WATER POLLUTION CONTROL RESKAiu'H SERIES
540M- FOR 03/72
REVERSE OSMOSIS DEMINERALIZATION
OF ACID MINE DRAINAGE
U.S. ENVIRONMENTAL PROTECTION A(,KN( Y
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes 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 water research program of the Environmental
Protection Agency, through in-house research and grants and
contracts with Federal, state, and local agencies, research
institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water}, Research Information Division, R&M, Environmental
Protection Agency, Washington, D. C. 20460
-------
REVERSE OSMOSIS DEMINERALIZATION
OF ACID MINE DRAINAGE
by
The Ecology Division
Rex Chairibelt Inc.
Milwaukee1; Wisconsin 53201
for the
Commonwealth of Pennsylvania
Department of Environmental Resources
Harrisburg, Pennsylvania 17102
and the
Office of Research and Monitoring
Environmental Protection Agency
Program No. 14010 FQR
March 1972
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.00
-------
iPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the -environmenta1
Protection Agency^ nor does mention of trade names or
commercial products constitute endorsement or recommend-
ation for use.
11
-------
ABSTRACT
The objective of this study was to determine the operational methods
and procedures necessary to successfully demineralize acid mine
drainage utilizing reverse osmosis (RO). The study was conducted in
two phases. Phase I consisted of laboratory bench scale investigations
to determine methods for controlling iron fouling and to select a
process flow sheet. Phase II was the field operation based on the
flow sheet selected in Phase I.
The field test site was located in Mocanaqua, Pennsylvania. The source
of acid mine drainage was the discharge from an abandoned underground
anthracite coal mine. Treatment prior to RO consisted of filtration
(lOy) followed by ultraviolet light disinfection. The brine from the
RO unit was treated by neutralization, oxidation and settling. The field
test phase spanned a four month period. Frequent samples were analyzed
to characterize the operation of the system.
The results obtained indicated that it was feasible to demineralize
acid mine drainage by reverse osmosis. Membrane fouling due to iron
was satisfactorily controlled. The recovery of product water was
limited to about 75% due to calcium sulfate fouling. Product water
was of potable quality in all respects except for iron, manganese,
and pH. Neutralization, oxidation and filtration would be required
to meet potable standards.
This study was performed under Contract No. CR-86-A with the Common-
wealth of Pennsylvania, Department of Environmental Resources.
Partial sponsorship was provided by EPA (Program No. 14010 FQR).
iii
-------
CONTENTS
Section Page
I CONCLUSIONS 1
II RECOMMENDATIONS 3
III INTRODUCTION 5
IV LITERATURE SEARCH 7
V LABORATORY INVESTIGATIONS 9
Iron Fouling Investigations 15
Investigation of Alternate Flow Schemes 20
VI DESCRIPTION OF FIELD DEMONSTRATION SYSTEM 35
VII DISCUSSION OF FIELD TEST RESULTS 43
Raw AMD Characteristics 43
Operation of the Pretreatment System 47
Operation of the Tubular RO System 50
Operation of the Hollow Fiber RO System 57
Operation of the Brine Treatment System 66
VIII GENERAL DISCUSSION 73
Discussion of Flushing Techniques 73
Economic Consideration for RO-AMD Operation 74
IX ACKNOWLEDGEMENTS 79
X REFERENCES 81
XI PUBLICATIONS 83
XII GLOSSARY OF TERMS 85
XIV APPENDICIES 87
I. Operating Data Hollow Fiber RO System 87
II. Tubular RO System Data 104
v
-------
FIGURES
1 TUBULAR RO MODULE CONFIGURATION
2 SPIRAL WOUND CONFIGURATION *~
3 HOLLOW FIBER MODULE
4 SCHEMATIC FLOW DIAGRAM FOR LABORATORY RO UNIT |£
5 IRON FOULING INVESTIGATIONS SIMULATED AMD 16
6 IRON FOULING INVESTIGATIONS (Type 300 Tubular Modules) 19
7 ALTERNATE AMD TREATMENT SCHEME - METHOD A, LIMITED
PRETREATMENT 21
8 ALTERNATE AMD TREATMENT SCHEME - METHOD B,
PRENEUTRALIZATION 27
9 OPERATIONAL DATA FOR PRENEUTRALIZATION (METHOD B) 28
10 ALTERNATE AMD TREATMENT SCHEME - METHOD C,
PRENEUTRALIZATION & SETTLING 32
11 SCHEMATIC DIAGRAM OF THE FIELD TEST APPARATUS 36
12 FIELD TEST APPARATUS 37
13 INITIAL MODULE ARRANGEMENT TUBULAR RO SYSTEM 38
14 MODULE ARRANGEMENT HOLLOW FIBER RO SYSTEM 40
15 HOLLOW FIBER MODULES 41
16 NEUTRALIZATION SYSTEM 42
17 RAW WATER QUALITY VARIATION 48
18 IRON OXIDATION RATES 49
19 MODULE ARRANGEMENTS UTILIZED FOR TUBULAR 310 MODULES 51
20 TUBULAR RO SYSTEM OPERATION WITH TYPE 31- MODULES 53
21 TUBULAR RO SYSTEM OPERATION WITH TYPE 610 MODULES
22 HOLLOW FIBER RO OPERATION PERMEATOR #691 58
23 HOLLOW FIBER RO SYSTEM OPERATION 59
24 COMPARISON OF CaS04 MOLAR SOLUBILITY PRODUCT AND PRODUCT
WATER FLOW 63
25 IRON OXIDATION STUDY 67
26 SETTLING RATE TESTS 68
27 FLOW SHEET USED FOR COST ESTIMATES 76
vi
-------
TABLES
No.
1 TYPICAL COMPOSITION OF SIMULATED ACID MINE WATER 15
2 SUMMARY OF THE OPERATING CONDITIONS FOR RO EXPERIMENTS
IN THE LABORATORY 17
3 DATA SUMMARY ON SIMULATED AMD 22
4 CHANGE IN BRINE CHARACTERISTICS WITH TIME 24
5 CHEMICAL ANALYSIS OF HOLLOW FIBER RO BRINE 25
6 ANALYSIS OF SOFTENED RO BRINE 25
7 TYPICAL WATER QUALITY DATA FOR THE IRON SLURRY
EXPERIMENT 29
8 MANGANESE REMOVAL WITHOUT pH ADJUSTMENT 30
9 OPERATIONAL DATA FOR SCHEME C 33
10 RAW ACID MINE DRAINAGE CHARACTERISTICS 44
11 COMPARISON OF FIELD AND LABORATORY ANALYSIS OF RAW
ACID MINE DRAINAGE 45
12 SALT REJECTION CHARACTERISTICS TUBULAR RO SYSTEM 56
13 SALT REJECTION CHARACTERISTICS HOLLOW RIBER ?C> SYSTEM 65
14 BENCH SCALE NEUTRALIZATION TESTS 70
15 SUMMARY OF NEUTRALIZATION SYSTEM OPERATION 71
16 SUMMARY OF ATOMIC ABSORPTION ANALYSIS OF NEUTRALIZATION
SYSTEM 72
17 LABORATORY RO MODULE CLEANING RESULTS 75
1-1 HOLLOW FIBER OPERATING DATA FOR SINGLE PERMEATOR 88
1-2 HOLLOW FIBER OPERATING DATA 2-1 ARRAY 90
1-3 OPERATIONAL DATA HOLLOW FIBER RO SYSTEM (2-1 ARRAY FIRST
STAGE MODULES) 93
1-4 HOLLOW FIBER SYSTEM OPERATION (2-1 ARRAY, 2ND STAGE
MODULE) 96
1-5 ANALYSIS DATA HOLLOW FIBER SYSTEM (AA DATA) 99
1-6 ANALYSIS DATA HOLLOW FIBER SYSTEM (AA DATA) 100
1-7 FIELD ANALYSIS DATA HOLLOW FIBER SYSTEM 101
1-8 FIELD ANALYSIS DATA HOLLOW FIBER SYSTEM 102
1-9 FIELD ANALYSIS DATA HOLLOW FIBER SYSTEM 103
II-l TUBULAR RO SYSTEM OPERATIONAL DATA 310 MODULES 105
II-2 TUBULAR RO OPERATIONAL DATA 610 MODULES 108
II-3 FIELD ANALYSIS DATA TUBULAR RO SYSTEM 109
II-4 LABORATORY ANALYSIS DATA TUBULAR RO UNIT 110
-------
SECTION I
CONCLUSIONS
The following conclusions were made based on the data obtained during
this study:
1. The feasibility of using Reverse Osmosis (RO) to provide potable
water from acid mine drainage was demonstrated.
