US Army Corps
of Engineers
Construction Engineering
Research Laboratory
Sponsored By
University of Pittsburgh
In Cooperation With
U.S. Environmental
Protection Agency
U.S. National
Science Foundation
Proceedings:
••••••I
FIRST INTERNATIONAL CONFERENCE
ON FIXED-FILM BIOLOGICAL PROCESSES
April 20-23,1982
Kings Island, Ohio
Edited by Y.C. Wu, Ed D. Smith,
R.D. Miller, and E.J. Opatken
VOl. II
-------
-------
UPGRADING ACTIVATED SLUDGE PROCESS WITH
ROTATING BIOLOGICAL CONTACTORS
Roger C. Ward. Project Manager, Howard'Needles Tammen
and Bergendoff, Indianapolis, Indiana.
James F. Goble. Superintendent, Crawfordsville Waste-
water Treatment Plant, Crawfordsville, Indiana.
INTRODUCTION '
The City of Crawfordsville, Indiana, is a community of
approximately 13,000 residents and has a diverse industrial
base. The major industrial wastewater sources are metal
plating, printing, and wire fabrication operations. Waste-
water flow from industry constitutes approximately 25% of the
total plant flow.
The original wastewater treatment plant, constructed in
1940, was a 1.0 million gallon per day (MGD) primary and con-
ventional activated sludge facility. Expansion projects
through 1970 increased plant treatment capacity to 1.8 MGD by
expansion of the primary and secondary tankage and aeration
blower system capacity.
In 1977, with Federal and State financial assistance, the
design of yet another improvement project commenced. Facility
planning recommended the expansion of the average daily treat-
ment capacity of the plant to 3.4 MGD and the incorporation of
advanced wastewater treatment facilities (i.e., tertiary fil-
tration). The most cost-effective approach for upgrading the
existing activated sludge process to accommodate the increased
617
-------
design organic loading was determined to be the installation
of a fixed-film biological "roughing" (pretreatment) process
so as to reduce the organic loading imposed on the existing
activated sludge aeration tankage. On the basis of, costs and
operational flexibility, mechanically driven rotating biologi-
cal contactors (RBC's), rather than trickling filters, were
selected as the biological roughing process. Other additions
and modifications to the facility included: expansion of the
primary and secondary tankage, raw sewage pumping and aeration
blower system capacity, addition of dual media filtration,
dissolved air flotation thickening of waste activated sludge,
and belt filter press dewatering of the anaerobically digested
sludge. A schematic of the overall treatment process is pre-
sented on Figure 1 and related design data is summarized in
Table I. RBC tank layout, is shown on Figure 2.
This paper presents:
1. The design methodology used for the sizing and layout
of the RBC units;
2. A comparison of the full-scale operational data
collected since November, 1979 to the performance
predicted by the design methodology;
3. The operational factors which affected the performance
of the RBC units; and
4. The enhancements to the overall plant performance
(e.g., nitrification and secondary clarification)
which are attributed to the RBC process.
RBC DESIGN METHODOLOGY
The surface area requirement of the RBC's was generally
based upon achieving a 50% reduction of the soluble five-day
biochemical oxygen demand (SBOD,-) of the wastewater ahead of
the activated sludge process. Design methods (published in
RBC manufacturer catalogs prior to 1977) for determining the
required surface area typically did not address the biological
roughing application and did not consider applications for
which the effluent SBOD,- would intentionally exceed 25 mg/1.
However, Antonie(l) proposed a design method and model equa-
tion for multiple stage RBC's that did address the cases for
which the effluent SBOD^ would be in excess of 25 mg/1. Fig-
ure 3 illustrates that design method, gives the model equation
for the original sizing of the RBC's, and shows an example of
predicting full-scale performance.
618
-------
fe
o
o
I
UJ
en '
t£
S"-|
619
-------
TABLE I
PLANT DESIGN DATA
Raw Wastewater Strength
Raw Wastewater Flow
Flow Equalization Basin
Raw Sewage Pumping
Primary Settling Tanks
Rotating Biological Contactors
Aeration Tanks
Air Blowers
Secondary Settling Tanks
Filter Feed Pumps
Dual Media Gravity Filters
Anaerobic Digesters
Dissolved Air Flotation Unit
Sludge Dewatering Belt Presses
180 mg/1 BOD5 (five-day
biochemical oxygen demand)
190 mg/1 TSS (total suspended
solids)
Average, 3.4 MGD (million
gallons per day)
650,000 gallon capacity
6 - 1400 GPM (gallons per
minute) pumps
8-10 Ft. x 35 Ft. units
4 - 96,000 Sq. Ft. units
4 - 112,000 gallon tanks
2250 CFM capacity
4-20 Ft. x 122 Ft. units
2 - 2850 GPM pumps
6 - 11.5 Ft. x 11.5 Ft. filters
2 - 35,000 Cu. Ft. units
1 - 35 Ft. diameter unit
2-2 meter units
620
-------
o
CM <
a <
o
m
o:
«
r •
, — i —
t. .
f
it
T
1
• .
»
if
n
t
it
1
i
v -.^* — ^ — ». ^
i
i
. -. t »
\- j
iit
O.U1
2
621
-------
U
O
o
'did oooi-AOT/sai)
SQOS anamos
622
-------
The actual design technique involved several trial and
error solutions. Four standard size 25 ft. x 12 ft. diameter
(100,000 sq.ft. of media per unit) RBC units were determined
to be necessary to meet the process requirements for 50% SBOD,.
reduction, if the RBC units were operated as a two-stage pro-
cess. System flexibility was maximized by configuring the
four units side-by-side and fabricating each unit with two
distinct media sections. See Figure 2. This configuration
permits the isolation of any one RBC unit (for maintenance)
and either a single or two-stage process. Approximately 4%
of the media is removed to affect the division of the RBC
media into two distinct media sections. As a result, the
actual surface area of the media of each RBC unit was 96,000
rather than 100,000 sq.ft. The actual total design surface
area of media was 384,000 sq.ft.
The single stage design example is presented on Figure 3.
The single stage, rather than the two-stage design example is
presented because the RBC units were started up as a single
stage process and have remained as a single stage process.
The performance of the RBC units has not yet required a change
to the two-stage process.
The final design effluent SBODc for a single stage
process was determined on Figure 3 as follows:
1. Calculate the design hydraulic loading.
Design Flow =3.4 MGD plus 10% for recycle streams-=
3.7 MGD
Design Media Surface Area =384,000 sq.ft.
Design Single Stage Hydraulic Loading =3,700,000
gallons per day (gpd) ? 384,000 sq.ft.
= 9.6 gpd/sq.ft.
2. Find the slope of the hydraulic loading line on
Figure 3, at the design hydraulic loading of 9.6
gpd/sq.ft.
The dimensions [gpd/sq.ft.] are equivalent to:
[Ibs. SBODr removed/(day - 1000 sq.ft.)]
Img/1 SBOD5 removed]
•x 1,000/8.34 Ibs./gal.
623
-------
Therefore, to express the design hydraulic loading
of 9.6 gpd/sq.ft. in terms of the slope dimensions
on Figure 3, the hydraulic loading is multiplied by
8.34 Ibs. per gal./I,000.
0 , ,, . 8.34 Ibs./gal.
9.6 gpd/sq.ft. x 17000 =
0.08 Ibs. SBOD5 Removed/(Day-1,000 sq.ft)
mg/1 SBOD5 Removed
A simple graphical display of the calculated slope
is determined by arbitrarily selecting the Y-co-
ordinate as 3.0 and computing the X-coordinate:
X-coordinate
Y-coordinate
slope
3.0
0.08
- 37
Therefore, the slope of the design hydraulic
loading line is arbitrarily shown on Figure 3 as
intersecting the Y-axis at 3.0 and X-axis at 37.
3. Find the single stage design effluent SBOD5, given
the designed hydraulic loading rate of 9.6 gpd/
sq.ft. and the design influent SBOD,- as 79 mg/1.
Shift the hydraulic loading line so that it
intersects the X-axis at 79 mg/1. The X-coor-
dinate of the point of intersection of the
shifted hydraulic loading line with the design
curve is the design effluent SHOD,- which is
shown as 49 mg/1.
Although a single stage configuration did not predict a
50% reduction in SBOD,- at the design single stage hydraulic
and organic loading, the two-stage configuration did predict a
conformance with the basic design requirement.
EVALUATION OF DESIGN METHODOLOGY
Table II lists the monthly operating data of the RBC
units, and contains two columns which readily indicate the
624
-------
11
CQ Cfl
W XI
~^
U CO O
•O O CM
>> i-J O
O O O . O CO
m m cn
-3-
-3-
in cn m CM
in r^
CM CM
m
1
CM
1
in vo in
co
I
co
I
in
I
O
rH
1
cn cn m
rH
o
2:
•u >
rj o
O K
I
3
^
ex 4J
01 U
ov O
> o
O 01
z o
G
00
•H
01
01 ,
a
a. ra
g "
to
a) r-i
u nj a)
•H g 4-J
EO ^ *H
O O 0}
D. C O
6
O 01
O >-i
01
3 3
O CO
CM E
to co
G M rH
o
C. 01 14-4
3 4J o
•a to
mo
to
to
B
o
a. g
E
-. C G
CJ -H
01 E
,4
to
C 3 O
MOO.
• O V4 *O
•rt 3 0)
4J O M
CO 4H tO
M XI
01 >-,
P- V4 0)
O 01 01
> 00
CO 01 to
00 14
to c oi
4-J 01 >
to ^ to
trj
01 4-1 >,
•H i-l G
« 0.0
00 to
c to ai
3 10
•O l-i 01
00 >
T3 O
.01 rH XI
4J tO CO
•J G
01 O T3
rH 'H 01
rH 4J 4J
O U tO
O O -H
CXrH
O) O
J-l M tO
0) CX 4-1
3 tfl
3 Q
0) O
^ iH
O U-l •
CO X cfl
•H -a
•a to to
01 VJ
4-1 14-1 3
cox:
01 H
to oo
01 c x:
J-l -H 00
O. 4J 3
to o
tO 'H ^4 •
4-1- to x: x:
CO G 4J 4J
o
to S
Q x: is
O O G U
pa to 3 01
to o) c/3 a.
625
-------
validity of the design technique: predicted effluent SBOD,-
and actual minus predicted effluent SBOD,-. An example of now
the predicted effluent SBOD,- values were obtained is illus-
trated on Figure 3 for the September, 1981 data set. The
slope of the hydraulic loading line was graphically depicted
as follows:
1. Express the monthly average hydraulic loading of 4.8
gpd/sq.ft. in terms of the slope dimensions on Figure
3.
/ o j/ t: 8.34 Ibs./gal.
4.8 gpd/sq.ft. x 1>OQO * -
0.04 Ibs. SBODr Removed/(Day - 1,000 sq.ft)
mg/1 SBOD,- Removed
2. Arbitrarily select the Y-coordinate as 3.0 and compute
the X-coordinate.
X-coordinate
3.0
0.04
= 75
3. Draw a line with a slope equal to the hydraulic
loading as intersecting the Y-axis at 3.0 and the X-
axis at 75.
The predicted effluent SBOD5 value of 27 mg/1 was
found by shifting the hydraulic loading line so
that it intersects the X-axis at the monthly
average influent SBOD,- value of 59 mg/1. The X-
coordinate of the point of intersection of the
shifted hydraulic loading line with the design
curve is the predicted effluent SBOD,- value,
which is shown as 27 mg/1.
In general, the actual performance of the RBC units has
been reasonably close to performance predicted by the design
methodology. The predicted effluent SBODr values of the 1981
data set averaged approximately 90% of the actual effluent
SBODc values. December, 1979 to February, 1980, data are not
considered representative of normal performance because the
plant received industrial cyanide spills during December, 1979
and January, 1980.
626
-------
RBC PERFORMANCE AND OPERATION
Figure 4 illustrates that the RBC units consistently
removed 40 to 50% of the influent SBOD5. The removal rates
did not appear to be significantly affected by hydraulic or
organic loading rates. For example: The average hydraulic
loading rate in May, 1981, was 8.5 gpd/sq.ft. and the percent
SBODn removal for that month was 41%; whereas, the average hy-
draulic loading rate in November, 1981, was 4.3 gpd/sq.ft. and
the percent SBOD^ removal for that month was 43%.
Figure 5 is a plot of the effluent SBOD5 versus SBODc
removal rate and the original design curve. Most of the 1981
data points are in close agreement with the original design
curve. Better control of the sludge handling/treatment sludge
recycle streams occurred in 1981 and is considered partly re-
sponsible for the improved performance of the RBC units. Al-
so, the industrial source of cyanide was controlled.
In general, the performance during the summer months was
better than during the winter months. Performance factors
other than wastewater temperature probably account for dif-
ference in performance. The wastewater temperature does not
fluctuate significantly (12.0 to 22.0°C). Furthermore, the
coldest monthly average wastewater temperature (12°C) was re-
corded in February, 1981, but the performance during that
month was a similar to the performance during August, 1981,
when the highest monthly average wastewater temperature (22°C)
was recorded. Plant operating personnel attribute the sea-
sonal performance differences to an increase of sludge hand-
ling/treatment recycle streams (e.g., anaerobic digester
supernate) during the winter months due to periodic interrup-
tions of the disposal operations of liquid digested sludge and
lower primary digester temperature.
The daily average SBOD,- loading exceeded a RBC manu-
facturer's (2) recommended limit of 4.0 Ibs. SBOD5/(day-1000
sq.ft.) for 10 out of the 26 months of operation listed.
Plant operating personnel daily have checked the bio-film of
the RBC units for patches of white growth (which is a visual
indication of undesireable forms of microbial life, presumed
to be beggiatoa), and have seldom noted patches of white
growth.
; Routinely, the plant operating personnel exercise buried
drain valves in order to assure valve operability. Signifi-
cant amounts of sludge had been noted to be withdrawn during
627
-------
to
q
CM'
0001 -AVQ/SQ1)
CO
LU
cr
o .
il£
z
Q
Q08 3190105
628
-------
o
z
Ul
o
UJ
< oo
(T-
UJ O
Q.UJ
OQ
< f-
IT- .
UJ O
Q- UJ
OQ
85s
10
ui
cc
0
O
CO
20
8i
oc
5
O
o
•8
..o
o
OJ
ID
rr>
O
to
CJ
Q
CJ
in
'(zidOOOi-AW/sai)
saoa 3iamos
629
-------
the routine exercise of the RBC tank drain valves. Impressed
by the amount of sludge withdrawn, the operating personnel de-
cided to withdraw sludge from the RBC tanks on a regular
basis. Coincidental with the decision to periodically with-
draw sludge from the RBC tanks, the plant operating personnel
also noticed that the white patchy growth on the bio-film
appeared less frequently. Sludge is withdrawn routinely 2 to
3 times per week and more often if white patchy growth
develops.
An attempt was made to correlate SBOD^ removal
efficiency with the practice of routine sludge withdrawal from
the RBC tanks. Sludge was withdrawn from only two of the four
RBC tanks during the period of November, 1981 through
February, 1982. The daily average effluent SBOD5 of the
drained and undrained tanks was nearly identical during the
trial period. Even though the data generated during the
trial period did not support the assumption that periodic
sludge withdrawal has a beneficial impact upon SBOD,- removal,
the plant operating personnel have maintained their periodic
sludge withdrawal operation and have seldom observed white
patchy growth.
AFFECT OF THE RBC PROCESS ON OVERALL PLANT PERFORMANCE
Table III lists the 1981 monthly operational data of the
RBC and activated sludge processes. The secondary effluent
typically was nitrified and had a total suspended solids con-
centration less than 20 mg/1. The biological roughing of the
RBC process not only reduced the SBOD,- load to the activated
sludge process, but was also likely responsible for the excel-
lent quality of the secondary effluent. Both the nitrifica-
tion and the excellent secondary clarification that occurred
would not have been anticipated for an activated sludge pro-
cess which operated at an F/M ratio greater than 0.20/days,
mean cell residence time less than 4.5 days, and a hydraulic
detention time less than 4 hours.
In addition to SBODc removal, the RBC process converted
the form of the suspended solids entering the RBC's from a
very non-descriptive particle type to noticeably long, dark,
stringy biological solids. This formulation of biological
solids ahead of aeration may be one of the enhancement factors
that an RBC process lends to the activated sludge process.
630
-------
C -J en ot
0 uj as 8
cj r; z ^
V
V V V V V
sa
CJ)
S
cu <
CJ H
gg
oil
Ed H
.
01 cl CO
cj oi o
T3 -^
a -H
CO CO 01
oi 0) 6
H
cil o »
" co cJ ^
ill I
Q CO
O C CO
-H O3 rJ
JJ S
CO £>l
os a co
. ••• co
iJ 01
*~* cu
c->
fc
O
C 000
o m c o
LJ rr\ a m
CO -3" CM rH . rH rH
m eg
's »
0 c
CO O
o m
_J rvl
ra
•a
c
o
o
ai
CO
•a .
C!
ra
0)
oo
13
3
01
"O
c
01
D.
CO
3
•a
ai
• M
u-i in co
O C'--^
O
B^3 rO *O
O (U
1^ rH 4-»
CTJ W
[>i *J TO
rH O 3
0) H
W CO
co *J *a
o
ooo
me o
a «H IH
o "a i
M CO >>
C/3 O CO
t-J a
o /—*
•H OttM
rH C "
3 -H fc
CO 73 "-^
U CO O
•O O OH
>v t-J O
-3-
-3-
0
rH
in
en
CM
o
in r~*
CM -T
CM CM
O O
OO
CM
CO
O
vo
rH
CM
O
O
O
o
CM
O
00
rH
O
in
00
tH
o
-3-
rH
O
ON
in
rH
O
rH
OO
rH
O
O
en
in
X 3, O
O rH c/3
^ U-l
O. «-l CO
D. U J3
CO rJ
u
CO CO ?-,
CO K .0
CD t-J
T3 4-1
01 CO CO
T3 B rH
C -H CO
01 X 3
o. o cr
CO ^ 01
3 O.
CO D. 01
CO €
a) -H
rH CO' H
•H -H
U 01
co in o
r-< a a
O O 01
> pa -a
en -H
- co
h u ai
o G os
3 0)
cr 3 rH
•H rH rH
rJ «-l 01
u-< O
•o in
01 B
cd
2
CO
3
co
o
631
-------
CONCLUSIONS
The design method for RBC sizing proposed by Antonie(l)
and as presented on Figure 3 is a valid design method for
sizing a RBC process as a biological roughing process ahead of
an activated sludge process.
Using a RBC process as a biological roughing process
ahead of an activated sludge process is a workable and cost-
effective method for upgrading an existing activated sludge
process.
Periodic sludge withdrawal from RBC tanks may help
prevent white patchy growth on the RBC bio-film.
Sludge handling/treatment recycle streams adversely
affect RBC performance.
The RBC biological roughing process enhances overall
plant performance by producing biological solids which en-
courage nitrification within the activated sludge process and
aid in secondary effluent clarification.
REFERENCES
1. Antonie, R. L., "Rotating Biological Contactor for
Secondary Wastewater Treatment," presented at the Gulp/
Wesner/Culp WWT Seminar on October 27-28, 1976, held at
South Lake Tahoe, Stateline, Nevada.
2. Autotrol Wastewater Treatment Systems Design Manual,
dated 1979, by Autotrol Corporation Bio-systems Division,
Milwaukee, Wisconsin.
632
-------
USE OF SUPPLEMENTAL AERATION AND PH ADJUSTMENT
TO IMPROVE NITRIFICATION
IN A FULL SCALE ROTATING BIOLOGICAL CONTACTOR SYSTEM
James L. Albert. U.S. Army Environmental Hygiene Agency,
Aberdeen Proving Ground, Maryland.
INTRODUCTION
Nitrification was substantially improved following the
installation of supplemental.aeration in a full scale rotating
biological contactor (RBC) system designed to treat domestic
wastewater to an effluent level of 10 mg/L of 5-day total
biochemical oxygen demand (BOD^) and 2 mg/L ammonia-nitrogen
(NH^-N). A slight improvement in nitrification was also
observed when the pH of the wastewater was adjusted from 6.6 to
3.4 with soda ash for an extended period. This RBC system is
the biological treatment portion of a 6 MGD wastewater treatment
plant (WWTP) serving a major U.S Army installation with an
effective population of 40,000. The RBC units are arranged in
6 treatment banks of 6 stages each with the first.3 stages in
each bank intended for BOD, removal and the last 3 stages for
NH -N removal. Performance3of the RBC system was evaluated
during summer conditions (wastewater temperature of 26° C and
system flow of 4.5 MGD) before and after the addition of 8 cfm
of supplemental aeration per lineal foot of RBC shaft in the
first 2 stages of each bank. The effect of pH adjustment was
evaluated by comparing a control bank to a parallel bank in
which up to 1200 pounds per day of soda ash was added to the
third stage for 7 weeks.
" The opinions or assertions contained herein are the
private views of the author and are not to be construed as
official or as reflecting the views of the Department of the
Army or the Department of Defense."
633
-------
Prior to the aeration of the wastewater, dissolved
oxygen (DO) limiting conditions (1.0 mg/L or less) existed
in the first 4 stages, the white sulfur bacteria, Beggiatoa,
predominated on the media, 75% of the BOD^-S (soluble) re-
moval occurred in the first 3 stages, and 59% of the applied
NH,-N was oxidized. With the supplemental aeration, DO was
never less than 1.5 mg/L, Beggiatoa was sparsely present on
only the first stage media, 96% of the BOD..-S removal
occurred in the first 3 stages, and 86% of the NH3~N was
oxidized yielding an effluent concentration of 2.1 mg/L.
The design nitrification rate of 0.3 pounds of NHo-N removed
per 1000 ft^ per day existed only when the DO level was above
2.5 mg/L. Although adjustment of pH produced questionable
results, an 11% improvement in NF^-N removal was briefly
observed as compared to the control bank with just supple-
mental aeration.
Conclusions are that NH3-N removal is dependent on
prior 6005-3 removal so that there is not competition be-
tween 3005 and NH3~N removal organisms for space in the
latter stages. Low DO spread BOD--S removal into stages where
nitrification was to occur. More importantly, the nitrifi-
cation rate was limited by low DO levels. The need to pro-
vide at least 2.5 mg/L of DO in the stages where maximum
nitrification is expected was clearly shown. Some benefit
may be gained through operation in higher pH ranges; however,
the design nitrification rate was achieved in the 6.6 pH
range.
BACKGROUND
Treatment Plant
The WWTP was upgraded in 1977 from a trickling filter
system to a RBC system in order to provide both secondary
treatment and nitrification. The upgraded plant was
designed to meet the National Pollutant Discharge Elimina-
tion System (NPDES) permit limitations shown in Table 1. A
schematic diagram of the unit processes is shown in Figure
1. Design capacity of the WWTP is 6 MGD, while maximum
hydraulic capacity is 18 MGD.
The RBC system, shown in Figure 2, consists of 36
mechanically driven RBC units arranged in a matrix of -6
treatment banks, each with 6 stages. The first 3 stages
634
-------
TABLE 1. NPDES Permit Parameters and Limitations
Parameter
pH
Chlorine Residual
Fecal Coliform (FC)
Suspend Solids (SS)
Five-day Biochemical
Oxygen Demand (BOD-)
Ammonia Nitrogen (NH_-N)
Dissolved Oxygen (DO)
Monthly Average, Summer
(May 1 through October 31)
6.0 - 9.0
Min cone to comply with
FC limit
200/100 mL
30 mg/L
10 mg/L
2.0 mg/L
>6.0 mg/L
635
-------
636
-------
2 u
a CL
•n o
637
-------
in each bank have regular density media (100,000 ft ) for BOD-
removal; the last 3 have high density (150,000 ft2) for
NH3~N removal. The RBC shaft with media measures 25 ft long
and 12 ft in diameter. Each RBC is positioned in a concrete
tank with an approximate volume of 16,500 gal. The stages
are separated by underflow baffles which provide plug flow
through the bank. Based on a dye study (1), the hydraulic
detention time across 6 stages is 2 hours and 30 minutes at
a flow rate of 5.5 MGD.
The BOD5 removal part of the RBC system was designed
based on hydraulic loading (gpd/ft2) versus BOD^ removal
(per cent) curves using an influent BOD5 concentration of
140 mg/L. The overall hydraulic loading is 1.33 gpd/ft2 at
the 6 MGD design flow. Specific removal rates were used to
size the NH3-N removal part of the RBC system. The design
was based on an influent NH3-N concentration to stage 4 of
15.8 mg/L. A removal rate of 0.28 pounds NH3~N removed per
1000 ft2 of media surface per day was used for NHg-N removal
down to 5 mg/L. Removal from 5 to 2 mg/L is to be done at
0.20 pounds NH3~N removed per 1000 ft2 per day(l).
Previous Studies
Summer and winter studies (August 1978 and January 1979)
were conducted by Hitdlebaugh and Miller (1, 2, 3) to evaluate
the performance of the upgraded WWTP, They found that the
RBC system performed at less than design expectations for
BODc and NH3~N removal. This was attributed both to DO
limiting conditions (less than 1 mg/L) in several RBC stages
and to relatively low pH (less than 7.0) in the latter RBC
stages. During the winter study when DO limiting conditions
did not exist, the RBC system actually removed more NH^-N than
during the summer study in spite of the winter wastewater
temperature of 13 C. Analyses of samples for BODj-S and
nitrification-suppressed BOD5 was found to be essential for
the evaluation of WWTP's designed for nitrification. Recom-
mendations for future RBC system designs called for the use
of supplemental aeration to overcome limiting DO levels and
chemical feed to maintain optimum pH levels.
638
-------
Supplemental Aeration
A supplemental aeration system was installed in February
1981 to improve BOD and NH_-N removal in the RBC system. An
equally important benefit was expected to be the physical
stripping of excess attached biological growth, thereby reducing
the possibility of further RBC shaft failures. Between May 1980
and August 1981, 4 stages had become non-operational due to
shaft failure. Diffusers were installed in the first 2 working
stages of each bank as shown in Figure 2. The circular coarse
bubble diffusers (4 per shaft) are offset from the shaft center
plumb line by about 2 ft. Clearance between the RBC media and
tank bottom ranges from 6 in. at stage 1 to 15 in. at stage 6.
Air is provided by 2 blowers, each with 1200 cfm capacity.
Shaft weight measurement devices called "load cells" were also
installed at this time on bank 4. Using these load cells, the
operators can make adjustments to the air flow rate to insure
that the maximum shaft weight specified by the manufacturer is
not exceeded.
METHODOLOGY
Objectives and Materials
The objectives of this study were to evaluate the effect
of the supplemental aeration on the RBC system performance and
to assess the potential benefit from pH adjustment by chemical
addition to levels considered optimum for nitrification. Since
the NH_-N discharge limit had always been exceeded in the month
of August (highest wastewater temperature and lowest DO) and
since the August 1978 study by Hitdlebaugh and Miller provided
an excellent baseline for conditions existing prior to the use
of supplemental aeration, August 19-25, 1981 was selected for
the study period. RBC bank 4 (see Figure 2) had suffered no
structural damage and continued, as in the previous study, to
be used as the primary bank to evaluate the internal performance
of the RBC system. Because bank 3most resembles bank 4, it was
used as the experimental pH adjustment bank. The decision to
add soda ash at the end of stage 3 (see Figure 2) was based
primarily on pilot scale RBC studies done by Stratta and Long(4).
Their studies indicated that pH adjustment with soda ash yielded
NH^-N removal as good as with lime and did not cause solids
precipitation problems. Their studies also showed that the
nitrifying organisms need 5 weeks to acclimate to a higher pH.
639
-------
A chemical feed system consisting of a 500 gal. tank with
flash mixer and a 30 gpm capacity centrifugal pump was operated
for 7 weeks before the August sampling period. Soda ash solu-
tions were made 4 times per day using either 300 Ibs. (7.2 %
solution) or 400 Ibs. (9.6 % solution) of dry soda ash. The
solution was pumped at a constant 1 gpm rate to 2 points
ahead of the effluent baffle in stage 3. The higher concentra-
tion solution was used during the times of the day when peak
wastewater flows occurred. An automatic pH control system,
later seen to be essential, was outside the scope of this
study. Despite equipment problems and washout by rain induced
high flows, the pH in stage 3 was maintained between 8.1 and
9.3 for 4 of the 7 weeks.
Sampling and Analyses
The sampling and analytical program from the 1973 study
was duplicated as closely as possible to permit accurate
comparison of results from both studies. Twenty-four hour
flow proportioned composite samples were collected for 7 days
at the RBC system influent and effluent, and at the WWTP
effluent. Grab samples of the RBC bank influent and waste-
water in each of the 6 stages of banks 3 and 4 were collected
at 5 times during the study to determine changes in waste-
water characteristics through the RBC system. Sampling times
were selected to correspond to those used in the previous
study. Grab samples of the wastewater in stage 6 of the other
4 banks were also collected. Temperature, DO, and pH data were
taken during each sample period using portable instruments.
Sample point locations are shown on Figures 1, 2, and 2a.
All sample analyses were conducted onsite by the Environ-
mental Chemistry Division of the U.S. Army Environmental
Hygiene Agency. A mobile laboratory was set up at the WWTP for
both studies. Nitrification was suppressed using the ammonium
chloride method (5) in order to determine the relative BOD.
exerted by carbonaceous and nitrogenous substances. Tests tor
soluble BOD_ and TOC were conducted on the filtrate passing
through a 0?45 micron filter. All sampling and analyses were
conducted in accordance with "Standard Methods for the Exam-
ination of Water and Wastewater" (6) or "Methods for Chemical
Analysis of Water and Wastes" (7).
640
-------
vO
(U
oo
CD
•60'
cfl
fl)
00
to
ao
to
CO
a)
oo
to
so
.u
02
-o
c
to
m
to
• c
CO
-I
CO
C
to
O
O
t,
tO
CO
60
C
CO
CNJ
C
C
• r-l
j_)
to
O
o
CJ
to
4J
E
Q
u
I — 1
tO
o
60
C
i-H
c
• r-*
O
e.
0)
! 1
O.
co
CO
to
^
OS
' CO
o
c
OJ
-a
641
-------
FINDINGS
The sulfur oxidizing bacteria, Beggiatoa, which had been
abundant during low DO conditions, was sparsely present on only
the first stage media under the new aerated conditions. The
biomass growth was much thinner and more uniform in the aerated
stages, possibly due to physical stripping action of the air.
Hydraulic and organic loading rates are shown in Table 2.-
Wastewater characteristics (from composite samples), in and out
of the RBC system, and at the WWTP effluent are shown in Tables
3 and 4. The performance (from grab samples) of individual
banks is shown in Table 5. Changes in wastewater characteristics
(from grab samples) as it passes through each RBC stage are
shown in Tables 6-8 and Figures 3 and 4. BOD.-S and NH -N
removal rates in each stage are shown in Figures 5 and 6.
Loading and operating conditions were essentially the same
for both studies (Tables 2 and 3). The RBC system effluent
NH--N concentrations were improved from 6.2 mg/L in 1978 to 2.1
mg/L in 1981 because DO limiting conditions were eliminated by
the supplemental aeration (Tables 3 and 6, Figure 3). Had all
36 RBC stages been operational, RBC system effluent NH..-N
concentrations would have been in the 1.3 to 1.8 mg/L range
achieved by indivdual banks 1 and 4 (Table 5). As predicted by
Hitdlebaugh and Miller (2,3), the impact of the aeration on
BOD--S removal was not to increase the amount removed (Table 3),
but to concentrate removal in the first 2 stages (Table 6,
Figure 3) at a higher rate (Figure 5). Because BOD_-S removal
was occurring more efficiently, more space for nitrifying
organisms was available in the early stages and nitrification
actually began in stage 1 (Table 6, Figure 3) with the peak
nitrification rate occurring in stage 3 (Figure 5). Although
BOD- and NH~-N removal organisms do compete for space, the
major factor limiting nitrification was the low wastewater DO
levels (Figure 6). A minimum DO level of 2.0 mg/L available at
the location where nitrification is expected to occur has been
suggested by others (8). The design nitrification rate (1,9)
was observed at the 2.5 mg/L DO level (Figures 5 and 6). The
largest DO drop across any of the stages was across the stage
with the maximum nitrification rate (Figure 6). This is expected
since NH.,-N removal requires 4.6 times as much DO as BOD--S
removal 18). Mechanical reaeration from turning RBC media does
not bring the DO level back up to 2.0 mg/L until 2 stages later
(Figure 6); therefore, supplemental aeration provided to stage 3
would improve nitrification rates and overall NH -N removal.
642
-------
The improved NH--N removal by the RBC system can be seen
in the components of total BOD,, in the WWTP effluent (Table 4).
Before aeration, the nitrification process was continuing out
the end of the plant indicated by a total BOD- of 11 mg/L and
nitrogenous BOD,, of 7 mg/L (total minus carbonaceous). With
aeration, nitrification shifted back up into the RBC system
indicated by a total BOD,, of 5 mg/L and nitrogenous BOD, of .
2 mg/L.
The effects of pH adjustment were disappointing because
it is well documented that higher pH levels (.8.0 to 8.5) are
optimum for nitrification (4), A side by side comparison of
banks 3 and 4, showed no improvement in total NH--N removed
(Table 7), although a different removal rate pattern was
observed. The set of grab samples on August 19 did show an 11%
improvement (Table 8); however, this was questionable because
N07/NO_-N levels did not support the NH -N levels. The major
reason for the poor performance was probably the wide pH fluct-
uations (8.1 to 9.3) inherent in the chemical feed system
which did not let the nitrifying organisms acclimate to a.
constant pH.
Amazingly, the maximum design nitrification rate of 0.3
Ibs NH_-N removed per 1000 ft per day was observed at a PH of
6.6 (Figures 4 and 5). The overall NH.,-N removal of 86% was the
same as that found by Stratta and Long (4), but the maximum
removal rate was less than half of their observed value. Under
design load conditions (6 MGD), pH adjustment may be needed to
achieve the desired NH -N removal with existing media area.
SUMMARY
The supplemental aeration of the RBC system eliminated
nuisance organisms, enhanced BOD5~S and NH--N removal, and
provided the operational flexibility to control the thickness
of biomass growth on the media. Placement of air diffusers in
not only BOD..-S removal stages, but also in NH.,-N removal
stages should be considered. DO levels should Be maintained at
2.5 mg/L in stages where maximum nitrification rates are
expected.
643
-------
ACKNOWLEDGEMENT
s
The author would like to thank Mr. John A. Hitdlebaugh
and Major Roy D. Miller for their helpful advice in the design
of this study. A special thanks is extended to Mr. Charles I.
Noss, Mr. Kenneth A. Bartgis, and Captain Edmund Kobylinski
from the US Army Medical Bioengineering Research and Develop-
ment Laboratory for their efforts in support of the study.
The data shown in the following Tables and Figures
from the August 15-21, 1978 study (without supple-
mental aeration) was extracted from the paper
entitled "Full-Scale Rotating Biological Contactor
for Secondary Treatment," fay John A. Hitdlebaugh
and Roy D. Miller, presented at the first National
Symposium/Workshop on Rotating Biological Contactor
Technology, Champion, PA (1980).
644
-------
TABLE 2. RBC System Loading Rates*
Location
August
15-21, 1978
August
19-25, 1981
Design**
Hydraulic Loading (gpd/ft )
Six Stages
First Stage
1.0
7.5
0.93
7.0
1.33
Organic Loading (Ibs BODs/lOOO ft^/day)
Six Stages
First Stage
0.60
4.5
0.68
5.08
1.56
11.7
Organic Loading (Ibs'BOD5-S/1000 ft~/day)
Six Stages
First Stage
0.18
1.31
0.15
1.11
4.0
*•**
*Based on data in Table 3.
'""'"Based on design flow of 6.0 MGD, RBC influent concentra-
tion of 140 mg/L BOD5, and RBC stage 4 influent concentra-
tion of 15.8 mg/L NH3-N.
-f-^U^j,
This level recommended to prevent DO limiting conditions
in the first stage (9).
645
-------
TABLE 3. RBC System Influent and Effluent Characteristics
August 15-21, 19782 August 19-25, 19813
Avg Flow = 4.5 MGD4 Avg Flow = 4.2 ""
Temp = 26°C Temp
25°C
D
Parameter
Conductivity (umho/cm)
Total Alkalinity
SS
BOD.
BOD^-soluble
TOC3
TOC-soluble
TKN
NH -N
NO /NO -N
Influent
960
158
69
72
21
42
23
21
16.0
0.05
Effluent
930
90
63
61
4
24
11
8.9
6.2
8.9
Influent
890
154
55
87
19
• 55
21
22
16.0
< .01
Effluent
855
84
41
34
4
27
11
6.8
2.1
13
^Values shown are average of 7 - 24 hr composite samples.
^Without supplemental aeration.
.With supplemental aeration.
Sum of STP influent flow (3.7 MGD) and recirculated flow
.(.8 MGD).
Sum of STP influent flow (3.9 MGD) and recirculated flow
(.25 MGD).
All units are mg/L unless otherwise noted.
646
-------
TABLE 4. Wastewater
2
Parameter
Treatment Plant Effluent Values
August 15-21, 19783 August 19-25, 19814
pH (Standard units)
BOD - total
BOD_ - soluble
BOD- - carbonaceous
NH3- N
SS
Flow (MGD)
6.7
11
2
4
6.2
9
3.7
6
5
2
3
2
6
3
.9
.6
.4
.9
^Values shown are average of 7 - 24 hr composite samples.
-All units are mg/L unless otherwise noted.
,Without supplemental aeration.
-With supplemental aeration.
Average of 7 daily measurements made at various times
with a portable instrument.
647
-------
co
CM
co
CO
14-1
8
±J S
CO
3 CM
CM
CM
CM
CM
CO
CO
in
a
-------
4
CU
CO
4-1
en
^t
f\
01
o
fi
2
j_j
o
s_4
_!)
^-«
"c
CO
O
05
•
^O
•Vj
1 1
pa
rJ
X) eu
ON CO
r-l 3
« 4-1
r-4 W3
CM 'CO
1 3
in
"""^
4-1 .
3
00
3
Z
|
CO
O
'Z*
CM
O
z:
z\
1
,co
Z
25
^J
£H
CO
i
ir
Q
O
M
O
O
p^«
r—i
•
H
IE
LO i^*\ ^^ -XJ* CNJ ..^ o^ ' N!)
._( .-^ ,—» r-^ -H r— i "iJ
U™\ W5
.CN ci3
i ^
0*1
rH
;.n m iJ
co co co r^ -o .u^ m co
• rJ
. <
. 4-J
S
3
i — i
M-l r^ CM CO <}• oo i^* in T^ en
CM CM r-l r-l CU CU
4J
o -o
CU r4
r-4 CO
•—4 "^3
0 C
O c3
4J
ON CM r^*» in in vO* ~sj* w M
CM rH !
•1) "3
rJ 3
p . m «
0) O
3 ' rr
. r-4 CU CJ
m i— i o-j co <)• m o soco
C 3 Ci)
1— 1 CU rJ
60 - (1) r4
'O CO > O
PS 4-1 <; u-i
ei en H
649
-------
-4
(0
r-l
O.
S
CO
J.J
rt
O
in
U-4
O
01
co
o
r4
0)
^.
^
1
JJ
c
•u
S
to
•'-i
"O
*— *
a,
U-4
Q
to
O
O
W-l
1- [
—3
^
r-
—2
-5
23
c_4
rjy;
1
CO
O
^
CN
O
Q Z
O
S
0
z
1
II co
1 ffi
"2
p2
i^
0)
c. >•> z
r-l W
S C E-
O O
r— 1
f-( £J
O
O 4-J CO
O (3 1
in 'H m
CM 0) Q
,M
3 ~" <
•* 3 •
i— 1 <3 H
CO 03
,— (
n
m _^
^l ~5-
ON
r-4
i j
vs
-J
ap
<^
i-H r-l , •— 1 CM
O O CM r* m O O ,r-4
O O •— I r^* in vO r-l f^* 'O '^ CO •
1 iQ vp v^Q yr\ v^n v,^ vf^ [— 1 . • i{— > ^^ vO
04
3
CO
• M
4J . CO iJ
a c
0) ^ - D
3 C 3
r-4 fl r-4
«-l •— 1 CM CO <(• m vO pi 14-4 i— 1 CM
3. C
—IS) t— 1 0;
CO M
CJ> S3 CJ nj
=3 ij ;= u
±i CO Ci V2
CO P^
^ r— t— i co
r-H r-l
r-l
o m co CM
r_^
CM ON
• #
m o r- m
r-l r-4
vO ,~O CO CO
~d* \o o i— i
i— 1 i— 1 CM CO
S0 r— i O '"*">
30 r--. m <•
r-4 r-4 r-4 r-4
vD CO O O
r*"^ f^ r^* r^
co -^- m O
CO
OJ
2
• r-4
•^
U-4
0
U
01
to
J)
g
C3
CO
Je
u
4J
C3
-a
0)
^j
a
(D
i—1
r-4
O
o
V} •
CJ sn
i-l 4J
Cu *r^
2 C
« 3
CO
a:
j^i c^
C3
M T3
JO
'4-1 "3
0 C
C3
V] 4J
CU 10
3
O rJ
UJ O
CO "4-
m —
C-
3 O
0 0
r4 r"x
U-l 0)
to — H
•U — ,
r-4 00
3 S
Cfl
D a)
)-l rJ
c3
'4-4
o w
^J
D -H
CO S
nj 3
U
I) r-4
> r-4
*3^
-------
.)
w
o
r— I
g
(3
OO
i-Q
ctj
J-4
u
in
u-i
O
4J
to
] j
o
to
4-1
CJ
U-I
(4-1
a
CO
Cd
S3
^_(
2
I
o1
^
OJ
O
Q
O
"**
CO p^
r^ 1
• m
ii gy
•
»M
c
(0
03
r-i >1 2!
"a
i
4J
to
3
S?
*c
i-l \O
Or— IvOO^O OOCOC7\O
O O r— 1 CO C?l CO in i— 1 CO in C COCNOO O ( — • m,
CN CN r-4 i— i i— 1 -r-l C-J CN CN r-l r-H
4-1
G
g. . ., . .,
4J
to
-^•co^-ooin-^cN 3 COCN if — vovomi~»i~—
CN CN -^ — 1 r-4 i-l CN CMCNCNrUr-lr-(CN
C
o
• ,-4
•u
CO
j^
CU
r^> -^O -^* ^O *xT co o '--O 'vO i ^o ^«o v^c i*^ j"v, \o r*** r*** *co GO r^*- r^*
~*-.
cn
.«
CO - • -(,
4J 4-1 '
<"• ' ^ * C
3 • "a • . a) . -
3 ' co 3
r-H 23 1— 1
i4_ir— icNco in co
651
-------
4.0 -r
3.0 _.
2.0 -•
DO
(ng/D
1.0
O - Without supplemental aeration.
A- With supplemental aeration.
(Supolemental
Aeration)
Pounds
:!H3-N
Removed
Per
1000 £t"
Per Dav
SBC End of
Inr Stage L
.30 -.
.20 ..
.10
Across
Stages: 1
Effects of Dissolved Oxygen Concentration on the
"H,-N Removal Race (Data from Table i).
•• f
652
-------
Pounds
BOD -S
Removed
Per
1000 ft"
Per Day
O — Without supplemental aeration,
Bank 4, Aug 1978.
A- With supplemental aeration,
Bank 4, Aug 1981.
Q- With supplemental aeration and
pH adjustment (soda ash),
Bank 3, Aug 1981.
(Supplemental
Aeration)
Pounds
SH.-N
Remov.ed
Per „
1000 ft
Per 'Day
Across
Stages: 1 2 3 45 6
Figure 5. Comparison of Removal Rates With and Without Supplemental
Aeration and pH Adjustment (Based on data in Tables } & 7).
653
-------
pH
6.8
6.7
6.6
6.5
6.4
6.3
160
150
125 .
Total
Alkalinity
100
Cmg/L)
TKN
6
75'
50
23
20 • •
15 ••
10
Cmg/L)
O - Without
Supplemental
Aeration
A - With Supplemental
Aeration
(Supplemental Aeration)
L\
TKN
NO,/NO--N
Figure
Comparison of pH, Total Alkalinity, TKN, NO,/NO,-N Levels
Through RBC Bank 4 With and Without Supplemental Aeration
(Data from Table £).
654
-------
4.0 -•
3.0 ••
DO
(mg/L)
300 -S
(mg/L)
(mg/L)
O -Without supplemental aeration
A -With supplemental aeration.
(Supplemental Aeration)
Figure 3. Comparison of DO, 300,-S, and NH--N Levels Through RBC Bank <*
With and Without Supplemental Aeration (Data from Table f).
Iff
655
-------
ABBREVIATIONS
BODc 5-day total biochemical oxygen demand
BOD--S Soluble BOD5
CaCO» Calcium carbonate
cfm Cubic feet per minute
DO Dissolved oxygen
Eff Effluent
FC Fecal coliform
2
ft Square feet
gal Gallon
2
gpd/ft Gallons per day per square foot
gpm Gallons per minute
in Inch
Inf Influent
Ibs Pounds
mg/L Milligram per liter
MGD Million gallons per day
ml Milliliter
u.mho/cm Micromhos per centimeter
N Nitrogen
NH,—N Ammonia expressed as nitrogen
N00/NO_-N Nitrite plus nitrate expressed as nitrogen
pEi Negative logarithm of hydrogen ion concentration
SS Suspended solids
T Alk Total alkalinity
Temp °C Temperature in degrees centigrade
TKN Total Kjeldahl nitrogen
TOC Total organic carbon
TOC-S Soluble TOC
656
-------
REFERENCES
1. Hitdlebaugh, J.A., "Phase I, Water Quality Engineering
Special Study No. 32-24-0116-79, Sewage Treatment Plant
Evaluation, Summer Conditions, 14-24 August and 25-29
September, 1978", U.S. Army Environmental Hygiene Agency
(1979).
2. Hitdlebaugh, J.A. and Miller, R.D., "Full-Scale Rotating
Biological Contactor for Secondary Treatment and Nitrifi-
cation", Proceedings: First National Symposium/Workshop on
Rotating Biological Contactor Technology, -.Champion, PAC1980).
3. Hitdlebaugh, J.A, and Miller, R.D., "Operating Problems
with Rotating Biological Contactors", Jour. Water Poll.
Control Fed., 53, 1283 (1981).
4. Stratta, J.M. and Long, D.A., "Nitrification Enhancement
Through pH Control with Rotating Biological Contactors",
Final Report, Institute for Research on Land and Water
Resources, The Pennsylvania State University (1981).
5. Siddigi, R.H., et al,, "Elimination of Nitrification in the
BOD Determination with 0.1 M Ammonia Nitrogen", Jour. Water.
Poll. Control Fed., 39. 579 (1967), '
6. "Standard Methods for the Examination of Water and Waste-
water", 15th Edition, American Public Health Association,
Washington, D.C. (1980).
7. "Methods for Chemical Analysis of Water and Wastes", U.S.
Environ. Protection Agency, EPA-625-16-74-003 (1974).
8. "Process Design Manual for Nitrogen Control", U.S. Environ.
Protection Agency Tech. Transfer (1975).
9. Autotrol Corporation, "Wastewater Treatment System Design
Manual", Milwaukee, WI (1979).
657
-------
APPLICATION OF ROTARY SCREENS, BIOLOGICAL
CONTACTORS, AND GRAVITY PLATE SETTLERS TO
TREAT WASTEWATERS IN HOBOKEN AND NORTH
BERGEN, NEW JERSEY
Joseph M. Lynch, P.E.
President
Mayo, Lynch and Associates, Inc. Hoboken, N.J.
U.S.A.
Jiunn Min Huang, P.E.
Project Manager
Mayo, Lynch and Associates, Inc. Hoboken, N.J.
U.S.A.
C. H. Joseph Yang, Ph.D.
Senior Environmental Engineer
Mayo, Lynch and Associates, Inc. Hoboken, N.J.
U.S.A.
INTRODUCTION
A flow-through compact system composed of
rotary screens, biological contactors, and gravity
plate settlers was tested by Mayo, Lynch and Asso-
ciates at a full-scale pilot plant in Hoboken, New
Jersey for wastewater treatment. The results of
the pilot plant study was applied to the North Ber-
gen Central Sewerage Treatment Plant, North Bergen,
New Jersey which was also designed by Mayo, Lynch
and Associates. The Utilization of this
flow-through system for municipal wastewater
treatment has been demonstrated successfully at the
North Bergn SPT. This is the first full-scale
plant, adopting the concept of flow-through compact
system in the world.
658
-------
Tne flow-through system is composed of
rotoscreen, biological contactocs and gravity plate
settlers. When wastewater flows into this system,
particles greater than 0.02 inch are removed by
rotoscreen first. The rotoscreen, which replaces
of grit chamber and primary settling tank in
conventional treatment system, has the function, of
primary -treatment. The effluent from the roto-
screen then flows.through the biological contactors
which could be either rotating biological
contactors (RBC) or trickling filters. The
evaluation of treatment levels at various rates of
hydraulic loadings and biological contactor stages
was also conducted during the course of this study.
The results of the tests were used to establish
guidelines for the design of the North Bergen
Central Sewage Treatment Plant.
Some advantages of a flow-through system
consisting of rotoscreens, biological contactor and
gravity plate settlers to wastewater treatment
include: reduced land requirements, reduced
capital costs as well as a reduction in operation
and maintenance costs when compared to a
conventional activated sludge process facilities;
ability to meet secondary treatment requirements-;
ability to withstand hydraulic and organic surges;
compatibility in good settling characteristics of
sloughed sludge with the gravity plate settlers;
ease of operation and elimination of wind and
thermal disturbances. These and other
characteristics of the systm were examined during
our evaluation of pilot the plant study, the
operation of the full scale North Bergen Central
Sewage Treatment Plant demonstrated it.
659
-------
The object of this paper is to present the
results o£ an investigation into the feasibility
study of a flow-through system for wastewater
treatment as shown by results from the Hoboken
Pilot Plant. The operation data of North Bergen
Central Sewage Treatment Plant, which was designed
and incorporates the results of the Hoboken Pilot
Plant study, were collected to support the concept
of the flow-through system. Reluctant data
concerning both, the pilot plant study and
full-scale treatment plant operation, are presented
in this paper.
660
-------
PLANT OPERATION
Hoboken Pilot Plant
The pilot plant, was established within the
site of the Hoboken Treatment Plant, Hoboken, New
Jersey. There were two flow-through treatment con-
figurations in the pilot plant tested in full scale
operations to establish treatability. The first
configuration consisted of a rotoscreen, two-stage
trickling filter and a gravity plate settler. The
second configuration was essentially the same
except that the two-stage trickling filter was
replaced by a rotating biological contactor. The
raw wastewater, which flowed to the full-scale
pilot plants by a splitter box shown in Fig 1, was
pumped from the head end of the treatment plant
just ahead of grit chambers. In this way no solids
would be 'removed prior to the rotoscreen. The
rotoscreen is a stainless steel drum of which the
periphery is covered with a stainless steel mesh
with 0.02 inch openings. The unit is so designed
that the sewage passes through the screen and the
solid particles larger" than 0.02 inches become
impinged upon the screen. As the drum rotates the
solids are scraped off the screen thereby cleaning
the screen. The cleaned portion of screen is then
rotated until it again comes into contact with the
raw sewage. *
The
divided
rotating
ter was
tions.
with B.F.
disperse
trickling
installed.
sewage passing through the rotoscreen was
between the trickling filters and the
biological contactor. The trickling fil-
constructed of steel containing four sec-
Each section has eight foot depth packed
. Goodrich vinyl core media. In order to
tne sewage evenly across the top of. the
filter, 16 evenly spaced nozzles were
The gravity of flow of sewage ovr the
media forms a biological slime which provides the
medium for the biological treatment. The biologi-
cally treated sewage then flowed into the LAMELLA
gravity settler manufactured by the Parkson Cor-
poration. This unit separates the biological
solids and other solids from the sewage by means of
sedimentation between inclined plates within the
unit.
661
-------
z
Hit
^
J
u.
u.
UJ
^
•-*
, w ®
~ *• ® z
0 > 5 ~
E « m 0>
_ . o
ai •• ,- ~
_l O "> O
1
•o
— **" "S
CO
UJ
^f«
• Vi»
i <
UJ
o
c
3 ® w
*•• •
^ = o
t. U. Z
I
»--
S >> L. c
— •; o —
Z —
S 2 • I
_i a co o
i
to
<£
cc
•o
••«
—— ^^*- ™~
GO
. 1
>
J—
UJ
II
u.
UJ
1
o
w
*j
2
/
^
1 3>
O C
O *
CO
k
\
o
c ,_
M 0) v~
o tZ °
~ u. z
i—
w*^ -
——•^B*»^
f
^ »
TO>
T3
O
I
T
1
A
T
h
UJ
iju
u.
in
•o
0
CO
j k
, L. |~
~ ** • —
~ -> Z *"
E
-------
The second means of biological treatment
tes.ted was the rotating biological contactor (or
bio-disc). The bio-disc consists of circular
plastic sheets revolving within a hemi-cylindrical
tank, through the sewage flows. The plastic sheets
or disks rotate through the sewage growing the
biological slime for treatment. The disk used for
this pilot plant was 11'-8" long, 12'-10" wide.
The unit was divided into five equal stages each
containing 8,190 ft2 of media surface. The
circular media sheets were 1'-9" in diameter. The
rotating bio-disc unit was manufactured by EPCO-
Horrnel Co. During the study various stages - two,
three, four and five - RBC' s were utilized.
The effluent from the rotating bio-disc then
flowed by gravity to the microstrainer or LAMELLA
gravity settler. The microstrainer was tested to
see if the biological solids produced by the rota-
ting oio-disc could be" removed or not. The mirco-
strainer has a drum covered with a fine mesh having
an opening of 35 microns. The microstrainer was
manufactured by ZORN.
The rotating bio-disc, microstrainer and elec-
trical controls were contained within a 24" by 16"
by 16' wood building to protect them from the ele-
ments . • - -
24 hour composite samples were analyzed for
BOD5, dissolved oxygen, settlable solids and sus-
pended solids. Temperature, and. pH were taken from
plant records and grab samples. Those analyses
were petforrned in accordance with Standard Methods
(2). -.'•',.
North Bergen Central Sewage Treatment Plant
The Central Sewage Treatment Plant- in North
Bergen, New Jersey is an application of rotary
screens, biological contactors, and gravity plate
settlers as a process for treatment of wastewater.
663
-------
The treatment plant consists of the following
major treatment units: bar screens, lift station,
self cleaning fine rotary screen for primary grit
and suspended solids reduction; rotating bio-disc
system for biological treatment; gravity plate
settlers for separation of the biomass and
suspended solids from the rotating bio-disc system
effluent; post aeration to increase dissolved
oxygen concentration in the stream; and
chlorination system for the final effluent
disinfection. This whole process is shown in Fig.
2.
The plant uses a rotary self-cleaning screen
in place of the conventional grit chamber and
primary clarifiers. The Hydrocyclonics
Corporation's Model RSA-36120 with 0.02" opening
was selected for design because it will pass more
flow and give an effluent with almost identical
levels of suspended solids.
The rotating bio-disc system was designed to
lower the soluble BOD to 10 mg/1. A hydraulic
application rate of 2.4 gpd/ft^. is expected to
accomplish this, using a two stage arrangement to
assure that the system be kept aerobic. At the
design flow rate, a minimum of 4.16 million square
feet of media surface area is expected to provide
this degree of treatment. This amount of surface
area can be provided by using 32 units arranged as
16 flow streams of two stages in which the area of
the first stage is 104,000 ft2 per shaft and the
area of the second stage is 156,000 ft2 per
shaft. Each shaft of discs is installed in a steel
tank with each such unit serving as a stage of
treatment. The shafts will be rotated at a speed
of 1.6 rpm.
The rotating bio-disc system installed in
the North Bergen Central Treatment Plant can be
summarized as follows:
664
-------
1-
z
Z
111
UJ
cc
H
Ul
CD
UJ
CO
cc
K
Z
UJ
o
z
111
cp
QC
Ul
00
CC
O
Z
o
t-
<
2
ui
i
o
CO
CM
*
0>
665
-------
Type:Autotrol Model 601-251
and 6b1-251 Total Number of Unit32 Surface
area/each Unit:
1st Stage 104,000 sq ft. and
2nd Stage 156,000 sq. ft.
Drive Speed: 1.b rpm
Diameter of Disc. 3.6 meter bio-surf HD shaft
Hydraulic loading: rate 2.4 gpd/ft2
The Claripak gravity plate settlers installed
in the North Bergen Central Treatment plant are
summarized as below:
Type
Total Number of Unit
Dimensions of
FRP Sheets
Inclination
Projected Settling
area/each Unit
Hydraulic Loading
Peabody Welles Series 3000A
7
24" wide x 120" long
55°
2,500 ft2 ea.
571 gpd/ft2
The gravity plate settler is suitable for
final solids separation following either a rotating
bio-disc or trickling filter system which is demon-
strated by the pilot plant study in this paper.
The operation of North Bergen Central Sewage
Treatment Plant was described according to the
parameters of 6005, suspended solids, volatile
suspended solids, ammonia nitrogen, dissolved
oxygen, chlorine residual, fecal coliform,
temperature, and pH which was :analyzed in the
treatment plant by plant personnel.
666
-------
RESULTS AND DISCUSSIONS '
Hoooken Pilot Plant Study
The pilot plant established in the Hoboken
Treatment plant was operated in a full-scale
facility. The wastewater flows to the pilot plant
varied between 13.7 and 56.8 gpm over the test
period. Sut the hydraulic loading rate on the
trickling filters or RBC were controlled at desired
values by pumping. The wastewater temperatures
•ranged from ." 1 1 ° to 26°C. The pH levels remained
fairly constant at 6.5 +0.5.
The BOD concentration of the wastewater in
Hoboken, New Jersey plant varied throughout the
experimentation period. Figure 3 shows the
influent BOD concentration of. the raw wastewater on
the in.dicatd months. Most of the raw wastewater
BOD concentration ranged from 75 to 175 mg/1. The
total suspended solids (TSS) concentration of the
raw wastewater is shown in Figure 4. TSS varied
from 14 mg/1 to 175 mg/1 averaging 58 mg/1 over the
test period. But most of them were still in the
range of 20 to 80 mg/1.
The performance of each unit in the pilot
plant study is described below.
667
-------
225-
200-
175-
150-
125-
n-
g50
5
e i
Z I
d
0'
DAH.Y EIO-DISC B.O.Q CONCENTRATION
influent
effluent
^ ^, +. ^ ^, ^-
AUG. SEPT. OCT. NOV. DEC. APR. JUL.
19 77
1978
Fig. 3 INFLUENT AND EFFLUENT BOD CONCENTRATIONS
BIO-DISC CONFIGURATION
668
-------
DAILY BIO-DISC T.S.S. CONCENTRATION
i
O "
I—
<
cc
>—
UJ
Z
O 20-
AUG.
1977
Jl
influent
effluent
"•*•«
SEPT. OCT.
NOV. DEC.
APR.
1978
V
JUL.
Fig.4 INFLUENT AND EFFLUENT TSS CONCENTRATIONS
BIO-DISC CONFIGURATION
669
-------
Rotoscreen (Rotostrainer)
The openings in the rotoscreen were 0.02
inches against a flow ranging from 13.7 to 56.8
gpm. All solids larger than 0.02 inches were re-
moved. The influent suspended solids concentration
ranged from 12 to 162 mg/1, averaging 75 mg/1 over
the testing period in April, 1978. The effluent
readings from the rotoscreen for suspended solids
ranged from 5 to 149 mg/1, averaging 65.6 mg/1.
The average percentage removal for suspended solids
is 7.7%. The actual levels for the influent sus-
pended solids should be somewhat higher than the
values indicated in the report. This condition
stems from the fact that the sampling devices used
to record the data were not capable of accepting
gross vegetable and fecal matter found in the raw
sewage influent samples. This restriction in the
size of particles accepted by the sampler would
produce lower than actual results in the influent
but would have little effect on the effluent re-
sults. Actual suspended solids removal of 15 to
20% have been documented elsewhere (3). It is
anticipated that the actual removals were higher
than the pilot plant test data indicated. The
North Bergen treatment plant design was based on
20% suspended solids removal by the rotoscreen.
The actual operations of rotoscreen in the North
Bergen Central Sewage Treatment Plant from October,
1981 to January 1982 are shown in Table 9 in which
it snows average suspended solids removal was
28.9%.
Compared to the primary settling tank, the ro-
toscreen can remove less suspended solids. How-
ever, it does not significantly affect the
performance of trickling filter or rotating
bio-disc. The BOD load applied to a trickling
filter or rotating bio-disc does not include the
fraction which is included in a settleable solids.
Therefore, the rotoscreen removes less suspended
solids than does a primary settling tank, this will
not signficantly increase the BOD loading to the
670
-------
trickling filter or rotating bio-disc process. All
influent settleable BOD is not required to be re-
moved in tne primary treatment unit as long as it
is not solids which can plug the media and inter-
fere with the biological activity. The settleable
solids which are not removed by rotoscreen will be
settled in the secondary clarifier.
used in
was
re-
in Tables
Trickling Filter
A plastic media trickling filter was use<
this study. The effluent from the rotoscreen
fed to the trickling filter. The BOD and SS „_
movals by the trickling filter are shown in Tables
3 and 4. The trickling filter effluent sample was
treated to simulate a final clarifier. The corn-
posit sample of the trickling filter was settled in
a one-liter graduated cylinder for 60 or 30 minutes
depending on BOD or SS analyses, It was found dur-
ing SS testing with the LAMELLA gravity settler
that the 30 minute settling test more closely simu-
lated the LAMELLA gravity settler, then the 60
• • - • used to .simulate the final
minute settling was used
n-ir»nf-ciC! aohf-linrT Fi"^y" SOD
lated
minute settling was
clarifier. Therefore/ 30
for SS analyses, and 60
analyses.
minutes settling for
As shown in Table 1 , at the hydraulic loading
rate of 1.4 gpm/ft2 the overall BOD removals were
more than 8b%. It can be concluded that the trick-
ling filter provides excellent removals under the
application of rotoscreen for the primary treat-
ment. The various BOD concentration of 8.8, 10.0,
18.9, and 16.2 mg/1 from the tower 2 shown in Table
1 exhibits the lower limit with biological treat-
ment. The pilot plant test result has shown that
the biological reaction in the trickling filter is
similar to the first order kenietic reaction. This
implied that the rate of SOD removal will increase
as the influent BOD concentration increases. This
phenomenon has been demonstrated on wastewater at
specific hydraulic loadings for a BOD5 concentra-
671
-------
lr*j jj 'o
o *
•o w s
c j-i &-<
'O 01
C 4-1
(0 -
Qj I
0? (
•o :
: 0) S-)
i CN JJ <0
I' fO C rH
0) O fXJ >.i O Jl' Q (0
D U-i T3 -C -H C' U T) .£ •*-»
3 vo AJ ->-i <2 6 vO aJ -^ 3J
" — fij 3 ^. EH *- ,-B 3 e
1
.s
*U_I
U
ra
e
c
4J
"i
n
a*
jj
c
t-i
fc
(u
jj
2
IT
S
u
1
s
ud
U)
u
u
It
u
•».;
U
2
OJ
i!
£u
E*
"«-H
O
U
Ss
y
•H
iJ
8'
(-H
a.
°
1
9
^
1
s
™
I
u
.-(
2
o
S
i—
§
Q
i
g
^~
Vi
S
01
jj
•M
!t
V
•* f.
" «
H) W
JJ
g
m tu
m a
C 4J
•rf m m u
*D C1 iJ «
§ as. gl
d Sslg:
0 < r- S S.
'^ !
1 >
O
±1
672
-------
tions ranging from 80 to 600 mg/1 (4). So the BOD
removals were more than 85% in the Hoboken pilot
plant study by using the rotoscreen as a primary
treatment.
Due to river inflow and infiltration, a high
chloride concentration, 50 to 770 mg/1 is found in
the influent to tne Hoboken Treatment Plant. How-
ever, no deleterious effect on the process was
observed during the pilot test period.
The effect of NaCl on biological film has been
reported by Lawton (5). He found that a step in-
crease in salt concentration to 20,000 mg/1 had a
deleterious effects on the film growth, but the
film growth recovered in one day.
As shown in Table 1, the effluent BOD concen-
tration from Tower 2 was lower than from Tower 1,
but with two towers, the trickling filter did not
significantly further reduce the BOD concentration.
This was because some of influent BOD concentration
were lower and the effluent BOD from Tower 1 had
reached the equilibrium BOD. Therefore, the BOD
removal efficiency was improved a little by
increasing the number of towers, under this,
wastewater characteristics and hydraulic loading
rate. The provision of the additional tower to the
treatment can secure the quality of the effluent
BOD concentration. In case the influent BOD
concentration is higher, the effluent of the first
tower can be treated by the second tower.
The overall suspended solids removals in this
study ranged from 61 to 82%. as shown in Table 2.
The effluent suspended solids concentration from
two-stage trickling filter was less than 20 mg/1
after 30 minute settling.
673
-------
CUV
04 OJ
-U CO 4-> "r-t
— JJ O r- jj
0) JJ ni . i
S O SI -
674
-------
Rotating-Biolog ical -Contactor-; , , - -,
In this study, another configuration was made
by using a rotating biological contactor (RBC)
in place .of the trickling filter to treat
wastewater which was primarily screened by a
rotoscreen. The treated effluent from the RBC,
then, was settled by LAMELLA^gravity settler. RBCs
like trickling filters are a fixed-film biological
treatment process. The settleable solids which
passed through the rotoscreen would hot effect the
RBC. The reasons are explained in the section on
trickling filters. The incoming settleable solids,
which will not plug the media, may be removed in a
final clarifier or LAMELLA gravity settler. The
low BOD and SS effluent concentration shown in
Tables 3, 4 and 5 implied the feasibility of
flow-througn system, by using a , rotoscreen for
primary .treatment and an RBC for secondary
treatment. Therefore, using an RBC or a trickling
filter in the flow-through system .resulted the same
efficiency on both BOD and S3 removal.
The rotating bio-disc (or RBC) was arranged
for two-stage operation at a hydraulic loading rate
2.5 gpd/r"t2 initially. The bio-disc was
during two periods. One was from
12/8/77. The other was from ,4/6/78
The result of the two periods are shown
operated
8/12/77 to
to 4/30/78.
in Tables 3
and 4. The wastewater
period ranged from' 1.2°
The major difference in
wastewater temperature
temperatures in the first
to 26°C averaging 19.6°C.
the second period was the
ranged from 12° to 15°C
averaging 14 C. The results from these two testing
periods indicates that at a hydraulic loading of
2.5 gpd/ft2 with a rotation rate of 1.6 rpm the
two-stage bio-disc would obtain 84% 8005 removal
and an expected 6005 effluent of 1-7.5 mg/1.
Suspended solids levels in the effluent are
expected to be 18 mg/1.
675
-------
TABLE 3
BOD5 and Suspended Solids Removal by Rotating
Biological Contactor during 8/12/77 to 12/8/79
Parameter
BOD5*
Suspended Solids*
0005 Removal
SS Removal
Effluent Range
4 co 48 mg/1
1 to 44 mg/1
61 to 100%
18 mg/1
16 mg/1
80%
60%
TABLS 4
6005 and Suspended Solids Removal by Rotating
Biological Contactors During 4/6/78 to 4/30/78
Parameter
Suspended Solids*
8005 Removal
SS Removal
Effluent Range
6 to 23 mg/1
1 to 57 mg/1
75 to 95.7%
Average
15 mg/1
22 mg/1
87. 8%
66%
After 60 minute laboratory settling for BOD analysis
After 30 minute laboratory settling for SS analysis
676
-------
.C
* 3 M
US\
,-i
5
•§ §
o s — '
So <#• CF\
iK — | CO
1
-.1
'— !
AJ
JJ r-1
^ £j
c -
•g as
U-l
§\ o
at AJ
^,
OT \
-S
5sg
O
in to
n't
M < T-
- oo
VD i •
«— ft ^d1
• in
en ' — ^?
|5 ^!
p o
T- GJ
m i- <: is ^
O 0
4J » r* , 4-)
r- TO > •-* r-
^D a*1. - 4J
o gj aj
• m
o •
• CM
-r~
33 % •"
O> rf, CO
O CD en
co > rvj
;-; s^
1 &
3 w SI cr.
- o w\ —
> £1)
sa
-co 05 co
\. o \ \ ON-..
cr» co
-------
In stage analysis, the RBCs were arranged for
a two, three, four and five stage operation at
hydraulic loading rates 2.0 and 2.5 gpd/ft2. The
RBCs were rotating at 1 .7 ' rpm for the period of
test. At two stage operation, the BOD removal was
89% as shown in Table 5. As the stages increased
from 3 stages to 5 stages, the BOD removals were
changed from 38% to 83%. The BOD removal
efficiency versus stages is plotted in Figure 5.
Prom the results, it was found the BOD removal was
higher at two stages than any other additional
stages. The performance of the RBC system was
related to the characteristics of the wastewater,
as it left the rotoscreen. The higher influent BOD
in two and three stages might result in higher BOD
removal efficiency. Although, on the average the
five stages did not remove any farther BOD, .it did
remove additional ammonia nitrogen. The four stage
system on the average lowered the ammonia nitrogen
concentration from 10.4 mg/1 to 5.2 mg/1, which is
a 50% reduction. An additional 1.8 mg/1 ammonia
nitrogen was removed by the fifth stage for a total
ammonia nitrogen reduction in the five stages of
67.3%. Apparently, BOD reduction in terms of
soluble BOD was nearly complete at a loading of 2.5
gpd/ft2.
In terms of net solids production, the two
stage system loaded at 2.5 gpd/ft2 gave a net
solids production of -17 mg/1; but to three and
four stages gave -6 mg/1 and 7 mg/1 at 2.5 gpd/2,
to five stages gave 9 mg/1 at 2.0 gpd/ft2 as
shown in Table 5. The net solids production was
calculated by substracting the suspended solids
concentration in the rotoscreen effluent from the
mixed-liquor suspended solids concentration in the
RBC. This difference represents the suspended
solids concentration changes through the RBC. The
sludge production per BOD removals for various
stages are shown in the last column of Table 5, and
also plotted in Figure 6. Sludge production was
calculated by substracting the suspended solids
concentration in the final clarifier effluent from
678
-------
o
.2
UI-
oc
Q
O
90
£ 85
O
•2.
yj
g
u_
u.
yj
8 0
75
70
O
2 3
STAGES OF RBC
Fig.5 EFFECT OF STAGES OF RBC ON
BOD REMOVAL EFFICIENCY
679
-------
.Q
«•».
JQ
Q
LU
O
2
ill
cc
Q
O
CO
CO
LOO
0.75
0.50
Q
111
O
§ Q£5
DC
Q.
LU
O
Q
JL
0
2 3
STAGES OF RBC
Fig.6 EFFECT OF STAGES OF RBC ON SLUDGE
PRODUCTION PER UNIT BOD REMOVAL
680
-------
the mixed liquor suspended solids concentration in
RBC. Dividing this sludge production by the
decrease in BOD concentration through the test unit
yields sludge production as sludge produced per
unit BOD removed. At the same hydraulic loading,
the smaller number of' stages produced low sludge
production per BOD removal. This is exactly what
would be expected from considerations of F/M where
F is food and M is biomass. In the RBC system, the
F/M ratio is directly to the F/A ratio, where A is
the area of active surface area in the system. The
more area in the first stage of the RBC system, the
lower is its F/A ratio and by comparison with F/M,
the lower should be the net sludge production.
When the hydraulic loading rate increased from 2.5
gpd/ft2 to 5.0 gpd/ft2 in a two-stage RBC, the
net sludge production per BOD removal would be more
as expected F/M ratio increase. As shown in Table
6, at the two stages of RBC the net solids
production and sludge production per BOD removal at
5.0 gpd/ft2 are larger than that at 2.5
gpd/ft2. .
.This observation of the performance of two
RBCs versus a larger number in a staged RBC system
gives one the feeling that future designs would be
optimized by spreading the surface area over a
smaller number of stages. For carbonaceous organic
removal, this would lead to a lower sludge produc-
tion, a more stable biomass, and a lower BOD in the
final effluent caused not by better soluble BOD
removal, but more by . the lower BOD of the more
stable biomass that might escape settling in the
final clarifier. The optimum number of stages
would appear to be two stages, for BOD removal and
for a given amount of tital supplied surface area.
When comparing with single state system, two stage
would reduce the impace of short circuit, toxic,
and shock loading due to low F/M.
681
-------
LAMELLA Gravity Settler
The LAMELLA gravity settler was used for final
clarification in the two treatment configurations
of pilot plant study. The testing results are
shown in Tables 7 and 8.
The hydraulic variation imposed on the LAMELLA
gravity settlers included loadings of 360, 720 and
1008 gpd/ft2 in the rotating bio-disc study. The
results are shown in Table 7. During the 360
gpd/£t2 loading priod the effluent quality
averaged 14 rag/1; the corresponding removal
averaged 30.5%. This removal efficiency reflects
the excellent floe development in the RBC process.
However, the 360 gpd/ft2 loading provided no real
challenge for the LAMELLA gravity settler.
Correspondingly, the hydraulic loading rates were
increased. At the 720 and 1008 gpd/ft2 loading
the effluent S3 was 23 and 14 mg/1 respectively.
Even the effluent SS were lower than 30 mg/1,
however, the removal efficiency were not shown
as good. This was due to the lower solids flux
through the rotating biodisc process during the
study.
In a two-day investigation, it was found the
effluent SS was 12 mg/1 at the 1530 gpd/ft2 load-
ing. The results indicate the LAMELLA gravity set-
tler can treat rotating bio-disc effluent at higher
surface loading rate. The intent of this quick
study was to examine draraat ically higher loadings
while the plant was relatively stable. Althouth
the LAMELLA gravity settler displayed good sus-
pended solids removals at the 1580 gpd/ft2 load
682
-------
ing, this test
full scale design.
720 or 1008 gpd/ft
was too short to be utilized for
However, higher design loading
are probable.
The performance of LAMELLA gravity settler foe
the effluent of the trickling filter process which
was operated in the 1.4 gpm/ft2 is shown in Table-
8.- The average effluent suspended solids was 8 and
11 mg/1 while the loading rate on the LAMELLA gra-
vity 'settler was 360 and 1008 gpd/ft2 respetive-
ly. These results exhibited that even when the
loading rate was up to 1008 gpd/ft2 the effluent
suspended solids was 11 mg/1 and removal effiency
;was 75%,
Therefore, the investigation indicates that
the LAMELLA gravity settler is suitable for final
solids separation following either a rotating
bio-disc or trickling filter system. The loading
rates on the LAMELLA gravity settler can be up to
1008 gpd/ft2. The investigation also found the
effect of biological Solids on the intervals of the
LAMELLA gravity settler was minor. Some growth
occurred on the wetted surfaces which only appeard
as a film and never affected the workings of the
LAMELLA gravity settler. The LAMELLA gravity
settler was cleaned twice in the six months of.
pilot operation and only then because of
changeovers in operations. Clogging problems are
existant in waste treatment application(2). .But
the application of LAMELLA gravity settler on
rotating bio-disc and trickling filter solids
exhibited a good solids separation. The effluent
solids of the rotating bib-disc and trickling
filter are unlike those produced by activated
sludge. The good settling characteristics of the
solids found in the rotating bio-disc or trickling,
filter effluent is due to the low mixed liquor
suspended solids concentration and result in
discrete settling which greatly enhance the solids
settling velocity. Zone, hindered or compression,
683
-------
Table 6
The Effect of Hydraulic Loading Rate
on wet Solids Production and BOD
Removal at Two-Stage RBC
Period
HLR
Temperature (gpd/
Net Solids
Production
(mg/1)
BOD
Removal
Sludye
Production
per BOD
Removal
(Ib/lb)
7/19/78
o
/28/78
24
5.0
21
82.2
0.81
4/b/78
to
4/30/78
24
2.5
1 .7
85.0
0.50
Taoie 7
The Performance of Lamella Gravity Settler for the
Effluent of Bio-Disc Process
Overflow Rate
(gjjd/ft2)
360
720
1008
Parameter
BOD5
SS
BOD5
SS
30 D 5
Effluent (mg/1)
14
9
28
23
21
14
Overall
Removal (%)
80.5
82.4
75.0
60.0
«4. 5
67.5
684
-------
and is a function
plate settler the
distributed. The
by outside forces
sudden hydraulic
settling does not occur in the secondary clarifier
following the biological contactor system. The
removal of discrete particles is independent of
tank depth and detention time
only of the overflow rate. In
flow is laminar and uniformly
suspended solids are not upset
such as convection currents or
change.
Microstrainer
A 35 micron microstrainer was directly applied
to the effluent from the two-stage RBC with
hydraulic loading rate of 2.5 gpm/ft^. The test
results of BOD and SS removal of two-stage RBC
system, which received wastewater from a roto
screen, by microstrainer and simulated clarifier
are shown in Table 9. The simulated clarifier is
mentioned before when settling RBC effluent in a
one-liter graduated cylinder for 60 and 30 minutes
for BOD and S3 tests respectively.
The results in Table 9 indicate that a
microstrainer achieved better SS removal than the
simulated clarifier, while the effluent 8005 was
ony 6 mg/1. However," after August 31, 1977 it was
found that the microscreen failed due to heavy
slime growths. Microorgamisms grew up quickly on
the screen's surface which caused clogging and
biofloc in the effluent. The use of chlorinated
water for cleaning gave only temporary improvement.
Those operation difficulties made it impossible to
apply a microstrainer for the removal of SS from
RBC effluent.
685
-------
Table 8
The Performance of Lamella Gravity Settler for the Effluent
of the Trickling Filter Process
Overflow Rate
(ypd/ft2)
360
1008
Parameter
BOD 5
SS
BOD 5
SS
Effluent (mg/1)
12
8
22
11
Overall
Removal (%)
90
81
' 86
75
Table 9
BOD and SS Jteinovals of Two-Stage RtC System
by Micrcscriner and Simulated Clarit'iec
Kiram&ter
No at
bampK-s
12
12
. fluent
(n.j/1)
IGo
44
%MK)val se
, Microstrainer (rtg/1) 'Miccostrainer Clarlfier '(rag/I) Clarifier
81.9
11
10
bO.6
77.3
IVSt
From a/ 1-3/77 to 8/31/77
686
-------
North Bergen Treatment Plant
The concept of design for the North" Bergen
Central Sewage Treatment Plant was based on the
results of Hoboken pilot plant, study which was
proven to be a reliable technology. This 10 MGD
treatment plant was finished in-1981. According to
the operation report of North Bergen Treatment
Plant/ the average data for the first 4 months,
October, Nqvember, and December of 1981 and January
of 1982 are summarized in Tables 10, 11 and 12.
The numbers shown on Tables 10, 11 and 12 are the
averages of data, based on daily sampling. Table
10 shows the flow rates and characteristics of raw
wastewater. The actual average flow was from 0.85
to 1.5 MGD which is much less than the design flow
rate 10 MGD. This is because -all areas have not
been hooked up the sewer line yet. - Therefore, part
of the facilities at North Bergen Central Sludge'
Treatment Plant are not in operation.
The application of rotoscreens, in the North
Bergen Central Sewage Treatment Plant, was for
primary treatment. The results of the operation
are shown in Table 11. The suspended solids
removals were from 24.9 to 33.6% averaging 28.9%
during the first four month operation. The
manufacturer's information showed about 15 to 20%
SS removal when applying rotoscreening to municipal
sewage. Therefore, 20% SS removal by applying/
rotoscreens for preliminary treatment can be
expected. Two advantages of rotoscreens are high
dry weight of solids produced and small floor
requirements, associated with low operation and
capital costs. The maintenance costs are also
lower due to fewer moving parts when compared with
primary clarifiers.
Two-stage rotating biological contactors were
applied to remove BOD in the North Bergen Central
Sewage Treatment Plant. The BOD removal efficiency
ranged from H4.1 to 89.7% averaging 86.8% during
the first four months. Compared with the data in
687
-------
00
CO
in
in
0)
C
(0
03 r-(
O 0<
•i-l
4J | '
U) C
•H CO
t I C
0) 4->
4J (0
O CO
(0 M
(0
.G CO
U 0i
-O I
C CU
CN
O g «-
CQ -^
EC
0)
r--
cn
LD
o
m
00
en
m
in
en
n
ex
o
in
m
oo
<=*
r-
Temper at u
(°C)
JC
4J
G
O
IS
• s
CO O
> 1-1
O
2;
^o
o^
^*
oo
•
0
CO
Q
g
cu
o
CO
Q
±
OO
o\
•
o
>1
J»4
(0
d
c
(0
1-3
688
-------
0\P
CO >
O
cy
on
ro
CTi
CM
r-
Is
0)
r-i
JD
(0
EH
O
53 4->
C
CD rt5
& r-H
C 4->
•H C
0)
C S
0) -U
0) <0
y> 1;
O H
03 &H
O
4J D
O en
Pi, 03
to
(0 rtf
> S-l
O 4J
e c
iD 0)
« O
CO
4J T-
c \
CO CM
C 13 OO
O T- C ov
0)
"o
4J
o
«- o
CU
O
^
0)
CJ
fl>
Q
Jj
m
a
c
CO
Jen
lea
la;
•>
689
-------
M r- r—
" S * °
— "o— o
r— I r_
! ^
690
-------
Hoboken pilot plant study, the average influent BOD
concentration was 109.1 mg/1 which is about the
same influent BOD concentration at the Hoboken
Pilot Plant study. The" average effluent BOD
concentration was 14.5 mg/1 which is close to or a
little lower than that in the Hoboken pilot plant
study. However, nitrification was already ocurring
in the North Bergen treatment plant. In Table 12,
it shows the .averaging ammonia nitrogen removal was
57.1%. The average ammonia nitrogen concentration
was reduced from 16.3 mg/1 to 6.9 mg/1. . '•
Suspended solids removal ranged from 77.1 to
88.9% averaging 82.5%. The average effluent,
suspended solids concentration of 18.7 mg/1 from
LAMELLA gravity clarifier implied that the
settlability of biological slime sloughed from RBC
was good. After four months operation the
biological solids did not grow on the surfaces of
LAMELLA gravity plate which also implied that the
characteristics of biological solids from RBC was
suitable to the use of LAMELLA gravity settler for,
the final liquid solids separation.
Some advantages of LAMELLA gravity settlers
includes low space requirements, low installed
costs, low maintenance, due to fewer moving parts
to wear, replace and adjust, and high efficiency.
In a well designed and properly sized LAMELLA
gravity • settler, the flow is laminar, therefore,
the suspended solids are not upset by outside
forces such as convention currents or sudden
hydraulic change. '
The successful operations in the North Bergen
Central Sewage Treatment Plant duri.ng the last four
months implied that the application of flow-through
system to municipal sewage is feasible. • The short
691
-------
detention time of rotoscreening in the system pro-
vides the great advantage to the biological treat-
ment system. Usually, the"hydraulic detention in
tne primary clarifier is two hours. But the deten-
tion in the rotoscreen is only 2 to 3 minutes.
Therefore, tne difference between the influent tem-
perature and the effluent temperature is not too
much. This is very important for biological acti-
vity during the winter time. The higher water tem-
perature makes the biological activity in the reac-
tor higher. The comparison of rotoscreen versus
primary settling tank and LAMELLA gravity settler
versus secondary settling tank to their detention
times and land requirements are listed in Table 13.
The detention time in the entire system including
rotoscreen, RBC, and LAMELLA gravity settler takes
less than an hour, compared with the detention time
of 6 hours in conventional activated sludge. This
compact system also provides another benefit, due
to its compactness, the treatment system in the
North Bergen treatment plant is housed in a build-
ing which protects the rotoscreen, RBC, and LAMELLA
gravity settler equipment from extreme temperature,
viariation, neavy rain and high wind. The con-
struction cost of RBC using shallow tanks (6 ft.
above ground) versus activated sludge tank with 15
foot depth underground is relatively lower as the
water in the activated sludge tank exerts consider-
able pressure on the soil, thereby requiring costly
pile foundation and dewatering, especially, if the
soil conditions on site are bad. Tne total con-
struction cost of the North Bergen Central Sewage
Treatment Plant in 1980 was 9.5 million dollars.
An equivalent secondary treatment plant would have
cost at least 22 million dollars, (figures obtained
by using data on the Constructions Costs for
Municipal Wastewater Treatment Plants (1978) (6)
updated to 1980 dollars). The low operation and
maintenance costs, minimum land requirements and
reduced capital cost inherent in this flow-through
system show great potential in the treatment of
municipal sewage, in the near future.
692
-------
ro
'—
'CD
r-i
o
to
EH
cn
4-J
G .
CD
e
CD T3
SJ C
•H fO
.3
CJ1 0)
CD Cn
D 4-1
T3 rH G
C CO "3 ,
rd rH
i r*1 rrj C\ j
CD
t3 4-) 4-J
G fCS G
•H e
cn 4-J 4->
CD O (0
•H ' SJ
EH rH EH
fO
C G G
O O CD
•H -.H cn
4-> 4-1 ^J
G C CD
. CD CD 00
4J >
OJ G ^
Q 0 4J
CJ M
4H O
o c z
CD
C CD
O S
CO 4-1
•H CD
M CO
rd
•a,
e
o
CJ
CN
HH
• —
no
c
4-J fd
C 1-3
m
i-H
•
4-> C
C "rH
0) S
5 • —
4->
03
CD EH
M *
EH Q
C
0)
cn
M
CD
PQ
jG CO
-1— 1 -I-*
Sn -H
O C
EH
-H •
•P Q
O
rH
rd
C
O
-H
4-J
C
CD W
> 4->
C -H
O C
£
, o
4J
o
1-H
V( i — I
tO 4-> ^
S 4-J C
•H i SH
•H - rH 4-1 '4J
4-1 CJ ' E CCS 4-J
O P3 (0 J-l CD
oc; aj J o co
0 0
O O
Q 0
V - , -^
•sO -r>
CN ' r-'
• " ' . o
in
O
4J
o
VO O
ro CTs
rH •••
fl5 'TD SH
C CD CD
O 4-> - >i-H
•H tO CD V4 MH
4-i > cn rd -H
G ' -H T3 t3 Vi
CD 4-1 D C (0
> O i-H O rH
C
-------
SUMMARY AND CONCLUSIONS
The Hoboken pilot plant was operated on a
full-scale facility by using rotoscreens for
primary treatment. Rotating biological contactors
or trickling filter for carbonaceous removal, and
LAMELLA gravity settler for biological solid and
liquid separation. The results showed this
flowthrough system can successfully treat sewage at
the effluent concentration of both BOD and
suspended solids less than 20 mg/1. The biomass
sloughed from the media of rotating biological
contactor or trickling filter into the mixed liquor
were easily settled down on the LAMELLA gravity
settler without causing any problem on the settler,
such as biological growth on the plater of the
settler.
In stage analysis, the operation of the rotat-
ing biological contactor at two stages showd the
best results, either in the BOD removal efficiency
or sludge production, while comparing with 3, 4 or
5 stages.
Therefore, the concept of flow-through system
and two-stage rotating biological contactor was
adopted in the design of the North Bergen Central
Sewage Treatment Plant.
Prom the operation data at the first four
months, it showed the flow-through system was
successfully applied to the North Bergen Central
sewage Treatment Plant. The rotoscreen removed the
suspended solids from 24.9 to 33.6% averaging 28.9%
which is less than that of the primary settling
tank. However,, the effluent of the rotoscreen
would not effect the whole process. The average
BOD and SS concentration form LAMELLA gravity
settler runs 14.5 mg/1 and 18.9 mg/1 respectively.
Due to the short hydraulic detentioin time of 2 to
3 minutes at the rotoscreen the temperature between
influent and effluent would not make much
difference. This short detention time provides a
694
-------
great advantage in biological treatment, in which
the biological activity is temperature relative/.
during the winter weather. The ' low construction
cost of 11 million dollars versus 22 million
dollars for the 'equivalent secondary treatment
plant, low operation and maintenance cost, low land
requirements and low energy cost implied the
flow-through system a great potential to the
treatment of municipal sewage in the near future.
When the physical conditions are limited by
available land size, poor soil conditions, and high
g^roundwater levels, the authors recommend that you
should consider the application of this
flow-through system ' for municipal wastewater
treatment. '
ACKNOWLEDGEMENTS
We wish to extend special thanks to Seamus
Cunningham, Vice President of Mayo, Lynch and
Associates. His administrative assistance has been
instrumental in the conduct of this work. We would
also like to thank Robert Androsiglio, Peter Lynch
and Anna Lynch, Engineers of Mayo, Lynch and
Associates for their technical and laboratory help
for the duration of the project.
695
-------
REFERENCE:
1. Germain, J.E. "Economical Treatment of Domestic Waste by
Plastic - Medium Trickling Filters "JWFGF Vol. 38, No. 2,
February, 1966, PP. 192-203.
2. "Standard Methods for the Examination of Water and
Wastewater." 14th Ed., Amer. Pub. Health Assn., New York,
N.Y., 1975.
3. "Rotoscreen" oy Hycor Corporation, Bulletin No. 1101 1077,
1977.
4. Autotrol Corp., "Application of Rotating Disc Process to
Municipal Wastewater Treatment, U.S. Environmental Protection
Agency Water Pollution Control Research Series, Project No.
17050 Dam, Nov. 1971 . ' " '
5. Lawton, G.W. and Eggert, C.V., "Effect of High Sodium Filter
Slimes." Sew. and Ind. Wastes, 29, 1226 (1956).
6. U.S. EPA Technical Report "Construction Costs for Municipal
Wastewater Treatment Plants: 1973-1977" EPA 430/9-77-013,
January, 1978.
