&EPA
United States Municipal Environmental Research
Environmental Protection Laboratory
Agency Cincinnati OH 45268
EPA-600/2-84-106
June 1984
Research and Development
Design Information on
Rotating Biological
Contactors
-------
EPA-600/2-8*t-106
June 1984
DESIGN INFORMATION ON
ROTATING BIOLOGICAL CONTACTORS
by
Richard C. Brenner*
James A. Heidman
Edward J. Opatken
Albert C. Petrasek, Jr.
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
*Project Officer
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been subjected to the Agency's peer and administrative review
procedures and approved for publication as an EPA document. Mention of trade
names or commercial products does not constitute endorsement or recommendation
for use.
ii
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. The complexity of our environment and
the interplay of its components require a concentrated and integrated attack
on pollution problems.
Research and development is the necessary first step in problem solution and
involves defining the problem, measuring its impact, and searching for solu-
tions. The Municipal Environmental Research Laboratory develops new and
improved technology and systems to 1) prevent, treat, and manage wastewater
and solid and hazardous waste pollutant discharges from municipal and com-
munity sources, 2) preserve and treat public drinking water supplies, and 3)
minimize the adverse economic, social, health, and aesthetic effects of
pollution. This publication is one of the products of the Laboratory's
research program and provides a vital communications link between the re-
searcher, the design engineer, and the user community.
This inhouse project was undertaken to evaluate and summarize within one
document relevant design information pertaining to the rotating biological
contactor (RBC) process. The information presented herein is intended to
supplement, not replace, commonly-used RBC design tools such as manufacturers
design manuals and catalogs and design procedures published in the technical
literature. Factors affecting process selection and design, equipment per-
formance and reliability, and power consumption have been addressed. It is
believed this document will be of considerable interest to regulatory and
municipal officials, design engineers, and plant operating personnel.
Information was originally presented on the equipment and design practices
of six U.S. RBC manufacturers. One manufacturer, referred to hereinafter
as Manufacturer X, has recently stopped marketing RBC units. Descriptions
of Manufacturer X's equipment have subsequently been removed from this
document, but published references to and discussions of its previous
design methods have been retained where appropriate.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
111
-------
ABSTRACT
The relatively rapid introduction of rotating biological contactors (RBC's)
into the United States for municipal wastewater treatment has resulted in the
widespread application of a technology with which many design engineers are
not intimately familiar. Of necessity, many RBC designs initially were based
solely on proprietary-generated empirical design methodology. More recently,
as interest in the process has increased, more complex, deterministic design
approaches have begun to appear in the technical literature. The purpose of
this document is to review and assess existing RBC design procedures and
provide more in-depth design information on critical features of the RBC
process and key parameters affecting its operation and performance than is
typically available to the design engineering community.
The information contained in this document is intended to supplement and
qualify that available from RBC manufacturers and in the published literature.
Topics addressed include process and design considerations for carbonaceous
removal, nitrification, and denitrification; equipment reliability and
service life; power requirements for air and mechanical drive RBC units; and
general system design considerations involving structural, hydraulic, and
operational flexibility. A major priority in the preparation of the document
was given to emphasizing practical, usable design information as well as
important theoretical concepts.
Office and/or manufacturing facilities of most of the current U.S. proprietors
of RBC equipment were visited during the course of preparing this document.
Extensive discussions were held with all the manufacturers. The majority of
the data used in evaluating the RBC process were obtained from the technical
literature, various conference proceedings, and the files of the manufacturers.
Field studies were conducted in the greater Cincinnati area to supplement
existing power and air flow data for mechanically and air driven RBC's,
respectively.
This report was prepared in fulfillment of inhouse Task No. COO/38/E01/B15b
of Decision Unit B-113, Program Element CAZB1B. Rigorous literature review,
data evaluation, and technical writing were carried out from September 1,
1981, to October 4, 1982. The year-and-a-half since then have been devoted
to internal discussions and negotiations within the Agency on several issues
addressed in the document, generation of more reliable RBC mechanical drive
power data than were available in late 1982, and revisions and additions to
the document stimulated by extramural reviews. Final work was completed on
April 9, 1984.
iv
-------
CONTENTS
Section Pa
Foreword i i i
Abstract iv
Contents v
Figures x
Tables xiv
Acknowledgements xvi
Metric Conversions xvii
Introduction 1-1
1.1 Purpose 1-1
1.2 Brief History of RBC's 1-1
1.3 References 1-5
Process Considerations 2-1
2.1 Introduction 2-1
2.2 Wastewater Carbonaceous Characteristics 2-1
2.3 Biology 2-2
2.4 Mass Transfer and Kinetic Considerations 2-2
2.5 Mass and Hydraulic Loading Variations 2-7
2.5.1 Effect on Organic Removal Performance 2-7
2.5.2 Effect on Nitrification Performance 2-8
2.6 Scale-up 2-10
2.7 Organic Removal 2-14
2.7.1 Active Biomass 2-14
2.7.2 Loading Limitations 2-15
2.7.2.1 Beggiatoa 2-15
2.7.2.2 Oxygen Transfer 2-16
2.7.3 Additional Factors Affecting Organic
Removal 2-22
2.7.3.1 Tank Volume-to-Media Surface
Area Ratio (Detention Time) 2-22
2.7.3.2 Number of Stages 2-22
-------
CONTENTS (continued)
Section Page
2.7.3.3 Wastewater Temperature 2-23
2.7.3.4 Influent Wastewater Concentration 2-25
2.8 Nitrification 2-25
2.8.1 Overview 2-25
2.8.2 Nitrification Reactions 2-26
2.8.3 Nitrifying Biofilm Characteristics 2-27
2.8.4 Factors Affecting Nitrification 2-28
2.8.4.1 Growth Kinetics 2-28
2.8.4.2 Kinetic Models for Ammonia
Nitrogen Removal 2-29
2.8.4.3 Influent Nitrogen Composition 2-31
2.8.4.4 Wastewater Temperature 2-32
2.8.4.5 Dissolved Oxygen 2-35
2.8.4.6 Alkalinity and pH 2-36
2.8.4.7 Inhibition 2-40
2.8.5 Interdependency of Factors Affecting
Nitrification 2-41
2.8.5.1 Combined Kinetic Expression 2-41
2.8.5.2 Comparison of Pilot- and Full-
Scale RBC Nitrification Rates 2-41
2.9 Denitrification 2-43
2.9.1 Background 2-43
2.9.2 Kinetics 2-44
2.9.3 Previous Studies and Empirical
Design Formulations 2-46
2.10 Secondary Clarification 2-48
2.11 Sludge Production 2-49
2.12 References 2-49
3 Equipment 3-1
3.1 Introduction 3-1
3.2 General Equipment Description 3-2
3.3 Equipment Components 3-2
3.3.1 Shafts 3-2
3.3.2 Media 3-3
VI
-------
CONTENTS (continued)
Section
3.3.2.1 General Description 3-3
3.3.2.2 Configurations 3-5
3.3.2.3 Density 3-7
3.3.2.4 Assembly and Support 3-8
3.3.3 Drives 3-13
3.3.3.1 Mechanical Drive Option 3-13
3.3.3.2 Air Drive Option 3-15
3.3.4 Bearings 3-15
3.3.5 Instrumentation 3-15
3.3.5.1 Load Cells 3-15
3.3.5.2 Flow Control 3-17
3.3.5.3 Dissolved Oxygen Monitoring
and Control 3-18
3.4 Equipment Performance 3-20
3.4.1 Overall Structural and Mechanical
Reliability 3-20
3.4.1.1 Failures of the Shaft Component 3-22
3.4.1.2 Failures of the Media Component 3-23
3.4.1.3 Failures of the Drive Component 3-24
3.5 Evaluation of Shaft Design Procedures 3-25
3.5.1 Design Criteria and Codes 3-25
3.5.1.1 General Considerations 3-25
3.5.1.2 Fatigue Design Procedures 3-27
3.5.1.3 Service Conditions 3-29
3.5.1.4 Summary of Stress Category
Recommendations 3-30
3.5.2 Estimated Shaft Fatigue Performance 3-30
3.5.2.1 Design Parameters 3-30
3.5.2.2 Matrix of Estimated Shaft Lives 3-31
3.5.3 Summary of Conclusions 3-37
3.6 References 3-37
VI1
-------
CONTENTS (continued)
Section Page
4 Power Consumption 4-1
4.1 Introduction 4-1
4.2 Fundamentals 4-1
4.3 Air Drive Systems 4-4
4.3.1 Basic Relationships 4-4
4.3.2 Blower Design 4-8
4.3.3 Field Measurements of Air Flow
Versus Rotational Speed 4-11
4.4 Mechanical Drive Systems 4-15
4.4.1 Clean Media Power Measurements 4-15
4.4.2 Field Power Measurements 4-15
4.5 References 4-19
5 Design Considerations 5-1
5.1 Introduction 5-1
5.2 Mathematical Models 5-1
5.3 Pilot Plant Studies 5-2
5.4 Organic Removal Design 5-4
5.4.1 Comparison of Available Empirical
Design Approaches 5-4
5.4.2 Comparison of Selected Studies with
Empirical Design Approaches 5-9
5.4.3 An Alternative Design Approach
(Second-Order Kinetics) 5-11
5.5 Nitrification Design 5-19
5.5.1 Comparison of Available Empirical
Design Approaches 5-19
5.5.2 Analysis of Available RBC
Nitrification Data 5-26
5.5.2.1 Observed Pilot-Scale Removals 5-26
5.5.2.2 Observed Full-Scale Removals 5-28
5.5.2.3 Influence of Soluble Organics 5-34
5.5.2.4 Comparison of Meausred Versus
Predicted Values for RBC Combined
Carbon Oxidation-Nitrification 5-35
vm
-------
CONTENTS (continued)
Section Page
5.5.2.5 Comparison of Measured Versus
Predicted Values for RBC
Separate-Stage Nitrification 5-38
5.5.3 Clarification Following RBC Separate-Stage
Nitrification 5-39
5.5.4 Summary 5-40
5.6 Denitrification Design 5-42
5.6.1 Introduction 5-42
5.6.2 Design Curves 5-42
5.6.3 Performance Data 5-44
5.7 Design Case History for Combined Carbonaceous
and Nutrient Removal 5-45
5.8 General Plant Design Considerations 5-49
5.9 References 5-51
-------
FIGURES
Number Page
1-1 Photograph of Clow Mechanical Drive RBC System
at Upper Mill Creek Wastewater Treatment Plant,
Butler County, Ohio 1-3
1-2 Photograph of Autotrol Air Drive RBC System at
Lower East Fork - Little Miami River Regional
Wastewater Treatment Facility, Clermont County, Ohio 1-4
2-1 Relative Concentrations of Oxygen and Substrate for One
Hypothetical Loading Condition and RBC Rotational Speed
as a Function of Media Location . 2-4
2-2 Effect of Flow Shocks on RBC Nitrification Performance 2-9
2-3 Effect of Concentration Shocks on RBC Nitrification
Performance 2-9
2-4 Effect of Diurnal Flow and Concentration Variations on
RBC Nitrification Performance 2-11
2-5 Effect of Daily Influent Concentration Variations on RBC
Nitrification Performance at Constant Flow 2-12
2-6 DO Limiting Conditions Related to Influent Organic
Concentration and Hydraulic Loading 2-18
2-7 Estimate of RBC Oxygen Utilization Rates 2-21
2-8 Manufacturer Temperature Correction Factors for
BOD5 Removal 2-24
2-9 Second-Generation Autotrol Ammonia Nitrogen Removal
Rate Curve for Full-Scale RBC's 2-30
2-10 Manufacturer Temperature Correction Factors for
Nitrification 2-34
2-11 Effect of pH on RBC Nitrification Rates 2-38
-------
FIGURES (continued)
Number
2-12 RBC Process Configuration for Denitrification Using
Methanol Addition 2-45
2-13 RBC Process Configuration for Denitrification Using
Primary Effluent as the Carbon Source 2-45
2-14 RBC Denitrification Rate for No-Flow Conditions as
a Function of Bulk Liquid Nitrate Nitrogen Concentration 2-47
3-1 Cross-Sections of RBC Shafts 3-4
3-2 RBC Media Configurations 3-6
3-3 Autotrol' High and Standard Density Media 3-7
3-4 Clow Media Assembly 3-9
3-5 Crane-Cochrane Media Assembly 3-10
3-6 Lyco (Series 300) Media Assembly 3-11
3-7 Lyco/Hormel (Series 200) Media Assembly 3-12
3-8 RBC Mechanical Drive Assembly 3-14
3-9 Air Drive RBC Schematic 3-16
3-10 RBC Shaft Bearing Assembly 3-16
3-11 First-Stage Beggiatoa Growth at Columbus, Indiana
RBC Plant 3-19
3-12 Allowable Fatigue Stress Ranges for Three Stress
Categories of Redundant Structures in Atmospheric Service 3-26
4-1 Typical Centrifugal Compressor Horsepower Versus Air
Flow Curve 4-10
4-2 Typical Centrifugal Compressor Efficiency Versus Air
Flow Curve 4-10
4-3 Summary of Air Flow Versus Rotational Speed Measurements
Made at Lower East Fork - Little Miami River Regional
Wastewater Treatment Facility, Clermont County, Ohio 4-13
-------
FIGURES (continued)
Number Page
4-5 Measured Power Consumption for Mechanically Driven Units 4-18
5-1 Effluent BODs as a Function of Organic Loading
for Selected RBC Manufacturers Design Techniques 5-6
5-2 Organic Removal as a Function of Organic Loading
for Selected RBC Manufacturers Design Techniques 5-7
5-3 Autotrol Organic Removal Design Relationships for
Mechanical Drive RBC's 5-10
5-4 Disappearance of Soluble BODs with Time at Cleves 5-16
5-5 Disappearance of Soluble BODs with Time at Enumclaw 5-16
5-6 Disappearance of Soluble BODs with Time at Lancaster 5-16
5-7 Disappearance of Soluble BODs with Time at Lower
East Fork 5-17
5-8 Disappearance of Soluble BODs with Time at Woodburn 5-17
5-9 Disappearance of Soluble BODs with Time at Glenwood
Springs 5-17
5-10 Disappearance of Soluble BODs with Time at Dodgeville 5-18
5-11 Disappearance of Soluble BODs with Time at West Dundee 5-18
5-12 Disappearance of Soluble 6005 with Time at Hartford 5-18
5-13 Comparative RBC Nitrification Design Curves for an
Influent Ammonia Nitrogen Concentration of 10 mg/1 5-21
5-14 Comparative RBC Nitrification Design Curves for an
Influent Ammonia Nitrogen Concentration of 15 mg/1 5-22
5-15 Comparative RBC Nitrification Design Curves for an
Influent Ammonia Nitrogen Concentration of 20 mg/1 5-23
5-16 Comparative RBC Nitrification Design Curves for an
Influent Ammonia Nitrogen Concentration of 30 mg/1 5-24
5-17 Estimated Ammonia Nitrogen Removal Rates for Pilot-Scale,
RBC Separate-Stage Nitrification 5-27
XII
-------
FIGURES (continued)
Number Page
5-18 Estimated Ammonia NNitrogen Removal Rates for
Pilot-Scale, RBC Combined Carbon Oxidation-Nitrification 5-29
5-19 Full-Scale RBC Nitrification Rates at Design Wastewater
Temperature (55°F) 5-30
5-20 Full-Scale RBC Nitrification Rates at High Wastewater
Temperatures 5-31
5-21 Full-Scale RBC Nitrification Rates at Low Wastewater
Temperatures 5-33
5-22 Autotrol Nitrification Design Relationships 5-37
5-23 Autotrol Denitrification Design Relationships 5-43
5-24 Autotrol Temperature Correction Factors for
Denitrification 5-43
5-25 Process Flow Diagram for Orlando, Florida (Easterly)
Iron Bridge Water Pollution Control Facility 5-46
xm
-------
TABLES
Number
Page
2-1 First-Stage Biomass and Organic Removal Relationships 2-14
2-2 Organic Removal and Estimated Biological SRT's for
RBC Systems 2-15
2-3 First-Stage RBC Data Reported by Hynek and Chou 2-20
2-4 Alkaline Enhancement RBC Nitrification Data from
Stratta and Long 2-39
3-1 Shaft Characteristics 3-3
3-2 Structural and Mechanical Performance of RBC Plants 3-21
3-3 Recommended Stress Categories for RBC Tubular Shafts 3-31
3-4 Estimated Fatigue Life for 30-in. Diameter Circular
Shaft 3-32
3-5 Estimated Fatigue Life for 28-in. Diameter Circular
Shaft 3-33
3-6 Estimated Fatigue Life for 16-in. Square Shaft 3-34
3-7 Estimated Fatigue Life for 24-in. Octagonal Shaft 3-35
4-1 Comparison of Full-Load Current, Minimum Cooper Wire
Size, and Wire Costs for Various 7.5-hp Motor Options 4-2
4-2 Change in Power Factor and Motor Efficiency as a
Function of Load for 7.5-hp Polyphase Induction Motor 4-3
4-3 Results of RBC Mechanical Drive Power Measurements at
the LeSourdsville Wastewater Treatment Plant, Butler
County, Ohio 4-5
4-4 Results of RBC Mechanical Drive Power Measurements at
the Upper Mill Creek Wastewater Treatment Plant, Butler
County, Ohio 4-6
xiv
-------
TABLES (continued)
Number Page
4-5 Selected Characteristics of Lower East Fork and
Indian Creek Plants 4-12
4-6 Power Measurements for Mechanical Drive RBC Plants 4-16
5-1 Wastewater Parameters and Required Media Surface Areas
for Example RBC Organic Removal Design Problems 5-8
5-2 Comparison of Measured, Second-Order Predicted, and
Autotrol Design Manual Predicted Disappearance of SBODs
for Air Drive RBC Systems 5-13
5-3 Wastewater Parameters and Required Media Surface Areas
for Example RBC Nitrification Design Problems 5-25
5-4 Effect of Soluble BOD5 on RBC Nitrification Rates
at Gladstone, Michigan 5-35
5*5 Predicted Versus Measured Final Effluent Ammonia
Nitrogen for Gladstone, Michigan 5-36
5-6 Predicted Versus Measured Final Effluent Ammonia
Nitrogen for Indianapolis, indiana 5-39
5-7 Performance Data from Cadillac, Michigan Separate-Stage
RBC Nitrification Plant 5-40
5-8 RBC Denitrification Performance at Miyazaki City,
Japan (1977) 5-45
xv
-------
ACKNOWLEDGEMENTS
The four authors of this document were assisted In their efforts by a
diversified, six-member peer review panel that met twice in Cincinnati for
2 days each time. The purpose of the first meeting was to review and
modify as needed the initial approach and topic selections of the authors;
an in-depth critique of the first draft of the document was conducted at
the second meeting. A list of the panel members is given below.
For a technology as new as RBC's and considering the range of topics considered,
a number of differing opinions was anticipated and encountered among the authors
and peer panel reviewers concerning specific recommendations such as allowable
organic and nitrogen loadings, the number of stages required for a given level
of treatment, air flow requirements for air driven units, etc. Nonetheless,
this document does represent a consensus of opinion of the authors and reviewers
on many, but certainly not all, of the topics that were addressed.
Two consultants, Mark D. Bowman and John T. Gaunt of Purdue University, were
retained to perform structural analyses related to RBC shaft fatigue design.
This material is presented in the latter part of Section 3. In addition, Meyer
I. Landsberg, a self-employed consultant from St. Louis Park, Minnesota, was
retained to evaluate RBC polyethylene media properties. Excerpts from his report
are also presented in Section 3.
The invaluable assistance of Walter T. Schuk of MERL, EPA, Cincinnati in con-
ducting field air flow and power measurements on air and mechanical drive RBC
units, respectively, and constructing and calibrating an accurate air flow
measurement device is greatly appreciated. The initial leadership of Jeremiah
J. McCarthy, formerly of MERL and now with the U.S. Army 10th Medical Laboratory,
in forming the document team and establishing guidelines for carrying out the
project is gratefully acknowledged. Sincere appreciation is also extended
to ASCE's Task Committee on Rotating Biological Contactors, Water Pollution
Management Committee, Environmental Engineering Division, for their thorough
review comments.
Peer Review Panel
1. Dale A. DeCarlo, Burgess and Niple, Ltd., Columbus, Ohio
2. Donald F. Kincannon, Oklahoma State University, Stillwater, Oklahoma
3. F. Michael Saunders, Georgia Institute of Technology, Atlanta, Georgia
4. 0. Karl Scheible, HydroQual, Inc., Mahwah, New Jersey
5. Ed D. Smith, U.S. Army Construction Engineering Research Laboratory,
Champaign, Illinois
6. Richard A. Sullivan, Autotrol Corporation, Milwaukee, Wisconsin
xv1
-------
METRIC CONVERSIONS
Customary Unit
cfm
cu ft
OF
fpm
ft
gpd/sq ft
gpm
gpm
hp
hp/100,000 sq ft
in.
in.
ksi
Ib/cu ft
1b/day/1000 sq ft
mgd
mgd
psi
sq ft
ton (short)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
4.719 x TO'4
2.832 x TO'2
0.556(°F - 32)
5.08 x TO'3
0.3048
4.074 x ID'2
6.308 x TO'5
6.308 x ID'2
0.7457
8.027 x TO-5
25.4
2.54 x 104
7.031 x 105
16.02
4.883 x TO'3
3.785 x 103
4.381 x ID'2
7.031 x 102
9.29 x 10-2
9.072 x 102
Metric Unit
m3/sec
m/sec
m
m3/day/m2
m3/sec
I/sec
kW
kW/m2
mm
ym
kgf/m2
kg/m3
kg/day/m2
m3/day
m^/sec
kgf/m2
m2
kg
xvii
-------
SECTION 1
INTRODUCTION
1.1 Purpose
The primary purpose of this design information document is to provide design
information for rotating biological contactor (RBC) treatment of municipal
wastewater. Many aspects of this document are equally applicable to industrial
wastewater treatment. This document is intended to supplement commonly accepted
RBC design methodology, such as manufacturers design manuals, by providing
appropriate qualifiers and/or information not readily available to the design
community. Important design parameters and relationships (or lack of them) are
discussed in order to promote a more rational RBC design approach. Topics
addressed include process and design considerations for carbonaceous removal,
nitrification, and denitrification; equipment reliability and service life;
power requirements for air and mechanically driven units; and general system
design considerations involving structural and hydraulic flexibility. Practi-
cal, usable design information has been emphasized as well as theoretical con-
cepts. Hopefully, the information in this document will provide the design
engineer with a more in-depth perspective on key RBC design considerations than
is normally available in existing manuals and reports.
1.2 Brief History of RBC's
The first commercial RBC system was installed in West Germany in 1960. These
first RBC units were constructed with flat, 0.5-in. thick, 6.5- to 10-ft di-
ameter, expanded polystyrene sheets. The process became popular in West Ger-
many for small installations.
The J. Conrad Stengel in Company of West Germany licensed the RBC process to the
Allis-Chalmers Corporation for manufacturing and sales in the United States.
The business was subsequently sold to the Autotrol Corporation* in 1970. Mar-
keting efforts met with limited commercial success until the development of
*Autotrol Corporation is referred to throughout this document. On September
17, 1982, Envirex Inc., a subsidiary of Rexnord Inc., acquired certain assets
of and manufacturing and future worldwide marketing rights to Autotrol's line
of RBC products.
1-1
(1)
-------
thin, corrugated, high density polyethylene media by Autotrol in 1972. This
development increased the available surface area per unit shaft length by 70 to
150 percent, thereby increasing the cost effectiveness of the process in com-
parison with other economically-feasible secondary treatment alternatives.
Polyethylene or plastic media RBC systems are currently offered by a number of
proprietors.
At the time of this writing (September 1982), over 550 RBC plants were treating
municipal wastewater in the United States (1)(2)(3)(4)(5)(6). The process is
used worldwide for both municipal and industrial wastewater treatment. Approxi-
mately 70 percent of the RBC systems operating in the United States and Canada
are designed for organic carbon removal only, 25 percent for combined organic
removal and nitrification, and 5 percent for nitrification of secondary efflu-
ent (7). RBC systems have also been designed for upgrading trickling filter
plants and for denitrification with methanol addition (with completely sub-
merged media). Approximately 25 percent of the existing RBC municipal facili-
ties in the United States are package plants (8). The largest municipal RBC
facility is the 54-mgd system at Alexandria, Virginia. In-depth historical
reviews of RBC development are available (9)(10).
RBC media are rotated slowly (1 to 2 rpm) in a wastewater bath, with about 40
percent submerged at any given time. Media rotation is most commonly accom-
plished by mechanical drives, although air driven units are being employed at
some recently completed installations. Typical mechanical and air drive sys-
tems with individually covered RBC stages are shown in Figures 1-1 and 1-2,
respectively.
A distinguishing characteristic of partially-submerged, rotating plastic media
is the alternating exposure of the attached biofilm to air and wastewater.
This characteristic, though attractive to many designers, is what makes its
rational analysis difficult. Microorganisms respond to the environment sur-
rounding them, and, in the RBC, that environment is continually changing. The
requirement for movement of organics and nutrients from the liquid phase into
the biofilm and oxygen from the atmosphere into the liquid film, biofilm, and
bulk liquid makes it necessary to consider physical mass transfer as well as
microbial reaction rates when rationally analyzing RBC performance. By and
large, the design community has relied on the various RBC manufacturers to
provide the biological design procedures necessary to determine appropriate
sizing and staging of RBC systems. These procedures have been and continue to
be modified as more process performance information becomes available. As
interest in and experience with RBC systems has grown, the related technical
literature has also expanded providing alternative design methods and/or
approaches to those advocated by the equipment manufacturers.
Problems with the mechanical reliability of RBC systems have been encountered.
Premature shaft failures, stub end failures, and media separation/degradation
have all been experienced in some installations. Each manufacturer takes a
1-2
(2)
-------
CO I
^-^ CO
Figure 1-1. Photograph of Clow mechanical drive RBC system at Upper Mill
Creek Wastewater Treatment Plant, Butler County, Ohio.
-------
Figure 1-2. Photograph of Autotrol air drive RBC system at Lower East Fork -
Little Miami River Regional Wastewater Treatment Facility,
Clermont County, Ohio.
-------
somewhat different approach to media configuration; media fabrication, attach-
ment and support; and shaft design, and, as expected, there are differing opin-
ions as to the suitability of alternative equipment. Considerable proprietary
effort has been expended in assessing and correcting the equipment deficiencies
that have surfaced since the technology's introduction to North America. Since
RBC equipment designs are frequently updated, manufacturers' brochures and
other information need to be reviewed carefully to assess whether they are
current.
1.3 References
1. Personal communication from R. A. Sullivan, Autotrol Corporation, Milwau-
kee, Wisconsin, to E. J. Opatken, USEPA, Cincinnati, Ohio, September 30,
1982.
2. Personal communication from S. Banerjee, Clow Corporation, Florence, Ken-
tucky, to E. J. Opatken, USEPA, Cincinnati, Ohio, September 29, 1982.
3. Personal communication from R. W. Hankes, Crane Company, Cochrane Environ-
mental Systems, King of Prussia, Pennsylvania, to E. J. Opatken, USEPA,
Cincinnati, Ohio,
4. Personal communication from G. Ornstein, Lyco Division of Remsco Associ-
ates, Marlboro, New Jersey, to E. J. Opatken, USEPA, Cincinnati, Ohio,
September 30, 1982.
5. Personal communication from representative of Manufacturer X to E. J.
Opatken, USEPA, Cincinnati, Ohio, September 29, 1982.
6. Personal communication from G. W. Davis, Walker Process Corporation,
Aurora, Illinois, to E. J. Opatken, USEPA, Cincinnati, Ohio, September 30,
1982.
7. Hynek, R. J. and H. lemura. Nitrogen and Phosphorus Removal with Rotating
Contactors. In: Proceedings of the 1st National Symposium/Workshop on
Rotating Biological Contactor Technology, Vol. I, EPA-600/9-80-046a, NTIS
PB81-124539, June. pp. 295-314.
8. Chesner, W. H. and J. lannone. Review of Current RBC Performance and
Design Procedures. Report prepared for USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, under Contract No. 68-02-2775
by Roy F. Weston, Inc., (Publication pending).
9. Antonie, R. L. Fixed Biological Surfaces - Wastewater Treatment: The
Rotating Biological Contactor. CRC Press, Cleveland, Ohio, 1976.
10. Grieves, C. Dynamic and Steady State Models for the Rotating Biological
Disc Reactor. Ph.D. Thesis, Clemson University, Clemson, South Carolina,
1972.
1-5
(5)
-------
SECTION 2
PROCESS CONSIDERATIONS
2.1 Introduction
Procedures for predicting organic removal, nitrification, and denitrifica-
tion relationships occurring in the RBC process continue to evolve as addi-
tional information becomes available. Relationships that were recommended
a few years ago have subsequently been modified or replaced with alternative
methodologies. Reported experimental pilot and field data exhibit sufficient
variation such that support for conflicting design approaches can usually
be found. The present state-of-the-art is such that no single best design
procedure or set of relationships is universally applicable.
2.2 Wastewater Carbonaceous Characteristics
The biodegradable materials in municipal wastewaters are exceedingly diverse,
both with respect to number of components and range of particle sizes. If
wastewater is filtered through membrane or glass fiber filters (which nor-
mally allow passage of particles up to about 0.3 to 2.0 ym, depending on
the particular filter used), that material passing is generally defined as
soluble material; this operational definition of soluble material will be
used throughout this document. References to BODs will mean total 6005
unless specifically indicated to the contrary.
Primary effluent soluble BODs typically represents from 40 to 60 percent of
the BOD5 loading to a secondary RBC system. Where fine screening is used
in place of primary clarification, the soluble BODs component may comprise
as little as 30 to 40 percent of the BODs loading to an RBC system. Munici-
pal wastewaters receiving large and/or concentrated industrial waste flows
may have soluble BODs percentages that differ from these values. An RBC
system must not only remove the soluble wastewater components through a
combination of biological oxidation and cell synthesis, but must also agglom-
erate/bio-precipitate/biosorb/metabolize a substantial fraction of the
incoming particulate material if a clarified effluent of acceptable
secondary quality is to be achieved.
A tendency has evolved to focus on the removal of soluble material in RBC
systems since this material is readily measurable from stage to stage, is
2-1
(6)
r
-------
more easily incorporated into various mathematical modelling approaches,
and is a key design parameter. The designer should also be cognizant that
influent particulate BOD components can potentially undergo hydrolysis by
exocellular enzymes and contribute to the biodegradable organic loading.
In some cases, the particulate materials remaining after final clarification
exert far more influence on effluent quality than would be anticipated
based solely on observation of soluble BODs residuals (1).
2.3 Biology
The RBC biofilm surface is highly adsorptive, partially due to its polysac-
charide nature. The biological population that inhabits each stage of an
RBC train reflects the environmental and loading conditions specific to
that stage; consequently, a succession of microfauna develops from stage to
stage consistent with the decrease in organic loading to each succeeding
stage through the train (2)(3). Simple visual observation reflects a grada-
tion in slime thickness and color in staged systems and alone can often
provide an excellent indication of process performance. The first stage in
a system operating within the proper organic loading range exhibits a
characteristic brownish-grey color, and terminal stages that are nitrifying
normally have a characteristic reddish-bronze color (4).
The dominant growth in the first stage(s) frequently consists of zoogleal
and filamentous bacteria. The protozoan population increases in diversity
and abundance from stage to stage as the stage loading decreases. If systems
are not covered, algal growths may be expected to develop. Certain unde-
sirable microfauna such as the snails that flourish in some trickling
filter installations may also infest RBC systems (5). In lightly-loaded
stages, the growth may not be uniform and patches of bare media reflecting
predation by metazoa, such as bristle worms, may appear. In single-stage
systems, the predominant organisms reflect the overall unit loading with
Vorticella and other stalked ciliates frequently being the dominant proto-
zoan population when reactor soluble BOD5 is less than 15 mg/1 (6).
2.4 Mass Transfer and Kinetic Considerations
The overall performance of fixed film processes including the RBC is influ-
enced by both mass transfer and biological kinetic considerations. Labora-
tory studies by Kornegay and Andrews (7) with simple substrates demonstrated
that the substrate removal rate became constant after the biofilm reached a
certain thickness and that further increases in film thickness need not
result in increased rates of substrate removal under constant-defined feed
conditions. Increases in dissolved oxygen (DO) levels did not affect the
maximum removal rates in these studies. Whalen et al . (8) used a microelec-
trode to measure DO's in excess of 5 mg/1 in the slime layers developed
from a 20-mg/l nutrient broth feed. When the nutrient concentration was
2-2
(7)
-------
increased to 500 mg/1, the slime DO dropped from greater than 7 mg/1 in the
bulk liquid to 0.24 mg/1 at a 150-ym depth in the 175-um thick film.
According to Famularo et al. (9), mass transfer resistances associated with
both the liquid phase and biofilm result in significant concentration gradi-
ents from the bulk liquid to reaction sites and generally control system
performance.
Before presenting the equations used to describe RBC performance, it is
useful to qualitatively examine the changes occurring during media rotation.
These changes can be visualized by an examination of Figure 2-1. Relative
values of substrate and DO are shown for one hypothetical condition. These
relative values will, of course, vary for any particular set of design
conditions. When the media is exposed to the atmosphere, the liquid film
boundary at the air interface immediately becomes saturated with DO as
shown for Point A in Figure 2-1. This saturation in turn results in an
increase in the mass of oxygen that diffuses into the biofilm. When the
media is submerged, oxygen transfer can occur either into or out of the
biofilm depending on the bulk liquid DO levels and the degree of mixing of
the liquid film with the bulk liquid. Although it is convenient to assume
that the liquid film completely mixes with the bulk liquid when the media
enters the liquid from the atmosphere, the studies of Zeevelkink et al.
(10) show that this is not really the case. The substrate concentrations
within the biofilm are also shown to vary with position on the media, with
the point of maximum substrate penetration into the biofilm occurring at
the point of minimum DO level.
Organic and nitrogenous materials must be transported from the bulk liquid
to the biofilm surface. The incomplete mixing of the bulk liquid with the
liquid phase immediately adjacent to the biofilm surface indicates that ex-
ternal mass transfer resistance is an important consideration. Mathematical
models commonly assume that a stagnant film layer exists and that the flux
of material through this layer can be calculated as:
where J is the flux (m/t), AS is the difference in concentration between
the bulk liquid and the liquid film at the biofilm surface (m/1.3), AL is
the thickness of the stagnant film (L), A is the biofilm surface area (L^),
D is the diffusion coefficient of the component of interest (L^/t), and m,
L, and t denote dimensions of mass, length, and time, respectively.
An alternate approach is to utilize a mass transfer coefficient, k], that
incorporates the estimates of diffusion, film thickness, and any convective
mass transfer effects into one parameter and results in an equation as
follows:
2-3
(8)
-------
T
HI
oc
LLJ
I
D.
CO
O
Q
D
O
CO
1
SUBMERGED
SECTOR
POINT
D
/ RBC
MEDIA
LIQUID
FILM
Figure 2-1. Relative concentrations of oxygen and substrate for one
hypothetical loading condition and RBC rotational speed as a
function of media location.
2-4
(9)
-------
J =
where k] has dimensions of m/t.
ASA
(2-2)
For that portion of the rotational cycle where the media is in the air,
oxygen is transferred into the liquid film adhering to the biofilm surface.
The liquid film thickness is a function of rotational velocity and surface
structure. In actual practice, liquid film thickness is not uniform during
the air portion of the cycle since visual observation reveals a flow of
water across the media/biofilm surface as the media rotates through the
air.
The overall rate of reaction within the biofilm is determined by both the
intrinsic reaction rates associated with the biological metabolism of the
diverse number of constituents in the wastewater and the diffusion of oxygen
and substrates within the biofilm. Mass transfer of either substrate or DO
in the interior of the biofilm is customarily modeled by assuming that
Pick's law of diffusion is applicable (9)(11)(12)(13)(14). The flux at any
point can be calculated from Equation 2-1 or 2-2.
Frequently, the rate of substrate removal is modeled by an expression in
which the rate is assumed to be a function of the substrate level. For
example, if the substrate removal rate was controlled by substrate concen-
tration, the Monod expression would be:
= k
(2-3)
where X and S are the biomass and substrate concentrations (m/l_3), respec-
tively, k is the maximum substrate utilization rate (t"'), and Ks is the
half-velocity constant (m/L3). In other instances, it is more convenient
to utilize either a zero-order equation:
(2-4)
a first-order equation:
= -k'S
(2-5)
2-5
(10)
-------
or an nth order equation:
where k' is the reaction rate constant (dimensions vary with order of reac-
tion) to describe the unit substrate removal rate [(l/X)(dS/dt)] as a func-
tion of substrate concentration. If S is much larger than Ks in the Monod
expression, the unit substrate removal rate is essentially equal to k,
i.e., zero order. Conversely, if Kc is much larger than S, the unit substrate
removal rate is effectively [(k/Ks)(S)], i.e., first order. Thus, in many
instances, either zero-order or first-order equations are good descriptors
of unit substrate removal rates in the concentration ranges of interest.
For biofilms where oxygen can also limit the rates of reaction, such as a
thick or heavily loaded biofilm, an expression such as:
(2-7)
where KQ? is the half-saturation constant with respect to oxygen may be
applied Tn which the unit substrate removal rate is taken as a multiplica-
tive function of both substrate and DO concentrations. For a biological
reaction such as nitrification where NH3-N, oxygen, pH, and temperature arev
all important variables, a general unit substrate removal rate expression
may take the following form:
DO \ (K?n 0(T-20)) (A pH&) (2-8)
irw I ^
where K2Q®~ describes the nitrifier response over varying temperature
ranges and A pH^ is a generalized expression describing the influence of pH
on the nitrification rate.
If substrate removal within the biofilm is assumed to follow the Monod
relationship depicted in Equation 2-3, the biomass concentration is assumed
to remain constant with biofilm depth, and oxygen concentrations are not
limiting, the equation for diffusional mass transfer (Equation 2-1) can be
combined with that for substrate removal (Equation 2-3) to yield the follow-
ing steady state equation for substrate concentration within the biofilm:
(2-9)
2-6
(ID
-------
This nonlinear second-order differential equation does not possess an expli-
cit solution, although numerical techniques can readily be employed to
evaluate the process. An analytical solution can be obtained for either
zero-order or first-order biological kinetics. When the equations are
applied for substrate utilization in this form, it is assumed that the
electron acceptor is not rate limiting. Oxygen transfer becomes rate limit-
ing and controls the overall reaction rate in heavily loaded systems. The
studies of Williamson and McCarty (13) in a special laboratory reactor
confirmed that the mathematical approach assuming stagnant liquid layers
between the bulk liquid and biofilm surface and a diffusion model for move-
ment within the biofilm could accurately describe nitrification with bio-
films of various depth and where either the electron donor or acceptor was
flux and substrate limiting.
Mathematical models have been developed that consider substrate transport
and metabolism and oxygen transport and depletion within the biofilm as a
function of liquid film thickness, biofilm thickness, rotational speed,
mixing with the bulk liquid, etc. (9)(11). The relative rates of both
transport and diffusion must be considered in any generalized attempt to
construct deterministic mathematical models that accurately duplicate RBC
performance. An excellent discussion of the relative influence of these
rates is provided by Grady and Lim (11). When all of the factors affecting
substrate removal and oxygen transfer are considered, it is apparent that
there is no a priori reason to expect substrate removal from the bulk liq-
uid in an RBC system to follow any simple mathematical model.
2.5 Mass and Hydraulic Loading Variations
2.5.1 Effect on Organic Removal Performance
The design approach recommended in most situations by the RBC manufacturers
to account for the impact of loading variations on organic removal is to
utilize anticipated design average flows and incoming 8005 concentrations
to determine media requirements. According to several current design man-
uals (15)(16), these average conditions can be used for RBC design for
organic removal if the daily peak-to-average flow ratio is 2.5 or less,
irrespective of the incoming mass loading pattern. If the above ratio is
greater than 2.5, the manufacturers recommend the use of predicted peak
hourly or daily flow for design, depending on site-specific conditions, or
the provision of flow equalization.
Limited data showing RBC effluent quality changes as hydraulic and mass
loadings vary are available (1)(17). In the absence of pilot plant data
for a particular design in question, the basic recommendations of the RBC
proprietors offer a reasonable first approach in deciding whether to utilize
average or peak loading conditions for sizing media surface area to meet
specific effluent limitations.
2-7
(12)
-------
2.5.2 Effect on Nitrification Performance
Compared to heterotrophic microorganisms, nitrifying bacteria have extreme-
ly long generation times, with reported doubling periods of 8 to 17 hr
(18). External accumulation and/or internal storage of substrate for delay-
ed metabolism has not been shown to take place with nitrifiers as they do
with heterotrophs. Consequently, the impact of flow and mass loading varia-
tions is more severe on the nitrification efficiency of biological systems
in general than on their organic removal performance.
In assessing the effects of loading variations on nitrification perfor-
mance, the attached biofilm of an RBC system represents both an advantage
and a disadvantage compared to a suspended growth system. The incorporation
of the nitrifying microorganisms within an attached slime resists biomass
"washout" and gross loss of nitrifying capability during periods of hydrau-
lic surging such as intense rainstorms. In contrast, hydraulic surging can
displace significant fractions of activated sludge biomass to the receiving
water depending on sludge settling properties and the peak hydraulic capacity
of the secondary clarifier. While increased rates of biofilm sloughing and
decreased nitrification efficiency are to be expected during high hydraulic
throughputs, nitrifying RBC systems normally demonstrate rapid recovery
following termination of the high-throughput event.
During the surge event itself, however, whether encompassing hydraulic
overload, mass ammonia nitrogen overload, or both, RBC's and other fixed
film processes can be anticipated to exhibit poorer nitrification efficiency
than a suspended growth reactor, provided the activated sludge biomass is
not "washed out" of the system. This discrepancy is attributed to the
longer nominal detention times typically provided in activated sludge units
and the fact that fixed film processes are assumed to be mass transfer
limited while suspended growth processes are not.
