WATER POLLUTION CONTROL RESEARCH SERIES • ORD- 17O4OEFQ12/69
REVERSE OSMOSIS RENOVATION
OF MUNICIPAL WASTEWATER
U.S. DEPARTMENT OF THE INTERIOR • FEDERAL WATER QUALITY ADMINISTRATION
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe the
results and progress in the control and abatement of pollution
in our Nation's waters. They provide a central source of
information on the research, development, and demonstration
activities in the Federal Water Quality Administration, in the
U. S. Department of the Interior, through inhouse research
and grants and contracts with Federal, State, and local agencies,
research institutions, and industrial organizations.
A triplicate abstract card sheet is included in the report to
facilitate information retrieval. Space is provided on the card
for the user's accession number and for additional uniterms.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Head, Project Reports System,
Planning and Resources Office, Office of Research and Development,
Department of the Interior, Federal Water Quality Administration,
Room 1108, Washington, D. C. 20242.
-------
REVERSE OSMOSIS RENOVATION OF MUNICIPAL WASTEWATER
Environmental Systems Division
Aerojet-General Corporation
El Monte, California 91734
for the
FEDERAL WATER QUALITY ADMINISTRATION
DEPARTMENT OF THE INTERIOR
Program #17040 EFQ
Contract #14-12-184
FWQA Project Officer, G. Stern
Advanced Waste Treatment Research Laboratory
Cincinnati, Ohio
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $1.50
-------
FWQA Review Notice
This report has been reviewed by the Federal
Water Quality Administration and approved
for publication. Approval does not signify
that the contents necessarily reflect the views
and policies of the Federal Water Quality Ad-
ministration, nor does mention of trade names
or commercial products constitute endorse-
ment or recommendation for use.
11
-------
ABSTRACT
A fifteen-month laboratory program has shown that all grades of
municipal wastewater may be significantly improved by the reverse
osmosis process.
Comparisons are provided on the behavior and response of the re-
verse osmosis process to carbon-treated secondary sewage, alum-
treated secondary sewage, secondary sewage, alum-treated primary
sewage, primary sewage, raw sewage, and digester supernatant.
High removals of dissolved minerals, organic substances, and sus-
pended matter have all been achieved in the same treatment.
The effects of a flocculant, dispersant, chelating agent, enzyme, and
acid on reducing product water flux decline are compared. The rela-
tive effects of reverse osmosis test-cell geometry on solids deposi-
tion and membrane performance are presented.
A phenomenological model is postulated describing the role of undis-
solved solids and organic substances in producing product water flux
decline and the subsequent maintenance of constant product water
fluxes.
A computer model of the reverse osmosis process, compatible with
the executive program written by the Federal Water Quality Adminis-
tration, has been developed to provide an accurate and rapid method
of determining the design and cost of reverse osmosis facilities.
This report was submitted in fulfillment of Program No. 17040 EFQ
and Contract No. 14-12-184 between the Federal Water Quality Ad-
ministration and the Aerojet-General Corporation.
Key Words: Reverse osmosis, sewage treatment, process model,
tertiary treatment, computer model, membrane
process, wastewater renovation, demineralization,
solids removal, organics removal.
111
-------
CONTENTS
Section Page
ABSTRACT iii
FIGURES vi
TABLES vii
L CONCLUSIONS AND RECOMMENDATIONS 1
II. INTRODUCTION 4
The Reverse Osmosis Process 5
The Program 6
HI. LABORATORY PROCEDURES 8
Sewage Feed Waters 8
Apparatus 8
Membranes 12
Operating Conditions 12
Measurements 18
Data Reduction 19
IV. LABORATORY RESULTS 21
Product Water Flux 21
Flat-plate Test Cells 21
Carbon-treated Secondary Sewage 21
Secondary Sewage 30
Primary Sewage 30
Raw Sewage 73
Digester Sewage 73
Tubular Membranes 73
Carbon-treated Secondary Sewage 73
Secondary Sewage 73
Primary Sewage 92
Raw Sewage 92
Product Water Quality 92
V. DISCUSSION OF RESULTS 119
Test-cell Geometry 119
Operating Pressure 121
Additives 125
Depressurization 131
Recovery Ratio 131
Advanced Membranes 134
The Fouling Mechanism 134
VL REVERSE OSMOSIS PROCESS MODEL 141
IV
-------
CONTENTS
VII. ACKNOWLEDGEMENTS 158
VIII. GLOSSARY 159
-------
FIGURES
Figure Page
1. Flat-plate Test Apparatus 10
2. Laboratory Membrane Configurations 11
3. Tubular Test Apparatus 13
4. Laboratory Test Apparatus Flow Sheet 14
5. - 93. See Table 2 16
94. Effects of Feed Water Type on Product Water Flux 120
95. Comparison Between Test-cell Geometry and
Typical Product Water Flux Decline 122
96. Effects of Pressure on Product Water Flux Decline 123
97. Effects of Pressure and Membrane Permeability
on Product Water Flux Decline 124
98. Effects of Additives With Carbon-treated
Secondary Sewage 126
99. Effects of Additives With Secondary Sewage 127
100. Effects of Alum Treatment and Additives With
Secondary Sewage in Flat-plate Test Cells 129
101. Effects of Alum Treatment and Additives With
Secondary Sewage in Tubular Membranes 130
102. Effects of pH on Zimmite 190 with Primary
Sewage 132
103. Optimization of Zimmite 190 Dosages For
Primary Sewage 133
104. Product Water Flux, Test 97 135
105. Product Water Flux, Test 98 136
106. Product Water Flux, Test 99 137
107. Photographs of Reverse Osmosis Membranes after
Processing of Municipal Wastewater 140
VI
-------
TABLES
Table Page
1. Average Feed Water Quality 9
2. Laboratory Test Schedule 16
3. Average Wastewater Constituent Rejections and
Product Water Quality for 68° Flat-plate
Membrane at 700 psig 114
4. Average Wastewater Constituent Rejections and
Product Water Quality for 0. 25-in. Diameter
Tubes at 700 psig 115
5. Average Wastewater Constituent Rejections and
Product Water Quality for 0. 56-in. Diameter
Tubes at 700 psig 116
6. Average Wastewater Constituent Rejections and
Product Water Quality for 44° Flat-plate
Membranes at 200 psig 117
7. Osmotic Pressures of Pure Solutions 146
8. Subroutine RO Program 152
9. Subroutine RO Variables and Parameters 154
10. Wastewater Constituent Rejections For Subroutine
RO 155
11. Subroutine RO Decision Matrix 156
12. Subroutine RO Stream Matrix 157
vn
-------
Section I
CONCLUSIONS AND RECOMMENDATIONS
Based on the conduct of nearly 90 individual laboratory-scale reverse
osmosis tests to determine the effects of municipal waste-water quality
and operating conditions on process performance, a number of con-
clusions are presented.
The product water flux approached within 20 days from commencement
of operations a relatively constant value for all of the municipal waste-
waters tested for at least that period of time. The level of stabilized
product water flux was directly related, within certain limits, to mem-
brane permeability and feed water quality. For a fixed set of operating
conditions, membrane material, annealing temperature and no precon-
ditioning other than pH adjustment, the same product water flux of
2 gal/(sq ft)(day) was obtained with raw sewage and primary sewage, but
increased progressively to 6 gal/(sq ft)(day) with secondary sewage and
to greater than 18 gal/(sq ft)(day) with carbon-treated secondary sewage.
Whereas a stabilized product water flux of 5. 5 gal/(sq ft)(day) was ob-
tained with the standard membrane material (degree of acetyl substitu-
tion, 2. 41) on alum-treated, sand-filtered primary settled sewage, a
higher stabilized product water flux of 17 gal/(sq ft)(day) was obtained
on the same feed water under identical operating conditions with an
advanced membrane (degree of acetyl substitution, 2. 64) possessing an
inherently greater product water flux capability. Thus it would appear
that the deposited material on the membrane is not the only limiting fac-
tor in achieving maximum product water flux, indicating that higher puri-
fied water production can be expected from the processing of properly
conditioned municipal wastewater with reverse osmosis membranes hav-
ing intrinsicly greater water transport properties.
High removals of most major pollutants contained in municipal wastewater
were accomplished by the reverse osmosis process. Although the rejec-
tions were dependent somewhat upon the feed water quality and the opera-
ting conditions, average removals for a 68° flat-plate membrane, ex-
cluding digester supernatant, ranged from 83 to 92 percent for total dis-
solved salts as measured by electrical conductivity (EC), from 79 to 94
for oxidizable organics (COD), from 71 to 92 for organic nitrogen, from
74 to 87 percent for ammonium, from 45 to 90 percent for nitrite, from
23 to 92 percent for nitrate, from 93 to 99 percent for phosphate, and
from 83 to 99 percent for methylene blue active substances (MBAS).
A continuous extended test of 2. 5-month duration demonstrated that a
constant product water flux of about 5. 5 gal/(sq ft)(day) could be main-
tained with alum-treated, sand-filtered primary settled sewage at product
water recovery ratios as high as 95 percent.
Continuous tests using an advanced cellulose acetate-cellulose triacetate
blend membrane produced a product water flux of around 17 gal/(sq ft)(day)
-------
with alum-treated, sand-filtered primary settled sewage during a
scheduled 12-day test period.
Of the three types of additives studied, the flocculating agents were
more effective than either a dispersant or chelating agent in reducing
the product water flux decline and in producing the highest stabilized
product water flux, when used in the concentrations chosen Pre-
treatment of the municipal wastewater feed by flocculation with alumi-
num sulfate (alum) followed by rapid sand filtration was the best of all
preconditioning methods used for maintaining product water flux at the
highest levels. It appears however that similar results may be ex-
perienced with organic flocculating agents.
The principal causative agent in the membrane fouling process appears
to be finely dispersed solids. Dissolved organic substances are of lesser
relative importance; whereas the effect of gross readily settleable matter
is negligible.
Development of a constant product water flux is believed to be the conse-
quence of an equilibrium established between the rate of solids deposi-
tion on the membrane surface and the rate of solids removal from the
surface. The position of the equilibrium is dependent upon the nature
and concentration of dissolved organic substances, which more than
likely provide a cohesiveness and adhesiveness to the solids, and the de-
gree of local turbulence at the membrane surface.
For solids-bearing wastewaters, low-pressure (200 psig) reverse osmo-
sis operation was associated with lower product water flux declines than
high-pressure (700 psig) operation. This relationship was not observed
with carbon-treated secondary sewage where the performance with high-
and low-pressure membranes was interchanged. Suspended solids are
the only pressure sensitive sewage constituents that could account for
this observation. Thus it would appear that high operating pressures
increase the density and stability of the flux-reducing, deposited layer
on the membrane surface.
Soaking of a severely fouled membrane in an enzyme-active solution was
Zu1?7ut0 beneficial to the restoration of product water flux, indicating
that the enzyme disrupts the adhesion or stability of deposits on the mem-
brane surface.
Differences in reverse osmosis performance between sheet membranes
supported in the Hat-plate test cells and tubular membranes cast inside
fiberglass tubes became manifest only in those instances where alum
pretreatment was provided. Calcium sulfate precipitation occurred in
the flat-plate test cells due to the existence of substantial areas in the
test cell where adequate turbulence was not provided and concentration of
the feed water stream to saturation was effected. The product water flux
in these situations was diminished by two factors: reduced effective mem-
brane surface area and reduced average effective pressure due to exces-
sive pressure drop through the test apparatus.
-------
In view of the observations made during this program and other informa-
tion bearing on the matter, it is quite evident that the renovation of even
relatively untreated municipal wastewaters by the reverse osmosis pro-
cess is technically feasible, although the ultimate productivity and eco-
nomic status of the process have yet to be determined. Further improve-
ments in, and studies with, newly developed and emerging membrane
materials will result certainly in improved product water fluxes and even
perhaps greater specific pollutant rejections with resulting higher quality
product water. Concomitantly, the development of an acceptably perform-
ing low-pressure reverse osmosis operation could reduce both capital and
operating costs.
The results obtained indicate that further investigation is necessary to
realize the goal of practical municipal wastewater renovation by reverse
osmosis systems. Laboratory-scale reverse osmosis test apparatus,
while serving an extremely important function, has certain limitations
that require the conduct of further tests under full-scale conditions. Of
most importance is the inability to operate small units at practical pro-
duct water recovery ratios without either violating minimum wastewater
flow conditions or employing wastewater recirculation. Low flows pro-
vide insufficient turbulence to control solids deposition and to prevent the
creation of excessively high salt concentrations and associated high osmotic
pressures adjacent to the membrane surface that reduce drastically the ef-
fective pressure and the resulting water flux. Recirculation can cause
modified and unrealistic wastewater characteristics that may influence test
results in a way not experienced in a full-scale plant where recirculation
need not be practiced.
To take full advantage of the technological progress and momentum of the
development of the reverse osmosis process related to other applications
and to substantiate the laboratory-scale observations made during the pro-
gram reported herein, the operation of an appropriately sized reverse os-
mosis pilot plant at a sewage treatment facility is recommended. The
pilot plant should be designed specifically to establish the necessary opera-
ting conditions and parameters for sustained performance and to provide
realistic full-scale cost data on the process. It also is recommended that
the field activity be supported with a series of laboratory-scale investiga-
tions designed to establish initial operating conditions for the pilot plant,
to explore rationally, quickly, and conveniently alternate operating cri-
teria, and to investigate in greater detail the nature of, and means of con-
trolling, the fouling of reverse osmosis membranes by municipal waste-
waters.
-------
Section II
INTRODUCTION
Rapidly increasing populations and expanding industrial activities are
placing greater and greater demands on fresh water supplies that are
relatively static in availability and in some cases even decreasing as
the result of pollution. Recognition by the Department of Interior of
the need to augment the nation's natural water resources through the
desalination of brackish and marine waters has resulted in major ad-
vances in the technology and development of suitable processes. It has
become apparent however that perhaps a much better source of water
for reclamation by these methods is municipal wastewater, since it con-
tains far fewer dissolved minerals and is always available relatively
near the intended use.
Several other water resource management objectives also can be achieved
concomitantly--more effective wastewater treatment and water pollution
control, and a general improvement in the mineral quality of the nation's
water supplies. Refractory materials, both mineral and organic, not re-
moved by conventional sewage treatment processes are effectively reduced
in concentration by desalination processes. Overdraft of underground res-
ervoirs and the resulting degradation of the groundwater due to irrigation
and other natural and artificial recharge practices could not only be check-
ed, but the quality could be enhanced by the planned replenishment of these
groundwaters with desalinated and renovated municipal wastewater. In-
deed, a net removal of salts and refractory organics from the nation's
water resources could be effected and the waters restored to their more
natural quality.
Of the many processes capable of demineralizing water, reverse osmosis
appears well suited to the renovation of municipal wastewater. Processes
requiring a change of phase--distillation and freezing—are better suited
for more saline waters since their performance and costs are relatively
independent of salt concentration. Of the comparatively low-energy de-
salination processes, utilizing semipermeable membranes, only the re-
verse osmosis process possesses the intrinsic advantage that the treat-
ment boundary or interface requires the transport of just purified product
water and not the pollutants.
Whereas conventional wastewater treatment processes require a multitude
of steps to perform partial and in a few cases very high removals of waste-
water constituents, the reverse osmosis process is capable of performing
a much superior treatment in fewer operations. Dissolved salts, organic
substances, and insoluble suspended matter are all removed in the same
procedure.
The potential application of the reverse osmosis process to the treatment
of municipal wastewater early in its development was appreciated by the
Federal Water Quality Administration. Through the aegis of this and other
-------
agencies of the Department of Interior, the reverse osmosis process
has been brought in a decade from the state of a laboratory curiosity
to the threshold of practical utility.
But where applied like other "tertiary" treatment processes that have
demonstrated technical feasibility when preceded by one or a combina-
tion of conventional treatment processes, the economic practicability
is difficult to establish. Because the reverse osmosis process effects
only a separation and concentration and not a conversion of wastewater
constituents, it must be accompanied by other processes—processes
which treat the resulting concentrated reject stream from the reverse
osmosis unit for subsequent disposal and which perhaps prepare the
wastewater for the reverse osmosis unit, and processes which are
specifically attuned to the attributes and which can take full advantage
of the reverse osmosis process. Thus the whole system from the head-
works to the outfall must be considered if a rational basis is to be pro-
vided upon which to assess the technical and economical utility of an
overall wastewater renovation system embodying reverse osmosis.
THE REVERSE OSMOSIS PROCESS
Basic elements of the reverse osmosis process consist of the membrane,
a means for providing a high-pressure differential across the membrane,
and a support for the membrane against this pressure differential. To
meet these requirements various geometries have evolved, including
circular flat-plate membrane stacks contained within cylindrical pres-
sure vessels, tubular membranes contained within porous tubes, and
flat membrane sheets spirally wound.
A number of different membrane materials possess the favorable osmotic
properties of relatively high product water flux and low or no solute trans-
port. Noteworthy among these are of course various forms of cellulose
acetate, which to date are the only formulations that have found extensive
use in the desalination of nonpotable waters.
Membranes can be prepared to provide a fairly wide range of known salt
rejection characteristics. The salt rejection and product water flux capa-
bilities of a reverse osmosis membrane are determined by, in addition to
formulation, the membrane annealing temperature and time employed dur-
ing its manufacture. Since solute rejection increases and product water
flux decreases with increasing annealing temperature and time, a high
salt rejection is associated generally with a low liquid flux, and vice versa.
The useful life span of a reverse osmosis membrane is that period of time
during which the membrane retains an acceptable product water flux.
Upon use a membrane exhibits a product water flux decline as a result of
two factors--intrinsic membrane compaction or reorientation and mate-
rials deposition on the membrane surface. The intrinsic flux decline is
a function of membrane formulation, operating pressure, and membrane
annealing temperature and is relatively small but significant in extent.
The magnitude of the flux decline associated with the deposition of mate-
rials from the wastewater varies from insignificant to excessive and is
-------
related to the nature and composition of the wastewater, the hydraulic
conditions prevailing near the membrane surface, the degree of solute
concentration or product water recovery effected by the process, and
other operating parameters.
Operating pressure of reverse osmosis units must be sufficiently high
to overcome osmotic pressure and to provide a driving force for ac-
ceptable water flux. Upper limits on operating pressure are imposed
by the economics of producing equipment capable of maintaining the
high pressure differentials, and the ability of the membrane to struc-
turally withstand the pressure. Since the magnitude of the osmotic
pressure is directly related to the salt concentration, typical opera-
ting pressures for desalination of sea water containing approximately
35, 000 mg/1 of dissolved solids and of brackish water containing about
4, 000 mg/1 are 1, 500 and 750 psig, respectively.
Reverse osmosis equipment used for the renovation of municipal waste-
water, which contains from 500 to 1, 500 mg/1 of total dissolved solids,
could be operated at pressures lower than have been used in other ap-
plication s--pressure s perhaps as low as 200 psig.
An important operating variable in reverse osmosis is the product
water recovery ratio, or the amount of water produced from the total
quantity processed. The highest recovery at which the process can be
operated is dependent usually upon the concentration of low-solubility
inorganic compounds and organic substances present in the feed water.
As the feed water proceeds through the processing equipment, concen-
tration of wastewater constituents occurs due to removal of water
through the membrane and substances thus formed can coat the mem-
brane and reduce the water flux. The effective concentration is even
greater at the membrane surface than in the bulk liquid due to the
buildup of a boundary layer. The magnitude of this concentration
polarization is a function of the hydraulic conditions in the wastewater
channel; turbulent flow produces a lesser effect than does laminar flow.
Also, the product water quality deteriorates quite rapidly as recovery
ratios are increased above 90 and approach 100 percent, due to the ex-
tremely large salt concentration effected in the wastewater stream.
Whereas a product water recovery ratio near 50 percent is maximum
for sea water desalination operations, recovery ratios of 80 or 90 per-
cent and higher appear feasible in the renovation of municipal waste-
water with proper preconditioning of the feed water.
THE PROGRAM
In general, this program was directed to the definition of a municipal
wastewater renovation system, utilizing the reverse osmosis process
and considering associated pretreatment and post-treatment operations,
which is optimum in terms of operating characteristics and costs. Spe-
cific objectives were to establish the relationships between wastewater
feed quality and reverse osmosis performance, to determine the effects
of selected feed water additives and conditioning on reverse osmosis
-------
performance, to investigate the relative effects of high-pressure and
low-pressure operation on reverse osmosis performance, and to
develop a computer model of the reverse osmosis process for use in
a wastewater treatment simulation program.
The response of the reverse osmosis process to municipal wastewaters
of different character was determined by laboratory-scale testing of con-
ventional cellulose acetate membranes with sewages ranging in quality
from raw to activated carbon-treated secondary and with digester super-
natant. The wastewaters were further modified with selected chemical
additives and operations. A limited number of tests were performed
with more advanced membranes to establish their performance charac-
teristics.
A computer model of the reverse osmosis process was developed and
programmed for use in the Federal Water Quality Administration
Executive Digital Computer Program for Preliminary Design of Waste-
water Treatment Systems.
