EPA-670/2-74-077
September 1974
Environmental Protection Technology Series
REVERSE OSMOSIS OF
TREATED AND UNTREATED
SECONDARY SEWAGE EFFLUENT
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-7H-077
September 1974
REVERSE OSMOSIS OF TREATED AND
UNTREATED SECONDARY SEWAGE EFFLUENT
By
Doyle F. Boen
Gerald L. Johannsen
Eastern Municipal Water District
Hemet, California 923^3
Grant No. WPRD ^-01-6?
Project YfOkO DSR
Program Element 1BBC&3
Project Officer
Gerald Stern
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Cincinnati, Ohio l|-5268
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 1*5268
-------
REVIEW NOTICE
The National Environmental Research Center—Cincinnati has
reviewed this report and approved its publication. Approval
does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
11
-------
FOREWORD
Man and his environment must be protected from the adverse effects
of pesticides, radiation, noise and other forms of pollution, and
the unwise management of solid waste. Efforts to protect the
environment require a focus that recognizes the interplay between
the components of our physical environment—air, water, and land.
The National Environmental_Research Centers provide this
multidisciplinary focus through programs engaged in:
• studies on the effects of environmental
contaminants on man and the biosphere, and
•a search for ways to prevent contamination
and to recycle valuable resources.
Reverse osmosis is one of the primary processes by which the
environmental component of water may be protected from contamination,
This text is an attempt to define some of the fundamental abilities,
requirements and limitations of reverse osmosis when applied to
various qualities of secondary sewage effluent.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
111
-------
ABSTRACT
A pilot study was conducted to determine reverse osmosis feasibility
on untreated and treated secondary effluents. Six commercially
designed reverse osmosis pilot units, with 3,000 to 10,000 GPD nominal
capacities and different module concepts, were tested.
Post treatment of secondary effluent feeds, using alum clarification,
sand filtration, granular activated carbon treatment, chlorine additions
and pH adjustment, in different combinations improves reverse osmosis
performance and significantly extends useful membrane life.
Membrane fouling occurs despite post secondary effluent treatments.
Enzymatic detergent solutions were moderately effective as membrane
rejuvenation treatments. Inorganic fouling (particularly with phosphates)
could be removed with solutions of the sodium salt of ethylenediaminetetra-
acetic acid.
Of the module concepts tested, one of the tubular makes and the spiral
wound had the best overall performance.
Based on the pilot plant data, the total reverse osmosis costs, excluding
brine disposal is estimated to be $0-78/1,000 gallons for a 0.9 MGD product
water facility and about $0.73/1,000 gallons for a 9 MGD product water
facility.
This report was submitted in fulfillment of Project Grant Number 170^0
DSR by the Eastern Municipal Water "District, Hemet, California, under the
partial sponsorship of the Environmental Protection Agency. Work was
completed as of April, 197^-
iv
-------
CONTENTS
Page
Abstract iv
List of Figures vii
List of Tables ix
Acknowledgements xiii
Sections
I Conclusions 1
II Recommendations 3
III Introduction 5
IV Data Collection 9
V Computational Methods 15
VI Secondary Effluent 19
VII Post-Treatment of Secondary Effluent 2k
VIII Reverse Osmosis Operations - Preliminary Discussion h6
IX Aerojet-General Corporation Reverse Osmosis Unit,
Tubular Membrane Design 63
X American Standard (Abcor) Reverse Osmosis Unit,
Tubular Membrane Design 78
XI E. I. Du Pont De Nemours & Co. Reverse Osmosis Unit,
Hollow Fiber Concept 93
XII Gulf General Atomic Unit, Spirally Wound Module
Design 11^
XIII Raypak, Inc. Reverse Osmosis Unit, Modified Tubular
Design 131
XIV Universal Water Corporation Reverse Osmosis Unit
A Tubular Design 137
-------
Contents Cont'd
No. Page
XV Inter-Unit Comparisons l^O
XVI Reverse Osmosis Costs 175
XVII References 186
XVIII Appendices
vi
-------
FIGURES
Ho. Page
1 Aerial View of Facilities 20
2 E.M.W.D. Location Map 21
3 Post Secondary Effluent Treatment Facilities 25
k Reactor-clarifier 26
5 Sand Filters and Activated Carbon Filters 27
6 Generalized Reverse Osmosis Unit Flow Scheme Vf
7 Flow Schematic, 10,000 Gallon Per Day R.O. Unit 1*8
8 Microphotograph of Slime Removed from Universal R.O. Unit 53
9 Slime Microphotograph 53
10 Flow Pattern "a" 56
11 Aerojet-General Reverse Osmosis Unit £l±
12 A vs. Time Plotted - Aerojet-General 75
13 A vs. Time Plotted - Aerojet-General 76
14 A vs. Time Plotted - Aerojet General 77
15 American Standard (Abcor) Reverse Osmosis Unit 79
16 Tubular Components, American Standard (Abcor) R.O. Unit 79
17 A vs. Time Plotted - American Standard 91
18 A vs. Time Plotted - American Standard 92
19 Du Pont Installation with B-5 Permeators in Place 93
m
20 IV Permasep Pilot Plant Flow Diagram 95
21 Simplified Internal Flow Scheme, B-5 Module 97
o
22 Cut Away Drawing of Permasep^- Permeator 97
23 A vs. Time Plotted - Du Pont 110
VI1
-------
Figures Cont'd
No.
2k A vs. Time Plotted - Du Pont HI
25 A vs. Time Plotted - Du Pont 112
26 Gulf General Atomic, Inc. Reverse Osmosis Unit 115
27 Details of Spiral Wound Module 117
28 A vs. Time Plotted - Gulf 129
29 A vs. Time Plotted - Gulf 130
30 A vs. Time Plotted - Raypak 136
31 Universal Water Corp. Reverse Osmosis Unit 138
32 Universal Reverse Osmosis Unit in Part 138
33 A vs. Time Plotted - Universal 155
3k A vs. Time Plotted - Universal 156
35 A vs. Time Plotted - Universal 157
36 A vs. Time Plotted - Universal 158
viii
-------
TABLES
No. Page
1 Chronological Orientation of Experimental Program 8
2 Constituent Analysis Methods 12
3 Analytical Results of Seasonal Variation in Secondary
Effluent 22
k Domestic Water Quality to Sewered Areas of EMWD
Compared to Secondary Effluent 23
5 Activated Carbon Physical Properties and Specifications 28
6 Post-Secondary Treatment Backflush Record 30
7-12 Post Secondary Treatment Unit Constituent Analyses 33-^-
13 Post Treatment Data Weeks 16-33 k2
Ik Daily Computer Printout 50
15 Record of R.O. Unit Feed pH 52
l6 Aerojet-General Reynolds Numbers 57
17 Estimated Minimum Reynolds Numbers per Modular Section 58
18 Program Output, "Water Permeability Studies" 59
19 Program Output, "Average Rejection Ratios" 60
20 Average Rejection Ratios Summary Output 62
21 Reverse Osmosis Process Information, Aerojet-General 66
22 Out-of-Service Record, Aerojet-General 67
23 Membrane Failure Record, Aerojet-General 68
2k Water Permeability Data, Aerojet-General 69
25 Aerojet-General Water Recovery Data 70
26 pH Adjusted Feed Water Quality, Aerojet-General 71
27 Product Water Quality, Aerojet-General 71
28 Brine Quality, Aerojet-General j2
ix
-------
Tables Cont'd
No.
29 Water Recovery and Total Rejection Ratios,
Aerojet-General f2
30 Average Rejection and Material Balance Ratios,
Aerojet-General 73
31 Out-of-Service Record, American Standard "2
32 Reverse Osmosis Process Information, American Standard 83
33 Membrane Failure Tabulation, American Standard 84
3^ Water Permeability Data, American Standard 85
35 Water Recovery Data Confidence Level, American Standard 86
36 pH Adjusted Feed Water Quality, American Standard 87
37 Product Water Quality, American Standard 8?
38 Brine Quality, American Standard 88
39 Water Recovery and Total Rejection Ratios,
American Standard 88
lj-0 Membrane Cleansing History and Product Flux Increases,
American Standard 89
kL Average Rejection and Material Balance Ratios,
American Standard 90
(S)
ii-2 Equipment Description, Du Pont Permasep^- Package 96
i»-3 Estimated Membrane Surface Area, Du Pont B-9's 98
kk Reverse Osmosis Process Information, Du Pont 100
lj-5 Out-of-Service Record, Du Pont 101
k6 Water Permeability Data, Du Pont 102
hj pH Adjusted Feed Water Quality, Du Pont 103
lj-8 Product Water Quality, Du Pont 103
k<$ Brine Quality, Du Pont 104
50 Water Recovery and Total Rejection Ratios, Du Pont iol|.
x
-------
Tables Cont'd
No.
51 Du Pont Water Recovery Data 105
52 Constituent Data, Courtesy of Du Pont 105
53 Average Rejection and Material Balance Ratios, Du Pont 107
5^ Membrane Rejuvenation Record, Du Pont B-5's 108
55 Membrane Rejuvenation Record, Du Pont B-9's 108
56 Membrane Rejuvenation Record, Du Pont B-9's 108
57 Du Pont B-9 Permeability Analyses, Colorado River
Water Test Run 113
58 Out-of-Service Record, Gulf 118
59 Reverse Osmosis Process Information, Gulf 119
60 Gulf Water Recovery Data 120
6l Water Permeability Data, Gulf 121
62 pH Adjusted Feed Water Quality, Gulf 122
63 Product Water Recovery, Gulf 122
6U Brine Quality, Gulf 123
65 Water Recovery and Total Rejection Ratios, Gulf 123
66 Average Rejection and Material Balance Ratios, Gulf 125
67 Gulf Record of Membrane Rejuvenation 126
68 Reverse Osmosis Process Information, Raypak 132
69 Water Permeability Data, Raypak 132
70 pH Adjusted Feed Water Quality, Raypak 1$±
71 Product Water Quality, Raypak 13l|-
72 Brine Quality, Raypak 1&
73 Water Recovery and Total Rejection Ratios, Raypak 135
xi
-------
Tables Cont'd
7^ Average Rejection and Material Balance Ratios, Raypak 135
75 Reverse Osmosis Membrane Performance Comparison,
Universal vs Hemet Tests ^
76 Out- of -Service Record, Universal
77 Study of Life Spans, Membrane Sets, Universal
78 Process Information, Universal
79 Water Permeability Data, Universal li(-5
80 pH Adjusted Feed Water Quality, Universal
8l Product Water Quality, Universal
82 Brine Quality, Universal
83 Recovery and Total Rejection Ratios, Universal
Qk Universal Water Recovery Data 150
85 Average Rejection and Material Balance Ratios, Universal 152
86 Membrane Rejuvenation Record, Universal 153
87 Maintenance and Allotment/100 Available Operating Hours l6l
88 - 92 Operational Variables and Values 163-167
93 Comparison of Membrane Configurations 168
9^ Average Per Cent Reductions of Constituents 170
95 Minimum Volume Increase Requirements per Unit to Meet
Specific Demand - 10,000 gpd @ 90$ Rejection 171
96 Estimate of Membrane Life Based on Product Water
Recovery Loss 173
xii
-------
ACKNOWLEDGEMENTS
The sixteen-month program reported herein was performed under the
joint auspices of the U. S. Environmental Protection Agency and
the Eastern Municipal Water District, Hemet, California, (Messrs.
Doyle F. Boen, General Manager and Chief Engineer; Claire A. Gillette,
Operations Engineer). Mr. Gerald Stern, U.S.E.P.A., was the Project
Officer and Mr. Frank R. Bridgeford (Consultant) was the Project
Engineer. The following participated in the test work:
Mr. Dean C. Rauscher - Reverse Osmosis Area
Mr. Richard K. Morton - Post-Secondary Treatment Area
Mr. Carson R. O'Dell - Chemist
Mr. Stephen A. Hays - Chemist (Author, Section A-6)
The various reverse osmosis manufacturers and representatives were
most helpful in giving both guidance and advice during the progress
of the work and in reviewing a portion of this report. It is
impossible to mention all those who have contributed to the success
of this project. Among those who deserve special mention are:
Mr. A.C. F. (Tom) Ammerlaan - Abcor, Inc.
Mr. Warren H. Bossert - Aerojet-General Corp.
Mr. Serop Manjikian - Universal Water Corp.
Mr. Victor Tomsic - E. I. Du Pont de Nemours & Co.
Mr. Randolph L. Truby - Gulf Environmental Systems Co.
Mr. Frank R. Shippey - Raypak, Inc.
xiii
-------
SECTION I
CONCLUSIONS
1. The use of the reverse osmosis process in the Hemet, California,
groundwater recycling program on treated or untreated secondary
effluent feed is relatively expensive ip comparison with importing
Northern California low total dissolved solids water.
2. Effective post secondary treatment sequences (for the feed to
the reverse osmosis process) are:
(a) alum clarification followed by sand filtration,
(b) sand filtration followed by granular activated
carbon treatment,
(c) though not directly tested, alum clarification
followed by granular activated carbon treatment
would also be an effective post secondary
effluent treatment,
(d) sand filtration alone was not as effective for
p'ost treating secondary effluents when compared
to the prior listed post treatment sequences;
however, sand filtration cost is considerably
less and can be used as the post treatment on
better quality secondary effluents.
Each of the treatment sequences noted above is followed by
chlorine addition (0.5 to 1.0 mg/1 chlorine residual) and pH
adjustment (5.0 to 5«5)«
3. Reverse osmosis cost can be significantly reduced with post
secondary effluent treatment and will significantly extend
the useful membrane life.
4. The choice of post secondary effluent treatment depends on
the quality of the secondary effluent and the reverse osmosis
module concept. Closely packed (high density) reverse osmosis
membrane surfaces are more subject to solids fouling than the
open tube membrane configuration. Organics can be a significant
factor in membrane fouling.
5. Despite the use of post secondary effluent treatments, reverse
osmosis membrane fouling is a critical problem in reverse
osmosis wastewater treatment. The most effective membrane
rejuvenation treatments were:
(a) Enzymatic detergent solutions
(b) For inorganic fouling, particularly with phosphates,
a solution of the sodium salt of Ethylenediaminetetraacetic
acid.
-------
6. The use of higher Reynolds number flow conditions, without
excessive pressure drops, is an effective approach for
retarding membrane fouling.
7. Of the module designs tested, the spiral wound and one of the
tubular makes had the best overall performance.
8. The use of higher product water flux and lower solute rejection
("open") membranes is not appropriate even with the use of
post secondary effluent treatments. This membrane type is
subject to severe compaction, and internal membrane fouling
which rapidly negates its initial advantages.
9» Based on the pilot plant data, taken at Hemet, California,
the total reverse osmosis costs including post secondary
effluent treatments and membrane rejuvenation, but excluding
blending and brine disposal costs, are estimated at $0.78/1000
for 0.9 MGD product water and $0.73/1000 gallons for 9 MOD
product water. These costs are based on 90$ total dissolved
solids rejection and up to 90$ product water recovery.
10. Based on a secondary effluent average TDS of 7l6 mg/1,
approximately 3^$ reverse osmosis product water can be
mixed with 66$ secondary effluent water to produce a
blended water with 500 mg/1 TDS. The cost for the R.O.
product water portion is estimated at $0.25/1000 gallons
for the 9 MGD facility.
-------
SECTION II
RECOMMENDATIONS
The primary objective of this pilot plant study was to determine the
feasibility of treating secondary effluent by reverse osmosis so that
the final effluent might then be used advantageously as an integral
part of the Hemet-San Jacinto closed basin ground water recycling
program. Several types of small (under 10,000 gpd nominal capacity)
reverse osmosis units were tested. Total costs for treating secondary
effluent appear higher as compared to importing Northern California
water. However, new developments in reverse osmosis are occurring
so rapidly that some of the data and costs shown in this report may
now be obsolete. It is recommended that reverse osmosis demonstration
studies be conducted to determine the following:
(l) The cost tradeoff between the type of post secondary
effluent treatment needed versus membrane life and
membrane rejuvenations for 90 per cent product
water recovery and 90 per cent salt rejection. The
smaller reverse osmosis units used in this study showed
a maximum of 75 Per cent product water recovery on a
once through use basis. The cost estimates based on
this study data were projected to 90 per cent product
water recovery.
(2) Additional experiences for determining operation costs
are needed. For example, the manpower cost is about
20 per cent of the total reverse osmosis cost based
on this study. This manpower cost is much greater
than that suggested by the manufacturer for the
spiral wound unit. This difference could be the
result of interpreting manpower needs from a research
and development study as compared to more routine.
operations.
(3) Brine disposal costs were not included because only
small amounts of brine were generated. Also brine
disposal was not an integral part of this study.
However, brine disposal could be a very critical
cost factor and could govern the type of post
secondary effluent treatments employed, and the
membrane replacements and membrane rejuvenations
needed.
In a closed ground water basin area or for other effluent discharges,
reverse osmosis would be used to control the total dissolved solids
and other specific constituent concentrations in the final effluent.
From a water management viewpoint, this control may be better
accomplished by using the reverse osmosis process on the raw water (supply)
side for the following reasons:
-------
^
(l) Higher quality water would be available for domestic uses.
Domestic users would share in the treatment costs, but
also would benefit by not needing individual demineralizers
which usually are more expensive to use. The wastewater
would benefit because it would contain less total dissolved
solids and other specific constituents (i.e., sodium)
introduced by the regeneration of individual domestic units,
thus avoiding extra buildup of salts which would be
ultimately detrimental in a closed ground water basin or
violate effluent discharge standards.
(2) Pretreatment of the raw water feed could be less costly
as compared to secondary effluents prior to reverse
osmosis treatment. For instance, organic fouling would
be less likely by using a raw water feed.
Despite the apparent advantages cited above for treating raw water, it
still may be necessary to use the reverse osmosis process on secondary
effluent to avoid specific discharges and to meet the 1985 no pollutant
discharge goal in the Federal Water Pollution Control Act Amendments
of 1972. Therefore, side-by-side demonstration studies are recommended
to determine the best cost effective use for the reverse osmosis
process.
It is recommended that future reverse osmosis studies use a computer
approach for compiling and analyzing data.
-------
SECTION III
INTRODUCTION
Purpose of the Study
,The primary objective of this pilot plant study was to determine the
feasibility of treating secondary effluent by reverse osmosis so that
the final effluent might then be used advantageously as an integral
part of the Hemet-San Jacinto ground "water recycling program. Reverse
osmosis performance depends to a great extent on feed water quality.
Particulate, colloidal, and dissolved substances in secondary effluent
are known to have an adverse effect on reverse osmosis performance.
Therefore, post secondary treatment processes were selected and operated
under various combinations to determine their effectiveness for removing
constituents that interfere with reverse osmosis performance. By
conducting a study of post-secondary effluent processes, it was hoped
that the most effective and economical process combination could be
found.
Historical Record
By accepting responsibility for disposal of wastewater from the entire
Hemet-San Jacinto Valley (a semi-arid region with an average rainfall
of 12 inches and serving 30,000 people) Eastern Municipal Water
District was thrust into an area of prime ecological importance.
Almost simultaneous with the start-up of its water reclamation facility,
EMWD launched a research project solely to study the effectiveness of
its ground water recharge program. The saline condition of the
reclaimed wastewater prompted the District to investigate current
methods of salt removal. The need to eliminate salt from the recharge
water is accentuated by local geolbgical data indicating a closed
underground water reservoir. Other than surface flow, little water
is believed to escape the Valley. At the time this was fully realized,
reverse osmosis was still in its early developmental stage, though it
showed much promise in economical salt removal from saline waters.
For this reason it was chosen for study by EMWD.
Assuming that reverse osmosis (R.O.) was economical and reliable, the
role of R.O. in a wastewater recycling and ground recharge program
would be the following:
(a) Reduce the concentration of total dissolved solids and
refractories in the final sewage effluent;
(b) Curtail long-term concentration buildup of total dissolved
solids and refractories in the ground water reservoir;
(c) Aid in maintaining the concentration of total dissolved
solids and refractories in the ground water reservoir at
levels acceptable to regulatory agencies;
(d) Minimize the future possibility of having to construct
water treatment facilities (demineralization) that might
be required at numerous locations over the ground water
aquifer.
-------
6
In December, 1966, the Federal Water Pollution Control Administration,
nov integrated into the Environmental Protection Agency, approved the
proposal made by Eastern Municipal Water District of Hemet, California
to -undertake a Reverse Osmosis Demonstration Project under Research
and Development Grant WPRD 4-01-6?. The original title of the project
•was "Reverse Osmosis to Remove Dissolved Solids From Reclaimed Water
Used in Ground Water Recharge Program" and the assigned contract
number was WPRD 170^0 DSR.
As of July, 1967, however, little research had been completed in the
reverse osmosis field and its application to wastewater. A 15,000
GPD unit was operating at this time but had encountered numerous
problems. It was apparent from these difficulties, that there were
still some basic questions of Reverse Osmosis needing answers. Among
these questions were, "What types of units are best suited for the
various possible qualities of treated and untreated secondary effluent?"
and "What are the economic advantages of using post-secondary treatment?"
or even "To what degree can a fouled membrane/module be rejuvenated?"
To the planners, the Hemet facility seemed well-suited to investigate
these problems. Final plans and specifications received the approval
of the Project Officer on September 11, 1968 and construction of the
building and facilities was completed in December, 1969. Testing of
the reverse osmosis units, was initiated on March 6, 1970, and concluded
on June 25, 1971, a total operating period of 69 weeks.
Experimental Program
Originally, plans called for a study of one large reverse osmosis unit
which was to operate on a feed of sand filtered secondary effluent. On
the reasoning that more could be learned from a broader program, the
plans were changed to include a study of five smaller units, each
representing one of the major concepts of R.O. These concepts were: the
flat plate design, the hollow fiber module, the spiral wound module, the
high flux "loose" membrane tubular design, and the high rejection "tight"
tubular design, (in the best interests of the study, a tubular design
was substituted for the near-obsolete flat plate design.) In addition,
a tubular design with the cellulose acetate membrane on the outside of
•fee tube was also tested.
Under the revised objectives., post treatment of secondary effluent (feed)
was expanded to include the following group of processes to be used in
various combinations:
A. Reactor-clarification with alum and polymer coagulation
B. Pressure sand filtration
C. Granular activated carbon filtration
D. Diatomaceous earth filtration
E. Pre-R.O. unit chlorination (mandatory)
F. pH adjustment (mandatory)
-------
(in subsequent references, capital letters as used above will be
used to designate post-secondary treatment processes.)
For two major reasons it was also decided to increase the daily
capacity of post-secondary effluent treatment equipment from 50,000
to 150,000 gallons:
(l) To allow for underdesign (manufacturers' stated unit
capabilities were in terms of constituted salt solution,
not wastewater).
(2) To allow for larger capabilities, should a magnified
follow-on study take place, based on the assumption that
the percent cost for a larger facility would be substantially
below the per cent capacity increase.
Operating under the various controlled conditions, evaluation criteria
were established based on total dissolved solids reduction observed
for each unit. The scope of the evaluation criteria was eventually
magnified to analyze each type of liquid flow for the maximum number
of important constituents within the limitations of a two-man
laboratory crew. The execution of these plans generated such a mass
of data that it was necessary to sort and analyze the available
information on an IBM 1130 Computer. This requirement was not
anticipated in the planning of the original scope of the work.
Table 1 demonstrates how the above processes were integrated into the
experimental program (capital letters designate the post-secondary
treatment sequence).
-------
8
Table 1. CHRONOLOGICAL ORIENTATION OF EXPERIMENTAL PROGRAM
Period Covered
Week No.
2-7
5-7
1-7
1-7
7-17
14-28
7-24
7-24
7-30
28-33
24-33
24-33
31-33
35-36
34-41
34-38
38-41
33-M
33-"H
41-47
41-49
M-146
41-48
49-53
45-57
48-57
61-64
57-61*
57-66
62-64
57-64
64-69
64-66
66-69
64-68
64-69
I Post-Treatment Sequence
(A^leaetor-Glarifier )
(B=Sand Filters )
C=Carbon Filters )
D=B.E. Filters )
E=Pre-R.O. Unit Chlorination j
F=pH Control )
A,B,C,E,F,
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
A,B,C,D,E,F
A,B,C,D,E,F
A,B,C,D,E,F
A,B,C,D,E,F
A,B,C,D,E,F
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
A,B,E,F
A,B,E,F
A,B,E,F
B,E,F
B,E,F
B,E,F
B,E
B,E,F
E,F
E,F
E,F
E
E,F
Reverse Osmosis
Manufacturer
Aerojet
Du Pont
Gulf
Universal
Aerojet
American Standard
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Universal
Aerojet
American Standard
Du Pont
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Universal
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Raypai
Universal
American Standard
Du Pont
Gulf
Raypak
Universal
Reverse Osmosis Type
fl.P.'s = Turbulence Promoters)
Tubular, Normal Flux
Hollow Fiber, B-5's
Spiral Wound
Tubular, Normal Flux
Tubular, Normal Flux
Tubular, W/T.P.'s
Hollow Fiber, B-5's
Spiral Wound
Tubular, Normal Flux
Tubular, W/T.P.'s
Hollow Fiber
Spiral Wound
Tubular High Flux
Tubular High Fltoc
Tubular, W/T.P.'s
Hollow Fiber, B-5's
Hollow Fiber, B-9's
Spiral Wound
Tubular, Normal Flux
Tubular, W/T.P.'s
Hollow Fiber, B-9's
Gulf Spiral Wound
Tubular, Normal Flux
Hollow Fiber, B-9's
Spiral Wound
Tubular, W/T.P.'s
Hollow Fiber, B-9'e
Spiral Wound
Modified Tubular
Tubular, High Rejection
Tubular, W/T.P.'s
Hollow Fiber B-9's
Spiral Wound
Modified Tubular
Tubular, High Rejection
-------
SECTION IV
DATA COLLECTION
Process Variables
The operational plan for the study required that the performance of
six reverse osmosis units be examined, when feasible, under six types
of feed flows:
1. Full Post Treatment: reactor-clarified, sand
filtered, granular activated carbon filtered,
diatomaceous earth filtered, chlorinated, pH
adjusted effluent.
2. Reactor-clarified, sand and granular activated carbon
filtered, chlorinated, pH adjusted secondary effluent.
3. Sand and granular activated carbon filtered,- chlorinated,
pH adjusted effluent.
h. Reactor-clarified, sand filtered, chlorinated, pH
adjusted effluent.
5. Sand filtered, chlorinated, pH adjusted effluent.
6. Chlorinated, pH adjusted effluent.
The first five post-treatment sequences were evaluated with respect
to their solute removal characteristics. The permeation performances
of the six reverse osmosis units were observed under conditions of
"new" and "in-service" membranes, various types of flow patterns,
membrane cleansing methods, etc.
Accuracy of Physical Data
Equation (l8), defined in Appendix A-l, was used as the osmotic
pressure correction for the reverse osmosis flux calculations. This
correction is based on a number containing only one significant figure,
but since the flux equation was used primarily to compare similar sets
of data, percentage deviations of osmotic pressure arising from small
computational inaccuracies, resulted in negligible error in the
computed terminal ratios. The exclusion of an osmotic pressure factor
would have caused substantial distortion in the terminal ratios.
All thermometers, water meters and pressure gauges were calibrated
within their normal range of use at the start of the project. Observed
data were adjusted for the small corrections at the time of measurement.
These instruments were checked at irregular intervals during the study.
At the study's conclusion, the instruments again were calibrated;
corrections were minimal.
-------
10
Random and indeterminate errors in instrument observations were estimated
using the following assumed deviations:
Pressure readings ± 5 psi.
Temperatures 1 1° P.
Conductivities It was assumed that the instrument was
accurate to i 2$ and that it could be
read to i 5 microrahos.
Liquid Flow Rates These were determined by reading the
difference in the integrator reading
during a stop watch measured interval
of one minute. The integrator difference
was recorded to ± Q.05 gal. and it was
assumed that the stop watch was accurate
to ± 1$.
Membrane Area None. The areas were furnished by the
manufacturer and were assumed to be correct.
Using the above assumptions within normal data ranges, the approximate
experimental errors for important performance ratios were calculated:
Recovery Ratio ± 5^
Total Rejection Ratio + 2
-------
u
samples were taken of the product and occasionally the brine flow from
each unit, at times selected to coincide with the time of recording the
physical data measurements. Starting in mid-July, 1970, the practice of
taking daily composites of all product water samples was employed whenever
feasible. In some instances this required the installation of gmg.ll
booster pumps on discharge lines. Brine samples were usually taken
weekly and were prepared by combining two grab samples at the end of
a 2k hour feed-product water compositing period. The resultant data
were used primarily as material balance indicators.
Peed and discharge streams were composited for testing by continuously
passing a portion of the total flow through small diameter black plastic
tubing to a central bank of three-port, time-sequenced solenoid valves.
At pre-determined intervals, each valve was activated by a Paragon
(Model 1015-ors) timer clock to discharge a limited quantity into an
individual sample bottle, housed in a household refrigerator. These
samples were collected daily for analysis.
Analytical Methods
Table 2 lists the various chemical analysis methods used in this project.
The results for the phosphorus and nitrogen constituents using the methods
listed in the table sometimes show wide variability.
In the case of phosphorus, using the ascorbic acid method, the
concentration as determined by the analyst was influenced by three factors:
1. The proper control of the reactor-clarifier (for a
time the coagulator had poor internal circulation
due to paint films stripping from sub-surface walls);
2. The care used in selecting and positioning matched
sets of tubes as required for operation in the 880
milli-micron wavelength field of the spectrophotometer
(in this case a Bausch and Lomb Spectronic 20).
3. The day-to-day performance and trends of the reverse
osmosis units, (e.g. - the phosphate concentration
in the product water tended to rise as phosphates
were deposited on the membrane surface and would
diminish again after the scale had been removed
by an EDTA flush).
As the study progressed, it was found that over 95$ of the total
phosphorous was in the ortho form. For this reason, analysis for
total phosphorous was dispensed with.
The concentration of nitrate nitrogen sometimes showed even more
variability than the ortho-phosphate. This was probably caused by
periodic changes in the nitrate level of the secondary effluent. It
was frequently noted that a grab sample collected at the end of a twenty-
four hour compositing period showed a twenty-fold difference from
-------
12
Table 2. CONSTITUENT ANALYSIS METHODS
Constituent
Acidity - CaC03(EsA)
Total Alkalinity-CaC03
AmTionia Nitrogen (ng/l)
Biochemical Oxygen Demand
Boron (rag/l)
Calcium (mg/l)
Chemical Oxygen Demand-
Total and soluble (mg/l)
Chloride (mg/l)
Chlorine-Residual (mg/l)
Coliform (MPN/100 ml)
Total Dissolved Solids
Fluoride (ag/l)
Hardness - CaCO, (mg/l)
Hltrate -Nitrogen (mg/l)
Nitrite-Nitrogen (mg/l)
Organic-Hi trogen (mg/l)
Phospnorous-P, Total and
Ortho
Potassium (mg/l)
Specific Conductance
@ 25° C. (ndcroahos/cm)
Sodium (mg/l)
Sulfate (mg/l)
Surfactants (mg/l)
Suspended Solids (mg/l)
Turbidity (J.T.U.)
Method
Methyl Orange -
pH 4.5
Phenylthalein &
Methyl Orange
Distillation and
Direct Hesslsrizatior
5-day 8 20° C.
Carmine Reagent
EDTA Titration
Dichromate - Low
Level
Mercuric Nitrate
Titration
Xodometric
Multiple Tube
Fermentation
Evaporation @l80° C.
SPADIIS (w/Distilla-
tion)
EDTA Titration
Bruclne Sulfate
FWPCA Manual
11
12
i
«•
-
-
19
-
-
-
257
-
165
Standard
Methods
12th Edition
p. No.
47
48
193 (B)
415
63 (B)
74
-
87 (B)
91
567
-
144 & 135 u
i47 (B) .
Transcribed from Treatment Plant Lab Records
Distillation and
Direct Nesslerization
Ascorbic Acid
Beckman Flame
Wheatstone Bridge
(Bedonan Model)
Beckman. Flame
Gravimetric
Methylene Blue
0.45 miron membrane
filter @ 105" C.
Hach Model 2100
225
-
-
-
-
-
275
403 + 392
-
238
238
287
296
540 (B)
-
Sample Filtered
Through Milllpore
0.45 Micron Filter
(+)=Filt'd.(-)=UnfilW
-
-
-
-
+
+
Total -
Sol. + w/inorganlc
binder
+
-
-
*
*
+
-
-
*
+
.
+
*
•t-
+
-
-------
13
related composites. In some instances it appeared that the nitrate
reduction occurring in the carbon towers showed an absolute
concentration rather than a percentage type decrease. Since the flaw
in the post-treatment plant was influenced to some degree by the
requirements of the reverse osmosis units, this could also result in
varying nitrate levels in the R.O. feed solution. At times the
concentration varied by a hundred-fold within the same week.
The nitrate analyses themselves possessed reasonably good accuracy.
For example, a set of twelve spiked samples with the additions ranging
from 0.0 to 10.0 mg/1 nitrate nitrogen, showed average discrepancies
of less than plus three per cent of the nominal values.
On the whole, all analyses appeared to be reasonably accurate. An
ionic equivalent balance made in August, 1970 showed 11.2 milli-moles
of cations and 10.7 milli-moles anions, or a balance within five per
cent.
Further comments on the phosphate and nitrate analytical methods will
be found in Appendix A-5•
Data Compatibility
The performance ratios used in evaluating the operation of the reverse
osmosis units are defined and when necessary, derived in Appendix A-l.
After each performance ratio was calculated for a particular time group
or set of process variables, it seemed desirable to examine their
mutual compatibility. This was done by making use of various material
balance and rejection ratio identities.
Where observed data are internally consistent, most of these
relationships yield agreement ratios equal to unity and the degree of
variation from that value provides an index of error.
It should not be expected that data obtained from a demonstration
study such as this, using plant scale equipment affected by numerous
known, unknown, and partially controllable variables, would show the
same degree of reliability or reproducibility as might be expected
from smaller laboratory-type experimentation with only a few variables.
It will be shown in later sections that most of the observed and
computed data obtained in this study are probably accurate to plus
or minus five to ten per cent. Where wide deviations occur, they may
be explained as due to insufficient data, the inclusion of several
incongruous data sub-sets into larger groups, probable errors in some
analyses (particularly nitrates, phosphates, C.O.D., etc. which sometimes
have low reproducibility ratios) or process disturbances (power outages,
equipment or membrane failures, rapid membrane fouling conditions,
ineffective membrane flushing operations, etc.)
-------
The apparent accuracy of some of the tabulated data and ratios included
in this report might have been improved if some data, inconsistent with
the rest, had been arbitrarily deleted. This practice was avoided.
Only about three data points out of many thousand were discarded and
this in one single instance where their inclusion would have grossly
distorted the result. It seemed preferable, instead, to present all
of the available information so that the degree of accuracy of the
study might be better evaluated.
-------
SECTION V
COMPUTATIONAL METHODS
Introduction
The primary data collected during this work include transient and
recorded instrument readings, results of laboratory analyses, and
reports of visually observed process conditions. These data
classes have been described in Section IV, -which also includes a
largely subjective estimate of their probable precision. It is the
purpose of this section to define how these types of information were
used to derive various performance ratios and to develop indices which
may be employed to obtain an objective estimate of the data accuracy.
For clarification it is wise to explain the basis of the data processing
methods used in this report. Since all of the work was concerned with
plant-scale operations with either unknown, uncontrolled or perhaps even
uncontrollable variables, and since it was felt that all data have
significance of either a negative or positive nature, all data with few
exceptions was included as observed or corrected.
As data accumulated, various material balance and statistical ratios
(which will be described below), were developed to indicate the internal
integrity of the grouped factors so that rational decisions on the
probable data accuracy might be made.
Mathematical Analyses
The performance of the equipment within the post-treatment area was
evaluated by determining, for each step, a "reduction ratio" applicable
to a particular chemical constituent or physical property. This ratio
is defined as the fraction of the particular feed constituent which was
removed by the passage of the fluid through the post-treatment step,
and is calculated by using Equation (l5>) which is defined and listed
with all other numbered equations and variables in Appendix A-l.
The reason why uniquely paired data sets must be used in this equation
and not the "period-averaged" analyses is discussed in detail in Section
VII under the sub-heading "Data Consolidation."
For the sake of making cost breakdowns, some effort has been made to
determine secondary post-treatment operating efficiencies, but extensive
refinement of data was considered to be unwarranted because of the
variability in the data.
When considering the permeability of the reverse osmosis units, the
important performance factors are the temperature-corrected water recovery
ratios (Rc), the temperature and pressure corrected A values, the tangent
of the log A x 105 vs log time (in hours) line (b), the total rejection
ratio (J.fc) and the average rejection ratio (Ja)«
15
-------
16
While the symbolic notation used in this report and the algebraic
equations employed in calculating the various performance factors from
the raw plant data are found in the Appendices, a number of these items
require further explanation.
The observed feed flow rate to a reverse osmosis unit (?„) and the
observed product flow rate (F ) were both measured values. The reject
flow rate (F ) was determinedly difference. After correcting F for
temperature by Equation (12) to obtain F , the corrected recover
at 25° C conditions was determined.
From Fpc, using Equations (12), (17) and (l) the most important eval-
uation parametric value, the water permeation rate (A), was obtained.
The latter value is proportional to the net effective pressure, the
available membrane area, and the flow rate, (A=gm H^O/sq. cm-atm-sec).
The value "A" was not arbitrarily chosen, but reflects a factorial
kinship to the Pure Water Permeability Constant (PWP) or "A" defined
by Souririjan in Reverse Osmosis (1970), page 179. The difference
between the (A) and (PWP) is associated with the applied water character.
Operating specifically with pure water, the constant (PWP) varies only
with pressure. The causative physical change in the membrane is known
as "compaction."
When a solution is made up of multiple solutes (organic and inorganic;
suspended and dissolved) the "A" value is never constant, but varies
with several conditions including pressure. A saline solution depresses
the "A" value and this is probably caused by "concentration polarization."
The adhesive character of the solute and turbulence at the membrane
surface will determine whether a time dependent decrease of "A" will
also occur. The physical change causing the latter "A" depression is
known as "membrane fouling." The complexity of treated and untreated
secondary effluent as used in this study would cause "A" to vary/decline
for all three reasons: compaction, concentration polarization, and
membrane fouling.
The "A" values computed in this report are usually lower and not easily
translatable into manufacturers' specified membrane values, since the
units of expression may not be the same or more importantly since the
manufacturers values were obtained under ideal applied-fluid conditions.
A few comments are required to indicate the precise definition and
meaning of the "A value vs time" slopes (b) plotted in this report.
Though some of the computed "A's" are based on "new" membrane data and
others on "used" membrane data following procedural changes, initial
time (t) of any data set was assumed to be the arbitrary hour one, to
obtain rational slopes of (b). Zero time could not be used because the
log of zero is negative infinity and is meaningless to graph.
While the above anomaly might have been avoided by redefining "A" as a
semi-log relationship it was determined that this would not give a
straight line plot from the test data. The data does indicate that the
-------
IT
effect of migration of the normally present constituents into the membrane
•wall is a logarithmic function of time and thus requires a log-log
presentation.
The amount of funds available and manufacturers' difficulties in supplying
new membranes made it impossible and impractical to change membranes in
accord with each feed type. An attempt was made, therefore, to minimize
the fouling effects on the "A" value by using the best quality feed first
and the lowest quality last.
In Sections IX through XIV, data and calculations for various consolidated
time groups are presented to depict the effects of special conditions.
With the effects of fouling minimized, each group should not be regarded
as a fraction of the composite group data, but should be considered as
individual with its own time continum.
Another area studied in this reverse osmosis project was the solute
rejection characteristics of each unit. Tables will appear later
depicting concentrations of the feed, product and brine flows for
each unit during various time study group periods. Both average
rejection ratios (Ja) and total rejection ratios (j^) were computed,
when data were available, for each of these time periods using Equations
(l4) and (l6). These results are tabulated in the reverse osmosis unit
discussion sections.
In conformity with what appears to be the usual practice, the average
rejection (Ja) was calculated for each individual data set and then all
of the individual Ja's were averaged to obtain the mean Ja value listed
in the tabulated data. Similarly the total rejection ratios given in
the tables were determined by taking the mean feed, product and brine
values for the period and then calculating the total rejection (J^.) for
the total period. (When Ja and J^ are based on conductivity data,
Equation (12) or its equivalent must be used.)
The reason why all available solute concentration data were included
in the summaries (as was also true for all permeation type information)
was that one of the objectives of this project was to indicate the
quality of the data obtainable from a carefully controlled demonstration
project of this nature. The elimination of some data, by either
statistical methods or through subjective opinions, would have destroyed
the integrity of the results. It became necessary, however, as a result
of this decision, to develop various internal mathematical indices which
would give an idea of the probable consistency and accuracy of the
included information. This was accomplished through the use of several
"accuracy indicators". (See Appendix for list and derivation of all
equations).
Within the reverse osmosis area several expedients were possible.
Standard deviations (%) - Equation (22) and (sz) - Equation (23) were
determined for the various b values and water recovery ratios, while
the computation of a data correlation coefficient (r) - Equation (2l)
provided an overall evaluation of the A vs time relationship.
-------
18
The analytical data for the feed, product and brine concentrations
vere audited by computing a material balance agreement ratio, (E),
using Equation (7), as well as the determination of various sz
values when required.
While agreement ratios E~ and E (the two are identical and are defined
by Equations (8) and (9Kwould riormally provide a means of cross-
checking the internal accuracy of the EQ and J^ ratios, they were
not wholly satisfactory on this project because of the slight difference
in computational methods for Ja and Jt, as explained above. They did,
however, provide reasonably close correlations in most instances.
Electronic Data Processing
The raw data assembled during this work include over 10,000 chemical
analyses and over V?,000 individual physical observations. The numerous
tables included in this report show only the selected period averages
and the various derived ratios prepared from this mass of data. These
calculations were made using six main programs and numerous sub-routines
prepared specifically for processing this data on an IBM 1130 Computer.
Since these programs were prepared to conform to this project's specific
data input and printout format they may not be suitable or adaptable
for use in other situations.
For the sake of brevity only one example of a daily printer output
appears in Section VIII; however, a copy of the complete stack of
"monthly statistic summary" printer outputs is included in Appendix A-6.
-------
SECTION VI
SECONDARY EFFLUENT
The source water used in this study was the secondary effluent from
the Hemet-San Jacinto Valley Water Reclamation Facility owned and
operated by Eastern Municipal Water District (see Figure l). Originally
the plant was designed for 2.5 million gallons average daily flow.
The design capacity was rapidly being reached just as the reverse
osmosis project was drawing to a close. Although average daily flows
were below design capacity, it was the effects of daily peak flows
which indicated a need for expansion. The most apparent effects in
the secondary effluent were (l) lack of nitrification and' (2) high
suspended solids content.
Located five miles northwest of the City of Hemet (see Figure 2) the
plant operates on the conventional activated sludge principle for the
treatment of primary clarified sewage. After biological treatment
and final clarification, there is an optional step of disinfection by
chlorination. The chlorine residual, which averaged 2.k mg/1 (range
0-7 mg/l) between March and November 1970 (weeks 1-3*0 > took on
particular significance since the program called for the discontinuance
of activated carbon adsorption at the beginning of 1971.
Because activated carbon removes low concentrations of residual chlorine
almost indefinitely, the aforementioned concentrations were never
considered to be hazardous to the reverse osmosis membranes. With
filters absent, however, the membranes could be damaged severely. In
anticipation of removing the activated carbon filters, (not actually
to occur until February 1971) treatment plant chlorination was terminated
in November 1970. After this procedural change no detectable short or
long term effects upon the membranes could be traced to the absence of
chlorination. What membrane damage did occur, is attributed to membrane
fouling and natural deterioration (hydrolysis) of the membrane with time.
The data, for the above reasons, was not separated into sets according to
"Chlorination" and "Non-Chlorination".
In addition to the treatment plant chlorination (disinfection), there was
the previously mentioned pre-R.O. unit chlorination which was practiced
throughout the operating period except during the first three weeks.
With a range limit of 0 to 1 mg/1, the attempt was made to maintain most
unit feeds at a 0.5 mg/1 chlorine residual.
Table 3 is a tabulation of constituent seasonal variation of the
secondary effluent. At one time it was thought that an analysis of
hourly variation would be helpful, but this seemed unnecessary as all
post-treated or untreated effluent water was passed through the 6,000
gallon clearwell before entering the R.O. Units. Variations of
constituent concentrations were thus "buffered" in the mixing chamber,
which held about one tenth of the total daily flow through the reverse
osmosis plant.
19
-------
CONTROL BUILDING
i_ji_ .
. DISTRICT L A BORA TORY
SECONDARY
DIGESTER
HCADWORKS P PRIMARY CLMl\F\LUS
CLNTRIFUGE
PRIMARY DlGf.STf.R5
SECONDARY
REVERSE. OSMOSIS BUILDING
REACTOR -CLARIF1ER
•• - -K ?