2. The flux declines observed were tolerable (slope of log-log plot
of flux and operating time less than 0.031) and flux rates can be
sustained with a minimum of membrane flushing.
3. Oxidation of iron (II) by bacteria can be controlled by ultraviolet
light disinfection or lowering the pH of the feed water.
4. The acid mine drainage should not be neutralized prior to RO
processing.
5. Feed water pH is critical with regard to iron fouling of RO
membranes. Iron fouling can be controlled completely at a pH of
•^2.8 or below.
6. Allowable product water recovery is strongly influenced by the CaSO^
concentrations in the brine.
7. Calcium sulfate fouling of the RO membranes was found to occur above
a CaS04 molar solubility product of 25 to 35 x 10 , as measured in
the brine stream.
8. Rejection of individual ions across the RO membranes was in the
range of 99.2 to 99.7 percent based on average brine concentrations.
9. Product waters of 25 mg/1 total dissolved solids (TDS) were produced
from a feed TDS level of 1319 mg/1.
10. The product waters produced did not meet the USPHS recommendations
for iron, manganese, and pH, hence, would require limited additional
treatment.
11. A high flux decline rate was observed for the tubular system when
low salt rejection (98.5%) membranes were utilized, while the
higher salt rejection (99.6%) membranes had significantly improved
flux stability.
12. Iron oxide precipitation on the RO membranes was successfully
removed using a solution of sodium hydrosulfite (Na^S^O,).
-------
13. Calcium sulfate precipitation on the RO membranes was success-
fully removed using a solution of ammoniated citric acid at pH 8.
14. No damage to the RO membrane desalting properties was observed
due to 2670 hours of sustained operation on acid mine drainage
or the various flushing solutions utilized.
15. Neutralization of the brine to a pH of 7.9, followed by oxidation
and settling did not produce an effluent which could be reprocessed
by the RO system.
-------
SECTION II
RECOMMENDATIONS
1. Additional studies on RO brine treatment be made to determine
the necessary treatment which will allow the brine to be recycled
and thus eliminate this waste disposal problem.
2. Studies should be initiated to investigate recovery of iron and
aluminum for use as coagulants at sewage treatment plants.
Successful recovery could lead to reduced operating costs.
3. In acid mine drainage where the majority of the iron is in the form
of iron (II), a small amount of iron (III) generally exists. The
role of this iron (III) with regard to membrane fouling should be
further evaluated.
-------
SECTION III
INTRODUCTION
The pollutional effects of acid mine drainage (AMD) as well as various
methods of abating this pollution have been well documented as a result
of federally sponsored projects (1) (2). Among the various pollution
abatement techniques proposed, the use of reverse osmosis (RO) to
purify AMD appears promising and has been under study since 1966. Most
of the previous studies have been of short duration, that is less than
1000 hours continuous operation, and many have been less than 200 hours
operation. These short term tests were not sufficient to Identify the
various operating problems which could occur in treatment AMD utilizing
reverse osmosis. Furthermore, specific problems developed from certain
studies regarding iron fouling (3) and calcium sulfate fouling (4) which
required additional study to optimize the flow sheet and operating
procedures for an AMD/RO treatment system.
The objectives of this study were:
1. Determine the causes of iron fouling previously encountered (3)
and formulate methods of controlling and/or eliminating this
type of fouling.
2. Investigate various alternate flow schemes for treating AMD
utilizing RO.
3. Operate an AMD/RO treatment system for a sufficiently long
operational period to establish reliable operating characteristics,
To accomplish the stated objectives, the project was divided basically
into two phases, 1) a laboratory investigation phase to evaluate the
iron fouling problem and investigation of various possible alternate
flow schemes, 2) a field evaluation phase to operate the selected flow
scheme for a period of 2400 continuous operating hours, evaluating such
parameters as water recovery rates, rates of membrane fouling, permeate
water qualities, specific operating procedures required to minimize
membrane fouling, and membrane cleaning techniques.
The source of acid mine drainage (AMD) was the Mocanaqua discharge in
Mocanaqua, Pennsylvania. This is the same discharge utilized in the
previous study (3).
-------
SECTION IV
LITERATURE SEARCH
The feasibility of utilizing reverse osmosis (RO) to recover high
purity product water from acid mine drainage (AMD) waters and to abate
pollution has been under investigation since 1966 (5). Riedinger and
Schultz (5) found that high quality water could be produced from acid
mine drainage via reverse osmosis. The membrane system which was
utilized was a spiral wound system marketed by Gulf Environmental
Systems. Feed water pH was 3 or less and contained approximately 100
mg/1 of iron. Water recoveries in excess of 90% were reported, but
some iron fouling of the membrane was found to occur, decreasing the
product water output. Other investigations had also indicated problems
with iron fouling of RO membranes and it appeared that iron fouling
and subsequent membrane cleaning was the most critical problem
encountered in applying this process to the treatment of acid mine waters.
Hill (9), however, reporting on the work with the acid mine waters in
Norton, West Virignia, indicated that no problems with iron fouling were
experienced. Salt rejections as high as 99% were reported. The majority
of the iron at this site was in the trivalent state. Kremen et al. (4)
reporting from work at the same site, concluded that reverse osmosis
could process acid mine drainage feed streams to high degrees of recovery,
could produce excellent permeate water, and posed no special or difficult
problems for reverse osmosis processing. They further stated that
membrane lifetimes had been demonstrated which permitted confident cost
projections for immediate technology and for reasonably certain near
future state of the art. Sustained reverse osmosis operation up to
75% recovery levels were reported. Increased recovery levels up to
92% for short periods did not show the anticipated difficulties with
calcium sulfate precipitation. Although fouling of the membranes at
the discharge end of the plant did occur, these calcium sulfate scales
could be removed by operating the unit at 50% recovery for short periods
of time, thus flushing them from the membranes.
At the same time Mason (3), reporting on the work done at Shickshinny,
Pennsylvania (Mocanaqua discharge), concluded that although a high
quality product water could be produced via reverse osmosis, a number
of operational problems needed to be investigated before RO could be
applied to treat acid mine drainage on a large scale. This work was
conducted on acid feed waters containing the majority of iron in the
ferrous state. A tubular RO configuration manufactured by Calgon-Havens
Industries was utilized for this study. The main problem emerging from
this study was the maintenance of high water permeation rates due to
membrane fouling by iron. It was also indicated that to utilize the
permeates for potable use further treatment would be required when the
iron content in the feed water exceeded 100 mg/1.
-------
The sustained membrane performance tests had so far been limited to
water recovery rates below 80% because of the fear of fouling the
membranes with calcium sulfate. A study (8) carried out at three
different mine drainage sites by Gulf Environmental Systems under the
sponsorship of EPA, concluded that the limiting factor in achieving
the maximum water recoveries was the calcium sulfate concentration.
No iron fouling was reported during operation of any of the three
sites investigated. To further increase the water recovery rates, a
combination of neutralization and reverse osmosis called the
'Neutralosis' process was proposed by the EPA staff (9). The process
utilized the operation of the reverse osmosis unit at maximum recovery
O90%). The brine was then neutralized and settled and the overflow
from the settling tank returned to the RO unit for reprocessing. It
was concluded that the Neutralosis process produced 98% water recovery
when operated on a predominantly ferric iron acid mine drainge. However,
these results were based on relatively short term testing (less than
100 hours).
In view of the conflicting observations discussed above regarding iron
fouling of the cellulose acetate membranes and current developments in
the reverse osmosis membranes and hardware technology, it was apparent
that many technical areas required further investigation in order to
successfully apply reverse osmosis to the treatment of acid mind
drainge.
