v>EPA
United States
Environmental Protection
Agency
Municipal Environmental Research EPA-600/2-78-117
Laboratory July 1978
Cincinnati OH 45268
Research and Development
Expanded Bed
Biological
Treatment
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. US Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interlace in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 Special Reports
9 Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards
This document is available to the public through the National Technical Informa-
tion Service Springfield Virginia 22161
-------
EPA-600/2-78-117
July 1978
EXPANDED BED BIOLOGICAL TREATMENT
E. Timothy Oppelt
John M. Smith
Walter A. Feige
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
U. S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Office of Research and Development,
U. S. Environmental Protection Agency, and approved for publication. Mention
of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
FOREWORD
The 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. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of our environment and the interplay between its components
require a concentrated and integrated attack on the pollution problem.
Research and development is the necessary first step in problem
solution; it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treatment,
and management of wastewater and solid and hazardous waste pollutant dis-
charges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the adverse
economic, social, health, and aesthetic effects of pollution. This publi-
cation is one of the products of that research; a most vital communications
link between the researcher and the user community.
This report describes the results of a pilot-scale evaluation of a
novel secondary wastewater treatment process, Expanded Bed Biological
Treatment. The process is designed to achieve rapid removal of organic
pollutants from settled wastewater by contacting it with high concentrations
of microorganisms in a three phase (wastewater, oxygen gas, inert media)
fluidized bed reactor.
Francis T. Mayo
Director
Municipal Environmental Research Laboratory
111
-------
ABSTRACT
A three-year pilot-scale research investigation at the EPA Lebanon Pilot
Plant was conducted to evaluate the feasibility of a unique biological second-
ary treatment process, designated the Expanded Bed Biological Treatment Process
(EBBT) .
The EBBT process is a three-phase (oxygen/ wast ewater/ sand media) fluid-
ized bed contacting system in which settled wastewater is passed upwards
through a series of two to eight columnar reactors partially filled with fine
sand particles. The velocity of the wastewater flow is sufficient to keep
the sand particles in suspension. The result is a fluidized bed of sand
which provides a large surface area upon which bacteria can grow. These
bacteria remove contaminants from the wastewater as it passes by. An aerobic
environment is provided by cocurrently feeding high-purity oxygen gas into the
base of each reactor or" preferably by diffusing it into the wastewater before
it enters each reactor.
The pilot-scale system was operated at flows ranging from 7.6 to 30.2
fc/min (3 to 8 gpm) and reactor rise rates of 163 to 816 £/min/m (4 to 20 gpm/
System empty bed hydraulic retention times ranged from 15 minutes to 52
minutes in the testing.
The EBBT process achieved an average TCOD removal efficiency of 75 per-
cent and an effluent TCOD of 48.8 mg/fl, (13 mgA/TBOD5) at an empty bed
retention time of 44 minutes and a TCOD volumetric organic loading rate of
6.4 kg/m3/day (400 lb/1000 ft3/day) . Secondary effluent guideline quality
effluent was achieved at empty bed retention times as short as 25 minutes.
The EBBT unit operated at mixed liquor volatile suspended solids concen-
trations in the range of 14,000 to 16,000 mg/Jl . Individual reactor concen-
trations ranged as high as 30,000 mg/Jl. The highly concentrated biomass
contributed to the rapid reaction rates experienced.
Net waste solids production for the process ranged from 0.26 to 0.57 kg
VSS/kg TCOD removed. The solids retention time ranged from 8.7 days to 5.2
days, respectively. The waste solids production values are in a range com-
parable to that reported for many suspended growth oxygen activated sludge
systems.
The ability to maintain stable fluidized bed conditions was found to be a
critical operational factor. The degree of system instability was a function
of the amount of sand support media lost to the process effluent daily and
transported between reactors in the system. Daily loss rates in excess of 1.0
-------
percent of the total system wet sand weight were found to cause severe plugging
of the process lines, excess washout of bacterial mass, and poor treatment
performance. The principal causes of sand bed instability were found to be:
1) transport of sand media through the system by attachment to rising undis-
solved oxygen bubbles in the fluidized beds; and 2) over-expansion of the
fluidized bed as a consequence of excess biological growths on the sand
particles which acted to reduce the effective particle density over a period
of process operation.
Physical and monitoring limitations of the pilot-plant system itself
resulted in frequent operational difficulties and periods of low process
performance. Problems encountered included clogging of process lines,
structural failure of system reactors under conditions of excess pressure,
and low oxygen gas utilization efficiency (measured at only 38 percent). The
low strength of the City of Lebanon wastewater often limited the ability of
the process to achieve higher calculated organic removal efficiencies.
The fluidized bed treatment concept employed in EBBT is believed to have
substantial potential for application in the wastewater treatment field. The
ultimate utility of the process, however, will depend upon the development
and demonstration of: an effective means of controlling the amount of bio-
logical growth per media particle; a means of dissolving high quantities
of oxygen gas external to the reactor to avoid three-phase conditions; and a
system for achieving oxygen utilization efficiencies in excess of 90 percent.
-------
CONTENTS
Foreword iii
Abstract iv
Figures viii
Tables ix
Acknowledgement x
1. Introduction 1
2. Conclusions . 2
3. Recommendations 4
4. Background 5
5. Pilot Scale Studies 7
Objectives and Project Organization 7
Preliminary Studies - Media Evaluation 7
Experimental Results - Two-Column System 9
System Description 9
Results .- 11
Experimental Results - Eight-Column System 17
System Description 17
Results 19
Experimental Results - Modified Eight-Column System . . 24
Description of Modifications 24
Results 27
6. Discussion 36
Treatment Performance 36
Carbonaceous Treatment 36
Nitrification 43
Solids Production 43
Oxygen Utilization Efficiency 46
Expanded Bed Stability 51
References 54
VII
-------
FIGURES
NUMBER PAGE
1 Schematic Drawing EBBT Two Column System 10
2 EBBT Bed-Expansion (Clean Sand) as a function of Rise Rate 12
3 EBBT Eight Column System 18
4 Column Settled Sand Bed Depths at the Completion of Run 8 23
5 Modified Eight Column System 25
6 Sand Recycle System 26
7 EBBT Bubble Trap Device 30
8 COD Removal Efficiency as a Function of Empty Bed 37
Retention Time
9 Total COD Removal Efficiency as a Function of Total COD 39
Volumetric Loading Rate
10 Soluble COD Removal Efficiency as a Function of Soluble 40
COD Volumetric Loading Rate
11 Effluent Total COD as a Function of Total COD Volumetric 41
Loading Rate
12 Effluent Soluble COD as a Function of Soluble COD 42
Volumetric Loading Rate
13 System Average Dissolved Oxygen Profiles for Run 7 48
Through Run 15
Vlll
-------
TABLES
NUMBER PAGE
1 Operational Summary - Two Column System 13
2 Summary of EBBT Wet Sand Loss for Run 3 Through Run 6 15
3 Operational Summary - Eight Column System 21
4 Summary of EBBT Wet Sand Loss for Run 7 and Run 8 22
5 Operational Summary - Modified Eight Column System 28
6 Operational Summary Six Column System 32
7 Summary of EBBT Wet Sand Loss for Run 12 Through Run 14 33
8 Operational Summary Six Column EBBT System 35
9 Summary of EBBT Solids Production Data 44
10 EBBT Mixed Liquor Volatile Solids Levels Run 15 47
11 Summary of Oxygen Addition Rates for the EBBT System 50
12 Summary of Wet Sand Loss from the EBBT System 52
IX
-------
ACKNOWLEDGEMENT
The authors wish to thank the operating staff of the Lebanon Pilot
Plant for their determination, patience, and contribution in the conduct
of this work. Particular recognition belongs to Richard Spellmire,
Selden Heath, Jerry Burden, Albert Oberschlake, and James Bond.
-------
SECTION 1
INTRODUCTION
In April 1970, a research investigation was initiated at the Environmental
Protection Agency's Lebanon Pilot Plant in Lebanon, Ohio, to evaluate the
feasibility of a unique secondary biological treatment concept. The waste-
water treatment concept which evolved from these investigations is the
Expanded Bed Biological Treatment System (EBBT).
EBBT is a three-phase (oxygen-wastewater-sand) fluidized bed contacting
system. Wastewater is passed upwards through a series of columnar reactors
partially filled with fine sand particles. The velocity of wastewater flow
is sufficient to keep the sand grains in suspension. The result is an
expanded (or fluidized) bed of sand which provides a large surface area upon
which bacteria can grow. These bacteria remove contaminants from the
wastewater as it passes by. An aerobic environment is provided by cocurrently
feeding high-purity oxygen gas into the base of each reactor or by diffusing
it into the wastewater before it enters each reactor.
The EBBT system is capable of removing a greater amount of contaminant
per unit volume of reactor than currently available suspended growth air
or high-purity oxygen biological treatment processes. This is possible
because of the significantly increased effective treatment surface area and
the higher concentrations of active organisms which are possible in the
system.
This report outlines the background of the EBBT concept and summarizes
the results of the pilot scale investigations at the Lebanon Pilot Plant.
-------
SECTION 2
CONCLUSIONS
The EBBT system tested at the EPA Lebanon Pilot Plant represents a
potentially valuable secondary wastewater treatment process. The system
was capable of producing an effluent wastewater meeting proposed secondary
effluent quality guidelines at an empty bed retention time (based on Q)
as short as 25 minutes. At an empty bed retention time of 44 minutes the
system achieved an average TCOD removal efficiency of 75 percent, while
producing an effluent TCOD of 48'. 8 mg/1 (13 mg/1 TBOD5) . This was at a
TCOD volumetric organic loading rate of 6.4 kg/m3/day (400 lb/1000 ft3/day).
The EBBT unit was capable of operating at mixed liquor volatile sus-
pended solids concentrations in the range of 14,000 to 16,000 mg/1.
Individual reactor concentrations ranged as high as 30,000 mg/1. The
highly concentrated biomass contributed to the rapid reaction rates
experienced.
Net waste solids production for the process ranged from 0.26 to 0.57
kg VSS/kg TCOD removed. The solids retention time ranged from 8.7 days
to 5.2 days respectively. The waste solids production values are in a
range comparable to that reported for many suspended growth oxygen
activated sludge systems.
The ability to maintain stable fluidized bed conditions was found to
be a critical operational factor. The degree of system instability was a
function of the amount of sand support media lost to the process effluent
daily and transported between reactors in the system. Loss rates, in
excess of 1.0 percent of the total system wet sand weight were found to
cause severe plugging of the process lines,excess washout of bacterial
mass, and poor treatment performance. The principal causes of sand bed
instability were found to be: 1) transport of sand media through the
system by attachment to rising undissolved oxygen bubbles in the fluidized
beds; and 2) over-expansion of the fluidized bed as a consequence of excess
biological growths on the sand particles which acted to reduce the
effective particle density over a period of process operation. The ultimate
utility of the fluidized bed treatment concept (EBBT) will depend upon the
development- of an effective means of positively controlling biological growth
per media particle and dissolving sufficient oxygen gas external to the
fluidized reactor to avoid three-phase conditions.
Oxygen utilization efficiency was measured for one of the fifteen
experimental runs. It was found to be only 38 percent, compared to the often
greater than 90 percent efficiencies reported for commercially marketed high
-------
purity oxygen activated sludge processes. The principal causes of low
efficiency were an unoptimized diffuser design and the inability to recycle
system off-gases from the pilot plant unit.
The calculation of treatment performance efficiency was often limited by
the low strength wastewater at Lebanon, Ohio.
-------
SECTION 3
RECOMMENDATIONS
The pilot scale experiments at the Lebanon Pilot Plant accomplished
their objective as a first examination of the feasibility of a fluidized bed
biological contacting concept. However, many additional technical con-
siderations require examination and resolution before the process can be
moved from development to demonstration. These include:
(1) Additional experimental testing on a reasonable strength wastewater
to establish the process relationship between SRT and treatment
performance, determine optimal particle biomass, refine yield
coefficients, and quantify the impact of temperature and diurnal
flow variation upon performance.
C2) Examination of advanced techniques for dissolution of oxygen into
the wastewater prior to entering the reactor environment.
s
(3) Testing of alternative methods of separating excess biological
growth from the suspension media for positive control of SRT and
percent expansion of the fluidized bed.
(4) Various alternates for final effluent solids separation should be
evaluated. The nature of the suspended solids effluent from the
final reactor stage in the experimental EBBT system at Lebanon
suggests that the requirements may be considerably different from
conventional or oxygen activated sludge plants. One alternate
would be to design a clarification stage for the top of the final
reactor stage. The moderate effluent suspended solids concentration
leaving the final reactor stage in the EBBT system (30 mg/1 to
50 mg/1) suggests the possibility of direct dual media filtration of
reactor effluent as a possibility to be examined.
