FINE PORE DIFFUSER SYSTEM EVALUATION FOR THE
GREEN BAY METROPOLITAN SEWERAGE DISTRICT
by
James J. Marx
Rusk Environment and Infrastructure, Inc.
(formerly Donohue and Associates, Inc.)
Sheboygan, Wisconsin 53083
Cooperative Agreement No. CR812167
Project Officer
Richard C. Brenner
Water and Hazardous Waste Treatment Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
-------
DISCLAIMER
Development of the information in this report has been funded in part by the U S
Envoynmenta Protection Agency under Cooperative Agreement No. CR812167 L the
American Society of Civil Engineers. The report has been subjected to Agency p^r and
administrative review and approved for publication as an EPA document Mentfonof trad
names; or commercial products does not constitute endorsement or recoimnenda on ?or™
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products and
practices frequently carry with them the increased generation of materials that, if
improperly dealt with, can threaten both public health and the environment The U S
Environmental Protection Agency (EPA) is charged by Congress with protecting the' "
Nation s land, air, and water resources. Under a mandate of national environmental laws,
the Agency strives to
formulae and implement actions leading to a compatible balance between human activities
and the ability of natural systems to support and nurture life. These laws direct EPA to
perform research to define our environmental problems, measure the impacts, and search
for solutions. ^«»vu
The Risk Reduction Engineering Laboratory is responsible for planning, implementing
and managing research, development, and demonstration programs to provide an
authoritative, defensible engineering basis in support of the policies, programs, and
regulations of EPA with respect to drinking water, wastewater, pesticides, toxic substances,
solid and hazardous wastes, and Superfund-related activities. This publication is one of the
products of that research and provides a vital communication link between the researcher
and the user community.
As part of these activities, an EPA cooperative agreement was awarded to the American
Society of Civil Engineers (ASCE) in 1985 to evaluate the existing data base on fine pore
diffused aeration systems in both clean and process waters, conduct field studies at a
number of municipal wastewater treatment facilities employing fine pore aeration and
prepare a comprehensive design manual on the subject This manual, entitled "Design
Manual -Fine Pore Aeration Systems," was completed in September 1989 and is available
hrough EPA'sCenter for Environmental Research Information, Cincinnati, Ohio 45268
IcA P°rt N?' EPA/625"1-89/<>23). The field studies, carried out as contracts under the
AbCE cooperative agreement, were designed to produce reliable information on the
perforMance and operational requirements of fine pore devices under process conditions
fhese studies resulted in 16 separate contractor reports and provided critical input to the
design manual. This report summarizes the results of one of the 16 field studies.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
ui
-------
PREFACE
In 1985, the U.S. Environmental Protection Agency funded Cooperative Research
Agreement CR812167 with the American Society of Civil Engineers to evaluate the existing
data base on fine pore diffused aeration systems in both clean and process waters, conduct
field studies at a number of municipal wastewater treatment facilities employing fine pore
diffused aeration, and prepare a comprehensive design manual on the subject This manual,
entitled "Design Manual - Fine Pore Aeration Systems," was published in September 1989
(EPA Report No. EPA/725/1-89/023) and is available from the EPA Center for
Environmental Research Information, Cincinnati, OH 45268.
As part of this project, contracts were awarded under the cooperative research
agreement to conduct 16 field studies to provide technical input to the Design Manual. Each
of these field studies resulted in a contractor report. In addition to quality assurance/quality
control (QA/QC) data that may be included in these reports, comprehensive QA/QC
information is contained in the Design Manual. A listing of these reports is presented below.
All of the reports are available from the National Technical Information Service, 5285 Port
Royal Road, Springfield, VA 22161 (Telephone: 703-487-4650).
1. "Fine Pore Diffuser System Evaluation for the Green Bay Metropolitan Sewerage
District" (EPA/600/R-94/093) by J J. Marx
2. "Oxygen Transfer Efficiency Surveys at the Jones Island Treatment Plants, 1985-
1988" (EPA/600/R-94/094) by R. Warriner
3. "Fine Pore Diffuser Fouling: The Los Angeles Studies" (EPA/600/R-94/095) by M.K.
Stenstrom and G. Masutani
4. "Oxygen Transfer Studies at the Madison Metropolitan Sewerage District Facilities"
(EPA/600/R-94/096) by W.C. Boyle, A. Craven, W. Danley, and M. Rieth
5. "Long Term Performance Characteristics of Fine Pore Ceramic Diffusers at Monroe,
Wisconsin" (EPA/600/R-94/097) by D.T. Redmon, L. Ewing, H. Melcer, and G.V.
Ellefson
6. "Case History of Fine Pore Diffuser Retrofit at Ridgewood, New Jersey"
(EPA/600/R-94/098) by J.A. Mueller and P.D. Saurer
7, "Oxygen Transfer Efficiency Surveys at the South Shore Wastewater Treatment
Plant, 1985-1987" (EPA/600/R-94/099) by R. Warriner
iv
-------
8. "Fine Pore Diffuser Case History for Frankenmuth, Michigan" (EPA/tfOO/R-94/100)
by T.A. AHbaugh and S J. Kang
9. "Off-gas Analysis Results and Fine Pore Retrofit Information for Glastonbury,
Connecticut" (EPA/600/R-94/101) by R.G. Gilbert and R.C. Sullivan
10. "Off-Gas Analysis Results and Fine Pore Retrofit Case History for Hartford,
Connecticut" (EPA/600/R-94/105) by R.G. Gilbert and R.C. Sullivan
11. "The Measurement and Control of Fouling in Fine Pore Diffuser Systems"
(EPA/600/R-94/102) by E.L. Barnhart and M. Collins
12. "Fouling of Fine Pore Diffused Aerators: An Interplant Comparison"
(EPA/600/R-94/103) by C.R. Bafflod and K. Hopkins
13. "Case History Report on Milwaukee Ceramic Plate Aeration Facilities"
(EPA/600/R-94/106) by L.A. Ernest
14. "Survey and Evaluation of Porous Polyethylene Media Fine Bubble Tube and Disk
Aerators" (EPA/600/R-94/104) by D.H. Houck
15. "Investigations into Biofouling Phenomena in Fine Pore Aeration Devices"
(EPA/600/R-94/107) by W. Jansen, J.W. Costerton, and H. Melcer
16. "Characterization of Clean and Fouled Perforated Membrane Diffusers"
(EPA/600/R-94/108) by Ewing Engineering Co.
-------
ABSTRACT
The Green Bay Metropolitan Sewerage District retrofitted two quadrants of their
activated sludge aeration system with ceramic and membrane fine pore diffusers to provide
savings in energy usage compared to the sparged turbine aerators originally installed
Because significant diffuser fouling was expected, the two diffuser types were closely
monitored over an 18-month period. The oxygen transfer efficiencies of the full-scale
systems were measured using off-gas techniques. The effects of diffuser fouling and the
effectiveness of cleaning procedures were evaluated in the laboratory using dynamic wet
pressure and steady-state clean water oxygen transfer tests. Although fouling was
significant on both types of diffusers, cost-effective cleaning procedures were developed The
ceramic disc diffusers provided better long-term performance than the membrane tube
diffusers, which irreversibly lost oxygen transfer efficiency with time in use. Collectively
the fine pore diffuser systems provided a 30% savings in electrical power usage compared to
the original sparged turbine aerators.
This report was submitted in partial fulfillment of Cooperative Agreement No.
CR812167 by the American Society of Civil Engineers under subcontract to Rusk
Environment and Infrastructure, Inc. (formerly Donohue and Associates, Inc.) under the
partial sponsorship of the U.S. Environmental Protection Agency. The work reported herein
was conducted over the period of 1986-1987.
VI
-------
CONTENTS
Foreword
D f 111
Preface
• •••....... ••••••••»••«•-•••••••••.........,.,,. |y
Abstract
_,. vi
Figures
Tables V!"
u
Introduction . 1
Description of Facilities 2
Study Methods „
Oxygen Transfer Efficiency .....' R
Diffuser Evaluation . 11
Foulant Analysis •-
Results and Discussion . . . 13
Operational Procedures and Goals 13
Oxygen Transfer Efficiency Versus Time \ 15
Off-Gas Testing 15
Efficiency Factor ^n
Oxygen Transfer Efficiency Versus Air Flow Rate 22
Apparent Alpha Versus Time 23
Pilot Diffuser Evaluations 3n
Summer 1986 " " " ' ,"
October 1986 '.'.'.'.'.'.'.'.'. 34
November 1987 35
Grid Diffuser Evaluations
November 1986 , 41
July 198? ;;;;;;;;;;;;;;;;; ;;;;;;;;;;;;;;;;;;; 45
Foulant Analysis --
Cost Analysis „
Capital s_
Operation __
Maintenance ' ^
Summary ,_
Conclusions '
Recommendations ,_
References ,q
Appendices
A. Efficiency Factor Calculation Description 70
B. Statistical Review of the Off-Gas Test Results 73
C. Operational Cost Data Summaries ' ] * ] 98
vit
-------
FIGURES
Number
1 Schematic of the Green Bay Wastewater Plant 3
2 Aeration Basin Zones and Grids 5
3 Hood Positions Used for Off-Gas Testing 9
4 Oxygen Transfer Efficiency (Off-Gas Method) Versus Time . . . 17
5 Contact Basin Efficiency Factor Versus Time 21
6 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Contact Basin Ceramic Diffusers Before Cleaning 24
7 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Contact Basin Ceramic Diffusers After Cleaning 25
8 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Contact Basin Membrane Diffusers Before Cleaning 26
9 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Contact Basin Membrane Diffusers After Cleaning 27
10 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Reaeration Basin Ceramic Diffusers Before
Cleaning - 23
11 Oxygen Transfer Efficiency Versus Position and Air Flow
Rate - Reaeration Basin Membrane Diffusers Before
Cleaning 29
12 Photographs of Fractured Ceramic Diffusers 39
13 Ceramic Diffusers Test Results - Grid Units Removed
November 1986 42
14 Membrane Diffusers Test Results - Grid Units Removed
November 1986 44
15 Ceramic Diffusers Test Results - Grid Units Removed July
19»7 46
16 Membrane Diffusers Test Results - Grid Units Removed July
1987 48
17 Photographic Comparison - New Versus Hose-Brush-Hose
Cleaned Contact Basin Membrane Diffuser 49
18 Photographic Comparison - New Versus Hose-Flush-Hose
Cleaned Contact Basin Membrane Diffuser 50
19 Photographic Comparison - New Versus Hose-Brush-Hose
Cleaned Reaeration Basin Membrane Diffuser 51
20 Photograph of Foulant From Contact Basin Ceramic Diffuser 53
21 Power Cost Versus Time m 61
Vlll
-------
TABLES
Number Page
1 Fine Pore Aeration System Design Criteria . . 4
2 Fine Pore Diffuser Configuration Summary 6
3 Clean Water Oxygen Transfer Test Results 7
4 Diffuser Preventive Maintenance Summary 14
5 Operating Data for May 1986 - October 1987 ]', ' ' 15
6 Statistical Analysis of Alpha'-SOTE Data 18
7 Variable Air Flow Rate Testing - May 14, 1986 23
8 Apparent Alpha as a Function of Time in Service 30
9 Diffuser Characterization Results - New Membrane Discs 31
10 Diffuser Characterization Results - New Membrane Tubes 32
11 Ceramic Diffuser Test Results - Pilot Units Removed
June 11, 1986 32
12 Ceramic Diffuser Test Results - Pilot Units Removed
June 22, 1986 . . . - 33
13 Ceramic Diffuser Test Results - Pilot Units Removed
October 21, 1986 35
14 Membrane Diffuser Test Results - Pilot Units Removed
October 28, 1986 36
15 Final Characterization of Used Ceramic Diffusers 37
16 Final Characterization of Used membrane Diffusers 40
17 Comparison of New and Used Membranes 52
18 Ceramic Diffuser Foulant Analyses Summary 55
19 Membrane Diffuser Foulant Analyses Summary 56
20 Summary of Capital Costs 58
21 Electrical Power Usage and Cost Summary 50
22 Estimated Maintenance Costs 53
23 Alternative Aeration Systems Cost Summary 55
IX
-------
INTRODUCTION
The Green Bay Metropolitan Sewerage District (GBMSD) operates an
activated sludge secondary treatment facility. Historically, about half of
the plant's total electrical energy needs were for operating the sparged
turbine aeration system. In the spring of 1983, Donohue & Associates
was retained to prepare a Predesign and Feasibility Report for Aeration
System Modifications.1 The report published in December 1983 was
supported by field testing conducted during the summer of 1983. Of the
alternatives studied, fine pore ceramic diffusers showed potential for
significant energy savings. However, a major concern with the fine pore
diffusers was fouling and the effects that fouling could have on the overall
system economics by increasing operating pressure, decreasing oxygen
transfer efficiency and increasing maintenance costs for diffuser
cleaning.
Because of the concern over fouling, the report recommended that fine
pore diffused air aeration be tested by installing a full-scale ceramic
diffuser system in one quadrant of the aeration complex. Since the 1983
field tests showed that a substantial portion of the foulant materials was
soluble in acid, the recommended ceramic system was to include an in
situ acid gas injection system. Preliminary calculations presented in
Appendix B to the 1983 Report showed that to operate the fine pore
system in a mode equivalent to the design year loadings, approximately
40 percent of the influent flow would have to be fed to the fine pore
system and 60 percent to one or more of the sparged turbine systems.
In the fall of 1984, Donohue was asked to evaluate the Wyss Flex-A-Tube
membrane diffuser as an alternate fine pore system. The manufacturer
claimed that the flexible membrane tube diffusers normally did not foul
and, if fouling occurred, it could be controlled by flexing the units. A
letter report dated November 30, 1984 concluded that the flexible
membrane diffuser system would also be worth evaluating in a full-scale
test. Based on the data available at the time, the ceramic diffusers were
expected to have a higher oxygen transfer efficiency than the membrane
diffusers when the systems were clean. However, since fouling was
expected to adversely affect the oxygen transfer efficiency, the flexible
membrane units could have produced better oxygen transfer efficiency
in the long run if in fact the diffusers did not foul.
Because the cost effectiveness of ceramic disc and flexible membrane
tube diffusers was believed to be similar, and the fouling characteristics
and maintenance procedures were potentially different, it was decided
to install and test both types of fine pore diffusers. Testing two fine pore
diffuser systems also provided the opportunity to try to treat all of the
-------
wastewater through the test quadrants and achieve substantial energy
savings during the test period. !
DESCRIPTION OF FACILITIES
The GBMSD activated sludge facility treats a mixture of metropolitan
(Metro), and pulp and paper mill (mill) wastewaters. The Metro
wastewaters comprise municipal and industrial wastes including a
seasonal contribution from vegetable canning industries. The Metro
wastewaters receive preliminary and primary treatment before entering
the activated sludge system. The mill wastewaters receive primary
treatment at the mills before they enter the plant through separate
interceptor sewers and are pumped directly to the activated sludge
system. Plant recycle streams which include decant liquor from a Zimpro
sludge heat treatment process are returned to the primary effluent
channels.
The activated sludge process is a contact stabilization system consisting
of four quadrants (Figure 1). Each quadrant includes a 73.3 feet by 244
feet by 20.5 feet water depth contact basin and a 36.3 feet by 244 feet by
22.5 feet water depth reaeration basin. Each contact basin was originally
equipped with twelve 125 HP sparged turbine aerators. The reaeration
basins had six 75 HP aerators. The process air is supplied by four 2500
HP centifugal compressors each having a capacity of 44,500 scfm at a
discharge pressure of 12 psig.
Quadrant number 2 was retrofitted with ceramic disc diffusers while
quadrant number 4 received the flexible membrane tube diffusers. The
design criteria for selecting the number and distribution of fine pore
diffusers in the aeration basins are presented in Table 1. The design
criteria were based on the results of the field studies conducted in 1983,
operating experience with the sparged turbines, and information
supplied by the manufacturers of the fine pore aeration systems. The
alpha values used for design, 0.68 in the contact basins and 0.90 in the
reaeration basins, were selected based on the results of the 1983 field
work and with consideration given to other design constraints such as
maximizing the power savings during the test period by treating the
entire plant flow through the two fine pore diffuser equipped quadrants.
However, this design strategy was only viable if the actual alpha values
were equal to or greater than the values used for design. In the event that
alpha was lower or plant loadings were higher than expected, additional
aeration capacity could be added by placing one or more of the sparged
turbine systems in service.
The contact basins were divided into three zones and subdivided into ten
grids. The number of diffusers per grid was the highest in the tank inlet
-------
I
!s
o
Q.
O
O
'§
CD
c
_o
"55
O
o
o
-------
X
Table 1 • Fine Pore Aeration System Design Criteria
Biochemical Oxygen Demand (BODs)1
Average Day (50th Percentile)
Maximum Month
Maximum Day (99th Percentile)
*y
Oxygen Requirement
Contact Basins
Size
Number
Oxygen Demand
Oxygen Demand Profile Zone 1
Zone 2
ZoneS
Alpha
Beta
Diffuser Submergence
Reaeration Basins
Size
Number
Volume
Oxygen Demand
Oxygen Demand Profile
Alpha
Beta
Diffuser Submergence
161,000 Ib/day
202,000 Ib/day
239,000 Ib/day
1.0 O2/lb BODs applied
733' x 244' x20.5' deep
4
75% of quadrant total
55%
30%
15%
0.60-0.75
0.95
19.1 feet
363'x 244'x 22.5'deep
4
1.5 xlO6 gal each
25% of Quadrant Total
Uniformly Distributed
0.90
0.95
19.1 feet
1 Calculated from data reported for January 1982 through April 1983. A total of 478
values were used.
Based on 1983 off-gas results.
zone and lowest in the tank outlet zone. The reaeration basins were
divided into six grids. Since the oxygen demand was expected to be
relatively uniform, the number of diffusers per grid was constant The
layout of the aeration zones and grids are shown in Figure 2.
The number of diffusers installed in each grid and the design air flow
rates per diffuser are summarized in Table 2. The design average air flow
rates used for the ceramic diffusers, 2.1 scfm/unit in the contact basin
and 1.9 scfm/unit in the reaeration basin, were relatively high compared
to normal practice of designing for about 1.25 scfm/unit. The high air
rates in the contact basin resulted from providing the maximum number
of diffusers that would physically fit in the basin inlet grids and then using
-------
Reaeration
Basin 4
Contact
Basin 4
Contact
Basin 2
Reaeration
Basin 2
RAS
MIXED
UQUOR
t
R4. 1
R4.2P!
