased and the oxygen penetration into the biofilm
can thus be increased. The objective of this study was to investigate the
operational characteristics of several bench-scale, enclosed RBC units
receiving oxygen as the feed gas.
EXPERIMENTAL STUDY
For the convenience of this study, a synthetic milk waste was used as
the influent feed. Three bench-scale, enclosed RBC units were built and
each of them was operated under a specifically designed condition. The
operational characteristics of these units were evaluated in order to assess
the advantages of using the oxygen-enriched RBC system.
Influent F e ed. The synthetic milk waste was prepared at two different
strengths, the higher one being twice as strong as the lower one. The
single-strength waste was prepared by adding 0.900 g of a commercial dry
milk powder (Kroger Brand, a product of Kroger Company, Cincinnati, OH)
and 0.084 g of K2HP04 to each 1 of tap water. The dibasic potassium
phosphate was added to both supplement the phosphorus content and buffer the
pH so that the feed solution would not drop to the isoelectric range of
casein to cause protein precipitation. The chemical oxygen demand (COD)
of this waste was about 1,000 rag/1 . Its pH was 7.2. During the experi-
mental study, a suitable quantity of this waste was prepared each day,
which was then placed in three 55-gal drums stored inside a 4°C walk-in
cooler. The waste was continuously stirred to keep undissolved milk solids
in suspension.
Rotating Biological Contactors. The three bench-scale RBC units were
built with plexiglas tubes, each 30.5 cm (12 in.) OD and 0.61 m (2 ft)
long with a wall thickness of 3.18 mm (0.125 in.). Each tube was made to
hold a series of discs submerged at one-half tube depth. The end plates
as well as the rotating shaft-bearing assembly were so constructed that the
entire unit could withstand a pressure of 68.95 kN/m^ (10 psig) without any
water or gas leaks. The detail construction drawing is shown in Figure 1.
The end plates, each 12.7 mm (0.5 in.) thick, were milled to have a
groove 1.5 mm (0.06 in.) deep to fit the tubular wall with a clearance of
0.64 mm (0.025). The groove was filled with a rubber-based silicone seal
to serve as an 0-ring gasket. The two end plates were pressed against the
tubular body and secured together with eight connector rods, each 4,76 mm
(0.188 in.) and spaced at 45° intervals. With this construction, the unit
was able to achieve a gas-tight condition, even with an applied pressure
of up to 68.95 kN/m2 (10 psig).
638
-------
I'J"
Conrnstor rod* at 45*
" OD * l/e" thick pleiiglat tub*
t 3/16" Stabilizer rod* at 120
Top 1/4 pipe thread for
0"30 psi prfltsure gouge
Top 1/4" ppe thread for Qg or
off influent Hnc
Tap 1/2" pip* thnad
for Q-5Q psl prtiiur*
relief *atve (i tnd only}
•Drill 3/8 Hfli* for l^*
effluent pleiiglos pipe
Countersink ihaft hale to seat
1/2" ID rubber 0-rmg
ENDPLATE DETAIL
7/16" oir
-------
Inside the tublar body was a total of 34 circular discs, each 26.7 cm
(10.5 in.) in diam x 3.2 mm (0.125 in.) in thickness. The discs were equally
spaced with wood spacers to result in a face-to-face clearance of 12.7 mm
(0.5 in.). Three threaded stabilizer rods, each 4,76 mm (0.188in.) in
diam and spaced at 120° were passed through holes drilled 12.7 mm (O.Sin.)
from the outer edge of each disc, and then bolted at both sides of the
disc to hold it in alignment.
The tubular body was divided at the midpoint into two equal sections
with a divider so that each RBC unit would function as a two-stage system.
The bottom 25.4 mm (1 in.) portion of the divider was cut off to facilitate
passage of the liquid and sludge. Numerous 11.1 mm (0.44 in.) holes were
drilled in the upper half of the divider to allow the passage of gas.
The divider was cut to fit exactly into the tubular body and was then secured
in place using hose clamp pressure around the circumference of the outer
wall of the tubular body.
In order to make the entire RBC system air and water tight, a special
shaft-bearing assembly had to be devised. The shaft holes of the end plates
were countersunk to seat a 12.7 mm (0.5 in.) ID 0-ring. A spring was then
positioned to push the 0-ring into the countersunk space and prevent leaks
around the shaft. However, due to friction the 0-rings were found to grad-
ually wear out and had to be replaced at 2-3 wk intervals.
The end plates were properly tapped and threaded for installations
of liquid influent and effluent lines, gas inlet and outlet, the latter
being a pressure relief valve, and a pressure gauge or mercury manometer
for accurately monitoring the internal operating pressure. Both influent
and effluent lines were connected to positive displacement pumps so that the
system could be maintained at any desirable pressure. Also, just before the
stage divider, a spigot was tapped to facilitate the withdrawal of the first-
stage effluent sample. The detail schematic of the three experimental RBC
systems is shown in Figure 2.
During operation, the water level was maintained at half depth of the
tubular body, and the rotating speed was set at 12 rpm, which provided a
peripheral velocity of 10 m/min (33 fpm). The flow rate was rigidly con-
trolled at 56 ml/min (2,105 gal/day). The following represents the pertinent
information about the system operation:
Liquid flow rate 56 ml/min (21.5 gpd)
No. of stages
in each RBC system 2
No. of discs
in each stage 17
Disc diam 266.7 mm (10.5 in.)
Rotating speed 10 m/min (33 fpm)
Face-to-face spacing
between discs 12.7 mm (0.5 in.)
640
-------
cr>
-£
Stoge Divider
ROTATING BOLOGICAL
CONTApTOR
^-•l
•lit Stogt Effluwt Tap
ROTATING BIOLOGICAL
CONTA'CTOR
(preuujlztd)
\
NBt
orlng
ROTATING BIOLOGICAL
H-
CONTAlCTOR
• EfflMKl
•Efflutnt
TQ ,.
•EffluMit
Figure 2. Detail Operational Schematic of the Three RBC Systems
-------
Liquid volume
in each stage
Gas volume
in each stage
Total disc surface
in each stage
Ratio of vol/surface
in each stage
Ratio of flow rate/surface
in each stage
Detention time
in each stage
10.58 1 (2.82 gal)
10.58 I (2.82 gal)
1.89 m2 (20.6 ft2)
5.60 1/m2 (0.14 gal/ft2)
42.7 l/m2-day (1.04 gpd/ft2)
3.15 hr
Experimental Approach. Three separate experimental runs were conducted
in this study. The operating conditions designed for each RBC system in
these three experimental runs are shown in Table I.
The first experimental run was conducted using air as the feed gas.
The gas flow rate was controlled at 420 ml/rain for all three RBC units.
This was equivalent to 1.0 ft^ air per gal of influent waste flow, similar
to that commonly used by an activated sludge plant. The first RBC system was
not operated under pressurization, and it was designated as Unit I. The
second RBC system was operated under a gauge pressure of 27.6 kN/m2 (4 psig).
Since this unit received the same influent waste (i.e., single- strength or
1000 tng/1 COD) as the first unit except that it was pressurized, this
second unit was designated as Unit I-P. The notation "P" indicates
operation under pressure. The third RBC system received a double-strength
waste (i.e., 2000 mg/1 COD) and it was also operated under a pressure of
27.6 kN/m2 (4 psig). Thus, this unit was designated as Unit II-P; that
is, the notation "II-P" indicates that the unit received the double-
strength waste and it was operated under pressure.
The first experimental run was started in late April of 1978 when the
three RBC systems were first placed in operation. During the first month
the synthetic milk waste was mixed with gradually decreasing concentrations
of sewage to provide microbial seedings and speed up the establishment of
biomass on the disc surface. Thereafter, only the synthetic milk waste
was used. By early August, all three RBC units had reached consistent
operation as evidenced by a fairly uniform reduction of COD in the day-to-
day operation. An intensive monitoring program starting August 7 and ending
August 18, was undertaken to evaluate the operational characteristics of
each RBC system. This included the measurement of such performance parameters
as soluble chemical oxygen demand (COD), suspended solids (SS), ammonia
nitrogen, nitrate nitrogen, and sludge settling rate. Also, the mixed liquor
suspended solids (MLSS) and the dissolved oxygen (DO) concentrations in each
stage of the three RBC system were also determined. All of these tests were
performed according to the procedures set forth in Standard Methods. 10
642
-------
After the detail operational characteristics of the first experimental
run were obtained, the three RBC units were switched to the new operating
condition designed for the second experimental run. That is, the feed
gas was switched to pure oxygen and its flow rate was reduced to 240 ml/min,
which was equal to 0.57 ft3 oxygen per gal of influent waste. A "transitional"
period of slightly more than a week was observed before the three RBC units
reestablish consistent operation. Thus, at the end of the third week
(i.e., beginning September 11, 1978), a 10-day intensive monitoring program
was again undertaken to assess the performance characteristics of the second
experimental run.
Similarly, at the conclusion of the second experimental run, the three
RBC units were switched to the operation condition designed for the next
phase, or the third experimental run. Actually, the only change in this
phase was to reduce the oxygen flow rate from 240 ml/min to 42 ml/min. This
low flow rate was attempted because in the second experimental run, the
DO levels in all three RBC units were found to be excessively high. It
took only a few days for the three RBC units to reestablish consistent per-
formance. Thus, the intensive monitoring work was again started October 1,
1978. It lasted for 10 days before this phase of work was concluded.
RESULTS AND DISCUSSION
The main objective of this research was to evaluate the operational
characteristics of parallel bench-scale RBC units under an oxygen-enriched
environment with various organic loadings. The following will discuss the
findings of various operational parameters examined in this study.
P.O. In this study, the main reason for using either pure oxygen
or pressurization in an RBC unit was to increase the partial pressure of
oxygen in the treatment system so that the oxygen availability would not
become a limiting factor in the biooxidation process. In the three separate
experimental runs, the first was conducted using compressed air, while the
latter two were conducted using pure oxygen. Since the rate of organic
loading on each specific RBC unit was the same for all three experimental
runs, it was reasonable to expect that the use of pure oxygen would yield
a higher D.O. level provided that the gas flow was adequate. This expecta-
tion was generally verified from the results obtained in this study (Fig-
ure 3).
In the first experimental run (conducted in August, 1978), the D.O.
levels in the first stage of all three RBC systems were generally low.
The two units under pressurization (Units I-P and II-P) were actually found
to have D.O. from 0.1 to 0.5 mg/i , which were less than the value observed
for the unpressurized unit (Unit I), approximately 2-3 mg/1 . Since Unit
II-P received twice as much organic loading as Unit I and the increment of
oxygen partial pressure through pressurization amounted to only 27 percent
(i.e., 4/14.7=27%), it was understandable that the D.O. in this unit would
be lower than that in Unit I. However, it was difficult to explain why
Unit I-P had a lower D.O. level than Unit I. The D. 0. in the second stage
did follow a reasonable trend, i.e., Unit I-P had the highest level, followed
by Unit I, and then Unit II-P.
643
-------
-AIR *4 \* OXYGEN >-J
UNIT I
INIT I-P.II-P
*
f°' "' °^
1^,.. we.,.,)
T II 14
AUO 1978
» 13 17 21 I a » 13
SEP 1978 OCT 1978
(a) First Stage - D.O.
("> "' "•)
\«|[ Wttl*/ V
7 n i»
AUO 1978
UNIT I- /
'
.
/ V'
(b) Second Stage - D.O.
Figure 3. Variations of D.O. in Each RBC System
in the Three Separate Experimental Runs
644
-------
In the second experimental run (September 1978), in which pure oxygen
was injected at a rate of 240 m 1/min or equivalent to 0.57 ft3 oxygen/gal
waste, all three RBC systems were found to have high D.O. (at least 20
mg/l ) in both the first and second stages. The day-to-day fluctuations of
D.O. in any specific units were partly due to variations of MLSS present in
the systems (as will be shown later) and partly due to possible experimental
errors encountered in the D. 0. measurement using the probe technique. Be-
cause of the excessively high D. 0. present in the withdrawn samples, they
had to be diluted first with deoxygenated water to bring the D.O. down to
a measurable range (maximum 15 to 20 mg/l ) by the probe technique. Through
this dilution, a portion of the over-saturated D. 0. could be easily released.
In the third experimental run (October 1978), in which the pure oxygen
flow rate was reduced to only 42 m 1/min or equivalent to 0.1 ft-^/gal waste,
the D.O. variations in all three RBC units generally followed an expectable
pattern. That is,Unit I-P had the highest D.O., followed by Unit I and
then Unit II. In fact, the D.O. in Unit II, particularly in the first
stage, was either zero or close to zero in many observations, indicating
that the supply of oxygen to this unit was not adequate to meet the biological
demand.
Biomass Growth The periodic sloughings of biomass from the disc
are normally due to the anaerobic activity developed at the deep layer of
the biofilm8. > 11 Thus, in any RBC system if the oxygen transfer is in-
creased (through the use of either pure oxygen or pressurization), the de-
velopment of biofilm would also be expected to increase. This expectation
was substantiated in this study. It was consistently demonstrated in the
three separate experimental runs that whenever the unit was supplied with
pure oxygen, particularly in conjunction with pressurization, an extra-
ordinarily heavy growth of biofilms was found. In fact, the biofilms
in the first stage were so thick that the clearance between many discs were
covered with biomass (Figure 4).
However, the observed heavy accumulations of biomass on the disc
surface must not be misinterpreted as large productions of sludge mass
from the RBC system. It has been reported^ that the use of oxygen-
enriched atmosphere tends to reduce sludge production in an RBC system.
In this study no attempts were made to measure the biofilm thickness
because of the difficulties involved in gaining an access to an enclosed
RBC unit. However, routine determinations of the^MLSS were made for all
liquid samples withdrawn from both stages of the three RBC units, and -the
results are shown in Figure 5. It was found that the MLSS observed during
the short monitoring period of each separate experimental run was quite
variable and therefore, did not truly reflect the exact quantity of biomass
present in the RBC system. This variation of MLSS was caused mainly by
inconsistent sloughings of the biofilms during the three experimental runs.
It appeared that the sloughings occurred in a somewhat cyclic manner. As
such, there were times when the measured MLSS was low, but the thickness of
biofilm on the disc surface was very high. Therefore, if it were intended
to determine the relative sludge productions amoung different RBC units, it
was necessary to conduct each experimental phase over a sufficiently long
period of time so that the collected MLSS data can be analyzed statistically
in order to obtain a realistic sludge production rate. Neverthless, the data
645
-------
Figure 4. Development of Thick Biofilm in the
Oxygen Pressurized RBC Unit
-------
[*«— OXYGEN -
(2i_llL°z)
^--UNIT
T II 13 19 9 13 17 21
AUG 1976 SEP 1978
3 9 13
OCT 1978
(a) First Stage - MLSS
OXYGEN -
VflO' WOiUV
OXYGEN
/O.I rf3 o.-i
'-UNIT I-P
/vwr H-P
. HIN
II 13 19 9 n 17 21
AUG 1978 SEP 1978
3 9
OCT 1978
(b) Second Stage - MLSS
Figure 5. Variations of MLSS in Each RBC System
in the Three Separate Experimental Runs
647
-------
shown in Figure 5 did indicate that the quantity of MLSS in the first
stage of each RBC unit was considerably higher than that in the corresponding
second stage. This was, of course, because there were not much organics
remaining in the second stage, and the growth of biofilms was limited by
the availability of organic substrate rather than by the oxygen penetration.
pH Throughout the three separate experimental runs, the pH values
in each" of the RBC systems were found to stay within 7 + 1 units. In
general, the pH level was slightly higher in the second stage than in the
first stage. This could be due to deamination of milk protein in the bio-
oxidation process. There was only one instance at which the pH did fall
below 6. This occurred in Unit II of the third experimental run when the
oxygen supply to this unit was not adequate to maintain a complete aerobic
condition (Figure 3). As such, a buildup of organic acids could have
occurred.
Soluble COD The reduction of COD in each RBC unit depended not only on
the quantity of biomass present in the system, but also on the diffusability
of organic substrate into the biomass. It was mentioned earlier that when
an RBC unit was supplied with an adequate flow of pure oxygen, particularly
under pressurization, it was able to develop an extraordinarily thick layer
of biomass in its first stage. However, this heavy growth of biofilms had
not always been accompanied by a higher level of COD removal compared to
the other units, as shown in Figure 6 (a). In fact, the unpressurized
Unit I showed a consistently better removal of COD than its pressurized
counterpart (Unit I-P). It was believed that the over development of the
biofilm and its subsequent bridging over the disc clearance actually re-
duced the available surface area for organic substrate to reach the bulk of
the biomass. This kind of problem was not observed in the second stage of
Units I and I-P, Figure 6 (b). Therefore, their relative COD reductions
in the second stage were quite comparable and both were likely limited by
the substrate diffusion. Because of this, in the future practical design of
an RBC system using pure oxygen, it is important to provide adequate spacings
(depending on the waste strength, but at least 1 in. for a waste having a
COD of 1000 mg/L) between discs to aviod biomass-bridging over the discs.
In the RBC unit receiving the double-strength waste (i.e., Unit II),
the COD reductions in both stages were considerably higher in the second
experimental run (the unit received an adequate flow of pure oxygen) than
those in the first and third experimental runs. This would apparently
suggest that, at a high organic loading rate, the limiting factor for the COD
removal was the oxygen flux. Since the second experimental run supplied more
oxygen than the other two, it was able to achieve a better COD removal.
Therefore, it is reasonable to conclude that use of pure oxygen in sufficient
quantities was able to increase the COD removal in a heavily loaded RBC sys-
tem. If the oxygen supply was not sufficient, such as that occurred in the
third experimental run, the system could become anaerobic (Figure 3) and the
COD removal would be drastically reduced.
Sludge Settleability and Effluent Suspend Solids. The overall treatment
efficiency of any biological system depends greatly on the settleability of
biological solids in the secondary clarifier. Since sloughings of biofilms
in an RBC system generally occur in a cyclic manner, its secondary clarifier
648
-------
—*,.—.
(10 IIJ Al.-v
\,a *,.,.]
UNIT n-p^ v
i /I
"\ / •
,Ac^/^-UNiT I-P
V-JJNJT 1 _„
. "
i~-
OXYGEN •
/•O. 17 fl1 0,\ \
^ Vgal Wait* 1
UNIT I-P— V
*, \
1
/OXYGEN »
/—UNIT II
/o_^ "
^ \ gal Wotti-'
UNIT 1-Pv.
UNIT 1-— \
3 9 1
AUG 1978 SEP 1978 OCT 1978
(a) First Stage - Soluble COD
(got Wo.i.')
,*-UNIT 11-P x
\ /"^
\ /UNIT 1-pA
UNIT II —^••X'
^•UNIT I
•' t--j*
' (.gil w«,l
_/
UNIT I-P
UNIT 1
^U
AUG 1978 SEP 1978 OCT 1978
(b) Second Stage - Soluble COD
Figure 6. Comparative Reductions of Soluble COD
in Different RBC Systems
649
-------
would be subject to periodic shock loadings of biological solids. Thus,
the sludge settleability would become critical in dictating the overall
treatment efficiency. It has been reported by Bintanja, _et jal_.12 that
the use of an oxygen-enriched atmosphere is able to improve the sludge
settleability. This finding had also been substantiated in this study.
Figure 7 shows a striking improvement of the sludge settleability
observed in Unit I when pure oxygen was used to replace air. Figure 8
further indicates that, even with pressurization alone, an RBC unit supplied
with air could also achieve a significant improvement of the sludge settle-
ability as compared to an unpressurized unit, i.e., Figure 7 (a) vs. Figure
8 (a). When pressurization was used in conjunction with pure oxygen, the
sludge settleability was extremely good, and the settled sludge also
compacted well, as shown in Figure 8 (b). In fact, the sludge produced
from such a unit appeared to consist of very dense granules, as shown in
Figure 9 . This figure compares the relative settling and compaction of
the three MLSS samples taken from the three "oxygenated" units designated,
respectively, as Units I (0 psi), I-P (4-psi) and II (0 psi).
In order to shed some light on the microbial nature of the sludges
obtained from different RBC units, photomicrographs were taken throughout
the course of the study. Typical results are shown in Figure 10 which in-
dicates that Unit I (with unpressurized air) contained a large quantity of
filamentous microorganisms, Photo (a). But when the air was replaced by
oxygen, the extent of filamentous growth was significantly reduced, Photo
(b). Pressurization of an RBC system also showed a reduction of the fila-
mentous growth, Photo (c) and (d). The sludge in the oxygen-pressurized
RBC unit appeared to be "chunky" and dense, as shown in Photo (d).
Because of the good sludge settleability, Units I and I-P almost
consistently had lower effluent SS (measured after 1-hr settling) in the
second and third experimental runs (i.e., using pure oxygen) than in the
first experimental run (Figure 11). The effluent SS of Unit II were high
in the third experimental run. This was because this unit had experienced
anaerobic conditions from time to time, which had a deteriorating effect
on the sludge settling quality.
Nitrification. Rotating biological contactors have been known to
be effective in achieving nitrification in the multi-stage operation. This
is possible because the slow-growing nitrifiers can grow separately from the
saprophytes in the latter stages of the RBC system. Torpey, jetjl.4 reported
that use of an oxygen-enriched atmosphere was able to accelerate the BOD
removal, thereby allowing nitrification to occur effectively in an earlier
stage. In addition, the rate of nitrification was also increased by the
oxygen enrichment.
In this study, it was found that when an RBC unit was supplied with
air, pressurization was able to enhance nitrification. Figures 12 and 13
show that in the first experimental run (August 1978), both stages of Unit
I-P had lower NHj-N and higher NOs-N than the unpressurized Unit I. Even
in the double-loaded Unit II-P, considerable extents of nitrification were
650
-------
(a) Unit I ~ Air at 0 psi
UNIT I! Pura Ox«jft-i (Op»i)
_„_. 010 fe1 o> /
-350 rng/1 MLSS
--588
-------
(a) Unit I-P - Air at 4 psi
UNIT I-P: Pur* Oxygen (4 pli)
Flow Polo!
0.37 II3 0, /jol *01U
o 10 (I3 02/gol won*
576 mg/l MLSS
^-238 mg/l MLSS
,179 mg/l MLSS
/ x-50, 65. a 142 mg/l MLSS
10 20 30 40 30
TIME (mini
(b) Unit I-P - Pure Oxygen at 4 psi
Figure 8. Improvement of Sludge Settleability by
Using Pressurization
652
-------
(a)
Beginning of Settling
(b)
One Hr After Settling
(c)
Secondary Clarifier
Figure 9. Relative Settling Qualities of Sludges
from Different RBC Units with Pure Oxygen
653
-------
in
(a) Air @ 0 psi
(b) Oxygen @ 0 psi
(c) Air @ 4 psi
(d) Oxygen @ 4 psi
Figure 10. Photomicrographs (100X) of Sludge Cultures Obtained from
Different RBC Units
-------
—OXYGEN
i OJ_j£jgi\
\ «ei wotf* /
v
T !i its I* i 13 IT 21
AUG i<9?8 SEP 1978
5 9 IS
OCT I9T8
(a) First Stage - Effluent SS
—-OXYGEN—
-I
T H 15 I* 9 13 IT 21
AU3 1978 SEF 1978
5 * IS
OCT lfT»
(b) Second Stage - Effluent SS
Figure 11. Variations of Effluent SS in the
Three RBC Systems
655
-------
CO 11^
gal w«i
II1 »J.\
-OXYGEN™
rn.n "3 e
x^— UNIT i
II 17 21
SIP I9T8
(a) First Stage -
fl.Q ft3 HiA
\4al wstli/
-OXYSEN-
li3 I
ff- tlNIT II
-HJNIT 1
/ai ,,3 o,N
(.Ti-SST.*) .
1 i-c.
H IS 19 9 l> IT
AUO 1978 SEP 1978
(b) Second Stage - NH.-N
Figure 12. Variations of NH--N in the Three
RBC Systems °
656
-------
—ASS-—
•10 II3«.V\
—OXYGCN-
/OS|n3 0,N
UNIT
UWT Ik
r n 15 19 9 a IT zi
AUG 1978 SEP 1976
5 9 IS
OCT 1978
(a) First Stage - NO,-N
-OXYGEN H
A
A
r_,
^-UNIT I-P"
«.». . Y
if 13 19 9 19 IT 21
&UG I5T8 SEP 19?8
3 9 IJ
OCT I3?g
(b) Second Stage - NO--N
Figure 13. Variations of NO,-N in the Three
RBC Systems
657
-------
also observed in both stages of the system. When air was replaced with
pure oxygen in the second experimental run (September 1978), both stages
of Units I and I-P appeared to have lower NH3-N than those in the first
experimental run. However, this was not always accompanied by greater
amounts of N03-N in the treated effluent. There were two possible explan-
ations for this. First, the heavy production of biofilms in this experi-
mental run could have assimilated a greater amount of NH3-N into the bio-
mass. Secondly, the excessively high concentrations of oxygen (well above
20 mg/1 ) present in these RBC units could have imposed some inhibitory effects
on the nitrifiers. The latter explanation may also apply to the data observed
in Unit II, which showed a lower NQ3~N content in the second experimental run
than in the first one.