(5) Any future experimental pilot plants should have at most two
stages which are at least 60 cm in diameter to reduce difficulties
in pilot operation. The test system should avoid introducing
gaseous oxygen into the sand/wastewater environment in the
fluidized reactor. A method for positive control of excess
biological solids is essential. Provision of a capacity for off-gas
recycle to the influent should be considered. The unit should
provide for a flow capacity consistent with an operating rise rate
of 8 to 20 gpm/ft2.
-------
SECTION 4
BACKGROUND
The use of attached biological slimes in wastewater treatment is widely
reported. Packed beds of various types have been employed for decades,
ranging from traditional trickling filter applications to Emscher filters(1),
contact aerators, forced aerated trickling filters(2), and the more recent
commercially marketed plastic media units. Of special note is a 1935 patent
issued to C. C. Hayes (3) which proposed an aerated upflow packed bed system
which was designed for both carbonaceous removal and nitrification. Recently,
columnar systems for denitrification of nitrified effluents have been
demonstrated (4, 5, 6, 7).
The use of a fluidized bed process for biological treatment of secondary
wastewater effluent has been reported(8). The process consists of the
cocurrent flow of air and wastewater upward through a bed of fluidized sand
media. Physical-chemical adsorption of soluble and colloidal material from
the dilute final effluent onto the sand particles creates high substrate
concentrations at the particle surfaces. This permits the development of
biological surface slimes which have been reported to effect 70 to 90 percent
removal of the organics in the final effluent. The process is designated the
Pulsed Absorption Bed Process (PAB) because of the periodic or pulsed motion
created in the fluidized bed as the air bubbles rise through the media.
The addition of coal particles in conventional activated sludge aeration
basins has been reported to increase system efficiency by a mechanism similar
to that employed in the PAB process(9). The suspended coal particles con-
centrate nutrients and organic pollutants from the mixed liquor while
providing surface area for microbial growth in the "enriched" microenvironment
adjacent to the particle.
Several reports have appeared describing an increase in the capacity of
activated carbon systems which was attributed to biological growth on the
coal particles. Hopkins, et al, reported the development of bacterial slimes
on activated carbon particles in expanded bed contactors(10). Some
biological removal of organics was accomplished, but the absolute amount
was not estimated.
In 1969, the addition of pure oxygen to carbon adsorption columns at
the EPA Blue Plains Pilot Plant and at the EPA Pomona Pilot Plant was
initiated to examine the potential for nitrification in the column environment
resulting from attached biological growths. Moderate nitrification was
-------
reported at Blue Plains where lime-clarified secondary effluent was
studied (11). Packed beds receiving secondary effluent at Pomona were
capable of oxidizing 3 to 4 mg/1 of ammonia to nitrate (4).
The use of pure oxygen in waste treatment is, of course, not a new
concept. The potential benefits of using oxygen gas in activated sludge
treatment were first recognized by Pirnie (12) when he proposed the so-called
"bio-precipitation" process over 20 years ago. Biological success with the
process was later achieved at bench scale by Okun (13) and finally at pilot
scale by Budd and Lambeth (14). In the past, low oxygen utilization
efficiencies for the process precluded its full-scale application. Since
1968, however, the development of "UNOX" process by Union Carbide Corpora-
tion (15), the "SIMPLOX" process by Cosmodyne Corporation (16), and the
AIRCO pipeline reactor, oxygen activated sludge treatment has progressed
rapidly from the drawing board to a full-scale marketed process. The UNOX
system, as an example, has been reported capable of attaining BOD5 and
suspended solids removals of 90 percent at a retention time of 1.4 hours
(based on Q) while achieving up to 90 percent oxygen utilization
efficiency (15).
Since the completion of this investigation, the results of other studies
of the treatment of wastewater in fluidized beds have been reported. Jeris,
et al., reported the results of testing on municipal wastewater at Nassau
County, New York (17). Separate pilot studies at flows ranging from
150,000 to 300,000 I/day (40,000 gpd to 80,000 gpd) were conducted for
carbonaceous BOD removal, nitrification and denitrification. BOD removals
of 90 percent were reported. Studies have also been reported by Scott and
Rancher at the Oak Ridge National Laboratory, using a tapered fluidized bed
reactor for biological denitrificaton (18).
The Expanded Bed Biological Treatment concept (EBBT) was designed to
combine the acknowledged benefits of oxygen aeration technology and attached
growth biological systems into one process. As it was initially conceived,
EBBT was intended to be a high-purity oxygen, fixed-film nitrification
system. It was hoped that such a system would improve upon the variable
nitrification performance of single sludge systems, while overcoming the
greater expense of the more reliable two sludge system for nitrification.
The intimate contact with surface-bound slimes, greater cell residence
times, and use of pure oxygen in EBBT were viewed as key factors to improved
performance.
When the expanded bed pilot test program was initiated at the Lebanon
Pilot Plant in September 1970, the project scope was expanded to also
consider treatment of primary effluent in a staged columnar system designed
to achieve both organic removal and nitrification. The research was conducted
over a three-year period ending in September 1973.
-------
SECTION 5
PILOT SCALE STUDIES
OBJECTIVES AND PROJECT ORGANIZATION
The primary objective of the research program at Lebanon was to deter-
mine the feasibility of pure oxygen biological oxidation of organic wastes
with subsequent nitrification of ammonia in a three-phase fluidized bed
system. The results were to be used to evaluate process performance in
comparison with the best state-of-the-art suspended growth reactor. Increased
efficiency over conventional treatment schemes was anticipated due to:
1) high effective mixed liquor biological solids concentrations possible
through attached films; 2) the use of high-purity oxygen; 3) shorter required
contact times; 4) benefit gained by having to gravity-separate only waste
solids from the system; 5) elimination of sludge recycle; and, 6) resistance
to shock loads.
The study was divided into four distinct phases. Phase I included a
literature review and desk top design of the experimental system. Phase II
provided for final design and construction of a 37.8 liter per minute
(10 gpm) pilot scale test facility. Phase III consisted of preliminary
studies to determine best fluidization conditions, fluidization media, and
process flow rates. Phase IV, with which the bulk of this report is
concerned, consisted of the actual biological testing. The object in this
phase was to determine process kinetics, maximum organic removal efficiency,
minimum contact time, oxygen utilization efficiency, and sludge production
rates.
PRELIMINARY STUDIES - MEDIA EVALUATION
The first step in the EBBT evaluation involved the selection of a
fluidization media suited for use in the reactors. The objective was to
provide a particle small enough to maximize the available surface area for
bacterial attachment, but with a specific gravity conducive to bed control
under fluidized conditions.
The physical test system used in the initial studies was a modification
of a pilot-scale system used previously for columnar packed-bed denitrification
research (7). The primary concern was ease of media evaluation and not
treatment optimization. The unit consisted of five 10.2 cm (4-inch) I.D.
columns connected in series. Each column was 3.66 meters (12 feet) high with
27.9 cm (11 inches) of gravel support media in the base. Primary effluent
-------
from the city of Lebanon, Ohio Sewage Treatment Plant was pumped into the
base of the first column with a variable-speed, positive-displacement pump.
Oxygen gas was fed into each column base through 7.6 cm (3-inch) diameter
stainless steel diffusers at rates ranging from 100 to 200 cc/min. The
system influent and effluent were routinely analyzed for TCOD, SCOD, TOC,
SOC, pH, TSS, YSS, temperature, NH3-N, and N03.
The first media examined was No. 1 anthracite coal. A sieve analysis
revealed that this coal possessed an effective size of 0.62 mm (mean size
1.0 mm) with a 1.77 coefficient of uniformity. This media is typical of
that used in many dual-media filter applications.
Each column was filled with 1.83 meters (6 feet) of coal. The system
was then operated for 30 days at a feed rate of 11.6 liters/min (1.75 gpm),
which corresponded to a column rise rate of 816 1/min/m2 (20 gpm/ft2), a
rate often employed in filter backwashing. Results were encouraging,
as moderate organic removals were obtained during this period (52 percent
removal of SCOD, and 32 percent of TOC). However, the coal media proved
inadequate as a fluidization media. Bed separation and system plugging were
frequent, and a considerable amount of the media washed into the effluent
daily. Reduction of column flow rates reduced media loss somehwat, but
only at the expense of an increased frequency of column plugging which
resulted from incomplete fluidization.
Other media were tested in a similar fashion, some in the 10.2 cm
(4-inch) I.D. column system, and others in a larger system (25.4 cm (10-inch)
I.D. column, which is described later. These media included:
Medium cinders - sp. gravity 1.1, mean particle diameter 2.0 mm,
size distribution 0.21 mm - 26.67 mm
Crushed cinders - mean particle diameter 1.2 mm, size distribution
0.21 mm - 0.373 mm
Hydraulically classified cinders - mean particle diameter 0.5 mm,
size distribution 0.420 mm - 1.41 mm
Filter sand - sp. gravity 2.65, mean particle diameter 0.5 mm, size
distribution 0.25 mm - 0.85 mm
The hope in employing the cinder material was that the numerous cavities
in each particle due to its porous nature would provide a greater surface
area for slime growth than the coal. The use of this material, however, was
found undesirable for several reasons: 1) the particle size distribution was
too extreme; 2) the specific gravity ranged too wide (0.4 to 1.1 g/cc); and,
3) the particles broke down from abrasive contact in the system. This
resulted in high rates of media washout and frequency of plugging of the
lines connecting the columns.
-------
The fine filter sand proved to be the most satisfactory of those media
tested. The physical properties of this sand included an effective size of
0.37 mm (mean particle diameter 0.5 mm) and a coefficient of uniformity of
0.36. The sand's high uniformity and greater density (sp. gravity 2.65 g/cc)
produced much more stable fluidized bed conditions. The greater upflow
velocity required to suspend the particles not only provided better fluid
mixing in the column, but also appeared to reduce the frequency of plugging
inside the column.
All tests employing filter sand as bed media were conducted in the
larger, 25.4 cm (10-inch) I.D. column system. Once the fluidization require-
ments and the hydraulic acceptability of the sand were determined, treatment
performance testing was initiated. A description of this system and the
results obtained are described below.
EXPERIMENTAL RESULTS - TWO-COLUMN SYSTEM
System Description
In November 1970, construction of a larger scale EBBT pilot system
was initiated. This system was designed to provide liquid retention times
similar to those experienced in the preliminary system, while avoiding some
of the hydraulic and mechanical problems experienced with the smaller unit.
The 25.4 cm (10-inch) diameter columnar-unit was completed in late
February 1971, and testing with cinder media of various types, as described
above was started in April. The media evaluation trials were completed in
September 1971, with the selection of the 0.5 mm filter sand as the most
promising media.
Figure 1 is a schematic drawing of the two-column system, prior to
the initiation of biolgoical studies in September 1971. The unit consisted
of two 3.66 m (12-foot) high, 25.4 cm (10-inch) I.D. clear acrylic columns,
connected in series. An 8.9 cm (3-1/2 inch) diameter stainless steel
diffuser was installed in the base of each column to permit addition of
oxygen gas.
The diffusers were mounted below a 1.9 cm (3/4-inch) thick circular PVC
flow distribution plate. Approximately 27.9 cm (11 inches) of support
gravel was placed above the plate, followed by 1.83 m (6 feet) of the sand
media. To limit media washout, a 45.7 cm (18-inch) diameter expanded
section was later added to the top of the second stage. It was anticipated
that the resulting drop in upflow velocity in the expanded section would
prevent suspended sand particles from being carried into the effluent.
Primary effluent from the Lebanon Sewage Treatment Plant was fed into
the bottom of the first stage through a variable-speed positive-displacement
pump. The effluent from the top of the first stage was then carried under
pressure to the base of the second stage through a 3.8 cm (1-1/2 inch) PVC
pipe. Stage No. 2 effluent continued to a 661.5 1 (175-gallon), 0.71 m^
-------
[EXPANDED SECTION
I WITH BAFFLE
EFFLUEN1
CLEAR COLU/VU
HEIGHT=3.66m.
DIAMETER=25.4
SAND MEDIA
JS-
km.
DISTRIBUTION AND
SUPPORT MEDIA.
X
PRIMARY EFFLUEN1
FEED PUMP/ fc
~~G/
r *
t
-
p
MHM^BIM
-
d
[o,
~~-~
\
rn
E
»*-
1
'U*** Vj
LHMV^^J
T
i
b
~T
*
COLUMN #2
EFFLUENT
FINAL EFFLUENT
i
\ "^FIWAI riAPIFIPP
1
1
^/SURFACE AREA=0.71m
WASTE SLUDGE
AND SAND
f\ CIIDDIY
'
EFFLUENT RECYCLE PUMP
FIGURE 1. EBBT TWO-COLUMN SYSTEM
10
-------
(7.67 ft2) final clarifier. A portion of the stafe No. 2 effluent could also
be recycled by a second variable-speed positive-displacement pump back to the
incluent stream. The purpose of this was to add rise-rate control flexibility
and limit depletion of the sand beds by recycling a portion of the sand media
being lost to the effluent.