696
-------
AN IN DEPTH COMPLIANCE AND PERFORMANCE
ANALYSIS OF THE RBC PROCESS AT MUNICIPAL
SEWAGE TREATMENT PLANTS IN THE UNITED STATES
Robert J. Hynek. Manager of Process Verification,
Autotrol Corporation, Milwaukee, Wisconsin
Richard A. Sullivan. Manager of Process Engineering,
Autotrol Corporation, Milwaukee, Wisconsin
INTRODUCTION
A number of recent governmental reports, most notably the
Government Accounting Office (G.A.O.) report entitled, "Costly
Wastewater Treatment Plants Fail To Perform As Expected", have
severely criticized the performance efficiency of existing
wastewater treatment plants. The plants evaluated by the
G.A.O. were reported to employ either activated sludge, trick-
ling filter or lagoon unit operations for biological waste-
water treatment. Autotrol Corporation, as part of.a routine
customer service, regularly checks the operational performance
efficiency of Rotating Biological Contactor (RBC) plants where
it has supplied equipment. This paper statistically compares
the performance efficiency of those treatment plants evalu-
ated by the G.A.O. with RBC plants supplied by Autotrol Corpor
ation. This report was originally written early in 1981 and
further modified after evaluating an additional 22 P^nts
early in 1982. It should be noted that the data from 1980 and
1981 was statistically similar and 1981 data verified original
conclusions presented in the initial report.
In general, while the G.A.O. report stated that "E.P.A statis-
tical report on plant performance show that between.50 and 75/o
697
-------
of the treatment plants in operation are violating their per-
mits in any given time", Autotrol's study on performance of
existing RBC plants revealed that these RBC plants were meet-
ing their discharge requirements for BOD removal 89% of the
time and were meeting their discharge requirements for suspend
ed solids 92.5% of the time. The data further suggests that^
long term process failures (violations occurring more than six
months per year) is 500% more likely in alternative technology
evaluated by G.A.O. as compared to RBC plants evaluated by
Autotrol.
When discharge violations did exist in RBC facilities, the
vast majority of the violations were minor excursions from
standards. Where major violations occurred, the prime cause
was equipment deficiency, not treatment reliability.
DESCRIPTION OF THE STUDY
In November of 1980, the General Accounting Office reported
back to the Congress the results of its, investigation into the
operation of Publicly Owned Treatment Works (POTW) that were
designed, built, and funded under the authority of the grants
program. The report stated that at any given time 50% to 75%
of the plants are in violation of their National Pollutant
Discharge Elimination System (NPDES) permits. The report con-
tinues that a random sample of 242 plants in 10 states found
87% of the plants in violation of their permits at least one
month per year with 27% in "serious" violation.
Autotrol Corporation decided to statistically analyze and com-
pare the results of its RBC facilities to those analyzed by th
G.A.O. Important to the confidence of the reader is the know!
edge that all the data displayed and the only data we would us
is that gathered by the municipality and reported to its appro
priate state agency.
Although every attempt was made to make the comparison as simi
lar as possible, i.e.,
Autotrol and G.A.O. evaluated plants classified by
E.P.A. as capable of providing secondary or better
levels of treatment.
Autotrol and the G.A.O. evaluated plants based on
the issued NPDES permit.
698
-------
Some differences that did exist are as follows:
G.A.O. evaluated 242 plants from the universe of
676 plants available for evaluation (36%). Autotrol
evaluated 46 plants out of a possible 160 plants (29%).
The G.A.O. plants were "randomly selected". Autotrol
plants are those for which sufficient data was avail-
able to make a meaningful analysis and comparison.
The G.A.O. report selected plants of flow ranges between
1.0 mgd to 50.0 mgd. Autotrol plants had a flow range
of between 0.2 to 8.0 mgd. •; ,
G.A.O. evaluated data for a one year period between
1978-1979. Autotrol evaluated annual .data for 1980
and for 1981. -- ' -
The G.A.O. report evaluated data for BOD5, TSS and
fecal colifbrm. Because of the limited, amount of
data on fecal coliform, Autotrol's evaluations were
based on BOD5 and TSS only.
RESULTS.
The summary of plants evaluated, the violations observed, and
the number of monthly sample periods are described in Table I
for the 1980 data base and in Table II for the 1981 data base.
The following charts show the compilation of both 1980 and
1981 data for all monthly data reviewed and the resultant
successful compliance ratio.
PERFORMANCE CHARACTERISTICS OF RBC PLANTS
SAMPLE PERIOD 1980-1981
Number of Number of-..
Monthly Monthly
Samples Violations
—. BOD5
% Successful
Operation
All
Plants
Surveyed,.
598
66
88.96%
TSS
All
Plants
Surveyed
598
45
92.47%
699
-------
The results indicate that regardless of load and flow condi-
tions, plant operation, or potential lab analysis error, the
RBC plants evaluated were meeting discharge requirements 89%
of the time for BOD removal and 92.5% of the time for TSS re-
moval. This high degree of compliance is considered very good
particularly in light of the G.A.O. report statement,"E.P.A.'s
statistical reports on plant performance show that between 50
and 75 percent of the treatment plants in operation are viola-
ting their permit at any given time".
In an attempt to directly correlate the G.A.O. report on the
number of plants with discharge violations with the Autotrol
report on RBC facilities, Tables III, IV and V were developed.
Table III reports the number of plants evaluated in various
regions by the G.A.O., the total number of plants in violation
and the number of these violations per year of operation.
Autotrol developed similar comparisons for the 1980 sample
period (Table IV) and for the 1981 sample period (Table V).
In addition, because certain plants (#3, 18, 33, 44, 45) had
data for both 1980 and 1981, worst year data for duplicate
plants were used to develop a composite summary of RBC facili-
ties. The composite RBC plant performance is shown in Table
VI.
The following conclusions can be drawn from a comparison of
Tables III, IV, V and VI:
1.
The plants evaluated by G.A.O. demonstrated that 49.2%
of the plants violated discharge permits for more
than 6 months out of the year. Less than 11% of
Autotrol RBC plants had these extended violations of
more than 6 months per year. (Table III versus Table
VI) •
2. While 25.6% of the plants evaluated by G.A.O failed
to achieve successful operation for more than 3 months
per year, only 4% of the RBC plants surveyed performed
this poorly. (Table III versus VI).
3 While in excess of 87% of the G.A.O. plants experienced
at least minor violations, only 37% of the RBC plants
experienced similar difficulties. That is, 63% of^all
RBC plants continually met discharge limits month in
and month out. (Table III versus VI).
700
-------
4. While only 34.7% of the G.A.O. plants surveyed performed
satisfactorily for more than 9 months per year, the
RBC data showed that 82.6% of the plants performed
satisfactorily for periods in excess of 9 months per
year. . . " - ,
5. The data developed for RBC plants were conducted over
a 2-year period. The comparison between 1980 and 1981
data basis was statistically similar and repetitive.
6. The chances of having no violation in any year is 4
times greater using RBC plants evaluated by Autotrol
as compared to employing those plants evaluated by
G.A.O.
7 The chances of long term process failure (i.e. greater
than 6 months) is 5 times as likely in those plants
evaluated by G.A.O. as compared to Autotrol RBC facil-
ities.
DISCUSSION OF VIOLATIONS OF REG FACILITIES
(COMPOSITE 1980-1981 DATA
Of the forty-six (46) RBC plants evaluated, seventeen (17)
plants experienced violation of wastewater NPDES permits.
Twelve (12) can be classified as non-serious violations wnile
five (5) can be termed serious violations. The definition
of serious violation is as defined in the G.A.O. report, when
one or more of the three parameters was violated for more than
four (4) consecutive months during the review period and aver-
aged more than 50% above the permit limit during the period
of non-compliance". The definition of serious violation is
similar in Autotrol's evaluation with the exception being that
fecal coliform data was insufficient as a parameter to be
evaluated. The following discussions will classify the extent
of non-serious and serious violations encountered during our
survey.
NON-SERIOUS VIOLATIONS
Of the twelve (12) plants experiencing non-serious vio-
lations, nine (9) of those facilities had yearly average
discharge BOD5 and TSS values lower than their monthly
permit allowance. (In general, the yearly average BOD5
discharge .value was 75% of the monthly discharge require-
ment) . Of those plants that exceeded the discharge re-
quirement on a yearly average basis, a highest yearly
701
-------
average value of 37 ppm 6005 and a highest yearly average
TSS value of 32 ppm was recorded. In general, a review of
those plants experiencing non-serious violations indicate
that violations are minor, of very short term, and appear
to have been corrected in the most recent operations.
SERIOUS VIOLATIONS
Five (5) plants in our survey experienced serious viola-
tion (10.9%). This compares to the G.A.O. survey where
sixty-six (66) plants (27.3%) had serious violations. Two
(2) of the five (5) RBC plants experienced mechanical prob-
lems which resulted in poor performance. One of the five
facilities experienced industrial waste loads with inadequate
pretreatment which resulted in inferior performance. It is
not known whether the mechanical problems were caused by
design or equipment deficiencies. The following chart
describes the plant, reason for non-compliance, average
effluent BOD5 values and average effluent yearly TSS yearly
values for the survey year.
Plant
#11 WI
#24 OH
#41 IA
#34 WA
#22 WI
Avg,Yearly Avg.Yearly
Manor Category Eff. BOD5 Eff. TSS
Industrial Waste Overload 61 ppm 40 ppm
Equipment Deficiency 15 ppm 3 ppm
Equipment Deficiency 41 ppm 42 ppm
O&M Deficiency 42 ppm 19 ppm
O&M Deficiency 8 ppm 26 ppm
It should be noted that even though these plants experienced
serious violations, two of the five facilities still provided
effluent quality classified as better then secondary treatment
by the E.P.A.
CONCLUSIONS
The above comparative analysis indicates that RBC facilities
performed significantly better than facilities evaluated by
the G.A.O. Conformance was better in terms of both non-serious
and serious violation categories.
702
-------
Plant
No.
3
4
5
6
8
10
11
13
16
17
18
19
21
23
24
26
27
28
30
31
32
33
38
39
41
44
45
46
48
TOTALS
MONTHLY
Mo
State No.
PA
OH
IN
MI
IA
MN
WI
WI
KY
MI
CO
OR
MI
CO
OH
MI
WI
NY
WI
WI
KY
WI
WI
KS
IA
NE
OR
WA
WA
Compliance Ratio
TABLE I
NPDES VIOLATIONS
TBODC
_/
. Violation Per ,
of Mo. Surveyed
1/12
0/12
0/12
0/12
0/12
1/12
12/12
0/11
0/12
0/12
0/12
0/12
0/12
0/12
7/12
0/12
0/12
0/12
0/12
1/12
0/11
3/12
2/12
0/12
8/12
1/12
1/12
0/10
0/11
37/343
89.2%
DURING 1980
TSS
Mo. Violation Per
No. of Mo. Surveyed
0/12
0/12
'0/12
0/12
0/12
0/12
12/12
0/11
0/12
0/12
1/12
0/12
0/12
1/12
0/12
0/12
0/12
0/12
0/12
0/12
0/11
0/12
2/12
0/12
9/12
2/12
0/12
0/10
0/11
27/343
92.1%
703
-------
Plant
No.
3
5
9
12
14
16
17
18
21
22
25
26
33
34
35
36
40
44
45
46
47
50
TOTALS
MONTHLY
Mo.
State No.
PA
IN
WI
WA
IL
KY
MI
CO
MI
WI
MI
MI
WI
WA
SD
IL
MI
NE
OR
WA
OR
WA
•
Compliance Ratio
TABLE II
NPDES VIOLATIONS
TBOD
Violation Per
of Mo. Surveyed
1/11
0/12
6/11
0/12
0/12
0/12
0/12
1/12
0/12
0/12
0/11
0/12
2/12
11/12
0/12
0/12
0/12
4/10
0/12
0/12
0/10
4/10
29/255
88.7%
DURING 1981
TSS
Mo. Violation Per
No. of Mo. Surveyed
0/11
0/12
0/11
0/12
3/12
0/12
0/12
0/12
0/12
7/12
0/11
0/12
1/12
1/12
0/12
0/12
0/12
3/10
1/12
0/12
0/10
2/10
18/255
92.9%
704
-------
d
o
•H
4-1
cd
>
r^
o
•H
d
CO -H
!3
o co
H 0)
EH -H
jj -H
O rH
M -H
> 0
CO I — - .
1
• k • •
C3 LO
o> r^ ^cf i — 1 co
CM
PM
H •
o- • -
PM CD M
W rH 0)
Pi CM ,0
6 e
" Ctf 0
o co ia
O CM O CM O
o
-------
1
B
O
•rl
4-1
cd
rH
CO O
0 >
M
<3 *rl
h4
O CO
M 0)
> -H
4_)
PH -H
33 -rl
H 0
K cd
O F*i
H
P> i-3
M PU
W O
•J pa
S
H nJ
CO
,0
4-1
C
O
S
m
o
j_i
CU
Q
B
3
CN
! — |
1
o
1-1
in
o m H
vD
1
H G
O O rH O -H
4-1
>H -rl
1-4 rH
o o -a- tn -H
CO
1
rH
EH CJ
CTi rH rH rH |Z! Cd
CO CO p ]
-3-
W O
rH -* OT Pi
EH . ^
CO
4-J
G
O
S
MH
O
^
0)
o
B
3
CN
rH
1
O
rH
^Q LO *>^
. * * *
i— i - CTi i — |
CTi
1
• * *
rH <
f£) [V|
CO CO
O rH
oo ?-.
OH M
rH Cd
g
g
3
CO
4J
!H G
O o
P4 -H
QJ 60
Pi CD
Pi
i — 1
O
f-l
4-J
O
4-1
^J
<£j
0) OO >,
r> CT» J-l
•H rH Cd
4—1 S
cd g
CO rH 3
B^S 3 CO
P 6
(1) 3 4J
CO > CJ }H G
CO -H O O
O 4-10) p, iH
M cd CO 0) M
O rH M Pi 0)
rH
CO 3 G 0
3 6-S CJ M t-i
0 4-1
•H 0
rl 4J
cd 3
J> CJ
CO -H
O 4-10)
!H cd CO
O rH SH
< 30)
s >
CO 3 B
3 &-S CJ H
O
•H
^
cd
"
706
-------
H
£>
w
1-1
PQ
in
X— N
rH
00
O>
rH
c£J
O
00
o**
rH
1
S_|
w
§
CO
£5
H
CO
E-i
<1
p-l
O
pq
pn,
o
H
CO
0
O
O
C
O
•H
4J
cd
i-H
O
•H
£>
Pi
•H
CO
cu
•H
4J
-H
1 — 1
•H
CJ
cd
PT,
cn
4-1
S
o
^r|
m
o
r(
cu
to
6
j2j
CN
i-H
O
rH
on o co
^ ^ ^ ^
CO
m r^ oo
CT*
1
• * *
CO ^> CNJ O
CO i— 1
j/^l
LO i — 1 <1" \&
\o
1
• * * .
CO ^-D VQ 1^*- CM
CN rH 00
v_x
**O vD O
CO
I
i—l
• • •
o^ o^ o~* r^-
i-H rH CO
4-1
cn ^d
cd 4-)
a) pi
J£
4-J
•>
cd
e
3
CO
4-1
M C
O 0
cu M
pi cu
Pi
rH
O
4-1
O
4-1
3
a)
•rl
1 1
cd '
CO rH
B^ 3 -
t= s
0) 3
cn > cj
cn -H
O 4-1 CU
M cd cn
a rH M
< 3 cu
e >
CO 3 fl
3 fr-S O M
O
•H
$-!
cd
707
-------
THE USE OF PLASTIC MEDIA TRICKLING FILTERS
TWO CASE HISTORIES
Felix F. Sampayo, Jones & Henry Engineers, Limited
Toledo, Ohio
INTRODUCTION
Over the last ten years, Jones & Henry Engineers has
investigated the use of fixed film biological processes at
a number of locations in the Great Lakes area.
Approximately half of these studies have been directed
toward the treatment of high strength waste and half
toward the production of nitrified effluents. In some
cases, the investigations have included rotating
biological contactors and trickling filters.
708
-------
This paper discusses two case histories on the use of
plastic media trickling filters. One of the case
histories deals with the treatment of high strength wastes
at Kalamazoo, Michigan and the other with nitrification of
the Lima, Ohio secondary effluent.
THE KALAMAZOO CASE HISTORY
The City of Kalamazoo, Michigan conditions sludge by
wet air oxidation.^ The conditioned sludge is thickened
in decant tanks, dewatered by vacuum filters, and
incinerated. The residue ash is landfilled. The decant
tank supernatant and the vacuum filter filtrate are
recycled to the head of the plant. Although the recycle
streams represent less than 1 percent of the plant flow,
they constitute 24 percent of the total organic load.
An alternative to direct recycling is treating the
supernatant separately. Jones & Henry Engineers tested
several processes for possible separate treatment of the
wet air oxidation recycle streams as part of the design
for advanced wastewater treatment. The practicality of
various separate treatment processes was ascertained
through desk top, bench scale, and pilot studies. Process
effectiveness was judged on BOD reductions. Process
viability was determined using additional paramaters
including color removal, odor control, process
reliability, operational simplicity, space requirements,
and economics. Bench scale studies of activated sludge
and physical-chemical methods showed these processes to be
ineffective. Pilot plant investigations demonstrated that
attached growth reactors would best be used for
pretreatment of the recycle streams. Design information
was developed during the tests that substantiated and
expanded previous research.
709
-------
CHARACTERISTICS OP RECYCLE STREAMS
The supernatant and filtrate have essentially the
same physical and chemical properties. The coffee-colored
liquors have an average temperature of 120°F (49°C), a
noticeable odor, and strong frothing tendencies.
Standard chemical analysis of wet air oxidation
by-product samples showed the liquor is acidic and
extremely rich in nitrogen, with substantial amounts of
chloride and sulfate and low levels of phosphorus and
suspended solids. The waste has high BOD, COD, and TOC.
Specific chemical characteristics are detailed in Table I.
THE STUDY
A plastic media trickling filter was pilot tested on
the supernatant of the wet air oxidation decant tank. The
unit was operated for nearly six months (December 17, 1974
to June 2, 1975). The temperature and oxygen demand of
the supernatant were controlled by dilution with ground
water. This was necessary as the temperature of the
recycle stream was too high for biological treatment.
The test parameters for the pilot program were total
and soluble BOD and COD, and total and volatile suspended
solids. Color and odor were noted but not measured. The
parameters were monitored on the influent and effluent to
determine process efficiency at various organic and
hydraulic loadings, and dilution ratios. Samples were
collected seven days per week; one sample every two
hours. Analyses were performed on daily composites of
grab samples.
Figure 1 is a diagram of the pilot facility. The
supernatant was diluted with ground water at the top of
the tower. This mixture flowed through a funnel to a
rotary distributor that controlled the hydraulic loading
to the filter. The effluent was collected in the
recycling reservoir at the bottom of the filter, and a
portion of it was returned to a tank at the top of the
tower. The constant level feed and recycle tanks were
equipped with outlets to ensure constant discharge. The
feed, recycle rates, and dilution ratio were varied to
study the filter's efficiency under different conditions.
710
-------
H
fa
O
o
o
co
o in
o , oo
oo r~
10-
VO
vo
o
00
oo
0>J
VD
o
M
u
'JS
m
P4
4J
W
H
~~* M — CJ ^^
1-1
O
in
o
o
o
oo
in
H
r-
m
vo
in
in
4J
CO
CO
M
TO
Q M 3
-^ M ^ J C
y * u O -H
JL> p coQ oo TO co
-O EH O M o
O .4 H-C
in EH 5 S o) ^ 78 b
M tHOTOQts>a—EH^-'TO'-'TO^'EH^—9^-'S~-'
711
-------
oo
CM
(-1
0)
•o
H Q)
a
CM
i-H
in
oo
10
§
§
!3 CO
§ &
3 $ HT
g 3 O m
EH o j
- ^-% "SO Q
H iH - O W O
w
Q CM Q O g
CM .-I « rM rtj i-H
W
cn
O>
Oil
m
i1
en — m —
712
-------
Heat Treated
Liquor
J
Feed Tank
Dilution
Water
Raw Waste Recycle
Pump Pump
p-
5
o
t
CD
0
CD
O
0
0>
tr
t
±
nip »
"A
^N
f^
o
cu
o
CD
t/>
CO
§
^
O3
oc
Effluent ^
r ^-
t •'.
Sampling Point
L
Plastic Media "
Trickling Filter
FIGURE 1
DIAGRAM OF PILOT FACILITY
_
713
-------
The pilot equipment included:
Pilot Tower Steel shell - 3 feet in diameter
- 30 feet high
Media - 21.5 feet high Surfpac
Media
Surface area - 7.07 sf
Volume - 153 cf
2 - 0.5 HP pumps for influent and recycle
1 - 1.5 HP pump for dilution
RESULTS
The filter removed less than 60 percent of the BOD
under all test conditions (Table II). The highest removal
of soluble BOD was 59 percent, achieved with low organic
and hydraulic loadings (85 lbs/1,000 cf/day and 407
gpd/sf) with a recycle ratio of 2.5. BOD removals in
excess of 45 percent were attained at the same hydraulic
load, but with higher organic loadings (282 and 469
lbs/1,000 cf/day) and a recycle ratio of 5.25. In
general, BOD removal percentages were reduced as the load
increased.
The highest BOD removals per unit volume of media
occurred at low levels of dilution with recycle ratios in
excess of 5.0. The low dilution levels had a higher
temperature which improved treatment efficiency and
minimized freezing problems; however, high recycle ratios
lowered the temperature. Therefore, if the process were
to be used, it would be necessary to achieve a balance
between recycling and dilution.
The performance of the pilot plant was characterized
by a mathematical expression similar to one proposed by
the National Research Council.4 The expression may be
used to predict the effectiveness of the unit under a wide
range of operating conditions or applied to full-scale
systems with varied loads and recycle rates. The formula
was not verified for high levels of treatment as removals
of BOD in excess of 60 percent were not attained.
714
-------
ra
rH
(0
§ §
o m
\ o T3
M O V.
H rH O
oo in co tx r~ oo in
OOrHrHOOOr-rH
inm'*ooincr»c>joj
rH CN
I
EH
4J
§§
dg
t-l
JJ
0> Q H
D O \
H ffl O
OH 03 ~~
CP> 00
S §5 g Q-H\O(0
8"e M WOrH-«IH rHrHCNrHCNCM1*
3 fa J rH O
a
H . -< & o i !••••••
H • fl) EH ininin^vor*
Ul -
US Q r-l V
rj!s g O.H ovocooooinin
OflJ \
O 0) pj K ca rH oj o csi c
P '-P S rH S i-H ^ 3
O
X
«
0) W
_4J\ OrHrHCNOppp
r^ id n3 ^j* vo vo ^*
fe (« ftl rH
715
-------
The performance of the unit at 20°C may be
described as follows:
E
SBOD
-0.249 + 0.0423 ^W/FB
where
ESBOD
W
V
F
R
= fractional efficiency of soluble
BOD removal
= soluble BOD loading to filter
(Ibs/day)
= volume of filter media (acre-ft)
1 + R
= recirculation factor =
= recirculation ratio
(l+R/10)
BOD removals during the investigation showed that the
process could be used as a roughing device prior to
recycling. The major difficulty encountered during this
study was the control of foam and odor. Extensive
facilities for foam and odor control would be required to
develop a viable process.
DISCUSSION
Treatment of wet air oxidation recycle streams using
plastic media trickling filters was not deemed to be
viable at Kalamazoo. The requirements for dilution water,
foam control, and odor control would have resulted in a
rather complex treatment scheme with a high potential for
creating nuisance conditions. The need to lower the
temperature would require either well water or high
volumes of primary effluent. The use of clean well water
would have increased substantially the flow through
subsequent treatment processes. The use of the warm
primary effluent would increase pumping requirements to
the towers.
The process selected for Kalamazoo was to continue to
return the wet air oxidation decant and filtrate directly
back to the wet stream treatment train.
716
-------
CONCLUSIONS
The following conclusions may be derived from the study.
1. Plastic media filters can be used to remove a
substantial portion of the soluble BOD in wet air
oxidation recycle streams.
2. The process could be used for pretreatment of wastes
before they are returned to the wet stream treatment
process.
3. The effluent from the process has high BOD and color
and is not suitable for direct discharge.
4. The use of plastic media trickling filters would
require a significant investment in temperature, foam,
and odor control facilities.
5. Treatment of waste oxidation recycle streams using
plastic media filters was found to be undesirable due
to the high potential for odor, extreme foaming
characteristics, and added complexity to the overall
, waste treatment scheme.
THE LIMA CASE HISTORY
The second case history deals with the use of plastic
media trickling filters for nitrification at the City of
Lima, Ohio Wastewater Treatment Plant. The pilot studies
leading to the design of these facilities have been
reported elsewhere^*3. iphe description of the plant and
early operating experience also have been reported
previously4. This paper summarizes the plant design
criteria and discusses operating experience for the five
years the complete plant has been in operation.
THE PLANT
The improvements to the wet stream facilities were
essentially completed in the fall of 1976. Upgrading of
the sludge treatment/disposal facilities was completed in
mid-1979.
717
-------
The plant is designed for an average dry weather flow
of 18.5 mgd and a peak of 53 mgd. The original design
concept called for the secondary and advanced treatment
portions of the plant to operate at a peak rate of 33 mgd
with the remaining flow receiving primary settling and
chlorination.
The upgraded plant includes screening, grit removal,
primary settling, aeration, final settling, nitrification
towers, chlorination, and phosphorus removal. Ferric
chloride and anionic polymer are used for phosphorus
removal. Sludge treatment and disposal includes gravity
thickening, anaerobic digestion, vacuum filtration, sludge
cake storage, and land spreading,. Normal sludge
treatment/disposal uses thickening, digestion, and land
spreading of liquid sludge. Vacuum filtration and sludge
storage followed by landfilling is used as backup to land
spreading. The design also includes a centralized
computer control system. The plant discharges to the
Auglaize River.
The design of the two 106 ft., diameter nitrification
towers was based on the results of pilot studies. The
media for the full-scale facility was supplied by
Goodrich. The basis of design, description of the
individual treatment units, and projected plant effluent
are shown in Table III.
718
-------
TABLE III
y OF LIMA, OHIO ••..-.
LTER TREATMENT PLANT
A AND DESCRIPTION OF PLANT .;
OH EH
W M
< &
13 0
rn
(H
CO
H
Q
cy Facilities: 33 mgd
53 mgd
Capacity and/or
ze Operating Conditions
(0 -H
•H W
4J «
VI -P
0) C
EH Ot
13 -B
•a
g1
o
ro
in
%%
-H -H
^ v
in vo
rH rH
C (0
•o 10 at
U.
in (0 >i
• "O vi
oo C n>
rH 0 S
0 -H
0) vi
•• ca ft
O _£ .C
rH en o>
O O ' " C^l
>1 VI VI *"
rH X! X!
•H H EH a
0 B o £
a> at ro
•6
e
0
vo
rH
O
x -a
o S
OJ
rH
o; •
ro
rH (>)
^~*
Ci
ca
c
•H
ca -
(0 ^
ffl CN
S CN
01
•P rH
•H
a
co ca co
•O "O "O
p. p. p.
Cn en Cn
o o o
o o in
^^ ^1 VD
•^ •*
H rH
"O "O "O
i^i"
o o in
ro ro oo
in ro rH
co
vo
O)
•»
CN
-
^•^
ca
G
EH
O^
C 'I i 'i ii *i-i
•H ca ca ca
Z» o ro o
4J 0 0 0
at 10 oo o>
ca «. «. «.
^i
a ^
g CN CN H
•rH
VI
*O 'O
(^ ^^
in o
oo ro
H ro
to ca
VI VI
rH O
rH
B
IH
O
o
in
0)
o
ro
^^
i°-
ca
1
EH
C
O
•H
4J
(0
VI
J!
T3
at
•H
1
(0
§
XJ
rH
VI
•rH
0
o
VO
H
' O
0
ro
ON
ro
^•>
2- S
CO O
vi CQ
a>
S o
O o
rH rH
m -
0
O
-H
-------
u
J
O '-P
\ -rH
•§"§
m o
t>i
4-> &>
••H C
m'-P
Pi (0
CJ OJ
o
<4H MH
ra ra
•O 13
CF* D^
«tf in
•O T3
CTI Ol
o in
CO 00
CO rH
ra
UH M
ra EH
S ra
pi JQ
CO CO
r- i-H
0 0
Cg; OJ
•a TJ
01 Dl
g g
in in
CO 00
rH rH
<4H
to
n EH
e ra
&X!
O^ rH
0 CM
CO CO
rH O
<2» <3f
T3 'O
tJi CJ*
e g
0 O
CO CO
CO CO
inutes contact
e
<*
CM
PM
rH
i1
in
CM
rH
s
CM
o
o>
f
rH
ra
•H
M
a
MH
ra
\
ra
X)
't
rH
CM
0)
If
3 rH
H m
ra
m m
0 -O
•g§
Q) U
r; d)
o ra
*W 'II
CO CQ
'^s^ 'v.
W CO
j3 ^Q
H H
^* CT^
CM ro
§
-H
4J
O
,|,J
V
"O
to
m
•o
rH
CM
rH
m
c>
4J
UH
O
o
o
o
CO
o
i-H
TABLE III
(Continued)
1
ra
••*
rH
X
a) m
N -rH
•H -O
Oi
o>
«
in
rH
CM
m
•rH
•O
CO
in
rH
JS
o
4J
14-1
O
o
fc
1^.
CO
A
O
m
o>
X!
CT>
0
rH
CM
A
U
m
A
a
0
rH
CM
rQ
fi.
rH
X
•
m
•rH
•o
w
in
t3
§
CM
CM
X
•
n}
•rH
•O
m
in
00
•O
>
ra
CM
CM
^
•
m
•H
•o
^
in
00
T3
S
ra
CM
CO
X
•
m
•H
•O
«.
O
r»
CM
CM
u
CM
-P
•H
c
P
U
1
•H
53
4J
0>
r- )
jj
•H
tion Towers
m
u
-rH
IW
-rH
[ 1
•rH
Contact Tan!
(U
c
•H
H
o
rH
O
s Removal
3
1-1
O
Oj
ra
O
Fu
ra -N
& 2.
§ c7
oi
rH CO 6
m rH J>|
o o H
•rH Q) O
g fe CM
0)
ja
0
ickeners (2
X!
EH
0)
tJ>
•0
3
rH
U)
Digesters
0
•rH
XI
o
o>
m
c
rH
??
^i m
H *O
m c
s o
•rH O
I-i 0)
CM tA
ra
1
EH
C
•H
•O
H
O
Hi
0)
•o1
3
rH
W
720
-------
M!
O 4J
\ -i-l
;|O *O
* o
o
4J cn
•H G
O -H
(0 -P
Q< (0
(0 M
JS
B
o
CO
O
O
o
o
0)
1
CO
II
10
•H
•o
CO
X
•H
•o
in
•o -o
I I
o o
ro ro
•O
cn cn
B S —
Cn
e
in in
• •
00 CO
rH rH
(!)
Cn
(0
M
§
(0
ki (0
C O
e; — ~--
ro
CO
fa
O
.
-------
THE NPDES PERMIT
The plant operates under a National Pollutant
Discharge Elimination System (NPDES) Permit issued on
January 21, 1981 that expires on January 20, 1985. The
Permit establishes limitations for three different flow
conditions. The pertinent limits of the Permit may be
summarized as follows:
At 18.5 mgd
At 33.0 mgd
At 53.0 mgd
30 Day 7 Day 30 Day 7 day 30 Day 7 Day
Parameter
Suspended
Avg. Avg.
Solids
BOD=
(mg/1)
(mg/1)
14
20
13
20
14
Avg.
30
19
Avg.
30
20
1,000 2,000 1,000 2,000
Fecal
Coli-*
form/
100 ml
NH3-N
(Summer)
NH3-N
(Winter)
P (mg/1)
*Summer
Only
The Permit also requires that the D.O. of the
effluent be not less than 5 mg/1.
The winter NE^-N limitation in the Permit is more
restrictive than the plant was designed for.
1.5
722
-------
OPERATING CHARACTERISTICS,
OPERATION PROBLEMS, AND CORRECTIVE MEASURES
The nitrification facilities were put on line in late
summer of 1976. Operation was interrupted shortly
afterwards to repair damage to the plastic media in one of
the towers. Damage resulted from mechanical failure of
one of the distributor arms.
The facility began nitrifying in about eight weeks and was
producing the expected effluent values by early November.
The time required for the start of nitrification was about
the same as in the pilot studies.
Operation of the nitrification facilities has been
remarkably free of problems. Early on, the towers
operated at almost 100 percent recirculation with no
attempts to optimize recirculation rates. For the last
couple of years, recirculation has been set at about 11
mgd with the rate varying inversely to flow.
The operational simplicity of the system is greatly
appreciated by the plant personnel. The operator simply
reviews the computer printouts in the morning and makes
any necessary adjustments. The plant personnel claim that
operation of the towers and recirculation system "take
about a minute a day".
The towers sloughed off solids late in the summer of
1979 for a period of approximately two hours. No decrease
in process efficiency was reported following slough-off.
No noticeable sloughing of solids has occurred since.
The towers have experienced no significant operating
problems during about 5.5 years of operation other than
icing two or three times during the winters of 1977 and
1978, two of the coldest winters on record for the area.
During these occurrences, ice built up along the filter
walls and stopped the distributor arms. The operators
broke the ice and the towers were put back into operation.
At the beginning of the winter of 1979, operating
personnel capped the end nozzle in each of the distributor
arms, eliminating ice formation from splashes on the
walls. No icing problems have been experienced since then.
In late October 1981, the bearings in the
distributing mechanism in one of the towers failed.
723
-------
RESULTS
The results for BOD, suspended solids, ammonia, and
dissolved oxygen during the first three full years of
treatment facilities operation have been reported
previously and are summarized in Table IV. The results
for 1980 and 1981 are presented in Tables V and VI. Raw
wastewater and air temperatures for the same period are
shown in Table VII.
The tables show that the quality of the effluent is
consistent and generally better than required by the NPDES
Permit. The one instance of high NH3-N in February 1980
has been attributed to analytical error. At the time the
plant was experimenting with different analytical
methods. The high suspended solids during January,
February, March, and April of 1980 were due to the plant's
inability to dispose of sludge with the consequent high
sludge inventory in the aeration system.
The flows shown in Table V are total flows to the
plant and include a small portion of the flow that
received primary treatment only during January, February,
March, and April of 1980. The actual average flow through
secondary treatment and nitrification for these months are:
January
February
March
April
13.14 mgd
10.64 mgd
22.98 mgd
18.80 mgd
Beginning with May of 1980, the computer program
governing storage in the sewage system began operating
properly. Storage in the system has reduced peak flows to
the plant and practically eliminated the need to bypass
partially treated wastes.
From 1979 through early 1981, plant personnel took
approximately four measurements per month of TK-N in the
nitrification towers influent and effluent. The average
for these values is shown in Table VIII. Also shown in
the table are the loading to the tower (Ibs TK-N/sf/day)
and the resulting effluent NH3~N concentration for the
years 1979, 1980, and part of 1981.
724
-------
s
en
•
en
ro
•
i-l
C !3 rH
O
M
g
en
r^
en
rH
r~
r»-
en
8
H
tl
03
M
3
H
m s
rH
fa
en in
o
ro
in
en
~ tJ Q| oo in
i1^
Q
5
m
rH
oo
o
vo
00
rH
in
AO
rH
in «>
01 r~-
o in
rH rH
10
00
o
• ,
vo
VI
0}
0)
en
00
en
en
r-
cn
725
-------
o
CO
4J -^
C J3 r-1
0) T \
3 ro O>
CC :
U CO ' .
-i Q —
0) O Oi
> « e
ril **•*
tncovowvocnovovocococv)
r-l r-( CT> ' O CN CM CO
i-liH rH rH • H i-t rrl H
o>
o> >J
•H (01
w Q|
oooooivoncot^oinroroo
caconocancqvooorHirii-i
rHoooincoinvorHcrioocTico
S
i
« • • • Q) ^ • 4-J • • •
c.QMM>ie.-ioo
726
-------
Q g i-HrHi-ii-Hi-ir-Hi-Hr-li-Hi-Hi-Hi-l
C 3 H CT»co«4'-if)ir)r»ifN)csr>jvoinro
DCOOIOOOOOOOOOOOO
MH J3 »-•
rH H •*'*«*inmo,ror-ooo\r-co
£j -H g OOOOOOOOOOOO
g .fe —
d. - « • • •
S1
S ss
g s
EH M
!Z K
SI H
O to O §
H ftf prt
}^rj P^ p^ *^
H • •• s o
H H H S jj
g . ^ g |; ! ' « p, rt
-------
TABLE VII
CITY OF LIMA, OHIO
WASTEWATER TREATMENT PLANT
RAW WASTEWATER AND AIR TEMPERATURE (1980-1981)
Avg. Raw Wastewater
Temp. (°F)
January 1980
February
March
April
May
June
July
August
September
October
November
December
January 1981
February
March
April
May
June
July
August
September
October
November
December
47
45
42
49
58
64
70
73
71
63
54
52
48
48
50
56
59
66
72
74
66
60
56
50.9
Air Temp. (°P)
Avg. Max. Avg. Min.
31
31
44
60
75
79
86
85
79
63
46
33
31
43
48
64
64
78
80
77
68
57
45
30.5
21
17
29
39
51
58
67
68
58
41
32
23
17
24
30
44
48
59
64
58
54
40
33
21.9
728
-------
en
-P
H
\
HrOi-HOHi-HHOOOrHp
oo
o
CO
sf
If
a
§
O ft
M
SI
H fa
t-3 «
Em
° w
H rf
H 3:
O Kl
*O y£5
W I KM
o « m
**;0
vi to >w
0) XI W
£ H '
OHi-IOrHOOOOOOO
oooooooooooo
,
c
•8 SB
S H
O M-i !S
EH C I
M M
in
CM
a a s: r-t
D VI O -H
0) (U
!i
3 -P O 0) <1>
cn a-u > o
3 0) U O 0)
«C CO O 53 Q
729
_
-------
H
OinPOOi-Hi-HrHOOOOO
o en
« •
r-l O
H
01
SB
in 10
• •
CM CO
w •""• 'ri
a?
5 i
o s*s
i4 E-i TO
M W . -
a xi w
o o
^** r i
OOr-IOOOOOOOOO
OOOOC3OOOOOOO
rf oo
o o
4-4
a o>
coir>oi«ijj
CM CNJ
• •
CT> H
m
u>
10
M
0)
IM
i
OO
o -H
.<3 Q)
0)
to
0> &
js
0)
4J
0) (U
00
M n)
(0 3
> O
3 3 3 0) O O (1)
§
•d
-p
0)
730
-------
SLUDGE PRODUCTION
The nitrification towers are not followed by settling
tanks, therefore no sludge is collected. The pilot
studies leading to the design showed a sludge collection
system would not be required. The findings of the pilot
studies have been largely confirmed during operation. The
towers have sloughed off solids noticably only once for a
period of about two hours during about five and one-half
years of operation.
IMPACT ON RIVER WATER QUALITY
Table IX shows the BOD and D.O. concentrations
measured upstream and downstream from the plant in 1980
and 1981. Downstream measurements are taken approximately
400 yards below the plant discharge. It is evident from
the table that the plant effluent has essentially no
impact on the Auglaize River water quality.
OPERATION AND MAINTENANCE COST
Operation and maintenance costs averaged $142.55 per
million gallons in 1978, $138.93 per million gallons in
1979, $178.50 in 1980, and $244.82 in 1981. Itemized
operation and maintenance costs for the last two years are
shown in Table X.
The plant operator reports that practically no
manpower is required to operate the nitrification towers
and appurtenant pumping-station. For all practical
purposes power and maintenance are the only operating
expenses associated with the towers.
731
-------
H
EH Pi
O el r-HHr-IHr-lr-(rHr-)rH
i-li-(H. HHH
0)
,-, M
H C
00 5
I O
O Q
CO
Q
8
Q SI rH r-l rH
i-Hr-lHi-Hr-HrHi-Hr-)
0)
U
(0
o
00
en
3 X!
i-l O
fl)
w a) ja e
>1 3 4J Q g
*33(UOO
h) >i
e "B a
-------
a e
CO CO ON
o CM en o co co
rH rH rH i-H i-H
0)
CQ
Q
inrH
Q
CN in co in «>
in ^i* in in in
I
&
O
D a a) u o
fe
J-.
4J
»
cr>
C
•H
•H
0)
CO
0)
rH
a.
CQ
733
-------
TABLE X
CITY OF LIMA, OHIO
WASTEWATER TREATMENT PLANT
OPERATION AND MAINTENANCE COST (1980 AND 1981)
Item 1980 1981
Payroll
Power
Chlorine
Chemicals**
Miscellaneous
$460,590.42
196,546.12*
5,208.38
102,870.05
162,265.17
$927,480.14
$ 510,036.08
191,114.18
3,771.59
179,192.08
172,260.34
$1,056,374.20
* The cost for power averaged 0.022/KWH.
** Ferric chloride and polymer.
DISCUSSION
Figures 2 and 3 show the data derived from the pilot
studies and the operating results for 1979, 1980, and part
of 1981; the only years that the plant has collected data
regularly on the TK-N concentration in the activated
sludge effluent. The results predicted by the pilot
studies have been confirmed under actual operation. This
strongly supports the concept of designing nitrification
towers on the basis of TK-N loads.
The nitrification efficiency of the total system meets
design expectations. For part of the year, the secondary
plant nitrifies well, as evidenced by the low TK-N in the
secondary effluent. During that time, the towers function
as polishing facilities. The towers always oxidize
substantial amounts of TK-N. The additional TK-N
oxidation in the towers results in a very stable effluent
with very low TK-N values. For the twelve-month period of
March 80 through February 81, the TK-N in the effluent
averaged only 2.65 mg/1. The organic Nitrogen in the
effluent averaged 1.47 mg/1.
734
-------
DC
01
5
ro
D
CO
O
1
Q
o
(O
CJ
in
(l/Bua)
JBMOJ. u| |\|-£HN
735
-------
O —
= 3
CL U.
0<
00
at
o
CO
o>
Q
O)
C
CO
111
oc
D
cc
LLI
N
Z Q
=?£
col-
X Z
2
C)
CO
q
CO
q
in
o o o o o
(I/Bui)
49MOj_ U| |\i-£|-|l\l
736
-------
The tower effluent is always better than the
influent. Even though substantial nitrification occurs in
secondary treatment, the towers are needed to consistently
meet NH3-N effluent requirements.
The towers are not generally loaded to their design
capacity. Plots of TK-N applied (Ibs/sf/day) vs. TK-N
oxidized (Ibs/sf/day) show a straight line relationship
indicating that loadings have not been the limiting factor
in TK-N oxidation.
During the summer months, the NH3~N concentration
in the plant effluent is about the same as for loadings
ranging from 0.01 to 0.15 Ibs TK-N/sf/day. During the
winter months, and over the same loading range, the
effluent NH3-N concentration increases substantially as
the TK-N load increases.
The plant has produced the desired results while
treating the highly .variable wastewaters generated by a
partially combined sewerage system. In any single month,
the flow can range from one third to almost three times
the design average. The monthly average for BOD in the
raw sewage has ranged from 53 mg/1 to 246 mg/1. The
monthly average for suspended solids in the raw sewage has
ranged from 83 mg/1 to 210 mg/1. It is unlikely that
single or two-stage activated sludge systems could provide
as reliable a treatment level under such adverse
conditions.
Stable performance is achieved with a minimum of
operational adjustment to the nitrification facility.
During 1981, the operators simply kept the recycle flow at
11 mgd.
CONCLUSIONS
The following conclusions may be derived from the
first five years of operation of the activated
sludge/nitrification tower process used at Lima:
1. Nitrification towers following activated sludge
consistently produce a high quality effluent with low
BOD, suspended solids, TK-N, and NH3-N, and very
high dissolved oxygen concentration.
737
-------
2. The treatment process is extremely reliable and able
to withstand shocks.
3. The plant effluent is normally saturated or super
saturated with oxygen. The combination of high D.O.
and BOD results in an effluent that exerts little or
no oxygen demand on the receiving stream.
4. The towers are easy to operate.
5. The performance of the full-scale facility has
confirmed the design criteria derived from pilot
studies. Very low ammonia concentrations can be
obtained even during the cold winters experienced in
midwestern United States.
6. Final settling following the towers has not been
necessary as the effluent contains very low suspended
solids.
738
-------
REFERENCES
1 Sampayo, Felix F., Hollopeter, Dail C., "Treatment of
Wet Air Oxidation Recycle Streams", Proceedings 7th Annual
WWEMA Industrial Pollution Conference, June 1979.
2 Sampayo, Felix F., "Nitrification Studies at Lima,
Ohio", 47th Annual Meeting of the Ohio Water Pollution
Control Conference, Columbus, Ohio, June 1973 (Conference
Reprint).
3 Sampayo, Felix F., "The Use of Nitrification Towers at
Lima, Ohio", Second Annual Conference, Water Management
Association of Ohio, Columbus, Ohio, November 1973
(Conference Reprint).
4 Sampayo, Felix F., "Nitrification at Lima, Ohio",
Proceedings International Seminar on Control of Nutrients
in Wastewater Effluents, September 1980.
739
-------
PART VII: NITRIFICATION AND DENITRIFICATION
NITRIFICATION OF A MUNICIPAL TRICKLING FILTER
EFFLUENT USING ROTATING BIOLOGICAL CONTACTORS
Frederic C. Blanc. Professor, Department of Civil
Engineering, Northeastern University, Boston, Mass.
James C. O'Shaughnessy. Associate Professor, Depart-
ment of Civil Engineering, Northeastern University,
Boston, Massachusetts.
Charles H. Miller. President, Haley and Ward, Inc.,
Waltham, Massachusetts
John E. O'Connell. Project Manager, Haley and Ward,Inc
Waltham, Massachusetts.
INTRODUCTION
This paper presents the results of a pilot plant inves-
tigation and design for achieving seasonal nitrification of
a secondary trickling filter effluent. The municipal treat-
ment plant in Milford, Massachusetts discharges into the
headwaters of the Charles River, a Class B stream. Conse-
quently, the Commonwealth of Massachusetts has required a
nitrified effluent, phosphorus removal, and effluent filtra-
tion. Table 1 lists the present and proposed discharge
limitations formulated by the U.S.E.P.A. and the Massachu-
setts Division of Water Pollution Control.
740
-------
Table 1. Present Discharge Limitations
Effluent Characteristic
Monthly
Average
Maximum
Day
Flow (MGD)
Biochemical Oxygen Demand
5-day 20° C
Total suspended solids
Settleable .Solids
Fecal Coliform Bacteria
Total Coliform Bacteria
PH
4.0
30 mg/1
30 mg/1
200/100 ml
1000/100 ml
50 mg/1
50 mg/1
0.3 ml/1
400/100 ml
2000/100 ml
Shall remain between 6.0 and 9.0
Proposed Discharge Limitations
Effluent Characteristic
Monthly
Average
Maximum
Day
Flow
Biochemical Oxygen Demand
5-day 20° C
Total suspended solids
Settleable Solids
Total Ammonia Nitrogen
Total Phosphorus
Dissolved Oxygen
Fecal Coliform Bacteria
PH
4.3
7 mg/1 11 mg/1
7 mg/1 11 mg/1
0.1 ml/1 0.3 ml/1
1.0.mg/1 1.5 mg/1
1.0 mg/1 1.5 mg/1
Not less than 6 mg/1
200/100 ml 400/100 ml
Shall remain between 6.0 and 9.0
After preliminary evaluation of alternate methods of
achieving nitrification, rotating biological contactors were
selected. A pilot plant study was conducted during the late
summer and fall of 1977 to determine both the performance of
such a system in nitrifying a trickling filter effluent and
the design and operating parameters. Specifically the ob-
jectives of the pilot plant operation were to determine
various design and operational parameters of the RBC process
741
-------
with regard to nitrification, including loading rates,
removal rates, removal efficiencies, solids production, and
solids settling characteristics.
Existing Wastewater Treatment Plant
The existing sewage treatment plant consists of primary
sedimentation, raw sludge dewatering, secondary treatment by
high rate trickling filters with a varying degree of recir-
culation, final clarification, and chlorination of the efflu-
ent before discharge to the Charles River. Flow enters the
plant through two pipe lilies, each equipped with a Parshall
flume and its attendent flow recording equipment. One line
receives the discharge from the Charles Street pumping sta-
tion through .variable speed pumping equipment, and that dis-
charge is therefore responsive to inflow variations. The
other line is the main gravity flow outfall from that portion
of the -collection system not draining to the Charles Street
pumping station, and is equipped with a comminutor in addi-
tion to the flow meter.
Grit removal is accomplished in the primary settling
tanks along with scum and settled sludge removal. The com-
bined grit and raw sludge are then dewatered by vacuum fil-
tration, \\fith the aid of lime and ferric chloride as sludge
conditioners. Dried sludge, with a pH of 11 to 12, is dis-
posed of as land fill in the discontinued sand filter beds.
The effluent from the primary settling tanks flows through a
flow control chamber to 3 trickling filter rotary distribu-
tors which uniformly disperse the wastewater onto stone
media filters.
Trickling filter effluent is pumped to the final clari-
fier and a proportionate amount of this effluent is.recircu-
lated to the trickling filter inflow, in order to provide
the hydraulic quantity necessary for proper operation of the
filter distributors during periods of low flow. Effluent
from the final clarifier is subsequently chlorinated in a
chlorine contact chamber before discharge.
The wastewater flow at Milford is predominately domestic
in character, and presently contains very little so-called
"industrial" wastewater. Analyses of the raw wastewater
reinforce this conviction. It is anticipated that this
situation will not substantially change in the future. Fur-
thermore, any future industrial wastes, which might affect
proper treatment plant operation will be required to receive
adequate pretreatment at the point of origin prior to their
acceptance by the municipal treatment facility.
742
-------
Plant records evaluated over the past five years in-
dicate the average daily flow to the treatment facility is
2.4 M.G.D. and the average influent BOD-5 is 140 mg/1.
Effluent BOD-5 averages 25 mg/1., which corresponds to a
BOD-5 removal rate on the order of 82%. The facility serves
a population of approximately 22,500.
Pilot Plant Description
During the late summer and fall of 1977 a pilot plant
was operated: on the Milford Wastewater Treatment Plant site.
Figure' 1, illustrates the flow scheme used for this pilot
plant operation.
As shown in Figure 1, the clarified trickling filter
effluent 'was pumped from the effluent trough of the clarifier
to the two RBC units by individual submersible pumps. The
rate of flow to each RBC unit was controlled by a valve
following each pump.
The RBC unit consisted of a semi-cylindrical tank 4.23
feet long and 4.0 feet in diameter which was divided into
four equal-volume compartments by means of 1/4-inch steel
partitions. Twelve 47-inch diameter rotating polystyrene
discs per compartment were mounted on a center drive shaft
so that approximately 29% of the .disc surface area was sub-
merged in the tank contents. Each unit contained a total of
1570 square feet of surface area and a net tank volume of
120 gallons with the discs in the tank.
Connections between the compartments or stages permittee
four-stage series operation. These connections were made by
means of external pipes which by-passed the compartment par-
titions on the outside of the tank. The wastewater therefore
flowed through the unit perpendicular to the center shaft.
Alkalinity additions were, made by means of peristaltic
pumps from two 55-gallon tanks to the first stage of each
RBC unit. The chemical solution was kept mixed by means of
a submersible pump in the bottom of the tank.
The effluent from the RBC unit then flowed by gravity
to a settling tank, where solids were permitted to settle
out and' the overflow was discharged to the control chamber
adjacent to the existing secondary clarifier.
Pilot Plant Operation
After a two week startup period in the beginning of
August 1977, the pilot plant acquired a reasonably full
growth of biomass. From that time until December 1977
743
-------
1-
5
o
m
o:
*
I
»
I
N
i
w
I
t
UJ
UJ
<
S g
i i
o
'O
u.
cc
o
UJ
744
-------
performance data was measured. .
It has been stated in the literature concerning nitri-
fication that any biological process for the conversion of
ammonia to nitrate will require a source of alkalinity in
the amount of 7.14 mg/1 of CaCO for each 1.0 mg/1 of
NH -N oxidized. (I, 2, 3). The RBC process is no exception
and since the influent to the RBC's contained an average of
only 51 mg/1 as GaCO-, an external source of alkalinity was
necessary to insure complete nitrification of the ammonia
present.
Alkalinity was supplied to the units in the form of
either bicarbonate of soda, soda ash or lime. Efforts were
made during the pilot plant study to maintain a pH of 7.0
to 8.5 ^nd an alkalinity of greater than 200 mg/1.
Sampling was arranged so that observations could be -
made under various conditions and at different times during
the day. A 24-hour sampler was utilized a number of times
to monitor the characteristics of the influent to the RBC
units during the course of a day. In addition, the regular
sampling procedure for monitoring the performance of the
units was arranged to provide data of both morning and after
noon operation, because the lowest influent concentrations
of NH--N were .observed between 5 A.M. and 10 A.M. and the
highest between 3 P.M. and midnight.
Operational Results
Table 2 illustrates the typical unchlorinated trickling
filter effluent for summer operation which served'as the
influent to the RBC pilot plant.
Table 2. Typical Unchlorinated Trickling
Filter Effluent
Parameter
Typical Value
TKN, mg/1
NH3-N, mg/1
NO -N, mg/1
•D
, mg/1
Suspended Solids, mg/1
Volatile Suspended Solids, mg/1
10
6
9
25
47
38
745
-------
Parameter
pH *
Alkalinity (as CaCO ) , mg/1
Typical Value
6.0 to 7.0
51
* Values are before adjustment by chemical addition for
pilot plant operation.
Due to variations in the strength of the raw wastewater
entering the plant there was variability in the trickling
filter effluent quality. Table 3 depicts the variation in
ammonia nitrogen concentration with time for three twenty-
four hour periods.
Table, 3. RBC Influent Concentration Variation
Over a 24 hour Period
Aug. 11-12
Time ' NH mg/1
Aug. 14-15
Time NH mg/1
o
Aug. 16-17
Time NH, mg/1
11:30 AM
12:30 PM
30
30
3:30
30
30
6:30
7:30
8:30
9:30
10:30
11:30 PM
12:30 AM
30
30
30
4:30
5:30
6:30
7:30
8:30
9:30
8.4
9.8
11.0
8.8
8.4
7.8
7.7
11.3
9.5
12.0
10.0
9.4
9.4
9.2
8.0
6.5
6.6
7.0
10.2
4:30 P
5:30
6:30
7:30
8:30
9:30
10:30
11:30 PM
12:30 AM
:30
:30
:30
:30
5:30
6:30
7:30
8:30
9:30
10:30
11:30 AM
12:30 PM
M..
1:
2:
3:
4:
8.8
12.0
9.4
11.3
7.7
10.0
8.3
8.9
8.9
7.4
9.8
8.2
7.4
7.0
6.9
6.8
6.3
6.6
11.0
7:30 AM
8:30
9:30
10:30
11:30 AM
12:30 PM
:30
:30
:30
1:
2;
3:
4:30
5:
6:
7;
:30
:30
:30
8:30
9:30
10:30
11:30 PM
12:30 AM
:30
:30
:30
:30
1:
2:
3:
4:
5:30
4.7
5.8
8.4
11.5
12.9
11.2
10.0
11.2
7.4
7.0
10.0
6.0
Note: For most loadings NH_ Ammonia Range 5 to 13 mg/1
746
-------
Analysis of the performance of the pilot plant was
begun after an adequate bibmass had been established on the
surface of the media. From this point and throughout the
course of the study it was apparent that the RBC units could
consistently produce an effluent with less than 0.5 mg/1 of
ammonia, providing that sufficient alkalinity was available.
The establishment of loading criteria therefore remainec
as the primary concern of the study. Figure 2 illustrates
the results of the analyses performed from mid-August througl
November, 1977, and indicates a linear relationship between
pounds of ammonia applied to the RBC units and pounds of
ammonia removed, for the range 0.04 to 0.4 pounds NH -N/1000
ft^-day. This linear relationship between pounds applied
and. pounds removed has been established in other studies.
(3) (4). ~ ' ' ; '
Figures 3 and 4 present the efficiency of the RBC units
with respect to ammonia removal. Figure 3 represents all
the data points while Figure 4 contains only those points.
for which the influent ammonia concentration to the RBC units
was greater than or equal to 6.0 mg/1. It is evident from
these efficiency plots that the removal of ammonia was
greater than 95 percent for loadings up to 0.2 pounds NHg-
applied per 1000 square feet of surface area-day.
In addition, for loadings up to 0.4 pounds NHL-N appliec
1000 s.f.-day, the removal rate was generally greater than
90 percent. r
During the course of the pilot study the temperature
ranged from 9° C to 19.5° C with the typical value being
16° to 17° C.
In this range, the temperature did not appear to have
a significant affect on the efficiency of ammonia removal in
the RBC units.
During the course of the study a number of sequential
sampling runs were conducted following an immediate increase
in the hydraulic load to simulate a peak flow. The sequen-
tial sampling was done in an attempt to trace a plug of
wastewater flow through the unit.
Figures 5 through 7 show the ammonia nitrogen concen-
tration levels through the pilot units for both pre-peak
equilibrium loading rates and peak loading rates. Table 4
indicates the ammonia loading rates for the same peaking
experiments.
At application rates lower than 0.2 Ib NH3-N/day-1000'
s.f. the amount of ammonia removed virtually doubled as the
application rate doubled. As indicated in Figures 5 thru 7
747
-------
O.4
0.3
o
I
o
o
o
tn
m
o
UJ
Q.
Q.
3-10 0.2
z
O.I
© UNIT I DATA
H UNIT 2 DATA
0 Q
©/©
O.I
O.2
O.3
0.4
NH, REMOVED, LB&/ IOOO FT. — DAY
O
FIGURE 2. NH3-N APPLIED VS NH3~N REMOVAL
748
-------
00
-Q
* 2
0 Q
Q
UJ
Q.
0.
a
0
©
/
/ Q
0
R
H/ Q 0
'
H/ Q
B0 © Q
E3 13©
© _
8
S
Q E
S
Q3AOH3H VINOMWV !N33U3d
O
CO
f
UJ
a.
a.
»
10
UJ
a:
o
q
"o"
749
-------
2 2
— CM
1- H
i §
0 Q
o/
B
© 0
O © E
O /
B/
/B
CO
o
X
UJ
o
UJ
o
o
o
•si
a
UJ
a.
a.
xw
UJ
UJ
cc
o
UJ
K
CM
'6
-------
10-
FIGURE 5 « TOR 9/27/77
lAK LOAD - 0.26 #/DAY-lOOOFT?
PEAK LOAD = 0.32 «/DAY — lOOOFT.
- AVE. LOAD * 0.13 «/D»— lOOOFT?
-Q ©
2 STAGE 3
m 5-
FIGURE 6' FOR 8/18/77
PEAK LOAD * 0.2»/DAY— lOOOFT.
AVE. LOAD * O.I «/DAY— IOOO FT?
2 STAGE 3
FIGURE 7 » FOR 9/21/77
PEAK LOAD = 0.42 «/DAY — IOOO FT?
ME. LOAD« 0.21 «/OAY —IOOOFT*
PEAK LOADING TEST RESULTS
751
-------
the effluent concentrations of ammonia were slightly higher
under peak conditions. With the exception of the run when
the peak loading factor was increased from 0.2 to 0.4 Ib-
NH,-N per day per 1000 ft.2 shown in Figure 7, the effect
of 4 stage operation attenuated decreases in removal effi-
ciency due to peaking. This was the only case where the
effluent ammonia concentration exceeded 1 mg/1.
Table 4. Effects of Peak Loading on NH,-N Removal
Q Detention NH -N Removed
(gpm) time (hrs) Applied „ #/day-
#/day-1000 ft. 1000 ft.
%
Removal
0.8'
1.6 *
1.1
2.2 *
1.3
2.6 *
1.3
1.85 *
1.3
2.25 *
0.8
1.6 *
0.9
1.85 *
1.0
1.9 *
2.5
1.25
1.82
0.91
1.54
0.77
54
08
1.54
0.88
2.5
1.25
2.22
1.08
2.0
1.05
0.1
0.22
0.16
0.33
0.13
0.25
0.1
0.26
0.13
•0.32
0.1
0.22
0.2
0.42
0.11
0.21
0.098
0.20
0.155
0.306
0.125
0.23
0.95
0.239
0.126
0.286
0.98
0.187
0.187
0.320
0.108
0.192
91
93
92
92
89
85
76
92
* Flow values from an instant flow peak and represent the
loading which gave the least efficient treatment.
Table 4, which summarizes the significant parameters
observed during the peak loading studies, indicates that
good removal efficiencies were maintained even when substrate
loading rates were doubled and hydraulic retention time
halved.
The results of the pilot study indicated that the eff-
luent BOD,, from the RBC units averaged approximately 8 mg/1.
752
-------
These samples were either collected from the 55 gal. drum
clarifiers, or samples collected from stage 4 .of the RBC's
and allowed to settle, for 30 minutes. The soluble portion
of the effluent was approximately 56% and averaged 4 to 5
mg/1 of BOD,.. The addition of alum to the last stage of
the pilot RBC units further reduced the effluent five day
BOD to an average value .of less than 2 mg/1.
The average suspended solids content of the influent
was 47 mg/1, the fourth stage of the RBC contained about
53 mg/1, whereas- the effluent from the units averaged 32
mg/1.., of which approximately 80 percent were, volatile.
For hydraulic loadings up to 1.80 gpm/s.f. (2.0 gpm)
there appeared to be little to no net solids production. The
settled effluent averaged 16 mg/1 of suspended solids and it
would appear that the suspended solids in the effluent from
a clarifier following the RBC can be maintained at under 20
to 30 mg/1 for low hydraulic loading rates. The solids
determinations during the pilot studies included suspended
solids from the existing tr.ickling filters which contributed
a relatively high percentage of colloidal matter.
Processes Selection and Design of the Wastewater Treatment
Facility
Four processes for the removal or conversion of ammonia
were investigated to determine their cost-effectiveness, in
accordance with EPA criteria. The four processes were as
follows: (a) nitrification using Rotating Biological Con-
tactors (RBC); (b) nitrification by aeration in a plug flow
reactor; (c) ammonia removal by breakpoint chlorination; and
(d) ammonia removal by selective ion-exchange.
Nitrification using RBC was selected as the lowest
total life-cycle cost process, although its capital costs
are somewhat higher than the other processes. In addition,
the overall advantages of process stability and flexibility
as well as substantially lower .operation and maintenance
other than occasional lubrication of the drives and it
requires significantly less attention during operation than
does the aeration system.
In order to utilize the existing facilities to the most
economical degree, and to keep the operation and maintenance
costs as low as practical, the present method of treatment
by primary sedimentation followed by trickling filters will
be retained. As shown previously, the performance of the
plant as of late has been of the highest quality to be
753
-------
expected of that ,type of process. In addition, the plant
personnel is thoroughly familiar with the.operation of the
existing treatment facility and could therefore maintain
the present standards of quality with little difficulty.
The proposed additions to the plant will provide ter-
tiary treatment to reduce the levels of phosphorus and
ammonia-nitrogen in the effluent. This secondary stage in
the treatment process ideally complements the trickling fil-
ter operation because trickling filters will produce an
effluent with sufficient BOD to maintain the nitrifying or-
ganisms (See Figure 8).
The pilot plant study established that the RBC process
is capable of achieving greater than 95 percent removal of
ammonia, up t,o an ammonia loading rate of 0."2 pounds NH,-N
per 1000 square feet of surface area per day, provided that
sufficient alkalinity is available in the wastewater. At
the projected ammonia load for the year 2000, some 3,200,000
square feet of surface area- would be required for this load-
ing rate. In order to provide this amount of surface area,
it was decided that four (4) trains consisting of six (6)
shafts, each 25 feet long, would be utilized. It was felt
that the four separate trains would provide the desired deg-
ree of flexibility in the system and six shafts per train
would insure that sufficient surface area was provided to
prevent ammonia breakthrough due to peak flow loadings.
The first two (2) shafts of media in each train will be fab-
ricated of so-called "standard density" media, that is,
approximately 100,000 s.f. per shaft. The remaining four (4)
shafts in each train will be fabricated of "high density"
media or about 150,000 s.f. per shaft. In this manner,
problems associated with "bridging" of the media due to over-
loading in the first and second stages can be avoided. In
addition, the six stages of media provide sufficient surface
area for carbonaceous BOD removal in the first two stages
of each unit.
Flexibility in each train will be provided through the
utilization of removable wooden baffles between shafts. The
amount .of surface area per stage and the number of stages
per train can then be varied to suit operating conditions.
Alkalinity will be provided immediately before the RBC
process in the form of hydrated lime. Hydrated lime was
chosen as the source of alkalinity due to its lower cost and
local availability. It is anticipated that the alkalinity
will be a minimum of 200 mg/1 in the influent to the RBC.
754
-------
AERATED SEPTAGE
SLUDGE
THICKENING
SLUDGE
DEWATERING
L.
GRIT REMOVAL
COMMINUTION
FLOW METERING
PRIMARY CLARIFICATION
FLOW CONTROL
TRICKLING FILTER
MAIN PUMPING STATION
-ALUM ADDITION
INTERMEDIATE CLARIFICATION
-LIME ADDITION
ROTATING BIOLOGICAL
CONTACTORS
ALUM ADDITION
FINAL CLARIFICATION
FINAL FILTRATION
ULTRA-VIOLET DISINFECTION
SCHEMATIC FLOW
DIAGRAM
.CASCADE AERATOR
FIGURE 8.
755
-------
Flexibility was also designed into the phosphorus
removal system, in that two different chemicals for the
removal of phosphorus can be stored and fed at one time, and
chemicals can be added and mixed with the wastewater at three
separate locations. Laboratory treatability studies estab-
lished that alum would provide sufficient removal of phos-
phorus to meet the discharge limitation of 1.0 mg/1 consis-
tently. In addition, the study established that the addition
of alum immediately following the trickling filters would be
the least costly method. This is due to the low pH (5.8 to
6.2) of the trickling filter effluent. Removal of phosphorus
with alum after the RBC process would require substantially
more chemical due to the elevated pH of the wastewater at
this point. However, chemical addition and mixing facilities
will be provided after the RBC process for removal of phos-
phorus in the final clarifiers. In addition, chemical
additions can be made to the headworks at the aerated grit
chamber for phosphorus removal in the primary clarifiers.
SUMMARY AND CONCLUSIONS
1. A linear relationship between pounds of ammonia
applied versus pounds of ammonia removed existed for ammonia
loadings ranging from 0.04 to 0.4 pounds NH -N per 1000 s.f-
day.
2. That the removal of ammonia was greater than 95
percent for loadings up to 0.2 pounds NH_-N applied/1000 s.f.
day, and generally above 90 percent removal for loadings up
to 0.4 pounds NH3~N applied/1000 s.f.-day.
3. That the RBC units were able to respond well to
peak hydraulic loading.
4. That there is little to no net production of solids
within the RBC units. The solids which were produced did
not exhibit good settling characteristics.
756
-------
References
1. "Process Design Manual for Nitrogen Control", U.S.
Environmental Protection Agency - Technology Transfer.
2. Saunders, F.C. and Pope, R.L., "Nitrification with
Rotating Biological Contactor Systems", Env. Resources
Center Technical Report (ERG 06-78), Georgia Institute
of Technology, Atlanta, Georgia, October 1978.
3. O'Shaughnessy,J.C., Blanc, F.C., Brooks, P.,
Silbovitz, A. and Stanton, R. "Nitrification of
Municipal Wastewater Using Rotating Biological
Contactors", Proceedings First National Symposium/
Workshop on Rotating Biological Contactor Technology,
University of Pittsburg, Vol. II, pp. 1193-1219,
June 1980.
4. Odegaard, H. and Rusten, B., "Nitrogen Removal in
Rotating Biological Contactors Without the Use of
External- Carbon Source", Proc. First National Symposium
on RBC Technology, University of Pittsburg, Vol. II,
pp. 1301-1317, June 1980.
,757
-------
IMPROVEMENT OF NITRIFICATION IN ROTATING BIOLOGICAL
CONTACTORS BY MEANS OF ALKALINE CHEMICAL ADDITION
James M. Stratta, U. S. Army Environmental Hygiene
Agency, Aberdeen Proving Ground, Maryland.
David A. Long, Department of Civil Engineering, The
Pennsylvania State University, University Park,
Pennsylvania
Michael C. Doherty, U. S. Army Environmental Hygiene
Agency, Aberdeen Proving Ground, Maryland.
INTRODUCTION
The need to achieve compliance with ammonia-nitrogen
discharge limitations and the current emphasis on energy
conservation have resulted in the utilization of RBC tech-
nology for the nitrification of secondary wastewater
effluents. By the end of the 1970s, four pilot scale
efforts, independent of the RBC industry, had been completed
which demonstrated that the RBC could nitrify successfully
secondary wastewater effluent (1, 2, 3, 4). In 1979,
approximately 70 percent of the RBC systems in the United
The opinions or assertions contained herein are the private
views of the authors and are not to be construed as reflect-
ing the views of the Department of the Army or the Depart-
ment of Defense.
758
-------
States were .designed to remove carbonaceous biochemical
oxygen demand (CBOD). Another,25 percent of the RBC systems
were designed to remove CBOD and for nitrification in the
same RBC units. The remaining 5 percent were constructed
to nitrify secondary wastewater effluents in order to
achieve ammonia-nitrogen effluent discharge limitations (5)
Initial evaluations of full .scale nitrifying RBC facilities
reveal that they have not been completely satisfactory (6,7)
Hitdlebaugh (7) reported that an RBC facility, built for
CBOD removal and nitrification, failed to meet design
specifications during both winter and summer operations.
The inability to meet CBOD and ammonia-nitrogen limitations
during the summer was attributed to relatively low dissolved
oxygen (DO) concentrations (less than 1 mg/1). in the initial
nitrifying stages and a low pH (less than ph 7.0) in the
latter nitrifying stages. The DO level increased during
winter operations and CBOD was removed sufficiently to
achieve design expectations. Ammonia-nitrogen removal also
improved during the winter but not sufficiently to achieve
design projections or effluent limitations. Recommendations
from this study included the use of alkaline chemical feed
systems to maintain optimum pH levels in order to improve
nitrification.
Nitrification within the RBC biofilm is essentially
a two-step microbiological process .which utilizes two groups
of autotrophic bacteria of the family Nitrobacteraceae. The
first group of bacteria oxidizes ammonia to, nitrite and the
second group of bacteria oxidizes nitrite to nitrate. The
Nitrosomonas and Nitrobacter genera are considered to be the
predominant nitrifying bacteria inhabiting the wastewater
environment. Heterotrophic nitrification also occurs when
nitrite or nitrate is produced from organic or inorganic
compounds by heterotrophic organisms. Over 100 heterotrophic
species (including fungi) have been identified which are
capable of heterotrophic nitrification. However, the overall
contribution to the oxidized nitrogen forms by heterotrophic
nitrification is considered to be relatively small (8).
The growth rates of nitrifying bacteria are much slower
than the growth rates o'f heterotrophic bacteria. This
important distinction accounts for the inability of nitri-
fication to proceed simultaneously with CBOD removal when
high concentrations of organic material (greater than 30 mg/1
of BOD) are present in the wastewater. Minimum doubling
times reported for the ammonia-oxidizing bacteria are from
8 to 17 hours (9). Because the growth rates for nitrite-
759
-------
oxidizers are greater than the growth rates for ammonia-
oxidizers (9,10), elevated nitrite concentrations normally
do not persist and the ammonia-oxidation step controls
the total amount of ammonia which is oxidized to nitrate
within the wastewater environment. Carbon dioxide (C02) is
the carbon source for these autotrophic nitrifying bacteria
(11). Although some nitrifying bacteria have been observed
to use organic compounds, they were not observed to utilize
these organic compounds as the sole carbon source for growth
(12). The generation of bacterial biomass per unit of
ammonia oxidized (cell yield) is quite small. The total
yield for both Nitrosomonas and Nitrobacter has been observed
to be from 0.06 to 0.20 gram of cells per gram of ammonia
oxidized (9). The nitrification of 20 mg/1 of ammonia-
nitrogen generates approximately 2 mg/1 of solids (10).
Therefore, the net amount of inorganic carbon required for
this amount of nitrification is quite low. McGhee(13)
reported that the inorganic carbon requirements for the
nitrite oxidation step could be met without inorganic
carbon being present in the bulk solution. The utilizable
source of inorganic carbon was the C0« generated from
endogenous respiration within the biorilm.
The simplified oxidative reactions below describe the
salient aspects of the microbial oxidation of ammonia.
The microorganisms derive energy from these reactions;
this energy is used for C09 fixation.
Ammonia Oxidation:
NH,
2HC0
(1)
Nitrite Oxidation:
0.50
-"•"NO,
(2)
Overall Reaction:
NH,
20
2 HC0
(3)
As can be seen from Equation 1, the nitrification process
results in the production of acid which neutralizes the
alkalinity in the wastewater. Theoretically, 7.1 mg/1 of
alkalinity is destroyed for each 1 mg/1 of ammonia oxidized.
The destruction of alkalinity results in pH depression.
760
-------
The actual pH depression is mitigated somewhat by the
removal of carbonic acid through the stripping of CO 'from
the wastewater surface (10). However, under low alkalinity
conditions, pH depression is enhanced due to the reduced
buffer capacity of the wastewater. The level of alkalinity
within wastewaters varies widely. The major factor influenc-
ing the amount of alkalinity present is the source of the
carriage water, or the drinking water supply. High
alkalinities normally are associated with ground water
supplies and much lower alkalinities are associated with
surface supplies. Domestic wastewater contributes from
50 to 200 mg CaCO /I to the natural alkalinity of the
carriage water (15). Therefore, the amount of alkalinity
in a domestic wastewater may range from less than 100 mg
CaCO /I to several hundred mg CaCO /I. The net.effect of
such variations in alkalinity is to provide a different
buffering capacity for each wastewater treatment system.
Domestic wastewaters normally contain from 12 to 25 mg/1 of
ammonia-nitrogen. The range of alkalinity destroyed during
the nitrification of these concentrations of ammonia is
85 mg CaCO,./! to 178 mg CaCO,./!. Obviously, the pR depress-
ion resulting from nitrification can be slight for low
ammonia-high alkalinity wastewaters or significant for
high ammonia-low alkalinity wastewaters. Wastewater pH
levels are typically around pH 7.5. However, pH depression
to below pH 7.0 is common for low alkalinity wastewaters.
The level,of pH has an important effect on the nitri-
fication process. There have been a number of researchers
since the turn of the century who have addressed the
subject of the effect of pH on nitrification. Those
researchers who have made contributions pertinent to
this study are listed in reverse chronological order in
Table 1. It is interesting to note the large variation in
the effect of pH on biological nitrification that is
reported in the literature. The variation in the effect
of pH on nitrification is due in large measure to the
nature of the experiments undertaken, i.e., the homogeniety
of culture involved (pure versus mixed culture); scale of
the experiment (laboratory to full scale); nature of the
biofilm (suspended versus fixed film) and a variety of
(and frequently unspecified) acclimation times utilized in
the experiments.
Several investigators (1, 3, 15, 16) within recent years
have attempted to provide more information on nitrification
within a fixed film mode. Haug and McCarty (15) utilized
761
-------
Table 1. Literature Review of Optimum pH
Values for Nitrification,.
Ref.
3
1
13
17
18
16
15
19
20
21
22
23,24
25
26
27
28
29
Author
Miller, et al.
Borchardt, et al.
McGhee
Srna & Baggaley
Button & LaRocca
Huang & Hop son
Haug & McCarty
Mulbarger
Wild, et al.
Loveless & Painter
Downing & Knowles
Andersen
Boon & Landelot
Engel & Alexander
Bushwell & Shiota
Hoffman & Lees
Meyer ho ff
Year
1979
1978
1975
1975
1975
1974
1972
1972
1971
1968
1967
1964
1962
1958
1954
1952
1917
Optimum pH
8.0-8.5
7.1-8.6
8.0-9.0
7.45
8.4-8.6
8.4-9.0
7.8-8.3
8.4
8.4
7.5-8.0
7.2-8.0
8.4-8.5
7.0-8.6
7.0-9.0
8.0-8.5
8.0-9.0
8.5-8.8
8.5-9.0
Organism or
System Studied
RBC (Pilot)
RBC (Pilot)
A.S. (Lab)
Sub. Filt. (Lab)
A.S.
Bio film (Lab)
Sub. Filt. (Lab)
A.S.
A.S.
Nitrosomonas
—
Nitrosomonas
Nitrobacter
Nitrosomonas
Nitrosomonas
Nitrosomonas
Nitrosomonas
Nitrobacter
762
-------
a laboratory scale fixed film submerged reactor and a
synthetic wastewater and performed a short term pH-
nitrification study (18 hours at each pH value) using a
biofilm developed at neutral pH .and observed essentially
the same rate of nitrification at pH 6.5 as at pH 9.0.
At pH 6.0, the observed rate of.nitrification was reduced
to approximately 42 percent of the maximum rate and nitri-
fication essentially stopped at pH 5.5. However, after
only 10 days of operation at pH 6.0, the submerged filter
was reported to have acclimated sufficiently to perform
at the maximum rate of nitrification. This finding
demonstrates the ability of nitrifying organisms to acclimate
to low pH condition. The reason for this unique finding
may be due in part to non-equilibrium conditions existing
within the submerged filter after the startup period. Huang
and Hopson (16) utilized a laboratory scale inclined fixed
film surface and a synthetic wastewater to evaluate the
effect of pH on nitrification. Their experiment examined
the short term (less than 10 hours) effect of pH on the
nitrification process and produced a maximum rate of nitri-
fication at pH 8.4 to pH 9.0 with approximately 25 percent
of the maximum rate occurring at pH 6.0. After three weeks
of acclimation at pH 6.6, the rate of nitrification was
approximately 85 percent of the maximum rate observed.
Borchardt (1) performed a short term pH-nitrifleation
study utilizing a 0.6 meter pilot RBC treating domestic
wastewater effluent from a trickling filter in a laboratory
where ammonia, alkalinity and pH were controlled. The
rate of nitrification was examined at eleven different
levels of alkalinity after a short but undefined acclimation
period. The results of this short term study revealed a
nearly constant rate of nitrification between pH 7.1 and
8.6. Approximately 25 percent of the maximum rate of
nitrification was observed at pH 6.5 and zero nitrification
was indicated at pH 6.0. Borchardt was careful to point
out the limitations of attempting to extrapolate his
short term data into the long term.
Miller (3) most recently reported on a pilot scale
0.5 meter RBC treating domestic wastewater effluent from
a pilot trickling filter wherein significantly greater
rates of nitrification were observed at elevated pH levels
(pH 8.0 to pH 8.5) than at neutral pH (approximately
pH 7.1). This nitrification study is unique in that lime
addition for phosphorus removal preceded the nitrification
process and the nitrifying RBC stages had acclimated at
763
-------
the elevated pH levels. A transition in biofilm performance
was observed when the elevated pH of the wastewater was
reduced to the neutral pH range. Nitrification performance
initially remained unchanged. After approximately four
days, the performance level started to deteriorate. In "
nine days, the performance had reverted to a lower nitri-
fication level. This latter finding was not discussed
fully by Miller; however, it is important because it
helps to establish potential physical differences between
the biofilms developed at neutral and elevated pH levels.
This situation was not observed by the other investigators
using fixed films mentioned above because none ever
attempted to acclimate biofilms at the elevated pH levels.
Such differences cannot be assumed to be purely indicative
of only the pH dependent rates of microbial nitrification.
These differences also are reflective of the entire hetero-
genous population developed within each biofilm which
dictate film development, cohesion, and retention
characteristics (sludge age). There is essentially no
information within these wastewater nitrification studies
which addresses changes in biofilm and microbial popula-
tions under various pH conditions. In general, this
important consideration has been ignored in such waste-
water research studies. However, current research efforts
such as those by Olem (30), LaMotta (31) and Characklis (32)
are starting to examine more closely the mechanics of
biofilm development and the characterization of microbial
populations.
The addition of alkaline chemicals to wastewater
treatment systems to increase pH and provide added buffer
capacity has been attempted with varying degrees of success.
Heidman (33) conducted a pilot study at the Blue Plains
WWTP using an activated sludge system which incorporated
pH controlled nitrification. This study was inconclusive
because it failed to demonstrate the relative nitrification
without chemical addition. Hutton (18) demonstrated
the feasibility of optimizing the nitrification of high
ammonia strength industrial wastewaters with alkaline
chemical addition. Lue-Ling (34) reported success in
using alkaline chemical addition to nitrify high ammonia
strength lagoon supernatant with EJBCs. Hitdlebaugh (35)
attempted to enhance the nitrification of domestic waste-
water with RBCs through alkaline chemical addition; however,
the results were inconclusive. The literature fails to
address the efficacy of optimizing domestic wastewater
764
-------
nitrification within the RBC system through pH control as
well as the use of alternative pH control'schemes.
OBJECTIVES AND SCOPE
The objectives of this research were to:
1. Establish the relative rates of nitrification
for domestic wastewater treatment within an
acclimated RBC fixed film system as a function
of pH.
2. Observe and characterize the relative changes in
the RBC biofilm as a function of" pH.
3. Evaluate the efficacy of chemical addition to
improve nitrification within an RBC fixed film
system through the maintenance of an optimum pH.
4. Evaluate alternative alkaline chemicals for pH
controlled nitrification for the RBC.
5. Develop design criteria, as appropriate, for pH'
controlled nitrification for the RBC.
EXPERIMENTAL PROCEDURES . ' .. '
Pilot scale 0.5 meter diameter RBC systems were used
in this research to nitrify high rate trickling filter
effluent from the Pennsylvania- State University (PSU)
wastewater treatment plant (WWTP). The effect of pH
on nitrification within an RBC was evaluated using four
single stage RBCs operating in parallel (Figure 1). The
pH of the RBC systems treating the PSU WWTP trickling
filter effluent was varied first from pH 6,. 3 to pH 7.5
and then from pH 7.6 to pH 8.8 and the relative levels of
nitrification at the various pH's then were observed.
High pH and low pH environments were created by adding
sodium hydroxide and sulfuric acid, respectively, to the
wastewater streams after clarification (Figure. 1). Each
observation period was started with no biofilm on the
RBC discs and lasted approximately ten weeks.
The effect of alkaline chemical addition was evaluated
utilizing five 2-stage RBCs operating in parallel (Figure
2). The level of nitrification of a low pH 2-stage
control RBC system (control) was compared against the
nitrification level of four other 2-stage RBC systems
receiving four different alkaline chemicals. The four
alkaline chemicals used were calcium hydroxide, sodium
carbonate, sodium hydroxide, and sodium bicarbonate.
765
-------
FLOW DIVIDER
RBC UNITS*-
CLARIFIER -
TEMPERATURE
CONTROLLER
PUMP
®H2S04or NaOH
J\\
../ \ X
\ X
pH CONTROLLER
Figure 1. Schematic Diagram of the Pilot RBC Units for
the Low pH- and High pH-Nitrification Study
766
-------
RECYCLE
V PUMP J
CLARI FIER -
TEMPERATURE
CONTROLLER
FLOW Dl V IDER
NaOH
r
NaHCOa
1-1
pH CONTROLLER
2-1
1-2
Ca(OH)2
3-1
2-2
4-1
3-2
4-2
PUMP
5-1
\
RBC - STAGE
5-2
Figure 2. Schematic Diagram of the 2-Stage RBC
Systems of the Alkaline Chemical Addition
Study.
767
-------
The alkaline chemicals were added to the first stage only.
The first stages of the calcium hydroxide, sodium carbonate,
and sodium hydroxide RBC systems were maintained at the
optimum pH level for nitrification (approximately pH 8.5).
The first stages of the sodium bicarbonate and the control
RBC systems were maintained at pH 7.5 and pH 7.0, respectively.
The low pH wastewater was created by adding sulfuric acid
prior to clarification and the alkaline chemicals then were
added directly to the first stage of each RBC. Ammonium
chloride was added to the wastewater during PSU break
periods to augment the low influent ammonia-nitrogen... There
was no biomass on the discs at the start of the test. The
observation period for this part of the study lasted
approximately 11 weeks.
Wastewater sampling was accomplished by compositing
grab samples on influents and effluents. Biofilm sampling
and analyses were performed in accordance with modifications
of the procedures reported by Olem (30). The most
probable numbers (MPN) of ammonia-oxidizing bacteria were
determined using a modification of the Nitrosomonas MPN
technique of Alexander and Clark (36) which was reported
by LaBeda and Alexander (37) as well as Rowe (38). The
nitrite-oxidizing bacteria MPN values were determined
using a modification of the Nitrobacter MPN technique of .
Alexander and Clark (36) which was reported by LaBeda and
Alexander (37) and Ghiorse and Alexander (39). The
enumeration of heterotrophic bacteria was accomplished by
spread plating serial dilutions of Modified Taylor's Media
(40). Detailed descriptions of all sampling and analytical
procedures are found in Stratta (41).
RESULTS AND DISCUSSION
Low pH-Nitrification Study
This research phase was devoted to the evaluation
of the relative rates of nitrification in single stage
RBC systems operated at pH 7.5, 7.1, 6.5 and 6.3. The
pH 7.5 RBC treated the unaltered wastewater and served
as the control. Data showing the operational characteristics
of the four RBC systems are presented in Table 2. The rates
of nitrification which initially developed for the
pH 7.5 and the pH 7.1 RBC systems were greater than
the nitrification rates of the two lower pH RBC systems.
Nitrification was not established in the pH 6.3 RBC system
768
-------
Table 2. Pilot Single-Stage Nitrifying
RBC Operating Characteristics
Secondary Clarifier
Surface Settling Rate (@6.8 m3/d) - 5.9 m3/d-m2
Detention Time (@6.8 m3/d) - 2.1 hr
RBC
Number of RBCs
Stages per RBC
Discs per Stage
Disc Diameter
Disc Area - Total
Rotational Speed
Peripheral Speed
Hydraulic Loading0
- 4
- 1
- 9
- 0.5 m
- 5.3 m2
- 13 rpm
- 0.34 m/sec
- 81 l/m2-d
aThe hydraulic loading for all four RBC units was nominally
81 l/m2-d (2 gal/d'ft2). The hydraulic loading calculation
is based upon the assumption that each RBC is the first
stage of a 4-stage RBC.
769
-------
during the first 25 days of operation. Two short-term
pH excursions may have had an adverse impact on nitrification
development on the pH 6.3 RBC. On Day 27, the pH of the
pH 6.3 unit was adjusted upward to pH 6.7 in an attempt
to obtain additional information regarding the nitrification
rate between pH 6.5 and pH 7.1; this RBC is referred to
hereafter as the 6.3/6.7 RBC. Based upon nitrification
performance, a period of relative equilibrium was established
by about Day 37 for the pH 7.5, 7.1, and 6.5 RBC units.
Data on the relative amounts of ammonia-nitrogen removed
by the RBCs are presented in Table 3. The rate of nitri-
fication of the pH 7.1 RBC system was 96 percent of that
observed at pH 7.5 and the rate for the pH 6.5 RBC was 80
percent of the rate for the pH 7.5 RBC. Because of the
long period of time required for nitrification to become
established in the pH 6.3/6.7 RBC, data for this unit are
not included in Table 3. Nitrogen balances for each
Table 3. Relative Rates of Nitrification for
RBC Systems Operating Under Low
pH Conditionsa
RBC pH
Ammonia-N Removed
g NH3-N/m2-d
Percent of
Maximum
7.5
7.1
6.5
2.5
2.4
2.0
100
96
80
Based upon data from Day 37 to Day 69
of the RBC systems are presented in Table 4. The slightly
lower nitrogen recoveries for the two higher pH systems
might reflect small nitrogen losses associated with
denitrification within the heavier biofilms as well as minor
losses due to ammonia stripping.
770
-------
Table 4. RBC Nitrogen Balances for the
Low pH-Nitrification Study
RBC
1
2
3
4
pH
7.5 .
7.1
6.5
6.3/6.7
Total Nitrogen ...
Influent
19.5(23)°
19.. 6 (23)
19, 7 (24)
19.6(23)
- mg/1
Effluent ;
19.3(23)
19.7(23)
20.3(24)
20.2(23)'
Percent
Recovery
99
100
103
103
aBased upon data from Day 37 to Day 69
Total nitrogen is total oxidized nitrogen plus total
Kjeldahl nitrogen (TKN)
CNumber in parenthesis is the number of samples utilized
in the total nitrogen determinations.
The four RBCs developed biofilms which could be
sensed by touch within 48 hours. .A noticeable bronze color
developed after five days of operation. By the tenth
day, all RBC systems had developed thin and highly uniformly
textured coatings which possessed a visually apparent
gradation. The heaviest biofilm appeared in the pH 7.5 RBC
and the lightest growth of biofilm was in the pH 6.3/6.7 RBC.
All four RBC units showed some degree of sloughing by
Day 19 with the greatest sloughing occurring in the
pH 6.3/6.7 RBC. The, biofilm color changed from bronze
to brown with increasing age and increased biomass. After
the loss of the-initial biofilm uniformity, the biofilm
became increasingly patchy with time and decreasing pH. At
the conclusion of this phase of the study, the non- :
uniformity of the biofilm was quite evident visually and
seemed to be related directly to the relative nitrification
rates recorded over the duration of the study. The .
771
-------
patchy appearance was attributed mainly to biofilm loss
resulting from hydraulic shear. However, biofilm sloughing
from the disc surface did occur.
The RBC disc biofilm development data for all four
RBC systems during the low pH-nitrification study are
presented in Figure 3. The RBC systems at pH 7.5 and
7.1 showed the best performance and had the most disc
biofilm. The pH 6.5 RBC showed a lower level of performance
and less biofilm. The RBC which was operated initially
at pH 6.3 and later adjusted to pH 6.7 had the lowest
performance level throughout most of the 69-day study
and also developed the least amount of biofilm. The
maximum ammonia-oxidation levels for the pH 7.5, 7.1, and
6.5 RBC's were achieved when the biofilm masses were
approximately 2.0, 2.2, and 1.5 mg/cm2 respectively. In-
creases in disc biomass did not enhance the nitrification
rates for any of these RBC systems. The pH 6.3/6.7 RBC
added biofilm during the first three weeks of operation
at a rate comparable to that of the pH 7.5 RBC yet
showed no nitrification capacity. This result indicates
that, at least initially, organisms other than nitrifying
bacteria were inhabiting the RBC discs.