The data of the third experimental run (October 1978) generally
repeated the pattern of the second experimental run, except that more-
extensive nitrification occurred in Unit II due to its "less than excessive"
oxygen content present in the system. All of these seems to support the
suggestion that an excessively high concentration of oxygen (20 mg/1 or
above) in an RBC system can impose an inhibitory effect on nitrifiers. If
this suggestion is indeed true, then in the future multi-stage design, use
of pure oxygen should be limited to only the front stages of the system.
The latter stages should use only regular air to allow effective nitrification
to occur. This conclusion must be considered preliminary, and further re-
search work has to be done to substantiate its fact.
CONCLUSION
Based on the findings of this study, the following conclusions can
be drawn:
1. Use of pure oxygen, especially in conjunction with a pressur-
ization of 27.6 kN/m2 (4 psig), was able to allow a thick layer of bio-
films to develop in an RBC system treating a synthetic milk waste having
a COD of 1000 mg/1 . The biofilms were thick enough to bridge the 1.27
cm (0.5 in.) clearance between discs in the first stage of the RBC system.
2. Mien the bridging of disc clearance was too extensive, the
available surface area for organic substrates to diffuse into the bulk of
the biomass became significantly reduced. This would result in a reduction
o£ the COD removal. Therefore, in the design of an RBC system using pure
oxygen, an adequate spacing ( at least 1 in,,) between discs must be provided
to avoid the biomass-bridging.
3. If there were no extensive biomass bridging and the substrate
concentration was high, use of pure oxygen in sufficient quantities was
able to improve the COD removal as compared to the use of air.
4. Use of pure oxygen in an RBC system showed a definite improvement
of the sludge settleability. The sludge appeared to be very dense and was
generally absent of filamentous growths. The use of pressurized air had
a similar, but less noticeable effect.
658
-------
5. Pressurization of an air-supplied RBC unit was able to enhance
nitrification. However, when air was replaced with pure oxygen, the ex-
cessively high concentration of oxygen present in an RBC treatment system
appeared to have an inhibitory effect on the nitrifiers,
REFERENCES
1. Williamson, K», and McCarty, P. L., "A Model of Substrate
Utilization by Bacterial Films." Jour,, Water Poll. Control Fed.,
£8, 9 (1976).
2o Williamson, L., and McCarty, P. L., "Verification Studies of
the Biofilm Model for Bacterial Substrate Utilization." Jour.
Water Poll. Control Fed., 48, 281 (1976)„
3. Famularo, Jo, Mueller, J0 A., and Mulligan, T., "Application of
Mass Transfer to Rotating Biological Contactors0" Jour. Water
Poll. Control Fed., 5CK, 652 (1978) „
4. Torpey, W0, Heukelekian, Ho, Kaplovsky, A. J0, and Epstein, L0,
"Effects of Exposing Slimes of Rotating Discs to Atmospheres
Enriched with Oxygeno" Proc. 6th Intlo Water Poll. Res. Conf.,
405 (1972)
5. Friedman, A. A,, Robbins, L. E., and Woods, R. C., "Effect of Disc
Rotational Speed on RBC Efficiency." ProCo 55rd Purdue Ind. Waste
Conf. (in press)
6. Chittenden, J. A0, and Wells, W. J., "Rotating Biological
Contactors Following Anaerobic Lagoons," Jour. Water Poll.
Control Fed., 45_, 746 (1971).
7o Welch, F. M., "Preliminary Results of the New Approach in the
Aerobic Biological Treatment of Highly Concentrated Wastes."
ProCo 25rd Purdue Indo Waste Conf., 428 (1968)»
8. Antonie, R0 L., Fixed Biological Surfaces-Wastewater Treatment,
Chemical Rubber Company (CRC) Press, Inc0, Cleveland, Ohio (1976).
9. Antonie, R0 L., Kluge, D. L., and Mielke, J. H., "Evaluation of
a Rotating Disc Wastewater Treatment Plant." Jour. Water Poll.
Control Fedo, 46>, 498 (1974).
10. "Standard Methods for the Examination of Water and Wastewaters."
14th Ed., Amer Pub. Health Assn., Washington, D. C. (1975).
11. Hawkes, H, A., "Film Accumulation and Grazing Activity in the
Sewage Filters at Birmingham." Jour. ProCo Inst. Sew Purif., 88-110
(1957).
12o Bintaja, H. H., Brunsmann, J. J., and Boelhouwer, C0, "The
Use of Oxygen in a Rotating Disc Process." Water Research, 10,
561 (1976).
659
-------
Page Intentionally Blank
-------
PART VI: INDUSTRIAL WASTEWATER TREATMENT
WASTEWATER TREATABILITY STUDIES FOR A
GRASSROOTS CHEMICAL COMPLEX USING
BENCH SCALE ROTATING BIOLOGICAL CONTACTORS
By
Joe C. Watt
Development Coordinator
Environmental Systems Division
Charles J. Cahill
Environmental Scientist
Environmental Systems Division
Catalytic, Inc.
Philadelphia, Pennsylvania, U.S.A.
Introduction
Pilot scale test units are normally recommended when evaluating the
applicability of Rotating Biological Contactor (RBC) systems for the treatment
of industrial wastes. However, insufficient samples, lack of funds and/or
tight schedules often create circumstances which make it impractical to
conduct treatability studies in the larger pilot test units. To accommodate
these situations, the Environmental Systems Division of Catalytic, Inc.
has fabricated bench-scale, modular design RBC systems and utilized them
for industrial waste treatability studies where the benefits of the RBC
concept appeared to have potential application.
This paper describes an application of this equipment on a project
where pilot-scale operation was not possible; and where the treatment system
design was part of a total design of a grassroots production facility.
Catalytic International in London, England was one of several contractors
working on different process areas and off-sites of a large chemical complex
for the client, Berol Kemi AB of Sweden.
661
-------
During the engineering design phase, Catalytic's Environmental Systems
Division provided consultation and laboratory services to Catalytic in
London for treatment of the wastewaters that would be generated from the
whole chemical complex. The plant was to be built in a virgin area on
Swedens western coast. The combined discharge was estimated to be about
120,000 gpd including contaminated stormwater.
The site was in an area previously classified as non-industrial in
Stenungshund, Sweden. Oxochemicals and phthalate plasticizers were the
principal compounds scheduled for production. The wastewater treatment
facilities had to be installed and operable at the time of plant start-up.
The test work included primary treatment of the phthalate ester waste
stream and of the partial oxidation wastes, and biological treatment of
the total combined wastewaters using a rotating biological contactor.
The development program was initiated on 10 June 1977 and completed on
12 August 1977. These tests and the resulting conclusions are summarized
herein.
APPROACH
A sample representative of the combined wastewaters projected for
the new facility was not available at any single site. However, some of
the chemical process routes included in the proposed design were in operation
at various locations around the world as part of other manufacturing operations,
Consequently, it was necessary to calculate the wastewater composition
using material balances and process experience. And then, where it was
possible to obtain access, a limited number of selected wastewater streams
were gathered at operating process units similar to those under design.
These samples were used for characterization by comparison with material
balances and for evaluation of pretreatment options.
A "recipe" was derived and the combined "wastewater" was synthesized
in Catalytic's Environmental Laboratory in the U.S. to conduct the concept
development study for RBC treatment of the wastes. The wastewater fed
to the units was prepared daily using purchased chemicals, where possible,
or from chemical combinations synthesized in the laboratory using the same
process technology as for the production plant design. After simulating
all the preferred pretreatment steps, the mixture was evaluated using two
parallel bench-scale (25 centimeter diameter disks) 5-stage RBC units.
The units were operated at steady state conditions under various loadings;
and performance data were collected for each stage.
Pretreatment studies on selected process streams, settling and thickening
tests on the biosludges, and screening for sludge dewatering were part
of the study. The need to have the treatment plant on line for the start-up
of the manufacturing facility created a tight schedule. Due to lack of
time the following activities were not evaluated:
662
-------
o Waste variability effect on treatment
o Evaluation for chronic toxicity of the waste
o Biological treatment plant limits
o Dilution effects on the biological treatment plant
o Sludge stabilization
Regulatory Limits
There was no permit available at the time of the study. The following
limitations were obtained from the presiding regulatory board, and these
numbers were used as design basis:
Ton/year
Organics 25
BOD 18
COD 37
Total Solids To be determined
Cyanide To be determined
Specific Metals To be determined
At design flow, this can be further summarized as follows:
Avg Kg/day Avg. mg/1 @ * Avg. Daily Flow
'7
BOD7 49 225 @ 450 m3/day 320 @ 317 m3/day
COD 101 110 @ 450 m3/day 155 <§ 317 m3/day
The Swedish regulatory board requirements stipulated 7-day BOD analysis
pposed to 5-day BOI
Waste Characterization
as opposed to 5-day BOD. All BOD's run during the study are BOD_.
Material balance data was obtained through the client from all the
contractors involved in the design of the whole complex. These were reviewed
and additional data were then requested more specific to waste discharges,
since some of the process materials balances were not entirely adequate
for this purpose. Additional information, derived from their existing
processes, was obtained -from the client, and also from the licensee of
another of the processes. Wastewater "formulas" were developed, including
process washes and process pad wash downs and stonnwater. Following sampling
and analytical characterization, treatability studies of certain process
streams were run in order to evaluate pretreatment. The formulas were
then refined in order to characterize the waste after pretreatment steps.
This was done in order to make a judgement on whether or not the final
concentration of any components would be detrimental to the biological
process at concentrations present in the combined wastewater, or would
exceed effluent limitations following biological treatment.
*450 m /day is the projected wet weather flow (includes storm water)
3
317 m /day is the projected dry weather flow
663
-------
Table A summarizes some of these formulations. Due to their possible
proprietary nature, the * components are not identified specifically.
As can be seen from the table (Column A), there was a considerable quantity
of "unidentified organics" that were derived from the material balances.
The rest of the columns have these quantities distributed proportionally
among the other organic components. Many of these unknowns may be byproducts
of the process reactions. Some of the components had to be chemically
synthesized in our laboratory using the proposed process technology. This
necessity had one possible advantage. By synthesizing the chemicals, some
of these unknowns are possibly included in the mix.
INVESTIGATIVE PROGRAM
Some of the process streams contained pollutants that could be harmful
to biological treatment; or that might be of regulatory concern, but not
significantly removed by biotreatment. Considerable characterization and
treatability screening was conducted on the samples that were obtained
from similar existing processes.
Acute microbiological toxicity studies were conducted separately on
most of these constituents, in the form that they were added to the synthetic
wastewater used in the "recipe". All the levels of acute toxicity were
significantly greater than the levels that were expected in the wastewater
derived from the material balance. Chronic toxicity was not observed,
although the study period was too short for a meaningful evaluation of
this aspect. Pretreatment studies were run, however, in order to gain
some assurance of the feasibility of removal of certain of these pollutants
from the process discharge, should it become necessary. These pretreatment
studies are discussed briefly followed by discussion of the RBC study.
Pretreatment Screening
During an on-site visit to two existing chemical plants, samples were
taken and jar tests were conducted on a "Sour Water" wastewater that was
a product of one of the process steps.
The chemicals evaluated for removal of the suspended solids, metals,
and sulfide in these two streams were lime (CaOH); aluminum sulfate (A1_(SO,)-.
18H_0); potassium hydroxide (KOH) in combination with ferric chloride (Fed-.
6H26); potassium hydroxide (KOH) in combination with ferrous sulfate (FeSO,.
7H»0); and lime (CaOH) in combination with ferric chloride (FeCl-.6H20).
The results of the jar tests conducted on-site were not conclusive.
These tests were repeated in the laboratory, and based on these repeated
tests and the improved treated suspended solids level, ferric chloride
(FeCl«.6H20) was found to be the best chemical treatment scheme for these
waters. However, the dosage of ferric chloride was not optimized for metals
or sulfide removal, and would not be until the plant was in operation.
A high-molecular weight polymer was also needed to improve the settleability.
The following table summarizes the results of settling without chemicals
and the ferric chloride treatment.
664
-------
TABLE A
Summary of Waste Characteristics Tables
en
en
UT
Flow, cubic meters per day
*Unidentified organics, ppm
*0xochemical, ppm
*0xoalcohol, ppm
*0xochemical, ppm
*Plasticizer intermediate, ppm
*Plasticizer5 ppm
Isopropanol, ppm
Butraldehyde, ppm
Naphtha, ppm
Sodium Formate, ppm
Hydrazine, ppm
Trisodium Phosphate, ppm
Sodium Sulphate, ppm
*Heavy Metals, Catalysts & Iron
Sulfide, ppm
Chloride, ppm
Cyanide, ppm
Ammonia, ppm
Settleable Solids, ppm
Dissolved Solids, ppm
BOD, ppm
COD, ppm
317
750
60
6
246
650
246
B
317
95
10
395
1035
395
4
54
11
11
3763
1-2
35
44
1
35
123
4000
5
86
11
11
3763
1-2
35
44
1
35
123
4000
Dry Weather
317
95
10
35
35
70
395
50
10
90
11
11
3763
0.2-0.4
10
150
0.2
35
1
4000
1100
1900
Wet Weather
450
67
7
25
25
50
278
35
7
63
8
8
2655
0.2-0.3
7
106
0.1
25
Trace
2818
775
1340
95
10
35
35
70
395
50
10
90
11
11
3763
0.4
10
150
0.2
35
1
4000
700
1200
A - Waste characteristics from material balance
B - Synthetic Waste Composition (no pretreatment)
C - Synthetic Waste Composition (with pretreatment)
D - Synthetic Waste Composition Fed to RBC
-------
Sample 1
Raw
Gravity
Settled 2 hrs.
BOD
COD
SS
TOG
NH,
CN
*Heavy Metal
*Heavy Metal
Fe
Phenol
Sulfide
8.0 mg/1
1412.0
128
62.0
mg/1
mg/1
mg/1
0.74 mg/1
0.6 mg/1
0.39 mg/1
5,8 mg/1
.01 mg/1
0.23 mg/1
5.0
39.0
43.0
73.0
62.0
1.7
0.2
BOD
COD
SS
TOG
NH,
CN4
*Metal
*Metal
Fe
Phenol
Raw
2.0
237.0
60.0
132.0
500.0
15.2
73.0
mg/1
mg/1
mg/1
mg/1
mg/1
7 mg/1
2 mg/1
.05 mg/1
0.9 mg/1
.01 mg/1
0.23 mg/1
Gravity
Settled 2 hrs.
Sample 2
FeCl, Treated
5 mg/1
23 mg/1
14 mg/1
31 mg/1
62 mg/1
0.13 mg/1
0.2 mg/1
0.1 mg/1
2.9 mg/1
.01 mg/1
0.28 mg/1
FeCl, Treated
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
20.0
256.0
34.0
109.0
500.0
10.5
1.9
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
7.0
169.0
13.0
23.0
500.0
13.2
0.2
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
0.98 mg/1
0.25 mg/1
7.25 mg/1
0.34 mg/1
0.25 mg/1
9.6 mg/1
0.55 mg/1
0.25 mg/1
7.0 mg/1
In addition to the jar tests conducted above, a sample was spiked
with NH,, sulfide, CN, heavy metals, and Fe. A portion of the spiked sample
was allowed to settle for one hour untreated and the supernatant was submitted
for analysis. A portion of the spiked sample was treated by the addition
of 400 mg/1 ferrous sulfate with the pH adjusted from 8.2 to 10.0 with
7 mg/1 of a 10 N solution of sodium hydroxide. Dow A-23 was also added
at 5 mg/1. The sample was allowed to settle for thirty minutes and the
supernatant submitted for analysis. Another portion of the spiked sample
was treated with the same chemicals, however, ten minutes was allowed to
elapse between each chemical addition to simulate separate mixing tanks.
The remaining portion of the spiked sample was treated by the addition
of 400 mg/1 ferric chloride with the pH adjusted from 8.1 to 10.0 with
7 mls/1 of a 10 N solution of sodium hydroxide. The results of these tests
are summarized in the following table.
Raw
NH, 47
Sulfide 3
CN 0
Metal 0
Metal 1
Fe 3
.4
.23
.35
.77
.9
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
Spiked
2,000 mg/1
0.9 mg/1
30.23 mg/1
60.05 mg/1
80.8 mg/1
46.9 mg/1
Settled
1 Hour
1,600
0.9
12.1
41.7
4.38
mg/1
mg/1
mg/1
mg/1
mg/1
FeSO,
Single
Addition
1,450
1.1
7.5
2.77
30..4
5.09
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
FeSO,
Multiple
Addition FeCl,
1,450
1.1
7.2
1.56
26.0
4.12
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
1,300 mg/1
1.1 mg/1
7.55 mg/1
1.63 mg/1
30.1 mg/1
5.68 mg/1
666
-------
Data from the preceding jar test indicate that either chemical treatment
procedure would be effective in removing the metals. The quantities of
chemicals used in the jar tests were not optimized; this would require
more testing after plant start-up.
During an on-site visit to one of the plasticizer plants, wastewater
samples were taken from the water phase of the plasticizer catch tanks
and submitted for analysis. The results of these analyses are contained
in the following table.
Tank f 1 Tank#2 Tank #3
BOD 16,000 mg/1
Soluble BOD 13,500 mg/1 8,000 mg/1 8,000 mg/1
(filtered)
COD 25,800 mg/1 15,000 mg/1 13,400 mg/1
Soluble COD 21,600 mg/1 13,600 mg/1 12,500 mg/1
-(filtered)
Suspended Solids 578 mg/1 338 mg/1 414 mg/1
Total Solids 2,600 mg/1 1,800 mg/1 2,300 mg/1
pH 6.7 5.5 6.7
Gravity separation tests were conducted on the plasticizer catch tank
wastewater. The results of a typical test where the water phase from the
catch tank was settled for 2 additional hours are shown in the following
table.
20°F 80°F 155°F
COD 17,008 mg/1 17,401 mg/1 15,951 mg/1
BOD 11,500 mg/1 9,600 mg/1 9,900 mg/1
TOG 4,600 mg/1 4,260 mg/1 4,200 mg/1
Distillation/stripping screening tests were conducted on the catch
tank wastewater, and also the use of vacuum was evaluated. This testing
confirmed that adequate allowance for ambient gravity separation should
be included in the design, and that further removal or recovery by enhanced
physical separation techniques would not be fruitful.
Biological Treatability
Evaluation of the rotating biological contactor was the main thrust
of the treatability studies. Two identical bench-scale RBC's were run
in parallel to gain the most data within a minimum time period.
Each unit consisted of five stages. The first stage contained five
circular styrofoam discs, 0.25 meters in diameter, with a total surface
area of 0.51 square meters; the second stage contained four discs with
a total surface area of 0.408 square meters; and the remaining three stages
each had three discs for a surface area of 0.306 square meters in each
stage. The unit was constructed using a 12-inch diameter PVC pipe, cut
longitudinally in half, as the basin, PVC sheet was cut and formed to
667
-------
section off each individual stage. The discs were fastened to a wooden
shaft. Each disc was spaced *j-inch apart on the shaft. The actual volume
of water contained in the combined disc sections of the unit was 11.5 liters.
Contact times through the 5-stages ranged from 5 to 11 hours for the different
loadings. The flow pattern through the biocontactor was plug flow; determined
by dye testing. The flow entered the head end of the shaft, and criscrossed
through the various stages through alternately placed overflow weirs.
The clarifier section of the unit was the last stage of the unit without
discs. Sludge was syphoned from the bottom of the clarifier section each
day. The drive unit was a variable speed motor and controller. Nutrients
were added in the form of ammonium phosphate to the synthetic wastewater
in a slight excess of what would be needed to support a biological culture.
The tests were carried out over a nine-week period. The first week and
a half was needed to start an active biological culture growing on the
discs. During this period, the synthetic wastewater fed to the discs was
a composition containing small amounts of the chemicals determined in the
material balance. After the acclimation and start-up period, the biosystems
were run at steady state using the "recipe", with samples being taken at
each stage to gain the information required to predict the surface area
necessary to meet the effluent limitations. Three runs were made at design
concentration at three different hydraulic loadings. The performance data
for these are summarized in Table B. The data were analyzed statistically
to ensure steady state and normal distribution. An additional run was
made at twice the design influent concentration at the same hydraulic loading
as the intermediate loaded system of the first 3 runs. The data from that
system are not summarized, but were used to test some of the design correlations
that will be discussed and graphically presented.
The use by others of peripheral disc speed as a design parameter has
led to higher RPM's and excessive aeration in smaller units which can lead
to scale—up problems. That is, full-scale performance something less than
bench or pilot—scale predictions. To avoid that problem, all our study
work is performed by operating the units at the minimum rotation speed
that will maintain a dissolved oxygen concentration of 1.0 mg/1 in the
liquid of the first stage. During the different runs the rotation of the
shaft varied according to loading from 6 rpm to 10 rpm.
Settling and thickening tests using a one-liter graduated cylinder
were conducted. The sludge that sloughed from the rotating biological
contactor during the test period did so in very large gelatinous pieces;
uncharacteristic of other industrial RBC sludges we have encountered.
Much of it had to be physically removed from each stage because it would
not pass through the small channels of the bench-scale biocontactor. In
addition, a whole stage would slough all at once. No two stages sloughed
simultaneously, however, and the units operated without any excessive loss
of efficiency through a sloughed stage.
The settling and compaction time for the sludge was approximately
one hour,and the sludge compacted to a concentration of approximately one
percent. The effluent TSS concentration in the biocontactor averaged about
20 mg/1. A typical settling time is shown in Figure 9.
668
-------
TABLE B
cr>
Influent
Stage 1
Stage 1,2
Stage 1,2,3
Stage 1,2,3,4
Stage 1,2,3,4,5
Influent
Stage 1
Stage 1,2
Stage 1,2,3
Stage 1,2,3,4
Stage 1,2,3,4,5
Influent
Stage 1
Stage 1,2
Stage 1,2,3
Stage 1,2,3,4
Stage 1,2,3,4,5
CODgm/m /day
75.1
41.7
31.3
25.0
20.9
49.0
27.2
20.4
16.3
13.6
114.3
63.5
47.6
38.1
31.7
RBC #1 RUN #1
32.98 liters/total m /day
CODmg/1 % Reduction
1160
556
352
173
73
45
12.24
1154
371
223
114
62
50
48.24
1210
585
463
274
165
77
_
52.1
69.7
85.1
93.7
96.1
RBC #2 RUN #1 _
9
liters/total m /day
_
67.9
80.7
90.1
94.6
95.7
RBC #1 RUN #2 .
/
liters/total m /day
_
51.6
61.7
77.4
86.4
93.6
BODgm/m/day
12.7
8.7
27.0
21.6
18.0
BODmg/1
708
% Reduction
17
706
687
125
71
22
97.6
99.2
81.8
89.7
96.8
-------
Sludge stabilization was not one of the items included for laboratory
evaluation, however, brief screening tests were conducted on unstabilized
sludge for dewatering applications. Centrifugation screening tests were
conducted to determine the feasibility of the application of centrifugation
for dewatering of sludges generated by the installation of a rotating biological
disc. A gravity thickened sludge with a suspended solids concentration
of 10,200 mg/1 was screened. The sludge could only be concentrated by
centrifugation to 30,600 tng/1 which is a sludge containing 96.3 percent
moisture. Gravity thickened sludge would not filter in a vacuum filtration
screening test. Separate chemical additions of lime, alum, Fed.,, nonionic,
cationic, and anionic polymers did not improve either process. At this
point the study ended, and a recommendation was made that due to the relatively
low volume of sludge projected, the design of any dewatering equipment
be done after the plant had been started—up, and the actual sludge generated
could be evaluated after stabilization, using on-site pilot equipment.
RBC DESIGN
Several of the design bases that were available at the time of the
study were evaluated, as well as a number of additional empirical approaches
of our own, based on fitting a line(s) to the data (rather than vice-versa).
The design chosen was coincidentally equivalent to one of the laboratory
runs. However, the data from that run falls on a straight line that includes
the data from the other loadings that were run. These several approaches
and the recommended design parameters will be explained in this section
and presented graphically.
For several reasons that will also be further delineated, COD was
used as the design parameter.
Design Parameters
Of the parameters mentioned in the regulatory limits section, dissolved
organics and the resultant oxygen demand parameters were those considered
for the RBC design.
Of the others, total solids (essentially total dissolved inorganic
species) would not be treated in a biological process and should future
limits require treatment, other unit processes and/or source controls would
have to be evaluated separately. Since the ultimate discharge of this
effluent is to the sea, it is not expected that the projected dissolved
solids concentrations would be in excess of any limits. (Also effluent
suspended solids are anticipated to be less than 30 mg/1).
Projected cyanide and heavy metals concentrations are low enough not
to effect the biological system, and would not be expected to be in excess
of future limitations. However, should the material balance projections
be inaccurate for these paramters, or if regulatory limits dictate, space
has been left near the major process sources of these pollutants for future
pretreatment to reduce these parameters before going to the RBC system.