-------
Section III
LABORATORY PROCEDURES
Accurate laboratory duplication of full-scale sewage treatment plant
facilities and sewages is difficult if not impossible. A good alterna-
tive is the use of sample sewages under controlled laboratory condi-
tions, so that test results then may be related to similar operating
conditions experienced in real treatment facilities.
SEWAGE FEED WATERS
Daily sewage samples were collected for use in the reverse osmosis
test equipment. The samples, representing qualities of effluent rang-
ing from carbon-treated secondary sewage to digester supernatant,were
collected at the County Sanitation Districts of Orange County Sewage
Treatment Plant No. 1, Fountain Valley, California and the Pomona
Water Reclamation Plant, Pomona, California. The Orange County col-
lections were made between 7:00 and 7:30 a. m. on weekdays and 10:30
to 11:30 a.m. Saturdays. Samples from Pomona were obtained at 11:00
a. m. weekdays and 4:00 p. m. Saturday and Sunday.
Table 1 presents average feed water qualities of the various sewages
used in this program.
Due to the summer discharges of sugar refinery wastes into the Foun-
tain Valley collection system, all sewages originating from Fountain
Valley after 8 April 1969 may have an unusally high organic content,
affecting tests between Tests 52 and 96. Also the carbon-treated secon-
dary sewage was of two qualities. In Tests 2 through 8 a relatively high-
quality sewage was used from four serial carbon columns wherein the
last was just regenerated. All subsequent tests with carbon-treated
secondary sewage utilized feed water originating from three serial par-
tially exhausted carbon columns while the fourth was undergoing regen-
eration of the carbon. This difference did not result in noticeable varia-
tions in the monitored feed water quality indicators, but nonetheless, may
be a factor influencing the reverse osmosis membrane performance.
Alum-treated, sand-filtered sewage from Fountain Valley was collected
from the pilot water reclamation plant operated by the Orange County
Water District. All alum-treated sewages from Pomona were processed
and filtered in the laboratory.
APPARATUS
Tests were conducted in assemblies representing two different membrane
geometries--sheet and tubular. The flat-plate test cells shown in Fig-
ures 1 and 2 had been used for many years in conjunction with sea water
and brackish water treatment and provided a readily available, proven
means of conducting the experimental program. Different types of mem-
branes could be tested easily in these flat-plate cells.
8
-------
Table 1
AVERAGE FEED WATER QUALITY
(mg/1)
Pomona Wastewater
Carbon-treated Secondary Sewage
Secondary Sewage
Primary Sewage
EC*
1125
1340
1560
Total
COD
15
43
236
NH4-N
11.0
7.6
23.6
Organic
N
1. 5
2. 3
8. 1
NO2-N
0. "~ '
0.
0.
074
046
236
NO3-N
2. ~~
8.
1.
\{.
45
34
Total
P04-P
10. 0
18. 5
18.6
MB AS
0. 13
1.42
2.41
Fountain Valley Wastewater
Alum-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Raw Sewage
Winter
Winter
Summer**
Winter
Summer**
Winter
Summer**
2161
2240
2290
2375
2673
2833
3120
195
144
115
.
426
318
458
21.6
28.2
26.4
35. 1
39.0
40.4
34. 1
6.0
5.8
3.9
8. 5
31.0
11.7
18.7
0.
0.
0.
0.
0.
0.
0.
004
053
090
042
064
259 .
062
0.
1.
0.
0.
0.
0.
1.
48
87
65
37
54
87
05
0.6
9. 3
7.6
7.2
9. 3
9. 3
6.0
1. 12
1. 14
0.92
2.10
-
2.60
0. 50
Digester Sewage
Winter
13100
3190
413
256
0.028
18.5
Electrical Conductivity, Airnhos/cm at 25" C
l(jt(
Summer sewage at Fountain Valley contains a waste discharge from a sugar beet refinery.
2.8
28.3
-------
Figure 1. FLAT-PLATE TEST APPARATUS
-------
Waste
Feed
1
1
1
\l
^
1
1
1
[ \
[\
V
c
00
o
o
c
I
— r
i
i
1
i
^>
I
i
I
Ii
JU
O-ring
Seals
1
* *^*
Rnrlf inn"
Paper
Membrane.
1 1
1 I
1 1
1 1
Product
Flat-Plate Test Cell
Feed
Fiberglass Shell
S S / S / S 7 .
Membrane — ^ ,
s s
' 0.
0.
^
25
56
' / y ^
in. or
in. pc
' /
iper
f s s
Backing
Waste
I * t
Product
Tubular Test Cell
Figure 2. LABORATORY MEMBRANE CONFIGURATIONS
11
-------
Tubular membrane configurations were tested with the support appa-
ratus pictured in Figures 2 and 3. These stands were used in the
evaluation of both the 0. 25- and 0. 56-in. diameter tubular membranes.
A flow sheet for the laboratory test apparatus is shown in Figure 4.
MEMBRANES
The reverse osmosis membranes used in this program were cast from
conventional cellulose acetate formulations. Of the two sheet mem-
branes, differentiated by their annealing temperatures in °C, the 68°
membrane produced during the first two hours of operation 30 ga.ll-
(sq ft)(day), or 29/*g/(sq cm)(sec)(atm) ,* of product water from a 0. 57-
percent NaCl solution at 700 psig with an 87-percent rejection of NaCl.
The 44° membrane produced 20 gal/(sq ft)(day), or 68>*g/(sq cm)(sec)-
(atm), with the same feed water but at 200 psig, and possessed lower
salt rejections of 24 percent. Different methods of casting tubular
membranes introduced small variations in membrane performance and
therefore the two types are considered to have different characteristics.
The 0. 25-in. tubular membrane delivered 23 gal/(sq ft)(day), or 20 yug/-
(sq cm)(sec)(atm), with the 1-percent NaCl solution at 800 psig, and
provided a NaCl rejection of 85 percent. The 0. 56-in. tubular mem-
brane provided 15 gal/(sq ft)(day), or 15/*g/(sq cm)(sec)(atm), and pos-
sessed an 81-percent NaCl rejection with a 1-percent NaCl solution.
Three 2. 5-in. diameter discs were used in series arrangement in the
flat-plate test cells, whereas single 14- and 10-in. lengths of 0. 25- and
0. 56-in. diameter tubes, respectively, were used in the tubular test
assemblies.
Virgin membranes were employed at the beginning of each test run.
OPERATING CONDITIONS
All tests were conducted continuously 24 hours a day, seven days a week
throughout their duration. All equipment was shut down routinely every
eight hours for several minutes to produce a depressurization and back-
flow of product water through the membrane to cleanse it and help main-
tain product water flux.
Two operating pressures were utilized in the laboratory program. The
68° sheet membranes and all the tubular units were operated at 700
psig. Some tests were performed on the 68° membranes at 200 psig.
The 44° flat-plate membranes were tested at 200 psig. Most of the
tests in this program were conducted at a product water recovery ratio
of 80 percent, so that the membranes experienced concentrated feed
similar to that found in a full-scale plant after 80 percent of the feed
1 Mg/(sq cm)(sec)(atm) is equivalent to 1. 01 gal/(sq ft)(day) at 700
psig.
12
-------
Figure 3. TUBULAR TEST APPARATUS
-------
Fresh Sewage
I
Pretreatment
Feed
Cylinder
Mixing
Chamber
Heat
Exchanger
-^ Overflow
^r > -
•^-J Concentrate Membrane
V
'rodu<
Pressure
Regulator
B t
Product
Figure 4. LABORATORY TEST APPARATUS FLOW SHEET
-------
water had been removed as product. Excessive salt build-up in the re-
circulating system was prevented by removal of concentrate from the
mixing chamber by means of an overflow tube or blowdown. Initial con-
trol of the recovery ratio was done by balancing the product and wasted
waters against the feed input. This system was eventually modified in
that at the time of concentration of the feed water as determined by
mass balance, the electrical conductivity of the concentrate was record-
ed and thereafter maintained at that level throughout the test. Concen-
tration of all feed waters to the 80-percent recovery condition was ef-
fected in the reverse osmosis apparatus with the same membrane uti-
lized for the conduct of the test.
Further modifications of the sewage quality, such as by acid addition,
were accomplished in the mixing chamber from which the feed water
flowed to the pumps, through the reverse osmosis units, and back again
in slightly more concentrated form.
Most tests in this program were conducted at a pH of 5, controlled with
sulfuric acid, to assure that no phosphate or carbonate precipitation
would occur. However, pH values as high as 8 were employed with one
of the more promising additives.
Each unit was operated at a constant throughput that provided calculated
nominal Reynolds numbers of about 3,000 for the flat-plate test cells
and 5,000 for both tubular cross-sections.
The introduction of additives to the various sewage feeds was intended
to prevent product water flux decline caused by the deposition of solids
on membrane surfaces. The additives chosen for this study were Zim-
mite 190 and 120, anionic flocculants; Calgon, a chelating agent;
Cyanamer, a dispersant; Biz, an enzyme-active laundry presoaking
agent; and alum, a cationic coagulating agent. These additives were
maintained at constant dosages throughout most of the test program with
the various grades of wastewater, except where noted otherwise. The
feed water concentrations of 2. 2 mg/1 of Zimmite 190 and lOOMl/1 of
Zimmite 120 were selected as the standard dosages based on the manu-
facturer's recommended range for maintaining clean sewer lines with
these products. Calgon and Cyanamer additives were employed at
10 mg/1 based on previous successful results on brackish water at this
concentration. Biz was used at 20 mg/1 on the basis of an arbitrary
upper economic level. The optimum dosage of alum was determined
daily using jar tests on fresh sewage; typically, these were 125 and
400 mg/1 for secondary and primary sewage, respectively.
An inventory of all individual tests performed in this program is pre-
sented in Table 2.
The recirculated wastewater quality was initially checked for pH and
total dissolved solids by electrical conductivity every hour for 16 hours
of the day, but was reduced eventually to bi-hourly sampling for 16
hours. The feed waters were chlorinated and the concentrate was test-
ed twice a day for chlorine residual, which was maintained at less than
5 mg/1.
15
-------
Table 2
LABORATORY TEST SCHEDULE
Test Membrane Feedwater Pressure Additive ^References
(psig)
2
3
4
5
6
7
8
9
10
11
12
13
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
44
45
46
47
48
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Sheet
Tube
Tube
Tube
Sheet
Sheet
Sheet
Sheet
Tube
Tube
Sheet
Tube
Sheet
Carbon
Carbon
Carbon
Carbon
Carbon
Carbon
Carbon
Secondary
Secondary
Secondary
Secondary
Secondary
Primary
Primary
Primary
Raw
Primary
Raw
Primary
Primary
Digester
Raw
Raw
Secondary
Digester
Digester
Digester
Secondary
Secondary
Secondary
Secondary
Secondary
Secondary
Secondary
Secondary
Secondary
Carbon
Carbon
Secondary
Secondary
Secondary
Carbon
Primary
Secondary
700
700
200
200
200
700
700
700
200
200
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
200
200
700
700
700
700
700
700
700
700
700
700
700
700
700
700
Alum
ABS
ABS
Calgon
Cyanamer
Cyanamer
Calgon
Zml90
Zml90
Calgon
Cyanamer
Zm 190
Calgon
Cyanamer
Zm 190
Alum
Alum
Alum
Alum-Cy
Alum-Zm
Alum-Zm
Alum-Zm
Alum
Alum-Cal
Calgon
Zm 190
Biz
Zm 190
Cyanamer
Zm 190
Figure
5
7
14
15
13
11
12
17
29
28
22
23
53
37
52
58
41
59
40
39
62
60
61
20
64
65
63
18
31
32
27
25
76
77
70
26
9
8
24
68
73
10
84
16
Page
22
24
32
33
31
28
29
35
47
46
40
41
71
55
70
77
59
78
58
57
81
79
80
38
83
84
82
36
49
50
45
43
96
97
89
44
26
25
42
87
93
27
104
34
16
-------
Table 2 (continued)
LABORATORY TEST SCHEDULE
Test Membrane
49
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
68
69
70
71
72
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
Sheet
Tube
Sheet
Sheet
Sheet
Tube
Tube
Tube
Tube
Sheet
Sheet
Tube
Sheet
Tube
Tube
Sheet
Sheet
Sheet
Tube
Tube
Tube
Sheet
Sheet
Tube
Tube
Sheet
Sheet
Tube
Sheet
Tube
Tube
Sheet
Sheet
Tube
Tube
Sheet
Sheet
Tube
Sheet
Tube
Tube
Sheet
Sheet
Sheet
Tube
Feedwater
Primary
Carbon
Primary
Primary
Primary
Primary
Carbon
Secondary
Secondary
Primary
Primary
Secondary
Primary
Raw
Raw
Primary
Primary
Primary
Primary
Secondary
Primary
Primary
Primary
Secondary
Primary
Carbon
Primary
Primary
Primary
Primary
Primary
Primary
Secondary
Primary
Primary
Secondary
Secondary
Primary
Secondary
Primary
Primary
Secondary
Secondary
Secondary
Primary
Pressure
(psig)
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
200
700
200
700
700
700
700
700
700
700
700
700
700
200
700
700
700
700
700
700
200
200
700
200
700
700
200
200
700
700
Additive
Biz
Zm 190
Zm 190
Zm 190
Zm 190
Zm 190
Zm 190
Zm 190
Calgon
Zm 190
Zm 190
Zm 190
Zm 190
Alum
Zm 190
Zm 120
Zm 190
Alum
Alum
Zm 190
Calgon
Zm 190
Zm 190
Zm 190
Zm 190
Zm 190
Cyanamer
Zm 190
Zm 190
Zm 190
Biz
Zm 190
Zm 190
Zm 190
Zm 190
Zm 190
C-31
References
Figure Page
54
66
49
43
47
82
67
69
74
42
50
75
46
92
93
55
38
57
78
72
81
51
44
71
79
6
45
89
56
86
83
48
21
90
87
36
35
91
30
88
85
34
33
19
80
72
85
67
61
65
102
86
88
94
60
68
95
64
112
113
74
56
76
98
91
101
69
62
90
99
23
63
109
75
106
103
66
39
110
107
54
53
111
48
108
105
52
51
37
100
17
-------
MEASUREMENTS
At the beginning of the program, samples taken for chemical analysis
were 24-hr composites of the feed, waste, and product waters collected
every two days. Once membrane performance with regard to waste con-
stituent rejection had been established, chemical analyses were perform-
ed as a check on gross membrane deterioration and concentrate quality
and were conducted on 24-hr composite samples every week. The follow-
ing constituents were monitored and analyzed in accordance with proce-
dures outlined in the Twelfth Edition of Standard Methods for the Examin-
ation of Water and Wastewaters, American Public Health Association,
Inc. , or FWQA approved alternate methods: pH, electrical conductivity,
total dissolved solids, chemical oxygen demand, biochemical oxygen de-
mand, organic nitrogen, ammonium, nitrite, nitrate, total phosphate,
and methylene blue active substances.
Electrical conductivity was determined by using a conductivity bridge
with a 4. 5-ml capacity cell.
Total dissolved solids were measured by weighing the residue from a
filtered (0. 45fi) and evaporated sample.
Total chemical oxygen demand was determined by potassium dichromate-
sulfuric acid digestion for two hours and ferrous ammonium sulfate
titration to the ferroin indicator endpoint; any chloride present in the
sample was complexed with mercuric sulfate.
Ammonium contained in samples was distilled from Kjeldahl flasks and
collected in boric acid solution. The amount of ammonia in the dis-
tillate was determined either colorimetrically at 467-m/i wavelength in
a spectrophotometer following Nesslerization (less than 5 mg/1 ammonia)
or volumetrically by titration with sulfuric acid using methyl red-alpha -
zurine mixed indicator (greater than 5 mg/1 ammonia).
Organic nitrogen content of samples was ascertained by difference be-
tween the results of total unoxidized nitrogen and ammonia nitrogen
analyses. Total unoxidized nitrogen concentration was determined by
conversion to ammonia through sulfuric acid digestion (in the presence
of potassium sulfate and copper sulfate), collection in boric acid solu-
tion by alkaline distillation, and detection either by Nesslerization or
acid titration.
Analyses for nitrite content were performed on samples by reacting
with sulfanilic acid, developing color with N-(l-naphthyl)-ethylenedia-
mine dihydrochloride, and determining optical density at a wavelength
of 550 in/* in a spectrophotometer.
Nitrate analyses were performed spectrophotometrically using a brucine-
sulfanilic acid and sodium chloride color development procedure by the
method of Kahn and Brezenski from "Determination of Nitrate in Estau-
rine Waters," Environmental Science and Technology, 1, 488-49(1967).
18
-------
Total phosphate analyses were performed on samples by digestion with
concentrated nitric acid and strong sulfuric acid, color development of
the orthophosphate with ammonium molybdate, sulfuric acid, ascorbic
acid, and potassium antimonyl-tartrate mixture, and measurement of
absorbance at a wavelength of 885 mpt in a spectrophotometer.
Methylene blue active substance was determined by extraction of the
methylene blue complex with chloroform and measurement of light
absorbance at 650-mjx wavelength in a spectrophotometer.
Hydrogen ion concentration of samples was measured with a pH meter
using a glass electrode and calomel reference cell.
Residual chlorine concentrations -were determined utilizing the thiosul-
fate-iodide titration method.
The standard 5-day, 20° C dilution bottle method was used for biochemi-
cal oxygen demand determinations; dissolved oxygen analyses were made
using the Winkler, azide modification, method.
Collected samples were refrigerated at 0°C prior to analysis to reduce
deterioration of various sewage constituents.
Based on ten samplings for each feed water, a ratio between total dis-
solved solids and electrical conductivity was determined. A value of
0. 6 may be considered representative of the TDSrEC ratios. In the case
where slightly ionized macromolecules, such as duodecal benzyl sulfon-
ic acid, were added to relatively high-quality feed water, the ratio in-
creased to as high as 0. 9 in the feed and recycle waters while remain-
ing near 0. 6 in the product stream.
A number of biological oxygen demand analyses were performed to deter-
mine the relationship between the BOD and COD values. The raw sewage
feed water and wastewater streams had a CODrBOD ratio of approxi-
mately 2. The digester, primary, and secondary sewage feed and waste
streams had ratios ranging from 3 to 7, with an average of 4, while car-
bon-treated and alum-treated, sand-filtered sewages provided ratios
ranging from 6 to 14, with an average of 6 for the feed waters and 13 for
the waste streams. The ratios for the respective product water streams
were double the values obtained for the wastewater streams. The in-
creasing difference between COD and BOD with progressing sewage treat-
ment is expected, as the biologically oxidizable materials are gradually
removed while the relatively nonbiodegradable constituents remain and
increase in relative proportion.
DATA REDUCTION
Membrane performance was measured at the 80-percent or higher re-
covery levels maintained during each test run. Data acquired during
the concentration of the feed to the test level were not included because,
at conditions other than fully concentrated feed, membrane behavior is
19
-------
nonrepresentative and data become more ambiguous with the addition
of a variable recovery ratio. Initial concentration time varied with
product water flux but usually ranged from eight hours to two days.
It was originally believed that a 2-week test period would be adequate
to establish long-term membrane performance. Also tests which re-
sulted in a greater than 50-percent flux loss were not considered
worthy of further study and were discontinued. Laboratory experience
gradually demonstrated that with a testing period of three weeks, ulti-
mate long-term flux could be established with reasonable certainty.
Moreover it was established that initial flux declines were not neces-
sarily indicative of future trends. For these reasons, test durations
are not the same throughout the period of study.
Initial examination of the data on product water fluxes indicated that
most test results may be presented as exponential functions of time.
Certain results however do not fit this description, but on closer exam-
ination may be seen as approaching a steady-state condition of zero flux
decline. In such cases initial decline data are still felt to be best repre-
sented in exponential form with the steady-state condition of no decline
needing no mathematical representation. Due to the short duration of
many of the test runs it would be difficult to ascertain when the mathe-
matical function is no longer consistent with the data. Therefore flux
decline data have been calculated using all of the data points presented
and are given in the form of the coefficients J , k, and a to fit the
general equation
T T kt ± a
J = J e
o
where J is product water flux in gal/(sq ft)(day), JQ is initial flux,
in gal/(sq ft)(day), upon attainment of specified concentration, k
is the flux decline coefficient in day~l, t is expressed in days of
operation at desired wastewater concentration condition, and a is
the standard deviation of the flux decline exponent.
Inasmuch as the ultimate aim of this program was to establish the best
process for wastewater treatment by reverse osmosis, the only truly
acceptable result, which has been achieved in several instances, is a
flux decline of zero or nearly zero. With such results available, the
direct comparison of flux declines markedly different from zero be-
comes a somewhat academic matter. However, the rate of flux de-
cline for a given set of conditions does indicate how rapidly the steady-
state flux will be reached and does provide for a standard of compari-
son when the final flux is not determined in a test.
The flux decline expressions were calculated from 8-hr composited
volumes of product, with three data bits per day. The fluxes were
converted to their logarithms and with their respective time periods
were used to find the best fit of the data by least-squares regression.
Membrane performance with respect to constituent rejection was
calculated on the basis of what actually contacts the membrane surface
at the particular recovery ratio.