The relative geometric mean data for the nitrifying
bacterial populations per unit disc area and per unit
volatile weight for each RBC system after the initial
month of startup are presented in Figure 4. These graphs
demonstrate clearly that the total number of viable
nitrifying bacteria on each RBC was related directly to
overall RBC nitrification performance. The higher pH
systems had larger populations of both ammonia-oxidizing
and nitrite-oxidizing bacteria. The sustained depressed
nitrification performance and the relatively low nitrifying
bacteria populations of the 6.3/6.7 RBC indicated that a
significant period of time was required for complete
autotrophic adjustment in response to system changes under
relatively low pH conditions. The ratios of ammonia-
oxidizing bacteria to nitrite-oxidizing bacteria for
the pH 7.5, 7.1, 6.5 and 6.3/6.7 RBC units were 16:1, 14:1,
9.4:1 and 3.3:1, respectively. This observation indicates
that the lower pH systems favor nitrite-oxidizing bacteria
relative to ammonia-oxidizing bacteria. This conclusion
is shown graphically in Figure 4 which shows that the
number of ammonia-oxidizing bacteria per dvg (dry volatile
gram) decreased with decreasing pH, while the number of
nitrite-oxidizing bacteria per dvg increased for the two
772
-------
0 10 20 30 40 50 60 70
DAYS OF OPERATION
LEGEND: RBC
1-PH7.5 -
3-pH6.5 -
2 pH7.1
-•-4-pH 6.3/6.7
Figure 3. RBC Disc Biofilm Development for the
Low pH-Nitrification Study
773
-------
'
1
1™
T"
OA
L.
|
a/v
i
U3JI
'xwa
.%Jw Q
e
j
3
...
_L
1
3
cn
a
h.
a
N
C
X
o
1
<
^ I
« i
n 1
CO >-l
a o
o
•rl 4-1
4J -H
CC) 0
§"§
0 cd
cd cu
O 4J T3
•H -H 3
^04-1
4J (b CO
0)
O O
O O 4-1
cd
a) cd o
i> -H -H
•H te <4-l
CU -rl
Cd 4J
iH O 4J
cu cd -H
0)
S
bO
774
-------
lower pH RBC units. The volatile contents for these biofilms
were 82, 83, 84, and 81 percent for the pH 7.5, 7.1, 6.5, and
6.3/6.7 RBCs, respectively.
High pH-Nitrification Study
This research phase was devoted to the evaluation of the
relative rates of nitrification in single stage RBC systems
operated at pH 7.6, 8.0, 8.5, and 8.8. The pH 7.6 RBC
treated the unaltered wastewater and served as the control.
The operational characteristics of the four RBC systems were
the same as reported previously in Table 2. On Day 36, a
pH excursion to approximately pH 11.0 for an estimated
two hours occurred within the pH 8.8 RBC with rather
dramatic results. This short-term transient condition
appeared to have little effect on the ammonia-oxidation
process. However, the RBC experienced an immediate loss
of nitrite-oxidation capability and a very slow nitrite-
oxidation recovery. Based upon ammonia removal, and not
complete oxidation, a period of relative equilibrium was
established after approximately five weeks of operation.
Data on the relative amounts of ammonia-nitrogen removed
by the RBCs are presented in Table 5. The pH 8.8 RBC
Table 5. Relative Rates of Nitrification for RBC
Systems Operating Under High pH Conditions'1
RBC pH
7.6
8.0
8.5
8.8
Ammonia-N Removed
(g NH3-N/m2-d)
2.0
2.6
3.1
2.9
Percent of
Maximum
65
84
100
94
aBased upon data from Day 38 to Day 71
775
-------
and the pH 8.0 RBC removed 94 and 84 percent as much
ammonia as did the pH 8.5 system, respectively; whereas
the control RBC removed only 65 percent as much ammonia.
Data on the nitrogen balances for the four RBC systems
for this time period are presented in Table 6. As noted
earlier, the slightly lower nitrogen recoveries at the
higher pH conditions may be due to small nitrogen losses
resulting from ammonia stripping and denitrification within
the heavier biofilms. These nitrogen recovery results
indicate that ammonia stripping was not a major factor
affecting the change in ammonia levels between pH 7.6 and
pH 8.8.
Table 6. RBC Nitrogen Balances for the
High pH Nitrification Study
RBC
1
2
3
4
pH
7.6
8.0
8.5
8.8
Total Nitrogen
Influent
21.7(21)°
21.5(21)
21.7(21)
21.8(21)
- mg/1
Effluent
21.1(21)
20.2(21)
20.6(21)
20.8(21)
Percent
Recovery
97
94
95
95
Nitrogen balances are based upon data from Day 38 to Day 71.
Total nitrogen is total oxidized nitrogen plus total
Kjeldahl nitrogen (TKN).
CNumber in parenthesis is the number of samples utilized in
the total nitrogen evaluations.
The four RBCs developed biofilms which could be sensed
by touch within 48 hours. A noticeable reddish-brown bio-
film was evident on all the discs by the third day of
776
-------
operation. By the eighth day, the four RBCs had developed
thin and highly uniformly textured biofilms which possessed
a visually apparent gradation. The pH 8.5 and pH 8.8 RBC
biofilms initially developed more rapidly than did the pH
7.6 and pH 8.0 RBC biofilms. This initial biofilm
gradation was not related to ammonia removal efficiency.
The 'biofilm color had changed from reddish-brown to tan
or bronze on all discs by the 10th day. By the 13th day
of operation, the two lower pH RBCs had the most uniformly
textured biofilms while the pH 8.5 and pH 8.8 RBCs were
developing a "dimpled" appearance associated with the
heavier biofilms. The pH 8.8 RBC developed a patchy
appearance and also had started to slough significantly
after only two weeks. As time progressed, the RBC systems
added biofilm, but their texture became less uniform.
The heavier biofilms appeared to be associated with the
pH 8.0 and pH 8.5 systems. The pH 8.8 RBC experienced
the greatest biofilm sloughing. Figure 5 presents the
RBC disc biofilm development data for all four RBG systems
during the high pH-nitrification study. After the initial
month of operation, the levels of disc biofilm for the
pH 7.6, 8.0, and 8.5 RBCs were related directly to their
relative level of performance. The pH 8.8 RBC experienced
an initially high rate of biofilm development; however,
it reached a peak mass per unit area concentration on Day 31
and then experienced a continuous biofilm loss. All four
RBC systems experienced a marked decline in disc biofilm
after the end of the PSU spring term on Day 52 when the
influent CBOD concentration decreased. The four RBC
systems appeared to achieve an initial maximum level of
nitrification performance in about three weeks. These
performance levels cooresponded to disc biofilm concentra-
tions of approximately 0.8, 1.0, 1.2 and 1.4 mg/cm2 for the
pH7.6, 8.0, 8.5, and 8.8 RBCs, respectively. As demonstrated
previously, the increase in disc biomass did not improve the
rate of nitrification for any of the systems.
The data on the relative geometric mean bacteria
populations per unit of disc area and per unit weight of dry
volatile biofilm for each RBC system after the initial 30
days of operation are presented graphically in Figure 6.
The total numbers of ammonia-oxidizing and heterotrophic
bacteria increased with increasing pH up to pH 8.5 and then
experienced a drop at pH 8.8. The total number of nitrite-
oxidizing bacteria was greatest at pH 8.0 and decreased at
pH 8.5 and pH 8.8. The number of ammonia-oxidizing bacteria
relative to the total biofilm population was similar for
777
-------
-------
a
O
m
cc
o
K
UJ
a
ui
o
ui
ocn
cd
•H
cu PLI
4J ^j
O
cd cu
,J3 (-1
* *
CJ CO
cd a
cu o
co a •
o d 4J
•H cu d
•u P
cd oo o
s-i C o
CU -H
e t=~. cu
p m 4-)
S -rt Cd
CU M i-l
W ft
H-l ,.;
o &
CO
4J
CU
CO
CO
0
o
:>> -H
•H cd
OT fi ni
S O CU
o u g
ft 9
S O C
CO CU
T3
0) !>. O
CO cd vH
CU O S-J
!-4 M 4-J
cd m o
CO CO CU
fj fj 4J
cd o cu
cu -H ,d
a a d
•H -H cd
!-i IH
4-J -H CO
CU CO CU
S co P
O cd i-t
CU ^1 cti
o o >
w
H
VD
CU
E
•rl
779
-------
the pH 7.6, 8.0, and 8.5 systems but greatest at pH 8.8.
Similarly, the heterotrophic population increased with
pH. However, the nitrite-oxidizing bacteria populations
were nearly identical at pH 7.6 and pH 8.0 but lower at
pH 8.5 and pH 8.8. The ratios of heterotrophic to ammonia-
oxidizing to nitrite-oxidizing bacteria based on the data
presented in Figure 6 for the pH 7.6, 8.0, 8.5, and 8.8 RBC
units were 12:15:1, 16:16:1, 37:24:1, and 43:48:1, res-
pectively. In general, these population figures indicate
that the heterotrophic bacteria and the ammonia-oxidizing
bacteria are favored over the nitrite-oxidizing bacteria
with respect to increasing pH. During this period, the
mean biofilm concentrations were 1.80, 2.26, 2.99, and
1.23 mg/cm2 for the pH 7.6, 8.0, 8.5, and 8.8 RBC units,
respectively. The volatile content was 86, 86, 83, and 75
percent for the pH 7.6, 8.0, 8.5, and 8.8 RBC units,
respectively. This lower volatile content of the pH 8.8
RBC was attributed to low level precipitation of calcium
carbonate and entrainment of the precipitate within the
disc biofilm.
At the conclusion of the 10-week high pH-nitrification
study, the two RBC systems which had been operating at
pH 7.6 and pH 8.5 were utilized in a short-term pH-nitrifi-
cation study wherein the two RBC systems experienced
simultaneous short term changes in pH, i.e. 2 hours
of operation at each pH level. The pH level started at
pH 9.0 and was decreased progressively downward to
pH 6.0 without interruption. Alkalinity and pH levels
were maintained in each RBC by direct feed of sodium
hydroxide and sulfuric acid solutions. This test was
run twice (Day 73 and Day 79) with similar results. The
average values of nitrogen removal data obtained from
these two runs are shown in Figure 7.
The shapes of the performance curves for the two RBCs
are similar yet the level of nitrification for the two
RBCs are markedly different. Clearly, the RBC which had
acclimated at pH 8.5, and had a history of elevated
performance, retained its higher performance level in the
short-term and continued to perform significantly better
than the biofilm acclimated at pH 7.6. The RBC response
to short-term changes in pH is relatively constant between
pH 7.9 and pH 9.0 but highly dependent upon the previous
acclimated level of nitrification for the given RBC biofilm.
Data on the alkalinity levels also have been included in
Figure 7. The amount of alkalinity present at the low
780
-------
12
0
1
S
ui
cc
I 4
Iw
PH7.6
5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5
pH
300
250
»150
50
pH 7.6
pH 8.5
5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5
pH
Figure 7. The Relative Rates of Nitrification of RBC Systems
Acclimated at pH 8.5 and pH 7.6 and Subjected to
Short Term pH Changes and Related Alkalinity
Levels
781
-------
pH levels demonstrates that even at 20 mg/1 of
significant amounts of nitrification were achieved. The
response of the lower pH system closely resembles the
data of Borchardt (1) and similarly reveals good nitrifica-
tion at very low alkalinity levels. This result tends to
reinforce the observation that the pH level is much more
important than the alkalinity level per se in terms of
effect on nitrification. The amount of time required for
an RBC system to adjust to an altered pH condition is
discussed further in Stratta (41).
Alkaline Chemical Addition Study
This phase of the research was devoted to the evaluation
of the rates of nitrification of 2-stage RBC systems main-
tained at elevated pH levels through alkaline chemical
addition. The pH and alkalinity levels within the first
stages of four RBC systems were adjusted upward and
maintained artificially with four different alkaline
chemicals. Calcium hydroxide, sodium carbonate, and
sodium hydroxide were used to maintain approximately
pH 8.5, the optimum pH, in the first stage of three
different RBCs. Sodium bicarbonate was used to maintain
pH 7.5 in the first stage of the fourth RBC. A fifth
RBC which treated low pH-low alkalinity wastewater
was used as a control. Data on the operational character-
istics of the five RBC systems are presented in Table 7.
The overall nitrification capacity of the 2-stage
RBC control system developed more slowly than did that
of the higher pH systems. However, the control RBC and
the alkaline chemical feed RBC systems were operating at
approximately the same level of nitrification performance
after a little more than three weeks of operation. The
levels of performance for all five systems were very
similar for about the next ten days. The overall per-
formance of the control system, operating at the lower
pH level, started to deteriorate after approximately 35
days of operation.
Based upon the overall nitrification performance
of the five 2-stage RBC systems, a period of relative
equilibrium was established by about Day 38. The data
on the relative amounts of total ammonia-nitrogen removed
by the five RBCs from Day 38 to Day 75 are presented in
Table 8. These data show that the overall ammonia-nitrogen
removals for the sodium hydroxide, sodium carbonate, and
782
-------
Table 7. Pilot 2-Stage Nitrifying RBC
Systems' Operating Characteristics
Secondary Clarifier
Surface Settling Rate (@ 8.6 m3/d) -
Detention Time (@ 8.6 m3/d)
RBC
Number of RBCs
Stages per RBC
Discs per Stage -
Disc Diameter -
Disc Area - Total
Rotational Speed
Peripheral Speed
o
Hydraulic Loading
7.4 m3/d-m2
1.7 hr
5
2
9
0.5m
10.6 m2
13 rpm
0.34 m/sec
81 l/m2-d
SThe hydraulic loading for all five RBC units was nominally
81 l/m2-d (2 gal/d-ft2). The hydraulic loading calculation
is based upon the assumption that each 2-stage RBC system
contained the first two stages of a 4-stage RBC.
783
-------
Table 8. Relative Rates of Nitrification for
the RBC Systems of the Alkaline Chemical
Addition Study
Alkaline Ammonia-N Removed Percent Percent of
RBC Chemical (g NH3-N/m2'd) Removed Maximum
1
2
3
4
5
NaOH
NaHCO
Na0CO,
2 3
Ca(OH)2
Control
2.52
2.40
2.54
2.55
2.14
86
82
87
87
73
99
94
100
100
84
aBased upon data from Day 38 to Day 75
calcium hydroxide RBC systems were nearly the same. The per-
formance level of the sodium bicarbonate RBC system was about
6 percent less than that of the other three high pH systems.
The control RBC, which was operated at the lowest pH con-
ditions, removed about 16 percent less ammonia-nitrogen
than did the three high pH alkaline chemical feed systems.
Table 9 presents the ammonia removal data for each
respective RBC stage. The data on the amounts of ammonia-
nitrogen removed clearly demonstrate that the greatest
removal occurs at the elevated pH conditions. The amount
of ammonia removed by the first and second stages of the
three high pH alkaline chemical feed RBC systems was
essentially the same. The sodium bicarbonate RBC system
had lower pH levels and lower performance in both stages.
The control had the lowest stage pH levels and the poorest
performance for both stages. Nitrogen balances during
this period for the stages of the five RBC systems are
presented in Table 10. These nitrogen balances follow the
same pattern previously reported. The percent recovery
decreased slightly as pH increased. Again, this lower
784
-------
Table 9. Relative Rates of Nitrification for the
Stages of the RBC Systems of the
Alkaline Chemical Addition Study
RBC-Stage
1-1
2-1
3-1
4-1
5-1
1-2
2-2
3-2
4-2
5-2
Alkaline
Chemical pH
NaOH 8.5
NaHC03 7 . 5
Na?CO- 8.4
Ca(OH)2 8.5
Control 7.0
7.9
7.7
8.0
7.9
6.9
Ammonia-N
Removed
(g NH3-N/m2-d)
2.53
2.33
2.55
2.57
2.14
2.50
2.46
2.52
2.54
2.14
Percent
Removed
43
40
44
44
37
77
72
78
78
59
Percent
of
Maximum
98
91
99
. 100
83
98
97
99
100
84
•Based upon data from Day 38 to Day 75
Based upon ammonia-nitrogen influent to each RBC stage.,
°Based upon maximum ammonia-nitrogen "removed by calcium.
hydroxide RBC stages.
785
-------
Table 10. RBC Nitrogen Balances for the Alkaline
Chemical Addition Study
RBC-Stage
INFLUENT
1-1
2-1
3-1
4-1
5-1
1-2
2-2
3-2
4-2
5-2
Alkaline Total Nitrogen
Chemical pH mg/1
6.5
NaOH 8.5
NaHC03 7.5
Na2C03 8.4
Ca(OH)2 8.5
Control 7 . 0
7.9
7.7
8.0
7.9
6.9
24.8(24)°
24.4(26)
24.0(24)
23.5(25)
23.7(25)
24.4(25)
22.8(24)
23.0(25)
22.3(25)
22.3(25)
23.8(25)
Percent
Recovery
Stage
-
98
97
95
96
98
93
96
95
94
98
RBC
-
-
-
-
-
-
92
93
90
90
96
SBased upon data from Day 38 to Day 75.
Total nitrogen is total oxidized nitrogen plus total
Kjeldahl nitrogen (TKN).
CNumber in parenthesis is the number of samples utilized in
the total nitrogen determinations.
Nitrogen balances for the stages are based upon stage
influent nitrogen.
786
-------
recovery is attributed to small ammonia losses due to
ammonia stripping and denitrification within the thicker
biofilms associated with the higher pH levels.
Biofilm which could be sensed by touch had developed
on the discs of all stages within 36 hours. Within
72 hours from startup, all first stage discs had developed
biofilms which were noticeably heavier than the second
stage biofilms. The characteristic tan color associated
with nitrifying biofilms had developed by Day 4 and
was more apparent in the first stage biofilms. All of the
biofilms were very uniform in texture. This initial
biofilm growth appeared to be heavier than the biofilms
developed during the previous research phases.. By Day 6,
the trough walls also had developed noticeable amounts of
biofilm. Although both stage biofilms got heavier and
darker with time, the heavier and darker biofilms were on
the first stage discs. The first stage biofilms became
brown while those on the second stage discs remained tan
to bronze in color. By Day 16, discs in all the first
stages had experienced some sloughing while those in the
second stages retained their uniformity. The loss of
uniformity in the second stages commenced about Day 21. As
time progressed, the loss of biofilm uniformity was greatest
for the control and the sodium bicarbonate RBC systems. The
first stages of the calcium hydroxide, sodium carbonate,'
and the sodium hydroxide RBC systems had the heaviest and
most uniform biofilm coatings. The biofilm uniformity
related directly to the ammonia removal performance levels
of the RBC systems. The loss of biofilm during the 75-day
period was attributed mainly to hydraulic shear starting
on the surface of the biofilm and progressing inward.
Biofilm sloughing from the bare disc outward did not occur
continuously. The former method of sloughing appeared to
be associated with relatively low and steady CBOD" loadings,
while the latter form of sloughing appeared to be associated
with relatively high and fluctuating CBOD loadings.
The data for the RBC biofilm concentrations, percent
volatile matter, and percent nitrogen are presented in Table
11. These data show that the highest biofilm concentrations
were associated with the higher pH levels. Only the
addition of calcium hydroxide resulted in an increase in
inert matter entrained within both the first and second
stage biofilms and a significant increase in effluent sus-
pended solids. Sodium hydroxide, sodium carbonate, and
sodium bicarbonate additions did not affect the biofilm ,
787
-------
Table 11. Mean Biofilm Concentrations, Percent Volatile
Matter, and Percent Nitrogen in the RBC Disc
Biofilm of the Alkaline Chemical Addition Study'
RBC- Alkaline
Stage Chemical
1-1 NaOH
2-1 NaHC03
3-1 Na2C03
4-1 Ca(OH)2
5-1 Control
1-2
2-2
3-2
4-2
5-2
Stage
pH
8.5
7.5
8.4
8.5
7.0
7.9
7.7
8.0
7.9
6.9
Biofilm Volatile Biofilm Nitrogen
mg/cm2 % %
2.45
2.08
2.51
3.03
1.57
0.97
0.84
0.96
1.88
0.87
86
86
87
66
87
89
90
89
70
90
5.7
5.7
7.5
7.0
7.6
5.4
7.6
6.4
6.0
6.8
Samples taken at weekly intervals from Day 36 to Day 75
DNitrogen percentages are based upon weekly samples from
Day 23 to Day 75.
volatile content; however, a slight increase in volatile
content in all the second stage biofilms was noted. The
addition of sodium hydroxide, sodium bicarbonate, and sodium
carbonate caused only a slight increase of from 1 to 3 mg/1
in the suspended solids in the RBC effluents; whereas the
use of calcium hydroxide increased the effluent suspended
788
-------
solids by approximately 20 mg/1. The observed increase in
the calcium hydroxide RBC biofilm inert content as well
as the increase in suspended solids is attributed to the
reaction between the calcium hydroxide and the carbonic acid
or carbon dioxide in the wastewater to form calcium carbonate.
The populations of ammonia-oxidizing, nitrite-oxidizing,
and heterotrophic bacteria were monitored for both stages
of each RBC system. Figures 8 and 9 present graphically
the data on the relative geometric mean bacteria populations ,
per unit of disc area and per unit weight of dry volatile
biofilm for the stages of each RBC system from Day 36 to
Day 74. This time period corresponds to the same period
over which the relative nitrification rates are compared in
Tables 8 and 9. The populations of all three groups of
bacteria per unit area were greater for the first stages
of the three high pH, high performance systems than for
the first stages of sodium bicarbonate and control RBC
systems. The first stages of the former were maintained
at pH 8.4 to pH 8.5 while the latter were maintained at
pH 7.5 and 7.0 for the sodium bicarbonate and control RBC
systems, respectively. In the second stages, where there
was less CBOD, less disc biofilm, and no pH control, the
population differences were not as dramatic. The ratios
of the populations for the three groups of bacteria for
both stages of each RBC system are presented in Table 12.
Results of this research effort had indicated through-
out the various phases that heterotrophic activity and
biofilm development were enhanced under elevated pH
conditions. In order to provide additional information
regarding this observation, approximately 400 cm2 of new
disc material was added to the first stage discs of the
control (pH 7.0), the sodium bicarbonate (pH 7.5), and
the sodium hydroxide (pH 8.5) RBC systems on Day 62. The
development of biofilm and the establishment of heterotrophic
populations on these discs were monitored through Day 77.
The resulting data are presented in Figures 10 and 11. The
data demonstrated that both the biofilm and the hetero-
trophic activity developed more rapidly as pH increased from
pH 7.0 to pH 8.5. During this 15-day test period, the
influent CBOD (soluble and inhibited) concentration was,
approximately 8 mg/1. However, significantly greater
amounts of CBOD, if present, may overshadow the more
subtle influence of pH.
789
-------
\
O
ow
o
a
Z
o
m
cc
8
M
(0
C5
Z
N
0
X
o
Ul
cc
Z
1
1
o
N
o
X
o
1
<
<
1
<
cc
UJ
1-
o
<
ffl
Q
Z
rrj
CU
CO
cO
rd
0)
CO
CO
d
cO
0)
B
O
•rl
M
4-1
0)
B
0
CU
w
EH
O
O
£
o
K
h-
0
CC
Ul
Ul
I
1
\\,
1
CO >i
}-i cO
cu n
3 0
d 4J
CU
VD
4-1 CO
0
CD CO
4-> PI
CU
CO B
O
> CO T3
•H CO
O rd
P-i P. O
O pq
d n Pi
0) O CU
S M rd
CU 4J
O 4-J
•H CU 4-1
r-l W O
4-1
cu id co
B d cu
O cfl 60
CU cO
O "4-1
60 CO
i> TI a)
•rl N ,d
4J -rl 4-1 >-,
CD *xJ *°d
i-l -rl rl 3
01 X O 4-1
pi o M-I co
oo
cu
CU
zwo/viuaiova 301
Ul
790
-------
V)
U
CO
oc
o
N
Q
X
O
1
UJ
E
Z
1
i
N
Q
X
O
1
o
0) '
O
CU
O r
O
S3.
U
X
Q.
O
oc
oc
UJ
1-
Ul
X
1
^
ro
1W
o >>
cO
CO P
4-J
-------
Table 12. Ratio of Heterotrophic:Ammonia-Oxidizing:
Nitrite-Oxidizing Bacteria for the RBC Stages
of the Alkaline Chemical Addition Study3
Heterotrophs:Ammonia-Oxidizers:Nitrite-Oxidizers
RBC Stage 1 Stage 2
NaOH
Na-CO-
2 3
Ca(OH)2
NaHCO
Control
14 :
14 :
12 :
13 :
11 :
16
8.6
7.8
7.5
9.4
: 1
: 1
: 1
: 1
: 1
7.1 :
5.5 :
' 11. :
8.1 :
7.3 :
8.0
4.7
7.9
4.0
6.1
: 1
: 1
: 1
: 1
: 1
Ratios are based upon the geometric mean of 6 sets of samples
taken at weekly intervals from Day 36 to Day 74.
A summary of the alkalinity destruction rates based upon
data obtained from both continuous and batch operations is
presented in Table 13. Except for the rate observed in the
first stage of the calcium hydroxide RBC, all the alkalinity
destruction rates were in the range of commonly accepted
values. The unusually low alkalinity destruction rate
observed in the bulk solution of the first stage of the
calcium hydroxide RBC is attributed to the buildup of
calcium carbonate within the biofilm which effectively
neutralized some of the acid generated by the nitrifying
bacteria within the biofilm. The net result was to reduce
the overall amount of alkalinity destroyed in the bulk
solution during the nitrification, process.
SUMMARY AND CONCLUSIONS
This research examined the short and long-term effect
of pH upon the nitrification of wastewater within RBC fixed
film systems. In the long-term, the rate of nitrification
792
-------
/ /'pH7.0
15
Figure 10. Relative RBC Biofilm Development under pH
Conditions from pH 7.0 to pH 8.5
793
-------
9-
CM
I
Ul
o
o
- a
O 8
cc
fe
£
UJ
X
o
Q
7 -.
pH 7.0
6 9
DAYS
12
Figure 11. Relative RBC Heterotrophic Bacteria Growth
Under pH Conditions from pH 7.0 to pH 8.5
794
-------
Table 13. RBC Nitrification and Alkalinity Destruction
During the Alkaline Chemical Addition Study '
RBC -
Stage
1-1
2-1
3-1
4-1
5-1
Alkaline
Chemical
TStaOH
NaHC03
Na2C°3
Ca(OH)2
Control
Alkalinity Destruction
(mg CaC03/mg NH^N)
Continuous Operation3 Batch Operation
6.2
— 6.8
7.4
3.8
7.4 6.6
1-2
2-2
3-2
4-2
5-2
6.1
7.7
7.9
7.0
7.2
aBased upon data for the continuous operation from,Day 38
to Day 75.
within an RBC fixed film system was dependent upon pH. The
rate of nitrification increased with increasing pH up to a
maximum at pH 8.5. Approximately five weeks of operation
were required to clearly observe these differences. The
response of a nitrifying RBC system to short-term changes
in pH was relatively constant from pH 7.0 to pH 8.5. Below
pH 7.0, the adverse effect of pH becomes more pronounced.
795
-------
However, the absolute level of nitrification was dictated
by the biofilm's previous history of nitrification
performance. RBO systems continued to nitrify at a
relatively high rate after the pH had been reduced suddenly.
The nitrogen balances for the various research phases re-
vealed that the relative amount pf nitrogen recovered for
each RBC system generally was slightly less for the higher ,
pH systems. This result was attributed to low level ammonia
stripping as well as the loss of nitrate due to denitrifica-
tion within the biofilm.
There was no significant difference in the performance
of the 2-stage nitrifying RBC systems which received calcium
hydroxide, sodium carbonate and sodium hydroxide. The
performance levels of the sodium bicarbonate and the control
RBC systems were 6 and 16 percent less, respectively, than
those of the other three systems. The use of alkaline
chemicals to maintain approximately pH 8.5 in the first
stage of a 2-stage nitrifying RBC resulted in the removal
of approximately 19 percent more ammonia than in the control
RBC system. Except for the first stage of the RBC receiving
calcium hydroxide for pH adjustment, the range of alkalinity
destruction for all RBC systems in the alkaline chemical
addition study was from 6.2 to 7.9 mg CaCO /mg NBL-N. This
result was attributed to a neutralization capacity which
developed within the RBC biofilm due to the entrainment of
CaCO,. The production of significant amounts of inert
material and suspended solids when calcium hydroxide is used,
favors the use of sodium carbonate and sodium hydroxide when
the nitrification is not followed by secondary clarification.
Higher levels of nitrification for the RBC systems were
associated with greater disc biofilm uniformity. In all
cases, except for the pH 8.8 RBC of the high pH study,
the higher pH RBC systems maintained greater concentrations
of volatile biofilm per unit of RBC disc area. The loss
of biofilm from the RBC disc surface did not follow the
traditionally accepted sloughing pattern. Biofilm did not
slough from the disc surface outward. The dominant
pattern of biofilm loss was from the biofilm surface inward.
This loss was due to hydraulic shear at the biofilm surface.
The RBC disc biofilm characteristics changed with time. The
initial biofilm was uniform in texture and tan to bronze in
color. The biofilm went through an aging process wherein
the biofilm became darker and the texture became less
uniform; the lower the pH, the less uniform the biofilm.
The disc biofilm was affected greatly by low level changes
796
-------
in CBOD. The maximum rates of nitrification for individual
RBC stages were not associated with the maximum biofilm
concentrations on the discs. Disc biofilm continued to
develop after the individual RBC stages achieved their
maximum rate of nitrification. The elevated pH RBC biofilms,
which had enhanced nitrification capacities, had higher
nitrifying bacterial populations than the lower pH RBC
biofilms. The ammonia-oxidizing bacteria generally were
favored over the nitrite-oxidizing bacteria with respect to
increasing pH. Greater heterotrophic growth and more
rapid biofilm development was observed to occur at elevated
pH levels. • ' :
ACKNOWLEDGEMENT
This research was supported by the U. S. Army Medical
Research and Development Command under Contract No. DAMD17-
79-C-9110. . . '
797
-------
10.
11.
LITERATURE CITED
Borchardt, J. A., et al., Nitrification^of Secondary
Municipal Waste Effluents by Rotating Bio-Discs, USEPA
Grant No. R803407, Cincinnati, Ohio, 1978.
Hewitt, T., Nitrification of a, Secondary Municipal
Effluent Using a Rotating Biological Contactor, Research
Publication No. 71, Ontario Ministry of the Environment,
1978.
Miller, Roy D., et al, Rotating Biological Contactor
Process for Secondary Treatment and Nitrification
Following a Trickling Filter, Technical Report 7805,
U.S. Army Medical Bioengineering Research and Develop-
ment Laboratory, 1979.
O'Shaughnessy, James C. and Blanc, Frederic C., Bio-
logical Nitrification and Denitrification Using Rotating
Biological Contactors, Publication No. 97, Water Resources
Research Center, University of Massachusetts, Amherst,
1978.
Hynek, Robert J. and lemura, Hiroshi, "Nitrogen and
Phosphorus Removal with Rotating Biological Contactors,"
Proceedings; First National Symposium/Workshop on
Rotating Biological Contactor Technology, Champion,
PA, 1980, pp. 295-324. ~~
Crawford, Paul M.,' "Use of Rotating Biological Contactors
for Nitrification at the City of Guelph Water Pollution
Control Plant, Guelph, Ontario, Canada," Proceedings:
First National Symposium/Workshop on Rotating Biological
Contactor Technology. Champion, PA, 1980, pp. 1247-1274.
Hitdlebaugh, J. A. and Miller, R. D., "Full-Scale Rotat-
ing Biological Contactor for Secondary Treatment and
Nitrification," Proceedings; First National Symposium/
Workshop in Rotating Biological Contactor Technology,
Champion, PA, 1980.
Painter, H. A., "Microbial Transformations of Inorganic
Nitrogen," Proceedings; Conference on Nitrogen as a
Wastewater Pollutant, Copenhagen, Denmark, August, 1975.
Painter, H. A., "A Review of Literature on Inorganic
Nitrogen Metabolism," Water Research, 4, 1970, pp. 393-
450.
USEPA, Process Design Manual for Nitrogen Control, 1975.
Watson, Stanley R., "Part 12 - Gram-Negative Chemo-
lithotrophic Bacteria," Bergy's Manual of Determinative
Bacteriology, 8th Edition, 1974, pp. 450-456.
798
-------
12. Delwiche, C. C. and Finstein, M. S. , "Carbon and
Energy Sources for the Nitrifying Autotrophic Nitrobacter,'
Jour. Bact., 90, 1965, pp. 102-107.
13. McGhee, Mary Frances,'Fundamental Studies of the Nitri-
fication Process, Ph.D. Thesis, Department of Civil Eng-
ineering, University of Kansas, 1975.
14. Metcalf and Eddy Inc., Wastewater Engineering: Treatment,
Disposal, Reuse, McGraw-Hill, New York, 1979.
15. Haug, Robert T. and McCarty, Perry L.:, "Nitrification
with Submerged Filters," Jour.' Water Poll. Control Fed.,
_44, 1972, p. 2086. , .
16. Huang, C. S. and Hopson, N. E., "Temperature and pH
Effect on the Biological Nitrification Process," Paper
Presented at N. Y. Water Poll. Control Ass'n Winter
Meeting, N.Y. City, 1974.
17. Srna, Richard F. 'and Bagley, Anne, "Kinetic Response
of Perturbed Marine Nitrification Systems," Jour. Water
Poll. Control Fed. 47, 1975, p. 473.
18. Hutton, W. C. and LaRocca, S. A., "Biological Treatment
of Concentrated Ammonia Wastewaters," Jour. Water Poll.
Control Fed., 47, 1975, p. 989.
19. Mulbarger, M. C., The Three Sludge System for Nitrogen
and Phosphate Removal, Advanced Waste Treatment Research
Laboratory, Office of Research and Monitoring, Environ-
mental Protection Agency, 1972.
20. Wild, Harry E., et al., "Factors Affecting Nitrification
Kinetics," Jour. Water Poll. Control. Fed. 43, 1971,
p. 1845.
21. Loveless, J. E. and Painter, H. A., "The Influence of
Metal Ion Concentrations and pH Value on the Growth of
a Nitrosomonas Strain Isolated from Activated Sludge,"
Jour. Gen. Microb. , 52, 1968, pp.' 1-14.
22. Downing, A. L. and Khowles, G., "Population Dynamics in
Biological Treatment Plants," Advances in Water Pollu-
tion Research; Proceedings of 3rd International Con-
.ference in Munich. Germany, Vol. 2., 1966.
23. Anderson, J. H., "Oxidation of. Ammonia by Nitrosomonas,"
Biochemical Jour., 95, 1965, pp. 688-698.
24. Anderson, J. H., "The Metabolism of Hydroxylamine to
Nitrite by Nitrosomonas," Biochemical Jour., 91, 1964,
pp. 8-17.
25. Boon, B. and Laudelot, H., "Kinetics of Nitrite Oxida-
tion by Nitrobacter Winogradski," J. Blochem., 85, 1962,
pp. 440-447.
799
-------
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Engel, M. S. and Alexander, M. , "Growth and Autotrophic
Metabolism of Nitrosomonas Europaea," Jour. Bact.,76,
1958, pp. 217-222.
Buswell, A. M. , et al., "Laboratory Studies on the
Kinetics of the Growth of Nitrosomonas with Relation to
the Nitrification Phase of the BOD Test, Appl. Microb..
2, 1954, pp. 21-25.
Hoffman, T. and Lee, H., "The Biochemistry of the
Nitrifying Organisms, Part 4. The Respiration and
Intermediary Metabolism of Nitrosomonas," Jour, of Bio-
chem., 54, 1953, pp. 579-583.
MeyerhofT 0., "Untersuchungen uber den Atmungsvorgany
Nitrifizierenden Bakterien," Pflugers Archges Physiol.,
166, 1917, pp. 240-280. _
Olem, Harvey, Rotating-Disc Biological Oxidation ot
Ferrous Iron in Acid Mine Drainage Treatment, Ph.D.
Thesis, Department of Civil Engineering, The Pennsylvania
State University, 1978.
LaMotta, E. J. and Hickey, R. F., "Factors Affecting
Attachment and Development of Biological Films on Solid
Media," Proceedings: First National Symposium/Workshop
on Rotating Biological Contactor Technology, Champion,
PA, 1980, PP. 803-828.
Characklis, W. G. and Trulear, M. G., "Dynamics of
Microbial Film Processes." Proceedings; First National
Svr
I U J_CL-L i. JL -l.lt! J- .i.w*^-'-*'-**—" 3 _ — cj
_ymposium/Workshop on Rotating Biological Contactor
Technology, Champion, PA, 1980, pp. 365-408.
Heidman, James A., et al., Carbon, Nitrogen, and Phos-
phorus Removal in Staged Nitrification - Denitrification
Treatment, USEPA 670/2-75-052, 1975.
Lue-Ling, Cecil, et al., "Biological Nitrification of
Sludge Supernatant by Rotating Discs," Jour. Water Poll.
Control Fed.,48, 1976, pp. 25-39.
Hitdlebaugh, John A., Phase I. Water Quality Engineering
- .ecial Study No. 32-24-0116-79, Sewage Treatment Plant
S£
24
Evaluation, Summer Conditions, Fort Knox. Kentucky, 14
August and 25-29 September, 1978, U.S. Army Envir.
Hygiene Agency, 1979.
Alexander, Martin and Clark Francis E. , Nitrifying
Bacteria," Methods of Soil Analysis - Part 2, C. A. Clark
(ed.), Am. Soc. Agron. , Madison, WI, 1965.
LaBeda, David P. and Alexander, Martin, "Effects of
on Nitrification in Soil," Jour. Environ.
523-526.
Rowe, R. , et al. , "Microtechnique for Most-Probable-
Number Analysis," Appl. and Environ. Microb., 33, 1978,
SO, and N0?
Qual.. 7, 1978, pp
800
-------
pp. 625-680.
39. Ghiorse, William C. and Alexander, Martin, "Nitrifying
Populations and the Destruction of Nitrogen Dioxide
in Soil," Microb. Ecol., 4, 1978, pp. 233-240.
40. Taylor, C. B., "The Nutritional Requirements of the
Predominate Bacterial Flora of Soil," Proc. Soc. Appl.
Bact., 14, 1951, pp. 101-111.
41. Stratta, James M. and Long, David A., "Nitrification
Enhancement Through pH Control with Rotating Biological
Contactors," Final Report, Institute for Research on
Land and Water Resources, The Pennsylvania State
University, 1981.
801
-------
SIMULTANEOUS NITRIFICATION AND DENITRIFICATION
IN A ROTATING BIOLOGICAL CONTACTOR
Sumio MASUDA, Yoshimasa WATANABE and Masayoshi ISHIGURO
Department of Civil Engineering, Miyazaki University,
Miyazaki 880, Japan
INTRODUCTION
In a Rotating Biological Contactor (RBC) process using a
fixed biological film, oxygen gas seldom penetrates into the
deepest part of the biofilm. Therefore, denitrifying bacteria
usually exist in the anaerobic and inner portions within the
biofilm of the RBC process. Denitrifying bacteria existing
in the inner anaerobic portion within the biofilm utilize
organic matter in the waste water as a source of organic .car-
bon. Nitrite or nitrate nitrogen produced within the aerobic
biofilm is partially converted to gaseous nitrogen (ft2 or N20)
by the denitrifying bacteria.
This phenomenon of simultaneous nitrification and denit-
rification (referred to as SND) sometimes undoubtedly occurs
in the RBC nitrification process. The authors have already
observed and reported this phenomenon in RBC pilot plants as
well as in a proto-type RBC system(l,2,3). However, it has
not yet been verified whether or not gaseous nitrogen is pro-
duced in the biofilm. Therefore, the authors have carried out
a series of experiments using a completely closed RBC unit to
investigate SND. The experimental variables were ammonia
loading, organic loading, ammonia concentration and mean resi-
802
-------
dence time. In this paper, the experimental results concern-
ing SND in an RBC are presented and discussed.
MATERIALS AND METHODS
The experimental apparatus consisted of closed-type reac-
tor and disks made of waterproof veneer boards. Fig.l shows
the single-stage RBC unit used in this experiment. The resi-
dence time distribution of water in the reactor without a bio-
film perfectly coincided with that of a single completely mix-
ed-flow reactor. In order to develope nitrifying bacteria on
the disk surface, artificial waste water (Table.l) containing
ammonia and inorganic carbon was fed into the RBC unit. Water
temperature and pH were fixed at 30°C and 8.0, respectively.
When the nitrifying biofilm developed, the artificial waste
water containing methanol as a carbon source for the denitri-
fying bacteria was added. After a week, the outer layer of
the biofilm consisted of heterotrophic bacteria, and a bio-
film consisting of the heterotrophic bacteria layer, a nitri-
fying bacteria layer, and a denitrifying bacteria layer was
formed.
At this point, an experiment was started to measure the
concentration of inorganic nitrogen (N0.3-K, TTO2-N and NH3-N)
in the effluent and the composition of the gas in the air
phase. After one Run was completed, the heterotrophic bacte-
ria layer was washed out by water jet. Then the same proce-^
dure was repeated by adding the artificial waste water,
depending on the experimental conditions shown in Table.2.
The nitrogen removal rate due to simaltaneous nitrification
and denitrification includes nitrogen utilized for the cell
synthesis of all bacteria concerned in the reaction.
RESULTS AND DISCUSSION
Conposition of the Gas in the Air Phase
Fig. 2(a), (b), (c) show the relationship between the
amount of gas in the air phase and the elapsed time after, the
vent holes were closed. Fig. 2(a) shows the relationship bet-
ween the reduction rate of oxygen gas in the air phase and the
concentration of ammonia in the bulk water. The reduction
rate of oxygen gas was influenced by organic loading.
In Run 3-1, when the amount of oxygen in the air phase
decreased to 500cc, ammonia appeared in the effulent water.
In this case, it is believed that the oxygen gas was not suf-
ficiently supplied to the nitrifying bacterial film, because
803
-------
Table. 1 composition of
artificial substrates
Conposition
NH3C1
NaHCO 3
NaCl
MgSOit 7H2O
KHaPOit
Cone, (mg/1)
382
1200
146
123
68
(When ammonia Cone, is 100 mg/1)
Table. 2 Experimental conditions
Run No
1-1
1-2
1-3
2-1
2-2
2-3
3-1
3-2
3-3
4-1
4-2
4-3
5-1
5-2
6-1
6-2
Flow rate
(cc/min)
5
5
5
12
12
12
24
24
24
48
48
48
10
10
20
20
NH3-N
loading
(g/m2d)
1
1
1
1
1
1
1
1
1
1
1
1
2
2
4
4
Organic
loading
(g/m2d)
3.5
7.0
9.0
4.9
7.3
14.7
6.0
8.0
14.0
5.0
8.3
11.3
8.3
14.0
14.6
26.0
NH3-N Cone.
(mg/1)
100
100
100
50
50
50
25
25
25
12.5
12.5
12.5
100
100
100
100
M R T
(hr)
13.0
13.0
13.0
5.5
5.5
5.5
2.8
2.8
2.8
1.4
1.4
1.4
6.7
6.7
3.3
3.3
804
-------
O
ff
1
ft
H-
O
H
(D
cn
(D
r • >
Disk number : 20
Disk diameter : 16 cm
Disk thickness : 0.5 cm.
Disk space : 1.0 cm
Disk area . : 0.82 m2
to
^J
o.
(Thermostat, Heater I
o
3
CD
hi
<
(D
3
ci-
ts1
O
I-1
, fl>
ft
H-
O
3
O
Hi
13
(D
l-i
H-
3
CD
rt
P>
d
3
H-
ft
1-3 ffi >-3 HH >
3" CD J31 H- H-
tD ft) CD iQ HI
rt K C
M
Hi
Mi
805
-------
Oxygen gas ( cc )
H
CD
M
P>
rt
H-
O
3
W
CC
tr
CD
IT
CD
CD
3
CD
P)
01
CD
PJ Oi
Hi £
ft" O
(D ft-
H H-
O
rt- 3
CD l-i
p)
< rt
CD CD
ft
p)
CD CC
0) CD
S3 CD
en
CD
en
CD
-N (mg/1) N2 gas ( cc )
Unknown gas (cc)
H
I-1 1
^-, cn N
CD
O pi co
~lp
CD 01
00
\ ^
806
-------
of the reduction in the partial pressure of oxygen. As the
organic loading increased in Runs 3-2 and.3-3, the oxygen fed.
to nitrification was not: considered to "be enough even when the
vent holes were opened. . Fig. 2(b) shows the relationship bet-
ween the cumulative amount of nitrogen gas.and elapsed time.
Fig. 2(c) shows the:relationship between elapsed time and
the cumulative amount of unknown gas, which could be nitrous
oxide (N20). The cumulative amounts of the total gas (sum of
nitrogen and unknown gas) were almost equal, but the composi-
tion of the gas was different in each Run. R.W.Dawson and
K.L.Marphy (U) argued that elemental nitrogen was the end
product of denitrification above .a pH of 7-3, while below a
pH of 7-3 nitrous oxide production increases. As pH was less
than 7-0 in Run 3-1, it seems that the predominant gas was
nitrous oxide.
Changes in the Water Quality .;,.,.,
Fig. 3(a), (b), (c) show the relationship between inorga-
nic nitrogens and elapsed time. Fig 3(a) shows the changes
in the water quality at an ammonia loadings of 1 g/m^d and an
organic loading of 3-5 g/m^d. The concentration of nitrate
decreased with the elapsed time. .On the other hand, the rate
of simultaneous nitrification and denitrification increased
with the elapsed time. It seems oxygen fed into the biofilm
for nitrification became insufficient to accomplish complete
nitrification. The relationship between the partial pressure
of oxygen in the air phase and the nitrogen removal rate due
to SWD under the same conditions is shown in Fig. U(a).
Fig. 3(b) shows the relationship at an ammonia loading of
2 g/m^d and an organic loading of 8.3 g/m^d. The concent-
ration of ammonia in the bulk water increased linearly with
the elapsed time. Before the experiment began (i.e., during
the time the vent holes were open), The partial pressure of
oxygen in the gas phase was 21%. Enough oxygen gas was supp-
lied to the biofilm to accomplish complete nitrification. As
time elapsed, the nitrogen removal rate decreased, because
there was a shortage of oxygen in the biofilm for nitrifi-
cation.
Fig. 3(c) shows the same relationship at an ammonia load-
ing of H.O g/m d and an organic loading of lU.6 g/m^d. The
bulk ammonia concentration increased with the elapsed time,
because the biofilm for organic oxidation became thicker com-
pered with that shown in Fig. 3(b). When the MRT was long
enough (i.e. 13 hr) and the concentrations of ammonia and
organic matter were low, the nitrogen removal due to SND in-
807
-------
H-
U)
(-3
O
Cone, (mg/1), S N D (%)
t->
o
ra
•a
w to
•&,.,
H-
UI
D-
M
to oo
O O
tt^ LTl O^ ~*J 00 ^O ^^
o o o o o o o
(D
tn
P-
3
ft
(D
Cone, (mg/1) , S N D (%)
oo
oo
cr i
3
(D
i-! cri
o
/
/
o
i
i
o
I
I
o
I
w
CO
w
tn
(D
O rt
O
3
Ln
I
\ ° PH °
Cone, (mg/1) , S N D (%)
I-1 to OJ *^
o o o o o
I—1
Ul CT> ~-J 00 VC -O
o o o o o o
808
-------
creased with the elapsed time. On the other hand, when the
concentrations of ammonia and organic matter were high and
the MRT was rather short, no change could be seen in the nit-
rogen removal rate due to SKD. The removal rate of nitrogen
due to SKD depended upon ammonia loading, organic loading,
MRT, and the pressure of oxygen in the gas phase.
The Relationship Between a Partial Pressure of .Oxygen and SED
Fig. if. shows the relationship "between a partial pressure
of oxygen in the gas phase and the nitrgen removal rate due
to SND. The experimental results shown in Fig. U can be
qualitatively explained by the biofilm models shown in Fig. 5.
The biofilm consists of a heterotrophic bacteria film for
organic oxidation, an autotrophic bacteria film for nitrifi-
cation, and an anaerobic bacteria film for denitrification.
An aerobic biofilm can be considered-to be much thicker than
the biofilm for nitrification or organic oxidation.
In the case of the attached biofilm shown in Fig. 5(b),
the biofilm for organic oxidation was so thick that both the
organic matter and DO were mostly consumed within it. Then
the,biofilm dominant for nitrification became very thin. On
the other hand, in Fig. 5(c), the biofilm for organic oxida-
tion was not so thick, therfore, DO penetrated more deeply
into the biofilm for nitrification. Therefore, the biofilm
for nitrification became thicker than that, in 5(b).
Fig. 5(c) shows the case in which the biofilm for organic
oxidation slightly covered the biofilm for nitrification. In
this case, DO completely penetrated into the biofilm for nit-:
rification. Fig. Ma) shows the case in which bacteria for
organic oxidation slightly covered the biofilm for nitrifica-
tion at an ammonia loading of 1 g/m^d and an organic loading
of 3-5 g/m^d. The nitrification rate was sharply reduced at
these loadings of ammonia and organics, when the partial
pressure of oxygen in the. air phase reached less than 10%.
On the other hand, when the partial pressure of oxygen increa-
sed to more than 10%, the nitrification rate became indepen- •
dent of the partial pressure.. However, the nitrogen removal
rate due to SND .increased with the decrease in the partial
pressure of oxygen in the air phase. This can be qualitai-
tively explained by the biofilm, model shown in Fig. 5(c). As
the partial pressure of oxygen in the air phase decreased, a
part of nitrification biofilm, became anaerobic.
In the case of Fig. U(b)f at an ammonia loading of 2 g/m^
d and an organic loading of 8.3 g/m^d, the biofilm for nitri-
fication became thicker than that in Ma). At these loadings
809
-------
The nitrogen removal rate due to
SND, Nitrification (»)
at _° H
rt o
p- PJ ^-~
§ -««= M
W "-"
tr
CD
CD
CD
3
P>
H
ft
fu P-
3 pj
O- M
rt *0
5* H
(D CD
CO
3 cn
H- C
ft H
K CD
O
vQ O
CD Hi
O
H X
CD ^<
S iQ
O CD
< 3
P)
l_i p.
3
K
P) ft
rt cr
CD CD
DJ iQ
C P)
CD 0)
ft >O
o er
0)
cn c/i
a CD
to
o
to
o
03
O
O
O
(Jl
to
o
13
O
\
Q
The nitrogen removal rate due to
SND, Nitrification (»)
to
O
a
—r~
oo
o
—I-
o
o
—r
\
Ul
i
The nitrogen removal rate due to
SND, Nitrification (%)
O
to
o
1
l<^
o
1
CTv
O
1
CD
O
c
c
to
O
• 0
•2
a
rt
p-
c
3
ft
p-
O
810
-------
hfj
H-
W
P-
O
Mi
H-
(0
DO Cone.
\\\\\\m\
V Auto-
\Heterc
X
Bulk
\\vv\\\-\
V
\Heter
\
Bulk
\ \ \ \ \ \ \ \ %
\
\
\
Bulk :
OC Cone.
>\\u\\\\ \
v
NH3-N
Disk
\\\\\ \\ >
L^A \\V\\\\\\\
Anaerobic biofilm
^trophic \ bacterial film
V \
troXphic \ bacterial film
\ \
\
\Diffusion layer
a o
O 0
liquid
\ V\ \ \ \ \ \ \ \ \ \ v \ \ t
V—
Biofilm
2
ffi
GO
1
2
\\\\\\\\\\\\\
Disk
Anaerobic biofilm
\ '• :
\ \ Autoterophic
\ \ bacterial film
3tro\phic \bacterial.film
\
\ •
\Diffusion layer
0 0
o n .
liquid
\\\\\\\ \\\\\ \\\ \\
Biofilm
§ . - -
CO
1
^ . .
1A^\\ \\\\\ \\V
Disk
Anaerobic biofilm
\ Autoterophic
\ \ bacterial film
\ \
\^eterbtr\opic bacterial film
V
N^iffusion layer
D O
O O
Liquid
2
ffi
U)
& . '•
Biofilm . ; '
i _
811
-------
of ammonia and organic, the nitrification rate became 100%,
when the pressure of oxygen in the air phase equaled 21%.
The nitrification rate decreased sharply because the penetra-
tion depth of oxygen became shallower as the partial pressure
of oxygen in the air phased decreased. This is explained by
the biofilm model shown in 5(b). The biofilm for organic ox-
idation became thicker and the biofilm for nitrification be-
came thinner compared with the biofilm shown in Fig. 5(a), so
that the penetration depth of oxygen became shallower than
that in 5(a). The partial pressure of oxygen in the air phase
decreased, DO did not penetrate deeply into the biofilm for
nitrification; and, therefore sufficient nitrification did not
occur.
Fig. U(c) is the case at an ammonia loading of ^.3 g/m^d
and an organic loading of lh.6 g/m^d. The biofilm attached to
the disks for organic oxidation became considerably, thicker in
comparison with 5(b) or (c), while the biofilm for nitrifica-
tion became very thin. As a result, the nitrification rate .
equaled HO%, which was an extremly low value. This experimen-
tal result can be explained by the biofilm model shown in Fig.
5(a). Both organic carbon and DO were mostly consumed within
the biofilm for organic oxidation, because the high concent-
ration of organic carbon the biofilm for organic oxidation be-
came quite thick. Therefore, the nitrogen removal rate due to
SND also decreased, because of the amount of nitrate or nit-
rite diffusing to the anaerobic biofilm decreased.
Fig. 6 shows one of the experimental results obtained in
the batch experiment using heterotrophic bacteria scraped from
the outer biofilm layer. This shows that the amount of the
nitrifying bacteria contained in the scraped biofilm can be
neglected,i.e. only organic oxidation will occuer in the
outer biofilm.
Fig. 7(Runs 1,2,3) shows the cases in which MRT and orga-
nic loading were changed at a fixed ammonia loading of 1 g/m2d.
Fig. 7(Run l) shows the case in which two organic loadings of
3-5 and 7-0 g/m^d were used in the experiment with a fixed in-
fluent ammonia concentrationof 100 mg/1 and MRT of,13 hrs.
The nitrogen removal due to SND became 100%, when the vent
holes were opened. It decreased with the elapsed time after
the vent holes were closed. The nitrogen removal rate due to
SND was sharply reduced, when the partial pressure of oxygen
in the gas phase decreased to less than 10%. On the other
hand, in the condition in which the partial pressure of oxygen
increased to more than 10%, the nitrogen removal rate due to
SND became independent of the partial pressure of oxygen.
812
-------
CH3OH , NH3-N ( mg/1
H-
o
i {a u>
2 O o
3 rt
H. p.
0) g
3 H- 0
DJ 3
S _,
fD g
< pj -—
CD 3
H O
Ml-1 (-"
C NJ
CO O °
0
if 3
fD O
OJ fD
S3- g
H- H! °
O 2J
3 rt ,_,
H- I"1
£§ §
H' p
3 co
(D
K)
O
M to
O O UJ ..
po oo g g
i i /rn i i
• o
/
- / ••'
/ /
/ / 0 ^
/ 0 / 0 S
/ a / s t-i
• "• o
/ 0 ^ W
a / £ w
/ 0
- • f'a
1 '
^ o
* t
\ /
/ / .. ..
/ /
/ f oo tn
* ° -o
/ ' Ln o
/ t
/ ; 1
1 sT^
A ' X.
' °«* H
•»
** •> ^
" •» ^
** o
I 1 I i i
O M to U) *» Ol
en — i oo
000
— i 1 r
o
S!
OJ
rt
fD -,
Hi J
rt
i
"g °?
\-<
S '
C *
t—t
!T> O
(D
U)
0 O
o
O
O
o
D O ( mg/1 )
813
-------
(D O
& O
C Hi
(D
O
rt X
O ><
iQ
cn o
Z 3
D
0> rt
H-
(1)
K
pj pj
3 en
D. rc
o <
H (D
P> K
3 C
H- IT.
a
ft
o e
D. 3?
H- H-
3 rt
IQ h!
O
n>
The
SND
nitrogen removal rate due to
, Nitrification (*)
The nitrogen removal rate due
SND, Nitrification (%)
The nitrogen removal rate due to
SND, Nitrification (%)
The nitrogen removal rate due to
SND, Nitrification (%)
O ^
I
N>
cn
CH3OH ( mg/1 )
•F 01
ro
CHsOH ( mg/1
to jr
O ~
to o
Hi
£d tn
C
3 V
• O -
ro o
co
CO
I
NJ
814
-------
In Runs 1-1 and 1-2, it seems that partial oxygen press-
ure of 10$ was enough to acomplish 100$ removal. The maximum
nitrification rate and SND were obtained, when the partial
pressure of oxygen became. 3%. Nitrification was rate-limiting
when the pressure of oxygen became less than 5$- Where the
partial pressure of oxygen increased to more than 10%, the
diffusion of organic matter to the anaerobic biofilm was rate-
limiting (organic matter limitation). On the other hand,
when the pressure of oxygen in the air phase decreased to less
than 5$, the diffusion of oxygen to the nitrifying bacterial
film was rate-limiting (nitrification limitation). The patt-
ern shown in Fig. 7(a), (b), (c) will be referred to as Patt-
ern A in this paper.
Fig. 7(Run 2) shows the case in which two organic load-
ings of 7.0 and 10 g/m2d were used with fixed conditions of
an ammonia concentration of 50 mg/1 and a MRT of 5-5 hrs.
Run 2-1 also adhered to Pattern A. When the pressure of oxy-
gen in the gas phase became 10$, organic matter appeared in
the effluent water. In Run 3-3, organic matter in the efflu-
ent water did not appear until the pressure of oxygen reached
12$, but the nitrogen removal rate due to SND increased
sharply below a pressure 8$. SND changed from organic matter
limitation to nitrification limitation at a partial oxygen
pressure of 10$. Nitrification and SND showed the same de-
creasing pattern, when the partial pressure of, oxygen in the
air phase became less than 10$. The pattern shows in Fig 7-
(d), (e) will be referred to as pattern B.
Fig. 7(Run 3) shows the case in which two organic load-
ings of 8.0 and 10 g/m2d were used with a fixed condition of
an ammonia concentration of 25 mg/1 and a MRT of 2.8 hrs. In
Run 3-1, on condition that the pressure of oxygen in the air
phase reached 20$, SND became organic matter limitation. On
the other hand, when the partial pressure of oxygen in the
air phase decreased to less than 20$, it became nitrification
limitation. In Run 3-2, nitrification and SND show the same
decreasing pattern. The pattern shown in Fig. 7(f) will be
referred to as Pattern C. As explained above, the patterns
of SND can be classified into three types depending on the
experimental conditions.
The Relationship between the Concentration Ratio of Methanol
to Ammonia and SND
The relationship between the concentration ratio of meth-
anol to ammonia (i.e., the C/N ratio) and the nitrogen removal
rate due to SND is shown in Fig 8. In this experiment, orga-
815
-------
100
90
1 *80
O C!
^•Svo
-a 40
o G
S id
$ 30
C
I 20
S-l
i)
•H 10
C
^ o,
[_ • o
•» W L-**«^B1Bii^^ Q ""
^^•^^ (2) o
- O:NH3-N loading 1 g/m2d ^'X^^ ^ ^\
. • :NH3-N loading 2 g/m2d /C NO
/ ^).
O :NH3-N loading 4 g/m2d / x
: Nitrification fS
' : SND /
/
O
/ C)v.
/ *l-*— ,**•» ^Ss^.
S ^ ^x.
/ ^ ^s
"/ ^ \
\ x
3 \
0 n
^ *•" \
-•— ""' 1 1 -i i— .. i i i , ,
5 6
C/N ratio
8
10 11
Fig.
8 The nitrogen removal rate due to SND
versus C/N ratio
816
-------
nic loading ranged from 3-5 to 36 g/m2d. Organic loadings of
3-5, 7, 9, and 10 g/m2d were used in the experiment with an
ammonia loading of 1 g/m2d. At these loading rates of ammonia
and organic matter, a nitrification rate equal to 100$ was ob-
tained, even when the C/N ratio was changed from h to 7- How-
eyer, when the C/N ratio became more -than 7, the nitrification
rate became lower. At the same loading rate, the nitrogen re-
moval rate due to SND increased significantly with an increase
in the C/N ratio. The maximum nitrogen removal rate due to
SND was ofbtained at a C/N ratio equal to about 9- Organic
loadings of 8.3, lU and l8g/m2d were used in the experiment at
an ammonia loading of 2 g/m2d. The, nitrification rate sharply
decreased. The nitrogen removal rate due to SND increased
significantly with the increase in the organic loading and the
maximum nitrogen removal rate due .to SND was obtained at C/N
ratio of around 8.. Organic'loadings of iV. 6, 28 and 36 g/m2d
were used in the experiment with a fixed ammonia loading of
k.3 g/m2d. The nitrification rate decreased with the increase
of C/N ratio. The' nitrogen removal rate due to SND became very
high as the C/N ratio increased and the maximum nitrogen remo-
val rate due to SND was obtained-at a, C/N ratio of 7-
Fig. 9 and 10 show the relationship between the C/N ratio
and-SND, when sodium formic acid and ethylene glycol were
added as" organic carbons. A nirtogen removal rate of 20% due
to SND was obtained even at a C/N ratio of zero, because a
landfill leachate (Haginodai) was used for the raw waste water.
When sodium formic acid was-added as an organic carbon, the.
nitrogen removal rate of 80% due to SND was obtained at a high
C/N ratio of 35-
These results demonstrate it was difficult for the hete-
rotrophic and anaerobic bacteria to use the 'sodium formic
acid. On the other hand', in the case where ethylene glycol
was added as organic carbon, a nitrogen removal rate o.f 90%
due to SND was obtained at the low C/N ratio of h. Based on
the experimental data, the authors belive the effectiveness of
the organic carbon as the carbon source of SND depends on the
diffusivity and biodegradibility of the organics which influ-
ence the distributions of heterotrophic, nitrifying, and de-
nitrifying bacteria concentrations in the biofilm.
Simultaneous Nitrification and Deriitrification in a Proto-type
RBC Plant
Since Novenber 1976, a Rotating Biological Contactor has
been treating leachate from the Miyazaki Citiy Haginodai Land-
fill (5). The concentration of total nitrogen in the effluent
817
-------
s
100
80
(0
4J
0)
>
o
6
3
C
O
-H
4J
60
C
0)
Cn
o
V4
•y
•H
c
•H 40
"M
4J
a 20
-Q O —
O: Nitrification
• : SND
Ammonia loading :
Hydraulic loading :
Water temperature :
0
10 15
C/N ratio
20
25
30
35
Fig. 9 The nitrogen removal rate due to SND versus the
elapsed time ( Sodium formic acid as organic carbon)
o •LUU
4J
0)
3
•o ^80
Q) d?
-P •—
a
"260
r-i -r-{
n3 -P
> (0
O O
a) in 40
i-l -H
C J->
Q) -H
tr> a 20
•P Q
•|H ^4
cw n
"" " O /-\
° ^ -^- ~n
S^ ' ^
•
/
/
/ 2
/ Ammonia loading : 1 g/m d
1 Hydraulic loading : 27.5 l/m2d
I Water temperature : 30°C
I
-
/
/
1
K*
i " 1
C/N ratio
10
15
Fig. 10 The nitrogen removal rate due to SND the elapsed
time ( Ethylene glycol as organic carbon )
818
-------
H-
ua
H
(D
g
o
cn t-i
rt 0)
Oi rt
vQ fD
fD cn
H- O
3 Mi
01 W
fD 3
O
tr MI
H- O
O H
>-> H-
fD 3
pj fi
O H-
rt <
O H-
l-j Qj
d
The nitrogen removal rate due to
SND ( % )
o
cn
rt
Oi
cn
rt
cn tr
rt <
Oi Ui
cn
rt
M
O
~r
NJ
o
Ul
o
T~
o
~r
Ul
o
CJl
o
~-J
o
fD
H
PI
rt
Oi
3
to
Ui
OJ
to
O
oo
819
-------
The nitrogen removal rate due
to SND ( % )
to
CTl
o
CD
rt
CD
to
CD
0)
rt
CD
CD
(D
o <
(!) CD
P, hi
H- cn
C3 0
iQ cn
PJ cn
w a
pj
CTl
00
DJ
rt
CD
CD *»
(Ti
d CD
I-!
CD —
n
to
to
to
to
V*,
\
820
-------
water has constantly been less than 10 mg/1. In this experi-
ment, the nitrogen removal rate due to SND was obtained from
aerobic RBC.
Fig. 11 shows the cumulative nitrogen removal rate due to
SND. The nitrogen removal, rate due to SND increased at higher
temperatures. SND occurred significantly in the first two
stages but hardly occurred at all in the latter stages.
Fig. 12 shows the relationship between SND and water temp-
erature, with hydraulic loading as a parameter. SND became
significant at higher temperatures, and lower hydraulic load-
ings. Hydraulic loadings of less than 60 to TO 1/m^d should
be used in this plant to obtain a. nitrgen removal rate due to
SND greater than hd% at 20 °C.
SUMMARY AND CONCLUSIONS
The phenomenon of SND in a RBC process was confirmed by
measuring nitrogen gas production. The SND in a RBC was ex-
perimentally studied in terms of. mean residence time, organic
loading, water temperature, and the partial pressure of oxygen
in the air phase. Most of the conventional biological proce-
sses for nitrogen removal consist of a series of unit pro-^
cesses which perform organic oxidation, nitrification, and de-
nitrification separately. On the other hand, in the case of
nitrogen removal using the RBC nitrification process, all the
organic oxidation, nitrification, and denitrification could be
accomplished in the same reactor. The obtained results are
summarized as follows:
1. In a closed RBC for nitrification, nitrogen gas in-
creased in the air phase, while oxygen gas decreased.
2. The nitrogen removal rate due to SND depended upon
the MRT, C/N ratio, water temperature, partial pre-
ssure of oxygen, and ammonia concentration. When
the other parameters were fixed, the optimum C/N
r at io; existed for the maximum nitrogen removal rate
due to SND.
3. The SND patterns were classified into three types
depending on the experimental conditions.
References
ISHIGURO, M., WATANABE, Y. and MASUDA, S. "Advanced
Wastewater Treatment by Rotating Biological Disk
Unit (ll)" Journal of Japan Sewage Works Association,
Vol.lU,No.l52,pp.32-Ul,Jan.(l977)
821
-------
MASUDA, S., ISHIGURO, M. and WATA1ABE, Y., "Nitro-
gen Removal in a Rotating Biological Contactor (l)"
Journal of Japan Sewage Works Association,Vol.l6,
No.l87,pp2U-32,Dec.(1979)
ISHIGURO, M., WATANABE, Y. and MASUDA, S. , "An
Basic Investigation of Aerobic Denitrification by
Rotating Biological Disk Unit" Presented at the 33rd
Anual Conference of JSCE.,pp213-2lU,Oct.(1978)
Dawson R. N. and Murphy K. L., "The Temperature
Dependency of Biological Denitrification" Water
Research Pergrame Press.,Vol.6,pp.71-83 (1972)
ISHIGURO, M., WATANABE, Y. and MASUDA, S., "Treat-
ment of Leachate from Sanitary Landfill by Rotating
Biological Contactor", Environmental Conservation
Engineering,Vol.7,No.6,PP.513-521, June.(1978)
822
-------
DEVITRIFICATION IN A SUBMERGED BIO-DISC SYSTEM
WITH RAW SEWAGE AS CARBON SOURCE
Rusten. Division of Hydraulic & Sanitary
Engineering, The University of Trondheim, Norway.
Hall yard 0degaard. Division of Hydraulic & Sanitary
Engineering, The University of Trondheim, Norway.
INTRODUCTION
An investigation of denitrification by biofilms has
been performed in two submerged bio-disc units with municipal
sewage as carbon source. The project is based upon the pro-
cess that the authors presented in the Proceedings of The
First National Symposium on Rotating Biological Contactor
Technology (1). The flow sheet of the process is shown in
Figure 1. Nitrate-rich effluent is recycled to the inlet of
an anoxic tank where denitrification takes place with raw
municipal sewage as carbon source.
In this paper results are presented from a later more
thorough study of the denitrification part, in order to esta-
blish the basic design-criteria for the process. The goal
was to find the denitrification rate, - temperature dependency,
- pH dependency, - energy consumption and - alkalinity pro-
duction.
EXPERIMENTAL ARRANGEMENT
The experiments were carried out in two plexi-glass bio-
disc units arranged in parallel . The units called RBC A
823
-------
r = rQ CNr
Figure 1. Proposed process for nitrogen removal in biodisc
plants.
and RBC B were identical, each containing one 300 mm diameter
disc with a total biofilm area of 0-15 m2. The discs were
rotated at 1.7.'5 rev./min., which equals a peripheral velocity
of 27.5 cm/sec. The tank volumes were 10.5 litres. The units
were equipped with double walls. About 24 hours prior to samp-
ling, the inner walls were removed, scraped, washed and rein-
stalled. This was done to avoid the wall-growth effects,
that otherwise can be significant in pilot-scale plants.
Ahead of each RBC unit a tank for pH and temperature
control was installed.
The pH was adjusted by means of an automatic dosing equip-
ment, adding sulfuric acid or sodium hydroxide according to
signals given by pH-electrodes. The temperature was adjusted
with water from water-baths circulating in copper tubing. In
addition the RBC units and the tanks for pH and temperature
control were heavily insulated. The flow sheet of the experi-
mental set-up is shown in Figure 2.
The raw sewage was presettled before entering the raw
water tanks. Samples were analysed for influent and effluent
NO -N, N02-N, alkalinity and.SBOD5. In addition SCOD
ana DOC were measured during the temperature and pH runs.
Flow rates, temperature, pH and dissolved oxygen concentrati-
ons were observed for each run. SBOD5 was chosen as the main
parameter for measuring the organic content, since it was be-
lieved to be the parameter that best describes the organic part
available for the denitrifying organisms.
824
-------
Nitrifier
Nitrified
water
Raw-
water-
tank
acid base
acid base
pH- and temp.control
Deni trif ier
Figure 2. Experimental set-up.
825
-------
Initially both RBC units were run at 15 C and pH 7. In
the temperature- and pH dependency experiments, temperature and
pH in RBC A were varied. RBC B acted as a reference unit at
constant temperature and pH.
RESULTS AND DISCUSSION
Denitrification rates
Several factors are influencing the denitrification rate,
such as oxygen concentration, pH, temperature, carbon source,
nitrate loading and organic loading.
Denitrification has to be carried out at anoxic conditions.
It seems as if true anaerobic conditions in the liquid is not
necessary, and that 1-2 mg 02/£ does not influence denitrifica-
tion in biofilms (2,3). In this experiment the oxygen concen-
trations were usually below 0.4 mg 02/&, and never exceeded
1.0 mg 02/£ . Thus oxygen is not considered to be a limiting
factor for the denitrification rates found in this study.
Temperature and pH were held constant. (15°C and pH 7).
Various proportions and amounts of raw water and nitrified
water were fed to the denitrifiers, in order to cover a broad
range of hydraulic, organic and nitrate loadings.
When modelling a denitrification system many researchers
use a Monod relation, applying excess methanol and taking nit-
rate as the limiting substrate. Our experiments showed that
the denitrification rates were very dependent upon the SBOD5
concentration. Supposing that the RBC units are complete-mix
reactors, the denitrification rates should be a function of
effluent NO -N and SBOD5 concentrations. Adding a "Monod"
term to the steady-state removal equation given by Kornegay
and Andrews (4) gives:
826
-------
,0,-,
(sw*-» + si «o«V \KS
SBOD5
SBOD5
i no*-*
(1)
where Q = hydraulic flow rate (L3/T)
0 HO--* '
1 SBOOe
= influent and effluent concentration of NOX-N ,
"°«"" respectively (M/L1)
•effluent concentration of SBOD, (M/L3)
fl= maximum specific growth rate (T )
y = yield
A = area of biological film (L2)
X = concentration of organisms in the biological film (M/L3)
d = thickness of the active biological layer (L)
= saturation coefficients for NOX-N and SBOD
respect i ve 1 y (M/LS )
S HOu-N • K S SBOD
5»
Assuming that y, Y, X and d are constants, which is only
partly true/ Equation 1 can be rewritten to give the deni-
trification rate:
'S NO.-II Sl NOv-»
1 SBOD,;
KS S80D5 + Sl SB005
(2)
where R^ ^ = deni trification rate (M/L2T)
C = constant (M/L2T)
Least square regression was used to find the equation
that best described our results (5). A lot of functions,
including the one described by Equation 2, were tested.
The results for the model with best fit and for the Monod-
type model are listed in Table I,
827
-------
Table I. Denitrification rate at pH 7 and 15°C.
Rate expression , mg/m2h
/ NO,-N \
nNOx-H 1-7<- ^ 335.3 + L^_J 50 SB005
and RHOX-H ^ LKOX-N
Degr. of
freedom
96
94
Resi-
dual
mean
square
4486
13950
r for RKOX-H
obs. versus
RNOx-N pred.
0.9166
0.7216
(3)
(2)
where L
S
S
S
NOX-N
NOX-N load , mg/m2h
0 SBOD,
= influent SBODs concentration , mg/l
= effluent SBODg concentration , mg/l
H0 _„ = effluent NOX-N concentration , mg/l
828
-------
Denitrification rates are plotted in Figure 3. Also
shown are the rates predicted by Equation 3. Figure 4 and
Figure 5 show observed and predicted denitrification rates
using "the best fit model" and the Monod-type model, respecti-
vely.
Using municipal sewage with variations in both composition
and strength, it is not possible to attain steady-state condi-
tions. Influent SBOD5 concentrations covered a range from 9
to 80 mg/£ and influent NO -N concentrations varied from 5.35
to 20.80 rog/£. In addition we used theoretical hydraulic
residence times varying from 0.93 to 3.98 hours, with 1.15
hours as a typical value. As shown in Figure 4, equation 3
describes reasonably well the observed denitrification rates.
This means that in our RBC reactors the denitrification was
influenced mainly by the influent organic strength (SBOD5)
and the nitrate load.
The Monod-type model shows considerably more spread in
the denitrification rates (Figure 5).One of the reasons may be
that our system was not steady-state. Equation 2 gives us the
saturation coefficients, l
-------
O
CO
CO
-t-J
c:
Ci)
•r- O
(1)
(D
O
oo
Q.
Q-
(O
CD
-M
O CD
oo
ra -Q. O
-O 4J
fO <1) tO
co n3 cr
>
(D
o
03 T- -I-
T3
rcf
CO
830
-------
Z 800 -
T 600 —
600 800
RKU _, obs. mg NOx-N/m'h
Figure 4. Observed versus predicted denitrification
rates for the model with best fit (Equation 3)
831
-------
200
600 800
R obs. mg NO -N/m2h
Figure 5. Observed versus predicted denitrification
rates for the Monod-type model (Equation 2)
832
-------
The relative denitrification rates in RBC A were deter-
mined according to Equation 4:
RD - "Rate RBC A"
' - ~. "Rate RBC B"
The Arrhenius relationship is often used to describe
temperature effects:
(4)
RT =-• A-e — : (5)
Equation 5 was rewritten to give an Arrhenius plot:
RD=k.e-E/RT
In (RD) = -E-
+ In k
(6)
Using least square regression (Figure 6) we found an energy of
activation of 39 670 J/mole. Figure 7 shows the Arrhenius
equation and the observed values for the relative denitrifica-
tion rates.
Temperature dependencies can also be expressed by:
T+10
(7)
or:
-0
'T2 ~ 'T1
Results from this study are listed in Table II.
(8)
833
-------
o
-*'
(aa)
CO
o
4-
•r—
s_
CU
-o
o
CD
s_
4J
fO
S_
U)
zs
01
834
-------
o
ffl
y
T 1.80
1.40
1.20
1 .00
0.80
0.60
0.40
0.20
4771
10
15
J_
_L
20 25
Temperature °C
Figure 7. The effect of temperature on the rate of
denitrification.
835
-------
Table II. Temperature dependency.
Temp, range °C
'10
0
15 - 25
10 - 20
5-15
1.74
1.78
1 .82
1 -057
1 .059
1 .062
Murphy et.al (7) reports an activation energy of 69 300
d/mole_for a submerged RBC, using methanol as carbon source
Inis gives a greater temperature dependency than we found in
our study. Davies and Pretorius (8) observed a great drop in
?n lC]B1Sr e^W ]n C' reP°rt1n9 a Qio-value of 1..38 between
10 and 30°C and a Q10-value of 13.06 between 5 and 10°C
Our experiments show that denitrification with municipal
sewage as carbon source can be achieved down to 5°C For de-
sign, a temperature coefficient (e) of 1.06 should be appro-
priate. ^K
pH dependency
•k
RBC B was kept constant at pH 7, and the relative denitri-
fication rate at this pH is put equal to 1.00. In RBC A the pH
was increased in increments of 0.5 pH-units up to pH 10. Then
the unit was acclimatized at pH 6.5 before a gradual decrease
to pH 5.
Visual observations showed that the disc started to loose
excessive amounts of sludge above pH 9 and below pH 6. When
the inner walls were taken out for wash, they showed no siqn
nL9r-Vith/VHu12 and pH 5' The ^nitrifying organisms were
not killed at pH 5. pH 4.9 gave a relative denitrification
rate of 0.11 (Figure 8). The following day we observed a
relative rate of 0.30 at pH 5.2, which is a pretty good reco-
Vt- 1 y •
This study demonstrates that the optimum pH for denitri-
fication lies between 7 and 8-5, which is in agreement with
the results reported by Davies and Pretorius (8)
836
-------
1.2
1.0
5 0.8
OJ
cc
f
0.6
0.4
0.2
I L
10
pH
Figure 8. The effect of pH on the rate of denitrification.
837
-------
Energy consumption
The energy needed for denitrification can be expressed
by the substrate consumption ratio, defined as mg organic
matter consumed/mg NO -N removed (9). Figure 9 shows a plot
of NO -N removed versus SBOD5 consumed, giving a substrate
consumption ratio of 2.4 mg SBOD5 consumed/mg NO -N removed.
Using the regression equations in TableIII,the substrate
consumption ratio in our study can also be expressed as 4,6
mg SCOD/mg NO -N or 1.6 mg DOC/mg NO -N.
X A
Table III.Correlation between different organic parameters.
Regression equation
SBOD5
SBODS
» 0,523 SCOD-11.4
= K524 DOC-12.9
Number of
observations
138
133
Corr.
coeff .
0.9611
0-9370
Narkis, Rebhun and Sheindorf (10) have published results
from suspended culture experiments where different carbon
sources (methanol, sodium acetate and chemically treated raw
sewage) were used. They concluded that by expressing the or-
ganic matter as SBOD5, a critical value of 2.3 mg SBOD5/mg
NO -N existed when 100% denitrification was to be reached
regardless of what carbon source was used.
Monteith et.al (9) have investigated different carbon
sources. They found substrate consumption ratios between 0.7
and 2.6 mg DOC/mg NO -N removed. Methanol had average values
of 5.41 mg SCOD/mg NOX-N and 1 .17 mg DOC/mg NO -N, The sub-
strate consumption ratios are influenced by the presence of
dissolved oxygen and the carbon requirements for cell synthesis,
A carbon source with high substrate consumption ratio, would
tend to generate larger volumes of sludge, which is undesir-
able.
The municipal sewage used in this study has a substrate
consumption ratio well inside the range found for other car-
bon sources (9,10). When a resirculation system (Figure 1)
is used, influent dissolved oxygen concentration may increase,
giving a slightly higher substrate consumption ratio. Using
a ratio of 2.4 mg SBOD5/mg NO -N removed will therefore be a
X'
838
-------
I 40
8 35
m
25 -
20 -
15 _
10 -
Y = 2.4058X
N = 115
, r = 0.9097
_L
10 12
14 16
mg NOX-N/1 removed
Figure 9. Energy consumption for denitrification,
839
-------
safe assumption when calculating the organic load on the
nitrifier.
• A •
Alkalinity production
Denitrification produces alkalinity. Theoretically 1
mole NOJ-N removed gives 1 mole OH~. This corresponds to an
increase of 0.0714 meq/mg NOa-N removed.
Figure 10 shows the alkalinity production. It is found
to be 0.0713 ..meq/mg NO -N removed, which is very close to the
theoretical value, Usifig a RBC pilot-plant with methanol as
carbon source, Smith and Khettry (11) observed a gain in al-
kalinity of 0.074 meq/mg N03-N removed (3.7 mg CaC03/mg
NOa-N).
SUMMARY AND CONCLUSIONS
Denitrification studies have been performed in two sub-
merged bio-disc units, using presettled sewage as carbon
source.
The conclusions are as follows:
1. The denitrification rate at 15°C and pH 7 could be de-
scribed by the equation:
''U
= 12.72
NO -
335.:3+L
NO _
X
SQ SBQD , adding the limit
NO -N
A
where
NO -N
X
= denitrification rate, mg/m2-h
NO -
A
= NOX~
load'
= influent SBOD5 concentration, mg/SL
2. For the system shown in Figure 1, denitrification can be
regarded a 0. order reaction with respect to NO -N concen-
tration, and a 1. order reaction with respect to" SBOD5
concentration.
840
-------
4J
o
Z3
-a
o
Q.
fd
S:
cu
uoijDnpojd Ajtuij
841
-------
3. Temperature dependency was modelled using an Arrhenius
equation. This gave an activation energy of 39670 J/mole,
corresponding to a temperature coefficient (0) of about 1.06.
4. Optimum pH lies between 7 and 8-5.
5. The energy consumption for denitrification has been found
to be 2-4 mg SBOD5/mg NO -N removed.
P\
6. The alkalinity production has been found to be 0.0713 meq/
mg NO -N.
X
ACKNOWLEDGMENT
The authors wish to thank The Royal Norwegian Council for
Scientific and Industrial Research for financial support.
Also we want to thank G0ril Thorvaldsen for assisting in some
of the analytical work.
842
-------
LIST OF SYMBOLS
0 =
ys,
Vf =
d .=
k =
r =
A =
A =
C =
DOC
E =
KS =
temperature coefficient
maximum specific growth rate
thickness of the active biological layer
constant
correlation coefficient
area of biological film
frequency factor in the Arrhenius equation
constant : ,
= dissolved organic carbon
activation energy
saturation coefficient
-N - NO -N load
/\
NOg-N
NO -N
A
Q =
R =
RNO -
RD =
S1 -
SBOD5
SCOD
T =
X =
Y =
= nitrite nitrogen
= nitrate nitrogen
- Z N02-N + N03-N •-
hydraulic flow rate
gas constant (8.314 J/mole-°K)
ft = denitrification rate >
denitrification rate at temperature T
relative denitrification rate
influent substrate concentration
effluent substrate concentration
= soluble 5-day biochemical oxygen demand
= soluble chemical oxygen demand (Cr)
absolute temperature
concentration of organisms in the biological film
yield
843
-------
REFERENCES
1
8,
10,
11
. 0degaard, H.,and Rusten, B., "Nitrogen Removal in Rotating
Biological Contactors Without the Use of External Carbon
Source", Proceedings of the First National Symposium on
Rotating Biological Contactor Technology, Champion,
Pennsylvania, Feb. 4-6, 1980, pp. 1301-1317.
. Christensen, M.H., and Harremoes, P., "Biological Denitri-
fication of Sewage: A Literature Review", Progress of Water
Technology, Vol. 8, No. 4/5, pp. 509-555, 1977.
. Cheung, P.S./'Biological Denitrification in the Rotatinq-
Disc System", Water Pollution Control, Vol.79, No.3
pp. 395-408,1979.
, Kornegay, B.H., and Andrews, J,F., "Kinetics of Fixed-Film
Biological Reactors", Journal of Water Pollution Control
Federation, Vol.40, No. 11, Part 2, pp. .R 460-R 468,1968.
Dixon, W.J., and Brown, M.B., "Biomedical Computer Programs,
P-SeriesJ977", Health Sciences Computing Facility, Depart-
ment of Biomathematics, School of Medicine, University of
California, Los Angeles, 1977.
Requa, D.A., and Schroeder, E.D., "Kinetics of Packed-Bed
Denitrification", Journal of Water Pollution Control Feder-
ation, Vol. 45, No. 8, pp. 1696-1707, 1973.
Murphy, K.L., et.al., "Nitrogen Control: Design Consider-
ations For Supported Growth Systems", Journal' of Water Pol-
lution Control Federation, Vol. 49, pp. 549-557, 1977.
Davies, T.R., and Pretorius, W.A., "Denitrification With a
Bacterial Disc Unit", Water Research, Vol.9, pp.459-463,
I Z7 / D •
Monteith, H.D., Bridle, T.R., and Sutton, P.M., "Industrial
Waste Carbon Sources For Biological Denitrification",
Progress of Water Technology,Vol.12, Toronto,pp.127-141,
1980.
Narkis,N., Rebhun,M.,and Sheindorf,Ch., "Denitrification
at Various Carbon to Nitrogen Ratios", Water Research,
Vol.13, No.1, pp.93-98, 1979.
Smith,A.G., and Khettry,R.K., "Nitrification/Denitrifica-
tion Studies With Rotating Biological Contactors",Proceed-
ings of the First National Symposium on Rotating Biological
Contactor Technology,Champion,Pennsylvania,Feb.4-6,1980,
pp.1319-1341.
844
-------
OPERATION OF A RETAINED BIOMASS NITRIFICATION SYSTEM FOR
TREATING AQUACULTURE WATER FOR REUSE
D._ E. Brune — Department of Agricultural Engineering,
The Pennsylvania State University, University Park,
Pennsylvania.
R. Piedrahita - Department of Agricultural Engineering,
University of Calilfornia, Davis, California. , •;
ABSTRACT
A series of experimental trials were conducted in which
a variety of polyurethane materials of differing pore size
were evaluated.as a media for nitrifying .filters used in
treating water in a trout hatchery. These filters were found
to be capable of ammonia oxidation rates ranging from 80-180
mg—N/day/liter of filter volumes. These rates represent an
order of magnitude increase over removal rates in conventional
rock filters operating at influent ammonia levels at or below
0.5 mg/liter. These rates were, however, 50 to 100% lower
than the rates observed for a similar filter design operating
under laboratory conditions. The difference was attributed
to increased heterotropic fouling experienced during the
field operations. The optimum design for the field unit was
found to be an initial stage of open pore media to accept the
heterotropic loading followed by a second stage of fine pore
material to allow for complete nitrification. Such filters
may be operated at detention times less than four minutes
achieving well over 90% ammonia removal.
845
-------
INTRODUCTION
In the majority of cases in which low levels of ammonia
or nitrite must be removed from water supply systems, one of
two techniques have been utilized, either biological nitrifi-
cation or breakpoint chlorination. Gauntlet (1) compared
biological nitrification to breakpoint chlorination as a means
of treating water for potable supply. He suggests that the
disadvantages of relatively long contact time required for
chlorination, compounded by interferences from organic compounds
and potential production of dichloramines, trihalomethanes•and
other organochlorine compounds, makes nitrification a more
desirable means of low level ammonia removal from potable water
supply. Short (2) suggests that the fluidized bed nitrifi-
cation process is cheaper than breakpoint chlorination when the
ammonia levels to be removed exceed 0.2 mg/1.
Brune and Gunther (3) suggest that biological nitrifi-
cation could, in fact, be used to economically remove low
levels of ammonia from recirculating aquatic animal culture
facilities. Studies conducted by Gunther et. al (4) indicate
that ammonia levels ranging from 0.05 to 0.50 mg/1 could be
expected in such systems. The lower limit on these concen-
trations appear to favor breakpoint chlorination.
Unfortunately, however, the dangers of chronic toxicity to
fish from dichloramines or the possibility of acute toxicity
from accidental chlorine overdose makes this an undesirable
system for treating aquaculture reuse water.
In spite of the reduced rate of biological nitrification
at these low ammonia levels, these sytems can be made to
operate at high efficiency if the lower bacterial growth rate
can be compensated for by carrying much higher levels of total
biomass within the filter units, and at the same time, using
a filter media which permits high water passage rates. Brune
and Gunther (3) proposed the use of a "Retained Biomass Filter"
for such systems. These filters consist of submerged, downflow
columns of polyurethane cubes or sheets contained within ridge
cells of plastic netting. When tested under laboratory
conditions, such filters were capable of oxidizing ammonia at
rates of 100-400 mg-N/day/liter of filter volume as compared
to 10-30 mg-N/day/liter for conventional aquaculture filter
des igns.
846
-------
The purpose of this study was to further examine the
behavior of these retained biomass filters. In particular,
this study was directed at examining the success of three
such filter units under actual field conditions and to select
the appropriate media pore size yielding optimum performance.
METHODS AND MATERIALS , :
This study consisted of three separate trial runs with a
bank of filter units installed at the University of California
- Davis trout hatchery previously described (4). Figure 1
illustrates the three filter units and the,containment building.
Figure 2 shows the placement of the filter units in relation to
the trout culture and water storage tanks and rapid sand
filters.
Each of the filter units consisted of a plexiglass tube 2
, ft. deep by 1 ft. in diameter, sealed at each end with a
removable plywood section. During the runs, the tubes were
filled with varying types and configurations of polyurethane
material.
Water exiting from the trout culture tanks would first
pass through the solids settling tank (with approximately 2
minute detention time) and then be distributed to each of
the three filter units. Flow rates to the individual filter
units were controlled by 1/2 inch plastic valves and ranged
from 3-7 liters/minutes giving filter detention times averag-
ing 3-4 minutes. Each of the filters units was equipped with
7 sample ports at 2 inch intervals for sampling purposes.
Total pressure drop across the filters was limited to 4 inches.