*
R4.2 !
R4.3
R4.4
R4.5
R4.6P!
*
R4.6 !
: ;
*C4.6SP C4.6NP*
i • i
i i
C4.5/6S C4.5/6N
Zone 3
C4.4S C4.4N
— —Zone 2 —
C4.3NP!
*
C4.3S C4.3N !
C4.2NP!
*
C4.2S C4.2N !
', C2. ISP
#
! C4.1S C4.1N
g»l
MIXED
LIQUOR
*
i
i
*C2.6SP , C2.6NP*
Zo e 3
C2.4S C2.4N
IC2.3SP
*
! C2.3S - C2.3N
« 1 1
1C2.2SP
*
! .C2.2S C2.2N
C2.1NP!
*
C2.1S C2.1N !
RAS
i
R2. 1
1R2.2P
*
! R2.2
R2.3
R2.4
R2.5
IR2.6P
*
! R2.6
Primary Effluent
•f-
Mil! Waste
Pilot Headers — * —
FIGURE 2
Aeration Basin Zones and Grids
-------
Table 2 - Fine Pore Diffuser Configuration Summary
No. of Diffusers Diffusers Diffuser
Zone Grids jerGrid per Zone Density1
Contact Basin No. 2 - Ceramic Disc Diffusers
1 4 805 3220 0.54
2 4 490 1960 033
3 2 474 948 0.16
Basin Total 6128
Reaeration Basin No. 2 Ceramic Disc Diffusers
1 6 358 2148 0.24
Contact Basin No. 4 - Membrane Tube Diffusers
1 4 616 2464 0.41
2 4 350 1400 0.23
3 2 378 756 0.13
Basin Total 4620
Reaeration Basin No. 4 - Membrane Tube Diffusers
1 6 233 1398 0.16
Density = diffusers per square foot of floor area
Design
Air Rate
scfm/unit
2.1
2.1
2.1
1.9
2.9
2.9
2.9
2.6
the same diffuser air rate for all the grids in the basin. The higher diffuser
air flow rates at design average loadings normally would not be used
because the higher oxygen transfer rate required to handle the peak
diurnal variations in organic loading would have to be achieved at
diffuser air flow rates outside the most efficient range of operation. It
was possible in this application because the mill wastes provided
relatively constant diurnal loadings.
If the number of diffusers had not been constrained by space, more
diffusers could have been provided for a minimal increase in the capital
cost. If the design average diffuser air flow rate had been 1.25 scfm/unit
rather than 1.9 scfm/unit, the clean water oxygen transfer efficiency
would have been 6 to 8 percent higher.
The membrane system was designed for an air flow rate per diffuser of
2.9 scfm per unit in the contact basin and 2.6 scfm per unit in the
reaeration basin. The manufacturer's recommended design range for the
tube diffusers was between 2.0 and 5.0 scfm per unit
Before the aeration equipment was approved for manufacture, shop tests
were conducted to determine the clean water oxygen transfer efficiency.
The tests were conducted in accordance with the ASCE Standard
-------
procedure and witnessed by Jim Marx of Donohue & Associates and
Mike Pierner of the GBMSD. The Sanitaire system was tested at their
test facilities in Glendale, WI.2The tests were conducted in a 34.3-foot
by 6- foot rectangular tank. The Parkson system was tested by Gerry Shell
Environmental Engineers in a 21-foot diameter test tank. The results
of the clean water tests are summarized in Table 3.
The Sanitaire aeration system met the specified oxygen transfer
efficiencies. The Parkson diffusers performed substantially better than
the specified standard oxygen transfer efficiency (SOTE) of 27.5 percent.
Since the number of diffusers was not reduced to account for the higher
SOTE, the projected design air flow rates per membrane diffuser were
reduced to about 2.0 scfm/unit in the reaeration basin and 2.5 scfm/unit
in the contact basin.
In addition to installing the fine pore diffusers, inlet flumes were built in
the contact basins to distribute the incoming wastewater across the entire
width of the basins. The flumes were deemed necessary because
grid-type diffuser systems are considered by many to provide poor lateral
mixing. The 4'-6" wide by 7'-10" deep flumes have two rows of one-foot
diameter holes on 2'-0" centers in the floor. The number and^ize of the
holes were selected to provide several inches of headloss. No holes were
Table 3 - Clean Water Oxygen Transfer Test Results
Zone
Specified Air Flow Rate Diffuser
(scfm) Density
/1000 cu ft /diffuser sq ft/unit
SOTE
Specified
SOTE1
Sanitaire Ceramic Discs - Contact Basin
1 57.3 2.2 1.9
2 32.4 2.1 3.2
3 16.2 2.1 6.4
Mean Weighted Average2
Sanitaire Ceramic Discs - Reaeration Basin
All 21.8 1.9 4.3
Parkson Membrane Tubes - Contact Basin
1 59.9 2.9 2.4
2 34.4 2.9 4.2
3 183 2.9 7.9
Mean Weighted Average2
Parkson Membrane Tubes - Reaeration Basin
All 20.8 2.6 6.3
31.0
35.7
38.3
33.6
36.2
31.4
33.8
31.7
32.2
34.9
26.8
35.4
31.9
30.2
34.6
27.5
275
27.5
275
27.5
Standard oxygen transfer efficiency, i.e., clean water at 20°C, barometric pressure of 29.92 in. hg and
zero dissolved oxygen.
Weighted by the air flow rate to the grid.
-------
provided at the center of the flume where the mixed liquor enters
through a single large opening.
Five removable pilot headers were provided in each contact basin and
two were provided in each reaeration basin so diffusers could be
obtained for visual observation and testing without having to drain the
basins. Each removable header was equipped with four diffusers at a
submergence of about 17 feet The headers were placed at the inlet,
middle and outlet of the contact basins and at the inlet and outlet of the
reaeration basins (Figure 2). The diffusers on the removable headers and
two diffusers per fixed grid were provided with pressure taps for
monitoring pressure drop across the diffuser elements, also known as
dynamic wet pressure (DWP).
A two-stage in-line air filter was provided on the main header to the
ceramic disc quadrant. The air filter was originally designed to remove
more than 99.9 percent of all particles greater than or equal to 0.3
microns in size. After startup, problems were encountered with excessive
headless across the filter so the second stage elements were changed to
a coarser medium. The filtration efficiency of the modified filter was 98
percent of all particles greater than or equal to 1.0 micron.
The centrifugal compressors did not require modification because the
12 psig discharge pressure was more than adequate for the 19-foot
diffuser submergence provided.
STUDY METHODS
Oxygen Transfer Efficiency
The oxygen transfer performance of the two fine pore aeration systems
was evaluated by conducting tests on the full-scale systems using the
off-gas method.4 The off-gas tests were conducted by the Ewing
Engineering Company (Milwaukee, Wisconsin). A typical off-gas test
was conducted in two days. One day was used to survey 24 positions in
each of the contact basins. The second day was used to survey 12 positions
in each of the reaeration basins. The sampling positions are shown in
Figure 3. So that data from the two systems would be comparable, several
sampling hoods were used and the analyzer was moved in a serpentine
pattern starting at the inlet of one basin and moving across to the second
basin before moving to the next set of sampling locations further down
the length of each basin. In this way, corresponding areas in each basin
were sampled and analyzed at about the same time during the sampling
day.
-------
•i
CO
m
c CO
1 *~*
0>
i
CI3
a>
a:
a
"
2
<
**
2
§
U)
8
1
N
U
a
in
(S
o
4
m
10
d
u
in
U)
o
a
Lt
4
^
a
o*
S
(M
cJ
a
01
u
z
u
m
m
'
i>
us
in
PI
0
a>
z
in
Cj
I
in
o
»
in
in
3
S
n
o
at
T
a
J5
£
u o u
111™
o 3 S 3
T n n 2
S 33 S '
?' 1 S §
3333
4 < 4 4
in ra in n
T n M . ^
35 3 S
at a a at
n » n
U
CO ^ CO
LU(£ 0!
§•§0
S§ ^
LL i «*-
•2 5
til
-------
An efficiency factor (EF) based on the contact basin BODs loadings, air
volumes used, and mixed liquor dissolved oxygen concentrations was
calculated to investigate how the oxygen transfer efficiencies may have
changed during the periods between off-gas tests. Efficiency factors were
not calculated for the reaeration basins because the loadings were
unknown and could not be assumed equal because at times during the
study, each of the reaeration basins was used to supply return activated
sludge to two contact basins. The efficiency factor calculations are
detailed in Appendix B and summarized as follows:
EF = BOD Loading / Qa / DO Correction
Where:
BOD Loading = BODs loading to the contact basin, Ib/day
Qa = air flow rate to the contact basin, cfm
DO Correction = dissolved oxygen correction factor
= (C*-C)/C*
C* = system saturation DO concentration, mg/1
C = contact basin mixed liquor
DO concentration, mg/1
While the efficiency factor is not an exact measure of the oxygen transfer
efficiency, it will provide a reasonable indication of how the oxygen
transfer efficiency changed, if the oxygen required per pound of BODs
remains constant, nitrification does not occur, and the dissolved oxygen
concentration used in the air flow rate correction is representative of the
weighted average concentration in the mixed liquor.
The constant oxygen requirement is not critical because the two test
quadrants are being compared and any changes in the oxygen.
requirement would be the same for both systems. Nitrogenous oxygen
demand was not observed during the study.
The DO correction factor is the greatest source of error. The value used
for C should be the air flow weighted average dissolved oxygen
concentration in the contact basin. The mixed liquor dissolved oxygen
concentrations used in the efficiency factor calculations are computer
logged values based on measurements using one in-basin probe per tank.
While the single location near the outlet could provide dissolved oxygen
concentrations substantially different than the air flow weighted average,
the normal operating procedures used at the Green Bay plant minimized
the chances of error because the dissolved oxygen uniformity was
checked several times a week and adjustments were made to the air flow
distribution if a substantial imbalance was found. Therefore, the average
of the single probe readings would normally be indicative of the entire
tank contents.
10
-------
Diffuser Evaluations
Diffuser elements were tested at various times during the study to:
• Define the initial characteristics of the diffusers used in the study.
• Quantify changes in the diffuser characteristics caused by fouling or
aging.
• Evaluate the effectiveness of preventive maintenance and restora-
tive cleaning procedures.
All field monitoring was done by GBMSD operations personnel.
Laboratory evaluations were performed by the Ewing Engineering
Company.
Both types of diffusers were characterized using dynamic wet pressure
(DWP) and steady state clean water oxygen transfer efficiency tests. The
ceramic diffusers were further evaluated using the bubble release
vacuum (BRV) test. Ceramic disc thickness and strength, and membrane
size, weight, hardness and modulus of elasticity were measured to
quantify any changes caused by aging.
DWP measurements were made in situ at the treatment plant and in a
laboratory at the Ewing Engineering Company. The in situ
measurements were made on the grid diffusers and pilot diffusers.
Diffusers for laboratory testing were obtained from the removable pilot
headers or from the in-basin grids. Diffusers were collected from the
grids when the basins were drained for inspection and cleaning.
The steady state clean water oxygen transfer efficiency tests were run in
the laboratory in a 30-inch diameter by 10-foot water depth tank. Steady
state conditions were established by feeding a constant rate of sodium
sulfite solution to maintain a dissolved oxygen concentration of 1.0 to 3.0
mg/1. The oxygen transfer efficiency was measured using the off-gas
method.
Although BRV tests can only be conducted on diffusers that are out of
service, the BRV test can be performed without removing the diffusers
from the basins. However, it was more convenient and more precise data
were collected by conducting the tests in the laboratory.
Indentation hardness of homogeneous materials ranging from soft
vulcanized rubbers to some rigid plastics is covered in ASTM
Specification D-2240. The specification covers type A (for softer
materials) and type D (for harder materials) durometers and the
procedure for determining the hardness. Indentation hardness was
measured using a type A dial durometer. Five locations on the
11
-------
membrane were measured and the results were averaged and reported
as Shore A Durometer hardness.
Modulus of elasticity is the ratio of the stress (stretching force per unit
area of cross-section) to the strain (elongation per unit length). It was
determined by measuring the diameter of the sheath at operating
pressures of approximately 5 and 15 inches of water gauge (w.g.). A
section of membrane between 2 and 2.5 diameters long was fixed on a
support with the ends sealed. An internal pressure tap was provided and
connected to a manometer to measure the static pressure within 0.5 in.
w.g. The diameter of the membrane section was determined at the
desired operating pressures by measuring the circumference of the
cylindrical section using a pi tape, i.e., a flexible metal tape with 1- inch
markings at intervals equal to 3.1416 inches and each interval divided
into 0.005-inch graduations. The tensile modulus is calculated as follows:
ET = (0.018) (Pis-PS) (Ds)2
(Di5-D5)t
Where:
ET = Tensile modulus of elasticity, psi
P = Pressure, either 5 or 15 in. w.g.
D = Membrane diameter at 5 or 15 in. w.g.,
inches
t = Membrane thickness, inches
Foulant Analysis
Foulant materials deposited on the diffusers were analyzed by the Ewing
Engineering Company as follows:
• About 1 gram of the foulant (dry solids) was removed from a known
area of the diffuser and analyzed for total solids.
• The dried solids were crushed with a glass stirring rod and divided
into two portions. One half of the dried solids were ignited at 600°C
for 15 minutes to determine the volatile solids content.
• The second portion of dried solids was mixed with about 10 ml of 14
percent HC1 in a Gooch crucible, steeped for about 10 minutes,
washed with an additional 10 ml of 14 percent HC1 and rinsed with
distilled water. The sample was then dried and the acid soluble frac-
tion of the original sample was calculated.
• The acid treated and dried solids were ignited at 600°C for 15 minutes
to determine the volatile solids content of the acid insoluble solids.
Several samples of foulant were analyzed by electron dispersive
spectroscopy to determine the elemental constituents. The test measures
12
-------
elements with atomic numbers greater than 10, i.e., carbon, nitrogen and
oxygen are not measured.
RESULTS AND DISCUSSION
The membrane tube diffuser system was installed first and put into
service in January 1986. The ceramic disc diffuser system installation was
completed in April 1986. To provide a clean start for both systems, the
contact and reaeration basins equipped with the membranes were
drained and the diffusers were cleaned by hosing, scrubbing with a stiff
bristle brush, and rehosing. The cleaning procedure was recommended
by the manufacturer. Both systems were then put into service in May
1986.
The results of the various field and laboratory tests are presented and
discussed in the following subsections:
• Operational procedures and goals
• Oxygen transfer efficiency versus time
• Oxygen transfer efficiency versus air flow rate
« Apparent alpha versus time
i
• Pilot diffuser evaluations
• Grid diffuser evaluations
• Foulant analyses
• Cost analysis
Operational Procedures and Goals
In situ acid gas treatment of the ceramic diffusers and flexing of the
membrane diffusers were evaluated as preventive maintenance
techniques for controlling diffuser fouling. For the first six months of
operation, half of the ceramic grids were treated with HC1 gas at
frequencies ranging from monthly to every three months. Half of the
membrane tube grids were flexed every three weeks. The frequencies
were recommended by the equipment manufacturers. The preventive
maintenance schedule is summarized in Table 4.
The HC1 gas treatments were performed according to the manufacturer's
instructions. Typically the acid gas is fed to the diffusers until the DWP
is reduced to a predetermined value. Since there was little or no increase
in DWP during the interval between treatments, it was decided to feed
approximately 0.1 Ib HC1 per diffuser.
13
-------
Table 4 - Diffuser Preventive
Quadrant 2 - Ceramic Disks
HC1 Gas Treatments
Frequency
Monthly
Monthly
Monthly
Every 2 Months
Every 2 Months
Every 2 Months
Every 3 Months
Every 3 Months
Grids
C2.1N
C2.2N
R2.1
C2.3N
C2.4N
R2.3
C2.5/6N
R2.5
Maintenance Summary
Quadrant 4 - Membrane Tubes
Flexing ;
Frequency
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Every 3 Weeks
Grids
C4.1N
C4.2N
C4.3N
C4.4N
C4.5/6N
R4.1
R4.3
R4.5
The flexing procedure was according to the membrane diffuser
manufacturer's instructions. One grid was flexed at a time, by closing the
downcomer air valve, bleeding off the air in the header system so the
membranes would collapse completely onto the frame, increasing the air
flow rate to about 8 scfm per diffuser for two to five minutes, and
returning the air flow rate to the previous operating level.
The operational goals throughout the study were to operate the two test
quadrants at equal organic loadings, equal solids retention-times, and
equal dissolved oxygen concentrations. Table 5 shows the monthly
average data for these operating parameters from May 1986 through
October 1987.
Initially, 50 percent of the flow was fed to each of the two test quadrants
and the loadings were essentially equal. In mid-June 1986, it became
apparent that the two systems would not be able to handle the entire
load. The air usage was very high and it became difficult to maintain the
desired dissolved oxygen concentrations. At first, the problems appeared
to be worse in the ceramic system, so the plant staff decreased the BODs
loading to that quadrant by putting one of the sparged turbine contact
basins in service. The resultant flow and loading split was 25 percent to
the ceramic system, 25 percent to the sparged turbine system, and 50
percent to the membrane system. After about a week, adjustments were
made to equalize the loadings to all three contact basins. The third
contact basin remained in service through April 1987.
Reaeration basin 2 (ceramic system) supplied return sludge to contact
basin 2 (ceramic system) and contact basin 1 (sparged turbines). The
hydraulics associated with this operating mode resulted in slightly higher
BODs loadings to the membrane system; however, the differences
between the two test quadrants averaged about 5 percent.