670
-------
As to the design parameters, the "organic" limitation is somewhat
elusive, however, the following approach was applied:
In Column C of Table A showing the projected wastewater constitutents,
a calculated BOD and COD are presented. The COD values were calculated
using theoretical oxygen demand. For the major components (those accounting
for 90% of the organic loading) the ratio of weight of calculated oxygen
demand to weight of organic was never lower than 2.4. That is, one gram
of "organic" exhibits a theoretical COD of 2.4 gms. From Column D of Table A
the actual COD was slightly lower. Even reducing the ratio by that much
still results in COD to organic ratio of 1.52. The ratio of "organic"
to COD as derived from the regulatory limits is 1.48. This is all based
on the influent to biotreatment, but it appears safe to assume that if
the required COD reduction is obtained, the "organic" will be sufficiently
removed. Also, although not discussed in detail here, should this parameter
ultimately prove to be regulated using "TOG" or "oil and grease" measurements,
the data also project adequate removal.
An examination of Table C and Figure 1, derived from that table shows
that BOD removed at all loadings never falls below 80%. At all reason-
able COD removals, BOD will be more than adequate to meet effluent targets
as discussed in the next section. Because of this, and as COD is a more
efficient operating analysis to obtain, the interstage data and the resultant
correlations were derived using COD as the design parameter. About 40%
of the COD analyses runs had accompanying BOD analyses. The samples were
not filtered, but were very low in suspended solids. Some data were taken
on filtered vs. unfiltered COD and the difference was not significant,
and is not of concern for the RBC design.
TABLE C
% COD
Removed
96.1
95.7
*77.4
*86.4
93.6
95.3
92.1
80.0
COD
Applied No. of
gms/m /day Stages
20.9
13.6
47.6
38.1
31.7
39.6
47.5
59.3
5
5
3
4
5
5
4
3
Eff.
mg/1 COD
45
50
274
165
77
102
171
431
Eff.
mg/1 BOD
17
6
125
71
22
22
71
221
BOD: COD
Ratio
.38
.12
.46
.43
.29
.22
.42
.52
% BOD
Removed
97.6
99.2
81.8
89.7
96.8
98.3
94.3
82.4
Design Targets
A summary of the data used to determine the design effluent targets
is shown in Table D.
671
-------
TABLE D
Design Parameters
Wastewater Characteristics
: •-'••"- «j
Flow - Dry weather 317 m_/day
- Wet weather (Design) 450 m /day
BOD 222 kg/day
COD 380 kg/day
Influent Concentrations
Dry Weather Wet Weather
BOD 845 mg/1 495 mg/1
COD 1200 mg/1 700 mg/1
BOD:COD ratio 0,58
Effluent Requirements
BOD? 18 metric tons/year (49 Kg/day-avg.)
COD 37 metric tons/year (101 Kg/day-avg.)
Extrapolated Effluent Requirement Concentrations
Dry Weather Wet Weather
BOD? 155 mg/1 110 mg/1
COD 320 mg/1 225 mg/1
BOD:COD ratio 0.485
Based on the projected numbers, an average of 73% COD removal and
78% BOD removal would appear to be adequate for meeting the regulations,
and would keep capital expenditures as low as possible. However, there
are several reasons, some rather obvious, why such a design basis would
be improprietas.
The primary reason is that such a design is at the limit of the require-
ments. Also, the resultant loading would surely place the system in a
log growth mode and any shift in efficiency relative to loading and varia-
bility would necessarily be drastic. Although the projected waste is quite
biodegradable, such a loading would be in excess of prudent and accepted
design loadings with no allowance for other adverse factors such as:
1. The raw waste load is a projection and some allowance might be
made for the probability that the actual concentration and loading
will not be exactly as predicted.
2. Although every effort was made to minimize conditions that might
adversely effect scale-up, some allowance might be made for this
aspect.
3. It might reasonably be assumed that the final regulatory permit
would require daily averages and maximums.
672
-------
As mentioned earlier, wastewater variability, chronic toxicity, and
low loading conditions were not part of the laboratory study either, and
any "allowances" for any of these or the factors listed above would be
purely subjective. However, a COD removal of 85% and the resultant BOD,
removal of 90% or greater was chosen as the lower limit of a reasonable
operating range, and recommendations were based on the minimum requirements
to operate at that average condition. Although the loading of BOD per
area of RBC at that condition may still appear high to some, the extent
of the "allowances" such as design will afford will be further detailed
as the design is discussed.
Design Evaluation
As a point of reference in the following discussion and figures, "loadings"
are expressed generally as gms COD/sq meter of surface area (eq. Kg/1000 m ),
and the hydraulics are expressed as cu meters/day. Among other evaluations
the following linear plots were prepared using the study data:
Applied COD
Figure 2 wt/cumulative area
Figure 3 wt/cumulative area
Figure 4 wt/cumulative area
Figure 5 wt/cumulative area
Figure 6 wt/cumulative area
Figure 7 wt/area for each stage
vs. COD
% removed
Concentration Remaining
Wt removed/area
(of 1st stage)
*Wt removed/cumulative
area
**Wt removed/cumulative
area
Wt removed/area
for each stage
Further explanation of the above summary table is contained in the
figures themselves, as well as the following text.
Figures 2 and 3 are variations of plots that were part of two design
approaches that were generally considered viable at the time of design.
The percent removal basis shows no real pattern. The concentrations plot
is not usable either. If the effluent limitations were concentration based,
it may have been of some use; however, the first stage data did not plot
in the realm of the other stage data and a single line did not emerge,
rather three separate lines. If substrate concentrations were dependent
of loading, one would expect a single line to emerge. Also, since the
effluent limitations were based on weight (mass), a "mass removal" approach
might be better suited if the data fit.
*Points for each run drawn.
**Points for each stage of the different runs drawn.
673
-------
In Figures 4,5,6, & 7, the mass applied is plotted against mass removed;
however, they are normalized in several different manners. In Figure 4
the removal at each stage is shown as a function of the loading applied
to the 1st stage (gms/cm /stage 1) as that load is spread out over a larger
area (loading is expressed gms/cumulative cm ). Again a family of curves
developed with the first stage data are askew from the others. In looking
at the percent removal at each stage it was noted that although percent
removal varied, the points from the same stages (except stage 1) of the
different systems when connected formed a line, and that all these when
extrapolated, intersected at a single point (Figure 4). This exercise
still did not provide a usable design approach. Figures 5 & 6 show the
same plots except, that the loading and removal are normalized based on
the cumulative area. Figure 6 was used for design. It shows that mass
removal is a function of mass loading, as well as number of stages, and
that the number of stages continues to diminish in importance down to a
level, in this case of 20 COD gms/cm applied, where the number of stages
would no longer appear to affect mass removal at a steady state. It also
shows that the enhanced effect of staging diminishes with increasing stages;
with 4 stages being the limit beyond which little is gained by increased
staging. This last piece of information, of course, is no discovery, but
it does show an agreement of this approach with established fact.
This plot was used for design projections. Beyond a single stage
system, removal can be projected for a design loading based on number of
stages. This plot was tested by using the data from the very highly loaded
fourth system to see if concentration difference may have an effect. (This
fourth run was at double the concentration of the previous runs). In Figure 6A,
the first and second stage data for that system does not fall on the curves,
but actual performance is better than projected. Third stage data fall
in line, but is again slightly better than projected. It cannot be determined
with certainty whether these differences are due to higher concentration
or to the overall higher loading, but in either case the design approach
seems to be a safe one. In lieu of such study data, an attempt to look
separately at removals at lower concentrations was made using Figure 7.
Here, each stage is plotted separately; gms applied vs. gms removed, or
in 7A, gms applied vs. gms remaining. There is some scatter particularly
on the higher loaded runs. A least squares regression was made using the
two lower loaded systems (see Figure 7B). This line had a correlation
coefficient of 0.961 and intercepted the Y-axis at 4.4 (less than 4.4 gms/m
applied per stage removes zero gms COD). Using the slope and plotting
3 stages in the line shows increased system removal as the initial loading
is decreased. Less mass is removed, but a greater percent of the applied
load is. Low loading could not be evaluated with actual data.
Design
Catalytic1s Environmental Systems Division advised Catalytic International,
who was preparing the final design, of our determination of the minimum
area and number of stages that should be included based on the scope of
our study and the discussed evaluation.
674
-------
Although Figure 6 indicates that a two-stage design could theoretically
meet the effluent requirements, there were two primary reasons why at least
3 stages were recommended.
These studies were at steady state and as loading might shift up and
down, considerable efficiency could be gained at higher loadings for the
same area of contactor surface using a 3-stage vs. a 2-stage design. Three
stages or greater would assure system operation even if one stage were
malfunctioning. This was of specific concern because of the way the sludge
sloughed in our study work. It appeared possible that a whole stage might
be "denuded" at once and although we saw no great reduction in overall
efficiency from this occurrence on the lab scale, it appeared prudent to
allow for 2 fully covered stages to be in operation at any one time.
Looking at Figure 6 again, 85% COD removal for a three stage system
plots out to be 30 gms/cm applied. If the system were at a constant steady
state, the average regulation limit would be just met at 40 gms/cm applied
or 32 gms/cm removed. Restated, using minimum area recommended (for 85%
COD removal) of 12,667 cm , 8 gms/cm remaining x 12,667 cm is 101 Kgs/day
in the effluent. The recommended minimum design would leave 4.5 gins cm
remaining at steady state conditions.
It was difficult to obtain quantitative sludge production data in
the confines of this study. Some of the difficulties with sloughing sludge
were discussed earlier. However, a range of numbers were obtained from
the study work. We used the conservative end of that range (the highest
sludge yields). The number used was still within the realm of experience
(0.33 gins of solids per gm BOD- removed).
The RBC study was conducted at temperatures ranging from 18°C to 28°C
and no data for evaluation of cold temperatures operation was collected.
Since the biological unit is preceded by considerable equalization and
the units have a rather long contact time, it was recommended that the
systems be covered and heated.
SYSTEM DESIGN
The final design as completed by Catalytic International in London
is summarized in Figure 8, and includes surge capacity, oil removal, equaliza-
tion, 2-stage neutralisation, RBC, and final clarification.
Other recommendations supplied from the laboratory data were clarifier
sizing, degree of pre—treatment required, and solids (bio-sludge) production.
The final design is a 3-stage system; two air-driven shafts with the
second shaft baffled to provide the second and third stages. The system
totals 20,000 sq. meters of effective area. The shafts are covered and
the influent water is heated as it comes from the equalization basin by
direct steam injection. The effluent then flows to a single circular clarifier
which overflows to a basin with about 8-hours hydraulic retention time
675
-------
where it will be continually monitored for flow and quality. If necessary,
it can be diverted from this basin back to the surge basin to go through
the system again.
The final design also provides 15 day heated aerobic digestion of
sludge followed by centrifugation for dewatering.
SUMMARY
The use of bench-scale RBC systems broadens the scope of applicability
for this unit process. It becomes another option in areas where completely
mixed systems could only be evaluated before, due to the logistic limitations
of pilot scale studies. Bench-scale units can establish feasibility and
provide design data, allowing technical and economic comparisons to other
bio-processes and a rational design when it is the unit process of choice.
The study herein discussed illustrates that there is a reasonable
approach to defining and testing wastewaters and process plant effluents
when they do not actually exist on an entity. This approach allows a treatment
plant design that can better meet the needs of the grassroots process plant
once it is constructed and operating.
676
-------
FIGURE 1
§
o
o
10
OS
a
o
ca
100
90
80
% Removal
70
60
50
677
-------
FIGURE 2
100
20
40 50
60 70
X Reduction
678
-------
ON
~4
VD
120
160
o
o
o
r-80
6)
a.
i-*
3
90
-60
a.
SB
40
20
X
RBC
RBC
RBC
//I
112
til
RUN
RUN
RUN
ill
til
100
200
300 400 500 600
COD Remaining mg/1
-------
000 0 0 0
-------
ON
00
8
n
*»
110
COD gins Removed/m of 1st Stage
-------
oo
120
130
COD gms removed/in of 1st Stage
-------
120
100
o
§
80
cr>
oo
H-
O.
1
a>
Q.
40
20
to
A
RBC
RBC
RBC
RBC
12
til
Run
Run
Run
Run
#2
£2_
//I
//I
m
u>
10
20
30
40
50
60
70
80
90
COD removed gms/m
-------
oa
120
100
I
*
•a
i.80
60
40
20
1st
2nd
•Srri
4th
5th
tage
tage
f"Apj*_
tage
tagc
i
s
10
20
30
40 50 60
COD removed gms/m
70
80
90
-------
00
Ui
120
100
a
o
a
I" 80
•o
j-*
H-
ro
o.
1
HI
60
40
20
1st
2nd s
trvl g
s|age
age
4th s
5th s
age
age
so
w
10
20
30
40
50
60
70
80
COD removed gms/m
90
-------
oo
120
100
o
o
a
ai
v 8C
M
H-
a.
-»
NJ
—»
w
Sa fn
w 60
(D
40
20
X®
%
A
A
X
a
*
X
1
B
k
fi
X
A
C3
X
o
RBCll
RBC//2
RBCtfl
Run)? 2
Run#l
Run//l
FIGOSE 7
10
20
30
40
50
60
70
80
90
100
COD gms removed/in /Stage
-------
OS
00
•vj
60
50
o
§
1
in
H
§ 4°
f*
3
M»
y
w
•>,
ra
o 30
o
X
w
rf
0
OS
n
20
10
S
•
X
o
A
©
O
X,
*x
X
4
A
i
e
A
X
9
RBC
BBC
RBC
RBC
112 Run
91 Run
»2 Run
i>l Run
#2
02
#1
#1
FIGURE 7A
20
30
40
50
60
70'
80
90
COD ems reraainltuz/m /Dav/Staee
-------
oo
c»
120
100
S
00
80
re
o.
s"
(O
cn
rt
to
39
n
60
40
2C
Least
RBC
//I
Squai e
& 2
Reg:
Run 0:
essioi
10
20
30 40 50 60
COT) omq rpmovf»ri/Tn2/Sfni)r<»
70
80
90
-------
PROCESS
UNITS
•r
8 OUR «AS
STRIPPER
COOLIN4
TOWER
REMOVAL
I
_[
ir....r....."..t.r...
M——FERROUS SULPHATE
O
ACID
ACID-
U--"
SUR«E BASIN
FLOW EQUALIZATION
OO
EQUALIZATION
S BTAfI
NEUTRALIZATION
t
NUTRIENT
i rr STA«I
i
t NO ST*8E" *
.Rjc.
I
1
1
1
1
L_
i
DISC
TO
«—
CONTROL SASIN
( MONITOR IN« STATION)
STEAM
LINE KEY A|R ANO
EMERSENCY SEWER STEAM """*
BYP
HUH
HARSE ' ' !f?
SEA CtE
COM
ROOUCTS
pH PROCESS WASTES
:ESS SEWER
AN WATER SEWER L^NDRLL10 *
•INED WABTP LANDFILL
i
CLARIPIER
1
I SLUO«E
AEROBIC OWESTER
I
i
OEWATERIN*
, *.T0 SURBE
^^ 8ASIN
FI6URE 8
-------
FIGURE 9
SETTLING STUDY
1000
Dace: Aug 1977
SamplerRBC Sludge
MLSS: 2000-2500
Supernatant SS: 30-50
SVI:
100
30 40
Time In Minutes
690
-------
THE TREATMENT OF SALINE WASTEWATERS
USING A ROTATING BIOLOGICAL CONTACTOR
By
Mark E. Lang
Graduate Research Assistant of Civil Engineering
Stanley L, Klemetson
Associate Professor of Civil Engineering
Colorado State University
Fort Collins, Colorado, U, S. A.
INTRODUCTION
Recently aquaculture has become increasingly useful for the production
of food for world-wide consumption. Aquaculture is the raising and harvest-
ing of aquatic organisms in a controlled environment. As the use of aquacul-
ture increases, the need for the development of treatment methods for the
wastewaters generated becomes more apparent. The method of treatment to be
evaluated by this study is rotating biological contactors (RBC).
The type of aquaculture of interest in this study is the raising and
harvesting of MaoTobraohiwn Rosenbergi-i (prawns). Prawn larvae are cultured
in a 30 percent sea water solution. As the prawns mature they are able to
survive in decreasing saltwater concentrations until, at the adult stage,
the prawn are able to survive in freshwater. In a recirculating aquaculture
facility, certain aspects of water quality must be kept within prescribed
limits in order to provide optimum conditions for prawn growth and develop-
ment. The most significant parameters include dissolved oxygen (DO), temper-
ature, pH, unionized ammonia, and organic substances. The rotating biological
contactor will be evaluated for its ability to maintain these parameters at
levels allowing maximum prawn develpoment.
The rotating biological contactor is an aerobic treatment process that
consists of a series of circular discs connected to a common horizontal shaft.
The discs rotate partially submerged within the wastewater. A biological
film (biofilm) is allowed to form on the discs as they rotate through the
691
-------
wastewater contained in the contactor basin. While submerged in the waste-
water, the biofilm adsorbs organic matter and as the discs rotate, the bio-
film containing the adsorbed organic matter is exposed to the atmosphere.
The microorganisms, comprising the biofilm, use the oxygen from the atmos-
phere and the organic matter from the wastewater in a biodegradation-biotrans-
formation process. Waste organics are entirely or partially reduced to their
basic components (including ammonia, carbon dioxide, and water), and the
microorganisms suspended in the water are subsequently removed from the
system by sedimentation.
Rotating biological contactors are becoming increasingly popular in the
field of wastewater treatment for a number of reasons which include: (1)
lightweight and compactness, (2) low power consumption, (3) combination of
functions as a trickling filter and an activated sludge process, (4) high ..
efficiency in oxygen transfer, and (5) ability to achieve nitrification easily.
Success has been found in using the contactors to treat freshwater, domestic
wastewaters. This study, however, has been designed to evaluate the treat-
ment efficiency of rotating biological contactors in a closed marine system.
Previous studies have been concerned with the treatment of saline waste-
waters, and one study performed at the University of Rhode Island dealt with
the use of rotating biological contactors to treat saline domestic wastewaters.
This study will be somewhat different than others previously performed in
that nitrification of saline wastewaters will be evaluated. Ammonia is toxic
to aquatic organisms at extremely low concentrations. These toxic concen-
trations are dependent upon pH, temperature, and ionic strength and will be
determined for each of the sampling programs in order to evaluate the amount
of nitrification necessary at the different salinity levels.
In summary, it has been hypothesized that rotating biological contactors
can be used effectively to treat wastewaters such as those generated in
an aquaculture facility. This pilot study will concentrate on the technical
feasibility to achieve nitrification in a closed saline system using rotating
biological contactors.
LITERATURE REVIEW
Biological Treatment Cost Comparison
As previously stated, one of the reasons for the increased popularity
of rotating biological contactors is the low power consumption when compared
to other biological treatment methods. Poon et al., based on work from a
pilot study,estimates rotating biological contactor power requirements to
be 45 percent lower than an activated sludge unit of equivalent capacity
(0.8 MGD).2 in a study of winery wastes, La Bella et al., found the capital
costs of a RBC and activated sludge to be equal. Labella, however, found
the yearly operational costs to be approximately $6,000 less for the rotating
biological contactor than the activated sludge for flows between 0.34 and
0.44 MGD.3
System Loadings
Rotating biological contactors have been sized using a number of methods
including hydraulic loading, detention time, and total organic loading. The
method utilized in this study was total organic loading.
692
-------
The organic loading is the total organic mass applied to the system over
a period of time and is determined by multiplying the hydraulic loading rate
and the influent organic concentration, yielding a unit of mass per time
per area. Cook and Kincannon, in evaluating trickling filters as a fixed-
film biological treatment process, found the total mass of BOD and/or COD
applied to the system to be of importance when designing a process. The
total mass applied takes into account both the flow rate and the organic
concentration of wastewater. The process performance evaluated as COD
removal efficiency was dependent upon the total COD applied to the system as
gram/hr/m^, rather than its concentration or flow rate.-3 The COD removal
efficiency remained constant at constant total loadings regardless of whether
the load was caused by high organic concentrations and low flow rates or low
organic concentrations and high flow rates.
Poon and Mikucki, while testing rotating biological contactors for the
treatment of saline wastewaters, agreed with Cook and Kincannon in concluding
that although the hydraulic loading is important in the rotating biological
contactor- process design, organic loading should be considered equally as
important. High hydraulic loadings applied to low BOD influents will yield
the same removal efficiency as low hydraulic loadings applied to high BOD
influents." The work performed at Colorado State University was based on
the total organic load in order to monitor the effect of both the hydraulic
load and organic concentrations on the rotating biological contactor.
NITRIFICATION
As previously mentioned, the main objective of this stidy was to obtain
an optimum loading rate for nitrification of a saline wastewater. Nitrifica-
tion is the biological transformation of ammonia to nitrate and nitrite.
Ammonia, even at low concentrations, can have acute toxieity effects on
aquatic organisms and deplete the dissolved oxygen concentration while being
converted to nitrates in receiving waters. The toxicity of aqueous solutions
of ammonia can be attributed to the NH3 species. The toxicity of ammonia is
dependent upon the pH and concentration of total ammonia (NH^-J- NH.,). ^ There
are other factors which also affect the concentration of the NH3 in solution;
the most important of which are temperature and ionic strength. The Red
Book states that the concentration of the NH-j increases with increasing
temperature and decreases with increasing ionic strength.°
In a review of the EPA Red Book, the American Fisheries Society states
that the NH^ /NH3 ratio is a function of the activity of the changed speices
and the total ionic strength of the solution. Thruston et al., state that
there is a slight decrease in the unionized fraction of the total ammonia as
the ionic strength increases in a dilute saline solution (less than 40 per-
cent sea water)."
The American Fisheries Society also states that a decrease in the dis-
solved oxygen concentration will increase the toxicity of ammonia. It is
hypothesized that a reduction in the dissolved oxygen concentration would be
accompanied by an increased ventilation rate by organisms, increasing the
exposure to unionized ammonia.10
Nitrogen in the form of ammonia is converted to nitrate in two steps
by autotrophic nitrifying bacteria: Nitrosomonas and Nitrobacter. The
693
-------
reactions as presented by Metcalf and Eddy, Inc. are as follows:
NH+ + 1.5 0
N02~ + 0.5 02 -Nitrobacter
These reactions can be combined to read:
NH/4" + 202 - -*• NO + 2H+ + H20
The nitrifying organisms needed to convert ammonia to nitrate are present
in almost all aerobic biological treatment processes. In many instances,
however, the numbers are limited. Nitrification is brought about or encour-
aged by suitable adjustment of the operating parameters, namely the reduction
of total applied loads.-'-1 Antonie et al., state that nitrification in a
rotating biological contactor begins when the BOD approaches 30 mg/£. At
this concentration, nitrifying organisms are able to compete with the more
rapidly growing carbon oxidizing organisms and establish themselves in the
process. With the establishment of nitrifying organisms, nitrification
is allowed to proceed rapidly until at a BOD concentration of approximately.
10 mg/2, the nitrification is complete. 12
Weng and MoloE found using a pilot scale RBC and an ariticial substrate,
that the chemical oxygen found (COD) must be below 50 mg/fc for nitrification
to occur. 13 Using the artificial substrate, this 50 mg/fc COD corresponded
to a BOD concentration of approximately 14 mg/£. This is less than half the
concentration stated by Antonie for nitrification (approximately 30 mg/fc).
Weng and Molof's results show that increasing the disc surface area increased
the rate of nitrification. In the tests conducted, Weng and Molof found
that nitrification took place only in the stages where the mixed liquor dis-
solved oxygen was greater than 2 mg/£. From a review of the literature, it
may be stated that nitrification can be accomplished using a rotating biolog-
ical contactor operated under the proper conditions.
EFFECTS OF CHLORIDE CONCENTRATION
Researchers have found that wastewaters with high chloride concentrations
may be treated satisfactorily. 14,l->»lo -j^g chloride concentration does,
however, play a major role in the treatment of wastes.
There is a marked difference between biodegradation-biotransformation
rates of the suspended fraction in saltwater and freshwater environments,
It is believed that, for a microbial population such as those present in
freshwater, domestic wastes, certain exoenzyme action may be inhibited in
an environment with high chloride concentrations. This exoenzyme action is
generally considered necessary for the metabolism of insoluble substrates,
and its inhibition may result from a change in the surface configuration of
the suspended material or in the configuration of the enzymes themselves. 1^ •
With regard to nitrification, Ludzack and Noran state that the nitrifi-
cation during high-chloride operation was approximately 10 percent of that
expected for the same operation at lower chloride concentrations . 19 This
lower rate of nitrification may bring about the need for longer hydraulic
detention times in order to achieve the desired levels of ammonia removal.