In the first experimental trial, each of the 3 units was
filled with a different media type consisting of: 1) 1/2 inch
square cubes cut from a dense polyurethane material with pore
size of 0.1 mm. The cubes were loosely packed in six two inch
deep ridge cells fabricated from "conwed" plastic netting
material, 2) 1/2 inch square cubes in a similar arrangement but
cut from a lighter polyurethane material of approximately 0.15
mm pore size, 3) a similar arrangement except the light poly-
urethane material (0.15 mm) was arranged as uncut circular
sheets 1 inch thick stacked 24 deep. These filter units were
operated for approximately three months. Inlet ammonia levels
ranged from 0.25 to 0.40 mg/1 depending on fish loading and
water flow rates.
847
-------
In the second experimental run, three different pore sizes
were evaluated. These were a 2.0 ram, 1.3 ram, and 0.5 mm pore
size in an open mesh polyurethane material ("Scott" brand
industrial foam obtained from Wilshire Foam Products, Inc.,
Carson, California). In all cases, the material was distri-
buted in the columns as 12 inch uncut circular disks 1 inch
thick with 2 disks per ridge cell with a total of 24 disks.
The ridge plastic screen was designed to prevent compaction
of the filter media during operation. These filters were also
operated for a period of 3 months with inlet ammonia levels
ranging from 0.25 to 0.50 mg/1.
In the last experimental run, the three filter units were
operated with a mixture of pore sizes: 1) arrangement A
consisted of 100% 2 mm pore size, 2) arrangement B consisted
of one half 2 mm pore size with one half of the 0.5 mm pore
size, 3) arrangement C consisted of one third 2 mm, one third
1.3 ram, and one third 0.5 mm pore sizes. In all cases, the
filter media were arranged with largest pore sizes toward the
top of the column. This last run lasted 2 months and ammonia
concentrated ranged from 0.10 to 0.35 mg/1.
During each run, water samples were routinely taken from
inlet, outlet, and across the columns. Samples were taken
every 3 to 7 days depending on operating conditions. Filter
performance was evaluated by monitoring ammonia removal rates
across the columns. Ammonia concentrations were determined by
the phenolhypochlorite method of Solorzano (5) with modifi-
cation of Liddicoat et. al (6). Reagent grade hypochlorite
solution was used to make up the oxidizating solution as out-
lined by Solorzano. Liddicoat et. al (6) suggested that more
consistent reagent blanks could be obtained by substituting
potassium ferrocyanide as the catalyst in place of sodium
nitroprusside. This modification of the phenolhypochlorite
method was used in the present study.
Water samples were assayed within one hour of sampling.
Preliminary tests showed that they lose about 1% of their
ammonia per hour when stored at 18°C., No correction factor
was used, however, in the calculation of ammonia concentration.
Samples were simply assayed as soon as possible. A calibration
curve was made for determining the ammonia concentration.
The standard deviation of a 25 x 10~6 molar (0.35 mg NH3-N/1)
ammonia sample was +_ 9% using one centimeter round curvettes
and a B & L Spec 20 spectrophotometer. Nitrite was measured
colorimetrically by the method of Strickland and Parsons (7).
848
-------
RESULTS AND DISCUSSION
Figure 3 and Table 1 give data previously presented by
Brune and Gunther (3) obtained from a laboratory nitrifying
filter. Figure 3 shows that these units using a 6 inch down-
flow filter with a 0.15 mm pore size media were capable of
ammonia oxidation rates as high as 250 mg-N/day/liter, and that
these removal rates could be correlated reasonably well to the
influent average ammonia concentration, if separated into
detention time groups. In contrast, typical field filters used
in aquaculture operation (Kramer, Chin, & Mayo; Table 1)
operated at rates an order of magnitude lower. In fact, the
previous laboratory results represented removal rates more
closely approaching rates observed by Haug & McCarty (9) in
filters operating at high ammonia input levels.
In comparison, the removal rates observed f.or the first
series of field units (Figure 4) in this study averaged only
around 40 mg-N/day/liter with total ammonia removal of 60-70%
across the entire length of the column. The reason for this
lower removal was obvious from visual inspection of the
filters. The retained bimass filter operates on the princi-
ple that essentially 100% of the accumulated bacterial biomass
is retained within the filter media. When these bacterial
levels become restrictive to the water flow rate, the filter
bed is washed, thus the biomass levels in the filter are
externally controlled, rather than depending on sloughing
rate, as in the case of conventional rock filters. Although
the low pore size material permitted high nitrification
rates and high flow rates in the laboratory, this was not
the case in the field operation. The smaller pore operating
under field conditions rapidly became clogged from hetero-
trophic growth as a result of the low levels of dissolved
organics (~ 10 mg/1) and finely divided particulates present
in the trout water. As a result, frequent washing of the
filter media was required, drastically reducing the total
nitrifying biomass. The 0.5-1.3 mm pore was judged in-
appropriate for field use and the run was abandoned after
3 months operation. However, the data (Figure 3) does give
a good indication of the length of time required for filter
start-up. As can be seen, approximately two months were
needed to establish a completely nitrifying filter. The
filters would typically rise to a high rate of nitrifi-.
849
-------
cation as the bioraass levels increased, thus requiring washing
of the media and afterward weekly washing of the media with
removal rates stabilizing around 30-50 mg-W/day/liter.
Figure 5 illustrates the performance of the 3 larger pore
size polyurethane media (see Figure 6 for a media comparison).
These three filters were seen to come up to full performance a
month earlier than the previous run, most likely, a result of
high level of bacteria in the system tanks and piping. Also,
these filters because of their more open nature, were able to
perform at a higher biomass level giving removal rates around
150 mg-N/day/liter and total ammonia reducations of 70-99%
across the filter depth. The finer pore filter media showed
greater fluctuations in removal rates since it tended to
develop higher biomass levels followed by a need for more
thorough cleaning. Figure 7 shows that removal rates could
again be correlated reasonably well with influent ammonia
levels.
When fully loaded at influent ammonia levels of 0.40 to
0.50 mg/1, these filter media were able to perform at 140-180
rag—N/day/liter. This rate is again roughly an order of
magnitude higher than previous fixed bed nitrifying filters
used at these low levels. However, these rates are only 1/2
of the rates observed in the laboratory. The primary differ-
ence between the laboratory and field performance is again the
fouling of the filter due to the added heterotrophic bacterial
loading.
In comparing the 3 pore sizes, one can see (Figure 8) that
the 0.5 mm media was much more effective in removing ammonia
requiring only 1/5 of the filter depth to oxidize the majority
of the inlet ammonia, while the 2 mm media required the full
depth to achieve significant removal. Perhaps, just as import-
ant is the lower level of nitrite reduction in the 2 mm media
(Figure 9). Although the output level of nitrite from the 2 mm
media was low in relation to the total ammonia and nitrite
oxidized (Figure 10), these levels are still reason for concern
since low levels of nitrite also present a chronic toxicity
problem to fish.
850
-------
Finally, Figures 11, 12, and 13 show data taken from run
three in which the media types were mixed. Filters B and C
represent the most successful modification. As seen in
Figure 11, removals ranged from 80 to 120 rag-N/day/liter.
This removal efficiency is at the same level as the previous
runs when consideration is given to the reduced influent
ammonia levels experienced during these trials as a result of ;
lower fish loading during this period (Figure 14). Figure 12
shows the same correlation between filter performance and
average influent ammonia levels as previously demonstrated.
The important advantage of the combination of pore sizes
in the filter were: 1) the filter could be maintained at
high flow rates with only a weekly or biweekly washing of the
filter media since the heterotrophic biomass was carried in
the upper, more open pored layers and 2) a high level of nitri-
fication including more complete nitrite removal could be
maintained in the lower, smaller pored media.
851
-------
SUMMARY AND CONCLUSIONS
As a result of a series of pilot tests conducted in a
typical trout culturing facility, a more optimum design of
a biological nitrifying filter for treatment of low level
ammonia laden water has been achieved. The data obtained
in this study indicates that a reasonably high rate of
ammonia and nitrite oxidation (140-180 mg-N/day/liter) can
be achieved with an approximately 3-4 minute detention time
when using a filter design consisting of 6-10 inches of 2
mm open polyurethane followed by 6—10 inches of 0.5 mm
polyurethane filter media. These filters act as complete
biomass capture systems and require weekly to biweekly solids
removal for optimum performance.
The initial larger pore section of this filter acts as a
partial nitrifying and complete organic reduction unit, while
the second smaller pore section allows for more complete
nitrification. The two stage design allows for a sustained
high water passage rate requiring a minimum of filter washing.
Additional work should be conducted to develop a low cost,
automatic design for controlling the biomass washing of these
filter media in large scale units. Secondly, a filter media
material is needed that is more resistant to degradation.
After one year of operation all polyurethane media used in
this study had undergone severe degradation, so much so, that
they would have needed replacement had the experiments con-
tinued for a longer period of time.,
852
-------
ACKNOWLEDGEMENTS
This research was made possible by funds provided by the
University of California - Davis Aquaculture Program. At the
time of this study D. E. Brune was Assistant Professor of
Agricultural Engineering at TJCD. R. Piedrahita was a Research
Assistant in Agricultural Engineering at UCD.
853
-------
REFERENCES
Gauntlett, R. B., "Removal of Ammonia and Nitrate in the
Treatment of Potable Water," In Biological. Fluidized Bed
Treatment of Water and Wastewater, Edited by P. F. Cooper
and B. Atkinson, Ellis Horwood Limited, 1981.
Short, C. S., "Removal of Ammonia from River Water,"
Technical Report, Medmenham, Water Research Centre, 1975,
52 pages.
Brune, D. E., and Gunther, D. C., "The Design of a New High
Rate Nitrification Filter for Aquaculture Water Reuse,"
Journal of the World Mariculture Society, In press, 1981
Edition.
Gunther, D. C., Brune, D. E., and Gall, G. A. E., "Ammonia
Production and Removal in a Trout Rearing Facility,"
Transactions of the American Society of Agricultural
Engineers, Vol. 24, No. 5, pp. 1376-1380, 1981.
Solorzano, L., "Determination of Ammonia in Natural Waters
by the Phenol Hypochlorite Method," Limmol. Oceanogr.,
14:799-801, 1969.
Liddicoat, M. I., Tibbitts, S., and Butler, E. I., "The
Determination of Ammonia in Sea Water," Limmol. Oceanogr.,
20:131-132, 1975.
Strickland, J. D. II., and Parsons, T. R., "A Practical
Handbook of Sea Water Analysis," Fisheries Research Board
of Canada, Ottawa., 310 pp., 1972.
Kramer, C. and Mayo, R. D., "A Study for Developing of
Fish Hatchery Water Treatment Systems," A report prepared
for the Walla Walla District Corps of Engineers, 1972.
Haug, R. T. and McCarty, P. L., "Nitrification with
Submerged Filter," Journal WPCF, 44(11):2086-2103, 1972.
854
-------
10
Inlet
Ammonia
Levels
(Mg/1)
Kramer,
BTF
UF
AUF
DDF
SPDF
Run 1
Run 2
Run 3
Run 1
Run 2
Run 3
Run 1
Run 2
Run 3
, 0.45
0.84
0.80
1.7,7
0 .82
Haug &
8.0
8.0
7.7
Forster
1.0
1.0
1.0
Brune &
0.5
0.5
0.5
Loading
Rate
Mg-N/D/L
Chin, & Mayo
37
85
73
90 .
46
McCarty (9)
1536 .
768
370
50
100
200
Gun the r
100
240
500
Removal
Rate
Mg-N/D/L
(8)
-' 11
17
- 25
30
10
1045
595
350
49
92,
170
99
220
450
Detention
Time
(Min.)
7.5
15.0
30.0
28.8
14.0
7.2
11.9
5.1
3.0
Temp,
°C
11
17
25
30
10
17.6
14.2
15.9
28.2
25 .8
Variable
Variable
Variable
Variable
Variable
25°
25°
25°
26
26
26
Variable
Variable
Variable
Table 1: Comparison of Various Filter Performances Under
Differing Operating Conditions. (From (3)).
855
-------
11
Plywood
housing
Sampling
ports
Ridge
plastic
netting cell
Inlet from
settling tank
- Water level
2" Plexiglass
filter unit
o overflow
outlet
Figure 1. Experimental filter units.
856
-------
12
o
857
-------
AMMONIA REMOVAL , mg/day/1
858
-------
AMMONIA REMOVAL , mg-N/day/1
t
CD
4>-
O
0.
H-
O
<3
P
l-h
O
i-i
(D
l-h
H-
8
CO
co
§'
8
859
-------
Pore Size
••••••2.0mm
"™~* 1.3mm
—-0.5mm
10 20 30 40 50 60 70
TIME , Days
Figure 5. Ammonia removal in large pore filters,
860
-------
» *" J Aw •"* f *^ ^-V # ^
^^, A >M W, «. «i*r ^ •**• t
*"***&*&*'.
t-!
CD
Ml <
H- P
H- H
rt H-
•8 §
(t> 13
PL o
H- M
p> <<
• rt
g"
X
0)
(a
o g
rt P
C rt
P CD
K P
P- I-1
N CD
(D
^ C
• cn
8.
Hi
O
861
-------
AMMONIA REMOVAL, mg/day/1
g 8 8 $ $ ? §
n>
^
Q
P-
bl O*
-------
2.0mm
FILTER DEPTH, Inches
Figure 8. Drop in ammonia concentration
across three different pore
size filters.
863
-------
o>
| .05-
2 .04-
W x-v-,
o .03-
o
0 .02-
LU
h*
S
2 .01-
t
2 i
2.0mm
.•**^^^ ^^w.
.^ X
// 1.3mm >.
•*» A
__
^^s*~ ^***+^^ 0.5mm
5 10 15
FILTER DEPTH , Inches
Figure 9. Levels of nitrite across
three different pore size
filters.
864
-------
0.50-n
PORE SIZE
.......
2.0 mm
— 1.3mm
— 0.5mm
T —T 1 T— F* T
10 20 30 40 50 60 70
TIME , Days
Figure 10. Levels of outlet nitrite
from large pore filters.
865
-------
AMMONIA REMOVAL, mg/doy/1
§ 8 o
C
n
(D
O
0
§
(U
0
8-
f5S
S. o
3 °
a w £
tf- O
Ml
H-
rt
n>
8-
0>
o
8 $
866
-------
200-i
180-
160-
140-
100-
80-
60-
40-
20-
Mixed Medio
® Unit A
® Unit B
® Unit C
Average detention time
3.6 minutes
0.2 0.3
AVERAGE INLET NH3-N , mg/1
0.4
Figure 12. Relationship between ammonia removal
and inlet levels in mixed media filters.
867
-------
NITRITE CONCENTRATION, mg/1
P 2 P P O O
P Q O Li L.
I 9 I 8 8
t
ro
u>
P
rt
(0
3
H-
rt
H
H-
rt
H-
0)
O
O
3
O
(D
3
rt
fu
rt
s-
3
§
8,
I?
H-
H
rt
8
en
•) (5)
: I
c c c
O CD J>
X
CD
Q.
Q.
o*
868
-------
INLET
AMMONIA CONCENTRATION , mg/1
OQ
d
to
M
.£>
H
(D
rt
£B
O
M
cn
P4
OQ
rt
(D
rt
fD
n>
rt
i-i
t-i
§
CO
869
-------
Nitrified Secondary Treatment
Effluent by Plastic-Media Trickling Filter
Jiumrn Min Huang and Yeun C. Wu
Department of Civil Engineering
Polytechnic Institute of New York
& University of Pittsburgh
Alan Molof
Department of Civil Engineering
Polytechnic Institute of New York
INTRODUCTION
Ammonia nitrogen plays a vital role in the synthesis of
microorganisms in the secondary treatment of wastewater.
Generally, there is abundant ammonia nitrogen present in muni-
cipal effluent streams. Through the process of nitrification,
this excess nitrogen in the wastewater treatment plant effluent
consumes the dissolved oxygen in the receiving water. This,
along with the ammonia toxicity, creates unhealthy conditions
for aquatic life.
Nitrification is a two-step biological process in which
ammonia nitrogen is oxidized to nitrite and further to nitrate
as shown below:
NH,
Step one
Nitrosomonas
NO,
Step two ..
~ } -
Nitrobacter
Nitrosomonas and nitrobacter are commonly the most responsible
for either of the two oxidation steps. These organisms are
classified as autotrophic or chemolithotrophic because they
870
-------
obtain carbon source from dissolved carbonate and use oxidiza-
ble substrate as their source of energy for growth and metabol-
ism. The two steps for the nitrification transformation can be
written as follows:
NHJ-+ 1.5 02 + 2 HC03
Step one
2 N02 +2
N0~ + 0.5 02 Step two -> 2 N03
(1)'
•(2)
In addition, the overall reaction is
NHj + 2 02 + 2 HC03
N03 + 2 H2C03
Biological nitrification of a secondary effluent can be
accomplished by both suspended- and fixed-growth processes
(1-3). However, due to many advantages associated with the
latter process, more studies have been made recently in the
fixed-growth systems (4-7). Also, the use of light weight
synthetic media allows the fixed-film biological filter
to be constructed much deep thus minimizing space requirement
and maximizing loading capacities. The synthetic media
has an increased specific surface area for greater biomass
attachement that results in a lower solids production in
the plant effluent.
The main objective of this study is to investigate the
feasibility of using plastic-media trickling filter for the
removal of nitrogen from the effluent of activated sludge
wastewater treatment plant. The efficiency of the trickling
filter plant was evaluated under different nitrogen loading
rates. And the relationship between alkalinity destruction
and nitrogen removal was also studied. The settling charact-
871
-------
eristics of the nitrified sludge was determined by sludge
volume index (SVI).
EXPERTEMNTAL PROCEDURE
Two laboratory-scale trickling filter towers were construct-
ed in the Environmental Engineering Research Center at the Univ-
ersity of Pittsburgh. The filters were 6" squares and 8' long.
A synthetic plastic media manufactured by Munters was employed
for the study. The media model is Munters Biodek 19060, which
has a surface area 140 m2/m3 or 44 ft2/ft3 and a void ratio of
greater than 95%. Plant specifications and media structure
are shown in Table 1.
The influent for twin towers, connected in series, was
stored in a 250 gallons tank. The influent feed solution was
taken from the effluent of an existing activated sludge pilot plant
located in the same laboratory. The desired amounts of ammonia
nitrogen (NH.C1) and sodium bicarbonate (NaHCOg) were added
to the solution and mixed in the feed tank.
The wastewater was pumped to the plastic-media trickling
filters by two variable speed pumps. The wastewater was first
fed to the top of filter No.l and the effluent was discharged to
clarifier No. 1. And then a second pump taken the feed from
the above mentioned clarifier, continuously pumped the waste-
water up to the top of filter No. 2 where it traveled down
and into clarifier No. 2.
Effluent samples were collected twice each week. Sampling
ports were located 2 ft, 4 ft, 5.5 ft, and 7 ft from the top
of filter media in each tower. Controlling parameters included
NHo-N, N09, NOV alkalinity, pH, dissolved oxygen (DO), suspend-
ed solidsf BOD; and SVI. NH3-N, N02, and NOs were measured by an
Orion Digital lonalyzer. Other parameters such as TSS and VSS,
alkalinity, BOD were performed in accordance with "Standard
Method" (8). The pH and DO were measured by Orion pH meter and
YSI DO meter.
RESULTS AND DISCUSSION
1. Influent Wastewater Characteristics. Table 2 is a list
of influent feed conditions proceeding into first filter. The
range of flow rates tested varied from 45 to 115 gallons per day.
The highest NHs-N concentration was 109.4 mg/1 while the lowest
concentration was 43,6 mg/1. The influent nitrite and nitrate
concentration ranged from 3.2 to 12.4 mg/1 and the alkalinity
fluctutated between 315,7 mg/1 and 750 mg/1 as CaCOs. The pH
range was small and it varied between 7,7 and 8.3. The dissolved
872
-------
Table 1. Trickling Filter Plant Specifications
Parameters
Filters #1 and #2
(A). Reactor Dimension:
Size
Height
Vol ume
Gross
Media
(B). Filter Media:
Type
Surface Area
Void Ratio
(C). Media Configuration:
6" x 6"
7'
2.0 ft3
1.75 ft3
Hunters Biodek 19060
44 ft2/ft3 (140. m2/m3)
> 9555
873
-------
Flow /"
Distribution
Plate
Effluen
sA
Figure 1. Trickling Filter Pilot Plant
For Nitrification Study
874
-------
Table 2. Influent Feed Conditions
Flow Rate Influent Concentration, ma/1 .-•'••
Gal. /Day
78.6
49.3
102.8
49.4
98.5
78.8
99.6
78.8
45.0
49.9
98.0
115.0
57.5
57.7
: r::-:3-N NC2+N03 DO
74.18
109.40 •
.50.1
79.3
. 39.3
45.5
26.0
28.4
54.1
43.4
19.5
15.2
13.6
12.6
6.6
6.1
7.3
12.4
9,6 '
6.4
9.1
8.1
11.0
6.6
9.2
5.1
7.2
3.2
"4.8
4.7
6.1
5.0
8.8
4.5
7.7
5.3
5.4
4.8
8.9
• 4.8
4.9
5.2
Alkalinity pH Temperature, °C
as CaCOa
431.0
750.0 ,
368.0
650.5
315.7
460.0
323.2
373,6
536.9
457.0
311.2
366.4
425.4
366.4
8.2
8.0
7.8
7.9
7.7
8.0
7.7
, 8.1
7.9
8.2
7.7
8.3
7.9
8.2
24.0
: 20..0
24.0
21.5
22.0
23.0
22.0
23.1 '
20.6
22.0
22.0
24.4
23.5
2.6.0 -
* BOD < 25 mg/1
875
-------
oxygen content of the feed was within the level of 4.7 to 8.9 mg/1
The pilot plant was operated at room temperature, 20.6 to 26.0°C.
2. System Start-Up. At the beginning of the study,
the trickling filters were operated separately at the flow
rate of 42.6 gallons per day and 56.4 gallons per day. Figure
2 shows the NHo-N and N02+ N03 concentrations at different
filter depth with respect to days of operation for each filter.
From the figure it can be seen that it required 60 days
for one filter and 67 days for the other filter to remove
the NHo-N down to approximately 1.6 mg/1 at a filter depth of
6 feet? Also, from the same figure, it is apparent that the
NHo-N concentration decreases as the filter depth and start-
up days increase. In addition, Figure 2 further shows that
nitrification increases with filter depth and start-up time
because there is an increase in N02 + N03 concentration.
As already pointed out, the start-up time for obtaining a
steady-state operational condition in trickling filter nitri-
fication process was long with no seed organisms used. In order
to shorten the start-up time, flow recirculation through the
filter with seed organisms should be employed.
3. Trickling Filter Plant Performance. The ammonia nitrogen
conversion to nitrite and nitrate as a function of filter depth
under different nitrogen loading conditions are shown in Figures
3, 4, and 5. It is apparent from these figures that as the
filter depth increases, the NHs-N concentration decreases with
increasing the production of N02 and NO, in filter 1. The same
relationship continuously existed in fitter 2 until the net
accumulation of nttrite became decreased. As a result, the
rate of nitrate arose sharply but the removal of NH3-N became
insignificant.
The nitrogen loadings presently empolyed for the operation
of both filters 1 and 2 are shown in tables 3 and 4. The
nitrogen loading was calculated based on the kg NH3-N used per
tower surface area In m2 per day (kg NH,-N/mz-day). According-
to tables 3 and 4, the nitrogen loading varied from 0.120 kg/m
-day to 0.959 kq/m2-day in filter 1 operation and varied from
0.016 kg/m2-day to 0.530 kg/m2-day in filter 2 operation. Since
filter 2 treated the effluent of filter 1, the nitrogen loadings
were lower.
The reduction of ammonia nitrogen in the plastic-media
trickling filter system is certainly dependent upon the rate of
nitrogen loading. Trickling filter 1 showed NH,-N removal up
to 86.4% for a nitrogen loading of 0.120 kg/m2-8ay. As the
loading was increased to 0.959 kg/m2-day, the % NH3-N was red-
876
-------
Flou •= 56.4 gallons/dav
5 60
Height, ft
Figure 2. Trickling Filter Start-up
156
877
-------
100
Filter *2
fV- Cumulative
4560 2
Height, ft
5 . 6
Figure 3. Ammonia Conversion as a Function
of Nitrogen Loading at 0.281 and
0.395 ko/m2-day, respectively.
878
-------
560
Height, ft
45 6
Figure 4. Nitrogen Conversion as a Function
of Nitrogen Loading at 0.421 and
. '0.632 kg/m2-day, respectively.
879
-------
Figure 5. Anmonia Conversion as a Function
of Nitrogen Loading at 0.631 and
0.875 kg/m2-day, respectively.
880
-------
Table 3. Trickling Filter 1 Effluent
Characteristics
Nitrogen
Loading
Kg/m2-day
0.959
0.875
0.837
0.630
0.632
0.578
0.421
0.361
0.395
0.349
0.311
0.280
0.129
0.120
Effluent
,NH3-N
rag/1
41.4
62.0
24.7
29.3
15.0
15.0
7.0
5.8
14.0
7.7
3.3
2.5
1.8
1.7
Removal
44.2
43.3
50.6
63.0
62.0
67.0
73.1
79.3
74.1
82.2
82.9
83. 1
86.7
86.4
+ N03
41.75
58.18
32.93
63.12
25.43
31.02
20.65
32,23
55.18
27.55
16.23
16.37
15.6
13.02
PH
8.0
7.8
7.5
7.8
7.7
7.8
7.8
7.8
7.7
8.0
7.7
8.2
7.7
8.1
DO
mg/1
4.1
3.1
6.2
4.6
7.2
4.0
7.3
4.7
•4.3
4.2
7.2
4.6
4.5
4.6
TSS
7
10
9
6.6
11,3
3.6
5.3
4.6
13.7
3.0
20.0
2.7
3.3
6.2
VSS
mg/1
4
9
7
4.9
8.6
2.3
3.0
2.7
7.0
2.5
11.0
2.2
2.6
2.5
881
-------
Table 4. Trickling Filter 2 Effluent
Characteristics
Nitrogen
Loading
kg/m -day
0.530
0.495
0.413
0.240
0.232
0.191
0.113
0.074
0.102
0.060
0.053
0.047
0.017
0.016
Effluent
NH3-N
mg/1
16.9
21.5
8.43
2.72
3.45
2.00
0.33
0.45
1.48
0.36
0.12
0.02
0.02
0.02
% NH3-N
Re"oval
59.1
65.3
65.9
82.0
83.2
86.6
95.2
92.3.
89.4
95.3
96.2
98.8
98.8
98.8
N02
+ NO,
mg/r
59.2
84.9
36.4
86.0
33.0
37.5
25.9
35.5
57.1
38.6
23.7
24.5
17.8
15.5
pH
8.0
7.7
7.6
7.6
7.7
8.0
7.7
8.2
7.8
8.2
7.7
8.5
8.1
8.2
DO
mg/1
4.4
3.3
6.8
5.3
7.5
4.3
7.7
4.9
4.9
4.4
7.4
4.9
4,8
4.6
TSS VSS
ing/1
4.0
11.0
8.0
14.3
21.3
3.8
9.0
1.6
7.3
2.5
6.0
2.0
2.0
4.0
3.0
10.0
6.5
8.3
13.3
2.3
4.6
1.0
6.7
2.0
3.0
2.0
1.8
1.5
882
-------
uced to 44.2. Trickling filter 2 achieved a higher nitrogen
removal efficiency, ranging from 59.1% to 98.8%, due to a lower
loading condition. According to the present study, two 6-foot
filters connected in series are capable of obtaining an effluent
NH3-N Concentration less,than 1.0 mg/1 at a loading of 0.42 kg/
m2-day for first filter and of 0.113 kg/m2-day for second filter.
It is clear that the effluent NHs-N remaining depends not only
the nitrogen loading but also the initial concentration of NH3~
N in the wastewater. It is expected that the effluent NH3-N
concentration will be higher if the influent concentration of
NH3-N and the rate of nitrogen loading are higher.
Table 5 summarizes the nitrification data obtained from
the present study and the other investigations using rotating
biological contactors. The nitrogen loading used in this case
was calculated based on pounds of NHo-N applied per plastic media
surface area in ft2 per day (Ib NH^-N applied/1000 ft2-day).
Table 5 shows that the nitrification of secondary treatment
effluent by trickling filter is comparable to or even better
than that achieved by rotating biological contactors.
By comparing the pH and DO of the influent and effluent
of filter 1, it is observed that both of these parameters fell
slightly as the wastewater traveled through the filter. These
results are normal because of nitrogen oxidation and destruct-
ion of bicarbonate alkalinity. The effect of nitrogen loading
on alkalinity destruction will be discussed later in detail.
Recovery of pH and DO was obatined after the wastewater pass-
ing through filter 2. The pH increase was very slight while
the DO increase was more substantial. It seems obvious that
the amount of oxygen dissolved in the wastewater is greater
than the quanity required for biological nitrification in fil-
er 2. No oxygen deficiency was found during the entire period
of this study. Both effluents of filters 1 and 2 contained
small concentrations of total and volatile suspended solids in
accordance with tables 3 and 4.
The effect of filter height on the plant performance under
various nitrogen loading conditions is shown in Figure 6. It
can be seen in Figure 6 that for all loading conditions present-
ly investigated the percentage of NhU-N removal increases as
the filter depth also increases. But such increase was reduced
markedly when the nitrogen loading exceeded 0.8 kg/nr-day in
trickling filter 1. The results for trickling filter 2 are
opposite. At the low nitrogen loading condition, the % removal
of NH,-N from the filter 1 effluent was only slightly affected
by the filter depth measured at 2, 4, 5, and 6 feet below the
top of plastic media. The influence of filter depth on trickling
filter 2 performance becomes profoundly after the nitrogen load
exceeding 0.5 kg/m2-day.
883
-------
Table 5. Nitrification of Secondary Effluent
By Trickling Filter and Rotating
Biological Contactors
Nitrogen
Ib NH3-N
o
1000 ft*
Loading Rate
Applied
- day
Trickling Rotating
Filter Biological
/ *-* \
(a)
0.067
0.072
0.176
0.185
0.198
0.206
0.225
0.241
0.332
0.362
0.364
0.482
0.504
(br
0.20
0.40
0.50
0.60
0.61
0.76
0.78
1.10
1.29
Contactor
( ^(10)
0.08
0.13
0.18
0.21
0.22
0.42
0.48
0.83
Ammonia Nitrogen
Reduction
(*)
Trie! king Rotating
Filter Biological Contactor
(a)
99.9
99.8
99.3
98.9
98.9
97.2
98.4
98.7
95.6
83.2
93.1
80.3
77.1
(b)
95.0
90.0
80.0
90.0
65.0
13.0
73.0
25.0
35.0
(c)
94.4
70.3
83.1
88.3
91.7
76.8 :
83.1
70.3
884
-------
Removal of Influent NH--N
.—'
rf &
ST8*:?.
T3
fD CD
3 fD
O r+
< X
a> n
—• m
rt z
O r*
^3
-*iiQ
ss
% Removal of NH,-N Remaining in
the Effluent of Filter #1
885
-------
4. Alkalinity Requirement. The pound of alkalinity
destoryed per pound of ammonia nitrogen oxidized to nitrate
normally equals to 7.2 (11). A plot of pound of alkalinity
consumed per pound of NH3-N removed is shown in Figure 7.
It is clearly shown in Figure 7 that the alkalinity require-
ment for trickling filters 1 and 2 is different. And the
normal ratio of 7.2 was not observed.
Trickling filter 1 shows a ratio of approximately 6.7
up to a nitrogen loading of 0.50 kg/m2-day. As the loading
exceeded the above mentioned value, the pound of alkalinity
utilized per pound of ammonia nitrogen oxidized became smaller.
For instance, at the nitrogen loading of 0.959 kg/m2-day,
the ratio is 4.6 Trickling filter 2 shows the same resulting
curve as filter 1. According to Figure 7, the ratio varied
from 7.8 to 4.25 as the nitrogen loading increased from 0.016
kg/nr-day to 0.53 kg/m2-day. The ratio decrease became so
apparent when the nitrogen loading exceeding 0.25 kg/m2-day.
It also appeared that the effect of nitrogen loading on
alkalinity demand for nitrification in the trickling filter
system was significant, in particular, when the nitrogen Ipad
exceeding 0,5 and 0.25 kg/m^-day in the first and second filter,
respectively. Although the organic content in the wastewater
was low (<25 mg/1 as BOD), ammonia nitrogen was utilized by both
nitrifying bacteria and heterotrophic organisms due to the fast
feed rate. This is the reason to explain why the alkalinity
demand per unit amount of nitrogen consumed was low at high nitro^
gen loading condition.
5. Sludge Settling Characteristics. Nitrifying organisms
produced from the-fixed-film trickling filter system have an
excellent property in settling.1 Although the SVI was high
between 199 and 319 during the strat-up period, its value
varied only slightly around 110 at all times after the system
reached the steady-state condition. Normal sludge floe is
shown in Figure 8. The size of the floe was large and it
can be separated from liquid phase quickly. However, regular
microscopic examination of filter sludge occasionally found
that long-length filamentous microorganisms existed. It is
believed that these filaments are transferred from the effluent
of the activated sludge plant instead of being developed from
the trickling filters. The morphological structure of the
long-length filaments is shown in Figure 9.
886
-------
Filter
- 9
•8
S.
i
•«
-------
Figure 8. Normal Trickling Filter Sludge
(200 X)
888
-------
Figure 9. Filamentous Growth in Trickling
Filter Sludge (200 X)
889
-------
CONCLUSIONS
Biological nitrification of secondary effluent by a
two-stage trickling filter was throughly investigated under
the optimum pH and temperature conditions. It was found
that the efficiency of the trickling filter plant was a
function of influent nitrogen concentration, nitrogen loading,
and filter depth. The quantity of ammonia nitrogen removed
was higher tn filter 1 than filter 2. The continuously
oxidized ammonia nitrogen resulted in the accumulation of
nitrite and nitrate in the system, however, the production
of nitrite started to decrease with increasing the nitrate
when ammonia nitrogen remaining in the wastewater approached
to its lowest level. More than 93% of ammonia nitrogen can be
removed by passing it through two six-foot plastic media trickl-
filters, operated at the nitrogen loading equal to 0.63 kg/m2
-day in the first filter and 0.24 kg/m2-day in the second
filter.
Both pH and DO decreased and increased after the waste-
water traveled through filter 1 and filter 2, respectively.
Dissolved oxygen never became the growth-!imitina factor because
its concentration was over 3.0 mg/1. The effluent total and
suspended solids were extremely low and the sludge settle-
ablity was very high with an averaged SVI = 110. Fila-
mentous microorgansims was seldom found in the sludge and they
were probably transferred to the filter from the effluent of
the activated sludge plant. That is why the outgrowth of fila-
mentous microorganisms never occurred.
The alkalinity requirement for fixed-film biological
nitrification is somewhat different from that observed from
suspended growth systems. The quantity of alkalinity destory-
ed per unit of ammonia nitrogen removed is always below normal
value of 7.2 in both, filters presently investigated if the
nitrogen loading was kept to exceed 0.25 kg/m2-day. Both
nitrifiers and heterotrophic microrganisms play an important
role in nitrogen assimilation when the feed rate and the
nitrogen load are high. This explains why the alkalinity
demand for biological nitrification is lower under the above
mentioned condition.
REFERENCE
1. Poduska, R. A., and Andrews, J. F., "Dynamics of Nitrifi-
cation in the Activated Sludge Process," Jour, of Water
Poll. Control Federation, Vol. 47. pp.2599, 1975
690
-------
2. Downing., A. L., et al., "Nitrification in
the Activated
Vol. 2, pp.130,
10.
11.
Rotating Bioloc
February 2-4,
O'Shaughnessy,
Sludge Process," Jour. Inst. Sew. Purif.
1964. "
Lijklema, L., "Model for Nitrification in Activated Sludge
Process," Environ. Sci. & Techno!., Vol. 7, pp, 428, 1973.
Stenquist, R, J,, Parker, D. S., and Dosh, T. J., "Carbon
Oxidation-Nitrification In Synthetic Media Trickling Filter,"
Jour, of Water Poll. Control Federation, Vol. 46, pp. 2327,
1974.
Ito, K., and Matsuo, T., "The Effect of Organic Loading on
Nitrification In RBC Wastewater Treatment Processes,"
Proceedings of the First National Symposium/Workshop on
n'cal Contactor Technology, Vol. 2, pp. 1165,
I960.
J. C., et al., "Nitrification of Municipal
Wastewater Using Rotating Biological Contactors," Proceed-
ings of the first National Symposium/Workshop on Rotating
Biological Contactor Technology, Vol. 2, pp. 1193, February
2-4, 1980.
Zenz, D. R., et al., "Pilot Scale Studies On the Nitrification
of Primary and Secondary Effluents Using Rotating Biological
Discs at the Metropolitan Sanitary District Of Greater
Chicago," Proceedings of the First National Syposiurn/Workshop
on Rotating Biological Contactor Technology, Vol. 2 pp. 1221,
February 2-4, 1980.
Standard Method for the Examination of Water and Wastewater,
14 th Edition, APHA-AWWA-WPCF, 1975 _ .
Zenz, .D. R., et al., "Pilot'.Scale Studies On the Nitrification
of Primary and Secondary Effluents Using Rotating Biological
Discs at Metropolitan Sanitary District of Greater Chicago,"
Proceedings of the First National Symposium/Workshop on
Rotating Biological Contactor Technology, Volume II, pp. 1221,
Champion, Pennsylvania, 1980
O'Shaughnessy, J. C., et al., "Nitrification of Municipal
Wastewater Using Rotating Biological Contactors," Proceedings
of the First National Symposium/Workshop on Rotating Biological
Contactor Technology, Volume II, pp.1193, Champion, Pennsylvania,
1980. ,
Clark, J. W., et al., "Hater Supply and Pollution Control, pp,
743, Harper & Row Publishers, New York, 1977,
891
-------
PART VIII: INDUSTRIAL WASTEWATER TREATMENT
UPGRADING SLAUGHTERHOUSE EFFLUENT WITH ROTATING
BIOLOGICAL CONTACTORS
Torleiv Bilstad, Department of Environmental
Engineering, University of Rogaland,
N-4001 Stavanger, Norway
INTRODUCTION
The purpose of this work was to extend the potentially
attractive rotating biological contactor (RBC) process to
upgrade treated slaughterhouse wastewater effluent from a
biological tower. The research was primarily concerned with
demonstrating the RBC potential for removing biochemical
oxygen demand (BOD) from 1ow-temperature wastewater.
Attention was concentrated on the time to reach process
stability and treatment performance with low-temperature
wastewater.
The RBC-pilot plant, described more completely below,
receives a constant wastewater flow of 1.6 1/min. of settled
effluent from a tower trickling filter (bio-tower). The
organic loading to the pilot plant varies over time; from a
soluble BOD/of 50 mg/1 to as high as 700 mg/1. The higher
organic loading does not negatively affect RBC-treatment
performance. In fact, percent removal of BOD increases
with increasing organic loading.
Process stability measured as percent BOD removed in
10°C wastewater, was reached in approximately three weeks
after start-up. The bio-tower preceeding the RBC is
892
-------
considered a "roughing" process as it reduces the organic
loading on the RBC to a level that allows "optimal" removal.
This combination, bio-tower and RBC, therefore, enables the
treatment plant to meet new secondary treatment requirements
and other possibly more stringent requirements adoptable in
the future.
According to the literature, the number of treatment
flowsheets that can be derived by combining various biologi-
cal treatment processes is nearly endless. Combinations of
trick!ing-filter and activated-sludge processes are the more
common flowsheets. These processes have be used successfully
for a number of years for the treatment of all types of
wastewater, especially combined domestic and industrial
wastewater (1).
The process microbiology for combined biological treat-
ment processes is essentially the same as for the individual
processes. The biological activity in the bio-tower, the
"roughing" filter in this research, will be somewhat diffe-
rent from the RBC because of the higher shearing action
resulting from the hydraulic flowrates applied to the tower.
The bio-tower moreover acts to reduce the organic loading on
the RBC, making nitrification a possibility, expecially
during periods with higher wastewater temperatures, i.e.;
above 10 C.
The bio-tower effluent is piped to an adjoining fjord
with the outfall-pipe extending to a depth of 43 m. The
fjord is showing signs of deterioration with algae growth
and anoxic zones due to the total communal load of domestic
and industrial wastewater. It is therefore encouraging that
the RBC-pilot plant performs well as this process may
be the future mode of "polishing" the existing bio-tower
effluent.
INSTALLATION AND START-UP
Figure 1 shows the location of the plant. The present
effluent is piped through two 0.16 m diameter pipes, a
length of 240 m along the side-slope of the fjord, to a
depth of 43 m. A schematic of the treatment plant is
shown in Figure 2.
893
-------
HORGE.FJORW
SAHDKES
Figure 1. The fjord-system surrounding the cities of
Sandnes and Stavanger in the south-west of Norway.
The slaughterhouse is situated at AGRO, in the
lower portion of the figure.
894
-------
EQUILIZATION
— =>
WELL
WET
WELL
SCREENS
-
REMOVAL
440 m3 TOWER
TRICKLING FILTER
BASIN 650 m3
SEDIMENTATION
BASIN 85 m3
—
ROTATING BIOLOGICAL
CONTACTOR 47 m2
CHEMICAL SLUDGE
CONDITIONING
FILTER
PRESS
Figure 2. Slaughterhouse wastewater treatment flow-diagram
The slaughterhouse operates on one eight-hour work shift per
day, beginning at 7:00 a.m. Treated effluent from the
440 m3, 7.8 m high plastic packed bio-tower is recycled at a
rate of 50% during slaughterhouse production hours, and 100%
when there is no flow from the equilization basin. The
recycling of treated effluent is performed in order to
dilute the raw wastewater from the slaughterhouse before it
reaches the bio-mass in the tower, and to ascertain fluid
to the bio-mass at any one time, especially during off-
production hours at the slaughterhouse, to prevent dry-out
of the microorganisms.
The pilot plant module is a 0.92 m diameter four-stage
RBC unit mounted in a fiberglas tank; S5 Rotordisk, manu-
factured by CMS Equipment Limited, Mississauga, Ontario,
Canada. Each stage is 0.3 m long and the total contact
surface area for the entire unit is 47 m2 (500 ft2).
The four-stage rotorzone volume is 0.41 m3 (14.5 ft3) and
gives a theoretical detention time of 4.3 hours at a manu-
facturer recommended wastewater:flow of 2.3 m3/d, equivalent
to a hydraulic loading of 0.05 m3/m2-d.
Approximately 40% of the RBC surface area is submerged
in wastewater and the rotational speed is three revolutions
each minute giving a peripheral speed of 0.14 m/s. The
S5 Rotordisk is a package wastewater treatment plant and
895
-------
contains in addition to the four rotorzones, a 1.4 m3 primary
clarifyer and a 0.6 m3 final clarifyer. The whole unit is
enclosed. Figure 3 gives the plan view and a sectional side
view of the RBC-pilot plant.
After start-up, time was allowed for a biological slime
layer to develop on the rotating media. The average waste-
water temperature during the start-up period was approximate-
ly 10 C throughout the RBC. The hydraulic loading was con-
stant at 0.05 m3/m2-d, whereas the organic loading varied
substantially over time with an average total BOD7 of
645 mg/1, equivalent to an organic loading of 32 g total
BOD7/m2-d.
Table I and Figure 4 indicate that process stability
was reached in approximately three weeks. In fact total BOD
removal was 71% and total chemical oxygen demand removal (COD)
was 61% after 16 days. Obviously, this period was too short
for nitrification to develop and hence, removal of ammonia
nitrogen (NH3) was not evident. The 10 C wastewater tempera-
ture and high organic loading did not exactly provide optimal
conditions for the nitrifying organisms. The pH was approxi-
mately 7.5 and the alkalinity was 50 mg/1 as CaC03 or higher,
over the same period.
SAMPLING AND ANALYSIS
Several series of experiments have been conducted during
the weeks after start-up. The wastewater temperature through
the RBC has been relatively cold; approximately 10 C. Grab
samples of influent to the bio-tower, of influent to the RBC,
of wastewater from the four RBC-stages, and from the RBC-
effluent were collected at various times in order to deter-
mine changes in wastewater characteristics through the
treatment system. Grab samples were periodically coupled
with flow-proportioned composite samples collected either
manually or by an ISCO Model! 2100 automatic wastewater
sampler. Wastewater flow, temperature, pH and dissolved
oxygen (DO) were measured at the plant during sampling. The
choice of laboratory parameters varied and included alka-
linity, nitrogen, phosphorous, BOD', COD and total organic
carbon (TOC). Unless otherwise stated, analyses were con-
ducted on filtrate passing a 1 ym. poresize glasfibre filter.
The TOC analyses were conducted on a Beckman, Model!
915-B Total Organic Carbon Analyzer. Continuous BOD versus
896
-------
1.80 M
(N
(N
SNITT A-A
SLUDGE
STOR
SLUDGE STORAGE
I
Figure 3. Plan and sectional side view of the RBC,
S5 Rotordisk.
897
-------
PERCENT
REMOVAL+
40 -•
30- •
20 --
10 •-
0
BOD
COD
4-
3 6 9
DAYS AFTER START-UP
12
15
Figure 4. Total BOD7 and COD removal during start-up of
the RBC.
Table I
RBC Treatment Characterises with 10°C Wastewater.
Days
after
start-up
3
7
9
16
Inf.
295
525
780
630
TBOD7 (mg/1)
Eff. %R
263 11
385 27
382 , 51
,185 71
.Inf.
568
785
900
971
.TCOD.(mg/l)
Eff.
550
690
640
384
%R
3
12
29
61
898
-------
time curves were developed on an automatic instrument by use
of electrolytic respirotnetry; Voith Sapromat, Model C 12,
J.M. Voith GmbH; 7920 Heidenheim, West Germany. A schematic
diagram of a measuring unit is shown in Figure 5. Each
measuring unit comprises one reaction vessel, one oxygen
generator and one pressure indicator which are interconnected
by plastic hoses. The sealed measuring system is not affect-
ed by barometric air pressure fluctuations.
The required oxygen for the microorganisms is at any
time available in the electrolytic cell and is always
supplied in sufficient quantity to the sample of wastewater
to be analyzed. The BOD value which can be measured is
limited as the maximum oxygen demand of the sample may not
exceed 90 trig/1-h. If BOD5 values higher than 3000 mg'/l are
encountered, it is good practice ta dilute the sample.
Therefore, in contrast to the conventional BOD dilution
method, a genuine respiratory process takes place in the
Sapromat.
7B9 limn
O O O Q O
Figure 5. Schematic diagram of a Sapromat measuring unit
where A.= Reaction vessel, B = Oxygen generator,
C = Pressure indicator, T = Magnetic stirrer,
2 = Sample, 3 = C02 absorber, 4 = Pressure indica-
tor, 5 -= Electrolyte, 6 - Electrodes, 7 = Measur-
ing and control unit with digital printer
899
-------
RESULTS AND DISCUSSION
A typical treatment performance of the bio-tower and RBC
system, four and five weeks after start-up, is summarized in
Tables II and III respectively. The RBC pilot plant reaches
equilibrium with respect to organic removal approximately
three weeks after start-up. However, as would be expected for
the 10 C wastewater, nitrification was not apparent until the
fifth week of operation. Impending nitrification is illu-
strated by a small production of nitrites (N02) and nitrates
(N03) as in Figure 6. The RBC pilot plant has not experienced
problems with low concentrations of dissolved oxygen (DO).
Figure 7 is typical in this respect, indicating adequate
oxygen mass transfer in all four RBC-stages.
Four Ueeks After Start-up
The soluble (filtered) long-term BOD concentrations ob-
tained from the bio-tower and RBC pilot plant during a typical
slaughterhouse production day are illustrated in Figure 8.
The analyses were performed on composite afternoon samples,
each comprised of five half-hourly grab samples. These long-
term BOD curves are especially illustrative in the case of
the RBC-effluent where the effect of nitrification showed up
very clearly for the non-inhibited nitrogenous BOD. The
Sapromat oxygen uptake reaction was performed on non-diluted
wastewater which, no doubt, was of great importance for the
slow-growing nitrifiers. The standard dilution BOD-method
often retards nitrification because of a population decrease
of the nitrifying bacteria (2).
The rate of nitrification, and therefore the rate of
growth and length of generation times of the nitrifying
organisms, is affected by several environmental factors
including temperature, pH and dissolved oxygen. Borchardt
(3) estimated the rate of nitrification at 9 C to be about
50% of the rate at 20 C. The RBC did not provide a long
enough sludge-age after four weeks of operation to provide
nitrification at 9 C. The 20 C water-bath in the Sapromat
coupled with six days of incubation provided such a sludge-
age for the RBC-effluent sample (Figure 8). The pH of the
wastewater was normally above 7.0.
Successful removal of carbonaceous material expressed as
BOD, COD or TOC is apparent from Figure 9 and Table II. The
hydraulic loading was constant at 0.05 m3 /m2-d which yielded
900
-------
1— i
1— 1
cu
1 '
_Q
1—
S-
•p
rd
CU
-P
_cp
.
Q.
S-
03
OO
^
-P r—
CI co •
•r- 0
i — O
fO CO
r~~ CW
=a: E
. c
cu o
£="£
tO -P
OO 00
cu
•1™ _^
t— 0
o
CD i —
i — O
0.-
E^
OO
*— CD CM CM
CD t — CD CD
CO CM •* t—
CTl Cn CD t—
CM OO i£)
=a- CM CD CM
un oo vo r~^
CM to LO co
•=— CO CM CM
'*-' "~ *~
t IT) CM CD
<^J* •<: — *c—
Ln un CD i-n
CM oo en vo
^— CM CM
i— i Ll_
i 2: u-
S_ i— i V— UJ
CU 1 1 1
3 CJ) C_3 C_5
O CO CQ CQ
H~* £t: c^ o:
CD
OO
» « — ~ ~
p^
CD
i — CM CM "=i- .
CD CD CD CD
r~ co oo <=d-
•* CD OO !*>•
. CO r^ ^t- «=l-
CD un r~~ T—
CTi *r— <^- <^t~
CM •rr-
CM CD Cn -^1-
oo 10 CD r-^
CD LO T-
CD CD CM •=}•
CD CT> UD OO
r- oo
ID CD LO 1-n
CO CM CM CD
CM CM CM
l—i li
i z: u_
S_ 1— 1 -r- LU
CD 1 1 1
3 o c_j o
O CQ CQ CQ
H~ 01 o: o:
CD
oo
• • ~ — ™
cj-
« —
CD i — i — CM
CD CD CD CD
CD OO l~>. UD
OO CD •«— tO
"^t* r*^ ^i" LO
«^- ' r-» r-. i^
to" CD co to
^ — V*"
to CD CM r~.
tO VD LT) <—
UO OO •<— r—
1 1^0
tn «d-
CD CD un ID
CD
-------
I—
01
to
0)
cu
o>
O)
u_
0)
o
c
(t)
q_
ai
a.
CU
O
o en
CJ E
I >> m
CO -»-» O
-i<£ -r- O
r— C (T3
«=C T- CJ
O.
o
ID O
E to
to -(->
OO 00
O)
CU O
oo
CD co cn oo
CM
oo oo co to r-.
CD CD oo CM
co c\j
LO
01 CD LO oo uo oo co cr> CM CM oo r~~ oo
CM ID CM LO UO CO CO •*-"
10
co oo
co co cri co
oo to
CD co
oo oo
oo oo •=— •=—
I oo 10
tn
cri en co vo 10
un LO LO un tn
CD
C7>
re
X
CU
cu
o
O)
CD co CD oo
«^- CM CD CD
CTl CO
un 01 CD cr> LO CM oo
OO CD CO I"-- CO CTl
un CD CD in CD un
co 10 uo m «d-
en CD
UO CM CD CD UO
• • • • •
CD ID I— |-~- l~-
LO
CD CD LO LO LO f>. CD
^J- I— VD l~~ LT) •* UO
I I
I I
oo •
CO LO
• •
• CD CD
S- I— 1
I—
oo
m
Lu.
Lu
UJ
o <_> o c_> o
ro
CU I I I
•
CM OO •
S_ l-H *— CM OO •* LLJ
cu i i i i i i
3 O C_3 O C_5 CJ O
O CO CO CO CO CO CO
CTl
90'2
-------
T3
OJ
13
O
O
O)
r—
_d
oo
O
CM
O
J-
o
O-
CD
Q.
cn
IN
o o
O cn
O E
O
: co
cn
CD --^
O Cn
O) O
(0 -(->
(>1 CO
CD O
r—
CD CM OO OO LO LO LO
OO •*- UO «— OO CM •=—
CD •<— OO «* UO VO UO
•=— co uo •«— cn to
UO CM tO OO OO CD
r> oo CD r~» CM co cn
cn uo oo LO LO -^- CM
CM cn LO <3- -=± •* "3-
cn
co
Q.
X
CD
c
un CM CM cvj «—
i LO un un Ln LO
CD OO CM -— up un un «d~ <^f
cn •«— co i-» to to
r~^ to un un un un
un —
un un CD CD un uo un
r^ un cn co to un un
*— CM
i il
i i i i i i i
i ii i i
-t->
E
O
P
o>
r—
CO
i. l—l •«— CM OO •=*• UJ
Q) I I I I I I
3OOOOOO
O CO CO CQ CO CO CQ
i— c£. ex. c£. o: o; c£
CD
OO
oo
903
-------
CM
O E
(O
-I- C_>
E rtS
-r- O
I— CD
CD
CU
O
cu
CU O
(13 4->
00 CO
cu
o
t— CO O
t£o
03 I
oo o
Ln CO OO OO CM OO CO OO OO CM O1
CD CD
r — ID
CO
IO
CD CD CD CM -
Ln
to co •
co i
•=!• oo
• cr>
oo co CM to in 10 «3-
IQ CD to LO in -=d- •=*•
CD CD CTl OO CM CM
cr> in
CD ou
- oo
10 to to
CM en
OO
CM |-~- OO t— CD
cu
en LO •
CM cr> co
•«— oo r>-
OO CM •«— •
, -T— ooooLnoomooco
OO CO CM CT> Ln CD "T- CO
CD LO CD LO ID un CD
Ln 01 Ln CD co r-« UD
Ln CD CD CD un CD
«^ oo r^ •* ^— CD
«— co •«—•«—•«—'—
oo oo
oo
CM oo -^r Ln T—
• • • * •
CM oo Ln to r^
in CD Ln Ln
• • • •
CD CD CT> C7l
U_ 1—1
S- 1-1
0)
CM OO
I
I I
OOOOOO
O CQ CQ CD CQ CQ CO
H- o; o; o: o; o: o;
S- H-l
CU
I
CM OO -
I
SOOOCJOCJ)
O CO CO CO CO CO O3
ce:
a: o;
o;
904
-------
oo
o
•
CM
O
J-
o
O-
CT>
Cu
CT>
QO
O CD
CJ> E
4-> O
-i- o
c: to
-t- c_>
i— 01
O
CO
-o
O)
c
o
o
CJ>
O
c
o> o
"a.4J
E eo
to 4->
co co
O) O
i— ID <=}- CM
tn ^— r-^ liD to i_n in
IO
«* LO
uo un
CM •«— UD CM cr> en un
CD o> r-^. r~« to to i£>
CO CD LO CO CO CD
p^ CD CM CD r~. tn
CM IO CM CM «— •<—
CD CD in CD CD CD CD
CT> <£> cn cri r-^ un •<—
CO ^~" ^~ ^~ ^^ ^~
to 10 r~. i— i
LO r>- CD CD to i
• * • • •
CD CO Ln to U3
CD LO CD CD CO I
• • • » •
*a- ^- T— .*- CD
S- l—• *— CM CO <* LU
O) I I I I I I
S o c_> o o c_) c_)
O CQ CQ CQ D3 CQ CQ
CD
CO
O-i
905
-------
NO 2
NO 3
mg
I
3 --
2 -•
TOWER-IN RBC-IN
—I 1 1 -t-
STAGE-1 STAGE-2 STAGE-3 STAGE-4
SAMPLING STATIONS
Figure 6. Impending nitrification depicted by a small
production of nitrite and nitrate five weeks
after start-up.
RBCHN STAGE-1 STAGE-2 STAGE-3 STAGE-4
Figure 7. Dissolved oxygen through the RBC system.
906
-------
200 ••
100 ••
2 U 68
INCUBATION TIME (DAYS)
Figure 8. Long-term soluble BOD analyzed on the Sapromat.
Nitrification is evident in the one RBC-effluent
sample where nitrification is not inhibited.
907
-------
an organic loading of 17.8 g BOD5/m2-d and 19.5 g BOD7/m2-d.
Maximum recommended loading values in Norway are 15 g
BOD7/m2-d at 10 C for 85% BOD7-removal, and a maximum of 7.5 g
BOC7/m2-d if nitrification is intended (4).
The removal of soluble BOD was more than 90% in the RBC
pilot plant. The effluent BOD5 and BOD7 were approximately
25 mg/1 and 35 mg/1 respectively, four weeks after start-up.
To give an indication of the solid separation efficiency in
the RBC-final clarifier, the total (unfiltered) BODg and
BOD7 were 38 mg/1 and 46 mg/1 respectively for the situation
mg
I
800 -
600
400 - -
200 •-
COD
BO D
TOC
TOWER-IN RBC-IN STAGE-1 RBC-EFF
SAMPLING STATIONS
Figure 9. Removal of substrate in the tower trickling
filter and rotating biological contactor four
weeks after start-up.
908
-------
depicted in Figures 8 and 9. The evidence furthermore indi-
cated that the soluble BOD removal percentage in the RBC pilot
plant increased with increasing BOD concentration in the
influent.
Five Weeks After Start-up
The data presented in Table III and Figure 6 suggest
impending nitrification. The increasing concentration of N02+
N03 through the RBC pilot plant was consistent during the
24-hour period making up the data. This was not unwarranted
as Figure 8 revealed the presence of nitrifiers in the RBC-
effluent four weeks after start-up. The BOD-test was per-
formed on fiberglass-filtered, undiluted and nonseeded waste-
water and, as mentioned earlier, the impending nitrification
would probably not have been detected by a standard BOD-
dilution test. In addition to dilution of the sample, the
standard BOD test also incorporates seeding with raw waste-
water containing heterotrophic bacteria thereby possibly nega-
ting nitrification.
The principle of biologically induced nitrogen removal
in wastewater treatment ficilities is based on the activity of
populations of autotrophic nitrifying bacteria and their
capability to oxidize ammonia (NH3-N) to N02 and N03. In
addition to nitrification, microorganisms other than the
nitrifiers require nitrogen for growth. The amount of nitro-
gen assimilated during oxidation of carbonaceous material has
been placed at 5% of the oxygen demand (C : N : P = TOO : 5 :
1). That means removal of NH3-N during biological treatment
of wastewater may be because of assimilation, not necessarily
nitrification. A production of N02 + N03 as in Figure 6,
however, indicates nitrification.
As mentioned earlier, nitrification is affected primari-
ly by pH, DO and temperature. Also, at neutral pH levels
there is usually insignificant nitrification until soluble
BOD has been oxidized (5). Hence, to evaluate the RBC per-
formance for NH3 removal, progression of treatment within the
RBC stages must be assesed. Oxidation of carbonaceous
substrate, expressed as soluble COD is presented in Figure 10
and Table III. The observed COD decrease coupled with the
N02 + N03 increase shown in Figure 6, clearly indicate that
the conditions were amenable for nitrification. The impend-
ing nitrification was also suggested by the decrease in alka-
linity and soluble Total Kjeldahl Nitrogen (TKN) in Table III.
909
-------
COD
200-
100
TIME OF DAY (O'CLOCK)
Figure 10. Reduction of COD through the RBC for a typical
slaughterhouse production day.
910
-------
There is no doubt that the BOD removed by the bio-tower
was the more readily biodegradable fraction of the wastewater.
The soluble COD in the RBC-effluent was approximately 100 mg/1
and probably would not decrease much more if further biologi-
cal treatment was provided. This study supports the evidence
that when wastewater is sent through a series of trickling
filters, or recycled several times through the same filter,
the removability of wastewater organics decreases as the
number of passe£ increases (6).
CONCLUSIONS
The application and start-up performance of a RBC pilot
plant unit for upgrading clarified trickling filter effluent
has been described. Although the operational experience has
been very short, the following remarks can be made from
treatment of a 10 C slaughterhouse wastewater:
1. RBC process stability with respect to BOD removal was
reached in approximately three weeks after start-up.
The RBC-biofilm became mature relatively fast due to
seeding of microorganisms from the bio-tower.
2. The hydraulic loading to the RBC was constant at
0.05 mVm2- d whereas the organic loading varied from
approximately 2 g soluble BOD7/m2-d to 35 g soluble
BOD7/m2-d. A typical slaughterhouse production day
organic loading is approximately 20 g soluble
BOD7/m2-d, resulting in a soluble carbonaceous BOD7
effluent concentration of approximately 35 mg/1.
3. The Sapromat analysis for soluble BOD did not require
seeding or dilution of the wastewater to be tested.
The nitrifiers present will therefore perform immediate-
ly if other environmental conditions are satisfactory.
4. The effect of low temperature wastewater on NH3-removal
could not be verified in the short time after start-up.
This will be a question to answer after prolonged RBC
operation.
ACKNOWLEDGEMENTS
This study was supported by funds provided by the
Norwegian Environmental Protection Agency (SFT), The author
wish to thank Damann Anderson, Pa-tti Hantz, Arild Lohne,
01av Nordgulen and Sissel R0ine for assistance.
911
-------
REFERENCES
1. Metcalf & Eddy, Inc., "Wastewater Engineering: Treatment,
Disposal, Reuse", McGraw-Hill, Chapter 9, 1979.
2. Dague, R.E., "Inhibition of Nitrogenous BOD and Treat-
ment Plant Performance Evaluation", Journal of the Water
Pollution Control Federation, Vol. 53, No. 12, December
1981, pp 1738-1741.
3. Borchardt, J.A., "Nitrification in the Activated Sludge
Process", Division of Sanitary and Water Resources
Engineering, University of Michigan, Ann Arbor, 1966.
4. Statens Forurensingstilsyn, "Retningslinjer for Dimen-
sjonering av Avlapsrenseanlegg", p 42, Oslo, 1978.
5. Miller, R.D., et al., "Rotating Biological Contactor
Process for Secondary Treatment and Nitrification
Following a Trickling Filter",, Techn. Report 7905,
U.S. Army Med. Bioeng. R & D Lab., June 1979.
6. Poon, P.C., et al., "Upgrading with Rotating Biological
Contactors for BOD Removal", Journal of the Water
Pollution Control Federation, Vol. 53, No. 4, April 1981,
pp 474-481.
912
-------
EVALUATION OF AN ANAEROBIC ROTATING BIOLOGICAL
CONTACTOR SYSTEM FOR TREATMENT OF A MUNITION
WASTEWATER CONTAINING ORGANIC AND
INORGANIC NITRATES
Leonard L. Smith, Hercules Aerospace Division,
Hercules Incorporated, Radford Army Ammunition
Plant, Radford, Virginia.
INTRODUCTION
Radford Army Ammunition Plant (RAAP), like most of the
Army propellant and explosive manufacturing plants, was built
in the early 1940s to supply munitions for World War II. In
1970, the Army initiated modernization programs at its ammu-
nition plants to replace obsolete facilities and improve the
safety of operations. As part of this modernization program
a continuous automated multi-base line (CAMBL) manufacturing
facility was planned for construction at RAAP to augment the
present labor intensive batch-process. This paper describes
the studies that were conducted to develop design criteria
for a facility to treat the wastewaters that will be gener-
ated in the CAMBL.
WASTEWATER CHARACTERIZATION
A wastewater characterization study was conducted for
the CAMBL manufacturing facilities. Samples of the waste-
waters were collected and analyzed during the evaluation of a
prototype CAMBL manufacturing line. These data were compiled,
and the expected characterization of the full-scale facili-
ties were determined. The quantity of water requiring
913
-------
treatment from the CAMBL- facility" was determined to be •approx-
imately 50,000 gallons per day. This wastewater"Will contain
acetone, ethanol, nitroglycerin (NG), nitroguanidine (NGu),
other propellant ingredients, and inorganic nitrates.
LABORATORY TREATMENT STUDIES.
Laboratory-scale treatment studies were conducted to
determine the feasibility of selected treatment methods and to
define the design parameters for pilot plant studies.
Laboratory studies were conducted to determine the bio-
degradability of NGu using a biochemical oxygen demand (BOD)
test kit. The tests were set up to determine the oxygen
uptake rate of the readily biodegradable organic solvents
alone, and then with various quantities of NGu added. This
testing showed that NGu is not biodegradable by itself, but
when combined with a readily biodegradable carbon source NGu
is biodegradable.
Studies conducted by Wendt (1) showed that NG is biode-
gradable, but it does exert .a.'toxic effect on the biological
metabolism.
The biodegradability of NG and NGu was. further studied
using a laboratory-scale rotating biological contactor (bio
disc) unit. The wastewater utilized for this study was a
mixture of wastewater from the manufacture of other propel-
lants, waste process water from the manufacture of NG, and
the CAMBL pilot line effluent. During the study, the bio disc
influent contained a chemical oxygen demand (COD) concentra-
tion ranging from 500 to 1000 mg/1, a NGu concentration
varying from 30 to 70 mg/1, and a NG concentration of approxi-
mately 5 mg/1. During this period, the COD removal was
approximately 90 percent; the NGu removal ranged between 50
and 90 percent while achieving 100 percent NG removal.
Based upon the wastewater characterization and laboratory
studies, two design concepts were considered for this proposed
xjastewater treatment facility: (a) design a ^completely new
chemical-physical treatment facility for the treatment of this
wastewater alone, or (b) expand on the aerobic rotating bio-
logical contactor (RBC) treatment plant under construction at
RAAP for the treatment of the wastewater from the existing
manufacturing facilities.
Alternative (b) was selected for the pilot plant
evaluation, based on the estimated savings of over $800,000
in capital costs and an annual savings of about $160,000 in
operating costs. The characterization of the wastewaters
from the existing manufacturing facilities, the proposed
914
-------
CAMBL facility, and the combined facilities are shown in
Table I.
Table I. Characterization of Waste Waters of
Existing and Proposed Facilities
Parameter
Increase Due
Existing Continuous Combined to Continuous
Facility Facility Facilities Facility
Flow (mgd) 1.245 0.058 1.303
COD (Ib/day) 7818 1886 6604
(mg/1) . 607
BOD (Ib/day) 1887 754 2641
.(mg/1) 243
N03 (Ib/day) 3024 144 3168
(mg/1) 304
NG (Ib/day) — 14.3 14,3
(mg/1) 1.50
NGu (Ib/day) — 57.4 57.4
(mg/1) 5.28
4.6%
40%
40%
4.7%
The permit issued by the EPA and Commonwealth of Virginia
for the wastewater discharge from the aerobic KBC treatment
plant was based upon this facility treating the wastewater
from the present manufacturing facilities only. This requirec
that any new manufacturing facility to be constructed at RAAP
must also provide facilities for treatment of the wastewater
generated by that facility to ensure the effluent quality is
not degraded.
The aerobic RBC plant hydraulic capacity, as designed,
will be adequate for the additional wastewater flow, but
additional facilities will be required for the removal of the
additional organics, NG, NGu, and inorganic nitrates. The
laboratory studies showed that an aerobic biological treat-
ment system appeared to be a suitable method for the removal
of the organics, NG and NGu. However, an alternate treatment
915
-------
method will be required for' the. "removal of the inorganic'
nitrates. This can best be accomplished by a biolo'gical de-
nitrification system. Since additional RBC units will be
required for the organic removal, the decision was made to
evaluate on a pilot plant scale the use of submerged RBC units
for the biodenitrification process.
PILOT PLANT EVALUATION
For the evaluation studies to develop the design criteria
for the treatment of the combined RAAP wastewaters, a one-half
meter bio surf pilot plant, capable of independent operation
of each stage as aerobic or anaerobic, was purchased from
Autotrol Inc. This bio surf pilot plant consisted of a series
of 36 corrugated polyethylene discs containing a total of 250
ft2 of surface area. The discs and tank were divided into
four stages, separated by removable bulkheads. Each bulkhead
consisted of a top and bottom section, whereby, each stage
could be operated either completely or 40 percent submerged.
The first phase of the pilot plant evaluation was con-
ducted with all four stages completely submerged to determine
the feasibility of the decomposition of. the organic solvents,
NG and NGu, under anaerobic conditions and to determine the
rate of nitrate reduction in a biological denitrification
system. An airtight cover was installed on the pilot plant to
prevent the diffusion of oxygen into the wastewater from the
atmosphere. The system was operated during this period at a
hydraulic loading of 1.6 gpd/ft2 of surface area', and average
organic loadings of 6.0 pounds COD and 2.4 pounds BOD per day
per 1000 ft2 of surface area. The NG and NGu concentrations
were both maintained between 1 and 5 mg/1. During this phase
of the evaluation, the unit averaged 84 percent COD, 90 per-
cent BOD, 94 percent NGu, and 100 percent NG removal. The
nitrate removal rate was calculated as a ratio of the-BOD
removal. The BOD/NOs removal ratio during this phase of the
evaluation was 0.39. This evaluation demonstrated the feasi-
aility of treating this wastewater by a biological denitrifi-
cation process to achieve the proposed discharge limitations.
During the second phase of the evaluation, the third and
fourth stages of the pilot plant were converted to aerobic
stages. This study was conducted to determine the effects of
anaerobic RBC units operating in series with the aerobic RBC
inits. This change had little or no effect on the organics,
and NGu, removal rates of the pilot plant. Figure 1 shows
the COD influent and effluent concentration for this phase of
:he study.
916
-------
OQ
-------
The"laboratory and preliminary pilot plant data indicated
that a biological denitrification KBC system followed by
aerobic KBC units is a feasible treatment method for the re-
moval of the organics, NG, NGu, and inorganic nitrates from
the CAMBL manufacturing facility. A preliminary design of a
system for the treatment of the combined wastewaters consisted
of four additional completely submerged EEC shafts preceeding
the eight aerobic RBC shafts under construction. To evaluate
the efficiency of this proposed system, the pilot plant bio
surf unit was converted to a four-stage system, the first stage
anerobic followed by three aerobic stages. The sample collec-
tion points for the evaluation were selected at the first stage
influent and effluent, and the third stage effluent; therefore,
simulating the results from the proposed full-scale facility.
During the first two weeks of this evaluation the COD
influent concentration was maintained between 400 and 600 mg/1,
the NG concentration approximately 1 to 5 mg/1, and the NGu
concentration between 10 and 20 mg/1. Figure 2 shows the BOD
and COD influent and third stage effluent concentration during
this period. The desired influent COD during this period was
450 mg/1. However, due to the constant mixing of wastewaters
and the volatility of the organic solvents, wide day-to-day
fluctuations occurred. It can be seen from Figure 2 that even
ith these influent fluctuations, the effluent remained quite
.onstant.
Studies were conducted during the last month of the
avaluation to determine if this RBC system could^, operate
effectively under the worst conditions expected in a full-
scale facility and still produce arr-^ffluent meeting the
required discharge standards. The system was operated at an
average organic loading of 1.3 times the design loading, NG
Loading of 3.5 times the design loading, and an NGu loading of
twice the design loading. The system was operated at a low
temperature of from 6 to 12°C during this period. See Figure
3 for the results of this evaluation. Figure 4 shows the
average BOD and COD remaining and the cumulative BOD and COD
removal efficiency after each stage of treatment for this
tudy. During this phase, the allowable daily average COD of
L90 mg/1 was exceeded on only two days; however, the maximum
laily COD effluent concentration of 290 mg/1 was never
sxceeded. These adverse operating conditions reduced the
iverage COD removal from 85 percent to 74 percent during this
>eriod. The NG and NGu removals during this period were near
LOO percent most of the time.
Thje data from the pilot plant bio surf evaluation were
malyzed to determine the ratio of organic removal rates to
918
-------
o
o
3*3
^: so
co
CJ1
00
m o
* g
CO
t
I
1
f
t
col
-\
919
-------
IS
— 05
_» IO
4k CJ1 O5
«S'
i
CO
I
o o
: f i
: &..
00
CO
o
1 1 1
i
i
--•1
- s,
e»
v>
^
I
I S
';
f S
920
-------
S3 "•
O oo
- «
0
00
s
b,
s
00
O
00
o
-------
bhe" inorganic nitrates removal rates under various operating
conditions. As stated above, during the first pha~s~e of the
evaluation the pilot plant cover was installed to provide a
completely anaerobic system. This evaluation showed the BOD/
N03 removal ratio to be 0.39, which would calculate to provide
an average nitrate removal rate of 13.6 lb/day/1000 ft2 for
the full-scale system. During the later evaluations, the bio
surf cover was removed and the wastewater in the anaerobic
stage was exposed to the atmosphere, allowing oxygen to dif-
fuse into the wastewater, greatly reducing the nitrate removal
rate (see Figure 5). Based on the results of this phase, the
nitrate removal rate was calculated to be 3.4 Ib N03/day/1000
ft2 for the full-scale system. The great differences in the
nitrate removal rate between a covered and uncovered system
can provide a method to control the nitrate utilization of the
submerged RBC stage in the full-scale facility. The system
can be designed with removable cover section to provide a
flexibility to compensate for low or high nitrate concentra-
tions in the facilities influent.
The results from these evaluations were analyzed to pro-
vide the data necessary for the preparation of the design
criteria for the proposed facility. The organic load applied
was plotted versus the organic load removed for the pilot
plant anaerobic system (see Figure 6). A similar graph was
also prepared for the first two aerobic stages (Figure 7).
These graphs can be used to predict the efficiencies of the
full-scale facility at various organic and hydraulic loading.
The design of the expanded system at RAAP was analyzed, using
these graphs.
Based upon the data from this evaluation, the design
criteria for an addition to the RAAP RBC treatment facility
were prepared. Figure 8 shows the flow diagram of the pro-
posed RAAP facility for the treatment of combined wastewater.
These design criteria were submitted to the Corps of Engineers
for the design of the addition to the facility. The final
design was completed and construction of the facility at RAAP
nas been initiated.
REFERENCES
1. Wendt, T. M., Cornell, J. H., and Kaplan, A. M.,
"Microbial Degradation of Glycerol Nitrates," Applied and
Environmental Microbiology, Nov. 1978, pp 693-699.
922
-------
— * ro
un o>
00 CO O
o
90
923
-------
as e»
IS3 <
•*•«» m
o a
924
-------
925
-------
926
-------
APPLICATION OF ROTATING BIOLOGICAL CONTACTOR (RBC)
PROCESS FOR TREATMENT OF WASTEWATER CONTAINING
A FIREFIGHTING AGENT (AFFF)
Susan Landon-Arnold, MS, R.M., Visiting Scientist,
University of Oklahoma, Norman, Oklahoma
Deh Bin Chan, PhD., P.E., Research Civil Engineer,
Naval Civil Engineering Laboratory, Port Hueneme, CA
INTRODUCTION
A firefighting agent, Aqueous Film Forming Foam (AFFF),
has been used for fuel/oil fire extinguishment at airports
and on shipboard since 1970's. AFFF has been found to be the
most effective fuel/oil firefighting agent ever to be
formulated. In accordance with firefighting performance
specifications used by the manufacturers, a 28-square foot
fuel fire can be extinguished within 45 seconds with a 6%
AFFF solution (by volume).
AFFF consists of fluorochemical surfactants, hydro-
carbon surfactants, ethylene glycol and its derivatives, and
about 70% water (by weight). In a firefighting operation,
the AFFF concentrate is diluted to a 3-6% solution (by
volume), and sprayed under pressure onto the fire. The foam
created during the spray covers and extinguishes the fire.
AFFF concentrate (FC-780) contains an organic load of
approximately 380,000 mg/L COD, or 110,000 mg/L TOG or
325,000 mg/L BOD (Ref 1). A toxicity test with fathead
minnows indicated that the 48-hour TLm (LCsn) concentration
was about 1800 ppm (FC-206, by volume) (Ref 2). A maximum
927
-------
loading rate to an activated sludge treatment process (with
acclimation) was found to be 250 ppm (FC-206 by volume)
without the addition of an antifoam agent (Ref 2).
The Rotating Biological Contactor (RBC) is considered
to be a most cost-effective wastewater treatment process due
to it being simple in operation, low in capital investment
and low in energy requirements. The RBC will fit well with
trickling filter systems, which constitute approximately 95%
of the sewage treatment systems on military bases, and will
upgrade the effluent water quality to meet the National
Pollutant Discharge Elimination System (NPDES) permit
standards.
Increased popularity in the use of the RBC on military
bases is anticipated. This, in part, is due to the successes
experienced by other researchers for using the RBC to treat
various organic compounds, such as formaldehyde and formic
acid as well as the explosives RDX, HMX and TNT (Ref 3).
Such research prompted the following experimentation for
determining the RBC's feasibility (technically and economi-
cally) for treating AFFF containing wastewater. The re-
search effort initially began with a chemostat study of the
parameters and microorganisms that were amenable to AFFF
bioconversion. This was followed by experimentation with a
four-stage bench top RBC system. The percent (%) conversion
in COD, BOD and TOG was monitored as a means of determining
removal of AFFF.
MATERIALS AND METHODS
Two types of experimental systems were used in this
study. One was an aerobic chemostat used as an approach to
determine initial feasibility, the other was a bench top
model of an RBC.
Chemostat
Physical Set-Up; A diagram of the chemostat physical
set-up is given in Figure 1. Influent was sterilized in a
two-liter reservoir and put on line aseptically. The flow
rate of sterile medium into the reaction vessel was 1
ml/min. The reaction vessel was a 4-liter aspirator bottle
(Kimax), which was continuously agitated via a stir bar/stir
motor arrangement (Corning Hot Plate Stirrer, PC-351).
Aeration was accomplished via filtered air (Acropore 0.45-ym
filter) bubbled into the bottom of the reaction vessel. A
928
-------
£
o
z
o
o
00
•H
O
DC
UJ
W
in
CC 0.
=: Q.
Q O
UJ W
2
HI
- • • — m
co
H UJ UJ <
00
<
I
Q. DC
CO O
CC
o _
CL Z
uj O
J H
CO <
I I
UJ U.
929
-------
constant volume of 2 liters was maintained within the reac-
tion vessel by a siphon overflow tube into a waste recept-
acle which was replaced and autoclaved when full. Samples
were taken by suction-draw from the sampling port.
Inoculum; The inoculum or seed for the start up of the
chemostat was 0.05 gm each of the following: dried bacteria
culture (Horizon Ecology Company) for degrading fats, oils
and greases (#245-40), for hydrocarbon degradation in fresh
water (#245-60), and 5 ml of activated sludge from the Buena
Ventura County Water Treatment Plant.
Media; Bushnell Haas Broth (Difco) was used as a
minimal salts medium to which specific amounts of known
carbon could be added. The carbon used in this experiment
were D-Glucose (Difco) and/or the aqueous film forming foam
designated FC-780 (3M) . This then comprised the sterile
influent.
Growth conditions: The chemostat experiment was
conducted at ambient temperature, under mild aeration and
agitation. pH was monitored but no attempt at adjustment
was made.
Procedure; The system start-up was as follows. The
reaction vessel containing two liters of sterile Bushnell
Haas Broth (BHB), 0.05% Glucose and 0.5% FC-780 was seeded
with the inoculum and allowed to grow as a batch system.
After 48 hours and an increase to 0.60 optical density,
sterile influent containing 0.05% Glucose and 0.5% FC-780
was put on line. On day 8, the influent was changed to
contain 0.5% FC-780 (approximately 2000 ppm COD) as the only
carbon source. Samples from influent & effluent concur-
rently, were taken three (3) times per week and analyzed as
follows:
A. Turbidity. Utilizing the sterile influent as a
standard or blank, turbidity of the effluent was
determined at 460 nm, utilizing a Beckman Spec-
tronic 88.
B. pH. The pH of the influent and the effluent were
determined immediately after sample withdrawal,
using an Orion Research Model 701A/Digital
lonalyzer.
930
-------
C. Biochemical Oxygen Demand (BOD). The 5 day BOD
determination was used as outlined in Section 507
in Standard Methods (Ref 4), utilizing an Orion
Research Model 701A/Digital ionalyzer and Mode
97-08-00 02 electrode.
D. Chemical Oxygen Demand (COD). COD was performed
according to the method outlined in Section 508 of
Standard Methods (Ref 4) and modified by Technicon
(Ref 5).
E. Total Organic Carbon (TOG). A variation on the
procedure given in Section 505 of Standard Methods
(Ref 4) was used. The variation, the acid sparge
technique, was performed with the Beckman 915B TOC
analyzer, and is outlined in the operation manual
(Ref 6). All samples used for TOG, COD and/or BOD
determination were filtered prior to analysis
through a series of graded membrane filters, i.e. 5
ym, 1.2 ym, 0.8 ym and 0.45 pm (Gelman). Each
filter was washed prior to use with 30 ml of double
deionized water to remove any organic wetting agent
on the filter.
F. Microorganism Identification. Bacterial and fungal
populations were identified and enumerated utiliz-
ing Nalgene Nutrient Pad Kits, of the following
media: Standard TTC - for total counts, Azide -
for enterococci and fecal streptococcus, Wort - for
fungi, filamentous and non-filamentous, Weman - for
slime forming mesophilic bacteria (e.g. Leuconostoc
mesenteroides).
Rotating Biological Contactor (RBC)
Physical Set-Up; A diagram of. the RBC physical set-up
is given in Figure 2. Influent was sterilized and asepti-
cally added to a 20-liter reservoir (5 gal. bottle, Kimax).
The reservoir was then put on line aseptically. The flow
rate of sterile influent into the aerobic RBC was 3.5
ml/min, and was controlled by a peristaltic pump (Cole
Farmer). The RBC was on loan from the U.S. Army Mobility
Equipment Research and Development Command (USAMERDC), Ft.
Belvoir, VA., and has been described in detail by them (Ref
3). Basically, it was a five chambered unit constructed of
Plexiglass. Each of the first four chambers contained six
931
-------
O
O
O
5
CD
CD
CO
UJ
U. UJ
(D
O CC
DC O
H- 0-
CO
I I
O Q UJ
932
-------
1/4-inch (0..6 cm) thick plexiglass discs, 9-1/2 inches (24
cm) in diameter mounted on a shaft 1/2 inch (1.27 cm) in
diameter. One hundred and twelve (112) holes, 1/4 inch (0.6
cm) in diameter, were bored into each disc to aid in microb-
ial attachment. The total disc area was 23.55 ft2 (2.188
m2). The last chamber was void of discs, acting as a 1-
liter capacity reservoir-clarifier. The total liquid
capacity of the unit was 14.5 liters. The discs were
rotated at 17.5 rpm, thus being equivalent to an edge
velocity of 0.73 ft/s (0.22 m/s). An additional clarifier
was added in the form of a modified Imhoff cone, which was
used to visually measure the amount of sedimentation pro-
duced in a 24 hour period. Samples were taken from within
all four stages and from the influent.
Inoculum; The seed for the start-up of the RBC was one
liter of activated sludge obtained from the Buena Ventura
County Water Treatment Plant, and was inoculated within one
hour of acquisition.
Media; BHB was used. Varying concentrations of
FC-780, D-glucose, and Nutrient Broth (Difco) were added as
outlined in the procedure.
Growth Conditions; The RBC experiment was conducted at
ambient temperature. Aeration was accomplished by the
revolution of the discs through the wastewater. pH was
monitored and adjustments were made, using IN NaOH or IN
HCL, when necessary.
Procedure: The system start-up was as follows. The
RBC was filled with 14.5 liters of BHB plus 0.1% glucose,
inoculated with activated sludge and allowed to run as a
static system for 24 hours. Sterile influent containing
0.1% glucose was fed into the unit at a rate of 3.5 ml/min.
After 2 days it was determined that this mode of addition of
the carbon source was inadequate to maximize colonization of
the discs and so glucose and/or nutrient broth was added to
each stage once daily to a total concentration of 0.1%
carbon. On day 29, FC-780 was added to the influent at a
concentration of 100 ppm in terms of COD. The concentration
of FC-780 was gradually increased until a level of 1000 ppm
COD was achieved. Samples were taken three times per week.
BOD, COD, TOC and pH analysis were performed as described
under the chemostat procedure. Other parameters measured
were:
933
-------
A. Temperature. Readings were taken three times per
week, utilizing a Wahl digital heat-prober ther-
mometer. The thermometer was placed directly into
each of the four stages of the RBC.
B. Microorganism Identification. Bacterial and fungal
populations were identified and enumerated utiliz-
ing: Nalgene Nutrient Pad Kits — TTC and Wort,
Bio Stix and Myco Stix test strips (Ames Company)
and/or Total Count Water Tester (Millipore Corp).
Microscopic qualitative observations were done
every 14 days to visually monitor changes in
predominant populations, i.e. protozoal, fungal,
and nematodal.
RESULTS AND DISCUSSION
Chemostat
After 7 days of continuous operation, an apparent
steady state condition was achieved within the chemostat in
terms of COD, TOC and BOD conversion or percent (%) removal
from the supernate. Approximately 70% COD conversion, 80%
BOD conversion and 60% TOC conversion were consistently
observed from day 11 onward to day 43 (Figure 3). From day
40 until shut down of the chemostat on day 63, the percent
(%) conversion dropped to approximately 45% COD, 50% BOD and
40% TOC. This was in part correlated with a rise in the pH
of the reaction vessel to a pH of 7.1 or greater. The
microbial populations observed in the chemostat changed
drastically with the increase in pH. That is, a greater
number of yeast and slime-forming bacteria were noted. No
effort was made to readjust the pH of the chemostat and so
the percent conversion in all three dropped to a level of
40-50% conversion. It was decided for future experimenta-
tion to adjust the pH of the RBC to 7.00.
As seen in Figure 3, the percent (%) conversion.values
exhibited some variance. This is partially due to technical
errors and machine failure. That is dilution and sampling
errors were committed during a turnover of technical assis-
tance. Equipment failure would occur and no new influent
would enter the reactor vessel for a 12 to 16 hour period.
This would result in microbial back contamination from the
reactor vessel into the influent reservoir, which would
result in an increase in the pH and a decrease in available
934
-------
CO
CO
O
M
O
O
§
co
c
u.
-------
carbon in the influent and thus a lower TOC, COD and BOD
conversion measurement. It should also be mentioned that
plate counts were performed infrequently and were more for
qualitative determination of different microbial popula-
tions.
From the chemostat data, it was decided that enrichment
for mixed microbial populations, that were able to utilize
FC-780 as their sole source of supplemented carbon, was
possible. The data from the chemostat also indicated that
it was possible to change an influent containing 1200 to
1500 ppm COD of FC-780 into an effluent containing 100-200 '
ppm COD.
Rotating Biological Contactor
At initial start-up, 0.2% glucose was added via the
influent to stage 1 of the RBC at a flow rate of 4 ml/min.
This proved to be too high of a concentration of glucose in
that the pH of the RBC rapidly became acidic and was thought
to endanger the not yet well established microbial popula-
tion. Therefore, the concentration of added carbon, in the
form of glucose, was dropped to 0.1%. However, the majority
of this carbon (97%) was used in stages 1 and 2, and stages
3 and 4 failed to exhibit growth on the discs. To achieve
colonization of all the RBC discs, 0.1% carbon-source, in
the form of a 10X concentrate, was added to each stage
daily. This also stopped the recurring back contamination
into the sterile influent reservoir, which now contained BHB
only.
It was noted that the pH of the effluents daily dropped
into the acidic range (6.0 - 6.9) and had to be chemically
adjusted. After 3 days, nutrient broth was added in the
form of a 10X concentrate, along with the 10X glucose, to
result in a final concentration of 0.1% carbon. It was
thought that whereas glucose was metabolized aerobically
into acids, the nutrient broth would be metabolized with the
resulting release of amino groups. This would help to raise
the pH, and the protein itself would also act as an addi-
tional buffer. This provided adequate pH regulation unless a
malfunction in the equipment or a laboratory error occurred
which resulted in a decrease in the pH of one or glucose was
metabolized aerobically into acids, the nutrient broth would
be metabolized with the resulting release of amino groups.
This would help to raise the pH, and the protein itself
would also act as an additional buffer. This provided
936
-------
adequate pH regulation unless a malfunction in the equipment
or a laboratory error occurred which resulted in a decrease
in the pH of one or more of the stages.
The resulting reduction in COD, BOD, and TOC are
presented in Figures 4, 5, and 6, respectively. In these
figures, percent removal is shown with respect to time given
in days. After 30 days of continuous operation, an apparent
steady state condition was achieved within the RBC, in terms
of COD, TOC and BOD removal. Approximately 97% removal was
achieved in all three parameters measured. As seen in
Figure 7, exposure of the RBC to FC-780 began on day 35 with
the addition of 0.025% FC-780, or 100 ppm in terms of COD."
By day 60, 1000 ppm COD of FC-780 was being fed. Simul-
taneously, the amount of nutrient broth, which was the only
other carbon source after day 52, was lowered to a level of
approximately 500 ppm COD. This level of carbon was main-
tained until day 80. The conversion rate at that time was
98% COD, 96% BOD, and 94% TOC. A one way completely ran-
domized analysis of variance was conducted on each stage
with respect to COD, BOD or TOC. These results are given in
Table I.
Table I. RBC, One-Way Completely .Randomized
Analysis of Variance Versus Bartlett's ,
Variance
Analytical
Form
COD
BOD
TOC
One-Way
F
441.846
175.475
53.871
Analysis
Significance
0.000
0.000
0 . 000
Bartlett
F
34.543
17.659
91.259
1 s Variance
Significance
0.000 .
0.000 •"
0.000
The calculated F
,
'
value for this test , would-be
3.47. The values shown in'Table I, being larger than the .
calculated F, are indicative of a significant variation
between treatment and non-treatment with the RBC. The
Bartlett's test of homogeneous variance indicates no viola-
tion of the homogeneity assumptions of ANOVA. The signifi-
cance levels show low probability of error within the tests.
937
-------
si-
0)
6D
•H
CD
Ifl
*
09
CO
1-
1-
o
\f>
-------
m
00
i
UJ
939
-------
G»
CO
-------
CO
CM
0)
M
a
W)
CO
i
UJ
z:
1-1
CCOO Udd
941
-------
It can be seen from these results that the use of the RBC
was very significant in treating wastewater containing up to
1000 ppm COD of FC-780. .