Borchardt et al. (19) conducted a carefully-controlled laboratory comparison
of the effects of mass loading surges caused by hydraulic shock versus con-
centration shock on RBC nitrification performance employing a 2-ft diameter,
polystyrene, six-stage unit. The results of this experiment are summarized
in Figures 2-2 and 2-3, respectively.
Examination of these two figures reveals that the pilot RBC system responded
more favorably to the concentration surges than to the hydraulic surges.
Increases in effluent ammonia nitrogen levels were on the order of 10 to 50
percent less for the concentration shock runs. Time for complete recovery
to pre-shock performance levels, however, tended to be somewhat shorter
after cessation of the hydraulic shocks.
2-8
(13)
-------
DURATION OF FLOW SHOCK
O)
d
z
O
O
z
t
CO
I
LU
D
_i
LO-
LL
LU
NOTE:
Influent NH3-N: 20 mg/l
Normal Flow: 1.25 gpd/sq ft
Shock Flow: 2.5 gpd/sq ft
LEGEND:
O 1-hr shock
n 2-hr shock
A 4-hr shock
0 12345671
TIME (hr)
Figure 2-2. Effect of flow shocks on RBC nitrification performance [from
Borchardt et al. (19)].
DURATION OF STRENGTH SHOCK
A
D)
O
Z
O
O
CO
I
z
LU
D
_i
LL
LU
LU
NOTE:
Flow: 1.25 gpd/sq ft
Normal Strength of Influent: 20 mg/l NH3-N
Shock Strength of Influent: 40 mg/l NH3-N
LEGEND:
O 1-hr shock
n 2-hr shock
A 4-hr shock
0 12345678
TIME (hr)
Figure2-3. Effect of concentration shocks on RBC nitrification performance
[from Borchardt et al. (19)].
2-9
(14)
-------
For any given dry-weather diurnal flow pattern, variations in effluent con-
centration of unoxidized nitrogen from an RBC system will mirror variations
in influent concentration, but at a lower level (dictated by the average
removal or oxidation rate) and with a lag time roughly corresponding to
nominal reactor detention time (approximately 2 to 3 hr for a four- to six-
stage RBC reactor). This trailing pattern is readily evident in the influ-
ent and effluent TKN concentration data of Figure 2-4, adapted from Murphy
and Wilson's report (1) of studies carried out on a 19.7-in. diameter pilot
unit operated in the combined carbon oxidation-nitrification mode. Based
on these results, they recommend that 35 percent greater media surface area
be provided for a full-scale combined carbon oxidation-nitrification RBC
system operating with a daily peak-to-average-to-minimum flow ratio of
2:1:0.5 than for a system with the same design hydraulic loading where
prior flow equalization is provided.
The influent-effluent mirror pattern created by hourly variations in RBC
mass loading (Figure 2-4) extends to daily average RBC nitrification perfor-
mance as illustrated in Figure 2-5. These data were generated over a 3-1/2-mo
period for the City of Indianapolis (20) on a 10.5-ft diameter nitrifying
RBC pilot plant receiving biologically treated effluent as system feed.
Influent ammonia nitrogen concentration varied as shown, but influent flow
remained constant during selected phases at hydraulic loadings of 1.35 to
3.0 gpd/sq ft. Spikes in daily influent concentration were generally
matched by time-correlated but smaller spikes in effluent concentration.
Daily effluent ammonia nitrogen patterns from nitrifying activated sludge
systems tend to be smoother, reflecting less correlation to influent varia-
bility.
Consideration of loading variation impacts is not critical for RBC systems
designed to achieve intermediate levels of nitrification. To obtain consist-
ently low effluent ammonia nitrogen residuals (1 to 2 mg/1), however, the
aforegoing discussion strongly suggests that RBC nitrification designs
should provide adequate media surface area to compensate for expected flow
and concentration variability or should consider prior flow and mass equali-
zation.
2.6 Scale-up
A common RBC design practice in the past has been to scale up small pilot
plant results to full-scale applications by setting the media tip or peri-
pheral speed at the same value for both size units, i.e., normally about 60
fpm for mechanically driven units. In doing this, media surface area was
presumed to scale up directly as the key design parameter. More recent
studies have shown that smaller systems exhibit greater removal capacities
per unit surface area under conditions of high organic or hydraulic loading
and low DO concentration. Reh et al. (21) estimated that the oxygen trans-
fer capacity of their 18-in. diameter unit was about 1.6 times greater than
2-10
(15)
-------
TKN CONCENTRATION (mg/l)
ro
ca
c
CD
5. m
c?£
3 i±
0> 3
3, w
~"* ^
3 o
o
CD
S
a
"S
CD
m
O
^ 0)
O =f. u-
C 05
-^ -i
T3 55"
3" -*
^< 6'
0) 3
3 cn
Q-0
II
TKN LOAD (lb/day/1000 sq ft
HYDRAULIC LOADING
(gpd/sq ft)
ro
oo
ro
05
5- CD O ^
oo
ro
-b.
O)
ro
CO
O
O
ro
p
o>
ro oo
en
CD
O -1
bo CD
TI
o
T3
a.
-------
30
25
r 20
'•— ro
—> i
—i —i
-—- ro
<
OC
LU
O
z
o
o
15
10
LEGEND:
Influent
Effluent
•FEB
25
5
10
10 15 20 25 30
MAR
10 15 20 25
-APR-
5 10 15 20 25 30 5 10 15
•MAY-
•JUNE
1974
Figure2-5. Effect of daily influent concentration variations on RBC
nitrification performance at constant flow [from City of
Indianapolis (20)].
-------
that of a 10.5-ft diameter unit they evaluated. Murphy and Wilson (1)
reported an average 17-percent higher mass removal of COD per unit surface
area for a "1.6-ft diameter pilot RBC than for a 6.6-ft diameter unit opera-
ted at the same peripheral speed and recommended that 25 percent more sur-
face area be provided when extrapolating 1.6-ft RBC test data to full-scale
design and 10 percent more when extrapolating 6.6-ft RBC test data to full-
scale design. Pilot-scale RBC nitrification data also often exhibit ammonia
nitrogen removal rates 50 to 150 percent higher (19)(22)(23) than the com-
monly accepted full-scale removal rate of 0.3 Ib NH3-N/day/1000 sq ft (24).
The above discrepancy in removal capacities of different size RBC units may
be partially explained on the basis of "relative surface renewal". "Rela-
tive surface renewal" is defined as the ratio of rotational velocity for a
given unit to that of a standard 12-ft diameter unit when both are operated
at the same tip speed. Depending on diameter, the angular or rotational
velocity of a pilot RBC system may be several times that of a 12-ft diameter
field unit at identical tip speeds. For example, the rotational velocity
of a 4-ft diameter pilot unit is 4.8 rpm at the industry-standard tip speed
of 60 fpm compared to 1.6 rpm for a 12-ft diameter unit, yielding a "rela-
tive surface renewal" factor of 3.0. A "relative surface renewal" factor
greater than 1.0 results in more frequent exposure of a given element of
biofilin to the atmosphere where the majority of oxygen transfer is indicated
to take place in the RBC process (see Figure 2-1).
The high "relative surface renewal" factors associated with high rotational
velocities produce the following effects in an RBC: 1) less time for deple-
tion of substrate in the biofilm during the atmospheric portion of the
cycle, 2) less time for depletion of DO in the biofilm during the submerged
portion of the cycle, 3) increased turbulence at the tank air-wastewater
interface, 4) lifting of thicker films of wastewater into the atmosphere,
and 5) greater flow or draining of wastewater over the media and attached
biofilm back to the tank during the atmospheric portion of the cycle. The
latter three effects all promote higher oxygen transfer rates into the
biofilm. Because there is less time for substrate depletion, the average
substrate concentration in the biofilm remains higher throughout the atmos-
pheric portion of the cycle. The higher average concentration creates a
higher demand for oxygen at the wastewater film-biofilm interface, thereby
increasing the oxygen transfer driving force into and across the wastewater
film. Increased oxygen transfer and reduced opportunities for substrate
and DO depletion undoubtedly contribute substantially to the higher substrate
removal rates typically observed in RBC pilot plants as compared with full-
scale RBC systems.
Insufficient information is presently available to accurately predict appro-
priate scale-up factors for pilot-generated RBC data (25). Very small-scale
pilot units (1.5-ft diameter or less) are more useful in determining the
basic treatability of a wastewater than in establishing full-scale design
parameters. However, if small-diameter units must be operated to collect
design data, it is important that each stage be loaded below the oxygen
transfer capability of a full-scale unit to minimize scale-up considerations.
2-13
(18)
-------
2.7 Organic Removal
2.7.1 Active Biomass
Data relating organic loading to the first stage of a full-scale RBC system
and the resulting soluble BODs removals and biofilm thicknesses are shown
in Table 2-1. The biomass in this table was conservatively estimated by
assuming a 5-percent solids concentration with a specific gravity of 1.0
for the biomass component of the attached biofilm. The generally low solu-
ble BODs removals obtained at the relatively low calculated food-to-micro-
organism (F/M) loadings (based on total BODs applied) illustrate that mass
transfer of DO and substrate are more important in controlling first-stage
removals than microbial kinetic considerations.
The active biomass depth has been estimated by a number of investigators at
between 20 and 600 ym (0.0008 and 0.024 in.). The shaggy biomass surface
makes the definition of the active depth somewhat imprecise. Nonetheless,
when these depths are compared to the biofilm thicknesses shown in Table
2-1, it is clear that the first stage(s) in a highly loaded RBC system
consist of large amounts of biomass that are not contributing to the removal
of organic materials in the influent wastewater. If it is assumed that the
biofilm reaches an equilibrium condition, i.e., the total film thickness
does not continually increase, it is possible to grossly estimate the proba-
ble range of solids retention times (SRT's) for the organisms in the rela-
TABLE 2-1. FIRST-STAGE BIOMASS AND ORGANIC REMOVAL RELATIONSHIPS*
Run
V-A
V-B
VI-A
VI-C
VII
VIII
IX
Biofilm Thickness
(in.)
0.057
0.062
0.080
0.057
0.075
0.110
0.113
Biomass/1000 sq ft
(lb)
14.82
16.12
20.8
14.82
19.5
28.6
29.38
F/M
Loading
(day-1)
0.60
0.56
0.26
0.48
0.60
0.32
0.76
Soluble BODs
Removal
(%)
57
52
57
59
22
21
34
*Taken from Hynek and Chou (26).
2-14
(19)
-------
tively thin biomass layer responsible for substrate removal from the bulk
liquid. Such estimates are presented in Table 2-2 where the ranges of
percent solids in the biofilm, yields, and active layer thicknesses used in
the SRT calculations reflect estimated values based on several literature
citations (7)(8)(9)(17)(27)(28).
Although the calculations summarized in Table 2-2 can only approximate
actual conditions, they do illustrate several important points. At the
highest removal rate shown, the average SRT of the biomass layer actually
removing substrate from the bulk liquid is estimated to be between 0.25 and
1.5 days. Particulate BOD that becomes attached to biomass has a short
average time to be degraded before the biomass particle to which it is
attached leaves the biofilm and reenters the bulk liquid. It is not until
the removal rate drops to between 0.5 and 1.5 Ib soluble BODs/day/lOOO sq
ft that the SRT begins to reach the range that would be required for the
growth of nitrifying organisms. These calculations indicate why heavily
loaded RBC systems do not nitrify and also that in lightly loaded stages
(<0.5 Ib soluble BODs/day/lOOO sq ft) any particulate BOD that becomes at-
tached to the biofilm should have a sufficient average residence time in
the biomass to exert a substantial oxygen demand.
2.7.2 Loading Limitations
2.7.2.1 Beggiatoa
Where DO is depleted within the biofilm, sulfate reduction and/or anaerobic
decomposition of the solids can occur. Anaerobic decomposition products
TABLE 2-2. ORGANIC REMOVAL AND ESTIMATED BIOLOGICAL SRT's FOR RBC SYSTEMS
Range of Conditions Considered
in SRT Calculations
Soluble BODs
Removal Rate
for any Stage
(lb/day/1000 sq ft)
Estimated SRT
of Active
Biomass Layer
(days)
% Solids
in Biofilm
Yield
g SS
g(BOD5)R
Active
Layer
Thickness
(ym)
2.5
1.5
0.5
0.1
0.25- 1.5
0.49- 3.2
2.0 -13.6
10-153
5-10
5-10
5-10
5-10
0.8-1.0
0.8-1.0
0.6-1.0
0.4-1.0
60-150
70-180
100-200
100-300
2-15
(20)
-------
can diffuse back from deep within the biofilm into the active biofilm depth,
in effect producing an additional oxygen demand from the media side. When
sulfate is reduced, the relative proportions of HS~ and H£S existing in
equilibrium are pH dependent. At pH 7.0, 50 percent will exist as HS~ and
50 percent as H2S for the temperatures and conductivities associated with
municipal wastewater. Overall, each mole of sulfate that is reduced to
sulfide and then reoxidized and deposited as elemental sulfur releases a
net 1.5 moles 02-
When sulfide is present, either
duction deep within the biofilm
Beggiatoa can and frequently do
tions of autotrophy in Beggiato
concentrations on growth rate a
grow readily as a heterotroph o
number of strains require H2S f
granules within the cell in the
Beggiatoa compete with heterotr
media surface. Their predomina
tration of biomass on an RBC un
tial reduction in organic remov
takeover of the first stage of
shift the load to the next stag
over of the entire system and s
In the absence of a biological
by Beggiatoa, an organically ov
the maximum substrate removal r
by oxygen transfer into the bio
ing organisms become predominan
occurs within the biofilm inter
expected to fall from the maxim
rate.
2.7.2.2 Oxygen Transf
Oxygen is transferred directly
water film and biofilm on that
atmosphere. In addition, oxyge
fer due to turbulence generated
to the bulk liquid of the attac
lifted into the atmospheric sec
Where organic removals in the f
been limited by DO availability
oxygen content in the gas phase
in the influent wastewater or by its pro-
sulfide oxidizing organisms such as
grow on the biofilm surface (29). The ques-
and the relative influences of H2S and DO
e not yet clarified. The organism will
a number of dilute organic substances. A
r growth (30). Beggiatoa deposit sulfur
presence of oxygen.
phic organisms for oxygen and space on RBC
ce can result in an increase in the concen-
t while at the same time causing a substan-
1 per unit area. In extreme cases, the
n overloaded RBC system by Beggiatoa can
leading to a progressive Beggiatoa take-
gnificant deterioration of effluent quality.
ulfide oxidation problem such as produced
rloaded RBC stage may well be operating at
te possible. The rate will be controlled
ilm. As Beggiatoa or other sulfide oxidiz-
and/or excessive anaerobic metabolism
or, the overall organic removal rate may be
m possible bulk liquid substrate removal
rom the atmosphere into the attached waste-
ortion of the RBC media exposed to the
enters the bulk liquid from direct trans-
by rotation of the media and by the return
ed wastewater film and other wastewater
or that flows freely across the media.
rst stage of heavily loaded systems have
it has been shown that increasing the
will increase the organic removal rate
2-16
(21)
-------
(31)(32)(33). Gas transfer from the air directly into the attached waste-
water and biofilms represents the major oxygen source for the organisms.
Some researchers have attempted to develop procedures to predict oxygen
transfer into the bulk liquid and to relate these transfer coefficients to
system geometry. Using clean water test data, Severin et al. (34) were
able to relate a dimensionless oxygen transfer parameter to a Reynolds
number based on net water velocity and tank parameters. Scale-up based on
tip speed or surface area turnover was shown to be inadequate. Increasing
tip speed from 1.2 to 1.6 rpm increased oxygen transfer from 2.3 to 3.8 g
02/min with the 66-sheet, 11.7-ft diameter media packs employed.
Measurements of oxygen transfer coefficients for gas transfer into the RBC
bulk liquid do not provide a meaningful or useful indicator of system poten-
tial. If oxygen transfer from the gaseous phase into the wastewater film
with subsequent diffusion into the biofilm is much more important than
transfer from the bulk liquid into the biofilm as suggested by Huang (35),
then bulk liquid oxygen transfer coefficients are not a useful measure of
oxygen transfer occurring in an RBC system. While low bulk liquid DO's or
falling DO's from stage to stage are indicative of an overloaded system
(36), increasing DO levels in the bulk liquid will not necessarily overcome
process performance difficulties that may develop. Two of the plants stud-
ied by Chesner and lannone (36) had nuisance organisms (Beggiatoa) present
in the first stages despite bulk liquid DO levels of 1.5 to 2 mg/1. A
similar observation was also made at an RBC facility in Edgewater, New
Jersey (17). A major advantage of the air drive RBC system is claimed to
be related to the increased turbulence and stripping of excess biofilm as a
portion of the air bubbles rise through the media. This aspect may be of
substantially more benefit to system performance in heavily loaded systems
than small increases in bulk liquid DO levels afforded by the air drive
mechanism.
A major constraint in the design of any RBC system is limiting the loading
to the first stage(s) to values compatible with the oxygen transfer capa-
bility of the system. Exceeding this capability will often result in the
proliferation of sulfide oxidizing organisms that leads to overall process
deterioration. Chesner and lannone (36) surveyed 23 RBC installations and
related the presence of sulfide oxidizing organisms to overloading caused
by high hydraulic loading and/or high influent BOD concentrations. Their
survey results are summarized in Figure 2-6. It can be seen that a first-
stage loading limit in excess of 6.4 Ib total BODs/day/lOOO sq ft was associ-
ated with the presence of sulfide oxidizing organisms. This loading should
correspond to a soluble BOD5 loading in the range of 2.6 to 3.8 lb/day/
1000 sq ft.
According to the Autotrol design manual (15), the first-stage loading limit
for a mechanically driven shaft is 4 Ib soluble BODs/day/lOOO sq ft. Cases
2-17
(22)
-------
5UU
400
S^"
D)
Z
o
1-
<
|E 300
z
LLJ
O
z
o
o
IT)
Q
O
CD
z 200
in
D
LL
z
CD
OC
100
L-J O£
LEGEND:
O Plants reporting absence
of nuisance organism
growths
n Plants reporting
problems
'. organism
with nuisance
growths
\ —•— Represents an organic
loading breakpoint of 6.4
\ 34Q Ib BOD5/day/1000sq ft
\
\ D33
\NOTE: Numbers correspond to
plant identifications
r provided by Chesner and
', lannone (36).
\
.
\
\ D36
\ 35
\ D n5
10r\ D3°
60 . 31
\D °9
30 v
140 • n D15
016 """^-^
2004 °13 ^^ -^
i i i i i i
29 D
^ 8
i i i
4 6 8 10 12 14 16 18
FIRST STAGE HYDRAULIC LOADING (gpd/sq ft)
Figure 2-6. DO limiting conditions related to influent organic concentration
and hydraulic loading [from Chesner and lannone (36)].
20
2-18
(23)
-------
have been reported, however, where RBC first-stage oxygen transfer capabili-
ties were exceeded above loadings of 3 Ib soluble BOD5/day/1000 sq ft (37).
Supplemental aeration of the bulk liquid is reported to increase allowable
loadings (37), but it does not assure elimination of nuisance organisms
even in the presence of adequate bulk liquid DO as established by the
observations of Chesner and lannone (36). It would be expected, however,
that the dominance of nuisance organisms would tend to decrease with
increasing bulk liquid DO levels. Supplemental aeration studies conducted
by Chou (38) using mechanically driven units, clean media, and tap water
indicated that the oxygen transfer efficiency for the coarse bubble system
installed was 2 to 2.5 percent.
f
The Autotrol design manual (15) also indicates that zero-order removal of
soluble BOD5 at a maximum rate of 2.5 lb/day/1000 sq ft is possible for
mechanically driven units and 3 lb/day/1000 sq ft may be achieved with air
driven units. In the zero-order removal range, BOD removal is controlled by
oxygen diffusion (15). Consider the first-stage data obtained by Hynek and
Chou (26) using 11.8-ft diameter discs (12,500 sq ft/stage) to treat munici-
pal wastewater at Milwaukee County's South Shore Wastewater Treatment Plant
and summarized in Table 2-3. A maximum soluble 6005 removal of about 2.6
lb/day/1000 sq ft was achieved. It is clear, however, that soluble BODs
removal was not'constant, nor was it a single-valued function of bulk liquid
soluble BOD5 concentration. The periods of lowest removal were partially
related to periods of greatest biofilm thickness.
Soluble COD removal data obtained at the overloaded Brookville, Indiana RBC
plant by Opatken (39) indicated zero-order removal of soluble COD in each
of the four stages at a rate of 2.5 lb/day/1000 sq ft. This removal rate
is about one-half the expected maximum rate (assuming a typical soluble
COD:soluble BODs ratio of 2); however, this plant is routinely plagued by a
proliferation of Beggiatoa growth. Use of a clarigester (sludge storage)
for pretreatment at Brookville may contribute sulfide to the RBC reactor
influent.
The total COD reduction across an RBC stage is a direct measure of oxygen
transfer capability, provided that nitrate or sulfate reduction is not
occurring and stage influent and effluent DO levels are the same. Full-scale
measurements of total COD removal across a stage are virtually impossible
to find in the literature. Scheible and Novak (40) measured total COD
removal and the COD equivalent of the settled solids to estimate the total
oxygen utilization rate. Their results are presented in Figure 2-7. These
data were obtained with full-scale media on 13.5-ft long shafts and indicate
a maximum oxygen utilization rate of 1.4 to 1.5 Ib 02/day/1000 sq ft with
this rate limitation also reflecting the maximum oxygen transfer capability
of the system.
2-19
(24)
-------
TABLE 2-3. FIRST-STAGE RBC DATA REPORTED BY HYNEK AND CHOU (26)
ffi
no
Run No.
V-A
V-B
VI-A
VI-B
VI-C
VI!
VIII
IX
X
Waste-
water
Temp.
(OF)
53
59
73
64
66
65
63
57
54
Total BODs
App 1 i ed
/ Ib/day \ /
\\(W sq ft/ \
8.85
8.98
5.50
-
7.08
11.7
9.16
22.4
_
Mechanical
App I i ed
Ib/day \ /
Drive Soluble
Removed
Ib/day \ in
1000 sq ttj ^1000 sq ft/
4.25
2.86
2.52
1.29
2.50
4.20
3.54
6.26
5.54
2.41
1.5
1.44
0.76
1.47
0.94
0.76
2.14
1.78
BOD5
Reactor
(n.g/1)
41
40
31
16
32
45
40
44
53
Total BOOs
Applied
/ Ib/day \
\ 1000 sq ft/
10.0
10.2
5.39
-
7.45
12.1
9.34
23.4
-
Air
Appl ied
/ Ib/day
\ 1000 sq f t
4.82
3.25
2.47
1.34
2.63
4.31
3.61
6.53
5.65
Drive Soluble
Removed
\ / Ib/day
) \IOOO sq ft
2.59
1.93
1.65
0.93
1.42
1.16
1.13
2.53
1.59
BOD5
)in Reactor
(mg/1)
44
34
24
12
36
43
35
41
56
-------
<
N
I-
D
LLJ
O
X
o
DC —-
O ^
-I CT
< in
II
UJ >
tr sT
Q 5
O £
O ~
U.
O
UJ
I-
cr
Q
LLJ
I-
o
D
LEGEND:
O 3/22-4/6/77
D 4/11-5/13/77
A 5/23-6/30/77
® 7/18-9/25/77
• 12/1/77-2/24/78
TCOD REMOVAL
RATE
ESTIMATED O2 UTILIZATION
RATE (MODEL PREDICTION)
A
ESTIMATED O2 UTILIZATION
RATE (COD BALANCE)
0 2 4 6 8 10
TCOD LOADING RATE (lb/day/1000 sq ft )
Figure 2-7. Estimate of RBC oxygen utilization rates [from Scheible and Novak (40)].
2-21
(26)
-------
In summary, an organically overloaded RBC stage can lead to heavy biofilm
thickness and/or proliferation of nuisance organisms such as Beggiatoa that
result in a net decrease in organic removal across the stage. When design-
ing a mechanically driven RBC system, research and field observations
indicate that safe, conservative loading figures for first-stage loading
are 2.5 Ib soluble BOD5/day/sq ft or 6 Ib total BOD5/day/1000 sq ft. These
figures may be increased, but the designer must recognize that this increase
may lead to heaver-than-normal biofilm growth, bulk liquid DO depletion,
development of nuisance organisms, deterioration of overall process and/or
mechanical performance, etc. A first-stage loading exceeding the above
figures may be justified depending on the degree of operational and mainte-
nance attention the plant will receive, the structural capacity of the
selected shaft, the ability to strip excess biomass from the media, the
levels of sulfur compounds in the RBC system influent, the media surface
area required in the remaining stages, and the ability to vary the opera-
tional mode of the plant. The loading, however, should generally not
exceed 4 Ib soluble BOD5/day/1000 sq ft or 8 Ib total BOD5/day/1000 sq ft.
2.7.3 Additional Factors Affecting Organic Removal
In addition to the performance variations observed when treating the same
wastewater with different diameter units at equivalent loadings, a number
of other factors affect the performance of RBC systems and influence the
results observed. Some of these factors are discussed below.
2.7.3.1 Tank Volume-to-Media Surface Area Ratio (Detention Time)
Based largely on work with 2- to 6-ft diameter RBC pilot units, Antonie (4)
observed that increases in the tank liquid volume-to-media surface area
ratio beyond 0.12 gal/sq ft did not increase percent BOD removal at a given
hydraulic loading rate (gpd/sq ft of media). Use of different tank con-
figurations and tank volume-to-media surface area ratios will result in
differences in oxygen transfer into the bulk liquid, the degree of mixing,
the amount of dead space in the tank, and hydraulic retention time at a
given flow rate, all of which can affect a comparison of the results obtained
from various pilot- and full-scale studies.
2.7.3.2 Number of Stages
Studies conducted by Antonie (4) with 5.74-ft diameter discs clearly demon-
strated that a four-stage RBC unit produced higher percentage BOD and sus-
pended solids removals than obtained with a two-stage unit having the same
overall surface area when treating the same municipal wastewater over a
hydraulic loading range of 1 to 5 gpd/sq ft. For hydraulic loadings below 5
gpd/sq ft, four- and six-stage operation produced virtually identical per-
formance. Soluble organic loading data were not collected during these studies,
2-22
(27)
-------
For any kinetic order higher than zero, overall carbonaceous removal for a
given media surface area will be enhanced by increasing the number of stages.
When selecting the number of stages for an RBC system design, however, the
primary consideration is to ensure that the organic loading to any individual
stage is not excessive, i.e., not greater than 2.5 Ib soluble BODs/day/lOOO
sq ft. If the design selection calls for a one-stage system at a low over-
all organic loading, there is no reason to automatically add several addition-
al stages. In reality, stage selection is an integral part of the overall
design process and the staging recommendations provided in the manufacturers'
design manuals should be used with discretion. The current staging recommen-
dations of three of the manufacturcturers are summarized below.
Autotrol (15)
Target
Effluent
Soluble BOD5
(mg/D
>25
15-25
10-15
Recommended
Minimum
No. of
Stages
1
1 or 2
2 or 3
3 or 4
Clow (16)
At least
four stages
per flow path
recommended
Lyco (41)
Target
Total
Reduction
up to 40%
35 to 65%
60 to 85%
80 to 95%
Recommended
No. of
Stages
1
2
3
4
Minimum of four stages
recommended for combined
and NH3-N removal
2.7.3.3 Wastewater Temperature
Studies reported by Antonie (4) indicated that organic removal efficiency
was unaffected by wastewater temperatures above 55°F, but process perfor-
mance deteriorated at lower temperatures. The temperature correction curve
developed by Autotrol (6) assumed that a theta (9) value of 1.05 should be
applied below 55°F since this value is commonly associated with biological
systems. However, because the organic removal rate in an RBC system is con-
trolled by mass transfer rates of oxygen and substrate in addition to bio-
logical kinetics, the utilization of a temperature correction factor based
solely on biological kinetic considerations is questionable. Below about
50°F, variations in the temperature correction factors recommended by the
manufacturers for carbonaceous removal become increasingly pronounced with
decreasing temperature as illustrated in Figure 2-8 (42). The manufacturers'
surface area correction factors are referenced to a wastewater temperature
of 55°F.
In some highly loaded RBC systems, lower wastewater temperatures do not
always result in decreased carbonaceous removal rates and, in fact, may
2-23
(28)
-------
CO
c
CD
cio
3D &
O o
"< c
CD CD
213
O T3
3 CD
>-- CD
O
O
CD
O
o
»-l-
o
-^
en
CD
O
a
01
CD
3
o
CD
(6Z)
n-z
SURFACE AREA CORRECTION FACTOR
-U
o
CO
tn
en
£ o
m
-U.
^sl
CJl
m
33
H
m
^
T>
m
33
c
33
m
3! tn
o
en
ro
tn
tn
tn
I I
ii
m
o
00
8
CD O.
x o
o
-------
enhance removals. When this phenomenon occurs, it can be attributed to in-
creasing DO saturation values with decreasing temperature, which promote
increased oxygen transfer, and possible reduction in the concentration of
sulfide oxidizing organisms. For example, Scheible and Novak (40) obtained
the following essentially equivalent overall performance data under summer
and winter conditions with 12-ft diameter RBC units:
Parameter Summer Winter
Hydraulic loading (gpd/sq ft) 2.1 2.0
Organic loading (Ib total BODs/day/lOOO sq ft) 2.3 2.6
Wastewater temperature (°F) 79 52
Effluent total BOD5 (mg/1) 28 33
Effluent soluble BODs (mg/1) 23 24
Others (6) have reported increased removals in wintertime conditions for
heavily loaded systems, such as at Alexandria, Virginia.
2.7.3.4 Influent Wastewater Concentration
In general, percentage BOD removal for municipal wastewater increases at a
given hydraulic loading as the wastewater strength increases. For high
strength industrial wastewaters, this often is not the case. These general
trends have been reported by the ASCE RBC Task Committee (43) and Chesner
and lannone (36). Thus, a comparison of systems on the basis of percent
removals where different wastewater strengths are involved can be misleading.
2.8 Nitrification
2.8.1 Overview
The internally-staged configuration of an RBC system, which promotes sequen-
tial or plug flow removal of substrate, is conducive to the development of a
nitrifying bacterial population. The degree of this development in any
stage depends primarily on the soluble organic concentration in the stage
bulk liquid. Population dynamics dictate that heterotrophs, i.e., bacteria
that derive their energy from the oxidation of organic carbon, will predomi-
nate in an RBC biofilm when the organic concentration is high and the aver-
age SRT is low. As the organic concentration decreases to a level where the
growth rate of the nitrifiers is greater than the rate of active biofilm
sloughing, the percentage of nitrifying bacteria in the biofilm will increase
to a point where efficient nitrification is possible. In the RBC process,
the organic concentration transition point where incipient nitrification is
generally observed is approximately 15 mg/1 of soluble BODs or 30 mg/1 of
total BOD5 in the stage bulk liquid (15)(16)(41)(44).
2-25
(30)
-------
In combined carbon oxidation-nitrification RBC systems for treatment of
orimary effluent, a small fraction of the incoming unoxidized nitrogen (10
to 20 percent) is utilized in the lead stages to satisfy the cell growth re-
quirements of the predominant heterotrophic population. For municipal waste-
water applications, the soluble BODs concentration is typically reduced to
15 mg/1 or less by the second or third stage. Depending on other environment-
al conditions, principally wastewater temperature, DO, and pH, significant
nitrification generally begins in the third or fourth stage. From these
stages on, disappearance of ammonia nitrogen is due mainly to oxidation to
nitrate nitrogen rather than to heterotrophic metabolism. The onset of
nitrification is noted by the appearance of nitrite and/or nitrate nitrogen
in the stage liquid. In most systems, the nitrification rate in the transi-
tion stage is less than in the subsequent stage where soluble BOD5 decreases
to 10 mg/1 or less. If sufficient stages are provided to achieve low ammonia
nitrogen residuals (1 to 2 mg/1), the nitrification rate in the final stage(s)
will be markedly reduced because of ammonia nitrogen limitations.
RBC's are also utilized for nitrification of secondary-quality effluent.
Separate-stage systems of this type (not to be confused with the internal
stages of an RBC reactor) are designed to protect the nitrifying population
from major competition with heterotrophs. In this application, the soluble
BOD concentration entering the second-stage RBC reactor is sufficiently low
that maximum nitrification rates are usually established on the lead RBC
shaft. First-stage options employed ahead of the RBC unit include a variety
of activated sludge and attached growth processes. Separate-stage RBC nitri-
fication is generally employed as an add-on step to existing biological
treatment systems that are faced with meeting an ammonia nitrogen or TKN
effluent limitation.
2..S.2 Nitrification Reactions
Nitrification is an autotrophic process, i.e., energy for bacterial growth
is derived by the oxidation of inorganic nitrogen compounds, primarily ammonia
nitrogen. In contrast to heterotrophs, nitrifiers utilize carbon dioxide
(inorganic carbon) rather than organic carbon for synthesis of new cells.
Nitrifier cell yield per unit of substrate metabolized is many times smaller
than for heterotrophs. This factor accounts for the small percentage of
nitrifiers in a biofilm developed on a substrate high in organic carbon.
Although a variety of nitrifying organisms exist in nature, the two genera
of importance in wastewater treatment are Nitrosomonas and Nitrobacter.
Nitrosomonas oxidize ammonia nitrogen to nitrite nitrogen. Nitrite nitrogen
is subsequently oxidized to nitrate nitrogen by Nitrobacter.
Excellent reviews of the energy and synthesis relationships associated with
biological nitrification are presented in the EPA Process Design Manual for
Nitrogen Control (45) and elsewhere (46)(47)(48)(49). Overall oxidation and
synthesis-oxidation reactions are developed in these documents.
2-26
(31)
-------
Based on the overall oxidation reaction, theoretical alkalinity destruction
in the nitrification process is 7.14 mg (as CaC03)/mg NH3-N oxidized. If
the minor effect of cell synthesis is considered, this value decreases to
7.07 mg (as CaC03)/mg NH3-N oxidized. The more conservative ratio of 7.14
is recommended for engineering calculations (45).
Similarly, based solely on the overall oxidation reaction, the theoretical
oxygen requirement for nitrification is equal to 4.57 mg 02/mg NH3-N oxi-
dized. This theoretical requirement decreases to 4.19 mg 02/mg NH3-N oxi-
dized if both oxidation and synthesis are considered. Again, a conservative
ratio of 4.6 is recommended for engineering calculations (45).
2.8.3 Nitrifying Biofilm Characteristics
Stratta and Long (50) in pilot studies conducted primarily to determine the
effect of pH on RBC nitrification also carried out correlative investigations
of biofilm development and microbial enumeration. Their 19.7-in. diameter
pilot reactors treated sidestreams of clarified trickling filter effluent with
a soluble BODs concentration that varied between 5 and 10 mg/1. Biofilm char-
acteristics reported by the investigators are considered to be representative
of typical separate-stage RBC nitrification.
Within 5 to 10 days after startup from clean media, Stratta and Long observed
the development of thin, highly uniform textured coatings of biofilm, tan to
bronze in color. Film thickness increased (and color darkened) with aging,
reaching equilibrium depths after 25 to 60 days of operation. Initial slough-
ing was noticed within 2 to 3 wk, with texture beginning to lose its uniform-
ity and becoming increasingly patchy with time. Maximum ammonia nitrogen
oxidation rates were established within 3 to 4 wk at 58°F, increasing no fur-
ther even though biofilm depth continued to increase in most cases for several
weeks.
Stratta and Long estimated the active depth of mature nitrifying RBC biofilm
at 50 to 300 vim (0.002 to 0.012 in.), or about half of the maximum active
depth cited earlier for heterotrophic growth in Section 2.7.1. Crawford's
estimates (22) of overall biofilm thickness for an eight-shaft nitrifying
system varied from 900 ym (0.035 in.) on the first shaft to 300 ym (0.012
in.) on the eight shaft, suggesting as with heterotrophic growth that a sub-
stantial fraction of the attached biomass in a nitrifying RBC is not active.
Based on the methodologies employed, the concentrations of heterotrophs and
ammonia nitrogen oxidizers were projected by Stratta and Long to be roughly of
the same order of magnitude, i.e., around 1 x 10^ bacteria/dry volatile gram
(dvg) of biofilm. The concentration of nitrite nitrogen oxidizers was approxi-
mately an order of magnitude less, i.e., around 1 x lO'O bacteria/dvg or
2-27
(32)
-------
slightly lower. Nitrifiers were enumerated by the MPN technique and hetero-
trophs by plate counts. These data reflect considerably higher concentrations
than Ito and Matsuo's work (51), which enumerated 2 x 10^ Nitrobacter/dry
gram (dg), 2 x 10? Nitrosomonas/dg, and 2 x 10^ heterotrophs/dg in nitrifying
RBC biofilm.
2.8.4 Factors Affecting Nitrification
2.8.4.1 Growth Kinetics
Nitrification growth kinetics and substrate utilization have been described by
various methods. The most frequently used method is the following Monod ex-
pression:
y =
(2-10)
where y is the growth rate of either Nitrosomonas or Nitrobacter (day1), ft
is the maximum growth rate of either Nitrosomonas or Nitrobacter (day1), Ks
is the half-saturation constant (mg/1), i.e., the substrate (ammonia nitrogen)
concentration at half the maximum growth rate, and S is the growth-limiting
substrate (ammonia nitrogen) concentration (mg/1). The above equation assumes
no mass transfer or oxygen transfer limitations; nitrifier growth rate is
controlled only by the prevailing nitrogen substrate concentration and the
microorganism's ability to metabolize or oxidize that substrate.
Values of Ks found in the literature for Nitrosomonas and Nitrobacter are
typically equal to or less than 1 mg/1 at liquid temperatures of 68°F or less
(45), although Downing and coworkers (52)(53) have reported a Ks of 1.25 mg/1
for Nitrobacter in river water at 68°F. The following estimates of ft for
Nitrosomonas and Nitrobacter can also be found in the literature:
ft Values (day1)
Nitrosomas Nitrobacter
1.08
0.37
0.85
0.7
0.3
0.5
0.71
0.57
0.40
1.44
1.1
Liquid
Medium Temp.(°F) Ref.
Activated sludge 73 54
Activated sludge 73 55
Activated sludge (washout) 70 49
River water 68 52,53
Activated sludge 68 52,53
Activated sludge 68 56
Activated sludge 68 57
Activated sludge 61 58
Activated sludge (washout) 54 49
2-28
(33)
-------
Most researchers have found ft and Ks to increase with increasing liquid tem-
perature. Estimates of ft and Ks for attached growth biological treatment
processes are conspicuously absent in the literature, but would be expected to
approximate those of suspended growth cultures where mass transfer of ammonia
nitrogen and DO are not limiting.
The above data generally indicate that the maximum growth rate of Nitrobacter
is somewhat larger than that of Nitrosomonas, and, therefore, the oxidation of
ammonia nitrogen to nitrite nitrogen is the rate-limiting reaction in nitrifi-
cation. This conclusion is supported by the lack of accumulation of nitrite
nitrogen in mature nitrifying systems at concentrations encountered in munici-
pal wastewaters. The low reported Ks values for both organisms suggest that
nitrifiers are reproducing at or near their maximum growth rates at nitrogen
concentrations equal to or greater than 1 mg/1 in the absence of limiting
factors such as low DO, low pH, etc. (59).
2.8.4.2 Kinetic Models for Ammonia Nitrogen Removal
Several investigators have modified the Monod growth equation to predict rates
of ammonia removal (oxidation) in the RBC process. Pano et al. (23) adapted
the Monod equation to the mass removal rate of nitrogen for nitrification data
generated on 15-in. diameter, polyethylene-disc RBC pilot plants treating the
equivalent of primary effluent:
where Z is the mass removal rate of ammonia nitrogen (lb/day/1000 sq ft), kN
is the ammonia nitrogen maximum removal rate (lb/day/1000 sq ft), KM is the
ammonia nitrogen removal half-saturation constant (mg/1), and C-j is the ammo-
nia nitrogen concentration in the ith stage (mg/1). This equation substitutes
ammonia removal rate terms for the nitrifier growth rate terms of Equation 2-
10. The linear regression curves (discussed further in Section 5) fitted by
Pano et al. to their data indicated higher removal rates at increased waste-
water temperatures.
Pilot plant studies conducted with 4-ft diameter polystyrene RBC discs by
Borchardt et al. (19) on trickling filter effluent also produced a Monod-type
prediction of ammonia nitrogen removal rate. Their data, which will also be
discussed further in Section 5, could not be segregated into separate curve
fits at different temperature levels.
Early literature citations by Antonie (4) suggested that ammonia nitrogen
removal followed a first-order kinetic relationship through at least four
stages of RBC media. Based on further pilot plant testing at Madison, Wiscon-
sin, and Guelph, Ontario, and correlation with full-scale performance data,
Autotrol revised its nitrification design basis in the late 1970's (24). The
2-29
(34)
-------
updated design procedure projects first-order removal (oxidation) of ammonia
nitrogen at concentrations in the stage liquid below about 5 mg/1. Above 5
mg/1 NH3-N, removal is claimed to proceed at a maximum zero-order rate con-
sistant with the prevailing wastewater temperature and other environmental
factors. Applied to a given RBC system, this design basis assumes that, if
soluble BOD5 is below 15 mg/1, incoming ammonia nitrogen will be oxidized at a
constant (zero-order) rate as it passes through succeeding stages down to
about 5 mg/1 and thereafter at'a first-order rate. Autotrol's updated curve
for full-scale RBC's (24), plotted in Figure 2-9, exhibits a zero-order re-
moval rate above 5 mg/1 NH3-N of approximately 0.3 Ib NH3-N oxidized/day/1000
sq ft at a wastewater temperature of 55°F.
Studies undertaken at the Canadian Wastewater Technology Centre (60) investi-
gated zero-order, half-order, and first-order kinetic expressions for describ-
ing nitrification data developed on a 1.6 ft-diameter polyethylene disc pilot
unit. The zero-order model was selected as best fitting the data, with a
constant TKN removal rate of 0.22 lb/day/1000 sq ft at TKN concentrations
above 4 mg/1.