694
-------
Treatment of Saline Wastewaters with RBC
Previous work has been performed on the use of rotating biological con-
tactors to treat saline wastewaters.^" This work, performed at the University
of Rhode Island, focused on the highly saline, domestic wastewaters generated
on the island of Kwajalein in the South Pacific. Poon and Mikucki performed
the plot studies to obtain a secondary treatment method to meet the provisions
of the 1972 amendment of the Water Pollution Control Act (PL 92-500), along
with regulations implemented by the US Environmental Protection Agency. -^
The major concern of the pilot plant study was to demonstrate that
rotating biological contactors could meet the standards set by the national
pollutant discharge elimination system (NPDES) while treating a highly saline
wastewater. In order to do this, Poon and Mikucki conducted a series of
experiments with and without sea water. Initial tests were conducted with-
out the addition of sea water and high chloride concentrations in order to
demonstrate that the pilot plant could successfully treat a wastewater of
Kwajlein's strength to meet NPDES standards. Poon and Mikucki then evaluated
the performance of the rotating biological contactors using various hydraulic
and organic loadings at different chloride concentrations.
The two major findings of Poon and Mikucki were that: (1) the change
in chloride concentrations had no effect on the RBC treatment efficiency
after the microorganisms were allowed to acclimate and (2) the performance
of the RBC is dependent upon the combination of the hydraulic and organic
loadings. This combination of loading results in a loading term of mass per
area per
From the literature review, it can be seen that there is some controversy
within the field as to the effects of various parameters on rotating biological
contactors. These differences do not mean that RBC's cannot be properly
managed; rather, that continued research is necessary in certain aspects of
design and operation.
This literature review has shown that although the objectives of this
study have been met in other studies, they have not been incorporated into
one. Studies have shown that the rotating biological contactor can achieve
nitrification and can successfully treat saline wastes. This study has been
designed to determine the ability of a rotating biological contactor to
achieve nitrification in saline wastewaters, such as those generated by a
closed aquaculture system.
METHODOLOGY
The section on methodology is concerned with the approaches used in the
pilot study in order to obtain values. The methodology of the study is
important since it will have direct effects on interpretation of the data
collected.
The section has been divided into four segments. The segments describe
the pilot rotating biological contactor, the entire pilot system, the tests
performed on the system, and the quality control program followed during the
secondary program. The segments have been arranged in order to describe the
entire system from its initial design to tests performed to analyze the
systems treatment efficiency. '
695
-------
Description of the Rotating Biological Contactor
The rotating biological contactor used for the study was a relatively
small four-stage unit having a capacity of 0.92 ft^ (0.026 m^) with discs
in place. Each stage contained five discs having a total surface area of
26.40 ft2 (2.45 m2), and a specific surface area of 28.7 ft2/ft3 (93.47 m2/
m3).
The rotating biological contactor was constructed at the Engineering
Research Center of Colorado State University. The tank was fabricated from
a section of 12 inch (30.5 m) PVC pipe. The tank ends, staging baffles,
and discs were all fabricated from pelxiglass. The discs have been placed
perpindicular to the flow with 36 percent of the discs submerged. The discs
were rotated clockwise using a 1/4 horsepower Dayton gear reduced motor.
The discs rotational speed was 14 revolutions per minute yielding a periph-
eral velocity of 40.8 ft/min (12.4 m/min).
Description of the RBC System
3 3
The RBC was fed from a 1.01 ft (0.03 m ) holding basin which was kept
at a constant head using an overflow wier and two 4.0 ft3 (0.11 m^) storage
tanks. This system was incorporated into the study so that the RBC could
be fed using gravity rather than a pump which could possibly fail after
extended use.
The two storage tanks were used to hold the saline synthetic wastes
to be treated by the system. The synthetic sea water was produced using
Instant Ocean Sea Salts obtained from Aquarium Systems of Eastlake, Ohio.
The system was fed a balanced minimal media with a carbon source, sucrose,
serving as the growth limiting nutrient. The composition of the synthetic
feedstock is shown in Table 1. Compressed air was introduced into the two
storage tanks to insure adequate mixing.
Table 1. Composition of Svnthetic Feedstock
Constituent Concentration
SUCROSE 100 mg/£
S04 25 mgA
7H20 10 mg/Jl
6 mg/Jl
Mn SO^- H20 1 mg/£
CaCl2 0.76
FeCl2 • 6H20 0.05
After leaving the rotating biological contactor, the effluent flowed to
the drain of the Research Center. No recirculation was performed during
testing. The main advantages of recirculation are the delay of plant expansion
696
-------
along with the ability to supress shock loads, neither of which fall into the
scope of this study. With regard to nitrification, Lue-Hing et al., found'
that recirculation of clarified secondary effluent did not significantly
r\ f\ * *"•"'
increase nitrification. J For these reasons, recirculation was not included
in the study.
Tests Performedon the_RBC System
A series of tests were performed on the rotating biological contactor
system in both fresh and saline waters. The testing was more extensive on
the saline wastewaters. The major tests conducted were as follows:
Ammonia
Chemical Oxygen Demand
Specific Conductivity
Dissolved Oxygen
Nitrate
Nitrite
PH
Salinity
Temperature
All tests, with the exception of ammonia, were performed in accordance with
Standard Methods for the Examination of Water and Wastewater, 14thJEdition.^^
Ammonia concentrations were determined using two methods: (1) Orion
specific ion electrode model 95-10 for ammonia was used on the freshwater
samples, and (2) the Nesslerization Method 418B of Standard Methods was used
on the saline samples. The ammonia probe was not used in the saline samples
because membrane fouling caused unstable results. These tests were conducted
in order to monitor the performance of the rotating biological contactor
under various ammonia loadings in both freshwater and 10 percent sea water
envi ro nmen t s.
LABORATORY QUALITY CONTROL
A quality assurance program should be an important component of any
water quality study involving chemical or biological analysis. In order to
analyze the quality of data being collected, replicate and spiked samples
were run where applicable on a minimum of 15 percent of all samples.
Although this program of laboratory quality control does not entail any
quality control charts, the program was able to act as a satisfactory warning
system when any problems developed in sampling or analysis. The significant
problems encountered during the sampling and analysis segments of the study
will be discussed in the results.
The sampling program was similar for both the freshwater and 10 percent
sea water wastes. The two major exceptions being the method of determining
ammonia and the number of samples taken. More tests were performed on the
10 percent sea water wastes than were performed on freshwater. The fresh-
water data was obtained to verify that the contactor was able to achieve
adequate nitrification treating freshwater. This could be done without
extensive sampling. The saline wastewaters, being the primary interest of
697
-------
the study, have been monitored more closely in order to observe the various
stages of treatment. A discussion of the information obtained in both fresh
and saline wastewaters will follow in the section on results.
RESULTS
As previously discussed, the sampling program was divided into two main
segments: freshwater and saltwater. The freshwater segment was undertaken
in order to show that the rotating biological contactor could successfully
nitrify a freshwater waste, and to achieve loading rate-removal efficiency
relationships to compare to those of saline wastewaters. The 10 percent
sea water saline segment was undertaken to determine the loading rate— removal
efficiency relationships for a wastewater similar to those generated in a
Macrobraehium Roseribergii (prawn) rearing aquaculture facility.
Freshwater Results
The rotating biological contactor was first acclimated under freshwater
conditions using the synthetic feed material. This testing was conducted
to monitor the contactors* ability to nitrify wastewater under "nonsaline"
conditions. The results obtained during the freshwater testing may be seen
in Table 2.
Table 2. Freshwater Results
Influent
Ammonia
(mg/fc)
5.77
7.28
7.68
9.70
12.13
12.46
13,42
14.15
16.29
Effluent
Ammonia
(mg/£)
2.29
2.36
3.68
3.10
4.39
4,53
7.16
7.71
8.35
Hydraulic
Loading
(liter/100 mZ/day)
6,100
6,100
6,100
6,100
6,100
6,100
6,075
6,075
6,075
Ammonia
Loading
(g/100 m2/day)
30
40
50
60
70
80
90
100
110
Maximum
Ammonia
Removal
(%)
60
67
52
68
64
64
47
46
48
The removal efficiency of the contactor was found to be dependent upon
the applied ammonia load. A plot has been prepared of ammonia load versus
maximum ammonia removal. This plot may be seen as Figure 1, Under the
freshwater conditions, the optimal ammonia loading rate (from Figure 1) was
found to be approximately 60 grams/100 m^/day. The 60 gram/100 m^/day load-
ing rate resulted in a maximum ammonia removal efficiency of 68 percent,
yielding an effluent total ammonia (NH^+ + NH3) concentration of 3.1 rng/fc -
NH3 (2.6 mg/£ - N). The pH of the system's effluent was 7.1 at a tempera-
ture of 18°C. This pH and temperature, in combination with the total ammonia
concentration of 3.1 mg/& yields an unionized ammonia concentration of
0.013 mg/£ - NH3- This unionized value has been calculated using a percent
unionized ammonia of 0.430 based on work performed by Thruston et al., at
Montana State University.2^ As previously stated, this data demonstrates the
contactors ability to successfully nitrify a "nonsaline" wastewater.
698
-------
70'
I
LJJ
cc ^
< I
-------
In order to more precisely determine the optimum ammonia loading, the
model used by Weng and Molof in their nitrification studies will be used.
The model presented by Weng and Molof is as follows:
Log F = Log k + Log L0 + b Log Q + c Log S + d Log t 4- e Log A +
f Log d + g Log T
Where:
F = fraction of influent loading remaining in the effluent
k = the intercept value
L = influent loading
Q = flow rate
S = rotational disc speed
t = detention time of the liquid in the BFFRD system
A = effective disc surface area
D = submerged disc depth
T = liquid temperature
a, b, c, d,
e> f> 8> = the partial regression coefficients
Since all of the variables, with the exception of influent loading,
remained constant (assuming the variation of flow rate to be negligible),
the following constant may be determined:
Log k2 = b Log Q + c Log S + d Log t 4- e Log A + F Log D 4- g Log T
Combining the two equations yields:
Log F = Log k + Log k2 + a Log L
If:
Log k + Log k2 = Log k ,
then ,
Log F = Log k + a Log LQ
In order to use the equation with the data collected, Table 2 has been pre-
pared.
Table 2. Freshwater Ammonia Results
Ammonia Loading
L (grams/100 m2/day)
30
40
50
60
70
80
90
100
110
Log L
o
1.48
1.60
1.70
1.78
1.84
1.90
1.95
2.00
2.04
% Removal
60
67
52
68
64
64
47
46
48
F
.40
.33
.48
.32
.36
.36
.53
.54
.52
Log F
-0.40
-0.48
-0.32
-0.49
-0.44
-0.44 :
-0.28
-0.27
-0.28
700
-------
-0.6
-0.5
-0.4
LL
0
o
-0.3
-0,2
-0.1
14O 1.50
G R AMS/1OO nrr/ DAY
1.60 1.7O 1.8O . 1.90
LOG LQ
2.00
FIGURE 2. LOG F VS LOG Ln IN FRESHWATER
o
701
-------
A plot of Log F vs Log Lo may be seen as Figure 2. The plot shows two
distinct straight lines, the intersection of which represents the optimum
loading rate. In this case, the optimal ammonia loading rate for freshwater
was found to be 54 grams/100 m^/day which yielded an ammonia removal effi-
ciency of 73 percent. This determination of the optimal ammonia loading
represents a more precise method than plotting ammonia loading versus ammonia
removal to determine the optimal loading.
TenPercent Sea Water Results
After completion of freshwater data collection, the RBC system was
allowed to acclimate to a 10 percent sea water salt concentration. As
anticipated, the system's nitrification capacity was altered by the addition
of saltwater.
2
At an ammonia loading rate of 60 grams/100 m /day (optimum for the fresh-
water system from visual inspection), the maximum ammonia removal was 56 per-
cent in the 10 percent sea water waste. This removal efficiency is 13 percent
less than that of freshwater at the same loading rate.
Using the model described by Weng and Molof, the optimal ammonia loading
for the 10 percent sea water was found' to be 60 grams/100 m^/day. This load-
ing rate resulted in a maximum ammonia removal of 56 percent. The data used
to obtain these values may be seen in Tables 3 and 4 and has been presented
graphically in Figure 3.
Table 3. Saltwater Results
Influent
Ammonia
(ing/ 1)
4.5
7.5
9.5
10.4
14.8
16.2
Effluent
Ammonia
(mg/£)
3.1
4'. 6
4.3
4.6
8.4
9.2
Hydraulic
Loading
(liter/100 m2/day)
5,470
5,470
5,470
5,470
5,470
5,470
Ammonia
Loading
(g/100 m2/day)
25
40
50
60
80
90
Maximum
Ammonia
Removal
(%)
34
41
50
56
43
40
Comparing these results with those from freshwater testing, it can be
seen that although the optimal ammonia' loading remained essentially the same
the contactor's ability to nitrify decreased in the 10 percent sea water.
This decrease in nitrification in the 10 percent sea water coincides with the
conclusions of Ludzack and Noran that the nitrification will decrease with
increasing chloride concentrations.28
The ammonia loading rate was the only parameter that varied throughout
the monitoring program. The artificial substrate described in methodology
702
-------
- 0.6 .
-0.5 -
-0.4
LL
CD
O
60 GRAMS/100rrf/DAY
1.40 1.50 1.60 1.70 1.80 1.9O 2.00
LOG L
o
FIGURE 3. LOG F VS LOG Ln IN SALTWATER
o
703
-------
section yielded an influent chemical oxygen demand of approximately 115 mg/£,
The 10 percent sea water system had the same organic removal efficiency as
the freshwater system. Both systems achieved a 70 percent reduction in
chemical oxygen demand.
Table 4. Saltwater Ammonia Results
Ammonia Loading Log L
L (grams/100 m2/day) °
25
40
50
60
80
90
1,40
1.60
1.70
1.78
1.90
1.95
% Removal
34
41
50
56
43
40
F
0.66
0.59
0.50
0.44
0.57
0.60
Log F
-0.18
-0.23 .
-0.30
-0.36
-0.24
-0.22
The chemical oxygen demand was monitored at each stage of the contactor
during saltwater testing. The results of this monitoring may be seen in
Figure 4. These results show a rapid decline in the COD in the first stage,
approKimately 50 percent, followed by lower removal efficiencies in the
remaining three stages. Based on these results, it can be said that the
design of a rotating biological contactor for saline wastewaters is limited
by the ammonia load to the system.
The 10 percent sea water concentration yielded a salinity of 3.5°/°>
(grains/kilogram). This corresponds to a chloride concentration just under
2,000 mg/£ (1,920 mg/£). This chloride concentration is considerably lower
than others used in previous saltwater testing. The 10 percent sea water
salt concentration was used in this study because it is x^ithin the range of
salt concentrations required by MaoTobicac'hium Roseribergii as they develop
from the juvenile to adult stage.
The specific conductance of the 10 percent sea water increased slightly
as it passed through the contactor. The influent specific conductance was
5,500 umho/cm @ 2-0°C. It increased to 5,800 umho/cm @°20 C in the first stage
and remained constant through the remaining three stages. This increase of
300 ymhos/cm represents an increase of dissolved ionic matter of approximately
4 percent based on the assumption that dissolved ionic matter in mg/Jt is
equal to the specific conductance multiplied by an empirical factor of 0.8.
The dissolved oxygen and temperature were also monitored within the
contactor during saltwater testing. The influent dissolved oxygen was 0 mg/£
and increased to a concentration of 7.6 mg/£ in the effluent of the fourth
stage. This increase represents a reaeration capacity of 1.6 mg/A/hour. The
temperature of the wastewater decreased as it passed through the contactor.
The influent temperature ranged between 28° and 30°C (82° and 86°F) and
dropped to between 18° and 15°C (64° and 59°F) in the influent of the fourth
stage. The dissolved oxygen-temperature profile may be seen as Figure 5.
704
-------
CD
Q
Z
Ld
Q
Z
Ld
6
U
2
LU
X
u
100
80-.
60.-
40
20
INFLUENT
i n
STAGE
IT
N
FIGURE 4. CHEMICAL OXYGEN DEMAND
PROFILE
705
-------
en
E
LLJ
0
6
o
LU
>i
S5
c/2
o
30..
25.,
20.
15,.
\
TEMPERATURE
Ld
o:
Z)
s
o:
UJ
0.
UJ
10.
DISSOLVED OXYGEN
INFLUENT
STAGE
FIGURES. DISSOLVED OXYGEN -
TEMPERATURE PROFILES
706
-------
In summary, the results show that although the increased salinity did
not seem to affect the contactors'" ability to remove organic material
measured as the chemical oxygen demand. The increased salinity did, however,
alter the system's ability to nitrify ammonia. This decreased nitrification
will become the limiting parameter in the design of a rotating biological
contactor for use in an aquaculture facility.
CONCLUSIONS
The following conclusions can be made based on observations and analyses
of the data collected in this plot study:
(1) The nitrification of ammonia can be accomplished in a 10 percent
sea water waste,
(2) The ability of the rotating biological contactor to nitrify
decreased with increased salinity,
(3) The organic removal efficiency of the contactor measured as COD
was not affected by increased salinity (to-10 percent sea water).
(4) In designing a rotating biological contactor to nitrify a saline
wastewater, the ammonia loading will be the limiting design parameter.
(5) The reaeration rate of the contactor was satisfactory to maintain
a level of dissolved oxygen required within an aquaculture facility.
ACKNOWLEDGMENTS
Special thanks are extended to Anibal Alarcon for his insight during
the period of testing. This investigation was supported by the Department
of Civil Engineering, Environmental Engineering Program, Colorado State
University, Fort Collins, Colorado and Klemetson Engineering, Fort Collins,
Colorado.
REFERENCES
(1) Poon, C. P. C. and Mikucki, W. J., "Rotating Biological Contactors
Treat Island's Saline Sewage," Water andSewage Works, pp. 62-66,
February (1978).
(2) Poon, C. P. C., Chao, Y., and Mikucki, W, J., "Factors Controlling
Rotating Biological Contactor Performance," Journal Water Pollution
Control Federation, Volume 51, No. 3., pp. 601-611, March (1979).
(3) Labella, S. A., Thaker, I. H., and Tehan, J. E., "Treatment of
Winery Wastes by Aerated Lagoon, Activated Sludge, and Rotating
Biological Contactor," Proceedings: 27th Industrial Waste Conference,
Purdue University Extension Service 141, p. 803, (1972),
(4) Cook, E. E. and Kincannon, D. F., "An Evaluation of Trickling Filter
Performance," Water...and Sewage Works, Volume 118, No. 4, pp. 90-95,
April (1971).
707
-------
(5) Cook, E. E, and Kincannon, D, F., "An Evaluation of Trickling Filter
Performance," Water and Sewage Works;, Volume 118, No. 4, pp. 90-95,
April (1971),
(6) Poon, C. P. C. and Mikucki, W. J., "Rotating Biological Contactors
Treat Island's Saline Sewage," Water and Sewage Works, pp. 62-66,
February (1978).
(7) United States Environmental Protection Agency, Quality Criteria for
Water. Washington, D.C., July (1976).
(8) United States Environmental Protection Agency, Quality Criteria for
Water, Washington, D.C., July (1976),
(9) United States Environmental Protection Agency, Quality-Criteria for
Water. Washington, D.C., July (1976).
(10) American Fisheries Society, A Review of the EPA Red Book: Quality
Criteria for Water, Water Quality Section, American Fisheries Soceity,
Bethesda, Maryland, pp. 6-11, April (1979).
(11) Metcalf and Eddy, Inc., Wastewater Engineering: Treatment Disposal
Reuse, Second Edition, Revised by George Tchabanoglous, McGraw-Hill
Book Company, New York, New York, (1979).
(12) Antonie, R. L., Kluge, D. L,, and Mielke, J. H., "Evaluation of a
Rotating Disk Wastewater Treatment Plant," Journal Water Pollution
Control Federation, Volume 46, No. 3, pp. 498-511, March (1974).
(13) Weng, C. and Molof, A. H., "Nitrification in the Biological Fixed-
Film Rotating Disk System," Journal Water Pollution Control Federation,
Volume 46, No. 7, pp. 1674-1685, July (1974).
(14) Ludzack, F. J. and Noran, D, K., "Tolerance of High Salinities by
Conventional Wastewater Treatment Processes," Journal Water Pollution
Control Federation, Volume 37, No. 10, pp. 1404-1416, October (1965).
(15) Kessick, M. A. and Manchen, K. L., "Salt Water Domestic Waste Treatment,"
Journal Water Pollution Control Federation. Volume 48, No. 9, pp. 2131-
2136, September (1976).
(16) Poon, C. P. C., Chao, Y., and Mikucki, W. J., "Factors Controlling
Rotating Biological Contactor Performance," Journal Water Pollution
Control Federation, Volume 51, No. 3., pp. 601-611, March (1979).
(17) Kessick, M. A. and Manchen, K. L,, "Saltwater Domestic Waste Treatment,"
Journal Water Pollution Control Federation, Volume 48, No. 9, pp. 2131-
2136, September (1976).
(18) Kessick, M. A. and Manchen, K. L., "Saltwater Domestic Waste Treatment,"
Journal Water Pollution Control Federation, Volume 48, No. 9, pp. 2131-
2136, September (1976).
(19) Ludzack, F. J. and Noran, D. K. , "Tolerance of High Salinities by Conven-
tional Wastewater Treatment Processes," journal Water Pollution Control
Federation, Volume 37, No. 10, pp. 1404-1416, October (1965).
708
-------
(20) Poon, C. P. C. and Mikucki, W. J., "Rotating Biological Contactors
Treat Island's Saline Sewage," Wate_r and Sewage Works, pp. 62-66,
February (1978).
(21) United States Environmental Protection Agency, Quality Criteria for
Water, Washington, B.C., July (1976).
(22) Poon, C. P. C. and Mikucki, M. J., "Rotating Biological Contactors Treat
Island's Saline Sewage," Water and Sewage Works, pp. 62-66, February,
(.1978).
(23) Lue-Hing, C., Obayashi, A. W., Zenz, D. R. , Washington, B,, and Sawyer,
B. M., "Biological Nitrification of Sludge Supernatant by Rotating
Disks," Journal Water Pollution Control Federation, Volume 48, No. 1,
pp. 25-40, January (1976),
(24) APHA, AWWA, WPCF, Standard _Me_tb.ods for the Examination of Water and Waste-
water,_Fourteent h_ Ed it i on» American Public Health Association, Washington,
D.C. (1976).
(25) United States Environmental Protection Agency, Quality Criteria for
Water, Washington, D.C. July (1976),
(26) Thurston, R. V., Russo, R. C., and Emerson, K., Aqueous Ammonia
EcLuilibrium_Tabulation of Percent Unionized Ammonia, United
States Environmental Protection Agency, Duluth, Minnesota, 428 pages,
August (1979).
(27) Weng, C. and Molof, A. H,, "Nitrification in the Biological Fixed-Film
Rotating Disk System," Journal Water Pollution Control Federation,
Volume 46, No. 7, pp. 1674-1685, July (1974).
(28) Ludzack, F. J. and Noran, D. K., "Tolerance of High Salinities by Con-
ventional Wastewater Treatment Processes," Journal Water Pollution
Control Federation, Volume 37, No. 10, pp. 1404-1416, October (1965).
(29) APHA, AWWA, WPCF, Standard Methods for the Examination of Wate_rand
Wastewater, Fourteenth Edition, American Public Health Association,
Washington, D.C. (1976).
709
-------
Page Intentionally Blank
-------
RBC FOR MUNITIONS WASTEWATER TREATMENT
By
P. Gail Chesler
Physical Scientist
Gerald R. Eskelund
Branch Chief, Environmental Technology Branch
US Army Mobility Equipment Research & Development Command
Fort Belvoir, Virginia
Introduction
The United States Army operates one of the largest industrial complexes
in the United States. Just as other large chemical industries, the Army is
subject to all Federal laws, including EPA regulations in the form of National
Pollution Discharge Elimination System (NPDES) allowances. A part of the
Army's chemical production is a group of products known as explosives. Since
most of the manufacturing operations involved with this group of products are
unique to the military industrial complex, it has been necessary for the Army
to pursue a vigorous pollution abatement program, often being on the fore-
front of technology development.
This study was instituted to determine if a rotating biological
contactor (RBC) could be used to treat wastes from a proposed new RDX-HMX
manufacturing facility. Several problems were posed for this study.
Initially, what would be the composition of the waste stream from the proposed
facility? Secondly, what types of treatment would be effective? Finally,
would there be any residual problems not handled by proper sequence of treatment?
711
-------
The question of chemical composition of the waste stream was handled by
Armament Research and Development Command (ARRADCOM), Dover, NJ. From pilot
plant studies at Holston Army Ammunition Plant (HAAP) and the plans and
material balances for the proposed X-Facility, two ranges of waste streams
were predicted. The two ranges, designated A stream and B stream, differ
only from dilution by condensate from a cooling tower. The predicted chemical
compositions of the two waste streams are shown in Table 1. Table 2 gives
the actual values chosen for use in the bench scale study. One concern was
that the toxic concentrations of the constituents of the waste stream might
not allow biological degradation to take place. Results of preliminary test-
ing showed BOD removal taking place Indicating biological degradation. A
result which had impact on further testing was the wide disparity between
anticipated and actual values of BOD and COD. Projected values were roughly
1/6 the initial readings of these two parameters. The stream was unique in
composition and clearly the need existed to determine the most effective
treatment system.