Microbial populations observed are given in Table II.
A strong, heterogeneous population was observed throughout
the experiment. Although some changes in densities
occurred, the organisms listed in Table II were seen
throughout the experiment.
CONCLUSION
The purpose of this research was to determine the
feasibility of treating wastewater containing aqueous film
forming foam (FC-780) by an aerobic RBC. The preliminary
data presented here demonstrates that FC-780 is conducive to
aerobic bioconversion and removal with a properly adapted
RBC unit. A significant reduction in COD, BOD and'TOC has
been achieved in synthetic wastewater containing up to 1000
ppm COD of FC-780. It is possible that FC-780 loading may
be increased further and that higher reduction of the
parameters may be obtained by changes in flow rates and
contact times. These possibilities are being actively
addressed in preparation of scale up for pilot plant
operation.
Table II. Microbial Groupings as Observed on
Suspended Microscope Slides in the
Aerobic Rotating Biological Contactor
Nematodes
Fungi
Filamentous
Ex: Aspergillus
Pennicillium
Non-filamentous
Protozoa
Sarcodina
Ciliata
Ex: Suctoria
Zoothamnia
Voticella
Paramecium
Bacteria
Gram Positive
Staphylococci
Streptococci
Bacillae
Gram Negative
Bacillae
:Filamentous
bacillae
Algae
Phaeophyta
942
-------
REFERENCES
1. Chian, E.S.K., Wu, T.P., and Rowland, R.W.; "Membrane
Treatment of Aqueous Film Forming Foam (AFFF) Wastes for
Recovery of AFFF Active Ingredients", Final Report, Georgia
Institute of Technology, October 1980.
2. LeFebvre, E. E.-, and INMAN, R. C., "Biodegradability and
Toxicity of Lightwater FC-206 Aqueous Film Foaming Foam",
Report No. EHL (K) 74-26, USAF Environmental Health
Laboratory, Kelly Air Force Base, Texas, November 1974.
3. Chesler, G. , and Eskelund, G.R., "Rotating Biological
Contactors for Munitions Wastewater Treatment", Report 2319,
U.S. Army Mobility Equipment Research and DeVelpment
Command, Fort Belvoir, Virginia, February 1981.
4. American Public Health Association, "Standard Methods
for the Examination of, Water and Wastewater", 14th edition,
APHA, Washington, D.C.
5. Technicon Industrial Systems. 1976. "TIS Education
Department. Customer Training Manual." Technicon
Instruments Corporation, Tarrytown, N.Y.
6. Beckman. 1979. "Model 915B Total Organic Carbon
Analyzer." Beckman Instruments, Inc., Fullerton, CA.
943
-------
OPERATION OF A RBC FACILITY FOR THE TREATMENT
OF MUNITION MANUFACTURING PLANT WASTEWATER
Leonard L. Smith, Senior Technical Engineer, and
Wayne C. Greene, Water and Wastewater Engineer,
Hercules Aerospace Division, Hercules Incorporated,
Radford Army Ammunition Plant, Radford, Virginia
INTRODUCTION
In 1970, the Army initiated an extensive pollution abate-
ment program at all of its ammunition plants. The treatment
of waste process waters from these plants required development
of new or modifications of existing technology because of the
unique nature of the pollutants in the wastewater. This waste-
water contains ether, alcohol, acetone, inorganic nitrates,
traces of nitroglycerin (NG), and other propellant ingredients,
The initial wastewater treatment studies(1) were conduc-
ted using an activated sludge process. This study demonstrated
that the activated sludge process was not a feasible treatment
method for the RAAP wastewater due to the high variability in
flow and organic concentrations. A successful rotating bio-
logical contactor (RBC) pilot plant evaluation(!) was
conducted to define the design parameters and develop the
design criteria for a full-scale facility.
FREATMENT PLANT DESIGN
A RBC wastewater treatment facility was constructed at
RAAP for the treatment of the process wastewater based upon
944
-------
the design criteria developed from the pilot plant evaluation.
This facility consists of a 5110 m3 (1,350,000 gal) equaliza-
tion basin and eight RBC shafts containing a total surface
area of 56,782 m2 (612,000 ft2). Since the equalization basir
would probably develop a dispersed biological growth, even
without the addition of -nutrients, four 15-hp floating aera-
tors were provided to mix the basin and prevent sedimentation
of suspended solids, and to provide adequate aeration to
satisfy the oxygen uptake rate of the dispersed growth.
The RBC system (figure 1) was constructed to provide two
separate parallel RBC systems, each 'system consisting of three
stages. Stage one of each system contains two RBC shafts
while the other stages contain one shaft in each stage. The
design parameters for the RBC facility are shown in table I.
Table I. RBC Design Parameters
Flow
Chemical
Oxygen
Demand
(COD)
Biochemical
Oxygen
Demand
(BOD)
4716
(1,250,000)
Avg Flow Rate - m-Vday
(gpd)
Design Load - kg/day .
- (Ib/day)
« - mg/1
Avg Hydraulic Loading
-. - m3/m2.d
-; - (gpd/ft2)
Avg Organic Loading
- kg/1000 m2.d
- (lb/day/1000 ft2)
Discharge Limitations, maximum
- Average daily (mg/1)
- Maximum daily (mg/1)
2140
(4718)
452
856
(1888)
181
0.08
(2)
37.7
(7.7)
195
290
15.1
(3.1)
60
120
945
-------
OQ
CD
CO
o
946
-------
FACILITY START-UP fl?
The facility was placed into operation -during December
1980. The RBC basins were filled with process wastewater from
the equalization basin, 3.8'm3 (1000 gallons) of waste sludge
from a local municipal activated sludge wastewater treatment
plant, and one liter of phosphoric acid. The RBC- system was
operated in a batch mode for eight days to allow a biomass
growth to develop on the. RBC media. During this time ethyl
alcohol, potassium nitrate, and phosphoric acid were added
each day as nutrients. Soda ash was added as required to con-
trol the wastewater pH. The initial wastewater parameters
following start-up were: pH 8,1, temperature 8 °C (47°F),
dissolved oxygen (DO) 10.8 mg/1, nitrates (N) 14 mg/1, and
phosphates (P) 4 mg/1. The biomass growth on the RBC media
developed very > s lowly due to the low temperature. Sufficient
alcohol, nutrients, and soda ash were added each shift to
maintain the chemical oxygen demand (COD) between 100 . to 200
mg/1, nitrates (N) 5 to 30 mg/1, phosphates (P) 1 to 5 mg/1,
and the pH between 6.5 and 7.8. On the fourth day of operation
the wastewater temperature increased to 12 °C (53°F) and a very
noticeable acceleration of biomass growth was observed.
The RBC operation continued in the batch mode :until the
eighth day, at which time an influent flow of 1.14 m3/min (300
gpm) was started. The flow rate to. the RBC units was steadily
increased over the next few days \m to 3 m3/min (800 gpm).
pH FLUCTUATIONS '
Shortly after start-up the facility encountered a .period
of pH fluctuations. During this period, the influent pH
varied from 5.3 to 10.7 (figure 2). The variations in the
influent pH were the. result of a new pretreatment facility
being unable to accurately control the acid feed rate for the
pH control system. Until this problem was corrected, an
attempt was made to adjust the pH in the equalization basin by
the addition of soda ash; however, due to the. absence of rapid
nix equipment, this method was not completely successful.
As a result of. these pH excursions, most of the biomass
the RBCs sloughed, off. However, the biomass recovered
without reseeding.
[NITIAL OPERATION .
During the next month of .operation, despite the cold
additional PH excursions, and a highly
947
-------
(O
e
•^
(D
N9
•H 3
en
CD
OO
948
-------
variable organic load, the biomass growth continued to improve.!
The hydraulic loading to the RBC units was maintained constant:
except for adjustments necessary to maintain the equalization :
basin between the 60 to 90 percent level. However, wide '•
variations in the organic concentration,of the wastewater i
caused the COD removal efficiency to consistently fluctuate
between 40 and 80 percent. Grab samples of the influent and
effluent were collected each .morning, five days a week.
Figure 3 shows the COD of 'these samples collected during the
second month of operation. The low influent CODs on Monday of
each week were due to no manufacturing operations on weekends.
FULL FACILITY OPERATION !
In an effort to provide better control of the organic
loading to the RBCs, an on-line total organic carbon (TOG)
analyzer was installed to monitor the RBC influent. The on-
line TOG analyzer-verified that, even with a one-day retention
time in the equalization basin, sharp fluctuations in the
organic concentrations of the RBC influent were occurring. A
typical shock load caused an increase in TOC of,; 330 percent in
less than three hours (figure 4); The results of the TOG
analyses over a 23-day period are shown in figure 5. It was
obvious from these data that a method of controlling the
organic loading to the RBC was needed. The equalization basin
was designed to be operated at between 60 and 90 percent
Capacity. Therefore, by maintaining the level of the basin at
the low level when high organic loadings were expected and at
the high level when low organic loadings were expected, the
influent flow rate could be varied to minimize these fluctua-
tions. A chart was prepared for use by the plant operators
to control the organic loading to the RBC, based upon the
influent flow rate and TOC value (figure 6). The instructions
provided with the chart were as follows:
1. Keep the loading to the RBCs in the same loading zone
whenever possible. Change the flow rate in small steps when-
ever it is necessary to change zones.
2. Decreas.e the basin level during periods of low
organic loadings. Decrease the level to 60 percent on'Sunday
of each week.
3. Increase the basin level during periods of high
organic loadings. Increase the level to 90 percent on Friday
of each week.
4. The operators should maintain a record of the
influent loadings by plotting the changes on the chart, using
a new chart each day, and recording the time of changes.
949
-------
CO
O
TO
TO
O
CO
950
-------
o
r»
T
(O
O
= 3
^ m
£ i
951
-------
TOC (mg/l)
CO
(TO
—I
CD
CJI
5O
CD
O
O
3*
952
-------
1800 -
1600 -
1400-
1200 -
1000
UNDESIRABLE
LOADING
UNDESIRABLE
LOADING
FLOW RATE
(m '/min)
I
FLOW RATE
(gpm)
450
33 50
100 150
100 200 300 TOC
300 600 900 COD
CONCENTRATION (mg/l)
Figure 6 RBC INFLUENT ORGANIC LOADING CHART
953
-------
The use of this chart and the TOG monitor helped control
the organic loadings to the RBCs, but did not eliminate .
periodic shock loadings. Figure 7 shows the influent loading
and effluent discharge for a two-month period. During this
time, both of the RBC systems (eight shafts) were being used
equally. The data for this chart were from analyses of grab
samples collected at the start of. day shift, five days a week.
Twenty-four hour composite samples are collected once a week
to verify compliance with the discharge permit. The results
from these analyses show that the facility is meeting the dis-
charge limitations.
The data from this same period were plotted as COD
applied versus COD removed (figure 8) to determine the COD
removal at various influent loadings. As can be seen from
this graph, the data appear to be very scattered. A straight
line represented by the equation y = 0.83 x -0.6 was drawn on
the graph to represent the normal operation of the RBC system.
This corresponds closely to the equation developed during the
pilot plant evaluation^1) of (y = 0.83 x -1.2). This pilot
plant equation was developed at steady state loadings which
were varied over a period of time from 60 to 90 kg/1000 m2-d
(12 to 18 lb/day/1000 ft2). The data points on figure 8 fall-
ing considerably below this normal operation line .are indica-
tions of stresses on the system due to shock loads.
PARTIAL FACILITY OPERATION
During the period of time the above data were collected,
it can be seen from figures 7 and 8 that the RBC COD loadings
werebelow the average design.loadings of 37 kg/1000 m2-d (7.7
lb/day/1000 ft2) most of the time. In order to evaluate the
facility at design loadings, one RBC system (four shafts) was
shut down in October 1981. The biomass growth on the RBC's
media became heavier within the first few days of operation of
only one system. The COD data from the grab samples, collect-
ed while only one system was operating, are shown in figures 9
and 10. The COD loadings during this period fluctuated as
greatly as in the previous study; however, it can be noted
from figure 10 that fewer data points fell considerably below
the normal operating line. It appears that by increasing the
organic loadings, thereby causing a heavier biomass, the
system was more tolerant to shock -loads. A detailed study has
not been conducted to determine the effects of temperature on
the organic removal efficiency. However, during the last 30
days of this study, the wastewater temperature varied from 5°
to 10°C (42° to 50°F). It was noted by visual inspection
954
-------
COD (lb/day/1000 ft )
GO
O
ISA
5 CO
CD
O .-.
o a
a >
55
O
ra
o
CO
CO
o
CJ1
COD (kg/1000 m2d)
955
-------
Q.
CD
O
i s
o o
50 >
-
cr
CO ^x
o a-
m.
^ 5
2 °
o <
o ^
7 °
N)
m .«
956
-------
COD (lb/day/1000
-------
„_ n
O. 73
Q m
2 —
o
m
> __k
m
Q. —4
(O
m
00
CO
—4
m
o o
11
s\
X >
CJ1
958
-------
that the colder wastewater 'caused the biomass growth to
increass to compensate for the decreased activity of the
biomass.
CONCLUSIONS
Based upon the one-year operation of the RBC facility on
the wastewaters from this manufacturing facility, the follow-
ing preliminary conclusions are drawn:
1. The biomass on RBC units .can stand extreme stresses
from pH, temperature, and shock organic loads without the
system failing to a degree that it does not recover when the
stress is removed.
2. A sudden-increase in organic loading will not immedi-
ately increase the organic removal rate. However, the biomass
growth will increase rapidly to increase the removal rate. The
biomass growth rate will be directly related to the wastewater
temperature because of the slower growth at lower temperatures.
3. Low wastewater temperature does not significantly
decrease the organic removal efficiency at normal operating
levels. However, at higher organic loadings, the RBC may not
have .the capacity for the additional biomass growth.
REFERENCES
1. Smith, L. L., and Zeigler, E., Jr., "Biological Treatment
of a Munitions Manufacturing Facility Wastewater,"
Proceedings of the 33rd Industrial Waste Conference,
Purdue University, May 1978, pp. 432-439.
959
-------
TREATMENT OF STARCH INDUSTRIAL WASTE
BY RBCs
Chun Teh Li and Huoo Tein Chen, Department Of
Environmental Engineering, National Chung Kung
University, Tainan, Taiwan, R.O.C.
Yeun C. Wu, Department of Civil Engineering,_
University of Pittsburgh, Pittsburgh, Pa.
INTRODUCTION
Although successful treatment of industrial wastewaters
by rotating biological contactors (RBC) has been earlier re-
ported by many researchers, one important factor limiting the
plant performance is the availability of oxygen in the system,
in particular, when treating a high-strength organic waste
(1-7). Oxygen transfer can become more effective by increasing
the disc rotational speed but this application, however, is
not practical because of: (a) high power consumption, and (b)
high biomass slough-off induced by a high hydraulic shearing
force on the surface of biodisc that reduces the overall treat-
ment efficiency.
One alternative that may be able to treat high-strength
organic wastes successfully under the normal operating cond-
itions, is to replace the source of conventional air by pure
oxygen. Bintaja et al. first employed oxygen rn RBC system
for treatment of Cheese waste (8). Later, Huang successfully
treated synthetic milk waste by oxygen-enriched RBC system.
960
-------
They concluded that use of pure oxygen in sufficient amount
not only was able to improve COD removal satisfactorily but
also increased sludge settleability (9).
The present study was aimed to investigate the feasibility
of using pure oxygen and conventional air RBC systems for the
treatment of starch processing wastewater and also to study the
effects of wastewater property, organic loading, pH, and dis-
solved oxygen concentration on ,RBC plant performance. The final
goal of this research was to determine the kinetic data for
future design of both RBC systems.
Starch Industry Wastewater. Starch consumption in the
Republic of China (Taiwan) now reaches one hundred thousand
tons annually, of which one half of this demand is supplied by
local manufacturers and the other half is imported from the
outside producers. As a result, Taiwan government has decided
to assist private daily food industry in expanding the existing
starch processing plants prior to 1984, so that the future
domestic starch demand can be met.
The raw materials used presently for starch production in
Taiwan are sweet potato and corn. Approximately 70% of the
total starch is made of sweet potato because it is available
locally. However, corn starch is more perferably used by
consumers even the cost is higher. It is predicted that the
use of corn for starch-manufacturing will increase consider-
ably due to the current market demand.
An expansion of the existing starch processsing facilities
is greatly concerned by environmental scientists and engineers
in Taiwan. Because normally each ton of corn used for strach-
making could produce 13.5 - 15.1 kg BOD and, more importantly,
most of the starch manufacturing plants located in Tainan and
Chiayi directly discharged their effluents into the small
receiving streams. Purification of the pi ant-effluent waste-
waters is essentially necessary. The present study was initiated
to investigate the treatment of corn strach manufacturer plant
effluent by RBC systems. The flow diagram as presented later
explains the process employed for corn starch manufacturing.
It can be seen in the flow diagram that although a closed-
loop system is designed for the purpose of eliminating the dis-
charge of process water, a considerable amount of wastewater
is preduced due to leakage, overflowing, and accidential spill.
Normally, a 0.4 - 3.0 m3 wastewater generated per ton of corn
used was. found in a corn-starch manufacturing plant.
961
-------
CORN
I Vapor
-1— stee r
SteeBatgr | Evaporated
Dewatering
Drying
Protein
Dewatering
1 Moist
I Stare
Drying
1
Starch
Flow Diagram of Corn Starch Plant
962
-------
Biokinetics
The model employed for this study has been earlier pro-
posed by Monod and Clark et al.(lO,ll). The basic equations
are given as follows:
V(ds/dt) = Q S0 - Q Se- (ua/Y)
- (US/YS)
-(1)
in which
3
V = reactor volume, m
ds/dt = rate of substrate removal, mg/1-sec
3
Q = waste flow, m /day
S = influent substrate concentration, mg/1
S = effluent substrate concentration, mg/1
u = specific growth rate of attached biomass, day
a _i
u = specific growth rate of suspended biomass, day
X = weight of attached biomass per unit disc
a surface area, g/m2
X = concentration of suspended solids, mg/1
s 2
A = total disc surface, m
w
Y = yield coefficient of attached biomass
(mass of biomass produced in kg/ mass of
substrate removal in kg)
Y = yield coefficient of suspended biomass
s (mass of biomass produced in kg/ mass of
substrate removal in kg)
In the fixed-film RBC system, substrate removal by
attached biomass is much greater than suspended biomass. So,
it is reasonable to assume that the term (US/YS) xs v in
Eq. 1 can be eliminated. And then Eq, 1 becomes
v(ds/dt) = Q SQ - Q
- (ua/Y)
•(2)
963
-------
According to Monod, the change in biomass under sub-
strate limiting condition can be expressed as
= um { S/ (Ks + S)}
(3)
in which
-1
u = maximum specific growth rate, day
KS = half saturation constant, mg/1
S = limiting substrate concentration,
mg/1
Eqs 2 and 3 can be combined under a steady-state
condition (ds/dt)= 0 and expressed in linear form as shown
below.
1A = (K/e) (1/S ) + 1/e
•(4)
in which
X = (Q/AW)(S0 - S*)» mg/m2-day
e = (u /Y) X* mg/1-day
* a
S = effluent substrate concentration
* at steady-state condition, mg/1
X = attached biomass concentration at
steady-state condition, mg/1
The values of K and e can be calculated by plotting
1/x versus 1/S . The ordinate intercept and slope of
the line are equal to 1/e and KS/B, respectively.
Materials and Methods
The chemical composition of the starch processing waste-
water employed for this study is shown in Table 1. Apparently,
organic carbon concentration as BOD or COD and solids content
were high and the wastewater pH was in the acidic condition.
Throughout the entire study, the BOD, COD, pH, DO were
monitored in each stage of the RBC system. The laboratory
procedures used to determine the above mentioned parameters are
specified*in "Standard Methods" (12). The % oxygen content in
feed gas was measured in each stage of the pure oxygen RBC
964
-------
system by Beckman Oxygen Analyzer.
Table 1. General Property of the Wastewater
Parameter
Concentration, mg/1
BOD
COD
Organic-N
Total Solids
Total Volatile Solids
PH
BOD/N ratio
N/P ratio
2,700-3,900
5,200-7,100
300-400
3,300-5,400
2,900-4,700
4.05-4.55
10: 1
5: 1
Two RBC pilot plants were constructed identically in size.
The main difference between them was that the biodiscs in one
system were exposed to air and in other system they were
constantly contacted with pure oxygen. The physical structure
of these pilot plants is described in Table 2.
A total of 47% disc area was submerged and all discs were
supported by a common shaft in each stage. Three shafts in
each system were driven by the same motor so that they were
rotated at the same velocity. However, this motor enabled to
rotate at the speed varying from 0 to 200 rpm.
To produce an airtight condition for pure oxygen RBC
system, a plexiglass cover was bolted down at the top of each
biological reactor. The shaft was made of a V steel rod that
was connected to motor drive system through the shaft holes
at where h" 0-rings were poisitioned. Replacement of 0-rings
was necessary when they were worm down from friction. The re-
965
-------
Table 2. Pilot Plant Specifications
(1)
Biodiscs
Diameter, cm
Thickness, cm
•Spacing, cm
Number
2
Effective area, m
% Submerged area
1st
26
0.2
4
18
1,894
47
Stage Number
2nd
26
0.2
2.8
25
2,630
47
3rd
26
0.2
2.2
31
3,261
47
(2) Reactor
Material
Cress Sectional Area.
Length, cm
Volume, liters
Gross
Net
(3) Enclosure
Material
Size, cm
Volume, liters
Plexiglass Piste
260 x 405
760 760
25.27
24.40
25.27
24.00
760
25.27
23.70
plexiglass cover
750 x 250 x 150
39 38.6 38.2
* in the pure oxygen system only.
966
-------
actor structure is presented in Figure 1 in detail. All experi-
ments, were performed at a room temperature - 28 ± 2°C.
50
160
cm
50
cm
Pure 0?
In
Wastewater
In
30 30
260
cm
Effluent
to next
stage
O E
LT) O
LO E
tn o
O E
01 o
30 30
Figure 1. Structure of Pure Oxygen RBC
967
-------
Results and Discussion
To study the feasibility of using the RBC systems for
the treatment of starch manufacturing wastewater, the effects
of influent substrate concentration, organic loading, DO,
pH, disc rotational speed (RS), and hydraulic retention time
(RT)on substrate removal efficiency were investigated.
(1). Disc Rotational Speed. Antonie earlier suggested
that the rotational speed of biodiscs should be controlled
to exceed one foot per second (18.3 m/min.) so that oxygen
limitation could be avoided (8). The effect of disc rotation-
al speed on treatment efficiency was tested under the influent
substrate concentration ranged from 370 mg/1 to 1,300 mg/1
as BOD and from 860 mg/1 to 2,900 mg/1 as COD. The rotation-
al speed was controlled at 15, 20, and 25 rpm, respectively.
Table 3 summarized the results of this study conducted in
the pure oxygen RBC system. It was found that for all substrate
conditions presently investigated, neither BOD nor COD
removal was significantly increased due to the increase in
disc rotational speed. In other words, there is no need to
maintain the rotational speed over 15 rpm (12.3 m/min.),
according to the present study.
Table 3. Effects of Influent Substrate
Concentration and Disc Rotational
Speed on % COD and BOD Removal
Influent
LUU tone.
(mg/1)
860
1,580
1,900
2,300
2,900
% Reduction
Rotational Speed,
15
91.9
94.6
93.0
88.3
87.2
20
91.4
94.3.
93.6
90.7
87.8
rpm
25
91.6
94.0
93.0
91.0
88.3
968
-------
Table 3. Continued
Influent
BOD Cone.
(mg/1 )
370
650
900
1,000
1,300
'% Reduction
Rotational Speed, rpm
15
99.1
98.9
98.0
97.1
96.2
20
99.2
98.7
98.2
98.0
96.5
25
99.0
98.5
98.4
98.0
96.8
(2). pH Level. To obtain an optimum treatment efficiency,
the acidic starch wastewater was first neutralized by NaOH and
then mixed with KH2P04 solution to increase buffering capacity.
The pH level after adjustment was controlled to be within the
range of 8.1 to 8.9. Because of the chemical nature of the
wastewater, the pH was reduced approximately one unit during
12 hours storage. Normally, the higher the organic content in
the wastewater the faster the pH drop. For this reason, the
pH of the feed solution was adjsuted in accordance with COD
concentration. Table 4 shows the results of pH changes in the
feed as well as in each stage of the RBC systems.
The data as seen in Table 4 indicated that in most cases
the reduction of pH occurred in the first stage of the system
and an increase was found in the subsequent stages. The pH
drop was probably due to the accumulation of organic acids
resulted from active decomposition of organic substances where-
as the pH increase was induced by CO? production, a metabolic
by-product in this case. Additionally, the present study shows
that the influence of substarte concentration, rotational speed,
and retention time on pH was not significant.
969
-------
It Is eyident fro.m this study that the buffer solution added
to the wastewater is adequate to maintain a proper pH between
7,0 and 7,6. However, it i$ important to investigate whether
the chemical addition of phosphate buffer can be reduced, and to
determine its minimum requirement. Further study is essentially
needed for the future operation of the RBC system.
Table 4. Effect of Influent COD, Disc
Rotational Speed, and Hydraulic
Retention Time on pH
RBC
System
Operating Condition
COD Rotational Retention
mg/1 Speed, rpm Time,
Oxygen 1
2
3
4
5
1
2
3
4
5
1
2
3
4
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
15
15
15
15
15
15
15
15
15
15
15
15
15
15
12
12
12
12
12
6
6
6
6
6
4
4
4
4
hrs
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
PH Val
ue
Stage Number
Feed*
3/6.
5/6.
8/6.
8/8.
9/7.
2/7.
3/7.
5/7.
7/7.
9/7.
1/7.
4/7.
6/7.
8/8.
5
7
7
3
8
1
6
6
3
9
4
3
5
0
1st
7.
7.
7.
7.
7.
6.
6.
6.
6.
7.
7.
7.
7.
7.
1
3
2
0
1
9
9
9
7
0
6
1
1
0
2nd
7.
7.
7.
7.
7.
6.
7.
7.
7.
7.
7.
7.
7.
7.
2
4
3
1
5
9
1
1
1
1
6
2
3
1
3rd
7.4
7.5
7.5
7.2
7.6
7.0
7.3
7.2
7.4
7.4
7.6
7.3
7.6
7.3
970
-------
Table 4. Continued
DDP
KbL
System
Oxygen
Operating Condition
' ' c
COD Rotational Retention
mg/1 Speed, rpm Time, hrs Feed*
2,000
3,000
4,000
5,000
20
20
25
25
12
12
12
12
8.4/6.4
8.9/6.8
8.9/7.2
8.9/8.4
pH Value
tage
1st
7.4
7.3
7.5
7.1
Number
2nd
7.3
7.4
7.6
7.2
3rd
7.5
7.6
7.7
7.2
* (initial pH/pH after 12 hrs storage)
(3). Oxygen Consumption and DO Level, the effect of
organic loading on oxygen consumption and DO level under dif-
ferent disc rotational speeds (15, 20, and 25 rpm) was studied
for both air and pure oxygen RBC systems. The results are
shown in Table 5.
A. In Pure Oxygen RBC System. The feed gas contained 99.5%
oxygen and the feed rate was constantly controlled at 1,000 cc/
min.. Table 5 clearly shows that the % 02 gas remaining and the
% 0? utilization decreases as the number of RBC stages increases.
But the influence of organic loading on oxygen utilization is
different, that is the % Consumption increases or the %' 02
remaining decreases when tne organic loading increases. These
results are expected because: (a) the removal of organic substrate
takes place rapidly in the first stage of the RBC system that
consumes more oxygen, and (b) in addition to the oxygen uptake
in each stage, the accumulation of C02 reduces the % 02 remain-
ing in the feed gas.
The DO level in the oxygen RBC system was highly affected
by the influent substrate condition. In two cases presently
investigated, the DO level reached zero in the first stage as
the influent COD and organic loading exceeded 5,000 mg/1 and
95 g. COD/m2-day, respectively, in accordance with Table 6, Add-
itionally, it was also found that the DO level increased as the
number of RBC stages increased.
971
-------
Q)
S-
4J
re c
s- o
.»-> -r-
C +J
cu re
o N
cu =3 .;-
re c E
&- cu cu
4-> C7>4->
to >> w
J3 X >,
=J O CO
to
c o
M- O CO
o cc
cn
•<-><= c
O -r- CU
cu TD cn
<4- re >>
«t- O X
UJ _J O
LO
CD
r—
re
C
o
'•£ "O
re s- s-
N CU CO
.— .0
43 Z E
13 CM
CU
O. cr
o re 4->
^S 4-» CO
CO i—l
CT S- T3
E CU S-
•i- £) CO
CU CU CM
rv ri"
ojre
O 4-> CO
&S CO i-H
cn re
E -o
•i- i
o -ooj
o re S
—l cn
E
CU r—
3 0 ^
i— O cn
«r- 0 E
E
re
O T3
•r- CU E
4-5 CU Q.
re o. s.
4-> CO
c?
[^
•
o
^.
*
T-H
i— 1
•
f-
CO
*
o
CTI
o
i-H
^J.
•
OJ
CTi
CTi
T-i
o
o
o
I— 1
LO
I-H
o
r— <
^
•
t-H
f*^
•
LO
CO
o
CO
l-H
CT>
CO
OJ
CTI
CTi
t— 1
§
o
t— 1
o
CM
ID
0
O
OJ
CTI
o
CM
•
CO
CO
CO
LO
•
CO
CO
CO
CO
o
0
o
OJ
LO
i-H
1^,
•
o
^
•
r-i
f^Hfc
•
cn
«^.
•
s.
CO
i-H
CO
CO
CO
*
CTi
CO
CO
CO
o
o
o
ft
OJ
o
OJ
LO
•
o
LO
•
OJ
LO
•
OJ
^
•
IO
CO
i-H
CO
LO
cn
CO
CO
CO
o
o
o
rt
CM
LO
OJ
CTi
•
o
CT)
•
1— I
LO
*
CO
CM
•
co
CO
i-H
CO
o
•
LO
CO
p*^
LO
o
CD
o
*t
CO
LO
I-H
t-H
•
t-H
CO
•
i-H
^^
•
CM
CTI
•
CO
CO
o
LO
CO
CO
ID
CO
p^
LO
o
0
o
ft
co
o
OJ
r^
•
o
CO
•
OJ
CTI
•
CM
LO
•
co
CO
co
CO
LO
•
LO
CO
l"x.
LO
o
0
o
rt
co
LO
OJ
CM
•
t-H
p^
•
r- 1
1—*
•
s
LO
•
CM
CO
r--
co
CO
«3"
•
LO
CO
LO
r^
o
o
o
*t
"^
LO
t-H
LO
•
t-H
LO
•
OJ
LO
•
LO
o
•
cn
•*
i-H
CO
0
^.
CO
LO
CTi
o
o
o
rt
LO
LO
l-H
. 1
CO
•H
I, 1
cu
JS
JJ
o
4J
C
•H
18
"u
fl)
"3
w
s
CO
fcC
13
GJ
o
m
£
fl)
00
>>
tn
o*\
o\
<
W)
n)
972
-------
Table 6. Effects of Substrate Concentration
and Loading on DO Level in Oxygen
and Air RBC Systems
Influent
COD
mg/1
1,000
2,000
3,000
4,000
5,000
7,000
COD
Loading
2
g/m -day
19
38
57
76
95
133
1
Oxygen RBC
stage number
1st 2nd 3rd
8.5 19.3 21.6
6.0 20.3 18.0
4.2 13.0 16.0
1.5 9.8 12.0
0 6.0 9.0
0 0.2 0.5
2
Air RBC
stage number
1st
4.6
1.6
0
0
0
0
2nd
8.0
4.0
1.5
0
0
0
3rd
8.6
7.5
5.0
0
0
0
1. Disc rotational speed= 15 rpm
2. Disc rotational speed= 20 rpm
B. In Air RBC System. The effect of substrate concent-
ration and loading on DO level in the conventional air system
is similar to the pure oxygen system. However, the DO levels
for all three stages in the air RBC system were considerably
lower. An anaerobic condition was found in the first stage
of the system when the organic loading was equal to 57 g. COD/
m2-day and in all stages when the organic loading exceeded
76 g. COD/m2-day.
Biological response to oxygen deficiency in RBC systems
shows no significant difference from the other types of waste-
water treatment processes. Two interesting evidences have
been observed, one is the conversion of sulfate compounds
973
-------
CO
•—» o
a* o
DO i—
O o
O P*
CL
CQ
o
o
O 0»
o o.
o -.
70 CQ
P>
ro
vo
o
o
en
i-*
VO
en
en
en
t-»
eo
00
»-J
en
co
o
VJ
t-t
00
00
1—"
ID
CO
ro
i— '
00
en
•
t—>
t-»
ro
en
en
ro
en
t-*
00
ro
co
en
0
*>
CO
o
o
-e»
-p.
-vj
CO
I—"
VJ
Cn
CO
co
en
ro
i— >
-e»
i—"
•-J
CO
ro
t— >
00
i-t
en
•£*
I—"
00
00
Cn
en
00
en
ro
t-*
t— •
00
§
CO
ro
CO
o
CO
-P»
ro
en
ro
CO
CT>
*>
-P»
^
ro
VO
vo
ro
ro
>-j
CO
H-*
CD
1— »
CTi
O
t— «
«-J
en
CO
CO
1— '
vj
CO
en
t— '
-P»
>-j
o
ro
CTi
.£»
O
ro
~-j
<
en
o
ro
Cn
en
en
en
en
o
o
ro
CO
ro
ro
-P»
o
4=«
-P»
I—"
O
•>J
00
V£>
^
CO
-&
00
1— >
co
co
en
ro
CO
CO
0
ro
en
^
*»
en
ro
en
en
•^i
00
en
ro
en
ro
^i
ID
ro
en
i— >
en
t— «
en
-P>
l£>
t— '
O
00
en
t— *
i—"
i— '
ro
CD
ro
0
t— >
0
ro
i— >
^
CO
CO
ro
en
2
CO
en
h->
ro
en
•£>
i-j
^j
-£>
ro
i-*
o>
•^j
VO
CO
co
CO
en
CO
i— *
O
vo
o
i— >
I
ro
o
•>j
CO
CT>
i— '
en
CO
en
••j
CT>
CO
t-»
en
o
i— >
en
en
ro
*e
en
CO
<£>
f*
0
CO
en
en
-e»
00
o
00
en
en
vo
i— •
i— *
en
-t>
en
VO
o
^j
CO
ro
-d
cf>
ro
CO
H-»
i— '
vo
i— '
•*j
£
CO
00
•^1
CO
4*
en
CD
CO
ro
CD
CO
^
l~>
en
«k
cr»
VO
i— •
en
CO
ro
vo
•»j
o
ro
vo
CD
CO
•— «
o^
£
vo
CO
•~J
*>
co
-P»
o
^J
en
o
en
en
i— «
-e»
-P>
en
CO
CO
CO
CO
ro
•c»
-»j
^
CO
t-»
I—1
en
CO
O
vo
•vj
ji
CO
VO
0
03 C
o :
o e
o ec
o *•
o -
<£)
•^ r
ro o
1 C
CL -
,EU =
*< 1C
3- -
3 1
>—>
O
3
5 3
'?
fD
! S
•Sfc
_f
O
".g
» Cu
i. 3
J* ~J»
5 O
5 *
Retentior
T-I mrt
6*1
§
5
sf
sfr
CO
o
n>
cr
ro
3 ft) -h
vo 3 -h
to o co
< o
O) fit r+
n> CL 73
rt- CD CO
ro -s o
rf < CO
_>. oj c+
O -5 SU
3 — ••<£)
o ro
ro
o
•s co
eo c
o> cr
3 CO
o'-s
D>
P" r+
o ro
Q.
974
-------
t— " 1— "
VO Cn
O VO
O O
co (o
en I—.
O O
o 5
•vl Cn
ro -p»
ro ro
i—1 •-•
CO 1—
CO >4
0 00
Cn Cn
CO CO
!-• 10
cn cr>
•**. ^»
ro co
en 5.
*-• fo
*» co
cn co
CD CO
co ro
>-> ro
CJ !-•
co en
Cn «vl
ro *>
^o *»
£ 0
•vj CO
>>l CD
Cn VG
CO OC
cn a*
*» vo
\O 4k
(-•
O
vj
0
1— >
00
8
-C*
o
««J
•>4
H- »
oo
Cn
ro
t— >
cn
§
O\
cn
0
o
00
•»J
CO
ro
*»
i
•-•
^1 4k
§ s
co co
CO I—*
Co ro
s s
o o>
Cft Cfi
ro tn
ro en
i—i t— '
ro tn
en ro
«-• >4
>~j >j
co ro
ro ro
ro S
O 00
cr> en
O o
0 0
^ CO
0) 4k
^J t—
V£) \o
co co
CO O
§|£
^^
00 CO
cn
en
o
I
ro
O
vo
(«)
>j
ro
o>
en
en
en
cn
t-f
O
VJ
t— '
-t»
i
00
en
C3
o
ro
CO
co
VD
co
en
i
CO
ro ro
en ro
0 0
0 0
-p» *»
en i-«
•C» co
0 0
t—i i—*
*» ro
ro en
en *.
ro ro
cn co
CD cn
CO •£>
4* *»
>-> t— •
ro en
o vo
»-•
VO H->
^J O^
t— • i— •
en co
en cn
^—'
ro vo
ro O"!
ro ro
-P» cn
«vi vo
ro ro
ro -P»
ro ro
-P* 4»
-fa. CTl
O CO
CO CO
co vo
co cn
i— •
vo
co
O
CO
en
cn
0
i— '
i— '
0
5
ro
o
ro
vo
-P»
i— •
CO
•^J
I—"
-ti
1— •
t— '
en
co
i— '
en
•^j
co
o
o
ro
-P»
O
en
t— '
CO
5
vo
t— 1
-e>
to
en
ro
•vi
CO
o
CD
en
ro
i-»
en
CO
cn
-c»
ro
VO
^
ro
CT»
co
ro
en
vj
ro
ro
-p»
ro
vo
en
"v.
ro
^»
en
Oi
co
ro
en
&>
00
i-» i— '
ro i-«
^j cn
0 0
ro ro
CO O
•^J CO
o o
-•j en
ro cn
•o i—
i— ' i—"
CO t->
cn co
t-> cn
-p> «>
ro co
vo ro
H- 00
ro ro
ro >-j
CO VO
ro ro
CO CO
co ro
-v» -v.
ro ro
en ro
-J ^j
en en
ro i-»
VO CO
CO CO
en i->
CO O
*vi ^j
*» cn
£ ^
cn cn
ro >-•
vo cn
cn
en
o
t— •
t— •
VO
o
CO
»>J
I— »
cn
«>i
CO
-e>
*>
«vl
*>l
CO
-p»
en
ro
CO
en
ro
no
ro
cn
t— '
tn
co
•vi
tn
CO
^
cn
^4
o
cn
ro
cn
en
-P»
CO
O
i— '
en
i— •
ro
V4
•P*
-e»
-P»
tn
co
co
cn
ro
*>j
cn
co
i— •
o
ro
CO
cn
cn
ro
tn
vo
ro
en
2
.p.
CO O
O S
0 0
•
O 3
O VQ
0 *•>
VQ
^^ r~
3 0
ro a>
I O
Q. -•
Q> 3
<< VQ
3- 3
CO CD
I— •
(/)
rt-
ro 3
3 cr
Q. n>
-«j
o
-t>
co
ft-
o>
IO.
eo n>
2.
o
CD
2
> i— i
3
-h
_J
fD
r+
j
o
" vQ
DJ
3
. -"•
• n
)f
Retention
a*
;o
ro
o
<
ta
*
ro
o
3
(0
fD
3
CO
O
fD
cr
n
o
o
fD
Q.
975
-------
co ro
vo vo
co ro ro ro
8
_ vo o
oo ro ro
oooooocncnoo
•"«J en co co
ro vo
o o
en *fc
8
co 10
en vo
oo en
s
00 I—'
8
CO
•
00
en co co »-»
en o O
co O en
4a» ro vo
co cr>
CO CO CO I-"
o oo
en vo
ro •—« »-« t-f
en 'io oo .r*
vo »-• ci «sj
OD 40 CO O
rocouiuivovovovovovovot—>t—>roro
>
en co co o
•P* i-« en ro
.
cnvororocooooooooiNJOo-tj
cncncnroroco
rorocoro.c»i\jroro
rvs
ro
CO CO I—> !-<
CD o oo -vj oo i-»
O o o .j-* o en
«-• en INJ ro co
^J 7^ en o en ro
^ vo 4* lo co o
OooooovovDvovDvovDvovocncno^*^!
inpovoo>Nivoi->co
ovovot-'rococovoocorovDvocooo
"J -vj Co
Oo vo en
co en en
oo vo vo vo vo vo vo
co ro co
co co o en
ro ro
2 es g 2
•*» vo
Cn
CO
en Co
ro
00—-
o 3
O U3 Ct
vo
(D
300«
ro n» 3
i a.--.
-•• o
n
ro
co
n>
oo
Pi-
o>
CQ
ro
976
-------
to an offensive Hj>S gas and the other is the overgrowth of
filamentous microorganisms, that reduces the sludge settle-
ability. Unfortunately, no identification of thread-like
organisms was obtained from the present study.
(4) COD and BOD Removal As a Function of Substrate and
Loading Conditions. The effect of influent substrate con-
centration and organic loading on R3C plant performance was
investigated at the disc rotational speed equal to 15 and 20
rpm, respectively, for the conventional air and pure oxygen
system. The results of the study were summarized in Table
7.
It is apparent from Table 7 that the % COD or BOD removal
is highly dependent upon the system operating conditions.
In general, the lower the influent COD or BOD concentration
and organic loading the higher the % substrate removal. The
permissible organic loading and retention time for 90% BOD and
80% COD removal are estimated for both RBC systems as follows-
System
Oxygen
Air
% Removal
BOD
90
90
90
90
COD
80
80
80
80
Retention Organic Loading
Time, hrs g./m2-day
4
6
12
6
BOD
19.0
31.0
38.0
28.0
COD
49.0
70.0
110.0
62.3
To meet the above requirements, the oxygen RBC system
was perferably operated under a retention time between 6
and 12 hours. At the retention time of 6 hours, the oxygen
RBC system can be loaded at approximately 9.6% on COD basis
and 11% on BOD basis higher than the conventional air system.
Additionally, it was also found that in the oxygen system, the
organic loading rate at 6 hours retention time was 30% and
38% higher than at 4 hours, respectively, for 80% COD and 90%
BOD removals.
The net % reduction of COD and BOD from each stage of
the RBC systems were also reported in table 7. It was •
observed that when the organic laoding was low (i e BOD
load < 19.6 g./m2-day and COD load<43.9 g./m2_day), effective
977
-------
removal of BOD and COD was obtianed in the first stage of RBC.
The subsequent stages become important as a result of increasing
the organic loading or shortening the hydraulic retention
time. The overall treatment efficiency as a function of BOD
or COD loading is illustrated in Figures 2 and 3.
The relationship between organic loading and substrate
removal per unit surface area per time is shown in Figures 4
and 5. A linear relationship was found for the conventional
air system when the organic loading was below 60 g. BOD or
COD/ m2-day, and for the pure oxygen system as the loading was
not in excess of 70 g. BOD or COD/m2-day. The equations used
to describe the above mentioned relationship are given as
Air RBC System:
$ = 0.8 + 0.82 to, if to < 60 g/m -day
Oxygen RBC System:
$ = 2.8 + 0.79 03 if co < 70 g/m -day
•(5)
•(6)
in which
$ = Organic Removal in g COD or BOD removed/
m2-day
2
to = Organic Loading in o COD or BOD/m -day
Figures 4 and 5 also show that the predicted organic
removal efficiency $ calculated from Eqs. 5 and 6 exceeds the
observed values when the organic laoding w was greater than
the limits already mentioned.
(5) Foamino and Brdiging. Throughout the entire stuay,
foaming never occurred in the pure oxygen RBC system but it
was a serious problem in the conventional air RBC system
(see Figures 6 and 7). It was found that the first stage
of the conventional air system was covered by foam at the
organic loading equal to 38 gCOD/m2-day. When the loading
was increased to exceed 50.2 g. COD/mz-day, the operation
of both first and second stages of the RBC system was inter-
ferenced. The low efficiency of the conventional air RBC
system could be partially caused due to the foaming that
resulted in a low oxygen transfer to attached biomass.
The maximum thickness of the biofilm was approximately
equal to 2.5 cm on the first-stage biodiscs of the pure oxygen
978
-------
5
s
o
I
g
aoo
979
-------
§
s
ooa
980
-------
120
100
O BOD data
9 COO data
RS= 20 rpm
«r
60
40
20
20
40
60
80
100
120
140
160
a, g COD or BOD/m -day
Figure 4. Realtionship Between Organic Loading
and Substrate Removal Efficiency In
Air RBC System
981
-------
120
100
80
60
40
20
O BCD data
9 CCD data
RS= 15 rpm
60
Figure 5.
80
100
120
140
160
u, g COD or BOD/m -day
Relationship Between Organic Loading
and Substrate Removal Efficiency In
Oxygen RBC System
982
-------
Figure 6. Foaming Problem in Air RBC System
Figure 7. Disappearance of Foaming Problem In
Pure Oxygen RBC System
983
-------
RBC system. No biomas,s bridging was found. This result means
that 4" disc spacing presently provided for the first RBC stage
can be reduced to increase the disc number as well as the disc
surface area. By doing s-o, the organic loading can also be
reduced to increase the treatment efficiency.
Determination of 3 and K . Eq. 4 was used to calculate
the values of 3 and K by plotting (1/x) against (1/S*) (see
Ffgures 7 and 8). The 3 and K values obtained from the above
calculations were summarized in Table 8 as follows:
Table 8. 3 and K Values
Parameter
Ks, mg/1
3, mg/l-day
Air RBC System
BOD Basis
151.4
58.8
COD Basis
890.4
153.8
Oxygen RBC System
BOD Basis
10.9
31.7
COD Basis
432.0
117.6
By substituting 3 and K values into Eq. 4, four RBC kinetic
models are formed
Air RBC System:
— = 2.574 (1/S ) + 0.017 on BOD basis
Q(so-s )
•(7)
Aw *
= 5.783 (1/S ) + 0.0065 on COD basis
Q(so-s )
•(8)
984
-------
1,600
1,400 -
1,200 -
S 1,000 -
aoo -
• Pure Oxygen RBC System
O Air RBC System
600 -
4000
1/S , 1/mg x 10
Figure 7. Determination of B and K On
BOD Data Basis s
985
-------
700.
600
500
I" 400
1
« 300
••*.
200
100
.0065
= 0.0085
J I
O Air RBC System
• Pure Oxygen RBC System
I
I
20
40
60
80
100
120
1/S 1/mg x 10'
,-4
Figure 8. Determination of e and K on
COD Data Basis
140
160
180
986
-------
Oxygen RBC System:
V * • •
W = 0.344 (1/S ) + 0.0315 on BOD basis
Q(so- s
•(9)
A. ••*"'••
= 3.673 (1/S ) + 0.0085 on COD basis
Q(so- s
•(10)
Conclusions
The following conclusions were formulated as a result of
this study: .
1. Proper pH control is essentially required for the
treatment of this acidic wastewater. By adjusting the waste-
water pH to nearly 9.0, the final effluent pH after passing
through three RBC stages would be within the range of 7.0-
7.6. The pH control was made by the addition of NaOH along
with phosphate buffer.
2. In the pure oxygen RBC system, the disc rotational
speed did not significantly affect the results of % COD or
BOD removal, oxygen consumption, pH and DO level in each
RBC stage. At the constant rotational speed of 15 rpm (12.3
m/min.), the RBC plant performance was closely related to
the influent substrate concentration and organic loading as
well,
3. Both air and pure oxygen RBC systems are capable of
removing COD and BOD sufficiently. The COD and BOD removals
were more than 90% when the organic loading in the conventional
air system was below 28 g. BOD/m2-day and 62.3 g. COD/m2-day
or it was less than 38 g. BOD/mz-day and 110 g. COD/m2-day in
the pure oxygen system. From the present study, it was found
that the latter system was possible to be operated at the
organic loading 1.7 times higher than the former system if
the efficiency of the RBC plant in terms of BOD removal equaled
80%. In general, the reduction of organic substrate was
high in the first RBC stage when the organic loading was below
19.4 g. BOD/m -day in both systems. However, the subsequent
987
-------
stages become important after the organic loading condition :
exceeding the above value.
4. The values of g and K in Eq. 4 are obatinable. It
is apparent from the present stady that the pure oxygen RBC
system has lower e and K as compared to the conventional air
RBC system on both BOD and COD data basis.
5. Two serious problems which may occur during the operation;
of RBC systems are: (a) foaming and (b)septic environment. The
foaming problem started at the organic loading = 38 g. COD/m2-day
and the anaerobic suitation was developed at the loading equal or
greater than 57 g. COD/m2-day in the conventional air RBC system.
Although no foaming problem was found in the pure oxygen RBC
system, the septic condition occurred in the first stage at the
organic loading = 95 g. COD/m2-day. Poor sludge settling charact-
eristics was resulted from the overgrowth of thread-like organisms
under anaerobic condition.
References
1. Birks, C. W., and Hynek, R. J., " Treatment of Cheese Processing
Wastes- by the Bio-disc Process", in Proceedings of 26th Purdue
Industrial Waste Conference, Purdue University, pp. 89, 1971.
2. MaAliley, J. E., "A Pilot Plant Study of a Rotating Biological
Contactor for Secondary Treatment of Unbleached Kraft Mill
Waste", Tappi, vol. 57, No. 9, 1974.
3. Labella, S. A., Thaker, I. H., and Tahan, J. H.."Treatment of
of Winey Wastes by Aerated Lagoon, Activated Sludge, and Rotating
Biological Contactors", In Proceedings of 27th Purdue Industrial
Waste Conference, Purdue University, pp. 803, 1972.
4. Watt, J. C., and Cahill, C. J., "Wasteater Treatability Studies
for a Grassroots Chemical Complex Using Bench Scale Rotating
Biological Contactors", In Proceedings of the First National
Symposium/Workshop on Rotating Biological Contactor Technology,
Vol. 1, pp. 661, 1980.
5. Chesler, 6. P., and Eskeland, G. R., "RBC for Munitions Waste-
water Treatment", in Proceedings of the First National Symposium/
Workshop on Rotating Biological Contactor Technology, Vol. 1,
pp. 711, 1980.
988
-------
6. Tanacredt, J. T,, "Removal of Waste Petroleujn Derived
Polynuclear Arjnattc Hydrocarbons by Rotating Biological
Contactors11, In Proceedings of the first National Sympo-
sium/Workshop on Rotating Biological Contactor Technology,
Vol. 1, pp.725, 1980. . ™*
7. Bracewell, L. W., and Jenkins, D., "Treatment of Phenol
-Formaldehyde Resin Wastewater Using Rotating Biological
Contactors1', in Proceedings of the First National Symposium/
Workshop on Rotating Biological Contactor Technology, Vol.
1, PP. 733, 1980.
8. Bintaja, H. H., et a!., "The Use of Oxygen In A Rotating
Disc Process", Water Research, Vol. 10, pp. 561, 1976.
9. Huang, J. C., "Operational Experience of Oxygen-Enriched
Rotating Biological Contactors", in Proceedings of the
First National Symposium/Workshop on Rotating Biological
Contactors, Vol. 1, pp. 637, 1980.
10. Monod, J., "Recherches sur la croissance des cutlures
bacterbacteriennes", Hermann & Cie. Paris, pp.211, 1942.
11. Clark, J. H., et a!., "Performance of a Rotating Biological
Contactor Under Varying Wastewater Flow", Jour. Water Poll.
Control Federation, Vol. 50, pp.891, 1978.
12. Standard Methods for the; Examination of Water and Waste-
water, 14th Edition, APHA-AWWA-WPCF, 1975.
989
-------
INHIBITION OF NITRIFICATION BY CHROMIUM
IN A BIODISC SYSTM
Shin Joh Rang, McNamee, Porter and Seeley
Ann Arbor, Michigan 48104
Jack A. BQjjehardt, Department of Civil Engineering
The University of Michigan, Ann Arbor, Michigan 48109
INTRODUCTION
A series of investigations was undertaken to determine
the acceptability of certain .industrial wastes containing
hexavalent chromium by a biodisc system providing both secon-
dary treatment and biological nitrification.
The initial objectives of the study were: to determine
the extent to which a biodisc system can tolerate chromium
(VI) without losing efficiency in either BOD removal or in
nitrification; to establish mechanisms of chromium removal
and the benefits of staging, and to understand differences
in short and long-term effects and steady state or shock load
conditions.
These objectives were established to serve a/number of
purposes. The data gathered will assist municipalities in
determining the quantity and characteristics of chromium
containing waste that may be accepted without causing inter-
ference in operation. The data gathered will also assist
design engineers in understanding and better defining the
process reliability between an activated sludge and a bio-
disc system and finally the benefits of staging in an RBC-
990
-------
Numerous studies on the effect of metals on biological
nitrification have been reported in the literature in recent
years, all of which have been confined to activated sludge(l)
(2)(3). For this reason, this study on the effects of hexa-
valent chromium on the biodisc system was undertaken. In
order to minimize differences due to other effects, a pilot
plant was used.
EXPERIMENTAL METHODS AND APPARATUS
The shape and dimensions of the pilot system used are
shown in Figure 1. Each biodisc tank was equipped with two
sets (4 each) 2 ft. dia. disc media, a partition that
separated the two adjoining stages, a drive with chain and
speed controller. Two parallel systems were built for this
experiment, each containing three, two stage units. Light
weight concrete fillets, 3" x 3", coated with paraffin were
placed at the bottom corners to prevent sludge accumulation
and to improve the mixing pattern in the tank.
The standard substrate or feed solution was prepared
such that the characteristics would closely simulate those
of a typical municipal wastewater; 6005 at 200 mg/1, COD at
300 mg/1, total nitrogen at 20 mg/1. Dextrine was selected
to be the major carbon source due: to its slow biodegradation
rate.
A stock solution of hexavalent chromium was prepared
from K/jiC^Oy and fed at a predetermined concentration either
at the first stage or the fifth stage depending upon the
purpose of the particular experiment. Typically, the slug
loads of chromium were fed at the fifth stage to test short-
term effect on nitrifying cultures. Long-term effects on
the other hand, were studied by introducing chromium at the
first stage.
A stock solution containing glucose was prepared and
fed to test the effect of high organic loads on nitrifiers.
Table 1. Standard Feed Solution
Dextrin
Urea
Na2HP04
CaCl2
KC1
NaCl
MgS04
150 mg/1
42 mg/1
15.9 mg/1
5.6 mg/1
5.6 mg/1
12.1 mg/1
4.0 mg/1
Ivory Soap 6.3 mg/1
Consume Soup 2.1 ml/1
Ann Arbor tap water
991
-------
LU
CO
o
CO
Q_
LU
CC
Z)
o
992
-------
RESULTS AND DISCUSSIONS
A. Long Term Effect of Chromium
Three basis problems were present in this investigation:
(1) the concentration at which the effects of a given metal
ion are felt when fed continuously; (.2) the concentration
necessary to have a definite effect on a plant and the time
required; (3) the fate of chromium in the system.
In all cases the two parallel biodisc systems were oper-
ating satisfactorily in both COD removal and nitrification
before chromium was introduced. From earlier experiments
it had been determined that the hydraulic loading rate of
0.65 gpd/S.F. was appropriate for full nitrification(4).
The effects of 3 concentrations (1, 3, and 10 mg/1) of
chromium fed at the first stage were studied in a preliminary
manner at a hydraulic loading rate of 0.65 gpd/S.F. The
chromium was fed at 1.0 mg/1 approximately for two weeks,
while at other concentrations less than a week.
The efficiency of nitrification was represented by its
end product, nitrate, N03, since the intermediate product
nitrite, NO-, was negligible throughout the experiment. The
nitrification at 1 mg/1 of chromium in Run 2 was slightly
hampered at the second, third and fourth stages.and yet
nearly completed at the fifth stage. The sixth 'stage picked
up the difference and completed nitrification, see Figure 2.
Effects of chromium at 3 and 10 mg/1 on nitrification
were immediate and definitely inhibitory. The nitrate con-
centrations were reduced by 65 and 75 percent at each dosage
in Runs 3 and 4, respectively.
On the basis of these preliminary findings, a long-
term investigation of the chromium effect on nitrification
began. The chromium concentrations chosen were 1 mg/1 in
Run 5 and 2 mg/1 in Run 6. In addition, the hydraulic
loading rate was doubled to 1.3 gpd/S.F. to expand the
breadth of the investigation.
As shown in Figure 3, for a system receiving 1 mg/1
of chromium, the concentration of ammonia N continued to
change for approximately a month before a relatively stable
performance level was reached for the system.
Concentrations of other parameters such as COD, N02-N,
and N03-N at this steady state are shown in Figure 4. The
following observations were made:
993
-------
z
<
u
g
14 -t
12 -
10 -
c 8 -
111
6 -
4 -
2 -
LOADING RATE = 0.65 GPD/S.F.
DETENTION TIME = 70 MIN./STAGE
DISC DIAMETER = 2 FT.
ROTATING SPEED' - 10 RPM
'FEED1
1
FIGURE 2. EFFECT OF CONTINUOUS CHROMIUM
FEED ON NITRIFICATION
994
-------
16
14
12
<
_!
(3
S
<
ir
z
o
o
RUN 5
Chromium = 1 mg/l .
Load Rate = 1.3 gpd/ft2 »
Dia.of Disc =2-ft.
N03-N
at 34 Day
NO2-N .
— "at 34 Qay
34 DAY
19 DAY
\ 10 DAY
5 DAY
S-sa NOCr
FEED 1 2 34 5
STAGES
FIGURE 3. LONG TERM EFFECT OF CHROMIUM
ON NITRIFICATION AT 1 MG/L
995
-------
18
CHROMIUM = 2mg/l
LOAD RATE = 1.3gpd/ft2
DIA. OF DISC = 2-ft.
01—
FEED
12345
STAGES
FIGURE 4. LONG TERM EFFECT OF CHROMIUM ON
NITRIFICATION AT 2 MG/L
3 DAY
NOCr
996
-------
- Most of the COD was removed in the first two stages.
- Ammonification of organic nitrogen became complete
at the third stage, compared to the first stage in
Run 1.
- Nitrification of ammonia began at the third stage and
continued in the subsequent stages
- Ho accumulation of nitrite was observed, indicating
that the Nitrobacter group of organisms were not
inhibited. Only Nitrosomonas was inhibited by
chromium.
- Chromium concentrations in the liquid decreased
rapidly in the first two stages (1.1 to 0.6 mg/1)
and slowly in the subsequent stages (to 0.4 mg/1).
- The major mechanism of chromium removal from the
liquid appeared to be adsorption to the biomass. The
chromium content in the biomass layer closely approxi-
mated the COD profile throughout the treatment system.
In the first stage for example, the chromium consti-
tuted approximately 2 per cent of biomass on a dry
weight basis. In the following stages, the chromium
content ranged between 0.6 and 1.0 percent.
- Staging definitely worked in favor of organisms in
the later stages. While organisms initial stages were
exposed to chromium, those in the later stages were
not.
For the parallel system receiving 2 mg/1 of chromium,
Run 6, similar observations were made, see Figures 4 and 5.
The time required to reach a steady state operating condition
however, appeared shortened; 21 days. The following observa-
tions were made: "
- COD was being removed substantially in the initial
two stages but slowly,in the middle two stages.Overall
removal was satisfactory.
- Ammonification of organic nitrogen also took three
stages to complete.
- Nitrification began at the third stage and continued
in the subsequent stages. . • •• :
- Chromium concentration in the liquid decreased rapidly
, from 1.9 to 1.2 mg/1 in the first stage and remained
between 1.2 and 1.1 mg/1 in the subsequent stages.
997
-------
1"
o
998
-------
10 12 14 16 18
NUMBER OF DAYS ELAPSED
20 22
24
26 28
FIGURES. SHOCK RESPONSES ON NITRIFYING CULTURES
999
-------
- The major mechanism of chromium removal is by
adsorption to the biomass. The chromium content in
the biomass was approximately 2 per cent in the first
stage and 0.8 per cent on a dry weight basis .in the
last stage.
Since the organic strength of the feed solution was
the same between these two systems, the resultant
biomass characteristics and quantity should be similar.
It was then not surprising to observe that the
adsorption of chromium by the biomass at steady state
exhibited a similar mass relationship. Due to a higher
concentration gradient in the more highly loaded system
however, steady state appeared to have been reached
sooner in the latter system (Run 6) than in the former
where 1 mg/1 of chromium (Run 5) was used.
- Among the organisms involved, it may be concluded
again that the more sensitive of the two groups of
nitrifiers is Nitrosomonas. Nitrobacter appeared to
be less sensitive to chromium and thus to oxidize
nitrite without much accumulation of nitrite.
- Benefits of staging were again observed in that inhi-
bition took effect by stage. When a load came, the
late stages would not be affected, even if the initial
stages were adversely affected.
From the data presented in the section, one could
summarize that:
Hexavalent chromium could be adsorbed to the biomass
upon contact up to its adsorptive capacity (2 per cent for
heterotrophic organisms and less for a mixture of heterotro-
phic and autotrophic organisms).. This is a higher level than
0.8 percent as reported on activated sludge(3). Since the
biomass density of biofilm is higher than activated sludge
floe, it appears that the adsorptive capacity improves in a
biodisc system.
Inhibition was controlled more by what was in the bio-
mass layer than in the bulk liquid. Even though the bulk
liquid carried chromium at or above 0.6 mg/1, the effect was
not immediate.
Staging helped the process as a. whole. Benefits could
be achieved in two ways: While an upstream stage received
and removed chromium, the downstream stages were spared from
the inhibitory impact or received a minimal quantity, all
of which increased the overall process reliability.
1000
-------
Since the chromium concentration in the initial stages
would be higher than the case where staging was not practiced,
the rate of adsorption to the biofilm would be faster. Under
a slug load, the chromium would concentrate in one stage
and upon sloughing, could be removed from the system in an
efficient manner.
Nitrosomonas showed a greater degree of inhibition by
chromium than Nitrobacter, which was consistent with observa-
tions elsewhere(2).
B. Short Term Effect of Chromium . , •
The problems generated in this phase of the study were
to investigate: (1) that concentration which would produce
definite inhibition -and the resulting reaction time.
(2) the rate of recovery, if' any, once the slug is passed.
(3) acclimation effect under repeated slug doses.
•(4) the impacts of heterotrophic organisms on nitrifiers.
In this study concerned with the effect of slug doses
of hexavalent chromium on the nitrification process, concen-
trations of chromium at 5 mg/1 for 2 days (Run 7), 10 mg/1
for: 2 hours (Run 8) and 50 mg/1 for 2 hours (Run 9) were
used in succession. Approximately a week of recovery was
allowed between runs. These concentrations were fed .at the
fifth stage since at this point the system was supporting
full nitrification. The efficiency as measured by the reac-
tion product, or nitrate, NOo, dropped immediately as shown
in Figure 6. As soon as the chromium feed was removed from
the system however, the nitrification was shown to resume
.and within approximately a week had returned to its original
level. The degree of inhibition appeared to be inversely
proportional to the mass loading. It was postulated that
this rapid resumption of nitrification was made possible by
three contributing factors: (1) the adsorption of chromium
to the biomass layer which appeared to have caused an
immediate inhibition of the existing culture and yet remained
within the biomass layer, where its presence caused no fur-
ther inhibition of new growth. The de.sorption of chromium
did not appear to be significant; (2) continuous seeding
from the fourth stage which contained nitrifying cultures;
(3) acclimation of cultures to chromium was a definite con-
dition which minimized the degree of inhibition in Run.8
(with 50 mg/1 chromium). In this case inhibition was not
as severe as in Run 7 where 10 mg/1 chromium was fed.
1001
-------
C. Impact of Presence of Heterotrophs on Nitrification
In an effort to investigate nitrifying organisms compe-
tetiveness in the presence of heterotrophic organisms,
50 rag/1 of glucose was fed to the fifth stage for three days
in a similar manner as chromium had been previously intro-
duced .
Results are shown in Figure 7 that indicate with the
growth of heterotrophs on the surface, nitrification
efficiency starts to decline. Once the glucose feed was
removed, there is no immediate increase in nitrification
activity.
It appears that nitrifying cultures are outgrown by
heterotrophs on the same surface, thereby they become buried
among the heterotrophs. With the increased oxygen demand
exerted by heterotrophs, the dissolved oxygen concentration
in the bulk liquid decreased to 1.0 mg/1 and as a result
was limited for nitrifiers. This data confirms a theory that
under a d.o. suppressed environment, nitrifiers cannot func-
tion properly and therefore a nitrogenous oxygen demand can-
not exist.
SUMMARY AND CONCLUSIONS
Chromium may enter municipal waste treatment plants
in many different ways. Perhaps most frequently it occurs
in plating waste, although it may have its source in tanning
operations, in waters given corrosion inhibition treatment
with chromate, or in aluminum-anodizing wastes.
Concentrations of hexavalent chromium up to 10 mg/1
were fed up to four months on a continuous basis to two six-
stage biodisc systems treating synthetic sewage, with con-
centrations of 300 mg/1 of COD and 20 mg/1 of nitrogen.
Initial stages of the discs were supporting heterotrophic
organisms, while the biological nitrification was achieved
by autotrophs in the later stages.
Concentrations of hexavalent chromium up to 50 mg/1
were also fed as slugs to the biodisc system to test the
effects of such doses.
In addition, the effects of the presence of heterotro-
phic organisms on nitrifiers was examined by feeding glucose
directly into a nitrifying stage.
1002
-------
36 _
01 2 3 4 5 6
NO. OF DAYS AFTER GLUCOSE ADDITION
I
FIGURE?. EFFECT OF GLUCOSE ADDITION ON NITRIFICATION
1003
-------
From the observations made in this investigation, the
following conclusions may be drawn.
On a continuous basis, 1 mg/1 of hexavalent chromium
exhibited a consistently inhibitory effect on
nitrifying cultures.
The major mechanism of hexavalent chromium inhi-
bition to microorganisms appeared to be by adsorp-
tion. When chromium is adsorbed to a discrete
active microorganism, the inhibitory effect sets
in quickly. Chromium also may be adsorbed to non-
active biological film or floes, thereby becoming
immobilized. The degree of adsorption or affinity
however, depends on the characteristics of the film
or floes in the system. For example, heterotrophic
cultures established in the initial stages, oxidizing
primarily carbonaceous materials exhibited chromium
retention of approximately 2 per cent, while the
mixture of nitrifying cultures containing a limited
quantity of heterotrophic organisms established in
the later stages, exhibited a concentration of approxi-
mately 0.8 per cent on a dry weight basis.
As previously indicated data gathered in this inves-
tigation showed that the rates of adsorption and
the resulting inhibition to microorganism were
fast.
In a matter of hours the inhibition at a stage was
well defined. When the adsorptive capacity of the
entire biomass film in a stage was reached, the
impact was shown by reduced COD removal for heterotrophs
and reduced nitrate generation by nitrifiers.
Data also indicated that between two major groups
of nitrifiers, Nitrosomonas was more sensitive to
chromium and thus ammonia oxidation was decreased,
while the nitrite oxidation to nitrate remained
unchanged.
Data further indicated that staging offered advan-
tages in process reliability and also in isolation
and efficient removal of affected sludge from the
system.
1004
-------
The impact of slug loads of chromium on nitrifica-
tion was tested at concentrations up to 50 mg/1. *
Since the biofilm at the fifth stage composed
mostly of nitrifying cultures, the chromium addi-
tion at that stage resulted in immediate reduction
in nitrification, even at 5 mg/1. The complete
recovery of the system however was rapid, taking place
within approximately a week, indicating that the
chromium retained in the film and floes did not
adversely affect the newly developing cells being
established. Desorption of these chromium compounds
appeared not to have generated a problem. The
RBC system also showed a benefit due to continuous
seeding from the upstream stage. '
An advantage due to previous exposure to chromium
was also shown. Acclimation was evidenced from the
studies of slug doses.
The fact was likewise confirmed that when hetero-
trophs start to grow on the same support surface,
nitrifiers cannot compete for either surface or
oxygen and thus become overgrown.
1005
-------
REFERENCES
Tornlinson, T., Boon, A., and Trotman, C. , "Inhibition
of Nitrification in the Activated Sludge Process of
Sewage Disposal," J. Applied Bacteriology, 29, (2),
1966, pp. 266-291.
Painter, H., "Review of Literature on Inorganic Nitrogen
Metabolism in Microorganisms," Water Research, Vol. 4,
1970, PP. 393-450.
Anon, Interaction of Heavy Metals and Biological Sewage
Treatment Processes, U.S. Public Health Service Publi-
cation No. 999-wp-22, Environmental Health series,
May, 1965.
Borchardt, J. , Kang, S., and Chung T., Nitrification
of Secondary Municipal Waste Effluents by Rotating
Bio-discs, U.S. EPA 600/2-78-061, June, 1978.
ACKNOWLEDGEMENTS
The support of Capital Consultants of Lansing, Michigan
in this research is acknowledged.
This research was conducted at the Sanitary Engineering
Laboratory, the University of Michigan. At the time,
S. J. Kang was a graduate student in the Department of Civil
Engineering.
100F
-------
PART IX: INDUSTRIAL WASTEWATER TREATMENT
SCALE-UP AND PROCESS ANALYSIS TECHNIQUES FOR
PLASTIC MEDIA TRICKLING FILTRATION
Thomas P. Quirk, P.E. Department of Civil and
Environmental Engineering, Vanderbilt University,
Nashville, Tennessee.
W. Wesley Eckenfelder, Jr. Department of Civil
and Environmental Engineering, Vanderbilt Univer-
sity, Nashville, Tennessee.
ABSTRACT
Reaction models for bio-oxidation using a sheet flow re-
actor with a fixed biological film similar to that used in
plastic media trickling filtration are developed. The models
utilize plug flow hydraulics and accept various descriptions
of BOD removal kinetics including: zero order, first order,
retardant and concentration dependent mechanisms. Hydraulics,
kinetics and film geometry are individually incorporated.
System model equations are arranged into linear expres-
sions which allow graphical determination of model applicabil-
ity and rate constants from plots of experimental data.
The design and operating characteristics of laboratory
simulation equipment are presented. Simulation equipment con-
sists of continuously fed inclined planes with effluent sedi-
mentation and recirculation. Effluents can be fed from and
returned to cold room storage. Heating tapes control tempera-
ture above ambient levels. System location in a cold-room
provides temperature control at below ambient levels.
Applicability of the models is verified using operating
data from laboratory and full scale studies of a number of
effluents including: municipal sewage, whey wastewater, kraft
1007
-------
mill effluent, sulfite mill discharges, hard-board mill_ef-
fluent, yeast fermentation effluents, pharmaceutical dischar-
ges and meat processing wastes. Additional verification is
presented from literature data utilizing a glucose substrate.
Scale-up calculations are developed to utilize laboratory
data to determine full scale performance for various packing
geometries. Design calculations are also presented including
determination of the influence of recirculation.
INTRODUCTION
The trickling filter is comprised of a bed of media on
which biological film growths develop. Removal of BOD is ob-
tained by aerobic processes at the film surface and by anaer-
obic processes within the film interior. Modifications to
the classic rock trickling filter introduced plastic geometric
media to obtain increased surface area and porosity. Current
practice generally employs a lattice type structure of verti-
cally oriented media which induces a sheet flow regimen.
Reactor operation is such that laboratory scale simula-
tion can be used for rate .constant determination and in para-
llel operation with pilot plant equipment to reduce data
collection requirements and extend interpretation of pilot
scale results. Simulation equipment utilizes an inclined
plane of up to 18' in length operating with thermal regulators
and evaporator control systems.
The theory of BOD removal by trickling filter slime over
a reaction surface similar to that of an inclined plane has
not been formulated in the existing engineering literature.
Full scale design equations in current use have been fre-
quently developed by emperical methods or by analogy to for-
mulations used to describe BOD exertion in general. The work
presented herein describes development and verification of
reaction models suitable both for laboratory studies and for
scale-up to prototype conditions.
Consideration is given to the analysis of laboratory
scale data, the interpretation of pilot and full scale results
and the design of prototype systems for each reaction model.
The theoretical development is presented in the following se-
quence:
° General reaction model for inclined planes
o' Specific models for alternative bio-kinetic
rate processes
o Influence of recirculation and temperature
1008
-------
0 Scale-up techniques for full scale
conditions
o Design Procedures
General Reaction Model
A general model for BOD removal on an inclined plane
surface is obtained by the solution of a material balance
statement in which the hydraulics of liquid flow and the kin-
etics of biological reaction have been separately included.
A schematic of an inclined plane system is shown in Figure 1.
Figure 1 Schemotlc Representation, BOD Removal
Over Slimed Surface
The material balance statement is written thus:
INPUT - OUPUT - REMOVAL = ACCUMULATION
INPUT and OUTPUT terms are self explanatory. The REMOVAL term
is defined by the geometry of the reactor and the kinetics of
the biological reaction. The ACCUMULATION terms accounts for
mop
-------
the change in the quantity of BOD stored in the reactor vol-
ume. For both inclined plane surfaces and trickling filters
this storage is negligible. The terms of the material balance
statement are mathematically defined as follows:
= (Q+R)Sa
= (Q+R)Se
KrVr
where:
INPUT
OUTPUT
REMOVAL
ACCUMULATION = 0
Q = Untreated flow (gal min"1)
R = Recirculation flow (gal min~ )
Sa= BOD concentration as applied to
reactor (mg/1)
Se= BOD effluent concentration-mg/1
Vr=Volume of reactor (gal)
Kr= A generalized BOD removal rate constant
for the reactor volume (mg/1 min~l)
The reactor is examined over a differential element of
plane height (dH) to define reactor volume:
dVr = (d) (W) (dH)
(1)
The material balance for BOD input and removal is then
stated as follows:
(dS) (Q+R) = (Kr) (d) (W) (dH)
which when rearranged into a differential statement becomes:
d§. _ Kr(d) (W)
dH
(Q +
(2)
The term, (Q + R)/W is conveniently grouped as a hydraulic
loading per unit of plane width (U) as follows:
_dS = (Kr) (d)
dH (TT)
(3)
1010
-------
Equation (3) then is the general statement for BOD removal
over a slimed surface.
First Order React_imi_JCine_ti_cs
At this juncture biological reaction kinetics may be in-
troduced to develop a specific reaction model, e.g., first
order, retardant etc. A first order reaction model has been
found in practice to apply most frequently to trickling fil-
ter performance.
The generalized rate constant (Kr) for a first order re-
action is expressed as the product of a specific velocity
constant and the amount of substrate present. The relation-
ship in generalized terms is:.
Kr = (k) (S)
(4)
The units used for k and S are dependent on the way in which
the reaction is described e.g. in activated sludge work S has
the units of concentration and k is expressed as a substrate
concentration change, per unit of substrate and organisms pre-
sent, per unit of time. Kr then has the net units of time"1.
In fixed film reactors the specific velocity constant is
expressed in terms of the unit weight of substrate removed per
unit weight of film present, per unit of time.
k =
where:
dt Mf
dMs = substrate weight change
dt = time change
Mf = film weight present
By expressing substrate weight in terms of concentration, i.e. dM =V dS, k
then becomes: - • s
VrdS = /dS\ 1
dtMf (dt) /fj
(5)
Mf/Vr can then be termed the equivalent concentration of or-
ganisms in the elemental reactor volume being analyzed. For
the inclined plane:
v
/gf\
\ V
ion
-------
where:
df = depth of fixed film
d = depth of flow q
g = specific gravity of film
These definitions then provide the expression for
the generalized reaction rate as follows:
wherein
Kr = (k') fS)
k' = (k) (df) (gf)
(6)
The effective slime concentration provides for an active
weight of organisms per unit of area and thereby implies that
a thickness of film is operating to effect BOD metabolism.
This is indeed the case and has been experimentally verified
by Hoehn and Roy (1). Figure 2 presents their data for COD
removal vs. film thickness.
o
E C
££
§§
o o
«- en
o E
o
cc
450
300
150
200 400
Mean film thickness (urn)
6OO
Figure 2
Rate of COD Removal as a Function
or Film Thickness (I)
1012
-------
The description of a first order reaction may then be
completed as follows:
dS
dH
k'S
U
(7)
It is worthwhile to note that the derivation of Kr pro-
ceded in such a way as to eliminate flow depth (d) from the
final expression. This in turn results in an exponent of 1.0
for the flow term (U). If depth remained in the final equa-
tion then the fact that depth over an unslimed plane varies
with (U)~l/3 would result in an exponent of 2/3 for the flow
term U. The elimination of depth also tacity implies that
the reaction at the slime/water interface controls rather
than the transport of substrate through the liquid depth to
reach the interface. Equation (7) thus describes a reaction
controlled model. The description of the active organism con-
centration(Mf/Vr)interms of a ratio of slime specific weight
divided by liquid depth also results in an averaging process
for the amount of biomass effectively operating to remove and
metabolize BOD. The definition of the amount of biomass rep-
resente.dthe first half of'the analysis. The balance of the
(Kr) breakdown used kinetic expressions which are independent
of organism concentration i.e. Kr~ k'S. The effect of this
two part definition is to effectively insert a linear aver-
aging technique into the overall rate constant Kr.
Thus, any other BOD removal and metabolism mechanism which
can be reasonably described by a linear average of a combina-
tion of liquid depth and bio-mass amount will be approximated
by equation (7). This may help to explain why the first or-
der formulation has been found to apply to performance data
over a wide range of process conditions. Integration of ex-
pression (7) provides the basic analytical relationship for
BOD removal over an inclined surface:
= e-k'H/U
(8)
In the application of equation (8) to laboratory analysis
an incline of up to 45° from the horizontal may be used with-
out concern for the non-formulated depth effects caused by
angle of inclination.
Incorporation of Recirculation and Temperature
g ,
The ratio -5^ describes the change in BOD as applied to
1013
-------
the film and, therefore, implicitly includes the effects of
recirculation. For design purposes BOD changes relative to
the undiluted wastewater are required. The change in reactor
BOD's is related to the undiluted wastewater by a material
balance around the plane as follows:
*a _
r 1-
= f
where:
(1+r) (1-E)
r = recirculation ratio R/Q
(9)
E = BOD removal,efficiency based on
raw wastewater
The factor (f) is introduced for topographical simplicity.
The effect of temperature on reaction rate is introduced
using the Arrhenius relationship.
where:
k' AT
200
(10)
k' = reaction constant at temperature t
k*2Q = reaction constant at standard .
temperature, 20°C
At = reaction temperature differential
°C-20
6 = constant, usually taken as 1.035
ANALYSIS OF LABORATORY PLANE PERFORMANCE
The analysis equation for first order performance is
completed as follows:
it H/U
f = e
. ,
k
(11)
A graphical solution to equation (11) is obtained by taking
double logarithms as follows:
log
H
2.Slog (f)
log U + log(l/k'7n) (12)
A plot of data on log paper will provide a linear correlation
with slope equal to 1.0 and an intercept of log (l/k^o0) at
1014
-------
U = 1.0.
It is important to note that the variable H, height of
plane, is portrayed as having a linear effect on plane per-
formance. This is so in that laboratory planes are short and
receive a controlled, uniform discharge. Both of these char-
acteristics support a linear relationship between height,
flow pathway, film presence and thickness etc. This condition
does not necessarily apply to full scale towers.
An application of first order kinetics to inclined plane
treatment of whey plus sewage, Quirk (2), is shown on Figure 3.
Test results were obtained using plane lengths of 9 and 18 ft.
operated at hydraulic loadings from 0.02 to 0.15 gpm/f t. Re-
cycle ratios ranged from 0. to 6.0. Film areas of 0.375 and
0.750 SF provided the data shown operating -over a range of
organic loadings of 0.08 to 0.95 Ibs BOD/day/SF Slime. An hy-
draulic coefficient of n =: 1.0 and k2Qo = 1-8 x 10~3 gpm/SF
were determined. Shultze (3) also experimented with a mixture
of whey plus sewage using vertical meshed screens 3' x 6' in
dimension as a fixed film reactor. His data correlated using
a first order reaction as shown in Figure 4. Hydraulic ef-
fects produced an n =: 1.0-with a reaction rate constant of
k
° = 2.5 x 10~3 gpm/SF. Hydraulic loadings ranged from
8 to 0.45 gpm/SF and were about three times the hydraulic
loadings used by Quirk (2) .
"2
0.
Additional Kinetic Models
Kinetic models other than a first order reaction have
been employed to describe biological oxidation. While less
popular than the first order assumption, these additional mo-
dels have been found to correlate bio-oxidation data in a
successful manner. Because of a cumbersome mathematical
structure when used for fixed film reactors,these models have
been applied primarily to fluid bed reactors such as activated
sludge, aerated stabilization basins, etc.
The relationships for laboratory planes have been extend-
ed to include:
1. Linear and exponentially retardant reactions
2. Concentration dependent reaction
3. Zero order or constant rate reaction
1015
-------
" '00
I 20-3I°C
BOD 270-b'O mg/1
= l.S x 10-3 gpm/SF
n * 1.0
Figure 3 wney ond Sewage Performonce on Inclined Plane (2)
S
",
Z ,00
Schultze i3)
Sewage Plus wney
vertical Screen Filter
T - 11 to 220C
n » 1.0 ,
k20 - 2.5 x IO-3gpm/SF
Figure
wney and Sewage Performance on Vertical Screen
1016
-------
Retardant Reaction Models
In a simple retardant reaction the rate of BOD removal
per unit weight of organism is proportional not only to the
BOD concentration remaining but also to the fraction of BOD
remaining. Rates of removal decrease, or retard,rapidly as
high efficiencies of removal are approached and the relatively
assimilable organics havebeen metabolized. The kinetic state-
ment is written as :
Kr = (k)(Mf/Vr)(S) (S/Sa) (13)
After integration the completed equation for slimed plane
analysis is written as:
(k)' • (H)
f = 1 +
U
(14)
An exponantially retardant reaction can be described by the
relationship: r i -il/N
(f) = 1 + ~it (15)
Concentration Dependent Model
A concentration dependent reaction assumes a linear reac-
tion of all substrate elements with the percent remaining as
the expression of retardancy. Unlike the retardant reaction,
the concentration dependent formulation does not relate remo-
val rate to concentration remaining but rather to a maximum
rate which would occur at zero removal. The equation for a
slimed plane operating with concentration dependency becomes:
k'H/USa
f = e
(16)
The combined parameters of H, U, and Sa represent the or-
ganic loading or F/M ratio per unit area of film. Thus:
YF/M
(17)
a linear plot is obtained by relating log(f) to k/(F/M).
Inclined plane studies using glucose were found to be
correlated by a concentration dependent reaction. The Maier(A)
data is correlated by a concentration dependent model as shown
on Figure (5). The separation of the 37.3°C data is apparent
while the balance of the results fit a single model. The re-
action rate for these glucose experiments was found from the
1017
-------
correlation slope at k*2Q = °-5'0 x 10~3 lbs/SF/Day. The exis-
tence of a vertical intercept a 1/(F/M) = 0 is not predicted
by theory and appears to represent the effects of reduced hy-
draulic rates and increased glucose concentrations used to at-
tain high F/M values.
A similar effect is evidenced by the data of Oleszkiewicz
and Eckenfelder (5). This inclined plane data correlated us-
ing a concentration dependent reaction as shown in Figure (6).
The wide variation in feed BOD concentration and recycle ra-
tios used in these studies were all encompased by a single
model. The phenomenon of a vertical intercept is again evi-
denced.
Zero Order Reaction
In a zero order reaction the rate of BOD removal per unit
weight of orgamism is constant and is not effected by sub-
strate concentration, degree of removal etc. The relationship
is:
(F/M)
(18)
Figure (7) illustrates the zero order reaction obtained on ef-
fluents from an Insulation Board mill operating on a wood and
mineral raw material, Quirk (6). Inclined planes of 8 and 10
feet were used and operated over a range of 23 to 30°C. Raw
wastewater BOD ranged from 1000 to 6500mg/l. Recycle ratios
of up to 20/1 were used. A ^Q value of 5 x 10~3 Ibs/SF/day
was obtained for the highly spread data. A second illustra-
tion of the zero order reaction was obtained in a study of
yeast fermentation effluents, Quirk (7),as shown in Figure 8.
A reaction rate coefficient of 9.0 x 10~3 Ib/SF/day at 20°C
was obtained. A comparison between this data and subsequent
pilot plant operation will be presented below.
SCALE-UP RELATIONSHIPS
Conversions to full scale tower conditions is made by ad-
justing plane performance for the following:
1. Hydraulic loading at full tower height.
2. Film thickness anticipated.
1018
-------
Holer (t) blucose
26.7 mg/l a 2H.K°C
«6.0 mg/
50.0 UK)/
300 mg/
27.5 mg/
65.5 mq/1
127.0 mg/!
_. . .• rag/I
65.5 rag/1
127.5 mo/1
k20 -.5 x
2H.i|°C
a 24.n°c
a 2n.i"c
a lo.6°c
a io,6°c
a lo.6°c
a 37.32C
a 37.3°C
a 37.3°C
10"5 Ibs/SF/Doy
10.«°C 4 24.4 °C
.1 .2 .3 .4 .6 .« .7 .> .8 1.0 1.1 1.2 1.3 1.4 1.6 1.« 1.7
Figure 5
fl AT / (F/u) - 3F/om/D«V
Glucose Removal on on Inclined Plane C4>
Sq - 215 to 5018 mg/l
e diluted feed
• reclrculated feed
• reclrculated feed
1000 to 5000 mg/l
1800 to 3<400 ma/1
0.2 to 2.0
250 to 700 mg/l
r - 2.75 to 8.50
Concentration'Dependant Kinetics
10-3
kt - 2.05 Ibs/day/SF
SO 1OO 18O ZOO 28O MO
1OS /(F/U) - SF/ Day /«>.
Figure 6 Pharmaceutical Waste Removal on an Inclined Plane (5)
imp
-------
e
§.
-i 1 1 r
HordBoorg Hostexoters (6)
Ternc 23.7° to 30°C
BOD 1000 to 6500 mg/I
Zero Order Reaction
k20 - 5 x 10-" IDS/SF/Doy
n - 1.0
• Mineral Board Hastewater
• wpoa Bogrd wgstewqter
minprnl/wnnn
4.0
3.1
J.O
Figure 7 Hordboard Wastewaters Treated on an
Inclined Plane (6)
1.0
0.01
1.0
Yeast Fermentation Wostewoter (?)
BOO = 1500 to 8000 mg/1
Zero Order Kinetics
* 9 x 10'3 Ibs/SF/day
Figure 8 Yeast Fermentation Wostewater Treated on an
Incline Plane (7)
1020
-------
3. Surface are? characteristics of packing media.
4. Hydraulic characteristics of packing media.
Hydraulic loading on a full scale tower is related to plane
hydraulics by geometry as follows:
where:
q = (U) (A"v)
q = application rate to tower in gal min~l
ft ~2 of tower area
U = application rate to plane in gal min~l
ft of film width
A"v= slimed area of tower packing media
supporting a film growth - ft^ft^S
Slime thickness reduces exposed surface area below that avail-
able from bare media. A knowledge of media configuration and
slime thickness can be employed to determine the correction
required as follows:
A"v = (A'v) (ft)
where:
A'v = wetted area of tower packing media
supporting a film growth - ft^ft~3
ffc = a factor for the reduction of slime
area below that of bare media due to
thickness of slime growth
Laboratory observations of film thickness indicate an ft range
of 0.80 to 0.90. Hydraulic characteristics of packing media
are introduced by relating the wetted surface area A1 and
detention time to hydraulic loading. Adjustments of this type
are required primarily for media shape and an "aspacked" geo-
metry other than that obtained with a vertically oriented
media. The adjustments account for the change in wetted area
which occurs as liquid impinges upon randomly packed media and
is splashed or otherwise diverted into contact with additional
media surface which would otherwise remain unslimed. Adjust-
ment can also be made for the change in reaction rate as a
result of a change in the rate of transport of BOD from the
flowing liquid to the slime surface. This can also include
the effects of removal of suspended BOD by agglomeration
1021
-------
processes. These additions considerably complicate the math-
ematical descriptions of the process while many times failing
to increase the accuracy of design calculations.
For randomly packed media,adjustments for hydraulic ef-
fects can be made using a mathematical form prevalent in chem-
ical engineering when packed towers are analyzed, i.e.
(q)'
and
ii)
(k)
(q)
n
Substitution of the above scale-up relationships into the
equation for plane performance under first order kinetics
yields the equation for full scale performance as follows
f =
where:
t H/(q)n (19)
C = a combined constant for all
hydraulic effects .
The value of the exponent (n) will vary from an expected min-
imum of 0.50 for randomly packed media to 1.0 for packing sim-
ilar to stacked vertical sheets. For vertical media a sheet
flow regimen dynamically similar to that of plane hydraulics
tends to be maintained. However, an adjustment is required
for use of less than total media area resulting from distribu-
tion hydraulics through the tower. A constant adjustment of
90 percent media utilization is employed as follows:
A v - r = QO
_ __ ~~ «TJ m J \J
AV
where: 2 _•}
A' s wetted surface area (ft ft J) ;
A = manufacturer's rating for dry
v media (ft2ft-3)
C = a coefficient for wetting efficiency
w
for vertically oriented media the value of the hydraulic co-
efficient (C) in equation (19) equals Cw and the design re-
lationship is stated as follows:
1022
-------
where:
f =, e
k20 =
(20)
' ft ' Cw
The above scale-up procedure assumes that the adjustment for
non-uniform slime growth can be made by using a single linear
correction factor (Cw). This is equivalent to assuming that
areas of non-slime growth occur with equal size and frequency
throughout the tower depth. This assumption may or may not
apply to all tower designs. The alternative approach is to
assume that the uniformity of hydraulic distribution will vary
in a non-linear fashion with tower height (H). In this cir-
cumstance the effective wetted surface area will be related to
height as follows:
Av
= C
(21)
where:
C = a correlation constant
(1-m) = a measure of the non-linear
distribution of film area.