/«
SLUDGE-SLURRY DISPOSAL BASIN
- - * > '
CHLORINt CONTACT CHAMBER
ro
o
Figure 1. Aerial view of facilities
-------
OSMOSIS
PILOT PLANT
LOCATION MAP
EASTERN MUNICIPAL WATER DISTRICT
COUNTY OF RIVERSIDE, STATE OF CALIFORNIA
MAY I, 1972
LEGEND
EASTERN MUNICIPAL WATER DISTRICT
Figure 2. E.M.W.D. location map
ro
H
-------
Table 3. ANALYTICAL RESULTS OF SEASONAL VARIATION IN SECONDARY EFFLUENT
(mg/1 except turbidity as J.T.U.)
Constituent
Summer 1970
July
Through
September
Total Acidity (CaC03) 21.0
Total Alkalinity (CaC03) 230.7
B.O.D. - 5 day @ 20° 34.7
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
UfHo-N
W03~N
Organic BF
Ortho-P
Dissolved Oxygen
Sulfate
Suspended Solids
T. D. S.
54.4
133.0
41.1
56.5
17.2
5.4
3.0
14.2
2.8
154.0
6.4
725.0
Total Hardness (CaC03) 208.0
Turbidity
4.6
Fall 1970
October
Through
December
21.7
224.8
-
72.3
138.6
27.8
42.3
-
6.6
-
15.1
-
164.0
11.1
694.0
232.0
7.7
Winter 1971
January
Through
March
31.9
225.2
22.5
65.9
146.5
35.9
66.4
17.0
10.0
4.0
13.8
-
175.0
16.0
723.0
241.0
lU.O
Spring 1971
April
Through
March
29.0
250.2
22.0
63.8
150.0
29.8
60.5
13.6
4.9
12.5
0.4
152.0
8.0
720.0
231.0
6.7
Yearly
Average
25.9
232.7
26.4
64.1
142.0
33.6
56.4
15.9
6.7
3.5
13.9
1.6
161.0
10.4
716.0
228.0
8.2
-------
23
Table k depicts the average quality of domestic water supply to the
severed areas of the Eastern Municipal Water District. The constituent
levels are weighted for the volumetric contributions from the three
main water sources in the Valley. Colorado River, wells and Lake Hemet.
It can be seen that the sodium and chloride constituent levels are
much higher in the secondary effluent than in the domestic waters and
this accounts for the percentage difference between TDS levels. Waters
enter the domestic supply system below the recommended levels of TDS,
and increases approximately 30$ by the time it reaches the treatment
facility. Through research, it is now believed that the incremental
increase of sodium chloride is due to the domestic and commercial
water softeners operating in the Valley to counter the effects of
hardness.
To have this highly saline water enter the closed basin underground
reservoir in increasing quantities would cause a deterioration of the
latter. Removing the salt before percolation would protect the
groundwater reservoir at least from domestic waste water refractories.
Table k. DOMESTIC WATER QUALITY TO SEWERED AREAS OF E.M.W.D.
COMPARED TO SECONDARY EFFLUENT
(mg/l)
C onstituent
Boron
Chloride
Fluoride
Hardness
Sodium
Sulfate
T.D.S.
Domestic
Water
0.3
62
0.5
245
57
122
493
Secondary
Effluent
0.4
142
0.6
241
147
161
716
-------
SECTION VII
POST-TREATMENT OP SECONDARY EFFLUENT
Introduction
Discussion of post treatment can be divided into two categories:
"physical performance characteristics" and "operating costs." The
first will be discussed presently. The second is more appropriately
reviewed in Section XVI, "Reverse Osmosis Costs".
Scope of Post-Treatment Operations
Facilities (Figure 3) were provided for the five major secondary
effluent treatments. Four of these are discussed in detail below:
1. Reactor-clarification with cationic polymer and alum
injection. Purpose: to aid in the removal of residual
suspended solids. This was accomplished in the 15=5 ft.
diameter, open-top coagulator with a side wall height of
10.5 ft. and a capacity of 15,000 gallons (Figure 4). With
this capacity, the retention time was 2.4 hours and the
overflow rate was 0.55 gpm/sq.. ft. Operating on the sludge
blanket principle, the tank was provided with a rim mounted
overflow launder, a slow-moving electrically-propelled
mixer blade and chambers for chemical mixing and sludge
separation. An influent controller, which was activated
on water consumption demand, had a discharge rating of
125 GPM, the maximum possible flow rate through the
clarifier. A single 10 HP electric motor Aurora pump
fed the clarifier. Equipment was also provided for
automatic sludge or slurry removal. Clarification was
accomplished in the clarifier with the aid of two chemicals:
alum and Calgon cationic polyelectrolyte #ST-260. Best
operation seemed to occur using these chemicals in the
following proportions: polymer, 0.5 to 0.75 mg/1; alum
70 to 100 mg/1. Both chemicals were added at the same
point in the mixing chamber. Routine jar tests helped
the operator to determine the appropriate dosages.
2. Sand Filtration (See Figure 5).Purpose: to remove
suspended solids. The set of three filters were designed
for a total flow rate of about 2 to 3 gal./sq. ft-min.,
but were normally operated at about one-third that rate
( 1 gal/sq. ft-min. or l6 gal. min. downflow rate). The
three 54 in. diameter 60 in. shell height pressure sand
filters were normally operated in parallel. Each was packed
from the bottom up with the following layers:
8 in. with 1 1/2 to 1 in. gravel
2 1/2 in. with 1 in. to 1/2 in. gravel
2 1/2 in. with 1/2 in to 1/4 in. gravel
24
-------
POST SECONDARY EFFLUENT TREATMENT AREA
FLOW DIAGRAM
NOT TO SCALE
(MINOR BY-PASS t, AUXILIARY LINES NOT SHOWN)
LEGEND: OEQUIPMENT SYMBOL
O SAMPLING POINTS
MAJOR BY-PASS LINES
Figure 3. Post secondary effluent treatment facilities
ro
vn
-------
ALUM AND POLYMtD INJECTION POINT
SLUDGE. ,/t ,
DRAIN <\r—-1
SCPAI7AT/0M
CHAMBER
SAMPLING COCK.
CATC. VALVC—
SLUDGE. VALVE.
CLA-VAL eiOO
ro
ON
TANK x1 1
DRAIN Xl '
Figvire 4. Reactor-clarifier
-------
Figure 5. Sand filters (right) and activated carbon filters (left)
ro
-------
28
3 in. with 1/4 in. to 1/8 in. gravel
3 in. "with #12 sand (.8 to 1.2 mm)
21 in. of fl6 filter sand (l45 to .55 mm)
Each filter vessel was fed at the top through a centrally
located 3 in. pipe and was drained from the bottom through 8
symmetrically located cross -chord 3 A in. laterals, with a
tank total of 60 - 1/4 in. holes on 6 in. centers.
The sand filters were backwashed, either when the discharge
pressure fell about 20 psi below the normal feed pressure of
100 psi or when the discharge turbidity or suspended solids
content increased abruptly. While in use, the sand filters
were backwashed with four types of wash water: reactor
clarifier effluent, activated carbon filter effluent and
diatomaceous filter effluent. The rate of backwashing varied
considerably but averaged 9 GPM/sq. ft.
Granular Activated Carbon Filtration . Purpose : to remove
residual organics. For this filtration, three J2 in. diameter
by 120 in. shell height pressure carbon towers were employed.
They were operated in series at a hydraulic loading rate of
one-third the design rate of 3-5 gal/sq. ft-min. Each tower
was packed the same as the sand filter through the $12 sand
layer, but on top, in place of the $l6 sand, there was a 68
in. layer of activated carbon, purchased under the trade
name of "Filtrasorb 400@" of the Calgon Corporation. The
activated carbon had the following physical properties and
met the specifications listed below:
Table 5. ACTIVATED CARBON PHYSICAL PROPERTIES AND SPECIFICATIONS
Total surface area
(N2 BET method) M2/g 950-1050
Bed density backwashed and
drained, Ibs/ft3 26
Particle density wetted in water g/cc 1.3-1.4
Pore volume cc/g ' 0.85
Effective size mm 0.8-0.9
Uniformity coefficient ....... 1.9 or less
Sieve Size U.S. Std. Series
Larger than No. 8 - Max $ 8
Smaller than No. 30 - Max % 5
Mean Particle Diam. mm 1.5-1.7
Iodine Number, Min 900
Abrasion Numer, Min. JO (ASTM D 2355)
Ash Max % 8
Moisture as packed, Max % 2.0
-------
29
The carbon filters were backwashed with various post-treatment
process effluents when vessel pressure drops exceeded 10 to 20
psi. Backwash flow rates were usually on the order of k-5 GJM
per sq. ft. A record of backflushes according to post-treatment
unit equipment appears in Table 6, while dates, volumes, and
duration of backflushes appear in the Appendix Section A-k.
k. Diatomaceous Earth Filtration, Purpose: to provide final
polishing prior to reverse osmosis units. For this two D.E.
pressure filters were operated in parallel, each having a 30
inch diameter and a 48 inch shell height. Both filters contained
thirty-one vertically-positioned polypropylene-covered cycolac
filter elements. Such elements were kQ inches long and 3 inches
in diameter. Estimated filtration area was 97 sq. ft. per vessel.
Accessory equipment to the D.E. filter included two pre-coat
preparation pots and a filter body feed tank. Of the post-treat-
ment processes, the D.E. filters seemed to present the most prob-
lems. Essentially all problems were related to pre-coating of the
filter element tubes: according to the person in charge of the
units, "OSiere was always a question as to whether coatings were
uniform and of the appropriate thickness." In addition,
prevention of pre-coat wash-outs during filter operations was
not always assured, as pressure fluctuations in passing from
"wasting" to "on stream" could facilitate a wash-out of filter
media. Potential for wash-out was reduced substantially by
manipulation of the discharge piping. Installation of observation
ports would have given additional insurance in forming uniform
filter coatings, but time did not permit as D.E. filters were
the first equipment to be removed from the post-treatment area.
The D.E. filters were operated for five months with considerable
"down" time. Except for brief periods of singular operation,
the D.E. filters were operated in parallel.
Equipment and Facilities
All of the equipment and facilities which were part of this project,
including the laboratory, a small office and storage area, but excluding
the reactor-clarifier, clear-well, flushing water storage tank and waste
ponds (which were located in the nearby area) were housed in a rigid frame
building with an adequately drained concrete floor. The building was
approximately 62 ft. by 50 ft. with 20 ft. high walls. (See Figure l)
The temperature of the building in .the operation area was uncontrolled
for the most part although two forced-air heaters were available to
prevent frozen pipes in winter.
It should be remembered that pH adjustment is a step in post-treatment,
but because of its close association and necessity to the reverse osmosis
units, it is discussed in sections specific for the R.O. units.
-------
30
Table 6. POST-SECONDARY TREATMENT BACKFLUSH RECORD
Week No.
2
3
5
6
7
8
10
12
13
15
16
17
19
20
21
22
23
2k
26
27
28
30
36
37
39
in
te
ii^
ii.ii.
ItQ
51
52
53
5!).
55
56
57
58
59
60
6l
62
63
Total No.
Flushes Per
Time on Stream
Sand.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
3X
X
X
X
2X
X
2X
X
2X
2X
X
X
3X
X
39/63 weeks
Carbon
X
2X
X
X
X
X
X
X
X
-
X
X
X
X
2X OFF
l6/kh weeks
D.E.
2X
X
X
X
X
X
X
X
X
2X OFF
12/30 weeks
-------
31
Immediately outside the building were three reservoirs: the reactor-
clarifier, clear-well and a backflush water storage tank. The latter
two were identical with 10.5 ft. high side walls, 10 foot diameters
and 6,000 gallon capacities. Except for the final period of no post-
treatment, the clear-well always contained the product of the reactor-
clarifier.
Alternatively, the backflush tank stored both post-treatment effluent
and Colorado River Water at various times during the project.
A fourth reservoir made of sand with vinyl plastic sealed floor was
constructed to receive reactor-clarifier sludge/slurry discharge. The
reservoir's design capacity was 110,000 gallons with dimensions of
50! x 100' x 3'.
Miscellaneous Post-Treatment Equipment Specifications and General
Comments
Cross connections and by-pass piping were installed to backwash the
sand, carbon and D.E. filters and to modify or skip one or more of
the sequential, post-treatment processes. Thus, while full treatment
of secondary effluent included reactor-clarification followed by
passage through 3 sand filters, 3 carbon filters and 2 D.E. filters,
there were times when each (either singly or in combination), was
removed from service. However, it was impossible to invert the
post-treatment sequence: for example, to send carbon filtered effluent
through the reactor-clarifier.
Backwashing was accomplished with a 2 X 2 1/2 X 9A -GBPA Aurora Pump
powered by a 15 horse power 3,500 RPM, 3 phase, 230/1^0 volt motor;
two inch Hendy flow meters at each step provided both total flow and
instantaneous flow rates. Total flows and instantaneous rates were
recorded manually. Alternatively the reverse osmosis units required
an accurate record of flow and pressure patterns. To serve these
purposes Barton Models 2^2 and 208 A (pressure and flow recorders
respectively) were employed.
For coating the D.E. filter elements and reactor-clarifier alum injection,
a diaphragm duplex type, BIF model 1210-05-9109 pump was installed. Its
head for D.E. transport, was 0-8 GPH at 125 pounds discharge pressure.
The alum injection head for the same pump was capable of 2 GPH at 125 psi
discharge pressure. The power source was a 1/6 horse power, 115 volt,
60 cycle motor through an adjustable V-belt.
At various times during the project, air bumps were needed to help remove
the more adhesive coatings of impurities from both the post-treatment
filters and the reverse osmosis units. A Speedaire Model 12991, 1 horse
power compressor provided the air to facilitate air bumping. Air bumping
was an irregular practice, and no single criterion was used to justify the
need for an air bump.
-------
32
With only a few exceptions, the process piping used was P.V.C. 1220
and performed adequately.
Data Consolidation
The post- treatment constituent data were averaged into weekly groups
to reflect the effects of the various post-treatment sequences.
These data are shown in the influent-effluent columns of Tables 7
through 12. Included in these tables are the applicable "reduction
ratios" based upon the change in concentration of a particular
constituent before and after a specified post- treatment step or steps.
These "reduction ratios", which are similar to the total rejection
ratios used in the evaluation of reverse osmosis, were calculated as:
where Jp and Jf are the concentrations of a given constituent in the
"product" (in this case the discharge) and the "feed" flows
respectively.
It is important to realize that the reduction ratio was not calculated
from the list of influent and effluent data. In order to obtain more
significant information, the reduction ratio was calculated only from
paired influent/effluent data from each individual week and the ratios
for all weekly periods were then averaged to obtain the period mean.
A copy of a speciment computer printout for weeks 16-33 (selected at
random from the complete set of about 130) is shown as Table 13. The
column headings, which were chosen for ease in the interpretation of
the printout, require some explanation.
Code: PT12 indicates a post- treatment area sampling point,
PT meaning that a secondary effluent feed sample
analysis (point l) 'is being compared with the
reactor-clarifier effluent analysis (point 2). The
numeral 3 would indicate the sand filters, h the
carbon filters and 5 the D.E. filters.
Run Analysis: PS04 signifies a post treatment process analysis
for sulfate.
Feed and
Product: First set, columns 3 and k', these show the average
levels for the number of paired sets available, as
indicated in column 6. Second set, columns 7 and 8;
show the averages of all sample analyses made in the
designated time period. NF is the total number of feed
samples and NP the total number of product (outflow)
samples .
Rejection: This is the calculated total rejection in per cent.
(J X 100) as defined by Equation (15).
Zero's indicate missing data, not the absence of the solute.
-------
Table 7. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: REACTOR-CLARIECER; INELUENT SOURCE: SECONDARY EFELUENT
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D
Total C.O.D.
NH^-N
NO^-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1-15
Influent
mm
208.6T
32.57
62.53
128.00
5^.59
62 .in
17-31
9.61*
-
11*. 00
159.20
9.50
781.85
232.1*1*
5.65
Effluent
.
-
-
-
-
29.55
29.17
-
6.30
-
_
-
-
690.. oo
-
2.36
Reduction
Ratio
.
-
-
-
-
.382
.505
-
.1*06
-
_
-
-
.151
-
.581
Weeks 16-33
Influent
20.36
232.59
39.25
55.62
133.00
lH. 52
55.97
17.06
6.12
2.50
18.75
153.00
7.31
713.60
211*. 19
1*.86
Effluent
27.18
193.00
-
5^.70
123.33
22.66
32.39
15.23
If .1*7
2.23
10.65
173.78
9.60
679.89
196.21
3.50
Reduction
Ratio
-.231
.17^
-
.085
.ot*o
.1*51*
.1*32
.01*6
.361
.153
.1*32
-.196
-.297
.039
.071
.293
(JO
U)
-------
Table 7 Cont'd. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: BEACTOR-CLARIFIER; INFLUENT SOURCE: SECONDARY EFFLUENT
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NH,-N
NO^-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1*2-1*8
Influent
30.99
218.86
-
72.00
138.00
32.33
57.63
-
6.85
-
li*. 67
186.20
10.33
738.33
6.85
Effluent
to. 90
177.57
-
69.00
-
26.27
to.05
-
9-53
-
7.26
223.25
5-33
785.00
238 .it
3.90
Reduction
Ratio
-.320
.189
-
.Oil*
-
.188
.305
-
-.39
-
.505
-.176
.1*81*
-.063
.022
.1*38
Weeks ll-9-57
Influent
31. 3t
223.67
22.50
65.22
1^9-75
38.65
69.05
17.00
11.55
1*.00
13.27
169.33
17.00
707.78
231* .89
16.69
Effluent
38.07
2O2.22
-
63. 1*
-
28.38
1*5.68
_
9.72
-
8.1*7
185.00
9.60
575.00
230.00
8.88
Reduction
Ratio
-.211*
.096
-
.027
-
.266
.338
_
.158
-
.362
-.076
.1*35
.135
.021
.1*67
-------
Table 8. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: SAND FILTER; INFLUENT SOURCE: SECONDARY EFFLUENT
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NH0-N
NOj-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 3^-lH
Influent
20.02
232.12
-
80.00
139.50
26.90
te.6i
-
6.97
-
lt.86
166.00
13. ^o
706.25
2314-. oo
9.15
Effluent
2h. 58
211-0.00
_
-
138.50
26.07
39.21
-
5.67
-
lk.20
161)-. 80
7.00
689.00
232.88
if.8o
Reduction
Ratio
-.227
-,03>K
_
_
.082
.01*6
.1014-
-
.201
-
.028
.011
.lf78
.007
.005
.Vf5
Weeks 58-63
Influent
30.65
21*6.83
_
61.83
155.33
25.72
60. 1£
—
6.60
-
13.20
114.5.0
-
721.67
223.17
9.65
Effluent
32.32
25^.83
_
61.33
172.00
23.76
1U.87
-
-
-
13.18
11*5.0
-
71^.17
221.50
3.53
Reduction
Ratio
-.05l»-
-.032
_
.008
-.01
.076
.308
-
«•
-
.001
.000
_
.010
.008
.631*
u>
-------
Table 9« POST-SECONDARY TREATMENT UNIT CONSTITUENT .ANALYSES
UNIT: SAND EELTERS; INELUENT SOURCE: REACTOR-CLARIITER
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NH.-N
NOj-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1-15
Influent
_
—
«.
128.00
29.55
29.17
-
6.30
-
_
-
-
690.00
-
2.36
Effluent
_
_
_
-
2^.95
29.23
-
-
-
_
_
-
703.57
-
1.30
Reduction
Ratio
_
_
_
-
.156
-.002
-
-
-
.
-
-
-.008
-
.^50
Weeks 16-33
Influent
27.18
193-00
5^.70
123.33
22.66
32.39
15.23
k.kj
2.23
10.65
173.78
9.60
679.89
196.21
3.50
Effluent
25.92
212.70
k6.hk
-
20.35
25.22
_
2.70
-
9.77
3.00
713.00
191.71
1.32
Reduction
Ratio
.012
-.022
.049
.102
.221
_
.090
-
.083
_
.688
-.080
.Oik
.621*
-------
Table 9 Cont'd. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALOGS
UNIT: SAND FILTERS; INIUJENT SOURCE: REACTOR-CLARIEEER
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
MHo-N
J
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1*2-1*8
Influent
1*0.90
177.57
-
69.00
-
26.27
1*0.05
_
9-53
-
7.26
223.25
5.33
785.00
238.11*
3-90
Effluent
39.66
181*. 71
-
69.50
125.00
17.50
33-55
_
9-77
-
9.00
-
3.00
738.00
232.1*0
2.32
Reduction
Ratio
.030
-.01*0
-
-.007
-
.33^
.162
_
-.021*
-
-.239
-
.1*37
.022
.021*
.1*05
Weeks 1*9-57
Influent
38.07
202.22
-
63.1*1*
-
28.38
1*5.68
_
9-72
-
8.1*7
185.00
9.60
575.00
230.00
8.88
Effluent
38.67
206.1*1*
2.70
62.67
1^3.57
25.62
35-1)2
_
8.10
-
8.18
189.38
3.20
690.56
227.38
2.60
Reduction
Ratio
-.016
-.021
-
.012
-
.097
.225
m.
.167
-
.031*
_
.667
-.201
.010
.713
-------
Table 10. POST-SECONDARY -TREATMENT TOUT CONSTITUENT ANALYSES
UNIT: CARBON FILTERS; INFLUENT SOURCE: SAND FILTERS
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NH--N
NO^-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 314- Ul
Influent
21^.58
214O.O
_
_
138.50
26.07
39.21
_
5.67
-
1^.20
161). !80
7.00
689.00
232.88
1^.80
Effluent
25.114.
25^.0
.
714-. 00
136.00
5.76
10.814-
-
0.86
-
llf.OO
165.63
O.ll-O
688.75
226.62
0.1*6
Reduction
Ratio
-.023
-.058
_
_
.054
.779
.7214-
_
.814-5
-
.0114.
-.003
.9^3
.025
.027
.905
-------
Table 11. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: CARBON FILTERS; INFLUENT SOURCE: REACTOR CLARIFIER AND SAND FILTERS
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1-15
Influent
-
2^.95
29.23
-
703.57
1.30
Effluent
"™
15.32
15.09
-
681.82
•53
Reduction
Ratio
-
.656
.638
-
4
.591
Weeks 16-33
Influent
25.92
212.70
20.35
25.22
2.70
9.77
3.00
713.00
191.71
1.32
Effluent
20.05
220.80
1.20
53.16
130.50
5.37
6.96
3-37
If .28
8.76
171.71
0.60
201.92
0.35
Reduction
Ratio
.299
-.068
-.037
.736
.72^
-.21*8
.103
.800
-.051
.766
to
vo
-------
Table 11 Cont'd. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: CARBON FILTERS; INFLUENT SOURCE: REACTOR-CLARIFIER AND SAND FILTERS
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NOo-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1)2-1)3
Influent
39.66
l&k. 71
69.50
125.00
17.50
33.55
9-77
9.00
3.00
739-33
232 Ao
2.32
Effluent
195.29
69.50
129.67
3.93
7.62
5.17
213.20
0.00
698.33
23^.75
0.35
Reduction
Ratio
-.057
0
-.037
.775
.773
ATI
1.000
.016
-.001
.853
Weeks l)-9-57
Influent
38.67
206 Ai).
2.70
62.67
1^3-57
25.62
35. te
8.10
8.18
189.38
3.20
690.56
227.38
2.60
Effluent
35.80
210.00
65.00
180.00
3.80
-
lSO.00
0.00
650.00
235.00
oAo
Reduction
Ratio
-.170
-.111
-.015
.910
_
1.000
.058
.967
-------
Table 12. POST-SECONDARY TREATMENT UNIT CONSTITUENT ANALYSES
UNIT: D.E. FILTERS; INFLUENT SOURCE: REACTOR-CLARIFJD3R, SAND AND CARBON FILTERS
(mg/1 except turbidity as J.T.U.)
Constituent
Acidity
Alkalinity
B.O.D.
Calcium
Chloride
Dissolved C.O.D.
Total C.O.D.
NH--N
NO^-N
Organic N
Ortho-P
Sulfate
Suspended Solids
T.D.S.
Total Hardness
Turbidity
Weeks 1-15
Influent
Influent
—
-
-
-
-
15-32
15.09
-
-
-
-
-
6.00
681.82
-
0.53
Effluent
Effluent
.
-
-
-
-
9.23
9.11
_
-
-
_
-
6.00
686.88
-
0.44
Reduction
Ratio
-
-
_
-
.266
.010
_
-
-
_
-
0
-.007
-
.126
Weeks 16-28
Influent
21.56
223.85
1.20
54.28
132.60
5.79
7.36
14.26
5.60
4.28
9.73
163.00
1.00
646.43
199.50
0.36
Effluent
20.22
216.88
0.90
55.22
134.29
5.38
7.10
14.46
4.38
2.76
9.90
169.00
0.00
662.54
204.00
0.29
Reduction
Ratio
.058
.004
.250
-.017
-.049
.025
.036
-.104
.017
.402
-.017
-.049
1.000
.024
-.002
.227
-------
Table 13. POST TREATMENT DATA WEEKS l6 THROUGH 33
PRETREATMENT DATA.
CODE
PT12
PT13
PT14
PT15
PT23
PT24
PT25
PT34
PT35
PT45
PT12
PTli
PT14
PT15
PT22
PT24
PI2&
PT34
PT35
PT45
PT12
PT13
PT14
PT15
PT23
PT24
PT25
PT34
PT35
PT45
RUN ANALYSIS
PS04
PSU4
P504
PS04
P5G4
PS&4
P504
PS04
PS04
PS04
PTOS
PTOS
PTJS
PTOS
PTtJS
PTUS
PTDS
PTUS
PTDS
PTOS
PTHARD
PTHARO
PTHARD
PTHARO
PTrtAKD
PTHARO
PTHAKi)
PTHARD
PTHARD
PTriARD
WEEKS 16
FEED
148.00
.00
146.80
157.00
.00
115.1^
169.83
.00
.00
169.50
713.60
742.00
636.44
727.00
743.00
662.50
6d0.53
.00
713.00
640.56
211.307
20*. 571
£07.916
212. bOO
194.428
197.916
19^.700
191.714
183. OCO
198.500
THRuUGH 33
PRODUCT
177.00
.00
174.60
167.50
.00
171.71
163.00
.00
.00
166.25
685.53
713.00
650.00
o68.56
713.00
640.42
659.53
.00
720.00
633.11
196.231
191.714
202.416
204.600
191.714
201.583
201.600
201.428
196.750
198.750
REJECTION
19.58-
18.93-
6.68-
1.96
4.02
.74
3.93
3.91
5.31
8.04
4^04
3.33
3.09
.97-
1.16
7.13
8.52
2.65
3.76
1.40
1.84-
4.61-
5.06-
7.50-
, .12-
PAIRS
6
0
5
4
0
7
6
0
0
4
15
1
9
12
1
12
15
0
1
9
13
7
12
10
7
12
10
7
4
8
FEED
153.00
153.00
153.00
153.00
173.78
173.78
173.78
.00
.00
171.71
713.60
713.60
713.60
713.60
679.89
679.89
679.89
713.00
713.00
640.42
214.187
214.187
214.187
214.187
196.214
196.214
196.214
191.714
191.714
201.923
PRODUCT
173.78
.00
171.71
168.10
.00
171.71
168.10
171.71
168.10
168.10
679.89
713.00
640.42
659.53
713.00
640.42
659.53
640.42
659.53
659.53
190.214
191.714
201.923
203.818
191.714
201.923
203.31U
201.923
<£33.dl8
203.318
NF
7
7
7
7
9
9
9
0
0
7
15
15
15
15
18
18
18
1
1
12
16
16
16
16
14
14
14
7
7
13
NP
9
0
7
10
0
7
10
7
10
10
18
1
12
15
1
12
15
12
IS
15
14
7
13
11
7
13
11
13
11
11
-------
43
Unlike chemical levels within the reverse osmosis test area, toe
consistency of which may be evaluated through material balance
correlations, etc. (as discussed in a number of other sections of
this report), the accuracy of the post-treatment analytical data
was inferred only by computing standard deviations for integral
sets. Random checking of such deviations has led to the conclusion
that under the existent test conditions the rejection ratio method
used yields information of greater internal consistency than ratios
of either the constituent discharge level from each stage to the
constituent level of the secondary effluent or the mol fraction reduction
of each item per unit flow rate through each post-treatment step.
However, none of these post-treatment area ratios can be of any service
in detecting operational, sampling or analytical errors. These may have
been present in some degree* and while the arbitrary omission of
apparently aberrant data from some weekly sets might have improved the
consistency of some ratios, the practice has been avoided in all, except
a very few, instances. It seemed more informative to let the observed
data stand on their own merits as indicative of what might "normally"
be expected from in-plant, rather than laboratory or research type,
operations.
Summary and Remarks on Post-Treatment
When the per formance of one secondary post-treatment unit is pitted
against that of another, more questions often arise than are answered.
Some example questions may be, "Was the appropriate polymer used at
the optimal concentration?" or "Should the filters be backflushed at
regular, frequent intervals or when differential pressure across a unit
demands a backf lush?" or even "How often was the D.E. pre-coating
successful?" These and other questions should be kept in mind reading
this summary section.
The following categorical evaluation should give an impression of the
efficiencies of the post secondary effluent processes.
Major effects of alum coagulation and polymer injection on secondary
effluent:
1. Acidity increased about 25 per cent.
2. Alkalinity decreased about 15 per cent.
3. Dissolved C.O.D. decreased about 35 per cent.
k. Total C.O.D. decreased about kO per cent.
5. Nitrate nitrogen decreased about 35 per cent.
6. Ortho-phosphate decreased about kO per cent.
7. Total dissolved solids decreased about 8 per cent.
8. Sulfate increased about 15 per cent.
9. Suspended solids decreased about 17 per cent with a
range between a 30 per cent increase and a 50 per cent
decrease.
10. Turbidity decreased about 50 per cent.
-------
Major effects of sand filtration on untreated secondary
effluent:
1. Total C.O.D. decreased about 20 per cent.
2. Turbidity decreased about 50 per cent*
3. Suspended solids decreased about 50 per cent.
Major effects of sand filtration on coagulated reactor-
clarified secondary effluent:
1. Dissolved C.O.D. decreased about 15 per cent.
2. Total C.O.D. decreased about 15 per cent.
3. Turbidity decreased about 55 Per cent.
4. Suspended solids decreased about 60 per cent.
Major effects of activated carbon filtration on reactor-
clarified, sand filtered secondary effluent:
1. Dissolved C.O.D. decreased about 70 per cent.
2. Nitrate nitrogen decreased about 45 per cent.
3. Turbidity decreased about 80 per cent.
4. Suspended solids decreased about 90 per cent.
Major effects of activated carbon filtration of sand
filtered (only) secondary effluent:
1. Total and dissolved C.O.D. decreased about 75
per cent.
2. Turbidity and suspended solids decreased about
90 per cent.
Major effect of D.E. filtration on reactor-clarified, sand
and activated carbon filtered secondary effluent:
1. Residual detectable suspended solids (l mg/l) removed
(lOO per cent reduction).
Reactor-Clarification vs. Sand Filtration
The performance of the reactor-clarifier was generally satisfactory.
Influent C.O.D. of which 75 per cent was soluble, was reduced by 40
per cent. The dissolved organic removal could be due to adsorption
on the alum floe and/or biological reactions. The B.O.D. concentration
in the secondary effluent averaged around 26 mg/l (Table 3) and the
low overflow rate of 0.55 gpm/sq. ft. and 2.4 hr detention time is
sufficient for biological reactions. Biological activity (despite
the chlorination of secondary effluent from March, 1970 to November,
1970) is also shown by the 35 per cent decrease in nitrate-nitrogen
which is probably due to denitrification. Sand filtration was capable
of 20 per cent C.O.D. reduction, due to the removal of particulate C.O.D.
-------
which is close to the 25 per cent particulate C.O.D. removed by alum/
polymer clarification. Alum (polyelectrolyte) clarification had three
advantages over sand filtration: (a) removal of colloidal and some
organic materials, (b) removal of phosphate, about kO per cent due to
the aluminum phosphate chemical reactions, and (c) reduction of pH to
reduce both membrane hydrolysis and precipitation of chemical compounds.
Sand filtration showed a greater capability of removing suspended solids
as compared to the reactor-clarifier treatment (50 vs 17 per cent).
This is probably due to biological reactions in the clarifier which
releases gases, making it difficult to maintain the sludge blanket
thereby causing floe to float and discharge with the effluent. A
negative aspect of reactor-clarification is the high cost as compared
to sand filtration.
Sand Filtration vs. Granular Activated Carbon Filtration:
Granular activated carbon filtration was used to remove dissolved organics
and particulate matter. Granular activated carbon treatment resulted in
70 per cent dissolved C.O.D. removal, k^ per cent nitrate nitrogen removal
(probably due to denitrification), a turbidity decrease of 80 per cent
and suspended solids decrease of about 90 per cent. Sand filtration on
untreated secondary effluent resulted in 20 per cent C.O.D. removal,
turbidity reduction of 50 per cent and suspended solids decrease of 50
per cent. The advantage of granular activated carbon treatment is the
greater removal of constituents from secondary effluents that can cause
membrane fouling. The disadvantage of activated carbon treatment is the
higher costs as compared to sand filtration.
D.E. Filtration
D.E. filtration appears to offer little advantage as a polishing filter if
sand filtration or the granular activated carbon filter is used prior. The
higher cost of D.E. filtration and the apparently slight benefits derived
as a polishing filter makes its use of questionable value.
In conclusion, it appears that substantial refractory reduction is feasible
using three of the four post-secondary treatment processes. They are:
alum reactor-clarification, sand filtration and granular activated carbon
filtration. It is difficult to form generalizations from the tabulated
data but it seems valid to infer from the data that if full treatment
(reactor-clarification, sand filtration and granular activated carbon) is
scaled at unity, lesser degrees of post-secondary effluent treatment would
have roughly the following ratios of solute removal:
Sand filtration and granular activated carbon filtration 0.90 to 0.95
Reactor-clarification and sand filtration 0.70 to 0.80
Sand filtration only °'5° to 0.60
One other combination not tested during this project was reactor-
clarification followed by granular activated carbon. Operating cost of
this combination,however, is predictably higher (l) because of normal
carbon expense and (2) rapid carbon fouling caused by alum carry-over.
-------
SECTION VIII.
REVERSE OSMOSIS OPERATIONS - PRELIMINARY DISCUSSION
Introduction
Although "Reverse Osmosis" is already a familiar term in the water
works field, there are presumably many who have a limited knowledge
of the process. It is fortunate, however, that there are numerous
sources now available describing reverse osmosis in various degrees
of detail. For this reason the process theory will not be discussed
in this report. An excellent introductory article appears in the
August 31, 19T2 "Reference Number" issue of Water and Sewage Works.
For greater detail, books by Merten or Souririjan should be helpful.
Most attention during this study was centered around the comparative
performance characteristics (flux rate, configuration influence,
membrane life, etc.) of commercially available R.O. (reverse osmosis)
units. The various feed conditions (already described) provided the
means to gain broad information of R.O. unit capabilities. Inspection
of individual unit performance using specific feeds was the means of
determining the range of limitations for each unit.
Sections IX through XIV describe individual unit capabilities and
characteristics, while Sections XV and XVI present the performance
and cost comparisons.
Equipment and Facilities
Five reverse osmosis units were connected in parallel to a feed manifold
permitting the entire group to receive the same feed or any fraction of
the group to be supplied with a feed from any other point of the
post-treatment sequence. ; .~
- •! .-i- »t ^
Each reverse osmosis unit had its own feed pump, pH control and sulfuric
acid make-up system, piping, valves, instruments and sampling points.
Figure 6 is a generalized process and instrument diagram applicable to
all units. Of course, there were slight differences between units;
Figure 7 shows a particular arrangement of a reverse osmosis unit
installation.(Universal).
Initially, brine and product waters from each unit were returned to the
terminal secondary effluent flume of the sewage treatment plant. Later,
in keeping with the basic project purpose (of salt removal), these brines
were sent to the District Salt Evaporation Pond. No problems with the
brine disposal pond were recorded since the pond operated substantially
below design capacity. Originally, the pond was constructed to accomodate
all water softener regeneration brines which would otherwise be discharged
into the sewage collection system.
k6
-------
POST-TREATED __
OrUIENT
pUiT
|ftlMP
•@ M IN I
BRINE
PRV
LEGEND
flow METER
[INFLOW INDICATOR
(PR> PRESSURE
INDICATOR
•l) CONDUCTIVITY
INDICATOR
TEMPERATURE
INDICATOR
To WA.TC,
(**
* w w ••» » • »
CMCASUKED)
FCCD
PUMF
£
PRV
RCVCJ23C.
OSMOSlft
UNIT
PRODUCT
WATER
> K WH METER
RELIEF
NlTROGCN DOTTLE
(s) SAMPLE POINT
COMPOSITE DIAGRAM
REVERSE OSMOSJS
(VARIES 3LICHTLY WITH UM1T)
Figure 6. Generalized reverse osmosis unit flow scheme
-------
IIMLtT WATER
(I5-£0 PSIG)
C-l "
ACID
CARBOY x
(I5Y CUSTOMER)
FM
—txl-i
WASTt WATER
ij-OOOCHjrOOOO
PS-
. TAPWATER
50-90
C -1 PRODUCT WATEJ3
LEGEND
AC-l
AC-Z
C-l
CV-1
CV-Ze3
M-l
p-1
PH-1
PI-1
PI-Z
PI-3
PR.-1
PS-1
RV-1
T-l
V-l
V-Z
V-3
V-4
ps-e
5V-1
FI-I
Low 'PRESSURE. ACCUWULiTOtt
HIGH PRESSURE. ACCUMULATOR
1 INCH MALE HOSE. CoupLiiua
CHECK VALVE. 1 INCH
CHECK VALVE VL IMCM
R.O. MODULE DANK
PUM.P
PH COMTUOLLtR
GdtlCC 0-100 P31G
GAUGE 0-1500 PSI£
PR.H3JUIZE G/1UCE. 0-1500 P5IG
QflCK PrjtSSUTJE RECULATOR.
Loaa Of PKIME SwtrcH
RtLIEF V/1LVE SET AT 72? PSIfi
ADJUSTABLE To 1100 PSlfi
ACID MlX(MG TANK
1 IWCH SHUT-OFF VALVE
Vf. INCH MODULE BY-PASS BALL VALVE
Vl IWCH RECULATOQ BY-PA.SS BALL VALVE
'/£ IMCU DBINE RECIRCUUATION BALL VALVE
PBESSUEE SWITCH SET (8 100 PSIG (DOWN)
'/2 INCH SOLENOID VALVE.
FLOW INDICATOR
FLOW SCHEMATfC
10,000 GPD REVERSE 05M03I3
DESALTING UMIT
Figxire T. Flow schematic, 10,000 gallon per day R.O. unit (Courtesy Universal Water Corp.)
-------
Operational Control °
All units operated twenty-four hours a day, seven days a week whenever
possible, and were under the control of one technician ten hours per
day, five days a week. Although the units were unattended at night
and on weekends, considerable overtime was required. The remainder
of the crew consisted of two chemists and an equipment operator for
the post-secondary treatment area.
Operating data were collected from each R.O. unit by making careful
instrument readings at least once each working day. These were
supplemented by near-hourly observations of the various indicating
and recording instruments, a practice used to detect abnormalities
of operation.
These data were recorded on key-punch data input sheets for processing
on the IBM 1130 computer. The computer printout consisted of "daily"
and "monthly" summary reports, both of which included raw data, various
ratios (e.g. - water recovery and rejection), ambient temperature flux
rates, A values, plus a limited amount of chemical analytical data and
operational comments. Over a thousand of these printout sheets were
generated during the experimental phase of the project. A copy of a
specimen daily printout is shown in Table 1^ while the monthly printouts
Can be found in the Appendix. As mentioned earlier, the algebraic
formula used in computing the various ratios and correction factors
are also found in the Section "Appendices." Though the computed ratios
were adequate for daily control and short term process evaluation, it
eventually became evident that methods for cross-correlating data and
computed ratios would be needed if incongruities of the data were to
be detected. Using various derived equations (Appendix A-l) additional
computer programs were written to develop more useful information and
to provide other indices of performance.
pH Control
At this time it seems appropriate to mention pH control of the unit
feeds as it cannot be classified as a membrane cleansing agent per se,
nor is it fitting to include it as a secondary post-treatment since
it was never removed during the experimental work phase. The evidence
for making pH control a requirement is substantial. In 1957 Breton (E.J.)
published his findings which indicated a correlation between feed pH and
membrane longevity. Prom this, pH control was assumed to control the
rate of some chemical reaction associated with membrane molecular struc-
ture. Breton submitted that this reaction was the hydrolysis of cellulose
acetate, an ester. This view was reinforced when Vos, Burris and Riley
derived a rate constant for the reaction based on their own experimental
evidence. Being pH dependent, the constant reaches a minimum between
pH *(•.5 and 5.0, well on the acid side (see Merten, p. 151). pH control
in the acid direction also retards scale formation.
-------
Table Ik. DAILY COMPUTER PRINTOUT
PASTERN MUNICIPAL WATER DISTRICT
DAILY STATISTICS DETAIL
AEROJET AMER.STND.
•••PHYSICAL DATA»»»
FFPD FLOW, TOTAL
PH SENSOR
TO DRAIN
NET* UNADJUSTED
PRESSURE
CONDUCTIVITY. UNADJ.
DETERMINED AT
TEMPERATURE IN LINE
BRINF FLOW. BY DIFFERENCE
PRESSURE
CONDUCTIVITY
DETERMINED AT
TEMPERATURE IN LINE
PRODUCT FLOW
CONDUCTIVITY. UNADJ.
DETERMINED AT
PH FEED. RAW
FEED. POST-ACIO
BRINE
PRODUCT
MEMBRANE AREA
FEFO METER READING
PRODUCT METFR READING
ELECTRIC POWER METER
TEST HOURS ACCUMULATED
OFF TIME (TODAYI
TIMF DATA TAKEN
PRETRFATMENT CODE
TREATMENT. KIND
START TIME
FND TIME
LENGTH
•••ANALYTICAL DATA***
TOTAL DISS.SOLIDS. FEED
BRINE
PRODUCT
0.00
0.00
0.00
0.00
0
0
0
0
0.00
0
0
0
c
o.oc
0
0
0.00
0.00
0.00
0.00
0.0
0.
0.
0.
0.0
0.
0
0
NONE
0
0
0
9.15
0.15
0.33
8.70
600
1450
82
82
2.15
400
4950
84
83
6.55
345
82
7.20
4.25
3.55
4. SO
791.7
63440.
785650.
7522.
2333.3
4.
945
31
NONE
0
0
0
710
2805
225
RAW DATA
OU PONT
8.75
0.20
0.35
8.20
630
1450
85
B6
1.90
495
3850
88
88
6.30
635
87
7.35
5.70
5.75
5.55
160900.0
681840.
349090.
12448.
3697.7
0.
1530
23
NONE
0
0
0
705
2505
285
GULF G.A.
8.95
0.20
0.00
8.75
600
1430
86
92
2.75
595
4400
93
93
6.00
130
92
7.00
5.00
5.15
5.35
900.0
456940.
872780.
1018.
4358.4
0.
1450
23
NONE
0
0
0
705
2530
40
REVERSE OSMOSIS TEST
FOR SEP 9. 1970
UNIVERSAL
9.95
0.10
0.00
9.65
640
GPM
GPM
GPM
GPM
PSI
1460M1CROMHOS/CM
87
87
5.40
600
OEG.F.
DEG.F.
GPM
PSI
1930MICROMHOS/CM
89
90
4.45
DEG.F.
OEG.F.
GPM
960MICROMHOS/CM
88
7.00
5.60
5.65
5.60
224.0
940300.
173230.
17910.
852.9
0.
1510
31
NONE
0
0
0
710
1080
440
DEG.F.
SQ.FT.
GALLONS
GALLONS
KWH
HOURS
HOURS
HOURS
PPM
PPM
PPM
TURBIDITY FEED
PRODUCT
TOTAL CHEM.OX.DEM. FEED
PRODUCT
DISS. CHEM.OX.DEM. FEED
PRODUCT
CHLORINE FEED
PRODUCT
AMMONIA NITROGEN FEED
PRODUCT
FEED
PRODUCT
FEED
PRODUCT
NITRATE NITROGEN
TOTAL PHOSPHORUS
•**CO"MENTS»*»
.0
.0
6.6
1.3
.0
.0
6.6
2.3
.0
.0
6.6
.1
710
1080
440
.8
.0
.0
6.6
3.9
PPM
PPM
PPM
J.T.U.
J.T.U.
PPM
PPM
PPM
PPM
PPM
PPM
PPM
PPM
PPM
PPM
PPM
PPM
AMER.STD. OFF 4 HRS PUMP REPAIR. RUPTURED MODULE 1435 - BYPASSED. (CHANGED OIL)
•••PERFORMANCE INDICES. CORRECTED FOR TEMPERATURE***'
RECOVFRY R
REJECTION. TOTAL J
REJECTION, AVERAGE K
WATER FLUX F
WATEP FLUX F
WATER FLUX F
0.00
0.00
0.00
69.35
76.20
89.04
66.27
57.13
75.99
53.59
91.49
95.57
38.33
34.96
43.27
PER CENT
PER CENT
PER CENT
0.00000000 0.00762106 0.00003377 0.00521092 0.01685708 GPM/SO.FT.