Pure oxygen could be supplied to all application points in the system
by a bank of three (84.9 m ) liquid oxygen cylinders. The gas feed
pressure was maintained at 28.1 x 10^ kg/m2 (40 psig) and the gas was metered
to the system through a panel of rotameters. Provision was also made for
adding oxygen gas to the suction side of recycle pump. Preliminary testing
indicated that up to 20 mg/1 dissolved oxygen could be added to the recycle
stream in this manner.
Composite samplers were located on the influent and effluent streams of
the system. Grab samples could also be obtained from 2.54 cm (1-inch) sample
taps located at the base of each column and in the free board space at the
top of each column. Provision was not made on the initial unit for measuring
system off-gas volume or quality.
Results
A total of six experimental runs were conducted with the two-column
system. Runs 1 and 2 were devoted entirely to optimizing system hydraulics
and determining the effect of a series of modifications to the expanded
section on the second stage. Both runs were of short duration. Significant
biological activity was not observed until the end of Run 2.
Run 2 also provided an indication of the hydraulic performance of the
sand beds under varying column rise rates. In Figure 2, percent bed ex-
pansion as a function of column rise rate is shown for clean sand (minimal
attached slimes) under conditions of a 200 cc/minute oxygen feed rate and
an initial unexpanded bed height of 190.4 cm (75 inches).
Treatment performance testing began with the initiation of Run 3 on
September 30, 1971. The purpose of this and subsequent runs was to first
establish a viable microflora on the sand media, and then vary basic
operational parameters in a predetermined manner to evaluate: 1) substrate
removal efficiency as a function of system empty bed contact time;
2) optimum column rise rate for maximum internal mixing with minimum media
washout; 3) the effect of temperature upon performance; 4) maximum effective
mixed liquor solids concentration; and, 5) required oxygen rates.
Table 1 summarizes average operating conditions and system performance
for Run 3 through Run 6, covering a period from September 30, 1971 through
April 7- 1972. Throughout this period the system was monitored 24 hours
per day, seven days per week.
From Table 1 it can be seen that the system was operated with empty-
bed contact times ranging from 15.5 minutes to 46.6 minutes (based on Q).
The column rise rate was varied between 163.2 1/min/m2 (4 gpm/ft ) and
816 1/min/m2 (20 gpm/ft2) by adjusting the effluent recycle rate. The
11
-------
NOTE:1.0g.p.m./ft.=40.8l./min./m2
4 8 12 16 20 24
COLUMN RISE RATE (g.p.m./ft?)
FIGURE 2. BED EXPANSION (CLEAN SAND)
AS A FUNCTION OF RISE RATE
12
-------
TABLE 1 OPERATIONAL SUMMARY TWO-COLUMN SYSTEM
Parameter
Unit
Run 3
Run 4
Run 5
Run 6
Waste treated
Influent temperature
Influent flow rate (Q)
Recirculation flow rate (R)
System empty bed contact time
(based on Q)
Column rise rate (Q+R)/A
Clarifier overflow rate
Oxygen addition rate (at S.T.P.)
(system average)
Dissolved oxygen concentration
influent
stage No. 2 effluent
- clarifier effluent
recirculation stream
Performance Summary
Influent TCOD
SCOD
TOC
SOC
TSS
Effluent TCOD
SCOD
TOC
SOC
TSS
,TCODin-TCODout , .,
TCOD removal ( TCODin *100)
"COD removal fSCODin-SCODout Q
Overall COD fTCODin-SCODout
overall LULI (. TCODin xiuuj
removal TOCin.TOCout
TOC removal l TOC in X1UUJ
__ , ,TSS in-TSS out . _-,
loo in
°c
1/min
1/min
min
1/min/m
1/min/m
cc/min
gms/m3
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
rag/1
rag/1
mg/1
mg/1
mg/1
mg/1
%
%
%
%
%
*
Primary
Effluent
19.6
30.2
7.6
15.5
816
42.6
1233
46.4
1.1
4.0
2.2
19.6
233.2
93.3
57.0
29.0
82.0
171
68.2
51.0
21.0
61.0
26.6
26.9
70.7
10.6
25.6
Primary
Effluent
18.5
15.1
15.1
23.3
652.8
21.3
750
56.5
1.3
3.5
1.5
16.1
230.0
115.0
71.0
37.0
179
63.0
52.0
19.0
.
22.2
45.2
72.6
26.8
Primary
Effluent
12.8
7.6
0
46.6
163.2
10.6
826
124.3
3.3
4.59
18..'
288.8
78.3
71.3
35.8
111.0
100.0
36.5
26.1
14.0
28.4
65.3
53.5
87.5
61.8
74.4
Primary
Effluent
12.2
15..1
0
23.3
326.4
21.3
1030
77.5
4.9
7.3
8.3
179.6
69.9
70.7
28.7
84.2
86.2
35.7
33.5
15.6
32.7
52.0
48.9
80.1
52.6
61.1
*l/min x 0.265 = gpm 0.0245
**l/min/m2 x 0.0245 gpm/ft2
***gms/m3 x 0.00835 - lbs/1000 gal.
13
-------
surface overflow rate on the final clarifier ranged from a low of 1.53 x
104 1/min/m2 (375.5 gpm/ft2) in Run 5 to a high of 6.13 x 104 1/min/m2
(1502 gpm/ft2) in Run 3, and produced total system suspended solids removal
of 74.4 percent and 25.6 percent, respectively.
The oxygen addition rates and dissolved oxygen concentrations for
various system locations represent average conditions. Daily and hourly
conditions varied somewhat from these values. Average clarifier effluent
dissolved oxygen concentrations did not fall below 1.5 mg/1. The dissolved
oxygen concentrations in the effluent of stage 2 averaged between 3.5 mg/1
(Run 4) and 7.4 mg/1 (Run 5) while daily values never fell below 0.8 mg/1
in any of the runs.
Table 1 also shows the average wastewater characteristics of the
influent and effluent streams. The COD data reveals the relatively low
strength of the Lebanon wastewater. Removal efficiencies for both total and
soluble COD (TODC and SCOD), organic carbon (TOC and SOC) , and total sus-
pended solids (TSS) are given. In terms of monitoring system performance for
control purposes, greater significance was attached initially to the removal
of soluble waste components, the purpose being to eliminate the effect of any
variation in solids settling efficiency in the final clarifier. Soluble COD
removal ranged from a high of 55 percent in Run 6 to a low of 26.9 percent
in Run 3, representing absolute effluent SCOD values of 35.7 mg/1 and
68.2 mg/1, respectively. In most instances higher percent removal values
were believed to be limited by the low influent waste strength. The effluent
quality during Run 5 and Run 6, for instance, was considered good (36.5 mg/1
and 35.7 SCOD, respectively), but on a percent removed basis, performance ap-
peared poor.
Influent and effluent samples were analyzed for five-day biochemical
oxygen demand (BOD,-) only periodically to determine the average relationship
between BODg and the other performance parameters. The following relation-
ships remained relatively constant throughout the research:
Ratio EBBT Influent EBBT Influent
TCOD/BOD5 2.1 3.47
TOC/BOD5 1.0 0.62
The calculated effluent BODs using these ratios ranged from approxi-
mately 32 mg/1 in Runs 3 and 4 to 16 mg/1 in Run 5 and 21 mg/1 in Run 6.
BODs removal efficiency ranged from a low of 44 percent in Run 3 to a high
of 77 percent in Run 5. Treatment efficiency was directly proportional to
the empty bed contact time employed (based on Q and the total empty column
volume) . Best removals were obtained at the empty bed contact time of
46.6 minutes while reduced treatment efficiency was experienced at the
shortest contact time (15.5 minutes).
14
-------
Process performance can be described in terms of a parameter defined as
overall COD removal:
TCODin - SCODout
% = TCODlHx 10°
A number of studies have suggested that effluent SCOD is a better indicator
of the quantity of residual organic matter in biological treatment process
effluents than is BOD5 (18) (19). The measurement of COD is recommended in
lieu of BODj- because of the recognized time and reproducibility limitations
of the standard BOD5 test. The calculation of overall COD removal acknowledges
that suspended material present in the system influent wastewater can be
regarded as potential substrate for the microorganisms and not as cells.
The use of effluent soluble COD is designed to exclude sludge cells from
effluent residual organics and to divorce process biological (kinetic)
performance from final clarifier efficiency.
One drawback to the use of overall COD removal as an indicator of
treatment performance is that the COD analysis measures wastewater organic
matter without regard to its availability to microorganisms. Jenkins and
Menar(19) found, for instance, that a nondegradable effluent COD amounting
to 15 percent of the influent COD existed for the city of Richmond,
California wastewater. At the same time, the contribution of COD contained
in effluent suspended solids must be recognized as important in terms of
total system performance. For the purposes of this paper, overall COD
removal values are included for comparison, because much of the available
data relating to the performance of pure-oxygen activated sludge systems is
presented in this manner. Overall COD removal for EBBT ranged from 70.7
percent in Run 3 to a high of 87.5 percent in Run 5.
An important factor in the experimental runs conducted with the two-
column system was the media carryover rate. While the fine sand employed
was the most acceptable of those tested in the preliminary studies, the
problem of media carryover was only reduced and not eliminated through its
use as the EBBT media. Sand bed depletion did occur in the media selection
runs but at a rate lower than with any of the other media under similar
conditions. Sand carryover data from Runs 3 through 6 is given in Table 2.
TABLE 2. SUMMARY OF EBBT WET SAND LOSS FOR RUNS 3 THROUGH 6
Run
3
4
5
6
Column Rise Rate
(1/min/nr)*
816
652.8
163.2
326.4
Wet Sand
Average
1.86 [.68%]
0.86[.31%]
0.39[.14%]
1.76 [.64%]
Loss (kg/day)**
Range
0.45-4.13[1.50%]
0.08-9.53[3.48%]
0.13-1.27[.46%]
0. 07-16. 80[6. 13%]
*1 gpm/ft^ = 40.8 l/min/nv'
**1 Ib/day = 0.454 kg/day
[-%] represents daily wet sand loss values expressed as a percent of the total
wet sand in the system
15
-------
Wet sand losses were determined two or three times per week by removing
all collected sand from the cone of the system final clarifier. Biological
solids were first dislodged from the sand by washing in hot tap water.
The excess water was drained away and the remaining sand mixture weighed.
Initially, a dry sand weight was also determined. However, because the
wet mixture consistently possessed a 21-24 percent moisture content, wet
sand weight was considered equally indicative of dry sand carryover
performance. In all subsequent experimental runs only wet sand weight
is reported.
In general, the rate of sand loss increased with the column rise rate.
(with the exception of run 6). The average daily loss rates shown in
Table 2 were calculated by dividing the total cumulative wet sand losses
for each run by the duration of the run in days. Each run exhibited a
similar pattern of sand loss. Reasonable system stability was possible
for the first 10 to 20 days of operation. During this period the losses
remained below 0.5 kg/day, i.e., less than 0.2 percent of the total system
sand weight/day. Then, over a period of 5 to 10 days, the rate of sand loss
would increase rapidly, until it reached the maximum values shown in the
range column in Table 2. This phenomenon was also accompanied by increasing
depths of sand bed expansion, which in some instances resulted in the complete
filling of the freeboard space between the top of the bed and the top of the
column. The resulting problems of system plugging and loss of microbial
mass would force the termination of the experimental run.
A number of techniques were tested to deal with sand losses. In Run 3
the effluent recycle pump was connected to the cone of the final clarifier
in an attempt to return carryover sand to the system via the recycle stream.
The 7.6 1/min (2 gpm) rate employed, however, was inadequate. The rate was
then doubled to 15.2 1/min (4 gpm] in Run 4 and a verticle baffle was
installed in the expanded section on column #2 to limit the amount of sand
being drawn into the system effluent. The column rise rate was also reduced
to 652.8 l/min/m2 (16 gpm/ft2]. This combination worked effectively until a
surge of grease-laden septic tank wastes was discharged into the Lebanon
Treatment Plant and subsequently entered the pilot system. The grease coated
the sand particles, causing excessive bed expansion and loss of media. The
run was terminated.
Subsequently, a skimming ring was installed in the system feed tank to
help remove grease and scum before it entered the columns. No further
grease problems were encountered.
In spite of continued attempts to recycle sand and the use of various
levels of column rise rate, Run 5 and Run 6 also failed to find the operating
combination that would prevent excessive sand losses.
Several conclusions based on the operation were made. First, as Ibng
as daily sand losses did not exceed 1.0 percent of the total system sand
weight C\>2.74 kg/day) reasonably trouble-free process operation was possible.
A column rise rate of 652.8 1/min (16 gpm/ft2) appeared to be the maximum
16
-------
possible rate. Operation at lower rise rates generally extended the length
of time the system could operate before the 1.0 percent daily sand loss rate
was exceeded. Even the lowest rate employed, 163.2 1/min/m (4 gpm/ft ),
did not prevent eventual rapid and uncontrollable sand losses.