-------
SECTION V
LABORATORY INVESTIGATIONS
The objectives of the laboratory investigations were twofold:
1. To investigate the mechanisms involved in the fouling of RO
membranes bv iron.
2. Evaluation of alternate flow schemes to determine the most
desirable method of utilizing RO to treat acid mine drainge.
These investigations were conducted with pilot scale RO equipment using
both synthesized and actual acid mine waters. Three tvpes of commer-
cially available RO systems i.e. tubular, spiral wound and hollow fine
fiber were utilized during these investigations.
The tubular RO system was manufactured by Calgon-Havens. Both the older
module (type 300 - as utilized in a previous study (3)) and the modified
modules (type 310 and 510) were utilized. There are two main differences
between the older and the modified modules. First, the method of inter-
connection of the individual tubes within each module was different.
The tubes in the older module were connected by separate turn arounds
at both ends while in the new module they were connected internally by
means of an integral head (sealed by o-rings between the tubes and the
head). Second, the fiberglass tubes in the new module were significan-
tly stronger than the older tubes and were expected to have considerably
better life. The new tubes had been strengthened by utilizing new
manufacturing procedures. Each module consisted of 18 porous fiberglass
tubes with an effective membrane area of 16.9 sq ft. The new modules
were also equipped with turbulance promoters within each tube.
The turbulance promoter, called volume displacement rod (VDR), was a
helically wound rod and was placed inside the individual tubes to mini-
mize the concentration polarization effects by increasing the effective
brine velocity through the tubes. A diagram of the new tubular module
and the turbulence promoter is shown in Figure 1. Two types of tubular
cellulose acetate membranes were utilized during the laboratory inves-
tigation phase. The membranes with higher flux rates and lower salt
rejection were designated as type 300 or 310, while the membranes with
comparatively lower flux rates and higher salt rejection were designated
as type 500 or 510.
The two other types of RO equipment utilized in these investigations
were the spiral round system (obtained from Gulf Environmental Systems)
and the hollow fiber system (obtained from E.I. DuPont). Both these
configurations have the advantage of a high membrane area to volume
ratio compared to the tubular system. The spiral wound module was
rolled about a center tube much like a scroll and uses a mesh spacer
for the feed flow distribution. The membrane used x^as a newly developed
-------
Single Tube
Turbulence
Promoter Rod
End View
18 Porous Fiber Glass Tubes
in Series
Product Water Shroud
Feed Water
Module Assembly
Product Water
FIGURE 1
TUBULAR RO MODULE CONFIGURATION
-------
high flux-high rejection, cellulose acetate membrane (Module Type 4001).
The module in this system consists of one or more leaves wrapped around
a product water take-off tube. These leaves consist of a membrane,,
porous incompressible product water side backing material, and a brine
side flow spacer. The membrane is bonded along the two sides, at the
end, and around the product water tube, forming a sealed envelope that
encloses the backing material except at the product water tube open end.
The brine side flow spacer la placed on the membrane, and several lay e •..<=.
are then wrapped around the product water tube to form a cylJrjrical
module. Modules are then placed in a pressure vessel which consists of
a standard 4" schedule 40 steel pipe which has been coated for corrosion
resistance. The pressure vessel utilized in this study was approximately
10 feet long and held 3 modules. Each module was three feet long aud
contained 50 sq ft of membrane surface. The product water tubes for
each module are interconnected utilizing sleeves with "0" ring seals.
Figure 2 presents a sketch of both the spiral wrapping configuration as
well as the module arrangement within the pressure vessel.
The hollow fiber modules termed B-9 permeators utilized a newly developed
polyamide membrane. This membrane is characterized with several
advantageous feasures such as: significantly improved product water
rates, lower operating pressures and higher salt rejection capabilities
compared to the earlier membrane version 'B-5'. The B-9 module is 5.5
inches in diameter and 4 feet long. In each module, the individual
hollow fibers (42 micron (u) inside diameter by 84y outside diameter)
are bound into a cylindrical bundle containing a nominal fiber surface
area of 1900 sq ft. The open ends are potted in epoxy to separate the
purified water from the brine stream. This entire fiber assembly is
installed in a tubular pressure vessel. The pressure vessel is normally
made of aluminum. However, pressure vessels made of stainless steel or
fiberglass have also been introduced by the manufacturer for special
applications. Feed water under pressure circulates around the fibers.
Pure water passes through the walls of the fibers and flows up the bore.
The contaminants remain on the outside of the hollow fibers. The concen-
trate and permeate exit through separate outlets as shown in Figure 3.
A flow diagram of the pilot RO system utilized for the laboratory inves-
tigations is shown in Figure 4. The system consists of feeding the
wastewater through the modules under high pressure by a Moyno pump.
The pump speed is controlled by a variable drive and this controls the
pump flow rate. A half inch diameter stainless steel coil using a recir-
culation of cold tap water was incorporated in the feed tank to control
the feed water temperature in the range of 55 to 65°F.
In a typical experiment the pretreated wastewater was pumped to the
membrane bank from the feed water tank. Both the concentrate, and the
permeate were recirculated to the feed tank. Measurements were recorded
for IDS, temperature, pressure, pH and flow rates for the feed, concen-
trate and product streams. In order to simulate higher feed water recov-
ery, the concentrate was continuously recirculated to the feed tank
while the permeate was wasted until the desired recovery level was achieved.
11
-------
Purified
Water
Outlet
Membrane
Module
Concentrate Outle
Feed Side Spacer
Seal
Tubular
Pressure Vessel
Roll to
Assemble
Feed Flow
Permeate Out
Permeate Side Backing
Material with Membrane on
Each Side and Glued Around
Edges and to Center Tube
Permeate Flow
fter Passage Throu
Membrane)
FIGURE 2
SPIRAL WOUND CONFIGURATION
-------
FEED
Snap Ring
End Plate
Flow Screen
Open End
of Fibers
Epoxy
Tube Sheet
Porous
Back-up Disc
Snap Ring
CONCENTRATE
Porous Feed
Dis-rributor Tube
End Plate
Fiber
FIGURE 3
HOLLOW FIBER MODULE
-------
Permeate Stream
Concentrate Stream
High Pressure
Pump (Moyno)
sift-
Back Pressure
Regulator
Manual By Pass
High Pressure
Safety Relief
Pressure
Out
Pressure
In
Membrane Module
1. Tubular
Spiral Wound
B. Hollow Fiber
High-Low Pressure
Control
FIGURE 4
SCHEMATIC FLOW DIAGRAM FOR LABORATORY RO UNIT
-------
The composition of the simulated acid mine water utilized during this
study was similar to the one obtained for actual acid mine drainage
at Shickshinny, Pennsylvania in an earlier study (3). Table 1 presents
a typical composition of the simulated acid mine water. It should also
be noted that these waste waters were simulated using Milwaukee tap
water for all laboratory studies. A summary of the operating condi-
tions for the three types of RO equipment utilized is shown in Table 2.
TABLE 1
TYPICAL COMPOSITION OF SIMULATED ACID MINE WATER*
Ion mg /1
Calcium (Ca) 140
Magnesium (Mg) 100
Manganese 16
Iron (Fe"1"1") 120
Sulfate (SO/,) 800
pH, units 3.6
*Based on the field analysis performed at Shickshinny,
Pennsylvania during Fall, 1969.
Iron Fouling Investigations
The possible factors which could influence iron fouling of the RO mem-
branes as observed in the field (3) were postulated to be as follows:
1. Chemical oxidation of iron (II) to iron (III) due to oxygen
present in the feed water and subsequent precipiration of iron
(III) compounds on the RO membranes.
2. Biological oxidation of iron (II) to iron (III) by iron
oxidizing bacteria in the presence of oxygen and precipita-
tation of iron (II) compounds on the RO membranes.