The literature citations of different kinetic models as best describing the
oxidation of ammonia nitrogen in the RBC process are not necessarily inconsist-
ent. As noted in Section 2.4, if S in the Monod expression is much larger
than Ks, substrate (in this case ammonia nitrogen) removal will approach a
zero-order rate. Conversely, as substrate is removed and S approaches the
value of or becomes smaller than Ks, further substrate removal will begin to
approximate a first-order rate. The limits at which these transitions occur
in a given nitrifying RBC depend on biology, scale, environmental conditions,
mass transfer, and other factors.
0.4
LU
ol^ 0.3
08
UJ >.
DC "3
"O
I
•z.
0.2
0.1
0
WASTEWATER TEMPERATURE = 55° F
1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0
STAGE NH3-N CONCENTRATION (mg/l)
13.5 15.0
Figure 2-9. Second-generation Autotrol ammonia nitrogen removal rate curve
for full-scale RBC's [from Antonie (24)].
2-30
(35)
-------
2.8.4.3 Influent Nitrogen Composition
RBC nitrification designs are often based on influent ammonia nitrogen concen-
tration rather than influent unfiltered or soluble TKN concentration. For
nitrification of biologically treated effluents, this approach is justified
and should present no sizing problems. In combined carbon oxidation-nitrifi-
cation applications, however, such an approach could lead to gross undersizing
of the required media surface area in the tail-end nitrification portion of
the reactor.
Total unoxidized nitrogen is determined by the TKN procedure. TKN in turn is
composed of soluble and particulate fractions. The soluble fraction contains
urea and other organic nitrogen compounds as well as the inorganic ammonia
nitrogen component.
In raw wastewater, it is not unusual for unfiltered TKN-to-soluble TKN ratios
to reach 2.0 or greater. Following primary clarification, the particulate TKN
fraction is normally reduced to a few mg/1. At this point in a plant treat-
ment train, it is still possible for the soluble TKN concentration to be much
larger than the ammonia nitrogen concentration. This situation occurs when,
for whatever reasons, large quantities of urea have not yet hydrolyzed to
ammonia nitrogen. Prior hydrolysis depends on a number of factors, including
temperature and holdup time in the sewer system. When primary effluent enters
a biological reactor, however, the hydrolysis is usually rapidly completed,
adding to the ammonia nitrogen load available for nitrification.
Consider the following primary effluent composition data for August 13 and 14,
1975, for the full-scale, six-stage RBC system at Gladstone, Michigan (61):
Parameter (mg/1) August 13 August 14
Unfiltered TKN 28.0 24.8
Soluble TKN 23.0 21.5
NH3-N 9.2 17.7
N03-N 0.6 0.0
On August 14, the soluble TKN concentration in Gladstone's primary effluent
was only 3.8 mg/1 higher than the ammonia nitrogen concentration. On the
preceding day, however, the difference in these two concentrations was 13.8
mg/1. In this example, the use of primary effluent ammonia nitrogen data
alone in establishing required media surface area for RBC nitrification would
obviously constitute a dangerous design practice and could result in severe
undersizing.
2-31
(36)
-------
A procedure recommended by Autotrol for designing combined carbon oxidation-
nitrification RBC systems is to assume that particulate TKN will pass reason-
ably unchanged through the process and that all the influent soluble TKN minus
some loss for heterotrophic metabolism will be available for nitrification
before exiting the reactor. Based on analysis of numerous data sets, Sullivan
(6) has suggested using a subtraction factor of 0.10 x soluble 8005 removed as
an estimate of the amount of nitrogen consumed in heterotrophic metabolism.
For separate-stage RBC nitrification, hydrolysis of soluble TKN to ammonia
nitrogen will have been essentially completed by the time first-stage effluent
enters the RBC reactor. In this case, incoming soluble TKN and ammonia nitro-
gen concentrations should be approximately the same.
A less empirical approach is to assume that all influent total TKN, excepting
a Small refractory portion, will become available for metabolism and nitrifi-
cation in an RBC system through hydrolysis and other enzymatic reactions.
Randtke et al. (62) found that soluble organic nitrogen (SON) in the effluent
of four full-scale activated sludge plants treating municipal wastewaters
averaged 1.5 mg/1. Approximately two-thirds of this amount was determined to
be refractory organic materials present in the untreated wastewaters, and one-
third was generated biologically during treatment. An approximate protoplasm
composition of C]Q6Hl80045^16Pl> which represents 9.2 percent nitrogen, has
been proposed by Barth and Bunch (63). Combining this percentage with a con-
servative yield factor of 0.6 x soluble BOD5 removed yields a nitrogen sub-
traction factor for heterotrophic metabolism of 0.055 x soluble 6005 removed.
With this approach then, media requirements for RBC nitrification would be
based on an equivalent ammonia nitrogen concentration (mg/1) equal to influent
total TKN (mg/1) minus 1.0 mg/1 refractory SON minus 0.055 x soluble BODs
removed (mg/1). Whatever procedure is utilized, it is essential that the
equivalent ammonia nitrogen load an RBC sysem is designed to nitrify includes
the anticipated unoxidized nitrogen recycle loads from sludge digestion and
conditioning processes.
2.8.4.4 Wastewater Temperature
It is universally acknowledged that decreasing wastewater temperature adverse-
ly affects the growth rates of nitrifying organisms and their corresponding
capabilities for carrying out cxidative reactions. Differences of opinion
exist concerning the magnitude of the temperature impact over the temperature
range of interest in municipal wastewater treatment.
Downing and coworkers (52)(53) developed the following temperature relation-
ships for the maximum growth rate, ^N, and the half-saturation constant, Kj\|,
of pure cultures of Nitrosomonas:
2-32
(37)
-------
^N = 0.47e[(0-098(T - 15)], mg/i as N (2-12)
KN = io(0.051 T - 1.158), day-l (2-13)
where T is the wastewater temperature (°C).
Using a 2-ft diameter RBC pilot plant fed synthetic wastes in a laboratory
environment, Borchardt et al. (19) developed a temperature curve that predicts
an approximate doubling in the nitrification rate for every 18°F rise in waste-
water temperature between 50 and 86°F. Pilot-scale RBC work at Utah State
University (23) resulted in a projected value of 1.103 for the nitrification
temperature coefficient, 9|\j, between 59 and 68°F. Their pilot reactor opera-
ting at 41°F produced no nitrification. Mueller et al. (64) chose a 9|\j value
of 1.10 in calibrating a model developed to represent RBC nitrification.
Murphy and Wilson (1) conducted RBC pilot evaluations of combined carbon oxi-
dation-nitrification using 1.6-ft and 6.6-ft diameter polyethylene media.
They established design loading criteria based on influent soluble TKN for an
effluent unfiltered TKN goal of 5 mg/1 as a function of wastewater tempera-
ture. Their design premise assumed an influent total BODs-to-soluble TKN
ratio of 6. Final recommended loading criteria in Ib soluble TKN/day/1000 sq
ft ranged from 0.05 at 41°F to 0.08 at 50°F to 0.13 at 59°F to 0.23 at 68°F,
or a 4.6-fold increase from the lowest to the highest temperature considered.
Nitrification temperature correction curves developed by four RBC manufactur-
ers, Autotrol, Clow, Lyco, and Manufacturer X, are compared in Figure 2-10 (42),
(Published curves were not available for Walker Process and Crane-Cochrane.)
The correction factors shown refer to the additional media surface area re-
quired to achieve equivalent nitrification efficiency at the indicated tem-
perature as at 55°F. All of the curves are very similar in the 45 to 55°F
range. Lyco's recommended correction factor increases precipitously from 45°F
down to 42.5°F and is double the recommended Clow and Autotrol corrections at
the lower temperature.
None of the manufacturers recommend decreasing the quantity of media surface
area provided for nitrification at wastewater temperatures above 55°F. In
light of the reported evidence that nitrification rates in pure cultures and
pilot-scale RBC systems increase with increasing wastewater temperatures up to
75 to 85°F, the manufacturers' temperature correction curves suggest that
above 55°F, full-scale RBC nitrification rates become increasingly dominated
by environmental factors other than temperature, effectively masking the tem-
perature effect.
2-33
(38)
-------
5r
OJ
i
LEGEWD:
——• Autotrol
Clow
——— Lyco
—— Manufacturer X
DC
O
h-
o
LL
I 3
o
LU
DC
DC
O
O
HI
DC
<
LU
O
<
LL
DC
\
42.5
45.0 47.5 50.0
WASTEWATER TEMPERATURE (°F)
52.5
55.0
Figure 2-10. Manufacturer temperature correction factors for nitrification [from
Roy F. Weston (42)].
-------
2.8.4.5 Dissolved Oxygen
Nitrification is comprised of aerobic reactions that are generally acknowledged
as being more sensitive to DO concentration than heterotrophic reactions. A
minimum desired DO level of 2 mg/1 is often quoted. Cases can be found in the
literature, however, where only 40 to 90 percent of the maximum nitrification
rate was achieved at 2 mg/1 DO (65)(66)(67).
The Nitrogen Control Process Design Manual (45) suggests that the effect of DO
on the growth rate (day~^) of Nitrosomonas, UN can be represented by the
following Monod-type expression, if it is assumed that DO is a growth-limiting
substrate:
where Kg? is the half-saturation constant for oxygen (mg/1). Kg? values rang-
ing from 0.15 mg/1 at 59°F (68) to 2.0 mg/1 at 68°F (69) have been proposed.
A commonly-used value developed by the British (65) is 1.3 mg/1 (temperature
unspecified). Based on Equation 2-14 and an assumed KQ- of 1.3 mg/1, a DO
level of at least 4 mg/1 would have to be maintained to achieve 75 percent of
the maximum growth rate of Nitrosomonas.
Borchardt et al . (19) reporting on six-stage, 4-ft diameter RBC nitrification
of trickling filter effluent found that a DO level of 3.5 mg/1 was necessary
in the first three stages to ensure no retardation in oxygen uptake rates.
Mass removal rates of ammonia nitrogen in the last three stages of a six-
stage, full-scale U.S. Army RBC installation (70) were observed to average 50
percent higher in January (55°F) at stage bulk liquid DO concentrations of 3.5
to 4.8 mg/1 than in August (79°F) at DO's of 1.3 to 2.2 mg/1, although possible
low pH inhibition of nitrification in both months weakens the case for a direct
comparison of removal based on DO levels alone.
If low bulk liquid DO levels are encountered in the nitrification stages of an
RBC, transient intrusion of increased levels of soluble BOD into those stages
should be investigated as a contributing cause. This situation is more likely
to be encountered in combined carbon oxidation-nitrification RBC systems than
in separate-stage nitrification reactors.
Decreased nitrification rates in stages experiencing modest organic influx
could be due as much or more to inadequate DO in the stage bulk liquid than to
nitrifier competition from heterotrophic growth. In these cases, provisions
for correcting the DO deficiency on a temporary basis by increasing media
rotational speed or providing supplemental air may represent a viable solu-
tion. Hydrogen peroxide, while capable of increasing DO, tends to be toxic to
nitrifiers and should not be added to nitrifying RBC stages even on a temporary
2-35
(40)
-------
basis. Air drive RBC units would appear to offer a means of eliminating or
greatly reducing the potential for DO deficiencies in nitrifying RBC stages.
Since air drives are normally rotated at only 65 to 75 percent of the standard
1.6-rpm angular velocity of mechanical drives, however, the beneficial effects
of the oxygen in the compressed air are largely offset by lower media oxygena-
tion capacity (6).
If an RBC nitrification system is sized to produce low effluent ammonia nitro-
gen residuals, i.e., 2 mg/1 or less, the increase in DO concentration at the
exit end of the train will normally become pronounced. Effluent DO's for the
six-stage Gladstone, Michigan, municipal plant routinely reach 6 to 8 mg/1
with corresponding ammonia nitrogen concentrations of 1 to 3 mg/1 (71).
High effluent-end DO concentrations combined with very low levels of soluble
organic material can reportedly lead to deterioration of nitrification rates
through proliferation of higher life forms that ingest nitrifying microorganisms.
Pilot studies conducted on separate-stage, 10.5-ft diameter nitrification
units for the City of Indianapolis (20) correlated reduced nitrification effi-
ciencies with large microscopically-observed quantities of rotifiers, nematodes,
and other bacterial predators. A 5- to 8-percent flow spike of primary effluent
to the lead RBC stage retarded predator activity, and ammonia nitrogen removal
rates increased an average 10 percent for the 10 days after the spike.
To discourage selective predation of nitrifying bacteria, Sullivan (6) recom-
mends maintaining stage bulk liquid DO concentrations of no more than 3.5 mg/1
and preventing soluble BOD5 from dropping below 6 to 8 mg/1 in the polishing
stages of an RBC nitrification train. Incorporating operations flexibility in
the design of an RBC facility in the form of variable-speed drives, a supple-
mental air system, and/or multiple, individually-valved feed points to the RBC
reactor will enable the operator to respond to transient DO and predator con-
ditions that can negatively impact nitrification efficiency.
2.8.4.6 Alkalinity and pH
Nitrification is an acid-producing biochemical reaction. Approximately 7.1 mg
of calcium carbonate alkalinity are theoretically destroyed per mg of ammonia
nitrogen oxidized. Depending on initial alkalinity and unoxidized nitrogen
concentrations, the process of nitrification can potentially reduce wastewater
alkalinity to the point where pH will drop to 6.5 and even to 6.0 or less.
A strong relationship between pH and nitrification rate is generally acknowl-
edged. Optimal pH values found in the literature range-from 7.0 to 9.0 (50),
with nitrification efficiency falling off dramatically as pH decreases form
7.0 to 6.0 in unacclimated systems. Downing and Knowles (72) developed the
following expression relating Nitrosomonas growth rate to pH values up to 7.2
for combined carbon oxidation-nitrification systems:
2-36
(41)
-------
UN = >frN [1 - 0.833 (7.2 - pH)] (2-15)
Downing and Knowles assumed the growth rate to be constant in the pH range of
7.2 to 8.0.
Borchardt et al. (19) examined the effect of pH on RBC nitrification at eleven
different alkalinity levels using a 2-ft diameter polystyrene pilot unit and
uniform influent ammonia nitrogen concentrations of 20 mg/1. Short undefined
acclimation periods were employed at each alkalinity level. Their results
indicated a constant nitrification rate between pH 7.1 and pH 8.6, 25 percent
of this constant rate at pH 6.5, and zero nitrification at pH 6.0.
A comprehensive evaluation of pH and alkalinity effects on RBC nitrification
was recently completed by Stratta and Long (50). Four single-stage, 19.7-in.
diameter pilot systems were operated in parallel using high-rate trickling
filter clarified effluent as feed. Two long-term (10-wk) experimental phases
were conducted sequentially to investigate nitrification first at pH levels of
6.3 to 7.5 and second at pH levels of 7.6 to 8.8. In both phases, the experi-
ments were begun with clean media on day 1.
Stratta and Long's work demonstrated increasing ammonia nitrogen oxidation
rates with increasing pH up to pH 8.5 as shown in Figure 2-11. The rate at pH
8.8 decreased slightly, averaging 94 percent of the maximum rate at pH 8.5.
The rates presented in the figure are for the second 5-wk period of each phase
after the units had reached equilibrium.
Stratta and Long's pilot RBC units operating at lower pH values developed
nitrification rates comparable to those of the higher pH units during the
startup phase of each experiment. Nitrification comparability ceased after
approximately 35 days of operation, with deterioration of nitrification per-
formance to lower equilibrium rates for the lower pH units. A possible ex-
planation offered by the investigators for this observed phenomenon was the
gradual development of secondary predator populations in the lower pH systems.
Long-term pH reversion experiments were conducted by Stratta and Long on RBC
pilot systems acclimated to pH 8.0 and pH 8.8. Alkaline addition was termin-
ated so that each system began operating at the control pH level of 7.5. A
total of 19 days was required for the pH 8.8 unit to completely revert to
normal pH 7.5 nitrification efficiency, while the pH 8.0 unit reverted in 1
wk.
Snails migrating into Stratta and Long's RBC pilot systems along with the
trickling filter effluent feed inhabited the trough walls of those units with
a pH of 8.0 or less. They were not evident on the trough walls of the pH 8.5
and pH 8.8 units. The snails at no time took up residence on the media sur-
faces or in the attached biofilm.
2-37
(42)
-------
ro
00 OJ
*-— oo
O
o
O
CO
£
LU
I-
cc.
O
^
LU
GC
0.8
0.6
0.4
0.2
6.0
PHASE B-
PHASE A
OPTIMUM pH
7.0
8.0
9.0
pH
Figure 2-11. Effect of pH on RBC nitrification rates.
-------
Stratta and Long also conducted parallel nitrification enhancement investiga-
tions utilizing different alkaline chemicals to adjust pH in four two-stage
RBC pilot units. Sodium hydroxide (NaOH), sodium carbonate (soda ash, Na2C03)
and calcium hydroxide (lime, Ca(OH)2) were added to the first stage of three
of the units to maintain pH in those stages at the previously determined opti-
mum level of 8.5. Sodium bicarbonate (N&HC03) was dosed to maintain first-
stage pH of the fourth unit at 7.5. The results of this work are summarized
in Table 2-4.
The data in Table 2-4 exhibit remarkably similar ammonia nitrogen removal rates
for the three chemicals (sodium hydroxide, soda ash, and lime) in which first-
stage pH was maintained at 8.4 to 8.5. These rates were approximately 19 percent
higher than that of the control unit. Second-stage alkalinity destruction
varied from 14 percent less to 11 percent more than the theoretical require-
ment of 7.1 mg CaC03/mg NH3-N oxidized. All of the chemically-dosed RBC pilot
plants except the lime unit produced only minimal increases in effluent sus-
pended solids of 1 to 3 mg/1. Effluent suspended solids in the lime-dosed
unit, however, increased 20 mg/1.
The potential for enhancement of RBC nitrification rates at elevated pH's (8.0
to 8.8) demonstrated by Stratta and Long (50) on pilot scale systems has yet
to be duplicated with full-scale units. Albert (73) attempted to confirm
their results at a field-scale U.S. Army RBC plant. Soda ash was dosed to the
third stage of a six-stage, combined carbon oxidation-nitrification train,
i.e., at that point in the train where the onset of nitrification was routinely
observed. Although the target pH at the dose point was the optimum 8.5 pH
value noted by Stratta and Long, actual pH varied from 8.1 to 9.3 because
automatic pH control had not been provided. A side-by-side performance com-
parison with an undosed control train showed no overall enhancement of nitri-
fication rates in the dosed train, only a slight alternation in the interstage
TABLE 2-4. ALKALINE ENHANCEMENT RBC NITRIFICATION DATA
FROM STRATTA AND LONG (50)
Alkaline
Chemical
NaOH
Na2C03
Ca(OH)2
NaHC03
Control
First-
Stage
pH*
8.5
8.4
8.5
7.5
7.0
Second-
Stage
PH
7.9
8.0
7.9
7.7
6.9
NH3-N Removed
Overall
(lb/day/1000 sq ft)
0.516
0.520
0.522
0.492
0.438
Alkalinity Destroyed
in Second State
(mg CaC03/mg NH3-N)
6.1
7.9
7.0
7.7
7.2
*Influent to first stage maintained at pH 6.5 and 150 mg/1 CaC03 alkalinity.
2-39
(44)
-------
ammonia nitrogen removal pattern. The investigator suggested the lack of
observed improvement was probably due to the wide pH fluctuations experienced,
which preprevented constant-pH acclimation of the nitrifying population.
As discussed in Section 2.8.5.2, other factors such as oxygen transfer and
wastewater temperature may exert sufficient impact on full-scale RBC nitrifi-
cation to cancel or substantially damp any potentially beneficial effects of
pH optimization. Full-scale pH enchancement of RBC nitrification is a worthy
subject of further research, however.
2.8.4.7 Inhibition
Research data specific to nitrification inhibition in RBC's are generally
unavailable in the literature. It has been established that certain organic
compounds and heavy metals are toxic to unacclimated cultures of nitrifying
organisms (46)(74)(75)(76). Allyl-thiourea, for example, is an organic com-
pound frequently used to inhibit nitrification in the BOD test. Heavy metals
that have been implicated in nitrifier inhibition or toxicity include chro-
mium, copper, mercury, nickel, silver, and zinc, among others. Where indus-
trial wastes containing very high levels of ammonia or nitrite nitrogen are
discharged to municipal sewer systems as slug loads, temporary nitrifier
toxicity can result if the concentration increase for either nitrogen form
is sufficiently large (77).
Sawyer (78) has reported that 10 to 20 mg/1 of some heavy metals can be toler-
ated by nitrifiers at pH values of 7.5 to 8.0 where ionic disassociation is
small. Other metals that precipitate as hydroxides cause relatively little
inhibition as long as they remain insoluble but become very toxic if dissolved,
such as can happen with falling pH. Silver has been found to be toxic to
plastic media trickling filter nitrification of secondary effluent at concen-
trations as low as 2 pg/1 (79).
Equally important to the type and concentration of potential nitrification
inhibiting substances encountered is the variability with which they enter the
treatment system. Biological acclimation enables nitrifiers to adapt to higher
concentrations of these substances when present at reasonably consistent levels
than can be tolerated in slug loadings.
When lower-than-anticipated RBC nitrification rates (or cessation of nitrifica-
tion altogether) are experienced, conventional environmental impact factors
such as wastewater temperature, DO, and pH; nuisance growths; and organic and
ammonia nitrogen loadings should be investigated first. If these factors can
be ruled out as contributory agents to the problem, a comprehensive wastewater
characterization study is usually the next step. The compounds and metals
included in a characterization screening should be predicated on the type and
magnitude of industrial and commercial wastes known or suspected to be entering
the municipal treatment facility.
2-40
(45)
-------
2.8.5 Interdependency of Factors Affecting Nitrification
2.8.5.1 Combined Kinetic Expression
Kinetic factors affecting nitrifier growth and nitrification rate were dis-
cussed in Section 2.8.4. In the absence of inhibitory or toxic wastewater
components, the major factors are ammonia nitrogen concentration, wastewater
temperature, DO, and pH. A general expression was presented in Equation 2-5
indicating the interdependency of these kinetic parameters on unit substrate
(nitrogen ammonia) removal rate. A similar combined Monod expression has been
proposed (45) to relate Nitrosomonas growth rate to the same factors:
(a PHb)
(2-16)
where N is the ammonia nitrogen concentration (mg/1).
Substituting Equation 2-12 as an estimate of the effect of temperature on y^,
Equation 2-13 for KN, the British-developed value of 1.3 mg/1 (64) for KQ?,
and Equation 2-15 for the general pH effect term yields the following specific
equation valid for pH ^ 7.2 and wastewater temperatures between 8 and 30°C (45)
uw = 0.47 [60.098(T - 15)]
DO
1.3 + DO
10(0.051 T - 1.158) + N
• *
[1 - 0.833(7.2 - pH)]
(2-17)
The above equations indicate that if all the kinetic terms considered approach
their maximum values, i.e., if N is large in comparison to K|\j, if DO is sub-
stantially larger than 1.3 mg/1 (say 4 to 6 mg/1), and if pH is in the range
of 7.0 to 7.2, UN will approach thaffr.. value dictated by wastewater tempera-
ture. (Downing and coworkers (52)(53) have estimated^ to be 0.47 dayl at
15°C (59°F).) Conversely, if any one factor or parameter becomes limiting,
even if all the rest are non-limiting, yN and the attainable nitrification
rate will be much lower, perhaps even zero if the limiting condition is severe
enough.
2.8.5.2 Comparison of Pilot- and Full-Scale RBC Nitrification Rates
As will be shown in Section 5, peak demonstrated nitrification rates for full-
scale RBC's generally fall in the range of 0.3 to 0.35 Ib NH3-N oxidized/day/
1000 sq ft media surface area. The updated Autotrol design procedure is based
2-41
(46)
-------
on a maximum zero-order oxidation rate of approximately 0.3 Ib NH3~N/day/1000
sq ft at ammonia nitrogen concentrations of 5 mg/1 or higher in the bulk liquid
as illustrated previously in Figure 2-9 (24).
Published pilot-scale RBC data typically exhibit maximum nitrification rates
1.5 to 2.5 times higher than full-scale peak nitrification rates. Rates as
high as 0.74 to 0.78 Ib NH3-N/day/1000 sq ft were observed by Borchardt et al.
(19) on a 4-ft diameter RBC unit at temperatures of 50 to 65°F. Pano et al.
(23) achieved rates of 0.45 to 0.68 Ib NHs-N/day/lOOO sq ft on 15-in. diameter
pilot discs in the temperature range of 59 to 68°F. Stratta and Long (50)
utilized pH enhancement to increase nitrification rates from a control plateau
(7.0
-------
Nitrification rates have been achieved in some stages of full-scale RBC systems
at or near the consensus maximum rate of 0.3 Ib NH3-N/day/1000 sq ft with bulk
liquid DO's of 2 to 4 mg/1 (61). The presence of measurable bulk liquid DO
when a stage's nitrification rate is presumably being controlled by oxygen
transfer appears to be contradictory. Although definitive literature on this
subject is lacking, the fact that only 2 to 4 percent of the assumed maximum
oxygen transfer rate (1.5 lb/02/day/1000 sq ft) is required to increase bulk
liquid DO from 0 to 4 mg/1 across an RBC system at typical hydraulic loading
rates may offer a partial explanation.
Oxygen transfer limitations and substrate depletion rates may also figure
prominently in the discrepancy of observed temperature effects between full-
and pilot-scale RBC equipment. As illustrated previously in Figure 2-10, the
manufacturers do not recommend the use of a wastewater temperature correction
factor above 55°F. The validity of this recommendation is supported by field
data presented in Section 5. Borchardt et al. (19), on the other hand, meas-
ured a 10-percent increase in nitrification rate at temperatures of 54 to 67°F
compared to temperatures of 45 to 55°F using 4-ft diameter polystyrene units.
Pano et al. (23) achieved a 50-percent higher rate at 68°F than at 59°F on
15-in. diameter pilot systems.
As indicated in the aforegoing discussion, several kinetic parameters strongly
influence RBC nitrification. The designer must have an appreciation for the
interdependency of these parameters to intelligently size RBC nitrification
systems. If pilot-generated data are to be used as a basis for full-scale
design, the designer will either have to establish or accept a previously-
established empirical scale-up technique or attempt calibration and utiliza-
tion of one of several available RBC deterministic models.
2.9 Denitrification
2.9.1 Background
When the DO level is near or at zero, many heterotrophic microorganisms are
able to use nitrate nitrogen as an alternate electron acceptor for dissimila-
tory nitrate reduction to nitrogen gas. This phenomenon has been widely used
in wastewater treatment systems for nitrogen removal. Unoxidized nitrogen
must first be biologically converted to nitrate nitrogen in an aerobic environ-
ment. Denitrification can be accomplished in separate-stage RBC systems where
carbonaceous oxidation and nitrification occur in the lead stage(s) of the RBC
process train and a carbon source, commonly methanol, is added to provide the
energy for microbial denitrification in the last (anoxic) stage. This approach
is illustrated in Figure 2-12. When this approach is employed, it may be
necessary to add a terminal aerated RBC stage to the system to oxidize any
residual methanol not utilized for denitrification.
2-43
(48)
-------
RBC systems can also be staged to achieve denitrification as indicated in
Figure 2-13. In this case, the anoxic unit is the first stage of the RBC
train and the organic carbon naturally present in the wastewater is used for
nitrate reduction; nitrate nitrogen must be introduced to this stage by recir-
culation of nitrified wastewater from the downstream stages. This technique
is well known for activated sludge systems, e.g., the Bardenpho process, but
has not been utilized thus far in any full-scale municipal RBC installations
in the United States. According to the Autotrol patent for this denitrifica-
tion alternative (80), the required recirculation rate is from 100 to 300
percent of the flow rate entering the plant and as much as 75 percent of the
recirculated nitrate nitrogen can be reduced to nitrogen gas. Since the
stoichiometric quantity of alkalinity produced/mg N03-N reduced to nitrogen
gas is 3.57 mg as CaC03, the recirculation approach is advantageous where pH
adjustment and/or control are required to promote efficient nitrification.
As no DO is desired in the denitrification reaction, RBC media should be com-
pletely submerged in the wastewater for denitrification applications. The
submerged media are mechanically driven at a rotational velocity of about 1.6
rpm according to the current design procedure (15). Because RBC denitrifi-
cation is an emerging technology with only one U.S. full-scale municipal
facility at Orlando, Florida (81), in operation as of September 1982, the
mechanical reliability and ramifications of submerged media and shaft opera-
tion are unknown.
2.9.2 Kinetics
As noted by Harremoes and Riemer (82), denitrification is for all practical
purposes a zero-order reaction with respect to the nitrate nitrogen concen-
tration. The Michelis-Menten constant, Ks, is on the order of 0.1 rng/1. The
change in reaction rate to first order at low concentrations can be ignored.
If the biofilm on the surface of the RBC media is fully penetrated by organic
carbon, e.g., methanol, and nitrate, the reaction will follow zero-order re-
moval and be independent of the bulk liquid concentration. In this case, the
shear forces caused by rotation of the media are sufficient to prevent the
development of a biofilm that is thick enough to allow depletion of nitrate
anywhere within the film. If the biofilm is not fully penetrated, the re-
action becomes half order with respect to the bulk liquid concentration (14).
Based on the observations of Jeris et al. (83) with a denitrifying fluidized
bed, biofilms do not appear to be particularly fragile in denitrification
systems; these observations are contrary to the findings of Sullivan (6),
however, for denitrification in RBC systems.
Total methanol consumption for denitrification has been calculated by Christen-
sen and Harremoes (84) as 2.42 mg/mg N03-N reduced. Methanol is also required
to reduce any initial DO present in the incoming wastewater. A commonly-used
design value for the required methanol dosage is 3 mg/mg N03-N to be reduced.
2-44
(49)
-------
CARBONACEOUS OXIDATION
AND NITRIFICATION
DENITRIFICATION
TERMINAL
OXIDATION
STAGE
(OPTIONAL)
PRIMARY
EFFLUENT
w
*-
*-
;
k
r
^,
1
L
FINAL
CLARIFICATION
L _ I
METHANOL
ADDITION
Figure 2-12. RBC process configuration for denitrification using methanol
addition.
EFFLUENT
>•
•^~~ ro
en i
o -i^
•~-~> en
DENITRIFICATION
CARBONACEOUS OXIDATION
AND NITRIFICATION
1
PRIMARY
EFFLUENT
RECIRCULATION
FINAL
CLARIFICATION
OEFFLUEr
—
Figure 2-13. RBC process configuration for denitrification using primary
effluent as the carbon source.
-------
Williamson and McCarty (12) employed a stoichiometric requirement of 2.5 mg
methanol/mg N03-N in conjunction with a ratio of methanol-to-nitrate diffusion
coefficients of 2.0 inside the biofilm to predict that methanol would have to
be supplied in concentrations approximately five times as large as the nitrate
nitrogen concentration if methanol is not to be flux limiting in fixed film
systems. They concluded, therefore, that denitrification in a biofilm reactor
will be flux limited by methanol and not nitrate when methanol is supplied in
stoichiometric quantities. This is an interesting conclusion since the kinetic
analyses of fixed film systems frequently ignore the methanol level, provided
it is present in excess of the stoichiometric requirement.
2.9.3 Previous Studies and Empirical Design Formulations
The Autotrol design manual (15) presents a series of denitrification design
curves showing predicted effluent nitrate nitrogen concentrations for various
influent concentrations and hydraulic loading rates. According to Sullivan
(6), these curves were developed from observations at several sites. An analy-
sis of these curves indicates that the claimed denitrification rate is inde-
pendent of bulk liquid nitrate nitrogen down to a concentration of 1 mg/1 and
that the denitrification rate for 55°F is approximately 0.9 Ib NO3-N/day/1000
sq ft. A companion design curve (15) provides wastewater temperature correc-
tion factors applicable over the range of 45 to 65°F, with the correction
factor at 65°F about three times that for 45°F, i.e., one-third the media
surface area is required at 65°F for a given hydraulic loading.
The approach utilized by Autotrol is consistent with the statement of Murphy
et al. (60) that their studies indicated that denitrification in the presence
of an adequate organic carbon source was independent of the bulk liquid ni-
trate nitrogen concentration. However, since the data on which this conclu-
sion was based were not presented, it is impossible to evaluate Murphy et al.'
results. A four-stage, 19.7-in., submerged RBC pilot unit rotated at 13 rpm
was utilized in their work. The observed denitrification rate at 55°F was
approximately 0.85 Ib N03-N/day/1000 sq ft. This rate is roughly double that
reported for submerged, high-porosity media denitrification systems (45).
In contrast to the conclusion of Murphy et al. (60) and the design approach
utilized by Autotrol (15), the studies reported by Blanc et al. (85) using
totally submerged 2-ft diameter discs rotated at 3-1/8 rpm with methanol addi-
tion to the nitrified effluent feed at the Marlborough, Massachusetts Easterly
treatment plant indicated a strong dependency of denitrification rate on
nitrate nitrogen concentration in the bulk liquid in the 0 to 6 mg/1 range
(Figure 2-14). A strong dependence of denitrification rate on nitrate nitrogen
concentration in the bulk liquid was also reported by Harremoes and Riemer
(82) when nitrified wastewater receiving 3 mg methanol/mg N03-N was fed to
downflow filters. The observed removal rate in any filter section was a
function of bulk liquid concentration below about 30 mg/1 N03-N. The analysis
by Rittmann and McCarty (86) of the denitrification results reported by Jeris
et al. (83) with expanded bed operation also noted a continual decline in
2-46
(51)
-------
(Q
c
CD
ro
NO3-N CONCENTRATION (mg/l)
on
r\i
ro
CD
00-
-t^
g- 3J
00
Q_
CD
IN3
CO
CJ1
cn
CO
•— a. s;
" I-i;
2 Q)
^* ^"
1°
o S.
O 3
CD =2
3 O
IS
11
w ^"
Q. o'
0) 3
T3 CO
(D 0>
Q. CO
-« 0>
If
gj=*
0) §
o o
O
CO
I
33
m
m
day
00
CO
.a
p
ho
p
'-u
O
bo
O D
I! II O
COO H
°
co ?
-vl
oo
t> no
•t>. co co
co r\5 ^
o o 10
OOO
0> Q] Q>
l-t- l-t- r+
tt> Q> 05
m
o
-------
methanol utilization rate with distance up the column.
In view of the extremely limited data available, it is impossible to make firm
conclusions about the general appropriateness of assuming that RBC denitrifica-
tion rates are independent of nitrate nitrogen concentration down to 1 mg/1 in
the bulk liquid as currently proposed (15)(60). It is clear that denitrifica-
tion is not independent of bulk liquid nitrate nitrogen concentration for
other fixed film systems. Mass transfer considerations and the results of
Blanc et al. (85) suggest that additional RBC denitrification studies are
required to define design loading parameters that can be used with confidence,
particularly for any configuration where nitrate nitrogen levels will be pre-
sent in concentrations less than 6 mg/1 in the denitrification section of the
RBC reactor.
2.10 Secondary Clarification
The concentration of suspended solids leaving the last stage of an RBC train
treating municipal wastewater will normally be less than 300 mg/1 if primary
clarification is not provided and less than 200 mg/1 where primaries are used,
The settling characteristics of RBC solids during secondary clarification,
therefore, will essentially be those of a dilute suspension with zone or com-
pression settling confined to the clarifier bottom. Settling analyses for
this type of settling were described by Camp (87). Whenever a nitrified
effluent is produced, settled solids should not be allowed to accumulate in
the final clarifier for any substantial period (primarily depending on tem-
perature) to avoid solids resuspension and flotation from denitrification
reactions.
Studies conducted by Scheible and Novak (39) at the Edgewater, New Jersey RBC
facility indicated that peak surface overflow rates had to be limited to 550
to 650 gpd/sq ft to achieve less than 30 mg/1 of suspended solids in the final
effluent. Based on pilot plant studies, Srinivasaraghavan et al. (88) recom-
mended an average design overflow rate of 740 gpd/sq ft for the full-scale
Pinners Point RBC plant. Murphy and Wilson (1) recommend surface overflow
rates less than 600 gpd/sq ft to maximize solids removal. According to Smith
et al. (89), the recommended range of secondary clarifier average overflow
rates for trickling filter upgrading applications is 500 to 800 gpd/sq ft,
with 800 applying if the effluent suspended solids objective is 30 mg/1 and
500 gpd/sq ft if the goal is 15 mg/1. These ranges are the same as recom-
mended by Autotrol (15). Clow (16) recommends 800 gpd/sq ft for a 30/30 efflu-
ent and not more than 500 gpd/sq ft for an effluent of 10 mg/1 effluent
suspended solids. DeCarlo (90) recommends that peak hydraulic loading rates
be limited to 1000 to 1200 gpd/sq ft.
Polymer addition is an effective technique for improving solids capture in
secondary clarifiers. Polymer storage, mixing, and feed equipment can be in-
corporated as rather inexpensive insurance in the design of an RBC facility or
2-48
(53)
-------
it can be installed later if a chronic effluent clarity problem arises.
2.11 Sludge Production
Sludge production in RBC systems is sometimes estimated by subtracting the in-
fluent suspended solids to the RBC unit from the suspended solids leaving the
last stage and then relating the net solids gain to the amount of soluble 6005
removed. This approach tends to estimate small values for sludge production.
As shown previously in Table 2-2, the SRT of RBC biomass in the exterior bio-
film layers, which is in pseudo equilibrium with the bulk liquid, is a function
of the 6005 removal rate, i.e., the loading the biomass is seeing. Since the
organisms that grow in an RBC system are the same types that populate activated
sludge, RBC sludge production will depend on the loading to each stage, the
average residence time of the biofilm on a stage, and the fate of suspended
material that passes from stage to stage in multistage units, i.e., whether it
can reattach to the RBC surface in subsequent stages, thus allowing further
time for degradation. The above sludge age concept for predicting sludge
production from RBC systems has also been applied to trickling filter systems
(91).
2.12 References
1. Murphy, K. L. and R. W. Wilson. Pilot Plant Studies of Rotating Biologi-
cal Contactors Treating Municipal Wastewater. Report SCAT-2, Environment
Canada, Ottawa, Ontario, Prepared for Canada Mortgage and Housing Corpo-
ration, July 1980.
2. Hoag, G. E., W. D. Widmer, and W. H. Hovey. Microfauna and RBC Perform-
ance: Laboratory and Full-Scale Systems. In: Proceedings of the 1st
National Symposium/Workshop on Rotating Biological Contactor Technology,
Vol. I, EPA-600/9-80-046a, NTIS PB81-124539, June 1980. pp. 167-187.
3. Rutgers University, Department of Environmental Sciences. Rotating Bio-
logical Disk Wastewater Treatment Process - Pilot Plant Evaluation.
Final report for Grant No. 17010 EBM, USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, August 1972.
4. Antonie, R. L. Fixed Biological Surfaces - Wastewater Treatment: The
Rotating Biological Contactor. CRC Press, Cleveland, Ohio, 1976.
5. Ingram, W. M. Snails Associated with Sewage Treatment. Sewage and In-
dustrial Wastes, 30(6):821-825, June 1958.
6. Personal communications from R. A. Sullivan, Autotrol Corporation,
Milwaukee, Wisconsin, to J. A. Heidman and R. C. Brenner, USEPA,
Cincinnati, Ohio, April 14, 1982.
2-49
(54)
-------
7. Kornegay, B. H. and J. F. Andrews. Characteristics and Kinetics of Bio-
logical Fixed Film Reactors. FWPCA-17050-00/70, NTIS PB-199 834/BE,
1970.
8. Whalen, W. J., H. R. Bungay, and W. M. Sanders. Microelectrode Determi-
nation of Oxygen Profiles in Microbial Slime Systems. Environmental
Science and Technology, 3(12):1297-1298, December 1969.
9. Famularo, J., J. W. Mueller, and T. J. Mulligan. Verification Studies of
the Biofilm Model for Bacterial Substrate Utilization. Journal WPCF,
50(4):653- 671, April 1978.
10. Zeevalkink, J. A., P. Kelderman, D. C. Visser, and C. Boelhouwer. Physi-
cal Mass Transfer in a Rotating Disc Gas-Liquid Contactor. Water Re-
search, 13(9): 913-919, 1979.
11. Grady, C. P. L., Jr. and H. C. Lim. Biological Wastewater Treatment.
Marcel Dekker, New York City, 1980.
12. Williamson, K. and P. L. McCarty. A Model of Substrate Utilization by
Bacterial Films. Journal WPCF, 48(l):9-24, January 1976.
13. Williamson, K. and P. L. McCarty. Verification Studies of the Biofilm
Model for Bacterial Substrate Utilization. Journal WPCF, 48(2):281-296,
February 1976.
14. Harremoes, P. Biofilm Kinetics. In: Water Pollution Microbiology, Vol.
2, Ed. by R. Mitchell, Wiley - Interscience, New York City, 1978.
15. Autotrol Wastewater Treatment Systems Design Manual. Autotrol Corporation,
Bio-Systems Division, Milwaukee, Wisconsin, 1978.
16. Clow Envirodisc Rotating Biological Contactor Systems Catalog. Clow
Corporation, Florence, Kentucky, 1980.
17. Gutierrez, A., I. L. Bogert, 0. K. Scheible, and T. J. Mulligan. Up-
grading Primary Tanks with Rotating Biological Contactors. EPA-600/2-80-
003, NTIS PB80-203771, March 1980.
18. Painter, H. A. Microbial Transformations of Inorganic Nitrogen. Pre-
sented at the Conference on Nitrogen as a Water Pollutant, Copenhagen,
August 1975; Available in: Progress in Water Technology, 8(4/5):3-29,
1977.
19. Borchardt, J. A., S. J. Kang, and T. H. Chung. Nitrification of Second-
ary Municipal Waste Effluents by Rotating Bio-Discs. EPA-600/2-78-061,
NTIS PB-285 240/8BE, June 1978.
20. Advanced Wastewater Pilot Plant Treatment Studies. Report prepared for
City of Indianapolis by Reid, Quebe, Allison, Wilcox, and Associates,
Inc., Indianapolis, Indiana, January 1975.