14
-------
Table 5 -
Month
05/86
06/86
07/86
08/86
09/86
10/86
11/86
12/86
01/87
02/87
03/87
04/87
05/87
06/87
07/87
08/87
09/87
10/87
Average
Operating Data for May
Quadrant 2 -
BODs
Loading
1000 Ib/day
56.5
55.5
57.0
58.8
56.0
58.1
49.7
51.2
52.9
64.7
55.3
54.1
31.0
36.7
35.4
52.6
44.1
43.2
50.7
Ceramics
SRT
days
3.00
2.36
3.23
2.65
2.52
2.66
3.14
3.21
2.90
2.89
2.82
2.73
4.26
2.96
4.72
3.24
3.50
3.91
3.15
DO
mg/1
2.4
1.6
1.8
1.2
1.3
1.5
2.3
2.5
2.5
1.6
1.7
2.1
2.4
2.2
2.0
1.6
2.1
2.4
2.0
1986 - October 1987
Quadrant 4
BOD5
Loading
lOOlb/day
54.7
65.2
59.4
62.8
58.7
62.1
55.3
53.8
54.2
64.8
58.1
59.1
38.2
39.5
38.3
46.0
45.2
44.1
53.3
- Membranes
SRT
days
w«
3.02
2.36
3.03
2.86
2.86
2.87
3.37
3.64
3.42
3.26
3.09
2.92
3.67
2.74
4.91
3.55
3.98
4.11
3.31
DO
mg/1
~
2.3
1.6
1.8
1.4
1.6
1.6
1.8
2.5
2.4
2.1
2.0
2.0
2.2
2.0
1.8
1.8
2.0
2.3
2.0
Throughout the study, the solids retention times and dissolved oxygen
concentrations in the two quadrants were nearly equal.
Oxygen Transfer Efficiency Versus Time
Off-Gas Testing
The standard two-day off-gas test was conducted on the fine pore systems
nine times over the 18-month study period. The detailed data summary
tables are provided in a report by the Ewing Engineering Company.5 The
first six off-gas tests were conducted between the initial start-up in May
1986 and December 1986. The last three off-gas tests were conducted in
June, August and October 1987.
The in-process oxygen transfer efficiencies measured by the off-gas
method were adjusted to standard conditions of 20°C, 1 atmosphere and
zero dissolved oxygen concentration. The adjusted in-process oxygen
transfer efficiencies are designated as alpha'-SOTE, where; the alpha'
15
-------
stands for apparent alpha. Alpha, defined as the ratio of the in-process
oxygen transfer coefficient to the clean water oxygen transfer coefficient,
is a function of the wastewater characteristics for a given system and
geometry. Apparent alpha includes other factors that affect performance
such as physical changes in the diffuser characteristics caused by fouling
and/or aging.
The results of the off-gas testing are shown in Figure 4. Each data point
is a basin average comprising the local data points weighted by the local
flux measurements as follows:
Alpha'-SOTE = SUM [(Fi« alpha'-SOTEi)]/SUM [Fi]
Where:
Alpha'-SOTE = off-gas weighted basin average oxygen
transfer efficiency, decimal
Fi = local off-gas flux, scfm/sq ft
alpha'-SOTEi = local oxygen transfer efficiency, decimal
Each contact basin data point comprises 24 local measurements while
each reaeration basin data point comprises 12 measurements. The
December 1986 data points are an exception. The contact basin point
consists of 12 measurements and the reaeration basin point consists of
10 measurements. The December sampling plan was reduced because
daylight hours were limited.
The lines connecting the data points are provided to show the overall
changes in the alpha'-SOTE between test results. The lines are broken
in November 1986 and July 1987 to indicate that the systems were
drained for inspection and diffuser cleaning.
The significance of the differences in the off-gas weighted basin average
(mean) alpha'-SOTE values were evaluated using statistical tests. The
analysis of variance was used for the contact basin data while the t-test
was used for the reaeration basin data. The tests determine the
probability that observed differences in the means are caused by random
rather than systematic variations. The analyses are based 'on a null
hypothesis that the means are equal. Then, if the test indicates a high
probability that the difference is not caused by random variation, the
hypothesis is rejected and the conclusion is drawn that the observed
difference in the means is significant. On the other hand, if the
probability is low, it is concluded that the observed difference is not
statistically significant A significance level of 0.05 was used in all the
16
-------
25
20
UJ
015
CO
5io
CONTACT BASIN
Cleaned
Diffusers
Cleaned
Diffusers
56789 101112 1 23456789 1011
1986 1987
25
20
LU
&
CO
0
REAERATION BASIN
^A A ; \^
A— ^
5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9 1011
1986 1987
Legend
o Ceramics
A Membranes
17
FIGURE 4
Oxygen Transfer Efficiency
(Offgas Method) Versus Time
-------
tests. The significance level is the maximum probability of being wrong
if the null hypothesis is rejected.
Since the statistical tests only provide an insight into the relationship
between the sample mean and the sample variability, all the available
information should be considered before a final judgment is made. In
the case of analyzing the alpha'-SOTE data, additional information
regarding the oxygen transfer efficiency of the two test aeration systems
is available in the form of efficiency factors. These will be presented and
discussed in the next section. Details of the statistical analyses are
presented in Appendix C. The results are summarized in Table 6.
The contact basin alpha'-SOTE data depicted in Figure 4 show
discernible differences in all but two data sets, the July 30,1986 data and
the June 18, 1987 data. The statistical analysis of the data, however,
indicates that only the last two data sets are significantly different at the
0.05 significance level. The results of the statistical tests conducted on
the reaeration basin data confirm that the apparent differences shown in
Figure 4 are significant at the 0.05 confidence level.
Table 6 - Statistical Analysis of Alpha'-SOTE Data
Basin
Test Date
Contact 5-13-86
5-15-86
7-02-86
7-30-86
10-30-86
12-03-86
6-18-87
8-05-87
10-28-87
Reaeration 5-12-86
'No =
Yes =
5-16-86
7-01-86
7-29-86
10-29-86
12-02-86
6-17-87
8-04-87
10-27-87
Quadrant 2
Alpha'-SOTE
14.8
14.7
17.1
9.7
12.2
19.1
11.8
20.2
163
18.2
17.0
21.2
143
11.6
19.6
11.8
232
193
Quadrant 4
Alpha'-SOTE
163
162
17.0
12.1
143
163
11.4
12.1
113
17.6
17.0
18.4
112
13.4
13.1
11.1
13.7
12.7
Difference
Value
1.7
13
0.1
2.4
2.1
2.8
0.4
8.1
5.0
0.6
0.0
2.8
3.1
1.8
63
0.7
93
6.8
gig/
No
No
No
No
No
No
No
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Differences are not significant at the 0.05 confidence level
Differences are significant at the 0.05 confidence leveL
18
-------
The lack of statistical significance in several of the contact basin
comparisons may be due to the inherent variability in the local
measurements caused by the high organic loadings, the high air flow rates
and the severe taper in the air flow rates between zones. The operating
data, presented in the next section in the form of efficiency factors,
support the conclusion that the discernible differences shown in Figure
4 are real. In the reaeration basins, the variability in the local
measurement of oxygen transfer efficiency was much smaller than in the
contact basins. The lesser variability was possibly due to lower organic
loadings resulting in more uniform air flow rates throughout the basins.
The two systems started out at nearly equal oxygen transfer efficiency.
The oxygen transfer efficiency of both systems decreased substantially
such that after three months of operation, the oxygen transfer efficiency
was about 75 percent of the initial value. Rather than drain the basins
and clean the diffusers so early in the study, it was decided to continue
to operate with three contact basins in service and gain more operating
experience. In October 1986, the fifth off-gas test was conducted. The
off-gas test results indicated that the oxygen transfer efficiencies were
still substantially below the values measured in May 1986. There was also
a problem with maintaining acceptable DO concentrations at the inlet
end of the contact basins, so it was decided to drain the basins for
inspection and cleaning of the diffusers.
The cleaning methods employed were based on the results of evaluations
conducted in the laboratory on diffusers obtained from the removable
pilot headers. The laboratory test results and proposed cleaning
procedures were discussed with and agreed to by the manufacturers
before the cleaning activities were started. The first pilot diffusers tested
were ceramic discs from the reaeration basin. The foulant found on the
diffusers was in two layers. The top layer was slimy, about a 1/4-inch thick
and could be hosed off with water at a pressure of about 30 psig. The
bottom layer was black, hard and firmly attached to the diffuser surface.
The black layer was soluble in 14 percent HC1 but the reaction was rather
slow. The cleaning procedure selected for use on the ceramic diffusers
was hosing from the tank top with fire hoses, partially filling the basin
with service water, gas cleaning with 0.1 Ib HC1 per diffuser, draining off
the service water, and rehosing from the tank top.
The foulant found on the contact basin diffusers was a single layer having
characteristics similar to the top layer found on the reaeration basin
ceramic diffusers. It was readily removed by hosing. Tests on several of
the contact basin pilot diffusers indicated that the acid treatment step in
the cleaning procedure was probably not necessary because only a slight
improvement hi the clean water oxygen transfer efficiency was observed.
19
-------
However, it was decided to be conservative and perform the rigorous
hose, acid gas treat and rehose procedure anyway.
The membranes taken from the reaeration basin also had two foulant
layers. Membranes taken from the contact basin had a single layer. All
of the membranes were cleaned by hosing from the basin floor, scrubbing
with a stiff bristled brush, and rehosing. The procedure removed the
slime layer and the black material found on the reaeration basin units.
Tank top hosing was not attempted with the membrane tubes for two
reasons. First, there was concern that the force of the water falling from
a height of about 25 feet would break the connecting pipe nipples.
Second, the foulant was on both the top and the bottom of the diffusers
so tank top hosing would not have removed all of the foulant
Once the cleaning was complete, the systems were put back into service
and a followup off-gas test was conducted. The alpha'-SOTE of the
ceramic systems was essentially restored to its original condition in both
the contact and the reaeration basins, while the increases in alpha'-SOTE
in the membrane equipped basins were small.
The next off-gas test was conducted in June 1987. The measured oxygen
transfer efficiencies of the systems were similar to the oxygen transfer
efficiencies measured in October 1986 before the systems had been
cleaned. Based on the off-gas test results, the decision was made to drain,
inspect, and clean the two test quadrants. The same cleaning methods
used in November 1986 were used in July 1987. The ceramic diffusers
were tank top hosed, acid gas treated, and rehosed. The membrane
diffusers were in-basin hosed, brushed, and rehosed. After cleaning, the
systems were put back into service and retested by the off-gas method.
The oxygen transfer efficiency of the ceramic system showed an increase
to above its original level. The membranes showed little if any
improvement.
The last off-gas test was conducted during the last week in October 1987.
The oxygen transfer efficiency of the contact basin ceramic diffuser
system decreased from nearly 25 percent to about 17 percent. A similar
reduction was experienced in the reaeration basin. The membrane
systems remained relatively constant at about 12 percent in the contact
basin and 13 percent in the reaeration basin.
Efficiency Factor
The efficiency factors for the two fine pore contact basins, calculated
from monthly average data, are plotted versus time in Figure 5. The
efficiency factor and the off-gas results show similar trends. The
efficiency factors indicate that both systems lost oxygen transfer
20
-------
r
d 6
o
m
5 4
I
o>
'o
e
LU
D
Cleaned
Diffusers
•'
Cleaned
Diffusers
5 6 7
8 9 10 11 12 1 2 345 6 78 910
1986 1987
Legend
o Ceramics-
A Membranes
Based on the BOD5 loading and air flow rate to the contact
basins and correcting the air flow rate to zero DO.
FIGURE 5
Contact Basin Efficiency
Factor Versus Time
21
-------
efficiency after start up. The loss in efficiency was greater and more rapid
in the ceramic system, so over the first six months of operation, the
average oxygen transfer efficiency for the membrane system was higher.
After the November 1986 cleaning, the efficiency factor for the ceramic
system showed a substantial increase similar to the increase indicated by
the off-gas test results. The efficiency factor then decreased at a relatively
constant rate until the system was cleaned again in July 1987. During the
same time period, the efficiency factor for the membrane diffuser system
showed several monthly increases and then gradually decreased until the
system was cleaned in July. The rather large increase in the efficiency
factor seen in February 1987 is probably an outlier. The BODs loading
that month increased to both systems by 20 percent, but no additional
air was fed to the membrane quadrant. There may have been some
changes in the weighted average DO in the basin that were not accounted
for in the computer reported basin DO value.
The efficiency factor data from before and after the July cleaning show
that cleaning of the ceramic diffusers resulted in a substantial increase
in oxygen transfer efficiency. The membrane diffuser system showed
little or no increase in oxygen transfer efficiency after cleaning. The
ceramic diffuser system maintained a higher oxygen transfer efficiency
through to the end of the study period in October 1987. These same
effects were indicated by the off-gas test results.
Oxygen Transfer Efficiency Versus Air Flow Rate
On three occasions, the air flow rates to the full-scale systems were varied
and off-gas tests were performed to investigate the relationship between
oxygen transfer efficiency and air flow rate per diffuser. The diffuser air
flow rates were adjusted by changing the total air flow rate to the basin
so any changes in the mixing pattern among grids would be minimized.
The number of hood positions sampled was reduced so several air flow
rates could be tested in one day.
The first test was performed on May 14,1986. One sampling position was
tested in each of the contact basins. The data are summarized in Table
7. The data show that the air flow rate per diffuser did not result in a
substantial change in alpha'-SOTE. However, because only one
sampling location was tested, the data are not conclusive. Changes in the
mixing among grids could have had a substantial effect on the results.
The second and third tests were more extensive. Tests were conducted
before and after the July 1987 cleaning of the diffusers. Before the
cleaning effort, both the contact and reaeration basins were evaluated.
Five hood positions were sampled in each contact basin and four
22
-------
Table 7 - Variable Air
Flow Rate Testing - Mav 14. 1986
Contact Basin 2 - Ceramics1 Contact Basin 4 - M^mhran^c2
Air Flow
Rate, scfm
Total
13,600
13,600
17,000
7,000
13,600
7,000
1 Grid C2
2 Grid C4
/Unit
2.2
2.2
2.8
1.1
2.2
1.1
3NA
3NA
Time
hr
1210
1343
1353
1408
1418
1433
1444
1459
2151
Alpha' Air Flow Alpha'
-SOTE Rate, scfm Time -SOTE
% Total /Unit hr %
15.1 13,000 2.8 1517 15.2
15.2 17,000 3.7 1531 14.8
14.5
14.9 7,250 1.6 1544 12.3
16.9 13,000 2.8 1603 15.0
14.0
15.4 14,000 3.0 2135 14.3
14.9
15.2
15.1 ;
16.8 '
16.7 ' , ~
positions were used in the reaeration basins. After the cleaning effort,
the testing in the contact basins was repeated.
The alpha'-SOTE data are plotted versus sampling position in Figures 6
through 11. The results show that sampling position has a definite effect
on alpha'-SOTE. Review of the data collected from each Campling
position indicates that alpha'-SOTE is not affected substantially by
changes in the air flow rate per diffuser.
Apparent Alpha Versus Time
The basin average values of alpha' are summarized in Table 8. Initial
values of alpha' for the contact and reaeration basins were in the range
of 0.46 to 0.53, substantially less than the design values of 0.68 and 0.90
for the contact and reaeration basins, respectively. When fouled, the
range was reduced to 0.34 to 0.44. The lower alpha' values ;were the
reason why the two test quadrants could not treat the entire plant load
and the third contact basin had to be operated. After the cleaning
activities in November 1986 and July 1987, the ceramic system alpha'
23
-------
25
20
15
CO
Q.
<
10
OO
A
a
o
C2.15
Legend
Basin Air Flow Rate
6-23-87
6,200 scfm
12,200 scfm
17,600 scfm
C2.2N
C2.35
Grid
C2.4N
C2.55
Mean
Weighted
Average
6-24-87
A 6,800 scfm
9 17,000 scfm
Diffuser Air Flow Rate
0.4 - 0.7 scfm/unit
1.3 - 1.7 scfm/unit
2.6 - 3.4 scfm/unit
0.9 - 1.2 scfm/unit
2.3 - 3.3 scfm/unit
24
FIGURE 6;
Oxygen Transfer Efficiency;
Versus Position and Air
Fbw Rate - Contact Basin
Ceramic Diffusers
Before Cleaning
-------
LLJ
30
25
20
co
jQ.
< ,
0
C2.IN
Legend
Basin Air Flow Rate
8-6-87
a 10,500 scfm
o 17,100 scfm
C2.25
C2.3N
Grid
C2.45
C2.5N
Mean :
Weighted
Average
Diffuser Air Flow Rate
0.9-2.6 scfm/unit
1.5 - 3.1 scfm/unit
25
FIGURE 7
Oxygen Transfer Efficiency;
Versus Position and Air
Flow Rate - Contact Basin •
Ceramic Diffusers
After Cleaning
-------
18
16
14
llf
CO 12
•
"co
.9-
< 10
O
C4.IN
C4.25
C4.3N
Grid
C4.45
C4.5N
Legend
Basin Air Flow Rate
6-23-87
A 7,400 scfm
o 25,500 scfm
6-24-87
• 10,000 scfm
• 21,400 scfm
D iff user Air Flow Rate
1.0 - 2.1 scfm/unit
4.4 - 7.2 scfm/unit
1.9 - 2.6 scfm/unit
3.8 - 5.5 scfm/unit
Mean
Weighted
Average
26
FIGURE 8
Oxygen Transfer Efficiency
Versus Position and Air
Flow Rate - Contact Basin
Membrane Diffusers
Before Cleaning
-------
18
16
14
I
CO 12
03
Q.
10
8
C4.IN
C4.25
C4.3N
Grid
C4.45
Legend
Basin Air Flow Rate
8-6-87
A 9,900 scfm
o 18,200 scfm
Diffuser Air Flow Rate
2.0 - 2.4 scfm/unit
3.7 - 4.3 scfm/unit
Cfc
C4.5N
Mean
Weighted
Average
27
FIGURE 9
Oxygen Transfer Efficiency
Versus Position and Air
Flow Rate - Contact Basin
Membrane Diffusers
After Cleaning
-------
18
16
o 14
uF
Q
CO 12
1
"co
Q.
<10
8
6
• n & 0.5
°1.3
O2.5
• " :
•
D2.2
02.9 24
DD^i.o
2.0
1
D 2.0
O1.4
131.85
A1X> CP2.7
D2.0 2.5 A 1.7
•
• • • • f| § • R ••
R2.1 R2.3 R2.4 R2.6 Mean
Grid Weighted
Average
Legend
Basin Air Flow Rate
A * 2,250 scfm
a * 4,200 scfm
5,100 scfm
* Air Flow Rate/Diffuser
28
FIGURE 10
Oxygen Transfer Efficiency
Versus Position and Air
Flow Rate - Reaeration Basin
Ceramic Diffusers
Before Cleaning
-------
18
16
14
o
CO
I
"ctt
cL
12
10
8
D3.0
1.6
Q4.7
£1.7
Q2.7
1.6
n 3.0
£1.9
3.9 O 4.5
4.7
4.6
Q4.8
A1.7
4.5 Da.o
Q3.4
R4.1
R4.3 R4.4
Grid
R4.6
Legend
Basin Air Flow Rate
A * 2,350 scfm
n * 4,240 scfm
o * 6,300 scfm
* Air Flow Rate/Diffuser
Mean
Weighted
. Average
29
FIGURE 11
Oxygen Transfer Efficiency
Versus Position and Air
Flow Rate - Reaeration Basin
Membrane Diffusers
Before Cleaning
-------
Table 8 - Apparent Alpha as a Function of Time in Service
Time in Service, months
Basin
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Contact
Reaeration
Total Since Cleaning Apparent Alpha
Ceram Memb Ceram Memh Ceram Memh
<1
2
2
3
3
6
6
7
7
13
13
15
15
18
18
4
4
6
6
7
7
10
10
11
11
17
17
19
19
22
22
<1
2
2
3
3
6
6
-------
Table 9 - Diffuser Characterization
New Ceramic Discs
Diffuser No.