At the completion of Run 6 the operation of the two-column experimental
system was halted. Based on the results of the first six runs the decision
was made to expand the pilot facilities into a more functional and flexible
eight-column system. This system and a summary of its treatment performance
are described below:
EXPERIMENTAT RESULTS - EIGHT-COLUMN SYSTEM
System Description
Following the completion of Run 6, construction of an eight-column
pilot-scale system was initiated, utilizing the two initial 25.4 cm I.D.
(10-inch) acrylic columns and six new PVC columns, which were identical
in size to the acrylic columns. The primary purpose of the new design was
to increase system flexibility with respect to wastewater retention time,
column rise rates, and oxygen addition points. At the same time the plug-
flow sequence of eight reactors in series was intended to increase system
biological performance and increase the opportunity for nitrification in
the final columns of the system. No noticeable nitrification occurred in
the first six EBBT runs because of the low wastewater temperatures and
because the short duration of the experimental runs did not permit
establishment of nitrifiers.
The initial eight-column configuration is shown in Figure 3. Settled
wastewater entered the system through the feed tank used previously on the
two-column system (not shown here). The skimming ring was retained to
protect the system against an influx of grease-laden material and to
routinely remove floatables. From the feed tank, wastewater was pumped,
as before, into the base of the first column. The two clear acrylic
columns were used as the number 5 and number 8 columns with the six new
PVC columns placed in the remaining positions. The eight columns were
connected in series with 3.8 cm (1.5-inch) diameter PVC pipe which carried
liquid from the top of each column to the base of the next. The system
was constructed so that any column or columns could be taken off-stream
without interrupting overall system operation.
The method of adding oxygen to the system was also modified. The
stainless steel diffusers were removed from the two initial columns and
they along with the six new columns were fitted with a new type of diffuser.
The new diffuser consisted of 3.8 cm (1.5-inch) diameter, 0.79 mm (1/32-inch)
thick disk of a flexible plastic material similar to "naugahide". The
material has very small pore openings (indistinguishable to the naked eye).
This material is used by a local firm in the manufacture of tubular diffusers
for use in full scale activated sludge systems. The disk of plastic was
sandwiched between two 3.2 mm (1/8-inch) thick rings of plastic used as a
support frame. These units were, in turn, secured in a tee at the bottom of
the connecting line betwen each column. A diffuser was also added to a tee
17
-------
EXPANDED SECTION
WITH BAFFLE /
oo
6 PVC
COLUMNS
CLEAR
COLUMN
PRIMARY
EFFLUENT
FEED PUMP
SYSTEM
EFFLUENT
EFFLUENT RECYCLE PUMP
FINAL
EFFLUENT
FINAL
CLARIFIER
WASTE SLUDGE
AND SAND
O2SUPPLY
FIGURE 3. EBBT EIGHT COLUMN SYSTEM
-------
on the influent line to the first column. The primary purpose of these
changes was to allow the flexibility of step aerating the system, and to
offer time for bubble contact with the wastewater prior to its entry into
the reactors. At the same time, the plastic diffusers were capable of
producing a much finer bubble than the stainless steel diffusers. Over
18 months of continuous use, the plastic material showed no significant
deterioration.
From the clear acrylic column in the eighth position, process water
was then carried to the original final clarifier. A recycle pump was
included to permit sand recirculation from the final clarifier to the first
column. The baffled expanded section was allowed to remain on the eighth
column initially, but was later removed when it was determined that it did
not help control sand loss.
No new provisions were made to attempt to control sand losses from
the system. It was anticipated that the greater flexibility of the eight-
column system would provide improved operating conditions for sand bed
stabilization as well as biological efficiency. At the time of construction,
however, no understanding had been reached regarding the true cause of sand
loss.
Results
Five experimental runs were conducted on the enlarged system. The
first two of these, Run 7 and Run 8, were conducted on the eight-column
system in Figure 3. Later runs were conducted on a modified version of this
system, which will be described later.
Construction of the enlarged EBBT system was completed in June 1972.
Support media and 2.29 meters (7.5 feet) of sand media were placed in each
column, leaving an unfluidized free board height of 0.76 meters (2.5 feet).
On July 5, the system was started at an influent rate of 16.6 1/min (4.4 gpm).
It was found that this system was much more difficult to operate than the
initial two-column system. It was a major undertaking to sequentially
fluidize the eight columns. Because the PVC columns were opaque the operators
could not determine when the sand became properly fluidized. Frequently,
lower sections of the bed would apparently fluidize while the top layers
would remain intact and rise as a plug as the column filled. If several
start-up attempts failed, an access plug at the top of the column was removed
and a long, flexible plastic rod was inserted into the column. The rod
was worked up and down to break up the sand as the column filled. Once the
blocked column was started, the top was resealed and similar steps were taken
to start the remaining columns in their numerical sequence. Frequently,
more than 46 kg (100 pounds) of sand was lost to the clarifier during the
start-up procedure.
Particular attention had to be given to column pressures during both
start-up and normal operation. Excessive sand carryover between columns
sometimes caused complete plugging of the lines between the columns. The
positive displacement feed pump was capable of producing high column pressures
19
-------
in these instances. Under average operating conditions the pressure in the
first column was in the range of 146.5 to 171 kg/m2 (30 to 35 psig).
Pressures in succeeding columns decreased linearly down to a value of 19.5 to
29 kg/m2 (4 to 6 psig) in the last column. Pressures were read from gauges
located at the base of each column. Pressures above 195 kg/m2 (40 psig)
generally caused considerable leakage at the flanges on top of the PVC columns
and in some cases, when system plugging was rapid, tops even ruptured.
After approximately one week of trial runs and several sand replacements
Run 7 was initiated at a feed rate of 30.2 1/min (8 gpm) with no effluent
recycle. This provided a calculated liquid retention time of 27.6 minutes,
or an actual empty bed retention time of 39.0 minutes. Upon start-up the
expanded bed depth in the clear columns was 2.74 m (9 feet), leaving 0.305 m
of free board above the bed surface.
The system was operated for two and one-half weeks with minimal sampling
while slimes developed on the sand particles. Dissolved oxygen, pH,
temperature, and flow rate were the only parameters monitored. Initially,
oxygen gas was added only to the first column. An unadjusted rate of
500 cc/minute would maintain dissolved oxygen levels of 1.0 mg/1 or greater
in all columns. As the biological population developed the rate was
increased and oxygen was also applied to additional columns. Sampling to
determine waste treatment efficiency was initiated July 25 and continued to
August 15, 1972, when Run 8 was started.
In Run 8, the system influent flow rate was reduced to 26.5 1/min (7 gpm)
with no effluent recycle. This increased the liquid detention time to 31.5
minutes, or an empty bed detention time of 44.4 minutes. Data were collected
under these conditions for 16 days until excessive media loss from the
system combined with a weekend power failure forced termination of the run.
The operating conditions and experimental results of Run 7 and Run 8
are summarized in Table 3.
The data show that the treatment efficiency was significantly better
than that achieved in previous runs. Average SCOD removal was 71 percent in
both runs compared with a previous maximum of 53 percent in Run 5. This
likely resulted from several factors including: higher wastewater temperatures;
higher soluble waste strength; and an increased degree of fluidization from
the doubling of the column rise rate over that of Run 5.
It was discouraging, however, that the increased detention time of Run 8
did not improve treatment efficiency over that of Run 7. While it was
possible that the limit of biological degradability of the wastewater had
been approached, further analysis revealed that the principle limitation to
increased treatment efficiency in Run 8 was the continued inability to control
sand media losses from the system. Sand losses for Run 7 and Run 8 are shown
in Table 4.
20
-------
TABLE 3. OPERATIONAL SUMMARY EIGHT-COLUMN SYSTEM
Parameters
Waste treated
Influent temperature
Influent flow rate (Q)
Recirculation rate (R)
Column rise rate (Q + R/A)
Liquid retention time
Empty bed retention time
Final clarifier
Oxygen addition rate (8S.T.P.)
(system average)
Dissolved oxygen concentration
influent
column No. 1 effluent
column No. 4 effluent
column No. 8 effluent
(system effluent)
Performance summary
Influent TCOD
SCOD
Effluent TCOD
SCOD
TCOD rcmovaHTCODin-TCODoutx 100)
"COD removal fSCODin-SCODout ,
aiuu removal i SGODin ±uuj
Overall COD JCODin-SCODout .- .
Removal ( TCODin '
Unit
-
°C
1/min*
1/min
1/min/m
min
min
1/min/m2
cc/min
gms/m3***
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
%
%
%
Run 7
Primary
Effluent
20.6
29.9
0
644.6
27.6
39.0
42.2
4508
172
1.3
1.5
4.4
5.7
161.5
101.0
29. S
71
81.7
Run 8
Primary
Effluent
22.1
26.5
0
571.2
31.5
44.4
37.2
6245
268.8
1.0
2.7
5.1
4.2
116.3
33.8
71
* 1/min x 0.265 gpm
** 1/min/m2 x 0.0245 gpm/ft2
*** gms/m3 x 0.00835 = lbs/1000 gal.
21
-------
TABLE 4. SUMMARY OF EBBT SAND LOSS FOR RUN 7 AND RUN 8
Run
7
8
Column Rise Rate
(l/min/m2)*
644.6
571.2
Wet Sand Loss
Average
4.82 [.45%]
4.15 [.44%]
(kg/day)**
Range
3.6-7.72 [
0.9-7.27 [
,72%]
.77%]
*1 l/min/m2 = 0.0245 gpm/ft'
**1 kg/day =2.20 Ib/day
The sand loss rates were high throughout both runs. The average daily
loss rates of 4.82 kg/day and 4.15 kg/day were more characteristic of the
final portions of Runs 3, 4 and 6.
There is little question that the continuous and significant loss of
media available to support biological growth affected treatment performance,
especially in Run 8. Better than 30 percent of the sand initially in the
system was lost by the end of Run 7 and an additional 20 percent was lost by
the end of Run 8. In addition to the loss of sand there was also a
redistribution of the remaining sand. This can be seen in Figure 4, which
indicates the amount of sand remaining in each column where Run 8 was
terminated.
Profiles of SCOD removal through the system were compared for Run 7 and
Run 8. Data from early in Run 7 (before large cumulative sand loss)
indicated first-order removal kinetics. Similar data for the latter portions
of Run 8 indicated that little removal was effected in the first two reactors
and almost no removal in the third and fourth reactors. The bulk of removal
occurred in first-order fashion in the final four reactors. The net effect
was reduced overall treatment efficiency, in spite of increased detention
time.
At the completion of Run 8, it was apparent that sand bed attrition
was the factor most limiting to steady state system operation and the
attainment of higher treatment efficiencies. The principle cause of
excessive sand loss was believed to be the reduction in particle net density
by uncontrolled surface growth of biological slimes. This, in turn, would
cause overexpansion of the fluidized bed and loss of coated sand media to
the effluent The fluid shear forces within the reactors were apparently
insufficient to effect sloughing of excess biological growth. Sand lost to
22
-------
12
^tl Q
I 7
0
Ul
X
4
O
r-
-
#1
#2
iiii
(
COLUMN
#3 #4
#5
SAND
EXPAN
SECTIO
#6 #
IM _
r\ci^ ! !
DcD ::
N **:
7 ;;
:::
...
::: #8
INITIAL
FILL
DEPTH
FIGURE 4. COLUMN SETTLED SAND BED DEPTHS
AT THE COMPLETION OF RUN 3
-------
the system effluent possessed heavy slime accumulations. Sieve analyses of
dried and cleaned samples of carryover sand indicated that sand loss was not
specific to any particular particle size. In fact, the size distribution was
almost identical to that of the bulk reactor sand.
EXPERIMENTAL RESULTS - MODIFIED EIGHT-COLUMN SYSTEM
Description of Modifications
At the completion of Run 8 a number of system modifications were made to
improve control of the fluidized beds. The modified system is shown in
Figure 5.
Because the principal objective of this phase of the pilot evaluation
was to test biological viability of the process, and not to optimize
materials handling considerations, a rather simplistic approach to con-
trolling sand bed depletion was adopted. While it was believed that there
was some yet undetermined maximum amount of biological growth that could be
supported per sand particle, it was also felt that there was a corresponding
maximum degree of sand bed expansion possible for given flow conditions.
Consequently, the unfluidized sand bed height was reduced to 1.83 meters
(6 feet) in order to provide an additional 0.305 meter (1 foot) of free -
board height.
It was also felt that some artificial means of stripping excess
biological growth from the sand would be helpful in controlling sand bed
expansion. Unfortunately, the eight column system made any sophisticated
control options too expensive and difficult to construct and operate.
Pilot plant staff reductions and the demands of other research investigations
had also cut system monitoring to one shift per day for only five days per
week since the beginning of Run 7. Sand recirculation was therefore chosen
as the simplest and most expedient method of maintaining media in the system.