0.0000 10.9743 0.0486 7.5037 24.2741 GPO/SQ.FT.
0.00000000 5.17549B02 0.02?93634 3.5387568511.44771006 GMS/SOM*SEC
A VALUE
0.00000000 C.15S72073 0.00061441 0.08982336 0.27335041 G/SM*SEC*AT
-------
51
One scaling agent, calcium carbonate, does not exist below pH 5.0 as
a result of removing carbonate alkalinity. Another, calcium phosphate,
rapidly increases its solubility at pH's below 6.0. In spite of pH
control shown in Table 15, membrane fouling by scaling still occurred.
Except for short periods of no pH control (malfunctioning acid pumps)the
pH was nearly always within the limits recommended by the manufacturer.
Membrane Fouling and Cleansing
It is believed that at least two, and perhaps three types of membrane
fouling occurred during the operations. A calcium .deposit chemically
identified as tri-calcium orthophosphate was found in membranes of
the American Standard unit about Week 23. Similar deposits were later
found in other units. These deposits could be removed by a 15,000 to
30,000 ppm (2 to k oz./gal.) solution of EDTA (Questex l»SW or Versene
100) in water, adjusted to a pH of 7 with sulfuric acid. Solutions
of pH 3 to 4 sulfuric acid also effected various degrees of permeation
improvement but this was probably due to calcium carbonate removal.
Chemically, the calcium carbonate could have occurred solitarily,
bonded to phosphate or both.
Organic slime also contributed to membrane fouling but a 15,000 ppm
(2 oz./gal.) solution of "BIZ", an enzymatic-detergent by Proctor and
Gamble, seemed moderately effective in the removal of slimes as there
was usually a marked improvement in product flux following a "BIZ"
soaking. After the carbon and sand filter treatments were removed, a
portion of this slime film, (assumed to be also on the other units)
was removed from a Universal unit module for analyses.
When placed in water, the slime material appeared to consolidate into
an amorphous gel; the filmy appearance returned when separated from
the water. Samples of the film were sent to most of the R.O. equipment
suppliers, and one commented on the material:
"...The foulant was very slimy and brown colored, (it was
a light tan). The appearance at 1000X (raaenification) was
that of a typical membrane deposit from a unit running on
polluted surface waters. The deposit appeared to be
composed of (an) aggregation of colloidal and particulate
solids held together in a biologically oriented slime matrix.
Present in the sample were large masses of a filamentous
fungi and large numbers of a rod shaped (10 by 20>0 bacteria...
There was no obvious life in the sample but (it was felt) that
the bacteria seen grew in situ and were not trapped or deposited
in large numbers from the feed water...More than 0.5 ppm of
residual chlorine are (probably) required to prevent biological
growth in sewage effluents..."
Figures 8 and 9 show two microphotographs made by Gulf General Atomic
at high magnification. The large groups of bacteria were difficult to
record photographically (without drying and staining) because of Brownian
motion.
-------
Table 15. RECORD OF R.O. UNIT FEED pH
Period Covered
Week Nos.
2-7
5-7
1-7
1-7
7-17
14-28
7-24
7-24
7-30
28-33
24-33
24-33
31-33
35-36
34-41
34-38
38-41
33-41
33-41
41-47
41-49
41-46
41-48
49-53
46-57
48-57
61-64
57-64
57-66
62-64
57-64
64-69
64-66
66-69
64-68
64-69
Post-Treatment Sequence
(A=Reactor-Clarifier )
(B=Sand Filters )
(C=Carbon Filters )
(D*D.E. Filters )
(E=Pre-R.O. Unit Chlorination)
(F=pH Control )
A,B,C,E,P
A,B,C,E,F
A,B,C,E,P
A,B,C,B,P
A B C D E F
A B C D 1 F
A,B,C,D,E,F
A B C D E F
A,B,C,D,E,F
A,B,C,E,P
A,B,C,E,F
A,B,C,E,F
A,B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
B,C,E,F
A,B,C,E,P
A,B,C,E,F
A,B,C,E,P
A,B,C,E,F
A,B,E,F
A,B,E,C
B,E,F
B,E,P
B,E,F
B,E
B,E,F
E,P
E,F
E,F
E
E,F
Reverse Osmosis
Manufacturer
Aerojet
Du Pont
Gulf
Universal
Aerojet
V
American Standard
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Universal
Aerojet
American Standard
Du Pont
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Universal
Du Pont
Gulf
Universal
American Standard
Du Pont
Gulf
Raypak
Universal
American Standard
Du Pont
Gulf
Raypak
Universal
Average pH
5.82
5.52
5.62
5-71
5.67
4.64
5.69
5.51
5.66
4.77
5-45
5-38
5.63
5.77
5-33
6.11
5.49
5.67
5-14
5.91
5.42
5.54
5.77
f • •
5.23
f w^»^^
5.65
5.54
5.57
f w f \
5.49
x w • s
7-45
5.70
5.01
5.66
^ W ^f^f
5.55
7.46
5.78
-------
53
«?;$*£%• at,
•_>i> •-<<-••.. \ f .•
~*-~4i2:*s».-* * •' ' v^"
j*l*- , ^*«i JP»•*•--'**"
fe^tge.:^ ^^/
Figure 8. Microphotograph of slime removed from
Universal R.O. unit (Courtesy, Gulf)
S £":
B
F
Figure 9- Slime microphotograph (Courtesy, Gulf);
B=bacteria, F=fungus, S=particulate solids
-------
Another form of fouling was particulate solids deposition. Exactly
when these deposits began to substantially affect membrane flux is
uncertain. For the most part the buildup was gradual except when a
post-secondary effluent treatment process was removed. By the time
untreated secondary effluent was being fed to the units, particulate
solids were probably the chief fouling agents. Evidence of particulate
fouling is graphically shown in the project report by Aerojet-General
Reverse Osmosis Renovation of Municipal Wastewater, (contract ih-kO-lBk)
page 1^0. The fouling depicted is probably a good representation of
membrane fouling as it occurred at Hemet. In light of the potential
for particulate solids fouling, one manufacturer (Du Pont) insisted
that the feed flow to their unit be passed first through Cuno filters
(See Section XI).
In addition to EDTA and BIZ, a number of other types of membrane
cleansing and flushing solutions were used, particularly during the
latter phases of the project when the feed quality had deteriorated
considerably due to the removal of post-treatment processes. All
cleansing solutions were used with the knowledge and consent of the
R.O. manufacturers. A number of weak acid solutions (pH 3 to 5) were
used some of which were sulfuric, phosphoric, sulfamic and hydrochloric.
Additionally, solutions of sodium perborate with Tritox-100, carboxymethyl
cellulose, and sodium hypochlorite were tested. Cleansing effects are
generally discussed in the particular unit sections (IX through XIV)
while copies of flushing data sheets appear in Appendix Section A-3.
The rate of membrane fouling plus the frequency and effectiveness of
the cleansing methods employed are factors which must be considered to
evaluate the reverse osmosis units. The usual procedure, in this project
was to flush a unit whenever the flux rate dropped to 85 or 90 per cent
of the initial rate. Consequently, the operating periods for the R.O.
units varied from several days to many weeks, depending upon the feed
conditions, membrane fouling susceptibility,flushing history, etc. In
the latter stages of the project, an attempt was made to determine optimum
flushing time more objectively by a mathematical approach using Equation
(6). Membrane fouling is considered to be inevitable when using a high
solute feed such as the treated and untreated effluent used in this study.
There were, however, two partially controllable factors which may have
increased the rate of fouling: 1 ) variation of Reynolds numbers stemming
from changes in basic flow patterns, and 2 ) the low levels of pre-R.O.
unit chlorination (less than 0.5 ppm residual) necessary to accomodate
the Du Pont nylon filament permeators. While none of the manufacturers
seemed to feel that low chlorine residuals would adversely effect the
performance of their equipment, it is possible that some were not then
thinking in terms of a gradually deteriorating feed quality.
Reynolds Numbers
It was not a part of this project to research the theoretical aspects
of reverse osmosis, but the role of Reynolds Numbers as encountered
requires some consideration. Reynolds Number is dimensionless as shown
in the equation below (Goel and McCutchan, October, 19J1).
-------
N.
re
X D
55
where: Nre = Reynolds Number (Dimensionless)
D = Diameter of tube (ft.)
v = kinematic viscosity (1.039 X 10" 5 ft.2
/sec) determined experimentally
?k = brine velocity (ft./sec.)
Because Vb (velocity) = Q (volume as ft3/sec), the Nre equation
A (area of tube in ft2
can be expressed in terms of volume rate (gpm). Combining all factors,
the expression becomes:
wre = 3281*. Q/d
•where d = tube diameter (inches) and
Q and d are unique for each calculation
The Reynolds Number is important as an indicator of turbulence within
the membraned tube. The effect of increasing the flow rate is to
increase the Reynolds Number but the rate can be increased only to
the extent that it doesn' t cause a detrimental pressure drop across
the R.O. unit. By reducing the flow rate and thereby the Reynolds
Number, the rate of fouling may increase. In addition, water recovery
and solute rejection may be affected adversely.
It was possible, with some degree of confidence, to estimate the Reynolds
Numbers for the Aerojet-General, Raypak and Universal R.O. units which
had relatively uniform geometric properties. The Reynolds Numbers for
the American Standard unit, with its internal spherical "turbulence
promoters", and the Du Pont and Gulf units, with their intricate internal
flow patterns, could only be examined through some analog function such
as the frictional energy loss per unit length of module, etc.
A communication from Abcor, with regard to the American Standard unit
(the Conseps division of American Standard, the R.O. unit manufacturer
was acquired by Abcor December 20, 1970) reads in part as follows:
"... In a tube with turbulence promoter spheres,
turbulence is not equally distributed so we are dealing
with an average Reynolds Number... Based on an idea that
the same amount of energy loss per length of tube means
the same amount of turbulence (as implied by the Reynolds
Number) an equivalent or apparent N^e can be established
for a tube with T.P. (Turbulence Promoters)..." Example:
TM 5-1^ with T.P.
at 0.2k gpm pressure drop is 22 psi - energy loss
0.25 x 22 =5-5 gpm psi =2.5 watts
TM 5-1^ without T.P.
at 0.7 gpm pressure drop is 8 psi - energy loss
0.7 x 8 =5.6 gpm psi"
-------
"... This means 0.25 gpm with T.P. is approximately
equal to 0.7 gpm without T.P., the latter representing
a Reynolds number of approximately 3500«»'"
Reynolds Numbers for the tubular units (Aerojet-General and Universal)
and the annular spaced Raypak unit were estimated from the feed and product
flow rates, by using the modified Reynolds Number expression, after making
some simplifying assumptions:
1. The kinematic viscosity (determined from brine density
and absolute viscosity) of the exit modular flows was
essentially constant regardless of the section location.
2. The parasitic head losses and induced turbulence in
the inter-unit piping and return bends, etc. were
negligible;
3. The product permeation rate in gal./sq. ft.-min. was
constant regardless of the module position. (This
implies a constant net effective operating pressure
throughout the unit and the absence of localized
internal fouling, obstructions, etc.)j
k. No tubes or modules were out of service.
Example:
Flow pattern "a" (Figure 10) for the Aerojet unit was an arrangement
of six modules in parallel followed by four modules in parallel
and terminated by two modules in parallel.
FEED
n^ ** ' ' * v \ I .—.
r\ v v v i i-T-f 533X111333, • r-i
' "J \IIVV II HJ I1 ' ' ' " x x x \ I
i v i v v v ryr o • •' > ' *-^ LZI___~J
. ivvvivvvv \\\'i'>'i\v\ Nin v^w vv f
U\v l v\i '•j ("*"" "-U j
""""J / i .
Figure 10. Flow pattern a .
Let:
(f) = volume rate of total feed entering
reverse osmosis unit (gpm).
(p) = volume rate of total product leaving
reverse osmosis unit (gpm).
Each module contained l6 - 9/l6" inside diameter tubes in series.
Feed rate (f) was 7-25 gpm and the product rate (p) was 2.25 gpni.
The brine flow rate for each section of parallel modules was
then calculated:
-------
First section: (f/6) - (p/12) = 1.021 gpm = ^
Second section: (6 Q.^) . (p/i2) = 1.3!^ gpm =
Third section: (1^/2) - (p/12) = 2.500 gpm =
In this case d = 0.5625 in. and Nre = 581|.OQ. Since there are
maximum and minimum brine flows within each section, three values
of Nre per modular section are given below:
Table l6. AEROJET-GENERAL REYNOLDS NUMBERS
First Section
Second Section
Third Section
Max. Nre
7,060
8,9110
15,700
Min. Nre
5,960
7,850
1^,600
Avg. Nre
6,510
8,390
15,150
The critical minimum Reynolds values for the R.O. units have
been summarized in Table 17 according to flow patterns. In the
case of Raypak, the hydraulic radius rather than the actual
tube diameter was used in computing the value.
Turbulet flow conditions are usually present when the average
Reynolds Number is above 3,000 or k,QOO, but concentration
polarization problems may still occur in the last tube section
while operating in this range. An average lower limit of
about 5,000 is probably desirable for most reverse osmosis
operations.
Although it is possible that some other dimensionless grouping,
such as the Schmidt or Prandtl Numbers might have had more
significance than the Reynolds Number, they were not investigated.
Data Reduction
Section V discusses the major aspects of the computational methods
used in this report. It is now desirable, in anticipation of the
presentation of the summary data sheets for the individual reverse
osmosis units in Sections IX to XIV, to discuss and give a few
examples of the IEM 1130 computer output sheets from which the
summary information was derived.
Various programs were written for the study of the reverse osmosis
water permeability and rejection data. Each had a number of
auxiliary programs developed for special purposes. Some examples
of final program print-outs are shown in Tables Ik, 18, 19, 20 and
in Appendix A-6.
-------
Table 17. ESTIMATED MINIMUM REFOLDS NUMBERS PER MODULAR SECTION
oo
Unit
Aerojet General
Raypak
Universal
Flow Pattern
a
b
c
a
o
p
<1
r
s
Data Date
5/6/10
5/22/70
10/28/70
nA/7o
5/H/71
5/6/70
9/1/70
9/28/70
12/18/70
Estimated Reynolds Number-
Section
I.
5,970
3,970
5,200
10, lt-50
6,950
4,100
9,750
10, 400
8,100
II.
7,870
3,^50
5,060
12,800
1,850
15, 100
III.
15,000
5,370
9,150
22, 600
1,250
IV.
/
19,200
-------
Table 18. PROGRAM OUTPUT, "WATER PERMEABILITY STUDIES"
UNIT
DUPT
DATE
11/18/70
11/19/70
11/20/70
11/23/70
11/24/70
11/25/70
11/27/70
11/30/70
12/01/70
12/02/70
12/03/70
12/04/70
12/U7/70
TUTALS
MtAftS
UNIT
DOPI
RUN
FROM
UP06 11/18/70
AtC.H*S
4.0
22.7
45.9
119.4
145.0
164.1
210.1
272.0
293.4
324.7
343.1
366.4
437.1
RUN
HOURS
1.0
19.7
42.9
116.4
142.0
161.1
207.1
269.0
290.4
321.7
340.1
363.4
434.1
FROM
DP06 11/18/70
THROUGH
12/07/70
A VALUE
0.3553912
0.3408401
0.3135730
0.3191512
0.3167549
0.3152807
0.3056597
0.2347167
0.2437620
0.2768708
0.2741742
0.2714952
0.2703951
3.9280633
0.3021587
THROUGH
12/07/70
EXCLUDING
X
0.000000
1.294466
1.632457
2.065952
2.152287
2.207095
2.316179
2.429751
2.462996
2.507450
2.531606
2.560384
2.637589
26.798210
2.061490
AVG A VAL
0.3021587
Y
-0.449293
-0.467449
-0.503661
-0.496003
-0.499276
-0.501302
-0.514761
-0.545586
-0.547045
-0.557722
-0.561973
-0.566237
-0.568001
-6.778312
-0.521408
SLOPE
-0.045930
XX
0.000000
1.675641
2.664915
4.268158
4.632342
4.871267
5.364686
5.903692
6.066349
6.287306
6.409028
6.555567
6. 956675
61.655815
INTERCEPT
-0.426728
XY
0.000000
-0.605096
-0.822205
-1. 02*719
-1.074586
-1.106422
-1.192280
-1.325640
-1.347371
-1.39d461
-1.422694
-1.449786
-1.498153
-14.267412
STO DEV
0.008306
YY
0.201664
0.218508
0.253674
0.246019
0.249276
0.251304
0.264979
0.297665
0.299258
0.311054
0.315814
0.320625
0.322625
3.552669
T
-5.529323
VJ1
VO
-------
ON
O
Table 19. PROGRAM OUTPUT, "AVERAGE REJECTION RATIOS"
GULF
ITEM
RCHLDE
RSCCNO
RS04
RTOS
RACIOY
RALKY
RT.COD
RO.CGD
RCRG.N
RTHARO
RNH3.N
RNQ3.N
RLRTOP
RTUR3
RCAL
91 WEEKS
FEED
141.18
1412.84
364.61
821.88
194.666
36.542
44.407
28.fcOO
2.950
228.529
13.7500
5.4967
10.3588
3.6094
62, 80C
49 THROUGH
PRODUCT
19.27
112.00
4.47
63.13
123.653
12.236
3.564
1.625
.3UO
3.029
.6100
2.8033
.2560
.4013
.727
69
STANDARD DEVIATIONS
BRINE REJECTION
242.50
2584.42
760.50
1428.13
184.444
52.333
65.000
.000
3.400
424.615
15.6003
10.1783
19.3066
.0000
130.000
88
94
99
94
29
57
93
90
98
95
55
98
99
.67
.32
.42
.07
.46
.15
.28
.00
.63
.93
.84
.05
.10
.00
.60
RECOVERY EFFICIENCY
46
46
46
46
46
46
46
46
46
46
46
46
46
46
46
..57
.57
.57
.57
.57
.57
.57
.57
.57
.57
.57
.57
.57
.57
.57
90.68
94.80
102.33
97.53
92.94
148.38
91.48
.00
87.19
101.42
84.54
118.55
98.21
.00
101.46
NF NP
11 11
19 19
15 15
16 16
15 15
12 14
15 11
4 4
2 1
17 17
2 2
6 6
17 15
16 15
15 15
MB
10
19
10
16
9
6
9
0
1
13
2
6
15
0
10
MR
10
19
10
16
9
5
9
0
1
13
2
6
15
0
10
NV NE REJECTION RECOVERY EFFIC'CY
19 10
19 19
19 10
19 16
19 9
19 5
19 9
19 0
19 1
19 13
19 2
19 6
19 15
19 0
19 10
0.1425
0.0124
0.0093
0.0409
0.1577
0.2196
0.0895
0.0000
0.0000
0.0131
0.0492
0.2040
0.0455
0.0000
0.0016
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1490
0.1250
0.0321
0.1743
0.1210
0.2451
0.9129
0.1701
0.0000
0.0000
0.0965
0.0200
0.2902
0.0959
0.0000
0.0798
-------
6l
The form in which the data and calculated results appear in Table 18
requires a few comments. Again, it should be mentioned, Zeros
indicate the absence of data, not the level of the constituent. The
X and Y designations refer to the common logarithms of the hours and
A values, respectively. Both the mean A values and the mean hours
are the anti-logarithms of the means of the groups of the listed X's
and Y's, while the average A is the mean of the original A values.
The slope (b) and the standard deviation of the slope are defined by
Equations (5) and (22) found in Appendix A-l. The "A value intercept"
refers to Equation (2) and represents the statistically calculated
value of A at hour "one", while the "intercept" is its common
logarithm. After Equation (3) is used to calculate the probable
value for A at 1,000 hours, the two points may be joined on a log-log
plot to show the A vs hours regression line.
Tables 19 and 20 should be considered together as part of a paired
set of reverse osmosis data and average rejection ratios. The latter
shows all of the feed, product and brine specific conductivity data
for the Gulf unit from weeks 2 to 69. The average rejection ratio
Ja (in percent) shown in the fifth column was calculated using
Equation (l4) and the material balance ratio ("E" - or "Efficiency")
in the last column by Equation (7).
Columns 2, 3 and k of Table 19 show the period average constituent
level for weeks 49 through 69. The W-series of columns indicate
the number of data points included in the period and refer respectively
to the number of feed, product, brine, rejection, recovery and
"efficiency" data sets included in the averages. The total rejection
ratios (Jt) Equation (l6), included in the unit discussions, were
calculated from data similar to that of Table 20.
-------
Table 20. AVERAGE REJECTION RATIOS SUMMARY OUTPUT
ON
ro
6UIF
KtfK
10
11
12
13
14
15
16
11
It
19
20
21
22
23
2<>
2S
26
27
2e
29
30
31
32
33
3*
as
105 RSC
FEED
1206
1230
1311
1197
1262
1100
1133
139".
1317
12SO
1299
1206
14CO
1313
1363
127*
1199
1317
1242
12*6
12SO
1155
1463
1235
1179
11»2
1239
112*
1151
1253
1367
127S
1373
1173
OHO
PRODUCT
131
131
120
90
64
90
84
107
96
96
101
«3
•4
• 7
91
92
89
67
92
101
120
101
131
119
93
104
101
99
90
122
147
134
131
114
BRINE
2420
1937
2420
2007
184S
2019
19S6
3110
3020
2760
3440
2«
0.06193
-------
SECTION IX
AEROJET -GENERAL CORPORATION REVERSE OSMOSIS UNIT
TUBULAR MEMBRANE DESIGN
Introducti on
This unit was obtained on a monthly rental basis from the Environmental
Systems Division of Aerojet-General Corporation of El Monte, California;
the divisional name was later changed to Envirogenics Company. As
originally installed, it was termed a "tight membrane" tubular type
with a nominal capacity of about 8,000 gallons of product water per
day with a 90% rejection factor.
Physical Configuration
The installed Aerojet-General reverse osmosis unit, shown in the
photograph (Figure 11), consisted in part of 192 vertical tubes
positioned over a product water collecting pan. Each tube was 9/l6
in. I.D. by l4 ft. 3 in. long and was made up of an outer fiberglass
sheath enclosing both the tubular membrane and its spirally wound
paper cover. These tubes were arranged in series-connected groups of
16 each to form module sets, each with a total membrane area of 33-12
sq. ft. (2.07 sq. ft. per tube). The modules were connected in groups
to form various flow patterns, which in turn produced various levels
of performance. The Reynolds Numbers applicable to each of these
configurations are listed in Section VIII.
The four flow patterns tested are described below using lower case
letters to identify the pattern:
"a" - Six modules in parallel, followed in series by four
modules in parallel and terminating with two modules
in parallel. The complete unit contained twelve
modules, 192 tubes, and had a total nominal membrane
surface of 397-44 sq. ft. Membrane Set No. 1 was used
in this configuration.
"b" - Five modules in parallel followed by four modules in
parallel and terminating with two modules in parallel.
The eleven modules, 176 tubes, had a total nominal area
of 364.32 sq. ft. The unit ran only a total of 39 hours
in three consecutive days using pattern "b". Because of
this, no significant data was generated.
"c" - Same sequence as pattern "a". Membrane Set No. 2 was
used in this configuration.
"d" - Three modules in parallel, followed by two modules in
parallel and terminating with two modules in series.
The seven modules, 112 tubes, had a total nominal area
of 231.84 sq. ft. Membrane Set No. 2 was used in this
configuration.
63
-------
Figure 11. Aerojet-General reverse osmosis unit
-------
Membrane Specifications
65
A nvuflber of different membrane formulations were supplied by Aerojet-General
for their unit. Membrane set No. 1 initially contained l6o tubes with
membranes formulated from cellulose diacetate using propionamide as the
swelling agent. Thirty-two tubes formulated from a "blend" (9-B) of
cellulose diacetate and cellulose triacetate and maleic acid swelling
agent, were also provided. These tubes were of the cast membrane type
with dacron sleeves inserted into fiberglass casings. Membrane set
No. 2 was made up initially from whatever tubes were available at the
time and included both old and new "high flux" (blend) tubes and
"normal-flux" propionamide tubes.
Numerous tube failures occurred during the operating period. Replacements
for the failures were made up from whatever tubes were available at the
time, resulting in a heterogenous mixture of various membrane formulations
at random locations.
Chronological Record
The following notes are taken from the plant data logs to show the major
events and changes in operation.
Project
Week & Day
7
1
12
13
IT
21
35
36
37
5
5
3
7
6
Start of data collection. Membrane Set No. 1;
flow pattern "a". Operated on post treated
secondary effluent with the reactor-clarifier,
sand and granular activated carbon filters in
operation.
Started recycling part of brine flow to the feed.
Began in-plant chlorination of reverse osmosis feed.
Discontinued recycling brine flow to feed.
Started D.E. filtration of feed.
Changed to flow pattern "b"•
End of useful data; too many tube failures;
no replacements available.
Unit shut down; no useful data since 13-3-
Reached an agreement with manufacturer that
rental agreement ceased on 15-1.
Installed replacement tubes at the manufacturer's cost.
3
6
Membrane Set No. 2; flow pattern c . Unit restarted
on sand and granular activated carbon filtered, pre-unit
chlorinated secondary effluent feed.
Changed to flow pattern "d".
Large number of tube failures since 35-3; no replacements
available; unit permanently removed from service.
Installation removed by manufacturer-
Data Groupings
The chronological history of the Aerojet-General unit shows that, from
-------
66
March 9 to November 12, 1970 (weeks 2 through 37) a number of changes in
post-treatment conditions, membranes, flow patterns, etc. were made.
Unfortunately the large number of tube failures coupled with the non-
availability of suitable replacements made it impossible to collect
significant operating information during most of this period. In
Table 21 the available data was divided into five major time periods,
weeks 2, 3, k-6, 7-13, and 35-37. While the last time periods had
mixed flow patterns they were grouped to make the number of data sets
per period as large as possible.
Table 21. REVERSE OSMOSIS PROCESS INFORMATION
AEROJET-GENERAL
Week Nos.
2
3
li-6
7-13
35-37
Treatment
A,B,C,E,
A,B,C,E
A,B,C,E,F,
A,B,C,D,E,F
B,C,E,F
Membrane
Set
1
1
1
1
2
Flow
pattern
a
a
a
a,b
c,d
Special
Conditions
_
Brine Recirculation
Brine Recirculation
-
—
Mechanical and Operational Problems
The Aerojet-General unit experienced a much greater number of tube or
membrane failures than the other R.O. units, although all units operated
under essentially identical test conditions. While it is not appropriate
to discuss either the design advantages or apparent deficiencies of the
reverse osmosis units used in this work, a few factual comments must be made,
1. Aerojet-General initially proposed to furnish a flat-plate
unit. In the fall of 19&9, they advised that it would be
in the study's best interest to substitute a tubular type
design because the manufacture of flat plate R.O. units
was to be discontinued. Model No. 12-Bll|~6p-R, one of
Aerojet-General's first tubular membranes, was submitted
for testing.
2. At the time of the initial installation, the manufacturer
provided 38 additional tubes which could be used as
replacements, indicating that 20$ of a batch of similar
tubes had failed when tested in their laboratory.
Thirty-eight is exactly 20$ of 192, the number of tubes
originally installed. Interestingly, however, there was
no provision for replacing replacement tubes that failed.
At one time Aerojet indicated that an improved type of
tube might soon be available. It never was.
3. Twenty tubes had failed by week ten (May 1, 1970). In a
letter dated June 11, 1970, (week 15) Aerojet-General
was informed that:
-------
6,
67
... A total of thirty leaking membrane tubes (have
been removed) since May 1. In each case this required
complete shutdown of the unit...
Usable operational data (were obtained) on only four
days since May 15 and on three of these... the
operation was at best marginal as (it was necessary)
to shut the unit down on three occasions for the
replacement of four tubes..."
This situation did not improve during subsequent
months even after a nearly complete set of replacement
tubes had been installed. (Erom March 9 to November 12,
1970 there were over a hundred tube failures).
The long vertical tubes of the Aerojet-General
installation were attached, at their ends, to fixed
headers by ferrules and usually vibrated while in operation,
Most of the observed tube failures occurred at or near
the ferrules.
frequently many of the failed tubes could not be replaced
without removing adjacent tubes to provide the necessary
access.
High velocity leaks from a single tube frequently caused
adjacent tubes to fail.
The area around the unit was very humid due to partial
evaporation of the product water, spray from leaking
tubes, etc. The canvas enclosure supplied by Aerojet-
General (shown in the photograph) was of little value
in correcting this condition.
9. Pinhole leaks occurring in the interior of module sets
were very difficult to detect. In large numbers,undetected
leaks tend to distort data. The data from the 7-13 weekly
group is suspected of being distorted because of undetected
leaks, as the log A-log time slope for the period is both
positive and steep.
Erom week 2 to week 13 and from week 35 to week 37, there were 2328
available operating hours. The unit actually operated for 2031 hours,
or 87.24$ of the total time. The major out-of-service hours were as
follows:
8.
Table 22. OUT-OP-SERVICE RECORD, AEROJET-GENERAL
Mechanical problems
Membrane cleaning
Membrane failures
Alterations , additions
Feed Treatment Problems
Total Down Time
Hours
_
3
271
9
Ik
"297
*
_
.13
ll.6l».
.39
.60
12.76
-------
68
The membrane failure record for the Aerojet-General unit is shown in
Table 23. "Failures" have to be shown instead of "replacements"
because substitute membranes were rarely available in the necessary
quantity.
Table 23. MEMBRANE FAILUKE RECORD, AEROJET -GENERAL
Weekly Group
2
3
k-6
7-12
13-35
35-37
Number of Failures
0
0
2
33
TO (about)
26
During weeks 13 to 3^> the unit could not be kept in operation long
enough to secure any useful operating data. During weeks 35 "to 37 > the
unit was operated on the site under the direct supervision and control
of the manufacturer's service personnel. Failures continued to be so
numerous that the unit could be run only intermittently and it was
finally shut down at their suggestion.
Water Permeability Data
Table 2h shows the permeability data ratios for the Aerojet-General
unit. The essential information includes the test parameters (post
secondary effluent treatment, membrane set, flow pattern), the average
A value, (gm. H20/sq. cm - atm - sec) } the log A versus log time plot
with its standard deviation, the data correlation coefficient and the
average gallons per foot per day of product water at about 500 psi .
The various symbols, indices, and ratios used in Table 24 and others
which will appear later, are defined and discussed in Sections V, VIII,
and the Appendix Section A-l.
The water permeability data show a number of unusual features:
The average A and the GFD values for weeks 35-37 were
both much higher than would normally be expected. It
is believed that many leaking tubes went undetected
during that period. Well over forty known leaks were
isolated and it may be assumed that many others went
un-noticed. The log-log slope, its standard deviation
and the correlation coefficient, on the other hand, were
not abnormal. It is possible that this is an indication
that the membranes which were not leaking were becoming
fouled at a rapid but relatively constant rate.
-------
Table 2k. WATER PERMEABILITY DATA, AEROJET-GEKERAL
Week No.
2
3
h-6
7-13
35-3T
Wo. Data
Sets
8
If
15
25
9
Avg. A
x 105
1.235
1.213
1.217
1.290
2.891
Log-Log
Slope
+0.0117
-0.0023
+0.0096
+0.05^5
-0.011-22
Std. Dev,
Slope
0.0086
0.0123
0.0099
0.0152
0.0178
Correl.
Coeff.
.1(67
.132
.259
.600
.668
Avg.
G.F.D.
8.91
8.75
8.78
9.31
20.86
Avg. Effective
Op. Pressure (PSl)
^93
^97
501
5^7
508
vO
-------
70
The data correlation coefficients for weeks 3 and k-6 are
low. During these periods a portion of the final brine flow
was recycled to the feed inlet. This was done at the
suggestion, and under the direction of the manufacturer's
service representative in an attempt to improve the product
water recovery ratio which was then close to 0.30. In a
letter dated October 1, 1969, (which became part of the
rental agreement for this unit) Aerojet-General stated
that "sufficient feed pump capacity (would be provided)
to support a membrane flux of 20 gal./day sq. ft. of
membrane at a 90$ recovery factor." At the time this was
mistakenly assumed to be an implied estimate of the unit's
capability. In any event the actual recovery ratio never
coincided with the expected recovery ratio. Additionally,
the low correlation coefficients obtained suggest that the
recycled brine, because of an inadequate process piping
design, was not being uniformly blended with the fresh
feed flow before entering the unit.
Water Recovery and Total Rejection Ratios
The product water recovery and the total rejection ratios are shown
in Tables 25 through 29- The recovery ratios are the values at the
actual operating temperatures. These, rather than the data adjusted
to 25°C. (used in the permeability calculations), are required in the
determination of the material balance ratios.
Confidence levels for the water recovery ratios were calculated, when
possible, for each weekly group:
Table 25. AEROJET-GENERAL WATER RECOVERY DATA
Weekly
P er iod
2
3
k-6
7-13
35-37
Membrane
Set
1
1
1
1
2
Average
R ecovery
Ratio
.319
.306
.368
.k2k
.556
S tandard
Deviation
_
.118
.116
.(M
No. Of
Data Pts .
1
1
3
6
3
80$
C onf idence
Level
_
.1^ - .69
.18 - .67
.k7 - .6k
The wide variability of the data to week 13, as indicated by the broad
confidence levels, probably reflects the occurrence of the leaking
membranes mentioned previously.
Average Rejection and Material Balance Ratios
The calculated performance factors for the average rejection and material
-------
Table 26. pH ADJUSTED PEED WATER QUALITY, AEROJET-GEHERAL
Week
Nos.
2
3
h-6
7-13
2-13
35-37
2-37
Constituent Levels (mg/1 and micromhos)
T.D.S.
810.0
710.0
786.7
7^3.3
758.2
800.0
76^.2
Spec.
Cond.
1076
1135
1^00
1195
1235
1295
12^
Total
C.O.D.
28.5
25.3
20.5
8.5
36.7
7.1
15.6
Table 27. PRODUCT WATER QUALITY, AEROJET-GENERAL
Week
Nos.
2
3
k-6
7-13
2-13
35-37
2-37
Constituent Levels (mg/1 and micromhos)
T.D.S.
80.0
130.0
110. 0
190.0
152.7
110.0
157-9
Spec.
Cond.
367
173
180
201
208
230
211
Total
C.O.D.
3.0
^.5
5.7
1.9
3.2
—
3-3
-------
Table 28. BRINE QUALITY, AEROJET-GENERAL
Week
Nos.
2
3
4-6
7-12
2-12
35-37
2-37
Constituent Levels (mg/1 and microtnhos)
T.D.S.
-
-
1200
-
-
-
-
Spec. Cond.
1VT7
1670
1739
1786
1718
2359
1812
Ca
-
-
-
83
-
-
-
S\
-
-
-
436
-
-
-
Table 29. WATER RECOVERY AMD TOTAL REJECTION RATIOS, AEROJET-GENERAL
Week
Nos.
2
3
4-6
7-13
2-13
35-37
2-37
Water
Recovery
Ratio
0.319
0.306
0.368
0.424
0.388
0.556
0.424
Total Rejection Ratios
T.D.S.
0.901
0.817
0.860
0.7^4
0.799
0.862
0.793
Spec.
Cond.
0.659
0.848
0.872
0.833
0.832
0.822
0.830
Total
C.O.D.
0.905
0.822
0.722
0.776
0.913
—
0.789
-------
73
balance ratios are listed in Table 30. The inter-period variability
is minor and this suggests that the various changes in operating
conditions had relatively little effect on the membrane's selectivity
characteristics. It should be noted that the volume of data was
limited. There are two reasons for this: Until July 27, 1970 (week
22) only one chemist was available to make all the laboratory analyses,
and numerous shutdowns (resulting from membrane failure) disturbed
flow conditions to the extent that it limited the number of represent-
ative samples which could be taken.
Table 30. AVERAGE REJECTION AND MATERIAL BALANCE RATIOS, AEROJET-GENERAL
WEEK
NOS.
2
3
k-6
7-13
2-13
35-37
2-37
Avg. Rejection Ratios
T. D. S.
,900
=900
,900
SPEC.
COND.
.878
.85!*
.891
.850
.861
.875
.8611-
Material Balance Agreement Ratios
T. D. S.
1.365
1.365
1.365
SPEC.
COND.
1.C&5
1.116
1.062
.898
,968
e920
.957
Section VIII, (subsection "Reynolds Numbers") lists the Reynolds
numbers of four representative operating periods for the Aerojet-General
unit. Although the numbers appear to be sufficiently high to retard
scale formation on the membrane, the long unsupported tubes vibrated
badly as a result of flow turbulanee and pressure fluctuation. Higher
flow velocities probably would have accentuated this vibration, which
was thought to have influenced the incidence of ferrule and membrane
failures. It has been reported that Aerojet-General has now abandoned
the long tubular design.
Membrane Fouling and Cleansing
This unit was flushed with a cleansing mixture only three times. In
each case a 30,000 ppm solution of "Biz" adjusted to pH 8 was used.
The first two soakings lasted 15 minutes while the last was completed
in 20 minutes. The product flux increased each time, but the first
which was most the dramatic (57$) was partially due to increasing the
feed pressure from 500 to 6lO psi. The flux increases (without changing
the feed pressure) for the last two soakings were 19$ and 16$ respectively.
Any relationship that existed between the membrane A value and membrane
cleansing is depicted in Figure 13. Flushing data and bench sheets for
all units can be found in Appendix A-3.
-------
A Value - Time Plots
Three plots were prepared, using the data shown previously under
"Water Permeability", to illustrate the relationship of log A vs.
log time for the Aerojet-General unit (Figures 12, 13 and l4). The
regression lines in each were determined by modification of Equations
(2) and (3) to accomodate the absicssa scale.
Extended comments are unnecessary. The first two plots have positive
"b" value slopes while the third shows abnormally high A values. All
of these anomalies probably reflect the presence of severe membrane
leaks.
-------
WEEK NUMBER
1 T
7 8
i
Ul
I
z
u
o
X.
v>
z
< 0.6
O
-------
WEEK NUMBER
11 13
14
CTS
U
O
w
x
(/>
s
15
< 0.8
< 0.6
(/)
z
O
(J 0.4
<
cc
CQ
S.
UJ
S 0.2
A -INDICATES
MEMBRANE CLEA1MFD
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M.—SEC) —4.71
2. °/o RECOVERY— 44
3. OPERATING PRESSURE — S63p.s.i.
0.1
100 2
TIME(HOURS)
8 1000
8 10000
Figxxre 13. A vs. Time plotted logarithmically, Aerojet-General, k/ik/JO - 5/22/70
-------
10.0"
T~
35
WEEK NUMBER
36
37
O 4.0
X
3E
I
U
UJ
t/>
i
s
o
O
SLOPE= -
0.0422
o _ o
t.o
en
z
O
U
o:
m
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M.-SEC) —0.97
2. °/o RECOVERY — 56
3. OPERATING PRESSURE - 555 p.s.i.
0.10
10
6 B 100 2
TIME(HOURS)
e > 1000
6 8 10000
Figure 1^. A vs. Time plotted logarithmically, Aerojet-General, 10/27/70 - 11/12/70
-------
SECTION X
AMERICAN STANDARD (ABCOR) REVERSE OSMOSIS UNIT,
TUBULAR MEMBRANE DESIGN
Introduction
This unit was leased monthly from the Con Seps Department of American
Standard, Incorporated of Eightstown, New Jersey. On December 30, 1970
Abcor acquired the Con Seps Division of American Standard after which
the lease agreement was assigned to Abcor. The unit will be discussed
as the "American Standard" unit, as it was manufactured by that firm.
The nominal capacity of the Model TM 5-l4 unit, as first installed was
stated by the manufacturer to be ten thousand gallons per day with a
90$ removal of dissolved solids.
The nomenclature and equations used in this section are listed and
discussed in Sections V, VTII, and the Appendix.
Physical Configurations
The American Standard unit, (Figure 15) initially contained one hundred
vertically-positioned clear plastic tubular shrouds (or modules) each
with$3l6 stainless steel top and bottom .end connectors. A module
consisted of fourteen five foot long fibre-glass tubes. Each tube
contained three components: a cellulosic liner, the osmotic membrane
(seamless type) and an indeterminable number (perhaps 130) of spherical
ceramic turbulence promoters (Figure l6), each with a diameter of about
0.4 inches. The inside diameter of a tube was approximately 0.5 inches.
The modules had two types of stainless steel end connectors which permitted
either parallel or series flow through the tubes composing a module. The
flow arrangement was modified on a number of occasions when tube failures
occurred and replacements were not obtainable from the manufacturer.
The six major flow patterns are defined below using letter symbols as they
will appear in the tables. When the term "parallel" is used, it means
the flow through a module is divided equally among the first seven tubes
and then after combining the brines, the flow is distributed to the
remaining seven tubes in parallel. The term "series" denotes a module
as having all fourteen tubes serially connected.
"e" A flow scheme of three parallel rows (each row with thirteen
serially-connected "parallel" modules), followed by three
parallel rows (each with three serially-connected "parallel"
modules), followed by two parallel rows (each with ten serially-
connected "parallel" modules), followed by twenty-four serially-
connected "parallel" modules, followed finally by four parallel
rows (each with two serially-connected "series" modules).
Membrane set No. 1 was used, with replacements as available.
78
-------
79
Figure 15. American Standard (Abcor) reverse osmosis unit
Figure 16. Tubular
components, American
Standard (Ab c or)
R.O. unit
OSMOTIC
MUMGRANL
CCLLULOS1C L/fJC/2
-------
80
"f" Same as "e" but with the last eight modules removed.
Membrane set No. 1, with replacements, was used in this
flow configuration.
"g" Two parallel rows (each with thirteen series-connected
"parallel" modules followed by three parallel rows
(each with three series-connected "parallel" modules),
followed by two parallel rows (each with ten series-
connected "parallel" modules), followed by twenty-four
series-connected "parallel" modules, and terminated
by four parallel rows, each with two series-connected
"parallel" modules. Membrane set Wo. 1 was used in
this configuration.
"h,i" Heterogeneous flow patterns used in attempts to keep
the unit in service when many membrane failures (with
few available replacements) required the elimination
of modules, or the transfer of modules from one point
to another. Membrane set No. 1 was also used in these
configurations.
"j" Three parallel rows (each with thirteen series-connected
"parallel" modules) followed by two parallel rows (each
with thirteen series-connected "parallel" modules),
terminated by fourteen "parallel" modules in series.
Membrane set No. 2 was used for this configuration.
No attempt has been made in the above descriptions to assign membrane
areas to any of the above nominal flow configurations. Such estimates
would have been meaningless because membrane failures often required
the shifting of modules from one place to another or even the
complete elimination of various sections within the basic pattern.
Accurate records were kept, however, of the total membrane surface
area in use on any given day and these data formed the basis of the
water permeability information shown in Table 3^--
With the lack of consistent membrane areas and because the turbulence
promoters introduced unknown variables, Reynolds Numbers were not
estimated for the American Standard unit.
Membrane Specifications
The manufacturer stated that "...The original American Standard
membranes were subjected to a standard test with 5000 ppm NaCl solution
at 600 psi and 77°F...Flux rates ranged from 9 to 13 gpd/sq. ft. with
an average of around 10.5 gpd/sq. ft. Rejection was found to range
from 89 to 95$ NaCl rejection with an average of 91$..." These
membranes were termed AS-90+(90$ rejection) membranes. Some AS-197
membranes were installed early in October, 1970.
The second set of modules, installed by Abcor in early April, 1971,
were "similar" to the first except that the membranes were narrow
"Eastman No. KP-96" cellulose acetate sheets rolled into a tubular
shape with the longitudinal edges sealed. They were protected from
the resin impregnated fiberglass tubes by thin strips of Du Pont
"Re-May" paper. (American Standard had reportedly used "some"
-------
81
Eastman membranes of "... a different type and a better quality..."
in the earlier installation.)
Chronological Record
The following notes have teen extracted from the plant data log sheets
to show the major events and changes in operation during the study:
Week
11
3
5
to
18
28
31
.35
36
38
38
47
56
61
64 4
69 7
Data Groupings
1
7
1
7
5
1
6
4
1
3
3
Portion of equipment received in Hemet. Balance
delayed by labor problems not associated with either
Eastern Municipal Water District or American Standard.
Balance of major equipment received.
Start of data collection. Membrane Set Ho. 1; flow
pattern "e". Started unit with reactor-clarifier,
sand granular activated carbon, and D.E. filters in
operation with pre-R.O. unit chlorination and pH
adjustment of feed.
Removed about 8$ of the modules because of membrane
failures; flow pattern "f". Membranes in last section
became coated with tri-calcium ortho-phosphate scale
sometime between weeks 18 and 23.
Discontinued D.E. filtration of feed.
More membrane failures. Changed to flow pattern "g".
Discontinued reactor-clarification.
Changed to flow pattern "h".
Changed to flow pattern "i".
Out of service - feed pump bearing failures.
Resumed reactor-clarification of feed.
Many more membrane failures; no replacements available.
Removed unit from service.
Equipment lease agreement terminated by mutual consent.