K-39-40-1
K-39-40-2
K-39-40-3
K-39-40-4
K-39-40-5
K-39-40-6
K-39-40-7
K-39-40-8
K-39-40-9
K-39-40-10
K-39-40-11
K-39-40-12
K-39-40-13
K-39-40-14
K-39-40-15
Average
0.5cfm
4.90
4.90
5.05
4.70
4.80
4.70
5.00
4.90
5.00
5.00
5.00
4.95
4.80
4.85
4.95
4.90
Coefficient of variation,
DWP,
0.75 cfm
5.20
5.20
5.35
5.00
5.10
5.05
5.35
5.25
5.25
5.35
5.30
5.25
5.10
5.10
5.20
5.20
in. w.g.
2.0 cfm
5.90
5.75
6.00
5.60
5.70
5.50
6.00
5.70
5.95
5.85
5.80
5.90
5.60
5.60
5.95
5.79
Results -
3.1 cfm
7,05
6.75
7.30
6.70
6.85
6.35
7.40
6.60
7.20
7.00
6.80
7.15
6.55
6.45
7.15
6.89
BRV^
Mean
5.98
6.06
6.09
5.59
5.67
5.67
6.03
6.03
6.09
6.01
6.24
6.03
5.80
5.71
5.78
5.91
in. w.g.
COV1
0.032
0.032
0.021
0.032
0.026
0.025
0.020
0.035
0.029
0.019
0.031
0.022
0.030
0.040
0.032
0.033
standard deviation/mean. :
Table 10 - Diffuser Characterization Results -
New Membrane Tubes
Diffuser No.
K-38-43-1
K-38-43-2
K-38-43-3
K-38-43-4
K-38-43-5
K-38-43-6
K-38-43-7
K-38-43-8
K-38-43-9
K-38-43-10
K-38-43-11
K-38-43-12
K-38-43-13
Average
l.Ocfm
6.15
7.50
7.10
7.80
7.20
7.30
7.20
7.00
7.60
7.60
7.50
7.00
7.30
7.25
DWP,
3.0 cfm
8.90 .
9.80
9.80
9.70
9.90
9.90
9.65
9.30
10.0
9.60
9.90
9.80
9.00
9.70
Shore A Durometer, average of five
in. w.g.
6.0 cfm
10.7
11.2
12.5
11.8
12.3
12.6
12.0
11.7
12.5
11.8
12.5
12.5
12.4
12.0
points.
!
Hardness1 Weight
9.0 cfm
13.5
13.8
14.2
13.2
13.7
14.5
13.5
13.2
14.1
13.2
14.2
14.3
14.0
13.8
Mean
64.8
_
64.0
60.5
64.0
_
60.4
63.7
65.2
61.8
_
63.1
64.4
63.2
Grams
_
118.7
_
_
_
118.7
—
—
_
.
^
_
118.9
118.8
i
31
-------
Table 11 - Ceramic Diffuser Test Results -
Pilot Units Removed June 11, 1986
DWP,in.w.g.
Diffuser
No.
K-39-40-3
K-39-40-3
K-39-40-6
K-39-40-6
K-39-40-7
K-39-40-7
K-39-40-8
K-39-40-8
K-39-40-11
K-39-40-11
Condi
tion
New
Used2
New
Used2
New
Used3
New
Used3
New
Used4
0.5
cfm
5.05
8.30
4.70
8.80
5.00
6.50
4.90
7.00
5.00
6.25
0.75
cfm
535
9.40
5.05
10.2
5.35
7.55
5.25
7.65
530
7.10
2.0
cfm
6.00
14.8
5.50
17.6
6.00
11.7
5.70
12.5
5.80
10.7
K-39-40-12 New 4.95 5.25 5.90
K-39-40-12 Used4 5.85 6.65 9.60
1 Coefficient of variation, standard deviation/mean.
2 From pilot header R2.2P.
3 From pilot header C2.6SP.
4 From pilot header C2.2SP.
3.1
cfm
7.30
25.6
635
33.8
7.40
22.1
6.60
23.5
6.80
20.9
7.15
16.8
BRV.iiLw.fi.
Mean
6.09
27.2
5.67
25.9
6.03
19.0
6.03
19.0
6.24
18.5
6.03
16.0
cov1
0.021
0.179
0.025
0.191
0.020
0.211
0.035
0.164
0.031
0.303
0.022
0334
flow rate to the pilot header was reduced to account for the decreased
number of diffusers.
Summer 1986
On June 11, 1986, the Sanitaire pilot headers were removed to install
new air control orifices. While the headers were out, several; discs from
the contact and reaeration basins were tested on-site. The results of the
tests are summarized in Table 11. A substantial quantity of foulant
material was found on the surface of the discs. The material appeared to
be a slime growth containing a gritty sand-like material. The BRV and
DWP data indicate that the diffusers were moderately fouled.
During this site visit, the plant staff reported that the flow rate of air to
maintain acceptable DO concentrations was 10 to 15 percent higher for
the contact basin 2 ceramic discs compared to the contact basin 4
membrane tubes. To evaluate the situation, several more pilot diffusers
were removed and tested in the laboratory. Two diffusers were tested as
received. One was then HC1 gas cleaned, retested, hosed off with a 40
psig water spray at about 20 feet, and tested again. The results of the tests
are summarized in Table 12.
32
-------
The units, which had been in service in the contact basin for about 105
days, showed a decrease of about 7 percent in clean water oxygen transfer
efficiency. A single cycle of acid gas cleaning produced no immediate
improvement in oxygen transfer efficiency but a decrease in DWP did
occur. Light hosing restored the clean water oxygen transfer efficiency
to approximately its new condition. Further acid treatment and hosing
produced no further improvement in oxygen transfer efficiency. Light
hosing of a fouled disc which had no prior acid treatment restored it to
near its estimated original condition.
For comparison, a membrane tube diffuser was also removed and tested.
The DWP of the diffuser was only slightly higher than the measured
average for the new diffusers. The clean water oxygen transfer efficiency
of the used diffuser which was in service for about 87 days was essentially
equal to a new diffuser.
The observed need for more air in the contact basin with the ceramic
diffusers was further investigated by removing another pilot diffuser in
August 1986. The diffuser was clean water tested at air flow rates of 1.0,
2.0 and 2.8 cfm to determine if the relationship between oxygen transfer
efficiency and air flow rate was different for a fouled diffuser than for a
clean diffuser. A change in the relationship was seen at Lansing,
Michigan. The test results showed that the relationship was essentially
the same for both a clean and the fouled diffuser. In both cases, the
Table 12 - Ceramic Diffuser Test Results -
Pilot Units Removed June 22,1986
DWP, in WG
Condition
Newz
^
As Found
/\
HC1 Gassed2
Hosed2'3
H-A-H2'4
0.5
cfm
4.80
7.85
_
4.80
0.75
cfm
5.10
11.8
_
6.20
5.20
2.0
cfm
5.60
-
—
7.85
6.10
3.1
cfm
6.55
.
_
11.3
8.00
BRV, in w.g.
Mean
5.80
_
8.61
5.68
COV
0.030
0.097
0.015
CWOTE
Ratio to
Control1
_
0.89
084
U.OH
0.98
i.m
New0
Hosed3'5
4.85
6.00
5.10
6.50
5.60
8.00
6.45
11.3
5.71
7.06
0.040
0.075
1.04
Clean water oxygen transfer efficiency at 1 cfm measured by the small-tank steady state method and
compared to diffuser "C" used as the control.
3 Diffuser number K-39-40-13 mounted on pilot header number C2.1NP.
Hosing done with 40 psig water from a distance of 20 feet.
Hose-acid-hose, 14% HC1 sprayed on the diffuser and left for 20-30 seconds before rehosing.
Diffuser number K-39-40-14 mounted on pilot number header C2.1NP.
33
-------
oxygen transfer efficiency was proportional to the air flow fate to the
-0.21 power.
October 1986
Three ceramic disc diffusers and three membrane tube diffusers were
removed from the pilot headers in October 1986 for routine testing. The
discs were removed on October 21 and the membranes were removed
on October 28.
The ceramic disc diffuser taken from the inlet end of the contact basin
(C2.2SP) was tested ias found, after an acid gas treatment in the
laboratory and after hosing with a 30 psig water stream for 30 seconds.
The results of the tests are summarized in Table 13. The diffuser in the
as-found condition was fouled with a slime layer. The DWP and BRV
data indicate that the diffuser was moderately fouled. The clean water
transfer efficiency was reduced by about 25 percent A single acid gas
treatment did not improve the clean water transfer efficiency. The light
hosing, however, resulted in a substantial increase in clean water transfer
efficiency. Similar tests were conducted on a diffuser from pilot header
C2.6NP located near the outlet end of the contact basin. The test results
again showed low pressure hosing was effective in restoring the clean
water transfer efficiency.
The ceramic diffuser taken from the inlet end of the reaeration basin
(R2.2P) was more fouled than the contact basin diffusers. DWP and
BRV measurements were slightly higher while the clean water oxygen
transfer efficiency was slightly lower. Hosing effectively removed the
loosely attached slime layer, but a black splotchy material remained on
the surface of the diffuser after hosing. Despite the black material, the
clean water oxygen transfer efficiency improved by about 28 percent.
Because the black material was not readily dissolved by treating with 14
percent HC1, an attempt was made to remove the material by wire
brushing. The clean water oxygen transfer efficiency of the wire brushed
diffuser was higher than the diffuser just after hosing but the DWP and
BRV remained moderately higher than a new diffuser.
The membrane diffusers were collected from the inlet (C4.2NP) and
outlet (C4.6SP) ends of the contact basin and from the inlet end of the
reaeration basin (R4.2P). The clean water oxygen transfer efficiency
tests were conducted in parallel with a new membrane diffuser
designated GBN. The results of the tests are summarized in Table 14.
The diffusers were covered with a slime material similar to that found
on the ceramic discs. The slime material was easily removed by low
pressure (30 psig) hosing, brushing, and rehosing. The diffuser from the
34
-------
Table 13 - Ceramic Diffuser Test Results -
Pilot Units Removed October 21, 1986
DWP
Diffuser @ 2 cfm
No. Condition in. w.g.
K-39-40-11^ New 5.80
As Found 21.8
After LGC
After H/30 9.0
K-42-3-13 As Found 21.2
After Stor
After H/60 10.5
K-39-40-34 New 6.00
As Found 29.1
After Hose
After WB/H 8.80
Typical New (Table 9) 5.79
A new diffuser designated C was used as the control
2 From pilot header C2.2SP.
3 From pilot header C2.6NP.
4 From pilot header R23P.
Key: LGC = Laboratory gas cleaned
BRV,
Mean
6.24
28.9
-
12.9
24.7
-
143
6.09
32.0
-
10.2
5.91
in. w.g.
COV
0.031
0.200
-
0.230
0.140
-
0.130
0.021
0360
-
0.15
0.033
CWOTE
Ratio to
Control1
@cfm
_
0.77 @ 1.0 ,
0.79 @ 1.0
0.70 @ 2.0
0.91 @ 2.0
0.81 @ 1.0
0.89 @ 1.0
1.07 @ 2.0
«•
0.86 @ 1.0
0.68 @ 2.0
0.88 @ 2.0
0.99 @ 2.0
-
H/30 = Low pressure (30 psig) hosed for 30 seconds
H/60 = Low pressure hosed for 60 seconds
WB/H = Wire brushed then low pressure hosed
reaeration tank had a black crusty layer under the slime layer which also
was removed by the hose-brush-hose procedure. The DWP of the dirty
diffusers was slightly higher than GBN and several inches higher than
the average of 13 new diffusers tested at the beginning of the study. After
cleaning, the DWP of the pilot diffusers was substantially less than the
DWP of new diffusers. The clean water oxygen transfer efficiency of the
fouled diffusers was significantly less than GBN. Clean waiter oxygen
transfer efficiency tests were not conducted at this time on the
hosed-brushed-hosed diffusers so no comparison between dirty and
cleaned diffusers could be made.
November 1987
At the end of the study, several ceramic and membrane diffusers were
taken from pilot headers and tested to determine the characteristics of
35
-------
Table 14 - Membrane Diffuser Test Results - Pilot Units
Removed October 28, 1986
Diffuser From
No. Header Condition
K-42-16-1 C4.2NP As Found
After H-B-H
K-42-16-2 C4.6SP As Found
K-42-16-3 R4.2P As Found
After H-B-H
Typical New (From Table 10)
GBN
DWP@
2cfm
in. w.g.
11.2
6.60
11.9
14.9
4.00
8.60
11.2
A new diffuser designated GBN was used as the control
Key. H-B-H = Low pressure (30 psig) hosed, brushed with
and rehosed.
CWOTE
Ratio to
Controll
<5>cfm
0.74 @ 1.0
0.85 @ 1.0
0.85 @ 2.8
0.71 @ 1.0
0.74 @ 2.8
-
1.0
a scrub brush
diffusers that had been in service for the entire 18-month period. The
diffuser test results provided some unique information because none of
the units had ever been cleaned.
The results of the tests conducted on three ceramic diffusers taken from
contact basin pilot header C2.6SP and three diffusers taken from
reaeration basin R2.6P are presented in Table 15. All the diffusers were
very fouled, based on DWP, BRV, and clean water transfer efficiency
measurements. Treating with HC1 gas substantially reduced the DWP,
but not to like-new condition. Several of the diffusers were cleaned using
a more rigorous three-step procedure. The units were hosed, sprayed
with a 14 percent solution of HC1, allowed to stand for 15 to 20 minutes,
and hosed again (H-A-H). This treatment further reduced DWP. The
BRVs of the H-A-H cleaned diffusers were substantially reduced but
remained moderately higher than typical new units, and the coefficient
of variation (COV) of BRV increased indicating a loss in uniformity.
The fouled ceramic diffusers had clean water oxygen transfer efficiencies
that were 17 to 34 percent lower than a typical new diffuser. After
36
-------
w
fc.
W3
a
Sr
5
'i
2
o
T3
U
v:
*s
e
o
*«3
racteriza
j5
U
••w
«
c
s
12
aj
s
H
Strength'
%
O
g
3'
U
1
S
^ *
fc
r
H
0.
J«-
. C
c
..^ t
^ <
ca 5
E
- t3
c
fj
ec
S'
|"S
Q
^
t
CJ
C ,M
^1
u
o
Jo
Q *~
1
08
o
"O
2
5
B
s
p
"O
>o
^
(Ni
vn
»
§
&
o
'
,
1
1
c
00
c
As Found
Cu •
S
d
/<
?
^
'
1
*
S
o
*
1
f.
^
o
si
As Found
eu
3
"?
f— <
s
*
§
2
0 • • i , , ,
S 2 2
• • • o 5 • S
52 ^ v"i
QO r* ^i
• o o • o
IS f2 *"
' ' ' O O ' O
8 •-> -
0 3 0
I I
•o £ 1= "He4?
g S I 1 < =
•s £ 6 fe £ J3 S
4? 52 *~2 *"• «rt ci <±2
z < < < < < <
0.
e
t—
o
1
1
,
§
U-1
5 H 1 0
_« i 52,5 B
w
•o
'"o
.c
u
O
v;
Ǥ
'•5
"2
n
•o
.C
S3
lie
O
.C
5
—
1
"Sc
2
_3
"x
—
C
•3
u
i
CL
c.
w;
e
u
t>
u;
^
c. °
^ O
O 0
a E
w a
= |
X Jr
ii S o
o I u
«6-5
go
*c3 T3
© ^ «
li] H iS
jr c a
O R —
**> ^
37
-------
cleaning by the H-A-H method, the clean water oxygen transfer
efficiency was restored to about 95 percent of a typical new diffuser.
Diffusers K-49-40-8 and K-49-31-2 were fractured and photographs of
the cross-section were taken at a magnification of 15X (Figure 12).
Entrapped residues were limited to near the surface. The penetration
was approximately 500 microns (0.02 in.).
The exponent of the clean water oxygen transfer efficiency versus
diffuser air flow rate relationship was negative for both the new and used
diffusers, i.e., the oxygen transfer efficiency decreased when the diffuser
air flow rate was increased. The absolute value of the exponent for the
used diffusers, either fouled or cleaned, was greater than for the typical
new ceramic diffuser. For a new diffuser, increasing the air flow rate/unit
from 1.0 cfm to 3.0 cfm would result in a 15 percent decrease in the clean
water oxygen transfer efficiency. Using an average value of-0.21 for the
exponent of the used diffusers, increasing the diffuser air flow rate from
1.0 cfm to 3.0 cfm results in a 21 percent decrease in the clean water
oxygen transfer efficiency. This shows the importance of operating the
diffusers at the lower end of their normal design range.
The results of the tests conducted on three membrane diffusers taken
from contact basin pilot header C4.6NP and three diffusers taken from
reaeration basin R4.6P are presented in Table 16. All of the membranes
were moderately fouled based on the DWP data. At 1.0 cfm, the DWPs
were not noticeably higher than a typical new membrane. At 3.0 cfm, the
DWPs were substantially higher than a typical new diffuser. Cleaning by
hosing, brushing with a stiff bristle scrub brush, and rehosing (H-B-H),
or by inside-out flushing of the slits with a 3 to 5 gpm of clean water while
brushing the entire outer surface with a stiff nylon brush (FB) resulted
in DWPs that were substantially lower than the typical new membrane.
Air flow uniformity was determined by measuring the discharge rate
from five sections along the length of the membrane while it was
operated at 2 scfin. The results are presented as the coefficient of
variation (COV) equal to the standard deviation divided by the mean.