In determining the type of treatment, current technology was examined.
Work had been conducted previously at Holston Army Ammunition Plant on wastes
similar to those proposed. It was determined at that time that a biological
system could be the most effective treatment for their operation, but the
system chosen was not a RBC. Radford Army Ammunition Plant had also investi-
gated a biological system and had used a rotating biological contactor for its
wastes, but these wastes were not similar to those expected at the proposed
facility. The designer of the new manufacturing facility reviewed both sets
of data and proposed a treatment system which contained, as a key element,
a RBC. The proposed treatment scheme is shown in Figure 1. To establish
the validity of the treatment scheme, it was decided to perform bench scale
studies at Mobility Equipment Research and Development Command (HERADCOM) and
pilot scale studies at Atlantic Research Corporation. The authors of this
paper were involved with the bench scale testing, though close proximity of
the two organizations allowed good communication. It was determined that a
complete treatment train at the bench scale level would be set up. Emphasis
would be given to the biodisc component since this would contribute the
major portion of the removal. Due to a number of system operational problems,
it was quickly discovered that operation of more than the aerobic biodisc
would prove fruitless, so work was limited to a detailed study of the aerobic
bench scale biodisc. The immediate experimental goal was the determination of
the optimum flow rate of formulated wastewater to the biodisc to achieve
maximum BOD removal in conditions A and B.
A critical concern in the system was the explosive components because
of possible residual toxicological and mutagenicity problems. For the past
several years, the Surgeon General has had an extensive toxicology study under-
way for TNT, RDX, and HMX among other explosives. These materials have been
found to be toxic in varying levels, depending on the conditions to which the
explosives have been subjected. TNT, for instance, is photolyzed by sunlight
and its composition is altered if the solution is basic. The RDX and HMX do
not photolyze but do undergo some hydrolysis reactions. In both cases, the
toxic properties are altered but not eliminated. The RDX and HMX went through
the system almost unchanged, while the TNT degraded before treatment by the
disc. From the data, it is clear that treatment, other than biodisc, of the
explosives in the contaminated wastewater will be necessary. Carbon studies
712
-------
have been done and if a properly designed carbon system is employed, tests
have established that the explosive components can be removed to acceptable
levels.
EXPERIMENTAL DESIGN AND OPERATION
As previously mentioned, the original flow scheme is illustrated in
Figure 1 and constituents of the hypothetical waste stream are set out in
Table 2. Initial mixing of the waste stream took place in a covered 1100
litre tank on a weekly basis. PH adjustment was made using ammonium
hydroxide on a batch basis to raise the pH to approximately 7.4. The waste-
water was pumped into a smaller feed tank, 120 litres, for controlled flow
into the biodisc system. Continuous additional pH adjustment was made in
this tank by an automatic control system. Uniform mixing in both tanks was
accomplished by use of submerged pumps.
From this feed tank, the flow went into the aerobic biodisc under control
of a Masterflex pump. A picture of the aerobic biodisc unit appears as
Figure 2. Effluent from the aerobic biodisc flowed into the anerobic biodisc
unit which was covered and airtight. Beyond the anaerobic unit, the flow was
pumped into an aeration chamber where air was bubbled through the effluent.
This process enhanced settling of sludge which consisted primarily of biomass.
A clarifier followed in the flow pattern, and effluent was pumped next into
a multi-media column and on through a carbon column. The liquid was then dis-
charged into the drain.
A number of alterations in the physical set up took place over the course
of the experiment. The first was removal of the multimedia column and carbon
column due to growth of microorganisms which appeared similar to those present
on the discs. Backwashing the columns was not effective in removing these
microorganisms, particularly in the multimedia column, as they clumped together,
forming flake-like particles which clogged the column and caused substantial
back pressure.
The second alteration in the system was the removal of the anaerobic unit.
At a point early in the testing, it was necessary to dewater the entire system
and it was noted that virtually no growth had occurred on the discs rotating
in the anerobic unit. This fact had been anticipated due to the lack of gas
evolution from the unit. Possible explanations include the relative delicacy
of anerobic organisms and the fairly wide fluctuations in pH and flow rate to
which the system was subjected at start up.
The third alteration was in feed tank size. After two months of opera-
tion, it became apparent that use of the 1100 litre tank for feed mixing was
not an experimentally sound procedure. Extensive biological growth had taken
place there, and the symptom of that growth was consumption of COD by these
organisms. That is, COD was significantly higher for the tank mixture immedi-
ately after mixing than it was later in the week. For this reason, mixing
was done in the smaller 120 litre feed tank on a more frequent basis, in the
hope that the shorter retention time would inhibit growth. As of this writing,
four additional months into the testing program, it appears that growth in
the feed tank, is again a problem, calling for further modification of the
feed flow system.
713
-------
At this point, then, the system has been pared down to functioning
units as illustrated in Figure 3.
It is noteworthy that very early in the testing program, growth on
the aerobic disc unit was limited, as was COD removal. It was necessary
to balance the ingredients required for good biological growth by the
addition of two chemicals. Addition first of nitrogen, as ammonium
hydroxide, and then of phosphorous, as sodium phosphate, as supplemental
nutrients proved to be the necessary changes, and substantial growth
appeared within a week; COD removal efficiency increased dramatically as
well. Amounts of nitrogen and phosphorus to be added were calculated
using a molar ratio of carbon:nitrpgen:phosphorus::106:16:1 which is the
approximate ratio at which the microorganisms are thought to synthesize
those elements.
At start up, projected pH levels were in the range 5.5-7.3; by contrast,
it was found that the actual pH of the mixture in the feed tank was approxi-
mately 3. It was necessary to add significant amounts of ammonium hydroxide
during the use of each batch. This was above and beyond the initial addition
and must be considered in the actual plant operational costs.
In bench scale testing, it was important that the three explosives be
completely dissolved to be sure that they were actually included in the feed
stream to the biodisc. In order to reduce times and facilitate complete mix-
ing, it was necessary to enhance the solubility of the RDX and HMX by the
addition of more cyclohexanone than was called for in the hypothetical waste
stream. Sufficient acetone was present in the formula to insure the solubility
of the TNT. For reasons to be discussed later, RDX, HMX and TNT were deleted
from the feed stream for the second half of the experimentation.
Once the waste stream composition was fixed, analytical methods were used
to follow its progress through the treatment process. Analyses of the waste-
water were performed using techniques described in Standard Methods, 14th
edition. The specific parameters which were determined and the procedures used
are shown in Table 3. Analysis of the microbiological growth was done at
Natick Research and Development Command, Natick, Massachusetts, using culture
techniques and visible identification. The Ames test was run by Atlantic
Research Corporation with the five tester strains of Salmonella typhimurium.
RESULTS AND DISCUSSION
Figure 4 shows a curve which represents the COD and BOD removal efficien-
cies from start up through attainment of optimum removal. The initial goal of
95% 6005 removal efficiency has been shown to be attainable consistently and
repeatably. In spite of the high levels of formaldehyde, it is safe to con-
clude that the microorganisms were able to survive and to degrade the organics
in the waste stream effectively.
For A stream with BODg near 1600 mg/L, optimum removal efficiency was
found to occur at a loading rate of 3.3 Ib BOD,- per day per square foot of
disc surface area. Experimenters on both bench and pilot-scale levels were
714
-------
reassured that very similar removal efficiencies were attainable with both
systems when comparisons were made using loading rates per square foot of
disc surface area. This allowed confidence in scale-up factors from bench
to pilot-scale and from pilot to plant-scale for design purposes,
^
Data generated at the bench-scale level was the result of feeding the
biodisc unit on a 4-stage basis, that is, flow from compartment one went to
compartment two, and so forth. When COD analysis was done on a stage by stage
basis, it became clear that a substantial majority of the removal (around 70%
of the total) took place in stage one. Atlantic Research modified the pilot
plant scheme to have influent flow directly into stages one and two. In so
doing, they simulated three stage operation and were able to increase COD
removal.
It is interesting that pH measurements in the chambers of the biodisc
unit gave indication of COD removal efficiencies. If the pH dropped precipi-
tously in the stage 2 chamber, COD removal could be expected to be less than
adequate. Investigation of this situation could yield an on-line monitoring
technique.
It became apparent as the testing progressed that the RDX, HMX and TNT
were not being processed effectively by the system. RDX and HMX were found
in identical concentrations in the influent and effluent and TNT was trans-
formed as previously mentioned. In both cases, it is clear that treatment
other than by biodisc will be necessary and it is likely that TNT will not be
part of the ultimate influent to the biodisc. For these reasons, the decision
was made to leave the explosives out of the stock mix for the second half of
the experimentation. It should be noted that though the TNT was transformed,
it is not reasonable to conclude that the contamination problem had been
dealt with. It is known that TNT is readily transformed into compounds as
toxic as TNT. It has been found subsequently that the mix without the explo-
sives is not as toxic. Clearly, disposal of the explosives is a problem yet
to be dealt with. Treatment by carbon will probably be used, but the
designers must allow for the fact that biological growth can hinder the
operation.
In the course of the investigation, it appeared that the microorganisms
were not as differentiated as is often found in sewage treatment plants using
RBC's. Evidence was sought that the strain found was not so pure in culture
as to be susceptible to total kill should some toxic agent enter the system.
The test results showed that this was not a pure culture, and in fact, con-
sisted of two strains of fungi and seven different colonial morphologies. Two
fungi identified were Fusarium sp. and Geotrichum sp. Three pseudomonads were
isolated, one from the pseudomonas genus and two pseudomonad organisms. Two
common bacillus organisms of a ubiquitous nature rounded out the lot of mi-
crobes found. No further analysis was done to classify the organisms once it
was apparent that they were typical of normal sewage system organisms, and
obtaining seed material would not be difficult should a massive kill take
place. Roughly 10-14 days were required to go from clean disc start-up to
optimum removal, though clean disc start-up is not a likely occurrence due to
the hardiness of the microbial population in the face of adversity.
715
-------
CONCLUSIONS
As shown previously, the biodisc is effective in reducing the BOD and COD
significantly. This is true even if there are explosives present which are
known to be toxic. Ames tests run on biodisc effluent show that the explosives
produce a mutagenic effect. When explosives are not included in the waste
stream, the Ames test results are negative. These results were anticipated
from previous work done by the Surgeon General's office. The most toxic of
the explosives is TNT and its conversion products. This leaves two choices,
either pretreat or posttreat to remove the explosives. Both solutions may
be feasible but must be evaluated from an economic and a safety standpoint.
Further investigation is needed concerning these toxic by-products and their
removal.
One production line in the X-Facility had been originally designed with
supplemental pollution abatement equipment to deal with high levels of nitrogen
in the effluent. Since the bio-system needs nitrogen, that separate pollution
abatement process can perhaps be eliminated; this would provide a substantial
cost savings for the plant. Before the process can be changed it will be
necessary to determine if the nitrogen is in a usable form for the biosystem
and whether that stream contains any potential toxicants.
The small bench scale model has proven to be linear in scale-up when
considering pounds of BOD per square foot of disc surface area. This means
that small bench scale systems can be used to perform tests which should be
valid for full scale systems. Many variations and conditions can be investi-
gated such as changes in food sources and flow rates, potential plant chemical
surges, and other anticipated problems. The effect on the organisms as well
as on the efficiency of removal can be determined. The system is cost
effective. A small scale unit takes less chemicals, requires less power to
run, and can be conditioned more rapidly to changing parameters. Therefore,
the bench scale unit is a mini-ecosystem which provides fast, cost effective,
and reliable data for the investigation of large scale operations.
716
-------
TABLE 1
PREDICTED QUALITY OF INFLUENT TO INDUSTRIAL
WASTEWATER TREATMENT PLANT - X FACILITY
CONDITION A
CONDITION B
Total Flow, Gallons/Day
Contaminants:
N03-N02
Ammonia
RDX
HMX
TNT
Acetic Acid
Hexamine
Cyclohexanone
Propyl Alcohol
Methyl Acetate
Propyl Acetate
Formic Acid
Nitromethane
Formaldehyde
Phosphate
Sulphate
Acetic Anhydride
Amine
Organic Nitrogen
Toluene
Stearic Acid
Acetone
5539,8QQ
IB/DAY
M6/L
993,600
LB/DAY
MG/L
230
19-46
60-147
20
64-155
420-1052
464-576
518-648
653-816
250-312
77-96
2246-2808
250-312
6912-8640
66
1102
400
60
69
38-48
12-24
566-696
18
2-4
5-11
2
5-12
33-82
36-45
40-51
51-64
20-24
6-7
175-219
20-24
539-674
5
86
37
5
5
3-4
1-2
43-54
225
19-46
60-147
20
64-155
420-1052
464-576
518-648
653-816
250-312
77-96
2246-2808
250-312
6912-8640
56
829
400
60
69
38-48
12-24
556-696
27
2-6
7-18
2
8-19
51-127
56-70
63-78
80-99
30-38
9-12
272-339
30-38
836-1045
7
100
48
7
8
5-6
1-3
67-84
Condition A: Total wastewater includes heat exchanger condensate
Condition B: Total wastewater without heat exchanger condensate
717
-------
CHEMICAL
TABLE 2
EXPERIMENTAL PARAMETERS
FOR CONDITIONS A AND B
CONDITION A
(MG/L)
INCLUDES HEAT
EXCHANGER CONDENSATE
CONDITION B
(MG/L)
INCLUDES HEAT
EXCHANGER CONDENSATE
Formal dehyde
Formic Acid
Sulfate
Acetic Acid
1-Propanol
Acetone
Cyclohexanone
Hexamine
Methyl Acetate
Nitromethane
n-Propyl Acetate
Phosphate
Toluene
Amines
Stearic Acid
TNT
RDX and HMX
COD
BOD
pH
674
219
86
85
64
54
51
44
24
24
7
5
4
5
2
12
13
1650
1390
3
1045
339
100
184
99
84
78
70
38
38
12
7
6
7
3
19
20
2300
1660
3
718
-------
PARAMETER!
TABLE 3
LIST OF PARAMETERS TO BE MEASURED
AND ANALYTIC METHODS OR PROCEDURES
METHOD
Biochemical Oxygen
Demand-
per Standard Methods
Chemical Oxygen
Demand
Dichromate Reflux method
(Standard Methods , p. 550),
Total Organic Carbon
Dohrmann TOG Analyzer
(Standard Methods1, p. 532)
pH
Beckman pH meter
Temperature
Thermomete x
Ammonia, Nitrites,
Nitrates
Hach tests
TNT, RDX, HMX
Waters Liquid Chromatograph
American Public Health Association, jStandard Methods for the Examina-
tion of Water and Wastewater, 14th edition, APHA, Washington, D.C.
(1975).
719
-------
EFFLUENT
FROM PLANT
EQUALIZATION
POND
AERATION (?)
ROTATING
BIOLOGICAL
CONTACTORS
pH CONTROL
-vj
N3.
O
DISCHARGE
ACTIVATED
CARBON
CLARIFIER
CHLORINE^?)
DUAITMEDIA
FILTRATION
FIGURE 1
-------
721
-------
r
AEROBtC.
BlOD'.SC
SCHEMATIC PHYSICAL LAYOUT
86MCH 5CAL£ MODEL-
722
-------
o
r-
o
VD
-------
Page Intentionally Blank
-------
REMOVAL OF WASTE PETROLEUM DERIVED POLYNUCLEAR AROMATIC
HYDROCARBONS BY ROTATING BIOLOGICAL DISCS
By
John T. Tanacredi*
Institute of Health Sciences
Department of Environmental Health Science
Hunter College, C.U.N.Y., N.Y. 10029
Abstract
A staged, partially submerged rotating biological disc system was
assessed to determine its performance in the reduction of concentrations of
polynuclear (PNA) aromatic hydrocarbons attributable to waste crankcase oils
(WCCO) in wastewater effluents. Removal of petroleum derived PNA hydrocarbons
is important because of their known toxicity and carcinogenicity which pose
potential public health risks. Samples of the influent, several successive
stages of treatment by the disc system and the final effluent were collected,
extracted with carbon tetrachloride (CCL4) and analyzed. Two UV-fluorescence
spectroscopic techniques provide qualitative evidence of the presence of WCCO
hydrocarbons. An IR-quantification method was utilized to determine the total
extractable organics at each stage. The UV-fluorescence techniques rely upon
the ability of aromatic hydrocarbons from WCCO to generate a "fluorescence
profile" differentiable from other petroleum entities. Excitation of ex-
tracts at a specific wavelength produces WCCO fluorescence emission profiles
which can be compared to known "standard oils." These fluorescence methods
were successfully employed by us to identify weathered petroleum products
AThe analyses for this investigation were performed under the auspices
of the analytical research facility of the U.S. Environmental Protection
Agency, Industrial and Environmental Research Laboratory, Oil and Hazardous
Spills Branch, Edison, New Jersey, through an approved program that supports
graduate level research study of the environment.
725
-------
using exposed stock oils, and therefore they were used in this project for
the detection of a gradient hydrocarbon response to the disc treatment
system. Fluorescence maxima profiles (FKP's) showed a significant decrease
in detectable PNA's attributable to WCCO through the successive stages of
the disc system. Preliminary statistical analyses of IR quantification
data revealed a direct relationship between the degree of treatment and the
level of PNA concentration. These results indicate that such biological
systems for the removal of WCCO aromatic hydrocarbons are viable alter-
natives to secondary treatment systems commonly being employed. Varying
the flows, loadings and recycling may further improve removal efficiencies.
Additional studies are warranted in light of the possible future need to
reuse wastewaters.
Introduction
Considerable attention has been directed toward the use of rotating
biological contactors (RBC's), and they have been shown to be an effective
means of treating wastewaters.1 Previous investigations^ into the polynuclear
aromatic hydrocarbon (PNH) character of wastewaters in New York City Water
Pollution Control Facilities (WPCF), attributable to waste crankcase oils
(WCCO) have exhibited a range of PNH's. Recent investigations-* have also been
directed at such compounds attributable to automotive WCCO not only because
several constituents of these compounds (i.e. napthalenes) have been shown
to "bioactivate" compounds into mutagens, but also because of energy con-
servation needs.5 The purpose of this preliminary work was by use of UV-
fluoreseence techniques, to qualitatively determine the performance of a
pilot RBC system in the reduction of initial concentrations of detectable
PNH's attributable to WCCO.
EQUIPMENT
The pilot plant at Newtown Creek is the same system operated by W.
Torpey at the Jamaica WPCF during the period July to November 1969, and was
comprised of three main component parts: (a) Ten stages of rotating disks
(all discs were 3' in diameter) for the removal of organics and for ox-
idation of ammonia to nitrate, (b) six stages of illuminated rotating discs
for the removal of nitrogen and phosphorus from the effluent of the pre-
ceding system by synthesis into attached algal cells, and (c) six packed
beds of granular activated carbon columns for the adsorption of refractory
organics from the preceding algal system. Sedimentation of 1.5 hours was
interposed between the effluent from the ten stage unit and the algal unit.
A mixed media filter preceded adsorption in carbon columns for removing the
particulates, which were mainly algal cells generated on the illuminated
disks.
The flow through the ten stages was 28.39 1/min. "t 10%, where as the
algal unit rate was 11.35 1/min. Flow to the pressure downflow carbon
columns was at a surface loading of 203.6 1/min/nf. Disks rotated opposite
to the direction of a flow through the successive stages. The theoretical
detention time was six minutes in each stage, measured when the disks were
devoid of slime. Actual time was somewhat less, depending upon the degree
of displacement of fluid volumes by the slime. Addition information of the
RBC system operating results can be obtained from Torpey, W., et. al. (1973).
726
-------
ANALYTICAL APPROACH
Because of the complex chemistry of petroleum, each petroleum sample
lends itself to differentiation from others. This passive-tagging approach
establishes specific qualitative parameters for oil samples in the form of
"profiles''or "fingerprints" to be compared to a "reference standard profile."
Thus, positive correlations for RBC pilot plant effluent samples are either
established or not established with reference standards depending upon those
portions of the petrochemical waste that exhibit themselves in fingerprints
and remain stable under physiochemical processes and environmental con-
ditions.
METHODOLOGY
Water samples were collected at successive stages along the system in
980 ml. wide-mouth, glass Mason jars with Teflon-lined caps. Samples were
taken from the raw influent, stages 1, 3, 6, 8, 10, settling tank, 13, 16
and the final effluent. Each sample was adjusted for pH 3 and refrigerated
throughout storage until analysis. Samples were extracted with 50 ml CCL4
in seperatory funnels and the bottom layer collected. Solvent was stripped
off and residue weight recorded. An infra-red (IR) quantification method
was used to determine total extractable hydrocarbons (mg/1) from each
sample.7 CCL4 extracts were jet-air evaporated, concentrated, and residuas
weighed and brought to volume in hexanes for UV fluorescence analysis.
Previous investigators have exhibited the ability of fluorescence
spectroscopy to detect trace quantities of petroleum derived hydrocarbons
in oceanic waters. Investigators^ have been able to differentiate between
a lubricating oil and a crude or fuel oil using fluorescence spectroscopic
techniques. All petroleum products fluoresce when excited by UV light be-
cause of the presence of aromatic hydrocarbons with multi-ring configurations
such as fused ring polynuclear aromatics.10 A UV-fluorescence spectrophotometer
with two independent monochromaters (150 watt xenon are light source),
and a constant temperature cell bath maintained a 10mm path length quartz
cell at 20° JT 0.5°C, was used for all fluorescence analyses. A synchronous
excitation fluorescence spectroscopic technique was utilized for all analyses.
A standard reference WCCO was excited at 290 nm while scanning the emission
spectrum from 240 to 540 nm, generating a maxima emission profile (MEP) for
that excitation wavelength. This MEP was then used to correlate presence
of WCCO in successive stages along the RBC system. In addition, each sampled
stage was excited at successive excitation wavelengths from 240 nm to 440 nm
(at 20 nm-intervals) while scanning for the maximum fluorescence emission at
that excitation frequency. (Figure 1) Each maximum peak was utilized as
a point to be plotted graphically, generating a "fluorescence maxima profile"
(FMP) for each sample. Correlation was determined by visual comparison of
maxima profile plots of the WCCO reference standard, to RBC stage maxima
profile plots. (Figure 2).
727
-------
70-
60-
50-
40-
30-
20-
W
u
z
O
s
u
320 34O
~380
460 420
2k)
^80 300 320 34O 360
EXCITATION FREQUENCY
Figure 1.Fluoresence maxima profile (FMP) for WCCO.
It should be noted that the correlation criteria utilized for these analyses
is essentially qualitative, in that source identification, (such as gas station
or individuals dumping WCCO into sewers) of the detected waste petrochemical
cannot be directly established by the technique. The Newtown Creek Sewage
Treatment Facility (STF) however, is in an area of high petroleum hydrocarbon
load.
There are several refineries and industrial facilities surrounding the
treatment plant. Investigation by Mueller, J.A., et. al., 12 revealed the
Newtown Creek STF to be the principal discharger of oils and grease in treated
wastewaters to the New York Bight (12.6 metric tons/day).
728
-------
80-,
70-
60-
50-
30-
30-
20-
10-
*-WCCO
•6- -raw influent (Sample Sensitivity-0.3)
• -Sample stage"(SS-1.0) *•
o -Sample stage w(SS-t.O)
#-Sample stage »-(SS~1.0)
• -Sample stage*
-------
TABLE 1
TOTAL EXTRACTABLE ORGANICS FROM RBC SYSTEM
Sample
Fluorescence
Sample
Sensitivity
Correlation with
FMP of WCCO
MG/1 (TEO)
Raw
1
3
6
10
Settling Tank
13
16
Final
.3
.3
.3
.3
.3
1.0
1.0
3.0
3.0
(+) 16.7
(+) 10.2
(+) 9.9
(+) 7.9
(+) 4.4
(+) 2.6
(+) 4.8
C-) 2.6
(-) 0.12
variations in operational modes so as to clearly establish whether such
petroleum derived Hydrocarbon loadings are consistently and effectively removed
by RBC systems.
DISCUSSION AND CONCLUSION
A number of PNH's are potent carcinogens in animals and man. WCCO has
been shown, along with other petroleum products to contribute significant
quantities of detectable PNH's to aquatic environments. Traditional wastewater
treatment facilities in major urban areas have been shown to be relatively
ineffective in elimination, or providing a significant reduction of these com-
pounds. RBC systems have on the other hand, been successfully employed in
treatment of wastewaters, * * anri i-iairo a-yhihu-eA h«i-A sm-nr i si no-l v
effective removal of PNA"
and have exhibited here surprisingly
s over suspended cultures.
With more investigation into variations in such operational parameters
as detention times, slime build-up, influent loadings, effective disk surface
area, submerged disc depth, wastewater flow rates and temperature, greater in-
sight into PNH removal efficiencies may be provided.
This preliminary investigation appears to provide some credence to RBC
systems as an effective means to treat sewage, and to prepare water for future
re-use.