Replacement membranes received. Set No. 2 installed;
flow pattern "j". Operated using sand filtration with
pre-R.O. unit chlorination, (also pH adjustment).
Discontinued sand filtration of feed. Operated on
pre-R.O. unit chlorinated, pH adjusted, secondary
effluent.
Stopped testing program.
The R.O. units used in this study were subjected to relatively stable
conditions between process changes. These process changes were mainly
limited to modification of flow patterns, module replacements and
variation in the degree of feed treatment. In the case of the American
Standard Unit, the conditions were often difficult to define. The unit
had numerous membrane failures and since replacements were often
unavailable, the flow pattern within the unit was seldom consistent for
any reasonable length of time.
-------
82
As individual modules were gradually removed from service or to otner
locations, it became necessary in some instances to by-pass a group^of
modules. When replacements were available, they were inserted^within
groups of older sets. Under these conditions, it should be evident
that a portion of the tubular data, computed or observed, may be. invalid.
The data time groupings for the American Standard unit were rather
arbitrarily separated into the weekly sets shown in Table 32.
Mechanical and Operational Problems
The operation of the American Standard unit presented problems in^four
major areas: the feed pump, the design features of the installation,
the membrane failure rate and the availability of replacement modules.
The feed pump was initially a duplex FWI (Frank Wheatley Industries,
Tulsa, Oklahoma) plunger pump. Within two weeks, it was replaced by
a FWI, P-220A, triplex plunger pump to eliminate backlash in the gears
of the U.S. Motors Vari-Drive. The second pump was shut down many
times for short periods to correct packing and bearing problems and
once, for about a week, to make a complete over-haul.
The basic design of the American Standard unit made it difficult to
replace or isolate leaking modules and to remove foreign material
(mostly turbulence promoters) in the product water lines and fittings.
Both American Standard and Abcor were frequently delinquent in providing
replacement modules as required, and this accounted for the unit being
out of service for one hundred days (weeks Vf to 6l) of the study.
Consequently, it was impossible to test the unit on reactor-clarified
sand filtered secondary effluent without carbon filtration.
During July and early August, 1970 (weeks 19 to 22) membrane deposits
of a scale, chemically identified as tri-calcium ortho phosphate, were
found in the last section of the unit. It is believed that the scale
deposits formed as a result of operating at nearly a 90$ product water
recovery. When the product water recovery was lowered to approximately
80$ and after cleansing the membrane with a 3.5 pH sulfuric acid/water
solution, the unit appeared to be free of the deposition problem.
There were 6868 available operating hours in weeks l4 to 46 and 6l to
69. The unit was on the line for 94.42$ of this time (6485 hours).
The major out-of-service periods were as follows:
Table 31. OUT-OF-SERVICE RECORD, AMERICAN STANDARD
Mechanical Problems
Membrane Cleaning
Membrane Failures
Alterations, Additions
Post-Treatment Problems
Total Down Time
Hours
158
32
162
11
20
383
%
2.30
.46
2.36
•17
.29
5.58
-------
Table 32. REVERSE OSMOSIS PROCESS INFORMATION, AMERICAN STANDARD
83
Week Nos.
14-18
19-22
^ ff
19-28
24-28
29-31
32-33
34-35
36
3T
38-40
4l— 46
l4— 46
61-63
64-69
61-69
Secondary
Post-Treatment
A,B,C,D,E,F
A,B,C,D,E,F
A,B,C,D,E,F
A,B,CA»
A,B,C,I,»
B,C,E,F
B,C,E,P
B,C,E,P
B,C,E,P
A,B,C,E,F
VARIOUS
B,E,F
E,P
VARIOUS
Membrane Set
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
Flow Pattern
e
f
f
f
f
g
g
h
i
i
i
VARIOUS
J
0
J
Post-Treatment Legend
A = Reactor-Clarifier
B = Sand Filters
C = Carbon Filters
D = D.E. Filters
E = Pre-R.O. Unit Chlorination
F = pH Control
-------
Qk
Table 33 shows the membrane failures according to time grouping:
Table 33. MEMBRANE FAILURE TABULATION, AMERICAN STANDARD
Weekly Group
Il).-l8
19-28
29-31
32-33
3*1-35
36
37
38-^0
1*1-1*6
61-63
61*-69
Number of Failures
5
13 (scaling problems)
5
Hi-
0
6
1
11
7
11
10
Total 89
Water Permeability Data
Inspection of data from Table 3!* reveals the log A vs log time slope for
weeks 19 to 28 to be quite different from that for weeks 19 to 22 and
2k to 28. There are two possible reasons for this difference. First,
the A value for weeks 19 to 22 was low because of calcium phosphate
build-up on the membrane. Second, the average A value for weeks 2k to 28
was unusually high because of one high daily value, (2.87*0 possibly
caused by an undetected leak in the membrane.
For weeks 1*1-1*6 the data correlation is very low. Log A versus log time
slope is nearly horizontal (tan = -O.OOOl*). There were a couple of causes
for these abnormalities of horizontal slope. An EDTA flush about midway
through the period was partially responsible for maintaining high A values.
Additionally, numerous unscheduled shut-downs relaxed the membranes.
Although never considered as a standard cleaning procedure, membrane
relaxation is thought to be helpful in maintaining cleaner membranes ;
through relaxation higher permeation rates and A values could have resulted,
Conceivably, the high A values and fluxes encountered after week 6l can
be correlated to the new membrane set put into service at week 6l and/or
to undetected leaks which occurred after week 6l. "New" membranes usually
show better fluxes and rejection ratios than "old" ones. Indeed, the
flux did increase, but the product quality and rejection ratios remained
about the same as in previous periods. This suggests the possibility
that undetected leaks contributed to the high fluxes especially if the
number of membrane failures is proportional to the number of undetected
leaks. From Table33* it can be seen that membrane failures occurred
at a rather constant rate throughout the American Standard study; assuming
the latter proportion exists, the quality of product would be maintained
-------
Table 314-. WATER PERMEABILITY DATA, AMERICAN STANDARD
Week
Nos .
14-18
19-22
19-28
24-28
29-31
32-33
34-35
36
37
38-40
4l-l»6
14-1*6
61-63
64-69
61-69
No. Data
Sets
IT
21
47
25
14
10
9
3
1
11
27
147
14
26
40
Avg. A
x lo5
1.796
1.394
1.547
1.621
1.533
1.559
1.477
1.559
l.it64
1.459
1.440
1.540
2.092
1.807
1.907
log -log
Slope
-0.1389
-0.0244
+0.0232
-0.0259
-0.0094
-0.0024
-0.0087
-0.0116
-0.0130
-0.0099
. -o.ooo4
-0.0627
-0.1034
-0.0669
-0.0920
Std. DBV.
Slope
0.0310
0.0135
0.0156
0.0043
0.0038
o.oo49
0.0050
0.0018
0.0038
0.0048
0.0041
0.0084
0.0119
0.0143
0.0129
Correl .
Coeff .
.757
.382
.216
.779
.582
.167
.549
.989
.839
.565
.022
.527
.928
.668
.756
Avg.
G.F.D.
12.96
10.06
11.16
11.70
11.06
11.25
10.66
11.25
10.56
10.53
10.39
ll.ll
15.09
13.04
13.76
Avg. Effective
Op. Pressure (P.S.I.)
409
491
455
444
538
499
505
508
513
498
505
479
523
4oo
c
429
-------
86
partially as a function of undetected membrane leaks. Thus, it is
reasonable to suspect both the "new" membrane and membrane leaks as
contributing to high fluxes after week 6l.
Water Recovery and Total Rejection Ratios
Feed,product and brine water quality plus product water recovery and
total rejection ratios for the American Standard unit are shown in
Tables 36 through 39 . Confidence levels for the recovery ratios can
be found in Table 35. The recovery ratios seem relatively consistent
until week 6k, three weeks after the second membrane set was installed,
but almost simultaneous with the application of untreated secondary
effluent.
Because of the reasonable interval between weeks 6l and 6k, the formulation
character of Membrane Set No. 2 was assumed to play a minor role in
reducing the recovery ratios after week 6k. Alternatively, the deteriorated
feed quality of secondary effluent could and presumably did magnify the
effects of fouling, giving lower recovery ratios and higher rejection
ratios. The slime growth as mentioned in Section VIII ("Membrane Fouling
and Cleansing") and particulate fouling was probably the major reason for
the rapid change in recovery ratios after week 6k.
Table 35 - WATER RECOVERY DATA CONFIDENCE LEVELS
Weekly
Period
14-18
19-28
14-28
29-31
19-31
32-33
29-33
34-35
38-40
34-40
ln-l»6
lk-k6
61-63
64-69
61-69
114-69
Average
Recovery
Ratio
.889
.805
.833
.739
.790
.765
.749
.754
.756
.774
.689
.782
.724
=527
.593
.741
S tandard
Deviation
.033
.051
.061
.020
.054
.016
.Oik
.026
.025
.052
.023
.072
.010
.oko
.10k
.111
No. Of
Data Fts.
5
10
15
3
13
2
5
2
3
7
6
33
3
6
9
k2
Range @ 80$
Confidence
Level
.Qk - .9k
.73 - .88
.75 - .92
.70 - .78
.72 - .86
.72 - .81
•73 - .77
.67 - .83
.71 - .80
.70 - .85
.66 - .72
.69 - .88
.70 - .74
™ i i
.47 _ .59
i * S S
.45 - .74
.60 - .89
-------
Table 36. pH ADJUSTED FEED WATER QUALITY, AMERICAN STANDARD
WESK
ros.
14-18
19.28
1..28
29-31
19-31
32-33
29-?l
34-35
36
37
33-40
34-40
41-46
14-46
61.63
64-69
61-69
14-69
T.D.S.
765-0
7W.5
753-2
740.0
746.5
715-0
T30.0
775-0
800.0
730.0
795-0
773-3
772.0
757-5
853-3
810.8
82J.O
773-1
SPEC.
COOT).
1235.U
Ilil3. 4
1370.k
1201-3
1361i.2
12611.7
1228-3
1271-3
1266.5
1384.6
1269.1*
1283.0
1233-5
1305-2
1313.0
201*8.6
1798-7
11*16.1
cON:;'fiTUmT
CHLORIDE
110.09
130.25
127-77
180.10
129.09
-
82.77
-
-
-
130-99
130.99
126.1*1*
127-1*0
155-80
138.39
171.87
133-21*
LEVtLS (ng/
TOTAL
C.O.D.
3.1*2
7-57
7-85
5-91*
7-19
5-85
5-89
10.25
-
8.70
14.62
11.1*1*
7-93
8.35
4.69
56.78
53-53
18.15
1 and mlcrc
DISS.
C.O.D.
7.80
5-1*7
6.06
l*.l*8
5-23
lt.l*8
4.43
6-33
-
5-20
6-56
5-72
3.61
5.1*8
-
32.08
32.08
9.89
mhos)
TOTAL
HAMHESS
133- !*9
201.51
189.26
201.23
200.37
211*. 23
205.92
213-97
235-1*2
222.22
283-33
222.82
233-49
207.76
229.62
232.38
238.18
215-1*1
OKTHO-P
-
4.79
4-79
7-91
5-55
6.71
7-50
10.1*2
-
14.95
14.04
13-49
7-46
7-43
12.08
12.00
12.22
8.65
TOTAL
ALKAL.
55-91
54.37
54-87
-
54-37
42.26
42.26
25-92
-
33-49
67.91
44.28
30.67
43-15
18.90
54.44
41.17
42.53
H03-H
3-90
4.18
4.14
4.90
4.28
1.30
3-10
-
-
-
-
-
-
3-91
-
3.81
3-81
3-87
SULMTE
387.26
341-17
348.93
373-29
350.30
346.15
359-85
398.50
-
363-63
383.18
385.71
370.23
365.08
444.67
379-09
416.67
378.18
Table 37- PRODUCT WATER QUALITY, AMERICAN STANDARD
VE3C
COS.
14-18
19-28
l«-23
29-31
19-31
32-33
29-31
34-35
36
37
38-40
34.40
41.46
14.46
61-63
64-69
61-69
14-69
T.D.E.
241.5
161.0
184.0
160.0
160.8
175-0
166.0
190.0
150.0
110.0
1JO.O
156.7
140.0
168.2
86-7
56-7
66-7
144.8
SPEC.
COXD.
527-0
367.5
420.7
276.3
346.5
322.5
294.8
314.0
461.0
270.0
298-3
322.0
265.2
352.4
131.3
147-5
142.1
307-3
CONSTITUENT
CHLORIDE
24.0
56.14
52.13
49.00
55-25
-
49.00
m
-
m
41.00
41.00
44.00
48.92
26.33
15-50
22.00
41.44
LEVELS (me
TOTAL
C.O.D.
3-64
2.34
2.77
1.30
2.10
2.10
1.62
1.65
-
4.10
4.65
3-34
0.92
2-37
1.50
3-35
2.73
2.45
/I and micrc
DISS.
C.O.D.
3.20
1-33
1.80
0.47
l.ll
1.65
0.94
0-95
-
1.40
2.40
1.62
0.65
1.49
•
3-08
3.08
1.76
ratios)
TOTAL
HARDNESS
57-00.
25-39
30.66
32.60
27.05
29-35
31.30
29.10
22.60
22.00
26.80
26.07
30.12
29-71
5-97
4.88
5,24
23-91
OKTHO-P
-
•91
•91
1.10
•96
•96
1.05
2.00
-
1.60
1.60
1.70
0.85
1.05
0.29
0.18
0.22
0.83
TOTAL
ALKAL.
20.80
18.65
19-37
-
18.65
19.10
19.10
11.95
-
14.20
14.60
13.46
12.30
15-36
13-53
12.25
12.68
14.46
H03-H
2.20
3.1fl
3-04
2.90
3-14
1.20
2.05
-
•
-
-
-
-
2.82
-
1.64
1.64
2-35
SULFATE
146.00
29.00
45-71
54.50
35-38
40.JO
47.50
53-00
-
28.00
41.00
43.20
48.50
46.00
6.67
4.17
5.00
33-28
-------
Table 38. BRINE QUALITY, AMERICAN STANDARD UNIT
WEEK
JI'.3.
i*-w
19-23
l*-23
29-31
19-31
32-33
. 89-33
3*-35
36
37
38-40
3*-*P
M-*6
l*-*6
£l-£3
6*-69
61-69
ll>-69
T.D.S.
.
-
-
3092
2996
2303
2798
2360
.
2818
2625
2732
220*
2633
2673
1705
2070
32*5
SPEC*
c«n>.
9731
7967
8756
Ii659
6375
1.266
1*1.91*
3935
5367
*395
3726
*097
2970
5528
3690
2691.
30*2
1*938
ag/1 exce
CHLORIDE
-
•
-
1*61.
500
-
*6*
-
-
-
387
387
266
*3S
357
281*
339
-
pt Bpee. eon
TOTAL
C.O.D.
-
69.0
69.0
25.3
31.*
10.9
19.6
35-5
-
30.0
25-7
29.7
21.1
27.2
100
87.2
90.9
83.3
d. as mlcr
TOTAL
HARD-
NESS
.
-
-
81.7
812
51*0
69*
795
-
857
93*
875
671
7*7
77*
51*
611
522
oofeoi *.
ORTKO-P
.
-
-
32.2
71.5
SU.7
3*.0
52.0
-
57.5
51-3
53.5
26.3
52.*
*6.3
25.*
33.3
23.0
ALKA-
LINITY
-
-
-
*9.0
-
29.0
1*2.0
6.8
-
306
156
125
7*0
93.0
39.0
1*1. .1
*1.5
82.0
CALCIUM
-
-
-
22*
222
197
215
•
-
-
.
-
-
217
-
210
210
210
K03-S
6.5
6.5
2.7
*.2
1.7
2.*
0.1
-
-
2.8
8.8
1*
-
-
6.U
6.*
*.7
Table 39. WATER RECOVERY AND TOTAL REJECTION RATIOS, AMERICAN STANDARD
VS3C
::os.
1*-18
19-28
l*-26
29-31
19-31
32-33
29-33
3*-35
36
37
38.1*0
3*-*0
1*1-1*6
H.-U6
61-63
fife-£9
61-6?
ll*-£9
WATE3
RK.-OVEK
RATIO
.889
.805
.833
•739
•790
•765
•7*9
•75*
.825
.819
• 756
•77*
.689
.782
..72*
•527
•593
.7*1
T.D.S.
.68*
•785
•755
.78*
•785
•755
•773
•755
.812
.81.9
.811
•799
.818
• 778
.898
•930
•919
.813
SPEC.
COND.
•590
.7*0
• 693
•770
•7*6
•7*5
.760
•753
.636
.805
•765
•750
• 785
• 730
•900
.928
.921
•783
CHLORIDE
.782
• 569
•592
•592
•572
-
•592
-
-
-
.687
.687
.652
.616
.831
.888
.852
.689
TOTAL REJECT
TOTAL
C.O.D.
.568
.691
.6*7
.781
.708
.6*1
•725
•839
-
• 529
.682
.708
.88*
.716
.968
.9*1
•9*9
.865
ION RATIOS
mss.
C.O.D.
•590
•757
•703
•895
.786
.629
.788
.850
-
•731
.63*
• 717
.820
.728
-
.901.
.901.
.822
TOTAL
HARDNESS
•573
.87*
.838
.838
.865
.863
.8*8
.86*
.90*
•901
.880
.883
.871
.857
•97*
•979
•978
.889
OHEKO-F
-
.810
.810
.861
.827
.857
.860
.818
-
•893
.886
.87*
.886
•857
•976
•985
•982
.901.
TOTAL
ALKAL.
.628
.657
.6*7
-
•657
.5*8
.5*8
•539
-
•576
•785
.696
• 599
.6**
.28*
•775
.692
.660
K03-H
.*36
•239
.266
.1*08
.267
.077
•339
-
-
-
*
-
-
.278
«t
• 570
•5TO
•393
SULFATE
.623
•915
.86?
.85*
.899
.883
.863
.867
-
•923
•893
.888
.869
.87*
•985
•989
.988
•912
-------
Accompanying these lower recovery ratios and higher rejection ratios
was the side benefit of greater brine flow rates which in turn gave
Reynolds numbers sufficient to maintain the membrane surface
relatively free of deposited "material.
Average Rejection and Material Balance Ratios
The average rejection and material balance ratios are shown in Table 4l.
The values were calculated from Equations (7) and (l^). The average
rejection ratios are higher after week 6l. The material balance ratios
are normally within an acceptable plus or minus 10$ except for total
C.O.D. It is evident that some of the latter analyses are in error.
Membrane Fouling and Gleaning Procedures
American Standard membranes were flushed twenty-two times with a Biz
solution, three times with a weak sulfuric acid water solution and twice
with EDTA solution. Some pertinent information relative to membrane
cleaning is given in Table ^Q, while summaries of the bench sheets of
membrane cleansing procedures can be found in Appendix A-3.
Table 40. MEMBRANE CLEANSING- HISTORY AND PRODUCT FLUX INCREASES
AMERICAN STANDARD
Weekly
Group
1^-18
19-22
2h-2Q
23-31
3^-35
38-40
lH-ltf
61-63
6^-69
Total No.
and Per Cent
Biz
Ho, Times ($)
1
3
2
1
1
1
1
k
8
22
8.7
k.6
5-5
3.0
_
_
5-6
23-3
11.7
Acid Wash
Ho. Times (%)
_
_
_
-
—
_
3
3
_
_
_
-
_
_
3-9
-
3.9
EDTA
Ho. Times ($)
_
1
-
-
_
1
_
-
2
_
22.2
_
-
_
_
_
-
22.2
A Value - Time Plots
Two plots prepared (Figures 17 and 18) using the original plant data
are summarized in Table 34.
Figure 17 > for weekly groups ik to ^6, shows the effect of the phosphate
scale formation discussed previously and the A value recovery after
corrective measures were taken. Figure 18, shows data for weeks 6l to
69, during which the unit feed was untreated secondary effluent
except for pre-R.O. unit chlorination and pH adjustment.
-------
Table 4l. AVERAGE REJECTION AND MATERIAL BALANCE RATIOS, AMERICAN STANDARD UNIT
vo
o
Week
Nos.
14-18
19-28
14-28
29-31
19-31
32-33
29-33
34-35
36
37
38-40
34-40
41-46
14-1)6
61-63
64-69
61-69
1*1-69
Average Rejection Ratios
T.D.S.
mm
.925
.925
•915
.920
.882
.902
.89^
.926
.935
.918
.914
.904
.910
.952
•955
.954
.924
Spec.
Cond.
.900
.903
.902
.892
.900
.871
.883
•8T5
.846
.901
.888
.880
.877
.890
.948
.951
•950
.903
Total
C.O.D.
—
.890
.890
.918
.897
.815
.877
.928
_
.790
.776
.840
.935
.886
.975
.953
.959
.904
Total
Hard-
Ness
—
.959
.959
•931
.945
.942
.936
.942
-
.959
•951
.949
.935
.942
.988
.987
.987
.958
Ortho-P
—
.936
• 936
.9te
•938
.941
.942
.936
-
.958
.954
.950
.942
.942
.990
.991
.990
.957
NO -N
3
—
.405
.405
.567
.459
.368
,l»68
-
-
-
_
-
-
.436
-
.662
.662
.572
Material Balance Agreement Ratios
T.D.S.
.
1.079
1.079
1.222
1.151
.924
1.103
1.081
.870
.785
••
1.081
.997
.989
1.036
•972
1.032
1.012
1.028
Spec.
Cond.
1.148
1.051
1.083
1.016
1.043
.885
.963
.910
.951
.696
.925
.892
.930
.997
.816
.958
.910
.978
Total
C.O.D.
.
1.325
1.325
1.294
1.317
.884
1.113
1.085
_
1.017
•774
.947
.961
1.130
.547
.776
.719
1.027
Total
Hard-
Ness
.
1.050
1.050
1.328
1.189
.973
1.151
1.008
_
.781
1.101
1.000
1.011
1.045
.950
1.034
1.006
1.032
Ortho-P
„
1.267
1.267
1.084
1.198
•938
1.026
1.383
.837
1.095
1.102
1.106
1.125
1.040
1.050
1.046
1.101
NO -N
3
1.038
1.038
.888
.988
1.147
1.017
_
_
-
—
_
_
1.028
-
1.134
1.134
1.092
-------
WEEK NUMBER
15
15
15 16
16
16
17
18 19
26
20
38 50 _
I
u
LJ
tn
i
5
u
o
0.8-
z
<
t-
(/)
z
o
u
o:
CD
PHOSPHATE SCALE FORMATION
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M.-SEC) —4.94
2. % WATER RECOVERY- 78
3. AVERAGE OPERATING PRESSURE- S04 p. s.i.
O.I
8 100
I 1000
TIME(HOURS)
Figure 17. A vs. Time plotted logarithmically, American Standard, 4/26/71 - 6/25/71
8 10000
VO
H
-------
WEEK NUMBER
u
LU
tn
i
5
O
o
(/)
x
to
< 0.8-
•z.
< 0.6
I-
to
z
o
U
UJ
cc
CD
0.2
0.1
61
61
61 61
63 64 65
62
66
AVERAGE CONDITIONS:
1. FLUX (GIW/SQ.M.-SEC) —5.61
2- % WATER RECOVERY — 63
3. OPERATING PRESSURE — 445 p. s. i.
100 2
Tl ME(HouRs)
6 8 1000
a 10000
Figure 18. A vs. Time plotted logarithmically, American Standard, 6/3/70 - 1/15/71
-------
SECTION XI
E. I. DU PONT DE NEMOURS & CO. UNIT
HOLLOW FIBER CONCEPT
Introduction
One unit used in this study was the Du Pont "Permasep"V^ Pilot Plant
furnished complete with permeators, pump and instrumentation. The
unit capacity was rated at 10,000 gal./day with a 75 to 90$ product
water recovery at 600 psi. Figure 19 shows the unit as installed.
Figure 19- Du Pont unit with
B-5 permeators in place
Physical Configuration of Modules
Two different "Permasep" modules were tested. The first, called a B-5
type, was a 15 in. O.D. by 10.5 ft. dished and flanged cylinder enclosing
a draped set of hollow nylon fibers (see next sub-section for their
description). The nylon fibers were suspended from an epoxy tube sheet
and confined by a nylon net wrapping. Within the bundle was a vertical,
centrally-positioned feed tube having numerous 100 micron (0.01 cm)
flow distribution holes.
Midway through the testing period the two B-5 permeators were replaced by
five smaller modules. These modules, called the B-9 type, were small 6063-T6
horizontal aluminum cylinders, about 5-5 in. O.D. by kj in. long.
93
-------
94
Each permeator type used the same basic test unit package, consisting of
the equipment shown in Figure 20 and described in Table 42.
In both cases, the feed enters the permeator casing or shell and
contacts the bundle of hollow membrane fibers. A portion of the water
permeates the fibers;the rest is rejected as brine.
Figures 21 and 22 show the general flow patterns through each of the
module types.
The two B-5 modules were operated in series, with the brine from the
first being used as the feed for the second. The product flows were
combined to give an average flow and product quality. This config-
uration was designated as flow pattern "k".
Two module configurations were tested using B-9 modules. The first,
designated as the "t" pattern, consisted of three modules in parallel
followed by two in parallel. The "u" pattern consisted of five modules
in parallel and was adopted (at Du Font's suggestion) with the
expectation that the product recovery might be improved, since each
permeator would receive the same high inlet pressure. The results,
which will be discussed in subsequent sub-sections, were not encouraging.
In order to equalize the input flow to all five modules, flow constrictors
(balancing "venturies") were placed at each brine port.
The venturies were stainless steel tubes, about 3-75 cm long by 0.1 cm
I.D. which occasionally became partially clogged. This condition was
accompanied by a substantial reduction in brine flow. If undetected
and uncorrected, the problem rapidly became worse, which introduced
still other problems.
With a restricted flow, the introduction of cleansing solutions was a
slow and difficult process. It was also necessary to estimate brine
pressure within the modules by calculation, rather than by average.
Such estimates, however, could not be confirmed readily. It was
eventually realized that the fouling tendency, as a function of lower
flow rates per module, was much greater in the "u" pattern (5 in
parallel) than in the "t" pattern.
Membrane Specifications
The B-5 modules contained hollow, semi-porous formic acid-treated nylon
(perhaps the 6.6 type) non-asymmetric fibers with the Du Pont name
"ZYTEL". The fibers-were approximately 25 microns I.D. and 50 microns
OoD. with total membrane surface area for the two permeator modules
estimated by the manufacturer at 160,900 sq.. ft.
The B-9 modules contained hollow formic acid treated asymetric
aromatic polyamide fibers (Du Pont "NOMEX") having the dimensions of
40 microns I.D. and 80 microns O.D. The actual surface area of the B-9
modules varied as shells (modules) were replaced (see sub-sect!on
"Chronological Record"). The manufacturer's estimated areas for the
four modifications are listed in Table 43.
-------
95
-PERMEATE
-BRINE
BRINE
CONDUCTANCE
iFLOW
JMETER
FLOW
METER
PROD.
CONDUCT.
Q CONDUCT.
TEMP.
COMPENSATOR
X BRINE
T PRESS.
SAMPLE
TAP
SAMPLE
TAP
BACK PRESS.
REGULATOR VALVE
FEED fx
TEMP. IA
CARTRIDGE
FILTERS
NITROGEN
SUPPLY
FEED
HYDRAULIC
ACCUMULATOR
-M
1
PUM
PR
Q
P Sl
ESSU
5
J
PERMEATE
PORT
H
PERMEATOR
FEED
PRESSURE
Hh-LO
PRESSURE
SWITCH
Figure 20. iV Permasep pilot plant flov diagram (Courtesy, Du Pont)
-------
V.R)
Table k2. EQUIPMENT DESCRIPTION, DU PONT PERMASEP PACKAGE
Item
Description
Pump
Hydraulic Accumulator
Flow Meter
Pressure Gauge
Pressure Gauge
Filter
Temperature Gauge
Back Pressure Regulator
Positive displacement, reciprocating,
triplex, Armco Model J-231-L.
Greer 3,000 psi bladder accumulators for
•water service. Model 30A-IWS.
Fisher and Porter Model No. 10A2735A.
Ashcroft Maxisafe Gauge Type 1020P, 0-100
psig range.
Ashcroft Maxisafe Gauge Type 1377TAS,
0-1500 psig.
Cuno Micro Klean Fiber Cartridge Filter
316, Model 3AxBl with 5 micron wool
cartridges.
Ashcroft Dial Thermometer Cat. No. 2-6360BH,
0-100° C.
Maratta Back Pressure Regulator Model TRY 533-1A,
Part No. 806325.
-------
97
.FEED
PERMEATE!
Figure 21. Simplied internal flow scheme, B-5 module
END PLATE
SNAP RING
PERMEATE
FIBER
SHELL
CONCENTRATE
•0' RING SEAL
POROUS FEED END PLATE
DISTRIBUTOR TUBE
Figure 22. Cut away drawing of B-9 permeator, (Courtesy, Du Pont)
-------
98
Table ^3. ESTIMATED MEMBRANE SURFACE AREA, DU PONT B-9's
Week of Modification Estimated Surface Area
38 7,^8 sq. ft.
53 8, tell- sq. ft.
58 8,196 sq. ft.
63 7,920 sq. ft.
These areas were used in calculating the A values.
The manufacturer also stated that the B-9 modules would probably
provide better flow distribution, permeability, less fouling tendency
and significantly better solute rejection at a lower membrane cost
per gallon than the B-5 modules. This claim will be commented on
later.
Chronological Record
The following notes, abstracted from the plant data logs, list the
major events and changes in operation for the Du Pont unit:
Week and Day
k 7 Placed in operation two B-5 permeators (Set 1A)
on flow pattern "k". Ran on feed of reactor-
clarified, sand and carbon filtered secondary
effluent (with pre-R.O. unit chlorination and
pH control).
7 5 Added D.E. filtration to feed treatment sequence.
18 6 Replaced the Wo. 2 B-5 permeator.
2k 3 Feed treatment sequence reduced to reactor-
clarification plus sand and activated carbon
filtration.
33 7 Discontinued reactor-clarification of feed.
38 5 Replaced both B-5 permeators with five B-9 modules
(Set IIA) using flow pattern "t".
lH k After carbon filtration, add "pre-R.O. unit
chlorination and pH adjustment.
^9 ^ Discontinued carbon filtration.
53 5 Replaced entire set of B-9 modules with new one
(set IIB); initiated flow pattern "u". Operated
temporarily using potable Colorado River water.
Installed brine port venturies.
5^ 6 Resumed study using reactor-clarified, sand filtered,
chlorinated, pH adjusted secondary effluent.
57 1 Reduced treatment sequence to sand filtration only
(plus pre-R.O. unit chlorination and pH adjustment).
58 5 Replaced one permeator with B-9 module from used set IIA.
63 7 Replaced a second permeator with another module from
set IIA.
-------
99
614- 3 Operated using untreated secondary effluent
(pre-R.O. unit chlorination and pH adjustment excepted).
6k 5 Removed venturies from brine ports.
66 k Reduced product recovery; increased brine flow rate 100$.
67 2 Discontinued testing program.
Data Groupings
Twelve time periods were selected for comparison as shown in Table kk
with four additional consolidated groupings (weeks 8-23, ^-37 > 38-53,
55-66) to facilitate comparison between specific and mixed process
conditions.
Log A vs log time plots prepared for three of these groupings can be
found in the latter part of this section.
Mechanical and Modular Problems
®
The Permasep package experienced relatively few mechanical problems
during the fourteen months it was tested. The mechanical problems
consisted of a broken drive belt on the feed pump (week 51), alleged
leaks in the epoxy tube sheets (set IIA), and occasional malfunctions
in the high-low pressure cut-out system. Most of the B-5 modular problems
seemed to arise from one basic shortcoming in the unit design. Because
the feed solution percolated both radially and longitudinally through the
fiber bundle, it was hypothesized that there were regional flow differences
within each permeator. It is assumed that the B-5 type, because of its
longer length, was affected by regional flow differences more than the
B-9 module.
Observations made during the testing period seem to support the poor flow
distribution hypothesis. The second (downstream) series B-5 module
suffered a permanent flux decline about week 16 in spite of Du Pont's
prompt assistance and the use of a variety of cleansing solutions. While
other R.O. makes required short periods of soaking (15-60 minutes) and
flushing to restore normal fluxes, the Du Pont modules often required
12 to 2k hours of soaking for effective membrane rejuvenation. If the
flow distribution was not uniform, it is likely that particulate solids
became trapped within the bundle of fibers and prompted the longer soaking
periods. Unlike the tubular cellulose acetate membranes, which tolerated
feeds with moderate amounts of particulate solids, Du Pont designated that
the feed to its unit pass first through Cuno 10 micron filters. After some
major post secondary treatment processes were removed, a set of ^0-50 micron
filters were installed ahead of the 10 micron size to eliminate serious
clogging of the fine mesh filters. Not only were the B-9 modules better
able to handle particulate solids, but the unit was smaller and easier to
handle and most importantly, the cost of producing a gallon of product was
reduced (see Section XVI).
Between weeks k and 66, there were 10,372 available operating hours. The
Du Pont unit operated for 95^3 (92.01$) of these hours. The out-of-service
record for the Du Pont unit is shown in Table k^.
-------
Table
REVERSE OSMOSIS PROCESS INFORMATION, DU PONT
Week Nos .
8-18
19-23
8-23
24-33
34-37
4-37
38-41
42 -48
49-53
38-53
5^
55-57
58-63
64-66
55-66
(A=Reactor Clarifier)
(B=Sand Filters )
(C=Carbon Filters )
(D=D.E. Filters )
(E=Pre-R.O. Unit
Chlorination )
(F=pH Control )
A,B,C,K,P
FULL (A,B,C,D,E,F)
FULL
FULL
A,B,C,E,F
B,C,E,F
VARIOUS
B,C,E,F
A,B,C,E,F
A,B,E,F
VARIOUS
NONE
A,B,E,F
B,E,F
E,F
VARIOUS
Membrane Set
IA
IA
IA
IA
IA
IA
IA
IIA
IIA
IIA
IIA
IIB
IIB
IIB
IIB
IIB
Flow Pattern
k
k
k
k
k
k
k
t
t
t
t
u
u
u
u
u
Special Conditions
-
Permeator Replacement
-
-
_
_
-
_
-
_
COLORADO RIVER WATER TEST
-
Permeator Replacement
Permeator Replacement
-
H
o
o
-------
Table 45. OUT-OF-SERVICE RECORD, DU POET
101
Mechanical problems
Membrane cleaning
Membrane failures
Alterations, additions
Post-Secondary Trtmt. Prob.
Total down time
Hours
B-5 &
243
Of
243
B-9
205
-
205
Total
163
443
116
102
829
Total
%
1.57
4.32
1.12
.98
7-99
Water Permeability Data
The average A values and GFD (gallon/sq.ft.-day) ratios shown in Table
46 should not be compared with similar values given for other units.
In the hollow fiber concept, low flux (gin/cm2*sec) is offset by greater
available membrane area.
The thirty-fold difference between B-5 and B-9 average A values
illustrates why the Du Pont values are almost useless for comparison
with those of other units.
The utility of the A values lies with how they vary with time as in the
log A-log Time Plots. In Table 46 the slopes for B-5 and B-9 values
are similar. There was only one period (weeks 24-33) when the slope
turned positive, which in turn gave a low data correlation coeffient.
The rejuvenation log sheets show that the No. 2 B-5 permeator was
flushed four times with EDTA during that period. If a correlation
exists between EDTA flushes and A values, then it can be assumed that
there was significant scale buildup before the flushes. There seems to
be no identifiable relationship between flux decline and post secondary
effluent treatment until about week 58, when reactor-clarification was
terminated and sand filtration only was used in post treatment of the
secondary effluent. The flux then declined 40 per cent in five weeks.
The flux declined 65 per cent while operating on minimally-treated
feed (secondary effluent with chlorination and pH control) between
weeks 6k and 66.
Water Recovery and Total Rejection Ratios
Tables 47, 48 and 49 show the feed product and brine water constituent
concentrations. The Du Pont product water recovery ratios shown in
Tables 50 and 51, are not temperature corrected, as opposed to the A
value computations for which it is necessary to correct for temperature,
The time periods are split to show differences (if any) between
the B-5 and B-9 performances. Standard deviations for the recovery
ratios and the estimated confidence levels are shown in Table 51.
-------
8
Table 46.
WATER PERMEABILITY DATA, DU POUT
Week Nos.
4-7
8-18
19-23
8-23
24-33
34-37
4-37
38-4l
42-48
49-53
38-53
54
55-57
58-63
64-66
55-66
No. Data
Sets
14
50
25
75
47
21
157
13
38
19
70
6
16
29
13
59
Avg. A
x 105
0.008457
0.007323
0.006711
0.007119
0.006038
0.005974
0.006762
0.302159
0.261217
0.219444
0.257481
0.255550
0.158479
0.111664
0.062497
0.113098
Log-Log
Slope
-0.0118
-0.0914
-0.0491
-0.0821
•K). 0026
-0.0069
-0.0982
-0.0459
-0.0179
-0.0060
-0.0813
-0.0200
-0.0875
-0.0780
-0.0270
-0.2250
Std. Dev.
Slope
0.00760
0.01426
0.00661
0.01069
0.00250
0.00282
0.00692
0.00831
0.00386
0.00551
0.00626
0.00766
0.00813
0.01528
0.01147
0.02472
Correl .
Cosff.
.410
.679
.840
.668
• 153
.491
.752
.858
.612
•255
.844
• 794
.945
.700
• 579
.770
Avg.
G.F.D.
0.06102
0.05284
0.04842
0,05136
0.04356
0.04310
0.04879
2.180
1.885
1.583
1.858
1.844
1.14s
0.8057
0.4509
0.8l6o
Average Effective
Operating pressure (P..S.I.
462
5^3
552
530
560
552
5^3
379
363
336
358
382
378
367
365
370
-------
103
Table lj-7. pH ADJUSTED FEED WATER QUALITY, DU PONT
WEEK
HOS.
k-7
8-18
19-83
8-23
2l*-33
3"*-37
*-37
35-M
te-td
"•9-53
38-53
54
55-57
58-63
6k-66
55-«8
l*-66
T.D.S.
780.0
686 .It
768.0
711.9
729.5
761.2
730.9
815.0
769.2
813.0
796.0
821.0
780.0
805.8
8UO.O
807.9
763.8
SPEC.
CQND.
1222.2
1267.9
1256.3
1261* .5
1250,1
121*6.6
1253.5
123>*.9
1215.6
1216.6
1218.1
1291.7
12W.1
128J.O
2020.9
1*59.1
1285.2
>
CHLORIDE
—
110.09
121.01
118.79
137-01*
-
126.86
130.89
129.92
161.35
H*3-2l*
83.33
137.1*6
150.33
1W.87
3A5.W
137.69
8/1 excep
TOTAL
C.O.D.
12.1*0
8.1)8
9.88
8.T2
5.87
9.08
8.09
12.61
7.59
3U.88
15.50
10.91
3S-30
ln.93
1*9-13
"13.30
16.61.
E ipec* ec
DISS.
C.O.D.
13.60
8.39
6.55
7.W
"*.57
5.32
6.28
6.32
-
1.87
8.10
10.00
23.68
28.18
31.90
27.88
8.61
ind. as mi
TOTAL
HARD-
NESS
-
208.33
211.11
210.31*
200.11
222.01
207.81
23"*.8l
236.98
231.00
237.29
1(09.09
229.06
223.56
23"*-15
226:i8
223.87
eronhoe
ORTHO-P
_
-
3.59
3.59
6.99
13.68
7.36
1U.13
7.2>*
8.07
9.25
-
8.96
13.18
12.56
11.99
9.07
ALKA-
IIHITY
.
60.33
1*3.66
52.01
32.10
37.25
ItS. 63
93.80
70.19
69.29
75.81
1*7.02
68.87
66.67
98.00
T5.<*
67.18
CALCHM
.
-
58.83
58.88
52.73
-
5"*."*9 .
79-5"*
72.58
6k.26
67.69
85.71
61.51*
60.71
6I*.12
61.88
63.36
HOj-H
.
5.72
3-97
V.9T
5.50
3.20
U.90
0.1»5
5.19
8.72
6.01*
0.1*9
6.61
.3.30
5.86
5.1*
5.11
«v
.
295 .IB
223.12
237.8I>
328.58
385 -5U
312.95
330.19
335.63
323-58
329.33
1*16.66
339.80
308.33
337.58
321."i3
38l*.32
Table W. PRODUCT WATER QUALITY, DU PONT
WEEK
SOS.
"»-7
8-18
19-23
8-23
2M3
3^-37
fc-37
'38-1*1
1.2J»8
ty-53
38-53
5*
55-57
58-63
ft-66
55-66
l*-66
" T.D.S.
277-5
309.8
269.0
297-1
296.0
265.0
290.7
55-0
87-5
110.0
86.3
1*3-0
85.0
126.7
196.7
133-8
207-0
SPEC.
COND.
529.2
651.7
51*9-0
619-6
5*6.3
5W.5
579-1
102-5
178.7
205.6
168.1
93-0
132-3
257-0
608.3
313-7
1*16.1*
CHLORIDE
.
2* -00
102.25
86.60
115.25
-
99-33
7-33
15-33
22-75
15-90
11.00
15-67
23.00
1*6.00
23-13
M*.6l
og/l ex
TOTAL
C.O.D.
8.10
*-53
1*.88
lt.6b
3-U.
l*.07
*-31
4.27
0-77
1-50
2.01
U.80
5-30
2.60
7-*7
M3
3-79
.cept spec.
nrss.
C.O.D.
11.20
3-92
3-10
3-51
1-27
1.68
2.68
1.10
0.00
1.80
0.73
0.90
0.90
3.10
5.20
3-06
2.27
sond. as m
TOTAL
EASD-
NESS
-
1*0.00
53-20
I*9.W
36.82
35-30
W-73
7-28
17-30
23-10
16.61
l*-50
7-33
16.32
28.80
17-19
26.61*
icroobos
OBTHO-P
-
-
1.08
1.08
2.05
3-67
2.11
O.H
0.62
0.88
0.66
0.01
O.i»9
1.81
1.86
1.U6
1-33
'SOESL
AIKA-
Lrmw
-
21.90
11.70
16.80
16.85
23-17
1B-93
37-52
29-1*1
38-9*
Ht.ua
15.00
33-*7
35-27
U5.U7
37-37
31-17
CALCIUM
-
-
13-*3
13-W
8.28
-
9-86
7.00
1*.50
6.01*
5-96
0.60
1.60
*.25
6-73
l*.21
6.70
H03-S
-
3-80
3-85
3-53
5-15
1.6o
3-65
Q-07
0.61*
1.70
1.01
0.01*
1-50
0-95
2.U3
1.78
2.32
S0»
.
1*9.00
20.75
26.1*0
50.00
32.00
38.18
17-50
29.20
39-80
29-61*
10.00
17-33
37-00
79-67
"•2-75
36.00
-------
Table
BRINE QUALITY, DU PONT UNIT
WEEK
•OS.
M
8-18
19-23
8U-33
3">-37
*-37
38-41
te-48
•9-53
38-53
5*
55-57
58-63
64-66
55-66
4-66
1.0.3.
-
•
-
2158
23U
2375
2325
2055
1677
2026
2590
1903
1502
1103
1502
1950
SPEC.
COND.
2781
3797
4778
3802
3166
3779
3500
2988
2419
2937
3097
2833
2296
1791
2330
3271
m
CHLORIDE
-
-
202
222
•
222
449
370
278
374
334
321
234
203
263
295
5/1 except
TOTAL
C.O.D.
.
-
56.7
16.1
22.5
28.1
22.3
23.1
63.0
29.2
35-0
59.0
70.5
59-3
64.9
37.2
spec. eond.
TOOL
HARD-
NESS
-
-
-
717
761
73*
834
674
51"*
674
1272
567
456
358
459
641
a< aleroah
ORTHO-P
-
-
67.0
46.6
48.0
48.2
47.0
20.2
17.0 •
-
0.0
20.0
24.5
20.0
22.3
31.6
as
ALKA-
LINITY
-
-
-
94.8
39-5
70.2
220
184
140
179
70.0
57.7
136
71
100
127
CALCIUM
•
-
-
193
193
216
203
138
172
293
15*
127
.
141
174
IK>rH
-
-
5-1
1.8
2-3
2.7
6.4
10.3
lfl.9
11.4
1.6
lfl.0
9-7
8.1
10.4
5-7
SOU
-
-
-
-
-
-
800
872
716
756
1481
828
584
-
735
802
Table 50. WATER RECOVERY AMD TOTAL REJECTION RATIOS, DU POUT
WEEK
•OS.
/ *-7
8-18
5 19-23
8-23
24-33
1 3U-37
\ 4-37
/ 38-41
42.43
38-53
9 54
55-57
58-63
64-66
\ 55-66
4-66
WATER
BSCOVERf
RATIO
.704
•793
•T99
•795
•770
•741
•771
.684
.673
• 565
.642
.703
.566
• 515
•J51
.437
.632
T.D.S.
.644
•549
.650
.583
•594
.652
.602
•933
.886
.865
.892
.946
.891
.843
•766
.834
.W9
SPEC.