The typical value of COV is 0.24. Even the most fouled diffusers had
relatively good uniformity, the highest value being 0.52. Cleaning the
units improved air flow uniformity with values of COV ranging from 032
to 037.
The clean water oxygen transfer efficiencies of the fouled membranes
were about 35 percent less than typical new membranes at 1.0 cfm, and
about 25 percent less at 3.0 cfm. The greater loss in clean water oxygen
transfer efficiency at the lower diffuser air rate is also indicated by the
change in exponents for the oxygen transfer efficiency versus diffuser air
38
-------
Contact Basin
Macrograph
C40687 (15x)
Cross-Sections Show Entrapped Residues to
a Depth of About 500 Microns (0.02 inches)
Reaeration Basin
Macrograph
C40688 (15x)
39
FIGURE 12
Photographs of Fractured
Ceramic Diffusers
-------
(A
1
CM
5
e
£
JS
£
CJ
j^*
_
u
DV3
O
s
Character!:
C3
_C
1-*
V
JS
«
jj J
S 1?
*-s
•g
is-
E
.2
Q '
.e
00
c
'5 P JL
->,= u
,
u x
u S
f-
P
" f
3
§11
IE
'c
t**m
$ g
•-i't
cOc
b
j
c
Condition
|j
ciz
^ • , °- <=l-
^o ' »o ' \S ' ' *
2 ' ' • e • • o' • 11
S 2-
P o S
O • • i o ' ' c • • •
a ?. ^
» s • • a
SiSfiKp-pgKS^-s;
O O C5 C C O O O O O O
=C r^ o m t—
Q p -: oo'
0 0 0 • • . o o • • •
T T + +
8 P: S ' ' 'PS1 ' '
— O C= C C
? 3C ^1 * ^ ' r^i r*^ t i
* S r~. ? S P p
-00 0 00 0
S*^1^ -;f*!T»-c-i — f. lose
e<^^O^Tr^^i/S^Tf'(/S
-" M — < fS (S) fsj
.a
''I ^ *» f^ 86 0 V . f; 10 r> « =
f^j ^ 2 ^ o r^ -v wS (-*•' -^ .5 .y .
"lii g .
S5 11
^^Jf^OO*^ »^M ^rfOvQO •-• *« ** S
-oo^o.^ooo^o;^ -fl |1|
I =? i| -all
li^e'^i'^e tf * ^ ""^J
ilslsllslls llsSSll5?
22%$%<3%Z2% -g-Bl°f |1|
aEc" BO'a|3
i i i fe % % is|2?g:l5
3 3 3 S 5 5 s.|2@|o||
s-a IBEV!||
S S 2 ' "? ^ " fa^nSiS s 5$
S 2 22 22 Ic°HHg:5<
&4 4 44 44 ji-Sllhi
^! ii iiii ^si 3-U.oiUL)UJOS
™" •*• •*• * — ~n v '•Vi w
-------
flow rate relationship. The exponent determined for the new diffusers
was negative which is typical for fine pore diffusers while the used
membranes (both as found and cleaned) showed positive exponents. The
exponent data indicate that there may be some advantage to operating
at slightly higher air flow rates, or, conversely, that operating at higher
diffuser air flow rates over the range tested would not result in losses in
oxygen transfer efficiency as would be expected from the new diffuser
test data.
Dimensional and weight changes over the 18-month period were small.
The tensile modulus, which is a measure of a material's stiffness
increased substantially, indicating that the material physical properties
had changed. Photographs of these units were not taken because several
grid units had already been photographed. The membrane photographs
are presented in the next section.
Grid Diffuser Evaluations
Diffusers were removed for testing when the basins were drained in
November 1986 and July 1987.
November 1986
The results of laboratory tests conducted on the ceramic diffusers
removed before and after the November 1986 cleaning are presented in
Figure 13. As previously described, the grid units were cleaned by hosing,
treating with HC1 gas and rehosing, labeled H-AG-H in Figure 13. The
units designated as being acid gassed were given the treatment in situ on
the frequency listed in Table 4.
The fouled diffusers from the contact basin showed moderate increases
in BRV and DWP, and substantial reductions in clean water oxygen
transfer efficiency. The diffusers taken from the inlet end had higher
BRVs, DWPs and lower clean water oxygen transfer efficiencies than the
diffusers taken from the outlet end. Because the foulant consisted of a
loosely attached slime-like layer, it was effectively removed by a water
spray directed from the basin walkways using fire hoses. Acid gas
treatment and rehosing provided little if any further improvement in
BRV, DWP or clean water oxygen transfer efficiency. Regular acid gas
treatment may have provided some control of DWP increase but did not
prevent losses in oxygen transfer efficiency.
Visual observations made after draining the reaeration basin indicated
that the diffuser fouling was two-layered, a slime-like layer similar to that
found in the contact basin over a hard black layer. The fouling was
relatively uniform from the inlet to the outlet end. Tank top hosing
41
-------
25
20
6)
Si 15
c
i10
CD
0
• UUIMI/UJI BASIN
A^
•
A
A
A
A .
A A *&
I
F C1 C2 F C2
REAERATION BASIN
^5 A
w - -
0 •
-"•— ~— -^— ^— — — — — — — — •"
New
IB • •
F C1 FC2 F
20
CO
0)10
$
CL
1 °
1.0
o
LU
O
co 0.5
UJ
§
0
A
i '
F 01 C2 F™C2
i
A
F C1 C2 F C2
C2.1 C2.5/6
' •
i
: ° o «!- •
1 NEW
F C*! F~~C2~ "p^
0 o 0 • New
•---•- •
F C1 F C2 ~F~"
R2-1 R2.2 R2.6
Grid & Condition
F = Fouled
C1 := H (Field)
C2 := H-AG-H (Reid)
9 := Not Acid Gassed
o:: Acid Gassed
42
FIGURE 13
Ceramic Diffuser Test Results -
Grid Units Removed November 1986
-------
removed the slime- like layer but not the black splotchy layer. The black
material contained substantial amounts of iron and it was slowly soluble
in 14 percent HC1. Treating with HC1 gas and rehosing from the tank top
improved the diffuser characteristics but most of the black layer
remained. The diffusers that received the monthly in situ acid gas
treatments appeared to have the least amount of the black foulant on the
surface of the discs.
The results of the tests conducted on the membrane diffusers removed
before and after the November 1986 cleaning are presented in Figure
14. The cleaning procedure was the previously described hose, brush and
rehose (H-B-H) method. The flexing of selected grids was done every
three weeks. ;
The diffusers in the contact basin were more fouled at the outlet than at
the inlet end. Many of the diffusers in the first aeration zone looked clean
while the diffusers in the second and third zones were covered with about
a quarter-inch of a slime-like material. Large quantities of grit had
accumulated at the inlet end especially where the diffusers were the
cleanest. The grit may have provided some cleaning by abrasion. The
high liquid velocity in this area of the basin (between zones 1 and 2) was
the result of the significant differences in air rates used in the two zones.
The air flow rates to zone 1 were typically more than 50 percent higher
than in zone 2.
Cleaning improved the clean water oxygen transfer efficiency of the
membranes taken from the outlet end but not the one collected from the
inlet end. Cleaning did not restore the clean water oxygen transfer
efficiency to a like-new condition. The inability to restore the clean water
oxygen transfer efficiency may have been caused by changes in the
membrane materials as indicated by the lower than new DWPs of the
cleaned diffusers.
The observations made and the data collected on diffusers from the
reaeration basin before and after cleaning indicate that the fouling was
relatively uniform from inlet to outlet. The foulant was similar to that
found on the ceramic diffusers. It consisted of two layers, a slimy material
over a hard black crusty material. Both layers were effectively removed
by in-basin hosing, brushing and rehosing as evidenced by the after
cleaning DWPs and visual observations.
The DWPs of the dirty diffusers were not much different than the
average new membrane. Cleaning of the membranes lowered the DWP
substantially below the typical value for a new membrane and the clean
water oxygen transfer efficiency after cleaning did not approach the level
attained by a new diffuser such as GBN.
-------
15
4=
o
CM
t3 10
O)
3:
cf
|- 5
O
0
" CONTACT BASIN
•
£ *f"
A
&
| | j |
F C F C
REAERATION BASIN
O
O
% 0,New
0
1 IB II I
F F C F C F
1 .U
2
CD
0
LLJ
H—
§ 0.5
Nt
UJ
h-
0
CO
0
" A* A A
&
! ^
^
•
I | II
F C F C
C4.1 C4.5/6
New
A
0 • ft ° •
• o •
(
I I 1 II 1
F F C F C F
R4.1 R4.2 R4.5 R4.6
Grid & Condition
F = Fouled
C = H-B-H (Field)
A • = Not Flexed
A o = Flexed
FIGURE 14
Membrane Diffuser Test Results -
Grid Units Removed November 1986
44
-------
Flexing of the membrane diffusers did not control the fouling in either
the contact or the reaeration basin. DWPs measured on the flexed and
unflexed diffusers were nearly equal. This is consistent with observations
that the air flow rate to a recently flexed grid, if left unthrottled, was
greater than the air flow rate to the unflexed grids because the
relationship between DWP and air flow rate is such that small differences
in DWP result in large differences in air flow rate. The clean water
oxygen transfer efficiency of the fouled diffusers both flexed and
unflexed were nearly equal.
Because the benefits of the in situ acid gas treatments of the ceramics in
the contact basin were minimal, the number of grids receiving the
preventive maintenance treatment was reduced. Starting in December
1986, only one grid (C2.2N) in contact basin 2 was given monthly in situ
acid gas treatments. On the other hand, five of the six grids in the
reaeration basin (R2.1 through R2.5) were acid gas treated monthly
because the acid treatments appeared to minimize the black foulant layer
which was encountered in this basin.
The number of membrane tube equipped grids that were flexed was also
reduced from half to two in the contact basin (C4.2N and C4.4N), and
one in the reaeration basin (R4.1).
July 1987
The results of the laboratory tests conducted on the ceramic diffusers
removed in July 1987 are presented in Figure 15. The diffusers taken
from the contact basin showed moderate increases in BRV and DWP
except for the diffuser that did not receive the monthly acid gas
treatments. The BRV measured for.the untreated diffuser was 23.6
inches of water gauge (w.g.). The unit that received monthly acid gas
treatments showed a smaller increase in BRV and DWP than the
untreated unit. The rigorous hose, acid gas and rehose cleaning
procedure restored the BRV and DWP of both units to a like-new
condition.
The units taken from the reaeration basin had higher BRVs and DWPs
than the units taken from the contact basin. The unit that received
monthly acid treatments was less fouled, having a lower BRV, DWP and
visibly less of the black deposits. The DWPs of both units were returned
to a like-new condition after cleaning by the hose, acid gas, and rehose
method. However, not all of the black foulant material was removed by
the cleaning effort
The fouled ceramic diffusers from the contact and reaeration basins
showed losses in clean water oxygen transfer efficiency of about 25
45
-------
CONTACT BASIN
REAERATION BASIN
CL CF
25
0,20
= 15
QC
m 10
5
0
30
1 25
to 20
6>
Si 15
c
^ 10
o
0
A
•
.
4
F 1 F
F CL CF
•
•
' A
• ^V
1 1 *
— \^, •**.•** i
0
New
i it pi
F F CL F CL
*
0 o
o 9
^New
\ * II * •
F CL
F = Fouled
CL= H-AG-H (Laboratory)
CF=H-AG-H (Field)
A • = Not Acid Gassed
A o = Acid Gassed
F CL
1 .U
0
UJ
§ 0.5
\
UJ
B
CO
O
,^jr
A
f | I
F CL CF
C2.2
New
o
.
Urn * ' I *t
F F CL F CL
R2.2 R2.5 R2.6
Grid & Condition
46
FIGURE 15
Ceramic Drffuser Test Results -
Grid Units Removed July 1987
-------
percent. In the contact basin, the acid gas treated diffuser had a higher
clean water oxygen transfer efficiency than the untreated unit. However,
the number of diffusers tested was small so it is not possible to determine
if the monthly acid gas treatments provided a real benefit in terms of
limiting losses in clean water oxygen transfer efficiency. For the contact
basin units, the hose, acid gas, and rehose cleaning procedure resulted
in an increase in the clean water oxygen transfer efficiency to about 95
percent of new.
The clean water oxygen transfer efficiency of the all diffusers taken from
the reaeration basin were essentially equal. The increase in clean water
oxygen transfer efficiency after cleaning was less for the diffusers from
the reaeration basin.
The results of the July 1987 laboratory tests conducted on the membrane
diffusers are presented in Figure 16. The DWPs of the fouled diffusers
taken from the contact basin were less than or equal to a typical new
diffuser. Fouled diffusers from the reaeration basin showed moderate
increases in DWP. Cleaning by the hose, brush, and rehose method
reduced the DWP of all the units to a level substantially less than a new
diffuser. The clean water oxygen transfer efficiency of the fouled and
cleaned diffusers were essentially equal. Flexing of the membranes did
not appear to provide any benefit over not flexing.
Several of the used membrane diffusers were further evaluated to
determine which membrane characteristics, if any, had changed. The
membranes were measured, weighed and tested for elasticity and
hardness. The results of the measurements and tests are summarized in
Table 17. All of the membranes tested showed changes in dimensions,
weight and elasticity. Hardness was essentially unchanged.
The average increase in diameter was about 0.11-inch or 5 percent while
the average membrane thickness was reduced by about 0.0025-inch or 8
percent. The membrane material became stiffer as indicated by the
relatively large increase in tensile modulus. These quantitative changes
are in addition to the visual observations made during the diffuser testing
which indicated that the slits in the used membranes were wider than the
ones in the new membranes and that the slits did not close completely
after the air was shut off. The changes in the slits were documented by
photographing new and used membranes operating at 10 in. w.g. The
used membranes were cleaned either by the hose, brush and rehose
method or by flushing the slits with clean water from the inside out while
brushing. The flushing method was suggested by the membrane diffuser
manufacturer. Several of the photographs, taken by Midwest Research
Microscopy Inc. (Milwaukee, WI), are reproduced in Figures 17,18 and
47
-------
15
,£
"o
H 10
CO
q>
^
H- 5
Q
0
4 f\
1 .U
m
O
LJJ
S 0.5
CO
x.
LU
w o
• UUNIAUI BASIN
.
4
A
•
f i • •
F C F C
A^_ AA A
•
i
1
F C F C
C4.2 C4.4
HtAtHAIKJN BASIN
o
New
• o
• • • »
F G F C
New
• • o0
••
F C F C
R4.4 R4.5
Grid & Condition
F = Fouled
C = H-B-H (Laboratory)
A • = Not Flexed
A o = Flexed
FIGURE 16
Membrane Diffuser Test Results
Grid Units Removed July 1987!
48
-------
3
n
« r
. ! t < 11
:.'},-, . | (
-------
New Membrane Operating at DWP of 10 in. w.g.
Used Contact Basin Membrane Operating at DWP of 10 in. w.g.
FIGURE 18
Photographic Comparison
New Versus Flush/Brush
Cleaned Contact Basin
Membrane Diffuser
50
-------
New Membrane Operating at DWP of 10 in. w.g.
Used Reaeration Basin Membrane Operating at DWP of 10 in. w.g.
51
FIGURE 19
Photographic Comparison -
New Versus Hose-Brush-Hose
Cleaned Reaeration Basin
Membrane Diffuser
-------
Table 17 • Comparison of New and Used
Grid/Flexed Weight Length Diameter
DiffuserNo. gm. in. in.
Typical New
Mean 120
Sample Size 8
COV2 0.02
"R4.4/No~ ""
K-45-65-1 112
R4.5/Yes
K-45-65-3 114
C4.4N/Yes
K-45-67-2 114
C4.4S/NO
K-45-67-3 113
G4.2N/Yes
K-45-67-5 111
C4.2S/NO
K-45-67-6 118
Average 114
263
5
0.01
25.0
25.1
25.7
25.4
25.7
26.0
25.5
2.25
5
0.01
236
237
233
239
239
233
236
Membranes
;
Thick- Tensile
ness Modulus Hard
in. psi ness1
0.0311 630
8
0.05
0.0280
0.0278 940
0.0280
0.0272 860
0.0263
0.0287
0.0277 900
Percent
Change3 -42 -2.7 +4.9 -83 +45
1 Shore A Durometer, average of five points on the membrane, from Table 10.
2 Coefficient of variation, standard deviation/mean.
3 100 x (Used - New)/(New)
63.2
10
0.03
63.1
62.5
61.4
63.1
635
62.7
633
-0-5
19. The slits in all the used membranes were wider than the slits in a
typical new membrane.
Foulant Analysis
Samples of foulant materials were collected from several ceramic disc
and membrane tube diffusers after they were removed from; the contact
and reaeration basins during November 1986 and July 1987.
The diffusers taken from the reaeration basins were covered by two
foulant layers. The outer layer was a loosely attached slime-like material
that could be removed by hosing with water at a nozzle discharge
pressure of 30 psig. Visual and tactile observations indicated that this
layer comprised biological slime into which a gritty, sand-like material
was incorporated. A photograph of a cross-section of the slime layer is
presented in Figure 20.
52
-------
Metallographically Polished Cross-Section of
Residues Showing a Variety of Particles
Similar to Soil or Rock.
Contact Basin
Macrograph !
C40689 (100x)
FIGURE 20
Photograph of Foulant
From a Contact Basin
Ceramic Diffuser
53
-------
The inner layer was black, hard, and firmly attached to the surface of the
diffusers. Analysis of the black foulant by electron dispersive
spectroscopy showed that the major constituents were iron (18.5
percent) and calcium (32 percent). On the ceramic diffusers, this layer
was not uniformly distributed. When the outer layer was removed, the
inner layer looked like splotches of black tar. The distribution of this
foulant on the membrane diffusers was not readily discernible because
the membrane material was also black.
The results of analyses conducted on the foulant samples taken from the
ceramic diffusers are summarized in Table 18. The mass of the slimy
foulant layer varied substantially from diffuser to diffuser, and the
variations were random, i.e., there was no clear pattern of more foulant
at a basin inlet or outlet end. The regular acid gas treatments did not
substantially reduce the slimy foulant mass deposited on a diffuser even
though it resulted in a marked reduction in DWP. The volatile fraction
of the slimy layer was relatively constant with values between 13 and 24
percent. The low volatile content is explained by the presence of the
gritty material which was probably sand particles. The acid gas
treatments did reduce the acid soluble fraction in the foulant layer but
this was not a substantial portion of the total foulant mass.