20
-------
Section IV
LABORATORY RESULTS
The test results obtained in this program represent a large number and
diversity of factors that influence the performance of the reverse os-
mosis process. Performance is measured by two characteristics—pro-
duct water flux and wastewater constituent rejection. The magnitude
and temporal variation in product water flux are dependent upon the type
or character of feed water, test-cell geometry, membrane type or per-
meability, operating pressure, product water recovery ratio, and type
and dosage of additives. On the other hand solute or pollutant rejection
for this application and under the test conditions studied appears to be
relatively independent of all the aforementioned factors with the excep-
tion of feed water type and membrane permeability.
Product water flux results are presented in the form of figures which al-
so indicate values for all variable test parameters and calculated coeffi-
cients for the product water flux decline. All feed waters, except carbon-
treated secondary sewage, originating at the Pomona Plant are so desig-
nated while feed water supplies from Fountain Valley carry no special
designation. Carbon-treated sewage was available only at Pomona.
PRODUCT WATER FLUX
Flat-plate Test Cells
The product water fluxes obtained with the flat-plate test assemblies un-
der a variety of operational conditions are presented according to the type
or quality of municipal wastewater.
Carbon-treated Secondary Sewage
Carbon-treated secondary sewage was the highest quality effluent obtain-
able for this program. This sewage established the best performance
in this program for the reverse osmosis process with municipal waste-
waters. Figures 5 and 6 represent control tests on carbon-treated
secondary sewage with the 68° membrane. The aluminum content of
4 mg/1 expressed as alum in the feed water of Test 2 was naturally pre-
sent. Subsequent tests were not analyzed for their aluminum content.
Test 77 was a retrial of Test 2 after several months of operation. The
lower product water flux in Test 77 may be attributed to both a practical
inability to fully clean the test equipment of the deposits from other sew-
ages and the previously mentioned feed water quality differences between
the two sampling dates.
Figures 7, 8, 9, and 10 present values obtained with alum, Zimmite 190,
Calgon, and Cyanamer. Again it should be mentioned that Tests 39, 40,
and 46 were conducted on equipment that had been used to process waste-
waters of much lesser quality. Figures 11 and 12 demonstrated attempts
to find the effects of methylene blue active substances on product water
flux.
21
-------
30
20
•
fe
a
r 10
N
'
N
i J
()
j
Q
O
11
a
K
-
_j i i i
"•^x
i i
^-o-o-^o-o^
I I I I
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
4 mg/( Alum
J = 18.9 k = -0.0060 a= 0.096
o
1 1 1
1
1
-
-
10
15
TIME, days
20
- 30
- 20
10 -|
5
— 4
— 3
— 2
10
O
• '
fl
DO
:
Figure 5. PRODUCT WATER FLUX, TEST 2
-------
30
20
K
D
~ 10
s
a
X
^ H1 5
cc
LLJ
t- 4
<
a
'
a
-I
—
K
^"***
-------
30
•
o
10
:V
I
•
-
>
ts)
Q
in
I
•'•
••
O 3
Q
;;
o
i i i
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
175 mg/f Alum clarified
J = 18.3
o
k = -0.0164 a = 0.118
J L
10
15
TIME,days
20
25
- 30
- 20
*
I
- 10 r
u i
T "Z.
3 <
a
"•
u
- 2 S
Figure 7. PRODUCT WATER FLUX, TEST 3
-------
30
>
•••
a
~ 10
s
-
«
i
n
"'
:
a
3
a
a
=
—
V
X«~
1 1 1 1
rfffw-1
s^
i i i i
<
\^N,
1 1 1 1
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
2.2 mg/f Zimmite 190
J =6.1 k = -0.0040 a = 0.126
o
1 1 1 1
1 I 1 1
1 1 1 1
20
10
— f.
— 4
- 3
— 2
if!
II,
TIME, days
2!)
30
'
•
; I ;
i
i
II
DO
Figure 8. PRODUCT WATER FLUX, TEST 40
-------
:f
u
Q
O
a
Q
30
10
5
4
3
2
1
^~
*•!
Q
\
V
I
0
l~°-^N^-*
I
flL <-A_
Z—-*
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH
10 mg/p
J = 6.7
o
1
5 80% Recovery
Calgon
k = -0.0057 a = 0.147
1 1
1 1
- 20
- 10
- 5
- 4
- 3
- 2
5 10 15 20 25 30
o
I-
LU
y
u
u_
HI
O
LU
z
0
a
LU
2 5
TIME, days
Figure 9. PRODUCT WATER FLUX, TEST 39
-------
K)
30
20
I
10
':
i
>
D
5
4
' •
i
D
.
'!
-
—
K
>*
V
*N
1 1 1 1
L
>x-^V.
>>— o
till
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
10 mg/f Cyanamer
J = 11.3 k = -0.0689 a = OJ57
o
1 1 1 1
1 1 1 1
1 1 1 1
- 20
- 10
— c
4
- 3
- 2
i
•:•
. :
i
'
• )
•i
ri
DQ
S
in
15
TIME, days
20
25
0
Figure 10. PRODUCT WATER FLUX, TEST 46
-------
30
8Z
PRODUCT WATER FLUX, gal/(sq ft ) (day)
— . N
_. rvj cj -e» on o
lo-o-o.
>o— <
0
»*-**^s\
l_ 1
J*—o -° -o— <
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
5 mg/f Duodecal Benzyl Sulfonic Acid
J = 15.9 k = -0.0043 a = 0.072
o
1
1
20
- ?
*->
TO
- ^
vt
- 10 -£
U
NJ OJ *k C71
MEMBRANE COEFFICIENT. ^g/(sq
5 0 15 20 25 30
TIME, days
Figure 11. PRODUCT WATER FLUX, TEST 7
-------
PRODUCT WATER FLUX,gal/(sq ft) (day)
— « NJ
-> rsjcojicji o O:
—
"°*^
±111
N>^-^
^
^o
Carbon Treated Secondary Sewage
Flat Plate 68° Membrane
700 ps\g pH 5 80% Recovery
20 mg/f Duodecal Benzyl Sulfonic Acid
J = 20.0 k = -0.0770 a = 0.093
o
- 20
— c
— ro
10 "p
D
5T
O)
^
— I—
•z.
LU
5rr
O
LL
O
O
LU
3 ~Z-
<
a:
CD
s
LU
2 ^
) 5 10 15 20 25 30
TIME, days
Figure 12. PRODUCT WATER FLUX, TEST 8
-------
The results of low pressure operation at 200 psig on 68°, 55° , and 44°
membranes are depicted in Figures 13, 14, and 15.
Secondary Sewage
Secondary sewage with its comparatively low total solids content is of
great interest in the processing of municipal wastewater by the reverse
osmosis process. The results of high pressure operation are first pre-
sented in Figures 16 and 17 which are control tests for the Fountain
Valley and Pomona feed waters. Figures 18, 19, 20, 21, 22, 23, and
24 present test results with continuous addition of alum, 1. 1 and 2. 2
mg/1 of Zimmite 190, Calgon, Cyanamer, and Biz. Typical membrane
deposits were noted in Test 26 (Figure 20), where the product water
flux experienced a steady increase with no apparent cause for the in-
crease in flux. Operating results with dual pretreatment of the feed
water by alum and then addition of Zimmite 190, Calgon, and Cyanamer
are presented in Figures 25, 26, and 27. In all the tests on secondary
sewage with alum pretreatment, the sheet membranes became thickly
deposited with calcium sulfate. This did not occur with carbon-treated
secondary sewage in Test 3 (Figure 7). An analysis of deposits on the
membrane from Test 9 with secondary sewage revealed 98-percent vola-
tile material.
The control tests for low pressure operation with the 68° and 44° mem-
branes were conducted at pH values of 5 and 6 as shown in Figures 28,
29, and 30. Low pressure testing with alum-treated or Zimmite 190-
enhanced feed waters is depicted in Figures 31, 32, 33, 34, 35, and
36. The testing with Zimmite 190 was conducted at a pH of 6, based
on other test results, and explored various concentrations of the addi-
tive to establish the optimum dosage.
Primary Sewage
The successful treatment of primary sewage by reverse osmosis could
substantially change the current methods of wastewater treatment by
replacing conventional secondary treatment processes. The following
tests were conducted with primary sewage at 700 psig. Figures 37,
38, 39, and 40 present the results of the control tests on primary sew-
age, including experiments to determine the impact of minor variations
in operating procedure. Tests 16, 21, and 22 (Figures 37, 40, and 39,
respectively) were relatively short-term due to the approach of holi-
days.
Most of the testing of additives in this program was conducted at fixed
dosages and a pH of 5. The possibility that Zimmite 190 might be more
effective at a different dosage or a higher pH value led to the series of
tests investigating variations in these two parameters. The data from
these tests are shown in Figures 41, 42, 43, 44, 45, 46, 47, 48, 49,
and 50. The comparative performances of the other additives, including
Zimmite 120, an anionic polyelectrolyte, are presented in Figures 51,
52, 53, and 54.
30
-------
•
a
•a
it
a>
'
)
Q
ill
( )
I
Q
a
30
20
10
5
•1
!
2
1
>V .f**-^-
v- ^"-tt*
n
-*» ^*»
"•^•O^V^
V
>
I | L .
~^°~O
*^°~
Carbon Tr»at»d Secondary Sewaae
Flat Plate 68° Membrane
200 psig pH 5 80% Recovery
J = 10.0
o
k = -0.0275
1
a = 0.106
1 1
r70
-60
-50
-40
-JU
— 20
-15
10
- 5
- 4
5 10 15 20 25 30
i '
a
<
'
• 3
a
CQ
:
TIME, days
Figure 13. PRODUCT WATER FLUX, TEST 6
-------
i
30 E
20
•
5
0
~ 10
8,
-
i
H
' '
O
cr
B
I
—
_1_ JL_
Sx^*-^
''''
~~~\/~
i i I
Carbon Treated Secondary Sewage
Flat Plate 44° Membrane
200 psig pH 5 80% Recovery
>
J = 18.0 k = -0.0141 a = 0.066
o
1 1
1 1 1
1 1
r70
-60
-50
-40
^30
-20
-15
1 n
— IU
- 5
- 4
10
IS
TIME, days
20
25
30
I
>
•
I
z
:.!
J
O
.
;
r
Figure 14. PRODUCT WATER FLUX, TEST 4
-------
Ul
'
m
-'<
a
in
o
D
Q
i )
a
a-
30
10
5
4
'1
2
1
-
Iv^
I I I
^
I I I
^x-^o^S
I I I
Carbon Treated Secondary
Flat Plate 55° Membrane
Sewage
200 psig pH 5 80% Recovery
J = 12.5 k = 0.0023 a = 0.101
o
1 1
1 1 1
1 1
r70
-60
-50
-40
-30
-20
-15
m
- 5
- 4
M)
20
r
-> •
ra
•
a
\
u
u
LU
z
<
(i
BQ
:
LJJ
30
TIME, days
Figure 15. PRODUCT WATER FLUX, TEST 5
-------
JU
•JO
1
PRODUCT WATER FLUX,gal/(sq ft)
-> M CO -b Ol
—
~°""*W
v
v-»^
1 1 1
I
-S^
*-**^>^S
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
J =9.2 k = -0.0349 a = 0.109
o
I
1 1 1
1 1 1
20
I |
^ H
h—
>- 10
E
ti CJl
EFFICIENT, Lj.(|'(sg
NJ CO
MEMBRANE CO
0 5 10 15 20 25 30
TIME, days
Figure 16. PRODUCT WATER FLUX, TEST 48
-------
30 r=
20
IB
a
r 10
ft
\- 4
c.
o
a
a.
—
—
^•s^
^>a
V
~
1 1 1
v
\
1 1 1 1
1 1 1 1
Pomona Secondary Sewage
Flat Plate 68% Recovery
700 psig pH 5 80% Recovery
J =9.8 k = -0.1145 a = 0.055
o
1 1 1 1
1 1 1 1
1 1 1 1
- 20
- I
.2
£
Vi
- 10 —
E
1
— (-"
z
UJ
— c —
LL
NJ CJ
MEMBRANE COI
10
15
TIME, days
20
25
30
Figure 17. PRODUCT WATER FLUX, TEST 9
-------
30 f=^
20
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Alum Treated/ Filtered
k =-0.1009 a = 0.120
TO
TJ
10
SI
x
u_
cc
U
3
Q
o
r- 30
- 20
10 -
5?
- 5
u
O
' !
Ill
3 z
ta
5
Figure 18. PRODUCT WATER FLUX, TEST 30
-------
X
)
_)
LL
o:
LU
u
Q
o
a
f-L
30 ,=
20
r 10
—
—
\ ^/\
v^
1 1
.
r v ^
i i i i
i i i i
Pomona Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 6 80% Recovery
1.1 mg/l Zimmite 190
J =6.1 k= 0.0150 a = 0.231
o
|
1 1 1
1 1 1 1
20
• 10
- 5
4
- 3
— 2
*
<;
I LI
O
HI
' '
in
z
i
!L
rn
;
LU
JO
15
TIME, days
20
25
Figure 19. PRODUCT WATER FLUX, TEST 95
-------
30 t=
.'II
10
S
i
•
X
00
0
Q
O
cr
Q
—
-
«/Sv
^r
A
A
/v
*s~*
s
**f
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
2.2 mg/l Zimmite 190
J =5.4 k= 0.0296 a = 0.166
o
MEMBRANES PHOTOGRAPHED
"AND DISTURBED
1
1 1
1 1
• 20
- 10
— 5
- 4
- 3
- 2
0 ' ' ' "5 0 15 20 25 30
E
'•
o
ff
•
1 '
O
•
TIME, days
Figure 20. PRODUCT WATER FLUX, TEST 26
-------
re
TD
-
D
w—i
\OI-L
QC
in
h-
•r
(-
' i
>
("i
o
o
a
30
20
r 10
*> CJl
—
—
1 1 1 1
^s*
***»• 0 ,,4
till
h-MV __
^*-
-------
(0
20
a
0
~ 10
-
3
> -
.
i. i
(j
a
a
—
i\
\
\
\
A
A
\ i/ ^
•
i i i i
x
>
- MEMBRAN:
till
S
CS PHOTOGRA
1 1 1 1
Pomona Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
10 mg/f Calgon
J = 10.0 k = -0.0777 a = 0.220
o
PHED AND DU
1 1 1 I
5TURBED
1 1 1 1
1 1 1 1
20
10
- 5
i — 4
ta o
2
10
15
TIME, days
20
25
30
' :
4
a
Figure 22. PRODUCT WATER FLUX, TEST 12
-------
•
g
a
30
20
10
•
i
D
111
I
<
u
.'
I I
a
Q
—
—
1 1
' Sewage
68 Membrane
pH 5 80% Recovery
Cyanamer
k = 0.0095 a = 0.295
D DISTURBEL
1
— JU
?o
10
A
r>
9
10
15
TIME.days
20
25
30
! )
111
I
U
HI
, '
•i
fi
i
Figure 23. PRODUCT WATER FLUX, TEST 13
-------
30
20
•O
10
1
X
3
fc=! 5
oc
LU
I- 4
<
u 3
D J
O
o
a
a.
-
—
\
\
\
\
\
>
till
JV
V V
V
1 1 1 1
^~,
1 1
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
20 mg/t Biz
J =8.6 k = -0.0849 a = 0.215
o
+*
1 1 1 1
1 1 1 1
1 1
-
-
-
- 30
- 20
E
S
I
kio r
u
5
Ill
u
a.
8
LU
;1
a
co
u
5
10
15
TIME, days
20
25
30
Figure 24. PRODUCT WATER FLUX, TEST 41
-------
30
20
10
*
s
a
X
d
ill
)
D
O
or
a
™
^v^
1 1 1 1
^V
i i i i
L
^•»
I 1 1 1
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered, 2.2 mg/f Zimmite
190
J = 15.5 k = -0.0662 a = 0.099
o
| |
| |
1 1 1
-
-
-
-
— 30
- 20
10 -
5
4 LLJ
O
1 -
LLJ
3 1
a.
CO
10
15
TIME, days
20
25
10
Figure 25. PRODUCT WATER FLUX, TEST 34
-------
-
i
Q
I
<
' >
)
O
j
DC
u.
30
20
I
10
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered, 10 mg/f Calgon
a = 0.356
r- 30
- 20
j
,
- 10 -=•
i
•
V
. >
— 3
— 2
111
;
-
u
ii
.
•i
Q
.
Figure 26. PRODUCT WATER FLUX, TEST 38
-------
30
20
n
•
~ 10
X
U1 LL
I
h
u
a
o
a
a
E
4
_,
E\
\
V^i
u* ^
_s
i i i i
»-*>
^^*<
i i i i
^—^
i i i
Secondary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered, 10 mg/ip Cyanamer
J =9.4 k = -0.0628 a = 0.158
o
Si
1 1 1 1
1 1 1
— JU
20
10
- 3
2
if:
E
*
IB
'
in
< i
' I
1
:u
'•••
a:
DO
20
25
10
TIME, days
Figure 27. PRODUCT WATER FLUX, TEST 33
-------
O
30
20
IU
5
1
\
\.
0
k~*~^^
| | | |
^O "HI
Pomona Secondary Sewage
Flat Plate 68° Membrane
200 psig pH 5 80% Recovery
j = 7.4 k = -0.0323 a = 0.060
1
-70
-60
-50
—
-40
-30
-20
-15
10
- 5
- 4
5 10 15 20 25 30
1
(1,
a
I
.'
in
U
'
;
-
TIME, days
Figure 28. PRODUCT WATER FLUX, TEST 11
-------
4
- 1
I
o
_j
0
o
DC
o.
30
10
5
A
3
2
—
—
~^^
1 1 1 1
n^^-^o^j
i
fc.^.^v,,^— 0
1 1 1 1
Pomona Secondary Sewage
Flaf Plate 44 Membrane
200 prig
J =6.7
o
pH 5 80% Recovery
k = -0.0218 a = 0.043
1 1 1
1 1 1 1
-70
-60
-50
-40
-30
-'20
-15
-10
- 5
- 4
10
15
TIME, days
20
25
u
s
'i.
Z
O
U
OQ
30
Figure 29, PRODUCT WATER FLUX, TEST 10
-------
oo
30
20
~ 10
ft
*
3
O o
D J
O
O
oe
a
-
K
i i
*NV_
x^^-o^^
1 1 1 1
t*'A\.
/ ^v
1 1
Pomona Secondary Sewage
Flat Plate 44 Membrane
200 psig pH 6 80% Recovery
J = 11.1 k = -0.0374 a = 0.140
o
^>0— °^»
1
1 1
1 1 1
r-7f)
-60
-50
-40
-30
-20
-15
-10
- 5
- 4
10
is
TIME, days
20
25
•:
>
h
Z
LIJ
u
o
UJ
Z
i
u
Figure 30. PRODUCT WATER FLUX, TEST 90
-------
30 F
20
D
•
to
•o
~ 10
"•
D)
-
oc
LU
I
! -
O
a
Q
O
a
a
E
—
•^s^
I
s*<
s^/^
_i_
n.^> n ii ^.
^s
l
Secondary Sewage
Flat Plate 68° Membrane
200 psig pH 5 80% Recovery
Alum Treated, Filtered
J - 2.8 k = 0.0063 a = 0.084
o
>
I
I
-70
-60
-50
-40
-30
-20
-15
— IU
- 5
- 4
10
15
TIME, days
20
25
:.'
u
II
CD
Figure 31. PRODUCT WATER FLUX, TEST 31
-------
30
20
I
~ 10
M
*
ST
o a. 5
QC
I
O
a
o
tr
a
-
-
/
\y
N^
i i
hv
\
\ **
XX
\r
i i i i
^""^
^Sc
>*u
1 1 1 1
Secondary Sewage
Flat Plate 44° Membrane
200 psig pH 5 80% Recovery
Alum Treated, Filtered
J =9.9 k = -0.0246 a = 0.186
o
>
1 1 1 1
1 1 1 1
1 1 1 1
-70
-60
-50
-40
-30
r20
-15
-10
- 5
- 4
10
i',
TIME, days
20
25
'•
1
Z
LL!
u
o
1 '
•a
tr
a
30
Figure 32. PRODUCT WATER FLUX, TEST 32
-------
30
20
-
a
•o
_ 10
:-.,
X
. n
in
<
U
I
Q
O
a
a
—
—
°"°b*^^
^•^
I I I I
-\
\
I I I I
I I I
Pomona Secondary Sewage
Flat Plate 44° Membrane
200 psig pH 6 80% Recovery
0.2 mg/t Zimmite 190
J =6.1 k = -0.0663 a = 0.191
o
1 1 1 1
till
1 1 1 1
70
-60
— 50
-40
-30
r20
-15
-10
- 5
- 4
io
E
.-
•
11
I
• '
O
U
•i
DC
00
:
15
TIME, days
20
25
Figure 33. PRODUCT WATER FLUX, TEST 94
-------
30
20
10
O
_ 10
X
i 5
CSJ
t 4
(J
a
O
a
a
—
—
~~
—
—
£
O 1 \
\ / \
v V
i i i i
JL A
^A^N
1 1 1 1
1
1 1 1 1
Pomona Secondary Sewage
Flat Plate 44 Membrane
200 psig pH 6 80% Recovery
0.6 mg/P Zimmite 190
j =7.6 k = -0.0304 a = 0.208
0
1 1 1 1
1 1 1 1
1 1 1 1
r~ 7n
-60
-50
-
-40
— 30
-2U
-15
— 1U
- 5
- 4
10
15
TIME, days
20
25
./
,,
I
LLJ
C
a
ca
:
n
JO
Figure 34. PRODUCT WATER FLUX, TEST 93
-------
30 F
20
-
to
•a
~ 10
x
3
i n
L.