CONS'
• 567
.486
•563
.510
• 563
.560
• 538
•917
•853
.831
.862
.928
.894
.800
.on
•785
.w.
CHLORIDE
-
.782
•15?
•271
•159
-
.217
.944
.882 .
•359
.889
.866
.836
.847
.691
.841
.676
TOTAL
C.O.D.
• 347
•«5
.474
.463
.470
•551
.468
.661
.899
•957
.870
•560
.854
•938
.848
•900
.77*
mss.
C.O.D.
.176
• 533
•527
• 531
•723
.685
• 573
.826
-
•931
•910
•910
.962
.890
•837
.890
.736
TOTAL
HARD-
NESS
-
.808
.748
•765
.810
.841
.804
•969
•927
•900'
•930
•989
•963
•927
•876
.924
.681
OREHO-P
-
-
•699
•699
•707
•732
•714
•969
•914
.891
•929
-
•945
.863
.852
.878
.847
ALKA-
LIBTH
.
.636
• 732
.677
.475
•378
•556
.600
• 581
.438
•5*6
.681
.514
•471
•536
• 502
• 535
CALCIUM
.
-
•TO
•TO
.843
-
.819
-912
•938
.906
•912
•993
•974
•930
•895
•932
.OJO
HOj-H
.
•336
.181
.291
.064
•500
.256
.844
.876
.805
.832
.918
•773
.712
•585
•653
.546
SUUATE
.
.834
•907
.889
.845
•917
•878
•947
•913
•877
.910
•976
•949
.880
•764
.867
.«»
-------
Table 51. DU POM! WATER RECOVERY DATA
105
Weekly
Period
U-7
8-18
19-23
2^-33
3^-37
38-fcL
te-W
^9-53
5^
55-57
58-63
6k-66
Membrane
Set
IA
IA
1A
IA
IA
I1A
IIA
IIA
IIB
IIB
IIB
IIB
Average
Recovery
Ratio
• 7(*
• 793
.799
•770
.7*H
.68^
.673
.565
.685
.566
.515
•351
Standard
Deviation
.050
.03!)-
.020
.023
.020
.02?
.036
.016
-
.03^
.029
.161
No. Of
Data Pts.
1^
11
5
10
k
k
7
5
1
3
6
3
8($
Confidence
Level
.62 - .79
.75 - .8^
.77 - .83
.7^ - .80
.71 - .77
.6k - .73
.62 - .72
;5^ - -59
.50 - .63
.VT - .56
.16 - .$k
The standard deviations are -uniformly low (except during the last three
week period) and it would seem that the limits could have been logically
set at about the 90$ confidence level. During the final weeks
of the study, severely fouled membranes produced erratic performance
and incongruous data. There were only minor variations in the total
rejection ratios before this period. The changes toward lower product
water and higher solute rejection ratios after week 38 (after B-9
installation) reinforces the probability that the new membranes were
"tighter". The observation that the recovery ratios decreased while A
values increased would indicate an improvement in membrane efficiency.
The total rejection ratios in Table 52 obtained through the courtesy
of Du Pont, are based on a single day's performance of B-5 modules. The
values check rather closely with the average data shown in Table 50 for
weeks 19 through 23.
Table 52. SOME CONSTITUENT REJECTION DATA, COURTESY OF DU PONT
G onstituent
Total Rejection Ratio
Ca
Mg
Na
K
Cl
F
Si0
.83
.86
.!£
•53
.83
.69
.15
• 37
.07
-------
io6
Average Rejection and Material Balance Ratios
The average rejection ratios uneorrected for temperature and material balance
ratios in Table 53 show correlation with the degree of post secondary
treatment. The material balance data are uniformly good (agree within
plus or minus ten per cent). The exceptions are probably the result
of sampling or analytical errors.
Membrane Fouling and Cleansing
The Du Pont unit exhibited marked fouling tendencies and required frequent
and prolonged membrane rejuvenations. Many types of cleaning solutions
were tried, either singly or in combinations ( reversed flow, high and
low pressure flushings, air bumps, extended soaking periods, etc. ) without,
in most cases, a substantial or permanent improvement in the flux. The
manufacturer worked very closely with the plant operator in suggesting
new techniques, materials, etc. and assisted in their application.
A partial list of the chemicals and materials used, includes weak acetic
acid, Proctor & Gamble's "Biz", citric acid, dilute hydrochloric acid,
EDTA (as the tri-sodium salt), weak caustic soda, sodium hexa-metaphosphate,
sulfamic acid, etc., and even a Du Pont proprietary product termed "Chemical
X" (possibly tannic or formic acid).
The membranes were flushed or cleansed about ninety times in fourteen
months of operation. During the last months of the study, the flushing
frequency was increased to every other day. An inclusive list of
rejuvenations, divided for convenience into the three permeator service
periods, is found in Tables 5^ through 56.
Because the flux increases were based only on data obtained before and
after rejuvenation, there is no indication of the duration of improvement.
However, during the final weeks, the improvements were short lived, lasting
not much more than a day. For a more complete picture of membrane cleansing
procedures, the reader is referred to Appendix A-3.
Comments
Some permeators were removed and returned to the manufacturer for inspection.
The fibers were reported to be in excellent condition without evidence of
over-chlorination or deterioration across the bundle. It is the manufacturer's
opinion that higher (perhaps 3 fold) chlorination dosages could have been
tolerated, but only if applied using a contact chamber with a 30 minute
detention time. Although the membrane material was in good condition,
internal modular fouling was noted.
The examination report on the B-5 module No. 2 removed at week 18, stated
that the unit was partially fouled with inorganic deposits, "about 95$
calcium phosphate and sulfate." In a report on a spent B-9 module, two
types of fouling were cited. A crystalline deposit of calcium sulfate
was found on modules arranged in the "five-in-parallel" pattern. No
-------
Table 53- AVERAGE REJECTION AND MATERIAL BALANCE RATIOS, DU PONT UNIT
Week
Nos.
/ 4-7
f 8-18
» 19-23
^ 8-23
« 21*- 33
\ 34-37
4-TT
/ 38-1*1
/ 1*2-1*8
49-53
10
ON 38-53
« 54
i 55-57
\ 58-63
\ 64-66
\55-66
4-66
Average Rejection Ratios
T.D.S.
.767
.816
.
.816
.798
.828
.806
.967
.938
.912
.937
.976
.935
.885
.802
.877
.877
Spec.
Cond.
.747
1?35
.800
.755
.775
.736
.758
.956
.916
.887
.917
.967
.933
.851
.747
.845
.818
Total
C.O.D.
—
-
.805
.805
.681
.746
.723
.748
.947
.968
.891
.792
.888
.957
.862
.918
.822
Total
Hard-
Ness
.
-
- -
-
.912
.928
.919
.986
.962
.939
.961
.995
.984
.945
.900
.939
.945
Ortho-P
—
-
.855
.855
.840
.880
.850
.985
.954
.932
.955
_
.966
.900
.884
.914
.888
N03-N
_
-
• 333
.333
.168
.850
.380
.973
.928
.878
.919
.963
.878
.815
.656
.746
.735
Material Balance Agreement Ratios
T.D.S.
.912
1.071
-
1.071
.955
1.052
.988
1.022
.938
.972
.972
1.053
1.107
.956
.964
.995
.985
Spec.
Cond.
1.034
1.005
1.028
1.012
.985
.938
.998
.948
.919
.962
.,94o
.925
.973
.923
.985
.951
.974
Total
C.O.D.
.
-
1.405
1.405
1.068
1.015
1.131
.907
•933
.758
.885
1.310
.834
.846
.832
.840
•971
Total
Hard-
Ness
.
-
-
-
.982
1.003
.990
1.084
.983
1.021
1.020
1.060
I»i48
.990
1.019
1.030
1.010
Ortho-P
—
-
.888
.888
1.004
1.140
1.026
1.079
•999
.993
1.017
«.
1.014
1.002
1.043
1.016
1.019
NO--N
—
-
.883
.883
1.007
1.824
1.180
3-712
.825
.822
1.305
1.115
1.231
.961
1.044
1.048
1^730
6
-------
108
Table 5U. MEMBRANE REJUVENATION RECORD, DU PONT B-5's
Type
Air bump
Acetic acid
Biz
"Chemical X"
Dilute HC1
EDTA
Total
No. Of Times
11
1
18
1
1
11
*3
Average Flux Increase
5
0
8
0
0
9
Average 6%
W)
Table 55. MEMBRANE REJUVENATION RECORD, DU PONT B-9's
{WEEKS 38-53)
Type
Biz
EDTA
Total
No. Of Times
7
7
Ik
Average Flux Increase ($)
k
2
Average 3$
Table 56. MEMBRANE REJUVENATION RECORD, DU PONT B-9's
(WEEKS 55-66)
Type
Biz
EDTA
Biz-EDTA
Other
Total
No. Of Times
16
8
6
35
Average Flux Increase ($)
15
5
k
0
Average 8%
-------
109
deposits of similar constituency were found on the modules involved in
the 3-2 pattern. All of the modules which were in service at week 66
were heavily plugged with organic slime. What these observations indicate
is the superior nature of the 3-2 flow pattern "t". Using flow pattern
"t", the flow rate was great enough to prevent buildup of inorganic
scale, and there was no need to install venturies to maintain high brine
pressures,. Without venturies, the organic deposit problem might have
been reduced if not eliminated. When the "t" pattern was finally resumed,
it was too late to be effective on the severely fouled modules.
Du Pont also had other helpful comments which are condensed below:
1. Du Pont type permeators are not likely to be highly
successful using feeds with high total and dissolved C.O.D.
or large amounts of colloidal material.
2. The minimum advisable secondary effluent post-treatment
would be sodium hexa-metaphosphate (to inhibit calcium
sulphate precipitation) followed by sand filtration,
chlorination (with adequate contact time) and ten micron
cartridge filtration.
3. Automatic dumping and flushing controls are needed in
the event of power failures: (this would be helpful
in preventing supersaturated solutions from depositing
their solutes during a pressure drop or a pH change).
A Value - Time Plots
Three A vs time (hours) plots were prepared:
Figure 23 - weeks 4-37
Figure 2k - weeks 38-53
Figure 25 - weeks 55-66
The discontinuity in Figure 23 occurring at 2107 hours (week 18) resulted
when the second B-5 permeator was replaced by a new module. The discontinuity
in Figure 25 at 1373 hours (week 64) indicates the effect of discontinuing
sand filtration and operating the unit on chlorinated, pH adjusted (only)
secondary effluent.
Colorado River Water Test
In order to obtain R.O. data on very high calcium, high sulfate water, the
second set of B-9 modules were initiated using Colorado River Water. The
test conditions were:
Five B-9 permeators in parallel;
Venturi type pressure controlling orifices close to the shell
on the exit line from each permeator, as suggested by the manufacturer;
Total membrane area - 8,424 sq. ft.;
Average pressure in psi:
Inlet - 407.7; Brine (at shell) - 372.2;
-------
.01
o
x
04
I
u
o
o
.02
.008
-------
1
0.8
WEEK NUMBER
_ 38
in 0.6
O
0.4
I
u
UJ
CO
o
o
CO
X
CO
0.1
CO
o
u
Ul
o:
CD
UJ
.04
.02
.01
39 39 39
39
41
42 43
39
50
62
44
SLOPE=- .off
o c o
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M.-SEO-O.062
2. % WATER RECOVERY—75
3. OPERATING FEED PRESSURE -404 p. s. i.
4. BRINE PRESSURE —369p. s. i.
5. pH —5.90
100 2
TIME (HOURS)
8 1000
8 10000
Figure 2k. A vs. Time plotted logarithmically, Du Pont 11/18/70 - 3/2/71 (B-9's)
-------
WEEK NUMBER
o
x
3E
I-
I
O
LU
I
O
o
V.
s
o
< .06
V)
o
U .04
cc
m
s
LJ
.01
54
54
55 55
55
57
58 59
66
1 I T
55
60
AVERAGE CONDITIONS:
FLUX (GM/SQ.M.- SEC) — 0.29
% WATER RECOVERY - 56
OPERATING FEED PRESSURE -410 p. *. i
BRINE PRESSURE — 360 p. s. i.
pH-5.66
DISCONTINUED SAND FILTRATION
too
i otio
TIME
-------
113
Inlet feed temperatures 57 to 60 degrees (F);
Feed pH - 6.0
Average total feed flow rate (for the five modules) - 8.8 gpm.
After adjusting to 25° C, the physical test data averaged:
Recovery ratio - 0.912
A value - 0.256 x 10 gm/sq. cm-sec-atm
The chemical data and the total rejection ratios are shown in
Table 57:
Table 57. DU PONT B-9 PERMEATOR ANALYSES
COLORADO RIVER WATER TEST*RTM (mg/l)
Constituent
Na
K
Ca
Mg
Cl
Nitrate -IT
SOi,. (Raw = 336)
Ortho-P
Total P
Total C.O.D.
Diss. C.O.D.
Total Hardness
A cidif ied
Feed
124
7
88
37
102
0.43
436
0.04
0.04
11.0
9.7
372
T.D.S. (Raw 752) 822
Si
B
Fe
Mn
F
7
0.4
0.02
0.00
0.5
Product
18
0
0.5
0.7
13
0.08
11
o.o4
o.o4
4.8
0.9
4.0
47
1
0.3
0.00
0.00
0.3
Total
Rejection Ratio
.85
-
-99
,98
.87
.81
.97
_
_
.56
•91
=99
.94
.85
.25
_
-
.40
-------
SECTION XII
GULF GENERAL ATOMIC UNIT,
SPIRALLY WOUND MODULE DESIGN
Introduction
This unit was purchased from Gulf General Atomic Incorporated of
San Diego, California, (in late 1970 the corporate name was changed to
Gulf Environmental Systems Company). The unit's characteristic component
is its spirally wound module. The six modules making up the unit were
rated by the manufacturer at a nominal capacity of 10,000 gallons per
day, with 60$ product water recovery and about a 0.90 rejection ratio.
Physical Configuration
Figure 2.6 is a photograph of a Model 50l6 Gulf reverse osmosis unit as
used in this study. The six ten ft. long pressure vessels in the left
background, house the spirally wound modules; the three vessels in the
foreground contain three multistage series-connected submersible
centrifugal pumps. The polyethylene container at the left is a sulfuric
acid make-up tank for pH control of the feed.
¥ithin each pressure vessel were three-3 foot long Model ^000 modules
connected in series. The feed was distributed to the first three pressure
vessels in parallel. The two following vessels received (in parallel)
the combined brine flows from the first three vessels. A final vessel
received the combined brine flows from the latter two. This configuration
is designated as pattern "m" (3-2-1), with eighteen modules for an overall
effective membrane area of 900 sq. ft. (based on the nominal membrane area
of 50 sq.. ft. per module).
The unit had a nominal total rejection ratio of about 0.9^5 when operating
on a 2000 mg/1 sodium chloride solution at 600 psi and 25° C.
The product water flow rate, over the lifetime of the unit, was estimated
by the manufacturer to be 8.3 gal./sq.ft./day when no "major" membrane
fouling conditions existed. (Note: this would give the unit an average
capacity of about 7500 gal./day). At a 0.6o recovery ratio, and a 95$
time-on-line factor, the brine flow was estimated at about 3.6 gpm.
No Reynolds numbers were estimated for the Gulf unit but it was the
manufacturer's opinion that at brine flows below 2.5 gpm, concentration
polarization accompanied by a severe fouling was possible. There was
relatively little other mechanical equipment description supplied by the
manufacturer.
Membrane Specifications
The manufacturer stated that the membrane used in the Model 4000 Gulf
-------
Figure 26. Gulf General Atomic Inc. reverse osmosis unit
'
-------
n6
module was made from "...cellulose acetate (Eastman chemicals) processed
by a modified Loeb technique, (it is) asymmetric (with a) dense surface
layer 1000-2000 Angstrom units thick and a more spongy support 3-^ mils
thick..." In a second communication, it was stated that the "...
membrane contains approximately two-thirds water by weight and is cast
on a drum from commercial grades of cellulose acetate, described as a
2.5 acetate with an acetyl content of approximately 39 to ^0$. After
casting, the membrane is annealed for a short time at 80 to 85°C..."
They also stated that "... in the sets of modules ... tested (at Hemet),
the membrane was supported by a backing of D-601 polyester sailcloth..."
In Gulf's "SP32ll|-13A 0106910" product bulletin, it is stated that the
membranes plus the backing material are used to form a sandwich. After
sealing, this envelope and a mesh backing material are wound around a
perforated plastic tube to form the spiral module. Figure 27 illustrates.
Chronological Record
The following notes, taken from the plant data logs, show the major events
and changes in operation:
Week and Day
2 if Unit placed in operation. Membrane Set No. 1, flow
pattern "m". Unit operating on feed of secondary
effluent, treated by reactor-clarification, sand
and activated carbon filtration. pH adjustment
also included.
5 4 Added pre-R.O. unit chlorination to feed treatment sequence,
7 6 Began D.E. filtration - feed of reactor-clarified, sand,
activated carbon and D.E. filtered secondary effluent
followed by pre-R.O. unit chlorination and pH adjustment.
2k 3 Stopped D.E. filtration of feed.
31 3 Shut unit down - pump failure.
32 6 Restarted unit after pump failure.
33 6 Removed reactor-clarification. Peed treatment:
sand and granular activated carbon, filtration 'followed
by pre-R.O. unit chlorination and pH control.
iH k Resumed reactor-clarification at head of secondary
post treatment sequence.
h6 k Removed activated carbon filters; feed consisting of
reactor-clarified, sand filtered secondary effluent
with pre-R.O. unit chlorination and pH control.
¥3 6 Shut unit down - feed pump failure.
^9 6 Restarted unit with new (No. 2) membrane set installed.
Feed: reactor-clarified, sand filtered, pre-R.O. unit
chlorinated, and pH adjusted secondary effluent.
50 1 Shut unit down - feed pump failure.
50 7 Feed pump restarted.
57 6 Feed treatment: sand filtered secondary effluent with
chlorination and pH control.
-------
117
DETAILS OF SPIRAL WOUND MODULE CONSTRUCTION
SEE DETAIL A
MESH SPACER
MEMBRANE
PRODUCT SIDE
BACKING MATERIAL
PERMEATE TUBE
GLUE LINE
Figure 27. Details of spiral -wound module, (Courtesy, Gulf)
-------
118
Week and Day
66
69
3 Shut unit down - feed pump failure.
6 Restarted unit using pre-R.O. unit chlorinated
and pH adjusted secondary effluent.
7 Stopped testing program.
Data Groupings
The data groups shown in Table 59 were selected primarily to emphasize
the effect of changes in feed quality on reverse osmosis performance.
Two consolidated sets, for weeks 2-^8 and ^9-6$, were included to
show the average operating conditions for the two Gulf membrane sets
used during the study. Because these consolidated periods include
mixed process conditions, their data may not compare well with that
of other groups. The primary purpose was to group data which appear
in the Gulf log A time plots.
Mechanical andOperational Problems
Of the mechanical equipment, only the Reda multistage submersible
centrifugal pumps gave any serious problems during the test period. The
three series-connected pumps (Figure 26) were mounted with their individual
motors inside pressure vessels similar to those housing the spirally wound
membranes.
Conduction to the surrounding water was the only means to dissipate heat
generated by the 2 H.P., 230 volt motors which may have been the prime
cause of pump failure. Although these pumps were repaired or replaced
without charge by the manufacturer, their numerous failures (9 times over
69 weeks) broke the continuity of testing and very likely contributed to
membrane fouling. It is difficult to separate the "mechanical" pump
problems from those termed as "operational". The latter group normally
includes membrane fouling and membrane flushing difficulties and frequencies,
but in this instance difficulties could be the result of poor circulation,
changes in the mode of treatment or even the inherent flow characteristics
of the spirally wound module.
Between weeks 2 and 69, there were 11,59^- available operating hours. The
Gulf unit operated for about 86$ of this time (10,016 hours). Table 58
lists the major out-of-service hours and the percentages they represent.
Table 58. OUT-OF-SERVICE RECORD, GULF UNIT
Mechanical problems
Membrane cleaning
Membrane failures
Alterations, additions
Pretreatment feed problems
Total down time
Hours
1233
129
_
16
200
1578
I
10.61*
l.n
.11*
1.72
13.61
-------
119
Table 59. REVERSE OSMOSIS PROCESS INFORMATION, GULF
Week Wo.
2-3
*
5-6
7
8-23
2li-33
oh _ ]i i
I[.P — |[TD
14.7 -l^S
2— lj.8
1^9-57
58-63
11-9-69
66-69
Secondary
Post-Treatment
A,B,C,P
A,B,C,P
A,B,C,E,P
A,B,C,E,F
A,B,C,D,E,P
A,B,C,E,P
B,C,E,P
A,B,C,E,P
A,B,E,P
VARIOUS
A,B,E,P
B,E,P
VARIOUS
E,F
Membrane Set
1
1
1
1
1
1
1
1
1
1
2
2
2
2
Flow Pattern
m
m
m
m
m
m
m
m
m
m
m
m
m
m
Post-Treatment Legend
A = Reactor-Clarification
B = Sand Filtration
C = Activated Carbon Filtration
D = D.E. Filtration
E = R.O. Unit Chlorination
F = pH Control
-------
120
Water Permeability Data
The Gulf A values and similar related ratios are shown in Table 6l»
Data for week k are shown separately to emphasize the effect of the
first Biz flushing. If data for weeks 2-4 were grouped together, A
value data would appear (erroneously) continuous to week 3^.
At week 34 (1*663 hours on Figure 29), the A values fell abruptly when
the reactor-clarifier was shut down. The A values continued at a low
level even after resuming reactor-clarification at week 42 (6l02 Hours).
This would suggest that the membrane irreversibly fouled in the absence
of reactor-clarification.
The use of the D.E. filters during the 8-23 weekly period did not change
water permeability. This agrees with observations made on the other units.
The A value again declined sharply, when the activated carbon filters
were removed at week 4j. During the final weeks of the study, the log
A-log time slope became relatively steep indicating a loss of permeability
and accelerated membrane fouling.
The low data correlation coefficient in the 42-46 weekly grouping can be
correlated to a very low log-log slope ratio. The same is true, to a
lesser degree, for the 24-33 weeks' set. Both conditions can be related
to effective membrane rejuvenation toward the end of the grouping.
Water Recovery and Total Rejection
Tables 62, 63 and 6k show the constituent concentrations in the R.O. feed,
product water and brine. The total rejection ratios in Table 65 vary
slightly except for a few random points which may be the result of sampling
or analytical errors. Standard deviations for the. recovery ratios and the
estimated confidence levels are shown in Table 60. The standard deviations
are uniformly low in most instances and confidence levels correspondingly
narrow except during the early periods (reflecting high initial membrane
compaction rate) and during the intense fouling conditions after week 49-
Table 60. GULF WATER RECOVERY DATA
Weekly
P eriod
2-3
2-7
8-23
24-33
34-M
te-lrf
47-1*8
49-57
58-63
66-69
Membrane
Set
1
1
1
1
1
1
1
2
2
2
A verage
R ecovery
Ratio
.475
.502
.678
.696
.623
= 565
.1*82
.586
.402
.290
S tandard
D. eviation
.066
.059
.050
.02^
.037
.052
.011
.101
.086
.034
NO. Of
Data pts.
2
4
16
10
8
5
2
9
6
4
8(#
C onf idence
Level
.27 - .68
.40 - .60
.61 - .74
.66 - .73
,57 - .68
.W - .64
.45 - .52
.44 - .73
.28 - ,53
w^^ * f *s
.23 - .35
-------
Table 6l. WATER PERMEABILITY DATA, GULF
Week Nos.
2-3
l*
5-6
1
8-23
2l»~33
3l*-Ul
l*2-l»6
1*7-1*8
2-1*8
1*9-57
58-63
1*9-69
66-69
No. Data
Sets
8
1*
12
5
79
39
35
21*
11
217
37
31
83
15
Avg. A
x HP
0.9829
1.21*1*3
0.9330
1.1273
1.0079
0.9313
0.8251
0.7676
0.5918
0.9190
0.9258
0.5377
0.691*6
0.1*l£5
Log-Log
Slope
-0.051^
-0.0272
-0.021*8
+0.011*0
-0.01*29
+0.001*2
-0.01*1*2
-0.0006
-0.0271
-0.0806
-0.191*1
-0.1095
-0.2807
-0.0702
S.td. Dev.
Slope
0.0101
0.0209
0.0117
0.0131
0.001*8
0.001*9
0.0051
0.001*1
0.0172
0.0077
0.0155
0.0278
0.0183
0.0210
Correl,
Coeff.
.902
.677
.556
.525
.717
.138
.831
.030
.1*65
.579
.901*
•591
.863
.679
Avg.
G.F.D.
7-09
8.98
6.73
8.13
7.27
6.72
5.95
5.5^
1*.27
6.63
6.68
3.88
5.01
3.21*
Avg. Effective
Op. Pressure (P. S.I.)
539
570
592
596
579
580
581*
593
601
581
5M
582
559
561
-------
122
Table 6.2. FEED WATER QUALITY, GULF, pH ADJUSTED
WEEK
DOS.
2-3
4-7
8-23
24-33
34-41
42-1*6
47-48
2-1(6
49-5T
58-63
66-69
1*9-69
2-69
T.D.S.
T6o.o
750.0
740.3
T37-8
778.8
762.5
792-5
752.7
795-7
81*0.0
81*6.7
821.9
770.8
SPEC.
COND.
1222.7
1215.2
1283.8
1252.7
1253-1
1200.0
1329.3
1253-2
1288.8
121*6.1
1365.9
2000.0
1301*. 5
CHLORIDE
-
-
118.70
11*2.1*7
136.23
120.25
131*. 72
130.11*
137.36
150.26
116.79
151*. 1*1
134.65
mg/1 ex
TOTAL
C.O.D.
28.75
22.36
8.83
5.88
10.85
8.28
31-15
11.1*9
35.08
42-35
1*1*.QO
60.62
19-56
cept spec-
nrss.
C.O.D.
-
23-53
7-27
4-55
5-73
1.30
22.1*0
7-99
25-71
27-91
15-93
61.1*0
10.29
eond. aa m.
•TOTAL
HARD-
NESS
-
-
200.0
200.0
221.67
227.50
262.00
138.93
230.00
225-29
216.1*3
259-23
223-75
Lcromhos
OKEHO-P
-
-
3.07
6.90
13-61*
7-50
6.92
8.18
8.18
13-38
8.83
12,,80
9-25
ALKA-
LINITY
-
-
35-56
1*0.12
5^.20
26.05
41.34-
1*1.59
26.12
59-15
15-45
44.92
39-98
CALCIUM
-
-
58.82
51-57
66.66
-
67-65
55-62
76.17
66.69
72.70
54.17
58.94
N03-N
-
-
4.36
4.00
0.741
-
6.60
4.37
7-63
5-90
4.31
2.66
2.87
S04
-
-
360.0
314.3
34o.o
363.6
388.8
339.0
345.0
416.6
496.7
333-3
347.3
Table 63. PRODUCT WATER QUALITY, GULF
(mg/1 except spec. coad. as micromhos)
WEHC
HO.S
a-3
4-7
8-23
24-33
34-41
42-46
47-48
2-48
49-57
58-63
49-69
66-69
2-69
T.D.S.
70.0
80.0
48.0
72.8
69.4
85.0
87.5
65-6
56.!+
65.8
73-3
63.1
65.0
SPEC.
COND.
134.5
96.0
95-0
114.0
142.6
151.2
163-5
117.8
103-1
94-7
112.0
158.0
116.1
CHLORIDE
-
-
14.60
16-67
22.75
19.00
26.00
19.00
12.50
29-00
19-27
21.00
19.12
TOTAL
C.O.D.
2.300
2-750
0.362
0.711
2.050
2.533
2.150
1-275
6.700
1-567
3-564
4.850
1.702
DISS.
C.O.D.
-
1.200
0.640
0.414
0.700
1.100
2.150
0.759
0.900
1.200
1.625
3-500
0.864
TOTAL
HARD-
NESS
-
-
2.00
3-80
3-99
5-46
6-55
3-89
2.30
3-83
3-03
3-37
3-58
ORTHO-P
-
-
0.046
0.069
0.147
0.090
O.'JJO
0.090
0.090
0.428
0.256
0.320
0.148
TOTAL
ALKA-
LINITY
-
-
16.43
19.98
21.68
18.44
18.85
19-30
10.29
13-84
12.24
15.05
16.83
CALCIUM
-
2.000
1.083
2.000
-
1.1JO
1.446
0-457
1.067
0.727
0.650
1.0&L
NOj-N
-
-
2.272
2.450
0.280
-
4.250
2.438
3.800
2.700
2.803
1.360
2.548
SC^
-
-
3-60
4.40
2.38
4.00
3-50
3-39
4.14
5-00
4.47
4.00
3-82
-------
123
Table 6k. BRINE QUALITY, GULF
WEEK
803.
2-3
*-7
8-23
2l*-33
3Ma
1*2-1*6
I*7-W
2-W
1*9-57
58-63
66-69
1*9-69
2-69
T.D.S.
.
-
-
2370
2011
1720
1592
1962
1736
13*
1030
1526
1794
SPEC.
COND.
1925
2289
3802
4318
3167
2550
2490
3368
2802
2073
1661*
2332
3062
ms/1 ex
CHLORIDE
-
-
31*
>*53
330
319
256
31*9
281
20)4.
-
252
312
cept spec.
TOTAL
C.O.D.
.
.
38.2
15.1
19-1
18.0
55.5
23.1*
62.9
69.1
67.7
66.7
35.6
oond. as ml
TOTAL
HARD-
NESS
_
.
-
7*
620
555
523
599
506
1*22
305
1*1*9
535
crorahos
ORTHO-P
_
.
7.1*
52.7
38.1
15.2
13.9
32.5
18.7
21.6
li*.3
19.1
27.6
ALKA-
LINITY
.
m
.
55.0
71.9
1*0.8
50.5
54.3
27.6
36.1
28.6
31.0
l*l*.6
CALCItM
—
.
190
176
»
146
173
138
ill
.
129
11*7
NOjK
•»
7.8
5.3
9.5
12.0
7.9
7.7
16.3
6.3
l*.l
9.1*
7.7
sok
.
,
.
750
_
_
750
812
625
_
750
750
Table 65. WATER RECOVERY AND TOTAL REJECTION RATIOS, GULF
WEHC
BOS.
2-3
l*-7
8-23
24-33
3l*-l*l
1*2-1*6
47-48
2-48
1*9-57
58-63
66-69
49-69
2-69
WATER
RECOVER!
RATIO
.475
.502
.61*8
.696
.623
• 565
.1*82
.618
.586
.402
.290
.466
• 574
T.D.S.
.908
.893
•935
.901
•911
.889
.890
•913
•929
.922
.913
•923
.916
SPEC.
COND.
.890
-921
.926
.909
.887
.874
.877
.906
• 920
•924
.918
•921
•9U
CHLORIDE
-
-
•877
.883
•833
.842
.807
.854
•909
.807
•835
.864
.858
TOTAL
C.O.D.
.920
•877
•959
.879
.811
.694
•931
.889
.809
•963
•919
.920
•913
mss
C.O.D.
-
•949
.912
.909
.878
•154
.904
•905
•965
•957
.898
•943
.916
TOTAL
HARD-
NESS
-
-
•990
-981
.982
•976
•975
.982
•990
•983
.986
•987
.984
ORTHO-P
-
-
•985
•990
•989
.988
•987
•989
•989
.968
•971
•975
.984
ALKA-
LIHm
-
-
• 538
.502
.600
•392
.544
.536
.606
.766
.208
.665
• 579
CALCIUM
-
-
.966
•979
•970
-
•983
• 974
•993
•983
•990
.988
.982
H03-H
-
-
•479
.388
.622
-
• 356
.442
.502
• 542
.349
.490
.461
SULFATE
-
-
•990
.986
•993
•989
•991
•990
.988
.988
•991
.988
•989
-------
12k
Average Rejection and Material Balance Ratios
The average rejection and' material balance ratios shown in Table 66
are important as accuracy indicators for the sampling and analytical
methods used in this study. The material balances show data agreements
generally within plus or minus ten per cent (total C.O.D. and nitrate
nitrogen ratios excepted) which agrees well with data from other units.
Membrane Fouling and Cleansing
The Gulf modules were flushed one hundred times with various cleansing
solutions. This very high frequency, when compared with similar data
for the other units, is probably the result of poor flow distribution
within the spirally-wound modules.
The Gulf unit cleansings are arranged in Table 67 according to feed treatment.
This type of presentation was used because of the relatively consistent
operational conditions, not possible with the other units.
Table 67 indicates that membrane rejuvenations were most effective when
operating on the least treated feed. These data are misleading, however,
when it is realized that most of the improved post-rejuvenation fluxes
lasted less than twenty-four hours.
The membrane fouling problem was discussed on numerous occasions with the
manufacturer and most of the flushing techniques and changes in the solution
compositions were used at their suggestion. At one time they suggested that
a principal foulant might be aluminum salts carry-over from the reactor-
clarifier. At no time, however, did the reactor-clarifier operate without
subsequent sand or activated carbon filtration. Additionally, post-reactor-
clarifier samples were analyzed numerous times and at no time did the aluminum
content exceed 0.12 mg/1. Most of the samples contained no detectable
aluminum. Analyses of two "spent" flushing solutions of weak phosphoric
acid also indicated minute retention,of aluminum by the modules.
Particulate fouling, however, was probably the major cause of poor fluxes
experienced during the latter stages of this study. It was speculated
that because of the module design, the membrane or module acted as a filter
for the insoluble constituents. Once particulate buildup began,the fluxes
became gradually worse, especially with poor feed quality. Indeed, Gulf
mentions in a training handbook, "... the spiral module is not adapted
to the treatment of water containing a high degree of particulate matter.
If suspended solids... are introduced... a slow buildup of particulate
matter... may result." Besides particulate matter, other constituents
(i.e. dissolved organics) were considered as possible foulants of the Gulf
modules.
Since slime had been found in some of the other unit modules, it was
speculated that humic acid from the slime may have dissociated at pH's
below 5.0. Although the feed pH did drop slightly below 5.0 on two
abbreviated occasions towards the end of the study, there was no apparent
correlation with the A values (indicators of flux).
-------
Table 66. AVERAGE REJECTION AND MATERIAL BALANCE RATIOS, GULF UNIT
Week
Nos.
2-3
4-7
8-23
24-33
34-41
42-46
47-48
2-48
49-57
58-63
66-69
49-69
2-69
Average Rejection Ratios
T.D.S.
—
.883
.959
.960
.950
.926
.926
.944
.955
.933
.923
.941
.943
Spec.
Cond.
.920
.942
• 959
' .954
.931
.916
.909
.944
.951
.940
.930
•943
.944
Total
C.O.D.
.
-
• 971
• 923
.858
.752
.951
.902
.860
.979
.919
.933
.910
Total
Hard-
Ness
.
-
-
.990
.990
.985
.984
.988
•994
.984
.988
.989
.989
Ortho-P
—
-
.994
•995
.994
.992
.992
•994
•994
.969
.972
.981
.989
NO -N
3
_
-
.579
.638
.884
_
.402
.580
.671
.557
.366
.550
.567
Material Balance Agreement Ratios
T.D.S.
_
.852
1.653
1.062
1.012
.946
1.098
1.044
1.030
.923
•951
.975
1.016
Spec.
Cond.
.983
.883
.978
.974
.932
.941
.935
.956
.935
.962
.957
.948
.954
Total
C.O.D.
.
-
2.595
.932
•953
1,123
.964
1.303
1.003
.907
.687
•915
1.197
Total
Hard-
Ness
—
-
-
1.117
1.028
•956
1.069
1.036
i.o4i
1.010
.943
i.oi4
1.028
Ortho-P
_
-
1.130
1.042
1.078
.955
1.174
1.050
1.018
•950
.953
.982
1.023
N00-N
3
_
-
.907
1.064
2.400
-
.926
1.148
1.190
.914
1.315
1.186
1.165
fc
-------
Table 67. GULF RECORD OF MEMBRANE REJUVEHATIOH
Post Treatment Sequence:
(A=Reactor Clarifier )
!B=Sand Filters )
C=Carbon Filters )
D=D.E. Filters )
(E-Pre-R.O. Unit Chlorination)
(F=pH Control )
A,B,C,D,E,F
No. Flushes/ 37 weeks
Avg. % Flux Increase
B,C,D,E,F
Ho. Flushes/8 weeks
Avg, % Flux Increase
A,B,E,F
Ho. Flushes/11 weeks
Avg. $ Flux Increase
B,E,F,
Ho. Flushes/6 weeks
Avg. % Flux Increase
E,F,
No. Flushes/ 3 weeks
Avg. % Flux Increase
Totals
Ho. Flushes/68 weeks
Avg. % Flux Increase
Biz
22
8
2
k
17
18
15
114-
3
^5
59
17.8
EDTA
31
6
2
7
1
3
-
-
-
-
6
5-3
Biz & EDTA
1).
^
1^
8
1
11
-
-
-
-
9
7-7
H.PO
3 4
-
-
-
-
2
9
-
<•
-
-
2
9
HC1
11
8
-
-
5
1
5
8
-
-
21
5.6
HOC1
Mixture
(sodium perborate
Tritox X-100
CarboxymethyJ.
cellulose
solution)
-
-
-
-
-
-
-
-
2
40
2
1*0
-
-
-
-
-
-
-
-
1
*5
1
^5
-------
127
As mentioned early, it was the manufacturer's contention that low-
brine flows (below 2.5 gpm) could result in membrane fouling. Between
weeks 18 and 69, there were thirty-four days in which this occurred.
Thirty-three of these values were in the range of 2.20 to 2.V? gpm,
while the thirty-fourth was recorded at 1-95 gpm. These "sub'-critical"
brine flows cannot be given full credit for fouling, as heavy fouling
symptoms (weeks 51 to 69) do not correlate entirely with the general
period in which the low flows and reduced Reynolds Numbers occurred.
A third likely contributor to fouling was inadequate chlorination of
the unit feed. Without sufficient chlorination, bacterial slime
buildup on the membrane may occur readily, irrespective of modular
design. High nutrient values of the feed would accelerate the slime
growth rate. (See Section VII for the discussion of "Membrane Fouling
and Cleansing.")
Comments in Retrospect
Gulf has indicated that some of the procedures used in this project were
inappropriate to their unit. Among these "faults", was the policy of
gradually reducing the feed treatment. The other was an alleged low
chlorine residual in the feed.
Comments to the above follow:
1. The general impact of using a gradually worsening quality
of feed was recognized in the planning stages and ample
opportunity to voice disapproval with the program format
was given. With intentions clearly stated, the decision
was made to proceed with all units (irrespective of design)
operating under very similar conditions. One to be tested
was the spiral wound design module. Thus, the Gulf unit,
as used in this project was primarily representative of the
spirally-wound design and secondarily representative of Gulf
Environmental and its results should be viewed similarly.
2. With respect to the "low" residual chlorine maintained in
the feed, it is pertinent to quote a number of communications
from the manufacturer.
From a letter dated September 12, 1969,
"... as a recommended pretreatment procedure for your unit,
it will be necessary to adjust the pH of the feed to about 5
and to maintain a chlorine residual of about 1 mg/1..."
In the manufacturer's specific operating instructions, which were
received somewhat later, it was said that:
"...If the feedwater does not consistently contain at least
0.2 ppm Cl2 residual, continuous chlorine addition at a level
of 0.2 to 0.7 ppm should be considered..."
This would give a range of 0.2 to 0.9 mg/1.
-------
128
Again, in a letter dated December 11, 1969, it was stated,
"... if the feed is to be chlorinated continuously, the
maximum recommended chlorine residual is 1 mg/1..."
Although the above comments represent three separate attitudes towards
chlorination of feed, the feed maintained in this pilot study conformed
remarkably well to these specifications. Therefore, the chlorine levels
which were alleged to be so low as to allow slime growth on the membrane
were in actuality within the guidelines set by Gulf.
A Value - Time Plots
Figures 28 and 29 show the log-log A value - time (in hours) plots for
the two Gulf membrane sets. Figure 28 shows the data for the period from
week 2 to 1*8. The abrupt rise in A at about 200 hours (week 3) shows
the effect of the first Biz flush. The sharp decrease in A which levels
off about 73^ hours (week 6) appears to be the result of the second Biz
flush. Up to the time of the second Biz cleaning, the unit was flushed
seven times with a pH 5-0 water solution. It is evident that these
acidic water solutions were ineffective. The effect of the reactor-
clarifier shutdown which was discussed earlier under sub-section "Water
Permeability Data", appears at 5300 hours (week 3^). The abrupt decline
in A at about TOGO hours, (week Vf) is probably the result of discontinuing
granular activated carbon filtration.
All of the data shown in Figure 29 were obtained using the second set of
modules. The carbon filters were not in service during any portion of
this period and this probably caused the steep slope. Three feed pump
failures at weeks 63 and 6k (I9l6 to 2027) necessitated turning the unit
off until week 67 when a new feed pump was installed. The unit was
restarted on chlorinated pH adjusted secondary effluent.
The erratic data between 1800 and 3000 hours in Figure 29 is probably the
result of pump failures plus efforts to keep the modules free of deposits.
-------
WEEK NUMBER
10.0
i
13
25
37 49_
I
u
LJ
00
I
2
u
o
oo
oo
< o.e
h-
00
z
o
Q.
CD
LJ
0-2
0.10
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M.— SEC)— 3.83
2. % WATER RECOCOVERY— 56
3. OPERATING FEED PRESSURE— 602 p.s. i.
4. BRINE PRESSURE— 59* p.s.i.
5. pH-5.47
DISCONTINUED REACTOR-
CLARIFICATION
DISCONTINUED
CARBON FILTRATION
100 2
TIME (HOURS)
8 1000
8 10000
Figure 28. A vs. Time plotted logarithmically, Gulf, 3/9/70 - 1/27/71
to
VQ
-------
WEEK NUMBER
49
49 50 50
50
50
51
52 53
61
H
00
O
55
i
O
I
2
(J
O
< 0.8
0.6
z
O
U 0.4
o:
m
AVERAGE CONDITIONS:
1. FLUX (GIW/SQ.IVI.-SEC)-2.5g
2. % WATER RECOVERY—43
3. OPERATING FEED PRESSURE — 584p.s.i.
4. BRINE PRESSURE—564 p.s.i.
S. pH— 5.36
FEED
PUMP PROBLEMS
LJ
0.2
0.1
S 100 2
TIME
8 1000
10000
Figure 29. A vs. Time plotted logarithmically, Gulf, 2/k/Jl - 6/25/71
-------
SECTION XIII
RAYPAK INCORPORATED REVERSE OSMOSIS UNIT, MODIFIED TUBULAR DESIGN
Introduction
This unit was obtained at no cost to the project from the Ecological
Systems Division of Raypak, Incorporated of El Cajon, California, later
relocated to Westlake Village, California. This unit was substituted
for the Aerojet General unit, removed in December of 1970.
Its nominal rating was given as about 3*000 gallons of product water
per day at about a 50$ product water recovery. Solute rejection was
quoted as 90$•
Physical Configuration
The unit which was designated as ROpak Model 00300*1-03 consisted of
eight parallel sets of modules (cells). Each set (bank) was made up of
four series-connected modules. Within each module were four sets of two
tubes (cores). A single tube was half the length of the module, and
therefore, two tubes were joined end to end to form a complete module.
The cellulose acetate membrane was on the outside surface of the axially-
positioned tube sets. The thirty-two modules had a total surface area
of 128 sq. ft.
Irom the above information and the manufacturer's statement that the
tubes occupied half of the modular cross section, it was calculated that
the tubes were about 0.6k in. O.D., the modules about 1.3 in. I.D. with
hydraulic radius about 0.08 in.
Only one flow pattern, designated as "o" was used. This flow pattern
was an eight parallel, four series modular arrangement. The Reynolds
Number was estimated as about 7>000 for one of the better days of
operation of this unit.
Membrane Specifications
Specific details of the membrane type were not available from the
manufacturer. It was implied that the membrane was basically similar
to the type used on the Universal unit, i.e. a cellulose acetate,
formamide modified film as developed by Lpeb and Manjikian at U.C.L.A.,
with a 0.90 total rejection ratio.
131
-------
132
Chronological Record
Negotiations preliminary to the acquisition of the Raypak unit began
about week ^0. The manufacturer initially indicated delivery to be
about week ^7, but delivery problems along with other delays arose
and the equipment did not arrive until week 50. The first unit leaked
badly. A replacement unit was installed, but it too leaked badly so
new modules were installed at week 62. The first set of process data
was obtained in the latter part of week 62. The unit was shut down
permanently at week 67. Throughout its five week period of operation,
the unit ran without pH adjustment at the manufacturer's suggestion.
Data Groupings and Water Permeability Data
The Raypak process information and water permeability data are shown
in Tables 68 and 69. Both the calculated A values and the data
correlation coefficients were low throughout the testing period.
Table 68. REVERSE OSflOSIS PROCESS INFORMATION, RAYPAK
Week
Nbs.
62-63
6^-67
62-67
Treatment
B, E
E
Various
Membrane Set
1
1
1
Flow Pattern
o
o
o
Special Conditions
No pH Control
it
H
Table 69. WATER PERMEABILITY DATA, RAYPAK
Week
Kos.