The mass of the hard foulant layer was small compared to the slimy layer.
While the buildup was much less during the second six months of
operation, the mass of solids deposited was reduced substantially during
both periods by the regular in situ treatments with the HC1 gas.
The results of analyses conducted on the foulant samples taken from the
membrane tube diffusers are summarized in Table 19. The foulant
samples collected during November 1986 were chosen by visually
selecting a representative area on the diffusers. Because of the geometry
of the tube diffusers, samples from the top and bottom of the units were
collected and analyzed in July 1987. The November results indicate
substantial differences in the fouling of the membrane and ceramic
diffusers. The mass deposited on the membranes was much less than on
the ceramics, while the volatile fraction was higher and the acid soluble
fraction was equivalent The differences are explained by the results of
the analyses of the top and bottom foulant samples. The mass of foulant
material makingup the outer slimy layer was not evenly distributed. Most
of the mass was on the top of the diffuser. The top portion of the foulant
is very similar to the foulant found on the ceramic diffusers indicating
that the fouling mechanisms for the top of the membranes and the
ceramics are similar.
The foulant material collected from the bottom of the diffusers had much
less mass per unit area, much higher volatile fractions and lower acid
54
-------
£,
A
£
3
CO
cc
§^^
2°o^&c3 °S9
VO f- SO ^- v-j
"^^SSCJ ... wjoffivd^
^ oo cs r^i c4 o f^
K o « .
*? ^ S S ?5 ... g \0 yg ^ j _i
05 " 6
1
Ur
"i
1
*™J Ov OO CO ^f O C*J ^5 *O
^^ CO V) CO 00 ^O ^5 ^ CO ^5 ^5 ^D CO
£ N^u^ ^^r. g % 5;
^ ^ *O *O ®i ^ vo o ^
cs o *o m o\ »Q oo *H j2 c^ t*~ c** ^
u
ONC^ON CJOC*^ ONO*H
"pf~ < oca; fc > < U33 o < & > <
55
-------
03
E
3
cn
t/3
C
J2
I
L.
c
o\ oo o i— •
z
6
oo oc
£
S3
oo
oe
f-s
CS *t£
0. o g. O O. O
O s O — O —
E- o f- . o f- o
oa CD ea
o,
§• 1
f- o
03
o;
.0
i
i '
ON
'™H
~
IB
P"
IM
_e
"«
CO
CQ
C
^o
2
^
u
Oi
Si
J
S
CQ
g
S.
>~
««
r
i>
p3
u
£
60
•0 "1
'C u
0 E
~£
I
t«
CQ
S
Is
_w
3
£
0
1
SR
•§
r-
§• 1 §• 1
F- o E- c
03 02
?R
SR
-
*
.
0
B
OQ
C. •
|
=
cS
2
"
.
CJ
:«3
-tO
ON
0 0
•
-------
soluble fractions. These differences indicate that gravity settling may
play a role in the deposition of inert solids in the slimy foulant layer.
Removal of the inert solids before they reach the aeration basins could
possibly reduce the mass of foulant that builds up on the top surfaces of
the diffusers. The clean water oxygen transfer efficiency tests conducted
on the ceramic diffusers before and after hosing (refer to Tables 12 and
13) showed that the slimy foulant layer was responsible for most of the
losses in oxygen transfer efficiency. Since removing the inert solids may
reduce the total mass of foulant deposited, there is reason to believe that
the rate at which the overall system efficiency decreases would be
reduced too.
Cost Analysis
To compare the sparged turbine, ceramic disc and membrane tube
aeration systems, the capital, and the operation and maintenance (O &
M) costs were put on a common basis comprising treatment of 55 million
Ib BODs per year (150,700 Ib/day), approximately the average plant
loading from 1983 through 1985. The costs were also adjusted to reflect
the hypothetical situation in which all four quadrants have the same
aeration system. This situation is used for the cost analysis because it
provides a more equitable evaluation of the capital costs such as the acid
gas feed facilities which are the same for one or four quadrants.
Capital
The capital costs for the two fine pore diffuser aeration systems are
summarized in Table 20. The capital costs shown for the retrofit of one
quadrant are the 1985 construction costs determined by competitive
bidding. The original bid included the fine pore aeration systems, the in-
line air filter for the ceramic diffuser system and the numerous pilot
headers provided for the comparative study. The acid gas storage and
feed building and the acid feed equipment were added to the original
contract by change order. The 1985 construction costs were not escalated
because prices have not risen appreciably over the last few years and
because the original construction package included extra features such
as the numerous pilot headers which would not be needed in the retrofit
of the remaining quadrants. In addition, the competitiveness among
manufacturers of fine pore aeration equipment has increased. To
provide some factor of conservativeness, twenty percent was added to
the estimated 1988 construction costs to cover engineering and
contingencies.
For the ceramic diffuser system, it was assumed that additional in-line
air filters or a central filtration system located at the inlet to the
compressors would not be necessary. Four ceramic diffusers were
57
-------
Table 20 - Summary of Capital Costs
One Four
Component Quadrant Quadrants
Ceramic Discs
Aeration System $386,400 $1,545,600
Air Filtration 101,300 .. Q
Acid Gas Feed Building 68,900 68,900
Gas Cleaning Royalty2 ' 59,200 236*800
Subtotal 615,800 1,851,300
Engineering & Contingencies 44,5003 370,3004
Total $660,300 $2,221,600
Membrane Tubes
Aeration System $416,300 $1,665,200
Engineering & Contingencies 44,5003 333,Q004
Total $ 460,800 $ 1,998,200
1 Based on Contractor's 1985 Schedule of Values plus change orders.
2 U.S. Patent No. 4,382,867 held by Water Pollution Control Corporation, Milwaukee
Wisconsin 53201.
3 One-half of the actual engineering design and construction fees.
4 Engineering and contingencies estimated at 20 percent of construction.
installed on a pilot header provided by the ASCE for an inter-plant
fouling study. The ASCE header was placed in quadrant 4 and was fed
air that did not pass through the in-line filter. After approximately 16
months of operation, there was no sign of air-side fouling based on BRVs
measured on the air side of the diffusers.
The air diffusion equipment cost estimate for the ceramic disc fine pore
diffuser system was about $ 120,000 less than the membrane tube diffuser
system. Even with the additional cost for the acid gas feed facilities, the
ceramic system was about $50,000 less. The cost item that calces the
ceramic system more expensive was the royalty paid for practicing the in
situ acid gas cleaning system. To clean all the diffusers in both the contact
and reaeration basins would cost nearly a quarter of a million dollars.
Operation
The costs for operation of the two fine pore aeration systems and the
sparged turbine aeration system include the labor for process monitoring
and adjustment of the activated sludge control parameters, and the
electrical power to run the air compressors and the turbine mixers. Since
all three aeration systems require an equivalent amount of effort for
58
-------
monitoring the activated sludge process and adjusting the air flow rate
to the basins, the costs were not included in the comparison.
A comparison of the electrical power usage before and after the two
quadrants were retrofitted with the fine pore diffusers is presented in
Table 21. In the years 1983,1984 and 1985, the sparged turbine aeration
system was used exclusively. The average annual electrical power usage
was 25.8 million kilowatt-hours (KWH) and the cost was $1,032,200
($0.04/KWH). The average unit cost was $18.56/1000 Ib BODs treated.
In comparison, the power usage during the first 12 months of operation
with the two quadrants of fine pore diffusers (and usually one sparged
turbine contact basin) was 18.0 million KWH at an total cost of $721,000.
The unit cost was $12.42/1000 Ib BODs treated. This was a reduction in
electrical power usage of 7.8 million KWH for a savings of $311,200 or
30 percent less than the three-year average of sparged turbine operation.
Based on the unit costs, the reduction was slightly greater at 33 percent
($12.42 versus $18.56/1000 Ib BODs treated) because more BODs was
treated during the first 12 months of the study period. The last six months
of the study period were not included in the comparison because the
BODs loading to the plant was substantially lower and at least for part
of the period, the activated sludge system was in an upset condition.
The electrical power usage and cost for each of the two fine pore aeration
systems were estimated based on the air volumes used. The sum of the
monthly average contact and reaeration basin air flow rates were
converted to electrical power usage as follows:
P = Q (0.032) (24) (N)
Where:
P = monthly average power usage, KWH/month
Q = monthly average air flow rate to the quadrant, scfm
0.032 = kilowatts per scfm at a discharge pressure of 9.5 psig
taken from the 1984 compressor test report7, KW/scfm
24 = hours/day
N = number of days operated during the month, day/month
The power used to compress air for operation of the quadrant with the
sparged turbines was calculated as the difference between the usage for
the fine pore systems and the total billed amount for the month taken
from GBMSD records. Using this method essentially assigned all of the
59
-------
Table 21 - Electrical Power Usage and Cost Summary
Parameter Mixers Compressors
1983 (56.8 x 106 !b BOD5 treated)
Usage, KWH 10.8 x 10* 13.4 x 106
Cost,$ 432,400 . 537,000
$71000 Ib BODs 7.61 9.45
1984 (58.4 Ib x 106 BODs treated)
Usage, KWH 1 1.9 x 106 16.6 x 106
Cost, $ 477,800 665,200
$71000 Ib BODs 8.18 11.39
1985 (51.6 x 106 Ib BOD5 treated)
Usage, KWH 10.5 x 106 14.1 x 106
Cost, $ 419,000 564,900
$71000 Ib BODs 8.12 10.95
Average 1983 - 1985 (55.6 x 106 Ib BODs/year treated)
Usage, KWH 1 1. 1 x 106 14.7 x 10*
Cost,$ 443,200 589,000
$71000 Ib BODs 7.97 10.59
5/86 - 4/87 (58.1 x 106 Ib BODs treated)
Usage, KWH 3.6 x 106 14.4 x 106
Cost, $ 144,600 577,100
$71000 Ib BODs 2.49 9.93
Total
24.2 xlO6
969,000
17.06
28.5 x 106
1,143,000
19.57
24.6 x 106
984,400
19.07
25.8 x 106
1,032,200
18.56
18.0 xlO6
721,000
12.42
Note: Average cost per kilowatt-hour from 1983 through 1987 was S0.04/KWH. ',
air fed to the reaeration basins to the fine pore systems although for most
of the study one of the reaeration basins provided return activated sludge
to both a fine pore diffuser equipped contact basin and a sparged turbine
equipped contact basin.
The power costs were calculated from the usage data using the rate of
$0.04/KWH. Because the quadrants did not treat an equal quantity of
BODs each month, the cost data were normalized by dividing the total
cost by the pounds of BODs treated. The electrical power usage and cost
data are tabulated in Appendix D.
. The power costs per unit of BODs treated are plotted versus time in
Figure 21. The ceramic diffusers provided more economical treatment
than the membrane diffusers in May 1986. After the first month,
60
-------
a
00
CVJ
CVJ
a co
CO 00
w o>
•2 *~
0) CO
O) CD
CO C
a
•P
/
a
a
a
o
CN
-------
however, the oxygen transfer efficiency of the ceramic diffusers
decreased resulting in higher operating costs compared to the membrane
diffusers. No further increase in treatment cost was experienced after the
third month of operation. During the same initial six months of
operation, the membrane diffusers showed more variation but no
increasing or decreasing trend in power costs. During the first six months
of operation, the membrane diffusers had lower unit power costs than
the ceramic diffusers, $10.64 versus $11.73/1000 Ib BOD5 treated,
respectively.
After cleaning in November 1986, the cost of operating the ceramic
diffusers was reduced to the May 1986 level, approximately $9.00/1000
Ib BODs treated. Cleaning the membrane diffusers also resulted in a
decrease in the unit power costs to about $9.00/1000 Ib BODs treated.
Between December 1986 and the end of April 1987, the two fine pore
diffuser systems operated at equivalent unit costs. Not including the
February 1987 value for the membrane diffuser system, which is
considered an outlier (see discussion of efficiency factors), the average
unit power costs for the ceramic and membrane diffuser systems were
$9.95 and $9.66/1000 Ib BODs treated, respectively.
In May 1987, the activated sludge system was upset so the air flow rates
were not a good indicator of the electrical power actually required to
treat the incoming BODs. By the middle of June, the activated sludge
system had recovered but the efficiency of the fine pore aeration systems
was low and the energy usage was high so the diffusers were cleaned. The
membranes were cleaned the last week in June and the ceramics were
cleaned in July.
After cleaning, the power costs for the ceramics were again reduced to
the original like-new condition. The average unit power cost for the last
three months of the study was $9.64 /lOOO Ib BODs treated. The unit
power costs for the membranes after cleaning were reduced but not to a
like-new condition. The unit power costs remained high, averaging
$15.40/1000 Ib BODs for the last four months of the study. The
substantially higher unit power costs for the membranes are consistent
with the losses in oxygen transfer efficiency measured during the clean
water oxygen transfer efficiency testing done on the pilot and grid
diffusers, as well as the measured air requirement per Ib of BODs in the
membrane quadrant
Maintenance
The annual average maintenance costs for the sparged turbine and the
two fine pore diffuser aeration systems are summarized in Table 22. The
costs are based on having all four quadrants equipped with the same
62
-------
Table 22 - Estimated Maintenance Costs
Aeration Estimated
System Maintenance Item/Activity Annual Cost
Sparged Compressors . $ 5,400
Turbines Turbine Mixers 51*600
Total • • • $57^000
Ceramic Compressors $5 490
Discs Diffusers:
1) Drain contact and reaeration basins twice annually
and clean the diffusers
a) Labor (hosing) - 24 hours/quadrant/cleaning
@$30/hour $5,800
b) Labor (HC1 gassing) - 16 hours/quadrant/cleaning
@$30/hour . _ 35800
c) Chemicals - HC1 gas @ $0.20/diffuser/cleaning and '
33,104 diffusers 13,200
2) In situ acid gas treat reaeration basin
diffusers monthly2
a) Labor - 6 hours/quadrant/cleaning @ $30/hour 7,200
b) Chemicals - HC! gas @ $0.20/diffuser/cleaning ''
and 8,592 diffusers1 17,200
Diffuser subtotal $47,200
Total $52,600
Membrane Compressors $ 8 100
Tubes Diffusers:
1) Drain contact and reaeration basins annually
and clean the diffusers
a) Labor - 72 hours/quadrant/cleaning ,
@$30/hour 8,600
b) Supplies - none . 0
2) Flexing not required 0
Total $ 16,700
6,128 diffusers/contact basin and 2,148 diffusers/ reaeration basin.
Ten monthly treatments/year in addition to the two treatments associated with the semi-annual
cleaning.
63
-------
aeration system. For the fine pore aeration systems, the basis is operation
of all four quadrants, each quadrant having the same number of diffusers
that were installed in the test quadrant.
For the sparged turbines, the compressor and turbine maintenance costs
are based on the data assembled and reported in the 1983 Predesign
Report escalated 5 percent per year to 1988.
The compressor maintenance costs for the ceramic diffuser alternative
are estimated to be equal to the sparged turbine alternative because the
air usage for the two types of equipment are equivalent. The
maintenance costs for the ceramic disc diffusers include draining each
contact and reaeration basin twice annually for inspection and cleaning
of the diffusers. The cleaning procedure comprises hosing the diffusers
from the basin walkways using fire hoses followed by an in situ acid gas
treatment immediately after the basin is put back into service. The labor
estimate for hosing the diffusers is based on two operators for 8 hours
each for a contact basin and 4 hours each for a reaeration basin. For the
acid gas treatments, the estimate is for two operators for one half hour
per grid. Monthly acid gas treatments are included for the reaeration
basin diffusers to minimize the effects of the black foulant layer.
The estimated compressor maintenance costs for the membrane tube
diffuser alternative is based on the diffusers having the oxygen transfer
efficiency measured at the end of the study when considerably more air
was required. Since more than one compressor would be required at
times to meet the additional air requirement, the maintenance costs
would be higher. An increase of 50 percent was used for the estimate.
When the diffusers were new and the air volume compressed was
equivalent to that for the sparged turbines and the ceramic disc diffusers
the compressor maintenance costs would be equal for all three
alternatives.
Since cleaning of the membrane diffusers did not provide a substantial
improvement in the oxygen transfer efficiency and the DWP of the
fouled diffusers was not excessive, a minimal diffuser maintenance effort
comprising an annual draining of the contact and reaeration basins for
inspection and diffuser cleaning is used. The labor estimate is based on
six operators working eight hours each to clean a contact basin and four
hours to clean a reaeration basin. The cleaning procedure would be
hosing from the basin floor and brushing with a stiff bristled scrub brush.
Flexing of the membranes is not necessary.
The sparged turbines have the highest estimated annual maintenance
costs at $57,000, followed by the ceramic disc diffusers at $52,600 and
the membrane tube diffusers at $16,700.
64
-------
Summary
The estimated capital and annual O & M costs are summarized in Table
23. The unit power cost for the turbines is based on the average for the
period from 1983 through 1985. A unit cost of $10.00/1000 Ib of BODs
treated was used for the ceramic system because it appears to be a
reasonable estimate for a 6-month run between diffuser cleaning. The
average cost for the period from December 1986 through April 1987, a
5-month period, was $9.95/1000 Ib of BODs treated and for the period
from August 1987 through October 1987, a 3-month period, was
$9.64/1000 Ib of BODs treated. The unit cost of $15.40/1000 Ib of BODs
treated was used for the membranes. This was the average cost associated
with the last four months of the study.
The ceramic disc diffuser system is estimated to save about $471,000 per
year in electrical power costs and a total of $475,000 per year in total O
& M costs compared to the sparged turbine aeration system. The payback
on the capital investment would be about four and a half years.
The membrane tube diffusers would have shown similar economics had
they not lost their initial high efficiency. Based on the new diffusers, the
annual savings in total O & M costs are estimated at $514,000 which
Table 23 - Alternative Aeration Systems Cost Summary
Cost Parameter
Sparged Ceramic Membrane Tubes
Turbines Discs New Used
0 $2,221,600 $1,998,200 $1,998,200
1,020,800J
57,000
0
550,000"
5,400
47,200
550,000"
5,400
8,600
Capital
Operation and Maintenance1
Electrical Power2
Mechanical Maintenance
Diffuser Cleaning
Total O&M
O&M Present Worth6
Total Present Worth7
Annual cost.
2 Based on treating 55 million Ib BOD5/year.
Based on a unit power cost of $18,56/1000 Ib BODS treated.