As the value of m increases, the non-uniformity of film growth
with height also increases. At m = 0 slime growth is uniform
throughout the tower at m = 1 there is no slime growth in the
tower. The full scale relationships when m > 0 is:
f =
Full scale design equations for the balance of the kinetic
models are summarized on Table 1.
Scale-up Calculations
2^-3
"Whey and Sewage - using a packing with Av = 27 ft ft
o0c plane = 0.0018 gpm/ft2 = k'p
= Cw x f t x AV
= 0.9 x 0.9 x 27 = 22 ft2ft~3
(TOWER)
Av
1023
-------
Y«_-asc Wastewater -
= 0.0018 gpm/f 2 • 22 ft2
ft?
= 0.04 gpm/f t3
01. served k20o= °-03 gpm/ft3
using a packing with = 27 ft2 ft~3
plane = 0.009 lbs/day/ft
\ = 27 ft2/ft3
A^ = 22 ft2/ft3
k20° TOWER = °-009 x 22 = 0.198 lbs/day/ft3
Observed kO = 0.180 lbs/day/ft3
1024
-------
CO
CM
CM
m
CM
CN
cr
cc
t-i
0)
4-1
i— 1
•iH
ft,
60
C
CX
•H
4=
CO
C
o
•H
4—1
CO
t-l
CU
C
f-l ^
I cr
R
e
<^>
V
c
CO
V-i '
•H
fn
.. ;
cu
tH
&
•H
CO
4-1
rj
«'.
J-(
CO
4J
a;
r-(
cd
•H
C
CU
C
O
ex
X
w
4-1
p*
CO
-a
}-i
cO
4-1
CU
Pi
;H
4-1
CO
4-1
C
cu
a
• pi
0
4-1
CU
C
cu,
ex
o
1025
-------
Full Scale Performance
For a given type of packing,tower volume can vary with
the following design parameters:
1. Liquid application rate
2. Recycle ratio
3. Tower height
4. Efficiency of BOD removal
The effects of variations in the first three parameters are
dependent on (1) the necessity to maintain a minimum wetting
rate and (2) the numerical value of the exponent (n). In all
cases, an increase in efficiency of removal requires an in-
crease in tower volume. In general, the effects of design
parameters can be described as follows:
Design Variables
Change in Variable
Change in tower
volume
H
r
q
E
Increase
Increase
Increase
Increase
Decrease or no ch
Increase or no ch
Increase or no ch
Increase
Structural requirements and hydraulic distribution problems
limit maximum tower heights. Heights of 20 ft are common with
maximums to 45 ft.
Commercial packing of the lattice structure type appears
to require minimum application velocities of + 0.5 min ft~2.
Operation below the minimum velocity can result in progress-
ively less utilization of tower packing.
In order to maintain commonly used heights and provide a
minimum application velocity, effluent recycle is usually re-
quired. The added tower volume required to accommodate the
recycle varies with both the efficiency of removal sought and
the recycle ratio finally employed. Because of the
1026
-------
non-uniform influence of process variables, design calcula-
tions can involve relatively complex manipulations.
Process design data from pilot plant operation are ob-
tained using relationships (23) to (27) together with graph-
ical correlation techniques. The type of correlations used
while mathematically identical can non-the-less effect the
value of the process constants obtained and, therefore, the
design of the full scale unit. Techniques which utilize, sep-
arate rearrangements of Equations (23) to (27) to calculate
m & n independently can have this tendency. A single correla-
tion equation can be obtained by taking logarithum's twice
and rearranging the result to provide a linear relationship.
When m = o the full scale first order equation is expressed
as:
log
H6
AT
•2-3 log (f)
= n log(q) + log 1/k
20
(28)
The slope of the correlation is (n) and the intercept at q =
1.0 is log (l/k20).
In order to provide for the possibility that m>0 a two
part technique is used and is illustrated Figure 9.
1. Let n - U-ro'.-C
2. Rearrange eajct;on UK)
3. Obtain .slope =• 1-m
1. Obtain intercept
values at Q/H =1.0
3.
Plot Intercept values
from Step 1 vs.
hydraulic rote (Q)
Obtain Slope •= C
Intercept - logn/j...'
l/K20j
Compute n value
1. Substitute IT. value
from Step 1 into
general eauotion
2. Rearrange eauailor. n»)
3. Obtain slope = n.
4. Obtain Intercept
« log f ' "
Figure 9
LOO (a)
Graotilcal Analysis for m greater than o
1027
-------
A correlation is first obtained for variable tower height
at constant hydraulic rate (q). Double logarithums of equa-
tion (24) are used together with a mathematical substitution
designed to eliminate the effects of the hydraulic coefficient
(n) from the correlation. With the influence of (n) thus
eliminated, the value of m may be determined. The elimina-
tion of (n) is obtained by making the following substitution,
n = (1-m) + C
(29)
where C = a constant whose value,determined from test data,
will be such as to make the above statement valid.
Using this substitution with equation (24) the following lin-
ear equation is obtained after logarithums are taken twice:
AT
log
e
2.3 log (f)J
= (1-m) log (q/H) + log (CONSTANT)
(30)
where:
Constant = (q) /k
20
Using variable height data with constant hydraulic rate a plot
of equation (30) will yield (1-m) as a slope. At this junc-
ture, no use of the correlation intercept is made as the pur-
pose of the substitution was to remove (n) from the slope and
relocate it in a constant value intercept as part of C. The
intercept value has no other use in this part of the correla-
tion.
If data are taken over a range of hydraulic rates then a
plot may be made separately for each (q) value. This will pro-
vide a number of intercept values. These intercept values
may then be related to hydraulic rate in such a way as to
yield the numerical value of C.
log
C log(q) + log l/k2Q (31)
Log (q)°/k20 is obtained as the intercept value from prior
use of equation (30). Tower performance on the treatment of
black liquor from Eckenfelder (8) is shown in Figure 10. Data
for an integrated kraft mill effluent, Quirk (9), is shown on
Figure 11. Again n = 1.0 and a k£Q = 0.056 gpm/CF is deter-
mined.
Experience of the authors has shown that unless tower
operation is below a minimum wetting rate or there is a
1028
-------
Pol van a (I)
H 3.3 to 18.V
q .75 to 3.0 gom/SF
m = o
n = 1
kin = .056 gpm/CF <2<(0C>
037 gom/CF (390O
Figure 10 Treatment of Black Llauor on a Polygrld Tower (3)
t;
1
x
I I i i i i
Integrated Kraft 19)
H - 21'
a « 1 to 3.25 gom/SF
T -
26 to
n « 1.0
k2Q - .019 9Dm/CF
-i 1—>—i i i i i i 1 1—i i i i i i
«-0 10
Figure II Treatment of Integrated Kraft Effluent on
Q SurfDac Toxer (5)
1029
-------
-1—I—I I I I I
Scnultze HO)
Sewage » wtiey
6' Filter
3/1"-1 1/2" Stone
Temperature Unknown
n * 0.51
kt « .056 gom/CF
.10
u
Figure 12 Hhey and Sewage'Treated on a Gravel Packed
Filter (10)
sooo
1000
Heat Processing wastewaters (11)
Unsteady State .Operation
Influent BOD changes every 30 mm.
Unsettled Influent •
Settled Influent —
Settled influent —
Settled influent —
• kt
-kt
.027 gom/CF
.017 gpffl/CF
-"22 9OT/CF
.020 qpni/CF_
"• n - 1.0
H - 20.0'
q - 0.85 9Pfn
-»oo-
I
3OOO
10OO
8* ng/l
10.000
1000
10.000
Figure 13
unsteady Stdte Treatment of Meat Processing
Hastewaters on a 20 ft. Tower (11)
1030
-------
malfunction in the hydraulic distribution, the value of m will
equal 0 and n = 1.0 when dealing with vertical sheet flow
packing.
When working with random packing, values of n<1.0 will
be obtained. The data of Schultze (10) treating sewage plus
whey on a 6'deep gravel packed filter are shown on Figure 12.
The 3/4 to. 1 1/2" stone media provided a specific surface
area computed at 36 SF/GF. Filter operation covered a range
in hydraulic loading of from 0.0375 to .375 gpm/SF. An (n)
value less than 1.0 is evident. A reaction rate of.056 gpm/
CF was determined.
The doctoral thesis data of Hoodie (11) allows an exam-
ination of the correlation technique when applied to unsteady
state performance. Using a 3 sq/ft tower, 20' high, Hoodie
experimented with meat processing wastewaters with a COD of
up to 5000 mg/1. Hydraulic rates varied from 0.56 to 1.27
gpm/SF and an intermediate effluent sampling point at 12' was
used. Every 30 minutes the COD o,f the influent was changed
while the hydraulic rate remained constant. The response of
the tower to these variations in loading is shown in Figure
13. Effluent concentration (S$) is related to influent concen-
tration (Sa) using a First Order Reaction and equation (24).
Even though there is considerable scatter in some of the data,
the general applicability of the correlation is evident.
A comparison of unsteady state reaction rates with hy-
draulic loading and influent concentration variations is pro-
ivded in Table 2 and demonstrates 'that with careful sampling
technique, stable performance data can be obtained from
unsteady-state operation. Where such control can not be ob-
tained, as in a field installation which is not used for a
Ph.D. thesis, much more variation in unsteady-state perfor-
mance can be expected. This is illustrated in Figure 14 for
yeast fermentation effluent treated in a 21 pilot tower re-
ceiving continuously variable flows and influent concentra-
tions, Quirk (7). The inclined plane data previously refer-
enced on Figure 8 were superimposed on the pilot plant data.
The scale-up calculations described above were used. The lab-
oratory k'of 9 x 10~3 scaled-up to 0.180 Ibs/CF/day. Figure
14 illustrates that on the average the unsteady-state field
data group around the scale-up line from the laboratory plane.
An additional aspect of hydraulic loading is illustrated
in studies of Eckenfelder (15). Eckenfelder utilized a pilot
plant equipped with a vertically oriented asbestos packing,
operating over 7 to 1 height change, and 3 to 1 variation in
hydraulic loading. His data are presented in Figure 15.
1031
-------
BO
40
30
20
Yeast Fermentation Hostewaters
-------
Table.2 •
, : Applications of First Order Reaction
Correlation to Unsteady-State Operation after
Hoodie (11)
Hydraulic
Rate
Tower
Height
COD
Range
Reaction
Rate
0.56 gpm/SF 20.0'
0.85 gpm/SF 20.0'
1.27 gpm/SF 20.0'
1750-4800 mg/1 0.017 gpm/CF
1400-4000 mg/1 0.022 gpm/CF
1750-3000 mg/1 0.020 gpm/CF
It is seen'that a volumetric loading rate of between 0.15
and 0.20 gpm/CF on abrupt change occurs in the value of the
hydraulic coefficient (n). Above this threshold value and (n)
of 1.0 fit the data quite well. Below it the n decreased to
.50. This indicates that at some limiting value of hydraulic
loading n will decrease below 1.0 in response to poor wetting
efficiency and flow channeling which prevents uniform slime
growth throughout the tower.
Process Design
The previous illustrations of correlation approaches and
data fitting procedures underscore the fact that determina-
tion of the proper model is of paramount importance. In a
similar manner, it is necessary that laboratory and/or pilot
scale studies incorporate the full range of design variables.
Extraplotation beyond the confines of measured data can be
risky at best.
The effect of recirculation on design capability is a
topic of particular concern in process design. Conflicting
data and opinions populate the literature. Recirculation de-
creases the detention time while increasing the velocity of
flow &turbulent transport of BOD to the slime/liquid interface
which are opposing effects. A minimum hydraulic loading is
necessary for through wetting and this can be provided by re-
cycle. In order to approach the recycle question properly it
is necessary that recirculation over a wide range be utilized
in treatability studies so that data exist for recycle ratios
in excess of those ultimately selected for final design. With
these data in hand, process models may be employed to determine
1033
-------
Che net effect of recirculation.
Using the first order model as an illustration, the pro-
cess design calculation re-express the recycle/efficiency
function (f) as follows:
f =
"
1 +
(1-E) = EC + r
(1+r) (1-E)
1+r
(32)
Where Ec relates to BOD removal efficiency i.e.,
Ec = (1/1-E)
and allows topography to be simplified as the design analysis
precedes. The concept of a unit volume of tower per unit 'of
raw wastewater flow is then introduced. Equation (33) is em-
ployed to obtain a ratio of these unit volumes with and with-
out recirculation. The final relationship between unit vol-
umes is shown below:
where:
^ = (1 + r) (C ) (33)
* o **
V = the unit volume when recirculation
is practiced
VQ= the unit volume without recirculation
C = a sensitivity constant for a given
' efficiency and recycle ratio
When the constant (CE?r) equals its maximum value of 1.0,
the unit volume required with recirculation varies directly
in proportion with the arithmetic effect of recycle i.e.-^^=
(1 + r). At constant values less than 1.0 the increase in
tower volume is reduced below that dictated by hydraulic
through-put ratio. This metigation effect is formulated for
a first order reaction as follows:
(Ec+r
log 1 + r
[log (Ec) J
1/n
(34)
Using a vertically oriented media and an n value of 1.0, the
value of (Cg r) will vary with design efficiency as illustrat-
ed on Figure 16. While the impact of recirculation is reduc-
ed on Figure 16 as recycle ratio increases, total tower volume
1034
-------
UJ
o
1.0
0.9
0.8
0.7
0.6
O.5
0.4
0.3
0.2
Recirculotlon Effects
on
Trickling Filter Volume
First Order Reaction
n = 1.0
0123466
<1 -f r)
Figure 16 Recirculation Effects on Trickling Filter
Volumes
1035
-------
experiences a net increase. Unit volume ratios are compared
with design efficiency requirements and recycle ratios in a
general manner on Table 3. The analysis shows that recircu-
lation effects are moderate only for low removal efficiency
or tower performance as a pretreatment or roughing unit.
At (n) values less than 1.0 the effects of recycle are
much reduced over these shown on Table 5 . For example, at
n = 0.50 the unit volume ratio approaches 1.0 even though ef-
ficiency is 90% and a recycle ratio of 2.0 is used. This may
be compared with a volume ratio of 1.7 under comparable con-
ditions when n = 1.0. However, this can be only an apparant
volume reduction in that at n values less than 1.0 tower
media is not being used effectively and the value of k2Q will
be reduced. This can offset the reduced effects of recycle
usage.
A similar process design approach is used for the alter-
native reaction models. Unit volume relationships for these
models are shown on Table A.
A summary of reaction rates and reaction models for in-
dustrial effluents is presented on Table 5. In each case
sited, tower media was of-the vertically oriented type.
1036
-------
Table 3
Effect of Recirculation
on
Tower Unit Volume
First Order Reaction
n = 1.0
Efficiency
90
90
90
90
50
50
50
50
Recycle
Ratio(r)
1
2
3
4
1
2
3
4
Unit Volume
Ratio (V/VQ)
1.48
1.80
2.04
2.25
1.18
1.23
1.29
1.30
1037
-------
Table 4
Effect of Recirculation
on
Reaction Models n = 1
1. Zero Order 1.0 1.0
2. First Order (1+r) (1+r)
3. Simple Retardant 1.0 1.0
4. Concentration (1+r) (1+r)
Dependent
Tower Unit Volumes
Basic Multiplier Efficiency Recycle Constant
n = n n=l n = n
1.0
log (EC)
. i.o
'/E^+r
i.o
" ,Ec-frv
log (- --[+r )
1/n
log (Ec)
1/n
log (Ec)
1.0
E +r
E (I+r)
1/n
1038
-------
c
0)
D
w
o
C--J
cx~
60
rH CN r-^ -cr OO
CO
C
O
o
CO
Pi
C
O
CO
CO
I-
o
l-l
a>
-a
4-1
CO
o
t-l
l-l
cu
"0
S-i
o
CO
l-l
•H
o
CU
o
E-i
0)
g
M
cu
TJ
4-1
CO
l-l
•H
PL.
t-l
cu
4-1
C
CU
^
r-H
T3
4J
>-l
60
0!
4-1
c
hi
i-H
f__{
•H
^
4-1
14-1
CO
l-l
< « O O Pd PM t
rH
iH
e
rH C
•H CU
rH
00 4-1
Cfl 14-1
on w
J-l CO
cfl cO
O &
O -H
t-l
o
cu
Nl
l-l
0)
"O
o
o
01
Nl
cu
C
co
C
o
•H
4-1
Cfl
±J
C
cu
e
01
p£t
4-1
CO
CO
cu
>"
^|
cu
4-1
Cfl -
^
CU
4-1
CO
cfl
c
o
•H
4J
CO
C
0)
£
£4
0)
PL,
4-1
CO
Cfl
CU-
l-l
CU
4-1
CO
S
cu
4-J
CO
cfl
S
W
1039
-------
u
o
00
o
tx.
o
p-
o
oo
00 ,
oo
00
fn
c/1
13
En
(U
C 13
Qj C
O 0)
C D.
O •
O O
N)
0)
P.,
u>
•H
(A
U CO
O -P
M C
PH 0)
cd
a)
•P ^
-------
References
1. Hoehn, R.C. and Roy, A.O. "Effect of thickness of Bacter-
ial Film" Journal of the Water Pollution Control Federa-
tion, 45, 1973.
2. Quirk, T.P. "Whey Effluent-Packed Tower Trickling Filtra-
tion" EPA Water Pollution Control Research Series No.
12130 DUJ, 1971.
3. .Shulze, V.L. "Experimental Vertical Screen Trickling Fil-
ter" Sewage and Industrial Wastes, 29.. 1957.
4. Maieir, W.J., "Mass Transfer and Growth Kinetics on a Slime
Layer, A Simulation of the Trickling Filter" Ph.D. Thesis,
Cornell University, N.Y., 1966.
5. Oleszkiewicz, J.A. , and Eckenfelder, W.W. , "The Mechanism
of Substrate Removal in High Rate Platic Media Trickling
Filters" Technical Report No. 33, Dept. of Environmental
and Water Resources Engineering, Vanderbilt University.
6. Quirk, Lawler and Matusky Engineers Technical Report, 1970.
7. Quirk, Lawler and Matusky Engineers Technical Report, 1970.
8. Eckenfelder, W.W. , and Barnhart E.L. "Performance of High
Rate Trickling Filter using Selected Media" Journal of
Water Pollution Control Federation, December, 1965.
9. Technical Association of the Pulp and Paper Industry-Pilot
Plant Studies, 1967.
10. Shulze, K.L. "Coal and Efficiency of Trickling Filters"
Journal of the Water Pollution Control Federation, 32,
~ '
11. Moodie, S.P., "The Design of High Rate Trickling Filters"
Ph.D. Thesis, University of Queensland, Australia, 1981.
1041
-------
TREATMENT OF COKE PLANT WASTEWATERS
IN PACKED BED REACTORS
MEINT OLTHOF. INDUSTRIAL WASTE SECTION
DUNCAN, LAGNESE AND ASSOCIATES, INC.
JAN OLESZKIEWICZ. INDUSTRIAL WASTE SECTION
DUNCAN, LAGNESE AND ASSOCIATES, INC.
WILLIAM R. O'DONNELL. LEOPOLD COMPANY
The treatment of wastewaters originating from coke
oven operations has been studied in a wide variety of
biological treatment operations. Most of these studies
have shown that this wastewater is biodegradable to a
large extent and that given enough acclimation, the
bacteria can develop to fully treat compounds like phenol,
cyanides' and thiocyanates. The work that will be
presented ,in this paper describes the results of a
treatability study of wastewaters originating from a
benzol plant in an upflow biotower (UBT). The wastewater
constituents are the same as for coke oven wastewater,
except that most of the constituents are present in some-
what lower concentrations. The biotower used in this
study is a biological treatment process developed by the
Leopold Company and operates more or less like a reversed
flow trickling filter. The tower is packed, with random
plastic medium, the influent flows upward and air is
supplied by aeration through a filter underdrain system.
This paper will present the results of the treatability of
benzol plant wastewater in this reactor. The main pur-
pose of this study was to determine'the loading at which
the phenol was virtually completely removed from the
influent. The data obtained in this study will be
1042
-------
compared with studies performed by other researchers with
similar wastewater in activated sludge and with other
types of fixed film reactors. In this way, it will be
possible to .compare the performance of the different
biological treatment reactors available.
EXPERIMENTAL SET-UP
The work conducted in this project was done with a
pilot unit of the upflow biotower (UBT). Figure lisa
schematic diagram of the pilot plant set-up. The tower
had dimensions of 2 x 3 x 10 ft. high and was filled with
39 ft. of random,, plastic medium with a specific surface
area of 30 ft /ft . The air was distributed through a
filter underdrain at a rate of 2-5 scfm/ft . The waste-
water was fed at varying flow rates (0.5-2 gpm) through
the tower. ~The. waste characteristics of the influent to
the UBT are shown in Table I. It shows that the•average
soluble organic carbon (SOC) concentration is 435 mg/1.
The COD is about 1500 mg/1. 'This strength is about half
of waste ammonia liquor. The average phenol concen-
tration was only 36 mg/1, which is significantly less than
in waste ammonia liquor. The fact that the ratio between
COD and phenol in this waste is so much higher than in
normal waste ammonium liquor, indicates that the waste
from the benzol plant contains many more organics in
addition to phenol, while waste ammonia•liquor, primarily
contains, phenols as the organic material. Other specific
organics analyzed to be present in this waste are benzene,
toluene and napthalene. The concentrations from grab
samples for these specific organics during this study are
also shown in Table I. The pilot plant was started up by
feeding activated sludge from an existing wastewater
treatment plant'treating coke oven wastewater. The phenol
removal was virtually complete a few days after adding
seed to the tower. During the pilot study, the reactor
was fed hydraulically with 0.5, 1 and 2 gpm of wastewater.
The influent values fluctuated throughout this period so
that the organic load also fluctuated from day to day.
RESULTS OF THIS STUDY
The results of this study showed that the organics
present in this wastewater are very well biodegradable and
that it is possible to treat this wastewater biologically
1043
-------
NUTRIENT
FEED TANK
( PHOSPHORIC
ACID)
RAW
WASTE WATER
55 GAL.
DRUM
55 GAL.
DRUM
PRIMARY
CLARIFIER
EFFLUENT
VENT
FINAL
EFFLUENT
1 l_S_._r_^-
• —
r
L^^TP
_T^ -^
FINAL
CLARIFIER
'I
BIO-
TOWER
39 ft'3
(6.5 ft H
PACKING)
6ft2
AREA
^
-co
-co
-co
t ^
AIP
2-5 SCFM/ffs
FIGURE I
LAYOUT OF THE PILOT PLANT
1044
-------
TABLfi I
INFLUENT CHARACTERISTICS
Parameter
Suspended Solids
Oil and Grease
Phenol
SOC (Soluble Organic Carbon)
SCN
CN - Free
CN - Total
BOD5
COD
NH4-N
Sulfides
Grab Sample
OF WASTEWATER
50 % Value*
50
32
36
435
63
11.3
21.8
700
1500
42
40 -
TO UBT
90% Value*
N/A
200
67
630
102
18
33.4
995
2300
85
64
Benzene
Toluene
Naphtalene
208
30
* 50% Value: Mean concentration from a probability plot.
90% Value: Indicates the concentration that is not
exceeded in 90% of the samples analyzed.
1045
-------
to a very high extent. The probability curve for the
percentage SOC removal is shown in Figure 2. Typical
removal percentages for the various parameters are shown
The hydraulic load was 1 gpm/39 ft or
The corresponding average organic
load was 120 Ib SOC/1000 ftj day or 400 Ib COD/
in Tableau.
37 gpd/ft reactor.
was
1000 ft -day.
TABLE II
PILOT BIOTOWER LOADINGS AND OBTAINED REMOVALS
DURING RUN III - 1 GPM
Parameter
SOC
CODf
BOD5,f
CN-Free-Filtered
CN-Total-Filtered
Phenols-Filtered
SCN-Filtered
NH4+
Sulfides
Mean Loading
lbs/1000 ft -d
135
462
216
3.5
6.7
11.1
19.5
13
4.2
Mean
Removal (%)
57
,63
•5.1
71
75.2
99.91
61
0
73
In Figure 3 the percentage SOC removal is plotted
versus the effluent phenol concentration. This shows that
as long as at least 40-50% of the SOC is removed, the
effluent phenol concentration is very low. The removal
of free cyanides was not complete at the loadings at which
the plant was operated. Based on performance of other
systems, however, it is felt that free cyanide is totally
1046
-------
LL)
£E
IOO-
90 -
80 .-
70 -
60-
O
8. 50'
40 -
30-
20+ l-l-l—I—I—I 1 1 1 1— I—I—I—I 1 1—I
O.OI O.O5 O.2 O.5 I 2 5 IO 2O 30 4O 5O 60 70 80 90 95
FIGURE 2
CUMULATIVE FREQUENCY DISTRIBUTION
OF PERCENTAGE SOC REMOVAL
DURING RUN HI
1047
-------
0 +
0
10 20 30 40 50 60 70 80 90
% SOC REMOVAL
FIGURE 3
EFFLUENT PHENOL CONCENTRATION
AS A FUNCTION OF
PERCENTAGE SOC REMOVAL
1048
-------
biodegradable, provided the loading is low enough-. The
same is true for;thiocyanates. (See Discussion section.)
:. The only nutrient required for this waste stream was
phosphorus. This was added at such a concentration that
the effluent phosphorus concentration was maintained at
;,1 - 2 mg/1. This required on the average only 5 mg/1.
The amount of sludge :generated in this system ;repre-
sente,d, a yield factor based on *a BOD basis 'of'about 50%.
This, "-is comparable to conventional nigh-rate biological
systems. The loading -,at which this- sludge yield was
obtained was quite .high (216 Ib BOD /1000 ft-/day).
The air requirements for this unit were comparable to
activated sludge. ^The DO concentration in the effluent
was normally high (6X7,, mg/1) and therefore in the scale-up
of the system, it is possible to reduce the amount of air
required under normal .operating conditions.
Throughout the course 'of this six-month _. study, it
never proved to be necessary to backflush the tower. The
excess sludge sloughed off at a sufficient-rate to main-
tain a good flow and aeration throughout the system. The
upflow flowrate through the tower was about 0.16 gpm/ft .
Throughout this project, the unit suffered several
shock loads as result of leaks or spills in the'plant from
where the waste originated. .:•..The unit showed remarkable
potential for quick, recovery. Figure 4 shows a few cases
where the SOC jumped up from one day to another and the
effluent SOC value also "jumped up. However, .after
stabilizing the, influent condition, the effluent of the
biotower recovered rapidly. Throughout this study, the
phenol concentration in the effluent was * only in two
effluent samples above 1 mg/1. In each case, the phenol
concentration was below"1 mg/1 the following day.
The kinetics of this system can be expressed by the
equation
S /S = EXP(-K/L)
e o
where S is the effluent concentration, S is the influent
e o
concentration, K is a constant, L ,is loading. By plotting
log S /S versus 1/L, it is possible to obtain the value
of the removal coefficient. This is done in Figure 5 and
it shows a removal coefficient of K = 0.18 Ib/cu. ft/day.
104?
-------
4
I—I —I—I—I—I—I—l-*f ~
ro
EWOOOI/sqi 9NIQVO1 DOS
^—i_l_i_4co —*•
Q o o o o
* S £ Q 8
iNarruda oos
LJ
a:
Ld
CC
CO
o
en
1050
-------
IOOO--
9OO -f
800 7
7OOT
600 j
500y
4OO-r
O
8
2
UJ
'
.
U.
LL)
O)
cn
200-
)OO
90
80
70
60
5O
REMOVAL
COEFFICIENT
K = O.I83lbs./ft3-d
(25,96)
FLOW RAW/RECYCLE
• 1.0:0 RUN HE
I.O:I.O RUN H
12
14
16 18
I/L (d-ft3/lbs)
: LOAD lb/ft3DAY
FIGURE 5
KINETICS OF SOC REMOVAL
IN THE PACKED BED REACTOR
1051
-------
DISCUSSION
It is possible to compare the performance of this
upflow biotower with results reported by other investi-
gators of other biological systems. Since the main
objective of various biological systems is to have the
highest organic loading and the smallest possible volume
and still provide a stable operation, the results of all
data that will be discussed in this paper are expressed in
terms of performance versus volumetric organic load.
Activated sludge data can be converted to this by assuming
that the activated sludge concentration is 3 g/1 of active
biomass. The results for COD removal are plotted in
Figure 6' and it shows that the performance of the UBT is
quite favorable. The performance of activated sludge is
very close to the packed bed reactor, but one of the
disadvantages of activated sludge, i.e. the sludge
settling, has been avoided. Also, for an industrial type
of operation, it is felt that a fixed film bioreactor is
more suitable to handle shock loads.
The removal of CN and SCN was only partial at the
organic loadings the plant was operated. Generally, these
components are biodegradable at lower loadings. Figures 7
and 8 show the percentages removal and the effluent con-
centrations for these parameters as a function of loadings
for several types of systems. These data show quite a
fluctuation between the various reported data. In
general, the breakdown of CN and SCN are thought to be
totally- degradable at lower loadings (<200 Ib COD/
1000 ft /day). The CN removal data are more scattered
than the SCN data because of inconsistency in reporting CN
as total or free. Only the free cyanide is biodegradable
and therefore in cases with a large fraction of fixed
cyanide, the percentage removal will be low. Why in some
cases SCN is not totally removed at low loadings is
uncertain. Some research conducted in this area seems to
direct the focus to certain environmental conditions (pH,
phenol, ammonia concentration) but no total picture is yet
available on SCN oxidation (refence 17).
The effluent phenol concentration is of importance
since BAT guidelines require very low limits (0.025 mg/1).
This is not easily met with a biological system. Figure 9
shows the effluent phenol concentration for various
1052
-------
LJ
OC.
Q
O
O
100-
90T
so
70-
6Oj
50 y
40--
REFERENCES
PBR-A-7HIS STUDY
A.S. - 1,2,35,6,7,9,II,I2S,13
RBC-3,12
TF- 5
•I-
IOO
2OO
•I-
3OO
4OO
COD LOADING
Ib / IOOO ft3/ DAY
5OO 6OO
FIGURE 6
COD REMOVAL VERSUS
ORGANIC LOAD FOR VARIOUS
BIOLOGICAL REACTORS
. 1053
-------
^
o
5
UJ
o:
"Z.
o
_,
X
o
!E
*^
u
,
2
UJ
_|
U.
u.
LU
IOO -
9O -
8O -
70 -
60 -
50 •
4O •
30 •
10 -
6 -
5 -
4 -
3 -
Z -
«c
1 -
O.8 •
O.6 •
O.4
0.2-
-
(2)®^ REFERENCES
Q PBR - A-THIS STUDY
Q® ff. AS - 1, 2, 3S, 7,9,11,125,15
• ©^D ^© RBC-3,12
® © © ®
©
.©©©©© © .
i 1.1 i
L ©
- (J)
CUD
®(A)
^^ISS^-N (Ss)
V->(9) V_x
i (D (TTiCD
0 ^^
CD
- © © ^-\
w
- dXD®, ©
D IOO 2OO 3OO 40O 5
DO
COD LOADING
Ib/ IOOO ft3/DAY
FIGURE 7
CN REMOVAL VERSUS
ORGANIC LOAD FOR VARIOUS
BIOLOGICAL REACTORS
1054
-------
_J
^
O
5E
tr.
z
a*
_J
O
s
2
&
i—
LU
Z>
_l
tj
UJ
7~T9J — x^v^vC\§/T9Tx§? x — v.
90 -@ k^S'Sg) A3*
I/[7\(^) ^^
vJy^-^ fin
80 -
70 •
6O •
5O •
40 •
3O •
20 •
640-
32O-
160-
80 -
40-
20 -
IO-
5 -
O -
\^s
s~\ ' ' f(b
(D
REFERENCES
PBR- A- THIS STUDY
AS - 3S,4, 6,7, 8, 9, II1, 125,14,15
V$ RBC- 3,12 . ,
1 i i i
r 1 1 1 |
®
' ® ©
•(D
©
(D
• © ® .-^
s~\ (5s) /~\ —
@®® ®
®cl)®(6) (D ,
0 IOO 2OO 3OO 4OO 5C
COD LOADING
Ib/ 1000 ftVDAY
FIGURE 8
SCN REMOVAL VERSUS
ORGANIC LOAD FOR VARIOUS
BIOLOGICAL REACTORS
1055
-------
too
REFERENCES :
PBR - A-THIS STUDY
AS - 1,2,35,4,7,8,9,10,11,125,14
RBC-3,12
© (5s)
•I-
200
30O
4OO
COD LOADING
Ib / IOOO ft3 /DAY
500
600
FIGURE 9
PHENOL REMOVAL VERSUS
ORGANIC LOAD FOR VARIOUS
BIOLOGICAL REACTORS
1056
-------
studies. The figure shows only 3 effluent values below
0.05 mg/1. It is difficult to predict what will be
required to, get these very low effluent phenol concen-
trations. Low loadings is obviously a prerequisite but
this seems to be no guarantee for low phenol effluent con-
centration. The .analytical techniques used for analyzing
phenol could have a large. ,: implication at these low
residual values. / ,
One factor that requires some additional study is the
anticipated improvement in overall performance by staging
UBT reactors. As with plug flow versus completely mixed
i't is felt that by operating 2 reactors in series, the
effluent quality will be, better than with 1 reactor with
the same volume as the 2 reactors combined. With fixed
film reactors, there is more, of an opportunity, for special
groups of bacteria to be growing in certain sections of
the reactors. In batch tests, it has been shown
(reference 16) that the sequence of bio-oxidation of the
pollutants in coke oven wastewater is. phenol-cyanide-
thiocyanate. By staging reactors, it is possible to
simulate this type of ;0peration.
Another factor that shows promise for optimizing the
performance of the UBT is by reducing the energy require-
ment per pound of COD removed by increasing the height of
the tower. By increasing the height, the air has a longer
detention time and the packing should ensure a tortuous
path of the air. Both factors seem to work toward an
energy efficient reactor. No data are yet available on
this but intuitively this approach should be advantageous.
CONCLUSIONS ' ' . ' . ':
The UBT proved to be an efficient reactor for
biological degradation of benzol wastewater. Organic
loadings as high as .300 Ib COD/1000 ft /day resulted
in 70% or higher COD removal.
The percentage removal of CN and SCN was about 60-70%
at the loadings at which the reactor was operated.
Phenol is virtually completely oxidized as long as
the percentage COD- or SOC removal was above 50%.
This does not necess.arily mean that the phenol is
completely oxidized but at. least the ring is broken
and the resulting organic does not register as phenol.
1057
-------
REFERENCES
10.
11.
12.
Adams, C.A., R.M. Stein, W. W. Eckenfelder.
Treatment of two coke plant wastewaters to meet
guideline criteria. Proceedings 29th Purdue
Industrial Waste Conference, 1974, 864.
Adams, C.A. Treatment of high strength phenolic and
ammonia wastestream by single and multi-stage
activated sludge processes. Proceedings 29th Purdue
Industrial Waste Conference, 1974, 617.
Bauer, G.E. Biophysical treatment of -coke plant
wastewaters. Zimpro, Technical Bulletin 3250-T.
Bridle, T.R., W.K. Bedford and B.E. Jank. Bio-
logical treatment of coke plant wastewaters for
control of nitrogen and trace organics. 53rd Annual
Water Pollution Control Federation, September 1980.
British Coke Research Association. The biological
treatment of coke oven effluent. Laboratory and
pilot plant studies using packed towers, Coke
Research Report 52, May 1969.
Cooper, R.L. and J. R. Catchpole. Biological treat-
ment of phenolic wastes. Iron and Steel Institute
Publication No. 128, 1970.
Cousins, W.G., A.B. Windier. Tertiary treatment of
weak ammonia liquor. Journal Water Pollution Control
Federation, Vol. 44, No. 4, April 1972, p. 607.
Ganczarczyk, J. , D. Elion. Extended aeration of
coke plant effluents. 33rd Purdue Industrial Waste
Conference, 1978, p. 895.
Jones, L. Experimental investigations on biological
treatment of coke plant ammonia still effluent. M.S.
Thesis, Carnegie-Mellon University, 1978.
Kostenbader, P.D., J.W. Flecksteiner. Biological
oxidation of coke plant weak ammonia liquor. Journal
Water Pollution Control Federation, Vol. 41, No. 2,
February 1969, p. 199.
Luthy, R.G. and J.T. Tallon. Experimental analysis
of biological oxidation characteristics of hygas coal
gasification wastewater. Department of Energy,
FE-2496-27.
Medwith, B.W., J.F. Lefelhocz. Single stage bio-
logical treatment of coke plant wastewaters with a
hybrid suspended growth fixed film reactor. 36th
Purdue Industrial Waste Conference, 1981.
1058
-------
13. Olthof, M. , E. Pearson, N. Mancuso, I. Wittmann.
Biological treatment of coke oven wastewater
including provisions . for nitrification. Iron and
Steel Engineer, June 1980.
14, Osantowski, R., et.al. Two stage biological treat-
ment of coke, plant wastewater. EPA 600/2-81-052,
April -1981. ' ' .-."" .-'V' . , .' :
15. Wong-Chong, G., S. :Caruso. Advanced .biological
oxidation of, coke plant wastewaters for the removal
of, nitrogen compounds. Carnegie-Mellon Institute,
April 1977. . ••' • •
16. Wong Chong, G. Design and operation of biological
treatment for coke plant wastewaters. Carnegie-
Mellon Institute, 1978.
17. 'Neufeld, R.D. Thiocyanate bio-oxidation kinetics.
ASCE, Journal Environmental Engineering. Division,
Vol. 107, No. 5, October 1981, p. 1035. '
1059
-------
TRICKLING FILTER EXPANSION OF POTW BY
SNACK. FOOD MANUFACTURER
Michael R. Morlinc^,
Manager, Wastewater
Frito-Lay, Inc.
P.E., Group
Development,
Samuel M. Frenkil, P.E. Technical
Manager, Environmental Systems,
Frito-Lay, Inc.
Paul Trahan, Superintendent, Killingly
Water Pollution Control Plant.
INTRODUCTION
One of Frito-Lay, Inc.'s largest corn and potato snack
food manufacturing plants was recently opened in
Killingly, Connecticut, about halfway between Hartford,
Connecticut and Providence, Rhode Island. The decision to
locate this plant in Killingly required the company to
install on-site primary treatment as well as an expansion
of the publicly owned treatment work's (POTW's) secondary
treatment system. This expansion consisted of
constructing a $1.6 million trickling filter and pumping
station. This paper describes the joint effort between
the town of Killingly and Frito-Lay to design these
facilities and obtain State approval to construct them,
and reports on system performance to date.
The construction of a trickling filter or a POTW
expansion is not unusual. The uniqueness of this plan
stems from the fact that this project was the first
industrially designed and constructed expansion of a POTW
in the State of Connecticut, and would essentially double
the capacity of the secondary system without a
comprehensive plant expansion.
1060
-------
There was initial doubt on,the.part of the regulatory
agencies concerning the success of this project ^ since a
vlnture of this kind had not previously occurred in the
-'Hate of .Connecticut. .AEter careful analysis of the
pfoiect, all agencies and parties approved this plan and
Se facilities were constructed. This paper will discuss
some of the unique features of this cooperative project
First, 'these facilities were generally designed to
industrial standards. Second, the contractor was selected
as a result of a public bid. Third, all the Processing
equipment was purchased ;by Frito-Lay. Fourth, _ the.
facilities were designed, "constructed; and placed into
operation under the supervision of Frito-Lay. Fifth, _no
federal funds were involved., Sixth, the POTW expansion
was financed be industrial revenue bonds.
project Scoping & Negotiations
The town of KillingLy, Connecticut is located in
Windham County in the northeast corner of the state. In
spring 1978, Frito-Lay selected Killmgly as the site of
one of its, new .plants, .Killingly offered convenient
access to the major Connecticut, Massachusetts, Rhode
island and New York markets/ a solid work force,,
acceptable environmental costs, and strong interest by the
town leadership in the project.
• One of the .more complex "issues to resolve was
treatment of Frito-Lay's 1 million gallon per day high
strength wasteloads. A number of options were studied
separately by Frito-Lay, and jointly with the Killmgly
Sewer Authority. Ultimately, a decision was made for
Frito-Lay to construct a full primary treatment plant
on-site, discharge the effluent to the sewer system, and
build an expansion to the, POTW. The back-up alternative
was for Frito-Lay to build on-site secondary treatment and
discharge to the sewer. This would be done if Frito-Lay
and Killingly could not reach agreement, unexpected
technical problems arose,, or if costs became
uncontrollable.
1061
-------
The Killingly Sewer Authority operates an 8 million
gallon per day secondary treatment plant. The POTW was
completed in 1976 with the intent of providing substantial
reserve capacity in order to provide for future population
growth and to attract industry. This plant has a raw
sewage lift station, primary settling tanks, a
conventional activated sludge system, secondary settling
tanks, and chlorine contact tanks. Solids handling
consists of thickening both primary and secondary solids
prior to dewatering on rotary vacuum filters. Sludge is
lime treated, then landfilled. An on-site .sludge
incinerator will be used as back-up, or when incineration
of large sludge volumes becomes economical. Frito-Lay's
discharge to the POTW would consist primarily of soluble
BOD, and some solids. However, even with full primary
treatment of Frito-Lay's wastewater, the POTW could not
handle the wasteload. Secondary BOD was the problem. An
expansion to the BOD capacity might provide adequate
capacity. A study was initiated to investigate such an
expansion.
If a full secondary treatment system were to be
installed on-site by Frito-Lay, not only might it cost
more than an expansion of the POTW but it would result in
the duplication of equipment already installed at the
POTW. In addition, Frito-Lay would have to cope with
secondary sludge disposal. Costs were believed to be
compatable or significantly lower, if the POTW was
expanded. Therefore, Frito-Lay agreed to design and
construct a secondary pump station, a biofilter tower,
instruments and controls, sludge piping and related pumps,
valves, necessary appurtenances, and modifications to
appropriate existing components at the POTW.
The proposal to take an essentially new and under-
utilized POTW, expand it, and change the process concepts
was received with some skepticism by both the Killingly
Sewer Authority and Frito-Lay's management. Major
negotiating and paperwork nightmares were envisioned.
Project delays would impact Frito-Lay's start-up.
1062
-------
Regulatory authorities were expected to question .the
project, even though no change in the NPDES permit limits
or requests for federal funds were involved.
POTW Inprovements
Parallel to examining these procedural questions, the
Killingly Sewer Authority, the town's consultants, and
Frito-Lay's engineers began unit-by-unit analyses of the
POTW. Numerous meetings and iterations were involved.
The final result was a decision to expand the POTW by
adding a pump station, a packed media trickling filter
tower, and assorted controls. The POTW would be converted
from activated sludge to a trickling filter/roughing
filter process. The BOD capacity would be almost
doubled. All other systems were analyzed to be adequate.
Long term marginal capacity in clarifiers and flotation
units might be a problem if the city and Frito-Lay both
reached ultimate capacity. However, since this event is
anticipated to be at least 10 years in the future, work on
these systems was deferred.
The selected option called for converting the POTW
from a conventional activated sludge system to a
biofliter/activated sludge system. These improvements to
the Killingly POTW almost doubled the BOD capacity of the
secondary system (from 8,000 to 15,700 Ib/day) and
retained its current peak hydraulic capacity of 24 mgd.
These benefits were achieved by constructing a minimum
number of units. However, because the original plant was
not designed with this type of expansion in mind, many of
the connections and modifications were difficult and
expensive.
The pump station utilizes five constant speed,
submersible pumps to lift the primary effluent to the top
of the biofilter tower. Capacity of the pump station
varies 4,700 gpm with one pump running, to 17,000 gpm
under peak flow conditions. This allows hydraulic wetting
rate to be varied from 1.2 to 4.2 gpm/sf of surface area
1063
-------
of the biofilter. The puinp station utilizes simple
controls consisting of On/Off switches coupled to low-
level and high-level sensors. One of the five pumps will
be connected to the treatment plant's standby generator so
that process integrity is maintained in the event of power
outage. The number of pumps in operation will vary
according to flow conditions, but under peak design flows
only four pumps will operate with the remaining pump held
in reserve to comply with State regulations.
The discharge manifold, gate valves and check valves
are contained in a vault adjacent to the pump statioti.
Access to this vault is via two large doors. The pump
station was constructed of reinforced concrete. The wet
well and effluent channel are open but are protected by
handrails. An overhead electric hoist was provided to
facilitate pump maintenance. All the pump starters and
controls are located in a NEMA 14 enclosure, installed
between the pump station and the biotower.
The biofilter tower is approximately 70 feet in
diameter and 28 feet tall. The biological growth media
consists of horizontal redwood slats. The media, is
supported by pressure treated fir stringers which are
supported by a concrete underdirain system. Media depth is
21.5 feet resulting in a net media volume of 80,000 cubic
feet. The tower walls were constructed of precast
concrete panels trimmed with face brick to match the
existing plant structures. The design loading is 200
pounds BOD per 1000 cubic feet and 65% removal efficiency
is projected. In sizing the tower, it was assumed that no
additional BOD from Frito^Lay would be removed across the
existing POTW primary treatment system.
Wastewater is uniformly distributed over the media by
a four-arm rotating distributor. As the wastewater
splashes through the biofilter, bio-solids (microbes and
bacteria) form on the media and reduce the BOD of the
primary effluent. Bio-tower effluent is returned via the
10P4
-------
underdrain system to the pump -station. A broad-crested
weir divides this flow with a portion being recycled
through the tower to "seed" the process at a typical
internal recycle rate of 3:1 to 5:1. This serves to
increase removal efficiency and dampen shock loads. The
remainder of the bio-tower effluent flows by gravity to
the aeration basin (existing) for further treatment.
In order to convert the Killingly POTW from an
activated sludge to a biofilter/activated sludge system,
substantial improvements, additions, and renovations were
made to existing components. More than half, the cost of
the expansion project involved such "remodeling" items as:
0 Removal and relocation of major slide gates,
valves, and piping connections (30 and 36-inch).
0 Removal of existing primary to aeration basin
piping, and routing a 24" pipe to the new pump
station. Several other pipes were rerouted to
achieve this.
0 Core bore gallery walls to allow for 36" piping
extensions to the new pump station, and then
reconstruct the walls.
0 Route 36" return pipe from biofilter through
existing pipe gallery to two aeration basin
division boxes, including 36" valves and other
miscellaneous connections. Stainless steel pipe
.was used to facilitate installation in very tight
quarters where future maintenance would be
difficult.
0 Extend and reroute various sludge piping systems to
the new pump station. Stainless steel pipe was
also used here.
0 Repair all damaged or modified basin and tunnel
walls to be fully watertight.
1065
-------
0 Upgrade electrical system to handle the increased
motor load and install such conduit runs, starters,
controls, etc. as were required for the new systems.
0 Remove and extend existing flood plain retaining
wall, drainage system, paving, fencing, and rip-rap
to provide space to install biofilter and pump
s tation.
0 Place 5,000 cubic yards of select structural fill
on which the biotower and pump station were
constructed.
CONSTRUCTION OF THE EXPANSION
Since construction required difficult tie-ins and the
POTW had to continue in operation, contractor selection
was critical. During the design period, approximately 75
contractors were interviewed,. From these, 25 contractors
prequalified and were invited to bid. Many of the
contractors were concerned about the risks of working for
a private company (Frito-Lay) to construct facilities on
public (Authority) property. Two actions were taken to
address these risks. First, the specifications _were
written in the Construction Specification Institute
format. They explicitly set-forth the General Conditions
which related to this project, especially as to how the
contractor was to interface with Authority personnel and
Frito-Lay. In addition, the specifications contained the
contract to be executed between the successful contractor
and Frito-Lay. These details allowed many questions to be
answered about the contract before bids were received.
Secondly, a comprehensive pre-bid meeting was
conducted on-site. The plans, specifications, and site
were thoroughly reviewed. All present reached agreement
on'the best way to handle specification addenda. Finally,
it was announced that the contract would be awarded to the
qualifying low bidder at the bid opening if all paperwork
was in order and the bids contained no exceptions. There
1066
-------
would be no lengthy delays, "backroom negotiations" or
pressure to lower bids. We hoped to eliminate
contingencies, padding of the bids, and project delays.
These actions resulted in the receipt of 12 lump-sum
bids. Four of them were under the engineer's estimate.
One contractor was disqualified for failure to provide a
bid bond. Within 30 minutes -after'the last bid was opened
and qualified, the contract was awarded to the low bidder,
R.H. White Construction Company of Auburn, Massachusetts.
Frito-Lay and Killingly were committed to making these
facilities work as designed and on .schedule. This.
required the pre-selection and pre-purchase, and expedited
delivery of all the process equipment. Frito-Lay directly
purchased the .biotower . media, rotary distributor,
submersible pumps and controls, all valves over 12 inches,
and the flow meter and instrumentation. The detailed
design and specifications were completed after these
purchases were made so that more exact details were worked
into the plans. There would be no "or equals" or
substitutions that might cost more to install, lead to
contractor bid contingencies, or not meet the process
requirements. As a result the contractor knew before
submitting a bid exactly what equipment had to be
installed and when it would arrive on-site. Finally, all
construction was supervised under the direction of
Frito-Lay, utilizing the field engineer on-site at the
production plant (4 miles away) and a representative of
the consulting engineer.
These actions resulted in a high quality project that
was completed on time and within budget. Killingly Sewer
Authority personnel participated in every phase of the
process design, equipment specification and selection, and
in construction coordination. Their involvement from an
operating standpoint was vital. Although designed to
industrial standards, the facilities augment the
architecture of the POTW and they are completely
acceptable to the Authority.
1067
-------
PROJECT FINANCING
The final unusual feature of this project is the
method of financing the improvements. Due to funding
priorities and the industrial nature of the POTW capacity
expansion, federal funds were not available. A decision
was made to obtain pollution control industrial revenue
bonds (iRB's) through the Connecticut Development
Authority. In total, $6 million in bonds were obtained to
pay for various pollution expenditures associated with the
new plant project. The $2 million allocated to the POTW
expansion could only be approved if Frito-Lay, Inc.
continued to own the improvements for the twenty-five year
life of the bonds. After considerable negotiations and
legal opinions, a contract was executed that allowed
Frito-Lay to own the facilities and depreciate them. The
Killingly Sewer Authority is totally responsible for
Deration maintenance and repair. By utilizing IRB's,
Frito-Lay reduced the annual interest rate to 6.3% from
the 10-11% rate prevailing at the time of the closing on
the bonds. Abailability of IRB financing encouraged
Frito-Lay to proceed forward with the POTW expansion
program in Killingly.
SYSTEM PERFORMANCE
The trickling filter was placed in operation during
April, 1981. The biotower acclimated quickly, even though
it ^was started in the roughing filter mode with no seed
addition. In mid-August, the decision to convert to the
ABF1 mode was made in order to obtain warm weather
operating data before the onset of winter. A license to
operate the Activated Bio-Filter (ABF) process was issued
by Neptune - Microfloc, Inc. when the redwood media was
puchased. Table 1 summarizes system performance to date.
The roughing filter performed as expected. However,
the biotower was designed to operate in the ABF mode.
Starting in August, the removal efficiencies increased
dramatically from 65 percent to a high of 75 percent in
January, 1982. Also in January, the removal rate reached
52.8 pounds BOD per 1000 cubic feet. Average wastewater
1068
-------
temperature
January.
declined from 21.0°C in August to 11.5°c
in
Table 1. Summary of System Performance
Month (Ib
June , 19 81
July
August
September
October
November
December
January, 1982 .
Loading
Rate
BOD/1000 CF)
47.6
48.7
46.8
50.4
72.1
42.5
59.1
70.3
Remova 1
Rate
(Ib BOD/1000 CF)
24.9
22.8
30.9
33.7
47.1
27.1
41.3
52.8
Remova 1
Efficiency
(Percent)
52.4
46.8
65.9
66.9
65.4
63.8
69.8
75.2
During the period August through November, biotower
internal recycle was kept to a minimum and the return
activated sludge rate applied to the tower was 100% of
plant influent flow.
Shortly after start-up, there was a marked improvement
in the effluent suspended solids discharged by the POTW.
In addition, solids thickening operations improved. Most
.importantly, sludge dewatering costs decreased from a high
of $57.33 per ton to a low of $19.30 per ton. Table 2
summarizes vacuum filtration chemical conditioning costs
during the study period.
. Table 2. Vacuum Filtration Chemical Costs
PERIOD
January through April, 1981
May through August, 1981
September through January, 1982
DRY TONS
DEWATERED
329
385
471
AVERAGE
COST/TON
•* 43.15
27.06
28.97
10F9
-------
During this time, there was a 15 percent increase in
the quantity of dry solids dewatered but a 33 percent
decrease in the chemical conditioning costs per ton of dry
solids. These cost savings amount to $15,500 annually.
Less time is required to process these solids. Labor and
power savings are also substantial.
Conclusion
Based on operation to date, the biotower is achieving
the design removal efficiency. Overall, the POTW
operation has improved. Average effluent TSS and BOD
values are lower now than before Frito-Lay started
production. There have been few permit violations, these
being minor suspended solids excursions above the 30 mg/1
level. Settling characteristics of the secondary solids
have improved. m fact, it costs substantially less now
to dewater sludge than two years ago in spite of inflation
and even though sludge quantities have increased
substantially.
The Authority is sufficiently pleased with the
biotower facility that Frito-Lay, Inc. was able to
negotiate an expansion of production. in fact this
expansion is currently underway and will make the
Killmgly plant Frito-Lay's largest.
1070
-------
THE EVALUATION OF A BIOLOGICAL TOWER FOR TREATING
AQUACULTURE WASTEWATER FOR REUSE
Gary L. Rogers. Department of Civil Engineering
Brigham Young University, Provo, Utah
Stanley L. Klemetson. Department of Civil Engineering
'BrTghattTYoung University, Provo, Utah
INTRODUCTION
During the 1950's, Dr. Shao-wen Ling was sent to
Southeast Asia by the Food and Agriculture Organization (FAO)
as a fisheries specialist. He became interested in a large
Malaysian prawn which he first observed in the marketplace.
Since that time there have been many species of freshwater
prawns identified that are distributed over the tropical and
semi-tropical waters of the world. Most are limited to
waters that maintain annual temperatures of 22 to 30 C. The
most popular species for culture is Macrobrachium rosenbergii
(1).
Adult prawns are capable of living in freshwater and may
be either intensively cultured in tanks or raceways at high
stocking density or extensively cultured in earthen ponds at
a lower density. Marketable size prawns may be harvested in
five to six months under optimal conditions. Glude (2)
presented an overview of freshwater prawn culture that
describes rearing requirements and technques. Figure 1
illustrates the adult prawn showing characteristic parts.
1071
-------
The development of prawn larvae to metamorphasis
requires temperatures of between. 24 and 32 C with preferred
temperatures around 28 C. These temperatures also apply for
commercial growth of adults in ponds.
Adults and broodstock may be maintained easily in either
freshwater or brackish water. The larvae require brackish
water of between 8 and 17 ppt salinity. Juveniles require
salinities of about 5 to 8 ppt decreasing to freshwater as
they grow to adults.
The toxicity of ammonia to aquatic life has been fairly
well documented. The EPA criterion of unionized ammonia for
fish and aquatic life is less than 0.02 mg/1 (3). The 144
hour LC 50 for Macrobrachiuni rosenbergii was determined by
Armstrong (4), to be 0.80 mg/1 ammonia at a pH of 7.6.
Additional work is necessary to verify and establish these
toxic limits of unionized ammonia for the Malaysian prawn.
Nitrite has also been shown to be toxic to freshwater
prawns (5). LC 50 values ranged from 500 mg/1 for the first
12-hour exposure to 5 mg/1 during a 168-hour exposure. The
maximum level of nitrite tested with no deaths, ranged from
9.7 mg/1 for 24 hours to 1.8 mg/1 at 168 hours.
There is a paucity of data concerning the concentration
and nature of wastes produced in intensively cultured
systems. Little has been presented about water quality of
effluents from culture systems. The problem is coraplexed by
the fact that organics and nitrogen in the water may be due
to metabolic products of: prawns and other secondary feeders
in the culture ponds as well as unused feed added daily to
the water. A summary of the reported water quality values
and the proposed levels for design of new production
facilities is shown in Table 1.
Several investigators have described water recirculating
systems using biofilters for intensive cultivation of salmon
smolts (6,7,8), catfish (9), trout (10,11,12,13), shrimp
(14,15,16), tilapia (17,18), polyculture (19), carp (20),
and the combination of fish culture with hydroponics in a
single recirculation system (21). Few of the authors,
however, discuss the basis for their choice of biofilter
design parameters. There is little indication that the
designs conform either to economic or to resource minima for
their declared purposes. Spotte recommends that surface area
of the filter be made at least equal to the surface area of
the culture.' This rule of thumb may favor successful
recirculation water quality but says little about keeping the
investment minimal (22).
1072
-------
Removal of the toxic hazard from nitrogenous waste is
the most important task of the biofliter. The predominant
form of waste nitrogen from aquatic animals is ammonia,
excreted mostly from the gills or in the urine, or produced
by mineralization of organic nitrogenous substances by
bacteria. Invertebrates also excrete nitrogen in the form of
ammonia. Bacteria of the Nitrosomonas group oxidize ammonia
to nitrate;
NH + OH 4- 1.5 0
2 H + NO + H 0
2 2
the reaction is rate limiting in the overall process of
nitrogen control. For this reason, nitrites seldom
accumulate in biofiltered recirculation water. Nitrite is
readily oxidized to nitrate by Nitrobacter ssp.;
NO + 1/2 0
2 2
= NO
3
and the nitrate is then available either for assimilation by
green plants or bacteria, or for reduction to nitrous oxide
and free nitrogen, usually by anaerobic denitrifying
bacteria.
The microbial population activity in the filter is
essentially ubiquitous in distribution, spontaneously
coloniing new filters, although inoculation can be expedited
by addition of gravel from preexisting filters (8). The
bacterial species which predominate at various depths of the
filter are presumably self-selecting, in response to
environmental factors, such as the nutrient content,
hydraulic flow, and aeration of the filter bed, although
their respective numbers are known to fluctuate until an
equilibrium is established (Kawai, et al., 1964).
Few approaches to analysis of design standards for
biofilters in aquacultural applications are given in the
literature. Hirayama reported on filter carrying capacity
for a 300 liter marine aquarium containing one or two sea
breams, Chrysophrys major (23). Spotte cautions that
Himayama's results need verification before accepting their
validity in freshwater, but there is reason to exercise a
certain amount of skepticism in generalizing them even for
marine aquaria (22). One should keep in mind the small
numbers of fish on which they are based.
1073
-------
Harris considered use of biological filters in hatchery
water reuse systems. Submerged filters (Flexring and Flocor)
were tested and found satisfactory in control of fish wastes.
The basis for design of the biofilters used was Speece's
method for finding substrate volume based on ammonia
production rate divided by the nitrification rate (10).
A. major limiting factor in intensification of
aquaculture is water quality. Biofilters need to be included
in aquaculture systems to insure pollution removal and good
quality water for use and reuse. Work is being done on
optimizing treatment systems for aquatic animals and
developing accurate design specifications for treatment.
More research is needed in this area.
Management of the filter is an important consideration
in obtaining nitrification and BOD removal from either fresh
or saline aquaculture wastewater. Removal rates may be
enhanced by increased temperature, recirculating effluent
back over the filter media, dosing the filter, ventilating
the filter, adding additional media and surface area, and
maintaining environmental requirements by careful monitoring.
Wheaton (24) describes in depth design criteria for
submerged filters with varied filter media. The equations
presented for design are restricted to cold-water
applications and care is advised in extending to treatment
systems for treating wastewater produced in warmwater fish
culture facilities.
The types of media that have been applied to aquaculture
include oyster shell, rock and gravel, plastic rings, poly
beads, sand, and styrofoam. A new possibility is the Tri-
Pack spherical media available from Jaeger Tri-Packs in
Costa Mesa, Calif (technical bulletin, 1981).
Trickling filters and upflow submerged filters may be
designed according to a simplified procedure presented by
Soderberg and Quigley (25). Their data is for perch culture
and probably should not be generalized for culture of all
warmwater species of fish and invertebrates.
This study is a joint effort between Dr. Stanley L.
Klemetson currently at Brigham Young University and Dr. Dan
Cohen of the Hebrew University. The U.S. portion of the study
deals mostly with engineering problems while the Israel
portion deals with production of the animals. It has been
important to coordinate efforts and share information to
improve each study.
1074
-------
METHODOLOGY
The filters used in this study were laboratory-scale
units having a capacity of 1.4 cu. ft. each. Each filters
was filled with 1 to 1.5 inch slag with a total surface area
of 82.0 sq.ft., and a specific surface area of 60.0
sq.ft./cu.ft.
Figure 2 presents a schematic of the two filters used in
this study. The facilities at Brigham Young University were
used to construct the colunms. Both filters were fabricated
from a section of 10 inch PVC pipe. The end discs were cut
from plexiglass.
The two filters were fed a uniform flow from a 4.5 cu.
ft. plexiglass constant head tank with overflow weir (Figure
3). Two 4.0 ft (0.11 m ) storage tanks filled with
synthetic wastewater was used to feed the constant head tank.
The two storage tanks were used to hold the synthetic
wastewater to be treated by the system. The synthetic
wastewater was compounded as needed from a balanced minimal
media which approximates actual wastewater produced by
prawns. Table 2 presents the composition of the synthetic
feedstock solution. Diffused air was introduced to insure
adequate mixing within the storage tanks. Tap water was used
to fill the tank to Its desired level.
The studies done at the Hebrew University in Israel
utilized 900-liter aquaria and submerged filters. The
submerged filters were filled with 20-30 cm gravel in a 700 1
tank. The flowrate was controlled at 0.5 cu. m./ hr.
Salinity of the rearing tank was maintained at 1.2% sea water
for growth of freshwater prawn larvae.
Water quality sampling was done prior to feeding and
drainage of sediments for the Israel study. Standard Methods
(26) were used to determine dissolved oxygen, biochemical
oxygen demand, pH, alkalinity, salinity, ammonia, nitrate,and
nitrite both at the Hebrew University and at Brigham Young
University. Water samples were taken at both the influent
and effluent of the biofliters.
RESULTS
The results of work done at Brigham Young University are
presented in Table 3. After acclimation of the biofilters,
data was collected to evaluate their performance. Profiles
of the parameters monitored are presented in Figures 4
1075
-------
through 7 for the Submerged and Trickling filters. Figure 4
presents a profile of ammonia levels. The average value of
ammonia removal for the trickling filter was 78%. In the
case of the submerged anaerobic filter, there was an average
increase of ammonia by 8% through the filter.
Figure 5 presents the nitrate profile for the two
filters. The nitrate concentrations decreased by 97% in the
submerged filter while there was an increase observed in the
trickling filter. The length of time required for start-up
of the filters was 3-4 weeks. At three weeks, the trickling
filter was effectively nitrifying the synthetic wastewater.
The submerged filter was anaerobic and functioning as a
denitrifying filter in about the same length of time. Other
parameter profiles are presented in Figures 6 and 7.
Data describing the results of the Israel study are
presented in Table 4 and 5. The data indicates that the
nitrification efficiency of freshwater prawn wastewater was
59% ammonia removal and 55% nitrite removal. Figure 8
presents the nitrogen profiles for a tank and biofilter
system in Israel. The filters were capable of maintaining a
stable environment, though further study is necessary to
determine sizing criteria for filter design.
The results of these studies are being reviewed as part
of an ongoing study to develop design criteria for
application to treatment of warmwater aquaculture wastewater.
The data obtained in this portion of the study suggests that
nitrifiction efficiency can approach 80% when the synthetic
wastewater is lightly loaded. The optimum hydraulic loading
has not been determined. The efficiency of ammonia removal
in brackish water is about 60%.
An important consideration in maintenance of the
biofilter is aeration. If aquaculturists are interested only
in nitrification, careful monitoring and addition of
dissolved oxygen is important. Denitrification is an
anaerobic process that requires little maintenance.
Aquaculturists should consider the coupling of the two
processes, especially where the recontioned water is returned
to the culture vessel.
DISCUSSION
The design and sizing of biofilters for application to
aquaculture is more of an art than a science. If facilities
are to be developed for intensification of fish culture, a
great deal of research needs to be done. Design criteria are
1076
-------
available for facilities that rear cold-water fish and
invertebrates but is limited to water temperatures up to L5
C. Since operating temperatures are high (27 to 30 C),
organic loadings are small, and flowrates are high, standard
design tables are not adequate. To complicate matters more,
the actual waste production rates for the live aquaculture
systenus are variable and ill-defined. Optimization of design
will require the continued evaluation of equipment and
processes under a variety of conditions.
ACKNOWLEDGEMENT
The authors wish to express their thanks to the B.t-
National Agricultural Research and Development Fund Grant No.
US-60-80, for providing a research grant, for support of this
project, and to Aquaculture Production Technology (Israel),
Ltd.
REFERENCES
Hanson, J.A., Goodwin, H.L., Shrimp and Prawn Farming In
The Western Hemisphere, Dowden, Hutchinson, and Ross,
Inc., Stroudsburg, Pennsylvania, p.198 (1977).
Glude, J.B., The Freshwater Prawn Macrobrachium
rosenbergii, consulting report, Jan 1978.
United States Environmental Protection Agency, Quality
Criteria for Water, USEPA, Washington, D.C.,July (1976).
Armstrong, D.A., Chippendale, D.v Knight, A.W., and Colt,
J.E., Interaction of Ionized and Unionized Ammonia on
Short-term Survival and Growth of Prawn Larvae,
Macrobrachium rosenbergii, The Biological Bulletin, Vol
154, p. ffrn7~YeVT9T8T~
Kawaratani, R.K., State of the Art: Waste Heat
Utilization for Agriculture and Aquaculture, Tennessee
Valley Authority, August (1978).
Burrows, R.E., Controlled Environments for Salmon
Propagation, Prog. Fish Culture, 30(3):123-136, 1964.
Meade,T.L. The Technology of Closed System Culture of
Salmonids, Animal Science/NOAA Sea Grant University of
Rhode Island, Mar. Technol. Rep. 30.
Risa, S. and H. Skjervold, Water Reuse System for Smolt
Production, Aquaculture, 6:191-19.5, 1975.
1077
-------
9. Broussard, M.C. and B.A. Sumco, High Density Culture of
Channel Catfish in a Recirculating System, Prog. Fish
Cult., 38:138-141,1976.
10. Harris, L.E., Nitrifying Requirements of Water Reuse
Systems for Rainbow Trout, Colorado Division of Wildlife
Special Report, No. 41, Feb 1977.
11. Fyock,O.L., Nitrifiction Requirements of Water Reuse
Systems for Rainbow Trout, Colorado Division of Wildlife
Special Report, No.41, Feb 1971.
12. Mayo, R.D.,A Technical and Economic Review of the Use of
Reconditioned Water in Aquaculture, Aquaculture, pp.508-
520,1976.
13. Speece,R.E., Trout Metabolism Characteristics and the
Rational Design of Nitrifiction Facilities for Water
Reuse in Hatcheries,Trans. Amer. Fish. Soc.,102(2):323-
334,1973.
14. McSweeny,D.S., Intensive Culture Systems, Edited by J.A.
Hansen and H.L. Goodwin, In Shrimp and Prawn Farming in
the Western Hemisphere, Dowden, Hutchinson, and Ross,
Inc. Stroudsburg, Penn. pp.255-272,1977.
15. Mock,C.R., Ross, L.R., and Salser,B.R.,Design and
Evaluation of Waste Removal Systems for Shrimp Culture in
Closed Raceways,World Mariculture Society 6th Ann.
Workshop, Jan 1975.
16. Siddal,J.,Studies of Closed Marine Culture Systems, Prog.
Fish. Cult.,36(1):8-15,1974.
17. Otte,G. and Rosenthal,H., Management of a Closed Brackish
Water System for High Density Fish Culture by Biological
and Chemical Water Treatment,Aquaculture,18:169-181,1979.
18. Allison,R., Rakocy, J.E.,and Moss,D.D.,A Comparison of
Two Closed Systems for the Culture of Tilapia,Presented
at the International Symposium for Advance in Food
Producing Systems for Arid and Semi-arid Lands, Kuwait
City, Kuwait,April 1980.
19. VanGorder,S., Small Scale Fish Culture Systems, Rodale
Press Research Report 80-12, May 1980.
20. Meske,C.H., Fish Culture in a Recirculating System with
Water Treatment by Activated Sludge, Aquaculture,pp.527-
532,1976.
21. Lewis, W.H.,Yopp,J.H., Schramm, and Brandenberg,A.M., Use
of Hydroponics to Maintain Quality of Rec.irculated Water
in a Fish Culture System, Trans. Am. Fish. Soc., 107:92-
99,1978.
22. Spotte,S.H., Fish and Invertebrate Culture, Water
Management in Closed Systems,Wiley, London, 145 pg, 1970.
1078
-------
23. Hirayama,K., Studies on Water Control by Filtration
Through Sand Bed in a Marine Aquarium with Closed
Circulating System,IV. Rate of Pollution of Water by Fish
and the Possible Number and Weight of Fish Kept in an
Aquarium,Bui. Japan Soc.Sci. Fish., 32:27-30,1966.
24. Wheaton, F.W., Agricultural Engineering, Wiley
Interscience, New York, 1977.
25. Soderburg, R.W. and Quigley, J.T., The Technology of
Perch Aquaculture, University of Wisconsin Sea Grant
Program, WIS-SG-77-416, 1977.
26. APHA,AWWA,WPCF,Standard Methods for the Examination of
Water and Wastewater, 14th edition, Washington, D.C.,
1976.
1079
-------
Figure 1. Freshwater Prawns (Macrobrachium rosenbergii) and
Artificial Habitats in Fish Culture Tank.
1080
-------
8 in. PVC
1-T/2 in.
slag
sampling ports
Figure 2. Schematic Diagram of Biological Tower and
Submerged Filter (BYU-Study).
1081
-------
Figure 3. Constant Head Tank Used to Dose Biological
Tower and Submerged Filter (BYU Study).
1082
-------
10
10
5 I
A
Vlnf 1 uent
Effluent
Day
B
Effluent
Inf1uent
Day
Figure 4. Ammonia Removals in Biological Tower (A) and
Submerged Biofilter (B).
1083
-------
10
Effluent
en
§
10
Day
Day
'B
Figure 5. Nitrate Removals in Biological Tower (A) and
Submerged Biofilter (B) .
1084
-------
-------
4-1
•rl
Submerged Fi1ter
•5200
-------
(1/6UJ) ir-.-T £QN
--— \\'
•—
cd
M
CD
cu
g
60
•u
Ctf
rH
O
•H
a
0)
rH
•H
P-i
g
00
O
4-1
•H
OO
CU
60
1087
-------
af
•o
§
C-,
o
jl
3
CJ
3
cr
*3
&.
o
^»
1
o-
.3
a
5-
U*
O
w
o
•a
o
t*
o
*
1
•3
0
•o
o
CO
1
CO
a
a
a
O
o
o
"3
to
.c
Z
•o
£2
c
P"-t
»rf«
O
j3
t.
u
•8
C-,
w
CO
p
t=.
o
•*-*
u
E
Cj
L*
RJ
a
co • t_ S
r_j Q} ' C"S CD CD ^Q
V V AVVVVV/vvVV V V V V V'VVV'V
o
CQ -<3< ^ CO | CD CO
T-HOOJ C3 C3CDCOOrH 'IOO CD UO^
C3 O O O O^OOO CD^ CQ CQ^
C) ^
, •
CO ^^ 5>P
in CQ | 01 o^
CQrHCD 10 CD "^g^ ' « CD S O ^
O O O C3 C3CDCOC3 CO lOi-H COCD
CD o
C-3 CO C3 C*3 C^ vO
O pq CD CD CD O "^ ^*l" ^ CD T— 4 CD CD ^
CD 1/3 CM CD m »— I CDCOCD CJ T— 1 ^ CO CD
CD IO
CQ CO C^3 pP
C*3 | IT3 T-* CO in | CD ^
CD *^J* ^3 CD CD CD i — 1 IO CD ^^
CDCD CD CD CDCD CDCOCD i-HCO
CQ ^
*-4 O5 C-3
C^ 1 I CD 0
^ °J '-J -^r m S S CD "^ S o in
O *— t CD i-H t— t-H i— 1 t-H «D O CJC^
"•8
• 1 3 S 1
c -2 ° w
0 w ^ ^T? TJ Ti'~s
*O &9 x-^-u-— CJ^-a*-3
*X x *3 c:ojc^^cf~1
c — ^ p ±i £GJ c O^S'010 ^^^
^C *^ U O CJ CJ CJ CJ CJ C-5 i^ D^ , 1 »^i «n rC p^ ^ pir y^ y^ [/3 ^M f-i £— i f-i £— i f-i ^
'§
?g
c >>
4-)
w U3
0.0-
a)
K «
03
"ho
e-"|
c °> ,2
•^ rH Q
rn -C
IfS
o £ j=
i ^ ^
g^^
£< »"
o CM t
CJ W 0
J2
S CO O
•< D es
CS -O CJ "
Cn
t-
"3
CJ
o
o
CJ
CO
w
E
o
£_.
"3
CJ
o
•c
o
o
o
"3
o
o
5"
CO
t"-
c
Cw
;t
o
CJ
c
G
c
1-1
0
^
(t
cl
o
c
bi
ei
Q
"3
_CJ
'5
o
"o
S
o
•J^
5
E
a
o
CJ
CJ
p
aT
1
bp
£
1) <.
CT)
in
rM
_
"o
o
XI
o
CJ
6
CO
>
"3
Gf
^
2J
^j
oT
•a
o
3
O
4~>
ronmen
>
c
i)
1088
-------
Table 2. Synthetic Wastewater Solution (BYU study),
DEXTROSE
YEAST EXTRACT
UREA
NA2HPOi}
NAHC03
100 M6/L
10 MG/L
5 MG/L
40 MG/L
125 MG/L
2 MG/L
AS REQUIRED
1089
-------
/**\
3
4J
CO
1
*s^
4J
ft)
o
^
4J
3
o-
.^
0)
4J
££
!E
a)
£5
I
CO
0
S3
CO
j-—
1
CM
O
53
rt
CO
§
,
CO
CO
CM
|
VO
VO
r-(
CM
*
vO
i
o
t— i
CM
OO
in
•
^
CTv
,— 1
,
m
0
j>^
in
•
CO
o
a\
CO
4-1
a
a)
3
i-l
H
s~\
VO
•
[^
o
CM
\^r
x-s.
oo
oo
,
CO
v-x
oo
OO
v-'
y™\
00
CXJ
1—1
V_^
y~ X
oo
»
o
\**s
s~^
r-t
VO
v«^
O
•
o
I— 1
CM
I
CM
VO
t— 1
CO
•
VO
1
CO
I— 1
•
00
-sf
•
l-s.
CO
in
i— )
i — i
1
•
.— i
00
,
CM
1
m
•
o
m
I
CM
m
•
o
60
•rl OJ
iH 4J
0 -rl
•rl P4
!-l
H
s~*
in
r--
•
oo
P*"»
T— t
V.^
^"N
O
CO
•
in
^s
/-^
i-H
CO
r~
>s-/
PS,
— /
X— s
VD
r-H
' — ^
4J
a)
,_ ]
14-1
m
w
oo
•
VO
CO
CM
1
vO
OO
i— 1
«sT
•
.—1
1
CM
0
i— 1
00
00
r~
vO
CT»
•
o
i
i— i
0
0
oo
CM
0
1
1 — 1
o
o
•
o
.-H
1
oo
m
13
a) ^
ta
-------
cO
4J
ni
P
>>
4-1
•H
rH
nj
o*
f-l
Q)
4J
ni
13
id
^
PM
"4-1
O
to
§
£
*O'
•s
CD
0)
00
-*
OJ
•s
^f
1
CO
o
^*«
1
CM
CD
•^r
I
CO
' .
—f"
CD-
CD
o
2:
i — i
s:
1
^
s
Lf>
O —
• co
CD LO
CD "^
• — -
0
( — 1
LO
CD LO
• i — i
O r~^
1 O
O •
r^ o
- LO -
•
O
OQ
CO - — ^
o cn
. .— i
O CM
i
f~- o
°'~
CM- — •
• CM
co co
1 O
• CO
LO ,
CM-— ~
• LO
cn =»-
• • '
cn
CM
0
CD
CD cn
1 0
CO CD
CD -
• O
0- —
LO
0
CD- —
. CO
o cn
1 CD
CO •
CM O
CD
O
LO. — *
CO 1^-.
1 CM
LO •
CO CO
^
LO - — -
CM CO
• LO
CO >— 1
LO r~-
CO
LO
.—1
CO
CM
CM
CO
LO *-"^
O LO
• 0
0 •=*•
1 ~— •*
o
»— f
LO
0
o
•
0 —
1 CM
LO O
r--
0 0
, • '
CD
LO
CO
0
• ,-— *.
CD cn
i cn
LO CD
CO
CM O
d^
o
CO- — -
• CO
CO LO
1 *~^
CO •
cn co
^
-• — *
O CO
. CO
cn «— i
co r-I
LO
cn •-
CM
o
CO
LO
CO
LO-
O"ID
. CO
0 OO
1
CO -cf
« - — -
cn
LO
o
0
• — " "^
o r^
I CM
LO O
OO .
O CD
• ' — ••
o
LO
CM
O
• .. — -.
CD t^-
1 CO
CD CD
cn •
,-H O
O
O
CO
co
• -— *w
CO CO
1 CD
LO t— i
r^ •
o oo
od"
LO
• -— *-,
co cn
i •
O CO
LO
CO
CM
cn
CM
CO
CO
uo-
oco
• co
o »—<
i
LO *3"
• •
CD
LO
O
0
• f^^^i
O CM
1 CO
0 0
LO •
i— i O
• *- — -
CD
LO
CO
O
• - — *
CD ^J~
i cn
0 0
CO
CM O
°
O
CO
• ^— •.
CO CD
1 CO
LO <— 1
r**^ •
cn co
^
LO
CM - — ^
• CD
OO CO
1 <-<
CM •
LO r~~
LO
CM
cn
CM
I-H
co
- — !•
LO r~-
CD ^t*
• •
CD -<^-
1 •*• — -*
0
,— 1
LO
O
o
9 **~~*
O LO
1 CM
O O
cn •
O O
• •- — -
O
CD '."
CM ,
O
• * — ^
CD cn
i co
0 0
LO •
1-1 O
0
LO
CM
CO
• . — *
co r^-
i i~" ^
LO, •
•d- co
co*^"
o
t— * -~x
• oo
00 CO
i •
o co
LO
0
CM
oo"
CM
CM
CO
1091
-------
4J
rt
H
m
(0
OJ
Lf)
OD
+1
CO
CM
uo
oo
CT>
CM
ID
OO
+1
(£3
OO
CM
LO
CO
LO
oo
ID
00
co
co'
LT)
oo
CM
IX
1092
-------
BIOFILTRATION OF TANNERY WASTEWATER
Ahmed A„ Hamz a,
Fahmy M. El-Sharkawi,
Mohamed A0 Younis,
Department of Environmental Health,
High Institute of Public Health,
Alexandria University, Egypt
INTRODUCTION
Industrial pollution on an unprecedented scale: has emerged
as one of the most pressing problems throughout the world.
While governments have been grappling with a pernicious combi-
nation of economic, social and political problems, they have
not paid equal attention to issues related to environmental
protection,, Improving the quality of the environment is no ...
longer a luxury measure; it is one which, in the long run,
will generate immeasurable benefits in terms of protecting
public health and natural resources, and indirectly contribut-
ing to economic growtho
Alexandria is the principal port of Egypt, the country's
largest industrial center and its prime resort„ The ever-
increasing discharge of heavily polluted industrial effluents
from tanneries of the Mex Industrial Complex (MIC)into the Me-
diterranean has had an adverse effect on public health, fish
production, navigation and the environmental quality of the
area0 The combined effluent of MIC averages 2 million cubic
meters annually,with an estimated population equivalent of '.
400,0000
1093
-------
Biofiltration has been widely recognized as a reliable
treatment process, which suits the needs of small to medium-
size industries, due to its versatility, ability to take shock
loads and relative ease of operation. Decreased popularity of
biofiltration in comparison to other treatment processes has
been attributed to inability of existing installations to meet
emission limitations and the ineffectiveness of biofilters for
treatment of concentrated industrial effluents,,
Renewed interest in biofiltration in Egypt is due to rela-
tively low power requirements,'system flexibility, and the
need for considerably less technical know-how for effective
operation compared to other treatment technologies,,
High-rate biofiltration was shown by Smith and Kates(*)
to be cost-effective and capable of reducing organic loadingp
Several plastic media were tested in the high-rate biofilter
units and each medium supported an adequate level of microbial
organismso
Problems with odor were encountered because of the sludge
production of the high organic strength wastewaters0
Hosono and Kubota(2) reported that the BOD removal rate
per unit of power consumption was shown to decrease with in-
creasing BOD loadingso High-rate filters gave higher power
economy values at higher BOD loadings, whereas standard-rate
trickling filters were limited to low BOD loadings„
Bailey et al'3^ have shown that high rate biofiltration
can relieve overloaded conventional filters by removing about
40% of BOD from leather processing wastewater,, Plastic media
were used in the roughing filters which were dosed with tannery
effluents containing vegetable tanneries at rates of 3.3 to
6o9 Ib BOD/yd3/d0
Pierce'^) reported that removal of BOD at two-stage filter
plants is significantly higher than at single-stage plants„
Chemical treatment with metal salts and polymers upgraded
single-stage filter effluents from an average of 36 mg/1 BOD
to 21 mg/lo Similarly, effluent suspended solids were reduced
from 32 mg/1 to 19 mg/1. The cost of chemicals is not prohi-
bitively expensive„
Previous research has indicated the need to assess high-
rate biofiltration at various loadings and flow patterns for
treatment of specific industrial wastes, in order to accurately
evaluate system performance. It is particularly important to
compare the performance of biofilters with other competing bio-
logical processes for treatment of tannery wastes and evaluation
of cost-effective modifications to improve effluent quality,,
Recognition of these needs prompted the undertaking of this
study„
1094
-------
BACKGROUND
The MIC tanneries are located east and west of the muni-
cipal slaughterhouse, as shown in Figure 1. At present, the
slaughterhouse and tannery wastes (including organic particles
and toxic chemicals) are collected in public sewers and dis-
charged directly into the western harbour through three sepa-
rate outfalls, without pre-treatmento Sewer clogging is fre-
quently experienced due to large residues discharged with tan-
nery effluentSo V
The General Organization for Industrialization (GOFI) is
placing MIC tanneries on the top priority list of the most pol-
luting industries which require Government technical and finan-
cial support for installation of waste treatment facilities.
The available options being studied are: (a) primary treatment
of combined effluent before discharge into public sewers for
further treatment with domestic wastes or (b) biological treat-
ment to meet Egyptian effluent limitations for direct disposal
into water bodies„
According to prevalent practices in MIC tanneries, about
28-36 cubic meters of wastewater are generated per ton of hide
processed^) 0
Studies performed on the six major tanneries during 1980-
1981 indicated that clean water pools of the beam house contri-
bute 24o8% of the total effluent and only 0^.28% of the BOD load,
while vegetable tannery generates 104% of the liquid wastes
and 43o3% of the BOD load (Table I)0 This suggests that ju-
dicious segregation of relatively clean process waters may ap-
preciably reduce the size and costs of treatment facilities0
The pollutional loads of tannery processes shown in Table II
indicate that both beam house and tan-yard generate higher
loadings than those originated from retin, color and fat liquor
processeso
Chrometan mixed wastes comply with EPA guidelines^ ' while
BOD, COD and Oil and Grease (0 & G) loadings of the vegetable
tan mixed wastes were higher than those suggested in the guide-
lines o A summary of the physico-chemical characteristics and
trace metal constituents of various process effluents are shown
in Tables III and IV respectively,,
MATERIALS AND METHODS
The experimental system consists, of two biofilters, recy-
cling pumps and clarifiers, as shown in Figure 20 Each filter
1095
-------
« os coin
0} OJ
-------
Table 1. Distribution of Volume and BOD load foi the Processing
Operations of The MIC Tanneries.
Process Waste Volume % of Total BOD % of Total BOD
Volume Load '
A. Beam-House Wastes
Soaking & Washing
Lime & Unhairing
Delime 4 Baiting
Clean Wash Water Pools
B. Tan - Yard Wastes
Pickling '
Chrome Tanning
Pre - Tanning
Vegetable Tanning
C. Retan - Color & Fatliquor
Neutralization
Bleaching
Color S, Fatliquor
Total
Table II. Pollutional Loads of
Pollutional
Ei fluent Stream . BOD TOC
A. Beam-House Wastes
Soaking {. Washing 11.8 5.4
Lime & Unhairing 26.6 15.5
Delime & Baiting 16.5 8.2
Clean Wash Water Pools 0.4
B. Tan - Yard Wastes
Pickling 4.5 2.3
Chrome Tanning 7.3 3.7
Pre - Tanning 21.2 10.6
Vegetable Tanning 92.8
C. Retan - Color & Fatliquor
Neutralization 3.0 2.0
Bleaching 18.7 15.1
Color & Fatliquor 11.8 7.3
Chrome-Tan Mixed Waste 92.6
Vegetable-Tan Mixed Waste 150.8
EPA ( Ref.6) 95
8-3 5.4
24-3 12.4 .
20.0 7.7 .
24.8 o.2
77.2 25.7
5-5 2.1
5.5 3.4
2.0 9.9
1-4 43.3
14.4 58.7
2.8 1.4
218 8.7
2.8 5.5
8-4 15.6
100 . 100
the MIC Tanning Processes.
Load Kg/ K Kg * of hides
TKN SR 0 & G T Cr
0.74 5.9 2.5
1.64 51.8 2.2
5.3. 27.7 0.4
1.6
0.7 9.1 0.1
0.75 9.5 0.6
0.6 116' 0.4
0.9 8.5 0.05
0.5 51.2 0.3
1.2 6.8 60.5
:109.2 8.6 3.5
72.1 9.4 0.0
140 19 4.3
Mean of 12 observations.
* K/KKg •= Kg/1000 Kg.
1097
-------
x:
a
a
5
•5
o
1 -**
r> s
rH
)
r oo
c- B
rH
CO *^.
CMCO
X S
1 rH
O CO
W E
r-4
H CO
O 5
F-t
5 "M
-< E
S ^
5= B
^
o -•*
O 00
-. E
ca **^
o oo
0 E
rH
Q ^
O 60
pa B
OS ^
iJ *^.
Q) rH
CO E
r-H
Si 60
,Q
3 H
H SS
X O3
O. ^^
to
(0
0)
O
o
V-)
PU
CM m ON m
ON r- o co
vO CO vD rH
rH vo m ON
rH O CM rH
^
CO • VO CO
- CM O OO
r^> rH CO *O
co m o vo
o CM co m
ON ON -0- CM
rH rH CO rH
0 rH 0 CO
vO OO vD O
•fc±
O rH
yz o PS
S li
8 3g
CO
0 CO
rH VD
CO
^
ON m
CO CO
ON in
vo m
CM CM
O CO
O r-
03 vD
0 rH
rH
VO
vo r-*
CO rH
vo m
VO rH
CM -a-
CM rH
i-H CO
35
PH OH
CO -J i— 1 rH
mm o CM
rH rH
TO CM CO CO
O O CO rH
CO
0 .H
O O
r- r-
vO CO
CM rH
CO rH
co in co co
co co co -a-
vo r- CM co
-a- co oo
rH
r- ON m oo
O CO ON O
ON ON r*- CM
rH
O O CM
m vo O O
ON CO ON rH
CM vO OO CO
CM CO
t~~ "a1 r-» m
•a- vO VO rH
r^. co CM i —
CM -3- CM O
CM CM CM CO
rH
r~~ CM r-- in
vO ON vO CO
in *a~ »n rH
CO CM
P-. rH r- OO
mm CM vo
CM . CM i-H
O vD CM CO
ON vO r-» rH
in m CM *a-
CO CM
m ON
CM -a-
CM
1 1
rH ON CM vO
-a* m m
ix; Q ix Q
2i [X]
H ,-J
% §g
H UJ S
(X, > H
rH •
r-. m
o m
rH CM
CM CM
r* oo
vo o
m CM
CO
oo m
r* m
r- co
rH
r- in
vo ro
m CM
ON in
r^ -a-
vO vO
rH
r- r-
CM m
OO CM
CO rH
rH CM
rH
m m
vo oo
co m
O CO
O rH
tH
1
CO vO
vo r^
IX 0
en
tH
§§
O r-<
w
K \
m vo
CO OO
VD m
rH rH
rH CO
O
CO vD
CO CO
ON VO
CM
m CM
VD r-
CO rH
vo r^
CM vO
CO ON
rH
rH CO
rH ON
O vO
CM i-H
CO CM
m - f-v,
*
o
-H
4-1
•H
a)
-o
•o
J-l
rt
td
en
0
CO
e
0
iH
LJ
ca
a)
CO
ji
o
3
O
UH
MH
o
c
S
R
IX
-------
TablelV :Trace Metal Analyses of Wastewator from Processing Operations at MIC
Tanneries.
Process
Liming &
Unhair
Pickling
Chrome-.