62-63
6U-67
62-67
No. Data
Sets
9
18
27
Avg. A
x 1CP
0.7922
0.5571
0.6355
Log-Log
Slope
-0.0156
-0.1137
-0.1202
Std. Dev.
Slope
0.1097
0.0739
0.0716
Carrel.
Coeff.
.05^
•359
.318
Avg.
G.P.D.
5.72
^.02
lv.58
Avg. Eff.
Op. Press. (P.S.I.)
751)-
799
790
Mechanical and Operational Problems
The Raypak unit had three main problems: (l) Plastic inserts which
connected the tubes to the tube sheets at the end of the modules
leaked badly, (2) the feed pump was of insufficient capacity and
(3) the membrane area was too small.
-------
133
The feed pump was rated at 5.6 to 5.9 gpm. After the pilot plant study
was completed, the manufacturer stated that it should have been at least
15 gpm and that the Reynolds Number should have been about 20,000.
Water Recovery, Rejection Ratios and Product Water Analyses
Performance summary data are shown in Tables fO-jk. The actual water
recovery ratios are far below the design estimate of 0.50 even though the
unit was operated within the 700-800 psi pressure range recommended by
the manufacturer. Most other R.O. units were maintained in the 500-600
psi range.
Instead of the estimated 3000 gal./day of product water, the actual
rate was never higher than 1100 gpd and averaged less than 605 gal./day.
During the six week test period the product recovery ratio averaged 0.0913
and the standard deviation was very nearly the same figure. With the
single exception of the nitrate nitrogen ratio, most of the other solute
total rejection ratios were close to the anticipated 0.90 value.
The material balance agreement ratios, in Table jk} were uniformly close
to unity. This indicates that the reported low recovery ratios were
valid and that the basic problem was not associated with erroneous data,
but with basic design deficiencies in the Raypak unit.
Membrane Fouling and Cleansing
The membranes of the Raypak unit were cleansed eleven times during its
six week testing period. Various flushing solutions were used, some
of which caused the product flow rate to nearly double. It was usually
only a few hours, however, until the flow rate returned to its original
level.
The rate of fouling in the Raypak study seems abnormally high compared
to other units. In all fairness to the Raypak unit, it should be mentioned
that the feed to which the Raypak unit was exposed had high fouling
potential, and this probably accounts in part for the rapid fouling.
Likely fouling contributors include low circulation rate, tight packing of
the tubes, and perhaps, the basic modular design feature of having the
membrane supported from the inside.
A Value - Time Plots
Figure 30 shows the log-log A vs time (in hours) plot for the Raypak unit.
The extreme heterogeneity of the data is evident and requires no further
comment.
-------
Table TO. pH ADJUSTED FEED WATER QUALITY, RAYPAK
WEES
HOS.
62-63
6l*-67
62-67
T.D.S.
61*5.0
71*8.8
728.0
SPEC.
COBD.
1295-8
l880.lt
1686.7
TOTAL
C.O.D.
l*l*.l*l*
53- 5T
53-87
(mg/l exci
rass.
C.O.D.
-
31-67
31-67
spt spec, coi
TOTAL
HARD-
NESS
218.1*2
230.00
230.95
id. as micro
OKEHO-P
11.81*
12.37
12.28
mhos)
TOTAL
ALKA-
LINITSr
265.71
256.66
259-88
CALCIUM
62.50
6U.28
63.80
^H^H^HHBBIHVBBVBBB>IBBBHBBIBIIBIHI
H03-N
6.19
5.1*0
5.56
SULFATE
156.25
166.66
161.76
I^M^MH^H^V^^BBWIM*
Table 71. PRODUCT WATER QUALITY, RAYPAK
WEEK
NOS.
62-63
6i*-67
62-67
T.D.S.
30.0
75-0
66.0
SPEC.
COKD.
155-5
833-0
607-2
TOTAL
C.O.D.
o.to
8.25
6.68
(mg/l except spec. cond. as micromhos)
fflSS.
C.O.D.
-
8-33
8-33
TOTAL
HABD-
HESS
8.30
32-85
27-9^
ORTHO-P
0.1*5
1.87
1.58
TOTAL
ALKA-
LIMITy
18.60
1*9.28
1*3.11*
CALCIUM
2.00
8.10
5-86
HOj-H
1.1*0
2-95
2.61*
SULFATE
5-00
2.00
2-75
Table J2. BRINE QUALITY, RAYPAK
mg/l except spec. cond. as micromhos
WEEK
NOS.
62-63
64-67
62-67
T.D.S.
730
777
768
SPEC.
COND.
1331*
1312
1319
CHLORIDE
122
ll*9
136
TOTAL
C.O.D.
1*0
1*0
TOTAL
HARD-
NESS
250
261
259
ORTHO-P
14.6
13.6
13.8
ALKA-
LINITY
269
252
257
CALCIUM
_
-
-
HO -H
5-9
6.2
6.2
-------
Table 73- WATER RECOVERY AND TOTAL REJECTION RATIOS, RAYPAK
WEEK
KOS.
62-63
64-67
62-67
WATER
RECOVER?
RATIO
•039
.10t
.091
T.D.S.
•953
• 900
.909
SPEC.
COND.
.880
•557
.640
TOTAL
C.O.D.
•991
.846
.876
DISS.
C.O.D.
-
•737
•737
TOTAL
HARD-
NESS
.962
•859
•879
ORTHO-P
.962
.849
.871
TOTAL
ALKA-
LINITY
•930
.808
.834
CALCIUM
.968
.87^
.908
N03-N
• 774
.454
•525
SULFATE
.968
• 988
•983
Table
AVERAGE REJECTION AND MATERIAL BALANCE RATIOS, RAYPAK
WEEK
NOS.
62-63
64-67
62-67
AVIvv\JE RiiJjiC'i'ION RATIOS
T.D.S.
.956
-904
.917
SPEC.
COND.
.882
.564
.670
TOTAL
C.O.D.
-
.881
.881
TOTAL
HARD-
NESS
.965
.856
.878
ORTHO-P
• 966
.864
.885
N03-N
•769
•552
• 596
MATERIAL BALANCE AGREEMENT RATIOS
T.D.S.
1.089
1.003
1.024
SPEC.
COND.
1.017
1.013
1.014
TOTAL
C.O.D.
-
.911
•911
TOTAL
HARD-
NESS
1.098
•996
1.016
ORTHO-P
1.200
1-033
1.066
M>3-N
•923
1.250
1.184
vn
-------
WEEK NUMBER
62
62
63 63
63
65
66 67
63
68
I
U
ui
(/)
I
2
O
o
c/)
< 0.8-
O
O
0.6 •
0.4
UI
z
<
o:
OQ
s
ui
2 0.2
AVERAGE CONDITIONS:
1. FLUX (GM/SQ.M. —SEC.)- 3.08
2. o/o WATER RECOVERY — S.I
0 3. OPERATING FEED PRESSURE—815
4. BRINE PRESSURE -782
5. pH- 7.47
= - -120
0.1
6 8 100
TIME(HouRS)
8 1000
8 10000
Figure 30. A vs. Time plotted logarithmically, Raypak, 5/6/71 - 6/lV?l
-------
SECTION XIV
UNIVERSAL WATER CORPORATION REVERSE OSMOSIS UNIT, A TUBULAR DESIGN
Introduction
This unit was obtained on a monthly rental basis from the Universal
Water Corporation of San Diego, California. Its nominal capacity, with
a full complement of seventy-two modules, was rated by the manufacturer
at 10,000 gallons of product water per day.
The symbolic nomenclature and equations used in this section are listed
and discussed in Sections V, VIII, and in the Appendix Section 'A-l.
Physical Configurations
The unit, as first placed in operation (shown in Figures 31 and 32),
consisted of two identical vertically -mounted banks of nine horizontal
racks with four modules on each rack. Each module contained eighteen
series -connected O.k in. I.D. tubes, and there was 7.0 sq. ft. of
membrane area per module .
Figure 7 shows an idealistic schematic flow diagram for a Universal
unit. Although the drawing indicates the principal lines and control
points, it varies from the configuration tested at Hemet in at least
two important points:
1. It shows only two pressure accumulators. Up to three
were required to control the pressure pulsations of
the test unit.
2. A flow configuration wherein two banks of five racks
in parallel were followed by four racks in parallel
was not tested in this project.
The four flow patterns used on the Universal unit are listed and
identified by letter below.
""
p" - Flow was split between two identical banks (of racks) of
modules. Within a bank the flow was first distributed, in
parallel, to four sets of modules, (a set consisted of four
serially-connected modules). The combined brine from the
latter, was distributed in parallel to 3 more sets of
modules in parallel. The brine from these was distributed
to two final sets in parallel. Between the two banks there
were 72 modules. At 7 sq. ft. per module the complete unit
membrane area was 504 sq. ft. Membrane sets 1 and k were
used in this configuration.
q" - Two banks, of four parallel sets, each set with four modules
in series. The total nominal area was 224 sq. ft. (thirty-two
modules). Membrane set No. 2 was used in this flow configuration
which was termed the high flux "open" type.
137
»„, ti
-------
138
Figure 31- Universal(¥ater Corp. reverse osmosis unit
Figure 32. Universal
reverse osmosis unit
in part
-------
139
"r" - Two banks, each containing two sets of modules in parallel,
followed by two sets of modules in series. Total nominal
area (thirty-two modules) was 22k sq. ft. Membrane Set
Wo. 2 was used in this flow configuration.
"s" - Two banks, each with two sets in parallel (each set
containing eight modules in series). Total nominal area
(thirty-two modules) was 22k sq. ft. Membrane Set No. 3
was used in this flow configuration.
The Reynolds numbers applicable to each of these configurations are
derived and listed, for the normal flow rates specific to each, in
Section VIII.
Membrane Specifications
Four sets of membranes were used on the Universal unit. The
manufacturer states:
•w.
"...UWC (Universal Water Corporation) tubular modules are
lined with membranes prepared from proprietary formulations,
specifically pyridine modified CA (cellulose acetate) solutions.
They are independently cast by the bob method, wrapped with a
nylon backing and Inserted into the tubular module and heat-
treated in site."
"Performance (membrane) characteristics are controlled through
casting solution formulation variations and heat treatment
conditions. Standardized membrane performance characteristics
range from high flow - low selectivity (rejection) ultra-
filtration systems to membranes capable of single pass (at
low flux-high selectivity rejections)."
"Compaction characteristics of the cellulose acetate membranes
generally yield a slope of 0.008 to 0.015 for the plot log flux
(mg/sq. cm - sec) x 10~5 vs iog time hours.
One of the purposes of this R.O. study was to investigate different
membrane characteristics. Sections IX through XIII discussed test
results with different membrane formulations (cellulose acetate vs
nylon or asymetric aromatic polyamide materials) and different membrane
configuration (tubular, spiral wound or hollow fiber) concepts. The
four Universal cellulose acetate membrane sets not only contributed
useful information to compare these aforementioned characteristics
but the formulations were altered so that low product water flux,
high solute rejection ("tight) membranes could be compared with the
high product water flux, low solute rejection ("loose") membranes.
The detailed operational performance for each of the four Universal
cellulose acetate membrane sets will be discussed subsequently.
Table 75 compares the TDS rejection ratios and product water flux
specifications reported by Universal and the performance of similar
membranes in this pilot study, using various post-treated secondary
effluent feeds. Membrane Set 2, which was the high product water flux-
low solute rejection membrane, performed relatively poorly as shown
-------
H
O
Table 75. • REVERSE OSMOSIS MEMERA1E PERFORMANCE COMPARISON, UNIVERSAL VS. HEMET TESTS
M,embrane
s:et no.
1
2 *
3
4
Universal Tests
Type
CA-89
CA-P-70
CA-P-88
CA-P-90
Total TDS
Rejection
0.935
0.725
0.955
0.965
Flux
(GFD)
19
31
17
16
Op. Press.
(P.S.I.)
-
-
-
-
Total TDS
Rejection
0.910
0.506
0.95^
0.964
Hemet Tests
Flux
(GFD)
15-5
26.4
14.9
11.0
Op.^ P.ress.
(P.S.I.)
651
675
593
599
Weeks O.n Stream
22
10
14
22
* High product water flux - low solute rejection membrane.
-------
by the substantially lower TDS rejections, and the rapid membrane
deterioration (which will also be subsequently discussed). The
"loose" membrane concept was rejected for further study in this
program.
Chronological Record
The following notes are abstracted from the plant data logs to show the
major events and changes in operation made on the Universal unit during
the progress of the study.
Week and Day
2 4 Start of data collection. Membrane Set No. 1;
flow pattern "p" . Operated using reactor-clarified,
sand and activated carbon filtered, pH adjusted
secondary effluent.
3 1 Unit out of service, severe vibration.
3 6 Unit back in service.
k k Started in-plant chlorination of reverse osmosis feed.
7 ^ Started recycling part of brine flow.
7 5 Feed treatment sequence modified to reactor-clarification,
sand and carbon filtration, followed by D.E. filtration,
chlorination and pH adjustment.
11 5 Stopped recycling part of brine flow.
l6 k Erratic operation - unit vibrating, rupturing
to membranes, etc.
19 3
22 4 No pH adjustment to feed.
23 3 Resumed pH adjustment.
23 5 Installed "loose" membrane Set No. 2; flow pattern "q".
28 it- Decreased feed flow rate by about k-Ctf) (to 8 gpm) to
improve product water recovery.
29 5 Changed to flow pattern "r".
30 7 Feed treatment sequence modified to include reactor-
clarification, sand and carbon filtration with
chlorination and pH adjustment.
32 3 Installed membrane Set No. 3; flow pattern "s".
33 7 Removed reactor-clarification from feed treatment
sequence.
4l k- Resumed reactor-clarification of feed.
k6 3 Unit offj awaiting new modules.
kQ 5 Installed membrane set No. k; returned to flow pattern
"p". Dropped activated carbon from feed treatment
sequence.
57 7 Discontinued reactor-clarification of feed. New
sequence: sand filtration, chlorination and pH adjustment.
6k- Ij- Discontinued sand filtration of feed.
69 7 Stopped testing program.
-------
Data Groupings
The Universal unit, like others involved in this stud$ experienced
a range of feed qualities, membrane types, flow configurations, etc.
The collected process data should reflect the effects of these
conditions and are grouped therefore according to process conditions.
Their respective time periods are shown in Table 78• The primary
time groupings such as weeks 2-3, k-6, 7-11 are representative of
relative uniform operating conditions. Some of the primary groups,
however, were combined in order to give a general idea of the
performances by the different membrane sets. Such data groupings
have narrow limits of value.
Mechanical and Operational Problems
The Universal reverse osmosis unit had very few mechanical and design
problems. Initial difficulties were of relatively short duration.
The unit vibrated so badly during the first months of operation that
it was shut down on many occasions. The vibration is believed to have
been caused by a misaligned feed pump drive shaft.
Although no serious problems arose because of the Universal membranes'
susceptibility to internal collapse, caution was exercised at all times
to avoid forming a partial vacuum within the tubes. Other operational
or membrane problems were minimal. From week 2 to week k6 and from
week 48 to week 69, there were 10,896 available operating, hours. The
unit operated for 94.31$ of this time. Tables 76 and 77 both are
indicators of the Universal membrane sets.
Table 76. OUT-OF-SERVICE RECORD, UNIVERSAL UNIT
Mechanical problems
Membrane cleaning
Membrane failures
Alterations, additions
Pretreatinent feed problems
Total down time
Hours
2lfO
7*
19^
5
107
620
%
2.20
0.68
1.78
0.05
_o.9JL
5.69
Table 77. STUDY LIFE SPANS, UNIVERSAL MEMBRANE SETS
Membrane Set
1
2
3
Weekly Group
2-22
23-32
1*8-69
Weeks On Stream
a
10
22
To
No. Of Replacements
0
0
tal 19
-------
Table 78. PROCESS IKFOEMATIOU, UNIVERSAL UNIT
Week Nos.
2-3
h-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
30-32
23-32
33
3^_1^1
U2-46
33-W
U8-57
58-63
6^-69
U8-69
Post Treatment
A, B, C, F
A, B, C, E, F
Membrane Set
A, B, C, D, E, F
A, B, C, D, E, F
A, B, C, D, E, F
Various
A, B, C, D, E, F
A, B, C, D, E, F
A, B, C, D, E, F
A, B, C, D, E, F
A, B, C, D, E, F
A, B, C, D, E, F
Various
Various
A, B, C, E, F
B, C, E, F
A, B, C, E, F
Various
A, B, E, F
B, E, F
E, F
Various
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
k
k
k
k
Flow Pattern
P
P
P
P
P
P
P
q
q
q
r
q, r
r
q, *
s
s
s
s
P
P
P
P
Special Conditions
-
Reject Recirciilation
Severe Vibration
No pH Adjustment
(Consolidated groupings,
see above)
-
-
Lowered GPM Feed
—
-
-
-
—
-
-
*»
-
-
-
-
—
Post Secondary Treatment Legend
A « Reactor-Clarification
B = Sand Filtration
C = Activated Carbon Filtration
D » D. E. Filtration
E = Pre-R.O. Unit Chlorination
F = pH Control
-------
Water Permeability Data
Table 79 shows the water permeability data ratios. The essential
information includes the test parameters (feed treatment conditions,
membrane set, flow pattern), average A value, the log A-time slope
of the A values and the GFD (gal./sq. ft. day). Although it was
"tighter", membrane set number 4 also required cleansing about every
100 hours. This is mainly attributed to the poorer feed quality
used toward the end of the study.
Having discussed R.0...data consistency and permeability generally in
Section V, it is now desirable to evaluate the specific reliability of
the Universal water permeability data. If we assume that correlation
coefficients greater than 0.7 represent good consistency and coefficients
in the range of 0.5 to 0.7 are marginal but acceptable, then the data
for weeks 4-6 and 7-11 should be questioned. Note that the slope is
positive in the 4 to 6 week grouping. This was caused primarily by
a sudden rise in A values (1.7 to 2.1 x 10'?) about halfway through the
period. Experience has shown, and this is probably no exception, that
unprecedented rises in A values are usually caused by undetected leaks
in the membrane. Two leaking tubes were removed from service shortly
after the time grouping,leaving the data substantially weighted.
Over-lap of the "leaking membrane" data into the next time grouping
(7-11) was responsible for a low correlation coeffient there.
Data for the 31-32 weekly time grouping may also be in error. Applying
standard deviation, there is an 80$ chance that the true slope was
anywhere between -0.0139 an(i +0.0033* However, some reliability can
be placed in the -0.0053 value, as operating conditions were normal for
the period.
A low correlation coefficient was also obtained for week 33 to 46
(membrane set 3, flow pattern "s"). The absence of the reactor-clarifier
from the post secondary treatment sequence (weeks 34.-4l) probably accounts
for this result. The log A-log time slopes and their standard deviations
appear to be normal in the sub-groups within the 33 to 46 week period.
The data for the 64 to 69 weekly period is both aberrant and erratic.
The former condition i.s the result of using low quality feed, which was
chlorinated, pH adjusted secondary effluent. The erratic nature of the
data stems from the numerous membrane rejuvenations during the period.
Water Recovery and Total Rejection Ratios
Tables 80, 8l and 82 give the feed, product and brine water qualities
respectively. It was from these values that the water recovery and
rejection ratios were computed. The water recovery ratios of Table 83
are data averages for the indicated time periods at ambient temperature.
Confidence levels for these were calculated for each membrane service
period in Table 84.
-------
Table 79- WATER PERMEABILITY DATA, UNIVERSAL UNIT
Week Nos .
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
33
34-4i
42-46 •
33-46
48-57
5S.-63
64-69
1*8-69
jsio. iata
Sets
5
15
20
41
4
85
65
24
6
30
7
37
7
14
44
8
35
23
66
47
30
27
104
Avg1. A
x 105
2.013
1.881
2.375
2.134
2.378
2.150
2.223
4.992
2.1(62
it.itse
1.890
3.995
1.839
1.865
3.652
2.179
20.030
2.082
2.066
1.691
1.391
1.360
1.519
Log A - Log T
Slope
-0.0397
+0.025!*
-0.0079
-0.0399
+-.01)30
+0.0197
-0.0359
-0.1191
-0.0252
-0.1780
-0.0171*.
-0.211-07
-0.0053
-0.0168
-0.2771
+0.0160
-0.0015
-0.0062
-0.0051
-0.0551
-0.0665
-0.0124
-0.0923
Std. Dsv.
Slope
0.0120
0.0179
0.0039
0.0079
0.0283
o.oioii.
0.0079
0.0187
0.0215
0.0305
0.0037
0.0363
0.006o
0.0036
0.0347
0.0174
0.0051
0.0057
0.0049
0.0118
0.0199
0.0155
0.0111
Correl.
Coeff.
.886
.366
.433
.630
.768
.204
.496
.805
•505
.740
.905
.746
.364
.807
.776
.352
.050
.230
.130
-571
.533
.158
.637
Avg.
G.F.D.
14.52
13.57
17.14
15.40
17.16
15.51
16. 04
36.02
17.76
32.37
13.64
28.82
13.27
13.46
26.35
15.72
14.65
15.02
14.91
12.20
io.o4
9.81
10.96
Avg. Etf.
Op. Hress. (P.S.I.)
669
537
506
536
525
533
523
490
6lO
513
574
530
571
580
539
517
514
534
520
516
537
545
544
-------
Table 80. pH ADJUSTED FEED WATER QUALITY, UNIVERSAL
WEEK
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
33
34-41
42-46
33-te
48-57
58-63
64-69
118-69
T.D.S.
-
786.7
2349.0
73^-5
720.0
1166.8
1238.1
71*3-0
710.0
733-6
7^.0
734-4
707.5
718-3
729-0
775-0
778.8
760.0
773-8
785-0
805.0
811.7
797-7
SPEC.
COHD.
1119.6
1256.5
1222.6
1309.2
1396-1*
1268.2
1287.8
1229-5
1192.0
1220.5
111*9-3
1211-3
1227.8
1201.6
1211*.!
1250.0
1212.8
1203-5
1203-9
1237-2
1242.8
1970.6
1^35-9
CHLORIDE
-
-
-
130.54
115-38
212.12
212.12
131-1*7
-
131- 47
-
131.1*7
-
-
131- 47
-
131.40
120.69
128.68
11*9.23
150.00
137-50
11*7-9
rag/1 ej
TOTAL
C.O.D.
28.96
22-53
12.30
9-15
5-80
13-61
9-76
6.88
6.00
6.63
6.70
6.64
-
6.15
6.52
6.09
10.83
8-33
9-71
34.0
41.6
56-9
43-7
ccept spec.
DISS.
C.O.D.
-
23-53
-
8.11
3.10
11.01
7-39
4.03
4.70
4.25
3-70
4.17
3-80
3-76
4.09
5-19
5-76
-
5-19
24.9
-
33-40
28.07
cond. as ml
TOTAL
HARD-
NESS
-
-
-
211.91
208.33
210.87
210.87
190-55
182-92
188.51
220.08
192.40
196.06
203.97
193-41
236.36
225.50
221.00
227-00
236.67
222.50
222.67
230.00
crorahos
OHTHO-P
-
-
-
2.50
0.86
2.17
2.17
6.27
6.70
6.41
10.09
6.86
6.26
7-51*
6-94
7-77
14.27
7-08
11.64
8.00
7-50
11.00
10-33
ALKA-
LINITY
-
-
-
46.97
-
46.97
46.97
57-65
85-76
64.64
-
64.64
32.14
32.14
58.24
-
81.45
32-33
65-24
41.83
68.05
51-53
52.08
CALCIUM
-
-
>•
60-57
55 -T4
59-15
59-14
54.10
43-01
50.93
58.03
53-16
50.99
53-35
51-71
-
-
-
-
65.44
57-14
69-50
60.75
HOj-N
-
-
-
4.16
4.80
4.25
-
-
-
-
-
-
-
5.09
-
0.60
10. 00
3-73
8.24
5-50
3-93
5-85
S<\
-
-
-
364.58
149-53
317-14
31T-14
290.85
321.21
290.85
3"*5-96
309.43
325-14
332-56
313-87
333-33
321.43
366.67
363-75
345-55
361.66
375-00
340.00.
-------
Table 8l. PRODUCT WATER QUALITY, UNIVERSAL
WEEK
DOS.
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
2-32
33
34-41
42-46
33-W
48-57
58-63
64-69
148-69
2-69
T.U.S.
-
100.0
122.0
100.6
70.0
104.5
lOJ.ll.
168.0
512.5
266.1*
530.0
299.4
605.0
580.0
360.5
-
30.0
1*0.0
25.0
35."*
35-5
21.7
25.0
28.9
-
SPEC.
COND.
103.0
86.7
238.4
170.2
155-0
167.4
190.6
434.4
832.0
548.0
893-0
591-1
1035-0
987.7
679-9
332-7
60.0
57-0
68.6
61.4
53-2
52.2
67.0
56.0
185-4
CHLORIDE
-
-
-
31.20
12.00
28.00
28.00
45-75
-
45-75
-
45-75
-
-
45-75
35-10
-
9-33
7.00
8.75
3-88
4.50
5-50
4.29
15-93
mg/l a
TOTAL
C.O.D.
1.100
1.667
1 .440
1.070
3-500
1-375
1-337
2-320
4.300
2.886
5.400
3.200
-
2-700
2.844
1.831
1.000
1.625
0.500
1.292
2.480
0-333
1-933
1-529
1.629
cept spec.
Diss.
C.O.D.
-
1.600
2,150
1.767
0.800
1.718
1.744
1.225
1.850
1-433
2.300
1-557
1-950
2.067
1.644
1.685
0.700
1.014
-
0.867
1.667
-
0.167
0.786
1.306
pond, ns m
TOTAL
HARD-
NESS
-
-
-
29-8
10.00
26.57
26.57
34.68
135-00
63-34
160.00
75-42
87-05
111-37
77-75
58-56
5-20
4.51
4.42
4.54
2-84
2.67
2.72
2.76
20.72
.crorahos
OKTHO-P
-
-
-
0.210
0.450
0.258
0.258
0.834
5.000
2.024
3.400
2.196
3-015
3-143
2.360
1.659
0.070
0.157
0.092
0.128
0.024
0.040
0.033
0.031
0.564
TOTAL
ALKA-
LIKIW
-
-
-
20.76
-
20.76
20.76
32.40
27-70
31.22
-
31.22
10.80
10.80
27-14
23-95
-
12-95
10.70
12.20
11.63
12.93
12.47
12.24
14-95
CALCIUM
.
-
-
5-33
3.40
4.85
4.85
9-36
28.30
14. rr
41.20
18.07
23.10
29-13
19.08
15.01
-
-
-
-
0.58
0.40
0.41
0.48
6.29
N03-N
.
-
-
2.25
3-60
2.42
2.42
l;10
-
1.10
-
1.10
-
-
1.10
2.27
-
0-33
2.60
1.09
1.66
1.10
0.71
1.14
1.56
SOl,
.
-
-
26.25
6.00
22.20
22.20
13-67
159-00
50.00
210.00
82.00
112.50
145-00
90.71
62.17
3-00
2.25
5-50
2.91
3-11
2.17
1.50
2.38
18.82
-------
Table 82. BRINE QUALITY, UNIVERSAL
WEEK
DOS.
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
33
34-41
42-45
33-46
48-57
58-63
64-69
48-69
T.D.S.
-
-
-
5330
-
-
-
1115
958
-
885
979
1155
993
1008
1275
1174
1243
1201
1458
1365
1258
1371
SPEC.
COND.
1962
1699
6035
3463
2285
3613
M82
2245
1738
-
1599
20lU
1504
1551
1955
2263
1932
1721
1898
2389
Sl8l
2007
2232
as/1
CHLORIDE
-
-
-
388
206
343
343
189
-
-
-
-
114
114
171
-
207
532
288
254
222
229
263
except spec
TOTAL
C.O.D.
-
-
-
123
10.6
94.8
94.8
5-3
4.0
-
5.4
4.6
15.3
12.0
7.2
17.2
13-3
11.8
13.4
58.6
68.0
71.8
66.9
. cond . as
TOTAL
HAfiD-
NESS
-
-
-
2200
-
2200
2200
359
219
-
-
275
309
309
289
337
352
365
365
446
4o4
390
419
micromhos
ORTHO-P
-
-
-
-
-
-
-
33.6
17.5
-
9-7
43-9
10.5
10.1
38.8
12.0
22.5
11.3
18.2
16.1
23.6
19.8
19.1
ALKA-
LINITY
-
-
-
-
-
-
-
34.0
-
-
54.0
50.7
34.5
34.5
44.2
-
78.3
52.3
69.6
61.2
80.7
88.8
74.1
CALCIUM
-
-
-
600
-
-
-
99.0
61.7
-
63.0
69.4
83.0
76.3
73.3
-
-
-
-
122
119
-
121
MOjH
-
-
-
8.2
6.2
7.7
7.7
3.9
0.8
-
0.8
1.9
1.6
1.2
1.8
1.2
2.7
.
4.3
14.0
7.8
7.0
9.6
sok
-
-
-
2700
-
2700
2700
-
-
-
-
-
-
.
.
-
-
.
-
739
593
-
704
-------
Table 83. RECOVERY AND TOTAL REJECTION RATIOS, UNIVERSAL
llj-9
WEEK
NOS.
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
33
34-41
42-46
33-46
48-57
58-63
64-69
fc8-69
WATER
RECOVER!
RATIO
.489
• 320
•799
.a3
•396
.608
.676
.430
.435
.428
.388
.1*23
.416
.407
.422
• 500
.363
•341
•365
.491
.424
• 366
•439
T.D.S.
-
•873
.948
.863
•903
.910
•915
• 774
.278
.637
.284
• 592
.145
.193
-506
.961
.949
•967
.954
•955
•973
•969
.964
SPEC.
CONG.
.908
•931
.805
.870
.889
.868
.852
.647
.302
• 551
.223
• 512
•157
.178
.440
•952
•953
•943
•949
•957
•958
•966
.961
CHLORIDE
-
-
-
.761
.896
.782
.782
.662
-
.662
-
.662
-
-
.662
-
•929
•942
•932
•974
• 970
.960
• 971
TOTAL
C.O.D.
•962
.926
.883
.883
•397 -.'
•899
.863
.663
.283
• 565
.194
.518
-
.561
.564
.836
.850
•940
.867
•927
•992
.966
•965
DISS.
C.O.D.
-
•932
-
.782
•742
.844
•764
.696
.606
.663
•378
.627
.487
•451
• 598
.865
.824
- .
•833
•933
-
•995
•972
TOTAL
HARD-
NESS
-
-
-
•859
•952
.874
.874
.818
.262
.664
• 273
.608
.556
.454
.598
• 978
.980
.980
.980
.988
.988
.988
.988
ORTHO-P
-
-
-
.916
•477
.881
.881
.867
.254
.684
.663
.680
.518
•583
.650
.991
•989
•987
•989
•997
•997
•997
•997
TOTAL
ALKA-
LINITY
-
-
-
•558
-
•558
•558
.438
.677
•517
-
•517
.664
.664
•534
-
.341
.669
.813
.722
.810
•758
• 765
CALCIUM
.
-
-
• 912
•939
.918
.918
.827
•342
.710
.290
.652
.547
.454
•631
-
-
-
-
.991
•993
•994
.992
N03-H
-
-
-
.458
.250
.430
.430
-
-
-
-
-
-
-
.784
-
.437
.740
.708
•798
.800
.819
.805
SULfATE
.
-
-
.928
•957
•930
-930
•953
.505
.834
•393
•735
.654
.564
-711
•991
•993
•985
•992
•991
•99*
•996
•993
-------
150
Table 84. UNIVERSAL WATER RECOVERY DATA
Weekly
P eriod
2-22
23-32
33-^6
48-69
Membrane
Set
1
2
3
4
Average
Recovery
Ratio
.608
.422
.365
.439
Standard
Deviation
.186
.053
.045
.079
No. Of
Data Pts.
21
10
14
22
80$
Confidence
Level
.36 - .85
.35 - .50
.30 - .42
.34 - .5^
The water recovery data were examined and it was found that 80$ of
the values for the weekly period 2-22 fell within the limits of 0.34
and 0.88, thus validating the statistically derived limits of 0.36
to 0.85. The median value was 0.56. There were two approximate modes
occurring at about 0.36 and 0.60. This would indicate a two-peaked
distribution curve with the first centering at about week 6 and the
second at about week 11.
It will be recalled that a portion of the brine flow was recycled as
unit feed from week 7 to week 11. This correlates to erratic material
balance ratios obtained during this same period.
Recirculation was undertaken for two reasons. First, it was hoped that
the overall water recovery ratio might be improved, even at the cost of
decreasing the A value; second, it was considered desirable to simulate
the conditions of using modules in a long series. The first objective
was realized, improving the recovery ratio from about 0.4 to 0.8.
Whether the second was attained is questionable.
There are three levels of interest for the total rejection ratios:
1. The ratio for a particular ingredient, such as T.D.S.,
for the entire test period of a given membrane set;
2. The variations observed within a particular membrane
group as related to the various types of pretreatment
applied to the feed;
3. The ratio of a particular total rejection to the equivalent
ratios for other impurities.
The total T.D.S. rejection ratios for membrane set 2 (the low solute
rejection, high flux type) were lower than Universal had anticipated.
The low ratios were particularly evident in weekly periods 28-29, 30 and
31-32 during which the feed pressure was increased from, about 500 to
600/650 psi in an attempt to improve the product water flux. The
manufacturer diagnosed the trouble as premature high-rate compaction
accompanied by multiple surface cracking in the thinner film. Under
these circumstances, the aforementioned "collapsing" would have occurred
even under minimal pressures such as 200 to 250 psi. It is of some
interest and to the credit of the Universal membrane that few changes
occurred in the total rejection ratios as reactor-clarification, sand
and activated carbon filtration were removed from the feed treatment
sequence.
-------
151
Average Rejection and Material Balance Ratios
Table 85 shows the calculated average rejection ratios for the major
chemical constituents and the corresponding material balance agreement
ratios. Deviations for the ratios are given in Appendix A-l.
The significance of the Material Balance Agreement Ratio (E) and its
use in detecting data inconsistencies .has been discussed in Section V.
Total C.O.D. and nitrate-nitrogen material balance agreement ratios
varied greatly between week 22 and week h6. Original record sheets
suggest that the brine samples and/or analyses were in error.
Uniform ratios among the other constituents, nearly discounts leaking
membranes as a cause for nitrate nitrogen and C.O.D. inconsistencies.
The fact that no membrane changes were necessary during this period
also supports this view.
Membrane Set No. 2, in service during weeks 23 to 32, showed high
constituent levels in the product water and low rejection ratios. Since
the set was designated as a loose, high flux type of membrane, this result
was not unexpected. The 2-32 and 2-69 "weekly groupings also reflect the
lower rejection conditions of weeks 23-32 but to a lesser degree, and
this should be remembered when analyzing the performance characteristics
of the Universal installation.
Membrane Fouling and Cleansing
The tubular Universal R.O. unit operated significantly longer between
membrane cleansings than the other R.O. units. Several reasons could
account for this operational result: (l) the product water recovery
ratios were generally low (Table 82). This lower product water recovery
ratio is partly due to reduced surface area for membrane sets 2 through !»-.
Also Table 75 shows that the product water fluxes in the Hemet tests were
substantially less than the product water flux specifications furnished
by Universal. Thus the brine flow rates were higher and probably well
into the turbulent Reynolds Number region (above 5,000 Reynolds Number).
(2) Pour different membrane sets were used. These rather frequent membrane
replacements helped reduce permanent membrane fouling and therefore reduced
the need for more frequent membrane rejuvenations. There may be some
operational advantage to the practice of replacing membranes at regular
intervals. However, more operational experience is needed to determine
whether there is a corresponding cost advantage to replacing membranes at
regular intervals.
The Universal unit's membranes were flushed forty-nine times with a Biz
solution, twelve times with a low pH water mix and once with EDTA. The
average per cent flux recovery for each membrane set and type of flush
is shown in Table 86.
-------
Table 85. AVERAGE REJECTION ARC MATERIAL BALANCE RATIOS, UNIVERSAL UNIT
H
VI
ro
Week
Nos.
2-3
4-6
7-11
12-21
22
2-22
7-22
23-27
28-29
23-29
30
23-30
31-32
30-32
23-32
33
34-41
42-46
33-46
48-57
58-63
64-69
48-69
Average Rejection Ratios
T.D.S.
.
-
.942
.954
-
.946
.946
• 775
.355
.565
.342
.521
.202
.243
.430
.971
• 959
.974
.964
• 971
• 979
.976
.974
Spec.
Cond.
.939
.943
-933
• 932
.922
.934
.932
.718
.370
.618
.271
."575
.202
.225
.500
.963
.962
.953
• 959
• 970
.969
.973
• 970
Total
C.O.D.
-
-
.973
.817
.942
.942
.512
.144
.389
.107
.349
-
.554
.430
.914
.850
.922
.873
.949
• 994
•970
.971
Total
Hard- •
Ness
.
-
-
-
-
_
_
.754
.309
.605
_
.605
•596
•596
.601
.983
.984
.985
.984
.992
.991
.991
.991
Ortho-P
.
-
-
-
-
_
-
.868
.306
.681
.61)4
.675
.598
.613
.658
• 993
• 991
.990
• 991
.998
.998
.998
•998
NO_-N
-
-
• 570
.374
.505
.505
_
-
_
-
-
-
-
—
.772
.752
.765
.868
.835
.864
.863
Material Balance Agreement Ratios
T.D.S.
-
.354
.881
-
.530
• 530
1.039
1.028
1.033
.997
1.026
1.028
1.018
1.027
.842
.964
1.006
.964
1.033
.963
•992
1.004
Spec.
Cond.
1.082
1.004
•917
1.007
1.143
•999
.987
1.057
1.008
1.043
•994
1.037
1.000
.998
1.030
.828
.987
.963
.967
.950
.972
• 959
.959
Total
C.O.D.
-
-
4.164
3.621
4.055
4.055
.653
.694
..667
.806
.686
_
1.674
.918
1.492
1.122
.966
1.128
.902
.877
.821
.864
Total
Hard-
Ness
.
-
-
-
-
_
_
1.144
1.009
1.099
*•
1.099
1.113
1.113
1.104
.824
.982
1.064
.996
1.038
1.019
1.036
1.033
Ortho-P
.
-
-
-
-
_
•
.909
•995
.938
.676
.900
1.051
.926
.934
.754
1.043
1.044
1.019
1.006
.996
1.060
1.019
NO--N
-
-
.865
1.140
• 957
.957
_
-
.
_
_
_
-
2.893
.838
2.208
1.016
1.029
1.177
1.105
-------
Table 86. MEMBRANE REJUVENATION RECORD, UNIVERSAL UNIT
153
Membrane Set:
No. 1,
No. Flushes/22 weeks
Avg. % Flux Increase
No. 2,
No. Flushes /10 weeks
Avg. % Flux Increase
No. 3
No. Flushes /l4 weeks
Avg. % Flux Increase
No. 4,
No. Flushes /22 weeks
Avg. % Flux Increase
Total No./l68 weeks
Avg. % Flux Increase
Biz
7
12
18.4
5
9.2
25
18.9
49
16.7
Acid Wash
•• -^^— ^— ^^^^— •••! , ,_.
8
1.6
1
0.9
0
"
3
25
12
1-9
EDTA
^^^•^•^••••••••••••iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii^^
0
~
1
0
0
"
0
"
1
0
U.S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460
202-566-0556
-------
15*
A Value - Time Plots
Figures 33 through 36 show the log-log A vs time (hours) plots for
membrane sets 1 through k respectively.
Figure 33 (membrane set No. l) shows a slight positive slope principally
because low A values were obtained between 166 and 670 hours (weeks 4-6).
Table 79 indicates that the A value during these three weeks averaged
only 1.9 x 10"' as compared with prior and later period averages of 2.0
to 2.^ x 10~5. The plant data shows that the pH acid feed pump was not
functioning properly on four separate nights. Thus the membranes may
have fouled. "Normal" fluxes were restored following the Biz flush at
week 7« Another reason for the increase in A values was the need to
replace membrane sets more frequently during the membrane 1 test period
(see Table 77-). These frequent membrane set changes were prompted by
membrane failures, probably the result of excessive vibration in the
Universal R.O. unit (discussed earlier). Frequent membrane replacement
may help in maintaining less steep log A-log Time slopes .
Figure 3^ for membrane set No. 2, a high flux type, shows that there was
an abrupt decline in A at about 685 hours (week 28). The manufacturer
was consulted and they expressed the opinion that the flux decline was
due to the migration of solute ions into the interior of the membrane.
They suggested that the No. 2 set be replaced by a tighter type. Shortly
before the membrane change at week 36, the flow pattern was changed (1013
hours) from "q" to "r". Although the flux decline was halted, the membrane
was irremediably fouled.
Figure 35 shows that membrane set No. 3 was much more suited to the study
feed conditions. Even the elimination of the reactor-clarifier from the
treatment sequence at 23^- hours had little effect in altering the A value.
Figure 36 indicates the performance for membrane set No. k, which used
both sand filtered, chlorinated, pH adjusted secondary effluent and
chlorinated, pH adjusted secondary effluent. The non-conformity of
flux data on Figure 3^ up to about 2000 hours (week 6l), is probably the
result of membrane fouling and subsequent cleansing operations. At
2*H9 hours (week 63) the slime-like material mentioned in Section VHI,
sub-section "Membrane Fouling and Cleansing", was found in the Universal
unit membranes. It is probable that the slime-like material had been
accumulating since about 2000 hours (week 6l), when it became a
major fouling agent and a cause for poor product fluxes.
As post secondary effluent treatment processes (reactor-clarification
granular activated carbon filtration, etc.) were removed, the quality of
the secondary effluent feed became more critical to the R.O. units
performance. Membrane set number 1 was tested on the highest quality
post secondary effluent treatment, i.e. reactor-clarification, sand and
granular activated carbon filtration, chlorination, and pH adjustment.
Removal of the reactor-clarifier did not result in ex tensive, permanent
flux decline. Similarly, the removal of granular activated carbon
-------
WEEK NUMBER
o
x
U
UJ
tf>
I
s
u
o
10.0
1
6
i
~
: 2
I
> 2 2
>
i <
3
, i ,
»
13 25
7
- 1.0
<
0.8
I-
< 0.6|-
UJ
z
<
o:
CD
UJ
S 0.2
O.I
AVERAGE CONDITIONS:
1. FLUX(GM/SQ.M.-SEC.)- 7.67
2. % WATER RECOVERY- 60
3. OPERATING FEED PRESSURE- 651 p.s.i.
4. BRINE PRESSURE - 460 p.s.i.
S. pH - 5.63
tOO 2
TIME (HOURS)
6 g 1000
« 10000
Figure 33. A vs. Time plotted logarithmically, Universal, 3/10/70 - 8/3/70
-------
WEEK NUMBER
10
o
«—«
X 4
I
U
LU
cn
I
o
V)
N
V)
s
0.8-
0.6
V)
z
o
Id
AVERAGE CONDITIONS:
1. FLUX(GM/SQ.M.-SEC.)- 12.9
2.% WATER RECOVERY- 34
3. OPERATING FEED PRESSURE 675p.s.i.
4. BRINE PRESSURE-518 p.s.I.
S. pH - 5.67
CD
0.2
o.t
tOO 2
TIME(HouRS)
8 1000
10000
Figure 3^. A vs. Time plotted logarithmically, Universal, 8/VjO - 10/5/70
-------
WEEK NUMBER
10"
i -
32
32 32 32
33
33
34 35 36
43
55
37
I
u
u
o
(/)
x
(/>
2
CP
< 0-8
H-
Z
< 0-6
Z
O
<-> 0.4
Ul
Z
<
cc
OQ
s
u
SLOPE=- .0051
AVERAGE CONDITIONS:
1. FLUX(GM/SQ.M.- SEC) - 7.33
2.% WATER RECOVERY- 39
3. OPERATING FEED PRESSURE- 539 p.s.i.
4.BRINE PRESSURE- 474 p. s. I.
5. pH - 5.62
8 100 2
TIME (HOURS)
I 1000
10000
Figure 35« A vs. Time plotted logarithmically, Universal, 10/6/70 - 1/11/71
H
VJ1
-------
10
WEEK NUMBER
48
48 48 49
49
49
50 52 53
54
60
H
VJl
CO
I
u
UJ
(/)
I
u i
a
1.0
0.6
06
O
o
0.4
o:
CO
Z
UJ
0.2
AVERAGE CONDITIONS'
1. FLUX (GM/SQ.M.- SEC) — S.61
2. % WATER RECOVERY-47
3. OPERATING FEED PRESSURE— 599p.s.i.
4. BRINE PRESSURE — 513 p.s.i.
5. pH— 5.69
0.1
8 100 2
TIME (HOURS)
I 1000
8 10000
Figure 36. A vs. Time plotted logarithmically, Universal, 1/27/71 - 6/25/71
-------
159
filtration did not result in extensive permanent flux decline. It
wasn't until both the reaetor-elarifier and granular activated carbon
post secondary effluent treatments were removed that extensive
permanent flux decline occurred (Figure 36) and membrane rejuvenation
treatments became less effective in recovering product water fluxes.