I
o
a
o
rr
a
1
~
""""
—
^~
—
—
•
Tn^^^
™*^
1 ! 1
jt
/\
/ v*
/ ^^
v ^
1 1 1 1
^x_^
^'Xt 0 0^
1 1 1 1
Pomona Secondary Sewage
Flat Plate 44 Membrane
200 psig pH 6 80% Recovery
1.1 mg/f Zimmite 190
J = 8.2 k = -0.0296 a = 0.159
o
^^ ^^
*^\^*^
vr
\ |
S, A
v\
w
1 1 1 1
•
1 1 1 1
— 70
-60
-50
-40
-30
-20
-15
-10
- 5
- 4
II!
15
TIME, days
2I i
30
•
::
II
'• i'
III
>
•1
0
.
Figure 35. PRODUCT WATER FLUX, TEST 88
-------
Pomona Secondary Sewage
Flat Plate 44 Membrane
200 psig pH 6 80% Recovery
2.2 mg/f Zimmire 190
k = -0.0146 a = 0.309
i
U!
'
1
I
',
30
Figure 36. PRODUCT WATER FLUX, TEST 87
-------
•-
D
-
-
Ul -J
(J\ "-
cc
'
: .
!"
it
Q
30
20
10
5
4
3
2
I
o
\
\
\
I I I I
I I
I
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
j =9.5 k = -0.1369 a = 0.106
o
1 1
1
1 1 1
20
- 10
- 5
- 4
- 3
2
M
•
<>
•'
I
Mi
•
O
' '
III
<
a
CD
10
IS
TIME, days
20
25
30
Figure 37. PRODUCT WATER FLUX, TEST 16
-------
JU
20
1
•o
~" m
PRODUCT WATER FLUX, gai/
-------
-
to
D
I ii
'
0
(
>
I I
[I
a
30
20
10
5
4
3
2
i
V
x
X
-^
I I
I
I
Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Double Pump Capacity
J =9,5 k = -0.1584 a = 0.026
o
1
1 1
- 20
- 10
- 5
- 4
- 3
- 2
IE
TIME, days
20
25
30
i-
* <
ce
:
i
,li
1
•i
'
u
<
oc
DQ
Figure 39. PRODUCT WATER FLUX, TEST 22
-------
20
;
r 10
s
-
00 U.
U
Q
O
Q
a
15
TIME, days
20
25
1,1
- ZZ
~ D*
—
\
5
80% PR
k,
\
\
1 1
1 I
1 1 1 1
III!
Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
Accumulator bypassed on shutdown
J = 12.0 k = -0.2212 a = 0.029
o
1 1 I 1
1 1 1 1
till
—
-
<
4—
•
10 -g
•
5
&
:J
LLJ
5
- 3
o
'
I!
z
4
DC
00
tu
.
Figure 40. PRODUCT WATER FLUX, TEST 21
-------
20
i
'
~ 10
:,;
0,
Q
ill
I
<
•
a
< '
ii
Q
1
-
V
X
_5
1 . 1 mg/f
1 1 1 1
S^_^
>**-s>— eT
2.2 mg/p"'
1 1 1 1
/>— *
~S/^
1 1 1 1
Primary Sewage
Flat Plate 68° Membrane
700 psig
1.1 -2.2
Jo = 3.8
**^o
1 1 1 1
pH 5 80% Recovery
mg/t Zimmite 190
k = -0.0075 a = 0.189
1 1 1 1
1 1 1 1
- 20
- 10
4
— 3
- 2
10
15
TIME, days
20
25
:
'
< '
;-
:/
nj
= I
' )
-'
a
OQ
in
Figure 41. PRODUCT WATER FLUX, TEST 19
-------
30
•jn
S-
CD
•O
09
PRODUCT WATER FLUX, gal/(sq ft)
-» N> CO -to CT1
=
—
1.1
ma A
^~
^N
D
9
H- <
2m
n/c
y/'
>-v
5
I 1
10
v
N
I
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
1.1-2.2 mg/f Zimmite 190, Unchlorinated
J =3.1 k = -0.0292 a = 0.082
o
K. y^
15
TIME, days
>^o.
"^
1
f—
20
25
20
- 1
— js
I 1
- 10 •§
o
I
O)
— h-~
•z.
LU
- 5 y
U-
vj to
MEMBRANE CO
30
Figure 42. PRODUCT WATER FLUX, TEST 59
-------
30
20
-
T3
10
-
>
0s LL
"- DC
11
I
•r
< )
-
i i
O
ii
a
I REPLACED 1 MEMBRANE
1 1 I
I I I
Pomona Primary Sewage
Flat Plate 68° Membrane
700 pisg pH 5 80% Recovery
2.2 mg/t Zimmite 190
I I I I
I
I I I
I I
10
15
TIME, days
25
r- 30
- 20
E
IB
1-10
- 3
• ii
i
in
' .1
II!
o
'• 1
-1
DC
CD
UJ
- 2
id
Figure 43. PRODUCT WATER FLUX, TEST 53
-------
30
20
»
•o
10
n
ft
u. 5
tr
i4
§ 3
Q
O
a
a
m^^
1 J
"^a -/
^8
IIII
V
>_ _*
X /^
x /
\
\ /
v
V
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 6 80% Recovery
2.2 mg/f Zimmite 190
J =9.2 k = -0.0213 a = 0.215
o
_f/>v
\_
TO <
IIII
r~
IIII
1 1 1
I— 30
- 20
I
* •
IB
10 ?
•
a
: •:,
I
5 5
. **-
4 LLJ
O
'
- 3
- 2
fl
'II
:
15
TIME, days
20
25
30
Figure 44. PRODUCT WATER FLUX, TEST 74
-------
30 ,=:
20
T5
10
Si
n
' h
Q
LLJ
•f
U
'
Q
( i
(i
Q
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 7 80% Recovery
2.2 mg/P Zimmite 190
k = -0.0236 a = 0.161
I
*^
IS
!
i
: i
' J
! ?
a
OQ
5
Hi
30
Figure 45. PRODUCT WATER FLUX, TEST 78
-------
20
I
10
&
-
^ 5
cc
o 3
a
o
a
n
—
—
s:
Vs
i i i i
l\ y-8^
i i i i
•%_ X
^Or
I I
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig
2 . 2 mg/f
J =8.9
o
_s^
111)
pH 8 80% Recovery
Zimmite 190
k = -0.0378 a = 0.122
1 I |
1 1 1
-
—
-
IG
15
TIME, days
20
- 30
- 20
- 10
E
.
I
U
:.
— 4
i j
c i
- ;
— 3 Z
cc
to
— 2
if)
Figure 46. PRODUCT WATER FLUX, TEST 62
-------
30 f=
20
>
T3
10
-•
<
5
o -1
01 "- 5
oc
f i
D
a
.
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
4.4mg/f Zimmite 190
k = -0.0416
'-
•
•
',-
:
i
i;;
!,
LU
o
:
ti.
co
;
LU
30
Figure 47. PRODUCT WATER FLUX, TEST 54
-------
30
or*
99
PRODUCT WATER FLUX,gal/(sq ft) (day)
— > s
_> PO CO -t* U1
*-— ^
0
^
^\XA
i i i
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 6 80% Recovery
4.4 mg/t Zimmite 190
J =3.9 k = -0.0174 a = 0.158
o
•^-S/
_L _L
i i
- 20
E i
- 10 E
u
sr
O)
Z
ai
- 5 o
LL
— 4 UJ
N) OJ
MEMBRANE CO
5 10 15 20 25 30
TIME, days
Figure 48. PRODUCT WATER FLUX, TEST 83
-------
30
20
IB
73
~ 10
I
•
a
i
-
LL 5
(X
LU
t 4
< >
a
o
a
a
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
6.6 mg/j? Zimmite 190
k =-0.0011
— 30
- 20
- 10
5
1
2
E
•
••
>
••
it
•-•
:
- 2 ^
10
Figure 49. PRODUCT WATER FLUX, TEST 52
-------
30
ON
00
I
10
a
9
X
LL 5
a:
LU
U
O
O
cr
a
sr
—
—
Zr—
5
I I I I
*^
^^Nr^V
I i I I
r >\
" N
O"iN
i l l l
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig
8.8 mg/(
J =5.1
o
/-v
1
1 1 1 1
pH 5 80% Recovery
Zimmite 190
k = -0.0245 a = 0.172
1 1 1 I
1 1 1 L_
- JU
20
• 10
— 5
— 4
— 2
10
IS
TIME, days
E
• i
;
.i
-
in
U
LU
O
LU
Z
•-!
a
00
a
20
25
10
Figure 50. PRODUCT WATER FLUX, TEST 60
-------
30 f=
20
-j
> i
O
a
a.
-
—
O— Ok^
^^°^t^
I /\»
V / \
\y
V~^\
v/^
Pomona Primary Sewage
Flat PI are 68° Membrane
700 psig pH 5 80% Recovery
]QQfil/l Zimmire 120
j =6.7 k = -0.0564 a = 0.203
o
^v.
0 5 10 15 20 25 30
j— 30
- 20
- 10
- 5
- 4
- 3
.
I
•
/
i
i
-;
u
a
DO
TIME, days
Figure 51. PRODUCT WATER FLUX, TEST 72
-------
30 f=
20
10
X
cr
LU
I
(J
3
Q
O
n
a
—
-
«V— _
^v
X
\
\
_J 1 1 1
s,
1
1 1
Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
10 mg/f Calgon
J = 10.0 k = -0.1954 a = 0.108
o
1 1
1 1
20
- 10
4
- 3
- 2
•
"
i
'
I
at
)
LI;
O
u
1,1
•1
a
00
10
15
TIME, days
20
Figure 52. PRODUCT WATER FLUX, TEST 17
-------
30
20
a
O
K
g
-
-o 3
D
' >
I
Q
O
0
a
10
~
-
\
V
\
V
0
I 1 1 1
Pomona Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
10 mg/P Cyanamer
J =8.8 k = -0.1606 a = 0.077
o
1 1
1 1
- JU
- 20
- ?
— jo
I 10 f
— i
- \-
z
01
- 5 o
NJ W *•
MEMBRANE COEFFI
5 10 15 20 25
TIME, days
FIGURE 53. PRODUCT WATER FLUX, TEST 15
-------
2L
PRODUCT WATER FLUX. gal/(sq ft ) (day)
-• ro GO .S* CJi O O i
—
—
\
\\
V
I 1 1 1
/\
/ V^
>^
1 I 1 1
Primary Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
20 mgA Biz, Unchlorinated
j =4.9 k = -0.0421 a = 0.146
o
1 1 1 1
till
1 1 1 1
20
- ?
4_J
_ra
I I
- 10 i
u
ST
01
— -4
^ tjl
EFFICIENT, ,
0 CO J
MEMBRANE COf
10
21;
25
TIME, days
Figure 54. PRODUCT WATER FLUX, TEST 49
-------
Low-pressure treatment of primary sewage is reflected in Figures 55,
56, and 57. These tests were conducted with Zimmite 190 at two
dosages and pH values of 5 and 6 in an attempt to produce a very low
flux decline.
Raw Sewage
The treatment of homogenized raw sewage was attempted in the flat-
plate test cells with the high pressure membranes. Figures 58, 59,
60, and 61 reveal the product water fluxes achieved while processing
the raw sewage. In most cases, the membrane material became per-
forated and leaked due to cuts and scratches on the membrane surface
resulting most likely from hard, suspended particles contained in the
raw sewage.
Digester Sewage
Figures 62, 63, 64, and 65 show that treatment of the material from
mixed anaerobic sludge digestion tanks is possible with reverse osmo-
sis, but that severe fouling of the membrane surface will occur and the
product water flux will be very poor. The test runs were quickly ter-
minated as being impractical, even with the addition of the fixed dosages
of Zimmite 190, Calgon, and Cyanamer.
Tubular Membranes
The following presentation includes all the tests conducted with tubular
membranes under various operating conditions and is presented accord-
ing to type or quality of feed water.
Carbon-treated Secondary Sewage
The product water fluxes obtained while processing carbon-treated
secondary effluent with tubular membranes are probably not representa-
tive of those obtainable, due to several factors. Figures 66 and 67 show
results that were probably influenced by residual deposits of waste in
the testing equipment from previous sewage feeds. The membrane sur-
faces were coated with a pale greenish scum that was not typical of
earlier tests conducted on carbon-treated sewage. Moreover, Test 51
(Figure 66) was conducted with a plunger pump which is believed to have
contributed to the system fouling by shredding of its packing. Also these
tests were performed on sewage feeds collected from the reduced capac-
ity, partially exhausted activated-carbon treatment columns.
Secondary Sewage
There were two control tests for secondary sewage with the tubular mem-
brane. Test 44, shown in Figure 68, was contaminated by shredded
pump packing and was rerun as Test 57, presented in Figure 69. The
pretreatment of secondary sewage by alum clarification produced rela-
tively stable product water fluxes as may be seen in Figures 70, 71,
and 72. Test 37 (Figure 70) utilized the plunger pump but with a differ-
ent packing material that extruded excessively under pressure. Test 70
73
-------
4(1
20
•
•1
o
_ 10
n
91
U
Q
O
a
a
=
-
a -
X^
1 1 1
'"^•^.
*^*
1 1 1 1
t 7^
^^V
i i
Pomona P
Flat Plate
200 psig
2.2 mg/t
J =8.2
0
^^
^>0
1
rimary Sewage
> 44 Membrane
pH 5 80% Recovery
Zimmite 190
k = -0.0132 a = 0.177
1
1 1 1
— 70
-60
-50
-40
-30
-20
-IS
-10
- 5
- 4
10
IS
TIME, days
25
30
I
LI I
' )
O
UJ
ir
QQ
Figure 55. PRODUCT WATER FLUX, TEST 65
-------
ie
a
s
•
>
< u
ii
in
4,
••
\
( .)
•
' >
( >
(i
0
JU
20
10
5
3
2
1
—
—
\
\
\ A
\y
V
I I I
^A^
s^X.
i i i i
i
X
till
Pomona Primary Sewage
Flat Plate 44° Membrane
200 psig pH 6 80% Recovery
2.2 mg/f Zimmite 190
J =9.8 k = -0.0587 a = 0.192
o
1 1 1 1
1 1 1 1
1 1 1 I
20
- 10
- 5
- 3
- 2
10
15
TIME, days
20
25
30
>
''
5
• n
I
C
MEMBRANE COE
Figure 56. PRODUCT WATER FLUX, TEST 80
-------
> —
I
X
0-
ae
LLI
Q
O
cr
a
JO
20
10
e
4
3
2
=
— flt- ....,,
%L
^«L
\
\
\
I | |
> -** n
-------
30 ,=
20
>
n
•a
r 10
- •
3.
•
i
'
a
UJ
O
>
Q
O
a
a
—
—
\
v
-tr--
1 1 1
^*— -eW—JS.
;=• — w
1 1 1 1
1 1 1 I
Raw Sewa
Flo* Pint.
700 psig
1 = 4.0
o
III!
36
68° Membran*
oH 5 80% Recovery
k = -0.1 102 a = 0.278
1 1 1 1
1 1 1 1
— 30
p- 20
- 10
- 5
4
- 3
2
•-
r.
'
'
•
=l
i::
•
:
u
M i
'•<
a.
03
10
15
TIME, days
20
25
Figure 58. PRODUCT WATER FLUX, TEST 18
-------
ill
•11
0
~ 10
a
i
•
3
oo
Q
O
DC
a
Raw Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
2.2 mg/f Zimmite 190
= 0.264
MEMBRANES PHOTOGRAPHED
AND DISTURBED
I I I I
I I I I
I I
I- 30
- 20
E
- 10 -p
I
5
i
h-~
in
< j
!1
: I
I
n
Z
•1
n
u
- 2
I I
10
1'.
TIME, days
20
25
30
Figure 59. PRODUCT WATER FLUX, TEST 20
-------
JO
20
-
ig
D
~ 10
;
";
»
i
i 5
.
3
•.-
a
—
—
A
s^ y >*-*'s^
I
MM
I I I 1
MEMBRANES
AND DISTUR
1 1 1 1
^\^
V°^
PHOTOGRAPHE
BED
1 1 1 1
Raw Sewage
Flat Plate 68° Membrane
700 psig pH 5 80% Recovery
10 mg/f Calgon
J =4.7 k = -0.0279 a = 0.164
o
i
D
1 1 1 1
1 1 1 1
1 I 1 1
—
-
15
TIME, days
20
25
- 30
- 20
- 10 r
E
- 3
— 2
!
' '
'
n,
z
•;
DC
DO
to
Figure 60. PRODUCT WATER FLUX, TEST 24
-------
30
I
10
i
K
oou? 5
o oc
UJ
4
o
B
n
™
—
—
(—
~~ m
\
\
\
\
^•V
\ / X
(
r
™—
1 1 1 1
N
-MEMBRANES F
AND DISTURB!
1 1 1 1
<.
V
^-s
'HOTOGRAPHED
ED
1 1 1 1
Raw Sewage
Flat Plate
68° Membrane
700 psig pH 5 80% Recovery
10 mg/i Cyanamer
J =9.2
o
*
1 1 1 1
k = -0.0836
1 1 1 1
a = 0.240
1 1 1 1
20
—
?
-£
I 1
- 10 ~
— 0
sr
a.
»
z
UJ
NJ CO A C
MEMBRANE COEFFIC
15
TIME, days
25
30
Figure 61. PRODUCT WATER FLUX, TEST 25
-------
CD
uu
20
">
•o
~ 10
•*-<
«*-
_§:
-C
PRODUCT WATER FLUX, ga
-> M CO .b Ol
—
—
\
\
1 1 1 \
••••1
—OVERNIGHTS
f 1 1 1 1
HUT DOWN
1 1 1 1
Digester Sewage
Flat Plate 68° Membrane
700 psig pH 5 50% Recovery
J =2.8 k = -0.2472 a = 0.455
o
1 1 1 I
1 1 1 1
1 1 1 1
20
- ?
4^
JE
¥
— (/)
-10 5
o
sr
O)
^.
I —
Z
LU
J .b C
NE COEFFIC
<
CC
CO
LU
2 5
10
20
25
;*.'
TIME, days
Figure 62. PRODUCT WATER FLUX, TEST 23
-------
iJW
20
(0
•a
— in
Z8
PRODUCT WATER FLUX, gal/(sq ft)
-• ro co ^ ui
—
—
\
\
\
1 1 1 1
1 1 1 1
1 1 1 1
Digester Sewage
Flat Plate 68° Membrane
700 psig pH 5 50% Recovery
2.2 mg/t Zimmite 190
j =4.4 k = -0.4598 a = 0.191
o
till
1 1 1 1
1 1 1 1
• 20
= 1
- 10
E
™- U
sr
0)
— (-"
Z
— c —
O
LL
O
u
LU
- 3 Z
NJ
MEMBRA
10
20
TIME, days
Figure 63. PRODUCT WATER FLUX, TEST 29
-------
30 c=
20
n
r 10
UX
Si
II!
5
Q
rr
r.i
—
—
t
\
s
1 1 1 1
1 1 1 1
Digester Sewage
Flat Plate 68° Membrane
700 psig
10 mg/f
o
1 1 1 1
pH 5 50% Recovery
Calgon
k = -0.2497 a = 0.487
1 1 1
1 1 1 1
- 20
ro
= 1
— 10 "c
o
sr
— h-~
z
UJ
~ 5 y
LL
LL
0
O
UJ
— 3 Z
NJ
MEMBRA
10
15
TIME, days
20
25
30
Figure 64. PRODUCT WATER FLUX, TEST 27
-------
30 f=
I
10
i
00
-
K
in
i
.
i
i i
O
ir
a
—
-
\
\
V
Digester Sewage
Flat Plate 68° Membrane
700 psig pH 5 50% Recovery
10 mg/f Cyanamer
J =3.7 k = -0.3764 a = 0.341
o
_L
i
i i i
• -
•
0 5 10 15 20 25 30
- 30
- 20
- '
•
10 ?
.-
z
5 5
- 4
- 3
- 2
o
u
ill
<
re
B
:
TIME, days
Figure 65. PRODUCT WATER FLUX, TEST 28
-------
I
10
z>
£ ^ 5
cr
!
•:
D
—
K
v-°s
I I I
_^v ^-
^^y
^0
I I
1
1 1
Carbon Treated Secondary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
J = 11.5 k = -0.0434 a = 0.133
o
1
1 1
1
- 20
• 10
- 5
4
- 3
- 2
io
IS
TIME,days
<
2!