2-50
(55)
-------
21. Reh, C. W., T. E. Wilson, and R. Srinivasaraghavan. An Approach to De-
sign of RBC's for Treatment of Municipal Wastewater. Presented at the
ASCE National Environmental Engineering Conference, Nashville, Tennessee,
July 1977.
22. Crawford, P. M. Use of Rotating Biological Contactors for Nitrification
at the City of Guelph Water Pollution Control Plant, Guelph, Ontario,
Canada. In: Proceedings of the 1st National Symposium/Workshop on
Rotating Biological Contactor Technology, Vol. II, EPA-600/9-80-046b,
NTIS PB81-124547, June 1980. pp. 1247-1273.
23. Pano, A., E. J. Middlebrooks, and J. H. Reynolds. The Kinetics of Ro-
tating Biological Contactors Treating Domestic Wastewater. Water Quality
Series UWRL/Q-81/04, Utah State University, College of Engineering, Logan,
Utah, September 1981.
24. Antonie, R. L. Nitrogen Control with the Rotating Biological Contactor.
Autotrol Corporation brochure, Milwaukee, Wisconsin, (No date).
25. Workshop on RBC Research Needs (Discussion). In: Proceedings of the 1st
National Symposium/Workshop on Rotating Biological Contactor Technology,
Vol. II, EPA-600/9-80-046b, NTIS PB81-124547, June 1980. pp. 1429-1461.
26. Hynek, R. J. and C. C. Chou. Development and Performance of Air-Driven
Rotating Biological Contactors. In: Proceedings of the 34th Industrial
Waste Conference, Purdue University, West Lafayette, Indiana, May 8-10,
1979; Ann Arbor Science, Ann Arbor, Michigan, 1980. pp. 805-815.
27. Hoehn, R. C. and A. D. Ray. Effects of Thickness on Bacterial Film.
Journal WPCF, 45(11):2302-2320, November 1973.
28. Sanders, W. M., III. Oxygen Utilization by Slime Organisms in Continuous
Culture. Air and Water Pollution, 10(4):253-276, April 1966.
29. Alleman, J. E., J. A. Veil, and J. T. Canady. Scanning Electron Micro-
scope Evaluation of Rotating Biological Contactor Biofilm. Water Re-
search, 16(5):543-550, 1982.
30. CRC Handbook of Microbiology. Vol. I: Bacteria. 2nd Edition, Ed. by A.
I. Laskin and H. A. Lechevalier, CRC Press, Cleveland, Ohio, 1977.
31. Torpey, W., H. Heukelekian, A. J. Kaplovsky, and L. Epstein. Effects of
Exposing Slimes on Rotating Discs to Atmospheres Enriched with Oxygen.
Presented at the 6th International Conference on Water Pollution Research,
Jerusalem 18-23, 1972; Available in: Advances in Water Pollution Research,
Permagon Press, New York City, 1973. pp. 405-415.
32. Huang, J. C. Operational Experience of Oxygen Enriched Rotating Biologi-
cal Contactors. In: Proceedings of the 1st National Symposium/Workshop
on Rotating Biological Contactor Technology, Vol. I, EPA-600/9-80-046a,
NTIS PB81-124539, June 1980. pp. 637-659.
2-51
(56)
-------
33. Binitanjia, H. H. J., J. J. Brunsmann, and C. Boelhouwer. The Use of
Oxygen in a Rotating Disc Process. Water Research, 10(6):561-565, 1976
34. Severin, B. F., H. Brociner, J. E. Dumanowski, J. T. Su, and M. M.
Gurvitch. Empirical Oxygen Transfer Relation in the RBC Process. In:
Proceedings of the 1st National Symposium/Workshop on Rotating Biological
Contactor Technology, Vol. II, EPA-600/9-80-046b, NTIS PB81-124547, June
1980. pp. 1077-1100.
35. Huang, J. C. Workshop on RBC Research Needs (Discussion). Ibid. p.
1445.
36. Chesner, W. H. and J. lannone. Review of Current RBC Performance and
Design Procedures. Report prepared for USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, under Contract No. 68-02-2775
by Roy F. Weston, Inc., (Publication pending).
37. McCann, K. J. and R. A. Sullivan. Aerated RBC's - What are the Benefits.
In: Proceedings of the 1st National Symposium/Workshop on Rotating Bio-
logical Contactor Technology, Vol. I, EPA-600/9-80-046a, NTIS PB81-124539,
June 1980. pp. 515-523.
38. Chou, C. C. Oxygen Transfer Capacity of Clean Media Pilot Reactors at
South Shore. Autotrol Corporation memorandum, Milwaukee, Wisconsin,
August 23, 1978.
39. Opatken, E. J. Rotating Biological Contactors - Second Order Kinetics.
In: Proceedings of the First International Conference on Fixed-Film
Biological Processes, Vol. I, EPA-600/9-82-023a, Kings Island, Ohio,
April 20-23, 1982. pp. 210-232.
40. Scheible, 0. K. and J. J. Novak. Upgrading Primary Tanks with Rotating
Biological Contactors. In: Proceedings of the 1st National
Symposium/Workshop on Rotating Biological Contactor Technology, Vol.
II, EPA-600/9-80-046b, NTIS PB81-124547, June 1980. pp. 961-996.
41. Lyco Wastewater Products - RBS Systems Catalog. Lyco Division of Remsco
Associates, Marlboro, New Jersey, 1982.
42. Roy F. Weston, Inc. RBC DIS Subtask on RBC Design Approaches. Internal
report prepared for USEPA under Contract No. 68-03-3019, Cincinnati,
Ohio, March 5, 1982.
43. Banerji, S. K. ASCE Water Pollution Management Task Committee Report on
Rotating Biological Contactor for Secondary Treatment. In: Proceedings
of the 1st National Symposium/Workshop on Rotating Biological Contactor
Technology, Vol. I, EPA-600/9-80-046a, NTIS PB81-124539, June 1980. pp.
31-52.
44. RBC systems brochure for Manufacturer X, 1978.
45. Process Design Manual for Nitrogen Control. EPA-625/1-75-007, USEPA,
Center for Environmental Research, Cincinnati, Ohio, October 1975.
2-52
(57)
-------
46. Painter, H. A. A Review of Literature on Inorganic Nitrogen Metabolism
in Microorganisms. Water Research, 4(6):393-450, 1970.
47. Haug, R. T. and P. L. McCarty. Nitrification with the Submerged Filter.
Final report for Grant No. 17010 EPM, USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, August 1971.
48. Notes on Water Pollution No. 52. Water Research Centre, Stevenage,
England, 1971.
49. Gujer, W. and D. Jenkins. The Contact Stabilization Process - Oxygen
and Nitrogen Mass Balances. SERL Report No. 74-2, University of Califor-
nia, Sanitary Engineering Research Laboratory, Berkeley, California,
February 1974.
50. Stratta, J. M. and D. A. Long. Nitrification Enhancement Through pH
Control with Rotating Biological Contactors. Final report prepared
for USA Medical Research and Development Command under Contract No.
DAMD17-79-C-9110 by Pennsylvania State University, September 1981.
51. Ito, K. and T. Matsuo. The Effect of Organic Loading on Nitrification
in RBC Wastewater Treatment Processes. In: Proceedings of the 1st
National Symposium/Workshop on Rotating Biological Contactor Technology,
Vol. II, EPA-600/9-80-046b, NTIS PB81-124547, June 1980. pp. 1165-1175.
52. Knowles, G., A. L. Downing, and M. J. Barrett. Determination of Kinetic
.Constants for Nitrifying Bacteria in Mixed Culture, with the Aid of an
Electronic Computer. Journal of General Microbiology, 38:263-278,1965.
53. Downing, A. L. and A. P. Hopwood. Some Observations on the Kinetics
of Nitrifying Activated Sludge Plants. Schweiz. Zeitsch f Hydrol,
26(2):271-288, 1964.
54. Poduska, R. A. and J. F. Andrews. Dynamics of Nitrification in the
Activated Sludge Process. In: Proceedings of the 29th Industrial
Waste Conference, Purdue University, West Lafayette, Indiana, May 7-9,
1974. pp. 1005-1025.
55. Balakrishnan, S. and W. W. Eckenfelder, Jr. Nitrogen Relationships in
Biological Waste Treatment Processes - I, Nitrification in the Activated
Sludge Process. Water Research, 3(1):73-81, 1969.
56. Lawrence, A. W. and C. G. Brown. Design and Control of Nitrifying
Activated Sludge Systems. Journal WPCF, 48(7):1779-1803, July 1976.
57. Loehr, R. C., T. B. S. Prakasam, E. G. Srinath, and Y. D. Joo. Develop-
ment and Demonstration of Nutrient Removal from Animal Wastes. EPA-
R2-73-095, NTIS PB-217 887/9BE, January 1973.
2-53
(58)
-------
58. Wuhrmann, K. Research Developments in Regard to Concept and Base
Values of the Activated Sludge System. In: Advances in Water Quality
Improvement, Ed. by E. F. Gloyna and W. W. Eckenfelder, Jr., University
of Texas Press, Austin, Texas, 1968. pp. 143-170.
59. Sanders, F. M. and R. L. Pope. Nitrification with Rotating Biological
Contactor Systems. ERC Report No. 06-78, Georgia Institute of Technology,
Environmental Research Center, Atlanta, Georgia, October 1978.
60. Murphy, K. L., P. M. Sutton, R. W. Wilson, and B. E. Jank. Nitrogen
Control: Design Considerations for Supported Growth Systems. Journal
WPCF, 49(4):549-557, April 1977.
61. Letter (data) communication from R. A. Sullivan, Autotrol Corporation,
Milwaukee, Wisconsin, to R. C. Brenner, USEPA, Cincinnati, Ohio, March
12, 1982.
62. Randtke, S. J., G. F. Parkin, J. V. Keller, J. 0. Leckie, and P. L.
McCarty. Soluble Organic Nitrogen Characteristics and Removal. EPA-
600/2-78-030, NTIS PB-281 727/8BE, March 1978.
63. Barth, E. F. and R. L. Bunch. Biodegradation and Treatability of
Specific Pollutants. EPA-600/9-79-034, NTIS PB80-106438, October
1979.
64. Mueller, J. A., P. Paquin, and J. Famularo. Nitrification in Rotating
Biological Contactors. Journal WPCF, 52(4):688-710, April 1980.
65. Water Pollution Research Board. Water Pollution Research. 1964; Her
Majesty's Stationery Office, London, 1965.
66. Letter communication from A. D. Schwer, Metropolitan Sewer District of
Greater Cincinnati (Ohio), to D. S. Parker, Brown and Caldwell Consulting
Engineers, Walnut Creek, California, March 9, 1971.
67. Wuhrmann, K. Effects of Oxygen Tension on Biochemical Reactions in
Sewage Purification. In: Advances in Biological Treatment, Ed. by W.
W. Eckenfelder, Jr. and J. McCabe, Permagon Press, New York City, 1963.
68. Stankewich, M. J., Jr. Biological Nitrification with the High Purity
Oxygenation Process. In: Proceedings of the 27th Industrial Waste
Conference, Purdue University, West Lafayette, Indiana, May 2-4, 1972.
pp. 1-23.
69. Nagel, C. A. and J. G. Haworth. Operatonal Factors Affecting Nitrifi-
cation in the Activated Sludge Process. Presented at the 42nd Annual
Conference of the WPCF, Dallas, Texas, October 1969; Available as a
reprint from the County Sanitation Districts of Los Angeles County.
70. Hitdlebaugh, J. A. and R. D. Miller. Operational Problems with Rotating
Biological Contactors. Journal WPCF, 53(8):1283-1293, August 1981.
2-54
(59)
-------
71. Letter (data) communication from W. L. Morley, Wastewater Treatment
Plant, Gladstone, Michigan, to R. C. Brenner, USEPA, Cincinnati, Ohio,
November 17, 1981.
72. Downing, A. L. and G. Knowles. Population Dynamics in Biological
Treatment Plants. Presented at the 3rd International Conference on
Water Pollution Research, Munich, September 1966; Available in: Advances
in Water Pollution Research, Vol. 2, Water Pollution Control Federation,
Washington, D.C., 1967. pp. 117-142.
73. Albert, J. L. Use of Supplemental Aeration and pH Adjustment to Im-
prove Nitrification in a Full-Scale Rotating Biological Contactor
System. In: Proceedings of the First International Conference on
Fixed-Film Biological Processes, Vol. II, EPA-600/9-82-023b, Kings
Island, Ohio, April 20-23, 1982. pp. 633-657.
74. Mulbarger, M. C. The Three Sludge System for Nitrogen and Phosphorus
Removal. Presented at the 44th Annual Conference of the WPCF, San
Francisco, California, October 1971.
75. Interaction of Heavy Metals and Biological Sewage Treatment Processes.
USPHS, Division of Water Supply and Pollution Control, May 1965.
76. Letter report from Brown and Caldwell Consulting Engineers, Walnut
Creek, California, to Valley Community Services District, Dublin,
California, March 1974.
77. Anthonisen, A. C., R. C. Loehr, T. B. S. Prakasam, and E. G. Srinath.
Inhibition of Nitrification by Ammonia and Nitrous Acid. Journal
WPCF, 48(5):835-852, May 1976.
78. Letter communication from C. N. Sawyer, self employed, Sun City, Arizona,
to D. S. Parker, Brown and Caldwell Consulting Engineers, Walnut Creek,
California, January 24, 1975.
79. Personal communication from G. W. Davis, B. F. Goodrich Company, Brecks-
ville, Ohio, to D. S. Parker, Brown and Caldwell Consulting Engineers,
Walnut Creek, California, August 1974.
80. United States Patent 3,869,380. March 4, 1975.
81. Dallaire, G. U.S.'s Largest Rotating Biological Contactor Plant to
Slash Energy Use 30%. Civil Engineering, ASCE, 49(l):70-72, January
1979.
82. Harremoes, P. and M. Riemer. Pilot-Scale Experiments on Down-Flow
Filter Denitrification. Presented at the Conference on Nitrogen as a
Water Pollutant, Cophehagen, August 1975; Available in: Progress in
Water Technology, 8(4/5):557-576, 1977.
2-55
(60)
-------
83. Jen's, J. S., C. Beer, and J. A. Mueller. High Rate Biological Denitri-
fication Using a Fluidized Bed. Journal WPCF, 46(9):2118-2128, Septem-
ber 1974.
84. Christensen, M. H. and P. Harremoes. Biological Denitrification of
Sewage: A Literature Review. Presented at the Conference on Nitrogen
as a Water Pollutant, Copenhagen, August 1975; Available in: Progress
in Water Technology, 8(4/5) :509-555, 1977.
85. Blanc, F. C., J. C. O'Shaughnessy, D. J. Connick, and D. Wood. Denitri-
fication of Nitrified Municipal Wastewater Using Rotating Biological
Contactors. In: Proceedings of the 1st National Symposium/Workshop
on Rotating Biological Contactor Technology, Vol. II, EPA-600/9-80-
046b, NTIS PB81-124547, June 1980. pp. 1275-1300.
86. Rittman, B. E. and P. L. McCarty. Design of Fixed-Film Processes with
Steady-State-Biofilm Model. Presented at the 10th International Confer-
ence on Water Pollution, Toronto, June 23-27, 1980; Available in:
Progress in Water Technology, 12(6) :271-281, 1980.
87. Camp, T. R. Sedimentation and Design of Settling Tanks. Proceedings
of the ASCE, 71:445-486, 1945.
88. Srinivasaraghavan, R., C. W. Reh, and S. Liljegren. Performance Evalua-
tion of Air Driven RBC Process for Municipal Waste Treatment. In:
Proceedings of the 1st National Symposium/Workshop on Rotating Biological
Contactor Technology, Vol. I, EPA-600/9-80-046a, NTIS PB81-124539,
June 1980. pp. 525-552.
89. Smith, D. A., C. P. Poon, and R. D. Miller. Upgrading DA Trickling-
Filter Sewage Treatment Plants. Technical Report No. N-102, USA Corps
of Engineers, Construction Engineering Research Laboratory, Champaign,
Illinois, May 1981.
90. Personal communication from D. A. DeCarlo, Burgess and Niple, Ltd.,
Columbus, Ohio, to J. A. Heidman, USEPA, Cincinnati, Ohio, April 14,
1982.
91. Kincannon, D. F. and J. H. Sherrard. Trickling Filter Versus Activated
Sludge - When to Select Each Process. Proceedings of the 28th Indus-
trial Waste Conference, Purdue University, West Lafayette, Indiana,
May 1-3, 1973. pp. 69-75.
2-56
(61)
-------
SECTION 3
EQUIPMENT
3.1 Introduction
Six manufacturers were marketing RBC units in the United States at the time
of document preparation and provided input for this section (1)(2)(3)(4)(5)
(6). As stated in the Foreword, one manufacturer (Manufacturer X) has since
stopped marketing RBC's and descriptions and evaluations of its equipment
components have been removed herefrom. This section, therefore, covers
only the equipment of the five current manufacturers: Autotrol (now Envi-
rex), Clow, Crane-Cochrane, Lyco, and Walker Process.
Each manufacturer designs the RBC with an individuality that is unique to
its firm. Consequently, RBC's differ from one another in practically every
component used in their assembly, including shafts, plastic media configura-
tions, methods for separating the individual discs, methods of supporting
the plastic media, bearings, and drives. Competition between the manufac-
turers encourages technical innovations that they expect to translate into
marketing advantages. Because of this competition, most of the manufacturers
have progressed through several generations of design, production, and
testing of the individual components that make up their finished products.
In spite of continuing industry-wide efforts to produce a highly cost
effective and reliable process, mechanical failures of RBC equipment have
been experienced at a number of locations (7). Failures have occurred with
shafts, media, bearings, and drives. Many of the causes of these failures
have been corrected by proprietary redesign of the failed component. It is
important to realize that RBC's were placed in service in the United States
following a somewhat abbreviated pilot development program that was not
well suited to assessing the long-term reliability of equipment. Mechanical
treatment plants must operate 24 hr/day, 365 days/yr, and equipment that
appears to be satisfactory during intermittent pilot plant studies may fail
when it is required to operate and perform under "real world" conditions.
This section illustrates and discusses the wide variation in alternatives
available in the makeup of an RBC unit. The developmental status of key
system components is also addressed. It is the responsibility of the
designer to evaluate the equipment options and determine which components
3-1
(62)
-------
and assembly are best suited for his specific application.
3.2 General Equipment Description
RBC's are cylindrical-type structures consisting of plastic media attached
to and/or supported by rotating shafts. The media of four of the manufac-
turers are formed as discs and aligned perpendicular to the shaft. The
media of the fifth manufacturer, Walker Process, are spirally wound onto and
aligned parallel with the shaft. Biomass adheres to and grows on the media,
thus categorizing the RBC process as a fixed film biological process. The
fundamental dimensions of an RBC assembly are shaft length and media diameter.
The shaft length and the disc spacing employed by a particular manufacturer
determines the number of discs on each shaft. Maximum shaft length is struc-
turally limited presently to approximately 27 ft, with 25 ft occupied by
media. Maximum media diameter is limited to 12 ft, set by the maximum allow-
able road vehicle height of 14 ft. These maximum dimensions are representa-
tive of the typical modular RBC unit. The discs are spaced on the shaft
according to various plastic configurations and have a thickness of approxi-
mately 50 mils (0.05 in.). The number of discs per shaft and the diameter
of the discs determine the media surface area available for attached biologi-
cal growth and biochemical reactions. Media surface area is the principal
determinant of the number of RBC units or modules required for a particular
wastewater treatment facility.
The RBC manufacturers offer a variety of shaft lengths less than 27 ft and
media diameters less than 12 ft. These smaller units are utilized where the
surface area requirement is less than that provided by the typical modular
unit and for pilot-scale evaluations.
3.3 Equipment Components
3.3.1 Shafts
RBC shafts are used to support and rotate the plastic media and expose the
plastic surfaces to alternating cycles of wastewater and atmospheric air.
The shafts are fabricated from steel and are covered with a heavy protective
coating suitable for water and high humidity service.
Proper protective coating procedure requires sand blasting of the steel
shaft prior to application to ensure acceptable metal preparation. A coal
tar epoxy is normally used as the protective coating with a minimum film
thickness of 14 mils (0.014 in.). The performance of coal tar epoxy coatings
has, in general, been satisfactory to date. Current data, however, are
inadequate to predict whether these coatings will protect shafts from corro-
sive effects throughout a 20-yr design life.
3-2
(63)
-------
Michanically-driven shafts are generally rotated at a speed of 1.6 rpm, which
equates to a peripheral or tip speed of 60 fpm. The maximum practical speed
of rotation is 2.2 rpm, equivalent to a maximum peripheral speed of 83
ft/min.
Air-drive RBC units operate at rotational speeds of 1.0 to 1.4 rpm, or tip
speeds of 38 to 53 ft/min. The higher rotational speeds are normally employ-
ed at the front end of an RBC train in the first and second stages. The
lower speeds of rotation are used in the latter stages of an RBC train where
oxygen demand is lower.
Each manufacturer designs its own shape, size, and thickness of shaft. The
thickness is governed by structural requirements, and the shape is highly
dependent on the method the manufacturer employs in supporting the plastic
media from the shaft. The five manufacturers each utilize a shaft that dif-
fers from the others in either thickness, size, or shape, or in some cases
all three. Structurally, these differences are readily apparent as identi-
fied in Table 3-1 and shown in Figure 3-1. Lyco currently manufactures
Series 300 circular shafts. The previous Lyco/Hormel Series 200 octagonal
shaft is also included in Table 3-1 and Figure 3-1 because of the large
number of installations still using it.
3.3.2 Media
3.3.2.1 General Description
The heart of the RBC process is the plastic media. The removal of organics
and/or oxidation of ammonia nitrogen are achieved by the rotation of the
TABLE 3-1. SHAFT CHARACTERISTICS
Manufacturer
Autotrol
Clow
Crane-Cochrane
Lyco
Lyco/Hormel
Walker Process
Shape
Square
Round
Round
Round
Octagonal
Round
Size
(in.)
16 x 16
30
30
28
24
30
Thickness
(in.)
1.00
0.625
0.75
0.75
0.75
0.75
3-3
(64)
Section Modulus
(in. 3)
282
415
492
426
344
492
Ref.
1
2
8
4
4
6
-------
CO
I
-pa
CLOW
LYCO
WALKER PROCESS
16-in. square
(outside
dimension)
^
/
Weld
Locations ^.
-1 in.
AUTOTROL
Figure 3-1. Cross-sections of RBC shafts.
CRANE-COCHRANE
outer face
24 in. to
outer face
3/4 in.
LYCO/HORMEL
-------
media through the wastewater, which enables the attached biofilm to contact
the substrate, and then through the air to achieve oxygen transfer into the
film of wastewater and then into the biofilm itself.
When RBC's were first installed in the United States, the technology employed
was a carry-over from Europe where plastic discs were fabricated from expand-
ed polystyrene. Each disc was approximately 0.5 in. thick, and the spacing
between discs was 1.33 in. This type of disc arrangement had an available
surface area of approximately 23,000 sq ft for an 18-ft long shaft and a disc
diameter of 12 ft. In 1972, the polyethylene disc was introduced as a cost
reduction alternative to polystyrene. The major advantage of polyethylene
is its ability to be formed into various corrugations that require a thick-
ness of only 40 to 60 mils (0.04 to 0.06 in.). This innovation enabled
100,000 to 180,000 sq ft of surface area to be provided on the same 27-ft
shaft with 12-ft diameter media. Today, all U.S. manufacturers of RBC's
utilize polyethylene as their plastic media.
Polyethylene is a complex organic compound of the polyolefin polymer family.
This family includes a complete range of low, medium, high, and extra high
density polymers whose specific weight ranges from 57 to 60 Ib/cu ft (7).
The feed stock, catalysts, method of manufacture, and other olefins added
to the polyethylene determine the type and degree of branching, which in
turn determine the physical characteristics of the final polymer. Although
ethylene can be polymerized to form various types of polyethylene, the norm
in current practice is to add butene or hexene to the ethylene to produce
polyethylene co-polymers, which enhance the material's physical properties
(9). The high density polyethylenes (HDPE's) used in present-day RBC units
are all co-polymers.
3.3.2.2 Configurations
Various media configurations or corrugation patterns have been selected by
the manufacturers, each with its own claimed advantages. There are several
reasons for using configurations. Configurations add stiffness to the HOPE
sheets and enable these sheets to be formed with diameters as large as 12
ft. Configurations increase the available surface area by 15 to 20 percent.
Configurations cause the wastewater to follow a tortuous path through the
media, thus increasing wastewater exposure time to the air for greater oxygen
transfer in the atmospheric sector of the rotational cycle. Configurations
also stimulate air turbulence in the atmospheric sector of the cycle, again
leading to improved oxygen transfer. Finally, configurations are used as
spacers to keep the sheets separated.
Each of the above reasons are qualitative; no comparative evaluation has
been conducted to determine quantitatively the incremental advantage of one
configuration over another, or if the advantage can be quantitatively
assessed. The different types of media configurations employed by the manu-
facturers are illustrated in Figure 3-2. All of the manufacturers use
3-5
(66)
-------
^B| lH|^ ^tfHf jBtt^
FROM WALKER PROCESS BROCHURE (6)
FROM AUTOTROL
DESIGN MANUAL (1)
cr>
COURTESY OF LYCO
COURTESY OF CRANE-COCHRANE COURTESY OF CLOW
Figure 3-2. RBC media configurations.
-------
corrugated or honeycomb configurations to achieve increased media surface
area and structural stability.
3.3.2.3 Density
Standard density media are normally used in the front stages of an RBC train.
Standard density media are defined as media with a surface area of 100,000
sq ft supported on or from a 27-ft long shaft in which the media diameter is
approximately 12 ft. By reducing the space required for the repeating plas-
tic configuration by 33 percent, the available surface area can be effec-
tively increased by 50 percent. Media with a surface area of 150,000 sq ft
assembled and supported on or from a 27-ft shaft with a media diameter of
approximately 12 ft is defined as high density media. Some manufacturers
are now also offering media with densities of 120,000 and 180,000 sq ft per
27-ft long shaft for increased design flexibility.
The increase in surface area achieved by reducing the spacing between plastic
sheets is displayed in Figure 3-3. High density media have been used prima-
rily in the middle and final stages of an RBC train. Experience with high
density media at these specific stage locations has been good, both in terms
of biological performance and structural reliability. The increased surface
area has not resulted in increased shaft loadings because biofilm growth is
considerably less in the middle and latter stages than that normally encoun-
tered in the early or front stages. Use of high density media in the lead
stages of an RBC train should be avoided except for second-step nitrifica-
STANDARD
DENSITY
HIGH
DENSITY
Figure 3-3. Autotrol high and standard density media [from Autotrol design
manual (1)].
3-7
(68)
-------
tion applications or where the organic carbon load is known to be very low.
A recommended "rule-of-thumb" is to limit slime thickness on high density
media to 50 mils (0.05 in.) to provide an acceptable margin of safety
against shaft overload and possible fatigue. All of the manufacturers
offer high density media so that the designer is not limited to the selec-
tion of a single supplier.
3.3.2.4 Assembly and Support
RBC manufacturers employ various methods for supporting plastic media from
their shafts. Three of the five manufacturers, Clow, Crane-Cochrane, and
Lyco, rely on a coated steel or stainless steel radial arm system to support
the plastic media. Clow welds rings to the shaft and then bolts a box-like
structure together to contain its plastic media. This technique allows for
field assembly of the plastic media and facilitates easy replacement of a
media section or wedge. Clow's media pack consists of four sections with
eight pie-shaped wedges per section. A schematic of the Clow support as-
sembly is shown in Figure 3-4.
Crane-Cochrane's media are supported by steel arms radiating from rings that
are bolted to clips welded to the shaft and are connected to these arms by
2-in. steel pipes penetrating the discs through pre-formed collars. A mod-
ular unit contains 36-pie shaped wedges divided into six wedges per section
and six sections per shaft as shown pictorially in Figure 3-5.
The Lyco (Series 300) design uses modular, coated steel support ring assem-
blies with radial members that bolt onto the central round shaft. Center
shaft rings are welded to the shaft at and near both ends of the shaft
where stresses are low and are attached by means of a grout adhesive in the
middle portion of the shaft where stress levels are higher. The support
rings are located every 3.5 to 4 ft to divide the media pack into six or
seven sections. Each section is comprised of eight pie-shaped wedges sup-
ported by 24 rods (three per wedge) that are positioned through integrally
molded flanges located in the media. Each section is clamped to the support
rings by elastomer grommets. The Lyco support assembly is depicted diagram-
matically in Figure 3-6.
The Lyco/Hormel (Series 200) design employed modular assemblies of coated
steel rings and radial arms for supporting its plastic media. The rings
were fabricated to fit onto Lyco/Hormel's octagonal-shaped shaft and were
located every 3.5 to 4 ft to divide the media pack into six or seven sections.
Each section was comprised of eight pie-shaped wedges supported by 24 rods
(three per wedge) positioned through the plastic media and carried by the
radial steel arms. The Lyco/Hormel support assembly is depicted diagram-
matical ly in Figure 3-7.
3-8
(69)
-------
*!»•*_ •<
^2>0
^
-------
-—- OJ
--J I
Figure 3-5. Crane-Cochrane media assembly [courtesy of Crane-Cochrane].
-------
-------
MEDIA SEGMENT
SUPPORT RODS
•-^ co
—i i
CO —•
GROMMETS
CENTRAL SHAFT
STUB END
Figure 3-7. Lyco/Hormel (Series 200) media assembly [from Lyco catalog (4)].
-------
In contrast to the radial arm approach, Autotrol employs plastic media hubs
that fit onto its square shaft. The plastic media sheets are thermally
welded to individual plastic hubs and spotwelded to each other to form a
unitized section of 2.5-ft length. As the sections are positioned on the
shaft, the plastic hubs are shimmed in place to prevent movement- Ten sec-
tions are used per shaft to produce a 25-ft media assembly.
Walker Process also does not utilize a radial arm support system. Rather,
plastic media are attached to its shaft with a heavy layer of an epoxy bond-
ing agent and stainless steel strips. The media are spirally wound onto the
shaft in 35-in. wide strips. Each spirally-wound media layer is heat welded
to the preceding layer forming a monolithic structure (see Figure 3-2).
Eight such strips are attached to each shaft to form either the 100,000-sq
ft standard density media contactor or the 150,000-sq ft high density media
contactor.
A report by Chesner and lannone (7) discusses the advantages and early prob-
lems of systems that employ radial support arms and pie-shaped media wedges.
These systems enable the media to be replaced in the field without removing
the shaft. They can also prove to be beneficial in transportation since the
RBC unit can be broken down into smaller components. Earlier designs of
this type had some problems with movement of the media and subsequent damage
as the wedges responded to the alternating forces of gravity and buoyancy.
More recent designs have reduced movement by tightening the media strap
supports or by designing for increased stresses at the interface of the
media and support rods.
In designing an RBC system, the engineer must be cognizant of potential
operation and maintenance ramifications associated with media replacement.
As described above, radial support arm assemblies combined with sections
of pie-shaped media wedges (Clow, Crane-Cochrane, and Lyco) are conducive
to field replacement of media without removal of the shaft from the RBC
tank. The shafts of the other two manufacturers (Autotrol and Walker
Process) must be raised and/or removed to accommodate replacement of the
media. In the case of Autotrol, damaged media sections must be slid off
the end of the shaft. For Walker Process, replacement of the media,
which is epoxy bonded to the shaft, can only be accomplished by the
manufacturer's personnel, either at the site or after return to the factory.
3.3.3 Drives
3.3.3.1 Mechanical Drive Option
Historically, RBC's have been driven mechanically. Without exception, the
mechanical drive packages offered by the manufacturers have provided con-
stant speed capability only. Restricted drive flexibility has traditionally
3-13
(74)
-------
been part of RBC system design philosophy to prevent operator manipulation
of rotational speed and possible process upset. The potential advantages of
variable speed operation, however, suggest that a change in approach to
allow the operator greater latitude in speed selection could improve overall
plant performance by achieving greater control of biofilm thickness and DO
concentration.
As reported by Chesner and lannone (7), RBC manufacturers specify factory-
assembled drive packages for all mechanical drive equipment, consisting of
motors, speed reducers, and drive systems. Reduction of motor output speed
down to approximately 1.6 rpm can be accomplished through the use of various
combinations of multi-V-belts, gear boxes, and chain-and-sprocket units.
Multi-V-belt and chain-and-sprocket units are susceptible to alignment prob-
lems, but those in operation have been relatively trouble free. Gear boxes
are gaining in popularity because of their efficient energy transfer, low
maintenance requirements, and ease of installation and replacement.
The electric motors used for mechanical drive RBC's are normally 3-phase,
60-hertz units. The motors, designed with protective coatings for high
humidity environments, are capable of providing long-term reliable service.
Mechanical drive power requirements are covered in substantial detail in
Section 4. A photograph of a typical mechanical drive package is presented
in Figure 3.8.
Figure 3-8. RBC mechanical drive assembly
(courtesy of Crane-Cochrane).
3-14
(75)
-------
3.3.3.2 Air Drive Option
A recent development offered by Autotrol is the air drive RBC unit. The air
drive assembly consists of plastic cups welded around the outer perimeter of
the media. The cups are 4 or 6 in. deep, depending on stage location, to
accommodate and collect the air flow. An air header is placed below the
media (Figure 3-9), and air is released at a pressure of 3 to 4 psig into
the attached cups, creating a buoyant force that causes the shaft to turn.
Approximately 20 to 30 percent of the air is not captured by the cups and
escapes into the radial passages where it flows upward through the corrugated
media. This air movement tends to aid in shearing excess biomass from the
plastic media, thereby reducing film thickness and corresponding loads on
the shafts.
The operator can vary the rotational speed of air drive shafts by adjusting
the air flow to each stage. The recommended procedure is to reduce speed in
moving from the first to the final stage. Air flow requirements to maintain
various speeds of rotation are covered in greater detail in Section 4.
3.3.4 Bearings
As indicated by Chesner and lannone (7), bearings are one of the components
of an RBC unit requiring periodic maintenance. Some early RBC designs experi
enced deflections of longer shafts causing unequal wearing of the shaft ends
and bearings. The use of self-aligning bearing units appears to have elimi-
nated this condition. Protection from the corrosive effects of wastewater
by the use of high moisture bearings and cover plates (idle-end) has mini-
mized another potential problem.
To permit easy access for lubrication and maintenance, the bearings can be
located outside the media protective covers. Oversized grease cups can be
mounted atop the bearing housings to reduce lubrication manpower. A typical
bearing assembly used for RBC's is shown pictorially in Figure 3-10.
3.3.5 Instrumentation
3.3.5.1 Load Cells
Hydraulic load cells are available and in use for periodically measuring
total shaft load, which in turn can be used to estimate biofilm thickness.
The shaft weighing device consists of a load cell bearing installed on
the idle end of a shaft. A hand-operated hydraulic pump is attached to
the load cell and used to lift the bearing off its base while the shaft
continues to rotate or is momentarily stopped depending on manufacturer.
3-15
(76)
-------
RADIAL PASSAGES-
AIR CUPS
Figure 3-9. Air drive RBC schematic [from Autotrol design manual (1)].
Figure 3-10. RBC shaft bearing assembly (courtesy of Crane-Cochrane).
3-16
(77)
-------
The resulting hydraulic pressure is read from a gauge. This reading can
then be converted to shaft weight (automatically corrected for buoyancy
effects) and/or biofilm thickness.
The electronic strain gauge load cell is a recent development that enables
shaft load to be measured continuously without lifting the idle-end
bearing off its base. A companion weight converter unit is available
that when plugged into this type of load cell provides a direct readout
in pounds, which again is readily convertible to biofilm thickness. An
alarm can be incorporated in the assembly to signal when shaft load
reaches a preset percentage of design load. Electronic strain gauge load
cells are primarily applicable to new installations as it is difficult
and costly to modify existing RBC shaft end walls to accept them.
Since a higher percentage of carbonaceous material is removed in the
initial stages, the first and second stages can be expected to experience
the thickest biofilm growth and heaviest shaft loads in an RBC train. Load
cells strategically located at these and other critical stages can provide
the operator with advance notice of a gradual buildup in biofilm thickness.
Corrective action can then be initiated if shaft design loads are likely to
be exceeded. Possible corrective actions include 1) increasing rotational
speed during periods of low loadings (usually in the early morning hours) to
increase the film shearing force, 2) adding supplemental air for the same
purpose, and 3) increasing media surface area in the affected stage (usually
the first stage) by removing the baffle separating it from the succeeding
stage to equalize load distribution. An upper limit on biofilm thickness of
75 mils (0.075 in.) is recommended for standard density media to provide
adequate protection against shaft overload. The installation and regular
use of load cells should be considered an essential feature of sound RBC
process operating strategy.
3.3.5.2 Flow Control
RBC trains are normally arranged in parallel modules. Historically, the
operator has had to rely on visual observation of semi-continuous weir over-
flows to maintain equal hydraulic loading to each train. Some operators
have expressed concern that lack of positive flow measurement and control
leads to unequal flow distribution and randomly overloaded and underloaded
trains. Flow measuring weirs or other flow measuring devices coupled with
mechanisms for adjusting flow would provide operators with information on
and the capability to maintain proper flow distribution. In addition to
contributing to improved process performance, balanced flow distribution
would automatically tell operating personnel to look elsewhere for the cause
of any process upsets encountered.
3-17
(78)
-------
3.3.5.3 Dissolved Oxygen Monitoring and Control
Strategic DO monitoring is another operational tool that can be used advan-
tageously in RBC process control. The critical locations for monitoring DO
are the first and second stages. Falling stage liquor DO levels, particularly
in these lead stages, may forewarn of changing conditions conducive to the
growth of undesirable Beggiatoa slime (Figure 3-11). Low DO's (less than 2
mg/1) may also indicate a gross increase in either hydraulic or organic
loading or both.
DO instrumentation employed for process control should be serviced frequently
to obtain consistently reliable data. Erroneous readings resulting from
out-of-calibration meters, ruptured membranes, etc., will soon lead to opera-
tor disenchantment with and discarding of the equipment. DO probes should
be placed in tank -locations that are readily accessible, yet representative
of bulk liquid DO concentration. A recommended location is adjacent to the
outlet baffle, one-third of the tank width from the idle-end sidewall, and
one-third of the liquid depth off the floor.
Equally important to furnishing instrumentation that measures DO is providing
the operator with standby cabability to increase DO concentration when neces-
sary. Equipment options that can provide this capability include variable
speed drives to allow the speed of rotation to be increased during periods
of low DO and supplemental air systems that enable compressed air to be
delivered directly to the affected stages. Other alternatives are temporary
hydrogen peroxide addition, stage size rearrangement through removal of
baffles, and step feeding of incoming flow to several stages. The latter
two alternatives, while not adding additional oxygen directly to the waste-
water, can decrease oxygen demand rate in an overloaded stage.
a. Variable Speed Drives
If the designer desires, the manufacturers can modify their standard drive
packages to include variable speed capability. The additional flexibility
thereby afforded would provide the operator with a mechanism to strip excess
biofilm growth from the disc media or increase stage liquor DO. Generally,
the need to reduce biofilm thickness and/or increase DO concentration will
be limited to the first and possibly second stages where biomass growth
rates and oxygen demand are highest. Variable speed drives may also have
application in the latter stages of an RBC train where biofilm growth and
oxygen demand are typically low as a technique for decreasing power consump-
tion through reduced rotational speed.
Acceptable methods for achieving variable speed drive capability for RBC's
include positive infinitely variable (P.I.V.) speed changers and variable
frequency controllers, among others. P.I.V. speed changers are hand-adjusted
3-18
(79)
-------
BEGGIATOA GROWTH
CLOSE-UP VIEW
Figure3-11. First-stage Beggiatoa growth at Columbus, Indiana RBC plant
(courtesy of City of Columbus).
3-19
(80)
-------
units and work on the principal of changing mechanical gear ratios to obtain
desired rotational velocity. Variable frequency controllers enable a.c. mo-
tor speed to be changed directly by varying the frequency of input current.
b. Supplemental Air
Supplemental air represents a second option for transitional correction of
excess biofilm growth or low DO conditions in the lead stages of an RBC
system. Supplemental air in a mechanical drive RBC plant should be used
only when any of the above adverse conditions are encountered. Continuous
use should be discouraged to avoid excessive power consumption.
3.4 Equipment Performance
3.4.1 Overall Structural and Mechanical Reliability
An EPA-sponsored survey and evaluation of design procedures for and process,
operating and maintenance (O&M), equipment, and power performance of munici-
pal RBC treatment facilities was conducted during the period September 1979
through November 1981 (7). A total of 36 plants was included in the survey.
On-site visits and questionnaires were utilized to collect pertinent O&M and
equipment-related data from 17 of the 36 plants. The structural and mechani-
cal performance results of that survey are summarized in Table 3-2.
Equipment performance is separated in Table 3-2 into the three major component
units of an RBC system:
1. Shaft - includes the horizontal support, bearing axles, and
bearings. In the case of structurally supported media,
the radial arms are also considered part of the shaft.
2. Media - the polyethylene sheets, formed into a variety of shapes
whose surface supports the growth of a biological film.
These sheets are either formed to support their own weight
or are supported by steel arms radiating from a shaft.
3. Drive - includes the motor and the speed reduction system that
connects it to the shaft. Speed reduction systems are
composed of various combinations of multi-V-belt reducers,
gear reducers, and chain-and-sproket reducers.
Some of the failures noted are endemic to any mechanical system; others are
major and cannot be considered routine replacement. Only three of the 17
3-20
(81)
-------
TABLE 3-2. STRUCTURAL AND MECHANICAL PERFORMANCE OF RBC PLANTS
00
co
Plant
Location
Rhinelander, Wise.
Wappingers Falls, N.Y.
Fdgewater, N.J.
Washington Twp., N.J.
Selden, N.Y.
Winchester, Ky.
Gloucester, N.J.
Thermopolls, Wyo.
North Muntington, Pa.
Gladstone, Mich.
Theboygan, Mich.
Ionia, Mich.
Hertford, Mich.