Tanning
Final
•Waste
R
X
SD'
R
X
SD
R
X
SD
R
X
SD
Total
Chromium
mg/1
560-
1400
919
372
50-
141
97
37.4
TRACE METALS (ug/1)
Pb
80-
116
99
18
98-
175
128
41.4
98-
120
110
11
Cu
118-
205
148
50
15-
90
52.5
5.3
14-
30
197
896
25-
70
53.7
• 25
Fe
1130-
2100
1680
500
1300-
1310
1300
70
950-
1820
1357
4,40
.8-
1.3
1010
200
Ni
" 140-
480
357,
188
380-
380
380
0.0
. 240-
510
350
114
100-
300
206
98.8
Cd
20-
50
37
15
32-
62
47
21
26-
51
78
82 :
24-
35
29.8
4.5
Zn
>250
0.0
80-
206
143
89
46-
200
• 128
109
68-
100
58
32
X Mean of four observations
R- Range
SD= Standard Deviation
is 180 cm tall with a cross-sectional area of 400 cm2,, The.
filters are provided with a perforated tray at the top to per-
mit even flow distribution., The plastic media used in the study
are made of polypropylene (filter pack, Mass Transfer, Kendal,
England)„ The physical characteristics of the media are:
specific surface area 118 m /m2, volume void ratio 0093 and
minimum irrigation rate 503 m3/m2d0 The media are packed
randomly in the filters. The activated sludge unit used in the
study is described elsewhere^-' ' „
A schematic of the experimental phases of the study is il-
lustrated in Figure 3, Preliminary screening involved removal
of particulate matter, using a 165 mm mesh screen* Following
plain sedimentation for 12 hours, the supernatant was used in
a series of jar tests to determine the optimum dose of coagu-
lants and pH level. The supernatant from the coagulation/
sedimentation unit was fed continuously to both the biofilters
and the activated sludge unit. The effluents of the biological
treatment units were further treated in a double-stage filtra-
tion system0 The filters comprise plexiglass columns 12 cm
in diameter and 185 cm in Iength0 The first filter contains
sand with Oo2. mm effective size and 605 uniformity coefficient.
1099
-------
•IOFILTER
FEED
PLASTC
WEDIA
ft, lot RECYCLE
RECYCLE
RECYCLE
SINGLE STAGE PLANT
CLARIFIES
FEED
PLASTIC
KEDIA
TWO STAGE PLANT
«••
-------
z
c>
S
te
H
CO
g
3
1
o
o
o
m
tc
1
0
o
Q.
Ul
O
F
tc
r
M
a
z
tt
I-
C
5
tc
CO
Ul
ai
z
o
I
3
UJ
s
I
5 S
8 3E "
s y P
5 H
z
o
11
o
o
UJ
CO
o
I
o
§
o
2
< cs
z z
Ul
_J
CD
<.
cc
t-
LJ _J
Gj r- ^*
or 5 H
o Q «f
CO CD 3
M
0)
4-)
CC
&
QJ
4-1
O)
CO
£2
>>
!-i
QJ
P
CO
H
.P
Q)
4-1
cn
P
•H
cO
TO
cn
4)
o
o
t-l
PM
o
-H
4-1
CO
.c
O
txi
QJ
M
OO
X O
1101
-------
too
00
eo
70
J? CO
o
S
A.) TURBIDITY
20
10
4.6
S SB
PH VALUE
e.B
»o
o 70
B) COLOR
ALUM DOSE
so
100
ISO
200
2SO
a g 300
4.3
6 9.S
PM VALUE
Figure A. Effect of pH and Alum Dose on the Removal
of Turbidity and Color of Tannery Wastes.
1102
-------
The second filter is packed with the same sand mixed with 1%
w/w powdered activated carbon.
The physico-chemical characteristics and trace metal ana-
lyses of the raw and treated wastewaters were determined accord-
ing to the procedures described in the Standard Methods
RESULTS AND DISCUSSION
Treatability of Tannery Wastewater
An initial survey indicated that iron salts are not suit-
able coagulants, due to formation of intense black color. It
is presumed that iron salts react with gallic acid in the tan-
nery_wastes to form this persistent color0 Formation of a
turbid, muddy-looking solution and non-settleable sludge pre-
cluded the use of lime for pre-coagulation of tannery waste-
waters. As shown in Figure 4, appreciable removals of color
and turbidity were achieved by alum (Al2(SOA)o, 18 HoO)in the
range of 200-300 mg/1 .at a PH range of 6-6.5. Alum coagulation
was also effective for removal of organic constituents and trace
metals (Table V)„
Recycling of heavily tannery wastewater during biofiltra-
tion is indispensable as it appreciably improves the treatment
performance. Figure 5 illustrates BOD removal, where recycling
20O
400
•00
800
M_ RECYCLE RATIO (%)
Figure 5. Effect of Recycle Ratio on Apparent and True
BOD Removal by Biofiltration.
1103
-------
c-4 co o\
CM co CT\
CO f-( CM
H B
CM Ot CM
r-i CN
8
1104
-------
increases both true removal (So-Se) and apparent removal (So—Se),
where So, Se and Sa are the influent, effluent and applied BOD
after mixing with recycled flow, respectively. An optimum re-
cycling ratio of 600% was selected for the biofiltration study.
Recycling of the precoagulated tannery wastewaters in the single-
stage biofilter produced moderate recovery of BOD, TOC,tannin,
phenol and trace metals (Table VI). However, the recovery of
TKN was comparatively low. The effluent of the single-stage
biofilter constitutes high levels of organics and chromium
which exceed the emission limitations for discharge into public
sewers and the sea. Improved recoveries of various pollutants
were achieved in the double-stage system as shown in Table VI
and Figure 6« The low recovery of TKN during biofiltration
is attributed to the presence of a high concentration of nitri-
fying NHg-N in the influent (100-136 mg/1) which is toxic to
bacteria and hence retards the nitrification process,, The
average hydraulic rate in the biofilters was 0.02 l/m2 which
is much lower than the adequate rate for wetting (0.062 1/m^ 0S)
as recommended by the manufacturer 0 The high organic loading
and the low hydraulic rate contributed to the observed low
recoveries of the biofilters even when using the double-stage
system. Application of a higher hydraulic rate in larger ins-
tallations is expected to enhance wetting and consequently im-
prove the overall treatment efficiency.)
To compare the performance of biofiltration with other
treatment processes, a concurrent study using the Complete-Mix
Activated Sludge (CMAS) system was performed,, Table VII shows
the results of treatment of tannery effluent by the CMAS system
using 24 and 48 hours detention periods„ The CMAS operated at
BOD loadings of 205-409 kg/m30d, while maintaining an average
Mixed Liquor Suspended Solids (MLSS) of 2300 mg/1 and Sludge
Volume Index (SVI) of. 67 mg/l0
Aeration for 24 hours resulted in moderate recovery of
BOD and TOC; high recovery of tannin, chromium and l^S; while
the TKN removal was Iow0 Extended aeration for 48 hours pro-
duced a slight improvement in removal of most pollutants,,
Doubling the aeration time will result in significant increases
in capital and operating costs which are not justified by the
minor improvement in treatment efficiency,, Figure 7 illustra-
tes the comparative effects of the biofiltration and the CMAS
processeso With the exception of COD,both processes produced
more or less similar recoveries of pollutants associated with
tannery wastewater0
An approach to the evaluation of removal of soluble orga-
nics from industrial effluents based on molecular size
. 1105
-------
100
oo 50
•jj 40
20
10
Double-Stage
Single-Stage
BOD
COD
TOC
0 15 30 45 60
0 15 30 45 60
0 15 30 45 60
100
80
60
oo
rt
0)
40
30
20
10
8
6
4
O&G
0 15 30 45 60
TANNIN
0 15 30 45 60
Days
SR
0 15 30 45 60
Figure 6. Choronological Effect of Single-Stage and Double-
Stage Biofiltration on Removal of Pollutants From
Tannery Wastewater.
1106
-------
100
70
50
40
30
20
10
^ 7
W)
I
| 3
£2
CMAS
Biofiltration
O&G
BOD
II III IV V
I' II III IV V
TO C
I II III IV V
I II III IV V
COD
I II III IV V
SR
I II III IV V
I. Plain Sed. , II. Precoagulation III. Biological Treatment
IV. Sand Filtration, V. Multi-media Filtration
Figure 7. Effect of CMAS and Biofiltration on Removal of
Pollutants of Tannery VJastewater.
_
1107
-------
Table VII : Effect of CMAS Detention Time on Removal of Pollutents from
Tannery Wastewater
Parameter -
BOD
COD
TOC
SR
NH3-N
TKN
H2S
Tannin
Phenol
O&C
T-Cr.
Pb
Cu
Fe
Hi
Cd
Zn
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(ug/1)
(ug/1)
(ug/1)
(ug/1)
(ug/1)
(ug/1)
Influent
24
hr.
2970
3166
383
407
132
137
98
243 ,
0.65
65
3.25
46
24.3
340
98
10
48
hr.
2050
4028
758
1276
152
163
5.0
213
0.33
60
2.0
103
29
270
173.3
20
74.3
24
(A)
70.5
50.0
63.4
45.5
24.2
11.7
99.7
75.3
96.9
80.8
88.9
16.3
17.7
24.3
15.3
0.0
Recovery
hr.
(B)
83.0
82.5
76.0
93.0
34.3
40.0
99.95
86.0
98.5
97.6
99.0
40.3
72.5
75.0
25.6
31.0
61.3
%
48 hr.
(A)
89.7
50.9
79.2
80.9
57.2
27.6
100
80.3
97.6
83.3
90.0
29.6
62.1
52.8
37.4
48.5
26.0
(B)
92.0
75.3
90.7
95.5
60.0
55.0
100'
89.0
99.0
98.4
99.6
69.7
87.0
77.3
59.0
66.0
73.6
(A) Recovery of the CMAS Process only.
(B) Overall Treatment Recovery.
Table VIII :ESTIMATED TREATMENT COSTS OF TANNING WASTEWATER (Costs in US dollars)
PARAMET
1. Capital Costs
a. Primary (Screening, Sed. and Precoagul-
ation)•
b. Secondary (biological).
c. Tertiary(mixed media filtration ).
d. Non-Component cost (piping, Instruments).
Amortizations of Capital Costs ( 20 years)
ACTIVATED SLUDGE BIOFJLTRATION
285,000 285,000
345,000 265,000
205,000 205,000
ISA.OOP 120,000
989,000 875,000
98,900 87,500
2. Operating Costs
a. Chemicals
b. Power
c. Labor
d. Maintenance
120,000
22,500
15,000
18,000
120,000
2,500
6,000
5,400
Percentage costs compared to activated sludge.
175,600
274,500
100
133,900
221,400
80.7
1108
-------
distribution using Gel Chromatographic (GC) technique has been
detailed by Hamza and Tambo(9)o xhe organic constituents
which are easily eluted by water (Group I) are amenable to bio-
logical .treatment, while organics with high affinity with the
gel (Group II) require elution by NH^OH. This portion can be
effectively removed by tertiary treatment. Schematic GC pat-
terns of raw wastewater and the effluent of the CMAS and bio-
filtration are illustrated in Figure 8,, The GC patterns indi-
cate the presence of relatively high Group II constituents which
require further treatment for complete treatment. Mixed media
filtration removed 80-85% of the residual Group II constituents.
Treatment Costs
The cost estimates given in Table VIII reflect costs appli-
cable to centralized treatment of tannery effluent to produce '
water suitable to discharge into the sea0 The estimates are
based on the projected discharge of 205 million cubic meters
annually after implementation of the renovation and expansion
plan of GOFIo Cost estimates assume 20 years' service life and
5% low interest rate provided by the Egyptian Govovernment for
public secto.r industries. Operating costs were based on prices
of 1982, and cost analyses were performed according to EPA
Guidelines v-^w 0 As shown in Table 9 the estimated cost of
biofiltration is less than that of the CMAS process„ Further
savings are expected if the clean water pool can be segregated
and discharged without treatment„.
In case of lack of sufficient financial support, it is pro-
posed to install pre-treatment and biofilters first, while ad-
ding the post-filters at later stage0
Toxicity of Raw and Treated^Tannery Effluents
The effect of raw wastewater on the survival of Cyprimus
carpio and Mugil cap;.to has been investigated as a complementary
part of the study^^) „ The mean survival time was 5 minutes
for hatched embryo, 1.4 hour for larval stage and 17„2 hours
for-juvenile stage of Cyprimus carpio and 1604 hours for juve-
nile stage of Mugil capitOo The high toxicity is attributed
to the presence of high concentrations of trace metals, orga-
nic constituents, l^S salts and ammonia„ The 96 hours LC5Q
of the hatched embryo and larval and juvenile stages of Cyprimus
carpio exposed to the biofiltration effluent was achieved in
treated effluent diluted with water to 301%, 502% and 803%, res-
pectively., Although the treated waste was less toxic than the
raw waste, it is estimated that treated tannery waste must.be
1109
-------
Raw Wastewater
Figure
Biologically
Treated Tannery Effluents,
1110
-------
diluted 40 times to reduce the fish toxicity to an acceptable
level. The toxicity of the treated effluent is attributed to
the high concentration of NH3 (77 mg/1) which causes severe
histopathological changes in gill structure„ Ammonia stripping
is being investigated as an option to reduce fish toxicity.
CONCLUSIONS
A centralized treatment system is proposed for the MIC
tannerieso The treatment train encompasses precoagulations, bio-
filtration and post-filtration0 Among the treatment alternatives,
biofiltration is the most adequate due to lower operating costs,
suitability for intermittent flow, versatility and relative ease
of operation,, The proposed treatment train complies with the
limitations for emission into the sea (except for NH3). Reduc-
tion of the toxicity of the treated effluent can be achieved
by ammonia stripping or by dilution when mixed with sea water,,
Segregation of non-polluted effluents, originated in clean water
pools, from polluted wastes is expected to reduce treatment
costSo Government subsidization of the centralized treatment
facility is necessary to encourage MIC tanneries to institute
the proposed treatment system,,
ACKNOWLEDGEMENTS
This study was partially sponsored by the USEPA under Grant
No. PL-3-542-4o The authors wish to express appreciation to the
staff of the IWRC at the High Institute of Public Health,
Alexandria, Egypt, for their valuable help0
nn
-------
REFERENCES
1. Smith, P., and Yates, D., "Experience of High Rate Biological Filtration
at Derby Sewage Treatment Workers." Water Poll. Control (C.b) 79,198o,87.
2. Hosono, Y., and Kubota, H., "Characteristic Evolution of Trickling Filter
Process." Water Res. (G.B.) 14,1980,581.
3. Bailey, D., Robinson, K., Collins, S. and Clarke, J., " The Treatment of
Effluents From a Chrome Side Leather Tannery on a Conventional Biological
FilLer ", J. Water Poll. Control Fed., 71, 1972,202.
4. Pierce, D., " Upgrading Trickling Filters." Environmental Protection
Agency, EPA 4309-78-004,1978.
5. Younis, M., " Study of Methods of Industrial Waste Treatment of Tannery.
Wastes." MPI1 Thesis, Alexandria University, 1982.
6. EPA, " Document for Effluent Limitations Guidelines and New Sources
Performance Standard for Leather Tanning and Finishing Point Source
Category." EPA, 440/1-74 - 016 a, 1974.
7. Hamza A., and Hamouda, F. " Multi-process Treatment of Textile Wastewater"
Proceedings of the 35th Annual Purdue Industrial Waste Conference,
Ann Arbor Science' Publishing, 198o-152.
8. "Standard Methods for Examination of Water and Wastewater". Amer Pub.
Health Assn., Washington, D.C., 15th ed., 1890.
9. Hamza, A. and Tambo, N., " Evluation of Dairy Waste Treatability by Gel
Chronotography". Memories of the Faculty of Engineering, Hokkaido Univ.,
Japan, Vol XIV,4, 1977,44.
10. EPA Guideling " Inovative and Alternative Technology Guidelines" EPA
930/9-78.009,1980.
11. Zaki, M. and Saad, S., " Toxicity Assessment of Raw and Treated Tannery
Hastes Using Fish Bioassy." Symposium on Acute Aquatic Ecotoxicological
Tests Methodology, Standardization and Significance CERTIA, France,
Nov. 1981.
1112
-------
PART X: INNOVATIVE RESEARCH
Effect of Periodic Flow Reversal
Upon RBC Performance
John T. Bandy. U.S. Army Construction Engineering
Research Laboratory, Champaign, Illinois.
Manette C. Messenger. U.S. Army Construction Engineer-
ing Research Laboratory, Champaign, Illinois.
Introduction
Rotating Biological Contactors are traditionally
operated as a series of units through which water always
flows the same direction. After the microbial film is
established on the RBC's, a characteristic pattern of growth
is seen across the stages: growth is heaviest on the first
stages and diminishes with each sucessive stage once the
organic concentration falls to the level below which removal
is a function of .organic concentration. The biological com-
munities on each stage change in response to their differing
environments. The total growth on a stage is roughly pro-
portional to the organic concentrations which normally occur
within it.
When shock loadings occur across multistage RBC units,
downstream stages receive higher organic loadings than
usual. Studies of RBC response to shock loadings have
revealed that downstream stages have only a limited capacity
to treat these short duration excess loadings. Much of a
shock load passes through an RBC untreated.
1113
-------
The purpose of this research was to explore the feasi-
bility of increasing an RBC's capacity to treat shock loads
by periodically reversing the direction of flow across two
or more successive stages of an RBC installation. This con-
cept was first suggested by Borchardt, et al., in June of
1978 as a result of their studies of RBC nitrification.*
RBC biofilms have been observed to grow more rapidly in
response to an incremental organic concentration than they
decay in response ,to an equal decrement. This characteris-
tic of biofilms suggests the possibility of increasing the
total inventory of film within several stages of an RBC by
periodically reversing the direction of flow across those
stages. The period of reversal would be sufficiently long
to permit the concentration gradient across the stages to
re-establish itself in each direction but not so long that
the biofilm could adjust fully through growth and decay to
the new distribution of nutrients. Under average flow con-
ditions, the RBC's performance shouldn't be significantly
affected. The former earliest stage would see a lower aver-
age concentration than under conventional flow and would
presumably perform less removal. However, the former latter
stages would see higher average concentrations and due to
the advantage of growth over decay they would develop
heavier growth which would remove those organics now passed
by the first stage. Far more reserve capacity would exist
to treat shock loads due to the greater inventory and more
even distribution of biofilm created by periodic reversal.
An analogous mode of operation was explored for trick-
ling filters in England forty years ago.2 Flow was reversed
periodically across two filters in series in order to
prevent ponding on the more heavily loaded filter. The
alternate heavy feeding and comparative starvation experi-
enced by the slime encouraged first growth and then
endogenous respiration and sloughing. Under the "alternat-
ing double filtration" mode of operation, much less buildup
of slime and consequent ponding was observed. When slowly
rotating distributer arms came into vogue, alternate feeding
and starvation became feasible within one filter. Ponding
was seldom a problem and the alternating double filtration
mode of operation died out. Its implications for treatment
efficiency and effluent variability were never explored.
1114
-------
This research will evaluate the flow reversal concept
in RBC's. Today we will report our preliminary results.
More elaborate follow-up experimentation is now underway.
Materials and Methods
A Clow pilot scale RBC was used in this research. The
13 foot (4 m) diameter disks had a total area of 11000 sq.
ft. (1022 m2). The media was set in a 2000 gallon (7.6 m3)
tank divided into four equal compartments to allow staging.
The media rotated at about 1.6 rpm.
For purposes of this experiment, the plant was config-
ured as two 2-stage RBC's in parallel. Only one pair was
actually sampled.
We assumed that for flow reversal to have a significant
effect, organic concentrations in the reversed stages would
have to be low enough so that removals achieved depend on
their fluctuations. In some plants, the first stage or two
is saturated with respect to organics. Higher organic con-
centrations do not produce appreciably,higher removals on a
mass basis. Percent removals fall. At these high organic
concentrations, the mass transport of oxygen into or waste
products out of the film is controlling. The source of
wastewater for this experimentation was chosen to be the
partially treated effluent of an Imhoff tank. This moderate
strength wastewater (8005 of approximately 100 nig/1) never
saturated the first stage of the Clow Corporation pilot
scale RBC plant used for the reversal experiments.
Wastewater was pumped through the pilot RBC at flow
rates between 28,800 and 72,000 gpd (272 m3/ day). The flow
was measured with an in-pipe flow sensor. Hydraulic load-
ings varied from 2.6 to 6.5 gpd/ft 2 (.107 to .267
m3/in2/day). The corresponding organic loadings were 1.67 to
4.17 Ib SCOD/1000 ft2/day. The soluble ultimate BOD to
soluble COD ratio was variable but averaged .6.
The influent and effluent wastestreams were sampled
every hour with ISCO automatic samplers. Two ml of sulfuric
acid were added to each sample bottle before sample collec-
tion to stabilize the .sample. No sample was more than 26
hours old when it was analyzed.
1115
-------
The Hach Reactor Digestion COD method was used to
assess the RBC's performance. This is an EPA approved
method which is convenient for the rapid analysis of large
numbers of samples. All samples were filtered through What-
man No. 5 before analysis. ThepH and temperature were meas-
ured daily while experimentation was underway. Dissolved
oxygen rose from about 4 mg/1 in the influent to near
saturation in the effluent at the lowest organic loadings.
The dissolved oxygen profile across the system was essen-
tially flat at the highest loadings.
Experimentation was performed in three phases. First
the RBC was operated in the conventional manner for a period
of almost two months. The flow rate during this period of
film establishment and maturation and base-line data collec-
tion was 28,800 gpd (109 m3/day). Good removals (about 60
percent of the influent COD) were obtained by three weeks
after startup. A severe thunderstorm then scoured off much
of the growth requiring us to install a cover and allow some
regrowth prior to experimental data collection.
The second phase of the experiment involved daily flow
reversal. During this phase, valves were opened and closed
immediately after the noon sampling to make the first stage
the second, and the second, the first. Flow rates of 28,800
gpd (109 m^/day), 57600 gpd (218 m3/day) and 72,000 gpd (273
m^/day) were employed during this phase to allow the iden-
tification of any interactions between hydraulic loading and
flow reversal. Reversal experiments continued for seven
weeks.
The final phase of experimentation involved a return to
conventional one way flow. During this phase, the higher
flow rates used for reversed flow studies were also applied
with conventional flow.
Results and Discussion
The results of the experiments performed are presented
in Figure 1 and summarized in Table 1. Flow reversal was
initiated after almost 2 months of consistent COD removals
with conventional flow. Percent removals rose steadily dur-
ing the first 5 weeks of reversal despite greatly increased
hydraulic loadings. This was a striking result since
1116
-------
hydraulic loading and percent, removal are normally either
constant or inversely related in a moderately loaded RBC.
When the hydraulic loading was reduced to the pre-reversal
level for the purpose of obtaining reversed flow performance
data directly comparable to the conventional flow baseline,
sloughing increased markedly. Percent removals fell shar-
ply. The earlier higher flow rate was restored and conven-
tional flow operation was resumed when this trend became
apparent so that the expected gradual return of the system
to baseline performance might be observed. Performance was
falling at approximately the same rate as it had earlier
increased when a violent thunderstorm blew the unit's canvas
cover into the tank and broke the coupling between the
driven gear and the RBC shaft. Before repairs could be made
and a suitable replacement cover procured, the unit froze up
for the winter, prematurely terminating experimentation.
Table 1
Results of Experimentation
Fraction Removed
•Type Data
one way,
reversal,
reversal,
reversal,
reversal,
reversal
one way,
one way,
20 gpm
20 gpm
40 gpm
50 gpm
40 gpm
20 gpm
40 gpm .
50 gpm
// Observations
64
55
40
71
54
42
36
72
.473
.575
.625
.636
.722
.488
.641
.574
sd.
.132
.149
.086
.045
.053
.095
.082
.073
During these preliminary experiments, an average of 55
percent of the applied COD was removed under conventional
flow operation and 62 percent was removed with reversal.
The difference in the means was significant at the 1 percent
level based on 436 influent/effluent sample pairs. Influent
concentrations showed no correlation with percent removals.
The reversed flow effluents were less variable than
were the conventional flow effluents, especially when the
1117
-------
initial conventional flow baseline data were compared to the
reversed flow data. However, the experiments were ter-
minated before the higher flow rate conventional flow
effluents reached a steady state. Therefore, these data do
not establish a difference in effluent variability.
Experiments are underway now in which two compartments
of the Clow pilot plant are operated as a conventional two-
stage RBC and the remaining two are reversed daily. The
conventional flow control and the reversal experiment will
receive an identical influent and will operate under identi-
cal climatic conditions for several months. An unequivocal
comparison of the two modes of operation will then be possi-
ble. If the benefits of flow reversal suggested by the
early research are confirmed by this follow-up work, optimal
reversal periods can be identified and economic feasibility
analyses can be performed.
References
1. Borchardt, J. A., Kang, S. J., and T. H. Chung, Nitrifi-
cation of Secondary Municipal Waste Effluents by Rotat-
ing Bio-Discs, EPA-600/2-78-061, June 1978.
2. Hawkes, H. A., The Ecology of Wastewater Treatment, Per-
gamon Press, Inc., New York, 1963.
3. Banerji, S. K., ASCE Water Pollution Management Task
Committee Report on "Rotation Biological Contactor for
Secondary Treatment," Proceedings: First National
Symposium/Workshop on Rotating Biological Contactor
Technology held at Champion, PA, Feb 4-6, 1980.
1118
-------
% COD REMOVAL
_row-£oio>->joo*oo
ooooooooooo
H-
era
c
l-t
(0
w
X
X)
fD
t-1
H-
3
rt
03
ro
en
c
en
"a
K>
OQ
n>
CD
o
ro
O
GJ
O
O
01
o
o
8
1 !• 1 1 III 1 1
o •* . ro
3 ' . • 0
I * * . 1
»< • 3
e
.1
-
a
i^
*
*
* . °
*
3
< • m
5 0
- (A •
2. *
-
1
*
• J^
•" °
*
8
I •
? 01
I - 0
' I
ill I i 1*1 i i I
nig
-------
1120
-------
AN ASSESSMENT OF
DISSOLVED OXYGEN LIMITATIONS AND INTERSTAGE DESIGN
ROTATING BIOLOGICAL CONTACTOR (RBC). SYSTEMS
IN
Warren H. Chesner, Engineering Consultants £ Associates,
Commack, New York
John J. lannone, Roy F. Weston, Roslyn, New York
Jeremiah J. McCarthy, U.S. Army 10th Medical Laboratory,
West Germany
INTRODUCTION
Over 260 RBC systems are presently in operation in this
country, with flowrates ranging from less than 0.1 to S^MGD (l).
RBC systems have received much attention in recent years, in-
cluding investigations by EPA of process, power, and equipment
performance (2), and a National RBC Symposium in 1980 (3).
Numerous reports in the 1iterature and RBC facility surveys
(1 , 4, 5, 6, 7, 8) have reported difficulties with the initial
stages of RBC systems reflected by heavy biofilm growth, the
presence of nuisance organisms (beggiotoa), and a reduction
in organic removal rates. These problems have been attributed
to excessive organic loadings which result in dissolved oxygen
deficiencies in the biofilm. Recent unpublished surveys have
reported results (9) which indicate that nuisance organism
growth may be precipitated exclusively by high sulfide concen-
trations.
Empirical design approaches proposed by RBC vendors are
presently used almost exclusively by engineers for RBC municipal
wastewater treatment design. Only one vendor design presently
makes reference to the aforementioned organic; overload condition
(10). A loading limitation is recommended and defined in terms
1121
_
-------
of a limiting organic loading (e.g. for a mechanical drive RBC
system, 4.0 lbs/day/1000 sf of soluble BOD loading on the first
by
literature (11, 12) have stressed the need for de-
sign techniques based upon fundamental principles of substrate
and oxygen, transport and utilization to help define oxygen
limitations and stagewise kinetic parameters. The development
of a mechanistically sound kinetic model for RBC's or any fixed-
film biological process, however, is extremely difficult due to
the complex relationships among transport and other kinetic
phenomenon which must account for both substrate and oxygen
transport and utilization. The use of these complex models
design engineers is not practical.
The empirical vendor approaches are simple to use, re-
quiring the input of flowrate and organic concentration to
establish a given design surface area, They are limited how-
ever in that they have no fundamental basis to account for
organic removal, oxygen limitations, staging requirements, step
feeding or recycle, and do not address interstage removal.
On the basis of a review of available design method-
ologies, it is apparent that a simplified approach is desireable
to define RBC performance on the basis of rational kinetics.
The approach described by Clarke et. al, (11) is a reasonable
method which defines interstage removal on the basis of Monod
growth functions, and a mass balance with respect to substrate
across a completely mixed reactor. No simple model, however,
can account for the interaction between oxygen requirements and
substrate utilization. The design of RBC systems at the present
time then, is best undertaken in a simplified two-step approach
Step 1,
Step 2,
Define limiting design conditions to pre-
vent organic overloading and dissolved
oxygen limitation.
Utilization of a rational design method
to define the substrate removal capacity
of the RBC surface area under conditions
which are not oxygen limiting.
Under an ongoing contract with the Municipal Environmental
Research Laboratory of the U.S. Environmental Protection Agency,
data which define organic overload (oxygen limiting) conditions
were sought and an interstage model calibrated from field data
collected from several operating facilities.
1122
-------
ORGANIC OVERLOAD CONDITIONS
Optimum RBC process design requires that the microbial sub-
strate utilization rate represent the rate limiting condition.
Under this condition, the process achieves maximum use of sur-
face area because substrate removal is limited by the ability of
the biomass to assimilate the waste and no extraneous factors
(e.g. lack of dissolved oxygen) limit the rate of substrate re-
,mova1.
Except for the aforementioned vendor recommendations con-
cerning organic loading 1imitat ions, little data is available to
define the conditions which induce dissolved oxygen limitations.
Williamson and McCarty (12), utilizing their fixed film model,
predicted that dissolved oxygen limitations wpuld occur at
soluble BOD (SBOD) concentrations of kO mg/1, adjacent to the
biofilm. This would correspond to a mixed liquor RBC reactor
concentration somewhat greater than ^0 mg/1 SBOD. Field obser-
vations, 1iterature reviews, and telephone interviews were made
to determine influent conditions which result in organic over-
loads. Overloaded conditions on an RBC were identified by the
characteristic colonization of the media by nuisance organisms
which gain a competitive advantage over other organisms under
oxygen deficient conditions. For each of the facilities sur-
veyed, information concerning nuisance ,organisms, influent
organic concentrations, and hydraulic loading were recorded to
determine wastewater characteristics which could be associated
with these conditions. The survey included a total of twenty-
three facilities. Results are tabulated in Table I.
A graphic presentation of these results is presented
Figure 1, which depicts a plot of total influent BOD on the
ordinate and first-stage hydraulic loading on the abscissa.
The graph presents a demarcation line that separates facilities
which experienced first-stage overloading problems from those
which did not (i.e. by the presence or absence of nuisance or-
gan i sm) .
The relationship is depicted as a hyperbolic function,
where the product of the variables is a constant:
(BOD)(Hydraulic Loading) = Constant Organic Loading
From the graoh, it can be seen that the organic loading that;
separates plants with overloaded conditions from plants without
such prcble-s is 6.^ pounds BOD/day/1000 sf.
i n
Organic overload and oxygen limitations are used interchangeably,
1123
-------
TABLE I. Organic Overloading Conditions Related to Influent
Organic Concentration and Hydraulic Loading
Plant1
ID No.
1
2
A
5
8
10
12
13
14
15
16
29
30
31
32
33
34
35
3
6
9
11
36
Average
RBC Influent
BOD
Concentration
(mg/1)
125
48
50
182
47
169
85
55
96
98
72
93
175
145
505
350
336
180
118
152
144
96
213
Average
1st Stage
Hydraul ic
Loadi ng
(gpd/sf)
7.4
3.6
4.0
10.1
15.5
4.7
12.5
9.1
5.0
9.6
6.6
18.5
7-2
6.2
1.9
3.9
5.9
6.8
5.1
3.9
7.9
9.0
6.5
Cal culated
Average
1st Stage
Organic Loading
(Ibs/day/lOOOsf)
7,4
1.4
1.7
15-3
6.1
6.6
8.9
4,2
4.0
7.8
4.0
14.3
10.5
7.5
8.0
11.4
16.7
10.2
5.0
4.9
9.5
7.2
2
DO-Li mi 1 1 ng
Condi t ion
P
A
A
P
A
P
P
P
A
P
A
P
P
P
P
P
P
P
A
A
P
p
11.5 : P
i
: Except for air driven plant no, 16, all are mechanical
dr-fve facilities with no supplemental air.
2
A: Plants experiencing no nuisance organism growth (absence)
In the First Stage,
2
P: Plants experiencing problems with nuisance organism growths
in the First Stage,
1124
-------
LEGEND
SOO
.. «.
E
z
2 3OO
K.
1-
z
u
0
z
o
o
o 2OO
O
ID
1-
X
1U
_j
z 100
o
ID
tc
C
| A PLANTS REPORTING PROBLEM
! . , NUISANCE ORGANISM GROWT
1 ADJACENT I.D. NUMBERS IDE
SURVEY PLANTS
\ REPRESENTS AN ORGANIC L
1 BREAKPOINT OF 6.4 LBS 30
; 1000 SF
1
l _j4
. \
\
\
.35 5
6* \ *31 *9
\ f
3® •>s^
VN^. A12
1 1 I 1 l II t | i
> 2 4 6 8 10 12 14 16 18 2
FIRST STAGE HYDRAULIC LOADING ( GPD / SF )
LOADING
FIGURE } : D.O. LIMITING CONDITIONS RELATED TO
INFLUENT ORGANIC CONCENTRATION AND
HYDRAULIC LOADING
1125
-------
INTERSTAGE MODEL CALIBRATION
The Clarke model previously mentioned is based upon a mass
balance with respect to substrate across an assumed completely
mixed RBC stage at steady-state which can be written as follows:
' = FSo - FS] - ^X A
reactor liquid volume (volume)
(Equation 1)
whe re,
V
-j— = change of substrate concentration with time
(mass/volume • time)
F = wastewater flow rate (volume/time)
S = influent organic concentration (mass/volume)
S. = effluent organic concentration (mass/volume)
y = specific growth rate of attached RBC microorganisms
(Vtime)
Y = apparent yield of attached RBC microorganisms
/ mass biomass produced %
mass substrate consumed
X = mass of attached microorganisms per unit area
(mass/area)
A = RBC surface area (area)
This equation assumes that the intensity of mixing in each stage
is sufficient for complete mixing, and that organisms decay is
small compared to the growth rate, and that all substrate re-
moval Is due to attached biomass.
Using the Monod growth function:
S,
V = V
1
max
-) and,
(Equation 2)
max
Y
A X
(Equation 3)
1126
-------
defi ni ng ,
_
-
and,
max
v ,
*> as tne area capacity constant
(i.e. maximum substrate which
could be removed per unit area
per unit time)
then,
(Equation k)
' as t'ie remova' coeffiecient (Equation 5)
(actual substrate removed
per unit area per unit time)
at steady-state V Pp) = 0
and, PS,
,R =
1
, or
(Equation 6)
Kc
P~
(Equation 7)
A major feature of the Clarke model is its use of a rational
approach for defining substrate removal converting the terminol-
ogy into design parameters which are readily used in the field
today: R, the removal coefficient which reflects the organic
removal rate, can be defined in terms of pounds/day/1000 sf, and
P, the area capacity constant, which represents the maximum re-
moval rate possible, can also be described in terms of pounds/
day/1000 sf.
To calibrate the Clarke model, interstage soluble BOD1 data
was collected from eleven RBC facilities selected with three to
six RBC stages where organic overload (i.e. greater than
6.4 lbs/day/1000 sf BOD organic loading) did not exist. Organi-
cally overloaded faci1ities were screened for two reasons:
1. the organic removal rate of these facilities are
highly variable as a result of nuisance organism
interference and the influence of DO deficiencies
on substrate removal rate; and,
2, the intent of the designer as presented is to avoid
this condition (i.e. insure that design is not or-
ganically overloaded by keeping the loading level
below the DO-1imi ting level).
Interstage data available at the eleven facilities was soluble
BOD (SBOD).
1127
-------
area, R, and the
the model was cali-
The eleven facilities provided influent values of soluble BOD
from 10 - 95 mg/1, with a 55 mg/1 average value. Hydraulic
loadings ranged from 0.4 - 1.5 gpd/sf with a 1.3 gpd/sf average
value.
Using both the organic removal per unit
soluble BOD concentration in the reactor, S,
brated as follows:
Values of ( /R) vs ( /S) for each stage of all eleven
plants were plotted to yield a straight line with a
slope of K /P and a y-intercept of /P, per Equation ?•
The maximum removal rate, P, and the half velocity co-
efficient, K , were computed from this graph.
This graphical analysis for the four consecutive stages is pre-
sented in Figures 2 and 3, respectively.2 The results of the
four consecutive stages are presented in Table 2.
TABLE 2. Calibrated Maximum Removal Rate, P, and
Half Velocity Coefficients, K (Soluble BOD)
Stage No.
1
2
. *
Maximum Removal Rate, P
(GPD/SF • mg/1)
1000
667
400
100
(Lb/Day/1000 SF)
8.34
5.56
3-34
0.33
Half Velocity
Coefficient, K
(mg/1)
161
139
82
25
Straight lines were drawn through the data, visually
weighting the distribution of data points and screening
data which were judged as outliers.
1128
-------
8
\ . . 1
\
\ .
- \ *
\ :
. \. ...
\
*^»
•^^ *
••^v
1 , ! 1 1 1 1 1 . I \
d
0
CM
0
—. e w
Z, ui
- <
d OT i—
^ CO
IO
o
d
f-»
o> ea . N- <' -
_J
z
<
1—
0
^
Ul
tr
C9
u_
) a/i
1129
-------
\
\ . •
V
\
\
\
» \ "
\ -
\
• :•• \- <
V '."
%
• \
o
o
o
0 „
o e
^^
o
3§
g
d
<-s
3 10 O « **
3 -. . - o
a 0 0 X
UI
Cf>
H
co
( BUI/I - Qd9/JS ) U/l
CO
o:
CO
CO
*
1 I
en CO
0 0
d o
• \ l
\
•» » —
\ '
. \
•\
".\
\- -
\
i i i i i i i \
•?
CM
d
o
o
d g 10
O ^ UJ
d ^j
CO |—
0 ~
d
r\
^ (D
111
o:
CO
u.
( &UJ/I • QdO/dS) H/l
1130
-------
Values of effluent concentration in each stage can be ex-
pressed as a function of flowrate, influent concentration and
surface area by setting Equation 5 equal to Equation 6 and
solving for the effluent concentration, S., as follows:
F.
K — ~i *» o ~
A. o
P-.S.
r\ r* «
KS+.S.
(Equation 8)
The effluent concentration for any stage can be expressed as:
S. =
CCHL.(SO- K.)>,P.
- K.)- p.
. x SQ)
2CHL.)
(Equation 9)
where,
S = influent SBOD concentration (mg/l)
o
HL. = hydraulic loading (gpd/sf)
S. = effluent SBOD concentration (mg/l)
T .. .
P. = area capacity constant (I/day)
K. = half velocity coefficient (mg/l)
F. = wastewater flowrate (gpd)
A. = area (square feet)
i = denotes RBC stage under analysis
To determine the hydraulic loading given the influent concentration
and the required effluent concentration, Equation 8 can be re-
arranged as follows:
(P.)CS.)
Hi -
HLi ~ CK.
S.)(S - S.)
110 i
(Equation 10)
1131
-------
To assess the accuracy of the calibrated interstage (4 stage)
model, a comparison between the observed field performance of six-
teen RBC facilities and the model was undertaken.1
The sixteen facilities were divided into influent concen-
tration ranges of 52, 97, 1^8 and 2\k mg/1,2 In order to compare
the model, which expresses BOD in terms of soluble BOD (SBQD) , an
SBOD:BOD ratio of 0.5 was assumed
The model is compared to regression lines of field perform-
ance in Figure 4. The top graphic of Figure 4 presents the data
as organic removal in lbs/day/1000 sf vs hydraulic loading in
gpd/sf; the bottom graphic as organic removal vs organic loading,
both in lbs/day/1000 sf. The regression line extends across the
range of loading conditions (hydraulic and organic) observed in
the field. From Figure 4, it can be seen that the model is more
accurate (i.e. with respect to the regression line) for the lower
concentration ranges, the lower hydraulic loading ranges and the
lower organic loading ranges.
The lack of correlation at higher concentrations and load-
ings can be attributed to the fact that:
a. the model uses the Monod growth function to account for
both substrate utilization and mass transfer. In fixed
film systems mass transfer, however, is dependent upon
reactor concentration (i.e. actually the concentration
adjacent to the film). Higher influent concentrations
would exhibit greater mass transfer driving forces
which would
a substrate
increase values of P and
deficient system.
K ca1i brated i n
the model was not designed to predict removal in
organic overloaded environments and at the higher
loadings where the model does not correlate well,
first stage loadings may exceed design levels
(e.g. greater than 6.4 lbs/day/1000 sf).
'Primary clarifier BOD removal was assumed to be 30% for each
facility with primary clarifiers,
"Detailed data concerning these sixteen facilities are presented
elsewhere (2),
1132
-------
t:
CO
8
o
o
o
03
CO
CD 3
>2
O '
s
UJ
K.
O «
o
K.
O
ACTUAL
PREDICTED
* CONCENTRATIONS REPRESENT
INFLUENT CONDITIONS
_L
JL
_L
_L
I 2 34
HYDRAULIC LOADING (GPD/SF)
ORGANIC REMOVAL VS. HYDRAULIC LOADING
o
§
0
o
m
23
|2
tu
K.
O
o:
o
,
—APPROXIMATION OF DESIGN
PREDICTIONS FROM DATA POINTS
« 52 mg/1 BOD INFLUENT
A 97 mg/1 BOD INFLUENT
« KB mg/1 BOD INFLUENT
B 214 mg/1 BOD INFLUENT
LINEAR REGRESSION, ACTUAL
J l__
,
,
1 2345678
ORGANIC LOADING { LBS BOD/D/IOOO SF )
ORGANIC REMOVAL VS. ORGANIC LOADING
FIGURE 4= COMPARISON OF ACTUAL PERFORMANCE
TO DESIGN PREDICTIONS
1133
-------
MODEL ANALYSIS
The major advantage of a model such as the one described and
calibrated here is its ability to predict the effect of changes
in system characteristics, such as staging and step feeding, upon
RBC performance. In performing these .analyses, it must be re-
called that the model was intentionally calibrated under condi-
tions which were organically underloaded. As a result, advan-
tages or disadvantages of various process variations apply only
to those conditions.
Staging
Assuming an influent BOD = 100 mg/1 (
and a flowrate of 1.0 million gallons per
model was used to predict the benefits of
the amount of surface area required to ach
50, 60, 75, 85, and 95 percent. The resu-1
Figure 5 indicate that multistage systems
area than single stage systems to achieve
removal, up to a point. For the influent
optimum staging would be as follows:
i.e. SBOD = 50 mg/1),
day, the interstage
staging by assessing
ieve efficiencies of
ts illustrated i n
require less surface
a given percentage
condition indicated,
Desired Percent Removal
50
60
75
85
95
Number of Stages
1
2
3
3
Caution, however, must be used when increasing the number of
stages to insure that an organic overloading condition is not
created (i.e. loading per stage exceeds 6,4 lbs/day/1000 sf),
Step Feeding
At the same influent condition, an analysis of the advan-
tages of step feeding was undertaken. The analysis was under-
taken by assigning P and K values to each stage to account for
the relative quantities of influent flow which are being diverted
to downstream stages, but which exhibit kinetics associated with
the initial stages of the system. For example, if 75 percent of
the flow entered the first stage and 25 percent the second stage
1134
-------
vf
e s
o 9
00
CD _l
OT U.
Ul
U.
z
j-
O)
tlj
o
<
CO
ll.
o
cc
111
CO
CO
z
UJ
DC
13
O
UJ
CC
U
or
hJ
o
<
U.
DC
r>
CO
2:
o
o
<
to
u_
o
o
UJ
u_
U.
UJ
o
o
in
to
O
o
o
10
o
o
to
CJ
o
o
o
CM
o
o
o
o
o
o
o
(coi
vaav Hovjans aaamoaa
IU
cc
13
1135
-------
of a k stage system, then 75 percent of the flow would experience
removal on the basis of running the model consecutively through
four stages, with values of PjKj, P2K2' P3K3 and P4K*t for the
four stages. Twenty-five percent of the flow would experience
the removal associated with a three stage system with values of
P K., P_K2 and P K for the three stages. The final effluent was
calculated as a material balance of the two flows. A schematic
representation of this analyses is presented in Figure 6. Vari-
ous combinations of step feeding were examined, with no advantage
indicated under any conditions and decreasing efficiencies noted
if too great a percentage of flow was diverted to the latter
stages.
CONCLUSIONS
1. A limiting organic loading to the first stage of
6.k lbs/day/1000 sf of BOD (total BOD) was found
to be the organic loading beyond which nuisance
organism growth and corresponding process per-
formance problems occur,
2. An interstage model was calibrated at organic
concentration in a range of 50 - 100 mg/1 BOD
for systems which were not overloaded. The
model illustrates the advantages of staging
when high efficiencies are desired. The model
did not Indicate any advantage to step feeding
in systems that are organically underloaded,
3. The design of RBC systems is best undertaken
with two independent criteria. The first to
establish the desired surface area and staging
arrangements, and the second to ensure that no
stage is organically overloaded. The organic
loading limitation presented here is considered
a good design criteria to prevent oxygen limi-
tations. Additional data, is required to
establish an accurate rational interstage model
to predict RBC performance.
1136
-------
o
UJ
UJ
u.
0.
UJ
I-
(O
vD
UJ
cc
1137
-------
REFERENCES:
(1) Chesner, W.H., and lannone, J.J., "Current Status of
Municipal Wastewater Treatment with RBC Technology in
the US," presented at the Feb. 4-6, 1980, First National
Symposium on RBC Technology, held at Champion, PA.
(2) Chesner, W.H., and lannone, J,J,, "Review of RBC Design
Procedures and Process, O&M, Equipment and Power Perform-
ance," Municipal Environmental Research Laboratory, U.S.
Environmental Protection Agency, unpublished,
(3) Smith, E.D., Miller, R.D., and Wu, Y.C,, Editors, "Pro-
ceedings of the First National Workshop on Rotating Bio-
logical Contactor Technology," held at Champion, PA,
February 4-6, 1980.
(4) Lagnese, J.F., "Use of Supplemental Air to Correct an
Oxygen Limitation Condition of an Operating RBC System,"
presented at the February 4-6, 1980, First National
Symposium on RBC Technology, held at Champion, PA.
(5) Selden Sanitary Corporation, "Report on Nitrification,"
prepared by Henderson and Bowdwell, Plainview, NY,
December 1977.
(6) Hitdlebaugh, J.A,, and Miller, R.D,, "Full Scale RBC for
Secondary Treatment and Nitrification," presented at the
February 4-6, 1980, First National Symposium on RBC
Technology, held at Champion, PA,
(7) Wood, Paul K., "Report: Survey of Sewage Treatment Plant,
Winchester, KY," State Compliance Report, Kentucky Bureau
of Environmental Protection, August 3-4, 1976,
(8) Dobrowolski, F.J., Brown, J.M., and Bradley, F.M., "Testing
Rotating Biological Contactors for Secondary Treatment in a
Converted Primary Tank," presented at the Water Pollution
Control Association of Pennsylvania, August 8, 1974.
(9) Personal communication, Richard Phillips, State of Vermont,
Environmental Conservation Agency.
(10) Auto&Lol VUJJQYI Manual, 1979.
(11) Clark, J.H., Moseng, E.M., and Asano, T,, "Performance of
a Rotating Biological Contactor under Varying Wastewater
Flow," JounnaJL WPCF 50, (5), 896, 1978.
(12) Williamson, K. and McCarty, P.L., "Verification Studies
of the Biofilm Model for Bacterial Substrate Utilization,"
louJinaJt WPCF, 48, (2), 281, 1976.
1138
-------
COMBINED BIOLOGICAL/CHEMICAL TREATMENT IN RBC-PLANTS.
Hall yard 0degaard, Division of Hydraulic & Sanitary
Engineering, University of Trondheim, NTH, Norway.
INTRODUCTION
Over the last 5 years, there has been an increased use of
rotating biological contactors (RBC's) in Norway. Out of a
total of about 500 sewage treatment plants in the country, 52
(~ 10%) are RBC-plants. '
Since eutrophication is a major pollution problem in the
Norwegian waterways, chemical treatment to remove phosphorous
is extensively used. Of the 52 RBC-pTants, 48 of them are com-
bined biological/chemical plants.
The main objective for establishing chemical treatment in
combination with biological treatment in RBC's is of course
to remove phosphorous. In addition it is experienced,however,
that the addition of a chemical precipitant (normally alum or
ferric iron) will improve the effluent quality by coagulation
of the fine fraction of biofilm that often is difficult to
settle. Therefore the chemical treatment serves a dual pur-
pose. •
Chemical treatment may be achieved in combination with
RBC's by principally three process designs (fig.1), hereafter
named:
- Simultaneous precipitation
- Combined precipitation
- Post precipitation
1139
-------
r
Precip it ant add! tion
>^»
— »-
ill.
ill
» * i i 1 *
HlHt
1!
1!!
Mi
1
I'rlt
i
I L
H
!
R8C
-*•
/ Final settlinc
SIMULTANEOUS PREC 1PI TAT 1 ON
i . i j i « ; ,' ,
iiiill
Precipitant addition
R8C Flocc. .
COMBINED PRECIPITATION
Final settling
—
t
i
<>i
ill
i!
• |
•ii
' J 1. 1 i :
i i ' 1 ' ' 1 '
1 , I'!! ':,',.
Iliiijii.
RBC
-:- - J<
^ • 4f
1
1
4-
f
*
fi
*l
3
j1
/ Settling Floccul.
— Precipitant addition
— 1»- .
f Final settl ing
POST PRECIPITATION
Fig. 1. Different combined biological/chemical treatment
processes.
1140
-------
In simultaneous precipitation the precipitant is added to the
RBC tank, the precipitation occurs here and the precipitated
matter is removed together with the biofilm suspended matter
in the following separation unit, normally a sedimentation
tank. Since flocculation occurs in the RBC-tank,a floccula-
tion tank between the RBC and the settling unit is normally
not included.
Post-precipitation plants consist of biological and chemi-
cal treatment completely separated from each other. The RBC
has its own settlingtank followed by the chemical step5consis-
ting of chemical mixing^flocculation and separation of floes.
As will be shown later, the major part of the combined
biological/chemical RBC-plants in Norway are designed for
combined precipitation. Compared to the traditional post-
precipitation plant, the settling unit for the RBC-sludge is
here omitted. The precipitant is added to the RBC effluent
and the whole suspension (biological and chemical sludge) is
then flocculated before combined sludge removal in settling
units.
In this paper, an overview over the Norwegian experiences
with combined biological/chemical treatment in RBC plants will
be given first, and thereafter a special project concerning
the comparison of simultaneous precipitation and combined pre-
cipitation will be reported.
THE USE OF RBS PLANTS IN NORWAY - AN OVERVIEW.
The general picture of sewage treatment in Norway in
Norway is as follows:
- Most of the plants are small. About 50% of the total
number of plants (about 500 plants) has a connection of
less than 500 personequivalents, and about 93% of the
plants are for.less than 10.000 person equivalents.
- Chemical precipitation is extensively used either alone
of in combination with biological processes.
- All the plants are built-in, either in houses (most of
the plants) or in halls blasted into the rocky mountains
(the bigger plants). This is so because of the strict
labour environment rules and the cold climate during
winter.
Based on a questionnaire to the environmental protection
authorities in all the counties, information about the RBC
plants was gathered.
1141.
-------
In table 1 the total number of RBS plants are grouped accord-
ing to their size and process design.
Table 1. RBS-plants in Norway grouped according to size and
process design.
Process
design
Without
precipitation
Simultaneous
•precipitation
Combined
precipitation
Post
precipitation
Total
Size in personequivalents
> 500
2
2
18
1
23
> 500-1000
'
8
1
9
>1000-2000
1
11
,1
1-3
> 2000
1
"'!,
5
1
7
Total
4
2
42
4
52
It is demonstrated that combined biological/chemical treatment
is the normal (48 out of 52 plants) and that combined precipi-
tation is the biological/chemical process design mostly used
(42 out of 48 plants).
The principal reason for the popularity of this process
- The pollution authorities has accepted this design to
give an effluent quality comparable to what is expec-
ted from post-precipitation plants.
- While post-precipitation with RBC has not been invest-
ment economically competitive with post-precipitation
based on activated sludge, combined precipitation has,
because of the savings by omitting the RBC settling
unit.
There has also been questions whether not simultaneous
precipitation also could give results comparable to combined
precipitation.- If so, the flocculation tanks could also be
omitted. Since all the plants are built in, it is obvious
that savings in the area of the plants will give considerable
savings in the total investment cost. The main objective of
the project reported later in this paper was to investigate
1142
-------
this matter.
Effluent quality
~The information about effluent quality where not complete
for all the 52 plants, partly because many of the plants are
so new that the pollution authorities have not started their
control program yet and partly because information from the
county authorities was incomplete. In table 2 is summarized
results of 24 plants where effluent quality has been analyzed
on flow proportional samples repeatedly taken over one year.
The plants have been divided into two groups (^ 1000 person-
equivalents) .
All the plants included in table 2 are combined precipitations
plants.
Table 2. Mean effluent quality from Norwegian biological/
chemical RBC-plants (combined precipitation)
Size group j Tot P
Person equiv. ppm
< 1000 pe
>_ 1000 pe
1,27
0,39
Number
Samples
42
70
of
Plants
12
12
BOD 7
ppm
24-
15
Number
Samples
35
70
of
Plants
12
12
It can be seen that the smaller plants have problems with mee-
ting the effluent standard of phosphorous for this group
(< 0,8 mg P/&). This is mainly because of operational diffi-
culties with the chemical equipment. In the bigger plants how-
ever, which are well operated, the average effluent quality is
below the effluent quality standard for bigger post-precipita-
tion plants in Norway(<_ 0,5 mg P/£, _< 20 mg BOD7/£).
Operational experiences
There are presently 10 different RBC-products represented
among the Norwegian RBC-plants. The smaller plants (< 500 pe)
are dominated by local products. In the larger plants, three
products dominates completely:
- Bio-surf (Aerosurf)
- Enviro.disc
- Nova
It is not possible from the data collected to state as to
whether any of there products give better effluent quality than
the other. In the bigger, well operated plants they all give
1143
-------
effluents quality results that is expected from post-precipi-
tation plants « 0,5 mg PA, < 20 mg BOF7/A).
We have in Norway, as in most countries, I guess, exper-
ienced mechanical failures with the RBC plants, and because of
this,I can say that the popularity of the RBC's has faded some-
what lately.
Another problem that is experienced,is that nitrification
in the RBC reduces the alkalinity so much that it is difficult
to maintain sufficiently high pH for chemical precipitation.
Alum is normally used as precipitant and pH is then normally
5,0-6,5 in the precipitation step.
RESULT FROM A RBC-PLANT WITH COMBINED PRECIPITATION
For a period of over two years the influent and effluent
quality en flow proportional samples have been monitored at
Vinstra RBC sewage treatment plant which is operating accord-
ing to the combined precipitation mode.
The plant is designed for 5100 personequivalents with a
design flow of 140 m3/d. The flow diagram for the plant is
shown in fig.2.
The plant receives municipal and dairy wastewater and in
addition,external septic sludge is dewatered at the plant.
Reject water from this septic handling contributes signifi-
cantly to the composition of the raw water. The precipita-
tion chemical is aluminium-sulphate of the AVR-quality
(a Swedish product consisting of a combination of alum- and
ferricwsulphate). The dosage of 130-140 mg AVR/£ (about
n-lzW) Al/O is fed flow proportional to the water down-
stream the biodisc units.
As will be shown later, the average BOD7 - concentration
in the raw water is about 320 g 0/m3 and the average daily
flow 1200 m3/d. Presupposed that a BOD-reduction of 30% will
occur in the presetting units, the organic area loading is
19 g BOD7/m2.d, which is about the Norwegian design criteria
for combined biological/chemical RBC plants. (< 20 g BOD7/m2-
The average treatment results from this plant over the
last two years is shown in table 2.
The data in table 2 clearly demonstrate that a well oper-
ated combined precipitation RBC-plant can give an effluent qual-
ity of at least the same quality as traditional post precipi-
tation plants based on activated sludge can.
d)
1144
-------
O
CO
pi
c:
o
. Q-
'•i—
O
.
CL'
-a
CD
o
o
(O
"o
1Z
•
CM
CD
1145
-------
Table 2. Average 2-year result from the Vinstra
combined precipitation RBC plant.
Parameter in,g/m-
out, g/m;
COD
BOD7
Total P
SS
578 ±240
322 ±155
8,49±4,2
267 ±148
30 ±15
10 ± 4
0,18+0,13
7 ± 4
94,8
96,9
97,9
97,4
24
23
23
24
AN EXPERIMENTAL COMPARISON BETWEEN COMBINED AND
SIMULTANEOUS PRECIPITATION.
An experimental investigation was carried out during fall
1981, with the objective of studying combined precipitation at
extreme loading and of comparing combined precipitation and
simultaneous precipitation. Since the before mentioned Vinstra
plant, has two paralell treatment lines, this plant was chosen
as experimental site. Two different investigation periods were
carried out. In the first period,all of the water was led
through only one of the treatment lines after the RBC tanks.
In the other investigation period, the flocculation tanks in
one of the treatment lines were short - circuited, so that com-
bined and simultaneous precipitation could be investigated in
paralell with each other on the same settled raw water.
The two treatment lines are shown in fig.3 where the samp-
ling points are marked. The samples were taken as flow-propor-
tional samples and were analyzed for total and soluble COD,
total and soluble PO^-P, and suspended solids. In addition
the secci-depth at the end of the secondary settling tanks were
monitored. Since we here are mainly interested in the funct-
ioning of the biological/chemical treatment system, the raw
water composition described later is that obtained from the
outlet of the primary settlers.
The first investigation period lasted for one week and
the second for three weeks. The samples were taken daily and
analyzed immediately at the plant.
In the second period the point of precipitant addition
was changed during the period in order to see if this influ-
enced the results.
1146
-------
ro
T3
c
o
TJ
41
t-
a
•o
c
o
-------
In the first period we wanted to investigate how high
hydraulic loading influenced the, treatment result. As can be
seen from table 3 the average daily overflow rate on the final
settling unit was 20 m3/m2-d, corresponding to an average max-
imum overflow rate of 1,6 m3/m2-h. The Norwegian design cri-
teria for this kind of process is 1,3 m3/m2-h.
If so happened that the settled influent in the first
period was also very concentrated, partly due to a significant
contribution of septic sludge dewatering reject in this period.
Table 3. Organic and hydraulic loadings (^ Calc. values)
Ave.org. area/load
gCODTot/m2-d
gCODsol/m2-d
gBODs /m2-d1'
Tot ,
gBOD5 -/m2-du
sol
Ave . mi n . deten ti on
time in floccula-
tors
min.
Ave. max overflow
rate in final
settlers
m3/m2'h
Norw.
design
criteria
18
20
1,3
Period 1 ! Period 2
Comb.prec. I Comb. Simult.
57
26
29
14
20
1,6
36 36
10 10
19 19
5 5
50 0
0,6 0,6
This led to a very high average organic area loading,
57gCOD/m2-d based on total COD and 26 g CODSQl/m2-d based
on soluble COD. We have a pretty good knowledge about the
correlations between BOD5 and COD on this water and based on
the monitored COD-values5 we can calculate the BOD-loadings
to have been 29 g BOD5/m2-d based on total BOD5 and 14 g
BOD,
;sol
/m2-d based on soluble BOD5.
1148
-------
This means that the plant was both hydraulically and orga-
nically overloaded. The organic load was actually consider-
ably over what we had anticipated it to.be, namely about the
maximum load adviced in the Norwegian design criteria (see
table 3). • .
In the second period the hydraulic loadings were lower
partly because the amount of raw water was lower in this per-
iod, but mainly because the incoming flow was divided into the
two lines. The organic area load was however approximately
what was expected, very near the Norwegian .design criteria.
In both periods the precipitant dosage was kept at its
normal value, 135 g alum/m3 added flow proportional to the in-
coming water.
RESULTS AND DISCUSSION
In table 4 are summarized the average treatment result from
the two periods.
Table 4. Treatment results.
Parameter
CODTot in
out
CODsol.in
out
P0*-ptot.in
out
P0lt-Psol in
out
Period 1
Combined
561 ±220
31+ 10
287+137
21+ 9
6,5+2,4
0,34±0,18
5,6+3,4
0,02+0,01
SS out 13+1
Secci depth
(cm) -
163+33
Period
Combined
409+104
!57+ 30
110+ 35
34± 19
6,7+1,6
0,39+0,21
2,0±0,56
0,04+0,03
27±9
85+39
2
Simultaneous
409+104
72+ 29
110± 35
34+ 22
6,7+1,6
0,59+0,13
2,0+0,50
0,05+0,02
. 37+12
50+18
1149
-------
In spite of the extremely high "loading in the combined preci-
pitation line in investigation period 1,the plant gave still
very good effluent, quality similar to what normally was achi-
vied at hydraulically only half the load previously. The COD-
values in the effluent corresponds to BOD5-values of less than
5 g BOD5sol/m3 and less than 10 g BOD5tot/m3.
The phosphate precipitation was also good,precipitation
was complete (P0<*-psoi = 0,02 ppm) and even with an average
maximum overflow rate as high as 1,6 m/h, the separation of
floes was good, leaving total POit-P = 0,34 ppm and suspended
solids = 13ppm in the effluent. The effluent was very clear
with secci-depth of 163 cm.
As may be seen from table 4 the effluent guality in per-
iod 2 was not so good as in period 1 in either of the treat-
ment lines.Why the concentration of both total and soluble
organic matter went up also in the combined precipitation
line, we don't know. It is probably however that the extreme
organic loading in period 1 had some impact on the biofilm in
such a way that the high loading resulted in a thick biofilm
that stripped of to a greater extent in period 2 resulting
in a decrease in total active biomass.
However, since the main objective in period 2 was to com-
pare the two different processes the absolute treatment result
is not so important.
The effluent quality in the combined precipitation line
was not bad, however, with average COD-values in the effluent
corresponding to BOD5-values of less than 10 BOD5 ,/m3 and
less than 20 g BOD5 /m3. S01
It was obvious, however, that separation of floes was
worse than in period 1,leaving 27 ppm of suspended solids in
the effluent and the secci-depth had fallen to 85 cm. Phos-
phate removal was still good.
When we compare the results from the two treatment lines,
two things are clear:
- The removal of soluble organic matter was equally good
in the two lines.
- Separation of floes was better in the combined precipi-
tation line.
The difference between the two processes lies in the floe
separation aspect,as one might expect. This demonstrates the
usefulness of the flocculation tanks.
It must be said, however, that the results in the simul-
taneous precipitation line may have been influenced by:
1150
-------
- the fact that the precipitant addition point was chan-
ged over the period.
- the fact that the precipitant dosing equipment for this
line was a provisorium which may have given a dosing
rate not as reliable as in the other line.
An other investigation period was actually also performed
in which all water was led through the simultaneous precipita-
tion line giving a relatively high hydraulic load (ave. max
overflow rate 1,2 m/h). The organic loading in this period
was relatively low and therefore the results are not included
in detail. The results where very similar to the ones obtai-
ned in period 2. Towards the end of the hydraulically high
loaded simultaneous precipitation period the sludge separation
was, however, very good (SS < 10 ppm, P(\ - Ptot < 0,3 ppm and
BOD5 tot < 20 ppm). This proved to us that good treatment may
also be obtained by the simultaneous precipitation process.
Based on the results we would, however, advocate that
flocculation tanks are used with a detention time of 15-20 min.
It is very important that these are constructed so that biofilm
settling is avoided.
With regard to precipitation addition point it cannot be
stated from this investigation whether this should be done be-
fore or after the RBC-tank with combined precipitation. Since
only the normal point of addition, after the RBC's was tested
here. It may be argued however that precipitant addition be-
fore or into the RBC-tank would precipitate some of the solu-
ble organic matter and thus actually reducing the organic load
on the RBC. The comparison between simultaneous and combined
precipitation did not, however, confirm this, since soluble
COD in the effluent was the same in the two lines.
It may also be argued that precipitant addition before
the RBC would give less precipitant consumption because the
suspended solids concentration are lower here and consequent-
ly less precipitant would be consumed in coagulating suspended
matter.
In simultaneous precipitation we feel therefore that the
most correct precipitant adding point is before the RBC, main-
ly because this is the point where thorough mixing is easiest
to obtain. In the investigation, we did not see much differ-
ence in the results, for the different dosing points. But
then we kept the dosage constant.
1151.
-------
CONCLUSIONS
1. Combined biological/chemical treatment can be obtained in
RBC plants by adding precipitant before or into the RBC-
tank (simultaneous precipitation), after the RBC-tank and
before a flocculation/sedimentation system (combined preci-
pitation) or with separate chemical step downstream the RBC-
settling tank (post-precipitation).
2. The experiences from Norwegian combined precipitation plants
and from the project reported in this paper,is that excellent
treatment results may be obtained by combined precipitation.
Since this process has one sludge separation unit less than
post-precipitation, combined precipitation is economically
favourable compared to post-precipitation.
3..Simultaneous precipitation may also give acceptable effluent
quality» but it seems that separation of floes is better in
a combined precipitation system as a result of better floc-
culation.
4. Adding precipitant (alum) directly to the RBC-tank in simul-
taneous precipitation does not give any adverse effect on
the capability of the biofilm to remove organic matter when
pH is kept above pH. = 6,0.
5. Good removal of both organic matter and phosphate was ob-
tained with combined precipitation at high organic and hyd-
raulic loadings. A design criteria of average organic loa-
ding < 20 g BOD7 tot/m2 d> average max overflow rate in
final settler < 1,3 m3/m2-h and 20 min detention time in
flocculators seem to be acceptable. This is a consider-
ably higher organic loading than can be accepted when chem-
ical treatment is not included.
1152
-------
TREATMENT OF DOMESTIC SEWAGE BY AQUATIC RIBBON SYSTEM
Chun-Teh' Li. Department of Environmental
Engineering, National Cheng-Kung University, Tainan,
Taiwan
James S. Whang. AEPCO, Inc., Rockville, Maryland
T.N. Chiang. Department of Environmental Engineering,
National Cheng-Kung University, Tainan, Taiwan
INTRODUCTION
Fixed-film biological processes have become popular
for treating organic wastewaters during the past decade
because of their low energy, and possibly low manpower
requirements. There are many types of fixed-film biological
treatment processes including trickling filters; bio-towers,
which are basically trickling filters that use light-weight
plastic media instead of gravel as a substrate; rotating
biological contactors (RBCs); packed-bed reactors (PBRs);
and fluidized-bed reactors (FBRs). The latter three types
of fixed-film processes have been successfully implemented
for: (1) removal of soluble BOD from wastewaters; (2)
nitrification, and (3) denitrification of various
wastewaters. These processes generally consist of a
fixed-film reactor followed by a liquid-solids separation
unit such as a clarifier, a filter, or other special
liquid-solids separation unit.
The fixed-film biological processing system reported
in this study represents a new concept which combines the
fixed-film reactor and the liquid-solids separatdr into
one physical unit. In this new fixed-film system (Aquatic
Bio-Ribbon Treatment System), specially designed synthetic
ribbons are used as a substrate for microorganisms in a
1153
-------
reactor system to achieve removal of 5-day soluble BOD,
nitrification, and denitrification.
SYSTEM DESCRIPTION AND THEORETICAL CONSIDERATIONS
The aquatic ribbon reactor system (Figure 1) consists
of a secondary treatment chamber, a nitrification chamber,
and a denitrification chamber
in series. In the secondary treatment chamber, zones for
distinctly different unit processes are maintained. The
upper zone is aerated minimally just to meet all oxygen
transfer requirements. The lower zone, which is designed
to facilitate sedimentation of sludge particles, is not
aerated. Similar to other fixed-film systems, there is
no sludge return. The dissolved oxygen (D.O.) level in
the secondary treatment chamber is kept above 2.8 mg/1 at
all times by artificial aeration. Heterotrophic
microorganisms are grown on the ribbon surface. The
heterotrophs metabolize the organic matter present in the
sewage and available oxygen to achieve the conversion of
soluble BOD to suspended BOD. Exess biomass growth on the
ribbon surface eventually sloughs off and settles into the
lower zone of the secondary chamber where it is periodically
removed.
The second and the third chambers have an arrangement
similar to the first chamber. Both chambers are divided
into two zones: the upper zone provides the actual
biological treatment process; the lower zone facilitates
sludge removal.
The upper zone of the second chamber is aerated to
encourage growth of autotrophic nitrifying bacteria
(Nitrosomonas and Nitrobactor) on the ribbon surface,
but activated nitrifying sludge is not recirculated.
Ammonia-nitrogen present in the wastewater becomes the
electron donor in the bio-nitrification process and is
oxidized to nitrite and eventually nitrate. The D.O. level
in the nitrification chamber is maintained at, or above,
1.35 mg/1 to promote the necessary nitrification process.
The third chamber is isolated from the outside
atmosphere to promote anoxic conditions. This facilitates
1154
-------
0)
4-1
CO
O
4-1
O
TO
c
o
.0
f>
•H
PS
O
•H
•U
TO
3
cr
to
CO
Q)
O
O
1-1
60
1155
-------
growth of denitrifying bacteria on the ribbon surface and
thus biological denitrification processes. Denitrifying
bacteria which include the genera Pseudomonas, Bacillus,
and Achromobactor are responsible for the denitrification
process. Nitrate- and nitrite- nitrogen under anoxic
conditions become the electron acceptors in the
bio-denitrification process.
In the denitrification process, the denitrifiers need
soluble carbon to meet the metabolic requirements. As the
denitrification process occurs, the carbon source in the
wastewater is depleted and becomes a limiting factor.
Supplemental carbon is always necessary to enable the
denitrification process to continue. In this experiment,
carbon was supplemented by adding methanol to the inlet
of the denitrification chamber so that the
carbon-to-nitrogen ratio was maintained at, or slightly
above, 1.10. The empirical methanol feed concentration
or requirement can be computed by the following equation:
[CH3OH]
where,
and
2.47x[N03-N]
[CH,OH]
[NQ,-N]
[Nof-Nj
[D.6.]
1.53x[NO -N]
+ 0.87x[D.O.]
= methanol cone, in wastewater (mg/1)
= nitrate nitrogen cone, (mg N03-N/1)
- nitrite nitrogen cone, (mg N02~N/1)
= dissolved oxygen cone, (mg/1)
In this study, methanol was manually fed to the third
chamber to satisfy the carbori requirement of the
denitrification process. A reactor system, which uses part
of the influent-soluble BOD as a carbon source for the third
chamber, is being studied.
The removal and conversion mechanisms involved in BOD
removal, nitrification, and denitrification in an aquatic
ribbon system are depicted in Figure 2.
The-sewage used in testing the reactor system was
primarily of domestic origin. Table 1 summarizes the
composition of the wastewater used in this study. The
aquatic ribbon system is currently being studied for its
capability to treat tannery wastewater.
1156
-------
09
c
o
.o
o
• rJ
JJ
3
o-
c
o
00
O
•H
g
!E
a
I
03
O
•H
ao
o
i— i
o
•H
PQ
§
T3
-------
Table I. Characteristics of Wastewater
Parameter
Concentration
Temperature
BOD
COD
BOD/COD Ratio
TSS
VSS
Total Nitrogen
Ammonia Nitrogen
Nitrate Nitrogen
Total Phosphorus
Alkalinity
(as Calcium Carbonate)
7.50
26^ ± ;
100
252
0.37
150
112
29.5
20.0
0.01 -
8.0
174
7. '60
I °c
248 mg/1
560 mg/1
0.45
186 mg/1
140 mg/1
34.0 mg/1
24.0 mg/1
0.05 mg/1
12.0 mg/1
210 mg/1
The hydraulic detention times in each reactor chamber
were varied by changing influent rate. The influent rates
tested varied from 7.2 I/day (1.903 gpd) to 21.8 I/day
(5.762 gpd). These influent rates result in hydraulic
detention times from 8 to 24 hours in the secondary
treatment process (First Chamber), from 4 to 12 hours in
the nitrification process (Second Chamber), and from 2 to
4 hours in the denitrification process (Third Chamber).
Tables II to IV summarize pertinent experimental conditions,
Table II. Experimental Conditions for BOD Removal
Influent Hydraulic
Rate Detention
(I/day) Time (Hours)
7.2
10.8
14.4
21.6
24
16
12
8
BOD Loading Rate
SurfaceVolumeD.O.
(gm/M^day) (gm/M3-day) (mg/1)
14.17
13.20
27.43
41.14
248
231
480
720
2.85
2.80
"3.20
3.50
1158
-------
Table III. Experimental Conditions for Nitrification
Influent
Rate (I/day)
7.2
10.8
. 14.4
21.6
Hydraulic
Detention Time
(Hours)
12
8
6
4
Surface Loading D.O.
of Ammonia-N (mg/1)
(gm/M 2-day)
1.03 . 1.35
1.61 1.35
1.80 1.60
3.00 1.80
Table IV. Experimental Conditions for Denitrification
Influent
Rate
(I/day)
7.2
10.8
14.4
21.6
Hydraulic
Detention Time
(Hours)
6
4
3
2
Surface Loading D.O.
of Nitrate-N (mg/1)
(mg/M-day) _____
1.11 0.15
1.80 0.15
1.94 0.10
2.57 0.40
RESULTS AND DISCUSSIONS
Experimental results for the aquatic ribbon treatment
system are summarized in Table V. As shown, soluble 5-day
BOD removal was 93.5% at a hydraulic detention time of 24
hours. The soluble 5-day BOD removal efficiency decreased
with decreasing hydraulic detention time as evidenced by
only an 85.4% BOD removal- efficiency when the hydraulic
detention time was reduced to 8 hours.
As shown in Table V, nitrification efficiency was 97.6%
at a hydraulic detention time of 12 hours. Like BOD removal
efficiency, nitrification efficiency also decreased with
decreasing hydraulic detention time as evidenced by an 86.5%
1159
-------
Table V. Results of Treatability Study of Domestic Sewage
Influent
Rate
(I/day)
7.2
10.8
14.4
21.6
BOD
Removal
Nitrification
(%) Rate t d (%)
93
91
88
85
.5
.2
.3
.4
14
13
27
41
.17
.20
.43
.14
24
16
12
8
97.6
94.0
86.0
86.5
Rate
1.00
1.52
1.70
2.58
td
12
8
6
4
Denitrification
(%)
89.7
88.6
87.1
87.3
Rate
1.00
1.59
1.69
2.25
td
6
4
3
2
Note: Removal rate is expressed in mg/tr -day
t^ is the hydraulic detention time (hours).
nitrification efficiency when the hydraulic detention time
was reduced to 4 hours. Nitrification efficiency was
essentially unaffected by the hydraulic detention time,
once the hydraulic detention time exceeded 8 hours.
As shown in Table V, denitrification efficiency was
89.7% at a hydraulic detention time of 6 hours.
Denitrification efficiency decreased with decreasing
hydraulic detention time as shown by an 87.3%
denitrification efficiency when the hydraulic detention
time was reduced to 2 hours. It should be noted that
denitrification efficiency is essentially unaffected by
the hydraulic detention time, once the hydraulic detention
time exceeded 2 hours.
Figure 3 presents the relationship between soluble
BOD removal efficiency and BOD surface loading rate for
a given influent BOD concentration. Figure 4 presents the
relationship between soluble BOD removal efficiency and
BOD volumetric loading rate for a given influent BOD
concentration. From these graphs, the following are
concluded:
• Soluble BOD removal efficiency increases with
decreasing BOD surface loading rate or volumetric
loading rate
• Soluble BOD removal efficiency increases with
1160
-------
100 -
0246 8 10 12 14
»OD SUBFACE LORDING KftTE
-------
increasing influent BOD concentration at a given
surface, or volumetric loading rate, especially
when the loading rates are relatively high.
• The aquatic ribbon treatment system is more
efficient at higher influent BOD concentrations,
with an upper limit not yet defined.
The overall system BOD removal efficiency, expressed
as a function of hydraulic detention time, is presented
in Figure 5.
The experimental results related to the nitrification
chamber are presented in Figure 6. Nitification process
efficiency increases with a decreasing surface loading rate
of ammonia-nitrogen. However, the actual nitrification
rate (unit mass of ammonia-nitrogen removed per unit surface
area and unit time) increases with an increasing
ammonia-nitrogen loading rate, and seemingly reaches an
assymptotic level. This assymptotic value can not be
clearly defined, because it exists beyond the range of the
experimental conditions.
The experimental results related to the denitrification
chamber are presented in Figure 7. Similar to the
nitrification process, the denitrification efficiency is
shown to increase with a decreasing surface loading rate
of nitrate-nitrogen. The actual denitrification rate (unit
mass of nitrate-nitrogen removed per unit surface area and
unit time) increases with increasing surface loading rates
of nitrate-nitrogen. It is believed that there is an
assymptotic value, which again, can not be determined
because it lies beyond the range of the experimental
conditions.
The nitrification and denitrification efficiencies
as a function of hydraulic detention time are presented
in Figures 8 and 9, respectively. Figures 8 and 9 also
show the actual nitrification and denitrification rates
as a function of hydraulic detention time. In brief, the
actual nitrification and denitrification rates increase
with decreasing hydraulic detention time. The reaction
efficiency (whether nitrification or denitrification)
increases with decreasing hydraulic detention time.
However, the nitrification efficiency drops sharply when
the hydraulic detention time is shorter than 4 hours.
1162
-------
100
95
90
65
80
75
70
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
SURFACE LOADING RME (gm HH3-B/M2-day)
Figure 5. Level of Nitrification Vs. Surface
Loading Rate of Ammonia-Nitrogen
100
*"
B 95
s 90
S 85
g
S 80
75
70
-| r
T T
I
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
SUWAO5 LOADIHG SMS
I
s
8.
Figure 6. Level of Denitrification Vs. Surface
Loading Rate of Nitrate-Nitrogen
1163
-------
0 100
I
fa
w 95
|
i 90
o 85
80
J
0 7 14 21 28 35 42
HYDRAULIC DETENTION TIME (Hours)
Figure 7. Overall BOD Removal Vs. Detention Time
1164
-------
„ 100
B 95
C 90
M
E 85
80
75
g
70
i
A
0 24 6 8 10 12
HYDRAULIC CCTEHTIOK TIME (Roues)
Figure 8. Level of Nitrification Vs Detention Time
g
8
a
85
80
75
70
i
1 t
f
CN
7 f
•f
5I
4 §
3 §
£
2 $
0 1 234 56
HYDRAULIC DETISTIOH TUB (Sae»)
Figure 9. Level of Denitrification Vs. Detention Time
1165
-------
The amount of alkalinity expressed as calcium carbonate
consumed per mg/1 of ammonia-nitrogen nitrified is presented
in Figure 10. The amount of alkalinity generated per mg/1
of nitrate-nitrogen denitrified is presented in Figure 11.
The net effect on the system chemistry is that every mg/1
of nitrogen removed from the wastewater results in an
approximately 4.2 mg/1 reduction of alkalinity. For
domestic sewage, which normally contains 150 to 220 mg/1
'of alkalinity and 15 to 25 mg/1 of ammonia-nitrogen,
nitrification and denitrification are not limited by
alkalinity availability. If alkalinity availability is
anticipated to be a problem, manual addition of alkalinity
using lime or calcium carbonate should be considered.
With regard to supplementary carbon for the
denitrification process, it was demonstrated that a methanol
concentration of approximately 46 mg/1 in the wastewater
before the denitrification chamber was sufficient for
denitrification to take place at 90% efficiency.
It is possible to increase the total surface area of
the aquatic ribbons within each chamber of the reactor
system to increase the reaction rate, and thus somehow
proportionally reduce the hydraulic detention time
required. The specific surface area values (i.e., ratio
of total surface area and effective liquid volume for each
chamber) for the pilot system used in this study are
summarized in Table VI along with specific surface areas
reported by other researchers (Ref. 1,2,3) using RBC systems
for treating domestic or municipal wastewater. It can be
seen that the specific surface area used in this study is
approximately 5 to 15 times less than those reported by
others. Therefore, it appears that the total ribbon surface
area could be increased to considerably increase the system
capacity and/or reduce the hydraulic detention time
required.
COMPARISON WITH OTHER STUDIES
Because an aquatic ribbon fixed-film treatment process
is a new concept, no comparable data are available for a
comparison study. One can only attemp to make a generic
comparison study using the data obtained from RBC systems.
A number of researchers (Ref. 2,4,5,6) have tried to
1166
-------
8
a
140
130
120
110
100
90
7.08 ag/1 Alkalinity Consumed Per
ng/1 of NH3-N Nitrified
LJ^I I I 1 I I I I
0 15 16 17 18 19 20 21
coHoncnuwiotj OP wi3-n REMOVED (ng/i>
Figure 10. Consumption of Alkalinity as a Function of
NH3-N Nitrified
_ 60
f
50
40
r*Xn
2.9 ag/1 Alkalinity Increase Per
•g/1 of NO3-N Denitrified
LJSr* ' > ' 1 1 1 1
O 10 12 14 16 18 .30 22
OP BOj-B MMOVBD (Kg/1)
Figure 11. Increase of Alkalinity as a Function of
N03-N Denitrified
1167
-------
Table VI. Comparison of Specific Surface Area of Aquatic
Ribbon System and Rotating Biological Contactor
(RBC) Systems
Reference
This Study
Marsh, et al
Poon, et al
Huang, et al
Secondary
Treatment
Chamber
18 1/M
200 1/M
620 1/M
180 1/M
Nitrification
Chamber
40 1/M
200 1/M
620 1/M
Denitrification
Chamber
70 1/M
establish the relationship between soluble 5-day BOD removal
and loading rates. The most commonly used method is to
correlate the BOD removal rates with surface loading rates.
These relationships by Poon, et al (Curve A, Ref. 2),
Lagnese (Curve B, Ref. 4), and Reh (Curve C, Ref. 5), along
with the data obtained from this study, are presented in
Figure 12. It can be seen in Figure 12 that the soluble
5-day BOD removal rates achieved by the aquatic ribbons
system are comparable with RBC systems.
In a nitrification study using a 4-stage RBC system,
Marsh et al (Ref. 1) suggested an empirical equation to
describe effuent ammonia-nitrogen concentration as a
function of influent flow rate, influent ammonia-nitrogen
concentration, influent total 5-day BOD concentration, total
media surface area available, and wastewater temperature.
The equation is:
K
[ AxT ]
where, No - Influent ammonia-nitrogen cone, (mg NH3~N/1)
Ne - Effluent ammonia-nitrogen cone, (mg NH3~N/1)
Q = Volumetric flow rate, (cubic meter/sec)
S0 = Influent total 5-day BOD cone, (mg/1)
A = Total media surface area, (square meter)
T = Wastewater temperature, (degree Centigrade)
and K = empirical constant
1168
-------
C8
PS
M
C
•r-l
"O
tfl
o
•J
01
o
CO
ol
- 3.
S
CO
0)
O
I
ffi
u-i
O
O
cc
o
03
C
0)
4-)
V
.£>
CX
•H
J=
to
C
o
•H
4J
CO
"oi
cr!
0)
M
1169
-------
According to the above empirical equation, the data
obtained from this study and data obtained by Marsh, et
al (Ref. 1) were plotted and are presented in Figure 13.
Figure 13 indicates that the aquatic ribbon systems provide
a level of performance similar to RBC systems with
approximately the same slope of K = 15,280.
Poon, et al (Ref. 7), in a nitrification study using
a 4-stage RBG system, suggested that the unit
ammonia-nitrogen removal rates and surface loading rates
are best-fit by a logistic-S curve, expressed by the
following equation:
R
where,
and
R
max
1 + m-e
b-L
Rmax = Maximum unit surface nitrification rate
R = Unit nitrification rate
m = Coefficient
b = Coefficient
L = Surface loading rate of ammonia-nitrogen
Table VII. Comparison of Rmax, b, and m Values for
the Nitrification Process with Other Study
Parameter
Temperature
Rmax
m
b
Note:
This Study
at 26°C
26°C
3.08
7.80
-1.23
This Study
Adjusted
to 11°C
1.37
7.80
-2.76
Poon, et al ,
at 11°C
1.54
10.28
-2.87
Rmax is expressed in gm/M^-day.
m is a dimensionless constant.
b is expressed in M^-day/gm.
Using this logistic-S curve fitting technique, the
Rmax» ™» an<* b values for wastewater temperature at
approximately 26 degree centigrade and temperature-adjusted
1170
-------
O
4-1
O
tB
£2
I I
O s
-00
60
e
•H
T3
cd
o
3
co
C
O
4-1
C
-------
m, and b values from the data obtained from this study
are presented in Table VII. Values reported by Poon, et
al are also listed in Table VII. From Table VII, it can
be seen that the aquatic ribbon fixed-film treatment process
is similar to RBC systems in level of treatment performance,
and can be predicted fairly accurately by the logistic-S
curve as suggested for RBC systems.
CONCLUSIONS
The aquatic-ribbon fixed-film biological treatment
process discussed in this paper is a newly developed
treatment technology, which is capable of achieving removal
of soluble BOD, nitrification, and denitrification from
domestic or municpal wastewaters at a level similar to
conventional rotating biological contactors (RBCs). Under
the experimental conditions in this study, the pilot aquatic
ribbon treatment system was capable of removing more than
91% of the total 5-day BOD at a hydraulic detention of 16
hours; providing 94% nitrification at a hydraulic detention
time of 8 hours; and achieving 87% denitrification at a
hydraulic detention time of 2 hours.
The combination of the liquid-solids separation process
with the aquatic ribbon reactor into one physical unit is
a unique feature of the aquatic ribbon fixed-film biological
treatment system. This feature effectively eliminates the
need for a separate clarifier and thus should result in
a cost-savings for aquatic-ribbon systems compared to
conventional RBC systems. Because the liquid-solids
separation zones in the secondary treatment and
nitrification chambers are connected to the upper reaction
zones, aerobic conditions can be maintained at all times.
This reduces the potential of rising sludge and bulking
problems which are commonly encountered, when septic
conditions occur in the bottom sludge of a conventional
clarifier.
There are areas for further improvements to the aquatic
ribbon treatment system. The pilot system used in this
study has an effecive specific surface area of approximately
5 to 15 times less than typical values for conventional
RBC systems. Consequently, there appears to te a great
potential for aquatic ribbon systems to achieve a level
1172
-------
of performance equal to or better than conventional RBC
systems cost-effectively.
Aquatic ribbon biological treatment systems are a
low-technology, low-energy alternative to other conventional
treatment technologies. Aquatic-ribbon systems can be
easily incorporated into existing lagoons and activated
sludge treatment plants without significant process
modifications to improve levels of treatment and reduce
energy consumption.
1173
-------
REFERENCES
1. Marsh, D. , et al., "Coupled Trickling Filter - Rotating
Biological Contactor Nitrification Process", Jour, of
Water Pollution Control Federation, Vol. 53, No. 10,
pp. 1469-1480, October, 1981
2. Poon, C.P.C,, H.K. Chin, E.D. Smith, and W.J. Mikucki,
"Upgrading with Rotating Biological Contactors for BOD
Removal", Jour, of Water Pollution Control Federation,
Vol. 53, No. 4, pp.474-481, April, 1981
3. Huang, J.C., and V.T. Bates, "Comparative Performance
of Rotating Biological Contactors Using Air and Pure
Oxygen", Jour, of Water Pollution Control Federation,
Vol. 52, No. 11, pp. 2686-2703, November, 1980
4. Lagnese, J.F., "Evaluation of RBC Used to Upgrade
Municipal Plant to Secondary Standards", Paper presented
at the Technical Conference, Water Pollution Control
Association of Pennsylvania, Pittsburgh, Pennsylvania,
April, 1978
5. Reh, C.W., et al., "An Approach to Design of RBCs for
Treatment of Municipal Wastewater", Paper presented at
the ASCE National Environmental Engineering Conference,
Nashville, Tenn., July, 1977
6. Hao, 0., and G.F. Hendricks, "Rotating Biological
Reactors Remove Nitrients", Water & Sewage Works, 121,
Parts I and II, 44, November-December, 1974
7. Poon, C.P.C., H.K. Chin, E.D. Smith, and W.J. Mikucki,
"Upgrading with Rotating Biological Contactors for
Ammonia Nitrogen Removal", Jour.of Water Pollution
Control Federation, Vol. 53, No. 7, pp. 1158-1165,
July, 1981
-1174
-------
ACTIVATED FIXED FILM BIOSYSTEMS
IN WASTEWATER TREATMENT
John W. Smith. Professor. Memphis State University,
Memphis, Tennessee.
Hraj A. Khararjian. Professor. University of
Petroleum and Minerals, Dhahran, Saudi Arabia,
I. OVERVIEW
i
Historically, waste containing organic materials have
been subjected to biological treatment processes to reduce
the impact of the waste on receiving environment. The most ...
widely used concept in urban areas has been the activated
sludge process, a fluid bed system. With waste of increasing.
complexity from municipalities due to industrial influxes
and changing life styles, the basic.biological waste treat- .
ment process has been severely stressed in many locations to
perform satisfactorily. The activated sludge system was
encouraged in the 1970s due to its efficiency of BOD removal
as compared to the fixed film biological reactors, i.e.,
trickling filters. The future of new construction of
activated sludge systems or wastewater treatment in general
is clouded due to the de-emphasis by the present administra-
tion in Washington. While the need for environmental
improvement and management remains, federal funds for the
construction have been severely eliminated. New systems will
have to be justified on .the basis of savings or on the basis
of least cost to the municipality. -With the emphasis on
local financing of wastewater systems, and the need for
stable operation due to shock loading and variable levels of
1175
-------
of toxic materials, the fixed film biological system again
appears to have certain advantages over the activated sludge
process.
The traditional fixed film reactor has been the
"trickling filter" with either a stone or plastic media.
These units were originally installed due to their simplicity
of operation and, their low energy requirements. Their
limited removal efficiency and susceptibility to shock load-
ing were accepted as trade-offs for their advantages. The
introduction of the synthetic or plastic medium greatly
improved the efficiency of operation of the trickling
filter but at the sacrifice of more sophisticated operation
and more energy input. Directionally, advances xrere made in
fixed film reactors with the development of rotating
biological contactors to provide a low energy input system
but also a system which has improved operational
characteristics. As indicated by Benjes (1), the
manufacturers of rotating biological disks claim that the
system offers the following advantages as compared to
conventional fluid bed systems.