E
\
••
'
'
ii
GO
30
Figure 66. PRODUCT WATER FLUX, TEST 51
-------
98
PRODUCT WATER FLUX, gal/(sq ft ) (day)
— * ro
-> fo oj *» cji O O :
—
~\
X
-^v^
i i
"***^-*^
U '
I I I
^ ft
Carbon Tre
0.25 in. Ti
700 psig p
ated Secondary Sewage
ibular Membrane
H 5 80% Recovery
k = -0.0390 a = 0.061
— JU
?o
- ?
— 5
~ I
- 10 -g
i CJ1
FFICIENT, u,g/(sq
N3 CJ -
MEMBRANE COE
1 5 10 15 20 25 30
TIME, days
Figure 67. PRODUCT WATER FLUX, TEST 56
-------
oo
-J
X
D
I
u
3
Q
O
DC
Q.
30
20
I
10
5
4
—
— n
\ .
\>^y
^>v
1 1
^
v
>v
A
i
A
yv
/ x_
/ ^N
Secondary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
JQ = 8.8 k = -0.027 a = 0.172
- 30
— /u
in
5
- 3
5 10 15 20 25 3
f
u
O)
h
in
u
u.
LI. I
O
f 1
•1
c
DQ
:
TIME, days
Figure 68. PRODUCT WATER FLUX, TEST 44
-------
s
I
x
Ooi
00 3.
LJJ
i
<*
o
o
K
D
E
20
10
5
4
3
2
1
—
f^V
-A-X
/ \
4 *^
/
0
f
— 5 HR. PU/
I I I
j
»
t
\
o
W
i
REPAIR
1 1 1
10
I
Secondary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
J =7.4 k = -0.0070 a = 0.257
o
15
1
1
20
25
M
- 30
IU
5
1 — 4
— 3
1
II
u
O
111
z
LU
-
TIME, days
Figure 69. PRODUCT WATER FLUX, TEST 57
-------
-
>
• >
>
a
o
u
o
30
20
I
10
;,;
5
4
—
—
1 1 1 1
'^"^•^o-o-*^
1
NH
Secondary Sewage
0.25 In. Tubular Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered
Jo=15.2 k = -0.0173 a -0.069
L-16 HR SHUTDOWN
1
1 1
1
1 1 1
— JU
- 10
5
T
io
15
TIME, days
•I!
25
30
i
•
IS
'
•
;;;
a
3
' I
u
u
HI
< I
'
-1
DC
CO
FIGURE 70. PRODUCT WATER FLUX, TEST 37
-------
30
(0
•o
10
4>
--^
a
X
D
— I
v£>LL 5
LU
S 4
H
CJ 1
D 3
Q
O
er
a.
=
—
—
~^~
—
—
« J
^^**^
h-0^^^
I
^ ^^^
*T
Pomona Secondary Sewage
0.56 in. Tubular Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered
j =7 1 k = 0.0218 a = 0.104
0
_,&
*~
I 1 1
-
• 20
-
E
TO
— ~y
m
— (/>
- 10
E
— o
- 1
4
— t-"
•z.
LU
-55
LL
— 4 UJ
O
o
-32
-------
JV
20
CD
-o
PRODUCT WATER FLUX, gal/(sq ft)
— > NJ 00 f Ul C
—
^^ ^^^^^"^
1
^
>ft--u^
k^^
V
Pomona Secondary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
Alum treated/ unclarified
J =12.3 k = -0.0336 a = 0.166
o
^-O O ^s^.
fr/^
*-
-------
(Figure 72) produced a surprisingly stable flux with alum-treated,
settled, and then thoroughly remixed secondary sewage.
The effects of the addition of Zimmite 190 and of Calgon are presented
in Figures 73, 74, and 75. Test 58 (Figure 74) is a retrial of Test 45
(Figure 73) which had experienced the packing problem. Relatively
high stable fluxes resulted from the combining of alum treatment with
Zimmite 190 and with Cyanamer. Figures 76 and 77 illustrate two of
the most promising tests of this program.
Primary Sewage
Figure 78 presents the control test with primary sewage. In sharp
contrast, Figure 79 shows a 77-day test that had stabilized at a flux
of 4 to 5 gal/(sq ft)(day). This test on alum-treated primary sewage
was conducted at 80-, 90-, and 95-percent recovery levels with no
apparent adverse effects from the higher recovery ratio. Just prior
to termination of the program a similar test was started with Dow C-
31 as the flocculating agent. The results are shown in Figure 80.
The effectiveness of varying the pH between 5 and 6 with several con-
centrations of Zimmite 190 is recorded in Figures 81, 82, 83, 84,
85, 86, 87, and 88. In Test 55 (Figure 82) a daily 2. 5-min tap water
flush of the 0. 25-in. tubular membrane at 12, 500 Reynolds number
was instituted in an attempt to restore product water flux. Tests 81,
86, and 91 (Figures 86, 87, and 88, respectively) experienced rather
stable fluxes with Zimmite 190 at a pH of 6.
The tests with Calgon, Cyanamer, and Biz, as shown in Figures 89,
90, and 91, resulted in at least one interesting observation. Biz,
which had previously been used as a continuous additive, was now used
for a 15-min membrane soaking at 2, 000 mg/1 concentration. Follow-
ing a 15-min flush at 1 gpm with tap water, the product water flux
through the membrane experienced a strong but temporary recovery.
Raw Sewage
Because of the previously observed membrane abrasion, only two tests
were conducted with raw sewage. Figures 92 and 93 show the results
of no pretreatment and the use of Zimmite 190 to control product water
flux decline. The heavier solids content of raw sewage may have in-
activated most of the Zimmite 190 causing no net benefit from the
additive.
PRODUCT WATER QUALITY
Average wastewater constituent rejections obtained in this program for
each of the membrane types are presented in Tables 3, 4, 5, and 6,
which also contain the average product water pollutant concentrations.
The 68° flat-plate membranes and the 0. 25-in. diameter tubular mem-
branes exhibited very similar wastewater constituent rejections. The
44° flat-plate membranes demonstrated poorer rejections at the benefit
92
-------
30 p=r
20
•o
10
;;;
.-•
'
i 5
->
O
a
a
—
— m ^L
\
\
r^
i i
\
Nw is*
<^™O^
1 I 1 1
fc^v
x
>*-0
1 1 1 1
Secondary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
2.2 mgA Zimhife 190
J =13.6 k = -0.1022 a = 0.182
o
1 1 I 1
1 1 1 1
I 1 1 I
- 20
- 10
5
- 4
3
- 2
10
15
TIME, days
••
25
*
ra
I
•
'
(i
i
• i
11
QC
30
Figure 73. PRODUCT WATER FLUX, TEST 45
-------
•
n
0
-
'
OC
LU
'•'
I
f '
3
Q
O
a
a
Ml
20
10
5
4
3
2
1
_l_1
—
5 In
700 psig
2.2 mg/
J =3.7
o
»
1 1 1 1
ry Sewage
Tubular Membrane
pH 5 80% Recovery
p Zimmite 190
k= 0.0845 a = 0.231
1 1 1 1
till
20
— 10
4
- 3
- 2
* -
'-
!
f
5
I
• •
O
•
•i
a
oa
LU
10
i',
TIME, days
20
25
30
Figure 74. PRODUCT WATER FLUX, TEST 58
-------
30
20
Secondary Sewage
0.25 In. Tubular Membrane
700 psig pH 5 80% Recovery
10 mg/l Calgon
J =8.2 k=-0.0659 a = 0.291
o
l- 30
- 20
- 10
5
4
- 3
- 2
*-j
03
u
:;;
01
U
LLJ
o
u
cc
CO
111
30
Figure 75. PRODUCT WATER FLUX, TEST 61
-------
96
PRODUCT WATER FLUX. gal/(sq ft ) (day)
-> NJ co *» cn o o :
|\
E V~^
I I I I
-^
till
" 0 » 0 0 <
I I I
Secondary Sewage
0.25 In. Tubular Membrane
700 psig pH 5 80% Recovery
r Alum Treated, Filtered, 2.2 mg/f Zimmite
190
J = 18.00 k = -0.0167 a = 0.107
o
1 1 1
- ou
20
1
= 1
o
sr
9.
z
UJ
o
LL
O CO J
MEMBRANE COE
05 15 20 25 30
TIME, days
Figure 76. PRODUCT WATER FLUX, TEST 35
-------
Z.6
PRODUCT WATER FLUX, gal/(sq ft ) (day)
-» M U 4X Ol O O if
=
—
1 1 1 1
*— 0-O^y^O^
h^j^-0^0
Secondary Sewage
0.25 In. Tubular Membrane
700 psig pH 5 80% Recovery
Alum Treated, Filtered, 10 mg/f Cyanamer
J = 14.8 k = -0.0103 a = 0.056
o
1 1
- ou
20
*-•
— -5
- 10 -§
u
Sf
en
& Ul
EFFICIENT, /
ro CO
MEMBRANE COf
1 5 10 15 20 25 30
TIME, days
Figure 77. PRODUCT WATER FLUX, TEST 36
-------
30 rr-
20
1
~ 10
!
x
*> 3
oo u. 5
K
111
I- 4
o
K
Q.
Primary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
= 0.245
I— 30
- 20
10
5
4
- 3
— 2
i
s
z
UJ
u
8
LJJ
I
UJ
5
Figure 78. PRODUCT WATER FLUX, TEST 69
-------
0
D
1
i"
*: 10
ft
•
X 7
3 .
5 .
8 3
E
— i
1111
^
-j- . i i i
iiii
Sr —
1111
' ' 1
to% "™
1111
—
1
-~
'
JX
1 1
~^ ' • • 1
1 1 1
1..
1 1 1
— J
1
1 1
•
1
Pomona Primary Sewap*
0.56 in
Tubular Mention
700 pig pH 5 80-95% Recovery
Alum T
«ai«J, Fllnnd
.7k- -0.0033 o - 0.<99
-^-
M% RECOVERY l\f
i i i
:
SUPERS'
TtO
1 1 1 1
J~— '
TURATION
1 1 1
-
•n
-
E 1
10 1
'
1
I
' i
- , t
Figure 79. PRODUCT WATER FLUX, TEST 76
-------
dU
20
i
10
001
PRODUCT WATER FLUX, gal/(sq ft
-» to GJ *k in
—
—
^**^5sfc%s
1 1 1 1
1
1 1 1 1
1 1
Pomona Pri
0.56 in. Ti
700psig p
Dow C-31
J = 10.1
o
1 1 1
mary Sewage
jbular Membrane
H 5 S0% Recovery
treated/ filtered
k = -0.0413 a»0.107
- 20
- ?
— * to
= 1
- 10 -£
~0>
Z
111
CJ
LL
LL
0
CJ
LU
3 z
tr
a
5
LU
2
) 5 10 15 20 25 30
TIME, days
Figure 80. PRODUCT WATER FLUX, TEST 96
-------
30
20
IB
0
r 10
s
i
X
-3
2- 6
ill
I- 4
! >
3
O
o
a
a
Primary Sewage
0.25 In. Tubular Membrane
700 psig pH 5 80% Recovery
2.2 mg/f Zimmite 190
J =4.1 k =-0.0339 a = 0.108
r- 30
- 20
- 10
i
u
5
'!
- 3
- 2
LLJ
5
in
u
,-
•\
a
00
I I I
10
20
25
30
TIME, days
Figure 81. PRODUCT WATER FLUX, TEST 71
-------
30
I
10
5j
i
••
& ^ 5
t\> DC
u
Q
O
a
n
-
—
•»
r^S
/ \
/ 3
/
i i i i
\ r"
V
i i i i
•*-*-,
i
s /
— START
I I
Pomona Primary Sewage
0.25 in. Tubular Membrane
700 psig
2.2 mg/f
s.
^^.
2 1/2 MIN PLUS
1 1 1 1
pH 5 80% Recovery
Zimmite 190
k = -0.0333 a =0.215
H @ 2 GPM
1 1 1 |
_j
-
-
-
- 20
1
*J
1.0
- 10 -P
- 3
- 2
u
i;
O
o
tr
oa
:
10
16
TIME, days
25
30
Figure 82. PRODUCT WATER FLUX, TEST 55
-------
30
20
•a
~ 10
s
I
X
-J
o "-
U) OC
UJ
O
Q
O
OC
a.
5
4
Pomona Primary Sewage
0.56 in. Tubular Membrane
700 psig pH 5 80% Recovery
4.4 mg/f Zimmire 190
J =2.6 k = -0,0212 a = 0.139
o
I- 30
- 20
*--
CO
I- 10
25
30
: -i
UJ
y
LL
LL
01
O
o
- 3 z
rr
00
TIME, days
Figure 83. PRODUCT WATER FLUX, TEST 82
-------
30
20
•
•a
r 10
*
»
o cc
o
i
1. 1
O
E
i
2.2 mg/f I 11 mg/t
±111
Pomona Primary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
2.2-11mg/f Zimmire 190
J =5.4
o
k = -0.0235 a = 0.358
10
15
TIME, days
25
I- 30
- 20
* •
,r.
•
111
O
•1
a
,i
30
Figure 84. PRODUCT WATER FLUX, TEST 47
-------
30 -=•
20
•
(I,
•o
10
:.
'•
-
«
DC
LJJ
;
f i
O
a
a
-
—
*"*">
V
mr
I I I I
^°N
till
L
\
\
I I I I
Pomona Pr
0..«5AIn. T
700 psig |
1 . 1 mg/f
J =8.4
o
till
imary Sewage
ubular Membrane
>H 6 80% Recovery
Zimmite 190
k = -0.0654 a = 0.1 38
1 1 1 1
1 1 1 1
20
10
— 5
4
- 3
- 2
I0
IB
TIME, days
20
25
i
'
:,:
i l
•
i
a
aa
:
m
5
Figure 85. PRODUCT WATER FLUX, TEST 92
-------
30 g=
a
0
~ 10
s
i
O U.
ON CC
111
O
3
n
O
QC
Q
Pomona Primary Sewage
0.56 in. Tubular Membrane
700 psig pH 6 80% Recovery
2.2 mg/f Zimmite 190
k = -0.0168
III
I I I
I I I
• •
UJ
O
UJ
O
O
111
z
<
rr
OQ
5
UJ
3
30
Figure 86. PRODUCT WATER FLUX, TEST 81
-------
30
20
„;
D
~ 10
s
i
•
-
>
[II
i
•1
3 3
a
D
i.
I II L
Pomona Primary Sewage
0.56 in. Tubular Membrane
700 psig pH 6 90% Recovery
2.2 mg/l Zimmite 190
J =4.7 k = 0.0022 a = 0.139
o
r- 30
- 20
i
4. ..
eg
10 ?
: :
LU
5 o
4^-
LU
O
f )
- 3
•i
oc
m
:
- 2
10
v,,
TIME, days
20
30
Figure 87. PRODUCT WATER FLUX, TEST 86
-------
I
:•
I
x
t- -I
o u-
oo a:
LU
<
y
PR
-
MM
y^
y \
\
/*-•<
y
v
''i'
Pomona Primary Sewage
0.56 in. Tubular Membrane
700 psig
11 mg/t
i
I
i i i i
^ V
V-<
1 1 1 1
pH 6 80% Recovery
Zimmire 190
k= 0.0123 a = 0.251
,/N
^/
i i i i
X
i i i i
—
-
- 30
- 20
- 10
tn
CJ
ro
1
a
•-'•
i
MEMBRANE COEFFICIEN
10
15
TIME, days
20
25
Figure 88. PRODUCT WATER FLUX, TEST 91
-------
30
20
•a
10
g
i
X
'
ll
III
1
I I
Primary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
10 mg/4 Calgon
J =3.2
o
k = -0.0382 a = 0.141
I I
10
15
TIME,days
20
25
r- 30
h- 20
j
!
- 10 i
:/
;
i
'
in
•
-I
OC
ffl
- 3
- 2
10
Figure 89. PRODUCT WATER FLUX, TEST 79
-------
10
20
I
10
X
')
o
O
a
a
-
1
•L
| I
•• X, ^
^•"•^
1 1 1 1
^-S.
1 1 1 1
Primary Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
10 mg/l Cyanamer
J =5.0 k = -0.0157 a = 0.079
o
1 1 1 1
1 1 1 1
1 1 1 1
?n
- 10
— 5
— 4
"™
-------
30 r=
20
(i,
•)
r 10
s.
-
j
i 5
DC
01
4
f '
o
D
Q
Pomona Primary Sewage
0.56 in. Tubular Membrane
TOO ptifl pH 6 80% Recovery
Biz flush
k = -0.0138 a = 0.313
- 30
- 20
- 10 -
u
a
i.
i
5
4
- 2
U
in
O
(L
ta
LU
S
Figure 91. PRODUCT WATER FLUX, TEST 89
-------
i >
Raw Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
J =5.6 k = -0.0623 a = 0.124
- 30
- 20
i
M. 10 •=
D
i
__ o
- 2
•
11!
' '
..I
II
UJ
c
1
i
DC
Figure 92. PRODUCT WATER FLUX, TEST 63
-------
30
20
r,
D
~ 10
s
a
-
>
OJ CC
LJJ
Q
r-
n
0
—
—
o
\ .
\ /
v/
I I I I
v
**s
Raw Sewage
0.25 in. Tubular Membrane
700 psig pH 5 80% Recovery
2.2 mg/t Zimmite 190
J =4.8 k = -0.0353 a = 0.123
o
>
1
- 20
- 10
- 5
- 4
- 3
- 2
0 5 10 15 20 25 30
•
i
.
HI
i '
UJ
'
;i
•I
a
OD
TIME, days
Figure 93. PRODUCT WATER FLUX, TEST 64
-------
Table 3
AVERAGE WASTEWATER CONSTITUENT REJECTIONS AND PRODUCT WATER QUALITY
FOR 68° FLAT-PLATE MEMBRANES AT 700 PSIG
EC*
Pomona Wastewater
Carbon-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Fountain Valley Wastewater
Alum-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Raw Sewage
Digester Sewage
Pomona Wastewater
Carbon-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Fountain Valley Waatewater
Alum-treated Secondary S ewage
Secondary Sewage
Primary Sewage
Raw Sewage
Digester Sewage
92.0
91.7
87.6
85.9
90.2
90.5
82.8
77.6
193
246
328
483
374
298.
668
2585
Total
COD
NH4-N
Organic
N
NO2-N
NO3-N
Total
P04-P
Rejections, %
83.6
93.2
92.7
78.9
91.0
94.2
88.2
97.4
87. 3
86.7
85.0
79.5
86.4
82. 1
73.6
82.9
92.0
86.8
82. 1
70.7
87.6
89.4
78. 3
98.2
Product Water
9. 16
5. 15
16.2
15.4
11.0
6.88
17.0
151
1.86
3. 13
5. 19
4. 47
5.67
9.70
11.4
116
0. 330
1.01
1.65
1.46
1.29
2. 03
3.55
10.8
67.4
44.7
82.8
89.6
57. 0
89.9
60. 1
99.4
47.0
61, 3
68.0
23.0
68.8
92.0
69.5
78. 1
98.3
99. 1
98.2
93.4
98.9
99.2
96.6
94.4
Quality, mg/1**
0.015
0.024
0.008
0.001
0.011
0. 003
0.011
0.001
2.28
1.68
0.520
1. 27
0.910
0. 047
0. 280
2.94
0.40
0. 18
0.27
0. 072
0. 19
0.097
0. 36
9. 36
MBAS
92.5
83.0
99.0
90.2
92.9
93.9
92.5
99.4
0.52
0.28
0.24
0. 12
0. 11
0. 072
0. 11
0. 37
Electrical Conductivity, ^mhos/cm at 25°C
#
Adjusted to product quality of total output of plant operating at 80% recovery
except Digester Sewage which is at 50% recovery.
-------
Table 4
AVERAGE WASTEWATER CONSTITUENT REJECTIONS AND PRODUCT WATER QUALITY
FOR 0. 25-IN. DIAMETER TUBES AT 700 PSIG
Pomona Wastewater
Carbon-treated Secondary Sewage
Alum-treated Secondary Sewage
Primary Sewage
Fountain Valley Wastewater
Alum-tr<-atod Secondary Scwago
Secondary Sewage
Primary Sewage
Raw Sewage
Pomona Wastcwator
Carbon-treated Secondary Sewage
Alum-treated Secondary Sewage
Primary Sewage
Fountain Valley Wastewater
Aliuii-tri-;iti'd Secondary Sewage
Secondary Sewage
Primary Sewage
Raw Sewage
EC*
88.9
89.7
90. 2
85. 6
92.4
89.3
89. 1
204
237
232
-15Z
245
245
523
Total
COD
NH,-N
4
Organic
N
NO,-N
2
NO7-N
3
Total
PO.-P
4
MBAS
Rejections, %
90.2
89.8
87.8
86. 1
91.5
94.4
60. 0
88.3
84.5
90.7
79.9
89.7
65.6
77. 1
92.3
93.0
84. 3
77. 0
91. 1
84.2
89.7
Product Water
8.15
4. 31
21. 3
12. 4
13.5
6.23
84.7
2. 33
1. 49
1.71
6. 08
3.06
6.07
5.59
0.250
0. 640
1.65
1. 51
0.660
1.55
2.00
93.7
40. 5
100
88. 1
76.6-
39.6
92. 6
Quality,
0.000
0.018
0. 000
0. 000
0.011
0.032
0. 004
67.4
34. 5
46.9
24. 4
54. 2
83.3
51.9
mg/1**
0.170
6. 14
0. 990
1.27
0.510
0.092
0.550
98.6
99.2
98.6
92.5
98.8
99.0
98.2
0.17
0.060
0. 094
0. Or, 1
0. 12
0.057
0. 12
94. 0
_
91.7
93. 6
94.8
-
94. 1
0.030
_
0. 19
0. 07^
0. 071
.