Initially, this was attempted by pumping a dilute slurry of settled
sand sludge from the cone of the final clarifier to the base of the first
column. This approach worked well, but it was exceptionally hard on the
various recycle pumps used. Consequently, a batch sand injection system was
adopted. This is diagrammed in Figure 6.
Operationally, carryover or make-up sand was hand charged into the feed
cylinder. Then, the incoming waste flow was diverted through the cylinder,
fluidizing the sand and carrying it into the lead column just above the
gravel support media. When the sand was gone the feed stream was diverted
back into the base of the lead column. This method worked well and was
employed throughout the remainder of the study.
Additional system changes were made at this time and are diagrammed in
Figure 5. One change included the conversion of the skimming tank to a final
clarifier. The purpose was to provide a larger settling area and cor-
respondingly lower surface overflow rates than were possible in earlier
runs. The smaller clarifier previously used as the final clarifier was
converted to an intermediate clarifier for capturing only carryover sand.
24
-------
SYSTEM EFFLUENT
NJ
Cn
i
PRIMARY
EFFLUENT
i
SAND
RECYCLE
CYLINDER
SKIMMING
TANK
\
^
v^^
*
>
-------
PLAN VIEW
DRAIN LINE
SAND RECYCLE
CYLINDER
VIEW SECTION
l'/2 FEED
LINE
NORMAL
FEED FLOW
FRONT VIEW
REMOVABLE CAP
FOR BATCH CHARGING
RECYCLE SANDi
NOTE: lin.=2.54cm.
llb.=0.454kg.
NORMAL
FILL DEPTH
APPROX. SOIbs.
SAND
RECYCLE
CYLINDER-
COLUMN #}
SAND
cci:
SU PORT
MEDIA
I'/V'FEED
LINE
©
FIGURE 6. SAND RECYCI.R SYSTEM
-------
A high-speed laboratory mixer was installed in the clarifier stilling well
to shear attached slimes from the passing sand and mix the tank sufficiently
to prevent deposition of biological solids which were to be removed in the
final clarification step. The modified system also possessed the ability to
recycle treated effluent.
To better observe the nature of biological growth and enable visual
observations of bed expansion at critical locations in the system, the two
clear acrylic columns were moved to the number 3 and number 6 positions in
the eight column sequence. The ineffective expanded section was also
removed.
Examination of the in-line oxygen diffusers during the shut-down period
revealed that the membranes were intact and not clogged. However, it did ap-
pear that they had not been operating efficiently during the previous two
runs because a significant portion of the oxygen gas had been passing around
the circumference of the support disk rather than through the membrane.
To correct the problem a new fitting was devised which incorporated a rubber
seal to force all gas to pass through the diffuser.
Results
The modified eight-column system was put on stream in early October.
Data collection for Run 9 was initiated October 27, under an operating
condition of 18.9 1/min (5 gpm) with no recycle. The resulting reactor rise
rate of 408 1/min/nr (10 gpm/ft^) did not provide adequate bed fluidization.
Numerous hydraulic problems developed, one causing a pressure build-up which
irreparably fractured one of the clear acrylic columns.
Subsequently, the system was operated with only seven reactor columns
and the column rise rate was increased to 571.2 1/min/m2 (14 gpm/ft^) to
provide improved fluidization. A summary of operating conditions for Run 9
is included in Table 5.
Treatment efficiency was quite low, the system achieving only 45.2
percent removal of SCOD, and 51.2 percent removal of TCOD. However, this
was principally due to the low wastewater strength entering the system.
In spite of low percent removal the soluble effluent quality was better
than that in any previous run except Run 7. The relatively high TCOD in
the effluent resulted primarily from the washout of solids during the
numerous hydraulic upsets of this run.
Sand loss rates during this run were the highest yet experienced. An
average of 17.3 kg (38 pounds) of wet sand was lost to the intermediate
clarifier daily. This was twice the rate in Run 8 and almost three times
the rate of Run 7. Much of the hydraulic difficulty in Run 9 was directly
attributable to the high sand loss rate. All sand collected in the inter-
mediate clarifier was weighed daily and charged back into the system.
27
-------
TABLE 5. OPERATIONAL SUMMARY MODIFIED EIGHT-COLUMN SYSTEM
Parameter
Unit
Run 9
Run 10
Run 11
Waste treated
Influent temperature
Influent flow rate (Q)
Recirculation rate (R)
Column rise rate (Q + R/A)
Empty-bed retention time
Final clarifier S.O.R.
Oxygen addition rate
(system average)
Dissolved oxygen concentration
influent
- column No. 1 effluent
- column No. 3 effluent
column No. 7 effluent
Performance Summary
Influent TCOD
SCOD
TSS
Effluent TCOD
SCOD
TSS
TCOD removal /TCODin-TCODout Q(n
TCOD removal ( TCODin xiooj
TOD removal fSCODin-SCODout .,
SCOD removal I SCODin xluu->
°C
1/min*
1/min
1/min/m2
min
1/min/m2
cc/min
gm/m3***
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
1 %
Primary
Effluent
16.4
26.5
0
571.2
44.4
18.0
1140
49.1
2.2
3.2
5.0
4.7
157.9
54.6
78.0
29.9
51.2
45.2
81.1
Primary
Effluent
15
18.9
7.6
571.2
52.0
12.8
1256
75.7
4.3
5.8
9.2
7.2
146.1
52.1
70.7
26.0
5.16
50.0
82.2
Primary
Effluent
13.7
18.9
7.6
571.2
52.0
12.8
1601
96.5
3.5
4.6
4.9
4.0
183.6
82.2
101.1
37.3
44.9
54.6
79.7
*l/min x 0.265 = gpm
**l/min/m2 x 0.0245 gpm/ft2
***g/m3 x 0.0083S = lbs/1000 gal.
28
-------
It is important to note that at no time during this run were the sand
bed levels in any reactor observed to reach any point above either the free
board space sample taps or the column effluent ports. Thus, it was difficult
to explain the high sand loss rates.
On December 4, 1972, Run 10 was initiated at an 18.9 1/min (5 gpm)
influent rate and a 7.6 1/min (2 gpm) effluent recycle rate. This provided
the longest empty-bed retention time yet tested, 52 minutes. As can be
seen in Table 5, there was little significant improvement in treatment
efficiency over that of Run 9, again largely because of low wastewater
strength.
The sand carryover rate was also nearly the same, 17.7 kg (39 pounds)
of wet sand per day. However, careful observation of sand bed behavior led
to a discovery of what was believed to be another cause of sand bed carryover
in the EBBT system. Oxygen gas bubbles rising in the fluidized beds appeared
to be causing the bulk of the carryover, rather than excess growth of
biological mass on the sand particles. As oxygen bubbles would rise through
the fluidized bed sand particles would attach themselves to the bubbles
through a surface phenomenon. The sand-encrusted bubbles, about the size
of peas, would collect at the top of the fluidized bed. When enough of the
sand particles would slough off, the bubbles would rise slowly from the bed
surface and pass into the column effluent line, carrying the remaining sand
with them.
In order to limit sand transport through this "bubble" mechanism an
entrapment device was designed and installed in the top of each column in
the system. The device is shown in Figure 7. It consisted of a short
length of 7.6 cm (3-inch) diameter PVC pipe installed over the exit opening
at the top of each column in the system. The pipe was cut to allow flow
passage from the side of the pipe rather than the bottom, which was out-
fitted with a deflector disk. The function of the ring device was to create
a pocket of oxygen gas at the top of each column without interfering
with the normal flow of water. Rising gas bubbles with attached sand would
collide with the gas-liquid interface at the top of the column, breaking
the bubbles and permitting the sand to fall back into the sand bed.
The ring devices were installed in each of the seven columns in the
last week of December, while the system was in operation. The effect of
these devices was seen rapidly. Within 10 days of continued operation the
daily wet sand loss rate fell from 17.7 kg (39 pounds) to 2.7 kg (5.9 pounds).
Expressed in terms of percent of total system sand lost per day this repre-
sents a decrease from an initial rate of 1.8 percent per day down to 0.34
percent per day.
The changeover to the sand capture devices, completed on December 28,
marked the beginning of Run 11. Operation was continued at the same flow
conditions employed in Run 10 in order to determine the potential increase
in treatment efficiency that could result from improved control of sand
carryover. Treatment performance for the one-month period covered by this
run is shown in Table 5.
29
-------
04
o
COLUMN
EFFLUENT LINE
EI
3"0-PVC
PIPE
GAS PHASE
GAS /LIQUID
INTERFACE
FLOW SLOTS
^
BUBBLE TRAP
FLOW
FREEBOARD
NOTE:lin.=2.54cm.
-3%
COLUMN ASSEMBLY
DETAIL
Of.
GO
CO
FIGURE 7. EBBT BUBBLE TRAP DEVICE
-------
Treatment efficiency actually declined in Run 11. Also, it was not
possible to maintain the low 2.7 kg per day sand loss rate throughout the
run. The final two weeks exhibited a slow increase in daily loss rate up
to a rate of 4.1 kg per day (9 pounds) on January 22. This increase was
accompanied by rising column sand bed heights which reduced the free board
space to between 17.8 and 30.5 cm (7 and 12 inches) in each column. Thus,
while the sand capture devices had significantly reduced the loss rate, they
were only a partial solution. They were not capable of limiting losses due
to excess biological growth on the sand particles. At the same time, the
increased sand control that was effected did not result in a corresponding
improvement in treatment efficiency, at least in the short run.
Based on the results of the eleven runs conducted and a recognition
that only five months remained in the study program, the research direction
was changed at this point. Three runs (45 days maximum duration) were
planned (Runs 12-14). It was felt that reasonable sand bed stability was
possible in the short run. All of the planned runs were to utilize the
44-minute empty-bed detention time. Emphasis was switched from hydraulic
performance to concentrate on gathering data on mixed liquor suspended
solids, waste solids' production, and oxygen utilization efficiency. Plans
also included "spiking" the system feed stream with digester supernatant to
evaluate performance at higher organic loadings.
The results of Runs 12 through 14 are shown in Table 6. It should be
pointed out that the system was reduced to six reactors for these runs as
a result of the structural failure of the one remaining clear acrylic
column at the end of Run 11. Treatment performance, in terms of COD removal,
was universally improved over that of Runs 9, 10 and 11. TCOD removal was
5 to 25 percentage points better while SCOD removal rates were 9 to 15 per-
centage points better. For the most part this appeared to be largely a
result of generally higher influent wastewater strengths and temperatures.
However, the effluent SCOD concentration was seen to deteriorate somewhat
from the values experienced in Runs 9, 10 and 11.
Waste sludge solids production was determined in Run 10 and Run 12
by conducting a mass balance around the system for VSS. For Run 12 an
average of 0.33 kg excess VSS was produced per kg of TCOD removed. The
system (six columns) average reactor VSS was 12,912 mg/1. The calculated
solids retention time (SRT) was 8.06 days. Run 14, which suffered con-
siderably from excessive sand loss problems, exhibited a nearly two-fold
increase in excess solids production, 0.527 kg VSS per kg TCOD removed.
The reactor VSS concentration was 14,863 mg/1 and the SRT was 5.2 days,
significantly lower than that in Run 10.
The reactors were very stable with regard to sand loss rates in Run 12
and Run 13. Overexpansion of the sand beds did not become prevalent until
the end of Run 13. Run 14, however, exhibited very poor stability with
the average daily loss exceeding 4 percent of the system total wet sand.
Table 7 summarizes sand loss parameters for these three runs.
31
-------
TABLE 6. OPERATIONAL SUMMARY - SIX-COLUMN SYSTEM
Parameter
Waste treated
Influent temperature
Influent flow rate (Q)
Recirculation rate (R)
Column rise rate
Empty-bed retention time (Q)
Final clarifier S.O.R.
Oxygen addition rate
(system average)
Dissolved oxygen concentration
Influent
Column No. 1 effluent
Column No. 3 effluent
Column No. 6 effluent
Performance Summary
Influent TCOD
SCOD
TSS
Effluent TCOD
SCOD
TSS
TCOD r moiral JCODin-TCODout ,._.
TCOD removal C TCOpin xloo->
-COD r maval fSCODin-SCODout QQ,
SCOD removal ( SCODin X10(JJ
n n Amru.ii fTCODin-SCODout 1001
T« -, i rTSSin-TSSoutxiom
Unit
°C
1/min"
1/min
1/min/m2
min
1/min/m2
cc/min
gm/m3***
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
%
%
%
%
Primary
Effluent
12.1
18.9
0
408
44.1
12.8
1744
105.1
6.0
5.6
5.4
6.8
195.5
83.1
86.5
85.1
36.7
32.4
56.5
55.8
81.2
67.5
Primary
Effluent
12.5
18.9
0
408
44.1
12.8
2408
145.3
4.3
3.1
6.8
10.4
201.6
131.6
75.1
53.4
62.7
59.4
73.5
.