Vorhees, Twp. N.J.
Hamilton twp., N.J.
Clevps, Ohio
Ppwaukee, Wise.
No. of
Shafts
10
2
4
3
12
4
4
2
4
6
8
12
?
6
48
6
8
No. of
Years In
Operation
3
5
6
4
5
4
6
2
5
6
1
3
2
2
1
3
9
No. of
Failures
?
--.
4
3
4
2
2
1
2
—
3
...
1
---
—
2
1
...
—
Shaft
Description
Shaft break
—
Radial arms
deteriorating
Radial arms
deteriorating
Seized bear Ings
Shaft break
Worn bearings
Shaft break
Worn bearings
—
Shaft break
—
Shaft break
—
...
Shaft break
Shaft break
...
...
Record of Failures
Media Drive
Age at
Failure
(y)
2
—
|
3
3
4
2
3
5
—
4
...
1
—
...
1
1/2
-__
.__
Corrective No. of Age at Corrective No. of
Action Failures Description Failure Action Failures Description
(yr)
Replace — — — — — —
1 Media shifting 1 Repair 5 Broken
(tighten) drive chain
Replace 4 Hub failures 1 Replace —
Replace 1 Slight media 3 None
shifting as yet
Repair (2) 1 Brlttleness 4 None
Replace (?)
Replace — — — — — —
Replace — — — — — —
Replace
—
Replace — — — --- — —
—
Replace — — — — — —
1 Media shifting 1 Brace 2 Broken
drive chain
Replace 6 Media breaking 1 Replace — —
Replace — — — — —
— — —
1 Electrical
1 Drive chain
Age at Corrective
Failure Action
— —
1 Repair
...
.--
...
-_.
—
1 Replace
...
—
—
8 Repair
6 Replace
-•• .... . .
Reproduced in part from Chesner and lannone (7).
-------
plants surveyed reported no failures with any of the three major component
units.
3.4.1.1 Failures of the Shaft Component
The most serious equipment problem that can impact an RBC plant is shaft
failures. A shaft failure involves a structural break in the horizontal member
itself, the loss of the unit, and damage to a portion of the media. Repair
requires that the damaged unit be removed and that a new shaft be installed
along with salvaged and/or new media. Depending on the site layout, this may
entail the removal of protective media covers or relocation of the entire shaft
assembly outside a process building.
Shaft failures were reported for seven of the 17 plants listed in Table 3-2.
None of the replacement shafts have experienced any failures to date. Of the
12 total shaft failures reported, all but one were located in the first stage
of the respective treatment train, suggesting that excessive biofilm growth is
contributing to overloadings of shaft carrying capability (7).
Another identified cause of shaft failures has been poor welding practice.
The municipal RBC plant of the City of Columbus, Indiana, has experienced this
type of failure (10). This plant, not included in Chesner and lannone's sur-
vey (7), uses biodiscs for secondary treatment and nitrification and is the
site of a cooperative RBC research project between EPA and the City.
The Columbus facility consists of 10 trains with eight shafts per train for a
total of 80 shafts. The plant has experienced six shaft failures to date
(September 1982), five in the first stage and one in the second stage. All
six failures occurred at or near the same shaft location, approximately 20 ft
from one end of the shaft. The cognizant RBC manufacturer investigated these
failures and has attributed them to poor welding practice. In fabricating the
shafts, 20-ft welding backup plates followed by additional 5-ft plates were
used to span the 25-ft media segment of the shaft. The failures occurred at
the juncture of the two plates in all six cases. The practice of employing
discontinuous welding backup plates has since been terminated by the manufac-
turer in favor of uninterrupted 25-ft backup plates. * Thus far, this welding
procedural change has proven to be successful in eliminating shaft failures
due to welding practice.
Bearings fail in RBC units when not properly maintained, when shaft deflections
cause undue stress on the bearings, and when broken seals allow wastewater
onto bearing surfaces eventually leading to corrosion. Replacement is per-
formed with the shaft in place and is estimated to be a 2-day operation for
two men (7). Eight bearings in three of the plants surveyed have been re-
placed as indicated in Table 3-2.
3-22
(83)
-------
Radial arm media support systems are used by three of the five manufacturers
currently marketing RBC's in the United States. These systems have undergone
design modifications to improve their structural integrity. As shown in Table
3-2, two of the 17 plants surveyed have had to replace these support members.
3.4.1.2 Failures of the Media Component
As noted by Chesner and lannone (7), media failures can occur directly due
either to degradation of the HOPE from prolonged exposure to heat, concen-
trated organic solvents, or ultraviolet radiation or to breakage resulting
from the excessive weight load of heavy biofilm growth. Plastic media can
also suffer indirect or incidental breakage if the radial arm support system
permits excessive media movement.
Six of the 17 RBC plants listed in Table 3-2 reported having had problems with
the media component, including hub failures, shifting media, media brittleness,
and media breakage from unspecified causes. Corrective action ranged from
simple bracing to replacement of the entire media pack.
Media failures due to excessive shifting are being combated by the manufacturers
by radial arm support system redesign and recommended inspection programs that
specify more frequent tightening of the support rods. Ultraviolet rays in
sunlight can lead to media brittleness. Precautions should be taken to avoid
this problem either by adding anti-oxidatants to HOPE formulations or providing
covers to shield the media from exposure to ultraviolet radiation.
Degradation from exposure to organic solvents is also an area where little
difficulty should be expected or encountered. In a consultant report prepared
for EPA, Landsberg (9) states that although HOPE is chemically inert to most
wastewater ingredients, including dilute acids and bases, human wastes, house-
hold bleaches, detergents, soaps, and a large variety of concentrated acids
and bases, certain aromatic hydrocarbons (organic solvents) such as toluene,
methyl ethyl ketone, and gasoline have a deleterious effect on the material.
He goes on to note, however, that lack of resistance of HOPE to aromatic hydro-
carbons is not considered serious in municipal wastewater treatment since it
would be rare for municipal wastewater to contain significant concentrations
of these chemicals. Even when an occasional spill occurs, the concentration
of the aromatic compounds would be diluted by the time they reached the plant
and the contact period with any individual element of media sufficiently short
so that no serious problems with HOPE degradation would be anticipated.
A serious and little understood cause of media failure is breakage due to
stress cracking. Landsberg (9) reported that media failures from stress
cracking have occurred as little as 6 to 9 mo following startup. Excerpts of
his discussion relative to this problem area are presented below.
3-23
(84)
-------
Stress cracking is an internal or external rupture in the plastic caused by
tensile stresses of lesser magnitude than the short-term mechanical strength
of the material. Although the exact cause is not known, the suspected mechan-
ism involves propagation of a crack through the weak chemical bonds of the
material. The weakest link of high density crystalline polyethylene is along
the crystallite or sperulite boundaries. Any conditions leading to stress and
strain along these boundaries would lead to lower stress crack resistance.
Stress cracking is highly dependent on the nature and level of the applied
stress and most readily occurs under the influence of high multi-axial stress-
ing. If, in addition to this type of stressing, a hostile environment is
encountered, the rate and severity of the stress cracking are magnified.
Landsberg (9) notes that the literature indicates that HOPE can withstand many
chemicals when not under stress, but that under stress the same materials have
a severe deleterious effect.
In RBC units, multi-axial stressing occurs to an appreciable extent. First,
the RBC media is rotating slowly. Second, as the biofilm increases in weight,
the polyethylene is subjected to increased loads, which result in added stress-
ing. Finally, with some units, a number of media packs are bolted together,
which generates stresses in a plane perpendicular to the direction of rotation.
The various methods of attaching the plastic media to the shaft determine the
stresses imposed on the media and should be taken into consideration when
estimating expected media life.
Another factor that should be considered in evaluating the expected life of RBC
media is the difference of the coefficients of expansion of HOPE and steel.
Steel has a coefficient of expansion of approximately 6-8 in./in./°C x 10~6,
whereas the plastic has a coefficient of expansion of 1.3 in./in./°C x 10~4.
This difference in the coefficients of expansion can lead to problems when the
plastic components are in direct contact with the metal shaft and support arm
structures.
In summary, the integrated media-shaft units form a rather complex mechanical
system whose expected life depends not only on the properties of the plastic
components and the shaft carrying capability, but also on the structural
integrity of the radial arm media support system (Clow, Crane-Cochrane, and
Lyco) or the interaction of the shaft and media at their interface (Autotrol
and Walker Process). All of these factors should be assessed and taken into
account in comparing the expected life cycles and cost effectiveness of RBC
equipment.
3.4.1.3 Failures of the Drive Component
According to Chesner and lannone (7), the drive assembly is the most reliable
component of a mechanical drive RBC system. Only three of 17 plants surveyed
reported problems with mechanical drives (Table 3-2), and these were minor.
Operating experience with air drive systems is too limited to date to make
3-24
(85)
-------
definitive projections of equipment performance. The more critical aspects of
air drive performance would appear to be air distribution to multiple shafts
and maintaining uniform rotational speed rather than reliability of the air
delivery equipment. Lack of positive rotational control with air drive systems
has become severe enough in several cases to completely stop rotation.
3.5 Evaluation of Shaft Design Procedures
The wide variation in shaft design procedures has aroused considerable concern
as to the adequacy of RBC shafts to provide 20 or more yr of reliable service.
This concern was precipitated first by reported shaft failures at industrial
wastewater treatment installations followed by similar observations at several
municipal RBC facilities. As indicated in Section 3.4.1.1, most of the re-
ported shaft failures to date have been associated with the lead stages of RBC
trains where the heaviest biofilm growths are normally encountered.
Assuming satisfactory welding practice, the most likely cause of shaft failure
is fatigue. Under EPA sponsorship, Drs. R. M. Bowman and J. T. Gaunt of Purdue
University, were retained to 1) examine the various fatigue design techniques
employed by the manufacturers of RBC shafts, 2) recommend a preferred code or
design technique, and 3) conduct a fatigue behavior evaluation for various
shaft cross-sectional shapes representative of those used by the several manu-
facturers. Pertinent excerpts from their report (11) are given below.
3.5.1 Design Criteria and Codes
3.5.1.1 General Considerations
Most building and bridge design codes for structural steel use a similar design
methodology to account for fatigue resistance. Extensive experimental testing
of laboratory specimens has shown that the logarithm of fatigue life, N, is
linearly related to the logarithm of the range of applied stress, SR, between
certain limits of the cyclic life, as shown in Figure 3-12 for Categories A and
B of the American Welding Society (AWS) Structural Welding Code (12) for Design
of Tubular Structures (Section 10) and Category C of the AWS Welding Code (12)
for Design of New Bridges (Section 9). The stress range at which a curve
becomes flat is known as the fatigue (or endurance) limit and represents the
value of the applied stress range for which no appreciable fatigue damage
occurs under atmospheric conditions. A member subjected to rotating/cyclic
loads with a maximum corresponding stress range less than the fatigue limit
will not fail (within a certain high level of confidence) under these condi-
tions.
Geometry, detailing, and fabrication quality can each have a significant effect
on the fatigue resistance of a member. Transverse or longitudinal welding,
3-25
(86)
-------
100
50
20
cr
c/^
LU
o
z
Cfl
C/3 in
LJLJ 10
DC
I-
C/)
O
_l
o
o
FATIGUE LIMIT FOR
CATEGORY A (example)
NOTE: Curves A and B from Reference 12,
Section 10, Figure 10.7.4;
Curve C from Reference 12,
Section 9, Figure 9.4B
I I I 1 1 III
I I I
I III
I I I.
Mill I I I
INI
04 105 106 1Q7 1Q
CYCLES OF LOAD, N
Figure 3-12. Allowable fatigue stress ranges for three stress categories of
redundant structures in atmospheric service.
3-26
(87)
-------
welded attachments, bolting, and weld toe treatment are geometrical factors
that influence fatigue behavior. Since structural detailing and weld fabri-
cation quality can affect fatigue behavior, the weldment fabrication provisions
of the AWS Structural Welding Code (12) should be adopted as an absolute mini-
mum requirement. Sketches of appropriate joint details and fabrication pro-
visions are fully treated in Reference 12.
The particular S-N curve that describes the fatigue resistance of a given mem-
ber or detail must be used to properly design for fluctuating loads. To con-
servatively account for the inherent scatter of fatigue data, a lower confi-
dence level curve, usually 95 percent, is generally used to describe fatigue
resistance instead of the best-fit S-N curve.
Obviously, it is very important to select the S-N curve that best describes the
structural member being designed. If a family of S-N curves from a particular
code is utilized in designing for fatigue, in lieu of costly definition of an
S-N curve from exhaustive experimental laboratory testing, careful subjective
judgement must be exercised in selecting the appropriate S-N curve.
3.5.1.2 Fatigue Design Provisions
Fatigue provisions of various design codes used to proportion tubular members
are not all similar. The fatigue provisions of the AWS (12) Section 10 (Tubu-
lar Structures) and Section 9 (New Bridges), American Association of State
Highway and Transportation Officials (AASHTO) (13), American Institute of
Steel Construction (AISC) (14), and American Railway Engineering Association
(AREA). (15) codes were considered herein for RBC shaft design. Most of the
AISC, AREA, and AWS (New Bridges) fatigue provisions, however, are based on
the AASHTO fatigue requirements. Further, since the AWS (New Bridges) fatigue
provisions are identical to the AASHTO fatigue requirements, in-depth code
evaluation was limited to a detail-by-detail comparison of the fatigue design
provisions of the AWS Section 10 (Tubular Structures) and AWS Section 9 (New
Bridges) codes. Only typical RBC structural details were considered.
Although an RBC shaft is a non-redundant system, failure would not likely
result in loss of life. Therefore, AWS fatigue provisions for redundant load
path structures were utilized in this comparative design evaluation of RBC
shafts. These provisions are believed to be more representative of actual
shaft fatigue behavior.
a. Plain Unwelded Tubes
The fatigue resistance of plain unwelded tubes is dictated primarily by tube
geometry. For tubes without sharp corners or notches, the use of the S-N
curve for circular tubes given by AWS (Tubular Structures) Category A (see
Figure 3-12) is recommended for fatigue-resistant shaft design.
3-27
(88)
-------
b. Tubes with Longitudinal Weld Seams
The fatigue resistance of tubes with longitudinal weld seams or continuously
welded longitudinal attachments is governed by tube geometry and weld location.
For welds placed in the flat portions of rectangular or octagonal tubes, the
use of the S-N curve for circular sections given by AWS (Tubular Structures)
Category B (see Figure 3-12) is recommended for fatigue-resistant shaft design.
c. Tubes with Transverse Weld Attachments
The fatigue resistance of tubes with transverse weld attachments is governed
primarily by the attachment detail. The fatigue provisions of the AWS (Tubu-
lar Structures) and AWS (New Bridges) design codes for transverse weld attach-
ments are significantly different. The S-N curves for AWS (Tubular Structures)
Category C2, transverse ring stiffeners, and Category D, miscellaneous attach-
ments, are notably lower than the S-N curve for AWS (New Bridges) Category C,
welded attachments with detail dimension parallel to the direction of stress
less than 2-in. The use of the S-N curve given by AWS (New Bridges) Category
C (see Figure 3-12) is recommended for fatigue-resistant shaft design in this
case, based primarily on the background and supporting data for the three
curves.
AWS (Tubular Structures) Categories C2 and D are limited to situations in
which nominal member stresses represent actual load transfer across the weld
(16). The loads on RBC shafts, however, are transferred in-plane through the
welds, rather than out-of-plane across the attachment welds.
AWS (New Bridges) Category C, on the other hand, is intended for situations
in which the attachment is used to transfer in-plane loads or to stiffen the
cross-section. The position and endurance limit of the S-N curve for AWS
(New Bridges) Category C are based on numerous laboratory fatigue tests re-
ported in Reference 17 for steel beams with welded transverse attachment
lengths less than 2 in. parallel to the direction of stress. The AWS (New
Bridges) Category C endurance limit is also consistent with recent fatigue
data reported for transverse welded attachments on A588 steel beams (18).
d. Tubes with Bolted Attachments
Little direction is available for tubes with bolted attachments. AWS (Tubular
Structures) Category B represents a reasonable lower bound for fatigue data
reported by Stallmeyer (19), Osman (20), and Wilson et al. (21) for plain
plate mild steel specimens containing a central circular notch. Therefore,
the use of the S-N curve given by AWS (Tubular Structures) Category B (see
Figure 3-12) is recommended for fatigue-resistant design of shaft tubes with
bolted attachments.
3-28
(89)
-------
3.5.1.3 Service Conditions
The curves A, B, and C in Figure 3-12 are basic guidelines for shaft design
based on structural detail. Two extraneous conditions may alter these
curves.
Under constant cycle loading, a fatigue limit (crack growth threshhold)
exists for any given shaft. Three fatigue limits are illustrated in Figure
3-12 by the horizontal portions of Curves A, B, and C. This is not the
condition under variable random loadings where the evidence is that the S-N
relationship continues to extend into the lower stress range regions.
The effects of corrosion would also tend to devalue fatigue limits.
Obviously, the extent of devaluation would depend on the severity of the
environment.
For random variable (wave action) loadings and a sea water environment, the
American Petroleum Institute (API) (22) has suggested extending S-N curves
for fixed offshore platforms at the same slope (similarly to the extended
dashed portions of curves A, B, and C in Figure 3-12), with no fatigue
limits in the region of those cycle numbers representing the expected life
of RBC systems (20 yr).
RBC shafts typically are not subjected to random variable loadings, as
defined by API (22), during operation. Regardless of the structural detail
category of the shaft, it is essential that imposition of random variable
loadings also be avoided during maintenance activities.
The RBC manufacturers recognize that shafts must be protected from possible
loss of fatigue strength due to the corrosive effects of wastewater. Epoxy
coatings are routinely used by the manufacturers to protect shafts from
potential corrosion. Available information from coating manufacturers
indicates that epoxy coating materials have an expected life of from 5 to
over 10 yr, although long-term field data are limited. In selecting or
specifying RBC shafts, the designer should require evidence from the shaft
manufacturer that satisfactory fatigue stress protection, including protec-
tion from possible corrosive effects, has been provided for the design life
of the system under expected loading and environmental conditions.
After consideration of the protection offered by an epoxy coating, the
designer may feel that a depression or lowering of a particular fatigue
limit associated with curve A, B, or C in Figure 3-12 is warranted or that
the stress curve should be extended below the fatigue limit at some
intermediate slope. There is no compelling evidence to indicate, however,
that the allowable stress for an RBC shaft operating in municipal wastewater
3-29
(90)
-------
would need to be lowered to the degree obtained from the extended dashed
lines in Figure 3-12.
3.5.1.4 Summary of Stress Category Recommendations
The stress categories recommended for fatigue-resistant design of RBC shafts
are summarized in Table 3-3. One design code, AWS, can be utilized by speci-
fying the S-N curve appropriate for a given RBC shaft detail.
3.5.2 Estimated Shaft Fatigue Performance
3.5.2.1 Design Parameters
The expected fatigue lives for various shaft geometric cross-sections were
estimated based on the recommended fatigue stress categories summarized in
Table 3-3. Once the design stress for a given cross-section was determined,
the range of applied alternating stress resulting from shaft rotation was
used in conjunction with the design stress category (Table 3-3) to estimate
the corresponding fatigue life. Each rotation of the shaft represents one
loading cycle, and the fatigue life equals the number of loading cycles until
failure. Geometric shapes evaluated included a 30-in. diameter circular shaft,
a 16-in. diameter circular shaft, a 16-in. square shaft, and a 24-in. octagonal
shaft. A number of shaft wall thicknesses were considered for each geometric
cross-section.
Stress analysis of each shaft section was conducted using consistent design
criteria. The 12-ft diameter plastic media were assumed to be 40 percent
submerged in wastewater. Both standard density RBC media (100,000 sq ft
surface area/27-ft shaft length) and high density RBC media (150,000 sq ft
surface area/27-ft shaft length) were considered with uniform attached slime
thicknesses ranging in 25-mil (0.025-in.) increments from clean media to
150 mils (0.15 in.). The specific gravity of the attached biomass was assumed
to be 1.0.
A typical 27-ft long shaft was assumed to be simply supported with a 26.5-ft
span length from centerline to centerline of the bearings. The maximum bending
moment was computed assuming that 1) a hub weight of 200 Ib was positioned 6
in. from the center of the support, 2) a uniform shaft weight acted along the
entire span length, and 3) a uniform effective media weight acted along the
middle 25 ft of the shaft. The uniform effective media weight was computed on
the basis of 60 percent of the combined uniform total media and attached slime
weights to account for the impact of buoyancy. A total media weight (neglect-
ing buoyancy) of 10,000 Ib was assumed for standard density RBC media, while a
weight of 15,000 Ib was assumed for high density RBC media. Total media weight
includes the plastic media weight and media superstructure support weight.
3-30
(91)
-------
TABLE 3-3. RECOMMENDED STRESS CATEGORIES FOR RBC TUBULAR SHAFTS
RBC
Stress
Category
Type of
Shaft Design
Corresponding
Stress Category from
Design Standards
(Atmospheric Conditions)
AWS(TS)*
AHS(NB)t
I Plain unwelded tubes
II Tubes with longitudinal seam
and/or continuously welded
longitudinal stiffeners
III Tubes with transverse ring
stiffeners or miscellaneous
attachments fastened by
fillet welds with length in
direction of stress less than
2 in.
IV Tubes with miscellaneous
attachments fastened by
bolting or tapping
* Fatigue provisions in Section 10 (Design of Tubular Structures) of
Reference 12.
t Fatigue provisions in Section 9 (Design of New Bridges) of Reference 12.
The maximum bending stress for a given shaft cross-section was computed by
dividing the maximum bending moment by the corresponding section "modulus.
Because an RBC shaft is subjected to both tension and compression during each
rotational cycle, the stress range was then obtained by doubling the maximum
bending stress. Maximum principal stresses were not utilized since torsional
shearing stresses for the closed sections were very small.
3.5.2.2 Matrix of Estimated Shaft Lives
Estimated shaft fatigue lives for the several cross-sectional shapes investi-
gated are summarized in Tables 3-4 to 3-7 as a function of detail category,
structural shape thickness, media density, and attached biological slime
thickness. Loadings in which the resulting range of applied stress is less
than the endurance (fatigue) limit are denoted in the tables as FL. As noted
earlier, a member subjected to an applied stress range below the fatigue
3-31
(92)
-------
TABLE 3-4. ESTIMATED FATIGUE LIFE FOR 30-IN. DIAMETER CIRCULAR SHAFT
Standard Density Media
Structural
Shape Estimated Life (yr) at Indicated Slime Thickness (mils)*
Detail Thickness
Category (in.)
II 0.500
0.625
0.750*
0.875
III 0.500
0.625**
0.750tt
0.875
0
Fit
FL
FL
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
FL
FL
FL
FL
FL
75
FL
FL
FL
FL
FL
FL
FL
FL
100
FL
FL
FL
FL
5.1
FL
FL
FL
125
FL
FL
FL
FL
2.8
5.4
FL
FL
150
5.0
FL
FL
FL
1.7
3.3
5.6
FL
High Density Media
Structural
Shape Estimated Life (yr) at Indicated Slime Thickness (mils)*
Detail Thickness
Category (in.)
II 0.500
0.625
0.750*
0.875
III 0.500
0.625**
0.750tt
0.875
0
Fit
FL
FL
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
FL
FL
FL
FL
FL
75
FL
FL
FL
FL
3.0
FL
FL
FL
100
3.9
FL
FL
FL
1.5
2.8
4.7
FL
125
1.7
4.3
FL
FL
0.8
1.5
2.6
4.0
150
0.9
2.2
4.5
FL
0.5
0.9
1.6
2.5
* 1 mil = 0.001 in.
f FL = fatigue limit
* Walker Process
Clow
tt Crane-Cochrane
**
3-32
(93)
-------
TABLE 3-5. EXPECTED FATIGUE LIFE FOR 28-IN. DIAMETER CIRCULAR SHAFT
Standard Density Media
Structural
Shape Estimated Life (yr) at Indicated Slime Thickness (mils)1
Detail Thickness
Category
II
III
(in.)
0.500
0.625
0.7504=
0.875
0.500
0.625
0.750
0.875
0
Fit
FL
FL
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
FL
FL
FL
FL
FL
75
FL
FL
FL
FL
FL
FL
FL
FL
100
FL
FL
FL
FL
2.71
FL
FL
FL
125
4.41
FL
FL
FL
1.51
2.87
FL
FL
150
2.24
FL
FL
FL
0.91
1.76
2.95
4.50
High
Density
Media
Structural
Shape Estimated Life (yr) at Indicated Slime Thickness (mils)*
Detail
Category
II
III
Thickness
(in.)
0.500
0.625
0.7504=
0.875
0.500
0.625
0.750
0.875
0
FL
FL
FL
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
FL
4.13
FL
FL
FL
75
4.81
FL
FL
FL
1.61
3.06
FL
FL
100
1.77
4.31
FL
FL
0.77
1.48
2.49
3.81
125
0.78
1.92
3.92
FL
0.42
0.82
1.38
2.13
150
0.39
0.97
1.99
3.61
0.25
0.49
0.84
1.30
* 1 mil
t FL =
* Lyco
= 0.001 in.
fatigue limit
Series 300
3-33
(94)
-------
TABLE 3-6. ESTIMATED FATIGUE LIFE FOR 16-IN. SQUARE SHAFT
Standard
Detail
Category
I
II
Structural
Shape
Thickness
(in.)
0.875
1.00
1.125
0.875
1.00*
1.125
Estimated
0
FLt
FL
FL
FL
FL
FL
Life
25
FL
FL
FL
FL
FL
FL
Density
(yr) at
50
FL
FL
FL
FL
FL
FL
Media
Indicated
75
FL
FL
FL
FL
FL
FL
Slime
100
FL
FL
FL
5.6
FL
FL
Thickness (mils)*
125 150
FL 2.4
FL FL
FL FL
2.5 1.3
4.0 2.0
5.8 3.0
High Density Media
Detail
Category
I
II
Structural
Shape
Thickness
(in.)
0.875
1.00
1.125
0.875
1.00*
1.125
Estimated
0
FLt
FL
FL
FL
FL
FL
Life
25
FL
FL
FL
FL
FL
FL
(yr) at
50
FL
FL
FL
FL
FL
FL
Indicated
75
FL
FL
FL
2.8
4.3
6.3
Slime
100
1.7
FL
FL
1.0
1.6
2.4
x
Thickness (mils)*
125 150
0.6 0.2
1.1 0.4
1.8 0.7
0.5 0.2
0.7 0.4
1.1 0.5
* 1 mil = 0.001 in.
t FL = fatigue limit
t Autotrol
3-34
(95)
-------
TABLE 3-7. ESTIMATED FATIGUE LIFE FOR 24-IN. OCTAGONAL SHAFT
Standard Density Media
Structural
Shape Estimated Life (yr) at Indicated Slime Thickness (mils)*
Detail
Category
II
III
Thickness
(in.)
0.625
0.750*
0.875
0.625
0.750
0.875
0
Fit
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
FL
FL
FL
75
"~ FL
FL
FL
FL
FL
FL
100
FL
FL
FL
3.1
5.1
FL
125
5.0
FL
FL
1.7
2.9
4.3
150
2.6
5.1
FL
1.1
1.8
2.7
High
Density
Media
Structural
Shape Estimated Life (yr) at Indicated Slime thickness (mils)*
Detail
Category
II
III
Thickness
(in.)
0.625
0.750*
0.875
0.625
0.750
0.875
0
Fit
FL
FL
FL
FL
FL
25
FL
FL
FL
FL
FL
FL
50
FL
FL
FL
4.7
FL
FL
75
5.5
FL
FL
1.8
3.1
4.6
100
2.0
4.1
FL
0.9
1.5
2.3
125
0.9
1.8
3.3
0.5
0.8
1.3
150
0.5
0.9
1.7
0.3
0.5
0.8
* 1 mil = 0.001 in.
t FL = fatigue limit
* Lyco/Hormel Series 200
3-35
(96)
-------
limit will not fail (with a high level of confidence) as a result of structural
fatigue.
The estimated shaft lives presented in Tables 3-4 to 3-7 were calculated
based on an assumed rotational speed of 1.6 rpm or 840,000 revolutions/yr,
representative of a typical mechanical drive RBC unit. A shaft rotating at
1.2 rpm or 630,000 revolutions/yr, more representative of an air drive RBC
unit, would have an extended estimated life of 1.6/1.2 = 1.33 times that
projected for the same shaft design operating at 1.6 rpm.
Examination of these tables indicates that a shaft equipped with high density
media would be expected to reach its fatigue or endurance limit at a lower
biofilm thickness than the same shaft operating with standard density media.
This difference is attributable, of course, to the nominal 50 percent higher
specific surface area of high density media. Accordingly, high density media
should not be specified for use in the early stages of an RBC train, with the
possible exceptions of Walker Process and Lyco whose shafts are projected to
have fatigue limits equal to or greater than those loads resulting from bio-
film 125 mils (0.125 in.) and 100 mils (0.1 in.) thick, respectively, on high
density media. Of the other three current shaft designs, all are estimated
to have satisfactory fatigue resistance with high density media up to slime
thicknesses of at least 50 mils (0.05 in.) and two up to at least 75 mils
(0.075 in.). Based on these estimates, a conservative upper biofilm thick-
ness limit of 50 mils (0.05 in.) is recommended for high density media, with
the exceptions of Walker Process and Lyco as stated before.
High density media has been used advantageously in the middle and latter
stages of RBC trains where decreased availability of organic carbon and the
frequent occurrence of nitrification combine to produce biofilms character-
istically less than 50 mil (0.05 in.) thick. A recommended general practice,
subject to site specific adjustment, is not to employ high density media
before the third stage of a sequentially-staged RBC module.
Current manufacturer shaft designs used in conjunction with standard
density media are projected (Tables 3-4 to 3-7) to be resistant to failure
from fatigue at slime thicknesses ranging from at least 75 mils (0.075
in.) to 150 mils (0.15 in.) or greater. Again, the Walker Process and
Lyco designs are estimated to have higher load carrying capabilities than
the other three shafts. Allowing for exceptions based on engineering
judgement and the use of higher load capacity equipment, it is recommended
that when biofilm thickness on standard density media reaches 75 mils
(0.075 in.), corrective action be initiated to maintain an adequate safety
margin against shaft overstressing. In some instances, biofilm growth in
the initial stages (primarily stages 1 and 2) may naturally equilibrate
at thicknesses greater than 75 mils (0.075 in.). Depending on the
particular shaft design being used, it may then become critical for the
plant operator to implement contingency procedures to either 1) shear
biomass from the media through temporarily-increased rotational speed or
3-36
(97)
-------
supplemental air stripping as discussed in Section 3.3.5.2 or 2) reduce
the organic loading to the affected stage(s) by step feeding part of the
incoming flow to other stages or removing stage-separation baffles to
modify load distribution. Although standard density media can be and has
been specified for an entire RBC facility, their most beneficial usage
today is in the lead stages of an RBC train. Not only can shafts outfitted
with standard density media better tolerate the higher biofilm thicknesses
encountered there, but DO deficiencies are less likely to occur. The
higher volumetric concentration and accompanying higher overall demand
for oxygen of biomass growing on high density media could more readily
exceed the oxygen transfer capacity of media operating in the first two
stages.
3.5.3 Summary of Conclusions
In summarizing their report, Bowman and Gaunt (11) reiterated the following
major conclusions of their analysis of fatigue design criteria for RBC shafts:
1. Shafts should be proportioned in accordance with the fatigue design
stress categories presented in Table 3-3.
2. Use of appropriate fatigue provisions of the AWS Structural Welding
Code will provide for adequate fatigue-resistant shaft design.
3. Shaft structural detail will dictate whether fatigue provisions of
the AWS Structural Welding Code from the section on Design of Tubular
Structures or the section on Design of New Bridges should be used.
4. Weld fabrication quality should satisfy the provisions of the AWS
Structural Welding Code as an absolute minimum requirement.
5. A matrix of estimated shaft fatigue lives is presented in Tables 3-4
to 3-7 as a function of biofilm thickness for the several shaft geo-
metric cross-sections evaluated.
6. The above estimates of shaft fatigue life assume no shaft deteriora-
tion due to corrosion.
3.6 References
1. Autotrol Wastewater Treatment Systems Design Manual. Autotrol Corporation,
Bio-Systems Division, Milwaukee, Wisconsin, 1978.
2. Clow Envirodisc Rotating Biological Contactor Systems Catalog. Clow
Corporation, Florence, Kentucky, 1980.
3-37
(98)
-------
3. Crane-Cochrane Rotating Biological Contactor Brochure. Crane Company,
Cochrane Environmental Systems, King of Prussia, Pennsylvania, 1981.
4. Lyco Wastewater Products - RBS Systems Catalog. Lyco Division of
Remsco Associates, Marlboro, New Jersey, 1982.
5, RBC systems brochure from Manufacturer X, 1978.
6. Walker Process BioSpiral Rotating Biological Contactors Brochure. Walker
Process Corporation, Aurora, Illinois, 1979.
7. Chesner, W. H. and J. lannone. Review of Current RBC Performance and
Design Procedures. Report prepared for USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, under Contract No. 68-02-2775
by Roy F. Weston, Inc., (Publication pending).
8. Personal communication from R. W. Hankes, Crane Company, Cochrane
Environmental Systems, King of Prussia, Pennsylvania, to R. C. Brenner,
USEPA, Cincinnati, Ohio, August 25, 1982.
9. Landsberg, M. I. Rotating Biological Surface Waste Water Treatment
Units. Internal report prepared for USEPA under consultant agreement,
Cincinnati, Ohio, August 1981.
10. Opatken, E. J. Personal observations at Columbus, Indiana RBC plant
during field monitoring of cooperative research and development project.
USEPA Cooperative Agreement No. CR807463, 1981.
11. Bowman, M. D. and J. T. Gaunt. Fatigue Behavior of Rotating Biological
Contactor Shafts. Internal report prepared for USEPA under Purchase
Order No. C2159NAST, Cincinnati, Ohio, April 1982.
12. American Welding Society. Structural Welding Code-Steel. 5th Edition,
ANSI/AWS Dl.1-81, 1981.
13. American Association of State Highway and Transportation Officials.
Standard Specifications for Highway Bridges. 12th Edition, 1977.
14. American Institute of Steel Construction. Specifications for the Design,
Fabrication, and Erection of Structural Steel for Buildings. 1978.
15. American Railway Engineering Association. Specifications for Steel Rail-
way Bridges. 1981.
16. Marshall, P. W. and A. A. Torpac. Basis for Tubular Joint Design.
Welding Journal, 53(5):192s-201s, May 1974.
17. Fisher, J. W., P. A. Albrecht, B. T. Yen, D. J. Klingerman, and B. M.
McNamee. Fatigue Strength of Steel Beams with Welded Stiffeners and
Attachments. National Cooperative Highway Research Program Report No.
147, 1974.
3-38
(99)
-------
18. Friedland, I. M., P. A. Albrecht, and G. R. Irwin. Fatigue of Two Year
Weathered A588 Stiffeners and Attachments. Journal of the Structural
Division, ASCE, 108(ST1):125-144, January 1982.
19. StaHmeyer, J. E. The Effect of Stress History on Cumulative Damage in
Fatigue. Ph.D. Dissertation, University of Illinois at Champaign-Urbana,
1953.
20. Osman, H. A. Evaluation of Local Plastic Strain History in Fatigue.
Ph.D. Dissertation, University of Illinois at Champaign-Urbana, 1966.
21. Wilson, W. M. and F. P. Thomas. Fatigue Tests of Riveted Joints.
Engineering Experiment Station Bulletin No. 302, University of Illinois
at Champaign-Urbana, May 1938.
22. American Petroleum Institute. API Recommended Practice for Planning,
Designing, and Constructing Fixed Offshore Platforms. 12th Edition,
API RP 2A, January 1981.
3-39
(100)
-------
SECTION 4
POWER CONSUMPTION
4.1 Introduction
Power consumption in mechanically driven systems results from the power
used to overcome internal resistances and losses in the motor, friction
losses in gear reduction and drive belts or chains, friction losses in
shaft support bearings, and the drag forces resulting from rotation of the
media and attached biofilm through the wastewater. The drag forces are
affected by the amount of media surface area, the shape of that surface
area, rotational speed, wastewater viscosity, and the type and amount of
biological growth. Since each manufacturer uses media with somewhat
different shapes and different media support structures or orientation of
media openings, some variation in power requirements among the various
units can reasonably be expected.
With air driven systems, compressed air is discharged beneath the RBC media
as shown in Section 3. The rising air is captured by the air cups, and the
resulting buoyant forces provide the torque necessary for media rotation.
'Power is required in air driven systems for losses in the motor, compressor,
air headers, control valves, and diffusers and also to overcome the static
head of wastewater in the RBC tank.
4.2 Fundamentals
Power measurements from a number of mechanically and air driven RBC plants
are presented later in this section. Before these data are reviewed, it is
appropriate to set forth some fundamental considerations related to power
measurement and power consumption for those who may not be familiar with
this information.
Several basic alternatives must be considered in the design of any power
distribution system. The relatively high cost of electrical power today is
placing additional pressures on the designer to make power distribution
systems as energy efficient as possible. Induction motors constitute the
major power demands in most wastewater treatment plants and are available
at different operating voltages; the most common operating voltages are
4-1
(101)
-------
120, 208, 240, 480, and 600 volts. Induction motors are also available for
single-phase or polyphase service.
The selection of the operating voltage of an induction motor is an
important decision. As the operating voltage increases, the current
decreases, which directly reduces the line losses in the system and permits
the use of smaller conductors. Comparative information for various 7.5-hp
motor options is presented in Table 4-1. The full-load current for single-
phase operation is significantly higher than for the three-phase options
with commensurately higher wire costs. A substantial difference in wire
cost is also apparent between three-phase, 240- and 480-volt operation, but
the difference between the three-phase, 480- and 600-volt options is less
significant. Based on capital and operating costs, three-phase, 480-volt
operation is the industry standard.
For a balanced Y or A circuit, the actual power, kW, being drawn by a three-
phase motor is computed by the equation given below:
kW = ~\/T El (cos 0/1000) (4-1)
where kW is the kilowatts drawn, E is the line voltage, I is the current in
amperes, and 9 is the phase angle between the voltage and the current.
The apparent power, kVA, for the same motor is given by:
kVA - Y3~ (El/1000) (4-2)
TABLE 4-1. COMPARISON OF FULL-LOAD CURRENT, MINIMUM COPPER WIRE
SIZE, AND WIRE COSTS FOR VARIOUS 7.5-HP MOTOR OPTIONS
No. of
Phases
1
1
3
3
3
EMF
(volts)
120
240
240
480
600
Full -Load
Current
(amps)
80
40
28
14
11
Minimum Copper
Wire Size
(AW6)
1
6
8
12
14
Representative
Wire Cost*
($/1000 ft)
437.90
198.27
130.73
51.29
37.24
* Wire costs are based on 7-strand THW.
4-2
(102)
-------
The power factor, cos 9, can then be calculated as:
Power Factor = Actual Power (4_3)
rower ractor Apparent Power
The power factor can also be read directly by using a power factor meter.
It is important to recognize that the polyphase wattmeters used to meter
power consumption measure actual power, kW, drawn, but that demand charges
are based on apparent power, kVA. A power factor less than 0.9 will result
in higher power bills because of increased demand charges, and many electric
utilities have penalty charges for customers with low power factors.
Low power factors result from the operation of equipment with inductive
loads that require reactive power. Reactive power, kvar, is essentially
lost in the system and, consequently, is synonymous with higher power costs.
The power factor of an induction motor decreases, increasing the relative
kVAR, as the load decreases. Additionally, the efficiency of an induction
motor decreases as the load decreases. The variations in both power factor
and efficiency as a function of load are shown in Table 4-2. It is clear
that the needless oversizing of electric motors is costly both in terms of
capital and operating expenses.
When uncorrected, the power factors at wastewater treatment plants are
usually low, typically ranging from 0.4 to 0.7. Most electric utilities
have demand charge schedules that penalize customers with power factors
less than 0.9; therefore, power factor correction can be a very important
factor in reducing operating costs at wastewater treatment plants.
Field investigations were made on one train each at the LeSourdsville and
Upper Mill Creek (Butler County, Ohio) RBC treatment plants. Both plants
TABLE 4-2. CHANGE IN POWER FACTOR AND MOTOR
EFFICIENCY AS A FUNCTION OF LOAD
FOR 7.5-HP POLYPHASE INDUCTION MOTOR
Percent of
Full -Load
25
50
75
100
Power
Factor
0.37
0.70
0.78
0.82
Efficiency
(percent)
72.0
86.7
87.9
87.7
4-3
(103)
-------
employ mechanically driven RBC units equipped with power factor correction
capacitors, which permitted measurement of both the uncorrected and cor-
rected power factors. These data and other pertinent information for
LeSourdsville and Upper Mill Creek are presented in Tables 4-3 and 4-4,
respectively. The last column in both tables is the difference in apparent
power, kVA, before and after power factor correction. Typically, a 2.5-kVA
reduction in apparent power was noted at both plants as a result of power
factor correction. However, the data obviously indicate that additional
power factor correction would be beneficial for the LeSourdsville plant.
Assuming demand charges of $6.92/month/kVA, power factor correction in the
range illustrated in Tables 4-3 and 4-4 will reduce monthly operating costs
$15 to $20 per shaft. Typical list prices (third quarter 1982) for 1-, 2-,
and 5-kvar capacitors are $94, $99, and $143, respectively. Based on an
approximate installation cost of $100, the payout period for a 2.5-kVA
power factor correction is about 12 months. Depending on the original
power factor and the degree of correction possible, the payout period could
vary from a few months to several years.
Significant variation was observed in both current draw and power factor at
both plants. The variability was attributed to out-of-balance shafts due
to uneven biofilm growth. The required power increased as the heavier
segment of the RBC rotated to the top of the arc and decreased significantly
as the heavier section rotated to the bottom.