3. The influence of concentration polarization on the rates of
oxidation both chemical and biological
A total :
-------
TABLE_2
SUMMARY OF THE OPERATING CONDITIONS FOR RO EXPERIMENTS IN THE LABORATORY
RO Configuration
Manufacturer
Membrane Type
Individual Module Size
Membrane area per Module
Modules Utilized
Feed Pressure
Feed Temperature
Water Output per Module
Minimum Brine Flow
Range of Feed Water pH
Dissolved Oxygen
Feed Water Recoveries
Tubular
Calgon-Havens
300, 310 and 510
3.5" dia. x 8"
16.0 sq ft
3 to 4
600 psi
55 - 65°F
0.1 - 0.15 gpm
0.75 gpm
3.6 - 7.0
7.8 - 8,5 mg/1
25 - 60%
Spiral Wound
Gulf Env.Sys.
4001
4if dia. x 3'
50 sq ft
3
600 psi
55 - 65°F
0.6 - 0.8 gpm
3.0 gpm
3.6 - 7.0
7.8 - 8.5 mg/1
35 - 45%
Hollow Fiber
DuPont
B-9
5.5" dia. x 4'
1900 sq ft
1
400 psi
55 - 65°F
1.2 - 1.4 gpm
0.5 gpm
6.7
7.8 - 8.5 mg/1
40 - 85%
-------
Brine |
Velocity |
1.9 fps )
CO
«x
c
o
14-J
w
OJ
iJ
D
,—I
P=H
16
14
12
10
Type 310 Tubular Module
I
Brine Velocity
2.9 fps
Brine Velocity
1.9 fps
Type 510 Tubular Module
Previous
Study (3)
100 200 300
Elapsed Time, Hours
400
FIGURE 5
IRON FOULING INVESTIGATIONS SIMULATED AMD
-------
reason for the initial drop in flux for membrane types 310 and 510 can
apparently be attributed to compaction of the membrane at 600 psi. These
declines could not have been due to iron fouling of the membranes as the
flux rates stabilized after the initial decline. This observation was
also confirmed by the consistent level of total iron and ferrous iron
concentrations in the feed water at various intervals during the course
of the experiment.
In an effort to study the effect of velocity on flux rates, the feed
flow was increased from 0.75 gpm (1.9 fps) to 1.15 gpm (2.9 fps) after
45 hours of operation at the former velocity. At 118 hours of operation,
the safety rupture disc failed and caused a brief shutdown of the system.
As expected, partial restoration of the flux rates was noticed for both
types of membranes (Figure 5) due to the depressurization effect.
However, the flux rates indicated that the increase in velocity (meaning
increased turbulence) had little influence on the decline of flux rates.
To verify this observation, the feed velocity was reduced to the initial
level of 1.9 fpm at 230 hours. No significant change in the flux rates
was noticed in an additional 100 hours of operation. Therefore, it
was concluded that there was no significant effect of velocity on the
flux rates within the range investigated.
Comparing the flux decline rates for membrane type 310 from experiment
//I above to the rates attained in a previous field investigation (3),
it was seen that the flux decline rates observed in the field were
significantly greater than in the laboratory (see Figure 5).
Although the composition of the simulated AMD was similar to the
Mocanaqua discharge and was also saturated with dissolved oxygen,other
factors that could affect the flux characteristics were:
1. New tubular modules with strengthened tubes containing turbu-
lence promoter rods.
2. Absence of an ambient atmosphere containing iron bacteria.
To investigate the effect of the above factors, two additional tests
were performed with old Havens modules type 300. Experiment #2 was a
duplication of experiment //I with the exception of utilizing the old
Havens without turbulence promoters. Experiment //3 was conducted with
a 4:1 combination of the simulated acid mine water and actual Mocanaqua
discharge to provide some iron bacteria in the laboratory test solution.
Figure 6 shows a comparison of the flux decline characteristics for the
experiments 2 and 3 under the present study as well as for the field
investigation during the fall of 1969 (3). Comparing the flux decline
curves for the membrane type 300 with or without the iron bacteria
(Experiments //2 and //3 respectively), the flux rate characteristics were
found to be quite similar. It could therefore be concluded that the
presence of the iron bacteria did not have any significant effect on the
flux rates. However, such a conclusion might not be justified in light
of the following factors:
18
-------
12.0
o
r-
* 10.0
CA
CL
C
O
8.0
6.0
4.0
Experiment "2 - Membrane Type 300
Experiment /-'3 - Membrane Type 300
Mocanaqua Investigation - Eall 1969
Membrane Type 300
80 120 160
Operational Hours
200
240
FIGURE 6
IRON FOULING INVESTIGATIONS (Type 300 Tubular Modules)
-------
1. Only 13 gallons of actual acid mine water was mixed with 50
gallons of simulated acid mine water. Such a dilution might
not have had a sufficient concentration of iron bacteria to
cause significant membrane fouling during the short test period.
2. It was found that some copper was being dissolved into the aicd
mine water from a bonze flow meter during the course of the
experiment. Within six hours, a copper ion concentration of
12 mg/1 was recorded in the feed water. Such a high level of
copper ions could have had a toxic effect on the. iron bacteria.
Although the cause of the Iron fouling (previously encountered (3)) could
not be pinpointed in the laboratory, it was concluded indirectly that
biological iron oxidation was the cause of the observed membrane fouling.
Such a conclusion was derived by the elimination of various other possible
factors that could have caused the membrane fouling. For example,
1. Chemical oxidation of the iron did not occur as illustrated by
the stable flux rates for experiments 1 and 2.
2. The new type 310 tubular modules did not have an effect on the
flux rates, since the flux characteristics for both the type
300 and type 310 were similar (Experiments 1 and 2).
3. There were no apparent concentration polarization effects noted.
with regard to chemical oxidation of the iron, as the brine
velocity was varied (Experiment 1).
Investigation of Alternate Flow Schemes
Various items relevant to AMD treatment were evaluated during this phase
of the laboratory studies. These included evaluating various types of RO
hardware, as well as methods of pretreating the AMI) prior to processing
via RO. Also investigated were RO brine treatment alternatives as well
as upgrading of the RO product water to meet USPHS potable water standards,
As a result of these investigations, three possible AMD treatment schemes
were considered. These schemes utilized similar supporting processes but
each placed the RO unit at a different point within the scheme This
placement had a strong influence on the operation of the RO unit. Each
of the three schemes was designed as a complete system to produce a
potable water and to provide for ultimate disposal of all residues.
Treatment Method (A) - Limited Raw Water Pretreatment
Figure 7 presents one of the treatment schemes. This scheme provided
filtration and bacteria control of the raw acid mine water prior to
treatment by reverse osmosis. However, these pretreatment steps were
not necessary for simulated acid mine water. The spiral wound and
tubular RO units were operated under this scheme. It was not possible
to operate the hollow fiber system on this feed water because the module
had an aluminum shell and the effects of the low pH on the B-9 membrane
20
-------
Raw
AMD
Profiltratlon
and
Bacteri.il
Control
Recycle
Ultimate Disposal
1. Chemical Recovery
2. Stream Disposal
3. Mine Disposal
Overflow
Reverse Osmosis
Brine
Neutralization
and
Oxidation
Settling Tank
or
Lagoon
Underflow
Product
Water
Neutralization
and Oxidation
of
Fe and Mn
Filtration
and
Chlorinatlon
Drinking Water
Ultimate Disposal
1. Strip Mines
2. Deep Mines
3. Dewater and Landfill
FIGURE 7
ALTERNATE AMI) TREATMENT SCHEME - METHOD A, LIMITED PRETREATMENT
-------
were not known at the time the testing was performed. (Note: later
field operation indicated the membrane could withstand pH as low as 2.6
without damage.) A summary of the data taken on the spiral wound and
tubular units is presented in Table 3. It was noted that the spiral
wound membrane provided both superior rejection as well as flux. The
greatest differences in rejections were in the iron and sulfate values.