Primary
Effluent
14.5
18.9
0
408
44.1
12.8
3138
189.2
3.4
4.1
11.1
13.0
170.7
114.7
56.0
34.8
67.2
69.7
79.6
.
TSSin
*l/min x 0.265 = gpm
**l/min/m2 x 0.0245 = gpm/'ft2
***gm/m3 x 0.00835 = lbs/1000 gal.
32
-------
TABLE 7. SUMMARY OF EBBT SAND LOSS FOR RUNS 12, 13, and 14
Column Rise Rate Wet Sand Loss(kg/day) Avg. % of Total
^"^ fi / ' / 2~\
1 Average Range System Sand/Day
12
13
14
408
408
408
4.1
7.7
27.5
0.45-20.4
2.3-22.7
12.1-49.9
0.6
1.13
4.03
1/min/m2 x 0.245 = gpm/ft2
kg/day x 2.20 = Ib/day
The high loss rates of Run 14 were believed to have resulted from the
"spiking" of the system influent with digester supernatant in Run 13. A
short time after supernatant addition was started a significant shift
occurred in the microbiological population on the sand in the reactors.
Spaerotilus and Thiothrix began to predominate. A phenomenon similar to
"bulking" in conventional activated sludge plants resulted in a rapid
increase in the degree of sand bed expansion and daily loss rates. Since
this had not happened in previous testing it was felt that some component
of the digester supernatant had encouraged the proliferation of the fila-
mentous growths.
As a consequence of the uncontrollable sand losses in Run 14, data
collection was stopped. Digester supernatant addition was also terminated.
The six-column system was operated on a tap water feed for two weeks to
reduce the filamentous population. During this time a number of modifi-
cations were made to prepare for the final test period, Run 15, which was
to include further monitoring of excess solids production as well as a
determination of oxygen utilization efficiency.
In order to more accurately monitor dissolved oxygen levels in the
system a Weston and Stack model 3000 Dissolved Oxygen Analyzer and a model
60 high pressure/high dissolved oxygen concentration probe were added.
A limitation of past dissolved oxygen determinations was that samples taken
from the pressurized system had to be analyzed at atmospheric pressure with
a bench D.O. meter. The new system was fed by a sampling manifold connected
into the freeboard space of each of the six columns. By opening the
appropriate valves a sample could be withdrawn from each column and passed
directly into the D.O. probe at the same pressure that existed in the
column. Dissolved oxygen levels measured in this way were on the average
1.0 mg/1 higher than identical determinations made with a bench meter at
atmospheric pressure.
33
-------
An off-gas collection system was also added. The stilling well of the
intermediate clarifier was capped and equipped with a gas collection line.
The volume of off-gases was measured with a wet test meter. The exhaust
from the meter was collected in 0.0283nr (1 ft )mylar gas collection bags
and regularly analyzed for 02, N2, and C02 gas concentration.
Wastewater feeding was instituted on May 31. The start-up period
suffered from numerous plugging problems which were a result of five separate
City of Lebanon power failures. Stability was finally achieved by mid-July
and preliminary sampling was started July 17. The final run, Run 15, was
instituted August 13 when the biological performance reached equilibrium.
A summary of performance from this point to the final day of operation,
September 10, appears in Table 8.
This was one of the most stable periods of the entire experimental
program. TCOD removal efficiency was 75 percent and SCOD removal efficiency
averaged 55.9 percent. The effluent TCOD of 48.8 mg/1 was also the best
obtained by the EBBT system. The effluent TSS concentration was 11 mg/1.
Sand bed performance was exceptionally stable throughout the entire test
period. The average daily loss to the intermediate clarifier was 2.95 kg/day
(6.5 Ibs/day) . The unit was free of plugging and associated hydraulic
problems.
During the period covered by Table 8 data, the average effective MLVSS
was 14,276 mg/1. Solids production data indicated that 0.26 kg excess
volatile solids was formed per day per kg of TCOD removed. Based on the
average MLVSS maintained this was equivalent to 0.115 kg excess VSS formed/
day/kg MLVSS, and 0.422 kg TCOD removed/day/kg MLVSS. The SRT was 8.7 days.
At the completion of Run 15 the EBBT study was terminated. The sig-
nificance of the operating results obtained from the 3 year study and
conclusions regarding the feasibility of the EBBT process are discussed in the
following section.
34
-------
TABLE 8. OPERATIONAL SUMMARY SIX-COLUMN EBBT SYSTEM
Parameter
Waste treated
Influent temperature
Influent flow rate (Q)
Recycle rate (R)
Column rise rate (Q + R/A)
Ei.tJty bed retention time
'final clarifier S.O.R.
Oxygen addition rate
(system average)
Dissolved oxygen concentration
influent
- column No. 1 effluent
- column No. 3 effluent
- column No. 6 effluent
Performance Summary
Influent TCOD
SCOD
TSS
Effluent TCOD
SCOD
TSS
TCOD runciral /TCODin-TCODout .
i\.uu removal i TCODin xiuu;
"COD removal fSCODin-SCODout .
SCOD removal (. SCODin xlOOJ
Overall removal fTCODin-SCODout
BBS ^v*! rTSSin-TSSoutnnm
Unit
°C
1/min*
1/min
1/min/m2**
rain
1/min/m2
cc/min
gm/m3***
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
%
%
%
%
Run 15
Primary
Effluent
22.1
18.9
0
408
44.1
453
3438
147.3
1.7
5.4
13.4
11.9
196.3
73.7
85.2
48.8
32.4
11.0
75
55.9
83.5
a 7
*l/min x 0.265 = gpm
**l/min/m2 x 0.0245 = gpm/ft2
***gm/m3 x 0.00835 « lb/1000 gal.
35
-------
SECTION 6
DISCUSSION
TREATMENT PERFORMANCE
Carbonaceous Treatment
A principal objective of the EBBT pilot-scale investigations was to
determine the system operating conditions which would maximize biological
treatment efficiency. The primary system control parameter in this regard
was the hydraulic retention time. To a lesser extent reactor dissolved
oxygen levels could be varied to alter system conditions. The solids
retention time was not purposely controlled at a set level. Rather it
was derived from biological solids wasting through two mechanisms: 1)
natural sloughing of biological slimes from sand particles in the reactors;
and, 2) mechanical shearing of excess growth from carryover sand either by
force of a sand recycle pump or by a high-speed mixer.
On the basis of average treatment performance data from each of the
experimental runs an effort was made to determine the relationship between
COD removal efficiency and empty-bed retention time. Little correlation was
found between TCOD removal efficiency
(TCODin - TCODout
TCODin
and empty-bed retention time because of the random nature of TCOD removal
data. To a great extent this was caused by the considerable range of
surface overflow conditions under which the final clarifiers were operated.
The numerous system modifications also introduced variability to clarifier
performance. Most importantly, TCOD removal efficiency appeared to be
closely tied to the degree of hydraulic stability in the sand beds. Runs
with high sand loss rates consistently produced final effluents high in
TCOD because of the increased sloughing of surface slimes and the resulting
increase in solids applied to the clarifier.
A relationship was determined for SCOD removal efficiency as a
function of empty-bed retention time. This is illustrated in Figure 8.
Performance was found to fall into two separate regimes based upon
influent wastewater temperatures. As would be expected greater removal
efficiencies were possible at the higher temperatures (>18 C). A maximum
removal efficiency of 71 percent was achieved at the higher temperatures at an
empty-bed retention of time of 44 minutes. In the lower operating-temperatures
36
-------
80
>-
18°C)
OVERALL COD REMOVAL
10
30
50
EMPTY BED RETENTION TIME (MINUTES)
FIGURE 8. COD REMOVAL EFFICIENCY AS A
FUNCTION OF EMPTY BED RETENTION TIME
37
-------
range, soluble COD removal did not exceed 55 percent, even at retention
times in excess of 50 minutes. This resulted from the combined effects
of lower reaction rate and the generally weaker influent waste strength
during the colder months of operation.
Also shown in Figure 8 is the relationship between overall COD removal
(TCODin - SCODout
TCODin x iUUJ
and empty-bed retention time. The curve indicates that a maximum overall
COD removal efficiency of 85 percent was approached at an empty-bed
detention time of 52 minutes.
EBBT treatment performance was also examined in terms of the volumetric
organic loading rate applied to the system, expressed as kg of COD applied
per day per m of reactor volume. The rates during pilot testing ranged
from 4.99 kg COD/day/m3_to 35.2 kg COD/day/m3 (312 Ib COD/day/1000 ft* to
2200 kg COD/day/1000 f t ) . In all cases the empty-bed reactor volume was
employed in the calculations.
The relationship between COD removal efficiency and the COD loading rate
which was determined for the EBBT system is shown in Figures 9 and 10 where
TCOD removal (%) and SCOD removal (%) are respectively plotted against their
corresponding volumetric loading rates. In both instances the effect of low
influent wastewater temperature and strength is apparent. The data appeared
again to separate into two groups, the upper curve in each figure representing
data collected at wastewater temperatures above 18 C and the lower curve that
collected at temperatures below 15°C. However, when absolute effluent
quality (based on total and soluble COD) was examined as a function of the
volumetric loading rate, no temperature differentiation was possible. This
may be seen from Figures 11 and 12, where effluent TCOD and SCOD are plotted
respectively against the TCOD and SCOD volumetric loading rate. It appears
that lower influent waste strengths in the colder months had a greater effect
upon COD removal efficiency than the lower wastewater temperatures. In
more typical wastewaters possessing higher waste strengths, however, it is
believed that the temperature effects would predominate. At Lebanon,
treatment efficiency calculations were sometimes limited mathematically by
the strength of the wastewater treated.
The curve in Figure 11 suggests that the relationship between effluent
TCOD and the TCOD volumetric loading rate is linear up to a rate of approxi-
mately 8 kg TCOD/day/m (500 Ib TCOD/day/100 ft ), after which an apparent
first-order relationship prevails. At this high (in comparison with
conventional activated sludge systems) rate the system produced an average
effluent quality of 96 mg/1 TCOD, 41 mg/1 SCOD, and a calculated average
TBOD- of 27.6 mg/1 (based on an effluent TCOD:TBOD5 ratio of 3.48).
One of the principle objectives of the pilot-scale studies at the
Lebanon Pilot Plant was to determine the optimum design operation
conditions for the EBBT process. Based upon the biological treatment
performance data collected the pilot scale system operated most consistently
38
-------
>- 80
U
u
LL.
LL.
60
S 40
o
o
u
< 20
O
LEGEND
TCOD REMOVAL EFFICIENCY (T<15°C)
O TCOD REMOVAL EFFICIENCY (T>18°C)
NOTE:1.0lb./1000ft.3/day=0.016kg./m3/day
400 1200 2000
TOTAL COD VOLUMETRIC LOADING RATE
(lb./1000ft?-DAY)
FIGURE 9. TCOD REMOVAL EFFICIENCY AS A FUNCTION OF TCOD
VOLUMETRIC LOADING RATE
39
-------
>- 80
U
60
40
Q
O
u
20
O
to
LEGEND
SCOD REMOVAL EFFIENCY (T< 15 °C)
O SCOD REMOVAL EFFIENCY (T>18°C)
NOTE:1.0lb./1000ft.3/day=0.016kg./m3/day
200
600
1000
SOLUBLE COD VOLUMETRIC LOADING RATE
(lb./1000ft?-DAY)
FIGURE 10. SCOD REMOVAL EFFICIENCY AS A FUNCTION OF SCOD
VOLUMETRIC LOADING RATE
40
-------
160
0)120
Q
O
u
80
40
NOTE:1.0lb./1000ft.3/day=0.016kg./m3/day
800 1600 2400
TOTAL COD VOLUMETRIC LOADING RATE
(lb./1000ft?-DAY)
FIGURE 11. EFFLUENT TCOD AS A FUNCTION OF TCOD
VOLUMETRIC LOADING RATE
41
-------
0>
Q
O
60
40
O
ts>
2 20
NOTE:1.0lb./1000ft.3/day=0.016kg./m3/day
200 600 1000
SOLUBLE COD VOLUMETRIC LOADING RATE
(lb./1000ft?-DAY)
FIGURE 12. EFFLUENT SCOD AS A FUNCTION OF SCOD
VOLUMETRIC LOADING RATE
42
-------
under the following conditions:
empty-bed retention time: 44 minutes (Q)
2
reactor rise rate: 408 1/min/m
TCOD volumetric loading rate: 6.4 kg/day/m
Under these conditions an average effluent quality of 80 mg/1 TCOD and
31 mg/1 SCOD was obtained. This corresponded to a calculated average
effluent TBOD5 of 23 mg/1. fe
It is not possible, however, to suggest that these necessarily represent
optimum conditions, due to the physical limitations of the pilot system
itself. Many of the complications encountered with the operation of the
multi-column system have been discussed, including sand bed stability,
dissolved oxygen control, and plugging. Each of these contributed to the
problem of maintaining or sometimes even achieving stable operating
conditions. In all cases, however, data from period of extensive system
upset were not used in the determination of average system performance.