Both the spiral wound and hollow fiber membrane configurations were
affected extensively by the removal of either the reactor-clarifier
or granular activated carbon post secondary effluent treatments.
These results suggest that the advantage of a tubular design over
the more densely packed spiral wound or hollow fiber membrane
configurations lies in the tubular configuration's capability of
tolerating a poorer quality wastewater.
A major upset occurred in the biological treatment plant between
weeks 59 and 60. Consequently, all R.O. units were shut down during
this period. Plant records indicate that the condition was not
fully rectified in the biological treatment plant following resumption
of R.O. studies at week 60. Thus, there is a strong possibility that
suspended solids were responsible for depressed fluxes toward the
end of the project.
-------
SECTION XV
INTER-UNIT COMPARISONS
Introduction
If a comparison is to be made between the R.O. units used in this study,
it should be in general terms. It would be inappropriate to compare
the units on a specific point-by-point basis since process scheduling
and even financial considerations, existing throughout the project, made
it impossible to operate all units under the same exact conditions for
more than a few brief periods. Some installations and membrane
replacement deliveries were delayed. Membrane failures sometimes
occurred at critical periods and in these cases replacement modules
had to be installed or, in the case of non-availability, the total
membrane area had to be reduced. At times, it was necessary to renew
modules on one unit while the other units continued to operate under the
same post secondary effluent treatment conditions, but with older, fully
compacted or partially fouled membranes.
One of the most important findings of this pilot study, however, was
that all unit membranes were subject, in various degrees to "Rapid
Membrane Fouling." "Rapid Membrane Fouling " is defined as short term
or temporary impairment of membrane water permeability (caused by
organic and inorganic deposition) which may be eliminated by chemical
or physical rejuvenation processes. The rate of fouling is influenced
by three factors: l) the degree of post secondary effluent treatment,
2) membrane formulation characteristics as determined by casting technique
and, 3) the flow characteristics within the unit module. It now appears
that the periodic rejuvenation procedure is necessary to all R.O. units
with the provision that some require it more frequently or for longer
durations. Rejuvenation requirements of the pilot study R.O. units are
compared below.
Unit Reliability
Table 87 indicates on a unit basis, the hours needed for cleaning and
membrane replacement per 100 available operating hours. The values are
derived from the maintenance records found in the specific unit discussion
sections and Appendix Section A-3. The maintenance time allotment is
probably one of the important parameters of reliability. Other out-of-
service causes such as mechanical problems were not included as they
could represent conditions not directly involved with water transport
across a membrane. Thus, if a pump failed, it was either repaired or
replaced. (This example does not suggest that finding suitable pumps
for reverse osmosis units is always a simple task, but it does assume
that high pressure pump technology is relatively advanced compared to
reverse osmosis..)
Table 87 can best be understood by using the following discussion as a
guide:
160
-------
Membrane cleansing - for membrane cleansing, the tubular designs,
American Standard and Universal, needed substantially less time than
the spiral wound or hollow fiber module designs. This indicates that
the tubular design tends to remain free of deposition and/or scale
and that buildup is curtailed probably because of better flow
distribution characteristics. The Aerojet tubular unit required the
fewest hours for cleansing, but it was also the unit with the shortest
operating span, next to the Raypak tubular unit. The Aerojet unit
operated only l4 weeks out of the 69 week study period because of a
high membrane failure rate. It would be expected that any R.O. unit
would require very little or no cleaning during the brief initial full
post-secondary effluent treatment period.
The denser membrane packing of the spiral wound configuration is more
susceptible to particulate fouling and more difficult to clean. It is
possible that the present hollow fiber cleansing time is less than that
indicated in Table 87. The B-5 modules (replaced later by the newer
designed B-9 modules) operated on the best post secondary effluent feed.
By the time the B-9 modules were installed, the R.O. feed water quality
was reduced because of the removal of post secondary effluent treatment
processes. However, the cleansing time was essentially the same for
the B-9 modules operating on the poor quality feed water as for the B-5
modules operating on the better quality feed. This suggests that the B-9
module could be subject to less fouling, especially from particulates.
Table 87. MAINTENANCE TIME ALLOTMENT/100 AVAILABLE OPERATING HOURS
R.O. Make and Type
AEROJET, TUBULAR
AMERICAN STANDARD'
TUBULAR MODIFIED
DU PONT, HOLLOW FIBER
GULF, SPIRAL WOUND
UNIVERSAL, TUBULAR
Membrane
Cleaning
0.13
0.46
4.32
1.11
0.68
Membrane
Failures
11.64
2.36
0.00
0.00
1.78
Total
Hours
-
2.82
4.32
1.11
2.46
-------
162
Another factor to consider in the hollow fiber cleansing experience
is the use of cartridge filters on feed water entering the modules.
These filters probably slowed the rate of fouling. For this reason
cartridge filters are included in the Du Pont reverse osmosis costs
of Section XVI.
The second category in Table 87 relates to the hours devoted to membrane
failure. Tubular membranes failed more frequently than either the
spiral wound or hollow fiber configurations.
The last column lists the total hours devoted to the combined forms of
maintenance. The spiral wound design required the fewest total hours
for overall servicing. (The Aerojet tubular design was omitted because
of the abnormally high membrane failures and early withdrawl from the
test study). It is largely because of cleansing that the hollow fiber
configuration had the highest total maintenance time.
Water Permeability and Solute Rejections
Tables 88 through 92 compare R.O. data for various post secondary
effluent treatment feeds. An outstanding feature is the stability of
the rejection ratios regardless of the post secondary effluent treatment.
This does not infer that the quality of the product water remains
constant. Any solute removed during post secondary effluent treatment
results in a lower concentration of solute in the feed and product
water; however, the per cent solute removal across the membrane is
fairly constant regardless of the feed water solute concentrations.
Table 93 is & summary tabulation that can be used to determine the
performance differences between module configurations. The average
performance for each membrane configuration was calculated from
Tables 88 through 92 for the various R.O. feed waters. The Du Pont
B-9 module represents the hollow fiber module concept. Erom the data
in Table 93> it appears that the tubular R.O. concept is less subject
to "rapid" and "permanent" membrane fouling than either the hollow
fiber or spiral wound configurations, especially in the presence of poor
feed conditions.
Table 93 also illustrates the advantage of the tubular designs of operating
at a low water recovery ratio. Universal and American Standard R.O.
units had relatively gentle A value slopes which is probably the result
of low water recovery ratios. The effect of the turbulence promoters
in the American Standard tubular unit cannot be determined since there
was not a control unit without the turbulence promoters. In providing
higher Reynolds numbers, the turbulence promoters through slight, but
constant shifting (abrasion) on the membrane surface could have been
the major cause of membrane failure and some inconsistent data.
-------
Table 88. OPERATIONAL VARIABLES AND VALUES
FEED TREATMENT: REACTOR -CLARIFICATION, SAND, AND ACTIVATED CARBON FILTRATION (W/CHLORINATION AMD pH ADJUSTMENT)
Unit
Aerojet General
American Standard
Du Pont
Gulf
Universal
Weekly
Period
7-12
19-28
»H-li6
21^-33
l»2-l»8
8-23
te-l»6
2-22
23-32
I|2-lt6
Membrane
Set
1
1
1
1
2
1
1
1
2
3
Average
Feed
P.S.I.
565
605
635
650
too
600
605
6^5
575
605
Water
Recovery
Rate
.U21*
.805
.689
.770
.673
.6k8
.565
.608
.1*22
.3M
TDS Total
Rejection
Rate
.7^
.785
.818
.59^
.886
.935
.889
.910
.506
.967
. A VB T
Slope
+.05^5
+.0232
-.OOOU
+.0026
-.0179
-.ote9
-.0006
+.0197
- .2771
-.0062
-------
Table 89. OPERATIONAL VARIABLES AND VALUES
FEED TREATMENT: REACTOR-CLARIFICATION PLUS SAND FILTRATION (W/CHLORINATION AND pH ADJUSTMENT)
Unit
Du Pont
Gulf
Universal
Weekly
Period
^9-53
55-57
k7-W
^9-57
W-57
Membrane
Set
2
3
1
2
1*
Average
Feed
P.S.I.
kuo
M5
620
560
600
Water
Recovery
Rate
.565
.566
.l£2
.586
A91
TDS. Total
Rejection
Rate
.865
.891
.890
.929
.955
A vs T
Slope
-.0060
-.0875
-.0271
-.19^1
-.0551
-------
Table 90. OPERATIONAL VARIABLES'AND VALUES
FEED TREATMENT: SAND AND ACTIVATED CARBON FILTRATION (W/CHLORINATION AND pH ADJUSTMENT)
Unit
Aerojet General
American , Standard
Du Pont
Gulf
Universal
Weekly
Period
35-37
38-^0
3^-37
38-41
3^-M
Sfc-lH
Membrane
Set
2
1
1
2
1
3
Average
Feed
P.S.I.
555
630
660
M5
600
590
Water
Recovery
Rate
-.556
.756
.7^1
.68h
.623
.363
TDS Total
Rejection
Rate
.862
.811
.652
.933
.911
.9^9
A va T
Slope
-.CA-22
-.0099
-.0069
-.0^59
-.OMl2
-.0015
VJ1
-------
Table 91. OPERATIONAL VARIABLES AND VALUES
FEED: SAND FILTER, CHLORINATED, pH ADJUSTED SECONDARY EFFLUENT
Unit
American Standard
Du Pont
Gulf
Universal
Weekly
Period
61-63
58-63
58-63
58-63
Membrane
Set
2
3
2
k
Average
Feed
P.S.I.
635
11-05
610
590
Water
Recovery
Rate
.72^
•515
.lj-02
.tefc
TDS Total
Rejection
Rate
.898
.81tf
.922
•973
A vs T
Slope
- .103U
-.0780
-.1095
-.0665
-------
Table 92. OPERATIONAL VARIABLES AND VALUES
FEED: CHLORINATED, pH ADJUSTED SECONDARY EFFLUENT
Unit
American Standard
Du Pont
Gulf
Universal
Weekly
Period
64-69
64-69
66-69
64-69
Membrane
Set
2
3
2
4
Average
Feed
P.S.I.
635
420
600
600
Water
Recovery
Rate
.527
.351
.290
.366
TDS Total
Rejection
Rate
.930
.766
.913
.969
A vs T
Slope
-.0669
-.0270
-.0702
-.0124
-------
o\
00
Table 93- COMPARISON OF MEMBRAJffi CONFIGURATIONS
(weeks 38-69)
R. 0. Unit
American Standard
Du Pont
Gulf
Universal
Membrane
Configuration
Tubular
Hollow Fiber
(B-9 Module)
Spiral Wound
Tubular
Avg. Recovery
Ratio
0.652
0.579
0.486
0.413
Avg. TDS Total
Rejection Ratio
0.879
0.884
0.937
0.971
Avg. A vs.
Time Slope
-0.049
-0.139+
-0.200
-o.o6o4
-------
169
There are compensating factors of non-tubular R.O. unit designs that
can offset what appears to be the advantage of the tubular configuration.
The hollow fiber and spiral wound designs' advantages could be: (a)
lower membrane failure rate, (b) a more compact unit, (c) smaller space
allotment, (d) lower overall production costs, and (e) a possible
reduction of permanent membrane fouling tendency.
Table 9^ shows the average per cent reductions of constituents for
the three main membrane configurations and types of membrane material
(formulations). The period from week 38 to week 69 was used for this
comparison since the ^u Pont B-9 modules were used after week 38. The
major post-secondary treatment sequences during the period were: (l)
sand filtration plus granular activated carbon treatment, (2) chemical
clarification followed by sand filtration, (3) sand filtration only,
and (10 no treatment at all. Each of the latter treatment sequences
received the usual pre-R.O. chlorination and pH adjustment.
The data indicates that the cellulose acetate membranes are more
efficient than the polyamide for removing TDS, dissolved C.O.D., total
hardness, calcium, ortho phosphate and sulfate. In all constituent
categories, the Universal cellulose acetate membranes equaled or
surpassed the rejection performance of the polyamide fibers. However,
the polyamide fibers were more efficient for rejection of total C.O.D.,
nitrates and chlorides than the Gulf cellulose acetate membranes.
Product Water and TDS Rejection Ratios
If reverse osmosis is used for ground water recharge purposes, some
blending of the product water with treated or untreated secondary
effluent is possible. The four reverse osmosis units listed in Table
93 were originally specified to produce 10,000 GPD product water.
However, the effects of compaction, concentration polarization and
membrane fouling quickly reduced the product water recovery to something
less than the nominal 10,000 GPD specification. A mathematical
"adjustment" of the membrane area can therefore be made in order that
the nominal 10,000 GPD specification be met. The "adjustment" is
necessary for making R.O. cost estimates. A second adjustment is also
possible for the TDS rejection ability of an R.O. unit (see Table 95).
The values in Table 95 are based on the pilot plant data and are valid
provided the following assumptions are made:
(a) 90$ TDS rejection
(b) the minimum volume of R.O. product water is blended
to produce a 500 mg/1 TDS water for ground water
recharge
(c) the selected product water volume rates are representative
of the best possible output by each R.O. unit (usually
when first installed).
The data in Table 95 represents the best estimate that could be made for
determining capacity discrepancies which may be encountered in reverse
osmosis applications. These values can and will probably differ from
-------
Table 94. AVERAGE PER CENT REDUCTIONS OF CONSTITUENTS
Reverse Osmosis
Configuration
Du Pont Hollow Fiber
B-9 Modules Polyamide
Fibers
Gulf Spiral Wound
Cellulose Acetate
Universal Tubular
Cellulose Acetate
TDS
88
<*
"
97
Total
COD
90
88
93
Diss.
COD
90
92
96
Total
Hardness
95
98
99
Ca
96
99
99
Ortho-P
93
99
"
99
N03-»
84
55
84
Ammonia-
Nitrogen
•"•
96
•
99
ci
91
88
97
so4
9*
99
99
NOTE: Per cent reductions chosen from weeks 38 to 69, to compare these membrane configurations on the same
feed water. Data extracted from program data, exemplified in Table 20.
-------
Table 95 • MINIMUM VOLUME INCREASE REQUIRED PER UNIT TO MEET SPECIFIC DEMAND - 10,000 gpd @ 90$ REJECTION
Unit
American
Du Pont
Gulf
Universal
Discrepancy Between
"Nominal"
Production
(10,000 gpd)
and Observed
Production
(GPD)
-1252
+2360
-26^5
-1288
(*)
-1^
+19
-36
-15
Additional Product
Necessary to Compensate
For Discrepancy Between .
"Ideal" (90$) and Observed TDS
Rejection
(*)
+13
+27
-
+ 5
Overall Volume
Increase Needed to
Produce a Product
Analogous to 10,000 GPD
At 90$ Rejection
»)
+31
+ 3
+36
+21
•pj
-------
172
other installations. Evidence shovs that a 90$ product water recovery
and 90$ solute rejection could be stretching reverse osmosis performance-
to an upper limit vhen operating on a non-recirculation basis.
Experience "with the nominal 10,000 GPD American Standard tubular unit
indicated severe fouling when attempting to achieve 90% product water
recovery operating on a once through basis (Section X). The fouling
was partially overcome by operating at a lower (near 80$) product
water recovery level. Calcium salts, particularly phosphates and
sulfates which have low solubilities, probably scaled the membrane
surfaces when operating over the 80$ product recovery ratio. However,
it is also possible that some 10,000 GPD reverse osmosis units do not
provide for sufficiently high Reynolds number after 80$ product water
is recovered. It would be helpful to have information on larger units
for each manufacturer to determine cost declarations and operation ease
on a magnified scale. The smaller R.O. units were closer in keeping
with the objectives and funds available for the study.
Relative Membrane Product Recovery Losses
Table 96 shows how long the unit membrane sets may be expected to
operate under various feeds. The life spans were based on the rate of
recovery loss per week. Each recovery loss value was determined using
two selected original values of recovery. To eliminate errors due to
process changes, the two values were never separated by a process change
other than one membrane flush. Because both base values were taken in
the 3-5 day period after a membrane "Biz" flush the rate of loss as
recorded in column A is considered to be the "permanent recovery loss
rate" parameter. After establishing the rate of recovery loss, it was
relatively simple to determine parameter B which was based solely on
Parameter A. The actual contribution of recovery losses to the total
operating costs will be discussed in the next section.
Table 96 is most valuable in evaluating relative membrane life and not
as indicating actual membrane life. The actual membrane life evaluation
should have more constant operating conditions. However, Table 96
relates the effect of various operating conditions to a theoretical
estimate of membrane life. For example, the Universal tubular unit shows
a relatively short membrane life for the highest quality feed water
(horizontal groups 1 and 2). The reason for this is Universal membrane
set No. 2 was used during the test interval. The high product water
flux-low solute rejection membrane showed very rapid compaction, fouling,
and disintegration which resulted in a rapid decrease in product water
flux (slope of A vs T » See Figure 3^). The effect of changing to the
low product water flux-high solute rejection membrane (set k) resulted
in several magnitudes of increase in the membrane life span even with
the application of a lower quality feed water.
Table 96 also shows that there was very little permanent loss of pro-
duction in the Gulf spiral wound configuration until granular activated
carbon filtration was removed from the post-secondary effluent treat-
ment sequence. During the episode of no post secondary treatment except
-------
Table 96. ESTIMATE OF MEMBEIANE LIFE BASED ON PRODUCT WATER RECOVERY LOSS
Post-Secondary
Effluent Treatment
Sequence and Unit
A,B,C,D,E,F
American Standard
Du Pont
Gulf
Universal
A,B,C,E,F
American Standard
Du Pont
Gulf
Universal
A,B,E,F
American Standard
Du Pont
Gulf
Universal
B,E,F
American Standard
Du Pont
Gulf
Universal
E,F
American Standard
Du Pont
Gulf
Universal
Rate of Permanent
Product Water
Recovery Loss Per Wk.
(A)
-2.38*
No data on B-9's
No change
-1.65)1
-0.78^
-2.1$
No change
-5«23# 3
No data
-5.87#
-2.00$
-0.87$
No data
-3-95$
-4.25$
-0.35#
-5-57$
-9-57$
-0.44$
-2.34$
TOTALS (Simple Averages)
American Standard
Du Pont
Gulf
Universal
-2.9056
-4.55$
-1.34$
-2.09$
Membrane
Set No.
1
-
1
1 & 2
1
2A
1
-,2 & 3
-
2A
1 & 2
4
_
2B
2
4
2
2B
2
4
Theoretical No. Weeks
For Membrane Set To
Experience 25$ Loss
of Product Recovery
(B)
11.50
_
_
15-15
32.89
10.42
_
4.78
_
4.26
12.50
28.74
_
6.33
5.88
71-43
4.49
2.61
56.82
10.68
8.62
5.49
18.66
11.96
Theoretical No. of Membrane
Sets Which Would Be Necessary
To Replace Per Week Using Maximum
25$ Permanent Recovery Loss Criterion
(c)
0.09
0.01
0.07
0.03
0.10
0.01
0.21
_
0.23
0.08
0.03
f
0.16
0.17
0.01
0.22
0.38
0.02
0.09
0.12
0.18
0.05
0.08
-------
pre-R.O. unit chlorination and pH adjustment (treatment, E, F on
Table 96), the Gulf unit appears to have experienced only slight loss
on product water recovery. The value is misleading, however, unless
it is realized that the level of recovery at that time was only 29
per cent.
-------
SECTION XVI
REVERSE OSMOSIS COSTS
Introduction
The feasibility of using reverse osmosis to remove constituents
from treated or untreated secondary effluents was the primary purpose
of this study. The study was planned so that the data might be used
to estimate the economics of reverse osmosis in a local ground water
recharge program. The many project variables and details made the
costs difficult to estimate. Nevertheless, three different cost
estimates will be presented: two were prepared by cooperating R.O.
manufacturers (Du Pont and Gulf) and one was prepared by the project
personnel at Hemet, California, based upon the study data and
experience elsewhere.
Feed Treatment Requirements
Two separate, yet related, factors need to be considered in any
reverse osmosis application; flux (water permeation) rate and solute
removal. The water permeation rate is affected by such variables as
membrane characteristics module configuration, impurities present in
the feed water, the operating pH level, the amount of feed water
chlorination, the designed product water quality, etc., in addition
to the more direct variables such as temperature and the R.O.
operating pressure.
Solute removal is influenced primarily by membrane formulation
characteristics. In most instances a low solute concentration in the
feed water to a reverse osmosis unit results in higher water permeation
rates, a better quality product water and diminished membrane fouling
problems.
A cost trade-off exists between the extent of feed treatment and
reverse osmosis unit maintenance. With the higher cost of treating
secondary effluent, there is a corresponding decrease in membrane
replacement, rejuvenations and fouling tendencies. An improved
feed quality is usually accompanied by better product quality,
quantity and lower brine volume, all important cost factors.
When treatment of secondary effluent is considered solely from the
standpoint of beneficiating the reverse osmosis process, opinions
become divided as to what and how much pretreatment is economically
justifiable. The spiral wound R.O. manufacturer (Gulf) suggested
that only sand filtration, chlorination and pH control are essential
post secondary effluent treatments. The hollow fiber manufacturer
(Du Pont) advises that granular activated carbon filtration,
polyphosphate addition and ten micron cartridge filters should be
175
-------
176
added to the post secondary effluent treatments. It is unfortunate
that no tubular unit manufacturer was prepared to specify a post
secondary effluent treatment sequence with a cost estimate.
The data from this study suggests that the untreated secondary
effluent quality has little to do with selection of the best secondary
effluent treatment sequence. Instead it was determined that a
continuously good quality feed with low rate of membrane replacement
can be maintained using reactor-clarification with chemical
coagulation and sand filtration and provided that Reynolds number
is maintained in the J-i-,500 - 5,500 range or its equivalent.
Reverse Osmosis Cost Factors
As mentioned in Chapter X? membrane life determination was not a
primary objective of this pilot study. It is necessary, however, to
have at least a "ball park" estimate in order to arrive at any R.O.
cost estimate. There are a number of reasons for assuming a two year
membrane life: (l) the spiral wound R.O. unit operated successfully
for over a year on only two membrane sets using high and low quality
feed: (2) it appears that the hollow fiber unit could have matched the
spiral wound unit for membrane longevity had B-9 modules been used at
the beginning of the study and (3) a municipal water improvement
facility applying R.O. to brackish potable water is assuming a two
year membrane life under guarantee. The last reason is valid because
the initial feed waters at Hemet were of better physical and chemical
quality than many brackish municipal supply waters. As a result of a
good quality feed, a low fouling tendency and long membrane life can
be expected. The lifespans as calculated in Table $6 of Section XV
were not used as cost indicators since conditions were so variable.
Instead they are most important for comparative analysis.
Table 95, however, was used to establish whether size increases are
always necessary to meet the nominal capacity of the R.O. plants. The
Du Pont hollow fiber unit fulfilled its obligation by operating within
3$ of the nominal capacity. Therefore, it was unnecessary to apply
a volume/TDS rejection adjustment to the following "study" cost
estimate.
Reverse Osmosis Cost Estimates
(These cost estimates are based on the first quarter of 1972 MR Index.)
I. Project Estimate:
An estimate was made using the project data for a minimum 0.8 mgd
and possible 0.9 MGD product water facility using 1 MGD feed water.
For this cost analysis, reactor-clarification plus sand filtration
was chosen over sand filtration plus activated carbon filtration
as the post secondary treatment sequence. The latter sequence is
somewhat more efficient but the former sequence is usually
substantially cheaper and still provides a high degree of treatment.
-------
ITT
A. Basic Assumptions
1. The secondary effluent reverse osmosis feed water will
first be chemically treated in a reactor-clarifier by
either alum or ferric chloride, sand filtered,
chlorinated and pH adjusted.
Note: It would be preferable to first install pilot
units of Du Pont, Gulf or Universal manufacture
and operate them for a period of six to twelve
months to determine:
(a) whether ehemical clarification or granular
activated carbon treatments are needed
continuously in the full scale plant and
(b) which make of unit should be selected for
the second stage expansion.
2. Plant Operations
a. Operate 2k hours per day, 7 days per week
b. An estimated 10$ down time (including membrane
cleansing, replacement, repairs, etc.)
c. Membrane replacement every 2^ months
d. Feed pressure ^00 psi
e. Complete automation of all repetitive operations
with flow and pressure recorders, and controllers,
chlorination, pH and conductivity recorders,
automatic sampling devices, time sequential valve
operators, automatic shut down and diversion
facilities, automated sulfuric acid and flushing
solution handling and make up systems, alarms, etc.
f. With the above listed automation, the following plant
and laboratory personnel would be required:
Supervisor - 0.2 man years $ 3,000/year
Chemist (l): 10,000/year
Plant Operators (2) 18,000/year
Relief Operator and Chemist (l) 10,000/year
Instrument Man 80 Mechanic - 0.3 nian years k,OOP/year
Subtotal $ ^5,000/year
Assume 20$ for overhead 9.OOP/year
Total Labor Costs $ 54,000/year
-------
178
g. All post secondary effluent treatment and reverse osmosis
equipment will be located out-of-doors on well drained
concrete pads around a paved heavy equipment access road.
Included in the complex will be a centrally located 200
sq. ft. roofed "and air conditioned building to serve as
a combined operators record and instrument room. (Based
on a estimated rating of 2,000 gpd per sq. ft. the paved
equipment area, exclusive of the operators room and
access road, will be about 6,000 sq. ft.)
h. The existing reverse osmosis building on the site will
house the warehouse, instrument repair room and shop,
the conference room and plant foreman's office,
laboratory, washroom, and lunch room.
i. A Pelton wheel driven pump will be provided on the
brine stream to recover about 10$ of the reverse
osmosis unit power. (Note: Although the pay-off on an
800 K gpd rate is only about 10$ per year, it is
included in this estimate because power recovery will
be increasingly desirable should the projected plant
be enlarged or operated at lower recovery,rates.)
j. No charges are included in this estimate for the
following:
(l) The existing reverse osmosis building
(2) Secondary effluent delivery system
(3) Waste brine disposal
(4) Land Cost (or rental)
Comments: Many of the items listed under (A-2) are
inadequately defined at this time. The
major cost items (reverse osmosis unit
costs, tested membrane replacement time
intervals - and costs, labor requirements,
etc.) are so affected by unestablished
design details (examples - trade off costs
between automation and labor costs), quantity
discounts, development of improved membranes,
etc., that they cannot be estimated at this
time. Thus the following cost estimate is a
rough approximation using information derived
from the EMWD pilot plant studies:
B. Capital Costs
Cost
1. Reactor-clarifier and clearwell $100,000
(overflow rate 0.5 gpm/sq. ft.)
2. Sand filters 30,000
3. Pumps and Piping 50,000
k. Secondary Effluent Treatment Setup 5,000
5. Reverse Osmosis and Field Setup 6,500
-------
,. 179
6. Site Preparation 2,500
7. Paving and Drainage 2,000
8. Utilities 23,000
9. Instrumentation other than R.O. 50,000
10. Power Regeneration 15,000
11. Chemical Handling 50,000
12. pH Control 15,000
13. Chlorination 10,000
Ik. Remodeling existing building 5,000
15. Reverse Osmosis unit 500,000
16. Product water storage (100,000 gal.) 12,000
17. Erection and Assembly 32,000
(l) Sub-Total $908,000
Engineering (approximately 7-65$, ASCE
Man. and Reports on Engineering
Curve A, 1971) 70,000
(2) Sub-Total $978.000
C. Other Expenses
1. Indirect Field Labor 5,000
2. Home Office Cost 2,000
3. Start up Costs 20,000
k. Contingencies - including working capital,
interest during construction, 5$ of
sub-total (2) 49,000
$1,05 *!•, 000
Less: First Set Membranes (Expensed) 150,000
Total Capital Costs $ 90^,000
-------
180
Amortization:
20 years @ 6% - yearly cost
Factor is 0.0672 $ 78,828/yr.
Cost per 1,000 gallons product water
based on 0.9 mgd product water: 24.00
Comment: If the amortization used is 10 years
at 6$, the yearly cbsl; factor is 0.1359
or $122,854/year and 37-40/1,000 gallons.
D. Operation and Maintenance Costs
Item Yearly Cost Cost/1,000 gal.
1. Labor - (ind. O'H) $ 54,000 16.40
2. Electric Power 14,800 4.50
(after credit for regeneration)
3. Membrane Replacement 75,000 22.90
($150,000 for 2 year membrane life)
4. Plant Chemicals 27,900 8.50
5. Consumable Supplies 6,500 2.00
Total $ 178,200 54.20
Grand Total of Costs - - 5^-20 + 24.00 = 78.20/1,000 gallons
Comment: If 10 year amortization life is used, total
cost is 37.40plus 5^-20 or 91.60/1,000 gallon
product water.
Notes: a) If the reactor-clarifier .were not installed,
it is estimated that the capital costs would
be reduced to about $804,000 and the total
cost to about 75.50/1,000 gallons (21.30 +
54.20/1,000 gallons). However, if activated
carbon filters were added to the post secondary
effluent sequence (reactor-clarifier, sand
filters, pre-R.O- unit chlorination and pH
adjustment) the capital cost for R.O. would
be increased to $982,000 and the total cost
would be 84.90/1000 gal. including the cost
of media at 4.60/1000 gallons.
-------
181
b) The following assumptions would apply to a
plant with a 9 million gallon capacity:
(l) Capital costs would be adjusted by an.0.8
exponent factor,
(2) The $5!!.,000 annual labor cost would be
no higher for a 9 MGD plant than the 0.9 MGD
plant.
(3) The power, membrane, chemical, and supply costs
remain constant per 1000 gallons of product
water. The estimated cost for the larger
facility would then be 19.205 capital cost
plus 5^.2^ O&M cost for a total cost of
about 73^/1000 gallons., An estimate for a
10 MGD plant would not apply to the Hemet
facility since the present daily flow rate
there is less than k MGD.
c) The estimated cost for blending secondary effluent
with R.O. product water at Hemet is estimated from
the TDS in the secondary effluent (Table 3) which
averages 716 mg/1 and 72 mg/1 TDS in the R.O.
product water to produce a blended water with 500
mg/1 concentration.
Let X = R.O. product water .
1-X = Secondary Effluent
Total = 1 unit
X(72) + (1-X) (716) = 1 (500)
72X + 7l6 - 7l6x =500
6lrtx = 2l6
X = 0.3^
Thus 3*$ R.O. product water can be mixed with 66$ secondary
effluent to produce a blended water at 500 mg/1 TDS. Assuming
no added cost for the actual blending operation, the R.O. cost
portion is 0.34 x 78.2^/1,000 gallons or 26.6$ per 1,000
gallons; and 0.9 mgd R.O. product water can be blended with
1.7 mgd secondary effluent to produce 2.6 mgd blended 500 mg/1
TDS water for groundwater recharge.
II. Gulf Environmental System Cost Estimate:
The cost estimate for the spiral wound module offered by Gulf
Environmental Systems based on this study data and the
manufacturer's experience follows:
-------
182
A. 1,000,000 gallons product water per day from secondary effluent
B. 75^ recovery, 97% conductivity rejection, at 1|-00 psi
C. Three groups of units, each with an individual feed pump
D. "Adequate" instrumentation
E. 90 to 9 $f> running time
P. Cost of first set of modules included
G. Three year membrane life (replacements chargeable to expense)
H. Wo power regeneration
I . Automated membrane flushing operations
J. Total labor-operation and maintenance (no chemists) 1000 hours
total per year
K. Ten year amortization period
L. Wo exterior piping
M. Brine disposal not included
The above assumptions gave the following estimates:
1. Capital Cost $14-50,000
2. Operating Costs
gallons
Power 8
Membrane replacement 15
Cleaning chemicals 6
Operating chemicals 6
Operating labor 1.5
Sub -Total 3o"3
Amortization 8.2
Total 45.Q0/1,000 gallons
III. E. I. Du Pont de Wemours Cost Estimates
The cost estimate for the hollow fiber module offered by E. I. Du
Pont de Wemours based on this study data and the manufacturer's
experience follows:
1. 10,000,000 gallons product water per day from secondary
sewage
-------
1&3
2. 80$ recovery, 90-95$ rejection at 1*00 psi
3. Ten 100 unit sets of permeators, each of the ten divided
into two 50-unit control blocks
k. Instrumentation
5. Cost of first set of permeators included
6. Monthly replacement of filter cartridges
T. No power regeneration
8. Membrane flushing controls unstated
9. Labor: four operators and one each of the following:
analyst, instrument and maintenance man, helper, clerk-
typist, supervisor. Total nine employees
10. Amortization -- see below
11. Wo exterior piping
12. Brine disposal not included
The estimated capital costs follow:
1. Permeators (1000 - 8 in. diameter) $200,000
2. Fiber 1,800,000
3. Supports and integral piping 100,000
h-. Pumps and motors (five sets) 200,000
5. Battery limits piping 200,000
6. Cartridge filters and cleaning equipment 200,000
T. Electrical 100,000
8. Instrumentation 100,000
9. C12, Acid and polyphosphate piping 30, OOP
Total direct cost $2,930,000
10. Engineering 370,000
Total reverse osmosis cost $3^300,000
-------
11. Pretreatment (including sand and carbon
filters) 500,000
Total capital cost $3,800,000
The amortization of this equipment would be:
For 20 years at 6% 50/1000 gal. product
For 3 years at 6% 180/1000 gal. product
Item
Power
Acid
Calgon
Chlorine
Detergent
OUT-OF-POCKET COSTS
Comments
pump efficiency, 10/KWH
$33.36/ton, 110 ppm
150/lb., 10 ppm
7.60/lb., 0.5 ppm
300/lb., 1.0 wgt $/month
500/lb., 1.0 wgt $/month
Replaced monthly
EDTA
Cartridge filters
Labor & Maintenance (As above)
Sub-Total
Sand filters, including coagulant aids
Carbon adsorbtion including regeneration
Total
The above total cost is:
1. Plant amortization
2. Fiber
i
3. Out-of-pocket
k. Pre treatment
Total
0/1000 Gal
Product
k.6
1.9
1.6
0.03
OA
0.6
k.o
2.2
15.330
5.00
10.00
30.330
50/1000 gal.
180
150
150
530/1000 gal.
product
-------
185
The manufacturer advised that the major cost difference between a ten
million gallon per day plant and a 100,000 gallon per day plant would
be the amortization charge. For the smaller plant, this would be
19^/1000 gallons.
A 100,000 GPD plant would cost about $100,000; $20,000 for the fiber
and $80,000 for the auxiliaries. This would give costs of:
1. Amortization 1901/1000 gal,
2 . Fiber
3. Out-of-pocket
if. Pre treatment
Total 69^/1000
Summary of Costs
The "Study" cost estimate is substantially higher than either the Gulf
or Du Pont estimates. Two items could partially account for the
differences; (l) Membrane life was assumed to be 2 years in this study
estimate; both Gulf and Du Pont used a 3 year membrane life and (2) the
manpower required, differs substantially between the "Study" cost
estimate and the Gulf and Du Pont cost estimates. More information is
required on larger size R.O. studies over an extended period in order
to obtain better information on R.O. costs.
-------
SECTION XVII
REFERENCES
1. Breton, E. J., Jr., Water and Ion Flow Through Imperfect Osmotic
Membranes, Office of Saline Water, Research and Development
Progress Report Wo. 16, PBl6l391 (195?)
2. English, John H., et al., "Removal of Organics from Wastewater,"
Chemical Engineering Progress Symposium Series, 67, Wo. 107,
pp W7-153 (1970). ^
3. Goel, Vinay and Joseph ¥. McCutchan, Systems Analysis and
Optimization of a Tubular Module Reverse Osmosis Pilot Plant for
Sea Water Desalination, Water Resources Center Desalination
Report No. 45, UCLA-Eng-7l63, PP 27-38 (1971)
k. Hagstrom, D. E., "Reverse Osmosis - A Tool for Modern Industry,"
Water and Sewage Works, 119, Reference Number pp R103-Rll6 (1972),
5. Merten, Ulrich, Desalination by Reverse Osmosis, The M.I.T.
Press, Cambridge, Mass, and London, England (1966).
6. Sourirajan, S., Reverse Osmosis, Academic Press, New York (1970)
7. Vos, K.D., F. Burris, and R. L. Riley, J. Appl, Polymer Sci. 10,
825 (1966).
186
-------
SECTION XVIII
APPENDICES
A — 1 Glossary^ Equations and Derivations 188
A - 2 Daily Record of Feed pH's (Omitted)
A - 3 Membrane Rejuvenation Data 203
A - k Record of Backflushes for Post-Secondary Filters 213
*
A - 5 Comments on Nitrate and Phosphate Analyses 218
A - 6 Computer Printout Summary (Omitted)
NOTE: Because of paper shortage, Appendices A-2 and A-6
have been omitted from this volume. These sections
may be obtained from the National Technical Infor-
mation Service, Springfield, Va. 22151.
Beginning with page 197 of this volume, a double
set of page numbers appear. One set reflects the
original page numbers; the bracketed numbers
reflect continuous numbering of this shortened
volume.
187
-------
APPENDIX SECTION, A-l
GLOSSARY, EQUATIONS, AND DERIVATIONS
Glossary;
A Membrane water permeation rate (gm/sq cm-sec-atm)
A Calculated value for A at end of first hour (gm/sq cm-sec-atm)
A_ Calculated value for A at end of 1000 hours (gm/sq. cm-sec-atm)
B Solute permeability factor (cm/sec)
b Tangent of log(A x 10 ) vs log time (hrs) regression line
C Avg. solute concentration of feed and reject (mg/l)
a
C_ Solute concentration of feed (mg/l)
C Solute concentration of product (mg/l)
C Solute concentration of brine (mg/l)
D Optimum time between membrane flushings (days)
d Diameter of the membraned tube (in.)
E Material balance agreement ratio
F Decrease in product flow rate 2k hours (gal./min)
F Net feed flow rate (gal./min)
F Product flow rate - observed (gal./min)
F Product flow rate - corrected to 25 C (gal./min)
pc
F Brine flow rate (gal./min)
r
f Same as F (used in ND estimation) (gal./min)
I fie
GFD Gal./sq ft-day, at 500 P (gal./sq ft-day)
J Average solute rejection ratio
d>
J. Concentration of feed to a post secondary effluent
treatment step (mg/l)
J Concentration of discharge from above step (mg/l)
188
-------
189
Jfp Post secondary effluent treatment step rejection ratio
J+ Total solute rejection ratio
M Total effective membrane area (sq ft)
Hjte Reynolds number (dimensionless)
n Number of data points in a set
Pe Net effective operating pressure (psig)
Pf. Feed pressure (psig)
PO Osmotic pressure (psig)
Pr Brine pressure (psig)
p Same as F (used in N estimation) (gal./min)
RC Product recovery ratio - corrected to 25° C
RQ Product recovery ratio - observed
r Sample correlation coefficient
&fo Standard deviation of b
sz Standard deviation of z
Tf Temperature of feed (°F)
v Volume rate of the brine flow from a specific section of a
reverse osmosis unit (gal./m.±n)
w Volume rate of the entire reject from a reverse osmosis
unit (gal./min)
X,Y Logarithms of (A x ICr ) ana time (hrs) respectively
x,y Values defined by Equations (2^) and (25)
Z Any variable
** Exponent indicator
-------
190
List of Equations
B
D
E
E
E
GFD
Jft
(0.998 Fpc) / (M
Antilog
(Zx)/n-b(£Y) /n
Antilog (Log A..j_ + 3b)
0.06T1 A Pe Fl - Ja) / J 1
£xy /Ex2
-0.0625
1 + (1 + 32G/F, )2
E, =
C_R + C (1 - R )
(J /J ) (2 - R) /
(2/RQ ) (Jt - J
2 + R0(Jt
">]
(CL - Cf)
L°
- c.
F 1.03**(T7 - Tf) /I.8 (Empirical)
7.215 (A x 105)
- 2Cp / (Cf + Cr) = 1 - Cp/Ca
1 - (Jp/Jf^
1 - (Cp/Cf)
r
|(pf + Pr)
(Post-Secondary Treatment Area)
(D
(2)
(3)
(5)
(6)
(8)
(9)
(10)
(11)
(12)
(13)
(HO
(15)
(16)
(17)
-------
0.008
(Cf + Cr) /2 - (L
(Empirical, based
on conductivity data)
Fpc /Ff
VFf
. b/ n .
(n - 2)£x2
n - 1)
I*2
(Ex)2 /n
/n
(&)(&)/„
Conversion Factors
Multiply
Gal./min
Gal./sq ft-day
Gal./sq ft-day
Gal./sq ft-day-psi
Gal./sq ft-min-psi
Gm/sq cm-sec-atm
Gm/sq cm-sec-atm
Psi
Sq ft
6.308 x 101
x 10"5
-k
x 10
6.930 x 10"
0.9979
1.0021
7.2150 x 105
6.805 x 10"2
9.2903 x 10
*
191
(18)
(19)
(20)
(21)
(22)
(23)
(25)
(26)
To Get
Cu cm/sec
Gm/sq cm-sec
Gal./sq ft-min
Gm/sq cm-sec-atm
Gm/sq cm-sec-atm
Gal./sq ft-min-psi
Gal./sq ft-day (at
500 psi)
Atmospheres
Sq cm
-------
192
Derivation of Equations
B - Equation
The solute permeability "constant" is the ratio of the rate of
migration of an impurity through a membrane, in mg/sec-sq cm, to
the concentration difference across the membrane. In this form
it is expressed in cm/sec. The word "constant" is placed in
quotation marks to indicate that, when derived as stated, it is
only an occasionally useful, easily calculated, yet variable index
because it does not take the net operating pressure of the feed
flow concentrations into consideration. The same comment would
apply, in some small degree, to the A values derived in this report,
From the above definition
c y (6.308 x lo1)
B =
M(9-2903 x 102) (Cf + Cr) /2 - Cp
After substituting Equations (l) and (l^-) we obtain
B = A Pe x 6.707 x 10"2 (1 - Ja) /Ja
D - Equation (6):
It is frequently desirable to estimate the optimum time between
successive membrane cleansings. To do this make the following
assumptions:
1. The rate of product flow decreases linearly with time during
the relatively short time intervals under consideration;
2. This decrease is due to fouling and not compaction or process
changes;
3. The total down time from the start to the finish of the
cleansing operation is ninety minutes.
-------
193
Using the following special notation:
G = Product flow at start of observation period (gal./min)
g = Product dlow one day later (gal./min)
tan a = g - G (negative) (gal./min -day)
Ga = Average product flow rate during optimum length run (gal./min)
t = Down time as a result of cleansing (min)
T = Optimum running time between cleansing cycles (days)
Then assuming t = 90 minutes
Ga = MMOTG + 1MQ T2 (g-G)/2j
Equating first derivative (dGa/dT) to zero
720 (g-G) T2 + Tt(g-G) + tG = 0
Solving r
1 r~ ~~i -
T = -0.0625 <1 - [1 - 32G/(g-G)J 2
L
And using standard nomenclature ___
D = -0.0625 1 - (1 + 32G/Fd )*j
E - Equation (?)
The material balance states
Cf = CpR0 + Cr (1 - EO), giving
)J
Eg - Equation (8)
T + 90 )
(6)'
Cf -
OT
- R)
-------
Then from Equation
T T _ op / fn j. p }
a ~ P ^ r
- 2Cp /JGf + (Cf - CpR0)
L - HO)
- R0)J
= 2C
= 2C (1 - Rj /
cf(l - R0) + Cf - Cp
=Cf(2 - Ro) - (CfCpRo)
Substituting Equation (l6) and rearranging
1 - J.
2(1 - Jt) (1 - RQ)
~2(1 - Jt) (1 - Rj'
- 2R + JtR)
(2 - Jt)J ,
= 2(1 - Jt) (1 - RQ)J /[_2 - RQ(2 - Jt
J = J (2 - R )/ |2 + R (J - 2)1 to give
a t o |_ o 4- —I
and
E
(2 -
(J
E - Equation (9)
E is developed from Equation (8) by noting
R j (J - 2) + j = 2(J - J ), and
o a t t t a
(2/RQ) (Jt - Ja) + Ja (Jt - 2) + Jt, hence
E3 = (2/RoO (Jt - Ja)/ [ja(Jt - 2) + J^J
E - Equation (10)
Since CR -CR = C^. - C . therefore
p o r o r r'
E^ = (cr - cf) /PRO (cr - cp)J
r - Equation (2l)
By definition
(8)
(9)
(10)
r =
-------
195
2 T 2 2
sb = ^ y - P = std. deviation of b
^ :2 (n - 2)
=\ xy// x = log -log slope
b1 -> .ay /> y2
Then/ y = (n - 2)/ x^ su
/ > ' / ... ' b
[(n - 2) Bb2 +
- 2
And bb- = r2
r
2 ,2/V..2/V^ = b2/Qn _ 2) s^2 + b2j (21)
-------
APPENDIX SECTION A-2
DAILY RECORD OF MED pH'S
(OMITTED, Sei Page l8T)
[196]
-------
APPENDIX SECTION, A-3
REJUVENATION DATA, AEROJET MEMBRANE SET 1
WEEK
8
10
11
«••••••
TYPE CLEAll
Biz
Biz
Biz
MMMMVHH^^H^^H^M
OZ../GAL.