Based on a unit power cost of $10.00/1000 Ib BODS treated.
Based on a unit power cost of $15.40/1000 Ib BODS treated.
Present worth based on 20 years @ 8-7/8 percent interest, present worth factor = 921
Capital + O&M Present Worth.
847,000-
8,100
8,600
$1,077,800 $ 602,600 $ 564,000 $ 863,700
$9,927,500 $5,549,900 $5,194,400 $7,954,700
$9,927,500 $7,771,500 $7,192,600 $9,952,900
65
-------
would have provided a payback period of about four years. However, the
lost transfer efficiency resulted in a substantial increase in the electrical
power requirements and actual O & M savings of only $214,000 per year.
The payback period under these circumstances would be more than nine
years. Based on a 20- year life of the equipment and an interest rate of
8-7/8 percent, retrofitting with the membrane tube diffusers would not
be economical because the total present worth of the retrofit is greater
than the cost of continuing to operate the sparged turbines.
CONCLUSIONS
The activated sludge system at Green Bay provided a severe test for the
ceramic disc and membrane tube diffuser systems. The fouling on both
types of diffusers appeared to be exacerbated by the inert solids which
probably entered the aeration basins with the mill wastewaters. The
plasticized PVC membrane material appears to have been adversely
affected by environmental conditions including the temperature of the
mixed liquor, which had summer monthly averages above 85°F.
In situ acid gas treatments of the ceramic diffusers and flexing of the
membrane diffusers did not control deposition of a loosely attached
slimy foulant layer which had a substantial adverse effect on oxygen
transfer efficiency. The in situ acid gas treatments of the ceramic
diffusers controlled increases in DWP in both the contact and reaeration
basins, but especially in the reaeration basin where an kon-orataining
precipitate formed on the ceramic diffusers, resulting in a marked
increase in backpressure.
The oxygen transfer efficiency and operating pressure of the ceramic
diffusers could be restored to a like-new condition by draining the basins
and cleaning the diffusers with a three-step procedure consisting of
hosing from the walkways, acid gas injection, and rehosing. The
maintenance procedure could be shortened without reducing the
effectiveness of the cleaning if after the initial hosing, the basin was put
back into service and the in situ acid gas treatment was performed
immediately.
After 10 months in operation, cleaning the membranes by hosing from
the basin floor, scrubbing, and rehosing, reduced the operating pressure
to values substantially below typical new membranes, but did not restore
the oxygen transfer efficiency. The unrecoverable loss in oxygen transfer
efficiency was about 25 percent based on clean water test results. Data
from the full-scale operation indicated that the permanent loss may have
been even greater.
66
-------
Retrofitting the Green Bay Metropolitan Sewerage District's aeration
basins with fine pore diffusers was cost-effective. A capital investment
of about $1 million resulted in a first year electrical power savings in
excess of $300,000.
Further savings would be expected if all the sparged turbine aerators
were replaced with ceramic fine pore diffusers or membrane diffusers
that did not lose oxygen transfer efficiency with use.
The oxygen transfer efficiency of the ceramic diffusers could be
increased by about 10 percent if the air flow rate per diffuser were
reduced to the normal design rate of 1.0 to 1.25 scfm per unit from the
over 2.0 scfm/unit operating rate used during most of the study. From an
oxygen transfer efficiency standpoint, this could be best accomplished by
retrofitting more basins with the fine pore diffusers and then operating
all available basins.
RECOMMENDATIONS
Based on the results of the 18-month evaluation of the two fine pore
diffuser systems, it would be very economical for the GBMSD to retrofit
quadrants 1 and 3 with ceramic fine pore diffusers. However, until the
GBMSD completes their facility planning effort and a decision can be
made regarding further retrofits, we recommend that the following
interim activities be pursued:
1. Continue to monitor the operation of the fine pore aeration systems
to build a data base for operation on the wastewater that has been
received since the mill changes encountered in May 1987. The ef-
ficiency of the two aeration systems can be monitored by calculating
efficiency factors (see Appendix A) and verifying the efficiency fac-
tors by conducting infrequent off-gas analyses.
2. Continue to in situ acid gas treat all the reaeration basin ceramic dif-
fusers monthly to minimize buildup of the black foulant material.
3. In situ acid gas treat one ceramic grid in the contact basin monthly to
investigate if there is a long term benefit to providing regular treat-
ments between draining of the basins for restorative cleaning.
4. As an alternative to using gaseous HC1 as the second step in the res-
torative cleaning procedure, try using liquid HC1. The procedure
comprises three steps: hosing of the diffusers from the walkways, ap-
plying 14 percent HC1 using stainless steel spray applicators and
rehose after 20 to 30 minutes. This method has proven to be effective
in a number of plants. If handling of the liquid acid is not a problem,
67
-------
the cost of acid cleaning would be substantially reduced because the
gas cleaning royalty would only have to be paid on diffusers installed
in the reaeration basins.
5. Test one or more alternative membrane tube diffusers in the reaera-
tion basin where one grid contains 233 difrusers. There are membrane
diffusers available that have the 3/4 inch pipe thread connector now
being used and are made of materials other than plasticized PVC. One
of the alternative materials may provide the advantages experienced
with the membrane difrusers but without the relatively rapid loss in
oxygen transfer efficiency.
6. Continue to run unfiltered air through a few ceramic difrusers at 2 to
3 cfm per unit and monitor the air-side BRV periodically to see if ad-
ditional air filtration is needed in the final design.
68
-------
REFERENCES
1. Donohue & Associates, Inc., "Predesign and Feasibility Report for Aeration System
Modifications", Report prepared for the Green Bay Metropolitan Sewerage District
December 1983. ^
2. Saiiitaire - Water Pollution Control Corporation, "Oxygen Transfer Shop Test Report,"
Report prepared for the R. P. Honold Company, Sanitaire job no. 85-1206, September 1985.
3. Parkson Corporation and Gerry Shell Environmental Engineers, Inc., "Oxygen Transfer
Evaluation of the Wyss Flex-A-Tube Aeration System," Report prepared for the R.P. Honold
Company, August 1985.
4. Redmon, D.T., Boyle, W.C., Ewing, L., "Oxygen Transfer Efficiency Measurements Using
Off-Gas Techniques," Journal WPCF, Vol. 55, No. 11, November 1983.
5. Ewiing Engineering Company, Inc., "Results of Off-Gas Testing at the Green Bay
Metropolitan Sewerage District - May 1986 through October 1987," Report prepared for
Donohue & Associates, Inc., June 1988.
6. Allbaugh, T.A., Benoit, DJ., Spangler, J., "Aeration System design Using Off-Gas
Oxygen Transfer Testing," Paper presented at the 58th Annual Conference of the Water
Pollution Control Federation, October 6-10,1985, Kansas City, Missouri.
7. Domohue & Associates, Inc., "Aeration System Predesign Test Program," Report prepared
for the Green Bay Metropolitan Sewerage District, September 1984.
8. Box, G.E.P., Hunter, W.G., and Hunter, J.S., Statistics for Experimenters - An
Introduction to Design. Data Analysis, and Model Building, John Wiley & Sons, New York,
A y I o*
9. Volk, W., Applied Statistics for Engineers. McGraw-Hill Book Company, New York, 1969.
69
-------
APPENDIX A
EFFICIENCY FACTOR CALCULATION DESCRIPTION
70
-------
APPENDIX A
Efficiency Factor Calculation Description
Contact basin efficiency factors are calculated as follows:
EF = BODs / Q. / FDO
Where:
EF = contact basin efficiency factor, Ib BODo/day/cfro
corrected to zero dissolved oxygen concentration
BODs = BODs loading to the contact basin, Ib/day
Qm = nominal air flow rate to the contact basin based
on the computer logged data, cfm
FDO = dissolved oxygen correction factor, decimal
fraction
=
-------
Table Bl - Clean Water Dissolved Oxygen Saturation
Concentrations as a Function of Time
Temperature
°F
67
68
69
70
71
72
73
74
75
76
77
78
Concentration
mg/1
3. 19
9.09
8.99
8.89
8.80
8.71
8.62
8.53
8.44
8.35
8.26
8.18
Temperature
op
79
80
81
82
83
84
85
86
87
88
89
9O
Concent r at i on
mg/1
8. 10
8.02
7.94
7.86
7.78
7.71
7.63
7.56
7.49
7.42
7.35
7.28
72
-------
APPENDIX B
STATISTICAL REVIEW OF THE OFF-GAS TEST
RESULTS
73
-------
APPENDIX B
Statistical Review of the Off-Gas Test Results
CONTACT BASINS
The statistical significance of the differences observed in
the basin average alpha'-SOTE data was evaluated by analysis
of variance and the F test.B Differences in the average
alpha'-SOTE data were considered significant if the
calculated variance ratios were greater than F at the 0.05
significance level, i.e., the probability that the average
alpha'-SOTE values are equal is less than 5 percent.
The analysis of variance test was performed on alpha'-SOTE'
values that were calculated as follows:
alpha'-SOTE' = (alpha' -SOTEi ) (Fi )7'Fb
Where:
alpha'-SOTE' = local value of alpha'-SOTE adjusted for
variations in the off-gas flux within the
bay, decimal.
alpha'-SOTEi = local value of alpha'-SOTE, decimal.
Fi = local off-gas flux, scfm/sq ft.
Fb = bay average off-gas flux, scfm/sq ft.
Each bay comprises one-sixth of the aeration basin. Usually,
four local measurements were made in each bay. The one
exception was December 1986 when the daylight hour were
1i mi ted.
The data were blocked using the six bays. In this way, the
analysis of variance can test the significance of variations
among the bays.
The alpha'-SOTE, off-gas flux, alpha'-SOTE' and various
parameter averages are presented in Tables Cl through C9. The
tables were prepared to provide the data necessary for
calculating the following quantities needed to assemble an
analysis of variance table for each test.
1. Sum of squares of deviations associated with the blocks
(bays), SB.
SB = ink SUMi CCY'i-Y' )*]
2. Sum of squares of deviations associated with the
treatments (aeration equipment), ST.
ST = mn SUMt C(Y't-Y')a3 ;
74
-------
3. Sum of squares of deviations associated with the
interaction between the blocks and the treatments, SI.
SI = m SUMt SUMi C=]
4. Residual sum of squares, SR.
SR = SUMt SUMi SUMj CCY'tij-Y'ti>aD
Where:
k = number of treatments, t = 2 aeration systems.
m = number of measurements in each treatment block,
j = 4 alpha'—SOTE' values per bay.
n = number of blocks, i = 6 bays.
Y' = average of all measurements.
Y'i = average value for a bay, usually comprising 8
values, 4 from each aeration basin.
Y't = average value for a treatment, usually :
comprising 24 values, 4 values from each of 6
bays.
Y'ti = average value for each bay within a treatment,
usually comprising 4 values.
Y'tij = individual alpha'—SOTE values, usually 48 local
measurements.
The analysis of variance results for the nine off-gas tests
are presented in Tables CIO through CIS. The results are
summarized as follows:
Significance of Differences @ O.05 Level
Test Date
5-13-86
5-15-86
7-02-86
7-3O-86
1O-3O-86
12-O3-86
6-18-87
8-05-87
10-28-87
Between
Aeration
Systems
No
No
No
No
No
No
No
Yes
Yes
Among
Bays
Yes
No
Yes
Yes
Yes
No
Yes
No
No
Interactions
No
No ;
No
No
No
No
No
No
No
75
-------
The analysis of variance indicates that,, except for the last
two tests, the differences in the average alpha'-SOTE' values
for the two aeration systems were not significant although
many of the differences were more than 10 percent of the two
basin mean value. Collecting alpha*-SOTE data at more
locations could have increased the sensitivity of the
analysis of variance test by increasing the number of degrees
of freedom for the residual sum of squares. This is a moot
point, however, because it was not practical to sample more
than 48 positions in one day.
To test the adequacy of the 24 position per basin sampling
plan, the 24 positions were divided into two data sets of 12
measurements consisting of alternate north and south grids.
For example, the two sets for contact basin 2 are:
Set No. 1 Set No. 2
C2.ISA , . C2. 1NA . . ;.
C2.1SB C2.1NB
C2.2NA C2.2SA
C2.2NB C2.2SB
C2.3SA C2.3NA
C2.3SB C2.3NB
C2.4NA C2.4SA
C2.4NB C2.4SB
C2.5SA C2.5NA
C2.5SB C2.5NB
C2.6NA C2.6SA
C2.6NB C2.6SB
Similar sets were made from the contact basin 4 data. Once
again, alpha'-SQTE' data were used. The December 1986 test
data were not used.
Two estimates of the basin average alpha'-SOTE' were
calculated for each of the contact basins from the data sets.
The two estimates were then evaluated using a t test on the
data pairs.* The data and the results of the t test are
presented in Table C19. The t test indicates that the two
estimates of the basin average alpha'-SOTE' were not
significantly different at the O.O5 significance level. Since
both of the 12-position estimates provide statistically
equivalent results,, it can be concluded that the 24-position
sampling plan would provide a more precise estimate. Also,
the 12-position test conducted in December should be
considered statistically sound.
76
-------
REAERATION BASINS
The alpha'-SOTE data for the reaeration basins were evaluated
using the t test. The data were not blocked because visual
inspection did not produce any indication of systematic
spatial variations. The results of the t tests are presented
in Table C20. The statistical tests indicate that differences
in the basin averages which are greater than 6 percent of the
two basin mean were significantly different at the O.O5
significance level.
77
-------
R S
I I
II
.5
' 2 ! ;r ss a? 53 o i
!§:§ llii
slsii* ISIl
'
B s sillis §si«
i ! |s|§
5 iI Si^g
li
§§§§!§ |=|§
i lilgjl §|§|
Si 5 —
sis
I
I
i
i Ililll iiii
-
* " " *"! *^ "
jo o o » o
i S
8 o
S 3
I 1
ji Illiil I
•» ooeo.o « o o »• i o rf o o
-------
• grlr
— o >- - i.
" srlr
•** ^ >T - >"
i
§ is
VCu !
e» _ _ S
gggi
ii'is
i ii
§
S*-* o> e*»
»•— n va i
§ " £ !
S Sgg ;
gsg
3
£: ~™ _
S ! S S SC S
S ! 2 ^ = = S
o i e> o o e> 9
§S;
S S S 2 2 : 15
s sisg:g
i a - s
§ i s s
2* ir» •» *si
S S 5? S S
s §111
I I!
i ls§§ E ipiji Ililjl isil
i i i •»
=- £ £ ^
s islili i2sSi? sis§
,• -i".^"1™ ""^"T*^!"^ *^^"^*—
» ^^^O»^ ^ ^ ^ Oi I ^ ^O^O
Jg » g « O
-: £ s s g
s IS1S-S SgSSil iSS
-I l^'^r*:(*:!^: ^ "^ ^ » ! "^ -^-^trt
— *r O «M i *•» iy»
|glg:s 2
o «=> S o i S »
O O O O I O "0
s ^
*.
B S 5 i " ^
o o o ! o o
g ^* g o> «sr j fft r^.
g g§i§;i g
O O O O O i o O
a nsss:« s
IS — S — 2 ! *n ™
« --I 13 -J ^ ! ^ ^
i iiiiii i
! fi C
ssss S5I5
79
-------
I i
i
2 S
I I
g »
s- j.
o o i o o1 2 o
LI I B5il j 3 | = Ss
'• • ioooo'o oo22
3 j g S I § i i g g g §
! — S3 ? S ! s S"S°
'oc>oo'o =; ^ I o
s ssiiis lill
0 0000 to 0000
= lilsil llliil lilgi
o oooo. o- o o o ! o- o- .- * o- !
liisjl Illlii lili
-
"^ — — ff
Ss
°
ssssri i
80
-------
R K
g *
tn to
§ I
- S E3g 5 ! 2 »s<"-
=?:sggg!g sill
w »• » o o A A ! A ^-i_?_
; 2 j
: o !
.3
s
— -»«s;sgj2
1 i ~ s. : «; o ,3 5 i o-
•S!5u.!2g:t
§ I » 5 S - 'S
fc»l— l™
10000
s s s s c
s =s r s
S *"" S2 *"" :
I ill
I iiii
<» ooooio ciooe
I i i §11i liil
^ O C> A A I n .^ ^ _• _?
SSlI ! I
S §IiU
,; MS n r, K
S- i
'SS ', 2 ?
iss : s i
= =qs§ i §
<» ^ e> •=> i o
.g g s s s : $
^ •» » s •? ! T
1 §
«» eooeie cj c.' o el ! c! «; '.
I !l!!li I J
5 MM 1 1 §1*5 1 8
_o__ i - ¥ o- «,- «i . «.
81
-------
Ip f»
i s
o o
J
s
!!!!!!
!!!i!
mil! mi
I — t S — o a> r» i „
* i ililii
III
S 5
5 —
S o.
mils lii
-
•s ;
SSS;
82
-------
i
— * i ™ *? *
I §
5
s
II
is
c*
i
s i 2K
•3 ! .2-S
I 8 • = S
s s sj s:
g IS IS!I I§ i §
=
£ |
i! ssi« 11 ix s
- !
II Ii II
00 10 00
II I I II
i
is
53 S : s 2 :
SS:i s:
oo
i!
HIS IS j I is IB ss
oo !o So !~ ~ . ! ~ — <••
»**^i«»^^ to ACSIA.;^
§
5 S SS
; & x
83
-------
i
'm
I
S
•5
I
3
•S.S
s
s
s §g;
s ts — s
2 55 2 2
-!
2 Si 2
S = ! R Z Z » s
s§ i s sssi
1 5 ! K2 S3 C tr> **»
I Z * *^ ** ** •*• *^
i liil ! - §i«S
22 : a
SS!~
ilil i I liil
— 82S£i:2 KKSS
s silill I nil
j i iiil 1
i liil II ilslii issgis ~§
«> ^....^jo ^0-^^:5 5252:5 2
~ ~- 5 -- -
j. g r
I |§£| i s 5SSI i I =sssig §
»- SlsSiS 2
ss ; s
*— r-» i »-»
IIJI lillis B
l
" S S S S !
*7 ~; —:
-------
I
i s
o o
5
•«•
s
Is
. :
s
I
it
ft
<• UJ
s-g
* t
co a
c> CT
•i* "*"*
s ?