The large difference in flux can be partially attributed to the fact that
the spiral wound unit had less than 10 hours of operation and the initial
compaction of the membrane had not yet occurred. According to Gulf
Environmental Systems, approximately 20% of the initial flux rate is
generally lost in the first 100 hours of operation. This would yield
a stabilized flux value of about 18-19 gallons per sq ft per day at
77 F compared to the tubular unit flux of 12 gallons per sq ft per day.
TABLE 3
DATA SUMMARY ON SIMULATED AMD
Iron
Calcium
Magnesium
Manganese
Sulfate
TDS
pH
Spiral Wound Unit
Feed Product
Quality Quality
. mg/1 mg/1
Tubular Unit
131
190
106
24
1250
1643
3.5
0.38
2.0
1.0
0.3
3.0
26
Feed
Quality
120
134
98.5
16.0
1200
1700
3.6
Product
Quality
. _ mg/1
1.9
2.2
1.9
0.3
35
60
3.4
NOTES:
Pressure
Feed Flow
Brine Flow
Membrane Area
Flux
Run
Recovery
Date Run
600 psig
5.06 gpm
3.17 gpm
150 sq ft
24.4 gsfd @ 77°F
15.3 hours
37.3%
16 Dec. '70
600 psig
1.15 gpm
0.69 gpm
69.4 sq ft
11.8 gsfd @ 77°F
325 hours
40%
September, 1970
22
-------
The product water stream from both these units contained irqn and
manganese in excess of the USPHS drinking water standards, and the
values recorded would be even higher if the product water recovery is
increased. (This test was at 40% recovery.) The product water will
therefore require post treatment to meet USPHS standards. Preliminary
post treatment investigations of the RO permeate from the tubular system
were made. The treatment included neutralization and oxidation
followed by sand filtration. > Two treatment methods were investigated.
In Method I, the RO permeate was passed through a granular limestone
bed for the neutralization and oxidation of iron (II), to iron (III) in
the presence of oxygen. The detention time in the limestone bed was
3.8-7.5 minutes. The precipitated ferric hydroxide was removed by
sand filtration. In Method II, the ferrous iron in the RO permeate was
oxidized by the addition of 1.5 mg chlorine per mg ferrous iron after
raising the pi! to 7.0 by the addition of 25 mg/1 lime. Oxidation time
was 5 to 10 minutes. The resulting hydroxide was filtered through a
sand bed. It was found that in both the schemes the sand filtered
effluent contained less than 0.05 mg/1 total iron. Additional data is
required to define the process variables. However, it was shown that
it is feasible to reduce the iron concentration well below the drinking
water standards by providing any of the post treatment schemes described
above. It will also probably be necessary to add a small amount of
potassium permanganate to insure oxidation and precipitation of the
manganese to the levels required bv the USPHS standards. It may also
be noted that conventional water treatment methods would be applicable
for treating the permeate, since the levels of iron and manganese are
similar to those found in many municipal water supplies.
The concentrate stream from RO is a potential pollution problem arid
further treatment is required. Typically this would involve neutrali-
zation and oxidation of the iron followed by solids/liquid separation
in a settling lagoon. This treatment would remove essentially all of
the iron and aluminum if operated at a pH higher than 7.5. Some CaSC>4
precipitation and removal can be expected, but the settling tank over-
flow would still be supersaturated with CaSO/. The theoretical
solubility of CaSO^ is generally assumed to be 2000 mg/1 but this can
vary considerably with ionic strength of the solution, temperature,
reaction kinetics and the concentrations of various other ions in the
solution. It has been reported (8) that CaSO^ precipitation within an
RO unit can be controlled if the CaSO^ concentration does not exceed
about 300-400% of the theoretical concentration. The concentrate
(brine), in this case, retains all the CaSO^ in solution until the brine
has passed out of the RO unit. Then if the brine is held for a period
of time calcium sulfate will precipitate. The time dependency of
precipitation was also verified in this study. Simulated AMD was
neutralized and settled to remove iron. This water was then passed
through the hollow fiber RO module operating at 85% recovery. The
resulting brine was supersaturated with CaSO^. The changes in brine
quality with time are shown in Table 4.
23
-------
TABLE 4
CHANCE IN BRINE CHARACTERISTICS WITH TIME
Time Total Hardness Calcium as Ca Sulfate
(days) (mg/1 as CaCO?) .JSS/1)
0 5800 1260 6000
1 4550 860 4500
4 3880 624 4200
It may be seen that after 4 days, the calcium concentration had been
reduced to 624 mg/1 which is extremely close to the theoretical level
of 590 mg/1 Ca which is equivalent to 2000 mg/1 CaSO^. These facts
indicate that after treatment of the brine and precipitation of iron,
aluminum, and CaSO^, it may be possible to recycle the brine back
through the RO unit, and hence eliminate a liquid waste stream. This
system of recycling treated brine was developed at the EPA-Mine
Drainages Field site and reported by Hill, et al. (9). The process
(called Neutralosis) was operated for only short periods of time, and
hence, additional data is required to fully evaluate this treatment
method. It should also be noted that the Neutralosis process will
not remove magnesium (Mg) or manganese (Mn) unless neutralization of
the brine is taken to pH 9.5 or above. Therefore, these compounds
would build up within the system until an equilibrium is reached
where the pounds of Mg and Mn leaving the system in the waste sludge
plus the product water would equal the pounds of Mg and Mn entering
the system in the feed water. While no problem with magnesium was
anticipated, since magnesium salts are quite soluble, the manganese
could cause problems resulting in excessive amounts of manganese in the
RO product water.
Because of the problem of various ion build-ups in the svstem when
brine recirculation is practiced, experiments in brine softening, i.e.
removal of Ca, Mg, and Mn were conducted. The brine produced with
the hollow fiber unit from preneutralized and settled AMD has a compo-
sition as shown in Table 5. This brine was then subjected to various
lime and soda ash dosages to determine how much chemical was needed
to achieve a given degree of softening. Lime was added first at
dosages of 0 to 2200 mg/1 and reacted on a Phipps-Bird stirrer for
forty minutes to provide contact between the sludge blanket and the
brine. After settling, the supernatant was analyzed for total hardness,
calcium hardness, and pH. In some cases sulfate analysis were also
performed. The results of these tests are presented in Table 6.
24
-------
TABLE 5
CHEMICAL ANALYSIS OF HOLLOW FIBER RO BRINE (1)
Constituent
Sulfate
Total Hardness (2)
Calcium Hardness (2)
Manganese
Total Iron
Total Dissolved Solids
(1) Raw AMD was preneutralized and settled
(2) As CaC03
Concentration- mg/1
6000
5800
3150
65
0.35
8085
TABLE 6
ANALYSIS OF SOFTENED RO BRINE
Lime
Dosages
(mg/1)
1500
1500
1500
0
Soda Ash
Dosage
(mg/1)
0
5000
3500
4000
10.5
Total Hardness*
(mg/1 as CaC03)
5450
450
1160
620
Calcium Hardness*
(mg/1 as CaCClj')
5450
80
1040
220
* As CaC03
It is apparent from Table 6 that 1500 mg/1 lime (as CaO) removed all
the magnesium hardness as well as all manganese but increased the
calcium hardness. It may also be noted that 4000 mg/1 soda ash was
almost as effective in softening as 1500 mg/1 lime and 5000 mg/1 soda
ash. From this it can be concluded that, it is not necessary to add
lime to remove the magnesium hardness, since MgC03 is being precipi-
tated. The addition of lime to precipitate magnesium hardness only
increases the calcium noncarbonate hardness, thus requiring a larger
soda ash dosage.
It is difficult to compare these laboratorv dosages with the theoret-
ical amount required due to the changing nature of the brine as
discussed previously. Obviously, calcium sulfate precipitated with
brine aging (Table 4), rendering it difficult to compare some of the
tests. No calcium sulfate was precipitated during the softening
25
-------
reactions, however, as the sulfate concentration did not change. The
insensitivity of the sulfate test (±10%) make accounting for all the
sulfate difficult. It is apparent from these tests that 4000-5000 mg/1
soda ash will remove up to 90% of the total hardness present in the brine.
It may not, however, be necessary to remove 90% of the Ca, Mg, and Mn to
successfully recycle the supernatant back to the RO system. There are
additional problems which must be evaluated regarding softening. These
include increased sludge volumes which complicates the ultimate disposal
problem, and methods of, controlling the system.
It was therefore concluded that traatment method A was feasible and the
question of recycling clarifier overflow requires additional field
operation data. It is also recommended that additional studies be
undertaken regarding brine softening.