In summary, the biological treatment performance of the EBBT pilot
system indicated that under stable operating conditions a high-quality
effluent could be produced at significantly higher waste loading rates
and shorter hydraulic retention time than the conventional activated sludge
systems. These results are encouraging, in light of the physical limitations
of the pilot test system, and indicate that the process could be feasible
for eventual full-scale application. However, optimum operating conditions
cannot be defined until the process is evaluated on stronger wastewater and
with improved methods of dissolved oxygen supply and sand bed control.
Nitrification
Ammonia and nitrate data were collected at a number of points during
the EBBT test program. The degree of ammonia oxidation was slight to
moderate and no consistent pattern of nitrification was experienced upon
which quantifiable conclusions could be drawn. It is believed that
transport of sand through and out of the pilot system prevented the estab-
lishment of a stable population of nitrifiers.
SOLIDS PRODUCTION
One possible benefit from the use of commercial oxygen in biological
treatment processes is a decrease in sludge production (yield) as compared
with conventional air activated biological processes. One hope in the EBBT
investigations was that even lower yields could be obtained by employing
commercial oxygen in a fixed film system.
Much attention has been devoted to determining the effect of dissolved
oxygen concentration on microbial yield. There is considerable disagreement
among researchers as to what the true effect is. Some have indicated that
there is a significantly decreased yield with oxygen aeration when compared
43
-------
to air aeration under identical conditions CIS,.21, 24), while others have
concluded that there is either no effect or that the reported differences
are caused by modified SRT values rather than direct oxygen effects (22, 23).
In one study comparing bench-scale air and oxygen activated sludge
systems, the high dissolved oxygen concentrations of the oxygen reactor
resulted in a 20 percent reduction in yield when compared to the air
reactors (24). Suspended growth and attached film biological reactors
were also compared. The yields in both the air and oxygen attached film
systems were lower than those obtained with the corresponding suspended
growth systems (11 percent for air and 24 percent for oxygen).
For the EBBT system, sludge production, or yield, was determined for
Runs 12, 14, and 15. Yield was determined by measuring the amount of sludge
wasted plus the amount of solids lost to the effluent over a given time
period and relating this to a specific organic loading condition. Because
in EBBT there was no sludge recycle, all sludge collected in the final
clarifier was waste sludge. The results of these runs are summarized in
Table 9.
TABLE 9. SUMMARY OF EBBT SOLIDS PRODUCTION DATA
Parameter
kg TCOD removed/day
kg excess VSS formed/ day
kg MLVSS (system average)
ke excess VSS formed/ day
Run 12
2.90
0.95
7.67
0.33
Run 14
3.30
1.69
8.83
0.53
Run 15
3.58
0.98
8.49
0.26
kg TCOD removed/day
kg excess VSS formed/day
kg MLVSS
SRT (days)
MLVSS (jng/-l) (system average)
Average wet sand loss (kg/day)
Dissolved Oxygen (mg/1)
Influent
Column #1 effluent
Column #3 effluent
Column #6 effluent
0.38
0.38
6.0
5.6
5.4
6.8
3.5
4.1
11.1
13.0
0.42
8.06 5.20 8.70
12,912 14,863 14,276
4.1 27.5 2.95
1.7
5.4
13.4
11.9
44
-------
The average rate of TCOD removal in terms of kg per day was seen to
increase with successive runs. This primarily resulted from the increase
in influent waste strength from spring (Run 12) to late summer (Run 15).
The average daily excess volatile solids formation was nearly identical in
Run 12 and Run 15, but Run 14 was significantly higher. As a consequence
the yield value for Run 14 was also higher than that for the other runs.
The yield was 0.55 kg VSS formed/day/m3 versus values of 0.26 and 0.33 kg
VSS formed/day/m experienced in Run 12 and Run 15.
The principal reason for the high yield in Run 14 was the high sand
loss rate (27.5 kg/day) caused by the proliferation of filamentous micro-
organisms. This resulted in the transport of much greater amounts of sand
surface-bound slimes from the reactors. These were sheared from the sand in
the intermediate clarifier and settled as waste sludge in the final clarifier.
The average system MLVSS decreased from 16,000 mg/1 at the start of the run
to 13,500 mg/1 at the end. Thus, the apparent high solids production of
this run was principally a result of unstable sand bed conditions. MLVSS
values were reasonably constant in the other two runs.
The yield values for the EBBT pilot system were essentially comparable
to those reported for suspended growth high-purity oxygen systems. The
significant reductions in waste sludge production rates postulated in the
beginning phases of the program were not experienced.
It is also important to note at this point that a considerable amount
of effort was involved in actually determining average MLVSS for the entire
system. First of all, a determination had to be made for each of the
reactors because the amount of biological growth per sand particle varied
with location in the system. Also, there was only one sample point in each
column from which a grab sample of the fluidized bed could be taken. The
was located at the approximate midpoint of the height of the bed when
expanded at a rise rate of 10 gpm/ft . This was assumed to be representative
of conditions in the entire bed. An obvious difficulty here is apparent
when one considers the known variation in the height of sand in each column.
One day the sample point might be at the midpoint of the bed, while later
in the week it would be perhaps at the one-third or two-thirds point. The
actual determination of the volatile solids concentration on the bed sample
was also complicated. Typically, a 2000 ml sample was drawn from each
column. In exiting the sample tap some sludge solids" were stripped from the
sand particles. Thus, VSS had to be determined for both the free solids and
the attached solids. A total mass of volatile solids was determined for
each sample and then the total divided into the total volume of the
initial sample to determine the effective concentration of MLVSS. Ideally,
the best practice would have provided for a series of sample points in each
column so that a sample composited with height could be obtained. In
general, however, the values for MLVSS obtained with the mid-column sampling
method proved consistent. They are felt to be representative of the actual
volatile solids content in the system. The value of MLVSS used in computing
excess sludge production was the average of all total system values.
Some concept of the distribution of MLVSS in the EBBT process in
terms of time and system location can be gained from an examination of
45
-------
Table 10. This displays data employed in the determination of system MLVSS
for Run 15. Throughout the run the mixed liquor volatile solids concentration
was consistently lowest in the initial column stages and highest in the final
stages. To some extent this pattern resulted from the distribution of sand
in the system. Transport of sand between columns generally caused sand to
collect in the final columns in the system flow sequence. Examination of
SCOD removal with time, however, revealed that the majority of treatment was
still occurring in the first two columns. On the average, over 70 percent
of the SCOD removal was achieved in the first two columns.
OXYGEN UTILIZATION
A limitation of the EBBT experimental system was the inability to
precisely control dissolved oxygen levels in the reactors. Oxygen gas was
fed to each reactor through a separate rotameter. The oxygen gas was
supplied by several 113.5 kg (250 pound) liquid oxygen cylinders connected
in parallel. The gas supply was regulated to 195.3 kg/m2 (40 psig) prior to
entering the rotameters.
The .dissolved oxygen control strategy was essentially to maintain the
average system dissolved oxygen concentration at high levels. The hope here
was to encourage oxygen penetration into the attached slime layers, and to
minimize excess sludge production. The maximum concentration employed,
however, was generally limited to below 10 mg/1, because the higher oxygen
addition rates resulted in exceptional bubble transport through the reactors,
causing buildup in the reactors.
While it was generally possible to maintain adequate dissolved oxygen
levels in the system reactors, it was difficult to maintain consistent levels.
Reactor dissolved oxygen concentrations varied widely and unpredictably from
day to day and sometimes from morning to afternoon. The fluctuations were a
result of changing system variables including: wastewater temperature and
organic strength, influent dissolved oxygen concentrations, column gauge
pressure, the rate of sand transport between columns, and the concentration
of biomass in the reactors. Dissolved oxygen determinations were made two
times in every 24-hour period during the week and not at all on weekends.
In general, the oxygen addition rate for the system was primarily a
function of the oxygen requirements in column #1 and column #2 where the
greatest percent of organic removal was effected. Under many circumstances,
oxygen was added only to these two columns and not to subsequent columns
in the sequence. Oxygen supplied in a sufficient amount to maintain a
dissolved oxygen concentration of 2 to 4 mg/1 in the first two columns
generally resulted in a dissolved oxygen concentration above 4 mg/1 in the
remainder of the system due to adsorption of oxygen bubbles not utilized
in the first stages. When oxygen gas was added in small amounts (1/10 the
rate applied to columns #1 and #2) to subsequent columns downstream of
column #2, as in Runs 13, 14 and 15, the average dissolved oxygen concen-
tration increased dramatically in the final half of the system. This
phenomenon may be observed in Figure 13 which plots average dissolved oxygen
profiles for Runs 7 through 15.
46
-------
TABLE 10. EBBT MIXED LIQUOR VOLATILE SOLIDS LEVELS - RUN 15
Mixed Liquor Volatile Solids (mg/1)
Date
8/14
8/16
8/21
8/28
8/30
9/4
9/6
Avg All Dates
Col. #1
6,228
6,564
8,457
6,885
5,419
6,219
5,955
6,532
Col. #2
360
3,949
10,758
12,386
17,618
11,090
14,045
9,965
Col. #3
20,905
18,478
12,464
13,434
12,984
14,058
13,420
15,106
Col. #4
12,991
13,456
11,214
18,352
12,153
14,513
14,749
13,918
Col. #5
22,145
23,049
28,615
29,870
16,488
18,977
19,493
22,662
Col. #6
14,826
18,965
18,202
29,487
16,262
19,990
18,967
19,527
System Avg
10,441
14,076
14,951
18,402
13,487
14,141
14,438
14,276
-------
00
COLUMN LOCATION
FIGURE 13. SYSTEM AVERAGE DISSOLVED OXYGEN
PROFILES FOR RUN 7 THROUGH 15
-------
Oxygen addition rates for all experimental runs are summarized in
Table 11. Rates employed ranged from 0.047 to 0.270 kg 0 added/m (0.388
to 2.246 pounds of 02 added/1000 gallons) of sewage treated. These represent
amounts of oxygen input to the system. This corresponded to an operating
range of 0.629 to 1.230 kg 02 added/kg TCOD removed.
System oxygen utilization efficiency was determined for Run 15. Oxygen
input was monitored by measuring weight changes in the feed liquid oxygen
cylinders. System off-gases were collected by capping the stilling well
in the intermediate clarifier. Gas volume was determined by a wet test
meter. A solenoid valve on the meter exhaust would periodically divert a
small amount of sample to a mylar collection bag. Samples were analyzed
for C02, N2, and 02 by a local lab within 12 hours of collection. In
addition, liquid sampling was expanded to permit a TCOD balance around
the system to enable validation of oxygen utilization efficiency on the
basis of TCOD destruction.
The test was conducted over a period of five days. Within the first
two days, problems developed with the wet test meter. The rate of gas
evolution was 2-3 times that anticipated. Repeated tests and recalibration
of the instrument failed to locate the difficulty. Consequently, oxygen off-
gas had to be determined by difference using the TCOD data. Oxygen utili-
zation efficiency was determined in the following manner.
1) Basic Data
Influent 0_ (Ibs) = 0.48 Ibs
Effluent 02 (Ibs) = 2.26 Ibs
from liquid clarifier
Cylinder 0 applied (Ibs) = 67.0 Ibs
Influent TCOD (Ibs) = 57.57 Ibs
Effluent TCOD (Ibs) = 11.50 Ibs
Waste sludge TCOD (Ibs) = 20.05 Ibs
2) Calculation
- Ibs influent TCOD 57.57
- Ibs effluent TCOD -11.50
- Ibs waste sludge TCOD -20.05
TCOD destroyed =26.02 Ibs
Ibs TCOD destroyed 26.02 Ibs
Nitrate Nitrogen 02 demand
Ibs effluent dissolved 0 2.26 Ibs
Ibs 02 supplied 67.48 Ibs
= Ibs exhaust 02 39.20 Ibs
From this information, the efficiency of oxygen utilization can be
seen to be only 38.6 percent. Existing commercially marketed oxygen treatment
49
-------
TABLE 11. SUMMARY OF OXYGEN ADDITION DATA FOR THE EBBT SYSTEM
in
O
Run
3
4
5
6
7
8
9
10
11
12
13
14
15
System Average Flow
Oxygen Addition (Q)
Rate (cc/min) (1/min)*
1233
750
826
1030
4508
6245
1140
1256
1601
1744
2408
3138
3488
31.2
15.1
7.6
15.1
29.9
26.5
26.5
18.9
18.9
18.9
18.9
18.9
18.9
kg 02 added**
m3
0.047
0.057
0.125
0.078
0.172
0.270
0.049
0.076
0.097
0.105
0.146
0.190
0.210
kg 02 Added
kg TCOD Removed
0.750
1.110
0.661
0.832
-
-
0.666
0.629
0.688
0.846
1.055
0.957
1.460
kg 0^ Added
kg SCOD Removed
0.643
1.080
2.970
2.270
2.410
3.260
2.900
2.900
2.160
2.270
1.860
2.370
5.100
* 1/min
** kg 02
x 0.265 = gpm
added x 8.53 =
Ibs Q? added
m-
1000 gallons
-------
systems have been able to produce better than 90 percent oxygen utilization
by employing off-gas recirculation in staged systems. Clearly, a greater
oxygen efficiency would need to be achieved in a full-scale system in order
for the process to be competitive.