0. 080
'Electrical Conductivity, ^mhos/cm at 25°C
#>!<
Adjusted to product quality of total output of plant operating at 80% recovery
-------
Table 5
AVERAGE WASTEWATER CONSTITUENT REJECTIONS AND PRODUCT WATER QUALITY
FOR 0. 56-IN. DIAMETER TUBES AT 700 PSIG
Pomona Wastewater
ATum-treated Secondary Sewage
Alum-treated Primary Sewage
Primary Sewage
Pornona Waste-water
Alum-treated Secondary Sewage
Alum-treated Primary Sewage
Primary Sewage
EC*
Total
COD
Organic
NH4-N N
NO2-N
NO3-N
Total
P04-P
Rejections, %
93.3
93.2
95.3
96.
79.
92.
0
1
4
97.
87.
94.
7
4
7
100
57
62
Product
101
166
86.1
1.
18.
10.
66
3
5
0.
1.
1.
270
01
02
0
1
1
.7
.5
Water
.000
.67
.56
11. 1
49. 8
39. 8
Quality,
0. 036
0. 015
0.041
41.
36.
55.
2
5
1
98.
95.
99.
4
1
2
mg/1**
4.
2.
1.
69
54
24
0.
0.
0.
094
088
100
MB AS
>'<
'Electrical Conductivity,^mhos/cm at 25° C
''^Adjusted to product quality of total output of plant operating at 80% recovery
-------
Table 6
AVERAGE WASTEWATER CONSTITUENT REJECTIONS AND PRODUCT WATER QUALITY
FOR 44° FLAT-PLATE'MEMBRANES AT 200 PSIG
Pomona Wastewater
Carbon-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Fountain Valley Wastewater
Alum-treated Secondary Sewage
Pomona Wastcwalur
Carbon-treated Secondary Sewage
Secondary Sewage
Primary Sewage
Fountain Valley Wastewater
Alum-treated Secondary Sewage
EC*
79. 1
77.1
77.6
78. 3
Total
COD
77. 1
82.5
91. 0
79.9
NH4-N
48.5
85. 5
75.0
68.3
Organic
N NO2-N
Rejections, %
97.4 55.9
88.7 28.7
82.6 74.8
67.3 0.0
Product Water Quality,
371
372
455
4.66
7. 11
-22.5
3.78
1.99
5. 20
0.058 0.064
0.990 0.012
1.51 0.011
NO3-N
48. 8
34.3
53.6
18.5
mg/1**
1.57
4.98
0. 860
Total
P04-P
99.4
94.2
94.4
94. 1
0. 13
0.97
0.98
MB AS
92.9
93.6
87.0
0.071
0. 100
508
12.0
4. 74
0.790
0. 016
0. 092
0. 068
0. 14
"Electrical Conductivity, Mmhos/cm at 25° C
**Adjusted to product quality of total output of plant operating at 80% recovery
-------
of higher product water flux. The 0. 56-in. diameter tubular mem-
branes produced high rejections for most ions except the organic and
oxidized forms of nitrogen.
The concentrations of the various wastewater constituents in the pro-
duct water, listed in Tables 3, 4, 5, and 6, have been adjusted from
the observed values to correspond to concentrations that would be ex-
pected in a nonrecirculating reverse osmosis plant operating over the
complete range of product water recovery ratios from zero at the in-
fluent to the maximum (in this case, 80 percent) at the discharge. The
calculated modifications were based on the assumption that the individ-
ual constituent rejections are the same throughout the plant. The equa-
tions used to determine overall plant performance were
(Product Concentration) , = f. (Product Concentration), ,
, f (l-R)r - (1~R)
where ^ = R|u^
R _ I" Total Quantity Product Water]
~ I Total Quantity Feed Water J
, I" Average Product Concentration
I Average Wastewater Concentration! i__t
, I" Product Concentration "j
I Wastewater Concentration I, ,
118
-------
Section V
DISCUSSION OF RESULTS
In this program many tests were conducted under a wide variety of
operating conditions that are coupled by complex, and often unknown,
inter-relationships. To separate the effects of the various param-
eters within the constraints provided by the schedule of laboratory
testing, pertinent tests have been selected and the results or data
points smoothed for clarity and easier comparison. Actual data
points are presented in the preceding section.
A comparison is made in Figure 94 of the product water fluxes obtained
with the various types of sewage feeds under the standard control test
conditions. In all cases except one, higher quality feed water produced
higher product water fluxes. The long term, stabilized fluxes obtained
with primary and raw sewages appear to be identical at 2 gal/(sq ft)(day).
The major difference between raw and primary sewages is the absence
of gross settleable solids in the primary sewage, indicating that per-
haps these materials are not a critical factor in membrane fouling or
at least product water flux decline.
The magnitude of the stabilized product water flux appears to be quite
dependent on the feed water quality. Carbon-treated secondary sewage
produced 18 gal/(sq ft)(day) with new, clean equipment and relatively
high-quality feed water and 8. 5 gal/(sq ft)(day) with used equipment and
relatively low-quality feed water; whereas digester supernatant fouled
the membrane beyond hope of ascertaining the long-term flux. The rela-
tive absence of product water flux decline after an initial period of
several weeks without any special pretreatment of the sewages suggests
that treatment of the various sewages by reverse osmosis at reasonably
high levels of flux may be feasible without the use of additive chemicals
other than acid.
Although the sewages used in this program were from two different
sources and had quite different characteristics, no difference in re-
verse osmosis performance as a consequence was apparent. As an
example, secondary sewage from Pomona was, except for nitrate, to-
tal phosphate, and MBAS (all relatively low-concentration constituents),
markedly superior in quality to the winter-season secondary sewage
from Fountain Valley; yet typical results obtained with these two feeds,
(cf. Figures 20 and 21) are not noticeably dissimilar. For this reason
no distinction has been made between feed water sources in the analysis
of performance data.
TEST-CELL GEOMETRY
The greater part of the test data reveal little difference in product water
flux decline as a consequence of different test-cell or membrane geo-
metry. Essentially stable fluxes were observed with both the flat-plate
119
-------
30
25
20
15
10
_J f,
_i 6
I— u.
g * 5
***' III
u
g 3
O
oc
a.
I I '' I I I I I I I I I I I I I I I I I I I I I I I I I
Carbon Treated, Test 2
(4 Carbon Columns, Pott-regeneration)
Carbon Treated, Test 77
(3 Carbon Columns, Pre-regeneration)
Secondary, Test 48
Digester, Test 23, (50% recovery)
I I I I I ' '' '' ' ' ' ' '
Primary, Tests 21,22,66
Flat plate, 700 psig
pH 5, 80% recovery
I I I I I I I
10
15
TIME, days
20
25
30
Figure 94. EFFECTS OF FEED WATER TYPE ON PRODUCT WATER FLUX
-------
and tubular test cells and most feed waters. Where comparisons can
be made under nearly similar operating conditions (cf. Figures 44 and
86, 47 and 83, and 16 and 68, for example), the flat-plate test cells
provided only slightly greater product water fluxes and flux declines
in most but not all cases. The higher flux magnitudes exhibited by the
flat-plate membranes are in part or wholly due to their inherently
greater product water flux capability. These comparisons moreover
are made more difficult by the fact that the tubular units were operated
at a Reynolds number of 5, 000 and the flat-plate test cells were opera-
ted at a nominal Reynolds number of 3, 000. The lower value was se-
lected to correspond to the standardized test-cell conditions representa-
tive of field-scale desalination of sea water and brackish water with flat-
plate reverse osmosis units, whereas the larger number was chosen to
provide turbulence outside of the transition region. Operation of flat-
plate units at higher flow rates and concomitantly higher Reynolds num-
bers are not practical due to excessive pressure losses across the mem-
brane stacks.
One particular type of feed water, however, did produce results highly
favorable to the tubular membranes and is stressed herein because of
its potential in reverse osmosis processing of municipal wastewater.
Figure 95 presents the observed behavior of the flat-plate test cells and
both the 0. 25- and 0. 56-in. diameter tubular membranes with alum-
treated, sand-filtered secondary sewage, which was the only feed water
that provided a noticeable distinction between the test-cell geometry.
Not only are the differences in product water flux declines sizable, but
the sheet membranes which provide a greater flux with brackish water
produced a lesser product water flux than did the tubular membranes.
Thick deposits of calcium sulfate were found on the sheet membranes
while far fewer deposits of an unidentified material lined the tubular
membranes. These observed differences in performance between the
flat-plate and tubular test cells are believed to be manifestations only
of the greatly different flow conditions and turbulence obtained in the
two test apparatus.
OPERATING PRESSURE
Solids-bearing sewages—primary and secondary sewage--quickly bring
out the differences in membrane performance at high and low pressures.
Figure 96 compares product water fluxes between the 68° membrane at
700 and 200 psig, which demonstrated that the flux decline resulting
from 700 psig operation was in excess of that obtaining with 200 psig.
Tests conducted on both Zimmite 190-treated primary sewage and alum-
treated, sand-filtered secondary sewage again reveal that the greater
flux decline occurred at 700 psig, as shown in Figure 97.
The 44° membrane has no apparent characteristic different from the 68°
membrane that would account for greater flux stability with solids-bear-
ing sewages at low-pressure operation, other than the pressure itself.
One explanation for this observation involves the layer of materials that
121
-------
30
25
20
15
« 10
1 9
8
1
tv oc 5
„.
5 4
u
i 3
O
I ' I ' I I I I I I I I I I I I I I I I I I I I I I I
Tubular Membranes, Tests 35,36,37,75
Flat Plate Test Cells, Tests 30,31,32,33
AI urn-treated, filtered secondary sewage
700 psig, pH 5, 80% recovery
I I ' ' '
' 1 ' ' I ' ' ' ' ' ' '
10
15
TIME, days
20
25
30
Figure 95. COMPARISON BETWEEN TEST-CELL GEOMETRY AND
TYPICAL PRODUCT WATER FLUX DECLINE
-------
30
25
20
15
-------
30
25
20
15
10
9
8
I I I I I I I
1
3,
3 6
LL
ec 5
O
£
I 1 I I I I I I I I I I I I I I 1 I I I I I
— ——• —• Alum-treated Secondary
————— Zimmite-treated Primary
200 psig, 44° Membrane, Test 65
200 psig, 44° Membrane, Test 32
psig, 68 Membrane, Test 53
700 psig, 68° Membrane, Test 30
Flat-plate, pH 5, 80% recovery
I I I I I I I I » I I
I I 1 I I 1 L_J I I I I I I I I
10
15
TIME, days
20
25
30
Figure 97. EFFECTS OF PRESSURE AND MEMBRANE PERMEABILITY
ON PRODUCT WATER FLUX DECLINE
-------
deposit on the membrane after a few hours of operation. The water, to
reach the membrane surface, first must pass through the solids de-
posited on the membrane. A pressure gradient develops due to the
pressure loss experienced by the water in penetrating the deposits and
provides a force on the solids normal to the membrane surface counter-
acting the shear force of the bulk stream. At higher operating pres-
sures, the layer of deposits becomes more compacted and provides a
greater resistance to water flow and resultant higher pressure gradient
across the deposits.
ADDITIVES
The use of additives to prevent product water flux decline was based on
the premise that the solids deposited on or interacting -with the membrane
surface are a major factor in the decline and that certain additives could
be helpful in preventing that deposition or interaction. In comparing the
test results of various additives with carbon-treated sewage, shown in
Figure 98, it appears that additives are detrimental to the performance.
The control test had a slight decline in flux and greater total production
of about 10 gal/(sq ft)(day) than with any additive. The test with Cyana-
mer developed a flux decline from 13 to 4 gal/(sq ft) (day) in 15 days.
Close inspection reveals however that the tests with Zimmite 190 and
Calgon produced negligible flux declines, although lower stabilized pro-
duction levels of 6 and 7 gal/(sq ft)(day) were experienced. Since it is
recognized that overdosing a sewage with polyelectrolytes can produce
less than the desired result for any particular application; the dosages
of additives were perhaps too high for this particular feed water.
Higher fluxes were obtained in the control tests performed with carbon-
treated secondary sewage at the initiation of this study, which are be-
lieved due to both the uncontaminated nature of the apparatus and the ex-
ceptionally high quality of the feed water; but these results were disre-
garded in the foregoing analysis in favor of the results from the control
tests obtained on the apparatus after it had undergone similar service
to that for the additive tests.
In the treatment of secondary sewage with additive, the results of which
are summarized in Figure 99, it was found that Zimmite 190 was best
in preventing product water flux decline in the flat-plate test cells. The
first test with Zimmite 190, Test 26, gave a positive product water flux
decline slope at 80-percent recovery conditions; the usual initial drop in
flux from the maximum capability of the membrane occurred prior to
the 80-percent recovery level and thus prior to the value reported for
the first day of operation. Nevertheless, this test with Zimmite 190 ter-
minated at a higher flux than any of the other tests under similar condi-
tions. A subsequent retrial (Test 84) of Test 26 resulted in a decline in
product water flux from 14 to 9 gal/(sq ft)(day) in 15 days, terminating
at the same level as the original test. Perhaps an undetermined slug
of pollutant was responsible for the initial drop in Test 26 and the rest
of the trial was the slow recovery brought on with Zimmite 190. The re-
sults from tests conducted with both no additive and Cyanamer indicated
a lower stabilized flux of 6 gal/(sq ft)(day). Continuous addition of
125
-------
30
25
20
15
1
£ 10
sr 9
<
S 4
o
§ 3
O
oc
a.
I I I I I I I I I I I I I I I I I ! I I I I I I I I I I I
Control, Test 77
Calgon, Test 39
Zimmite 190, Test 40
Cyanamer, Test 46
Flat-plate, 700 psig, pH 5, 80% recovery
I I I I I i I I ' I ' ' I—I—I—I—I—L
I I t I I I I I I
10
15
TIME, days
20
25
30
Figure 98. EFFECTS OF ADDITIVES WITH CARBON-TREATED SECONDARY SEWAGE
-------
I I I I I I I I I I I
M
-*J
I I I I I I I I I I I I I I
_ Zimmite 190, Test 84
_ _ Zimmite 190, Test 26
Control, Test 48
Cyonomer, Test 13
xilgon, Test 12
Biz, Test 41
Alum, Test 30
Flat plat*, 700 psig, pH 5, 80% recovery
I I I I I I
i i i I I I i I I i I I I I I I I—LJ_J—L
20 25 3°
10
15
TIME, days
Figure 99. EFFECTS OF ADDITIVES WITH SECONDARY SEWAGE
-------
Calgon and Biz appear to be of no value in preventing flux decline and
possibly are detrimental as evidenced by production rates lower than
that obtained with no additive.
Comparison of the results of alum-treated, sand-filtered secondary sew-
age in flat-plate test cells, presented in Figure 100, reveals significant
initial flux declines in all tests from a high of approximately 15 gall -
(sq ft)(day) with ultimate stabilization of fluxes at less than 7 gal/(sq ft)-
(day) for employed additives. The combination of alum treatment and
dosage with Zimmite 190 provided a stable flux at 7 gal/(sq ft)(day) which
was slightly better than with no additive whatsoever and substantially
better than with Calgon, Cyanamer, or alum alone. It would appear that
Zimmite 190 has a definite ability to reduce membrane fouling caused by
calcium sulfate deposition, the presence of which was noted earlier in
the discussion of test-cell geometry. The consistently superior per-
formance of Zimmite 190 is again evidenced in Figure 101, which pre-
sents results of processing alum-treated secondary sewage in tubular
membranes. The product water fluxes in all tests were quite high
coupled with no product water flux decline after a small initial drop.
Zimmite 190 in Test 35 achieved a stabilized flux of 16 gal/(sq ft)(day)
which was slightly higher than observed with Cyanamer or no additive.
These test results indicate that an anionic flocculating agent without sub-
sequent solids removal is more effective in the maintenance of higher
fluxes than are a chelating agent or a dispersing agent in the concentra-
tions utilized. Also the results clearly indicate the benefits derived
from the use of a cationic flocculating agent (alum) with subsequent
sand filtration. Pretreatment with alum achieved stable fluxes superior
to those observed with the other additives alone. A test was performed
with no removal of the alum floe (Test 70) to ascertain whether the effec-
tiveness of alum treatment was associated simply with removal of solids
or was the result of a conditioning or modification of feed water charac-
teristics. A stable flux of 6 gal/(sq ft)(day) which was quite similar to
those achieved with and without other additives was obtained in this test,
indicating that the alum acts more to remove suspended and finely dis-
persed solids than to condition or modify the characteristics of the waste-
water process stream.
A primary function of maintaining a fixed additive dosage with the various
feed waters was to determine optimum additive dosage by varying the
quality of the feed water and not the quantity of additive. Under the stan-
dard test conditions, primary, raw, and digester sewages, with their
greater solids contents, were apparently too concentrated for the fixed
additive dosages used and no improvements in flux were observed in
either flat plates or tubes with continuous addition of Zimmite 190, Cyana-
mer, Zimmite 120, Calgon, or Biz. However, at a pH of 6 (standard test
condition was pH of 5), the standard dosage of 2. 2 mg/1 of Zimmite 190
in primary sewage appeared adequate by providing a product water flux
comparable to that obtained with alum-treated, sand-filtered primary sew-
age (cf. Figures 87 and 79). The dosages of additives used with secondary
sewage, i. e. , 10 mg/1 for Cyanamer and 2. 2 mg/1 for Zimmite 190, appar-
ently were adequate for that strength of sewage. The fixed dosage of Calgon
128
-------
tX)
re
•o
3.
x
D
tr
LLJ
30
25
20
15
10
9
8
7
6
5
I I I I I I I I I
I I I I I
I I I I I I I I I
i i r
5 4
u
O
O
cc
o_
Alum & Zimmife 190, Test 34
Control, Test 48
Alum & Calgon, Test 38
Alum & Cyanamer, Test 33
Alum, Test 30
700 psig, pH 5, 80% recovery
I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I
10
15
TIME, days
20
25
30
Figure 100. EFFECTS OF ALUM TREATMENT AND ADDITIVES WITH
SECONDARY SEWAGE IN FLAT-PLATE TEST CEILS
-------
30
25
20
15
I
£ 10
sr 9
^ 8
X
uo
C 5
S 4
u
S 3
O
I I
I II I I I I I I I I I I I I I I I I 1
0.25-in. tubular membrane
700 psig, pH 5, 80% recovery
Alum & Zimmlte 190, Test 35
Alum & Cyanamer, Test 36
Alum, Test 37
I I I I I »
10
15
TIME, days
20
25
30
Figure 101. EFFECTS OF ALUM TREATMENT AND ADDITIVES WITH
SECONDARY SEWAGE IN TUBULAR MEMBRANES
-------
producing 10 mg/1 concentrations was perhaps only effective with carbon-
treated sewage.
To substantiate earlier results with alum treatment, a 0. 56-in. diameter
tubular membrane was tested with alum-treated, sand-filtered primary
sewage. After 15 days operation, the flux decline became very small
at 5 gal/(sq ft)(day) and the flux stabilized and never dropped thereafter
below 4 gal/(sq ftj(day). After 60 days operation, when the recovery
ratio had been increased to 95 percent, the product water flux increased
to 5. 5 gal/(sq ft)(day) for a total of 77-days operation. This performance
was significantly better than the 2 to 3 gal/(sq ft)(day) flux obtained from
0. 25-in. diameter tubular membranes with primary sewage and various
additives other than alum.
Because of its generally superior performance with secondary sewage,
Zimmite 190 was investigated further to determine the effects of pH and
additive concentration on additive effectiveness. Primary sewage was
selected for these tests to provide a reasonably high solids-bearing feed
water. Figure 102 illustrates the results of testing one concentration
(2. 2 mg/1) of Zimmite 190 at various pH levels in flat-plate test cells.
Similar flux declines were observed for all pH conditions, but the high-
est flux was achieved at a pH of 6, indicating that this pH was optimum
for this additive and sewage combination.
The effects of different Zimmite 190 concentrations on product water flux
with primary sewage are shown in Figure 103, which indicates for this
wastewater and set of operating conditions that an optimum dosage of ad-
ditive was obtained. A wastewater concentration of 2. 2 mg/1 of Zimmite
190 produced much higher product water fluxes than did a lower concen-
tration of 1. 1 mg/1, and also greater fluxes than were observed at the
higher concentrations of 4. 4, 6. 6, and 8. 8 mg/1. It appears therefore
that Zimmite 190 concentrations of 4. 4 mg/1 and greater, although pro-
ducing greater product water fluxes than little or no additive, were ex-
cessive in dosage, were less effective than smaller dosages, and pro-
duced on occasion more erratic results.