Treatment Method B - Preneutralization (Figure 8)
In this scheme, the AMD was neutralized to a pH of 6.7 to 7.0 and
aerated to provide oxidation of the iron and manganese. Th,e objective
was to reduce the soluble iron and manganese in the pretreatment step,
to such a level that USPHS standards could be met after the reverse
osmosis treatment. The resultant slurry would contain all the iron and
manganese in an insoluble colloidal state which would be rejected 100%
by the RO unit. All 'other soluble ions (Ca, Mg, and 804) would be
rejected to the same extent as in untreated AMD. The product water from
the RO unit would meet USPHS standards and following chlorination could
be used as a potable water supply without post treatment as in the
previously discussed treatment method. The brine stream from this
system could then be routed to a lagoon for further concentration.
Disposal of the overflow and underflow from the lagoon would be identi-
cal to the previously discussed treatment method (Method A, Figure 7).
To evaluate this scheme the simulated acid mine feed water was
neutralized with 155 mg/1 of lime (as CaO) to a pE of 6.7 and was
aerated continuously to keep the ferric hydroxide slurry in suspension.
The ferrous iron content of this slurry was less than 0.05 mg/1 and
total iron content was about 125 mg/1. The iron slurry was then fed
to the tubular RO unit, since a slurry of this kind can only be treated
by a tubular RO system. Both the concentrate and the product water
were recirculated back to the feed tank and the flux and the water
quality data were monitored at various intervals. Figure 9 shows the
variation of the flux rates with the operational time. The flux rate
characteristics observed in this experiment were found to be very
favorable. The flux rate dropped from 10.6 to 9.6 gallons per sq ft of
membrane per day (gsfd) in the first six hours as would normally be
expected due to initial compaction at 600 psi, but significantly
higher flux rates were recorded at later time intervals. The flux
rate increased to a value of 10.8 gsfd at the end of 24 hours of opera-
tion and then stabilized at a value of 11.2 gsfd for a continuous test
duration of 235 hours. The reason for such flux characteristics could
26
-------
Raw
AMD
Neutralization
and
Oxidation of
Fe and Mn
Fe & Mn
Slurry
Reverse Osmosis
Product
Water
Possible
Recycle
Concentrate
Slurry
J'
Lagoon
Chlorination
Disposal
1. Chemical Recovery
2. Stream Disposal
3. Mine Disposal
Ultimate Disposal
1. Deep Mines
2. Strip Mines
3. Dewater & Landfill
Drinking
Water
FIGURE 8
ALTERNATE AMD TREATMENT SCHEME - METHOD B, PRENEUTRALIZATION
-------
Ni
CO
200
T3
01
O -^
03 60
% e 150
O CO
•O
•H -H
CO i-l
o OT 100
H
12
IM 1
03
CO -H
05
X CL
i-l C
fr< C
50 100 150
Elapsed Time (Hours)
200
250
FIGURE 9
OPERATIONAL DATA FOR PRENEUTRALIZATION (METHOD B)
-------
possibly be attributed to a scouring or brushing effect of the ferric
hydroxide precipitate on the cellulose acetate membranes. This scouring
apparently provided a continuous cleaning of the membranes and produced
higher sustainable flux rates. Moreover, all the ferrous iron had
already been converted to the ferric hydroxide precipitate and therefore
no fouling of membranes because of the in situ oxidation of the ferrous
iron was possible. During the experiment a slight increase in salt
passage through the membrane was observed. This increase is also shown
in Figure 9. It may be seen that the product water TDS increased stead-
ily from 68 hours to 163 hours and then appeared to stabilize. This
increase could indicate that some damage to the membrane desalting layer
had occurred, and this could have been the cause of the observed flux
increase.
Table 7 presents the typical feed and product water data obtained during
this test. Also shown in Table 7 is the calculated product water at a
91% feed water recovery. It may be seen that even at high recovery (91%)
the product water contained only 0.12 mg/1 of iron. However, it may be
pointed out that although the calculated total dissolved solids and iron
concentrations at high recoveries were estimated to be well below the
USPHS drinking water standards, a significantly high amount of manganese
permeated the membranes. Table 7 indicates manganese concentration of
2.3 mg/1 at 21% recovery and 4.8 mg/1 at 91% recovery. The manganese
in the feed waters was present in the manganous state; its oxidation to
the manganic state is extremely slow below a pH of 9.0. Therefore, even
when all the iron (II)had been converted to iron (III) at a pH of 6.7,
most of the manganese was in the soluble manganesus state only, and
hence, poorer removals were recorded for manganese ion by the RO membranes,
The allowable concentration of manganese in drinking water is only 0.05
mg/1 (USPHS standards). This means that even with a 99% rejection of
the manganese ion by the membranes, its concentration in the permeate
would exceed the drinking water standard limits for any feed waters
containing more than 5 mg/1 manganese.
TABLE 7
TYPICAL WATER QUALITY DATA FOR THE IRON SLURRY EXPERIMENT
Water Quality
Feed Water Recovery
Analysis Feed 21% 91%*
pH, units 6,7 6.2
Total Solids 2100 140 305
Total Hardness as CaC03 1200 85 197
Calcium 280 23 49
Magnesium 120 8.5 18
Manganes e 22.5 2.3 4.8
Sulfate 1300 99 213
Total iron 125 0.05 0.12
Ferrous iron <0.05 <0.05 <0.10
NOTES s All quantities expressed in mg/1 except where noted.
Membrane Type 300 tubular. Calculated values
29
-------
Additional laboratory tests were performed to investigate further
removal of manganese in the pretreatment step, prior to RO. A sample
of simulated AMD was neutralized to pH 7.4 with 170 mg/1 lime as CaO .
After aeration for 30 minutes and settling, the AMI") contained less
than 0.05 mg/1 total iron. Manganese was also reduced to 14 mg/1. It
was therefore demonstrated that although effective iron removal
could be obtained by this treatment method, the removal of manganese
was not very effective at neutral pH values. It was also determined
that the oxidation of manganese (II) was extremely slow without the
aid of an oxidizing agent other than molecular oxygen. The reduction
of manganese to the level of about 5 mg/1 is necessary to obtain a
product water from RO meeting the USPHS Standard of 0.05 mg/1 Mn.
This is assuming a 97-99% rejection of manganese by the RO membrane.
This level of rejection is possible with existing RO membranes. In an
effort to reduce the manganese level to the desired value of 5 mg/1
and also keep the pH below seven (the upper limit of cellulose acetate
membranes), a series of chlorine oxidation tests were performed.
Samples of previously neutralized and settled AMD were treated with
varying amounts of calcium hypochlorite. The samples were then mixed
for two hours, filtered and analyzed for residual manganese. The
results are shown in Table 8. It can be seen that a large dosage of
hypochlorite was needed to oxidize most of the manganese in two hours.
However, the high chlorine residual remaining makes this type treatment
undesirable.
TABLE 8
MANGANESE REMOVAL WITHOUT pll ADJUSTMENT
Hypochlorite
Dosage
(mg/1)
5
10
15
20
44
Residual
Manganese
10
9.5
7.6
6.5
1.0
Final
pH
Units
5.8
5.5
5.6
Additional tests using potassium permanganate demonstrated that a dosage
of 10 mg/1 followed by 1/2 hour aeration at pH 7.0 would reduce the
manganese to less than 1 mg/1. It was also found that neutralization
to a pH of 9.5 would also reduce the manganese to about 1 mg/1. The
permanganate method is preferred to the lime method, since pH 9.5 is
above the working range of the RO membranes and would require addition-
al pH adjustment, and addition of excess lime would present potential
scaling problems in the RO unit.
30
-------
Another limitation on the treatment of acid mine drainage bv this
process scheme may be the attainment of higher feed water recoveries.
It was found that when the iron slurry feed was concentrated fourfold
by recycling the brine and wasting the permeate, calcium sulfate precip-
itate was observed in the brine. It is indicated that the maximum
permissible feed recoveries without the calcium sulfate precipitation
might be limited in the range of 75-80% in actual full size plant
operation.
It was concluded that this scheme was not feasible based on the fact
that possible membrane damage occurred and potential CaSO^ scaling
problems were produced.