730
-------
ACKNOWLEDGEMENTS
Special appreciation to W. Torpey for permission to sample the pilot
RBC system at Newtown Creek, to M, Alavanja for the preliminary statistical
analysis, and to M. Gruenfeld, U. Frank and Dr. G. Kupchik for their comments
and support of this project. Discussion of results was provided by W.Torpey
REFERENCES
1. Torpey, W. et. al. (1971) Rotating Disks with Biological Growths Prepare
Mastewater for Disposal or Reuse, Jour. Water Poll. Control _Fed. > 43
(2181).
2. Tanacredi, J.T. (1977) Petroleum Hydrocarbons from Effluents: Detection
in Marine Environment. Jour. Water Poll. ControlFed., March: 216.
3. Sieger, T.L. and Tanacredi, J.T. (1979) Contribution of Polynuclear
Aromatic Hydrocarbons to Jamaica Bay Ecosystem Attributable to
Municipal Wastewater Effluents in proceedings 2nd Conference
NPS/AIBS Research in the Natural Parks, San Francisco, CA., 26-30
Nov. 1979.
4. Payne, J.F. and Martins, I. (1978) Crankcase Oils: Are They a Major
Mutagenic Burden in the Aquatic Environment? Sc1ence, 200: 329.
5. Tanacredi, J.T. (1979) "Waste Oil Recycling: Are There Any Incentives
Out There?" Unpublished Manuscript (Hunter College, Sept. of Env.
Health Sciences, CUNY).
6. Torpey, W. , et. al. (1973) "Effects of Exposing Slimes on Rotating Discs
to Atmospheres Enriched with Oxygen", in: Advances in Water Pollution
Research, Sixth Annual Conference held in Jerusalem, 8-23 June, 1972.
7. Gruenfeld, M. (1973) Extraction of Dispersed Oils from Water for
Quantitative Analysis by Infrared Spectrophotometry. Env. Sci &
Tech., vol. 7, No. 7, July.
8. Keizer, R.D. and Gordon, D.C. (1973) Detection of Trace Amounts of Oil
In Sea Water by Fluorescence Spectroscopy, Jour. Fish. Res. Brd. of
Canada, 30, 1039.
9. Goldberg, M.C. and Devonald, D.H. Ill, (1973) Fluorescent Spectroscopy -
A Technique for Characterizing Surface Films Jour. Res. U.S. Geol.
Survey, 1, 714.
10. Reicher, R.E. (1962) Amer. As&oc. JPetrol. Geol. Bull. 46, 60.
11. Frank, U., and Gruenfeld, M. (1978) Use of Synchronous Excitation
Fluorescence Spectroscopy for In Situ Quantifications of Hazardous
Materials in Water; in: proceedings 1978 Nat. Conference on Control
of Hazardous Materials Spills, Miami Beach, API/USCG/USEPA.
731
-------
12. Mueller, J.A.. et. al. (1976) "Contaminant inputs to the New York Bight,"
NOAA Technical Memorandum ERL—MESA—6, Marine Ecosystems Analysis
Program Office, Boulder, Colo.
13. Cecil Lue-Hing, et. al., (1976) "Biological Nitrification of Sludge
Supernatant by Rotating Discs," Jour. Water Poll.Cont. Fed.,
48(1) P- 25-46.
14. Davies, T.R. and W.A. Pretorius, (1975) "Denitrification with a
Bacterial Disc Unit", Water Research 9(4), 459-463.
15. Weng, C. and Molof, A. (1974) "Nitrification in the biological fixed-
film rotating disk system" Jour. Water Poll. Cont. Fed.,
46(7), 1674-1684.
732
-------
TREATMENT OF PHENOL-FORMALDEHYDE RESIN WASTEWATER
USING ROTATING BIOLOGICAL CONTACTORS
BY
Lloyd W. Bracewell PhD
Associate Engineer, Swanson-Oswald Associates, Inc.
San Francisco, California
David Jenkins PhD
Professor of Sanitary Engineering
University of California, Berkeley
Berkeley, California
Wilfrid Cameron
Biochemist, Swanson-Oswald Associates, Inc.
San Francisco, California
INTRODUCTION
Phenol-formaldehyde resin manufacture produces a very high strength
waste (COD approximately 200,000 mg/1), containing 60,000 mg/1 phenol,
20,000 mg/1 formaldehyde and 15,000 mg/1 methanol (PFM waste). Economical
and environmentally acceptable disposal of this waste is a major problem.
The problem is emphasized because of recent regulatory activity
which places phenols on the EPA list of priority pollutants. A survey of
local sewer use ordinances for 40 municipalities (POTW's) [1] revealed
phenol discharge limits of 0.005 - 3000 mg/1 with a mode of 0.5 mg/1.
Thus, to discharge PFM waste to most POTW's would require pretreatment
and/or dilution to reduce phenol by 99.99992%. Because of such stringent
requirements, PFM waste has been destroyed by incineration. While this
method results in zero discharge, the economic attractiveness is decreased
as the price of energy increases.
733
-------
After researching the techniques available to treat and/or dispose
of PFM waste including chemical oxidation [2,3,4], recovery systems
[5,6,7], and biological treatment [8,9,10], a multi-staged process
scheme was selected. Although all unit processes involved established
technologies, the proposed combination of treating and disposing of the
PFM waste was unique. A rotating biological contactor (BBC) was an
essential part of the process train. This paper focuses on the use of
the 1BC in treating PFM waste; discussion of the other components in the
process can only be covered briefly here.
PROCESS SELECTION
The selection of processes was largely influenced by the stringent
local sewer ordinances for phenol (J..O mg/1) , by the location of the
plant in Northern California, and by the low waste volume (less than
10,000 gpd). The stringent phenol limit meant that a treatment plant
treating the PFM waste would have to attain greater that 99% phenol
removal. The combination of Northern California location and climate,
low waste volume and available land area made attractive a zero dis-
charge approach, using terminal facultative evaporation ponds. Prior to
discharge into the ponds, it was necessary to consider additional bio-
logical treatment so that pond area for treatment could be reduced and
the need for pond aeration could be eliminated along with the possi-
bility of stripping PFM into the atmosphere. Rotating biological con-
tactors were selected for the pre-pond biological treatment step because
of their ease of operation, their relative freedom from sludge settling
problems compared to activated sludge, the ability to cover the system
for odor control, and their ability, compared to trickling filters, to
operate at high loadings without clogging.
A survey of the literature on phenol and formaldehyde removal by
biological treatment revealed that up to 2000 mg/1 of phenol is amenable
to biological treatment with 91% phenol removal;formaldehyde concentra-
tions of up to 670 mg/1 have been treated biologically with removals
generally in excess of 90%. The results of this review (Table I) indi-
cate that the usual phenol and formaldehyde levels subjected to aerobic
biological treatment are in the range of 100-500 mg/1 phenol and 100-300
mg/1 formaldehyde.
Organic removal rates have been reported for some phenol-containing
wastes in fixed film biological systems. Porter and Dutch [15] report
the use of a plastic media trickling filter tower with effluent recycle
to dilute influent phenol to approximately 100 mg/1. Phenol removal was
98% when the filter was loaded at approx. 22 Ib phenol/1000 ft3 of media
(approx. 2 Ib COD/1000 ft2/ day.) Jenkins [16] indicated that a phenol-
formaldehyde resin waste was diluted 40-fold and treated at the rate of
5000 Imp. gal/yd/day on a rock media trickling filter (assumed specific
area of 15 ft2/ft3) at a rate of 45 Ib COD/ft2/day with approximately
25 percent phenol removal.
734
-------
TABLE I
BIOLOGICAL TREATMENT ' OF PHENOLIC/FORMALDEHYDE WASTES
Method
Activated Sludge
Oxydation ditch
Lagoon
Trickling filter
Trickling filter
Trickling filter
and lagoon
Activated Sludge
Activated Sludge
Activated Sludge
Trickling filter
Trickling filter
Reference
Eisenhauer, 1968
[11,- 12]
11
it
it
Wolnak, 1971
[13]
n
Biczysko, 1969
[14]
n
n
Porter-Dutch,
1960 [15]
Jenkins, 1957
[16]
Initial
Phenol
Concentration
(mg/D
100-750
400-900
100-250
200-500
_
_
2000
1000
300
100
1000
Phenol
Reduction
(V\
\/o)
90-100
87.99
90-95
85-98
_
-
91.0
97.6
99.4
98
25
Initial
Formaldehyde
Concentration
(mg/1)
-
200
200
670
330
100
-
-
Formaldehyde
Reduction
(%)
-
75-90
99
93.2
99.6
99.6
-
-
-J
w
Ul
-------
While these literature values for phenol waste loading and removal
efficiency served as a guide for the the dilution requirements and
loadings, as well as removals expected in an RBC plant, no specific
information was available on the performance of RBC systems treating PFM
wastes. It was also known that facultative fonds can act as clarifier,
sludge digester and a source of recycle water, and are able to effi-
ciently treat low phenol concentrations (25-100 mg/1 [11,12,17]). The
RBC/ facultative pond recycle system was thus to be the basic treatment
approach; however, we also discovered in laboratory scale testing that
PFM waste, itself the distillate from a high temperature distillation at
the resin manufacturing site, could be further distilled at low temp-
erature (40-70°c) reducing phenol and formaldehyde concentrations
significantly. The pot residue of the low temperature distillation could
be incinerated at close to self-supporting combustion, as the residue
had almost 60 percent of the energy content of #2 fuel oil.
A pilot study was needed to test the feasibility of this combined
treatment process, to provide data that would establish the effective-
ness of the approach, and to allow an attempt at optimizing the sizing
of a distillation/RBC/faeultative pond recycle system, or just an RBC/
pond system should the distillation step not prove cost-effective. The
pilot study described in this paper had this objective.
MATERIALS AND METHODS
Pilot Plant Description
Two parallel but separate pilot plants were constructed during
September and October 1977, and operated from November 1977 through
March 1978. The biological treatment sections of both plants were
identical (Figures 1 and 2), and consisted of two stage RBC units
followed by a settling pond from which a recycle stream was drawn. The
recycle stream passed through a small RBC that was installed to provide
a source of seed organisms, in the event the larger two-stage RBC was
inactivated by toxic shock loads of PFM waste.
Pilot Plant A. A 4-stage, 250 ft2 surface area RBC pilot plant was
leased from the Autotrol Corporation (Milwaukee, Wisconsin) and sub-
divided into two 2-stage, 125 ft2 units by blocking the overflow weir
between stages 2 and 3. RBC Unit A consisted of the first two stages of
the Autotrol unit. It was fed by a calibrated scoop which rotated coax-
ially with the contactor. The discharge was taken from the second stage
of this unit. The combination of raw waste and recycle flow rates,used
in the pilot plant operation provided a hydraulic residence time of 3
hours in the RBC.
The settling pond was constructed from a 15-ft diameter, 4-ft deep
plastic pool. With a water depth of 3 ft, the pond contained 4000 gal.
Recycle flow was taken from a point 6 inches below the water surface by
an 0.2 hp submersible pump, and discharged to a sump with an overflow
736
-------
DILUTED RAW
WASTE RESERVOIR'
MIXING
CHAMBER
PRELIMINARY
RBC UNIT
FACULTATIVE
0 N D '
4000 GAL,
15' diom. x 4' deep
11 A II
FIGURE I PILOT PLANT "A
737
-------
MAIN RBC UNIT
MIXING
CHAMBER
FACULTATIVE
POND "B"
FIGURE 2 PILOT PLANT "B1
738
-------
back to the pond. The recycle rate was controlled by a manually-operated
valve on the sump outlet. The recycle flow then passed through a locally-
fabricated RBC containing six 2-foot diameter roughened plastic discs,
to provide 35 ft^ of contact surface.
Raw PFM waste, diluted 4-fold with domestic primary treated sewage
was fed into the RBC feed chamber, where it mixed with the recycle
stream from the preliminary RBC. When required, nutrients (ammonium
phosphate) were added directly to the diluted PFM waste reservoirs to
provide a 90:3:1 ratio of C:N:P.
Pilot Plant B. In addition to the units described for Pilot Plant
A, this plant had two 25-ft^ solar stills prior to the biological system,
fabricated from galvanized sheet metal, with glazing consisting of two
3-ft x 5-ft x 1/4" glass plates, set at 22° to the horizontal. The
design of these solar stills was based upon those currently used for
desalination of seawater [18,19].
Analytical and Sampling Methods
Analytical Methods. All samples were preserved and analyzed in
accordance with Standard Methods for the Analysis of Water and Wastewater
14th Edition [20], except as noted. COD was by the method of Ryding and
Forsberg [21]. The method was validated by running parallel samples by
Standard Methods techniques and obtaining values which were 95-100% of
those determined by the Standard Methods technique. Phenol was by gas
chromatography with a Varian Aerograph, Model 1200, with a 3-ft x 1/8
in. SS, 10% FFAP on Chrom W column. Parallel and additional determi-
nations were by the 4-amino-antipyrine method (Standard Methods).
.Formaldehyde and methanol were by gas chromatography, using a 5-ft x 1/8
in. SS Poropak N column. Parallel and additional analyses were attempted
by the method of Jephcott [22]. Ortho-phosphate was by the stannous
chloride method; Nitrate was by the zinc reduction method; Ammonia
Nitrogen was by the method of Zadorogny, et al [23]; Dissolved Oxygen
was in-situ, using a YSI specific oxygen electrode; pH was in situ
using a Corning No. 6 pH meter.
Sampling and Data Collection. The pilot study was conducted during
Fall and Winter of 1977-1978. Mean daytime temperature averaged 13°C.
In the pilot-scale distillation studies, distillation was discon-
tinuous, with make-up volumes of raw PFM waste added as required to
maintain a certain minimum volume in the still. The reported results of
low temperature PFM distillation are weighted averages of fractions
collected.
In the biological treatment studies, all samples were grab samples.
At a minimum, a bi-weekly sample was collected and stored on ice until
it was returned to the laboratory for analysis. Grab samples were deemed
generally representative because of the controlled hydraulic and organic
loading rates.
739
-------
PROCEDURE
The pilot plant program is subdivided as follows:
Start-up and acclimatization
Phase I — Operation at moderate PFM waste loadings
Phase II - Operation at high PFM waste loadings
Phase III(a) - Treatability of pond recycle flow
Phase III(b) - Treatability of PFM solar distillate
Start-up and Acclimatization
Primary treated domestic sewage was fed into each pilot plant at
the hydraulic loading rate recommended by the maufacturer (1.6 gal/ft^/
day, or 200 gal/day) for a period of two weeks.When an observable
biomass had developed on the RBG surfaces, PFM waste was introduced in
steadily increasing concentrations, while maintaining the hydraulic
loading constant at the above stated value.
Increases in PFM waste concentration were introduced first to
Plant A and then one week later, to Plant B, as a precaution in case an
overdose inactivated Plant A. This proved to be an unnecessary pre-
caution, as no toxic effects ascribable to increasing PFM waste levels
were observed during the study.
The PFM waste feed rates were increased to produce phenol incre-
ments of 50 to 75 mg phenol/1 at approximately weekly intervals, until
influent phenol concentrations of 300 to 400 mg/1 were reached in each
system. At this influent phenol level, the influent COD concentration
was approximately 1500 mg/1. After 6 weeks of acclimatization the do-
mestic sewage was replaced by recycle from the ponds.
Phase I - Moderate PFM Waste Loading
Influent COD levels greater than 1200 mg/1 in BBC Unit A were
obtained on December 1, 1977 (day 1 of the pilot program, and aver-
aged 1543 mg COD/1 (21 Ib COD/1000 ft2 media/day) over the next 48
days. RBC Unit B, the back up or control unit, did not reach influent
COD levels over 1000 mg/1 until December 23, 1977 (day 14) and averaged
1219 mg COD/1 (16 Ib COD/1000 ft2 media/day) over the next 34 days.
During this period it was attempted to keep the influent COD loadings as
constant as possible because at these COD levels phenol concentrations
averaged 350-400 mg/1, a value that from the literature was estimated to
be readily treatable.
Dilution water was continually provided to both 1BC units by
recycling pond effluent. Nutrient sources present in the pond recycle
stream left over from the RBC treatment of domestic primary effluent
during start-up and acclimatization were sufficient to support bio-
logical growth during this phase. As a result additional sources of
nutrients were not required.
740
-------
Phase II-High PFM Waste Loading
The application of PFM waste was increased dramatically to RBC Unit
A during the period from the forty-eighth to the fifty-eighth day and
then gradually decreased over the next 19 days to test the stability of
the RBC treatment process to shock loading. During this period influent
COD and phenol averaged 3061 mg COD/1 (41 Ib COD/1000 ft2/day) and 624
mg/1 respectively with peak values on day 58 at 6228 mg COD/1 (83 Ib
COD/1000 ft2/day) and 1200 mg phenol/1. It was attempted to use RBC Unit
B as a control during this phase and it was thus operated at a fairly
consistent COD loading of 1442 mg/1 (19 Ib COD/1000 ft2/day), almost
the same influent loading as applied to RBC Unit A in Phase I.
Nutrients in the holding ponds were exhausted during this phase and
nitrogen and phosphorous in the form of ammonium phosphate were added to
both RBC waste feed systems in the C:N:P ratio of 90:3:1.
Phase Ill(a) - Treatabilj-ty of Pond Recycle Flow
During the high PFM waste loading of Phase II, Pond A became over-
loaded with effluent COD and phenol concentrations over 600 mg/1 and 100
mg/1 respectively. To determine the comparative treatability of this COD
and residual PFM waste in the holding pond, RBC Unit A continued to
receive the recycle flow from Pond A but raw PFM waste was no longer
added to the influent,i.e., RBC Unit A was used to treat the water in
Pond A. This phase lasted 24 days.
Phase III(b)__-_ Treatability of PFM Solar Distillate
The solar distillation rates of PFM waste during November through
February for the size of stills used were inadequate to provide a suffi-
cient distillate volume to adequately load an RBC Unit at 1200 mg COD/1
on a daily basis. Therefore, solar distillate was collected and stored
until March 1 and then fed for 24 days to RBC Unit B at the average
influent COD loading level of 1235 mg/1 (17 Ib COD/1000 ft2/day) and
phenol level of 219 mg/1.
Solar distillation of PFM waste produces a distillate with a higher
relative percentage of methanol compared to phenol and formaldehyde than
in the raw PFM waste. This operating test run was undertaken to evaluate
the comparative treatability of solar distilled PFM waste versus the
previous runs treating raw PFM wastes.
PILOT PLANT RESULTS
Waste Streams
The chemical composition of the PFM waste stream to the pilot
plants is summarized in Table II. Raw PFM waste was applied to RBC Units
A and B during Phases I and II. Solar distillate from the PFM waste was
applied to RBC Unit B during Phase III(b). Raw PFM wAste has a pH of 4.6
and the chemical composition of its solar distillate is shown in Table
741
-------
II. When the pH of the raw waste is elevated by the addition "of NaOH or
lime to pH 9 or 10.5, the phenol-formaldehyde removals increase substan-
tially. Although this distillate was not used in the biological pilot
system the chemical composition is presented in Table II for comparative
purposes.
Pilot Plants A and B
Concentrations, loadings and removals of COD and phenol by each
pilot plant unit are tabulated in Table III. Figures 3 and 4 present the
time course of COD influent and effluent for each RBC unit. The effect
of COD loading on COD removal rates by each unit is presented in Figures
5 and 6. The effect of COD loading on percent COD removal for both pilot
plants is presented in Figure 7.
Results of COD, phenol, formaldehyde and methanol removal rates
attained by the RBC units and pond system for Phases I through III are
summarized in Table IV.
DISCUSSION OF RESULTS
It had been intended to conduct the pilot plant feasibility study
during June through December; however, delays and time limitations
required the program to be conducted from October through March during
the coldest part of the year. Water temperatures in the small pilot
plant averaged only 13-14 °C during the day and dropped to as low as
10 °C at night despite enclosing the RBC. Thus biological performance of
the RBC was not optimal at these low temperatures and was reduced
somewhat over what would be expected in a larger facility operating
year-round. Despite the adverse weather conditions and organic loading
extremes, the RBC pilot plant units showed operational stability and
treatment reliability.
Biological Treatment of Waste
During Phase I when COD loadings were relatively steady and of
moderate intensity RBC Unit A removed 61 percent of the influent phenol
and 63 percent of the influent COD at average influent concentrations of
427 mg/1 phenol and 1543 mg/1 COD. On a media surface area basis COD
loadings were approx. 21 lb/1000 ft2/day with COD removals at 13.1
lb/1000 ft2/day. The performance of RBC Unit B was slightly inferior -
54 percent phenol removal and 54 percent COD removal. Influent concen-
trations were 355 mg/1 phenol and 1435 mg/1 COD. The influent COD
loading rate was 16 lb/1000 ft2/day with COD removals of 8.8 lb/1000
£t^/ day. The reason for this small difference in performance of the two
1BC Units is not understood; in general, the results of the entire pilot
program suggest that RBC Unit A may have performed slightly more effi-
ciently than Unit B at the same organic loading rates. It should also be
noted that the interpretation of the phenol results for Phase I of Pilot
Plant B is limited as these are estimated values only consisting of two
actual data points combined with extrapolations from the much more
extensive COD data from Phase I.
742
-------
TABLE II
PFM WASTE STREAM CHARACTERISTICS
PFM WASTE STREAM
COD
mg/1
PHENOL,
mg/1
FORMALDEHYDE,
mg/1
METHANOL,
Raw Waste
180,000 - 200,000
45,000 - 65,000
20,000 - 24,000
15,000 - 22,000
-j
*»
Pilot Scale
Solar Distillation
pH 4.6
123,000
32,000
17,000
14,000
Laboratory Scale
Distillation
pH 9
pH 10.5
31,300
18,600
9,000
2,700
5,000
800
6,900
7,600
-------
TABLE III
COD AND PHENOL REMOVALS BY RBC UNITS
COD
2 min
^ l! \ max
ave
Effluent min
(mg/1) max
ave
Percent
Removal min
max
ave
Removal
Ib min
,_3_ 2, max
10 ft -day
ave
RBC UNIT A
PHASES
I
1220
1842
1543
340
769
563
46
76
63
8.6
20.0
13.1
II
1550
6228
3061
630
5568
2007
11
59
34
4.4
18.7
11.1
III
295
1070
694
111
315
195
47
84
72
2.4
10.0
6.7
RBC UNIT B
PHASES
I
138
2082
1219
62
744
561
30
78
54
1.0
19.3
8.8
II
1031
1830
1442
349
1572
812
14
66
44
3.4
13.6
9.0
III
800
1659
1235
310
973
455
44
72
63
6.5
14.8
9.8
PHENOL
2
Influent min
max
ave
Effluent min
max
ave
Percent
Removal min
max
ave
220
660
427
28
320
161
20
91
61
325
1200
624
61
1181
414
1.5
83
34
<10
210
127
<1
145
51
47
84
60
55
600,
3551
175
325
263
110
300
219
20
175..
1651
47
183
128
19
100
82
64
71
54
31
83
51
19
83
63
1.
2.
Estimate based on 2 data points and extrapolation from more extensive data.
Diluted with pond recycle.
744
-------
TABLE IV
SUMMARY OF PILOT PLANT RESULTS
Parameter
Phases
COD
RBC Inf.
RBC Eff.
Pond Eff.
TOTAL :
PHENOL
RBC Inf.
RBC Eff.
Pond Eff.
TOTAL :
FORMALDEHYDE
RBC Inf.
RBC Eff.
Pond Eff.
TOTAL:
METHANOL
RBC Inf.
RBC Eff.
Pond Eff.
TOTAL:
PILOT PLANT A
Avg. Cone.
(mg/1)2
I
1543
563
308
-
427
161
371
_
147
25
20
-
100
43
na
_
II
3061
2007
1112
-
624
414
139
-
112
56
12
-
323
205
124
-
Ill
694
195
298
-
127
51
22
-
47
14
10
-
72
41
26
_
Removal
Percent
I
63
46
79
61
73
90
83
20
86
57
-
>57
II
64
45
64
34
66
78
50
79
89
37
40
62
III
72
C-)53
57
60
57
83
48
29
63
43
37
64
PILOT PLANT B
Avg . Cone ,
(mg/1)2
I
1219
561
292
-
3551
1651
781
-
94
37
10
-
118
38
na
-
II
L442
812
429
-
263
128
92
—
70
36
18
-
181
144
115
-
Ill
1235
455
293
-
219
M
22
-
83
42
15
_
318
170
39
-
Removal
Percent
I
54
48
L 76
54
44'
76
61
73
89
68
_
>68
II
44
47
70
51
28
65
49
50
74
20
20
36
III
63
36
76
63
73
90
49
64
82
47
77
88
1.
Estimate based on 2 data points and extrapolation from more extensive COD
data.
o
'Diluted with pond recycle.