DEPRESSURIZATION
Standard test conditions included depressurization of the recirculating
system at 8-hr intervals. This procedure resulted in small but noticeable
increases in the product water flux usually on the order of 1 to 2 gal/-
(sq ft)(day) above the daily average which then declined to below average
during the subsequent 8-hr period. All of the sudden increases in flux,
such as shown in Figure 51, commenced from a depressurization event
which therefore is an important catalyst in restoring product water flux.
RECOVERY RATIO
In Tests 76 and 86 (Figures 79 and 87, respectively) recovery ratios of 90
and 95 percent were achieved with primary sewage without degradation of
the relatively stable product water flux. The stable fluxes associated with
these recovery ratios indicate that within reasonable limits the recovery
131
-------
OJ
ts)
30
25
20
15
I
•o
£10
W 9
^ 8
u.
oc 5
UJ
Q 3
O
EC
0.
I I I I I I I I I I I I I I I I I I I I I I I I I I I I
pH 6, Test 74
pH 5, Teit 53
pH 8, Test 62
pH 7, Test 78
Flat plate, 700 pstg, 80% recovery, 2.2 mg/l ZImmlte 190
I I I I I I I III I I I I I I I I I I
I I I
I I 1 I
10
20
25
15
TIME.days
Figure 102. EFFECTS OF pH ON ZIMMITE 190 WITH PRIMARY SEWAGE
30
-------
X
30
25
20
15
10
9
8
7
6
U)
Oo
o
Q 3
O
oc
0.
I I I I I I I I I I I I I I
Flat-plot., 700 psig. pH 5, 80% recovery
I I I I I I I I I I I I
2.2mg/f , Test 53
6 mg/f
8.8
, Test 52
•ng/f , Test 60
mg/f / Test 55
ontrol, Test 66
1 I I » I I » I I
10
15
TIME, days
20
Figure 103. OPTIMIZATION OF ZIMMITE 190 DOSAGES FOR PRIMARY SEWAGE
-------
ratio has little effect on product water flux. The rise in product water
flux associated with operation at a 95-percent recovery ratio in Test 76
(Figure 79) and the lesser product water flux decline and flux variations
observed at 90-percent recovery in Test 86 (Figure 87) in contrast to
80-percent recovery in Test 81 (Figure 86) hint at beneficial effects
from unknown factors in the more highly concentrated feed water.
ADVANCED MEMBRANES
After the termination of the regular laboratory effort, newly developed
tubular membranes of cellulose acetate-cellulose triacetate blend became
available for testing. With an untrained skeleton crew and minimum
supervision of the apparatus, alum-treated, sand-filtered primary sewage
was again treated by reverse osmosis in the laboratory. The results of
these tests are presented in Figures 104, 105, and 106, which indicate
very high product water fluxes in the neighborhood of 15 to 18 gal/(sq ft)-
(day) with strong signs of incipient flux stabilization at those levels. The
advanced 0. 56-in. diameter blend membranes had initial product water
fluxes with brackish water of 30 gal/(sq ft)(day), at an operating pressure
of 700 psig, or 29^g/(sq cm)(sec)(atm). By comparison, the 0. 56-in.
diameter cellulose acetate membrane employed during the regular test
program, characterized by an initial flux with brackish water of 15 ga.ll-
(sq ft)(day), at 700 psig pressure, or 15/xg/(sq cm)(sec)(atm), produced
over the 77-day test duration (Test 76) an average product water flux of
about 5. 5 gal/(sq ft)(day) on alum-treated, sand-filtered primary sew-
age. Therefore it would seem that the level at which the product water
flux stabilizes is highly dependent upon the inherent permeability of the
membrane in addition to the feed water characteristics and other opera-
ting parameters.
THE FOULING MECHANISM
Of the many constituents contained in municipal wastewaters, the apparent
participating species in the membrane fouling process can be broadly
classified as gross settleable solids and particulate matter, finely dis-
persed solids, dissolved organic substances, and inorganic precipitates.
Careful analysis of all accummulated results from this investigation has
provided a qualitative assessment of the role that each of the aforemen-
tioned classes of foulants has in contributing to the membrane fouling
process.
The effects of gross settleable solids, which are those materials removed
by conventional primary sewage treatment, on the membrane fouling mech-
anism and the flux decline phenomenon appear negligible. Although there
is a large difference in settleable solids concentrations between primary
and raw sewages, tests conducted with both feed waters tended to stabilize
at the same product water flux levels (cf. Figures 38 and 58, 50 and 59).
Had gross settleable solids been contributary to membrane fouling, the
final stabilized fluxes of the two sewages should have been different.
Finely dispersed, nonsettleable material appears to contribute greatly to
the membrane fouling process by producing marked decreases in the level
134
-------
30
20
re
0
~ 10
X
OO LL
^ CC
Hi
i
Q
o
Q
Q
5
r\
in LI
i
4
i i i i
^
i i i i
i i i i
Orange County Primary Sewage
0.56 In. Tubular Membrane
700 psig pH 5 80% Recovery
Cellulose Acetate - Cellulose Triacetate
Alum Treated, Filtered
1 1 1 1
1 1 1 1
1 1 1 1
- 20
— 10
- 5
4
- 3
- 2
to
15
TIME, days
25
l
* •
I
30
LLJ
'
is:
!
: '
<
DC
i.]
Figure 104. PRODUCT WATER FLUX, TEST 97
-------
20
— m
9£T
PRODUCT WATER FLUX, gal/(sq ft)
-» M CO .& CJI
•*
L I L I
I
Pomona Primary Sewopt
0.56 In. Tubular Membrane
700 psig pH 5 80% Recovery
Cellulose Acetate - Cellulose Triacetate
Alum Treated, Filtered
20
*d
_ro
E
in
z
LU
(J
LL
O
CJ
LU
3 z
c
CD
01
2 2
i 5 10 15 20 25 30
TIME, days
Figure 105. PRODUCT WATER FLUX, TEST 98
-------
!>
a
:
<
i
1 ,i
1
PRODUCT WATER
30
2.0
10
2
_ ,
k A
IV-VY
v-^v
Pomona Pri
O..V> In. Ti
700 psig
Cellulose >
Alum Treat
mary Sewage
jbular Membrane
pH 5 80% Recovery
Acetate - Cellulose Triacetate
ed, Filtered
I
I
on
— E
-S-
~ 1
-,0 |
en
Z
LLJ
— R r,
NJ OJ *»
MEMBRANE COEFFK
0 5 10 15 20 25 30
TIME, days
Figure 106. PRODUCT WATER FLUX, TEST 99
-------
of stabilized product water fluxes. Tests conducted with alum-treated,
sand-filtered secondary sewages processed in the tubular membranes
produced at 700 psig relatively high product water fluxes of 15 gal/-
(sq ft)(day) with clarified wastewaters (cf. Figures 70 and 72). In the
test with the unclarified alum-treated secondary sewage, visual in-
spection indicated that the flocculated solids were reconverted to the
finely dispersed state by the rather severe agitation and turbulence
created by.the pump and other appurtenances in the wastewater recir-
culation system.
The contribution of dissolved organic substances to the overall reduction
in product water flux levels •would appear small in comparison to the non-
settleable solids. The most promising results obtained in this program
were with primary and secondary sewages that had been subjected to
alum treatment consisting of alum addition, flocculation, sedimentation,
and sand filtration, which effects little removal of dissolved organic
materials. At the beginning of the test program when the laboratory
apparatus had not been exposed to wastewaters of any other quality, tests
with carbon-treated secondary sewage, which contained practically no
suspended solids and relatively low concentrations of finely dispersed
solids and dissolved organics, produced product water fluxes equivalent
in terms of ultimate membrane capability to those obtained with alum-
treated, sand-filtered secondary sewage. Subsequent tests with
carbon-treated secondary sewage produced notably lower product water
fluxes, which is attributed partially to visible residual materials in the
system from earlier tests with lower grades of municipal wastewater
and to undetected differences in feed water quality. Although no signifi-
cant differences could be noted between the compositions of the carbon-
treated secondary sewages collected at the different times of the year,
the second set of reverse osmosis tests were conducted just prior to and
during regeneration of the carbon columns.
Under the proper hydraulic conditions, in addition perhaps to the avail-
ability of minute amounts of solids, the presence of ions in municipal
wastewaters that may precipitate on the membrane surface does not
appear to present a critical factor in membrane fouling and reduced
product water fluxes. Near the termination of Test 76 (Figure 79), which
employed alum-treated, sand-filtered primary sewage and a tubular mem-
brane, the system was purposely supersaturated with calcium sulfate and
operated for four days without the occurrence of measurable product
water flux decreases.
In general, an initial product water flux decline is observed when process-
ing municipal wastewater by reverse osmosis that, depending upon the
wastewater characteristics, extends over a period of from several days to
several weeks, after which time no further flux decline is evident and a
stabilized product water flux occurs. A phenomenological membrane foul-
ing model is postulated that accounts for these observations.
The effect of intrinsic compaction of the membrane during operation is
ignored in this model because the relative product water flux decline asso-
ciated therewith is negligible, particularly over the real time considered
138
-------
herein. Both the membrane characteristics and the operating pressure
are considered to be the same for all feed water conditions.
For purposes of developing the postulated model and for ease of pre-
sentation, the principal membrane fouling agents will be placed into
two categories — finely dispersed, nonsettleable solids and dissolved
organics that produce cohesiveness of the solids.
At the outset of processing a solids-bearing water by reverse osmosis,
the solids deposit on the extremely smooth membrane surface increasing
the surface roughness and concomitant local turbulence until an equili-
brium is established bet-ween the rate of solids deposition and the rate
of solids removal. Both the magnitude of the flux decline during the
initial deposition and the subsequent stabilized value of flux are depen-
dent upon the finely dispersed solids concentration, the cohesiveness of
the deposited solids and the local turbulence. The rate of initial flux
decline is greater and the stabilized flux lesser as both dispersed solids
concentration and cohesiveness or dissolved organics concentration in-
crease.
The degree of deposited solids cohesiveness is related to not only the con-
centration of dissolved organic substances but to the physical, other
chemical, and electrical properties of the finely dispersed solids, which
determine the agglomerative tendency and capacity to form larger and
more dense, less permeable deposits. Thus the presence of strongly
charged polyelectrolytes can significantly alter the properties of the dis-
persed solids by both counteracting the adhesive capacity of the organic
materials and reducing the ability of the solids to intensify on the mem-
brane surface by producing either highly repellant similarly charged
particles or floes that are easily swept along.
Figure 107 shows photographs of membranes after treatment of municipal
sewages for varying periods of time. Figures 107a and 107b demonstrate
the change in appearance of a membrane used for processing secondary
sewage after 6 and 18 days, respectively. The dark circles near the
center of the membrane are reinforcing patches that were placed under
the inlet and outlet ports of the test cells to prevent membrane damage
and have shifted position. Some of the deposited solids appearing in
Figure 107a have been displaced in Figure 107b. A thick layer of solids
was deposited on the membrane from raw sewage as shown in Figure
107c0 Figure 107d, depicting a cross-section of the membrane used in
the 77-day stabilized flux test run, shows that even though heavy deposits
are apparent on the membrane a steady, appreciable flux can be achieved.
139
-------
a. Secondary Sewage, Test 13
6-days operation
A
b. Secondary Sewage, Test 13
18-days operation
it. *«w bwwBge, T«st 29 d. Alum-tr«ated Primary Sewage, Test 76
10-days operation 77-days operation
Figure 107. PHOTOGRAPHS OF REVERSE OSMOSIS MEMBRANES
AFTER PROCESSING OF MUNICIPAL WASTEWATER
-------
Section VI
REVERSE OSMOSIS PROCESS MODEL
A mathematical model of the reverse osmosis process to determine
performance and costs from influent waste-water characteristics and
specified operating conditions and equipment features has been pre-
pared for use as a subroutine in the Federal Water Quality Adminis-
tration's "Digital Computer Program for Preliminary Design of
Waste-water Treatment Systems. "
The reverse osmosis system is based on a plant in the capacity range
of from 1 to 100 mgd containing tubular membrane units, placed in
parallel-flow configuration, that decrease in number downstream or
as the wastewater proceeds through the process. In addition to in-
fluent wastewater quality and quantity supplied by the executive pro
gram from the immediately upstream process, the operational param-
eters that must be specified are:
Overall plant product water recovery ratio, i. e. , the
ratio of product water flow rate to feed water flow rate.
Maximum total operating pressure of plant.
Membrane coefficient applicable to wastewater feed quality.
Excess plant capacity factor.
The preliminary plant design is based on the following conditions or
a s sumption s:
The wastewater velocities, hence Reynolds numbers, are
constant throughout the plant.
The osmotic pressure is determined from the total dissolved
solids concentration and not from the summation of individual
ionic species.
The total average product water flux of the plant is based on an
average effective pressure obtaining through the length of the
plant, i. e. , average effective pressure is the total operating
pressure less both the average osmotic pressure and the
average frictional head loss.
The permeation or rejection of individual species is constant
throughout the plant and is a function only of the membrane
character.
The adjustment of wastewater pH by acid addition is con-
sidered an external pretreatment process and is not
141
-------
included in the reverse osmosis model particularly since
it may not be necessary to the successful operation of the
reverse osmosis process.
Energy is recovered from the waste stream by means of a
turbine.
PROCESS MODEL, DEVELOPMENT
A basic flow sheet for the reverse osmosis process or plant is shown
in the accompanying sketch,
REVERSE OSMOSIS
PLANT
' C
R
Wp, V,p
where Q , Q_, and Q represent stream volumes or flow rates of the
feed water, tfie purified product water, and the reject wastewater,
respectively; and CF, Cp, and CR signify the corresponding concentra-
tions of a wastewater constituent or solute in the process streams.
As the feed water progresses through the plant, the solute concentration
on the wastewater side of the membrane increases continuously due to a
much higher transport rate of water through the membrane than of solute.
Both the water and solute fluxes in the membrane vary to some extent in
the plant due primarily to the presence of nonuniform flow conditions,
but for the purposes of this development they are assumed to be con-
stant throughout the plant for any given set of operating conditions and
plant configuration.
The solute flux through the membrane is expressed conveniently as a
permeation or rejection, defined as
p = 1-r = C /C
(1)
where p and r are solute permeation and solute rejection by the mem-
brane, respectively, and C is the average bulk solute concentration in
the wastewater throughout the entire process or plant.
The product water recovery ratio, R, which is the ratio of feed water
flow into a section of the process to the product water flow from that
142
-------
section, increases from zero at the inlet of the plant to a maximum
value which is the overall recovery ratio for the complete plant, de-
fined as
F (2)
From the materials balance on the water and the solute, respectively,
and QFCF = QpCp + QRCR (4)
and the substitution of the product water recovery ratio parameter, the
inter-relation between the solute concentrations in the process streams
in terms of the product water recovery ratio becomes
CR = (CF - RCp)/(l-R) (5)
The concentration of solute in the product water from a section, desig-
nated by the primed values, is
L-r)c'dR (6)
R
, i
where R is the section recovery ratio and C is the concentration of
solute on the wastewater side of the membrane.
Substitution of Equation 6 into Equation 5, applied^o a section with
recovery R , provides an integral equation in C , namely
t
C=-S-r CF-I (1-DCdR (7)
(l-R) L J0 J
I
Differentiation of Equation 7 with respect to R gives
dC*/ c' = rdRf/(l-R!) (8)
which upon integration between the limits of 0 and R1 and CF and
C1 provides the concentration of solute on the wastewater side of the
membrane at the end of the section with recovery R , namely
143
-------
I
-r
C - CF (1-R) (9)
The solute concentration of the discharge from the plant is
CR = CF(l-R)"r (10)
The overall plant average solute concentration on the wastewater side
of the membrane is obtained from the integration of Equation 9 between
the recovery ratio limits, or
C = RL CF (1'R) ^ (H)
The solutions to Equation 11 are
C = D/1 _ whenr * 1 (12)
R(l-r)
-CF In(l-R)
and C = - when r = 1 (13)
R
The solute concentration in the overall plant product water as a function
of feed water solute concentration is found by substitution of Equation
10 into Equation 5,
cp=ir
The production of purified water from the reverse osmosis process is
determined from the product of the membrane surface available and the
product water flux associated with the membrane and other operating
conditions. Therefore the plant size required to provide a specified
production capacity is determined by the particular membrane charac-
teristics.
Product water flux is a function of the intrinsic water transport proper-
ties of the membrane, the nature of the feed wastewater, and the applied
pressure on the wastewater side of the membrane. The intrinsic rate at
which water permeates through the membrane is described by the intrin-
sic membrane coefficient and is dependent upon the membrane formula-
tion, casting procedure, and annealing conditions. The membrane
144
-------
coefficient for a particular application is measured in a test facility
under the same hydraulic conditions and pressure that would be ex-
perienced in the plant and with the wastewater to be processed. Thus
the effects of both the feed water quality and the greater than bulk
solute concentration that obtains at the membrane surface are incor-
porated into the membrane coefficient and can be disregarded from
further consideration.
The product water flux can be expressed by combining the membrane
flux coefficient and the operating pressure, or
J = 1.45xlO"3 A P (15)
o
where J is expressed in gal/(sq ft)(day), A? is expressed in
Mg/(sq cm)(sec)(atm), and P is expressed in psig.
The effective pressure responsible for the transport of water through
the membrane is a function primarily of the total operating or plant
inlet pressure reduced by the osmotic pressure exerted by the solutes
in the wastewater and the frictional losses encountered in the reverse
osmosis tubes and fittings. For the overall process or plant, the
average effective pressure can be approximated by
=PF-L (16)
where P is the plant average effective pressure^ P^ is the maximum
operating pressure obtaining at the plant inlet, PQ is the average os-
motic pressure occurring throughout the plant, and PL is the average
pressure drop experienced across the plant due to frictional losses.
Osmotic pressure of a dilute solution is a function of the total solute
concentration, or simply
P0 = klRTCS = k2°S
where P.- is the osmotic pressure, kj^ and k2 are proportionality
constants? R is the universal gas constant, T is the absolute tem-
perature, and GS is the solute concentration. In dilute solutions kj
is very nearly unity, but in more concentrated solutions it becomes
quite dependent upon the type and only slightly dependent upon the con-
centration of solute, as shown in Table 7. Because the concentration
of individual solute species is usually unknown and variable with time
and source, the osmotic pressure of municipal wastewater can be de-
termined sufficiently accurately from the total dissolved solids con-
tent If the values presented in Table 7 for CaCl2 are taken as being
typical of municipal wastewater in its many states, the osmotic pres-
sure constant, k,, at 25° C becomes 0.00866 psig/ppm. By comparison,
145
-------
Table 7
OSMOTIC PRESSURES OF PURE SOLUTIONS
(psi at 25° C)
Concentration NaCl Na2SO4 CaCl2 MgSO4 MgCl2
ppm
500 5.65 2.67 4.33 1.90 5.36
10,000 113 53.4 86.8 38.0 107
50,000 565 267 433 190 536
sea water with a total dissolved solids concentration of 34, 500 ppm
exerts an osmotic pressure of 25. 1 atm at 25° C, which results in a con-
stant, k2, equal to 0.0107 psig/ppm. In this development, a value of
k2 = 0. 010 psig/(mg/l) of total dissolved solids will be used to include
the increased concentration occurring at the membrane surface due to
boundary layer conditions. Thus the osmotic pressures associated with
the feed water to and the reject wastewater from the reverse osmosis
process can be related to the total dissolved solids content of the feed
water by the expressions, respectively,
PQF = 0.010 TDSF (18)
and PQR = 0. 010(l-R)"r TDSF (19)
where PQF and POR are osmotic pressures in psig of feed and reject
wastewaters, respectively, and TDSp is total dissolved solids concen-
tration in feed water as mg/1. The average osmotic pressure on the
wastewater side of the reverse osmosis membrane can be calculated
from Equation 12 in terms of feed water total dissolved solids concentra-
tion, and becomes
P, ^•01°J1-i1-R'''r3JTDS,. for r i 1 (20)
o'\ ETT^T ( '""F
The total head or pressure loss through the plant is a function of total
series tube length, tube diameter and roughness, number and type of
flow constrictions and disturbances, and the velocity or quantity of
flow through the plant. Because water is removed from the wastewater
stream as it progresses through the process, and because it is necessary
to maintain the wastewater stream at a turbulent condition above some
146
-------
minimum, the cross-sectional area of the reverse osmosis plant
must decrease downstream if recirculation of reject wastewater is
not practiced. In the plants of large capacities considered for this
development, i. e. , 1 to 100 mgd, it is possible to provide, for all
practical design purposes, identical flow conditions in all flow chan-
nels or tubes. This feature of constant Reynolds number and constant
individual tube velocity simplifies the plant design to a single tapered
configuration which is believed most applicable for the intended use of
this reverse osmosis model.
Total plant head loss can be estimated from the relation
where PL is overall pressure drop in the plant due to frictional losses,
k3 is the constant of proportionality, Ls is the total tube length for
series flow, vt is the cross -sectional flow velocity in a tube, and d is
the tube diameter. For a reverse osmosis plant possessing a constant
Reynolds number throughout, the total length of tubes in series-flow
configuration is
k^a.v, k . IT d v. k . d v.