Treatment Method C - Preneutralization and Settling
The third and final flow scheme preoxidizes the iron and manganese as
discussed in Method B, but also provides for the removal of the
precipitated compounds in a settling basin. This scheme is presented
in Figure 10. The overflow from the settling basin is then processed
by the RO unit. As in the previous method, the RO product Xv?aters
would be suitable for potable use after chlorination. The concentrate
from the RO unit may then be recycled to a limited extent as discussed
for Method A, Table 9 presents a summary of the data taken on neutral-
ized and settled AMD. All three units were operated on this feed water.
It may be seen that the hollow fiber and spiral wound units gave
comparable product water quality. The tubular unit gave considerably
poorer water quality due to previous damage to these membranes while
operating on an industrial waste. The expected water quality from the
tubular unit would be very close to the tubular data shown in Table 3
if undamaged membranes had been utilized. If potassium permanganate
were utilized during neutralization the product waters from this treat-
ment method would meet USPHS standards.
Treatment of the brine would not be required as in Method "A", Figure 7,
since the brine would already be neutralized. Recycling of any liquid
streams back through the RO to reduce the volume of liquid for ultimate
disposal would require the same consideration as disucssed for Method A.
Method ''C', Figure 10, also presents a potential CaSO^ scaling problem,
if lime is used to neutralize. This has to be considered a disadvantage
for this method.
31
-------
Raw
AMD
LO
NJ
Underflow
Neutralization
and Oxidation
of Fe and Mn
Settling Tank
Overflow
Lagoon
Overflow
Granular
Filter
Disposal through Lagoon
for CaSO^ Removal
Concentrate
Reverse Osmosis
Excess Overflow
to Disposal
1. Stream
2. Deep Mines
Product
Water
Underflow
Ultimate Disposal
1. Strip Mines
2. Deep Mines
Chlorination
FIGURE 10
Drinking Water
ALTERNATE AMD TREATMENT SCHEME - METHOD C, PRENEUTRALIZATION & SETTLING
-------
TABLE 9
OPERATIONAL DATA FOR METHOD C
(FIGURE 10)
Spiral Wound Unit Tubular Unit
Feed Product Feed Product
Quality Quality Quality Quality
mg/1 me/1 mg/1 mg/1
Hollow Fiber Unit
Feed Product
Quality Quality
mg/1 mg/3
Iron
Calcium
Magnesium
Manganese
Sulfate
TDS
pH
0.2
272
106
14
1200
1691
6.7
0.05
4.0
1.0
0.2
8
29
—
0.2
280
116
14
1250
1709
6.7
0.05
20
9.2
1.4
88
153
—
0.2
280
106
13.6
1300
1701
6.7
0.02
4.4
2.0
0.2
20
25
—
NOTES:
Pressure 600 psig
Feed Flow 4.62 gpm
Brine Flow 2.76 gpm
Membrane Area 150 sq ft
Flux 22.8 (? 77°F
Run Length 3.0 hours
Recovery 40.4%
Date 16 Dec. '70
600 psig
1.08 gpm
0.71 gpm
42 sq ft
11.7 gsfd (3 77°F
4.0 hours
34.4%
17 Dec. '70
400 psig
2.65 gpm
1.55 gpm
1900 sq ft
1.2 gsfd (? 77°F
4.5 hours
41%
18 Dec. '70
Note: Simulated AMD was neutralized and settled
33
-------
SECTION VI
DESCRIPTION OF FIELD DEMONSTRATION SYSTEM
The laboratory investigation phase provided the necessary process
information which was the basis for the field evaluation system design.
The most significant conclusion from the laboratory work was that the
cause of the previously encountered iron fouling was a result of
bacterial oxidation of ferrous iron and subsequent precipitation on
the membrane surface. It was theorized that to eliminate this source
of fouling disinfection of the AMD was required. After considering
various process problems based on the laboratory work and the work of
others (4), a flow sheet was selected. A schematic diagram of the flow
sheet may be seen in Figure 11. A photograph of the system is shown
in Figure 12.
The feed water was pumped into the treatment system utilizing a PVC
centrifugal pump (Item 1, Figure 12). This pump had a capacity of
15 gpm at 30 psi discharge pressure. This pressure was sufficient to
route the feed water through the pretreatment system. This system
consisted of a pressure sand filter which utilized 18 inches of filter
sand with an effective size of 0.45 to 0.55 mm supported on 8 inches
of gravel plus 1/4 inch, minus 5/16 inch size. From the sand filter
the water flowed through two standard 9V' x 2^"cartridge filters in
parallel (Item 2, Figure 12). The function of the dual filtration
system was to remove any suspended material, including bacteria, which
might foul the RO membranes. The feed AMD after filtration was passed
through an ultraviolet light sterilizer (Item 3, Figure 12). The
sterilizer was a standard model 1000-S manufactured by Ultradynamics
Corp. This unit provided an excess of 30,000 micro-watt seconds per
sq cm of 2537 angstrom ultraviolet light and meets all U.S. Department
of Health requirements for UV light purification equipment. The unit
was equipped with two 15 watt UV bulbs. The volume of the radiation
chamber was 8.05 gallons. Maximum depth of the radiated liquid was
three inches. The unit was constructed of stainless steel. After the
pretreatment system the flow was pressurized utilizing a moyno screw
pump. This pump provided pressurized water to the two RO systems,
i.e. tubular and hollow fine fiber.
The tubular system was manufactured by Calgon-Havens and utilized
type 310 integral head modules. The basic system is the same one
utilized in a previous study (3). The module arrangement, however,
was modified for this study. The module arrangement utilized may be
seen in Figure 13. It is basically a 6 x 4 x 2 array. Each row
contained 5 modules in series. Each module has 16.9 sq ft of membrane
area. The last two modules in each row of bank 2 (Figure 13) and the
last three modules in each row of bank 3, contained volume displacement
(turbulence promoter) rods (VDR). The purpose of these rods was to
insure turbulent conditions even through the brine floi-j rate was being
reduced. Turbulent conditions are desirable to prevent concentration
35
-------
AMD
Source
Feed Sand
Pump Filter Duii
Cartridge
Filters
pH Monitor
Ultraviolet
Light
Disinfection
High Pressure
Pump
Pressure
Reduction
Valve
U)
ON
Overflow
Lime Feeder
0
Sludge
Underflow
Settling Tank
Aeration Tank
Alternate Points
of Discharge
FIGURE 11
SCHEMATIC DIAGRAM OF THE FIELD TEST APPARATUS
-------
a. Overall Arrangement
b. Pretreatment
1. Feed Pump
2. UV light
3.
A.
Cartridge Filters
Tubular modules
FIG_y_RE_12
FIELD TEST APPARATUS
37
-------
Bank 1
Pressurized
U)
00
Feed
Bank 2
-*»•
L
V
vh
EH
V
vh
V
v|-
V
VH
Bank 3
Waste
Brine
60 modules @ 16,9 sq ft membrane
3 banks„ 5 modules in series, each row
modules marked with a V have volume
displacement (tubulence promoter) rods
FIGURE 13
INITIAL MODULE ARRANGEMENT TUBULAR RO SYSTEM
-------
polarization induced fouling of the RO membranes. At times during
the study the module configuration was changed to study various
parameters. These changes generally consisted of altering the module
configuration as well as the location of the (VDR) modules within the
system. These changes and results of these changes are discussed in
Section VII. The tubular modules are shown in Figure 12s Item 4.
The hollow fiber system which was utilized was manufactured by E.I.
DuPont, The module arrangement for this system may be seen in Figure 14
and a photograph in Figure 15. The DuPont module contains approximately
1900 sq ft of membrane area in the form of hollow fine fibers. The
fibers have an 85p outside diameter and a 42p inside diameter. The
fiber is designated as B-9 and is an aromatic polyamide polymer. It
is an ansitropic membrane with a O.ly skin. The fibers are packaged
in a module called a permeator which measures 5^ inches outside diameter
and is 47 inches in length. The rated water capacity of these permeators
is 2000 gpd at 68°F and 400 psig pressure. The initial configuration
(Figure 14a) contained a single permeator. Later in the study two
additional permeators were added and utilized in a 2-1 array as shown
in Figure 14b. Details on the operation of the hollow fiber system
may be found in Section VII of this report.