1
1
1
•••••^••••MBPBM^
02. WT.
USED
50
38
38
•^IHHBMHIHMPBHMMBM
PH
8
8
8
MH^BMBMMB
PRODUC
PRE
2.15
2.10
2.25
mBMMflHBBMMW
TGPM
POST
3.to
2.55
2.60
••••••••••••••••B
PRESSURE
PSI
PRE
500
510
520
••HMM^MMB
POST
610
500
520
•MOM^MIBHHi^H
CONTACT TIME
SOAK
15 min,
15 min.
20 min.
••••••••••••••HHBIIIIi
RECIEU
No
No
No
•HM^MBMBMnVIB
REJUVENATION DATA, RAYPAK MEMBRANE SET
WEEK
«»•••••
62
63
6k
6k
6k
65
65
66
66
67
67
68
•*
TYPE CLEAN
MMW^l^BM^aHMMMW
Colo. H00
Sxz
Biz
Biz
Biz
Biz
NaCOL/Biz
NaOCl
Biz/CL_
ClzA«7BizAl
CL2 Biz |
OZ./GAL.
^••••••••••^••••••i*
2
2
2
2
2
60/2
60
2°
OZ. WT.
USED
•IMVHIIIIHMIIMHIMMIBB^M
76
76
76
nu
76
ltgal/76
.25/gal
•
PH
7±
9.5
10.5
10.5
10.5
10.5
10.2
PRODUCT GPM
PRE
.2k
.75
.78
.29
.23
.27
.18
POST
.90
1.12
.76
.61
1.10
.3k
.60
PRESSURE
PSI
PRE
697=5
753-5
725
732.5
790
805
836
POST
7^
67U.5
766
725
797.5
717.5
7^2.5
CONTACT TIME
SOAK
30 min
RECIR.
33 min.
35 min.
§0 min.
50 min.
kQ min,
15/15/1
1
^5 niin.
203
[197]
-------
REJUVENATION MIA, AMERICAN STANDARD MEMBRANE SET 1
WEEK
17
18
19
21
23
26
27
31
34
39
42
47
TYPE CLEAN
Biz
Biz
Biz
Biz
Biz
EDIA
Biz
Biz
Biz
Biz
Biz
1DTA
OZ./GAL.
1
1
i*
1
1
4
1
1
1
1
OZ. WT.
USED
76
76
76
38
76
200
108
108
108
76
pH
6.5
7.0
7.0
7-5
8.0
5.5
8.0
7-5
7-5
7.5
PRODUCT GEM
ERE
5-75
5.80
5-85
6.45
6. to
5.00
4.70
POST
6.05
6.50
7.15
6.75
6.75
5.15
PRESSURE
PSI
ERE
(720)
(660)
430
535
565
64o
POST
(530)
(6to)
487.5
490
555
600
CONTACT TIME
SOAK
Yes
Yes
No
No
No
RECIR.
^5 min
45 min
45 min
60 min
50 min
45 min
65 min
SET 2
WEEK
9
10
10
10
10
11
12
12
13
14
15
15
TYPE CLEAN
Biz
Biz
H20
H20
H20
Biz
Biz
Biz
Biz
Biz
Biz
Biz
OZ./GAL.
1
1
-
-
-
1.5
1.5
2.5
2.3
2.3
2.3
OZ. WT.
USED
114
lit
.
pH
8
8
3-5
- 2.8-5.8
3-5
152 8.5-7=5
152
114
114
114
114
8
8
8.5
PRODUCT GEM
PRE
6.80
6.59
5.44
-
_
5.72
POST
7.42
6.74
5-65
_
-
PRESSURE
PSI
ERE
500
512. ;
501
_
POST
492.;
495
510
<•>
497.5 -
CONTACT TIME
SOAK
RECIR.
-
[198]
-------
REJUVENATION DATA, DU PONT
MEMBRANE SET IA (2 B-5'S IN SERIES, 1 AND 2)
205
WEEK
7
7
7
7
8
8
9
9
lU
Ik
Ik
Ik
Ik
Ik
15
15
16
16
16
16
18
18
18
19
20
21
21
21
22
25
26
27
28
^\j
29
^ s
29
PERMEATOR
1 or 2
1
1
1
1
1
1
1
1
1
1
TYPE CLEAN
Biz
Biz
Biz
Biz
Biz
OZ./3AL.
2
2
2
1*
k
1 , Inn ect, Chemical "X"
2
2
Biz
2
2 Inject Chemical "X"
2
2
2
2
2
2
2
2
2
2
Biz
Biz
Biz
Vel
Vel
HCL Dilute
Biz
k
1*
2
1
1
OZ. WT.
USED
100
100
100
228
228
111*
228
228
76
2k
1*56
PH
ioi
10±
ioi
•
10*
10±
ioi
10*
10±
5.2
5.0
2.0
2.0
RENEWED # 2 PERMEATOR (SET # 1 MOD.)
1
1
2
2
2
2
2
2
2
b
2
2
Biz
Biz
Biz
Biz
EDTA
EDTA
EDTA
Biz
EDTA
EDTA
2
2
2
2
3.2
1.6
3.2
2
I*
k
76
38
76
76
5 Ibs.
10*
10±
10±
10±
Hi
5 Ibs. 7
10 Ibs. 7
76
10*
12.5 Its. 7
12.5 Ibs. 7
PRODUCT GPM
PRE
5.25
5.85
6.30
6.UO
6.05
.6.05
5.35
5.1*0
5.70
.85
I*. 85
.85
6.70
6.1*5
2.10
6.50
2.10
6.62
1.69
6.10
1.69
5.95
6.80
6.32
1.76
6.2k
1.77
6.21
POST
5-80
6.UO
6.90
6.65
5.55
5.55
5.35
5.20
5.55
PRESSURE
PSI
PRE
620
620
610
615
665
660
61*5
POST
620
620
610
630
63(5
635
No Improvement
|
No Improvemt
art
No Improvement
No Improvement
7.15
6.95
2.3k
6.86
NA
6.65
1.93
6.1*5
1.90
6.UO
2.05
6.77
1.89
6.6U
1.82
6.22
650
660
660
660
635
61*5
650
630
665
625
650
635
635
625
625
630
650
610
CONTACT TIME
SOAK
yes
no
90 min
10 min
2 hr.
l| hr.
1 hr.
2 hr.
RECIR.
yes
yes
2l* hr.
2l* hr.
90 min
5 hr.
2 hr.
1? hr.
UO min
1 hr.
1 hr.
[199]
-------
206
SET IA (coat)
WEEK
32
3U
35
SET
1 or 2
2
1
2
TYPE CLEAN
EDTA
Biz
EDTA
OZ./GAL.
It
2
It
OZ. WT.
USED
12.5 Ibs.
76
12.5 Ibs.
pH
7
7
PRODUCT GPM
PRE
1.59
6.26
1.U8
5.^0
POST
1.65
5.81
1.56
5-56
PRESSURE
PSI
PRE
660
POST
650
CONTACT TIME
SOAK
RECIR.
3 hr.
REJUVENATION DATA, DU PONT MEMBRANE SET IIA
(3-2 PATTERN B-9'S)
WEEK
39
39
tti
Ul
1+5
U5
U9
U9
50
50
50
50
52
52
••••••••••••••i
STAGE
1 or 2
1
2
1
2
1
2
1
2
1
2
1
2
1
^^^^^^•^••••••ffWMHW
TYPE CLEAN
Biz
EDTA
Biz
EDTA
Biz
EDTA
Biz
EDTA
Biz
EDTA
Biz
EDTA
Biz
EDTA
••••••^••••••••••••••••••••IM
OZ . /GAL .
U
It
U
U
U
U
It
U
U
U
It
U
It
U
••••••••••••••••••••••^IWi^^BBVIIIIIIIII
OZ. WT.
USED
152
128
152
18U
152
18U
152
18U
152
I81t
152
128
152
18U
• in •..••.—•
PH
lot
7.5
10*
7.0
10*.
7.0
10±
7.0
10*
7.0
10*
7.0
ioi
7.0
^HVriBHM^
PRODUCT GPM
FRE
5.17
3.13
lt.09
2.60
3.23
1.81
3.17
1.81
3.18
1.76
2.93
1.57
••MBM^HBHIIIIII
POST
5.22
3.13
U.07
2.58
3.20
2.20
3.56
2.15
3.01
1.98
•KMHPVtfHBH^^^m
PRESSURE
PSI
PRE
It05
380
U05
375
395
360
U25
325
It25
3U5
U20
Uoo
315
•••••••^^••MM
POST
UOO
377
U12
388
395
It 00
355
390
355
U10
UOO
355
M^H^^MMMI^H^MHMHH^BflHM^^^^^^^^^^B
[200]
-------
MEMBRANE SET IIB
(5 IN PARALLEL, B-9'S)
207
WEEK
55
56
56
57
58
58
59
59
60
61
61
62
62
62
63
63
63
63
64
64
64
64
64
64
64
64
64
65
65
65
76
66
TYPE CLEAN
Biz^DTA
Biz/feDTA
Biz/EDTA
Biz/EDTA
Biz/feDTA
Biz
Biz
Biz
Biz
Biz-R
Biz
EDTA
Biz
H20
Biz
NaHOH
Biz
Biz
Biz
Biz
Biz
EDTA
Biz
EDTA
Biz
EDTA
H20
Formaldehyde
HzO/Air
EDTA/Air
Sulfamic
OZ./GAL.
4/4
4/4
4/4
4/4
\t / ii
Ji
|i
3.2
4
4
4
4
4
-
4
4
4
4
4
4
4
4
4
OZ. WT.
USED
152/128
152/128
152/128
152/128
152/128
152
190
152
190
190
190
200
152
-
190
190
152
-
-
190
192
190
200
PH
10/7
10/7
10/7
10/7
10/7
10
10*
10±
lot
ioi
lot
7
10^
6t
io|
12.9
lOi
io|
io|
lOf
io|
7
io|
7
_ \
3-4
PRODUCT GPM
PRE
3^98
4.13
4.19
3.80
3.73
3.66
2.90
2.15
3.00
2.62
2.70
2.02
2.48
-
2.14
2.00
POST
4.58
4.40
4.38
3-98
3.94
3.68
3.04
3.15
3.17
2.71
-
1.76
2.14
PRESSURE
PSI
PRE
386
400
417
379
377=5
377.5
298.5
400
377.3
364.5
296
403.5
-
417.5
403.5
POST
388.5
380
372.5
367.5
381.9
370
371-9
378.9
377.8
376
-
400
CONTACT TIME
SOAK
RECIR.
63/60
95/95
49/70
50/63
90/40
65
249
60
60
70
60
60
[201]
-------
208
REJUVENATION DATA, GULF MEMBRANE SETS 1 & 2
WEEK
3
4
5
5
5
5
5
6
6
6
7
8
8
8
9
9
9
10
11
11
13
14
15
16
16
17
18
19
23
26
27
28
29
30
33
35
35
36
37
37
38
ho
4i
42
SET.
1 or 2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
FOB CLEAN
Biz
HgO
H^O
HdO
HlO
HgO
HgO
HgO
Biz
Biz
Biz
Biz
Biz
Hj£
HaO
Biz
H,*)
Biz
Biz
Biz
Biz
Biz
Biz
Biz
H20
Biz
Biz
Biz
Biz
EDTA
Biz
EDTA
Biz
EDTA
Biz
Biz
EDTA
Biz
EDTA
Biz/EDTA
Biz /EDTA
Biz/EDTA
Biz/EDTA
Biz/EDTA
OZ./GAL.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
2
4
2
4
2
2
U
2
U
oz. wr.
USED
76
76
76
76
76
76
76
76
76
76
76
76
76
76
76
76
76
76
200
76
200
76
200
76
76
200
76
200
76/64
76/64
76/6U
76/61*
76/64
pH
7
5I
5*
5*
5l
5*
5±
4-5
7
7
7
7
7
U-5
4-5
7
U-5
7
7
7
7
7.2
7.2
7.6
7
7
7
7
7
7
7
7.5
7
8.5
8
7
8
7*
7.5
7*
7-
7
8
PRODUCT GPM
PEE
4.35
5.00
4.55
4.70
U.90
5.10
5.10
4.65
U.80
5.80
5.65
5.60
5.60
5.75
5.60
5.60
5.65
5.65
5.55
5.65
5.85
5.80
5-70
5.60
5.50
5.95
5.90
6.00
6.05
5.85
6.10
6.10
5.70
5.80
5.65
5.45
5.0
4.65
4.60
4.35
4.35
4.03
4.10
3.92
POST
6.30
5.50
4.95
5-25
5.50
5.50
6.00
6.40
6.20
5.70
5.90
6.15
5.65
5.70
5.90
5.85
5.80
5-95
6.15
6.20
5.95
6.10
6.00
6.10
6.10
6.40
6.30
6.30
6.50
6.10
6.10
5.90
5.70
5.35
5.10
4.60
4.70
4.70
4.40
4.20
4.05
PRESSURE
PSI
PRB
567,5
602.5
592.5
592.5
610
605
605
597.5
610
607.5
610
597=5
597.5
600
610
595
600
601.5
592.5
600
610
612.5
592.5
615
600
612.5
607.5
607,5
597.5
622.5
597.5
587.5
597.5
620
597.5
612.5
622.5
607.5
614.5
607.5
582.5
587,5
611
608.5
POST
517.5
592.5
592.5
592.5
610
600
595
597.5
595
592.5
595
595
595
607.5
597.5
597.5
597,5
597.5
597.5
587.5
587.5
595.
607.5
605
597.5
587.5
597.5
598.5
590
590
591
591
CONTACT TIME
SOAK
KUUXK.
30 min.
30 min.
30 min.
30 min.
30 min.
30 min.
30 min.
40 min.
40 min.
60 min.
40 min.
30 min.
30 min
40 min
30 min
•
•
•
40 min.
40 min.
40 min.
30 min.
30 min
30' .min
•
•
30 min.
30 min.
30 min.
30 min.
45 min.
60 min.
55 min.
50 min.
65 min.
60 min.
50 min.
60 min.
70 min.
53 min.
50 min.
50 min.
75 min.
53 min.
70 min.
71 min .
[202]
-------
MEMBRANE REJUVENATION, GULF (cont)
209
WEEK
44
45
46
46
Jl*7
L*7
47
48
51
51
52
52
53
53
53
ch
c|i
C||
*5«i
55
55
55
56
56
56
56
56
57
57
58
f^^
58
58
59
XX
59
59
Xx
60
w
60
6l
>^A*
61
6l
SiU.
62
Vb
62
62
62
63
:;J
63
SET.
1 or 2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
p
^B
2
2
£»
2
2
CH
2
^B
2
2
2
2
^B
2
TYPE CLEAN
Biz /EDTA
Biz
EDTA
Biz
Biz
HoO
Biz
Biz
Biz
Biz
Biz
EDTA
Biz
Biz /EDTA
HdO
Biz
HoO
Biz
Biz
Biz
Biz
H20
Biz
Biz
HgO/Biz
Biz
Biz
Special
Biz
Biz
Biz
Biz
Biz
Biz
Biz
Biz
Biz
HoO
Biz
Biz
HoO
H20
HgO
H20
G.
Biz
OZ./JAL.
2
4
4
2
2
2
2
2
2
4
2
2A
2
2.3
2.3/3.2
2.3
2.3
1.5
1.5
A-5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
OZ.WT.
USED
76/64
76
200
152
76
76
76
76
76
76
160
76
76/160
76
114
266
Tl li
Tl It
76
76
/76 3
76
76
132
76
76
76
76
76
PH
7.5
8
7
7.8
8
3
8
8
8
8
8
7
8
7/7
3A
7
3A
8
7
8
8
2-3
8
4
6
-.7/8
8
8
8
8
8
8
6.8
Plant Shut Down
1.5
1.5
1.5
1.5
2.3
3.8
76
76
76
76
190
8
8
8
2.5
8
8
2.5
2.5
5.5
2.5
8.0
PRODUCT GPM
PRE
3.77
3.68
3. BO
2.95
2.87
3.20
2.78
3.25
4.70
^.55
4.30
4.76
3.72
4.07
sieo
4.02
3.72
3.83
3.42
3.65
3.85
3.75
3.75
3.77
4.08
4.08
3.87
3.86
3.72
3.87
3.69
3.19
3.55
3.37
3.20
3.20
3.26
3.H
2.60
2.55
2.21
2.10
2.00
2.04
POST
4.10
3.98
3.95
3.73
3.77
3.20
3.47
3.68
5.78
5.90
5.40
4.90
4.50
4.40
^.37
4.07
4.15
3.96
3.97
4.10
3.88
4.25
3.76
4.07
4i50
4.45
4.18
IL OQ
ii QQ
3.92
3.72
3.80
3.50
3.68
3.55
3.36
3.33
2.84
2.55
2.40
2.61
2.05
2.15
PRESSURE
PSI
PRE
597.5
615
608
630
636.5
638.5
607
605
465
452.5
568.5
557.5
539
581
606
595
580
598.5
595
590
600
592.5
583.5
577.5
587.5
581.5
592.5
600
591
597=5
588.5
597.5
593.5
586.5
585.5
600
590.5
624
600
601.5
625.5
615
571
575
POST
599
600
600
603.5
591-5
602.5
597.5
400
472.5
530
553.5
560
577.5
587.5
580
588.5
585
598.5
582.5
581.5
583.5
570
586
581.5
596
587.5
579
587.5
565
590
590
581.5
592.5
592.5
598
595
589
592.5
585
610.5
583.5
603.5
568.5
565
-CONTACT TIME
SOAK RECIR.
51 min.
78 min.
72 min.
48 min.
55 min.
*5 min.
77 min.
50 min.
15 min.
45 min.
32 min.
25 min.
34/40 min.
>8 min.
*3 min.
58 min.
>2 min.
79 min.
62 min.
72 min.
62 min.
27 min.
36 min.
94 min.
27/29 min.
62 min.
71 min.
45 min.
39 min.
*5 min.
55 min.
51 min.
47 min.
50 min.
60 min.
20 min.
50 min.
80 min.
30 min.
30 min.
30 min.
40 min.
35 min.
[203]
-------
210
MEMBRANE REJUVENATION, GULP (cont)
WEEK
63
63
6V
6k
65
SET
1 or 2
2
2
2
2
2
TYPE CLEAN
Biz
Biz
Biz I
Biz I
NaOCL
OFF FROM 5-18 to 6-3-71
67
67
67
68
2
2
2
2
Biz-CL
Biz-I
Biz-I
NaOCL
OZ./GAL.
1.5
2.3
2.3
2.3
2.3
2.3
2.3
100ml/50g!
OZ. WT.
USED
76
11U
111*
UA
11U
114
11U
al.
pH
8
8
8
8
8
8
8
PRODUCT GPM
PRE
2.00
2.03
1.95
1.73
1.98
2.35
2.09
POST
2.62
2.75
3.5*
2.70
3.32
2.87
2.55
PRESSURE
PSI
PRE
620.5
603
578.5
POST
583-5
562.0
Pump Bad
ii i
5^2.5
557.5
596
585
i
570
585
577.5
582
CONTACT TIME
SOAK 1 BECIR.
80 min.
80 min.
90-130 min.
90-170 min.
58-3U min.
57 min.
62 min.
[2043
-------
EEJUVENATIQN DATA, UNIVERSAL MEMBRANE SETS 1,2,3,&1*
211
WEEK
7
8
9
9
10
10
10
10
11
11
11
11
13
15
22
23
23
21*
25
26
26
27
27
27
28
28
29
29
*** f
30
32
33
37
39
1*1
1*5
SET
HO.
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
TYPE CLEAN
Biz
Biz
HgO
H20
HdO
Biz
HoO
H?0
HoO
H20
Hop
Biz
Biz
Biz
Biz
Biz
Biz
Biz
Biz
Biz
H20
Biz
Biz
Biz
EDTA
Biz
Biz
Biz
Biz
REMOVED MODULES
3
w
3
3
*/
3
_s
3
Biz
Biz
Biz
Biz
Biz
Biz
OZ./GAL.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
OZ. WT.
USED
100
100
111*
111*
111*
111*
72
72
76
76
76
76
76
76
76
76
76
76
38
76
76
76
76
76
76
152
PH
lof
10±
Id*
10
10-
io|
10-
10
10
10
10
10
10
10
10
10
10
10
10
i
lot
10±
10*
10±
10+
8*
PRODUCT GPM
PEE
U.75
5.15
5.60
5.90
5.70
5.1*5
5.55
5.95
5-85
5-75
5-95
5.20
5.75
5.60
5.60
5-50
5.1*5
5.60
5-70
5.10
5.10
MS
Ms
3.80
3.30
3.15
2.55
2.1*5
2.1*0
2.M*
2.08
2.17
1.90
POST
5.65
5.70
5.70
5.90
5.95
6.25
5.70
8.10
8.10
7.1*5
6.65
6.60
5.75
5.50
5.65
li.55
l*.25
3.80
3.20
3.30
3.05
2.80
2.20
2.57
2.25
2.1*3
2.11
PRESSURE
PSI
PRE
582
5.25
502
502
POST
570
1*60
1*60
502.5 1*72.7
505 1*80
502.5 1*80
515
500
585
630
605
635
590
590
1*97.5
560
560
587,5
620
595
632.5
527:5 507.5
550
555
505
1*80
515
**92.5
507-5 515
560
572.5
CONTACT TIME
SOAK
RECIR
no 30 min.
20 min.
60 min.
50 min.
50 min.
60 min.
60 min.
75 min.
30 min.
1*0 min.
55 min.
55 min.
55 min.
75 min.
60 min.
30 min.
55 min.
50 min.
60 min.
65 min.
75 min.
63 min.
1*1 min.
[205]
-------
212
MEMBRANE REJUVENATION, UNIVERSAL (cont)
WEEK
51
51
52
52
53
53
54
54
54
5!*
55
55
56
*r *^
57
58
59
60
62
62
63
63
64
65
66
67
67
68
SET
NO.
4
4
4
4
4
4
4
4
4
4
4
4
k
k
k
T^PE CLEAN
Biz
Biz
Biz
Biz
Biz
Biz
HaO
Biz
HgO
Biz
Biz
Biz
Biz
Biz
Biz
Main Plant Off
4
4
k
k
4
4
U
k
k
k
k
Biz
Biz
H^
Biz
Biz
Biz
Biz
Biz
Biz
Biz
Biz
OZ./GAL.
2.3
2.3
2
2.3
2.3
2.3
2.3
2.3
2.3
2.3
2.3
2.3
1.5
1.5
1.5
1.5
2.3
2.3
2.3
2.3
oz. we.
USED
11U
114
76
11U
114
114
114
114
114
114
114
114
76
76
76
76
114
228
114
114
pH
8
8
8
7=5
8
8
5
8.5
4-5
8|
8±
si
8±
7.8
8.2
8
7.4
2.75
7.5
8.5
8.5
8.8
8.6
PRODUCT GPM
PRE
3.05
3.48
3.37
3.97
3. 40
3.77
3.88
3.75
4.05
3.87
4.15
4.17
4.45
3-98
4,08
3.69
3.11
2.65
3.10
POST
4.45
5.10
3.95
4.47
3.98
4.30
4.05
4.20
4.07
4.50
4.57
4.55
4.52
4.50
4.66
4.30
3.48
3.92
PRESSURE
PSI
PRE
560
545. '
572.5
570
560
552.5
572.5
562.5
550
551.5
585
567.5
550
550
570
527.5
549
551
582.5
POST
522.5
540
562.5
537.5
540
552.5
550
555
545
562.5
565
550
540
545
550
538
539
CONTACT TIME
SOAK RECIR.
40 min.
45 min.
35 min.
45 min.
33 min.
36 min.
40 min.
48 min.
40 min.
48 min.
52 min.
63 min.
30 min.
40 min.
70 min.
70 min.
73 min.
85 min.
65 min.
[206]
-------
APPENDIX SECTION A-k
PEED TREATMENT PROCESS REJUVENATIONAL RECORD
SAND FILTERS (S-l, S-2, S-3) BACKFLUSH HISTORY
WEEK
NO.
2
3
5
6
8
10
10
10
13
13
13
15
15
15
17
17
17
20
20
20
22
22
22
21*
2k
2k
26
26
26
28
28
28
30
TO
O ITVUMTV
• oo m oo co
UNIT
S-2
s-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
s-3
S-l
S-2
S-3
S-l
S-2
s-3
S-l
S-2
S-3
•MHBM^^M^V^^
DURATION
(MINUTES)
15'
15'
15'
15'
10'
9'
10'
20'
25 '
20'
15'
15'
15'
15'
15'
15'
_
-
.
20'
13'
17'
18 '
18'
21'
19'
18'
50'
17'
19'
19'
IV
10'
10'
15'
11'
10'
— — — •
VOLUME
(GALLONS)
2150
2800
1700
1600
1600
1300
1100
1000
3000
2600
2600
2700
3100
2300
2kOO
1800
2600
3000
3000
3000
3000
2000
3000
3000
3000
3000
3100
3000
3150
3000
3000
3000
1500
1500
1900
1800
1800
1650
: — -^-^— " ^— ^—
PRESSURE DIFFERENTIAL
PRE BACKFDUSH
Ik
18
17
6
Ik
8
7
8
16
16
17
Ik
13
15
19
17
9
20
20
20
15
15
15
15
15
15
17
17
17
20
20
20
18
18
18
__— — — — — — ^—
POST BACKFLUSH
3
1
1
0
0
0
0
0
c
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1,
4
2
0
0
0
0
0
0
7
7
7
FLOW
CONFIGURATION
Single
Single
Single
3 in Parallel
ii n n
ii n n
n n n
n n n
n n it
n n n
n n n
it n ti
nit n ~~
PI ii
it it it
nit it
ii n
n« n
¥1 if
Uff ff
II II
II It II
II tl II
II II It
II II II
II II II
II It It
II It It
II II II
It II It
II It It
II tl It
n n n
n n n
n n tt
n tt n
it tt n
it it n
n n tt
n n n
it n n
n n tt
>^ •«.
213
[207]
-------
SAND FILTERS (S-l, S-2, S-3) BACKFLUSH HISTORY
WEEK
NO.
41
41
in
44
44
49
49
49
49
49
49
51
51
51
52
52
52
53
53
53
54
54
54
55
55
55
55
56
56
56
57
57
57
58
58
58
58
58
58
59
59
59
UNIT
S-l
S-2
S-3
S-l
S-2
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
DURATION
(MINUTES)
17'
18'
21'
22 '
22 «
20'
20'
20'
.
_
_
18'
18'
22
20'
17'
20'
20'
20'
20'
20'
19'
19'
60'
25'
25'
11'
20'
17'
18'
23'
25'
2k t
11'
9'
11'
27'
27'
27'
32'
33'
34'
VOLUME
(GALLONS)
3000
3000
3400
3900
5500
3000
3000
3000
4000
4ooo
5700
3000
3000
3000
3100
3000
3000
2900
3000
3000
2500
2400
2400
10000
4800
48oo
900
3000
3000
3000
3500
3500
3500
1000
800
1000
3500
3800
3500
3500
3500
3500
PRESSURE DIFFERENTIAL
PRE BACKFLUSH
24
24
24
14
14
10
10
10
6
6
10
16
16
16
9
9
9
8
8
8
10
10
10
0
13
13
13
5
5
5
8
8
8
7
7
7
6
6
6
18
18
18
POST BACKFLUSH
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
SLOW
CONFIGURATION
3 in Parallel
n n n
ti it n
ii n n
n n n
it n n
n ti n
n n n
n n n
n n n
n ti n
n n it
it ti ti
n n ii
it it n
n ti »
» it n
n n it
n it it
n n n
n it ii
n it n
II N It
ti it n
it n it
n n n
» n n
it ti it
ti n n
n » it
n ii it
it n it
n it it
it it it
it tt it
ti it it
» n it
it ti it
n n it
it ii it
it ti n
it ti »
[208]
-------
SAND FILTERS (S-l, S-2, S-3) BACKELUSH HISTORY
215
WEEK
NO.
59
59
59
60
60
60
61
61
61
62
62
62
62
62
62
62
62
62
63
63
63
63
63
63
63
63
63
64
MMHMMI^BBI
UNIT
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
s-3
S-l
S-2
S-3
S-l
S-2
s-3
S-l
S-2
S-3
S-l
S-2
S-3
S-l
S-2
s-3
S-l
S-2
S-3
•MHM^MMMM^Mi
DURATION
(MINUTES)
-
_
27
28
27
.
—
.
«•>
.
«,
—
_
—
-
_
_
40'
40'
4of
S31
33'
36'
30'
33 '
32'
SO1
30|
.-
VOLUME
(GALLONS)
3000
3000
3000
3000
3000
3000
3500
3300
3700
2600
3200
3200
6200
4000
3500
4000
3700
3600
4000
4700
4500
4300
4300
4300
4200
4200
3200
4200
4200
4200
-»— ^— — ^— ^— «^«—
PRESSURE DIFFERENTIAL
PRE BACKFLUSH
-
_
n
n
n
17
17
17
15
15
15
9
9
9
6
6
6
2
2
2
3
3
3
4
i
4
3
3
3
.1 •
POST BACKFLUSH
-
_
0
0
0
2
2
2
9
9
9
3
3
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
FLOW
CONFIGURATION
3 in Parallel
n n it
n n it
n n it
tt tt tt
n n n
ti n ti
tt n it
it ti n
n ti ti
tt it it
nn tt
ii ••
Vmm H
IT II
tt IT It
nil ft
ii ii
n it 11
nn if
ii ii
nit n
n ii
nit tf
if 11
nn tt
ii ••
It N II
It It II
II It It
II II II
tt n n
tt n ii
it tt n
ti n n
n n n
— «^^—
[209]
-------
216
CARBON FILTERS (C-l, C-2, C-3) BACKFLUSH HISTORY
WEEK
NO.
5
6
6
6
7
7
12
13
13
16
22
27
30
37
39
1*2
1*1*
1*8
1*8
U8
UNIT
C-2
C-l
C-2
C-3
C-2
c-3
C-l
C-2
c-3
C-l
C-l
C-l
C-l
C-l
C-l
C-l
C-2
C-l
C-2
C-3
DURATION
(MINUTES)
60'
Uof
i+o1
1+5'
15'
10'
60'
1*0'
1*5*
120'
30'
60'
50'
60'
60'
55'
65'
•
VOLUME
(GALLONS)
3200
2800
31*50
2950
2800
ll*00
91*00
6300
7900
12UOO
7900
5250
7000
8750
9600
9UOO
9500
_
PRESSURE DIFFERENTIAL
PRE BACKFLUSH
n
12
1
3
1*
5
l*
3
3
10
ll*
29
22
16
28
23
ll*
9
12
2
POST BACKFLUSH
1
0
0
0
0
0
2
0
1
1
0
0
0
0
1
0
1
--
mm
FLOW
CONFIGURATION
a ingle
B4rieS^~3
ii
11
n
n
n
n
n
ii
n
n
n
n
n
n
n
n
n
it
[210]
-------
217
D. E. FILTERS (D-l, D-2) PRBCOAT RECHARGE HISTORY
WEEK
NO.
7
7
13
13
13
13
15
15
17
17
19
19
20
20
21
21
23
23
21*
2k
30
30
30
••••••••MBI^^B^Hi
UNIT
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-2
D-l
D-l
D-2
^MP—HBMPIMMM
DURATION
(MINUTES)
15'
10'
63'
63'
75|
_
_
_
M
15'
15'
5'
5'
15'
15'
35'
67'
35'
35'
60'
-
. i
^^^••••^"••"••'•^•^•^•^^••••••l
VOLUME
(GALLONS)
5300
2000
1300
1300
2800
1300
1900
3500
lUoo
1500
2000
2000
2000
2000
1300
1300
1600
2300
1300
1300
11000
•k
•MMMHHBIMMMMHMBBMWI
••••••••••HIIIMMIHWMII^^
PRESSURE D
PRE BACKFLUSH
22
2k
20
22
-
-
-
-
16
16
9
10
3
k
3
3
h*T
Ii7
\t
2
2
P»^««— ••••— *— «^— «-»^—
•*MWMHM«^M^HM«V^M^^B^^B^~»
IFFERENTIAL
POST BftCKFLUSH
17
17
0
5
2
1
-
-
1
1
3
k
3
k
3
3
10
10
3
*
^HMWMV^nM^— •~«~^— •«
HHTV^M^HHBIVWwai^B^BMVHMBHMBMM
CONDITION
CONFIGURATION
Pre-Coating
Each Time
n n
n n
n n
ii n
It n
II it
n it
n it
n it
it it
n n
it it
n it
n ti
n ti
it it
ntt
ii
nif
i?
nif
ii
Single
ft
II
^^••MMIWWMWMIiMH^^HBMl^HMM^
[211]
-------
APPENDIX SECTION, A-5
COMMENTS ON NITRATE AND PHOSPHATE ANALYSES
Introduction
Nitrate concentrations have proven to be the most variable of the
common analytical determinations used to monitor the results of
wastewater treatment and the various tertiary treatments, including
reverse osmosis, encountered in this project.
The newly issued 13th Edition of Standard Methods listing five
"Tentative" methods for nitrates in polluted waters, concedes the
difficulties involved and though avoiding a specific selection lists
enough precision and accuracy data to stress the likelihood of wide
variations in duplicates, especially at concentrations below one
mg/1 NOo-N.
A Brucine Method similar to that in the 1969 FWQA Manual was used in
this work and is one of the "Tentative Standard Methods". Organic
matter in high concentrations is listed as a probable interference.
It was concluded from experimental results on this project, that both
the organic suspended solids and the dissolved organics usually present
in secondary effluent are important interferences in the brucine nitrate
analysis. Discussions with others indicated that extreme efforts, such
as walk-in refrigerator titrations, have been employed to improve the
precision of the analyses. The potential use of nitrate analyses for
process plant control or pilot plant development comparisons make it
desirable to have a more economical, faster, and precise (that is, more
reproducible) analysis, even perhaps at the expense of less absolute
accuracy. An attempt was made to satisfy that need by carefully studying
results on standards, regular and spiked samples.
Theory
The deviation within brucine nitrate analyses seems to include:
(a) an absolute deviation independent of the sample and
(b) a variable deviation dependent upon sample interferences
The magniture of the absolute deviation depends not only upon the
photometric equipment and analytical technique but also on extreme care
in precise duplication of reaction times and temperatures.
In typical wastewater samples the variable deviation appears to be a
combination of a fixed first order function of the measured absorbance,
plus a second order function of the sample size. The first, when multiplied
as a constant by the concentration, is a direct percentage variable,
independent of the original sample size, and per se should not influence
218
[212]
-------
219
the choice of sample concentration. It gives the least percentage
error with the largest sample, and it therefore encourages use of
undiluted samples. The second order function probably represents
a reaction between the nitrate and the interference, in which
(Absorbance)equals (NO ) X (interference) or a (sample size) X b
(sample size) or ab (sample size squared). This gives the greatest
variation, both absolute and percentage-wise, with the most concentrated
sample,^and thus encourages use of the maximum dilution. It is suggested
that this second order function represents more error than deviation,
that is, it is consistently high (or low) depending upon the nature
of the interference, and yields good precision though less accuracy,
and poorer spike recoveries, when large numbers of samples are averaged.
This may explain why some analysts prefer the maximum practical dilution
to obtain accuracy even at the expense of higher labor costs (or fewer
analyses) and less precision. It is believed that the aspects of both
cost and precision justify the adoption of the simpler procedure (with
experimental corrections which can reduce the errors below those due
to normal deviations of the standard method) for repeated analyses of
the same process streams - as in process control or development studies.
It would be a substantial improvement if duplicates could agree within
ten per cent instead of the twenty-five per cent common by standard
methods.
Experimental Results
Uniformly good precision was attained with a simplified procedure relative
to previously published variations of the brucine nitrate method. It
is still necessary, however, to use "grab samples" obtained at the same
time from the feed, product and brine of the reverse osmosis units in
order to obtain the desired agreement on material balances.
Samples spiked with secondary effluent or its concentrated brines gave
low results (averaging about 90$ agreement). Filtration of secondary
effluent and brine samples gave poorer factors, suggesting that a positive
error from the suspended solids was partly offsetting the negative error
from the dissolved solids in the wastewater effluent. Filtration therefore
did not justify the extra time and labor involved, since it merely gave
a less accurate answer.
A brief experiment with an activated carbon preliminary separation,
virtually eliminated the nitrate detectible by the proposed method which
explains the reduction in nitrate between post-reactor-clarifier samples
and post-carbon samples during post-secondary treatment. The latter
effect may disappear .upon "nitrate saturation" of the carbon, but this
aspect was not investigated. Preliminary separation with chloroform,
which has been recommended as a stabilizing agent for nitrate samples,
showed some promise for improved accuracy but at the expense of greater
deviation (probably due to the residual chloroform). ^However, a counter-
current chloroform extraction followed by a "kerosene" extraction might
improve accuracy though at the expense of a higher labor cost.
[213J
-------
220
The ma^or interference appeared to be organic material, both dissolved
and suspended, and the results suggest that there -was a "minimum" percent
agreement of nitrate versus organic interference concentration, so that
the positive and negative errors were partially offsetting, but were not
affecting results by the same slope of absorbance curve of concentration.
As a statement of observation, the lower the nitrate concentration (such
as in the products) the greater the correction needed; the greater the
nitrate concentrations (brines with high interference character) the
less need for correction.
Procedure
The proposed method is extremely simple, though it requires the typical
safety precautions due any Brucine Method. Since the sample is diluted
less, extra care is recommended in adding concentrated sulfuric acid.
Tubes must always be pointed away from the face and instruments when
shaking. The sample plus reagent mixing is accomplished in a "matched
set" of "test tubes" suitable for the spectro-photometer used. The
following procedures, which may be modified to suit available equipment,
is recommended:
A. For One Inch Diameter Photometer "Test Tubes":
1. Pipet five ml sample into one inch diameter test tube
in a suitable rack
2. Add ten ml of acidified brucine reagent rapidly from a
Fisher automatic pipet
3. Promptly shake or vibrate the tube for ten seconds. A
combined swirling and tapping motion is effective. The
next tube may be filling while the first is being mixed
4. About ten minutes should be allowed for cooling. Readings
may be taken on a suitable photometer (The Bausch and Lomb
Spectronic 20 was used in Hemet) whenever convenient during
the next three days.
5. A blank is needed to set the photometer at 100$ transmittance,
and at least one standard should be read with each group of
samples. (The curve is not straight; at least four standards
should be used to plot the shape. The number of standards and
the frequency of their use depend upon the precision required.)
The nitrate concentration is read from the graph. The answer
is obtained by multiplying by the appropriate dilution and
correction factors.
B. For Half Inch Diameter Tubes:
1. Pipet one ml of sample into a half-inch diameter test
tube in a suitable rack
2. Squirt two (2.0) ml of acidified brucine reagent rapidly
into the sample from a marked measuring autopipet
3. Shake the tube for ten seconds by tapping sharply against
a gloved finger
k and 5. Same as for larger tubes
C. Dilutions:
For the five ml samples, with one inch diameter tubes, dilutions
are easily made by pipetting in one to five ml of sample, then
adding the remaining distilled water needed for the five ml total
[21U]
-------
221
from a 50 ml reservoir buret. (if greater dilutions are needed,
add the -water first and rinse the micropipet in the diluted sample
twice. Separate dilutions are satisfactory if preferred.)
D. Acidified Brucine Reagent:
The brucine reagent and the sodium arsenite solution, as described
in both Standard Methods and the "FWQA Manual", are mixed with
concentrated sulfuric acid with cooling to minimize reactions.
This mixed reagent gradually turns yellowish, even when refrigerated.
Setting the blank at 100$ corrects partially for these reagent
color changes, but does not correct for the slight shift of the
standards curve. The useful life of the reagent may be monitored
by checking the blank against distilled water set at 100$, and
changing at some selected level, such as 80$ transmittance of blank
versus water at 100$. (The specific percentage chosen will depend
upon the precision required and the number of samples to be analyzed.)
In order to minimize heating and extend the useful life of the
mixture, the acid was cooled and the other reagents added in
several increments with intermediate cooling. Specifically,
2-| ml sodium arsenite was added by pipet at the bottom of a 500
ml graduate of sulfuric acid, which was then covered and chilled
in the freezing compartment of a refrigerator. Then 10 ml of
brucine was added by pipet, near the bottom, slowly, stirring with
the pipet. After chilling again, twenty ml of brucine was added,
and another chilling preceded the final 20 ml addition. After
thorough mixing, the acidified reagent is ready for use.
E. Range of Method:
Whereas the FWQA. method suggests a range of 0.1 to 2.0 mg/1
UO_-N, but is only reasonably precise between 1 mg/1 and 2 mg/1
NOo-N, this method is convenient for a range of 0.5 to 7 mg/1 NOo-N
with one inch diameter tubes, or for 1 to Ik mg/1 NO -N with half
inch diameter tubes. Higher concentrations can be used if extra
standards are used and special curves and readings taken, setting
a medium concentration standard as 100$, and keeping all samples in
those runs above that concentration. Thus the entire range in most
typical well or wastewater samples may be run without any dilution
step. Spiking samples with standards is recommended occasionally
to double-check the accuracy.
Summary:
Precise nitrate analyses can be obtained in as simple a procedure as is
possible photometrically: "Pipet sample, mix in one reagent, read on
photometer, read off graph". Correction factors can improve the accuracy.
Standard spikes and different dilutions help to indicate the probable
accuracy. Thermal control is provided by the heat of sulfuric acid dilution
under repetitive but simple mixing conditions. Wo heating or cooling baths
are required. (Cooling is used in reagent preparation, which can be done
weekly.) The average absolute error of individual corrected results on
wastewater samples is believed to be lower than by various published
Brucine Methods.
[215]
-------
Summary Comments on the Phosphate Ascorbic Acid Analysis
The ascorbic acid method is an excellent photometric analytical method.
It is quite precise and accurate for waste water analyses. The persulfate
digestion is effective but not usually justified on secondary effluent
due to biological "digestion" in the activated sludge process. The
greatest weaknesses of the method are the high sensitivity to glassware
rotation or changes at the high 880 Mju wavelength, and the short "shelf"
life of critical reagents. The mixed reagent should be made fresh daily.
The ascorbic acid solution may be made fresh weekly if refrigerated. The
ammonium phosphomolybdate is much more sensitive to degradation in this
method than in the ANSA. method, and must be refrigerated. The curve is
very reproducible, and failure of the standards to reach their customary
absorbance indicates the possible need for fresh reagents. Whenever this
happens, re-reading the samples after sufficient time delay to permit
full color development will provide more accurate results. Usually
the color will stabilize within three hours and remain relatively stable
several days, unless phosphate and total solids concentrations are so
high that the dye precipitates. When color development is incomplete,
the color development rate varies widely according to temperature,
concentration of salts, etc. so that wide differences result between
"digested" and "ortho" samples. These differences are eliminated by
waiting until the color stabilizes.
[216]
-------
APPENDIX SECTIOU A-6
COMPUTER PRINTOUT SUMMARY
(Omitted, See Page 187)
[2173
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-670/2-74-077
3. RECIPIENT'S ACCESSIOWNO.
4. TITLE AND SUBTITLE
REVERSE OSMOSIS OF TREATED AND UNTREATED
SECONDARY SEWAGE EFFLUENT
5. REPORT DATE
Sept. 1974; Issuing Date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
Doyle F. Boen and Gerald L. Johannsen
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
Eastern Municipal Water District
P. 0. Box 858
Hemet, California 92343
10. PROGRAM ELEMENT NO. 1BB043;
ROAP 21-AST; Task 05
11.K8&W8XKT/GRANT NC
WPRD 4-01-67
12. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A pilot study was conducted to determine reverse osmosis-feasibility on
untreated and treated secondary effluents. Six commercially designed
reverse osmosis pilot units, with 3,000 to 10,000 GPD nominal capacitie:
and different module concepts, were tested. Post treatment of second-
ary effluent feeds, using alum clarification, sand filtration, granular
activated carbon treatment, chlorine additions and pH adjustment, in
different combinations improves reverse osmosis performance and signif-
icantly extends useful membrane life. Membrane fouling occurs despite
post secondary effluent treatments. Enzymatic detergent solutions were
moderately effective as membrane rejuvenation treatments. Inorganic
fouling (particularly with phosphates) could be removed with solutions
of the sodium salt of ethylenediaminetetraacetic acid. Of the module
concepts tested, one of the tubular makes and the spiral wound had the
best overall performance. Based on the pilot plant data, the total
reverse osmosis costs, excluding brine disposal, is estimated to be
$0.78/1,000 gallons for a 0.9 MGD product water facility and about
$0.73/1,000 gallons for a 9 MGD product water facility.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Sewage, *Effluents, ^Filtration,
Sewage filtration, *Coagulation,
^Flocculating, Microorganism
control (sewage), ^'Activated
carbon treatment, pH control,
*Cost analysis, Membranes
*Secondary effluents.
*Reverse osmosis
13B
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
232
RELEASE TO PUBLIC
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
E218]
U.S. GOVERNMENT PRINTING OFFICE: W.-657-586/5312 Region No. 5-'l I
------- |