Other factors directly affect the power consumed to operate mechanical
drive RBC installations. If properly maintained, the drive train, con-
sisting of gear reducers, chains and sprockets, and V-belts and pulleys,
can be reasonably efficient. A good quality gear reduction unit should
have an efficiency of 95 to 97 percent, and a chain-and-sprocket drive
should be about 95 percent efficient, provided neither the chain nor
sprockets are badly worn. A V-belt drive with new, matched belts and with
proper tension will have an efficiency of 93 to 94 percent. Under optimum
conditions, therefore, the drive train should operate at approximately 86
percent efficiency with respect to power transmission. If appreciable wear
occurs in any of the three components, efficiency will decrease with the V-
belt drives being most susceptible to wear.
4.3 Air Drive Systems
4.3.1 Basic Relationships
Air driven RBC's are proprietary systems manufactured by the Autotrol
Corporation. A typical air driven RBC unit contains 24 longitudinal rows
of inverted air trays with 11.5 trays per row attached around the periphery
of the media. Each tray is 2-ft long by 14-in. wide and is partitioned
4-4
(104)
-------
TABLE 4-3. RESULTS OF RBC MECHANICAL DRIVE POWER MEASUREMENTS AT THE
LESOURDSVILLE WASTEWATER TREATMENT PLANT, BUTLER COUNTY, OHIO
Field Measurements at Motor
IT j*
0 I
en tn
Drive
No.t
11
12
13
14
15
Rotational
Speed
(rpm)
1.52
1.68
1.63
1.61
1.52
Disconnect
Measurements at Motor Starter
with Power Factor Correction*
Apparent Actual
EMF
(volts)
476
476
476
476
476
Current
(amps)
9.5 - 9.7
5.4 - 6.3
5.2 - 8.8
6.8 -10.0
9.2 - 9.5
Power
Factor
<0.1 - 0.33
0.18 - 0.50
0.12 - 0.64
0.23 - 0.66
0.34 - 0.40
Power
(kVA)
7.91
4.82
5.77
6.93
7.71
Power
(kW)
<1.70
1.64
2.19
3.08
2.85
EMF
(volts)
478
478
478
478
478
Current
( amps )
5.9 - 6.3
2.5 - 3.7
2.2 - 5.9
4.3 - 8.3
9.4 - 9.7
Power
Factor
0.15 - 0.41
0.45 - 0.74
0.26 - 0.84
0.31 - 0.77
0.33 - 0.37
Apparent
Power
(kVA)
5.05
2.57
3.35
5.22
7.91
Actual
Power
(MO
1.41
1.53
1.84
2.82
2.69
Difference
in
Apparent
Power
(kVA)
2.86
2.25
2.42
1.71
-0.20
* Power factor corrected with 2-kvar Westinghouse Dyna-Vac capacitors.
t Drives 12 and 13 are 5-hp units; all other motors are 7.5-hp units; all motors are 480 V, 60 Hz, 3 phase, 1140 rpm, TEFC, S.F. = 1.15;
shafts 11 and 12 are equipped with standard density media; shafts 13 to 15 are equipped with high density media.
-------
TABLE 4-4. RESULTS OF RBC MECHANICAL DRIVE POWER MEASUREMENTS AT THE UPPER
MILL CREEK WASTEWATER TREATMENT PLANT, BUTLER COUNTY, OHIO
Field Measurements at Motor
Drive
No.t
Bl
82
B3
B4
B5
B6
87
Rotational
Speed
(rpm)
1.52
1.50
1.52
1.49
1.52
1.50
1.50
EMF
(volts)
478
478
478
478
476
472
476
Current
(amps)
5.8 - 5.9
5.8 - 6.0
5.8 - 6.9
6.2 - 7.6
6.0 - 7.7
6.0 - 6.9
5.8 - 7.7
Power
Factor
0.40 - 0.45
0.33 - 0.47
0.34 - 0.64
0.55 - 0.66
0.60 - 0.66
0.50 - 0.62
0.26 - 0.67
Disconnect
Apparent
Power
(kVA)
4.84
4.88
5.26
5.75
5.65
5.27
5.57
Actual
Power
(kW)
2.06
1.95
2.58
3.48
3.56
2.95
2.59
EMF
(volts)
480
480
480
476
480
480
476
Measurements at Motor Starter
with Power Factor Correction*
Current
(amps)
2.2 - 2.7
1.9 - 3.2
2.0 - 4.6
5.3 - 6.9
2.8 - 5.3
2.9 - 4.4
1.8 - 5.2
Power
Factort
0.99L - 1.00
0.98L - 0.99L
0.97L - 1.00
0.99L - 1.00
0.99 - 0.98L
0.99L - 1.00
0.99 - 1.00
Apparent
Power
(kVA)
2.04
2.12
2.74
5.03
3.37
3.03
2.89
Actual
Power
(kW)
2.04
2.10
2.70
5.00
3.34
3.02
2.88
Difference
in
Apparent
Power
(kVA)
2.80
2.76
2.52
0.72
2.28
2.24
2.68
* Data on size and manufacturer of power factor correction capacitors not available.
t All motors are 7-1/2 hp, 480 V, 60 Hz, 3 phase, 1165 rpm, TEFC, S.F. = 1.15; shafts Bl to B3 are equipped with standard density media;
shafts 84 to B7 are equipped with high density media.
T L indicates leading.
-------
into multiple air cups. Normally, 4-in. deep air cups are employed,
although 6-in. deep cups are available where increased torque is needed.
The radius of air driven RBC media is reduced from 5 ft-11 in. to 5 ft-7 in,
so that attachment of the 4-in. air cups results in the same 11 ft-10 in.
media diameter as with Autotrol's standard mechanically driven unit. The
available angle, a, over which the buoyant force can act with air drive
media is calculated below:
14 in.
12 in.
sin 9i
^WATER
SURFACE 0
sin 92
92
a
1/5.5833 - 0.1791
10.3°
1.1667/5.5833 = 0.2089
12.06°
90° - (9i + Q2) = 67.64°
AIR INTRODUCED
The 14-in. wide air trays provide 28 lineal ft around the media periphery
(24 rows x 1.1667 ft) to capture rising air bubbles. When multiplied by
shaft length, this peripheral footage represents approximately 75 percent
of the circumferential area of an 11 ft-10 in. diameter cylinder. Assuming
a 2-ft total gap between air trays on the longitudinal axis decreases the
total circumferential area available to trap rising air bubbles to roughly
69 percent of the surface area presented by a 25-ft long, 11 ft-10 in.
diameter cylinder.
As the orientation of the air cup openings changes during rotation through
the angle a, the volume of initially entrapped air bubbles that can be
retained within the cups continually decreases. The effective cup volume
per 4-in. deep tray at the 10.3° air introduction angle, QI, is 0.653 cu
ft (1). The total effective volume of air that can be trapped at a rotation-
al speed of 1 rpm is 180 cu ft (276 trays x 0.653 cu/t/tray). If air was
uniformly distributed through the 47 diffusers in the air distribution
system for a 25-ft shaft, the minimum total air volume that would need to
be supplied to fill the 4-in. air cups at the point of air introduction is
240 cu ft (180 cu ft/0.75), assuming no air is lost between trays within
each of the 24 rows. The static wastewater head in the air cup at the
point the air is introduced is normally 2.06 psig; therefore, to fill each
cup at the point of air introduction requires 240(14.7 + 2.06)/14.7 or
273.6 scfm at 1 atm. barometric pressure and a wastewater temperature of
68° F.
4-7
(107)
-------
If each cup is immediately filled at the point of air introduction, the
total torque around the sector bounded by angle a is 7355 ft/lb and the
work that would be performed at a rotational velocity of 1 rpm would be
1.40 hp. Autotrol (2) recommends that centrifugal blowers be designed for
at least 3-psig output and positive displacement blowers be designed for a
somewhat higher pressure rise because of losses from the inlet and outlet
silencers. At 68°F and 3 psig, 273.6 scfm is equivalent to 3.34 adiabatic
hp. If the compressor and motor have a combined efficiency of 50 percent
for a 3-psig air supply, the overall power efficiency per horsepower of
work performed when each cup is filled immediately at the point of air
introduction is 21.0 percent [1.407(3.34/0.50)]. Even if the air flow were
cut in half and the system rotated at 1 rpm, the output work of 0.91 hp
would only represent an overall power efficiency of 27.2 percent [0.91/
(1.67/0.50)] at the same assumed 50-percent combined efficiency for the
compressor and motor. These calculations illustrate the well known fact
that an air drive system cannot compete in energy efficiency with a mechan-
ical drive system at equal rotational speeds. In actuality, air driven
systems are designed to rotate from 1.0 to 1.3 rpm versus the recommended
1.6 rpm for mechanical systems and, hence, an actual energy comparison
between air and mechanical systems depends on the overall system designs.
4.3.2 Blower Design
Autotrol's design manual (2) recommends that 250 cfm of blower capacity be
provided for each shaft. According to Sullivan (1), these recommendations
have been revised and the current recommendation calls for 350 cfm of op-
erating capacity plus additional standby capacity. Maximum normalized air
flow at ambient conditions (68°F) with 4-in. air cups for various rotational
velocities is presently indicated by the manufacturer (3) to be as follows:
Air Flow (acfm*) at Indicated
Rotational Speed
Rotational speed (rpm) 1.0 1.2 1.4
Standard density media lOOt 150t 220t
High density media 145t 225t 350t
* acfm = actual cfm
t Wastewater temperatures below 68°F or use of 6-in. air cups
will increase these values.
The power requirements for air drive plants are more difficult to predict
than those for mechanical drive plants because the designer has a more
direct impact on power consumption. The sizing and design of the air
distribution system will directly affect the line losses and, therefore,
the discharge pressure at which the compressor must operate. In designing
4-8
(108)
-------
the distribution system, the designer is faced with the classical trade-off
of higher capital costs for the distribution system versus reduced power
costs resulting from decreased line losses. The usual industrial practice
is to design for headlosses in the range of 0.1 to 0.6 psi/100 ft of line.
Design of centrifugal compressors is a rather precise technology, and most
equipment manufacturers can provide units that are relatively efficient at
the design conditions. It is extremely important that the design conditions
stipulated by the engineer match actual operating conditions as closely as
possible. Compressor operation at conditions other than those for which
the compressor was designed can result in substantial loss of efficiency.
Figures 4-1 and 4-2 are modified performance curves for a centrifugal com-
pressor extrapolated from typical performance curves normally supplied by a
vendor, i.e., the volume versus discharge pressure and the horsepower versus
pressure curves. It is apparent in Figure 4-1 that the input horsepower
required is a nearly linear function of the inlet air flow. The difference
in input power requirements for 2.6- and 3.3-psig discharge pressures is
minimal.
Figure 4-2 illustrates why a disproportionately small reduction in input
power is achieved at reduced operating pressures. The efficiency of the
compressor shown decreases rapidly as the output pressure drops below the
design point of 3.3 psig. This characteristic of centrifugal compressors
requires that the designer carefully analyze system operation to minimize
power consumption and coordinate the design process with the compressor
manufacturer.
Consider an example of 200 ft of 16-in. steel pipe designed to carry 3200
scfm (Point A), which would produce friction losses of 0.8 psi or roughly
24 percent of the blower discharge pressure of 3.3 psig. If the compressor
inlet is then throttled to back the air flow down to 1600 scfm, the friction
losses in the line would drop to 0.14 psi. This action would decrease the
compressor discharge pressure 0.66 psi below the design point to approximate-
ly 2.6 psig (hypothetical Point B), and the compressor's efficiency would
be reduced from 72 percent to about 42 percent. Consequently, the attempt
to save power by a 50-percent throttling of the compressor would have
resulted in only a 19-hp or 33-percent decrease in power consumption.
For those facilities in which significant variations in air flow require-
ments are anticipated, consideration should be given to using multiple
compressors with different design points, both for air flow and discharge
pressure. An additional recommendation is that curves similar to those
developed in Figures 4-1 and 4-2 be required submittal data from the manu-
facturer along with the standard performance curves.
4-9
(109)
-------
POINT A,
60
55
cr 50
LU
O
Q.
5 45
40
35
1000 1400 1800 2200 2600 3200 3600
INLET AIR FLOW (scfm)
Figure 4-1. Typical centrifugal compressor horsepower versus air flow curve.
c
0)
O
80
70
60
O
z
LU
o 50
LU
U_
LU
40
30
POINT A
2.6 psig
-h
POINT B
1000 1400 1800 2200 2600 3200 3600
INLET AIR FLOW (scfm)
Figure4-2. Typical centrifugal compressor efficiency versus air flow curve.
4-10
(110)
-------
4.3.3 Field Measurements of Air Flow Versus Rotational Speed
As previously discussed, the power consumption for an air drive RBC system
can only be accurately estimated once the motor and blower characteristics,
line losses, and operating range have been addressed. The single-valued
curves of air flow versus rotational speed for standard and high density
media in the Autotrol design manual (2) were developed from measurements at
several pilot- and full-scale installations. These curves fall within the
range of values that were summarized in several internal Autotrol documents
made available to the writers.
Since information on speed versus air flow relationships is essentially
nonexistent to the design community at large, a study was made at two air
drive installations (Lower East Fork and Indian Creek) in the Cincinnati
area. Flexible hoses and a sharp-edged orifice plate inserted in a special-
ly constructed flanged PVC pipe apparatus were used to tap into air supply
headers and measure total air flow to individual RBC shafts. The apparatus
was transported intact to and calibrated by the Cox Instrument Company of
Detroit, which certified the accuracy of the measuring device to having
less than 3 percent error.
The main characteristics of the two plants studied and the results obtained
are presented in Table 4-5 and Figures 4-3 and 4-4. The Autotrol design
curves for high density media (3) are superimposed on Figure 4-3 for 6-in.
deep air cups with assumed heavy biological growth and 4-in. deep cups with
assumed normal biological growth for wastewater temperatures of 50 and
53.6°F, respectively. The Autotrol design curves for standard density
media (3) are superimposed on Figure 4-4 for heavy growth at a wastewater
temperature of 50°F and normal growth at 53.6°F. In general, the manufac-
turer's recommended design curves underestimate the actual air requirements
measured. While data from two plants are not sufficient to describe the
rotational responses that may prevail at other RBC air drive installations,
they do indicate that the recommended design relationships are not always
applicable. Furthermore, the load cell readings for the Lower East Fork
plant do not indicate that the biofilm being carried by these units is
excessive. The first-stage units at Indian Creek run with the inlet air
valves wide open at all times. The rotational responses of the first-stage
units at the highest air flow rates measured reflect,throttling back of the
other stages beyond their normally throttled conditions during the period
of air flow measurement. This plant is operating at 40 percent of design
flow. These results are discussed further in Section 5.
4-11
(HI)
-------
TABLE 4-5. SELECTED CHARACTERISTICS OF LOWER
EAST FORK AND INDIAN CREEK PLANTS
Parameter
Shaft length
Media density
Air cups
No. of Trains
Shafts per train
Shaft numbers measured
Lower East Fork Indian Creek
Air Cup
Curve Designation Depth
in Figures 4-3 and 4-4 (in.)
A
B
C
D
E
F
G
H
I
J
6
6
6
4
4
4
4
4
4
4
4 in
Stage
1
1
1
3
5
9
1
2
3
1
25 ft 20 ft
High Standard
. and 6 in. 4 in.
3 2
9 3
1,3,5,9 1,2,3
Date
Measured Biofilm Thickness (in.)
(1982) Measured* Observed**
2/11 0.050
3/15 0.044
3/17 0.042
3/15,3/17 0.038,0.037
2/23 0.045
2/23 - A/B
3/2 - C
3/11 - B
3/2 - A
3/11 - C
* From manufacturer's load cell calibration curve.
** A = light growth; B = normal growth; C = heavy growth.
4-12
(112)
-------
500
400
300
o
CO
DC
<
200
100
LEGEND:
Autotrol curve for high
density media, heavy
growth, 6-in. cups, and
50° F
— —Autotrol curve for high
density media, normal
growth, 4-in. cups, and
53.6° F
*Letter designations from Table 4-5
0.5 0.75 1.0
ROTATIONAL SPEED (rpm)
1.25
Figure 4-3. Summary of air flow versus rotational speed measurements made
at Lower East Fork-Little Miami River Regional Wastewater
Treatment Facility, Clermont County, Ohio.
4-13
(113)
-------
400
300
_
, i
E
«+—
o
tr
<
200
100
LEGEND:
Autotrol curve for std.
density media, heavy
growth, 4-in. cups, and
50° F
— -^Autotrol curve for std.
density media, normal
growth, 4-in. cups, and
53.6° F
NOTE: Autotrol design curves
lowered 20 percent to
yield values equivalent to
20-ft long shafts
*Letter designations
from Table 4-5
0.5
0.75 1.0
ROTATIONAL SPEED (rpm)
1.25
1.50
Figure4-4. Summary of air flow versus rotational speed measurements made
at Indian Creek Wastewater Treatment Plant, Hamilton County,
Ohio.
-------
4.4 Mechanical Drive Systems
4.4.1 Clean Media Power Measurements
Clean media power measurement data for mechanically driven RBC units manu-
factured by Clow and Autotrol Corporations have been reviewed and summarized
by Chesner and lannone (4). The results of these clean media tests
indicated the following:
1. The power required to rotate the media increased as a cubic
function of the rotational velocity.
2. Power losses from the motor, drive, and gear box were small
in relation to the power required to overcome the drag imposed by
the media.
3. The ratio of power consumed to media surface area at a given
rotational velocity was reasonably constant for both the Clow and
Autotrol units. The particular ratio observed varied with the
manufacturer.
4. Power consumption for units with rotational velocities of 1.4 to 1.7
rpm varied from 1.56 to 2.04 hp (1.16 to 1.52 kW) for shafts with
100,000 sq ft of media surface area and from 1.98 to 2.60 hp (1.48
to 1.94 kW) for shafts with 150,000 sq ft of media surface area.
According to Sullivan (1), the power requirements for field applications
is not a perfect cubic function of the rotational velocity for all mechani-
cal drive RBC systems. Other energy-consuming factors exist, independent
of speed, that can lead to a less-than-cubic relationship between power
and speed. Accordingly, the cubic relationship should not be used indis-
criminately.
4.4.2 Field Power Measurements
Field measurements of the power requirements for mechanical drive RBC systems
were made by Environmental Resources Management, Inc. (ERM) for EPA's Office
of Water Program Operations using an electric power/demand analyzer (5), by
the Roy F. Weston Company for EPA's Municipal Environmental Research Labora-
tory using polyphase wattmeters (4), and by EPA staff personnel using an
industrial power analyzer (6). The Roy F. Weston data were challenged by
the manufacturers. Since these data could not be subsequently verified,
they are not considered in this document. Selected results of the ERM and
EPA staff studies, however, are summarized in Table 4-6.
4-15
(115)
-------
TABLE 4-6. POWER MEASUREMENTS FOR MECHANICAL DRIVE RBC PLANTS
CT>
Motor
Plant Site/ Stage Size
WW Temp. Measured (hp)
Cheyney, Pa./
t°F
Pennsville,
N.J./68°F
King of Prussia,
Pa. (Matsunk
Plant)/61°F
Philadelphia,
Pa. (Northeast
Plant)/54°F
Marshall ,
Wisc./52°F
1
2
1
1
2
2
3
3
1
2
3
3
5
1
1
1
1
1
1
1
1
2
3
5
5
7.5
7.5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
Biofilm
Growth
M
L
H
H
M
M
M
M
L
L
L
L
L
M
M
M
M
M
M
M
M
M
L
Surface
Area
(sq ft)
128
165
120
120
150
150
180
180
100
100
150
150
150
100
100
100
100
100
100
100
100
100
150
,250
,750
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
Observe<
Speed
(rpm)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
i Recorded
Power
(kW)
1.21
1.22
2.60
2.48
2.32
2.00
1.78
1-64
1.64
1.64
2.39
2.31
2.27
2-33
2.53
2.05
2.83
2.41
2.12
2.33
2.48
2.04
2.15
(hp)
1.62
1..64
3.49
3.33
3.11
2.69
2.39
2.20
2.20
2.20
3.21
3.10
3.04
3.12
3.39
2.75
3.80
3.23
2.84
3.12
3.33
2.73
2.88
Recorded
Power
Factor
0.26
0.26
0.59
0.59
0.57
0.54
0.50
0.48
0.36
0.36
0.52
0.48
0.46
0.62
0.65
0.58
0.71
0.66
0.61
0-61
0.48
0.43
0.43
Manufacturer/
Reference
Lyco/5
Lyco/5
Walker
Process/5
Clow/5
Walker
Process/5
(continued)
-------
TABLE 4-6. (continued)
Motor
Plant Site/ Stage Size
WW Temp. Measured (hp)
Canonsburg, Pa.
(Canonsburg-
Houston
Plant)/61°F
St. Clairsville,
Ohio/62°F
Fairmont, W.Va./
62°F
1
1
2
3
3
3
4
5
5
1
1
2
3
3
4
5
1
1
2
3
4
4
5
5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
Biofilm
Growth
M
M
M
L
L
M
L
L
L
M
M
M
L
L
L
L
M
M
M
M
M
t
L
t
Surface
Area
(sq ft)
100
100
100
100
100
100
100
150
150
100
100
100
150
150
150
150
100
100
100
100
100
100
100
100
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
Observec
Speed
(rpm)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
\ Recorded
Power
(kW)
1.68
2.00
2.32
1.45
1.57
1.77
1.36
1.71
1.92
1.85
3.13
2.10
2.67
1.95
2.21
2.13
1.80
1.77
1.62
1.48
1.45
1.60
1.45
1.63
(hp)
2.25
2.68
3.11
1.94
2.11
2.37
1.82
2.29
2.57
2.48
4.20
2.82
3.58
2.61
2.96
2.86
2.41
2.37
2.17
1.98
1.94
2.15
1.94
2.19
Recorded
Power Manufacturer/
Factor Reference
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.28 Autotrol/5
.31
.38
.26
.27
.29
.24
.30
.35
.37 Clow/5
.54
.42
.50
.40
.45
.42
.53 Clow/5
.53
.51
.51
.50
.52
.49
.53
(continued)
-------
TABLE 4-6. (continued)
00
Plant Site/
WW Temp.
Mount Pleasant
Mich./58°F
Holt, Mich.
(Delhi
Charter Twp.
*. Plant)/58°F
i
00 Birdsboro, Pa.
60°F
Johnson Creek,
Wisc./47°F
Lake Mills,
Wisc./55°F
Stage
Measured
, 1
2
3
4
5
6
1
1
1
2
/ 1
1
2+4
2+4
2+4
1+4
1+4
1
2
2
3
4
5
Motor
Size
(hp)
5
5
5
5
5
5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
5
5
5
5
5
5
Biofilm
Growth
H
H
M
M
M
L
H
H
H
H
L
L
L
L
L
M
M
M
M
t
L
L
L
Surface
Area
(sq ft)
120,000
120,000
120,000
170,000
170,000
170,000
100,000
100,000
100,000
100,000
100,000
100,000
138,000
138,000
138,000
125,000
125,000
100,000
100,000
150,000
150,000
150,000
150,000
Observed
Speed
(rpm)
1.5
1.5
1.5
1.5
1.5
1.5
1.7
1.7
1.7
1.7
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Recorded
Power
(kW)
2.06
1.81
1.81
3.08
2.76
2.81
2.28
2.33
2.99
2.16
1.95
1.62
1.56
1.85
1.83
2.39
2.38
2.19
2.16
3.02
2.40
2.29
2.09
(hp)
2.76
2.43
2.43
4.13
3.70
3.77
3.06
3.12
4.01
2.90
2.61
2.17
2.09
2.48
2.45
3.21
3.19
2.94
2.90
4.05
3.22
3.07
2.80
Recorded
Power
Factor
0.58
0.55
0.54
0.68
0.67
0.66
0.35
0.35
0.45
0.35
0.21
0.18
0.16
0.22
0.19
0.82
0.82
0.46
0.43
0.53
0.46
0.43
0.41
Manufacturer/
Reference
Walker
Process/5
Autotrol/5
Autotrol/5
Clow/5
Walker
Process/5
(continued)
-------
TABLE 4-6. (continued)
Plant Site/
WW Temp.
East Washing-
ton, Pa.
(WEWJA
Plant)/57°F
LeSourdsville,
Ohio/t°F
Cincinnati,
Ohio (Upper
Mill Creek
Plant)/t°F
*L = growth <
A = visually
->• indicates
t indicates
Stage
Measured
1
1
1
2
3
4
5
6
6
6
1
2
3
4
5
1
2
3
4
5
6
7
0.031 in
Motor
Size Biofilm
(hp)
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
.; M =
light growth; B
multiple
stages
not observed or
Growth
M
M
M
M
M
L
L
L
L
L
t
t
t
t
t
C/B
C/B
B
B
B
A
A
growth of 0
= visually
per shaft.
recorded.
Surface Observed Recorded
Area Speed Power
(sq ft) (rpm)
100
100
100
100
100
150
150
150
150
150
100
100
150
150
150
100
100
100
150
150
150
150
.031-0.
normal
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
063 in.;
growth ;
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
H
C
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.66
.66
.54
.46
.50
.5
.5
.5
.5
.5
.5
.5
_
=
(kW)
1.80
2.27
2.23
2.78
2.09
2.80
2.19
2.71
2.61
2.50
2.55
2.11
3.69
3.80
3.63
2.9
2.7
2.9
3.1
2.6
2.7
2.2
growth of
visually
(hp)
2.41
3.04
2.99
3.72
2.80
3.75
2.94
3.63
3.50
3.35
3.42
2.83
4.95
5.10
4.87
3.89
3.62
3.89
4.16
3.49
3.62
2.95
Recorded
Power Manufacturer/
Factor Reference
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.064-0.125
heavy
growth
.25 Autotrol/5
.32
.34
.43
.32
.40
.30
.38
.41
.36
.21 Autotrol/6
.34
.38
.44
.37
.42 Clow/6
.40
.49
.60
.63
.56
.46
in.;
-------
Four of the five current RBC manufacturers are represented in Table 4-6.
Shaft lengths of all units reported in this table were confined to a range
of 23 to 27 ft, while media diameters varied from 11 to 12 ft. Rotational
speeds for most units were measured at 1.5 rpm; several units, however,
were clocked as high as 1.8 rpm.
According to Chesner and lannone (4), manufacturers' estimates of power
requirements for mechanical drive RBC's range from 2.7 to 3.4 hp/shaft (2.0
to 2.5 kW/shaft) for standard density media. The manufacturers' estimated
range for high density media is 3.5 to 4.2 hp/shaft (2.6 to 3.1 kW/shaft).
The 80 drive units field tested by ERM that were selected for inclusion
in Table 4-6 by the authors plus the 12 units evaluated by EPA had a mean
power requirement of 2.98 hp/shaft (2.22 kW/shaft) with a standard devia-
tion of 0.71 hp/shaft (0.53 kW/shaft). The highest recorded power usage
was 5.10 hp/shaft (3.80 kW/ shaft) at LeSourdesville, Ohio, and the lowest
was 1.62 hp/shaft (1.21 kW/ shaft) at Cheyney, Pennsylvania.
Of the above 92 shaft and drive assemblies, a total of 55 were equipped
with media having a surface area of 100,000 to 128,250 sq ft- Media having
a surface area in this range are generally regarded in the industry as
representing standard density media. The media surface area of the other
37 units varied from 138,000 to 180,000 sq ft, a range characteristic of
high density media. The average measured power consumption for the stand-
ard density units was 2.80 hp/shaft (2.09 kW/shaft) with a standard devi-
ation of 0.62 hp/shaft (0.46 kW/shaft). For the high density units, the
average power requirement recorded was 3.22 hp/shaft (2.40 kW/shaft) with
a standard deviation of 0.79 hp/shaft (0.59 kW/shaft). The above values
agree well with the manufacturers' estimates of mechanical drive power
requirements for standard density media and are slightly lower than the
manufacturers' estimates for high density media.
When field-measured power levels exceed the means indicated above for
standard and high density media by one to two standard deviations or more,
the operator should investigate whether the higher power consumption is
being caused by equipment problems, heavier-than-normal biofilm growth,
or both. Potential equipment problem areas include improper alignment,
inadequate lubrication, excessive rotational speed, excessive belt tension
or belt slippage, and general wear and deterioration of the drive components,
4.5 References
1. Personal communication from R. A. Sullivan, Autotrol Corporation,
Milwaukee, Wisconsin, to J. A. Heidman, USEPA, Cincinnati, Ohio,
April 14, 1982.
4-20
(120)
-------
2. Autotrol Wastewater Treatment Systems Design Manual. Autotrol
Corporation, Bio-Systems Division, Milwaukee, Wisconsin, 1978.
3. Aero-Surf Energy Requirements. Autotrol Corporation internal report,
Milwaukee, Wisconsin, April 1, 1981.
4. Chesner, W. H. and J. lannone. Review of Current RBC Performance and
Design Procedures. Report prepared for USEPA, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, under Contract No. 68-02-277b
by Roy F. Weston, Inc., (Publication pending).
5. Environmental Resources Management, Inc. Evaluation of the Energy
Requirements for Rotating Biological Contactors (RBCs). Report
prepared for USEPA, Office of Water Program Operations, Washington,
D.C., under Contract No. 68-01-6622, (Publication pending).
6. Heidman, 0. A., W. W. Schuk, and A. C. Petrasek. Field Measurements
of Power Consumption and Air Flow at RBC Installations. Internal re-
port, USEPA, Municipal Environmental Research Laboratory, Test and
Evaluation Facility, Cincinnati, Ohio, May 5, 1982.
4-21
(121)
-------
SECTION 5
DESIGN CONSIDERATIONS
5.1 Introduction
RBC systems can be employed for organic removal, nitrification, organic
removal plus nitrification, and/or denitrification. Several design approaches
are available to achieve the above objectives, including the use of pilot
plant studies, mathematical models, and empirial procedures, all of which are
discussed in subsequent sections. Pilot studies should be conducted where
economic considerations and feasibility warrant such efforts or where atypical
municipal wastewater characteristics are anticipated. If desired, calibration
of some of the more complex mathematical models can be combined with an RBC
pilot program to obtain estimates of model coefficients. Where pre-design
pilot plant evaluations are not or cannot be undertaken, the designer must
use empirical design approaches, exercising technical judgement regarding
their applicability and adaptibility to site-specific conditions.
5.2 Mathematical Models
The previous discussion in Section 2.7.1 has shown that not all the biofilm
on an RBC necessarily contributes to observed organic removals. Hence, the
mathematical models that attempt to duplicate all the factors contributing to
substrate removal in a fixed film system must combine equations for mass
transfer for both electron donor and acceptor with equations for microbial
metabolism. In RBC systems, mass transfer resistances associated with both
the wastewater film and the biofilm result in significant concentration
gradients from the bulk liquid to biological reaction sites. Consequently,
the overall rates of reaction may be controlled by metabolism or by diffusion.
Mathematical models that take a completely deterministic approach by
attempting to incorporate all of the factors affecting RBC performance
provide considerable insight into those variables and ranges of variables
that impact the RBC process. Employing this approach in design, however,
entails making certain assumptions about the wastewater film thickness in the
atmospheric portion of the cycle, mixing of this wastewater film with the
bulk liquid, the effect of RBC surface shape on mixing and surface biofilm
depth and uniformity, biofilm density, diffusion coefficient(s) within the
biofilm and possible variation with depth and/or the type of organisms that
predominate, and biochemical kinetic parameters. A mathematical model
5-1
(122)
-------
approach has been applied in some design situations in conjunction with pilot
plant* studies carried out on the wastewater in question. Calibration of the
model coefficients can be incorporated in the pilot plant program and is a
necessary step in the use of complex mathematical models for design purposes.
5.3 Pilot Plant Studies
The best source of RBC design information is a comprehensive on-site pilot
plant evaluation. The use of full-scale diameter media are recommended to
avoid scale-up problems. As previously indicated, tip speed is not a suitable
scale-up parameter. In developing an RBC pilot study, consider the following
points:
1. The influent wastewater should be thoroughly characterized. The
parameters of interest vary with the degree of treatment required. For
example, influent TKN, ammonia nitrogen, pH, and alkalinity are all important
parameters when nitrification is required, but are of lesser importance
(assuming the values are in the normal range for municipal wastewaters) for a
design requiring carbonaceous removal only. Measurement of influent suspended
and volatile suspended solids, total 6005, soluble 6005, and sulfide are of
importance in any characterization. Both the hydraulic and organic diurnal
patterns should also be examined.
2. Either raw wastewater screening or primary clarification should be
utilized ahead of the RBC pilot unit depending on the anticipated choice
for the final design. Where no final decision has been made, the effective-
ness of both options on RBC performance should be addressed.
3. The upper loading limit for any stage should be 2.5 to 3 Ib soluble
BOD5/day/1000 sq ft.
4. The change in total COD across a stage (unsettled samples) directly
measures oxygen transfer, provided that no nitrification, denitrification,
sulfide oxidation, sulfate reduction, or change in bulk liquor DO levels is
occurring.
5. Measurements of changes in soluble components across a stage provide
no information on settling characteristics of the suspended solids leaving
the RBC reactor. Settling tests conducted in 1-liter cylinders provide
relative settling information for comparative purposes, but will not duplicate
results obtained in full-scale clarifiers. Settling tests carried out in
large columns (6 to 8 in. in diameter and 6 to 8 ft deep) are always prefer-
able to 1-liter cyclinder tests. Where column tests are used to determine
final clarifier loadings, it is essential that the anticipated effectiveness
of the clarifier, i.e., the degree of departure of actual clarifier hydraulics
from idealized conditions, be considered (1).
6. Nitrification is slow to develop in cold temperatures, and 8 to 10
wk may be required before a nitrification system approaches equilibrium
5-2
(123)
-------
conditions. Seeding with a nitrifying sludge, if available, or temperature
enhancement should be considered. Either option may be feasible at pilot
scale. Where seasonal standards for nitrification are required for the final
design, the transition time and temperatures to develop an adequate nitrifying
population must be considered.
7. Where more than one RBC stage is included in a pilot plant investi-
gation, interstage data should normally be collected on the parameters of
interest. In addition to the parameters cited in item 1, interstage data on
DO levels will often prove informative.
8. In any lightly loaded stage, partial nitrification may occur and
provide a source of nitrifying organisms and residual ammonia nitrogen for
subsequent oxidation in a 6005 analysis. Inhibited 6005 analyses should be
routinely conducted if 6005 measurements are meant to measure carbonaceous
oxygen demand only.
9. Unless flow equalization is provided, a treatment plant will operate
with diurnal flow variations. Dry and wet weather peak load conditions
should be considered in the pilot study. Normally, dry weather peak loads
are of major concern only where the daily peak-to-average flow ratio exceeds
2.5.
10. Where wastewater temperatures less than 50°F are expected, a pilot
study should include operation at these low temperatures if at all possible.
Present empirical approaches predict large variations in needed media surface
areas at low temperatures (refer to Figures 2-8 and 2-10).
11. Incoming sulfide exerts an additional oxygen demand. Where sulfide
is known to be present in the influent wastewater, prechlorination or pre-
aeration may represent potentially cost effective means of eliminating the
oxygen demand posed by incoming sulfide.
12. Where process sidestreams (especially heat treatment liquor) are
expected to be recycled back through the RBC reactor, the additional loadings
from these materials should be incorporated in the pilot study experimental
program. Where nitrification is required, potential additional ammonia
nitrogen loadings from anaerobic digestion should be considered.
13. In conducting soluble BOD and/or COD analyses, care should be taken
to employ uniform procedures for sample filtration and storage. Wastewater
samples filtered immediately are apt to have higher soluble organic concen-
trations than samples filtered after several hours of storage; the converse
is true for sludge samples.
14. Solids production data as a function of total and soluble BOD5
removal and/or total and soluble COD removal should be collected. Sludge
thickening characteristics may also require investigation depending on the
options under consideration for sludge processing. Leaf tests, press tests,
and other analyses related to sludge dewatering may also be appropriate.
5-3
(124)
-------
CJ1
OOencr-hfD-'-SOX-h'— '
^~^
— i cn
INi 1
cn -P>
^-^
3"
fD
3
— j.
rt
«j.
en
3
O
rt
-a
o
en
en
_^.
cr
^-j
fD
Cu
3
CL
***^
0
-s
-h
fD
Cu
en
— i.
cr
^_i
fD
rt
O
f~~
3
CL
fD
-5
rt
CU
^~
fD
O
0
3
73
-s
fD
3"
fD
3
en
•j.
<
fD
^3
DO
£— )
-a
— j.
— >
0
rt
.
f^.
cn o
-5
-^. ea
CU
1— > 3
— i.
n
o
o TO
3 ro
-a 3
CU O
-s <
->. CU
en — i
O
3 O
fD
O en
—ft — ''
ea
3=- 3
<
CU
1.
1
CU
cr
_^
fD
m
3
73
-s
O
Cu
^_j
O
fD
en
i.
ea
3
^>
-a
73
-5
o
CU
0
3-
fD
en
CuOrtfDOCU3CU-5 1 C -h
3-SCuO-SO en cu3
-SeQO 3" rt rt C-'- X O rt
O"fDfD3CL 3" fD C — i- rt 3"
fD — 'enfD— '-enfDSenen— <-fD
CU • — h rt cu rt O
O rt cu — b CU fD rt C: 3 "5
OfD fD CO -h fD ea 73 fD
3 *" * — 1 — h ~" $ fD ~" h "^ "O O 3
en 3-cfD — ' CLrt-hO
— '• S fD 3 3 — '* C" — h " — O *-Z
CL — '• Ort3fD — 'en encu
fDrtCrt 3 O -Q Cu O — '
-53-en-i-enr+rtS —-en
fD fDOrt3" -henc
CLon 3cu-"--h3rtrtCrcu
• fDO eoen-sfD- Cu — 'O
rt-hOfD OO ea-5
rt -henfD3fD fD O
— ' rt X en — 1 o en
fD3"eoSCueuen3"Ocuen
Cu fD r"i~ — J* 3 3 Cu fD C ~5
Cr cu — '"d "< -5 -S cr fD
— '-CueQ— ' — ' <
encuOO3fDfDOO O eQ
O O 3 -S rt« 3-73 O73
3" — i — '• cu fD O CL
rt "< O -S cu -— '•*•»
3"-S
rt — '• CuT3O"SfD-fDX
en fD en 3" rt ><
— '• — '• en 3" en [>
3 rt-'-C/lCTCu O fD .-, rtg
733"3O — ' — '-5-iCUca
CucuQ. — 'fDencrfD ea^.
O rt fD C fD 13" — hirj fD — i
rt 73 CT rtT3 — ' O :?5 - -
enfD — ' 3" cu fD ~S X
O O 3 fD cu "5 fD ^^ CU cu
-h— 'CL rtCuO» JT3-S
STfDO rtCu .ci^fD
-hO-3-SrtfD-S-S
— '• — ' rt eQ 3" O" fD — K — ' 73
3fD CUfDOO73_J-'- — i
CU O3 C3-S3Q
— ' 3 -h -1- -s -s ro x n> <-h
fD O fD < O en (-+
O Cu r*i- 3 ft) O fD p^. CL f[>
— >en3 SO 33^-SQ.
CuC:fDfD<-'-OrttgCU
-j-fD3O — ' 3 ^ 3 cn
— h 3 C < en r+ <~h
-'•fD 3 cu O 3"~S 3" |!> -h o»
O 3 (3~ — ' d O Cu fD ^** "^
Cuc-l-fDen-5Sr+ 3o
rten-s <3-J- "^S
-j. rt fD O ^
OCTO3"en-h3 — 'rt
3 fD -h fD O 3-
3 -S fD
REMOVAL
qft
C
O3
r~
m
03
O
p
o
O
D
CQ
IV)
en -s rt cr < 3;
rt fD O fD CU C
CU CU -53
to en CL -s << -"•
fD o fD fi> -1- o
• 3 U3 3 3 -••
>• ~5 O tQ ~O
CU < CU
o -"• a. ro 3 — •
3 rt fD O- O
fD — • S
->• o — "• a> cu
c to o 3 O en
en 3 c rt
fD -"-73 rt — ' fD
-h 3 O 3- cu £
C -h c: fD -5 cu
— ' O 3 rt
-5 CL -h en fD
-h 3 en _i. -i. -5
o cu -s N
-5 rt S en (ti — '•
3 _j. _i. <-(- • en
< — i
-h fD — 'en eu
O rt — 1
~5 rt- rt CU 3" 3
O fD tQ fD — '•
"O 3 fD X
fD fD Q-- — - en rt
-5 X en —i. c
-h 0) rt 3 -S
o 3 o -am
-S-.. O — '
3 3 T3 -h fD O
->• fD -S -S -h
3 fD CU «•
. -h
<< << — '
en CL rt en *< ->•
->• CU 3- rt 3
on r+ fD fD CL CL
Cu 3 fD -J.
_j. — i tQ <
en en cu cu "5 — '•
fD rt 3 cu CL
-"•-O fD CL CL c
— ' cu ~5 cu cu
— • -5 rt CT — '
c cu en 3~ — '
en rt rt fD fD O
rt fD cu O
-S — 'to 3 O 3
CU << fD O O T3
rt en -s 3 o
fD -h • fD T3 C
CL O O =3
-5 CL C CL
O" TI — i. 3 on
fD fD O -h CL
— ' cu -S -h en o
O O ->• -h
S 3- rt 0 S
3" C -J-
ro en r+ — '
en -s
rt fD
cu ea
CO CU
fD -i H->
CL O>
en " •
~*.
N r+
fD 3" 3="
en fD
^ , —^
3 cu
en c 3
3- 3 ea
o cr fD
c: fD
—•-so
CL -h
O
cr -h en
fD <<
en en
CU e+ c+
CL cu fD
cxea 3
-5 fD
fD en — i
en o
en o Cu
fD -S CL
CL -"•
• r*- 3
3-ea
fD en
-S en
fD 3-
— ' O
cu c
rt — •
_.. CL
<
fD Cu
— J
— ' S.
O cu
Cu ^<
CL en
— *•
3 CT
U3 fD
rt fD
O <
Cu
fD — i
CU C
O Cu
3" rt
fD
en CL
rt •
CU
ea
fD i— i
3
*• —
c rt
3 3"
fD ->•
JD en
c
CU
Oi O
-s cu
fD rt
— i.