745
-------
-j
45.
cr>
10 20 30 40 50 60 70 80 90 100
DAY OF OPERATION
FIGURE 3 RBC UNIT A INFLUENT AND EFFLUENT
C.O.D. CONCENTRATIONS
-------
2000-
o
"O
O
O
O
Q
a
d
o
60
70
80
90
100
110
DAY OF OPERATION
FIGURE 4 RBC UNIT B INFLUENT AND EFFLUENT
C.O.D, CONCENTRATIONS
-------
03
15
o
o
o
10
w
IE
O
O
d
C.O.D. (mg/l)* 75 x C.O.D. (Ib./IOOO ft /day)
I
_L
I
10 20 2 30
INFLUENT C.O.D. LOADING (Ib./IOOO ft/day)
40
FIGURE 5
COMPARISON OF INFLUENT C.O.D. LOADING
WITH REMOVAL RATES FOR RBC UNIT A
-------
o
o
o
o
Ul
tr
o
d
15
10-
FIGURE 6
PHASES I
1.6 goi./IOOOfi/doy
PHASES I »I •
PHASE 1C A
C.O.D.(mg/l)« 75 x C.O.D.(lb./!OOOf!.2/doy) '
10 2° 2
INFLUENT C.O.D. LOADING (Ib./IOOO ft./day)
30
40
COMPARISON OF INFLUENT C.O.D. LOADING
WITH REMOVAL RATES FOR RBC UNIT B
-------
-4
Ul
o
100
80
2 60
(E
O
O
o
Ul
o 40
UJ
Q_
20
INFLUENT COO CONCENTRATION (mg/l)
1500 3000
4500
I
I
RBC UNITS A a B
* O
to
20 30 40 50
INFLUENT CO D LOADING (Ibs./IOOOft.^day)
60
70
FIGURE 7 COD REMOVAL EFFICIENCY
-------
During Phase II the influent COD concentration fed to RBC Unit A
was increased to 3000 mg/1. The average percent COD removal dropped to
34 percent but the COD removal rate decreased only 15 percent from 13.1
lb/1000 ft2/ day to 11.1 lb/1000 ft2/day. At these high loadings phenol
removals also decreased to 34 percent. During this period the biological
film on the biodiscs which previously had a greyish-brown appearance,
rapidly became white and filamentous. Microscopic examination of the
biofilm showed that Beggiatoa spp. was predominant in the film .to the
extent of nearly excluding all other organisms. When the COD loadings to
Unit A were again lowered in Phase III the amount of Beggiatoa spp. in
the biofilm decreased as the biofilm regained its original grey-brown
appearance concurrent with the increase in percent COD and phenol removals,
During this Phase II RBC Unit B continued to operate as a control
to Unit A. The average COD loading to Unit B was increased 18 percent to
19 Ib COD/1000 ft2/day to approximate the COD loading applied to Unit A
in Phase I. Comparing Unit B performance between Phases I and II phenol
removal efficiency apparently decreased only 6 percent while COD removal
efficiency decreased 19 percent; however, absolute COD removal rates
were essentially unchanged at 9.0 Ib COD/1000 ft2/day. The comparison
of phenol efficiency removal between Phases I and II is limited by the
uncertainty level associated with the Phase I phenol data for Pilot
Plant B, but the more extensive COD data indicate that the increase in
loading was sufficient to significantly reduce COD removal efficiency.
As a control RBC Unit B did not perform well in comparison to Unit A as
there appeared to be consistent treatment efficiency differences between
them.
In Phase III(a) RBC Unit A treated only the recycle stream from
Pond A which had become overloaded during Phase II. As the pond contents
were treated, the COD levels in the pond were reduced and the COD loadings
to the RBC decreased gradually to an average of 9.3 lb/1000 ft2/day
with COD removals averaging 60 percent. It would appear that this
recycle stream may have been somewhat more resistant to biological
oxidation than the diluted raw PFM waste fed during Phase I because,
although COD removal efficiency was 14 percent higher than in Phase I,
the average COD loadings in Phase III were only 45 percent of those in
Phase I.
When influent COD loading is plotted against COD removal for
RBC Unit A (Figure 5) a maximum COD removal rate at approximately 14
lb/1000 ft2/day is observed at a COD loading rate of 22-23 lb/1000
ft2/day. For RBC Unit B (Figure 6) a maximum COD removal rate is not
clearly defined because COD loadings were not increased to values high
enough to establish the shape of the curve at COD loadings above 20
lb/1000 ft2/day. Percentage COD removals for RBC Units A and B were then
combined and plotted against influent COD loading as shown in Figure 7.
The result is an almost linear relationship with 80 percent removals
found at COD loadings of 11-12 lb/1000 ft2/day and 20 percent removals
found at COD loadings of 38 lb/1000 ft2/day. The general relationships
established between COD loadings and removals as shown in Figures 5,6,
and 7 can now be used to assist in sizing an RBC to treat a PFM-waste.
Various COD loadings and required removals can be plugged in to optimize
the sizing and performance of the RBC unit in the overall process train.
751
-------
Biological Treatment of Distilled PFM Waste
In Phases I and II it was clearly demonstrated that both RBC Units
were able to remove 50-60 percent of the COD from diluted PFM waste with
COD loadings of 16-20 lb/1000 ft2/day. In Phase III(b) the impact of a
distillation pretreatment step on the biodegradability of PFM waste was
investigated. Here, RBC Unit B received solar distillate from raw PFM
waste which had an undiluted COD of 123,000 mg/1 but contained a much
higher percentage of COD attributable to methanol than the raw PFM waste
(Table II).
During biological treatment of the solar distillate COD removals
in RBC Unit B were 63 percent at influent COD levels that were almost
the same as in Phase I and only slightly lower (17 percent) than in
Phase II. The higher percent COD removal in Phase III (63 percent)
compared to Phases I and II (54 percent and 44 percent) indicate that
the distilled PFM waste was somewhat easier to treat. Also a plot of COD
removal versus influent COD loading in Figure 6 seems to indicate that
absolute removal rates in Phase III were higher than in Phases I and II.
Phenol removal in RGB Unit B was also higher in Phase III than in Phases
I and II by an amount that would largely account for the higher COD
removal. This result suggests that distillation may remove some poorly
oxidizeable phenolic compounds with the more biodegradeable phenolics
being more distillable. Thus, distillation of raw PFM waste besides
reducing COD loading by up to 90 percent [if the pH is adjusted to 10.5
(Table II)] may also provide a distillate slightly more amenable to
biological treatment.
Formaldehyde and Methanol Removal
Problems were encountered especially prior to Phase II with the
analytical methods for formaldehyde and methanol. The colorimetric
method of Jephcott was unsatisfactory and non-reproducible in our lab,
even with standard solutions, so that a gas chromatographic method had
to be adopted in the study. Data for formaldehyde and methanol are
presented in Table IV together with the COD and phenol data for both
pilot plants and all phases of operation.
Based on only 2 measurements, formaldehyde removal by RBC Unit A
in Phase I was 83 percent; in Phases II and III formaldehyde removal
decreased to 50 percent for both Units A and B. The only supportable
conclusion that can be made is that formaldehyde removals of 50 percent
were obtained independent of influent formaldehyde concentrations bet-
ween 72 and 323 mg/1. This percentage removal rate is lower than some
data reported in the literature for activated sludge. It is speculated
that some removals of PFM in activated sludge are probably due to
stripping by aeration.
In general, methanol removals in Phase I were 60 percent and about
40 percent in Phases II and III.
752
-------
Pond Performance
Table IV presents COD and PFM results for Recycle Ponds A and B,
and overall removals for the RBC units and the recycle ponds together.
The pond data should not be considered representative of full-scale
facultative pond performance because the pools were shallow and had a
low hydraulic residence time. In particular during the heavy loading of
RBC Unit A during Phase II, the Recycle Pond A was highly overloaded so
that in Phase I only pond effluent was recycled to RBC Unit A. RBC Unit
B was not so heavily loaded as RBC Unit A so that the performance of
Recycle Pond B did not change markedly during the pilot program. The
pond recycle retention times were 20 days so that changes in pond per-
formance between various phases of operation were difficult to distin-
guish. On the average, it can be said that these shallow pools removed
approximately 45 percent of the influent COD (largely through settling
since very little algae grew during the winter months of operation) to
produce an overall pilot plant removal of 70-75 percent. Pond phenol
removalyle Pond A was 60-70 percent and 30-70 percent in Recycle Pond B
to produce an overall plant phenol removal of 80-90 percent in Pilot
Plant A and 70-90 percent in Pilot Plant B.
Formaldehyde removals in the Recycle Ponds varied substantially
about an average removal of approximately 60 percent. Overall pilot
plant removal for formaldehyde was generally greater than 80 percent. No
methanol data was available for the Recycle Ponds in Phase I but in
Phases II and III the Recycle Ponds averaged 40 percent methanol removal
to produce an overall pilot plant methanol removal of greater than 60
percent.
Degradation of PFM Waste in RBC
Effluent COD values reported for both RBC units are for filtered
samples in order to trace the removal of soluble COD. Unfiltared ef-
fluent COD values were also obtained throughout the pilot study to
evaluate the relative percentage of influent COD biologically oxidized
and converted to organic material. The unfiltered COD results for both
RBC units showed that at average influent COD levels of 1500 mg/1, 32
percent of the influent COD was oxidized and 22 percent was converted to
suspended COD (Table V). The suspended COD in the RBC effluent will then
settle in the facultative pond and be anaerobically digested while the
soluble COD will need to be treated aerobically in the pond. The pond
soluble COD will be treated further by recycle of the pond effluent
through the RBC.
753
-------
TABLE V
SOLUABLE COD REMOVAL AND CONVERSION TO BIOMASS
Influent COD, unfiltered (mg/1)
Effluent COD, unfiltered (mg/1)
Effluent COD, filtered (mg/1)
COD removal (%)
COD conversion to bionass (%)
minimum
1030
375
345
1
1
maximum
1566
1906
1818
76
66
average
1445
982
715
32
22
Air Quality
Considerations of State of California Air Resources Board and OSHA
regulations made the emission of phenol and formaldehyde into the atmo-
sphere from the RBC units a major concern. Air in the enclosed main
contactor section of the RBC units was sampled for both phenol and
formaldehyde utilizing a micro-impinger, when influent COD concentra-
tions were at 1500 mg/1. No formaldehyde was detected in 1 m-* of
sampled air; phenol levels were 0.013 mg/m^ of sampled air. By com-
parison, the TLV (Threshold Limit Value) for phenol established by
OSHA is 10 (19 mg/m3) and the NIOSH recommended limit for phenol is
20
RBC Performance
The RBC part of the pilot plant performed well overall removing 60
percent of the phenol and COD at moderate loadings i.e., influent phenol
levels of 300-400 mg/1 and COD levels of 1200-1500 mg/1. The RBC's
demonstrated excellent treatment stability and showed no toxic effects
even at very high influent phenol levels (over 1000 mg/1) . By comparison
to the literature values shown in Table I, however, the pilot plant RBC
units did not achieve consistently the 85-95 percent phenol removals
reported. It is possible that RBC's are not as efficient at treating PFM
wastes as activated sludge or trickling filters; but, it should be noted
that the aeration process in activated sludge and the cascading of
sewage in the trickling filter dosing process certainly strips volatile
organics from the water adding to the overall removal efficiencies. No
data were found to indicate the level of phenol or formaldehyde in the
air over activated sludge or trickling filter units treating high
levels of these wastes and thus no relative weight can be given to the
contribution of this process in the overall efficiency of treating PFM
754
-------
waste. In this study in an enclosed RBC unit no formaldehyde and less
than 0.1 mg/m^ of phenol could be detected in the enclosure, indicating
that very little stripping of these organics occurred in the enclosed
RBC pilot system.
• The temperature of the pilot system also probably played a signi-
ficant role in that optimum pilot plant performance can not be expected
when the temperature is only 13 or 14°C. If the operation of the pilot
plant had been continued through warmer temperatures biological activity
would have increased and removal efficiences would no doubt have improved
somewhat, perhaps to 80 percent but probably not higher unless loading
rates were reduced significantly from the average values used. Despite
the apparent inefficiencies of RBC's in treating PFM waste compared to
activated sludge or trickling filters, it is likely that RBC's are
actually equally efficient if all conditions were equal and air stripping
of organics was included in an overall mass balance. This could only be
proven though by comparing parallel pilot systems.
SUMMARY AND CONCLUSIONS
Economical and environmentally acceptable disposal of high strength
phenol-formaldehyde waste (PFM waste) from resin manufacturing is a
major problem. Phenol is on the EPA list of priority pollutants and
local sewer use ordinances are generally stringent with respect to
phenol, usually allowing less than 1 mg/1. With the phenol level at
60,000 mg/1 in PFM waste, conventional biological treatment systems even
designed for 99 percent removals can not achieve 1 mg/1 effluents. Thus a
no discharge approach was selected as the most appropriate solution.
Currently PFM waste is disposed of by incineration and, while this
method results in zero—discharge, its economic attractiveness decreases
with the increasing cost of energy.
The selection of an alternative zero discharge treatment process
was largely influenced by the local climate, availability of land, and
the volume of waste to be treated - less than 10,000 gpd. The biological
system selected consisted of a rotating biological contactor (RBC)
followed by an evaporative facultative ponding system. A solar dis-
tillation unit was also evaluated as a pretreatment step for the PFM
waste. The pilot program demonstrated that a strong phenol-formaldehyde-
methanol waste from resin manufacturing or its solar distillate could be
treated by a combination of rotating biological contactor and recycle
ponding system,
The pilot system as a whole and the rotating biological contactors
in particular performed well but less efficiently that expected when
compared to literature values for activated sludge and trickling filters.
However, the pilot program was carried out during winter conditions and
the resulting low temperatures in the small pilot system undoubtedly
reduced bacterial activity and thus prevented optimum removal rates from
being attained. Furthermore, the extent to which air stripping contri-
buted to the literature removal efficiencies quoted is not known. The
RBC's did operate effectively throughout the study period under varying
755
-------
climatic and loading conditions and exhibited excellent stability in
withstanding periodic shock loadings. Percentage and absolute COD
removal rates for the RBC's were found to be a function of the influent
COD loading and this data provided sufficient information to allow the
sizing of full-scale rotating biological contactors for the treatment of
the PFM waste involved.
The choice of treatment system and the relative sizing of the
components would depend mainly upon the relative availability of land
area, capital, solar insolation and process waste heat. The solar dis-
tillation of the PFM waste has the advantage of producing a distillate
that is 15-20 percent more biodegradable and of reducing the COD loading
from 40-90 percent depending upon the level of pH adjustment of the raw
PFM waste. These benifits of solar distillation must be traded-off
against the capital and operating costs of the solar still, as well as
the decrease in sizing of the RBC and ponding system. Still size and
hence cost is dependent most strongly upon solar insolation values and
upon the amount of of process waste heat that could be made available
to supplement solar energy,
The relative sizing of the RBC/pond system depends upon the land
available for ponding and the waste stream volume if evaporative and zero-
discharge are treatment goals. The overall sizing of these units depends
on whether or not solar distillation is used and to what extent pH ad-
justment is made during solar distillation. After detailed cost-benifit
analysis comparing all of the above variables in a capital and operating
cost matrix and with the goal of an evaporative-zero discharge system,
it was determined that the least cost system for capital as well as
0 & M was a 1.1 acre still, operated at pH 9, followed by 2 RBC units
with 100,000 ft^ of contactor surface each, and four 1-acre ponds. Two
of these ponds would be deep facultative primary ponds, one would be
a second pond and one just for evaporative and blowdown.
756
-------
1. U.S. Environmental Protection Agency, Pretreatment of
Industrial Waste, Seminar Handout, p. 89, 1978
2, Hugh R. Eisenhower, The Ozonization of Phenolic Wastes,
Journal of Water Pollution Control Federation, 40,11, Part
1, 1968
3. Hugh R. Eisenhower, Increased Rate of Efficiency of
Phenolic Waste Ozonization, J. Water Pollution Control
Federation, 43,2, 1971
4. Oil adn Gas J., Phenols in Refinery Waste Water can be
Oxidized with Hydrogen Peroxide, V73 N3, Jan. 1975
5. Chester R. Fox, Remove and Recover Phenol, Hydrocarbon
Process, Vol. 54, No. 7, 1975
6. Rohm and Haas Company, Process for the Recovery of Phenol
From Aqueous Streams, Formaldhyde Regeneration System.
7, Vargiu et al., United States Patent 3,869,387, Process
for the Extraction of Phenol from Wastewaters in the
From of Urea-Formaldehyde-Phenol Condensates.
8. G. Brigmann and W. Schroder, Grosstechnische Biologische
Entphenolung der Abwasser eines Kunstharzbetriebes nach
dem Nocardia-Berfahren, Heft 7 (81, Jahrgang 1960)
Gesundheit—Ingenieur
9. Belgian Patent BF 807-425, Anaerobic Degradation of Phenol, Issued
July 5, 1974, Derwnent Belgian Patent Report, Vol. 5, No. 22, p.D2,
July 1974
10. Richard E. Rosfjord et al., Phenols, A Water Pollution
Control Assessment, Water and Sewage Workks, Vol. 123,
No.3. 1976
11. Hugh R. Eisenhower, Bephenolization of Water and Wastewater,
Water and Pollution Control, Sept. 1968
12. Hugh R, Eisenhower, Oxidation of Phenolic Wastes, J. Water
Pollution Control Federation, 36, 9, 1964
13. Bernard Wolnak and Associates, An Evaluation of the Potential
Effect of Jet-0-Matic and Monomatic Waste Inputs on Sewage
Treatment, unpublished.
757
-------
14. Jan Biczysko, Bakania Ned Oczyszczaniam sciekow Z Produckcji
Zywic Fenolo-formaldehydowych, Przemyse Chemiczny, 48,10,1969.
15. Porter and Dutch, J. Water Pollution Control Federation, vol.
32, p.622, 1960
16. Jenkins, S.H., Treatment of Trade Waste Waters and the Preven-
tion of River Pollution, Ed. Isaac, 1957
17. Bach, H., Disappearance of Phenol in Water,Gesundh, Ing. 52,
796, 1929.
18. United Nations Department of Economic and Social Affairs, Solar
Distillation as a Means of Meeting Small-Scale Water Demands,
1970.
19. National Academy of Sciences - National Research Council,
Pub. 568 - Saline Water Conversion, Proceedings of a Symposium,
46 November 1957
20. Standards Methods for the Examination of Water and Wastewater,
American Public Health Organization, 14th Edition
21. Ryding and Forsberg, A Mercury-Free Accelerated Method for
Determining COD; Water Research, Vol 11:pp. 801-805, 1977
22. Jephcott, Analyst, 60:588, 1935
23. Zadorogny,et al., Journal of the Water Pollution Control
Federation, vo. 45, No. 5, May 1973
758
-------
ENERGY RECOVERY FROM ANAEROBIC ROTATING BIOLOGICAL CONTACTOR
(AnRBC) TREATING HIGH STRENGTH CARBONACEOUS WASTEWATERS
A.A. Friedman, Associate Professor
Department of Civil Engineering
Syracuse University, Syracuse, N.Y. 13210
S.J. Tait, Sr. Research Engineer
International Paper Company
Tuxedo Park, N.Y. 10987
The continuing development of ever more stringent industrial effluent
discharge standards coupled with rapidly escalating energy costs encourage
the development of new approaches for the treatment of high strength carbon-
aceous wastewaters. Aerobic treatment designs for high strength wastewaters
are usually constrained by energy intensive mixing, oxygen transfer or sludge
handling processes. Previous experimental work with overloaded conventional
rotating biological contactors (1, 2, 3) led to the concept of the anaerobic
rotating biological contactor (AnRBC) for the low cost treatment of high
strength wastewaters. Based on the bench scale pilot plant data presented
below, it appears that the AnRBC process can attain high quality effluents
with relatively low energy inputs and can even result in net energy yield
under appropriate loading conditions.
AnRBC CONCEPT
Conceptually the AnRBC is similar to conventional aerobic rotating bio-
logical contactors in that microorganisms become attached to and grow on ro-
tating discs that are partially submerged in the wastewater as shown schemat-
ically in Figure 1. Groups of discs, separated into sequential compartments
called stages, are partially immersed in the wastewater and rotated continuous-
ly both to provide mixing within each stage and to facilitate product gas
transfer to the anoxic atmosphere maintained above the water surface. Adjacent
stages are separated by baffles to minimize short-circuiting. The flow is
passed from stage to stage through holes in the baffles below the waterline.
However, unlike conventional RBC units, the AnRBC is enclosed in an
airtight housing with an anoxic atmosphere maintained above the liquid. Also,
the depth of disc submergence is substantially greater than in conventional
aerobic units. Microorganism attachment to the rotating surfaces provides
long mean cell retention times in the reactor, which in turn encourages the
development of the slow growing methanogenic bacteria responsible for the con-
version of low molecular weight compounds to methane. Sloughed excess bio-
mass is carried from stage to stage through openings in the baffles and
leaves the reactor in the effluent. Conceptually, the AnRBC process couples
the advantages of the short hydraulic detention times typical for the fixed
film horizontal flow RBC process with the high strength, carbonaceous degra-
dation capabilities of anaerobic systems.
AnRBC PILOT PLANT AND EXPERIMENTS
The primary purpose of this study was to ascertain the conditions whereby
759
-------
anaerobic non-methanogenic and methanogenic microorganisms could successfully
be grown on rotating disc surfaces. Secondary purposes included (a) the de-
velopment of the models jfor predicting the removal of soluble organic substrate
as a function of both wastewater organic strength and feed flow rate, and (b)
the preliminary evaluation of process net energy requirements or yields.
To accomplish these goals, two identical four-stage reactors were con-
structed from plexiglass as shown in Figure 1. The inside diameter and disc
diameters were 13.97 cm (5.50 in.)& 12.70 on (5.00 in.)5respectively. The
1.270 cm (0.500 in.) horizontal shaft was supported by external end bearings
and was rotated by attachment through a pulley and belt arrangement with a
variable speed drive motor. Successive stages were separated by fixed baffle
plates with three 1.91 cm holes below the water!ine for the passage of solids
and the wastewater carrier stream from stage to stage. Ten 0.318 cm thick
discs were contained in each stage, providing a total reactor disc surface
area of 1.013 m2 (10.908 ft,2). The first stage of each reactor was preceded
by a small mixing chamber that contained a flat bladed impeller to distri-
bute the influent feed evenly through the numerous holes in the baffle A
separating the mixing chamber and the first stage. The water level in the
reactor was controlled by a dynamic head tube resembling a vented inverted
siphon on the effluent line. Valved liquid sampling ports permitted grab
sampling from each stage as well as from the effluent. Sas collected in
each stage was vented through a common manifold to a wet test gas meter.
The AnRBC pilot plants were operated with about 70 percent of the disc
area submerged for the studies reported herein. The reactor liquid volume
for these conditions was about 5.27 liters. Both reactor systems were housed
in a controlled temperature room with the internal reactor temperatures main-
tained at 35 +_ 0.5°C.
Experimental Parameters
The feed stock used for these experiments is similar to that used in
earlier loading studies with aerobic RBC pilot plants (1, 2, 3). The syn-
thetic wastewater contained dissolved sucrose as the sole organic carbon
source and was mixed with the soluble inorganic constituents shown in Table 1.
Ammonium salts served as the sole nitrogen source for these experiments to
avoid devitrification conditions. These soluble wastewater constituents are
biochemically similar to those that might be expected to result from food,
bottling and some organic chemical processing industries.
Soluble organic carbon was measured with a Dohrman DC50 Organic Carbon
Analyzer on samples that had been filtered through a 0.45 micrometer Mini-
port HA filter. Total alkalinity at a pH end-point of 3.7 was determined
as suggested in Standard Methods (4). Volatile acid akalinity (VAA) was
measured by the procedure developed by DiLallo and Albertson (5). Because
the analytical results of this test are not exact, the volatile acid alkalin-
ity results reported below should only be interpreted qualitatively. Since
influent organic substrates were entirely soluble, effluent solids (measured
as total residue, volatile residue, total nonfilterable residue, and volatile
nonfilterable residue as suggested in Standard Methods) (4) represent either
solids grown and sloughed in the reactor or solids precipitated in the reactor.
Product gas generation rates were measured with a wet test gas meter and gas
composition was determined with a gas chromatograph.
A total of eight experiments with variable loading conditions was con-
ducted and they can be considered portions of a single factorial experiment
760
-------
as indicated in Table 2. Loading conditions for each steady state experiment
are multiples of the influent flow rate (H) and the influent TOC concentration
(F) used in Experiment 1. Thus, the 3F-2H loading notation shown in Table 2
for Experiment 8 indicates that the influent feed concentration (C..) was about
three times stronger than that used in Experiment 1 and the flow rate
(Q) was twice as large.
Operati ons
Prior to start-up, tracer study experiments indicated that these four-
stage reactors can be described hydraulically as four complete-mix reactors
in series. The initial start-up procedures have been described elsewhere (6).
The reactors were operated for a given loading condition until effluent
soluble TOC concentrations and gas production rates varied by less than +5%
for three successive days. Operation for a minimum of three weeks following
a step change in loading was required to achieve these quasi-steady state
operating conditions. Intensive sampling over the next day was used to obtain
the data reported below. Continuous operation was extended for 218 days
until an 0-ring failed and reactor fluids began leaking from a bearing.