2-
J5 j*;
J2~ S
*C l «
o w
iiii
o O o o
— - -- o _
O O 0 O
iiii
SO O O
•-o jyj — ^
iiii
OO '»» <*l —
r-~ r*« — ^
« 12 *^ ***
0000
iiii
o o o o
ills
— •=••=> -^
1 O
II
i
! —
-
M
O
a
§
o
I
i
o
?
i
o
i
**
5 S S I
0 0 S S
0000
— o 2 «
o o o o
o — ^ •*•
0000
0000
—
asss
0000
ggs s
0 0* 0 0
t CO
i ^
1 O
1
! =.
i O
Ol
g
0
a
I
o
I
1
p-
o
«
0
c
o
o o o o
iBsi
o o o o
Ml CO — 03
S S SK
O O O O
i iss
iiii
O O O O
£j — "^ ^
R S3 5 «
O O O O
Ss|g
O O O O
0000
*O 0 hO f>
0 O O O
II
i
: 1
r o
CO
i
o
g
c
s
o
§
s
0
en
i
<£
8
•*
iiii
0 O 0 0
§SS 2
0000
o 5; S o
— i O O •-
0000
0000
^ 0 O O
Iiii
0 *> 0 0
ill!
Iiii
0000
§§ s §
O O O O
i
o
o>
i
i
o
>-
i ^
f O
S3
o
i
i
g
x
»•
rf
6
0
iiii
0 O 0 0
§ § s §
o o o o
Hi §
o o o o
E? m Irt Ǥ
o o <=> o-
sse s
§ s s s
ISIg
223 <;
gggg
0000
ggfg
0000
1 •
I
! i
I O
j 3
I "7
1 O
•5
t 0
i ?
t O
i
: e
: s
: o
2-
i I
i 2
: s
! **
lill
0 0 0 O
gs ^ s
- s s i
o o o o
1111
O O O O
IIII
O O O 0
iiii
0000
ssgss
o 2 22
iSii
o o o o
c» o o ^
O O 0 O
:= ! S S
? S
11
85
-------
5
£
e
"M
s
s
i I!
?_§
if
°. <=: ° •=: : ° s s s s : s Soggig
ooooio oo'o'ole; o-oS2s2
: 5 2 S !
: I §g !
i.
Missis SsSaiS Rgssis sss
— <5 o — i «— «— -i S -- i — S22""**"*1"^ r?? <> O
« ^ o
s Bslejl §!§§
I Iliils
•= .o ooo.e ooo'oic;
i 1 1
i isslll §11111
o oodoio" «5eJo-o'!^
i §
s assess E
s: zssa i S 5
IliiiS §5==i§ §ss5!§ 5
oo-55 io 225SiS 5555 I 5 5
86
-------
Table CIO - ANALYSIS OF VARIANCE TABLE - Test Conducted on 13 Nay 1386
Source of Sui of Degrees of Mean Square Ratio of
Variation Squares (SS> Freedoi (D.F.) (SS/D.F.) Mean Squares F(0.05,D.F.n,D.F.d)
Blocks SB = 0.0184 n-1 = 5 S'B = 0.0037 S'B/S'e = 2.9066 F(0.05,5,36) = 2.48
(Bays)
Treatients ST = 0.0034 k-1 = 1 S»T = 0.0034 S'T/S'e = 2.6583 F(0.05,l,36) = 4.11
(Equipment)
Interaction SI = 0.0024 (n-D* S11 = 0.0005 S'l/S'e = 0.3862 F(0.05,5,36) = 2.48
(k-1) =5 :
Error Se = 0.0455 nk(r-l) =36 S'e = 0.0013
Total S = 0.0697 N - 1 = 47 ; ;
NOTE:
k = Nuiber of treatments, i.e., tvo types of fine pore diffusers
• = Nuiber of replicates, i.e., four saipling positions per bay
n = Nuiber of blocks, i.e., six bays
N = Total nuiber of leasureients
• If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
87
-------
Table Cll - ANALYSIS OF VARIANCE TABLE - Test Conducted on 15 Hay 1986
Source of
Variation
Blocks
(Bays)
Treatments
(Equiptent)
Interaction
Error
Sui of
Squares
SB = 0
ST = 0
SI = 0
Se = 0
(SS)
.0147
.0015
.0032
.0446
Degrees of
Freedoi (D.F.)
n-1 = 5
k-1 = 1
(n-l)t
(k-1) = 5
nkd-1) = 36
Mean Square
(SS/D.F.)
S'B = 0.0029
S'T = 0.0015
S'l * 0.0006
S'e = 0.0012
Ratio of !
Hean Squares F(0.05,D.F.n,D.F.d) ;
S'B/S'e = 2.3797 F(0.05,5,36) = 2.48
S'T/S'e = 1.2489 i F(0.05,l,36) = 4.11 ,
S'l/S'e = 0.5122 ' F(0.05,5,36) = 2.48 ;
.
Total S = 0.0640 N - 1 = 47
MOTE:
It = Nuiber of treatients, i.e., tvo types of fine pore diffusers
• = Nuiber of replicates, i.e., four saipling positions per bay
n = Nuiber of blocks, i.e., six bays ;
IN = Total nuiber of leasureients
If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
-------
Table C12 - ANALYSIS OF VARIANCE TABLE - Test Conducted on 2 July 1986
Source of Sui of Degrees of Mean Square Ratio of
Variation Squares (SS) Freedot (D.F.) (SS/D.F.) Hean Squares F(0.05,D.F.n,D,F.d>
Blocks SB = 0.0962 n-1 = 5 S'B = 0.0192 S'B/S'e = 3.5394 F(0.05,5,36) = 2.48'
(Bays) i
!
Treatments ST = 0.0019 lc-1 = 1 S'T = 0.0019 S'T/S'e = 0.3585 ' FCO.05,1,36) = 4.11:
(Equipment) ;
Interaction SI = 0.0088 (n-D* S'I = 0.0018 S'l/S'e = 0.3250 F(0.05,5,36) = 2.48
(k-1) =5
Error Se = 0.1957 nkd-1) = 36 S'e = 0.0054
Total S = 0.3027 N - 1 = 47
NOTE: i
k = Nuiber of treatients, i.e., t«o types of fine pore diffusers '
• = Nuiber of replicates, i.e., four saipling positions per bay
in = Nuiber of blocks, i.e., six bays >
N = Total nuiber of teasureients
If the Ratio of Hean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
89
-------
Table C13 - ANALYSIS OF VARIANCE TABLE - Test Conducted on 30 July 1986
Source of Sui of Degrees of Mean Square Ratio of
Variation Squares (SS) Freedoi (D.F.) (SS/D.F.) Hean Squares F(0.05,D.F.n,D.F.d):
Blocks SB = 0.0286 n-1 = 5 S'B = 0.0057 S'B/S'e = 2.7473 F(0.05,5,36) = 2.48
(Bays)
Treatments ST = 0.0074 k-1 = 1 S'T = 0.0074 S'T/S'e = 3.5291 F(0.05,l,36) = 4.11
(Equipment).
Interaction SI = 0.0034 (n-D* S'I = 0.0007 S'l/S'e = 0.3281 F(0.05,5,36) = 2.48;
(k-1)'=5 ;
Error Se = 0.0750 nkd-1) = 36 S'e = 0.0021
Total S = .0.1144 N - 1 = 47
MOTE: ;
k = Nuiber of treatients, i.e., two types of fine pore diffusers :
• = Nuiber of replicates, i.e., four saipling positions per bay
n - Nuiber of blocks, i.e., six bays :
i - Total nuiber of teasureients
If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and i
the leans are considered significantly different at the 0.05 significance level.
90
-------
Table C14 - ANALYSIS OF VARIANCE TABLE - Test Conducted on 30 October 1986
Source of
Variation
Blocks
(Bays)
Treat lent s
(Equipient)
Interaction
Error
Total
iiui of
Squares (SS)
!>B = 0.0181
!5T = 0.0031
i •
SI = 0.0098
Se = 0.0480
S = 0.0791
NOTE:
Degrees of
Freedoi (D.F.)
n-1 = 5
k-1 = 1
(n-D*
(k-1) = 5
nkd-1) = 36
N - 1 = 47
Mean Square
(SS/D.F.)
S'B = 0.0036
S'T = 0.0031
S'l = 0.0020
S'e = 0.0013
Ratio of
Hean Squares F(0.05,D.F.n,D.F.d)
S'B/S'e = 2.7219 F(0.05,5,36) = 2.48
S'T/S'e = 2.3542 F(0.05,l,36) = 4.11
S'I/S'e= 1.4744 F(0.05,5,36) = 2.48
k = Nuiber of treatments, i.e., tvo types of fine pore diffusers
• - Nuiber of replicates, i.e., four saipling positions per bay
n = Nutber of blocks, i.e., six bays
N = Total nuiber of leasureients
If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
91
-------
Table CIS - ANALYSIS OF VARIANCE TABLE - Test Conducted on 3 Deceiber 1986 ;
Source of Sui of Degrees of Mean Square Ratio of !
Variation Squares (SS) Freedoi (D.F.) (SS/D.F.) Mean Squares F(0.05,D.F.n,D.F.d)
Blocks SB = 0.0213 n-1 = 5 S'B = 0.0043 S'B/S'e = 0.8926 F(0.05,5,12) = 2.48
(Bays) '
Treatients ST = 0.0045 k-1 = 1 S'T = 0.0045 S'T/S'e = 0.9388 F(0.05,l,12) = 4.1l!
(Equipment)
Interaction SI = 0.0047 (n-l)t S'I = 0.0009 S'l/S'e = 0.1975 ; F(0.05,5,12) = 2.48
Error Se = 0.0572 nkd-1) = 12 S'e = 0.0048
Total S = 0.0877 N - 1 = 47 • ;
P i
NOTE:
k = Nuiber of treatments, i.e., tvo types of fine pore diffusers
m = Nuiber of replicates, i.e., four saipling positions per bay
n = Number of blocks, i.e., six bays
N - Total number of measureients
If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of,equal variances is rejected and
the means are considered significantly different at the 0.05 significance level.
92
-------
Table CIS'- ANALYSIS OF VARIANCE TABLE - Test Conducted on 18 June 1987
Source of Sui of Degrees of Mean Square Ratio of I
Variation Squares (SS) Freedoi (D.F.) (SS/D.F.) Mean Squares F(0.05,D.F.n,D.F.d)
Blocks SB = 0.0255 n-1 = 5 S'B = 0.0051 S'B/S'e = 4.4108 F(0.05,5,36) = 2.48
(Bays)
Treatments ST = 0.0002 k-i = 1 S'T = 0.0002 S'T/S'e = 0.1757 F(0.05,l,36) = 4.11
(Equipient) •
Interaction SI = 0.0005 (n-D* S'I = 0.0001 S'l/S'e = 0.0893 F(0.05,5,36) = 2.48
(k-1) = 5
Error Se = 0.0417 nkd-1) = 36 S'e = 0.0012
Total S = 0.0679 N - 1 = 47 ; ;
NOTE:
k = Nuiber of treatients, i.e., two types of fine pore diffusers •
• = Nuiber of replicates, i.e., four sampling positions per bay
n = Nuiber of blocks, i.e., six bays
N = Total nuiber of leasureients
If the Ratio of Mean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
93
-------
Table C17 - ANALYSIS OF VARIANCE TABLE - Test Conducted on 5 August 1987
Source of Sui of Degrees of Mean Square Ratio of
!!-™!" SqUareS (SS> Freed°' (D ST= °'1238 k"1 = I S S'T = 0.1298 S'T/S'e = 21.3557 FCO.05,1,36) M.ll
Interaction SI = 0.0166 (n-l)*_ S' I = 0.0033 S'l/S'e = 0.5471 F(0.05,5,36) = 2.48
vK~U ~ 0
Error Se = 0.2189 nk(t-l) = 36 S'e = 0.0061 ,
Total S = 0.4128 N - 1 = 47
HOTE:
I: = Nuiber of treatments, i.e., two types of fine pore diffusers '.
n = Nmber of replicates, i.e., four saipling positions per bay
n = Nuiber of blocks, i.e., six bays • ' ;
M = Total number of leasureients ,'
If the Ratio of Hean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
94
-------
Table CIS - ANALYSIS OF VARIANCE TABLE - Test Conducted on 28 October 1987
Source of Sui of Degrees of
Variation Squares (SS) Freedom (D.F.)
Blocks SB = 0.0273 n-1 =5
(Bays)
f
Treatients ST = 0.0410 k-1 = 1
(Equipient)
Interaction SI = 0.0056 (n-D*
(k-1) = 5
Error
Se = 0.1049 nk(i-l) = 36
Mean Square
(SS/D.F.)
S'B = 0.0055
S'T = 0.0410
S'l = 0.0011
S'e = 0.0029
Ratio of
Mean Squares
S'B/S'e = 1.8711
S'T/S'e = 14.0616
S'l/S'e = 0.3872
F(0.05,D.F.n,D.F.d)
F(0.05,5,36) = 2.48
F(0.05,l,36) = 4.11
F(0.05,5,36) = 2.48
Total S = 0.1787 H - 1 = 47
NOTE:
k = Ninber of treatients, i.e., two types of fine pore diffusers
• = Nuiber of replicates, i.e., four sampling positions per bay
n = Nuiber of blocks, i.e., six bays
N = Total nuiber of teasureients
If the Ratio of dean Squares is > F(0.05,D.F.n,D.F.d), the hypothesis of equal variances is rejected and
the leans are considered significantly different at the 0.05 significance level.
95
-------
Table C19 - Analysis of Two 12-Position Saipling Plans
Contact Basin 2 - Ceraiics Contact Basin 4 - Membranes
Alpha' -SOTE'
Est. 1 Est. 2
5-13-86
5-15-86
7-02-86
7-30-86
10-30-86
12-03-86
6-18-87
8-05-87
10-28-87
0.1548
0.1552
0.1811
0.0977
0.1254
Insufficient
0.1098
0.2160
0.1807
0.1375
0.1425
0.1813
0.0936
0.1206
data
0.1182
0.2310
0.1622
Diff.
2 - 1
-0.0173
-0.0127
0.0002
-0.0041
-0.0048
0.0084
0.0150
-0.0185
Alpha'
Est. 1
0.1655
0.1584
0.1689
0.1189
0.1492
0.1075
0.1214
0.1083
-SOTE'
Est. 2
0.1609
0.1620
0.1680
0.1221
0.1291
0.1122
0.1175 .
0.1178
Diff.
2 - 1
-0.0046
0.0036
-0.0009
0.0032
-0.0201
0.0047
-0.0039
0.0095
Note: Data froi Tables Cl through C9.
Est, 1 comprises data froi positions ISA, 1SB, 2NA, 2NB, 3SA, 3SB, 4NA, 4KB, 5SA, 5SA, 6NA, 6NB.
Est. 2 cotprises data fro§ positions 1NA, 1NB, 2SA, 2S8, SNA, 3NB, 4SA, 4SB, SNA, 5NB, 6SA, 6S8.
t Test (reference 9)
Mean difference of all 16 pairs of data, d' = - 0.0026
Standard deviation of the differences, s'(d) = 0.0103
t = !d'!/£s'(d)/(n)«0.5] = 0.0037/10.01/4] = 0.064
t(0.05,15) = 2.15
Since t < t(0.05,15), accept null hypothesis that the two leans are equal.
96
-------
Table C20 - Statistical Analysis of Reaeration Basin Alpha'-SOTE Data (t test)
fiuad.
5-12-86 82
84
Oiff.
5-16-86 82
84
Oiff.
7-01-B6 62
84
Diff.
7-29-86 82
84
Diff.
10-29-86 82
84
Oiff.
12-02-86 62
84
Diff.
6-17-87 82
84
Diff.
8-04-87 62
84
Diff.
10-27-87 82
84
Alpha'
-SOTE
0.1822
0.1757
0.0065
0.1696
0.1697
-0.0001
0.2118
0.1838
0.0280
0.1426
0.1122
0.0304
0.1155
0.1336
-0.0181
0.1955
0.1310
0.0645
0.1178
0.1108
0.0070
0.2321
0.1365
0.0956
0.1953
0.1268
S(I>
0.0219
0.0204
0.0138
0.0168
0.0173
0.0290
0.0164
0.0271
0.0092
0.0157
0.0194
0.0113
0.0127
0.0075
0.0406
0.0150
0.0212
0.0216
U0.05,
D.F. s'd) D.F.p t D.F.p) Result
11 0.0212 22 7.975 2.074 SO r
H . !
11 0.0154 22 0.016 2.074 USD •
11 ! ' !
11 0.0239 22 2.872 2.074 SD
11
11 0.0224 22 3.325 2.074 SD !
H ;
11 0.0129 22 3.446 2.074 SD ;
11
9 0.0159 18 9.071 2.101 SD
9 ;
11 0.0104 22 1.644 2.074 USD
11 ; i
!
11 0.0306 -22 7.651 '• 2.074 SD •
11
11 0.0153 22 10.986 2.074 SD
11
Oiff. 0.0685 ,
s(x) = standard deviation of the teasureients
D.F. = degrees of freedoi for each set of basin •easureients
s'(x) = pooled estiiate of the standard deviation of all leasuretents
D.F.p = degrees of freedoi for the pooled estitate of the standard deviation
t = !x'i-x'2!/[s'(x)(l/nl+l/n2)«0.5J
t(0.05,Ofp) = value of t for probability of 0.05 and D.F.p degrees of freedoi
Result = if t > t(0.05,DFp) then the leans are significantly different (SD)
= if t < t(0.05,DFp) then the leans are not significantly different (NSD)
97
-------
APPENDIX C
OPERATIONAL COST DATA SUMMARIES
98
-------
I
.*>
«•
1
•to*
•S
S
I
s
I
i
I
E S =
*•» •+* •> ** — S •» -"
f? m <^* M di M »^ •
fc- 5-, *» o o« •• ++
r«^s5^s "°
Sillllll*
W M««ai-«fc>^^»^»^>S
» SJS^M^S^^w
99
-------
w> r— s* r—
^ <•* "•• —««-« .-I
r— r—
i
to
«
s
« £
«« I -S
AJ ka
a 5 a *4 -o
—< -<5 n o v
•> 4» O
g«g|SS
15-ss !.•
kM M-» •—I S ^^ *S
ww2S3I
«S££££
100
-------
«*•» se t^> «d-i o
-
s
e * e M -o
— « J5 K> o ni
n M M tc 0
CO M O O . -to
4^ *» J-» CJ «» «»
0 *O *^ *>
O *» *) fcJ *O
101
------- |