— ' O t->
->• 3 cn
3
— i. — i.
rt en
fD T3
CL -S 1C
fD
CT^3 cu
^< C 3
_i. CL
•a -s
n: ro cu
. CL — '
7T
CU CU
3 — '
CL -••
3
3 -••
CU rt
<< ^<
cr cu
fD CL
C_i.
n c:
O en
en rt
rt 3
fD
fD 3
-h rt
-h
fD en
n 3-
rt 0
-"• C
< — '
fD CL
->. cr
-ti fD
3 O
-i. O
rt 3
-S en
«j. _i.
-h CL
-•• fD
o -s
CU fD
rt CL
— j.
0 S
3 3-
fD
-5 3
fD fD
CU <
0 fD
rt ~5
— i.
0 3
3 ->•
rt
-5 -S
QI _j.
rt -h
fD -••
en I
-------
plant programs, empirical design curves and procedures developed by the
manufacturers of RBC equipment and information developed by others that is
available in the technical literature must be relied on as the basis for
design. In these situations, detailed characterization of the raw wastewater
(and primary effluent if available) should still be carried out if at all
possible.
The depth of information provided on various aspects of RBC organic removal
design differs considerably among the various manufacturers (3). Loading and
performance predictions vary depending on the particular set of design
conditions that are applicable. For example, according to the Autotrol (4)
and Clow (5) manuals, the design loading is the same in the presence or
absence of a primary clarifier, but Manufacturer X (6) applied a correction
factor of 1.2 to the computed media surface area when a primary clarifier was
to be omitted. Lyco (7) utilizes applied total 8005 as the controlling de-
sign parameter, whereas Autotrol and Clow use applied soluble 6005. Manufac-
turer X also used applied soluble 8005. Differences in other items such as
recommended wastewater temperature correction factors were presented previ-
ously in Section 2.
To provide perspective on the range of predicted performance that results
from using various manufacturers' design methods, predicted effluent quality
was examined for a range of loadings for the case where both influent and
effluent soluble-to-total 8005 ratios were assumed to be 0.5. The results
of this exercise are presented in Figure 5-1. The three Lyco predictions are
based on total influent 8005 concentrations of 100, 150, and 225 mg/1. The
numbers in parentheses for Manufacturer X represent the influent soluble 8005
concentration used to determine the indicated data point. Both Autotrol and
Clow predict identical relationships for the loading ranges examined. Pub-
lished definitive design procedures for Crane-Cochrane and Walker Process
were not available when this document was prepared (September 1982).
It is also informative to reexamine the various manufacturers predictions
when the results are plotted as mass of BODc; removed per mass of BODs applied
as shown in Figure 5-2. These organic loadings were also based on a soluble-
to-total BODs ratio of 0.50 for Autotrol, Clow, and Manufacturer X. For a
given mass of BODs applied, each manufacturer predicts that mass removal
increases with increasing wastewater concentration, although individual
predictions vary. The same general relationship between mass of BODs applied
and removed has been observed to hold in various studies reported in the
literature (8)(9)(10)(11).
Assumed raw wastewater and primary effluent characteristics for four example
organic removal design problems and comparative required media surface areas
calcualted using the manufacturers design curves are summarized in Table 5.1.
Again, it can be seen that the results vary with the largest difference (Lyco,
Example No. 4) reflecting a more conservative temperature correction factor
for 45°F.
5-5
(126)
-------
40
ro
O1
o> 30
in
D
O
CO
<
O
z
UJ
LJ_
UJ
20
10
AUTOTROL
AND
CLOW
NOTE: Soluble BOD5/total BOD5 = 0.5
LYCO(100)
h
(100) (30)
LYCO(150)
(/ (100) (50)
I I—
,LYCO(225)
MFR X
(30)
—I
x<20)
NOTE: Numbers in ( ) are
influent soluble BOD5
concentrations for
Manufacturer X and total
BODs concentrations
for Lyco
23456
ORGANIC LOADING RATE (Ib total BOD5/day/1000 sq ft)
Figure5-1. Effluent BODs as a function of organic loading for selected RBC
manufacturers design techniques.
-------
ro
00
C/>
O
CO
T3
0)
O
E
Q
O
CD
15
+J
O
uj 2
DC
O
^
111
DC
O
<
O
IT
O
NOTE: Soluble BOD5/total BOD5 = 0.5
LYCO (225 Total)
-LYCO (150 Total)
AUTOTROL AND
CLOW (70)
MFRX (30)
AUTOTROL AND CLOW (30)
LYCO (100 Total)
NOTE: Numbers in ( ) refer to
influent soluble BOD5
concentrations except as
noted
ORGANIC LOADING RATE (Ib total BOD5 applied/day/1000 sq ft)
Figure5-2. Organic removal as a function of organic loading forselected RBC
manufacturers design techniques.
-------
TABLE 5-1. WASTEWATER PARAMETERS AND REQUIRED MEDIA SURFACE AREAS FOR EXAMPLE
RBC ORGANIC REMOVAL DESIGN PROBLEMS
Design Example No.
Parameter
Average flow (mgd)
Raw wastewater TBODs (mg/1)
Raw wastewater SBODs (mg/1)
Wastewater temperature (°F)
Required TBOD5 removal (%)
Effluent TBOD5 (mg/1)
Effluent SBOD5 (mg/1)*
1
1.0
125
50
55
85
18.8
9.4
2
1.0
125
50
55
85
18.8
9.4
3
1.0
70
28
55
85
10.5
5.3
4
1.0
125
50
45
85
18.8
9.4
Primary clarifier TBOD5
removal (%)
30
None
None
30
Required Media Surface Area (sq ft x 1000)
Design Example No.
1
2
3
4
Autotrol(4)
286
286
227
375
Clow(5)
289
289
226
384
Mfr X(6)
276
331
310
373
Lyco(7)
327
400
476
654
*Based on an assumed SBOD5:TBOD5 ratio of 0.5 in the final effluent.
This discussion of manufacturers' design procedures serves to illustrate that
a range of required media surface areas can easily be computed for various
organic removal design situations and that care and judgement must be exer-
cised by the designer in determining appropriate media requirements.
Since predictions of effluent quality depend on empirical correlation of
observed results with the various design parameters it is hardly surprising
to find variations in predicted performance when using generalized design
curves and guidelines. In the absence of pilot plant data for the design in
question, any manufacturer's guidelines should be used with discretion.
Differences in predicted performance for various mechanically driven RBC
5-8
(129)
-------
systems should be viewed in the context of what assumptions were selected in
predicting final effluent quality, as well as any differences that may, in
fact, represent true differences in performance capabilities between the
various units being marketed.
5.4.2 Comparison of Selected Studies with Empirical Design Approaches
Autotrol has been a leader among manufacturers in publishing alternative RBC
design procedures, and its organic removal design curves (4) for mechanically
driven units are presented in Figure 5-3. These curves serve as a useful
backdrop against which several field studies conducted with Autotrol units
can be compared.
Murphy and Uilson (12) treated raw degritted wastewater at the Burlington
(Ontario) Skyway Water Pollution Control Plant with a four-stage, 6.6-ft
diameter RBC unit. The wastewater had an influent 8005 of 120 mg/1 and
influent suspended solids of 230 mg/1. Hydraulic loading to the RBC unit was
varied, but flow to the 4-ft diameter pilot clarifier was kept constant at
350 gpd/sq ft. Effluent soluble 6005 concentrations were consistently less
than or equal to 10 mg/1 at total BODs loadings up to 2 lb/day/1000 sq ft.
Effluent suspended solids, however, were usually in excess of 30 mg/1 (range
of 20 to 70 mg/1) even at these low loadings. It was also observed that 30-
min settling tests of RBC effluent in an Imhoff cone consistently under-
estimated the solids removals attained in the pilot clarifier. In this case,
assuming total effluent BODs values twice the soluble concentrations would be
in error. Data comparing a 1.6-ft diameter pilot unit with the above 6.6-ft
diameter unit demonstrated consistently better overall suspended solids
removal when a primary clarifier was used, suggesting that primary clari-
fication for this wastewater would produce better effluent quality. On the
other hand, plants at Georgetown, Kentucky (13), and Rhinelander, Wisconsin
(14), treat raw (screened only) wastewater and produce effluents as follows:
Hydraulic Total 6005 Suspended
Loading (mg/1) Solids (mg/1)
Plant (gpd/sq ft) Inf. Eff. Inf. Eff.
Georgetown 1.6 216 16 203 15
Rhinelander 1.9 145 15 172 20
These results indicate that while primary clarification is not a mandatory
requirement, there are obviously cases where it is desirable. Whenever high
grease concentrations (greater than 100 mg/1 hexane soluble fraction (15))
may be encountered, primary clarification should always be used with appro-
priate skimming devices on the clarifier surface.
5-9
(130)
-------
50 r 25 r
o>
40
in
Q
S 30
o
£ 20
z
LJJ
Lt 10
LU
WASTEWATER TEMPERATURE ^55°
INFLUENT SOLUBLE BOD5 (mg/l)
150 130 100 90 80 70 60
50
I1
CD
O
C/5
I-
LU
D
_i
LL
U_
LU
20
Q
O
S 15
10
01 234567
HYDRAULIC LOADING RATE (gpd/sq ft)
Figure5-3. Autotrol organic removal design relationships for mechanical
drive RBC's [from Autotrol design manual (4)].
Srinivasaraghaven et al. (9) treated primary effluent in a four-stage, 10.4-
ft diameter air driven RBC. The hydraulic loading ranged from 1 to 3 gpd/sq
ft, and although low effluent soluble BOD5 values were always obtained they
were always higher than predicted by the Autotrol curves (4) for air driven
units (not shown in this report, but predict slightly higher allowable
loadings for given effluent BODs concentrations than the mechanical drive
curves in Figure 5-3). Observed and predicted soluble values are shown below:
Parameter
Observed Effluent SBODs (mg/l)
Predicted Effluent SBOD5 (mg/l)
9
7
8
5
Phase
12
7
9
5
9
5
8
5
Full-scale RBC studies conducted at Alexandria, Virginia (16), compared
mechanical drive units with and without supplemental air addition utilizing
primary effluent feed that averaged 32 mg/l soluble BODs over 150 days of
data collection. For this mechanical drive system, an overall system loading
up to 1 Ib soluble BODs/day/lOOO sq ft would still meet Autotrol's recom-
5-10
(131)
-------
mended limit to the first stage of 4 Ib soluble BOD5/day/1000 sq ft. The 30-
to 40-mg/l soluble BODs concentrations typically encountered in the influent
would correspond to hydraulic loadings of 3 to 4 gpd/sq ft (at 1 Ib soluble
BOD5/day/1000 sq ft overall). The Autotrol design curves (Figure 5-3)
predict soluble effluent BOD's of 5 mg/1 resulting in 83- to 87-percent total
6005 removals. These predictions can be compared to measured soluble 6005
removals of less than 50 percent of the soluble 8005 applied in the loading
range in which the Autotrol curves are applicable. The measured effluent
soluble 8005 values averaged three to four times higher than the predicted
values. Removals in the supplemental air system were substantially better
because of biofilm thickness control.
It is concluded from the above cited studies that the inherent simplicity
offered by using any manufacturers' empirical design curves for organic
removal should be tempered with the realization that they are not always
accurate and in some cases can substantially overestimate attainable removals.
This discrepancy is to be expected since it is highly unlikely any set of
universal design curves will ever be developed that predict RBC effluent
quality as a single-valued function of some design parameter for all waste-
waters. Available data suggest that the manufacturers' design curves be used
with caution whenever first-stage loadings exceed 2.5 Ib soluble BODs/day/lOOO
sq ft (5 to 6 Ib total BOD5/day/1000 sq ft). These data also indicate that
the effluent total-to-soluble BOD5 ratio for RBC's may often exceed 2.0.
Predictions of suspended BOD removal in the final clarifier should be based
on clarifier overflow rates and not an automatic doubling of predicted
effluent soluble BOD.
5.4.3 An Alternative Design Approach (Second-Order Kinetics)
A number of empirical models have been developed and are currently being
used by manufacturers and others for predicting the organic removal per-
formance of RBC's. In attempting to devise a method for improved per-
formance estimation, Opatken (17) inserted soluble organics interstage data
(18)(19) into a second-order kinetic expression and found good correlation
between the rate of disappearance of soluble organics with the square of
the concentration. Field data were then collected on interstage soluble
COD (SCOD) at three RBC facilities. The data from two of these facilities
correlated well with the second-order kinetic expression. There were in-
dications of inadequate oxygen transfer at the third facility, and the
removal rate of SCOD for this system appeared to be oxygen transfer rate
limited rather than biochemical reaction rate limited. The reaction rate
for the third plant simulated zero-order removal of soluble organics rather
than second-order removal. A similar conclusion was reached by Opatken in
assessing data generated by Hynek and Chou (19) on an RBC system where the
hydraulic loading was doubled, i.e., the removal rate was zero order rather
than second order at the higher hydraulic loadings because of oxygen trans-
fer limitations.
5-11
(132)
-------
Opatken applied the Levenspiel equation for second-order kinetics (20) to
compare predicted interstage and final effluent soluble BODtj (SBODs) values
against measured values obtained at nine full-scale, air drive RBC plants
(21). The equation utilized for determining the concentration of soluble
organics at any stage was:
+ 4kt (Cn_i) , n
Cn = ikt
where Cn is the concentration of soluble organics in the nth stage (mg/1),
k is the second-order reaction rate constant (1/mg/hr), t is the average
hydraulic residence time in the nth stage (hr), and Cn_i is the concentra-
tion of soluble organics entering the nth stage (mg/1).
The second order reaction-rate constant, k, used in Equation 5-1 was de-
rived from full-scale, air drive, interstage data generated by Hynek and
Chou (19) during a comparative evaluation of air and mechanical drive RBC's.
The k value so derived is 0.083 1/mg/hr and is assumed by Opatken to be
constant when treating municipal wastewaters with RBC's.
The residence time in hours for any stage was calculated from hydraulic
loading data and application of the volumetric factor of 0.12 gal tank
liquid volume/sq ft media surface area. The primary effluent SBODs con-
centration, Cn_i, was used as the influent concentration to the first stage
to determine the concentration in the first stage. The SBODs concentration
of the first stage was then used as the influent concentration to the sec-
ond stage. This mathematical process was then repeated through each en-
suing stage.
Measured interstage SBOD5 data for nine air drive RBC systems are compared
with predicted interstage concentrations obtained by applying the second-
order rate expression and with predicted interstage concentrations obtained
by using the Autotrol design curves (4) in Table 5-2. The disappearance of
SBOD5 with time is displayed graphically in Figures 5-4 to 5-12 for measured
and second-order predicted values.
Good agreement was obtained between second-order predicted and measured
SBODs concentrations at seven of the nine plants, with substantial dif-
ferences exhibited at Enumclaw and Lancaster. Opatken (17) modified the
analysis of the Lancaster data by assuming an inadequate oxygen transfer
rate in the first stage due to the high organic loading and then applying
the second-order kinetic equation to the following stages. By using the
value of 78 mg/1 SBODs measured in the first-stage effluent as the initial
concentration for the rest of the train and then calculating SBODs ^n the
ensuing stages, he was able to obtain good agreement for Lancaster between
second-order predicted and measured SBODs concentrations. No plausible
5-12
(133)
-------
TABLE 5-2. COMPARISON OF MEASURED, SECOND-ORDER PREDICTED, AND AUTOTROL
DESIGN MANUAL PREDICTED DISAPPEARANCE OF SBOD5 FOR AIR DRIVE
RBC SYSTEMS
Cleves, Ohio
Shafts/Stage
Time/Stage (hr)
= 1-1-1
= 2.5, 2.
5, 2.5
Enumclaw, Washington
Shafts/Stage = 3-1-1-1
Time/Stage (hr) = 1.4, 0.46, 0
Predicted
Cin
Cl
C2
C3
Measured
40
8
5
3
Second
Order
40
12
5
3
Autotrol
Design
Manual
40
<5
Lancaster, Wisconsin
Shafts/Stage = 1-1-1
Time/Stage (hr)
= 1.4, 1.
-1.5**
4, 1.4, 2.2
r • *
C2* =
C3* =
C4* =
Measured
168
14
9
7
6
Lower East
.46, 0.46
Predicted
Second
Order
168
34
20
13
10
Fork, Ohio
Autotrol
Design
Manual
les"1"
Shafts/Stage = (3-2-2-2)**
Time/Stage
(hr) =
Predicted
Cl
C2
C3
C4
Measured
= 218
78
22
14
8
Second
Order
218
39 78t1"
15 22
8 10
4 5
Autotrol
Design
Manual
218t
Cl =
C2 =
C3 =
C4 =
Measured
20
11
6
5
5
0.97, 0.64,
Predi
Second
Order
20
11
8
6
5
0.64, 0.64
cted
Autotrol
Design
Manual
20t
<5
(continued)
5-13
(134)
-------
TABLE 5.2 (continued)
Woodburn,
Shafts/Stage
Time/Stage (hr)
Measured
Cin = 226
Ci = 28
C2 = 7
C3 = 7
€4 = 7
Dodgevill
Shafts/Stage
Time/Stage (hr)
Measured
Cin = 37
C] = 9
C2 = 7
C3 = 4
Washington
= 4-2-1.5**-!. 5**
= 1.69, 0.84, 0.63, 0.63
Predicted
Autotrol
Second Design
Order Manual
226 226*
37
17
11
8
e, Wisconsin
= 2-1-1
= 2.6, 1.3, 1.3
Predicted
Autotrol
Second Design
Order Manual
37 37
11 <5
7
4
Glenwood Springs, Colorado
Shafts/Stage = 1
Time/Stage (hr) =
Measured
Cin = «
Ci = 20
C2 = 14
C3 = 4
04 = 5
West Dundee
Shafts/Stage =
Time/Stage (hr) =
Measured
Cin = 101
Ci = 33
C2 = 15
C3 = 8
-1-1-1.5**
0.56, 0.56, 0.56, 0.84
Predicted
Autotrol
Second Design
Order Manual
43 43
22 13
13 <5
9
6
, Illinois
1-1-1.5**
0.76, 0.76, 1.2
Predicted
Autotrol
Second Design
Order- Manual
101 101
33 >25
16 10
9 <5
(continued)
5-14
(135)
-------
TABLE 5.2 (continued)
Hartford, Michigan
Shafts/Stage = 1-1-1-1
Time/Stage (hr) = 0.25, 0.25, 0.25, 0.25
Predicted
Measured
C2 =
C3 =
C4 =
17
13
12
9
8
Second
Order
17
13
11
9
8
Autotrol
Design
Manual
17t
-
-
<5
cin» Cl» C2, 03, and C4 are SBODs concentrations (mg/1) in the reactor
influent and first, second, third, and fourth stages, respectively.
tQutside influent SBODs limits (30 to 150 mg/1) given in Autotrol air
drive design curves, Figure C-1A (4).
**
High density media.
t1"Assume first stage is overloaded and determine concentrations of
in succeeding stages based on measured GI value.
explanation was offered by Opatken for the observed discrepancy at Enumclaw.
An analysis similar to Lancaster was not considered to be valid because the
measured concentration of SBODs in the first stage was considerably below
the second-order predicted value, and, therefore, oxygen transfer at Enumclaw
could not have been limiting. The interstage concentrations predicted with
the Autotrol design curves (4) are consistently lower than the measured
values.
The above empirical data indicate that Levenspiel's second-order kinetic
expression may offer an improved basis for predicting interstage soluble
organic removals in RBC systems. The lack of published, full-scale inter-
stage data for proprietary equipment other than Autotrol air drive units
5-15
(136)
-------
40
LEGEND;
O Measured values
A Predicted by 2nd-order
kinetics
02468
TIME (hr)
Figure 5-4. Disappearance of soluble BODg with time at Cleves.
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
80
024
TIME (hr)
Figure 5-5. Disappearance of soluble
with time at Enumclaw.
60
o>
Q
O
m 40
HI
_J
CD
D
_J
O
C/3
20
0
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
0246
TIME (hr)
Figure 5-6. Disappearance of soluble BOD5 with
time at Lancaster.
5-16
(137)
-------
LEGEND:
O Measured valves
A Predicted by 2nd-order
kinetics
TIME (hr)
Figure 5-7. Disappearance of soluble BOD5 with time at Lower East Fork.
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
60
024
TIME (hr)
Figure 5-8. Disappearance of soluble BODs with
time at Woodburn.
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
0 2 4
TIME (hr)
Figure 5-9. Disappearance of soluble
BODs with time at
Glenwood Springs.
5-17
(138)
-------
40
D)
ID
Q
O
CO
LU
_l
m
D
_l
O
CO
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
TIME (hr)
Figure 5-10. Disappearance of soluble BOD5 with
time at Dodgeville.
a 20
o <
m ^
LU —
_i cn
CO £
~™^ "*—*
O
CO g
LEGEND:
r- O Measured
> A Predicted
V. kinetics
"Qyv.^^
•^^-^_
•». _^
i
0 2
TIME (hr)
values
ay 2nd-order
i
4
Figure 5-12. Disappearance of soluble BOD5
with time at Hartford.
LEGEND:
O Measured values
A Predicted by 2nd-order
kinetics
TIME (hr)
Figure 5-11. Disappearance of soluble
BODs with time at West
Dundee.
5-18
(139)
-------
prevents comparison of the second-order technique with other empirical
models. The utility of the second-order kinetic procedure described herein
should continue to be evaluated for both mechanical and air drive RBC op-
tions as additional interstage data become available.
5.5 Nitrification Design
5.5.1 Comparison of Available Empirical Design Approaches
Published empirical procedures are available for RBC nitrification design
from the same four manufacturers, Autotrol (4), Clow (5), Lyco (7), and
Manufacturer X (6), cited under organic removal design (Section 5.4.1).
RBC nitrification systems are also marketed by Crane-Cochrane and Walker
Process, but published design methods were not available for these firms at
the time of this writing (September 1982).
As with organic removal design, most RBC nitrification design recommen-
dations vary considerably from manufacturer to manufacturer. Organic in-
fluent contraints where the recommended nitrification design procedures
become applicable, however, are similar for all four: 15 mg/1 soluble
or less for Autotrol, Clow, and Manufacturer X and 30 mg/1 total BODs or
less for Lyco. Manufacturer recommended temperature correction factors
that inflate required media surface area for nitrification at wastewater
temperatures below 55°F were shown previously in Figure 2-10.
In terms of internal staging, Clow recommends a minimum of four stages and
indicates that the addition of two more stages will increase nitrification
efficiency by 5 percent. Autotrol states that staging in its predicted
zero-order removal range, i.e., bulk liquid ammonia nitrogen concentrations
of 5 mg/1 and above, is ineffective and unnecessary. Autotrol does
recommend staging in the first-order zone of removal, i.e., below 5 mg/1
NH3-N. Lyco separates staging requirements into two categories: separate-
stage nitrification and combined carbon oxidation-nitrification. In the
former case, staging recommendations are identical to those for BODs
removal: one stage for up to 40 percent NH3-N removal, two stages for 35 to
65 percent removal, three stages for 60 to 85 percent removal, and four
stages for 80 to 95 percent removal. For the latter case, a minimum of
four stages is recommended. Staging for nitrification was not addressed by
Manufacturer X.
If peak daily flow is not anticipated to exceed average daily flow by a
factor of more than 2.5, Autotrol and Clow recommended using average flow
for design. Lyco indicates RBC systems are resistant to peaking and pro-
vides no specific guidelines. Manufacturer X advised using average daily
flow for design, without specific reference to a peak flow factor.
5-19
(140)
-------
To permit convenient comparison, all four of the above empirical design
methods were translated to a common predictive basis: effluent ammonia
nitrogen concentration versus applied hydraulic loading (3). Four com-
parative plots are presented in Figures 5-13, 5-14, 5-15, and 5-16, for
influent ammonia nitrogen levels of 10,. 15, 20, and 30 mg/1, respectively.
It was necessary to extrapolate between influent concentrations provided on
Lyco's basic design curves (7) to construct the portions of the comparative
figures applicable to that firm. In the case of Clow, the midpoint of the
loading ranges given for each design effluent concentration (5) was used in
developing its curves in the comparative figures.
Utilization of Figures 5-13 to 5-16 for separate-stage nitrification is
straightforward. The procedure is somewhat more involved for sizing com-
bined carbon oxidation-nitrification RBC systems. Each manufacturer's
procedure in the latter design application is or was based on first deter-
mining the required media surface area to reduce the incoming organic con-
centration to a prescribed level: 15 mg/1 soluble BODc for Autotrol, Clow,
and Manufacturer X and 30 mg/1 total BODg for Lyco. Carbonaceous media re-
quirements for the lead portion of the combined reactor should be estimated
separately for each manufacturer using its recommended design methods.
Figure 5-1, presented earlier in Section 5.4.1, compares mechanical drive
carbonaceous media requirements for the several manufacturers for the case
where reactor influent soluble-to-total BOD5 is assumed to be 0.5.
The next step is to determine media surface requirements for the nitri-
fication section of the reactor. The influent ammonia nitrogen concen-
tration to the first carbonaceous stage, not the ammonia nitrogen con-
centration at the stage where nitrification is presumed to begin, is used
by all four manufacturers for this purpose. The required media surface
area estimated for nitrification is then added to the carbonaceous media
requirement to yield overall system design surface area.
If the effluent target is for a soluble BODs of less than 15 mg/1, an addi-
tional step is necessary. Autotrol and Clow recommend checking back to
their respective carbonaceous design bases to assess whether the lower
soluble BODs goal will result in a greater estimated media requirement than
the sum of the two surface areas described above. The larger estimate con-
trols the design. Manufacturer X advocated calculating the media required
to lower soluble BODs from 15 m9/1 to tne target level and then adding
this surface area to that already determined by summing the above two
requirements. Lyco does not address this topic.
Three example nitrification design problems (mechanical drive) were solved
using the manufacturers' organic removal design procedures, Figure 5-15 (20
mg/1 influent NH3-N), and the temperature correction curves in Section 2
(Figures 2-8 and 2-10). The first two examples compare RBC combined carbon
oxidation-nitrification for two different wastewater temperatures; the
5-20
(141)
-------
O)
E 5
<
DC
LU
O
z
o
o
CO
I
LU
LEGEND:
Autotrol
Clow
Manufacturer X
NOTE: Influent NH3-N = 10 mg/l
(Lyco curve is not
identifiable at 10 mg/l)
1234
HYDRAULIC LOADING RATE (gpd/sq ft)
Figure 5-13. Comparative RBC nitrification design curves for an influent
ammonia nitrogen concentration of 10 mg/l [from Roy F. Weston
(3)].
-------
ro
ro
D)
E
<
DC
ai
O
z
O
O
I
m
I
I-
LU
U.
UJ
LEGEND:
Autotrol
— Clow
Lyco
Manufacturer X
NOTE; Influent NH3-N = 15 mg/l
HYDRAULIC LOADING RATE (gpd/sq ft)
Figure 5-14. Comparative RBC nitrification design curves for an influent
ammonia nitrogen concentration of 15 mg/l [from Roy F. Weston
(3)].
-------
en
O)
E
< 5
cc
Ul
O
z
O
o
CO
X
LU
D
LEGEND:
Autotrol
Clow
Lyco
Manufacturer X
NOTE: Influent NH3.-N = 20 mg/l
12345
HYDRAULIC LOADING RATE (gpd/sq ft)
Figure 5-15. Comparative RBC nitrification design curves for an influent
ammonia nitrogen concentration of 20 mg/l [from Roy F. Weston
(3)].
-------
en
0)
Z
HI
o
z
o
o
CO
I
ID
D
LLJ
LEGEND:
Autotrol
Clow
Lyco
Manufacturer X
WOTE: Influent NH3-N = 30 mg/l
0.5
1.0
1.5
2.0
2.5
3.0
HYDRAULIC LOADING RATE (gpd/sq ft )
Figure5-16.Comparative RBC nitrification design curves for an influent
ammonia nitrogen concentration of 30 mg/l [from Roy F. Weston
(3)]-
-------
third encompasses separate-stage RBC nitrification only. Assumed raw waste-
water characteristics and the calculated required media surface areas are
summarized in Table 5-3.
Carbonaceous media requirements for Lyco in Example Nos. 5 and 6 are ex-
trapolated values since the assumed raw wastewater and primary effluent
characteristics fall outside the band of its total BOD5 design curves.
TABLE 5-3. WASTEWATER PARAMETERS AND REQUIRED MEDIA SURFACE AREAS FOR EXAMPLE
RBC NITRIFICATION DESIGN PROBLEMS
Design Example No.
Parameter
Average flow (mgd)
Raw wastewater TBOD5 (mg/1)
Raw wastewater SBOD5 (mg/1)
Raw wastewater NH3-N (mg/1)
Wastewater temperature (°F)
Required TBODs removal (%)
Required NH3-N removal (%)
Effluent TBOD5 (mg/1)
Effluent SBOD5 (mg/l)t
Effluent NH3-N (mg/1)
Primary clarifier TBOD5
removal (%)
5
1.0
125
50
20
55
85
90
18.8
9.4
2.0
30
6
1.0
125
50
20
45
85
90
18.8
9.4
2.0
30
7
1.0
<30*
<15*
20
55
-
90
-
-
2.0
N/A
Required Media Surface Area (sq ft x 1000)
Design Example No.
5
6
7
Autotrol(4)
742
1193
532
Clow(5)
745
1215
538
Mfr X(6)
667
1175
500
Lyco(7)
687
1375
431
* Denotes concentration entering RBC separate-stage nitrification reactor.
t Based on an assumed SBODiTBOD ratio of 0.5 in the final effluent.
5-25
(146)
-------
Surface area requirements calculated using the four manufacturers' pro-
cedures are in close agreement (_+ 10 percent) only for Example No. 5. Sig-
nificant variability is evident for the low wastewater temperature problem
(Example No. 6), indicating the lack of uniformity in low temperture cor-
rection factors for nitrification.
The above examples are intended only to illustrate RBC nitrification media
sizing techniques employed by the equipment vendors and the degree of vari-
ability among techniques that can reasonbly be expected. The solution of
actual design problems, although ultimately the responsibility of the design
engineer, should include consultations with one or more manufacturers to
utilize their experience concerning such factors as unusual wastewater
characteristics, atypical flow variations, pretreatment options, possible
staging arrangements, standard density/high density media split, etc.
5.5.2 Analysis of Available RBC Nitrification Data
5.5.2.1 Observed Pilot-Scale Removals
Pilot studies have been employed to generate RBC nitrification performance
data, both for research purposes and to aid in the design of site-specific
full-scale facilities. Borchardt et al. (22) evaluated the effect of a
broad range of loadings on both interstage and overall removals in a 4-ft
diameter, polystyrene, six-stage pilot unit at Saline, Michigan. This
system was operated in a separate-stage nitrification mode on effluent from
Saline, Michigan's high-rate trickling filter plant.
Interstage data generated at Saline are plotted in Figure 5-17 as stage
ammonia nitrogen concentration versus the calculated removal rate for that
stage. Lower and upper curves bounding the Saline data exhibit maximum
ammonia nitrogen removal rates of approximately 0.4 and 0.75 lb/day/1000 sq
ft, respectively. No correlation was evident for data collected at dif-
ferent wastewater temperatures.
Antoine (23) conducted separate-stage nitrification studies on five pilot
plants in the mid-1970's. His composite removal rate curve, constructed
irrespective of temperature, is also shown in Figure 5-17 along with a
curve developed at 64°F on a separate-stage, 19.7-in. diameter pilot unit
at Guelph, Ontario (24). Antonie's composite curve shows no indication of
approaching a maximum zero-order nitrification rate at bulk liquid ammonia
nitrogen concentrations above 5 mg/1; the Saline boundary curves and the
Guelph curve follow the rapid transition from approximate first-order re-
moval to zero-order removal that is the basis of the current Autotrol de-
sign (4)(25). The knees of the Saline and Guelph curves occur at about 5
and 7 mg/1 NH3-N, respectively.
5-26
(147)
-------
en
0.8
CT
(O
O
| 0.6
TJ
~Q
UJ
O
eo
I
Z
0.4
0.2
n
n
o
D
OA
O
LEGEND:
O Saline (45-60° F)
a Saline (55-65° F)
A Saline (63-73° F)
•Estimated limits of
Saline, Mich, data (22)
-Curve (64°F) for
Guelph, Ontario (24)
-Curve (mixed temp.)
from Antonie (23)
6 8 10 12
STAGE NH3-N CONCENTRATION (mg/l)
14
16
18
Figure 5-17. Estimated ammonia nitrogen removal rates for pilot-scale,
RBC separate-stage nitrification.
-------
Pilot plant data for RBC combined carbon oxidation-nitrification are less
numerous. The results of work carried out at Utah State University (26) on
15-in. diameter pilot units are presented in Figure 5-18. Predictive equa-
tions based on the Monod relationship were developed with linear regression
techniques to fit the plotted data. Based on the relatively few data points
provided, a case could also be made for constructing a first order-zero
order rapid transition curve of reasonably good fit for the 68°F plot with
a knee at approximately 3 mg/1.
The high removal rates observed in RBC separate-stage nitrification pilot
studies are repeated in Figure 5-18. In contrast to the non-temperature
related data in Figure 5-17, the Utah State data indicate a definite temp-
erature dependence. The maximum ammonia nitrogen removal rate at 68°F
approximated 0.65 lb/day/1000 sq ft compared to 0.45 lb/day/1000 sq ft at
5.5.2.2 Observed Full -Scale Removals
Available nitrification interstage data at a wastewater temperature of 55 _+
2°F are plotted in Figure 5-19 for three full-scale RBC plants. Gladstone
and Cleves are combined carbon oxidation-nitrification systems, while Guelph
is a separate-stage nitrification unit treating activated sludge effluent.
Gladstone utilizes mechanical drives; Cleves and Guelph are equipped with
air drive equipment. All three are Autotrol systems.
Each point in Figure 5-19 represents data for one day for a given stage.
For the combined carbon oxidation-nitrification systems, ammonia nitrogen
removal data prior to the stage in the train where maximum nitrification
rates were observed were omitted on the assumption that organic removal was
still heavily influencing nitrifier growth in those stages. An estimated
curve drawn through the plotted points essentially duplicates the current
Autotrol design curve (25) for a wastewater temperature of 55°F (refer to
Figure 2-9). The zero-order removal rate above bulk liquid ammonia nitro-
gen concentrations of 5 mg/1 in Figure 5-19 is projected at 0.3 Ib NH3-
N/day/1000 sq ft, the same as the Autotrol design. The knee of the curve
is at approximately the same concentration for both curves also.
Nitrification data for five full-scale Autotrol facilities operated at
wastewater temperatures in the range of 60 to 70°F are presented in Figure
5-20. All but the Indianapolis facility are combined carbon oxidation-
nitrification systems. The Indianapolis data were generated on a 10.5-ft
diameter pilot unit and are included in the plot because of the close prox-
imity in media diameter to that of field units.
5-28
(149)
-------
LU
<
CC
N
0.5
0.4
0.3
5 e-
2E o
LU 3T
tr >. 0.2
0.1
0
0
o
LEGEND:
O Measured data at 59° F
•Predictive equation,
Z = 0.478 (Cj - 0.4)/[0.45 + (Cj - 0.4)]
4 6 8 10 12 14
Cj, STAGE NH3-N CONCENTRATION (mg/l)
16
18
LU
I-
<
cr
cr
in
LU ^
CE ^,
N
0.7
0.6
0.5
0.4
0.3
§ 0.2
0.1
0
O
o
o
o
o
LEGEND:
O Measured data at 68° F
Predictive equation,
Z = 0.766 Cj/(2.8 + Ci)
46 8 10 12 14 16
Cj, STAGE NH3-N CONCENTRATION (mg/l)
O
18
Figure5-18. Estimated ammonia nitrogen removal rates for pilot-scale, RBC
combined carbon oxidation-nitrification [adapted from Pano etal.
(26)].
5-29
(150)
-------
•^ en
O] (^
—' o
CT
O
CO
T3
.D
LU
O
2
LU
DC
CO
I
z
0.8
0.6
0.4
0.2
LEGEND:
O Gladstone, Mich. (27)
O Guelph, Ontario (24)
A Cleves, Ohio (28)
O
O
o
o
O
o
o
O
O D
-o
D
O
D
O
n
NOTE: Temperature = 55±2°F
4 6 8 10
STAGE NH3-N CONCENTRATION (mg/l)
12
Figure5-19. Full-scale RBC nitrification rates at design wastewater
temperature (55°F).
-------
U.B
cr
CO
o
° 0.6
CO
T3
jQ
HI
£ 0.4
cc
^
o
LJJ n 0
£C U-<1
Z
eo
I
Z
0
LEGEND:
O
D
A
e
•
-
t
o
o
A^
0 2
Gladstone, Mich. (27)
Lancaster, Wise. (28)
Lower East Fork, Ohio (28)
Columbus, Ind. (28)
Indianapolis, Ind. (29)
_ •
• O •
0 • A • 0 •
• A O
°0 0
0 ° 00 o°° °
o 9vP ° 8 °
°^D n O NOTE: Temperature = 65±5°F
8 D° e
4 6 8 10 12 14 16 18
STAGE NH3-N CONCENTRATION (mg/l)
Figure 5-20. Full-scale RBC nitrification rates at high wastewater temperatures.
-------
The data in Figure 5-20 are obviously much more scattered than in the 55 _+
2°F plot (Figure 5-19). No attempt to draw a representative curve through
the data was made. The plotted ammonia nitrogen removal rates for the
lower bulk liquid stage concentrations are less than those observed at 55 +_
2°F. At higher concentrations, the removal rate roughly centers around
0.3 Ib l\IH3-N/day/1000 sq ft with widely divergent points above and below.
If the Indianapolis data, which were collected at constant flow under care-
fully controlled pilot conditions, are ignored, the average removal rate
above 5 mg/1 NHa-N is well below 0.3 Ib NHs-N/day/lOOO sq ft. Considering
the data in Figures 5-19 and 5-20, there is no justifiable basis to predict
that full-scale RBC peak nitrification rates at temperatures well above
55°F will exceed the maximum design rate at 55°F of 0.3 Ib NH3-N/day/1000
sq ft.
Low temperature (47 +_ 2°F) interstage nitrification data from Gladstone are
shown in Figure 5-21. These data exhibit an increasing ammonia nitrogen
removal rate from the third stage through the sixth stage, suggesting that
the low wastewater temperature and perhaps also companion organic concen-
trations contributed to lower nitrifier development in those stages (4 and
5) where the highest rates would be expected at warmer temperatures. Ef-
fluent ammonia nitrogen residuals averaged approximately 9 mg/1 for the 6
days represented in Figure 5-21.
The stage-average ammonia nitrogen removal curve plotted for stages 3, 4,
5, and 6 in Figure 5-21 is tilted in the opposite direction from the warmer-
temperature plots. The removal or oxidation rate continued to increase
through stage 6. It is difficult to determine based solely on an examina-
tion of the stage-average curve whether nitrification ever reached its
maximum temperature-compatible rate, even in the sixth stage. The following
Gladstone data (27) collected at 47 + 2°F and 55 + 2°F were compared to
further examine this point:
Wastewater No. of Avg. Final Avg. NH3-N Removal Rate
Temp. Days Eff. NHs-N (Ib/day/1000 sq ft)
( OF) of Data (mg/1) Stage 3Stage 4Stage 5Stage 6
55+2 11 1.5 0.240 0.310 0.180 0.128
47+2 6 8.9 0.073 0.103 0.170 0.196
The ratio of the maximum removal rates for the two temperatures (stage 4
for 55°F/stage 6 for 47°F) is 1.58. Examination of the temperature cor-
rection curve for nitrification (Figure 2-10) reveals that the recommended
surface area correction factors for 47°F referred to a base of 55°F range
from 1.5 to 1.6 for the four manufacturers represented. This observation
provides a measure of indirect evidence that nitrification had reached its
temperature-compatible limit in stage 6 for the 47 + 2°F Gladstone data.
5-32
(153)
-------
•_j en
O
o
o
CO
.D
LU
I-
LU
QC
co
Z
0.5 r
0.4
0.3
0.2
0.1
NOTE: All data are measured
values from Gladstone,
Mich.(27)
Temperature = 47±2°F
SBOD5 = soluble BOD5
-CURVE PROJECTED TO
ZERO CONCENTRATION
O
STAGE 6-0
(AVG. SBOD5 = 5 mg/l
O
O
STAGE 5-D
(AVG. SBOD5 =6 mg/l
STAGE 4-A
(AVG. SBOD5 = 8 mg/l
STAGE 3-©
(AVG. SBOD5 = 13 mg/l
STAGE-AVERAGE
CURVE
10
12
14
16
18
20
STAGE NH3-N CONCENTRATION (mg/l)
Figure 5-21. Full-scale RBC nitrification rates at low wastewater temperatures.
-------
The Gladstone plant is equipped with 258,000 sq ft of RBC media. If the
stage-6 nitrification rate in Figure 5-21 is projected at zero-order re-
moval down to an ammonia nitrogen concentration of 5 mg/1 then at first-
order removal to zero concentration, the amount of additional media surface
area required at 47°F to reach the same 1.5-mg/l effluent residual achieved
at 55°F is estimated at about 130,000 sq ft. The surface area correction
factor calculated in this manner is 1.50, which would provide approximately
20,000 sq ft less media to reach 1.5 mg/1 NH3-N at 47°F than the 1.58 factor
based on a comparison of maximum nitrification rates only. The above ex-
ercises support the general applicability of the manufacturers' recommended
temperature correction factors for RBC nitrification occurring within the
range of 47 to 55°F.
5.5.2.3 Influence of Soluble Organics
The manufacturers' design procedures for nitrification (4) (5) (6) (7) are
based on the premise that soluble 6005 must be reduced to 15 mg/1 or less
before significant ammonia nitrogen oxidation can take place in an RBC
reactor. Although not implicitly stated, the manufacturers' literature
implies that once soluble BODs is reduced to 15 mg/1, nitrification will
commence immediately at maximum rates. As shown in Figure 5-21, this is
not necessarily the case. The highest nitrification rate was not achieved
until the sixth and last stage where the average soluble BOD5 concentration
for the 6 days of 47 + 2°F Gladstone data represented had reached 5 mg/1.