The poor oxygen utilization efficiency was primarily due to the limita-
tions of the oxygen dissolution method employed in the pilot system. In
future testing a more effective oxygen diffuser should be employed. The
analyses of the system off-gases in Run 15 indicated an average gas compo-
sition of 70.3 percent oxygen, 21.5 percent nitrogen, and 9.3 percent
carbon dioxide. The high oxygen content suggests the possibility of also
employing off-gas recycle to the beginning of the system or even around
each stage.
The most promising approach to improved oxygen utilization efficiency
would be dissolution of oxygen gas into the wastewater prior to entry into
the column. This would permit large amounts of oxygen to be dissolved
without the problem of sand bed instability created by oxygen bubbles in
the fluidized bed environment.
EXPANDED BED STABILITY
Through out the EBBT experimental program, considerations of expanded
bed stability were involved in nearly all decisions regarding system design,
operation, and modification. Treatment performance was directly related to
the degree of expanded bed control achieved. The system was considered to
be stable when the depth of bed expansion remained relatively constant, and
when the rate of daily sand loss was low. It was generally found that
reasonably consistent operation was possible at daily loss rates less than
0.4 percent of the total system sand weight expressed in terms of wet sand
(20 percent moisture content).
Unacceptable sand loss rates, expecially those greater than 1.0 percent
of total system sand/day, were commonly accompanied by severe hydraulic
problems and reduced treatment efficiency. Sand was easily transported
between columns, but excessive rates of sand movement caused plugging in the
connecting piping and flow distribution components. High reactor pressures
would develop in these instances, necessitating system shutdown to prevent
structural failure of the reactors. Inability to control sand losses
resulted in at least 10 column failures and numerous system shutdowns during
the three years of experimentation.
In addition to hydraulic difficulties, the loss of media from unstable
reactors also reduced treatment efficiency through the attrition of active
biological slimes from the system.
Sand losses for each of the thirteen experimental runs are summarized
in Table 12. Also shown are the corresponding average values for column
rise rates and oxygen addition rate.
Based on total data and observations made during each experimental run,
a number of judgements are possible with regard to EBBT bed stability. First
51
-------
TABLE 12. SUMMARY OF WET SAND LOSS FROM THE EBBT SYSTEM
Run
3
4
5
6
7
8
9
10
11
12
13
14
15
Column Rise
Rate 2
1/min/m
816.0
652.8
163.2
326.4
664.6
571.2
571.2
571.2
571.2
408.0
408.0
408.0
408.0
Average Oxygen
Addition Rate
(S.T.P.)
cc/min kg/m
1233 0.047
750
826
1030
4508
6245
1140
1256
1601
1744
2408
3138
3438
0.057
0.125
0.078
0.172
0.270
0.049
0.076
0.097
0.105
0.146
0.190
0.210
kg/ day
1.86
0.86
0.39
1.76
4.82
4.15
17.3
17.7
3.4
4.1
7.7
27.5
3.0
Average Wet
Sand Loss
% of Bed Weight
0.63
0.29
0.13
0.60
0.50
0.44
1.7
1.8
0.36
0.60
1.13
4.03
0.31
of all, the rate of sand loss was more sensitive to the amount of attached
biological slimes and the volume of oxygen gas applied to the system than
it was to changes in column rise rate. The principal importanc^ of the
column rise rate over the 163.2 to 816 1/min/m2 (4 to 20 gpm/ft ) range
examined was the determination of the degree of bed expansion.
Rise rate was related to the loss rate only in the sense that it
aggravated" the effects of excess biological growth on the sand particles
and the "bubble effect"- High rise rates were not the causative factor in
sand loss, as long as sufficient free board was available.
In the case of EBBT, the bubble ring devices placed at the reactor
effluent proved effective protection against media losses. Treatment
efficiency and allowable reactor biological mass concentration were therefore
not limited by oxygen dissolution considerations.
The question of what constitutes "excess" biological growth on the
media particles is a more fundamental one. The EBBT experimentation was
not extensive enough to provide the data necessary to establish an optimum
ratio of bacterial mass to media mass. Throughout the research the principal
52
-------
goal with regard to effective mixed liquor biological mass was to maintain
the highest concentration possible, allowing natural sloughing to determine
the limit of permissible growth. Based on the hydraulic problems which were
encountered throughout the study as a consequence of this approach it is
apparent that maximum mixed liquor biological growth is not necessarily
consistent with maximum treatment efficiency.
Clearly, some positive means of controlling the amount of growth per
sand particle is necessary. Growth could be controlled through an external
means of forced sloughing, such as a small "blow-down" tank into which a
portion of the expanded bed would be discharged periodically from each
reactor. If the shearing forces exerted during the transport were insuf-
ficient a mixer could be used to assist in dislodging the necessary amount
of bacterial growth. After a very short settling period the sludge mass
could be decanted to waste and the sand media returned to the reactor.
Rather than controlling sand loss by limiting the amount of bacterial
growth per media particle it may be possible to solve the problem merely
by using a denser media material such as garnet. In the same range of rise
rates employed with sand media, the denser particles would have a decreased
rate of expansion (i.e., greater resulting freeboard depth) and possibly
a reduced sensitivity to varying degrees of biological growth on the
particles. These attributes might permit either the use of higher reactor
rise rates, increased growth per particle, or the use of more media (number
of particles) per reactor than was possible with sand of a similar size
distribution. Any of these possible benefits would have to be weighed
against both the increased cost of the media and the greater resulting head
losses through the fluidized bed.
53
-------
REFERENCES
1. Bach, H., "The Tank Filter", Water Works and Sewerage, October 1937.
2. Ponninger, R., Supplement to Gesundheitsing, Series II, No. 18, Munich,
1938.
3. Hayes, Clyde C., Sewage Treating Process, U.S. Patent No. 1,991,896,
October 15, 1931.
4. English, John N., "Oxygen Addition Study", F.W.P.C.A. internal summary
report, Pomona Pilot Plant, Pomona, California, March 26, 1969.
5. Tamblyn, Thomas A. and Brain A. Sword, "The Anaerobic Filter for the
Denitrification of Agricultural Subsurface Drainage" Presented at the
24th Purdue Industrial Waste Conference, Purdue University, Lafayette,
Indiana, May 7, 1969.
6. McCarty, Perry L., Louis Beck, and Percy St. Amant, "Biological
Denitrification of Wastewaters by Addition of Organic Materials."
Presented at the 24th Purdue Industrial Waste Conference, Purdue
University, Lafayette, Indiana, May 7, 1969.
7. Smith, John M. Arthur N. Masse, Walter A. Feige, and Lawrence J. Kamphake,
"Nitrogen Removal from Municipal Wastewater by Columnar Dentirification",
Environmental Science and Technology, Vol. 6, No. 3, March 1972, pg. 260.
8. Johnson, Robert L., and E. Robert Bauman, "PAB Process for Advanced
Waste Treatment" Completion Report 1SU-ERI-AMES-9100, Iowa State Water
Research Institute, December 1970.
9. Biospherics Research Inc., "Use of Coal to Enhance Metabolic Treatment
of Sewage" Final Report project PB-184-445, Department of the Interior,
May 1969.
10. Hopkins, Charles R., Walter J. Weber, Jr., and Ralph Bloom, Jr., "A
Comparison of Expanded-Bed and Packed-Bed Adsorption Systems", Report
No. TWRC-2 F.W.P.C.A., Cincinnati, Ohio, December 1968.
11. Monthly Report, EPA Blue Plains Pilot Plant, Washington, D.C.,
October 1969.
12. Pirnie, M., Presentation at Twenty-First Annual Meeting of Federation
of Sewage Works Associations, Detroit, Michigan, October 18-21, 1948.
54
-------
13. Okun, D.A., "System of Bio-Precipitation of Organic Matter from Sewage",
Sewage Works Journal, Vol. 21, pg. 763, 1949.
14. Budd, W.E., and G.F. Lambeth, "High Purity Oxygen in Biological Sewage
Treatment", Sewage and Industrial Wastes, Vol. 29, pg. 253, 1957.
15. Union Carbide Corporation, "Investigation of the Use of High Purity
Oxygen Aeration in the Conventional Activated Sludge Process", Final
Report, F.W.Q.A. Contract 14-12-465, May 1970.
16. EPA Demonstration Grant S802356, City of Las Virgenes, California.
17. Jeris, J.S., R.W. Owens, R. Hickey, and F. Flood, "Biological Fluidized
Bed Treatment - BOD and Nitrogen Removal in Less than One Hour
Biological Treatment Time", Presented at the 48th Annual Conference of
the Water Pollution Control Federation, Miami, Florida, October 1975.
18. Scott, C.D. and C.W. Haucher, "Use of a Tapered Fluidized Bed as a
Continuous Bioreactor," Biotechnology and Bioengineering, Volume XVIII,
pp. 1393-1403 (1976).
19. Eckhoff, D.W., and D. Jenkins, "Transient Loading Effect in the
Activated Sludge Process". In "Advances in Water Pollution Research"
Proc. 3rd International Conference on Water Pollution Research,
Munich, 1967.
20. Jenkins, D. and A.B. Menar, "The Fate of Phosphorus in Sewage Treatment
Processes. I Primary Sedimentation and Activated Sludge", Sanitary
Engineering Research Laboratory Report, University of California,
Berkeley, 1967.
21. Stamberg, J.B., D.F. Bishop, A.B. Hais and S.M. Bennett, "System
Alternatives in Oxygen Activated Sludge", presented at the 45th
Annual Water Pollution Control Federation Conference, October 1972.
22. Bull, J.E. and M.J. Humenick, "High-Purity Oxygen in Biological Treatment
of Municipal Wastewater", Journal Water Pollution Control Federation,
Vol. 44, No. 1, 1972, pg. 65.
23. Sherrard, J. H. and E.D. Schroeder, "Variation of Cell Yield and
Growth Rate in the Completely Mixed Activated Sludge Process",
Proceedings of 27th Industrial Waste Conference, Purdue University,
May 1972.
24. Jewell, W. J., and S.E. MacKenzie, "Microbial Yield Dependence on
Dissolved Oxygen in Suspended and Attached Systems". Presented 6th
Water Resources Symposium, University of Texas, November 13, 1972.
55
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/2-78-117
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
EXPANDED BED BIOLOGICAL TREATMENT
5. REPORT DATE
July 1978 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
E. Timothy Oppelt, John M. Smith, Walter A. Feige
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Municipal Environmental Research Laboratorycin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
10. PROGRAM ELEMENT NO.
1BC611
11. CONTRACT/GRANT NO.
In-house
12. SPONSORING AGENCY NAME AND ADDRESS
Same as above
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
15 SUPPLEMENTARY NOTES
Project Officer: E. Timothy Oppelt 513/684-7606
is ABSTRACT
performance of a three-phase fluidized bed biological contacting system
for secondary treatment of settled wastewater was investigated in a 37.9 1/min. pilot
plant. The system consisted of a series of two to eight reactors partially filled with
fine sand. Wastewater was passed upwards through the reactors, fluidizing the sand,
while providing a large surface area upon which bacteria could grow. Aerobic conditions
were maintained by supplying high-purity oxygen gas to the influent of each reactor
stage.
The process achieved an average TCOD removal efficiency of 75 percent and an efflu-
ent TCOD of 48.8 mg/1 (13 mg/1 TBOD ) at an empty bed retention time of 44 minutes and
a TCOD loading rate of 6.4 kg/m /day. Secondary effluent guideline quality was possible
at a retention time as short as 25 minutes. System MLVSS concentrations ranged from
14,000 to 16,000 mg/1, with net waste solids production ranging from 0.26 to 0.57 ka
VSS/kg TCOD removed at solids retention times of 8.7 days and 5.2 days respectfully.
System performance was found to be directly proportional to the ability to control
excess biological growth on the sand and prevent sand particles from washing out of
the system.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
oxygen demand
fluidization
h.lDENTIFIERS'OPEN ENDED TERMS
secondary treatment
water pollution control
BOD removal
oxygen diffusion
c. COSATl I icld'Group
13B
18. DISTRIBUTION STATEMENT
Release to public
19 SECURITY CLASS (This Rtport)
unclassified
66
20 SECURITY CLASS IThis page/
unclassified
22. PRICE
EPA Fern. 7270-1 (R.» 4-77)
56
ft U.S. GOVERNMENT PRINTING OFFICE: 1978 757-140/1365
------- |