Experimentation was discontinued at this time and both reactors were dismantled
for inspection. Microbial solids were observed to coat all disc surfaces and
decreased in thickness from the first through the fourth stages, with only a
thin attached film layer present on the fourth stage discs.
RESULTS
Figures 2, 3, and 4 display the general data patterns seen as a result of
these experiments. The influent volatile acid alkalinity (VAA) was nearly
negligible relative to the values observed in the reactors for all feeding
conditions. Even though stage data points have been connected with straight
lines, it should be remembered that each stage is acting as a complete-mix
reactor and that there are hydraulic and reaction discontinuities between
adjacent stages.
Figure 2 illustrates that the VAA generated in the first stage was ac-
companied by simultaneous minor depressions of pH and carbonate alkalinity.
About 80% of the TOC removal occurred in the first stage. Both TOC and VAA
were further reduced in succeeding stages while the pH and alkalinity increased
in the downstream direction. Note that the fourth stage did not contribute to
TOC removal for this loading condition and apparently the hydraulic residence
time in the reactor could have been reduced from 17.5 to about 13 hours with
no loss in soluble TOC removal performance.
The results from Experiment 6, shown in Figure 3, are similar even though
the organic loading rate has been increased by a factor of four and the reactor
hydraulic detention time reduced to one-half that of Experiment 1 conditions.
For this loading situation, the first stage removed about one-half the influent
TOC. However, volatile acid production was more significant in the first
stage. Note also that the fourth stage TOC and VAA concentrations were sig-
nificantly higher, and that the overall soluble TOC removal was reduced from
the 96% found in Experiment 1 to about 79% in Experiment 6. It is probable
that additional stages would have resulted in more complete TOC removal.
Figure 4 illustrates the effects of a high loading rate and short hydraulic
detention time on system performance. The organic loading rate in Experiment
4 was increased by a factor of about 8.5 compared to Experiment 1, with the
761
-------
detention time being reduced to one-eighth that of the original conditions.
The sharp drop in pH and alkalinity along with the high production of VAA
suggests that acid fermentation occurred at a rate that overwhelmed the
slower-growing methanogenic bacteria in the system. Only 46% of the soluble
TOC was removed under these loading conditions. The increase in VAA in the
fourth stage may suggest that the fermentation process was inhibited by high
concentrations of soluble substrate in the first three stages.
When the data from these experiments are arrayed as shown in Figures 5
and 6, the effects of influent feed concentration (C.) and flow rate (Q) on
performance can be observed. Figure 5 shows that increasing the flow rate
(reducing the hydraulic detention time [0]) for constant C. conditions results
in higher TOC concentrations at most points within the reactor. Under rel-
atively low mass loading conditions (Experiments 1 and 2), effluents contained
relatively low volatile acid concentrations. At the higher mass loading rate
(Experiments 3 and 4), the effluents contained relatively high volatile acid
alkalinity concentrations and the product gas had a lower methane content.
Thus, an increase in the influent flow rate has the apparent effect of moving
methane fermentation downstream in the reactor. Alternately, these data sug-
gest that there might be a critical reactor hydraulic detention time required
to sustain effective methane fermentation. Based on these data, it appears
that this critical time period is between 4.4 and 8.8 hours. Of special
interest is the fact that the first stage substrate removal rates shown in
Figures 5 and 6 are substantially higher than those reported for conventional
aerobic RBC systems.
The data shown in Figure 6 can be used to compare TOC removal as a func-
tion of influent substrate concentration for constant flow rate conditions.
As expected, the system requires more stages (area) to achieve a specific
residual organic effluent concentration with increasing C^ values. Again,
the first stage organic removal rates are substantially higher than those
reported for conventional RBC treatment.
Mass loading and removal data are shown in Figure 7. Up to a loading of
about 22 g TOC per day (21.7 g TOC per m2 -day), soluble TOC removal appears
to be independent of either influent concentration or flow rate and averages
95+ percent removal. Although a smooth curve could have been passed through
all the data points, a sharp discontinuity in soluble TOC removal exists near
this limiting mass loading and it is evident that both additional increments
of TOC removal and the overall percent removal are highly dependent on mass
loading conditions.
Due to the small flow rates, gas production was difficult to monitor
with the available equipment. However, as shown by the dashed line in Figure 8,
a good linear correlation exists between soluble TOC removed and total gas pro-
duction. When data from Experiments 4 and 8 are omitted because high VAA
effluent concentrations indicate incomplete reactions, an even better corre-
lation exists (solid line). Hence, total gas production can be conservatively
estimated as 1.76 m3/kg TOC removed. Methane and carbon dioxide were present
in about even quantities for Experiments 1, 2, 5 and 7. The carbon dioxide
content increased from 54% in Experiments 3 and 6, to about 60% in Experiments
4 and 8. These, along with the VAA, pH, and TOC data, provide evidence that
methanogenic activity was unable to proceed to completion under the higher
loading conditions.
Effluent solids production is a critical consideration in the design of
biological treatment systems. As expected, gross solids production was depen-
dent on the mass of TOC removed as shown in Figure 9, with more net solids
being generated for the higher loading conditions. The observed yield (Y b ),
762
-------
defined as the mass of solids generated per mass of soluble TOC removed, appears
to be related to the logarithrh of the loading factor (0C-) as shown by the
solid line in Figure 10. Nearly as good a correlation and probably a more
meaningful relationship was found when Y . was compared to the logarithm of
0AC as shown by the dashed line in Figure 10. It seems rational to
expect lower observed cell yields under conditions of long detention times
and more complete carbon conversion to stable end-products.
PREDICTIVE MODELS
When stage TOC data from the eight experiments are displayed on a semi-
logarithmic plot as shown in Figure 11, apparent pseudo-first order relation-
ships between soluble TOC remaining and time are observed with slopes and
intercepts dependent on C.. For a given influent substrate concentration C.,
TOC remaining can be approximated by
In C = In C1 - KfT, (1)
where C is the soluble TOC (mg/1), Kf is the pseudo-first order rate constant
(hr1), and T is the hydraulic residence time (hr). When the slopes from the
three lines are compared with C-, a strong correlation between the pseudo-first
order rate constant (K^) and C- is observed (Figure 12). Both linear and loga-
rithmic relationships provide good predictions for IC over the range of C-
data evaluated. Using the empirically determined linear relationship
Kf = 0.3744 - 7.96 x 10"5 Ci (2)
for simplicity with Equation 1 yields
In C = In Ci - (0.3744 - 7.96 x lO^^T. (3)
As shown in Figure 13, Equation 3 provides reasonably good predictions for C
when compared against the measured data. The data scatter is suprisingly small
when^on^considers that a simple pseudo-first order expression is being used to
descrT&e the results of a complex series of anaerobic reactions that were only
partially complete for some of the loading conditions employed in these
studies.
Of course, the empirically determined constants indicated in Equations 2
and 3 are only valid for the experimental parameters employed in this study.
One would expect these constants to change as a function of substrate type, disc
diameter, rotational speed, immersion depth, etc. Additional studies will be
required to develop these relationships. However, a pseudo-first order ap-
proximation of the form indicated in Equation 3 may prove useful for some
design situations.
A more useful model for applying the AnRBC concept to design situations
is based on an empirical model originally developed by Schroeder (3) that
considered mass flux concepts for the removal of organic material on the face
of an RBC disc. Friedman, Woods, and Wilkey (2) and later Friedman, Robbins,
and Woods (3) applied Schroeder's model to aerobic RBC systems by considering
each RBC stage as a completely mixed reactor. They proposed the mass of organic
material removed per stage could be described by
763
-------
K C.2
where:
M = mass or organics removed per stage per unit time, mass/time
K = a removal rate constant, volume/mass-time per stage
C. = substrate concentration entering the stage, mass/volume
k1 = a constant.
If in the first stage C.»k', then Equation 4 can be closely approximated by
Mn = KCi . (5)
A K value was calculated for each set of loading conditions by using Equation
5 and first stage data for each of the eight experiments (Table 3). Using
these K values, k1 was calculated with Equation 4 for each experiment using
the last stage data for which significant TOC removal occurred. This results In the
average k1 value of 0.0426 g/1. Despite significant scatter for individual
k1 values, this average value for k' was used for the succeeding calculations
since k' has little effect in the early stages where most soluble TOC removal
occurs.
Under identical mass loading conditions, first stage mass removal rates
vary significantly as seen by comparing M and K values for Experiments 2 and 5
as well as Experiments 4 and 6 in Tables. The observed differences are
dependent on both the stage hydraulic detention time (0 ) and the influent
organic concentration C-. The natural logarithm of the product of these two
parameters, ln(6 C.)» has been termed the "loading factor" and appears to
be related to K as shown in Figure 14. A straight line provides a reasonably
good fit of these data considering that pH and VAA values varied significantly
for each first stage loading condition. Again, future studies will probably
show K to be a function of disc diameter, rotational speed, substrate type,
and other operating parameters as well as the loading factor. However, since
a reasonable relationship does appear to exist between K and ln(0 C.)>
trade-offs between hydraulic detention time (flow), influent concentration,
and desired effluent concentration can be predicted with the use of an empir-
ically determined relationship such as that shown in Figure 14.
In order to yegrifythis empirical model, the measured soluble TOC data
from the eight experiments were compared to calculated model values for each
set of loading conditions. The linear relationship shown in Figure 14 was
used to predict the K value for each set of loading conditions. Equation 4
was next used with the resulting K value and k1 average to predict the mass
removal for each stage. The C^ for the following stage was back calculated
from the mass remaining in the previous stage. This iterative process was
used to calculate the results shown in Figures 14, 15, and 16. The experimental
and predicted values agree closely for the loading conditions investigated.
ENERGY CONSIDERATIONS AND PROCESS SCALE-UP
Estimating anticipated prototype performance :based on bench scale data
obtained under ideal laboratory conditions is subject to numerous errors and
unforeseen problems. However, the following two examples will describe the
764
-------
net energy requirements and yield that may be obtained with the AnRBC designed
to meet specific effluent requirements. It is assumed that conventional twelve
foot diameter plastic discs will be used for the prototype installation. Based
on torque measurements made with clean, full size discs with an additional 45
percent allowance for the mass of microorganisms adhering to the discs as well
as motor and gear box losses, it is assumed that 4 HP will be more than suffic-
ient to rotate a shaft carrying 9300 m2 (100,000 ft2) of media. As will be
seen later, net energy requirements are relatively insensitive to motor horse-
power requirements. The following analysis also ignores external pumping
requirements, product gas compressor requirements and heat losses from the
reactor. This latter loss is assumed to be offset by a heat exchanger oper-
ating between the AnRBC effluent and the influent streams. Because the direct
scale-up of bench scale aerobic RBC data has been previously shown to lead to
poor prototype design3, the effective area used in three examples has been
reduced by 25 percent from 9300 m2 to 6975 m2 per shaft.
Example 1 - High Quality Effluent Required (95 percent Removal)
Using the critical loading of 21.7 g/m2/day from Figure 7 as a basis
for design, the mass of TOC that will be removed per day can conservatively
be estimated as
21.7 g/m2/d x 9300 m2 x 0.75 x 0.95 = 143,800 g TOC day.
Using the gas production rate indicated by Figure 8 and a value of 50 percent
methane in the product gas yields
VCR /day = 143.8 Kg TOC/day x 1.76 m3/Kg TOC x 0.5 = 126.5 m3/day CH4.
Therefore, the potential energy, E , available from this methane is
E = 126.5 m3/day x 35,800 KJ/m3 = 4.529 x 106 KJ/day.
Assuming that 79 percent of the methane energy content can be converted to
useful heat, the net available energy becomes 3.57 x 106 KJ/day/reactor.
Energy input to the reactor will be required to turn both the discs and
to heat the wastewater to a temperature of 35°C. The 4 HP will require an
input of about 258,000 KJ/day, far below the energy equivalent of the methane
produced by the reactor. However, the energy required to heat the influent
wastewater to a reactor operating temperature of 35°C is dependent on both the
influent temperature and the concentration of biodegradeable TOC. For this
design example, the product of the flow rate and influent TOC concentration is
21.7 g/m2/day x 9300 m2 x 0.75 =151,358 g/day.
Thus, an influent concentration of 2000 mg/1 yields a flow rate of 75,679 I/day.
The minimum energy (Q) required to heat this flow for a 20° temperature increase
is
Q = 75,679 Kg/day x 4200 J/Kg-°Cx 20 °C = 6..357 x 106 KJ/day
Figure 18 demonstrates the interactions between influent substrate con-
centration, required temperature change and net energy requirements. Thfe
765
-------
solid portions of the curves represent the region where experimental evidence
is available. The dashed portions of these curves represent projections for
a well buffered wastewater. The vertical difference between any point on a
curve and the 3.6 x 106 KJ/day line represents the net energy yield or input
requirement for the reactor. The breakover points for energy equivalence for
5°, 10° and 20°C heating requirements appear to be around 800, 1700 and 3000
mg/1 TOC, respectively. The region below the horizontal line and above the
appropriate heating curve represents the potential energy recovery available
from hicjh strength wastewater treatment with the AnRBC process.
Example 2 - Lower Effluent Quality Required (68% Removal)
This example illustrates an AnRBC energy balance for conditions when a
lower effluent quality is acceptable. Based on the data obtained from Experi-
ment 6 (shown in Figures 7 and 8) and a 45 percent methane content in the
product gas, it can be shown that a loading of 33.4 g TOC/m2/day will yield
about 123.8 mVday of methane with the same area! safety factor used in the
previous example. Using the same procedure for estimating energy availability
as in the previous example, gives a useable energy yield from product methane
of about 3.32 x 106 KJ/day/reactor. Again, it is conservatively estimated that
4 HP of electrical energy will be required to rotate the shaft. Figure 19
illustrates substrate concentration and heating requirement effects on total
energy requirements. This family of curves has been shifted to the right
relative to those shown in Figure 18 due to the larger volumes of water that
must be heated.
CONCLUSIONS
Based on the data and discussion presented above, it can be concluded
that
1. The AnRBC treatment is both feasible and practical for high strength
wastewaters.
2. Microorganisms, including the difficult to culture methanogenic bac-
teria, will readily adhere to and grow on rotating surfaces.
3. Because reaction phases or components can be separated by feed strength
and flow rate adjustments, AnRBC pilot plant studies should prove
useful for describing the complex interactions and removal kinetics
for anaerobic systems. Similarly, AnRBC pilot plants should prove
advantageous for the study of presumed toxicants on anaerobic proc-
esses.
4. The AnRBC process appears to be ideal for the pre-treatment of high
strength wastewaters with energy recovery potentially available as a
result of the production and utilization of methane in the product
gas.
766
-------
TABLE I
AnRBC FEED*
CONSTITUENT
C11H22°11
MgS04 ' 7 H20
KC1
MgCL2
CaCl.
CoCl.
Fed.
6 H20
6 H20
4 H20
(NH4)C12
(NH4)2HP04
NaHCO-,
CONCENTRATION
Variable
60.0 rag/1
120.0 mg/1
300.0 mg/1
100.0 mg/1
14.2 mg/1
90.0 mg/1
**
**
***
BODr
TOC
= 1.72,
COD
TOC
= 2.79,
COD
BODr
= 1.62
* Tap water was used to provide trace nutrients.
** Amount added depended on influent carbon concentration.
The N and P concentrations were maintained in excess of
a C:N:P ratio of 150:15:1.
*** 3000 mg/1 for Experiments 1-4, 6000 mg/1 for Experiments 5-8.
767
-------
TABLE 2
AnRBC LOADING CONDITIONS. 12 RPM
Unit Hydraulic
Detention Time
(0), hr
17.50
8.75
4.39
2.19
Stage Hydraulic
Detention Time
(6 ), hr
S
4.38
2.19
1.09
0.55
Influent TOG Concentration, mg/1
]n75
1F-1H (1)
1F-2H (2)
1F-4H (3)
1F-8H (4)
2320
2F-1H (5)
2F-2H (6)
—
—
3050
3F-1H (7)
3F-2H (8)
—
—
Flow Rate
(Q), 1/hr
0.30
0.60
1.20
2.40
Experiment number indicated in parentheses.
768
-------
TABLE 3
CALCULATED FIRST STAGE K VALUES
Experiment No.
1
2
3
4
5
6
7
8
Mn*
g/day
6.18
10.2
8.35
14.1
12.2
15.7
7.98
9.75
Ci*
J5/1
1.05
1.05
1.10
1.10
2.32
2.32
3.05
3.05
K
1/g-day
5.89
9.67
7.58
12.8
5.25
6.75
2.62
3.20
LN
(0Ci)*
8.43
7.74
7.10
6.40
9.23
3.53
9.50
8.81
*Measured or the product of measured parameters.
769
-------
REFERENCES
1. Wilkey, R.C. and A.A. Friedman, "Response of Rotating Biological Contactors
to Shock Loadings," Proceedings Fifth Annual Environmental Engineering and
Science Conference, University of Louisville, Louisville, Ky., March 1975.
2. Friedman, A.A., R.C. Woods and R.C. Wilkey, "Kinetic Response of Rotating
Biological Contactors," Proc. 30th Industrial Waste Conference, Purdue
University, May 1976.
3. Friedman, A.A., L.E. Robbins, and R.C. Woods, "Effect of Rotational Speed
on RBC Efficiency," Proc. 33rd Industrial Waste Conference, Purdue
University, May 1978 - also J. Water Poln. Control Fed., Nov. 1979, p. 2678.
4. APHA, WPCF, AWWA, Standard Methods for the Examination of Water and Wastewater
13th Ed., 1976.
5. DiLallo, R. and O.E. Albertson, "Volatile Acid By Direct Titration," Journal
Water Pollution Control Federation, Vol. 33, p. 365, 1961.
6. Tait, S.J, and A.A. Friedman, "Anaerobic Rotating Biological Contactor
(AnRBC) Treatment for High Strength Carbonaceous Wastewaters," presented
at 52nd An. Conf. Water Pollution Control Federation, Houston, TX,
October 11, 1979.
7. Schroeder, E.D., Water and Wastewater Treatment, McGraw-Hill Book Company,
p. 307, 1977.
NOTE: At the time the research was conducted, S.J. Tait and A.A. Friedman
were, respectively, graduate student and Associate Professor, Depart-
ment of Civil Engineering, Syracuse University, Syracuse, New York.
S.J. Tait is currently a Senior Research Engineer, International
Paper Company, Tuxedo Park, New York. A patent is pending for the
AnRBC process.
770
-------
VARIABLE
SPEED INFLUENT
PUMP
GAS FLOW
METER
GAS
COOECTiON
MANIFOLD
DISC
(Ten per Stage)
BEARING
SHAFT
IMPELLER
MIXING
CHAMBER
i-SAMPLE
PORTS'
1ST
STAGE
2 ND
STAGE
3RD
STAGE
4TH
STAGE
DISC
FIGURE I
ANAEROBIC ROTATING BIOLOGICAL CONTACTOR
-------
X
Q.
(T>
1
i
00
i
O
\
1^-
!
C£
1
D
^
0 >x
ro
LU
CM
Jr
•yT
'O
O
O
I
8
in
\
c
(HOOO£HD)AlINilV>nV QDV 3HI1V10A
*DOI 3iam os
UJ
o:
UJ
o_
x
LU
CM
UJ
o:
z>
CD
Ii.
772
-------
Q =0.60 I/HR
Q=2320rngA
$ =8,75 HR
L = 33.41 g TOG/DAY
^\
X_
A-""
-A
\
8
PH
485 mg/I
(79% REMOVAL)-
FIGURE 3 EXPERIMENT 6t 2F-2H
-------
0
2000
0
r
\
-\
L = 63.35 g TOC/DAY
Q = 2.4 4/HR
8 = 2.I9HR
Cj= IIOOmg/1
8
PH
X
(46% REMOVAL)
1
IN
FIGURE 4
I 2
STAGE
EXPERIMENT 4, IF-SH
-------
EXP
NO.
01
02
A3
©4
a
(HR)
175
8.8
4,4
2.2
FIRST STAGE REMOVAL,
LB COO/IOOOFT2-DAY)
13.3
21.8
17.9
30.2
4/HR
IN
FIGURE 5
I 2
STAGE
EFFECT OF FLOW RATE ON An RBC TREATMENT
-------
300(7
EXR
NO.
0 i
D 5
A 7
Ci
(mg/tt
1050
2320
3050
FIRST STAGE .
LB COD/IOOOF"PDAY
13.3
26.3
17.2
Q = 0.30 J/HR
S = i7.5 HR
1000
I 2
STAGE
SUBSTRATE CONCENTRATION EFFECTS
-------
o
o
O
m
001 3180105 lN30U3d
O Q O
CM
1
o.
!
o
o
OJ
O
O
O
O
O
o
o
Q
CL
a.
8 ^
H ^
LJ
o:
o
o
LLl
CD
15
iw
LJ
Z>
'Q3AOIAl3d 001
777
-------
70
60
I 50
LU
•5 40
00
U.
<
H
Q
o
cr
a,
30
20
10
P = I ATMOSPHERE
T -
1 HI6H\AA
K K EXPA
EXV8 s
i
/
EXR6
AEXP.3
NEGLECTING EXR NOS. 4 a 8
GAS YIELD = 1.76 J/g TOC
R2 = 0.953
'ALL DATA
GAS YIELD =2.42 l/g TOC
R2 = 0.871
5 10 15 20 25 30
SOLUBLE TOC REMOVED, g/DAY
TOTAL GAS PRODUCTION
35
-------
779
-------
0.6
0.5
Q
_i
o
<
u
o
J_
< 0.3
CO
02
O.i
0.135 in eq +1.237
* 0.947
EXP.8
/DEXP7
YOBS = -0.113 In 0AC + 1.059
R2 - 0.877
I 1 I
8
In 0Cj' In 0AC
FIG, 10 OBSERVED YIELD AS A FUNCTION OF LOADING
CONDITIONS
780
-------
1.00
0.50
0.10
0.05
0.03
0.02
0.01
A
EXP. 1-4
C; = 1075 mg/l
Kp=-0.2830
7,8
3050 mg/l
1216
i i -H
EXP. 5,6
2320 mg/J
0.2052
NEGLECT
0 2 4 6 8 10 12 14 16 18 20 22
0, HR
FIGURE II SOLUBLE TOC REMAINING
781
-------
0,30
0.20
3 i
10 (T
X
0.10
KF-0.3744-7,9Sx!0-5Cj
R2~ 0.9773
KF = 1.2894-0.1432 In Cj
R2 = 0.9192
FIGURE 12
1000 2000
r., «v» AT
^J» l«iy/ +
KFDEPENDENCE ON Cj
3000
4000
-------
4OOO
I I
2000
= lnCj-(0.3744-79Sx
IO"5Cj)T
1000
~ 400
IA.J
330
t-LJ
m 100
3 80
10 60
40
30
©
• EXP. 1-4
AEXR5S6
"oEXR758
j i
0 2 4 6 8 10 12 14
T1ME,HR
FIGURE 13 PREDICTED TOC VALUES
16 18 20
783
-------
14
10
8
6
4
0
O
o
K = 29.91- 2.82 (In aC)
R2 = 0.8064
O
C =mg/l TOC
a =HRS
I
1
1
O
1
6.5
FIGURE 14
7.0
9.0
9.5
7.5 8.0 8.5
lneC,rng-HR/l
DEPENDENCE OF FIRST STAGE K VALUES
ON THE LOADING FACTOR
784
-------
30005
f
Q =0,3 I/HR
e * 175 HR
FIGURE 15 MASS TRANSPORT MODEL,
EXPERIMENTS 1,5,7
785
-------
300
2500
2000
1500
o
e
1000
O
CO
500-
234
STAGE
FIGURE 16 MASS TRANSPORT MODEL,
EXPERIMENTS 2,6,8
786
-------
1500
1000
E
«*•>
u
DJ 500
o
CO
0
IN
EXP.4
Q=2.4I/HR
=2.19 HR
EX P. 3
Q=i.2 J/HR
fl=4.375HR
1
1
2 3
STAGE
o
FIGURE 17 MASS TRANSPORT MODEL,
EXPERIMENTS 3 AND 4
787
-------
14
12
o
x
1*10
LU
S
LU
^
r>
a
yj
a:
o
a:
LOADING = 21.7g/m2/day
A = 9300m2
EXCESS ENERGY
REGION
Ql i i i i
J 1 I ,
0 5 10 15 20 25
INFLUENT BIODEGRADEA3LE TOC,g/£
FIGURE 18 INFLUENT SUBSTRATE CONCENTRATION
AND TEMPERATURE EFFECTS ON ENERGY
REQUIREMENTS ,95% REMOVAL
788
-------
X
SH
a
CO
z
llj
s
UJ
£=
ID
a
a: 4
CD
(T
LU
LU
o
LOADING = 33.4 g TOG/m2/day
A s 9300 m2
3.3xl06^KJ/day-
EXCESS ENERGY
REGION
05 10 15 20 25
INFLUENT BIO DEGRADE ABLE TOC,g/P
FIGURE 19 INFLUENT SUBSTRATE CONCENTRATION
AND TEMPERATURE EFFECTS ON ENERGY
REQUiREfv1ENTSs68c/o REMOVAL
789
------- |