L = t =-1- _ * = 4 t (22)
_
Js 4Jjrd 4J
where k4 is_the unit conversion factor, a^ is the cross -sectional area
of the tube, J is the average product water flux, and s is the specific
membrane surface area. Replacing the velocity term in Equations 21
and 22 with the appropriate Reynolds number, NR = vtd/i/ , and substi-
tuting Equation 22 into Equation 21, the head loss is
PL = k3 V R ' <4J*)
where v is kinematic viscosity of the wastewater. The average fric-
tional pressure loss for this flow configuration is simply
PL = PL 12 (24)
Collecting terms into a single constant of proportionality, the expres-
sion for average head loss through the reverse osmosis plant as a func-
tion of design and operating parameters reduces to
) (25)
147
-------
where
(26)
and kf, is a factor incorporating frictional losses due to bends, valves,
contractions, and expansions in the flow channels and is assumed to
equal 2, i. e. , these losses are equal to the frictional losses resulting
from flow in straight tube lengths (45 ft in the Aerojet-General modules);
v is evaluated at 60°F and equal to 1. 2 x 1CT5 (sq ft)/ sec; p is the den-
sity of the wastewater and taken as 62. 4 lbm/(cu ft); f is the friction
factor which is assumed constant in this development for simplicity and
equal to 0. 031, a value representing a smooth pipe at a Reynolds num-
ber of 10,000; and gc is the gravitational constant equal to 32. 2 (lbm)-
(ft)/(lbf)(sq sec). Evaluation of k5 and insertion into Equation 25 pro-
duces
PT = 5. 82xlO'14N 3/(Jd3) (27)
J_i -K.
where PL is expressed in psig, J in gal/(sq ft)(day), and d in ft.
The commercially available reverse osmosis tube internal diameters
are limited to sizes ranging from about 0. 4 to 0. 6 in. To reduce the
opportunity for misuse of the computer program by the input of un-
realistic diameters and Reynolds numbers and to provide a diameter
compatible with the cost expressions incorporated, a value of 0. 56 in. ,
corresponding to the Aerojet-General internal tube diameter, and a
Reynolds number of 10,000 are substituted into Equation 27 resulting
in
PT = 573/J (28)
J_j
The relations between average product water flux and the known opera-
ting parameters can be summarized by appropriate substitution of
Equations 15, 20, and 28 into Equation 16, namely
1.45xlO"3Ao F
!0.010 [l-U-R/"^^^ 573
R(l-r) | F j
= p . __ (29)
The appropriate root from the solution of Equation 29 is
- _ B + (B2 - 4AC)0< 5 (30)
J —
2A
148
-------
where A = 1/(1. 45xlO~3 AQ) (31)
for r / 1 (32)
C = 573 (33)
The total surface area requirement for the reverse osmosis plant can
be determined according to the relation
S = Qpxl06/J=RQFxl06 /J (34)
where S is expressed in sq ft, Qp and QF in mgd, and J in
gal/(sq ft)(day). This expression can be converted to the total length
of all tubes contained in the plant by applying the specific membrane
surface area,
LT = S/s (35)
where LT is given in ft and s is equal to 0. 147 (sq ft)/ft for the
0. 56-in. diameter tube.
The power consumed to operate the reverse osmosis plant with a
pumping efficiency of 0. 8 is
KW = 0. 379 Qppp (36^
where KW is expressed in kw, Qp in mgd, and Pp in psig. The
power recoverable with a turbine of 0. 7 efficiency placed at the outlet
of the reverse osmosis plant can be calculated as follows,
JW(PF-PL) = 0.212(1-R)QF(PF-PL) (37)
Therefore the net power requirement is
0 379P^-0. 212(1-R)(P.,- PT ) (38)
F s! J-i I
PROCESS COST DEVELOPMENT
Major capital cost elements for a reverse osmosis plant consist of sup-
port structures for the membraned tubes, fittings and valves, high-
pressure pump, turbine generator for energy recovery, instrumenta-
tion power substation, and site improvements and housing. Membraned
tubes are not included in the construction cost since they are disposable
149
-------
after their useful life, which is assumed to be two years in this de-
velopment; hence tube cost is included as an annual operating ex-
pense.
Costs used in this model are based on the modular design of the Aero-
jet-General reverse osmosis system, which contains 11,520 lineal ft
equivalent to 1, 700 sq ft of membrane surface area. The modules are
self-supporting and contain all necessary connections between indi-
vidual tubes, which can easily be placed either in series or parallel
configuration, all valves and fittings, and product water collectors.
The estimated cost in dollars of the total number of required modules
without membraned tubes as a function of total plant membrane sur-
face area is
CMODU = 2. 65S (39)
Capital costs of items common to the modules are dependent upon
plant capacity and other operating parameters and are based on sup-
pliers quotations and estimates for equipment associated with plants
of several different sizes in the 1- to 100-mgd range. A scaling
exponent of 0.7 is employed on those parameters where economy of
scale can be realized.
The estimated capital costs in dollars for high-pressure pumps are
based on the use of a minimum of three separate pumps and can be
determined from
CPUMP = 224 (QF PF)°' 7 (40)
In these applications a turbine-generator for the recovery of flow
energy from the reject stream can be considered as simply a motor
pump running backwards at a lesser efficiency, and thus its capital
cost in dollars can be estimated from
r i0-7
CTURB = 224 I (1-R)QF (Pp - PL) (41)
Required instrumentation for a reverse osmosis plant would consist of
sensing and automatic data scanning and logging equipment for pressure
and flow rate readings and for conductivity measurements on the product
water from each module. The estimated capital cost in dollars of this
instrumentation is
CINST = 0. 2S 4 20, 000 Q °' ? (42)
150
-------
Because of the large power consumption by a reverse osmosis plant,
an electrical substation is provided, whose capital cost in dollars can
be estimated from
CSUBS = 850(KWN)°' ? (43)
The capital cost of land acquisition, site improvement, plant housing,
and ancillary services and facilities is related directly to plant capac-
ity for modular design and can be estimated in dollars from
CSITE = 0. 4S (44)
Upon collecting all the cost elements, the total capital cost in dollars of
a reverse osmosis plant becomes
CCOST = 3. 25S 4 224
4 850KWN°' ? (45)
Total annual recurring or operating costs, exclusive of capital amorti-
zation which is performed outside the reverse osmosis subroutine, in-
clude labor and labor overhead, general supplies and maintenance
materials, taxes and insurance, power, and plant remembraning. Es-
timated annual costs in dollars can be determined by the following ex-
pression
COSTO = 0. 055 CCOST -t- 61. 3 KWN + 1. 75S (46)
where the first term is comprised of labor and labor overhead, esti-
mated to be equal to 4 percent of capital costs; general supplies and
maintenance, equal to 0. 5 percent of capital costs; and taxes and in-
surance, representing 1 percent of total capital costs: the second term
is based on a rate of 0. 7 cents/kwhr: and the third term comprises the
costs for the biennial replacement of membraned tubes at a cost of
$3. 50/(sq ft) with a useful life of 2 yr.
SUBROUTINE RO
The reverse osmosis subroutine listing is presented in Table 8, A des-
cription of variables and their typical values are given in Table 9. Speci-
fic wastewater constituent rejections that can be expected are listed in
Table 10. The appropriate decision matrix and stream matrix are pre-
sented in Tables 11 and 12, respectively.
151
-------
Table 8
SUBROUTINE RO PROGRAM
SUBROUTINE RO
DOUBLE PRECISION COMPAR
DOUBLE PRECISION PRO
DOUBLE PRECISION PROCSS
DOUBLE PRECISION IPRO (50)
INTEGER OS1, OS2
COMMON /MATRIX/ SMATX(20, 50), DMATX(22, 20), ISi, IS2, OS1, OS2, N
COMMON /COSTS/ CCOST(20, 5), COSTO(20, 5), ACOST(20, 5), TCOST(20, 5)
COMMON /MISC/ PRO(50),COMPAR(20),FRPS(50),URPS(50),GPS(50),APS (50)
1,DEGC,CAER(50),CAER20(50),DO(50),DOSAT(50),AEFF20(50),URSS(50),
2XRSS(50),GSS(50),CEDR(50),VAER(50),VNIT(50),MLSS(50),MLASS(50),
3MLBSS(50),MLNBSS(50),MLDSS(50), AFS(50),FOOD(50),RTURN(50),
4MLISS(50),CNIT(50),CKWH(50),CFPGAL(50),CAIRP(50),BSIZE(50),TD(50),
5TDIG(50),C1DIG(50),C2DIG(50),VDIG(50),CH4CFD(50),CO2CFD(50),
6VFL(50),TVF(50),CFECL3(50),FECL3(50),WP(50), AVF(50), CCHEM(SO),
7TRR(50),GTH(50),GSTH(50),ATHM(50),ERR(50),WRE(50),GE(50),GES(50),
8AE(50),SBL(50),ASB(50),NN(10),TSMATX(20,50),ECF(50),
9BOD2(50), BOD5(50), CCI, AF, CTRP, CTGO, GLAND, TOTCC, TOTTC, TACOST, CCR,
XTCOSTO, CENG, ECF1, ECF2, ECF3, ECF4, AIRCFP
REAL MEMB,NETKW,REJ(20)
NAMELIST /LOCAL/ OFLUX, TOTSA, TOTHL, NETKW
OREC=DMATX(1,N)
TOTPR=DMATX( 2, N)
DO 5 K=3, 20
REJ(K)=DMATX(K, N)
CONTINUE
MEMB=DMATX(21, N)
ECF(N)= DMATX(22,N)
-------
Table 8 (continued)
SUBROUTINE RO PROGRAM
A=l. /(1.45E~3*MEMB)
B=TOTPR-(. 01*(1. -(1. -OREC)**(1. -REJ(15))))/(OREC*(1. -REJ(15)))*
1SMATX(15,IS1)
C = 573.
OFLUX=(B+SQRT(B*B-4. *A*C))/(2. *A)
SMATX(2, OS1)=OREC*SMATX(2, LSI)
SMATX(Z,OS2)=(1. -OREC)*SMATX(2,IS1)
DO 10 K=3,20
SMATX(K,OS1)=(1. -(1. -OREC)**(1. -REJ(K)))/OREC*SMATX(K,IS1)
SMATX(K,OS2)=(1. -OREC)**-REJ(K)*SMATX(K, IS1)
10 CONTINUE
TOTHL=1146./OFLUX
TOTSA=OREC*SMATX(2, ISl)*l. E6/OFLUX
NETKW=SMATX(2,IS1)*(. 379*TOTPR-. 212*(1. -OREC)*(TOTPR-TOTHL))
CCOST = 3. 25*TOTSA+224. *SMATX(2,IS1)**. 7*(TOTPR**. 7+((l. -OREC)*
l(TOTPR-TOTHL))**. 7-1-89. 3)+850. *NETKW**. 7
-------
Table 9
SUBROUTINE RO VARIABLES AND PARAMETERS
Symbol
CCOST
COSTO
ECF
ME MB
NETKW
OFLUX
OREC
REJ(K)
TOTHL
TOT PR
TOTS A
Typical
Value
1.1
10. to 25.
8 to . 95
300. to 800.
Description
Capital cost of plant, $
Operating cost, $/yr
Excess capacity factor
Membrane coefficient,
Mg/(sq cm)(sec)(atm)
Net pumping requirement, kwhr
Average overall plant product water
flux, gal/(sq ft)(day)
Overall plant product water recovery
ratio
Constituent rejections by membrane
Total friction energy loss through
plant, psig
Total operating pressure of plant,psig
Total membrane surface area of plant,
sq ft
'See Table 10 for specific rejections.
154
-------
Table 10
WASTEWATER CONSTITUENT REJECTIONS FOR SUBROUTINE RO
Description
Solid Organic Carbon
Solid Nonbiodegradable Carbon
Solid Organic Nitrogen
Solid Organic Phosphorus
Solid Fixed Matter
Solid BOD
Volatile Suspended Solids
Total Suspended Solids
Dissolved Organic Carbon
Dissolved Nonbiodegradable Carbon
Dissolved Nitrogen
Dissolved Phosphorus
Dissolved Fixed Matter
Alkalinity
Dissolved BOD
Symbol
REJ(3)
REJ(4)
REJ(5)
REJ(6)
REJ(7)
REJ(8)
REJ(9)
REJ(IO)
REJ(ll)
REJ(12)
REJ(13)
REJ(14)
REJ(15)
REJ(16)
REJ(17)
Typical
Value
1.00
1. 00
1. 00
1.00
1.00
1.00
1.00
1.00
0. 84
0.91
0.70
0.95
0.89
0.89
0. 88
155
-------
Table 11
SUBROUTINE RO DECISION MATRIX
DMATX(1,N)
DMATX(2,N)
DMATX(3,N)
DMATX(4, N)
DMATX(5,N)
DMATX(6,N)
DMATX(7,N)
DMATX(8, N)
DMATX(9,N)
DMATX(10,N)
DMATX(11,N)
DMATX(12,N)
DMATX(13,N)
DMATX(14,N)
DMATX(15,N)
DMATX(16,N)
DMATX(17,N)
DMATX(21,N)
DMATX{22,N)
OREC
TOT PR
REJ(3), Solid Organic Carbon
REJ(4), Solid Nonbiodegradable Carbon
REJ(5), Solid Organic Nitrogen
REJ(6), Solid Organic Phosphorus
REJ(7), Solid Fixed Matter
REJ(8), Solid BOD
REJ(9), Volatile Suspended Solids
REJ(IO), Total Suspended Solids
REJ(ll), Dissolved Organic Carbon
REJ(12), Dissolved Nonbiodegradable Carbon
REJ(13), Dissolved Nitrogen
REJ(14), Dissolved Phosphorus
REJ(15), Dissolved Fixed Matter
REJ(16), Alkalinity
REJ(17), Dissolved BOD
ME MB
ECF
156
-------
Table 12
SUBROUTINE RO STREAM MATRIX
SMATX(l.ISl)
SMATX(1,OS1)
SMATX(1,OS2)
SMATX(2, I)*
SMATX(3, I)
SMATX(4, I)
SMATX(5, I)
SMATX(6, I)
SMATX(7, I)
SMATX(8, I)
SMATX{9, I)
SMATX(10, I)
SMATX(11, I)
SMATX(12, I)
SMATX(13, I)
SMATX(14, I)
SMATX(15, I)
SMATX(16, I)
SMATX(17, I)
Feed Water Stream
Product Water Stream
Wastewater Stream
Volume Flow, mgd
Solid Organic Carbon, mg/1
Solid Nonbiodegradable Carbon, mg/1
Solid Organic Nitrogen, mg/1
Solid Organic Phosphorus, mg/1
Solid Fixed Matter, mg/1
Solid BOD, mg/1
Volatile Suspended Solids, mg/1
Total Suspended Solids, mg/1
Dissolved Organic Carbon, mg/1
Dissolved Nonbiodegradable Carbon, mg/1
Dissolved Nitrogen, mg/1
Dissolved Phosphorus, mg/1
Dissolved Fixed Matter, mg/1
Alkalinity, mg/1
Dissolved BOD, mg/1
^1 designates stream number, i. e. , IS1, OS1, or OS2.
157
-------
Section VH
ACKNOWLEDGMENTS
The fifteen-month program reported herein was performed by the En-
vironmental Systems Division, Aerojet-General Corporation at El
Monte, California, under the direction of Messrs. Gerald Stern and
Robert Smith, FWQA Project Officers. Aerojet-General personnel
participating in the program were Dr. D. L. Feuerstein> Program
Manager; Mr. T. A. Bursztynsky, Project Engineer; Dr. R. W.
Lawrence, Project Chemist; Mrs. I. P. Thomason, Analyst; Messrs.
B. J. McGrath, A. J. Patak, W. A. Barham, P. A. Tullius, and
R. E. Smith, Jr. , Laboratory Technicians; Mr. G. K. Haas and Mrs.
G. M. Hill, Programmers; and Mrs. M. D. Robinson, Secretary.
The complete cooperation and assistance of the County Sanitation
Districts of Orange County, the Orange County Water District, and
the County Sanitation Districts of Los Angeles County in providing
the municipal wastewaters used in this program is gratefully appre-
ciated and acknowledged.
158
-------
Section VIII
GLOSSARY
Text
Symbols
Computer
Symbols
B
C
C
R
CS
CCOST
CINST
CMODU
COSTO
CPUMP
CSITE
CSUBS
CTURB
d
A
ME MB
B
C
CCOST
COSTO
ECF
Definition
Standard deviation of membrane flux de-
cline coefficient
Cross-sectional area of individual reverse
osmosis tube, sq ft
Quadratic coefficient
Membrane water permeation coefficient,
Mg/(sq cm)(sec)(atm)
Quadratic coefficient
Quadratic coefficient
Average waste-water solute concentration
throughout reverse osmosis plant, mg/1
Feed water solute concentration, mg/1
Product water solute concentration, mg/1
Reject stream solute concentration, mg/1
Solute concentration, mg/1
Capital cost of reverse osmosis plant, $
Capital cost of process instrumentation, $
Capital cost of reverse osmosis process
modules, $
Annual operating and maintenance cost
of process, $/yr
Capital cost of process high-pressure
pumps, $
Capital cost of site improvements and
housing for process, $
Capital cost of electrical substation for
process, $
Capital cost for energy recovery turbines
for process, $
Diameter of individual membraned tube, ft
Excess capacity factor
Hydraulic friction factor
159
-------
T ext C ompute r
Symbols Symbols
J
J
•K i j • • • j
KW
KW
N
KWr
N.
R
P
P
_O
S
o
OF
OFLUX
NETKW
TOT PR
TOTHL
Definition
Factor relating laboratory to plant
product water quality
Gravitational constant,
(Ib )(£t)/(lbf)(sq sec)
m i
Product water flux through membrane,
gal/(sq ft)(day)
Average product water flux in process,
gal/(sq ft)(day)
Product water flux on first day at speci-
fied concentration condition, gal/(sq ft)-
(day)
Membrane flux decline coefficient, I/day
Constants of proportionality
Power consumption for process pump-
ing, kw
Net power requirement for process pump-
ing, kw
Power recovery from process turbines, kw
Total membraned tube length of series-
flow configuration in process, ft
Total overall membraned tube length in
process, ft
Reynolds number
Solute permeation through membrane
Net effective pressure causing transport
through the membrane, psig
Average net effective pressure in process,
psig
Operating pressure at process inlet, psig
Total pressure drop of wastewater stream
in process, psig
Average pressure drop in process, psig
Osmotic pressure of solution, psig
Average osmotic pressure of wastewater
stream in process, psig
Osmotic pressure of feedwater to pro-
cess, psig
160
-------
Text
Symbols
Computer
Symbols
Definition
OR
Q]
Q
R
r
R
t
T
TDS,
t
P
REJ
OREC
TOTSA
Osmotic pressure of reject wastewater
from process, psig
Feed water flow rate, mgd
Product water flow rate, mgd
Reject wastewater flow rate, mgd
Solute rejection by membrane
Overall product water recovery ratio of
process
Specific surface area of individual mem-
braned tube, (sqft)/ft
Total membrane surface area in process,
sq ft
Time from start of measurement at
specified concentration condition, days
Absolute temperature, °K
Turbidity of feed water to process, JTU
Total dissolved solids of feed water to
process, mg/1
Wastewater velocity in individual mem-
braned tube, ft/sec
Solution density, Ib /(cu ft)
Solution kinematic viscosity, (sq ft)/sec
161
-------
1
Accession Number
5
2
Subject Field &. Group
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Aerojet-General Corporation, El Monte, California
Title
Reverse Osmosis Renovation of Municipal Wastewater
1 f\ Authors)
Feuerstein, D. L.
Bursztynsky, T. A.
16
21
Project Designation
FWQA Program 17040 EFQ.
Contract#14-12-184
Note
22
Citation
23
Descriptors (Starred First)
Reverse osmosis, sewage treatment, process model, tertiary treatment, com-
puter model, membrane process, wastewater renovation, demineralization,
solids removal, organics removal.
25
Identifiers (Starred First)
27
Abstract A fifteen-month laboratory program has shown that all grades of municipal
wastewater may be significantly improved by the reverse osmosis process. Com-
parisons are provided on the behavior and response of the reverse osmosis process
to carbon-treated secondary sewage, alum-treated secondary sewage, secondary
sewage, primary settled sewage, raw sewage, and digester supernatant. High re-
movals of dissolved minerals, organic substances, and suspended matter have all been
achieved in the same treatment. The effects of a flocculant, dispersant, chelating
agent, enzyme, and acid on reducing product water flux decline are compared. The re-
lative effects of reverse osmosis test-cell geometry on solids deposition and membrane
performance are presented. A phenomenological model is postulated describing the
role of undissolved solids and organic substances in producing product water flux de-
cline and the subsequent maintenance of constant product water fluxes. A computer
model of the reverse osmosis process, compatible with the executive program written
by the Federal Water Quality Administration, has been developed to provide an accurate
and rapid method of determining the design and cost of reverse osmosis facilities.
Abstractor
D. L. Feuerstein
Institution
Aerojet-General Corporation
VR:102 IREV. JULY 1969)
NRSI C
SEND TO: WATER RESOURCES SCIENTIFIC INFORMATION CEN^
U S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240 j_
« U. S. GOVERNMENT PRINTING OFFICE : 1970 O - 410-161
------- |