1. Simpler operations
2. Capability of meeting a 10/10 effluent standard
without subsequent treatment
3. Final clarifier under flow concentration from
2 to 3 percent
4. Lower energy requirements per pound BOD removed
The standard rotating disk operation is a once through flow
process with no recycling involved. Obviously, this
presents a significant operational advantage ov-»r other
processes. The capability of the biological 5;;. 3tem to meet
oxygen demand requirements in the first stages of the
process has been questioned. Zero dissolved oxygen will
probably occur and potential odors will result. The claim
of achieving medium and low effluent standards without
additional treatment is probably true only of lightly loaded
systems. Excellent quality has been shown to be achievable
of operating plants; however, the ability of the system to
reach extremely low levels of effluent BOD and suspended
solids is questionable. The final underflow concentration
does reach the 2 to 3 percent level which offers an
advantage in the sludge handling facilities. The primary
advantage claimed for the system is the lower power require-
ment. Although this is probably true, based upon pilot
plant and operating studies, for reasonable effluent
criteria (30/30), this does not appear to be a valid claim
1176
-------
as the effluent concentrations become lower. The power
requirements appear to approach those for activated sludge
systems as these lower effluent concentrations are required.
A second type of fixed film system has also been
advocated" by certain manufacturers and engineers as a
competitor for the conventional activated sludge system. As
described by Richter (2), the activated bio filter (ABF)
system is actually a combination of a. fixed film reactor and
a fluid bed system. Biological solids which have been
clarified from the fluid bed system, are recycled back tp the
top of a fixed film reactor and allowed to floxtf through the
reactor into the fluid bed unit. This combination of
reactors has been utilized in potato processing,waste in
Idaho with very satisfactory results. The combination of
the two processes is reported to provide a very stable system
when receiving highly variable influent loads and to
provide a very rapidly settling biological floe. Benjes (1)
also evaluated the ABF system as a competitor to the
traditional activated sludge unit. Although this analysis
(1) was not based upon any one location or operation, the
same general advantages were discussed, i.e., for reliable
operation under varying loads, lower capital and operating .
costs, and simpler operation. It was suggested that each of
the claims had to be evaluated on a site specific basis
and should not be accepted as generalizations.
Research performed at Memphis State University over the
past several years has advanced the information available
for both rotating biological contactors and the ABF process
with not only city of Memphis municipal wastewater but also
synthetic wastewater utilizing glucose substrate. The
results of the MSU investigations will be presented in the
following paragraphs for both pilot plant field studies and
laboratory studies. Under a research grant from the city of
Memphis, Division Public Works, pilot plant studies were
Conducted over a one-year period at the T.E. Maxson waste-
water facility in Memphis. Comparable results were obtained
on an ABF system, a trickling filter utilizing plastic media,
'and a conventional contact stabilization activated sludge
systems. A cost effective analysis of the proposed
expansion of the T.E. Maxson facility prepared by Black and
Vetch engineering consultants in Kansas City, Missouri,
indicated that not only would the ABF system present a lower
capital cost alternative but would also be less energy
intensive and actually reduce the energy consumption of the
existing plant. The laboratory studies utilizing rotating.
1177
-------
contactors constructed of wood followed by a fluid bed
aeration system have indicated that the process can be
operated at high F/M ratios with reasonably consistent
performance. The activated RBC system has also been shown
to be resistant to toxic loads through the development of a
balanced eco system to degrade resistant pesticides.
II. ABF PILOT STUDIES
The ABF pilot plant, as shown in Figure 1, received
wastewater from the aerated grit chamber of the T.E. Maxson
wastewater treatment plant. The degrited wastewater passed
through a circular primary clarification basin then into a
mixing sump where the clarified effluent was mixed with
return sludge from the secondary clarifier and bottoms from
the biotower and pumped to the top of the biotower itself.
From the biotower, a portion of the flow which was not
recirculated was transferred to a short term complete mixed
aeration basin and then through a secondary clarifer where
the biomass was separated and either recycled or wasted to
an aerobic digestor. A similar unit was operated at the
city of Memphis North Wastewater Treatment Plant. The
biofilter was obtained from the Neptune Microfloc Company on
a loan basis for the pilot plant evaluation. The filter, a
24-foot high, 4-foot square unit, contained 21 feet of
horizontal wood medium consisting of wooden slats on one-
inch horizontal spacing with flights arranged every six inches.
The operation in the biotower was controlled by the organic
loading (pounds of BOD per cubic foot of medium) as well as
hydraulic loading (gallons per square foot of medium). The
effluent from the biotower contained high concentrations of
biological solids, both those entering the tower and those
which had fallen from the horizontal wooden medium. The
aeration basin was also provided by Neptune Microfloc and
was considered as an integral part of the fixed film treatment
concept. The flow from the biotower was subjected to an
aeration period of approximately two to four hours in order
to stablize the biological solids and allow for any
additional removal of soluble BOD. Air was supplied to the
system from the main air supply of the treatment plant
through diffusers (coarse bubble) located in the bottom of
the aeration basin. The effluent from the aeration basin was
clarified in a seven-foot diameter secondary clarifer. This
unit was not adequate for the flows placed through the
system as indicated by the surface over flow rate exceeding
1178
-------
DEGRITTED
INFLUENT
PRIMARY
SLUDGE
WASTE
PUMP
PRIMARY
CLARI-
FIER
AEROBIC
DIGESTOR
MOYNO
PUMP
o^
£=1
AIR
Figure 1
ABF PILOT PLANT LAYOUT
1179
-------
1700 gallons per day per square foot at times. The higher
than reasonable overflow rates caused the lower quality
effluent than would have been realized through an adequately
designed and sized clarifier.
The operational theory of the ABF system is relatively
simple but yet not normally experienced in waste treatment
systems. By recycling the underflow from the secondary
clarifier to the top of the biotower, a high microbial
solids level is achieved within the tower itself. The ABF
system utilizes a wooden horizontal medium as opposed to a
plastic medium. The horizontal medium not only allows the
microbial solids growing on the media to remain active
longer because of the moisture content of the wood medium,
but also provides finer droplet formation within the tower
due to the flow pattern around the horizontal wooden boards.
The flow rate through the system was initially set at 1.5
gallons per minute per square foot wetting rate on the
tower. This proved to be an unstable operation condition due
to uneven sloughing of solids from the tower. The unit was
operated slightly over three weeks in this mode and then
the wetting rate increased to 2 gallons per minute per
square foot at which point a uniform, constant sloughing
rate was achieved. At this wetting rate the raw wastewater
flow into the tower was 16 gallons per minute with a recycle
of return activated sludge of 7 gallons per minute and a
recycle from the tower of 9 gallons per minute providing the
2 gallon per minute per square foot wetting rate. Hourly
grab samples and 24 hour 'composite samples were utilized to
evaluate the performance of the system. The evaluation
parameter in the biotower was the loading in terms of pounds
of BOD per 1000 cubic feet. It is believed that the fixed
film reactor (ABF tower) operated as both an absorption
medium for collodial solids and colloidal BOD and as a
biological oxidation region due to the high surface area to
which the wastewater is exposed. The aeration basin is
utilized to allow time for microbial stabilization of the
remaining BOD coming from the bottom of the biotower. The
dissolved oxygen level and detention time in the aeration
basin were varied in this study to evaluate the minimum and
maximum values which could be utilized. Also, the loading
rate across the system in terms of a system food to micro-
organism ratio was observed and correlated with percent
removal as will be discussed in later paragraphs. The
concept of a system F/M ratio is a valid one for this type
of a biological system.
1180
-------
Although the influent was highly variable in BOD and
suspended solids, an equalization basin was not provided
ahead of the ABF tower. The reason for this exclusion'was to
provide a severe evaluation or test of the ABF unit by'itself
to equalize load fluctuations or conversely to absorb shock
loads. The assumption was that if the system performed
adequately without an equalization basin, it would present
many advantages to the city of Memphis or others who were
investigating this type of unit as a retro-fit to an
existing plant.
Beginning in March and April, a B.F. Goodrich plastic
media filter was replaced with the ABF tower supplied by the
Neptune Microfloc Corporation of Coirvallis, Oregon,
containing horizontal wood media. The operational mode was
changed because of the nature of the ABF system. Because the
biotower was operated at a relatively constant hydraulic
loading, wide fluctuations in the organic loading in terms of
pounds per thousand cubic -. foot per day of BOD were
experienced. The biotower, because it contained a high
population of biological solids on horizontal medium, was
also monitored for parameters related to a normal aeration
system, i.e., oxygen uptake rates and sludge volume indices.
The biotower performed extremely well and much better than
was originally anticipated when it was installed. As shown
in Figure 2, the organic loading varied from a high of
452 pounds per day per thousand cubic feet to a low of less
than 110 pounds per day per thousand cubic feet. Even with
this wide fluctuation in loading, the biotower removed a ,
consistent level of soluble BOD. The biotower functions very
similarly to the high rate plastic media filter in the sense
that a significant soluble BOD removal is anticipated. The
effluent from the biotower contains high levels of active
biological solids which render a total BOD analysis not
applicable. The oxygen uptake rates at the bottom of the
biotower were relatively high when compared to the pilot plant
complete mix aeration basin following the biotower (see
Figure 3). The solids settled reasonably well as indicated
by the sludge volume index values. Several studies were
performed by taking hourly samples of the influent and
effluent from the biotower to evaluate the ability of the
biotower to absorb shock organic loadings. These studies are
summarized in Table 1. As can be seen by analysis of the
data in this table, the soluble removal across the tower was
generally greater than 80% and often times reached as high
as 97%.
_
1181
-------
:XA! -"•
'QLZ
'
-------
'09S
i i
'OZ£ "082
'OTiZ '002
W3/3W) .3103 3>IOidT
'091
! N33AX0 3
"021
IJI33dS
i i i
*08 "OTi
1183
-------
TABLE 1
Bio-tower Evaluation
Soluble BOD Removal
Soluble BOD
Sample #
1
2
3
4
5
6
7
8
9
10
11
Influent
270
294
274
276
270
279
363
330
310
315
282
Bio-Tower Eff.
68
42
59
90
45
42
50
51
40
9
24
% Rem.
75
86
79
69
83
85
86
85
87
V /
97
J 1
92
Note:
Flow conditions were 14 gpm primary effluent, 10 gpm bio-tower
recycle, and 7 gpm return sludge.
TABLE 2
Effect of Aeration Time
on Soluble BOD Removal
Elapsed
Time
(hours)
0
1
2
3
4
5
6
7
8
10-17-79
Soluble BOD
Remaining
98
24
16
8
9
5
13
10-2-79
Soluble BOD
Remaining
11
1
6
1
1
3
5
5
5
1184
-------
The amount of aeration which was required to stabilize
the underflow from the biotower was an unknown entity. A
laboratory study was performed with the flow from the tower
bottom to determine the optimum aeration time. This study
was performed by pulling samples of the biotower bottom flow,
aerating it for a prolonged period of time while pulling
samples :of the mixed liquor at various time increments.
Analyses of laboratory and field studies indicated that an
aeration time of less than four hours and probably less than
three hours would be adequate to remove most of the
carbonaceous BOD (see Table 2). As shown by an analysis of
the data in Figure 4, the aeration basin was operated at
varying mixed liquor suspended solids levels with the mean
cell residence time at or about 4% to 5 days. The mean cell
residence time was calculated based on the amount of mixed
liquor solids in the aeration basin. An alternate procedure
using total solids inventory in the system was not utilized.
The F/M level was an arbitrary point as far as the study
was concerned but it tended to provide an indication of the
stability of the system. The oxygen uptake rate in the
aeration basin was relatively low (around 40 mg/1 per hour)
and stable even with a highly variable uptake ratio in the
tower bottom (see Figure 3). The low oxygen uptake rate was
an indication of the low level of soluble BOD entering the
system. With the ABF process, it is almost inappropriate to
speak of the biotower without speaking of the activated
sludge portion. Using a systems analysis approach where the
system considers the total load of the biotower as the food
and the microorganisms in the aeration basin ,as the amount
of microbes, a range of system F/M between .3 to greater than
2 was observed. The total system performance was found to
be less influenced by the loading on the biotower in terms of
pounds per day per thousand cubic feet than on the detention
time in the aeration basin.
Based on an assumed 80% removal of soluble BOD across
the biotower, the soluble loading onto the aeration basin in
terms of pounds per thousand foot of aeration volume per day
became relat ;.vely low. The aeration time proved to be a
critical factor in the level of BOD in the effluent. It
is normally recognized that the operation of a system is
satisfactory when the soluble BOD level is consistently less
than 10 mg/1 in the effluent. The total BOD in the effluent
of the aeration basin was at or above 30 mg/1 for most of the
test, period. It was determined toward the end of the test
program that nitrification was occurring in the effluent
1185
-------
*os
:,n
*se
*oe
-52
'02
-51
-or
'OS!
'SEI
'021
'SOI
'06
- --
Si "09
N39JLXO
'OS
•st
OOOn
'009E '0025
"0082
'OOTiZ '0002 "0091
X H/3H) SS1M NOIJ.b»a3b
'0021
•QOh
•oi
'L
'9
i i
'S '
DH 3-oanis
1186
-------
samples and steps were taken to alter laboratory procedures
to compensate for the nitrogenous BOD. When this
compensation was made, the pilot plant unit achieved the
desired effluent quality of less than 30 mg/1 BOD and
suspended solids.
III. LABORATORY ARBC STUDIES
To advance further the observations made with the ABF
system, a laboratory fixed film unit was developed as shown
in Figure 5. Using wooden dowels for the support media, the
.rotating contactors were exposed to a mixture of wastewater
and underflow solids from the secondary clarifier. Normal
treatment parameters (BOD, COD, SS) were measured on the
influent and effluent of the system. Mixed liquor suspended
solids x\rere also determined for each wheel and the aeration
basin. Influent to the North Treatment Plant as well as a
synthetic wastewater using glucose as a carbon source were
used in the laboratory studies. Shock load conditions
(organic and pesticides) were evaluated as well as normal
operating conditions. The pesticide shock load condition was
evaluated since the North Treatment Plant frequently receives
variable amounts of pesticide type compounds from one
industry.
The results obtained to this point are preliminary in
nature; however, they do tend to confirm several aspects of
the field studies. Detail comparisons are difficult to make
at this time due to the different waste materials and
difference in size of units. Directionally, the laboratory
studies tend to confirm the following observations made in the
field studies.
1. Oxygen uptake rate. The oxygen uptake rate in the
aeration basin in both studies was below 50 mg/l/hr.
Analysis of data from respirometor studies
performed with biofilm removed from the aeration
basin and the wheels indicated a stabilization
phenomenon as opposed to a rapid growth condition of
a fluid bed system. This substantially lower uptake
rate represents a significant savings in operational
costs over a fluid bed system. Supplemental
aeration in the first contactor was necessary in the
laboratory unit to prevent anaerobic conditions
from developing on the first contactor. The
recycling of high concentrations of settled MLSS
resulted in a high uptake in the first contactor.
1187
-------
Figure 5
ACTIVATED ROTATING
BIOLOGICAL CONTACTOR SCHEMATIC
RECYCLED SLUDGE
(INFLUENT
^
ft*
CON-
TACT
WHEEL
NO 1
CON-
TACT
WHEEL
NO 2
CON-
TACT
WHEEL
NO 3
— ^
EFFLUEN1
\
10°
/ "X^
9 /\ l/4" DOWLS ON 1/2" CENTERS
' P'
Q '
l P
//
/ /
1188
-------
2. Stability Under Variable Loads. Using the synthetic
feed, the organic load on the system was varied to
produce a F/M ratio of 0.2 up to 5.0 based on COD
into the system and MLSS in the aeration basin. The
system consistantly achieved better than a 90%
organic removal with F/M ratios greater than 0.8
while at lower ratios, the effluent deteriorated
markedly as shown in Figure 6.
3. Resistance to toxic loads. During the field studies,
the Maxson plant and the ABF pilot plant both
received a shock load of a phosphate based
pesticide. The, fluid bed system exhibited a decrease
in performance efficiency as was expected; however,
the performance of the pilot plant was not affected.
-: To evaluate further this phenomena under controlled
conditions, a series of shock load studies using
chloro-carbon intermediates from the manufacture of
endrin was performed with the laboratory (3). Not
only were the pesticide-type compounds absorbed onto
the biofilm and thus removed from the liquid phase
but biodegradation took place due to the hetrogenous
growth in the biofilm. Two destinctly different
gram negative bacilli and various yeasts were
identified through enrichment culture studies using
the chloro-carbon compounds as the sole carbon
source. Neither pure cultures isolated from the
primary enrichment media nor various reconstituted
mixed cultures would use the chlorinated carbon
compounds as a carbon source. However, four :
successive transfers of mixed cultures from the
primary enrichment media resulted in heavy growth
and chlorinated compound breakdown.
IV. SUMMARY
The results to date, both laboratory and field pilot
plant, indicate that the activated biological contactor (ABC)
concept offers several advantages over fluid bed systems in
wastewater treatment. Considerable energy savings appear
possible due to the lower oxygen uptake rate of the biomass.
Although the oxygen uptake rate of the ABF tower bottoms was
relatively high (^200 mg/l/hr.), the rate in the aeration
basin following the tower was less than 50 mgl/hr. All of
the reactors (contactors and aeration basin) in the
laboratory study exhibit uptake rates less than 50 mg/l/hr.
1189
-------
- O
Q
O
CJ
0)
C/2
VD
60
CO
00
VD
% •JBAOUI3H QOD
1190
-------
The hetrogenous biomass developed in the ABC system provides
significant operational stability to shock loads of organic or
toxic orgin. Several chlorinated compounds normally considered
nonbiogradable and somexvhat toxic xvere assimilated by the
biofilm in the laboratory unit. Similar results were observed
with a phosphate based pesticide shock load on the pilot plant
system.
Work is presently continuing of Memphis State University
to define further the mechanisms involved in the ABC.
Additional data will be developed to explore the application
of a mass transfer model to the system similar to that
evaluated by Famularo (4).
REFERENCES
1. Benjes, EL, "Evaluation of Biological Wastewater Treat-
ment Process," Waste Water Treatment and Reuse Seminar,
South Lake Tahoe, California, 1976.
2. Richter, G. A. and Guthrie, M. D., "ADF/Activated Sludge
Process Control," 42nd Pacific Northwest Pollution
Control Federation, 1975.
3. Redfield, J.; Smith, J. W.; Khararjian, H.; and Peterson,
G., "Pesticide Addition to Wastewaters Treated in Pilot
Activated Sludge, Activated Carbon, and Wood Rotating
Biological Contactor (RBC) Systems," Submitted to
Developments in Industrial Microbiology, 1981.
4. Famuloro, J.; Mueller, J. A.; and Mulligan, T.: "Appli-
cation of Mass Transfer to Rotating Biological Contactors,"
JWPCF 50, 653, 1978.
1191
-------
COMPARISON OF FIXED-FILM REACTORS WITH A
MODIFIED SLUDGE BLANKET REACTOR
Andre* Bachmann, Virginia L. Beard, and Perry L. McCarty,
Department of Civil Engineering, Stanford University,
Stanford, California 94305
INTRODUCTION
Over the last one hundred years, the anaerobic treatment
process has been developed beginning with "Moura's Automatic
Scavenger" and progressing to the conventional complete mixed
and fixed-film anaerobic reactors (1). The conventional
process is generally used for treatment of municipal sludges
and other concentrated wastes. The advantage of this process
is its simplicity in design and operation. The disadvantages
are that a long hydraulic detention time is required for
process efficiency and that reduced effluent quality results
from a high concentration of suspended solids unless some
means of effluent solids separation is provided.
To avoid some of these potential problems, the anaerobic
contact process has been developed (2) where effluent
suspended solids are settled and recycled back into the
reactor. This leads to a longer solids retention time and
therefore permits a significant volume reduction for a given
treatment efficiency.
To be able to treat relatively dilute organic wastes,
care must be taken to obtain a sufficiently long microorgan-
ism retention time within the system. Several alternatives
have been proposed, such as the "anaerobic filter" by Young
and McCarty (3), the "anaerobic attached-film expanded bed"
1192
-------
reactor by Switzenbaum and Jewell (4), the "upflow anaerobic
sludge blanket" reactor by Lettinga (5), and the "anaerobic
rotating biological disc" reactor.
Excellent results have been achieved with the "anaerobic
filter" in laboratory investigations (3,6-7). The advantage
of the system is, its high reliability combined with high
loading capacity and efficiency. ' However, the reactor
requires filling material which influences the economics of
the process. Furthermore, clogging problems may arise with
the filter which can influence its reliability if run over a
long period of time. >
The "anaerobic attached-film expanded bed" reactor has
the advantage of being relatively free of clogging, as the
waste passes in an upward direction through a bed of suspend-
ed media to which the bacteria attach. However, the
disadvantage is the rate of recycling generally required to
keep the media in suspension.
The "upflow anaerobic sludge blanket" reactor is a
modified version of the contact process and is based on an
upward movement of the waste through a dense blanket of
anaerobic sludge. This provides a greater surface area
between the gas and the liquid which is advantageous in
keeping the floating solids from clogging gas ports.
Nevertheless, a large risk with this reactor is the
possibility of further bed expansion and excessive loss of
microorganisms to the effluent. The sludge blanket process
also requires special granular sludge, which is difficult to
develop.
The "anaerobic rotating disc" reactor has also been
proposed for anaerobic treatment of wastewater (8). In this
study, it proved to be a reliable and stable operating react-
or with little potential for clogging while providing a high
void volume. Its major disadvantage is high capital cost due
to its relatively complicated construction.
Perhaps the major obstacles to wider application of the
anaerobic process for industrial waste treatment are the
relative difficulty in .operation and the absence of a simple
and cost-effective design.
The objective of this paper is to present a new process
termed the "anaerobic baffled reactor" which is simple in
form and may offer an economical solution -to the treatment of
intermediate and low strength industrial wastewater.
Furthermore, a unified model for the mathematical description
of fixed-film reactors and sludge blanket reactors will be
presented.
1193
-------
BAFFLED REACTOR CONCEPT
The baffled reactor (BR) for anaerobic treatment is
essentially a series of upflow sludge blanket reactors, but
because of its unique characteristics, it requires no .special
granular growth of bacteria which is difficult to obtain.
The baffled reactor's construction allows a high void volume
and, therefore, clogging problems are essentially elimi-
nated. A schematic diagram of an anaerobic baffled reactor
is shown in Figure 1. This process evolved from initial
studies with an "anaerobic rotating biological disc" reactor
from which it was found that no rotation of the discs was
necessary to obtain reliable reactor performance.
MATERIALS AND METHODS
Reactors. A comparison was made between the performance of
three laboratory-scale anaerobic reactors: the anaerobic
filter, the rotating biological contactor, and the baffled
reactor. All reactors were constructed from plexiglas. The
anaerobic baffled reactor (Figure 1) was 19.3 cm in length
and had a total volume of 1040 cm and an effective volume of
680 cm^. The liquid passed horizontally and around baffles
which served to maintain micoorganisms within the reactor.
Sample ports were placed at 2.5 cm intervals along the
reactor with an additional tap near the effluent port. The
baffled reactor effluent was passed through an inverted
siphon to separate the gas from the liquid. A low speed
peristaltic pump fed the reactor. After the comparitive
reactor evaluation, the anaerobic baffled reactor was scaled-
up to a 6.3 liter liquid volume (Figure 2). Several
modifications were made. The downflow chambers were
narrowed, thus widening the upflow chambers where most of the
cell mass had been found to collect. The lower edges of the
baffles were slanted to route the flow to the center of the
upflow chamber to achieve greater mixing of feed and
solids. The number of ports was increased to facilitate
sampling and wasting of solids should clogging occur.
The anaerobic filter (Figure 3) consisted of a bed of
stones with average diameter of 12 mm through which the
liquor was passed continuously in an upward direction. The
filter had a porosity of 0.42 and a liquid void volume of 400
ml.
The anaerobic rotating biological contactor (Figure 4)
consisted of circular plates connected to a slowly rotating
horizontal shaft. Organisms attached to the surfaces of the
1194
-------
a;
t-i
en
>•
co
i-i
o
•u
o
«fl
a)
OS
PO
o
•1-1
0)
cd
CB
!J
bC
«
•H
03
0)
O
CO
s
• H
1195
-------
CO
60
• H
ft,
1196
-------
FEED
BOTTLE
PERISTALTIC
PUMP
INVERTED,
SIPHON"""
i-GAS ,
WET-TEST
GAS METER
EFFLUENT
.SAMPLE
TAPS
FILTER
EFFLUENT
COLLECTION
BOTTLE
Figure 3. Schematic Illustration of Anaerobic Filter
System.
FEED
BOTTLE
PERISTALTIC
PUMP V
BAFFLES
ROTATING
/ DISCS
C
ORGANISMS
GAS
INVERTED
SIPHON c=a=>
EFFLUENT
COLLECTION
BOTTLE
Figure 4. Schematic Illustration of Anaerobic Rotating
Biological Contactor System.
1197
-------
plates. The liquor passed through in a horizontal direction
and contacted the microorganisms on the plates. The
reactor's liquid volume was 700 ml. The gas outlet was on
top of the reactor; the liquid outlet was on the side. The
outlet level was controlled by an inverted siphon.
Experimental Procedures. A complex protein-carbohydrate
mixture was selected as the substrate for the study.
Nutrient broth, an almost pure protein mixture, and glucose,
a pure carbohydrate, were combine'd in equal chemical oxygen
demand (COD) quantities to make up the feed with tapwater.
Sufficient nitrogen and phosphorous were available in the
nutrient broth for anaerobic growth. A sodium bicarbonate
buffer solution was sterilized separately and then added to
the feed solution to maintain the pH between 6.7 and 7.4.
The reactors were seeded with anaerobic sludge from a
municipal treatment plant, and were operated at a constant
temperature of 35°C ± 0.5°C in a walk-in controlled-tempera-
ture chamber. At the same hour each day, gas production was
recorded at atmospheric pressure with a wet-test meter
(Precision Scientific Co., Model 63115).
Twice a week, routine laboratory measurements of samples
withdrawn from the various levels of the reactors included
COD, pH and total volatile acids. Gas composition was deter-
mined weekly by gas chromatography. Effluent alkalinity was
occasionally monitored. Standard analyses were carried out
according to Standard Methods (9).
Detailed results are presented first for the baffled
reactor, and then its performance is compared, with that of
the other reactors.
RESULTS FOR BAFFLED REACTOR
Start-up and Loading. Initially, the organic loading of the
baffled reactor was kept at 2 kg COD/m d and then was
increased gradually up to a loading of 20 kg COD/m "d.
Reactor performance was constantly monitored. Thereafter,
the organic loading was stepwise reduced back to 5 kg
COD/m3'd. In a third phase, the hydraulic loading was
varied, while maintaining the organic loading at a constant
level of 5 kg COD/m d. The loading scheme, together with
the influent COD concentrations are shown in Figure 5.
Treatment Efficiencies. While increasing loading between 3
and 7 kg COD/m3'd, the COD removal was essentially constant
at about 78%, and decreased to 55% at a loading of 20 kg
1198
-------
10
u_
2
Q
S
0
4
10
tr
Q
£ o
20
q E '
Q
o
^Q fO
O
1 1
\
0 50 100 150 200
DAYS OF OPERATION
Figure 5. Loading Schedule for Anaerobic Baffled Reactor.
1199
-------
COD/m *d (Figure 6). The COD removal was linear up to an
organic loading of 12 kg COD/m d. The maximum attainable
COD removal rate with this small reactor was 10 kg COD/m3*d
for organic loadings above 16 kg COD/m d.
The decreasing loading range showed a significant devia-
tion from the increasing loading range operation. It appear-
ed that this was caused by the formation of gelatinous growth
of bacteria at the head end of the reactor which caused
short-circuiting of substrate through channels in the sludge
blanket.
Gas Production Rates and Methane Content of Gas. The gas
production rate (Figure 7) increased linearly with organic
loading up to 10 kg COD/m3'd. Above that, there was a slight
downward deviation from the linear increase. The percentage
of methane in the gas was 70% up to a loading of 7 kg
COD/m d and methane content then decreased to 50% at a
loading of, 20 kg COD/m3"d. The performance during the
decreasing loading range showed only a small deviation from
the increasing loading range for the gas production rate,
whereas this difference is more evident for the methane
percentage. Overall, the gas production rate and the methane
content of gas coincide well with the COD removal rate and
the treatment efficiency.
Modified Baffled Reactor. In order to improve the perfor-
mance of the baffled reactor, a reduction in the gelatinous
growth of bacteria that occurred in the head end of the
reactor was needed. Also, high volatile acid concentrations
occurring with high substrate concentrations had to be
controlled to minimize buffer additions. , Recirculation of
effluent to dilute the influent waste concentration to about
o
5 to 10 kg/m significantly reduced both of these problems.
This is shown by the intitial results with the enlarged and
modified baffled reactor (Table I). The treatment
efficiencies and gas production rates are higher than with
the smaller baffled reactor without recycle. Still higher
loadings should be possible.
COMPARITIVE EVALUATION OF REACTOR PERFORMANCE
The three different types of laboratory-scale, high-rate
reactors were investigated and their performance is compared
in Table II. All data were taken at steady state, and the
loadings and rates are based on void volume. Two of the
reactors, the anaerobic filter (AF) and the anaerobic
1200
-------
ORGANIC LOAD kg COD / m-d
Figure 6. Baffled Reactor COD Removal Rate and Treatment
Efficiency, Increasing Loading Range " (Circles),
Decrerasing Loading Range (Triangles), and
Influent Substrate Concentration was 8.6 kg
COD/m3.
1201
-------
"O
to-
10 -
O
h-
o
0 5
O
£
V)
100
o^
QJ
LU
50
0
10
20
ORGANIC LOAD kgCOD/m3.d
Figure 7. Baffled Reactor Gas Production Rate and Methane
Percentage of Gas. Increasing Loading Range
(Circles), Decreasing Loading Range (Triangles),
and Influent Substrate Concentration was 8.6 kg
COD/m3.
1202
-------
Table I. Initial Results With Modified Baffled Reactor
Data Set
o
Influent COD Concentration, kg/m
O O .
Hydraulic Loading, m /m day
Recycle Ratio: Or/0, m /m
v o
Organic Loading, kg COD/m day
COD Removal Efficiency, %
O Q '
Gas Production Rate, m /m day
Percent Methane, %
Effluent Volatile Acids, kg/m3
7.3
0.5
0.0
3.5
90
2.3
70
0.34
7.6
1.1.
0.4
8.3
82
4.5
56
0.80
8.1
1.1
2.3
9.0 .
78
4.3
56
0.70
8.3
1.3
2.0
10.6
91
6.9
53
0.40
Table II. Comparison of Reactor Performance
Reactor type
BR
Mod BR AF
ARBC
Influent COD Concentration, kg/m
Organic Loading, kg COD/m day
O O
Hydraulic Loading, m /m day
Percent Efficiency, %
0 O
Methane Production Rate, m /m day
Percent Methane, %
Effluent Volatile Acids, kg/m3
7.1
7.1
1.0
79
2.0
70
0.8
7.6
8.3
1.1
82
2.5
56
0.8
8.0
8.0
1.0
92
2.6
80
0.4
8.0
8.0
1.0
90
2.7
78
0.5
rotating biological contactor (ARBC) are considered to be
fixed-film type reactors. The third type, the baffled
reactor (BR) and the modified baffled reactor (mod BR), as
already described, may be considered as complete mixed sludge
blanket reactors (10). However, as shown in Table II, their
treatment behavior under identical conditions appeared to be
similar. For organic loadings between 7 and 8 kg COD/m d,
the treatment efficiencies varied between 80 and 90% and the
o o
methane production rates ranged between 2.0 and 2.7 m /m d.
1203
-------
This finding was confirmed in a study by Frostell (ID,
who compared an AF system with a sludge blanket reactor.
Over an organic loading range of 2 to 11.5 kg COD/nr "d and a
hydraulic loading range of 0.3 to 1.2 m3/m3'd, the behavior
of the two reactors proved to be essentially identical,
providing a good solids retention was achieved. This
similarity in behavior led to an attempt to model two reactor
types on a unified basis.
MODELING OF BAFFLED REACTOR PERFORMANCE
A fixed-film model was used to evaluate the hypothesis
that the baffled reactor performance can be modeled as a
fixed-film reactor. This evaluation is based upon the fixed-
film model of Williamson and McCarty (12) as further
evaluated by Rittmann and McCarty (13). The model provides a
closely approximate, explicit solution for the flux of a
limiting substrate into a "deep" fixed bacterial film. The
model incorporates concepts of liquid-layer mass transport,
Monod kinetics, and molecular diffusion. The derived
equation provides an explicit solution for the flux into a
bacterial film. The model has been applied with reasonable
success to the AF (12). The reason this model was thought
reasonable for the baffled reactor is that the sludge
particles within the sludge blanket may be considered as
fluidized spheres with surface area through which the solute
must diffuse for bacteria conaumption. The reactor may be
considered as a sequence of five separate chambers connected
in a series as illustrated in Figure 8.
s° s01
0°
o°+on
S1
V,
S2
^
S3
V3
-~ s*
v*
55
V5
ss^
RECYCLE R. S5, 0R
Q°
Figure 8. Flow Diagram of a Model of a Baffled Reactor.
1204
-------
For a complete-mix reactor with specific flow, rate
Q(T ) and specific surface area a (surface area per unit
reactor volume, L ), a mass balance on substrate gives
= -aCS
dt
03 - OS
'
where S is the bulk-liquid substrate concentration (ML."),- S°
the influent concentration, C the variable-order reaction
coefficient and q the variable-order.reaction order. For the
steady-state case, Eq. ] can be solved algebraically for S:
S = S° - (|)CSq (?.)
The model was applied by estimating the specific surface area
a for each of the five chambers from profile-data of profile
1 and applying these values of a to predict the reactor
behavior for other loadings (Profiles 2-4). This approach
assumes a constant specific surface area a for each reactor
chamber over the course of the study, which may not neces-
sarily have been true. Equation (2) must be solved
iteratively for S for each compartment of the reactor and due
to the recycling, another iteration cycle needs to be done
over the overall reactor, assuming ~a certain recycle
substrate concentration. A mass balance of influent and
recycle streams was used to determine the concentration in
the stream entering the first reactor.
Figure 9 presents experimental results and model predic-
tions for the kinetic coefficients and model parameters of
Table III. The kinetic coefficients are based on the values
of Lawrence and McCarty (15), and the model parameters on
values from Williamson (16) and Williamson and Chung (17).
Overall, the predictions were good, although the model as
applied, resulted in a lower rate of removal for Profiles 2,
3 and 4 than actually found. This probably was partly due to
use of a constant diffusion layer depth CL) for all cases.
This depth would, however, decrease with increased gas mixing
at higher loadings, leading to higher removal rates as
noted. Also, the increased mixing at high loading rates
would probably increase the area exposed to the substrate.
Both of these corrections would lead to a better fit between
predicted and measured performance.
A similar evaluation was performed assuming a series of
completely-mixed dispersed growth reactors and using Monod
kinetics. Values for active microorganism concentration were
1205-
-------
PROFILE !
R = 0
0 = 2.4 day'
S°= 7.3 kg/m3
PROFILE 2
R=0.37
0= 7.7 day'1
S°= 7.6 kg/m3
-t-
PROFILE 3
R=2.3
0= 18.2 day-'
S°= 8.1 kg/m3
R=2.0
0=19.2 day'
S°= 3.3 kg/m3
SAMPLE LOCATIONS
Figure 9. Comparison Between Steady-State Predictions Using
Fixed-Film Model for Baffled Reactor and Experi-
mental Data. T = 35°C, Kinetic Coefficients Used
are Shown in Table III. Refer to Figure 8 for
Flow Diagram of Baffled Reactor Model.
1206
-------
Table III. Kinetic Coefficients and Model Parameters
Process
Methanogenesis
Limiting Substrate
Temperature, °C
k, mg COD/mg VSS-da.y
Kg, mg/cm
3
(acetate + proprionate) - COD
35
8
-0.2
Xf, mg VSS/cm5 for
chamber 1 to 5
L , cm
DW, cm "/day
D£, cm /day
10, 8, 8, 8, 5
0.01.
0.8
0.64
determined in each chamber for one loading case, and these
values were used to predict substrate profiles for other
loadings. .The results as summarized in Figure 10 were
poor. Such a model does not give realistic interpretations
of the data as diffusional limitation in bringing substrate
to bacteria is not considered.
SUMMARY AND CONCLUSIONS .
An anaerobic sludge blanket process, termed the baffled
reactor, has been developed which shows excellent promise for
industrial wastewater treatment. It combines . the advantages
1207
-------
01
e o
Q o
O 8
O
1
PROFILE 1
1 1—
R = 0
0 = 2.4 day-'
S°=7.3 kg/m3
PROFILE 2
R = 0.37
Q= 7.7 day"'
S° = 7.6 kg/m3
PROFILE 3
R=2.3
0=18.2 day'1
S°=8.1 kg/m3
O
—I-
^-^^ 0=19.2
— S°=8.3
day-'
kg/m3
-
i i i i 1 5
-01 S1 S2 S3
SAMPLE LOCATIONS
Figure 10. Comparison Between Steady-State Predictions Using
Dispersed Growth Model for Baffled Reactor and
Experimental Data. Refer to Figure 8 for Flow
Diagram of Baffled Reactor Model.
1208
-------
of the anaerobic filter (3), which has a high stability and
reliability due to attachment of the biological solids onto
and between the filt.er media,, and, the upflow a,naerqbic sludge
process (5) in which the microbial mass itself functions as
the support medium for organism .attachment, leading to a high
void volume.
The baffled reactor's construction, however, avoids
certain significant limitations of these other reactors.
Specifically, the risk of clogg'ing and the risk of sludge bed
expansion with resulting high microbial losses have been
minimized. The baffled reactor maintains a high ;void volume
without the need of expensive and operationally work
intensive gas. collection systems or sludge .separation
systems. The over and under liquid flow reduces bacterial
washout considerably, and does not require unusual settling
properties for the microbial culture.
Although scale up factors are difficult to predict, the
influence of gas stirring will be more important in large
reactors due to the fact that gas is produced throughout the
whole column height in the reactor. This should .lead to
greater evolution of gas per unit horizontal cross-sectional
area, leading to more complete mixing in the upper portioii' of
the chamber. It should also result in greater turbulence and
resulting increase in mass transfer rates. These aspects
should lead to better efficiencies in large scale reactors..
However, other scale factors may decrease these advantages.
Thus, large scale experiments are now needed for better
evaluation of the baffled reactor patented.
The biofilm model appears to be generally applicable to
sludge blanket reactors. This suggest a unified approach can
be used in modeling several of the high rate anaerobic
reactors. This aspect is of importance as the model is able
to predict the performance from fundamentals of bacterial
kinetics and mass transport. Although this preliminary work
seems promising, future research is needed to confirm these
observations and to better include effects, of .turbulence on
mass transfer rates to biofilms.
ACKNOWLEDGEMENTS ' . ,
This research was supported by Research Sub-Grant No.
XR-9-8174-1 from the Solar Energy Research Institute, a
division of Midwest Research Institute.
1209
-------
SYMBOLS
The following symbols are used in this paper:
a =
C =
Df
k =
L
Q
q
R
S
o
t
V.
specific surface area per unit reactor volume (L
variable-order reaction coefficient (-);
2,p-l i.
1
);
= molecular diffusivity in bulk liquid (L T
= molecular diffusivity in biofilm (L T );
L~ );
maximum specific rate of substrate
utilization (MgM^V"1) ;
half-velocity coefficient (Mg
length of effective diffusion layer (L);
specific flow rate, (T "> >
variable-order reaction order, (-);
recycle ratio, (-);
bulk-liquid substrate concentration, (ML"-');
influent substrate concentration, (ML );
time, (T);
biofilm density, (ML~3);
volume of individual reactor chamber, (L )
REFERENCES
of Anaerobic
International
Travemuende,
McCarty, P. L., "One Hundred Years
Treatment," Presented at the Second
Conference on Anaerobic Digestion,
Germany, September 7, 1981.
Schroepfer, G. J., et al., "The Anaerobic Contact
Process as Applied to Packinghouse Wastes," Sewage and
Industrial Wastes, Vol 27, 1955, p 460.
Young, J. C. and McCarty, P. L., "The Anaerobic Filter
for Waste Treatment," Journal Water Pollution Control
Federation, Vol 41, 1969, R 160.
Switzenbaum, M.S. and Jewell, W. J., "Anaerobic
Attached-Film Expanded-Bed Reactor Treatment," Journal
Water Pollution Control Federation, Vol 52, 1980, p
1953.
Lettinga, G., et. al., "Use of the Upflow Sludge
Blanket (USB) Reactor Concept for Biological Wastewater
Treatment, Especially for Anaerobic Treatment,"
Biotechnology and Bioengineering, Vol 22, 1980, p 699.
1210
-------
6. Jennet, J. C., and Dennis, N. D., Jr., "Anaerobic
Filter Treatment of Pharmaceutical Waste," Journal
Water Pollution Control Federation, Vol 47, 1975, p
104.
7. Baugh, K. D., et. al., "Characterization and Methane
Fermentation of Soluble Products from Staged
Autohydrolysis of Wood," Proceedings of the Third
Symposium on Biotechnology in Energy Production and
Conservation, Tennessee, 1981.
8. Tait, S. J., and Friedman, A. A., "Anaerobic Rotating
Biological Contactor for Carbonaceous Wastewaters,"
Journal Water Pollution Control Federation, Vol 52,
1980, pp. 2257.
9. Standard Methods for the Examination of Water and
Wastewater. 14th ed. , American Public Health
Association, New York, NY 1975.
10. Hearties,. P. M. and Van Der Meer, R. R. , "Dynamics of
Liquid Flow in an Up-flow Reactor Used for Anaerobic
Treatment of Wastewater," Biotechnology and
Bioengineering, Vol 20, 1978, pp 1577-1594.
11. Frostell, B., "Anaerobic treatment in a sludge bed
system compared with a filter system," Journal Water
Pollution Control Federation, Vol 53, 1981, p 216.
12. Williamson, K. and McCarty, P. L., "A Model of
Substrate Utilization by Bacterial Films," Journal
Water Pollution Control Federation, Vol 48, 1976, p 9.
13. Rittmann, B. E., and McCarty, P. L., "Variable-Order
Model of Bacterial-Film Kinetics," Journal of the
Environmental Engineering Division, ASCE, Vol 104, No.
EE5, Proc. Paper 14067, 1978, pp 889-900.
14. Rittmann, B. E,. and McCarty, P. L., "Design of Fixed-
Film Process with Steady-State-Bio.f ilm Model," Progress
in Water Technology, Vol 12, 1980, pp 271-281.
15. Lawrence, L. A. and McCarty, P. L., "Kinetics of
Methane Fermentation in Anaerobic Treatment," Journal
Water Pollution Control Federation, Vol 41, 1969, Rl.
16. Williamson, K. J., "The Kinetics of Substrate
Utilization by Bacterial Films," Ph.D. thesis presented
to Stanford University, Stanford, California, 1973.
17. Williamson, K. J., and Chung, T. M., "Dual Limitation
of Substrate Utilization Kinetics Within Bacterial
Films," presented at March 19, 1975, 49th National
Meeting of the American Institute of Chemical
Engineers, Houston, Texas.
1211
-------
PART XI: AEROBIC AND ANAEROBIC TREATMENT-SUBMERGED MEDIA
REACTORS
TREATMENT OF HIGH-STRENGTH ORGANIC WASTES BY SUBMERGED
MEDIA ANAEROBIC REACTORS
STATE-OF-THE-ART REVIEW
Yeun C. Wu, Department of Civil Engineering, University
of Pittsburgh, Penna
John C. Kennedy, Department of Civil Engineering, University
of Pittsburgh, Penna
A. F. Gaudy, Jr., Department of Civil Engineering,
University of Delaware, Newark, Delaware
Ed. D. Smith, Environmental Division, U. S. Army
Construction Engineering Research Lab
INTRODUCTION
The anaerobic filter is basically an oxygen-free, media-
filled bed reactor. Anaerobes grow not only in the void
spaces between the media but also on the entire surface of
the media. The wastewater can be distributed from the top
(stationary type reactor) or it can be fed, across the bottom
of the filter (suspended type reactor). The latter type, the
upflow reactor, is more popular than the downflow reactor.
However, both reactors have completely submerged filter media
that is arranged in either a packed bed or fluidized bed
(Figure 1).
If anaerobic filters are classified on the basis of
flow pattern, there are two main types: plug flow and com-
plete mix. When the wastewater passes through an anaerobic
plug-flow filter reactor, the pH decreases initially as a
result of acid fermentation, and then increases in the
direction of the process flow, due to the biological removal
of the generated fatty acids, formation of ammonia and
reduction of sulfates. Since the acidic pH in the bottom
1212
-------
Recycle Line
Out
Out
IN
Figure 1.
IN
Plastic
Media
(A)
(C)
IN
Stones,
Rings, or
Granular
Media
(B)
(D)
Out
Out
IN
Schematic of Anaerobic Filters- (A) and (B)
Downflow Reactors; (C) and (D) Upflow Reactors
1213
-------
section of the filter can potentially inhibit the methane-
forming bacteria, substantial amounts of buffer solutions
are added to the influent waste stream to prevent such pH
decreases.
A completely mixed anaerobic filter would not experience
the pH decrease observed in plug-flow units, since the
mixing maintains a fairly uniform pH throughout the depth of
the filter. The mixing of the filter is achieved by recir-
culating the effluent into the filter at a large recycle: feed
ratio. This in turn would eliminate the need for adding
costly buffer solutions. If the effluent has a sufficient
bicarbonate buffer capacity, it is even able to neutralize
feed solutions with an acidic pH.
The decomposition of waste water by the anaerobic
filter has traditionally been considered to involve two
stages. In the first stage, complex materials such as fats,
proteins, and carbohydrates (COD) are hydrolyzed, fermented,
and converted to simple organic acids and alcohols by
facultative and anaerobic acid-forming bacteria. There is no
waste stabilization during the first step because there is no
methane production. Waste stabilization occurs in the second
stage when the volatile organic acids are converted to carbon
dioxide and methane by a special group of bacteria termed the
methane formers. The raethanogens are the most important group
of bacteria because they carry out the final step in the
overall process. They have slow growth rates and their
metabolism is usually considered rate-limiting in the
anaerobic stabilization of waste. However, the success of
the process is dependent on the presence of both acid-
producing and methane-producing bacteria, with the gas produc-
tion stage responsible for stabilization of the organic
materials.
Figure 2 represents the overall process of anaerobic
digestion according to the traditional concept. While the
distribution of carbon through the various types of inter-
mediates may be correct, it is now recognized that the
methane-forming bacteria do not utilize a variety of organic
acids as substrate. Methane is formed from acetic acid, formic
acid, CO 2 and H2. Thus, the methane fermentation occurs only
1214
-------
COMPU*
WASJS
OTHER
INTERMEDIATES
Figure 2. Metabolic Pathway IP. Methane Fermentation
of Complex Organic Hastes [ taken from
McCarty ( 1 )]
1215
-------
in the lower part of the figure and other organisms are
responsible for the center portion, i.e., the conversion of
various low-molecular weight acids and alcohols to the sub-
strates useable by the methane-formers.
A major advantage of the anaerobic filter is its ability
to produce methane gas. In general, the percentage of methane
in the gas evolved from.the filter is between 70 and 80%.
Theoretically this gas could be used to heat the incoming
wastewater or the reactor, thus increasing the efficiency of
the filter and decreasing power requirements. The ability
to use the methane gas in this way will depend on the quantity
of gas produced. The volume of methane gas produced is
approximately 4.3 to 8.0 ft3 per Ib COD stabilized.
The major factors influencing anaerobic filter performance
include organic loading, hydraulic detention time, temperature,
pH, alkalinity, wastewater characteristics, flow pattern, and
type of filter media. These physical and chemical controlling
factors generally have the same effects on the anaerobic
filter as on the conventional suspended-growth anaerobic
digestion system, except that the fixed-film anaerobic filter
can be operated with higher organic loading, COD:N:P ratio,
and metal concentrations.
Start-up of an anaerobic filter probably is the most
difficult period of operation. Start-up times in experimental
full-scale units have ranged from 10 to 180 days, with the
shorter times corresponding to the use 'of large amounts of
active seed while the longer times were associated with the
use of light seeding.
Analysis of the action of the process during start-up
has indicated three factors of importance. First, the slow
growth of anaerobic micro-organisms, especially at low
waste concentrations and at temperatures below 30°C, does
not permit rapid build-up of biological solids. Consequently,
a large seed mass is needed for rapid start-up. Secondly,
decreases in the buffering capacity of the waste, so that the
pH drops below about 6.5 at any point within the filter for
even short periods of time, increases the starting time
significantly. A third factor affecting start-up time is
related to the physical characteristics of the biological
1216
-------
suspended solids within the filter. During the early stages
of operation a significant fraction of the biological solids
remain finely dispersed throughout the liquid phase _and a
significant proportion washes out with the filter effluent.
At some time after initial seeding , flocculation of the
biological solids occurs in the filter and the solids washout
rate decreases, thereby increasing b9th the rate of active
biological solids accumulation and the rate of waste treatment.
The optimum method of seeding a filter is not known.
Large seed volumes help to start the filter more rapidly by
providing a large viable microbial population, and the large
amounts of suspended solids help to promote the surface
adhesion and flocculation which seem to be essential to good
operation. However, using large seed volumes may also con-
tribute significant amounts of volatile and non-volatile
solids which tend to plug the filter and reduce its effective-
ness for treating wastes.
CASE HISTORIES
Upflow anaerobic filters have been built in England
since 1876 to purify sewage but the organic removal was mainly
thought to be due to adsorption (2). Coulter, et al. (3) and
Witherow, et al. (4) .employed an anaerobic rock filter
following an anaerobic sludge contactor. The combined processes
produced a 65% reduction of BOD with most of it occurring
in the first unit. Winneberger et al. (5) employed an
anaerobic filter following a septic tank and noted a 65%
BOD removal and a 70% suspended solids removal at a 5-day
detention time. Research on anaerobic filters was published
in 1968 by Young and McCarty (6). Loadings ranging from
26.5 Ib COD/day/1000 cubic feet (0.424 kg/day/m3* to
212 Ib COD/day/1000 cubic feet (3.392 kg/day/m3) were tested
with theoretical detention times from 5.4 to 72 hours. COD
removals ranged from 68% to 98%. Young and McCarty found that,
at the same organic loading, the percentage•of COD removal
increased when the concentration of the influent COD increased.
*Kg COD/day/m3 = 0.016 Ib COD/day/1000 ft3
1217
-------
Ham and Boyle (7) found that anaerobic treatment could
effectively stabilize a raw leachate of approximately 10,000
mg/1 COD with a detention time of 10 days and loading less
than 32 Ib COD/day/1000 cubic feet (0.512 kg/day/m3). This
system reduced COD by about 90%. A system with a 12.5-day
detention time and loading of 13.0 Ib COD/day/1000 cubic feet
(0.208 kg/day/m3) increased the COD removal efficiency to
93%. Poree and Reid (8) obtained a COD reduction of 96% for
leachate with a COD of 12,900 mg/1. They obtained a higher
degree of treatment at a loading of 80.2 Ib COD/day/1000 cubic
feet (1.283 kg/day/m3). Chian and DeWalle (9) concluded that
a high strength wastewater with an acidic pH can be success-
fully treated using a completely mixed anaerobic filter. Table
1 summarizes results of studies done with anaerobic filters.
Anaerobic treatment processes are very effective in
removing heavy metals from waste streams by adsorption and
precipitation. Digester studies have shown that heavy metals
are present primarily in the solid phase as opposed to the
aqueous phase. High metal removals result from the separation
of the solids. A municipal digester study by Rudgal (32)
found the influent copper concentration at 226 mg/1 while the
concentration in the effluent supernatant was 11 mg/1. The
concentration in the sludge was as high as 500 mg/1. Chian
et al. (33) investigated the anaerobic filter and obtained up
to 95.5% metal removal. When these removals were calculated
with respect to soluble concentrations, the percentage increased
to 97.1%. This indicates that significant quantities of heavy
metals are associated with the suspended solids leaving the
filter. They concluded that their completely mixed anaerobic
filter was effective in removing heavy metals. And as the
metal concentrations increased in the effluent, the effective-
ness of the filter increased. The metals are removed from
the filter as a slurry in the bottom of the filter as they
precipitate. They also indicated that with decreasing hydraulic
detention time the metal removal percentage decreased while
the metal content in the bottom slurry increased.
SOLIDS PRODUCTION AND EFFLUENT CHARACTERISTICS
^A remarkable advantage of the anaerobic filter is its
ability to retain active biological solids for long periods
of time. There is a continual build-up of solids in the
1218
-------
T3
CD
O
O
CD
_a
ro
4-> t—
E ^
CD CD
3 E
t| — •
CH- CO
LU •
to
ro
t * >
&s o o
C-J E
CD
E
o
•i- CD •
4-> E to
E -r- J-
O) 1 — -E
CD
00
<+-
o
COO
E CD
a T- i— i
o -a --^
CJ) ro >
O ro
i— "t!
r-H
CO
CJ
•1 —
4-5
to
CD •!-
4_> £_
1 CO CD
rc 4-J
3: o
ro
ro
JE
r "i
S~
O
ro
CO
., —
[ S
to
CD
E
t — i
O
LO
O-i
1
o
OJ 1
CO LO
co cr>
i i
LO O
LO r--
co co
•^i- <-H
LO
r--
i
i--.
LO OJ
OJ •*
en
O
+J O
E O r—
O") ^^ d")
S_ 'E
CD 1
4-5 O
CD CD CD
C3 CD CD
Or
O " 1—1—1
•i— OJ ro t-H
E O
O II CD II
•i- 5-
E O 4-5 O
O CD CD O
^: <-) s: c_>
, — . , — ,
co to r^
r — i — i i — '
cn ^ — - o3 ^ — ^
CD TD
•— -r- 0
r— C|_ o
CD ro CO
. _a JE -a
O. CO O
E 10
ro i—i—
CJ) LU 03
0
CD
1
O
i— 1
co
i
<^j-
cn
CO
<=*•
i
OJ
r-l
0
OJ
OJ
1
«^f-
1 —
-^^
cn
E
0
CD
r— CD
ro CD
O * — '
•r- 1
4—*- CD
3 LO
CD OJ
O r-l
rO
E CD II
S- 4->
ro co CD
JE ro CD
a. 3: c_>
4->
CD.
E CO
E r-l
CD^
l""3
o3
CO
•r—
E
E
CD
Q
1
fe« ro
CO >
LO O
E
1 CD
o^
r-.
OJ Q
o
CQ
LO
CT>
1
cn
CO
LO
1--.
co
1
LO
OJ
t-H
-^
CD CO
*[—
M- CD
i— CD
rs co
C/O LO
CD 1
CO O
ro O
3: co
r-H
CD S-
4-J O II
13 E> LO
r— 0-0
•r- -I- CD
o _i ca
E
rO * — ^
E * — '
jr--
oS
E
O
to
r—
• I —
S
1
CO
r*«.
r
«^j-
CD
CD
<^f
1
r^-«
t-H
o
CO
t — i
i
LO
CO
O co *^ i
O "TD" r™ "^^-^
i — -i- O CD CO
•r- -0. CD
E CO OJ
ro CD "
cn-o i— i
t- >, 0
CD JE O O
CD >->O
O -O r— CD
•i — i — CD **
-l-> eC OJ
CD •>
-E " CD II
4-5 CO E
Ei — -r- CD
>i 0 E O
CO -E
0
rn
1219
-------
T3
CU
3
C
-P
E
o
0
1— 1
cu
o
rd
t-H
1 ' ^^
E cn
cu E
3
51 (/)
LU tO
rO
Q >
SSO S
cu
C£
E
O
•r—
4J CU
E E to
0> -r- $-
+J 1— -E
QJ
QC
CO
M-
cro
E O
Q -r- O
^3 ' U *~~^
CO rd ^
0 2?
i — rd
•O
to
O
4J
to
CU T-
rd 4->
3 O
rd
rd
_E
0
s-
a
rd
en
•r-
| ^
to
CU
E
1 — 1
O
o
LO
1
0
[
.— 1
^
01
,
0
CO
o
CO
CO
1
LO
t— 1
o
o
i
o
LO
i- o -^
o o cn
3 0 E
cr •>
Jj CM
in i
to
cu o
s- o
o CD
S-
cu u
CU Q
S- 0
03 O
»— 1
• CM
r— ^ — ^
-{-}
CU
CU
cu
S-
o
u_
1
^
cn
i
CO
""•
co
«xi"
i
CM
t— 1
O
LO
<^j-
1
CM
<£)
0 r-
i— CD ~-^
i — o3 t-H Cn
SO tD
CU 'r- r— 1
to S- 4-> «
O CU CU ~O 1
O Q- O T—
3 rd >=C O C3
•— D. , rd •!- CM
E S- <^ i —
O O) CU O II
«4-> S T3 C->
Q. O) E >>O
CU S_ O i— O
D. 03 C_> CD (—
CM
CM
•• — ^
E
rd
E
•o
^
1 I
r^ LO
cn cn
CO
CO O
<^t- o
to * — '
i
rd II
3 O
O T-
S— -P
•P- rd * — ^
o s- o
CU OJ
0 $_ E ••
r -^i o «— i
•p—
E
rd
O CU
•i— en to 4-5
13 E T3 rd
CU 1 •!-•!- T- S-
4-> ~-^ CJ E O "O
rd Cn ^
^r E rd -E
O JE 4-5 O) O
rd O 4-> E i — -Q
CU O cn O -i- S-
_J O E CJ 4-5 rd
«> CU rd O
r— CD i- CU i—
, — OO 4-* •*-* O X
•t- t/1 rd > CU
^- ll ^: •—
-o ^: o cu Q-
E C2 cn rd O) E
rd CD -r- CU S- O
i
CU O)
• — co i — cn
rd CM rd •> — '
s ** — ^ s
cu cu
Q Q
o25 o3
E E
rd rd
•r- *t—
JZ J^
0 CJ
cn
to
0 0
••- 0
E cn
rd "
cnCM
S- t-H
O
It
cu
•r- O
r— CJ
1220
-------
CD
E
03
E
O
q-
S-
CD
Q.
CD
• 1 —
U-
O
•i —
.a
o
i-
cu
03
E
CD
1 ^
q-
o
CO
E
O
4_5
03
cn
•r—
[ ^
CO
CD
>
E
l — i
I-H
cv
f'*m
&SO O
o E
CD
a;
E
O
4-> CD CO
E E S-
j. —
cu
CO
q-
cno
E 0
Q -i- O
o -a i— i
O 03 --^
0 >,
i — 03
-a
••^,
i— i
co
CJ
^o
CO
CD -r-
4-* £—
CO CD
03 4->
2 CJ
03
03
f*
CJ
5-
0
03
en
• ! —
4_5
CO
CU
>
r*
l — i
O
LO
CM
1
CO
CO
cn
i
CO
LO
CM
[•^
I
LO
,
"*
CM
f-H
CM
1
LO
•
CO
CM
1
O
s_
O-
r— 1
•a cu -^.
E •»-> en
03 03 E
-a -o CD
•r— >^j CD
O c~ O
•=C O CM
-Q i— 1
CD S- 1
i — O3 LO
•r- cj r--
-t-> CM
03 E II
r— -i- Q
O CD CD
> -l-> CJ
1 ^
€. ^ — «
03 VO
CJ
O
^~
o?S
cn
E
O
o
LO
CO
I— 1
1
o
LO
•*
CO CO CD
00 CO .1—
<*"1
1 1 3
r—
O LO O
CO LO CO
CO
CO
1
CO
I— 1
CD
<^-
co
1
CD
0
t— 1
1
o
S-
03
CJ
r—
cn ~-^
E cn
•r- E
co
CO O
CD O
O LO
O *•
S- CD CO
Q- -M
03 II
"0 S-
o -a a
o >>o
U- .E CJ
03
E
•i — *
.— CD
03 r-H
^" *• — '
oQ
s-
cu
E
E
, —
QL.
1
O
CO
1
0
CO
CM
I— i
CD
1^
CO
^~
"- — _
en
E
"O O
••- 0
CJ «^f
•^C **
CO
CJ
•r- II
•4-3
CD Q
O O
0 0
0 0
S- 0
Q- *•
CO
o
4-> OJ II
03 -i->
4-1 CO ' '
O 03 O
0- 12 CJ
.
r—
O3 * *
CM
•<-> i—i
CD — '
f\
CL. '
c^
O
-1 ^
^~-
•r—
03
Q-
o
CO
CO --—.
CO CD
• 1 CO
O CO
a i-\
CM
Q
1 CD
r- CJ
0
CO CD
i —
^^ ^
LO CO 3
CO 1 — •
CM
CM
r*^
CO
CM
, ,
**^^
cn
E
OJ
-M 0
CO CD
03 O
~^Z r*
CO
C~ |
0 O
S- 0
03 O
00 CM
4-> 11
03
OJ O
.E CD
3 CJ
^~
03
^_>
CU
rt r—
S_ 1—
CD S- 03
•4-5 O -M
-Ei — CO
CJ >> S-
•r- 03 O
a; i— Q
t__
03
>
O
E
CD
t_
.^ — %
LO
i-H
1
CO
1— 1
•» — ^
1221
-------
CU CH
=3 E
LU
03
O)
D:
to
01
CO
CO
CO
-t-> a)
c E
(U-
OJ
CO
oo
CM
oo
cno
a ••-
to >>
o to
CO
CM
CM
CM
00
OO
CM
CO
(LI 10
-(-> -I-
tO S-
03 (!)
3 -M
o
(O
ro
CD
O O
to O
cu cr>
• CM
-o
c a
ro o
_l C_3
•o
i— i— i-
o
O CD
O E
O) 3
i— CT
-Q-i-
fO_l
O)
i— O
+J-1-
+->-!->
ai a»
C/7-
r— -a
CD
cu
en
CO ||
o >>
o
O CD CO
•=C i-H
CO II
co en ii
+•> "o >>Q
to 13 O O> O
tO i— O <~ c ~>
3 OO CJ 3 —-
'QJ -~^ E <£)
T3 CJ) S- i— t—I
s e oo
u_ c: i
03
t3 4-5 O
•r- 13 CD
O CQ CD
D_ CD
CO «
co
co a
CJ CM
CU -r- CD
o o o
Q)
ta
CO
O)
o;
O)
O)
eo
00
oo
•4-J
CD
(O
^x:
to
.CO
N
1222
-------
CD
O
CJ
CD
JD
rO
-p I—
E ^>,
CD CD
3 E
<4- in
LU .
CO
ro
&s o o
CJ E
CD
a;
o
•i — •
4-> CD CO
E E i-
-M P "^
CD
Cd
CO
Cf-
o
o
o>o
E i-H
o -5 >,
CJ ro ro
0 T3
_Q
i —
C/)
CJ
•i —
4-)
CD CO
00 £
ro CD
-rg _^_)
O
to
&.
ra
e~
CJ
i-
o
CD
4J
CO
CD
E
I — i
1
^. •
cn
i
CO
00
^-f.
CXI
0
o
LO
1
CD
oo
oo
1 1
CD CD i—
ra >-, CD
CD E » — i
U •!-
=C >• 1
"CJ . — O
•r- 0 0
O CJ CD
•,
r— F*"-.
O CD
•r- II
4-> CD CD
cu -o 4-> a
O >-, to CD
0
co"
CXI
^
4J
•i —
1
CXI
CTl
1
CXI
r~-
^-J-
CXI
o
I — 1
CTl
1
CD
CTi
CO
-a
•1 — 1^
O CD O-^
=3; TD ocn
>> CDE
0 JE «
•r- CD "=3-
4-5 "O CXI
CD i —
O ra 1
«C E
S- O
•a o o
•r- u_ o
CJ
^^ r~~ r**1^
O i-H
O E
•i— ra II
E -E
S- 4-> Q
O CD CD
U_ 2! CJ
1
r--
cn
j
^j-
^
S-
CJ
^
^
0
o
cn
*— — ^ i
"=3" O
cn oo
5-j. ^-j.
CXI CXI
0
LO
1
o o
CXI LO
LO «d-
-a
O-^ CJ CD O^
t__j(_» c^ ~^^ f~~^rn
OE XOE
(_) _E «
LO -i— CD <=J- i —
co 4-> -a cxi ^.
CD i — CD
i cj ra i E
CD
f O *• IT) **
CTl j II
CD E -E C
4-> Q S- 4-> Q -r- Q
co CD O CD O ro O
LU cj u_ 2: cj a cj
1
CD
oo
1
o
LO
CO
cn
1
•=3-
,_j!
t-H
o
LO
CXI
i-H
1
CO
LO
t— 1
c
o
zs
cr
1 -^
CD
4-> E
E
CD 0
E ^1"
4-> LO
ro "
CD r—l
^- r-4
I—
II
ro Q
CD CD
HZ CJ
cn
CXI
to
O
C
E
O
1223
-------
cu cr
=3 E
CO
cu
CT>
in
CO
+J CU
CU T- S-
cu
o;
CO
oo
co
co
co
CM
CD t—1
Q T- >,
CJ CO •
O '
i— JQ
co
en
O)
o
10
0) S- r-
O) to
CO O)
3 -)->
O
CO
cu
4->
•M
CU
CD -
-M
(O
.
CT)
-(-> o
fO O)
•t-> Z!
> to un
CU
O
LU
01
CO
r— O
(O CO
cu
CO
O 3
ai
CO
i— O)
CU
CU
rsl
1224
-------
filter due to biological synthesis and no appreciable loss
occurs until the filter becomes filled with highly concentrated
biological solids. The filter can be operated for long periods
of time before sludge wasting is needed because a low percentage
of the COD removed is synthesized into biological solids. A
material balance done by Chian and DeWalle (9) indicated that
93% of the COD removed could be accounted for by the methane
gas formed.
Observation of the physical characteristics of the sludge
within the filter indicates that the solids lie loosely in
the interstitial spaces rather than becoming attached to, the
surface of the media. Plummer et al. (10) noted that solids
in their filter units were not attached strongly to the
media or to the sides of the unit.
It has been recognized that several factors determine the
amount of solids leaving the anaerobic filter. -Dennis and
Jennett (18) observed that the solids concentration was mainly
determined by the hydraulic detention time. Young and McCarty
(6) observed a gradual accumulation of solids in the anaerobic
filter, during which time the effluent suspended solids remained
low. Only after the filter reached its maximum storage capacity
would the effluent solids show an increase. When the influent
waste contains solids, no net removal may be observed. In
addition, it is noted that the porosity of the filter may have
a large effect on the solids concentration which is to be expected
since a lower void ratio will increase the collision frequency
between the solids and the filter media.
The effluent characteristics cannot be categorized 'for
all anaerobic filters. The properties of the effluent will be
dependent on on-site conditions such as influent concentrations
and loading rate. Most effluents contain a rather low concentra-
tion of suspended solids, a portion of which is readily settle-
able. • All filters will have an effluent with a very low
dissolved oxygen content 1 This is a disadvantage of the filter
because the effluent cannot be discharged to the environment
until the D.O. is raised to minimum requirements. The pH of
most reactor effluents should'range between 7.0 and 9.0.
1225
-------
BEHAVIOR UNDER ADVERSE ENVIRONMENTAL CONDITIONS
Anaerobic filters are much more resistant to variations
in waste load and environmental factors such as pH and tempera-
ture than originally thought. Laboratory scale filters have
shown rapid adjustment to four-fold surges in influent load.
There are four major types of transient loading or opera-
ting conditions that can affect filter performance: (a)variations
in loading as a result of changes in flow rate or waste strength,
(b) intermittent operation, (c) changes in pH, temperature, and
waste composition, and (d) influx of organic toxins or heavy
metals.
(a) Variations in Loading
Tests by a number of investigators (6,14,18) have shown
that anaerobic filters can readily accept variations in load
caused by changing either or both the flow rate or the waste
strength, without being upset permanently. The following equa-
tion proposed by Young (34) suggests that changing the flow
rate and waste strength simultaneously so that the organic load
remains constant will not cause the effluent BOD to change.
S =
e
PV
(1)
where
S = effluent BOD concentration
S = influent BOD concentration
K- = proportionality constant
K_ = proportionality constant
Q = flow rate
P = porosity of filter media
V = volume of reactor tank
L ~ organic loading to the filter
The equation also indicates that if the flow rate is held
constant the "steady state" effluent BOD concentration will
vary directly with changes in influent BOD. Data from El-
Shafie and Bloodgood (17) and Dennis and Jennett (18) support
this conclusion.
1226
-------
Short-term loading increases having a1 duration of one or
two hydraulic detention times can be expected to produce only
slight, short-term changes in effluent quality or gas pro-
duction. Long-term changes, however,' will cause the COD and
volatile acid profiles, and the population dynamics and solids
concentrations, to shift until a new "steady state" level of
performance is reached. Four-fold instantaneous increases in
loading have caused no permanent adverse effects on filter
performance.
Chian and DeWalle (9) tested their complete mix anaero-
bic filter for its ability to withstand shock loads. .When the
detention time, based on feed stream flow alone, was reduced
from 42 to 7 days for a ,1-day period, only a small change in
the pH was observed. The pH decreased from 7.2 to 6.9 when
the detention time was reduced to 4.25 days for a 1-day period.
The gas production did not show a corresponding increase, as
high concentrations of organics were present in the effluent
of the unit, which reduced the organic removal efficiency
to 54%. Solids in the filter were resuspended at the higher
flow rates which was indicated by the large differences in
the values for the filtered and unfiltered COD and the increase
in suspended solids. When the detention time was restored
to 42 days, after the shock load, the effluent COD and suspended
solids concentrations returned to values only slightly higher
than those observed before the shock load. Based on these
tests, it was concluded that the buffer capacity of the unit .
is sufficient to prevent large pH depressions at relatively
short detention times. However, at detention times as short
as three days, a large portion of the organics leave the unit
in the effluent stream and-the suspended solids experience an
increase.
(b) Intermittent Operation .
The second major type of operating condition that can
affect filter performance is intermittent operation. The
possibility has been tested by investigators (6,18) by
stopping all flow and load to filters for several days, as
might be used in practice for weekend operation, and there
was' essentially no loss in COD removal capacity or gas
production efficiency upon restarting at full load. After
1227
-------
fourteen days of down-time without feed, the COD removal
efficiency decreased to a greater extent, but full COD
removal capacity and gas production were achieved after
only three to four days of operation. Longer periods without
feed might be expected to produce a low quality effluent
for quite a long period of time after restoring waste load.
(c) Changes in pHr Temperature, and Waste Composition
Anaerobic filters, once "steady state" operation is
achieved, become quite resistant to pH changes. Rapid
recovery has been observed in filters exposed for a twelve
hour period to pH levels as low as 5.4. Exposure to pH levels
of 9.3 for as long as four days has caused only temporary
loss of treatment efficiency. While gas production and COD
removal were impaired'at these extreme pH values, the filters
recovered completely within one to two days after restoring
pH to normal levels.
In general, anaerobic filters are expected to perform
best at temperatures greater than 25° C. Filters have been
used successfully to treat potato processing wastes at
temperatures as'low as 19° C, but too little information is
available from which to draw significant conclusions about
the effect of lower temperatures on filter performances.
Variations in waste composition are expected to produce
little adverse response in anaerobic filters unless there is
an associated influx of toxic materials. However, the
composition of the waste significantly affects the solids
produced in the system. Biological solids will accumulate
much faster when treating a carbohydrate waste because synthe-
sis of biological solids is greatest with carbohydrates, and
at high loadings problems such as plugging or solids washout
might be encountered. However, in studies to date, no filter
has been reported to have become plugged beyond use.
(d) Organic Toxins or Heavy Metals
It has been generally assumed that anaerobic processes
are unable to cope with waste streams containing toxicants and
therefore are unsuitable for treatment of many wastewaters.
1228
-------
Toxicants do alter the kinetic parameters of methanogens and
thus increase their generation time and decrease pollutant
removal efficiency. However, these adverse effects can be off-
set by proper attention to solids retention time. Proper
acclimation procedures can also increase the threshold concen-
tration of toxicants which cause inhibition. The magnitude of
the toxic effect generated by a substance can be reduced sig- \
nificantly if the concentration is increased slowly. This
involves a process of acclimation which represents the adjust-
ment of the biological- population to the adverse effect of the
toxin. The acclimation process in a mixed microbial popula-
tion may involve any or all of three mechanisms: (1) mutation
of one or more species in the population; (2) selection of the
least sensitive species in the population; (3) alteration of
the metabolism of one Or more species to overcome the metabolic
block produced by the toxic material. All these mechanisms
may interact. In any case, resistance to a toxic substance
often involves an increase in the concentration of the substance
which can be tolerated rather than acquisition of total
resistance to the substance at any level. When the concentra-
tion of a toxic substance is increased slowly, the microbial
population can acquire increased resistance through all of the
mechanisms available to it. However, if a large concentration
of toxic material is introduced suddenly, the effects are quite
different than when the same concentration is reached after
an adequate series of acclimations because no time is allowed
for any of the available mechanisms to operate, and most of the
population will be destroyed. In evaluating data from toxicity
studies for design purposes, the engineer should consider the
test conditions used and whether toxic materials may be intro-
duced into the waste stream to be treated as a slug dose of
high concentration or as a constant component to which a
population may become acclimated. Speece, et al. (35) showed
that methanogenic bacteria could acclimate to toxicant con-
centrations that were 100 times greater than the concentra-
tions which caused inhibition of unacclimated cultures. They
found that continuous, increments of nickel chloride could be
added to the feed of a filter with no adverse effect on gas
production. An increase from 200 to 400 mg/1
decreased gas production but it resumed when nickel additions
were stopped. They were able to acclimate the filter to a
sulfide concentration of 1000 mg/1. Long term acclimation of
122P
-------
a filter to increasing levels of continuous sodium additions
showed no adverse effect up to additions of 7500 mg/1 as Na+.
Formaldehyde added to the feed of the anaerobic filter caused
no inhibition of gas production up to 400 mg/1. Additionally
they were able to acclimate the filter to 600 mg/1 of acrylic
acid and to acrolein at a concentration of 100 mg/1.
Parkin, et al. (36) found that cyanide and ammonia
toxicity were fairly reversible while chloroform, formalde-
hyde and sulfide exhibited some irreversible toxicity. Nickel
showed signs of irreversible toxicity depending on the concen-
tration. •
The early warning of possible metal toxicity is given by
a gradual decrease in gas production and an increase of the
effluent COD. This can be anticipated if strict attention is
given to the influent wastewater so that the operator will
know if inhibitory concentrations of metal ions are entering
the filter.
KINETIC ASPECTS
Because of high cell mass concentration and immobilization
of cells within the attached film, the substrate utilization
rates per unit volume of the biological reactor are high;
displacement of the culture composing the film by inactive cells
in the influent to the reactor is less likely than displacement
of a culture in dispersed/suspended growth, and a fixed-film
reactor is less susceptible to upset by shock loadings of sub-
strate and/or toxic metals.
As a result of these advantages inherent in the attached-
growth wastewater treatment systems, many researchers have
attempted to model growth and substrate removal by biofilms.
Earlier studies described the removal kinetics in terms of sub-
strate concentration existing in the bulk liquid phase. Recent
investigations have generally described substrate removal
kinetics in terms of simulated substrate concentrations in each
layer of the biofilm. The response of the entire film is pre-
dicted by the sum of layers. The most complete biofilm models
were developed by Williamson and McCarty (37) and Dewalle and
Chian (38) on the basis of two competing mechanisms, diffusion
1230
-------
(O
5-
13
OO
Biofilm
Interface
c • • • - - - - ., ,
-, Unit Surf ace'.;r
>V\ -Ar_.e-a,-.A . '•;'.;
Diffusion
Limited
Surface
F1ux '
|*- :;i."- >>/-.;;
... ., Ina'cf'ive
Substrate In lActive Mass Magg .
Bulk Solution r[a^ Li ^
u_ Biofilm
Lc
Figure 3. Substrate Concentration Profile Within
A Biofilm [ taken from Williamson &
McCarty (37)]
1231
-------
and metabolism. Basically, all models considered mass trans-
port from the bulk liquid to the biolayer by diffusion
through a boundary layer of stagnant liquid covering the bio-
layer. (See Figure 3.) Williamson and Mccarty's model is
represented by a second-order, non-linear ordinary differential
equation.
d2S
dz
dS
= - (-
X
dt
(2)
The equation states that the second derivative of substrate
concentration with respect to biofilm depth z is directly re-
lated to the utilization rate of the rate-limiting substrate
(~ <3Sc/dt) and biomass concentration (X) , but is inversely
related to the diffusion 'coefficient (Dc) within a biofilm.
Additionally, in Eq. 2 it is also assumed that the rate of
utilization of substrate at any depth within the biofilm follows
the Monod relationship; that is -(dSc/dt) = (qS )/(K.
which q is the maximum utilization rate of the rate-limiting
Sc),
in
substrate
Sc and KS is the Monod half-velocity coefficient.
This equation does not possess an explicit solution.
However, it^can be solved for the two limiting cases of the
Monod equation. When the substrate concentration S (at z = 0)
is much greater than the half-velocity concentration, K , Eq. 2
becomes a zero order kinetic equation: S
d2S
dz
(3)
D
and the biofilm substrate concentration (S ) can be computed by
using the equation given below: c
i
jX
But, when S is much less than K
order kinetic equation: s
qxs.
> (4)
Eq. 2 becomes a first
(5)
1232
-------
Under the above condition, the relationship between the
biofilm substrate concentration, S , and the other controlling
parameters such as q, X, K , L , and z is defined as:
S C
s = s
,c s
Cosh [(qX/D K ) (L -z)]
c s • c
Cosh (qX/D K )T L
OS C
(6)
Obviously, the problem in using this biofilm model is the
need to determine the thickness of the diffusional boundary
layer (L ) and the values of maximum substrate utilization
rate (q) and diffusion coefficient (D ). These are difficult
to measure in a fixed-film biological system.
Because of .these reasons, model modifications were sug-
gested by DeWalle and Chian (38) . According to Pick's law
of molecular diffusion, the mass transfer irate (3M/3t)
through a surface area A is proportional to the concentration
gradient of the substrate at the interface:
3.M
3t
= -AD
as
c_
c 3.z
(7)
By substituting the term 3S /3z in Eq. 7 into the integrated
C
form of Eqs. 3 and 5 for a unit cross-sectional area and at
z = L , DeWalle and Chian were able to define the. .rates of
•mass transfer as follows:
and
dM
dt
r~
dM
dt
= S
b/
c
if S » K
s s
if S « K
s s
(8)
(9)
where S,_ 'is' the substrate concentration in the bulk liquid.
L in Eq. 8 can be approximated using the equation proposed
by Pirt (39) and Saunders and Bazin (40) :
1233
-------
L =
C v
2r> -
c
qX
(10)
Eq. 8 indicates that the rate of mass transfer is indepen-
dent of substrate concentration, but directly proportional to
the thickness of the biolayer (L ) and the concentration of
attached biomass (X). If cell attachment is uniformly
distributed, the rate of substrate removal is also proportional
to the specific surface area of solid medium because Eq. 8 is
derived from a unit cross-sectional area. More importantly,
Eq. 8 further indicates that at very high substrate concentra-
tion the rate of substrate removal is highly dependent upon
the specific surface area of solid medium within the system
due to the zero order kinetics of the reaction. This is true
in particular when submerged filters are employed for the
treatment of high-strength organic wastes.
Eq. 9 states that the rate of mass transfer is independent
of biofilm thickness, but directly proportional to the bulk
substrate concentration (S,) and the square root of the bio-
mass concentration (X). Since for a given substrate, q, D ,
X, and K are not expected to vary greatly, Eqs. 7 and 8 can
be reduced to :
1 dM
V dt
Sb
and
1 dM A
V dt K2 v b
(11)
(12)
where V is the reactor volume and A/V is the specific surface
area in the reactor. K^ and K_ are coefficients based on
zero and first order kinetics, respectively.
The effects of effluent characteristics, specific surface
area and flow velocity on the substrate removal rate were dis-
cussed by the same investigators. They concluded that : (a) At
low substrate concentrations the removal rate as predicted by
Eq. 11 increased linearly with effluent concentration. The
effluent concentrations have a finite value when the removal
1234
-------
rate approaches zero. On the contrary, at high substrate con-
centrations a satisfactory linear relationship was not obtained
from Eq. 11 when the reciprocal substrate removal was plotted .
versus that of the substrate concentration, (b) An increase in
Kp was always observed as a result of increasing specific sur-
face area although no definitive correlation between the rate
of substrate removal and the value of the increasing specific
surface area could be established, (c) The calculated K2 (V/A)
values appeared to have a straight relation with flow velocity
when plotted on log-log paper. .This means that the rate of
substrate removal can be improved by minimizing the diameter
of the submerged filter column or maximizing the height of the
column in order,to obtain highar velocities. Practically
speaking, a tall column has one obvious disadvantage, i.e.,
a relatively larger volume in the last portion of the.column
is in contact with a low substrate concentration, which in
turn tends to reduce the substrate removal rate.
Kinetic models are presently not supported, by sufficient
analytical data and it is necessary that more research be
done in this area. ...
SUMMARY -..-••
The advantages inherent ,in the anaerobic filtration pro-
cess suggest that it is worthy of consideration as a basis for
full scale waste treatment facilities. The fixed film
anaerobic process is well suited .to handling large organic
loads. High COD removals can be .achieved, particularly with
high strength wastes, without the high operating costs that
are associated with other treatment operations which use aera-
tion or physical-chemical methods. Anaerobic digesters also
product a useable methane gas which could.be used to heat the
reactors. The methane production is almost certain to be well
in excess of the necessary requirements for heating purposes
and could possibly be used for additional heating of the treat-
ment facility. Research has shown that-the filter .can success-
fully treat many types of waste efficiently. More quantitative
•data from field installations and laboratory pilot plants are
heeded to establish design criteria for this treatment process.
1235
-------
REFERENCES
McCa^ty, P.L.? "Anaerobic Treatment pf. Soluble Wastes,"
Advances in Water Quality Improvement, The University of
Texas Press, Austin, Texas, p. 336, 1968.
Truesdale, G.A. et. al., "Journal Institute Public Health
Engrs.," Vol. 60, p. 273, 1961.
Coulter, J.B., Soneda, S., and Ettinger, M.B,, "Anaerobic
Contact Process for Sewage Disposal," Sewage Industrial
Wastes, Vol. 29, p. 468, 1957.
Witherow, J.L., Coulter, J.B., and Ettinger, M.D., "An-
aerobic Contact Process for Treatment of Suburban Sew-
age," Journal Sanitary Engineering Division ASCE, Vol. 84,
1958.
Winneberger, J.H., Menar, A.B., and McGauhey, P.H.,
"A Study of Methods of Preventing Failure of Septic Tank
Percolation Fields," Second Annual Report. Sanitary
Engineering Research Laboratory, University of California
Berkeley, 1961.
Young, J.C., and McCarty, P.L., "The Anaerobic Filter for
Waste Treatment," Journal Water Pollution Control Federa-
tion, Vol. 41, R160, 1968.
Ham, R.K., and Boyle, W.C., "The Treatability of Leachate
from Sanitary Landfills," Progress Report to the Environ-
mental Protection Agency, Office of Solid Waste Management
Washington, D.C., University of Wisconsin, 1971.
Foree, E.G., and Reid, V.M., "Anaerobic Biological Stabili-
zation of Sanitary Landfill Leachate," University of
Kentucky, Technical Report, UKY TR 65-73-CE 17, 1973.
Chian, E.S.K., and DeWalle, F.B., "Treatment of High
Strength Acidic Wastewater with a Completely Mixed Anaero-
bic Filter," Water Research, Vol. 11, p. 295, 1977.
Plummer, A. H., Malina, J.F., and Eckenfelder, W.W., "Sta-
bilization of a Low Solids Carbohydrate Waste by an Anaer-
obic Filter," Proceedings 23rd Purdue Industrial Waste
Conf., Purdue Univ., P. 467, 1968.
Clark, R.H., and Speece, R.E., "The pH Tolerance of Anaer-
obic Digestion," 5th International Water Pollution Research
Conference, 1970.
Pailthorp, R.E., Richter, G.A., and Filbert, J.W., "Anaer-
obic Secondary Treatment of Potato Process Wastewater,"
presented at 44th WPCF Convention, San Francisco, Oct., 197L
Richter, G.A., and Mackie, J.A., "Low Cost Treatment for
High Strength Waste," Cornell, Rowland, Hayes, Merryfield,
and Hill, Consulting Engineers, Oct. 1971.
10.
11.
12.
13,
1236
-------
14. Taylor, D.W., "Full Scale Anaerobic Trickling Filter
Evaluation," Proceedings 3rd National Symposium of Food
Processing Wastes, EPA Report R 2-72-018, Nov. 1972.
15. Dorstall, K.A., "Centennial Mills Starch-Gluten Plant,
Spokane Washington," unpublished report, EPA, Pacific
Northwest Environmental Research Laboratory, Oct. 1974.
16. Campbell, G.R. ^ Hydroscience, Inc., communication with
John L. Mancini, 1973.
17. El Shafie, A.T., and Bloodgood, D.E., "Anaerobic
Treatment in a Multiple Upflow System," Journal Water
Pollution Control Federation, Vol 45, p. 2345, 1973.
18. Dennis, N.D., and Jennett, C.S., "Pharmaceutical Waste
Treatment with an Anaerobic Filter," Proceedings 29th
Purdue Industrial Waste Conf., Purdue Univ., 1974.
19. Wilson, A.W., and Timpany, P.L., "Anaerobic Contact
Filter for Treatment of Waste Sulphite Liquor,"
presented at the 59th Annual Meeting of the Canadian
Pulp and Paper Association, Montreal, Quebec, Jan. 1973.
20. Hovius, J.C., Fisher, J.A., and Conway, K.A., "Anaerobic
Treatment of Synthetic Organic Wastes," EPA Report, Pro-
ject 12020 Dis., Jan. 1972.
21. Foree, E.G., and Horsley, E.E., "Advanced Studies of
the Submerged Anaerobic Filter for Brewery Press Liquor
Stabilization," Technical Report UKY 48-72-CE-14,
University of Kentucky, 1972.
22. Tadman, J.M., Dow Chemical, personal communication with
James Mueller, 1973.
23. Chian, E.S.K., and DeWalle, F.B., "Evaluation of Leach-
ate Treatment; Vol. 1: Characterization of Leachate,"
Environmental Protection Technology Series, EPA-600/
2-77-186a, Sept. 1977.
24. Khan, A.N., and Siddiqi, R.H,, "Wastewater Treatment by
Anaerobic Contact Filter," Journal Environmental Engi-
neering Division, ASCE, EE2, p. 102, 1976.
25. Haug, R.T., "Anaerobic Filter Treats Waste," Water and
Sewage Works, Feb. 1977.
26. Hakannson, H. et. al., "Anaerobic Treatment of"Whey;and
Whey Permeate in Submerged Filters," Inst. Vatten^-
Luftuardsforsl, Pub. B, 4Q6,:.1977.
27. Switzenbaum, M.S., "Anaerobic Expanded Bed Treatment of
Wastewater," Proceedings of the Seminar/Workshop on
Anaerobic Filters. Howey-In
-------
rial Waste Conf., 1979.
29. Donovan, E.J., "Treatment of High Strength Wastes by an
Anaerobic Filter," Proceedings of the Seminar/Workshop
on Anaerobic Filters. Howey-In-Hills, Florida, 1980.
30. Arora, H.C., Chattopadhya, S.N., and Routh, T., "Treat-
ment of Vegetable Tanning Effluent by the Anaerobic
Contact Filter Process," Journal Water Pollution Control
Federation, p. 584, 1975. (G.B.)
31. Jewell, W.J., Switzenbaum, M.S., and Morris, J.W.,
"Municipal Wastewater Treatment with the Anaerobic At-
tached Microbial Film Expanded Bed Process," Journal
Water Pollution Control Federation, Vol. 53, p. 482,
1981.
32. Rudgal, H.T., "Effect of Copper Bearing Wastes on Sludge
Digestion," Sewage Works Journal, Vol. 18, p. 1130,
1946.
33. Chian, E.S.K., DeWalle, F.B., and Brush, J., "Heavy
Metal Removal with Completely Mixed Anaerobic Filter,"
Journal Water Pollution Control Federation, Vol. 51,
p. 22, 1979.
34. Young, J.C., "Performance of Anaerobic Filters Under
Transient Loading and Operating Conditions," Proceedings
of the Seminar/Workshop on Anaerobic Filters. Howey-In-
The-Hills, Florida, 1980.
35. Speece, R.E. et. al., "Metane Fermentation Toxicity
Response," Proceedings of the Seminar/Workshop pn Anaer-
obic Filters, Howey-In-The-Hills, Florida, 1980.
36. Parkin, G.F., Speece, R.E., and Yang, C.H., "A Comparison
of the Response of Methanogens to Toxicants: Anaerobic
Filter vs Suspended Growth Systems," Proceedings of the
Seminar/Workshop. Howey-In-The-Hills, Florida, 1980. .
37. Williamson, K., and McCarty, P.L., "A Model of Substrate
Utilization by Bacterial Films," Journal of Water Pollu-
tion Control Federation, Vol. 48, p. 9, 1976.
38. DeWalle, F.B., and Chian, E.S.K., "Kinetics of Substrate
Removal in a Completely Mixed Anaerobic Filter," Bio-
technology and Bioengineering, Vol. 18, p. 1275, 1976.
39. Pirt, S.J., J. Appl. Chem. Biotechnol., Vol. 23, p, 389,
1973.
40, Saunders, P.T., and Bazin, M,J., J. Appl. Chem, Bio-
technol., Vol. 23, p. 847, 1973.
1238
-------
-------
-------
-------
------- |