The nitrification rate in the third stage (average soluble BOD5 of 13 mg/1)
was less than half that in the sixth stage.
The lag in achieving maximum nitrification rates with decreasing soluble
BOD5 values below 15 mg/1 is not restricted to cold wastewater tempera-
tures. The above 47 + 2°F data (February to April 1976) are compared in
Table 5-4 with 3 consecutive days of data for Gladstone in August 1976 (27)
when the wastewater temperature each day was 68°F. Negligible increase in
nitrate nitrogen was observed in stage 2 at 68°F, even though average solu-
ble 6005 in that stage was 10 mg/1. The maximum nitrification rate, as
indicated by the maximum ammonia nitrogen removal rate, was not reached
until stage 4 where soluble BODs had again decreased to 5 mg/1.
The above data suggest that the influence of soluble organic concentration
on nitrification rates in RBC's is relatively independent of temperature.
As seen by the absolute ammonia nitrogen removal rates and as discussed
previously in Section 2.8.4.4, temperature does have considerable impact on
the level of nitrification obtained below 55°F. The point at which the max-
imum rate was observed, however, occurred at the same soluble BOD5 concentra-
tion in both examples. The above discussion is not intended to imply that
soluble BOD5 must be reduced to 5 mg/1 to achieve maximum nitrification
rates in all situations, but does serve to illustrate that nitrifier-hetero-
trophic population dynamics continue to play an important role in RBC nitri-
fication down to concentrations substantially less than 15 mg/1
5-34
(155)
-------
TABLE 5-4. EFFECT OF SOLUBLE BOD5 ON RBC NITRIFICATION RATES AT GLADSTONE,
MICHIGAN
Stage Stage Stage Stage
Parameter _ 3 _ 4 _ 5 _ 6
2/10-4/15/76 (6 days of data)
(47 + 2°F)
SBOD5 (mg/1) 13 8* 6* 5*
NH3-N (mg/1) 16.7 15.0 12.2 8.9
(mg/l)t 1.4 3.5 6.4 9.5
NH3-N Removal Rate
(lb/day/1000 sq ft) 0.073 0.103 0.170 0.196
tNo significant N03-N appearance in stage 2 with average SBODs = 23 mg/1
8/12-8/14/75 (3 days of data)
(68°F)
SBOD5 (mg/1) 6* 5* 4* 5*
NH3-N (mg/1) 13.0 9.5 6.5 4.2
-N (mg/l)# 1.4 4.5 7.6 10.3
-N Removal Rate
(lb/day/1000 sq ft) 0.199 0.277 0.217 0.179
#No significant NOs-N appearance in stage 2 with average SBODs = 10 mg/1
inhibited SBODs values
5.5.2.4 Comparison of Measured Versus Predicted Values for
RBC Combined Carbon Oxidation-Nitrification
In addition to extensive interstage sampling conducted at the Gladstone RBC
facility by Autotrol, the plant staff has maintained excellent monthly
records of plant operation and performance (30). Measured nitrification
performance is compared with predicted performance for 2 yr of monthly-
average data at Gladstone in Table 5-5. The Autotrol curves (4) utilized
in this analysis included Figure 2-8 (temperature correction factors for
6005 removal), Figure 2-10 (temperature correction factors for nitrifi-
cation), Figure 5-3 (organic removal design relationships for mechanical
drive units), and Figure 5-22 (nitrification design relationships).
5-35
(156)
-------
TABLE 5-5. PREDICTED VERSUS MEASURED FINAL EFFLUENT AMMONIA NITROGEN
FOR GLADSTONE, MICHIGAN
Month
1977
Jan.
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
1980
Jan.
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
*Values
which
Pri.
Eff.
TBOD5
(mg/1)
155
144
106
90
121
117
94
86
93
90
76
88
77
85
91
88
86
108
128
68
71
94
115
99
estimated
extend only
Pri.
Eff.
NH3-N
(mg/1)
23.9
20.1
14.0
11.6
17.4
19.1
15.0
12.2
13.8
11.5
10.0
13.4
14.0
17.1
18.4
14.2
17.4
15.6
15.3
14.0
16.3
18.8
19.3
18.3
Hydraulic
Loading
Rate
(gpd/sq ft)
0.90
1.00
1.33
1.60
1.21
1.04
1.30
1.57
1.70
1.64
2.04
1.52
1.03
0.90
0.87
1.20
0.98
0.96
1.00
1.01
1.00
0.89
0.80
0.77
Waste-
water
Temp.
(°F)
47
43
44
47
54
61
65
66
64
60
56
52
46
46
46
47
54
59
64
66
65
61
55
50
by extrapolation of Autotrol nitri
from 1 to 6 mg/1 effluent NH3-N.
5-36
(157)
Measured
Fin. Eff.
NH3-N
(mg/1)
9.0
7.0
5.6
4.4
3.6
2.1
1.0
1.0
1.6
2.0
2.0
2.7
1.4
2.0
2.3
2.3
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
1.5
fication curves
Predicted
Fin. Eff.
NH3-N
(mg/1)
*7±
*9±
5.5
2.9
1.2
<1.0
1.1
<1.0
1.6
<1.0
1.2
1.5
1.1
1.2
1.6
1.7
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
(4),
-------
WASTEWATER TEMPERATURE > 55°F
INFLUENT NH3-N (mg/l)
0)
-§ 4
30 25 22 2019 18 17 16
15
CO
I
z
111
LLI
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
HYDRAULIC LOADING RATE (gpd/sq ft)
5.0
Figure 5-22. Autotrol nitrification design relationships [from Autotrol design
manual (4)].
For 1977, primary effluent total BODs averaged 105 mg/l, primary effluent
ammonia nitrogen averaged 15.2 mg/l, and the hydraulic loading rate aver-
aged 1.40 gpd/sq ft. The corresponding values for 1980 were 93 mg/l, 16.6
mg/l, and 0.95 gpd/sq ft, respectively.
The Autotrol organic removal curves were developed for use with influent
soluble BODs as one of the two input parameters. A primary effluent solu-
bie-to-total BODs ratio of 0.5 was assumed for Gladstone to enable utili-
zation of Figure 5-3. No temperature correction factors were employed for
either BODs removal or nitrification for wastewater temperatures >_55°F.
The predicted final effluent ammonia nitrogen concentrations, although
slightly optimistic overall, compare favorably with the measured field
values. Considering the assumption made for the primary effluent soluble-
to-total BODs ratio, the correlation is excellent. Based on these 2 yr of
data from Gladstone, the Autotrol empirical design procedure for combined
carbon oxidation-nitrification RBC systems appears to reasonably predict
actual performance.
5-37
(158)
-------
Judging the validity of a design procedure on the basis of one plant can be
misleading. Single-day data were provided (27) for three other RBC combined
carbon oxidation-nitrification plants: Hartford, Michigan; Lower East Fork,
Ohio; and Columbus, Indiana. Predicted and field-measured values for these
3 days of data are summarized below:
Plant
*Hartford
flower East
Fork
Pri.
Eff.
SBOD
(mg/1
28
19
27
Pri.
Eff.
NH3-N
17.4
17.0
16.3
#Columbus
*3/16/79, t8/16/79, #4/24/79
Hydraulic
Loading
Rate
(gpd/sq ft)
1.98
1.34
1.32
Waste
water
Temp.
45
68
66
Fin. Eff. NHj-N (mg/1)
Measured Predicted
7.5
1.6
4.2
7.5
Predicted and measured effluent values are in close agreement for two of
the above days of data; the Columbus data do not correlate well. The
danger inherent in making decisions based on 1 day of information, how-
ever, is obvious. Due to the slow response of nitrifiers to changing
environmental conditions, it is believed that correlation of predicted and
measured concentrations will improve by averaging data over a reasonable
time period, provided wastewater temperature does not vary more than 3 to
4°F during that period.
5.5.2.5 Comparison of Measured Versus Predicted Values for
RBC Separate-Stage Nitrification
Very little field data are available that can be used for comparative
purposes for this RBC application. Indianapolis (29) conducted a series
of separate-stage nitrification evaluations with a 10.5-ft diameter RBC
pilot plant. Diurnal flow variations were not imposed. Unchlorinated
activated sludge effluent was utilized as system feed. Although the
size of the unit employed was somewhat smaller than the 11.5- to 12.0-ft
diameter media sold commercially today, it was sufficiently large to be
representative of field-scale performance. Predicted and measured av-
erage final effluent ammonia nitrogen values for seven phases of the
study are presented in Table 5-6. Autotrol's nitrification design curves
(Figure 5-22) were used for determining the predicted values.
Correlation of predicted to measured data is reasonably good for four of
the seven phases and is not good for the other three. The authors of
the Indianapolis pilot study report noted periodic problems with predation
in the latter stages of the unit where soluble 6005 was routinely re-
5-38
(159)
-------
TABLE 5-6. PREDICTED VERSUS MEASURED FINAL EFFLUENT AMMONIA NITROGEN FOR
INDIANAPOLIS, INDIANA
Phase
4
5
6
7
8
9
10
Inf.
SBOD5
(mg/1)*
9
6
3
8
9
8
10
Inf.
NHa-N
(mg/1 )
18.1
14.0
11.0
14.1
12.1
8.1
11.5
Hydraulic
Loading
Rate
(gpd/sq ft)
1.35
2.65
2.64
3.00
1.87
1.79
2.94
Waste-
water
Temp.
(OF)
61
59
58
61
67
68
71
Measured
Fin. Eff.
NHa-N
(mg/1 )
6.6
5.0
1.4
5.7
1.9
0.5
1.9
Predicted
Fin. EFf.
NHs-N
(mg/1 )
<1.0
1.9
1.1
2.6
<1.0
<1.0
1.9
inhibited SBODs values
duced to 1 to 3 mg/1. Bypassing of small quantities of primary effluent
to the lead RBC stage was reported to reduce predation and improve ni-
trification efficiency. Whether the higher measured final effluent
ammonia nitrogen concentrations (Phases 4, 5, and 7) corresponded to
periods of increased predator activity is not known.
5,5.3 Clarification Following RBC Separate-Stage Nitrification
Gravity clarification is obviously required following RBC combined carbon
oxidation-nitrification where suspended solids concentrations leaving the
reactor are typically several hundred mg/1. The need for clarification
and/or the type of clarification following RBC separate-stage nitrification
are less clear.
Consider the data (27) shown in Table 5-7 for the Cadillac, Michigan me-
chanical drive RBC installation that is operated during warm-weather months
to nitrify effluent from the City's activated sludge plant. These data
indicate that the processes of nitrification and supplemental BODs removal
had little effect on suspended solids entering, passing through, and leav-
ing the reactor. The effluent from the last stage would meet most effluent
permit requirements.
The ammonia nitrogen loading at Cadillac is fairly low. Higher loadings
should have a greater impact on nitrifier growth and sloughing. In such a
5-39
(160)
-------
TABLE 5-7. PERFORMANCE DATA FROM CADILLAC, MICHIGAN SEPARATE-STAGE RBC
NITRIFICATION PLANT
Month
Avg.
NH3-N (mg/)
In Out
4.8 0.5
Total BODs (mg/1)
In
Out
17
TSS (mg/1)
In
16
Out*
Aug. '80
Sept.
Oct.
June '81
July
Aug.
Sept.
Oct.
6.8
2.0
2.0
6.7
2.9
5.8
3.6
8.5
0.5
0.4
0.2
0.4
0.4
0.4
0.4
0.9
18
15
14
20
11
23
8
-
7
9
5
7
4
5
6
-
17
30
20
14
8
10
15
-
17
30
15
13
10
10
13
-
15
*Unclarified suspended solids directly from last reactor stage.
situation, it is likely that additional solids would be generated, but
perhaps not to a degree that would necessarily require follow-on clari-
fication. Similarly, the removal of additional BODs (an average 11 mg/1 at
Cadillac) will also contribute to solids generation and sloughing. These
solids will ultimately pass through the system and, in some cases, may have
a sufficiently high concentration to warrant removal, although this obviously
is not true for Cadillac.
Existing installations employing RBC's for separate-stage nitrification
have provided either follow-on gravity clarification or follow-on granular
media filtration. It is suggested that clarification of any type not be
specified "out-of-hand" for RBC nitrification designs. A comprehensive
analysis of the factors affecting this decision should include consid-
eration of anticipated or demonstrated incoming suspended solids, organic
carbon, and unoxidized nitrogen loads and their anticipated variability as
well as the anticipated removals of these materials, anticipated solids
generation, and applicable final effluent limitations.
5.5.4 Summary
Based on the process and design considerations addressed in this document,
the following conclusions are offered regarding RBC nitrification:
5-40
(161)
-------
1. Pilot units (4-ft diameter and smaller) operated at field-equiva-
lent tip speeds of about 60 fpm produce nitrification rates unachievable
with full-scale RBC's for valid and explainable reasons.
2. Ammonia nitrogen oxidation in RBC pilot plants exhibits classic
Monod response with varying degrees of asymptotic approach to zero- and
first-order removal depending on concentration and other pilot study factors.
3. Pilot-scale results generated at or near tip speeds of 60 fpm
must be correlated to anticipated full-scale performance through the use of
empirically-confirmed scale-up factors or calibration of deterministic
models before they can be utilized with confidence for field design.
4. Field-scale RBC's nitrifying at or near 55°F can be described by
a dual-kinetic pattern: zero-order oxidation of ammonia nitrogen at approxi-
mately 0.3 lb/day/1000 sq ft down to bulk liquid concentrations of roughly
5 mg/1, followed by rapid transition to first-order oxidation thereafter.
Additional interstage field data should be collected to further verify this
observation.
5. Oxygen transfer capability is the dominant factor controlling
full-scale RBC maximum nitrification rates at wastewater temperatures above
55°F and bulk liquid ammonia nitrogen concentrations above 5 mg/1, ef-
fectively overriding increased nitrification potential at higher temp-
eratures.
6. Wastewater temperature is the major factor controlling full-scale
RBC nitrification below 55°F, becoming increasingly dominant as tempera-
tures of 40°F are approached.
7. Bulk liquid DO's of 3.5 to 4.0 mg/1 appear to be sufficient to
prevent significant retardation of nitrification rates and efficiency.
8. The potential for achieving enhanced nitrification rates via pH
adjustment to pH 8.5 has been shown conclusively at small scale but remains
to be demonstrated at full scale.
9. Soluble BODs plays an important role in RBC nitrifier-heterotroph
population dynamics down to concentrations of approximately 5 mg/1.
10. Predictability of full-scale effluent ammonia nitrogen concen-
trations using Autotrol's empirical design approach is good based on very
limited data. Additional data sets representing a broader spectrum of
plant and climatic conditions need to be evaluated, however, to determine
the method's range of applicability. Evaluation of other empirical design
procedures will require field-generated interstage performance data on
other proprietary equipment.
11. Clarification following RBC separate-stage nitrification should
not be mandated as a matter of course, but should be based on a thorough
evaluation of all contributing factors.
5-41
(162)
-------
5.6 Denitrificat ion Design
5.6.1 Introduction
As discussed in Section 2.9, the design basis for RBC denitrification is not
well established. Autotrol is the only manufacturer marketing RBC denitri-
fication systems at this time (September, 1982). Their design procedure,
developed primarily from RBC pilot studies conducted by Murphy et al. (31),
assumes that in the presence of an adequate organic carbon source denitri-
fication is independent of nitrate nitrogen concentration in the bulk liquid
down to 1 mg/1. The zero-order denitrification rate determined by Murphy et
al. with methanol addition is approximately 0.85 Ib N03-N/day/1000 sq ft at
55°F.
Conversely, pilot-scale RBC data developed by Blanc et al. (32) also using
methanol addition exhibit a first-order relationship between denitrification
rate and bulk liquid nitrate nitrogen concentration in the range of 1 to 6
mg/1 N03-N. As shown previously in Figure 2-14, their observed denitrifi-
cation rates increased from about 0.08 Ib N03-N/day/1000 sq ft at 1 mg/1 to
approximately 0.65 Ib N03-N/day/1000 sq ft at 6 mg/1.
Denitrification rates reported for attached growth systems other than RBC's
vary widely (33). For high-posity packed bed systems with various media
types including Koch flexirings, intalox saddles, Raschig rings, and Surfpac
plastic media, observed rates range from 0.026 to 0.24 Ib N03-N/day/1000 sq
ft at wastewater temperatures of 50 to 75°F. Higher rates have been reported
for high-porosity fluidized bed systems including sand and activated carbon,
varying from 0.66 to 9.8 Ib N03-N/day/1000 sq ft at temperatures of 60 to
75°F. The zero-order design removal rate recommended by Autotrol (4) for
RBC's of 0.92 Ib N03-N/day/1000 sq ft at 55°F falls between the ranges
reported for packed and fluidized bed column systems.
5.6.2 Design Curves
Autotrol has developed a series of RBC denitrification design curves for
domestic wastewater relating hydraulic loading to influent and effluent
nitrate nitrogen concentrations at 55°F (4). These curves are reproduced in
Figure 5-23. As shown, the maximum influent concentration considered is
30 mg/1 N03-N, while the minimum design effluent value for any condition is
1 mg/1 N03-N. The mass removal rates represented by the eight curves are
essentialy the same, i.e., 0.92 Ib N03-N/day/1000 sq ft.
For wastewater temperatures other than 55°F, Autotrol recommends the appli-
cation of an Arrhenius correction curve (Figure 5-24), also based on the work
of Murphy et al. (31). Figure 5-24 was developed using a single four-stage,
19.7-in diameter RBC pilot unit at flows of 0.65 to 1.05 gpm and may require
5-42
(163)
-------
10
WASTEWATER TEMPERATURE >55°F
INFLUENT NO3-N (mg/l)
30 25 20 18
6
LLJ
D
IU
2 4 6 8 10 12 14
HYDRAULIC LOADING RATE (gpd/sq ft)
16
Figure 5-23. Autotrol denitrification design relationships [from Autotrol
design manual (4)].
2.00
1.75
DC
O
O
LL
z
o
£
DC
O
O
1.50
1.25
1.0
0.75
0.50
0.25
0.0
40 45 50 55 60 65
WASTEWATER TEMPERATURE (°F)
Figure 5-24. Autotrol temperature correction factors for denitrification [from
Autotrol design manual (4)].
5-43
(164)
-------
further adaptation when full-scale denitrification performance data become
available.
5.6.3 Performance Data
No known full-scale municipal RBC denitrification systems had been in
operation for a sufficient period when this document was completed (September
1982) to produce representative equilibrium data. One plant, the Orlando,
Florida (Easterly) Iron Bridge Water Pollution Control Facility (34), was in
its startup phase but had yet to establish routine operation.
An RBC system has been used to treat leachate from a garbage dump landfill at
Miyazaki City, Japan, since 1977 (35). The leachate is characterized as low
in 6005 (10 to 30 mg/1), moderate in suspended solids (30 to 100 mg/1), low
in phosphorus (non-detected), and high in ammonia nitrogen (100 to 140 mg/1)
with sufficient alkalinity (800 to 1100 mg/1) to support nitrification.
Design influent concentrations were selected as 50, 100, and 200 mg/1,
respectively, for 6005, suspended solids, and ammonia nitrogen with effluent
objectives of 20 mg/1 6005, 25 mg/1 suspended solids, and 50 mg/1 N03-N.
A sequential RBC train with three separate functional RBC units was installed
to treat the Miyazaki City leachate. The first unit consists of a conven-
tional four-stage RBC reactor designed to reduce the incoming minimal BOD5
load to a low residual level and oxidize the highly-concentrated ammonia
nitrogen load to nitrate nitrogen. The second unit is a submerged four-stage
RBC whose function is to denitrify the nitrate nitrogen produced in the first
unit. Methanol is added to this unit. A final, short-detention (34-min),
four-stage RBC unit oxidizes residual methanol and reaerates the final
effluent. Clarification is provided only after the RBC reaeration unit.
Design flow and design wastewater temperature for this RBC system are 0.092
mgd and 59°F, respectively. Phosphoric acid is added ahead of the first RBC
unit to prevent phosphorus deficiency from developing within the biofilms.
Based on metabolic consumption of 20 mg/1 NH3-N in the nitrification unit,
the denitrification reactor was designed to handle 180 mg/1 N03-N at an
application rate of 1.51 lb/day/1000 sq ft. For an effluent requirement of
50 mg/1 N03-N, the design removal rate is 1.09 Ib N03-N/day/1000 sq ft. It
is not known if the Autotrol denitrification design procedure for municipal
systems was a major consideration in the selection of design criteria at
Miyazaki City.
The Miyazaki City RBC facility was started up in the fall of 1976. Published
performance data are available for 7 of the first 11 mo of 1977 (35). For
these 7 mo, BODs removal averaged 70 percent (20 mg/1 in, 7 mg/1 out), sus-
pended solids removal 89 percent (45 mg/1 in, 5 mg/1 out), and total nitrogen
removal 88 percent (129 mg/1 NH^-N in, 14 mg/1 N^-N + N03-N out). Detailed
performance data for the denitrification unit are presented in Table 5-8.
5-44
(165)
-------
TABLE 5-8. RBC DENITRIFICATION PERFORMANCE AT MIYAZAKI CITY, JAPAN (1977)
Waste-
Month
Jan.
Feb.
May
June
July
Aug.
Nov.
Avg.
% of
Design
Flow
21
19
34
47
53
43
71
41
water
Temp.
(°F)
48
46
72
75
86
86
73
69
In
(mg/1)
102
108
26
70
68
67
77
74
N03-N
Out
(mg/1)
15
Tr
11
13
5
14
2
9
Rem.
(%)
85
100
58
81
93
79
97
85
Loading
Rate
( Ib N03-N ^
\daylOOO sq ftj
0.18
0.17
0.07
0.28
0.30
0.24
0.46
0.24
Removal
Rate
/ Ib NO^-N \
/
\daylOOO sq ft/
0.15
0.17
0.04
0.23
0.28
0.19
0.45
0.21
Nitrate loading rates documented in Table 5-8 were not sufficiently high to
stress the Miyazaki City denitrification reactor to its design removal
capacity. The data do indicate that submerged biodiscs operating in the
presence of an adequate organic carbon source can effectively reduce nitrate
nitrogen to nitrogen gas. Substantiation of design criteria will require
additional operating and performance results, particularly from plants with
much lower nitrogen concentrations typical of municipal wastewater.
5.7 Design Case History for Combined Carbonaceous and Nutrient Removal
The Orlando Easterly plant is an advanced secondary treatment facility
designed to achieve efficient removal of nitrogen and phosphorus as well as
organics. Design flow is 24 mgd. The plant's liquid stream process units
are shown schematically in Figure 5-25 (34). RBC's comprise the biological
treatment components of the plant and are used for carbonaceous oxidation,
nitrification, and denitrification. Mineral addition capability is provided
at several feed points to precipitate phosphorus and improve solids capture.
Effluent filtration assures low residual suspended solids in the plant
discharge. Post aeration is utilized to raise effluent DO to permit levels.
Combined carbon oxidation-nitrification is accomplished in an air drive RBC
reactor. Nineteen parallel trains of nine shafts each (171 total) are
5-45
(166)
-------
SODIUM
ALUMINATE
+ POLYMER
ALUM
+ POLYMER
METHANOL
RAW
WASTE WATER
CARBONACEOUS
AND
NITRIFICATION
RBC's
PRIMARY
CLARIFIERS
DENITRIFICATION
RBC's
(SUBMERGED)
SECONDARY
CLARIFIERS
cr>
01
i
SAND
FILTERS
CHLORINE
CONTACT
CHAMBER
POST
AERATION
(SURFACE
AERATORS)
FINAL
EFFLUENT
ALUM
+ CHLORINE
CHLORINE
Figure5-25. Process flow diagram for Orlando, Florida (Easterly) Iron Bridge Water Pollution
Control Facility [adapted from Dallaire(34)].
-------
provided for this purpose. The first three shafts in each train are equipped
with standard density media (100,000 sq ft/shaft) and the last six with high
density media (150,000 sq ft/shaft). The nine-shaft trains are divided into
five stages using a 3-2-2-1-1 configuration. Primary efflent can be step
feed to the first two stages.
The original design specified 180 mechanical drive units for carbonaceous
oxidation and nitrification. A 5-percent reduction in required media surface
area was achieved by the switch to air drive.
Each air drive shaft is outfitted with a load cell. An attempt will be made
by the plant staff to distribute organic loading and air supply to limit
biofilm thickness on the carbonaceous stages to 50 mils (0.05 in) (36). Each
shaft is also equipped with an individual air throttling value. Two 400-hp,
16,000-cfm blowers, one operating, one standby, supply air to drive the
carbonaceous and nitrification units. The media on the first four shafts of
each train have 6-in. deep air cups, while the depth of the cups on the last
five stages is 4 in.
Denitrification is achieved with 42 submerged, mechanically driven RBC's
consisting of six trains of seven shafts each. An under-over-under baffle
arrangement effectively divides each shaft into a separate stage. The shafts
are rotated with 5-hp motors, although normal power draw is expected to be
closer to 3 hp (36). The media surface area provided on each shaft is
108,000 sq ft. The denitrification media are rotated countercurrent to flow
except for the last stage, which is rotated with flow to help lift sloughed
solids over the last baffle weir.
Organic carbon necessary to expedite the anoxic denitrification reactions is
supplied by methanol addition. Methanol dose is automatically controlled by
feedback signals from a flow meter and an autoanalyzer that continuously
monitors incoming nitrate nitrogen concentration.
The denitrification tanks are covered with 100-mm diameter, high density
polyethylene, floating spheres formulated to include carbon black. These
spheres cover approximately 92 percent of the tank surface and reduce
atmospheric reaeration to 10 to 12 percent of that of an open tank. Nitrogen
gas can escape to the atmosphere and rain can fall through openings between
the spheres. Alternatives to the floating-sphere concept that were considered
included floating plank styrofoam with nylon straps and neoprene membranes
with rigid support frames (36).
An interesting feature of the Orlando Easterly physical layout is that the
RBC trains (both aerobic and anoxic) are arranged in parallel rows of two. A
gap has been provided between each set of rows extending the entire length of
the trains to permit crane access for shaft removal, if necessary. With this
5-47
(168)
-------
configuration, the required crane reach is limited to one shaft length or
approximately 30 ft and a 40- to 50-ton crane will suffice. If the required
crane reach is extended to two shaft lengths, a 150-ton crane is necessary (36)
Roughing primaries, designed to achieve 30 to 50 percent suspended solids
removal, were selected for Orlando Easterly. Wastewater temperatures always
exceed the industry-established design temperature of 55°F, above which media
surface area correction factors are not applied. Peak dry weather design
flow is 30 mgd. Raw wastewater characteristics used as the basis of design
and the plant's discharge permit requirements are summarized below (34)(37):
Parameter
TBOD5
SBOD5
TSS
Total N
Total P
DO
Raw
Wastewater
230
75
230
30
Tr
8
Nil
Discharge Permit
Requirements
5
3
1
1
7
The plant's discharge permit requirements are very stringent and will require
excellent process control on a day-to-day basis to meet standards. Using an
average design flow of 24 mgd and based on a primary effluent soluble BODs
concentration of 75 mg/1, a primary effluent equivalent NH3-N concentration
of 24 mg/1, and a denitrification reactor influent N03-N concentration of
22 mg/1 (37), the following process design loading rates can be calculated:
Carbonaceous-Nitrification Reactor
Hydraulic loading: 1.05 gpd/sq ft
Organic loading: 2.65 Ib SBODs/day/lOOO sq ft (first stage)
0.66 Ib SBOD5/day/1000 sq ft (overall)
Ammonia nitrogen loading: 0.28 Ib NH3-N/day/1000 sq ft (stages 2 to 5)
0.21 Ib NH3-N/day/1000 sq ft (overall)
Denitrification Reactor
Hydraulic Loading: 5.3 gpd/sq ft
Nitrate nitrogen loading: 0.97 Ib N03-N/day/1000 sq ft
5-48
(169)
-------
Secondary Clarlfler
Hydraulic loading: 600 gpd/sq ft
At the time of this writing (September 1982), the Orlando Easterly plant was
still undergoing startup and shakedown. No routine operating and performance
data had yet been generated. Since it is the first-of-a-kind and because of
its high performance expectations, progress at this advanced secondary
facility will be followed closely by the municipal wastewater treatment field.
5.8 General Plant Design Considerations
Proper design of an RBC installation entails,, among other things, an adequate
determination of influent and sidestream loadings and the selection of an
overall plant layout that ensures the RBC system is compatible with the other
unit processes selected. In addition, RBC system design should provide
sufficient flexibility to promote good operation and maintenance practices.
In this regard, there are a number of design considerations germane to RBC
installations that should be addressed.
In general, housing RBC units within a building (as opposed to using individual
covers for the RBC units) is undesirable because of the corrosive atmospheric
conditions associated with I^S release and the high humidities that result.
Condensation problems have been encountered on interior building walls in
cold climates, and the associated high humidities and ventilation requirements
increase heating costs. Where buildings are chosen, the building design must
provide for removal of a shaft/media assembly should repair or replacement
prove necessary. In contrast to individual fiberglass covers, full building
cover normally provides more convenient access to RBC's for routine mainte-
nance and visual observation.
In all RBC designs, access to individual shafts for repair or possible removal
must be considered. Bearings should also be accessible for easy removal and
replacement if necessary. The weight of a 27-ft long shaft and media assembly
may be expected to range from 18,000 to 25,000 Ib for clean media depending
on whether standard or high density media is specified. A fully loaded
100,000-sq ft shaft with a 0.1-in. thick biofilm has a dead weight of about
70,000 Ib. Some manufacturers assemble their units in such a way that the
media can be removed from the shaft while the shaft remains in the RBC tank
(Clow, Crane-Cochrane, and Lyco), whereas the media of another manufacturer
(Autotrol) can only be removed nondestructively once the shaft has been
lifted from the tank. The Walker Process media is epoxy bonded to the
shaft. Where all units in a large installation are physically located
very close together, it has been necessary to utilize large off-the-road
cranes for shaft removal. Crane reach, crane size, and the impact of being
able to drain RBC tankage and dry a unit prior to shaft removal should
all be considered when designing the RBC layout.
5-49
(170)
-------
Whenever multiple process trains are employed, provision for positive and
measureable flow control to individual trains is essential. Use of single,
long influent channels with slide gate control for individual trains makes it
difficult for the operator to locate flow maldistributions and implement
appropriate corrective procedures. Splitter boxes and/or weirs are low cost
solutions to this problem. Provision of adequate flow control equipment is
especially important if individual trains are fed from a single channel with
rotation of some trains with and other trains against the direction of plant
flow.
Feeding and discharge flexibility should be considered in RBC design. Step
feed capability can relieve overloaded first stage(s) and potentially decrease
or eliminate excessively thick biofilm growth. Removable baffles may also be
effective in this regard. In underloaded plants, the final stages frequently
must be operated to keep suspended solids in suspension rather than to provide
additional treatment to meet effluent standards; capability to temporarily
bypass these terminal units would result in energy and operation and main-
tenance savings. If sufficient flow flexibility is available, loss of an
individual unit need not result in shutting down an entire process train.
Adjustment of total flow/loading distribution may be an operational necessity,
again emphasizing the need for positive and measurable flow control and/or
splitting.
Load cells, especially in the first stage(s), can provide useful operating
and shaft load data. Where parallel trains are in operation, they can
pinpoint overloaded or underloaded trains. Stop motion detectors, rpm
indicators, and clamp-on ammeters are also potentially useful monitoring
instruments.
The use of deep channels leading to and exiting from RBC tanks has resulted
in solids deposition and subsequent accumulation at a number of installations,
Providing for channel aeration (3.5 scfm/linear foot) or employing channel
configurations that promote adequate scouring velocity should overcome this
problem.
Most RBC designs are based on the units operating at a submergence of
approximately 40 percent based on total media surface area. To avoid
possible shaft overstressing and inadequate media wetting, the RBC
manufacturers strongly recommend never dropping the liquid operating
level below 35-percent submergence. They also recommend a clearance of 4
to 9 in. between the tank floor and the bottom of the rotating media to
maintain sufficient bottom velocities to prevent solids deposition in the
tank.
Depending on media formulation, e.g., the use of carbon black, media
strength can be severely degraded by exposure to sunlight (ultraviolet
degradation). When RBC units are stored on-site for an extended period
5-50
(171)
-------
of time prior to installation, provisions must be made to ensure that
they remain protected from direct sunlight. Media can also be severely
impacted by high wastewater temperatures (>95°F); this is a potential
problem in some industrial applications or in municipal installations
that receive large industrial flows or are located in desert-like
environments.
Equipment warranties can be negotiated with the manufacturers, and, in
some cases, extended equipment warranties have been obtained. This
possibility should be thoroughly considered in equipment specifications.
Although RBC manufacturers continue to make improvements in their equipment,
major equipment problems and failures have occurred at some installations.
The impact of sidestreams from other unit processes on RBC performance must
be considered. Anaerobic digesters increase ammonia nitrogen loadings, and
sludge conditioning processes such as heat treatment contribute increased
organic and ammonia nitrogen loadings. Whenever septic tank discharges
comprise part of the influent wastewater or any unit processes are employed
that may produce sulfide ahead of the RBC units, the additional oxygen demand
associated with sulfide must be considered in system design. High influent
grease loads require the use of primary clarifiers instead of screens.
Most existing air drive installations do not have provisions for measuring
and controlling air flow to individual RBC units. Furthermore, in some
plants (see Figures 4-3 and 4-4), the possibility that some of the air drive
diffusers have plugged cannot be easily verified. Operating an air drive
facility under these types of "blind" conditions adds to the difficulty of
appropriately responding to operational problems that may arise. The
question of how much plant blower capacity should be provided is open to
debate. A critical need exists for more information on the normal range of
air flows that should be expected with air driven units.
5.9 References
1. Fair, G. M., J. C. Geyer, and D. A. Okun.
water Treatment and Disposal. Vol. 2, p.
York City, 1968.
Water Purification and Waste-
25-13, John Wiley and Sons, New
2. Clark, J. H., E. M. Moseng, and T. Asano. Performance of a Rotating
Biological Contactor Under Varying Flow. Journal WPCF, 50(5):896-911,
May 1978.
3. Roy F. Weston, Inc. RBC DIS Subtask on RBC Design Approaches. Internal
report prepared for USEPA under Contract No. 68-03-3019, Cincinnati,
Ohio, March 5, 1982.
5-51
(172)
-------
4. Autotrol Wastewater Treatment Systems Design Manual. Autotrol Corpora-
tion, Bio-Systems Division, Milwaukee, Wisconsin, 1978.
5. Clow Envirodisc Rotating Biological Contactor Systems Catalog. Clow
Corporation, Florence, Kentucky, 1980.
6. RBC systems brochure for Manufacturer X, 1978.
7. Lyco Wastewater Products - RBS Systems Catalog. Lyco Division of Remsco
Associates, Marlboro, New Jersey, 1982.
8. DuPont, R. R. and R. E. McKinney. Data Evaluation of a Municipal RBC
Installation, Kirksville, Missouri. In: Proceedings of the 1st National
Symposium/Workshop on Rotating Biological Contactor Technology, Vol. I,
EPA-600/9-80-046a, NTIS PB81-124539, June 1980. pp. 205-234.
9. Srinivasaragharan, R., C. W. Reh, and S. Liljegren. Performance Evalua-
tion of Air Driven RBC Process for Municipal Waste Treatment. Ibid. pp.
525-552.
10. Orwin, L. W. and C. D. Siebenthal. Hydraulic and Organic Forcing of a
Pilot-Scale RBC Unit. Ibid. pp. 119-136.
11. Stover, E. L. and D. F. Kincannon. Evaluating Rotating Biological Con-
tactor Performance. Water and Sewage Works, 123(3):88-91, March 1976.
12. Murphy, K. L. and R. W. Wilson. Pilot Plant Studies of Rotating Bio-
logical Contactors Treating Municipal Wastewater. Report SCAT-2,
Environment Canada, Ottawa, Ontario, Prepared for Canada Mortgage and
Housing Corporation, July 1980.
13. Chou, C. C., R. J.. Hynek, and R. A. Sullivan. Comparison of Full Scale
RBC Performance with Design Criteria. Bio-Surf Process Technology Book,
Autotrol Corporation, Milwaukee, Wisconsin, (No date).
14. Existing Plant Summaries. Bio-Surf Process Technology Book, Autotrol
Corporation, Milwaukee, Wisconsin, (No date).
15. Personal communication from R. A. Sullivan, Autotrol Corporation,
Milwaukee, Wisconsin, to J. A. Heidman, USEPA, Cincinnati, Ohio,
January 13, 1982.
16. Srinivasaraghavan, R., C. W. Reh, and J. Canady. Plant Scale Investiga-
tion of RBC Process Supplemental Aeration. In: Proceedings of the 1st
National Symposium/Workshop on Rotating Biological Contactor Technology,
Vol. I, EPA-600/9-80-046a, NTIS PB81-124539, June 1980. pp. 575-598.
17• Opatken, E. J. Rotating Biological Contactors - Second Order Kinetics.
In: Proceedings of the First International Conference on Fixed-Film
Biological Processes, Vol. I, EPA-600/9-82-023a, Kings Island, Ohio,
April 20-23, 1982. pp. 210-232.
5-52
(173)
-------
18. Friedman, A. A., R. C. Woods, and R. C. Wilkey. Kinetic Response of
Rotating Biological Contactors. In: Proceedings of the 31st Industrial
Waste Conference Purdue University, West Lafayette, Indiana, May 4-6,
1976; Ann Arbor Science, Ann Arbor, Michigan, 1977. pp. 420-433.
19. Hynek, R. J. and C. C. Chou. Development and Performance of Air-Driven
Rotating Biological Contactors. In: Proceedings of the 34th Industrial
Waste Conference, Purdue University, West Lafayette, Indiana, May 8-10,
1979; Ann Arbor Science, Ann Arbor, Michigan, 1980. pp. 805-815.
20. Levenspiel, 0. Chemical Reaction Engineering. John Wiley and Sons, New
York City, 1972.
21. Letter (data) communication from J. lannone, Roy F. Weston, Inc., Roslyn,
New York, to E. J. Optaken, USEPA, Cincinnati, Ohio, December 7, 1981.
22. Borchardt, J. A., S. J. Kang, and T. H. Chung. Nitrification of Secondary
Municipal Waste Effluents by Rotating Bio-Discs. EPA-600/2-78-061, NTIS
PB-285 240/8BE, June 1978.
23. Antonie, R. L. Nitrification of Activated Sludge Effluent with the Bio-
Surf Process. Presented at the Annual Conference of the Ohio Water
Pollution Control Association, Toledo, Ohio, June 7-13, 1974.
24. Crawford, P. M. Use of Rotating Biological Contactors for Nitrification
at the City of Guelph Water Pollution Control Plant, Guelph, Ontario,
Canada. In: Proceedings of the 1st National Symposium/Workshop on
Rotating Biological Contactor Technology, Vol. II, EPA-600/9-80-046b,
NTIS PB81-124547, June 1980. pp. 1247-1273.
25. Antonie, R. L. Nitrogen Control with the Rotating Biological Contactor.
Autotrol Corporation brochure, Milwaukee, Wisconsin, (No date).
26. Pano, A., E. J. Middlebrooks, and J. H. Reynolds. The Kinetics of
Rotating Biological Contactors Treating Domestic Wastewater. Water
Quality Series UWRL/Q-81/04, Utah State University, College of Engineer-
ing, Logan, Utah, September 1981.
27. Letter (data) communication from R. A. Sullivan, Autotrol Corporation,
Milwaukee, Wisconsin, to R. C. Brenner, USEPA, Cincinnati, Ohio, March
12, 1982.
28. Letter (data) communication from R. J. Hynek, Autotrol Corporation,
Milwaukee, Wisconsin, to E. J. Opatken, USEPA, Cincinnati, Ohio, July 22,
1980.
29. Advanced Wastewater Pilot Plant Treatment Studies. Report prepared for
City of Indianapolis by Reid, Quebe, Allison, Wilcox, and Associates,
Inc., Indianapolis, Indiana, January 1975.
5-53
(174)
-------
30. Letter (data) communication from W. L. Morley, Wastewater Treatment
Plant, Gladstone, Michigan, to R. C. Brenner, USEPA, Cincinnati, Ohio,
November 17, 1981.
31. Murphy, K. L., P. M. Sutton, R. W. Wilson, and B. E. Jank. Nitrogen
Control: Design Considerations for Supported Growth Systems. Journal
WPCF, 49(4):549-557, April 1977.
32. Blanc, F. C., J. C. O'Shaughnessy, D. J. Connick, and D. Wood. Denitri-
fication of Nitrified Municipal Wastewater Using Rotating Biological
Contactors. In: Proceedings of the 1st National Symposium/Workshop on
Rotating Biological Contactor Technology, Vol. II, EPA-600/9-80-046b,
NTIS PB81-124547, June 1980. pp. 1275-1300.
33. Process Design Manual for Nitrogen Control. EPA-625/1-75-007, USEPA,
Center for Environmental Research, Cincinnati, Ohio, October 1975.
34. Dallaire, G. U.S.'s Largest Rotating Biological Contactor Plant to Slash
Energy Use 30%. Civil Engineering, ASCE, 49(l):70-72, January 1979.
35. Hynek, R. J. and H. lemura. Nitrogen and Phosphorus Removal with
Rotating Biological Contactors. In: Proceedings of the 1st National
Symposium/Workshop on Rotating Biological Contactor Technology, Vol. I,
EPA-600/9-80-046a, NTIS PB81-124539, June 1980. pp. 295-324.
36. Personal communications from D. W. Bouck, Dawkins and Associates, Inc.,
Orlando, Florida, to R. C. Brenner, USEPA, Cincinnati, Ohio, November 16,
1981.
37. Personal communication from R. J. Hynek, Autotrol Corporation, Milwaukee,
Wisconsin, to R. C. Brenner, USEPA, Cincinnati, Ohio, September 20, 1982.
5-54
\ I /5/ * U 8 OOVERNMENT PRINTING OFFICE 1S84- 759-102/0973
------- |