The field demonstration system also included a brine treatment unit
(Figure 11). This consisted of neutralization and oxidation of the brine
followed by sedimentation. Photographs of the neutralization system
are shown in Figure 16. The aeration tank was 5*2 feet diameter round
polyethylene tank which was operated at a 20 inch water depth for a
total volume of 296 gallons. Lime was added to the brine flow utilizing
a dry lime screw type feeder. The feeder was controlled by a pH meter
to maintain the desired pH level. After aeration and conversion of the
ferrous iron to ferric iron, the slurry was routed to a settling tank
for solids liquid separation. The settling tank was a portable swimming
pool 10 feet in diameter and was operated at 24 inch water depth. Water
volume was 1172 gallons. An inlet baffle was provided (Figure 17,
Item 4) to dissipate the inlet velocity and prevent short circuiting.
The operating depth was held constant regardless of flow rate by a
float operated throttling valve (Figure 16, Item 5). Settled sludge
was removed manually utilizing a swimming pool vacuum cleaner type
device (Figure 16, Item 6). The neutralization system could be
operated on either tubular or hollow fiber RO brine. Details on
neutralization system operation can be found in Section VII ol this
report.
-------
Feed x~N
(J ^
Pressure
Reduction
Valve
B-9 Permeator
// 0691
— — — — - — -«»» Brine
Product Water _
a) Initial Permeator Arrangement
Brine
Feed
Pressure
Reduction
Valve
B-9 Permeator
// 1131
B-9 Permeator
// 1129
B-9 Permeator
// 0691
b) Final Permeator Arrangement
Each Permeator Rated @ 2000 gpd - 68°F - 400 psig
with 1900 sq ft Membrane
FIGURE 14
MODULE ARRANGEMENT HOLLOW FIBER RO SYSTEM
I
Product
Water
40
-------
a. Initial System
b. Modified System
FIGURE 15
HOLLOW FIBER MODULES
-------
a. Aeration Tank
b. Settling Tank
1. pH probe 2. Air Header 3^
4. Inlet Baffle 5. Effluent Float 6.'
FIGURE 16
NEUTRALIZATION SYSTEM
Neutralized brine inlet
Sludge Remover
42
-------
SECTION VII
DISCUSSION OF FIELD TEST RESULTS
The field operation covered many phases and utilized different RO
equipment and configurations. A single hour clock was used as a
reference for all field operations. This clock was not resetiileable.
At the start of this study the clock reading was 1140 hours. All
figures and tables appearing in this section refer to elapsed
operating hours which is the final clock reading minus the initial
clock reading. In order to compare individual tables or figures with
regard to absolute time a reference to the hour clock reading has
been included in each table or figure where necessary. This allows
comparison of tubular and hollow fiber RO system operation,, The 'tour
clock was connected directly to the high pressure purnp and registered
hours of pump operation. The field test phase started on April 30,
1971 and was completed on August 27, 1971. The total houts available
during the test period was 2856. The elapsed operating time recorded
was 2794 hours. The difference in these times -cas a res '.It of power
outages and system flushing.
Raw AMD Characteristics
The characteristics of the AMD utilized for this study are shown in
Table 10. Analysis were run in the field using a Hach water analysis
kit No. EL-DR. All analysis were run .-..tcording to the instruments
provided with this kit. However, all volumetric measurements for
dilutions etc. were made using glass volumetric pipets and graduated
cylinders instead of the less accurate plastic measuring devices
provided with the kit. Measurements for pH were made with a Beckmann
Model P-2 pH meter. Total dissolved solids (TDS) were measured with
a Myron-L hand-held (TDS) meter.
Samples were also shipped to Milwaukee for laboratory analysis. Metal
ion analysis was performed on a Perkin-Elrner Model 403 Atomic Absorption
Unit. Laboratory TDS measurements were made using Standard Methods
(10). Host samples were shipped to Milwaukee via airplane and the
analysis completed within 24 to 30 hours after sampling. Some samples,
however, were shipped parcel post. These samples were preserved with
nitric acid as recommended by Perkin-Elmer.
Table 10 presents both the field and laboratory analysis, the mean
values, and the 95% confidence range i.e. the range in which 95% of
the analysis would be expected to fall. These analysis compare favor-
ably with data taken from this site during a previous study (3) except
for the iron values which were generally lower. All analysis were not
run in the field, however Table 11 provides a direct comparison of
43
-------
TABLE 10
RAW ACID MINE DRAINAGE CHARACTERISTICS
Home Office Laboratory Analysis
Analyjiis _ __
Calcium rag /I
M ag r. a s i urn ;n g / 1
Manganese mg/1
Total Iron mg/1
Silica mg /I
Aluminum mg/1
IDS mg/1
No. of
Analysis
12
12
12
12
11
12
11
Mean
Value
110,9
82,6
14.0
70.1
10.8
8.3
1319
95% Confidence
Range
105 - 117
78 - 83
13.5 - 14=5
64 - 76
10.3 - 11.3
8.0 - 8.6
1234 - 1404
Analysis
Calcium mg/1
Total Hardness mg/1
Total Iron mg/1
Ferrous Iron rag/1
Sulfate mg/1
TDSJ mg/1
pH units
Field Analysis
No. of
Analyjsis
Mean
Value
34
40
40
37
29
23
114.6
610
68
63
1185
3.38
95% Confidence
Range
109 - 120
583 - 637
64 - 73
59 - 66
757 - 839
1151 - 1219
3.27 - 3.43
^ Ion analysis by Ator.dc
2 Utilizing Hach Field T
3 With a "Myron-L
st Kid #EL-DR
IDS meter
44
-------
TABLE 11
COMPARISON OF FIELD1 AND LABORATORY2 ANALYSIS
OF RAW ACID MINE DRAINAGE
Elapsed Time
Field
1230
1301
1395
1632
1898
2051
2314
2528
3051
3140
Clock
Laboratory
1234
1302
1380
1590
1899
2051
2314
2528
3074
3162
Total Iron
Calcium
mg/1
Field
78
74
71
54
55
56
57
70
76
83
Laboratory
71
74
75
54
57
59
69
74
84
80
Field
130
130
152
96
122
120
116
120
124
100
mg/1
Laboratory
115
115
115
97
97
102
107
112
125
123
Total
Hardness
as CaC03 mg/1
Field
648
728
600
480
465
550
560
650
700
690
0
Laboratory
835
884
851
707
710
746
780
808
927
895
Average
Range at 95%
Confidence Level
67.4
±7.6
69.7
±7.1
121
±11
111
±1
607
± 65
814
±55
1 Using Hach Kit //EL-DR
2 Using Atomic Absorption
3 Calculated from Ca, Mg, Mn, Al, and Fe analysis
-------
results from the laboratory and field. Comparisons were made for
total iron, calcium and total hardness. Paired comparison tests were
run (11) on each set of field and laboratory data. The results of
this test indicated there was no significant difference between the
field and laboratory analysis for iron and calcium at the 95% confidence
level. Stated another way the apparent differences between the field
and laboratory analysis for iron and calcium were not significant
statistically. The conclusion can be made, therefore, that the field
test procedures gave results for iron and calcium comparable to labora-
tory analysis.
The total hardness values for the field and laboratory however, were
not consistent. The total hardness values for the laboratory analysis
were calculated by summing the hardness producing ions (Fe, Al, Ca,
Mgs Mn) and expressing this sum as total (CaC03> hardness. The labor-
atory total hardness values were always significantly higher than the
field value. This indicated that the total hardness test in the field
was not accurately measuring all the hardness producing ions. Further-
more, Standard Methods (lO)states that the levels of certain ions
(Fe, Mn) present in these samples will cause a low total hardness
reading. Manganese and aluminum can also be analyzed with the Hack
kits however, they were not measured during this study. Based on
previous experience, reasonable accuracy is expected on acid mine
waters for these analysis. Since the total hardness value is used to
calculate the magnesium level it must be concluded that magnesium cannot
be accurately determined on acid mine waters using a Hach kit.
The dissolved oxygen levels in the AMD feed water were measured peri-
odically. These values are shown below.
Dissolved Oxygen Levels in the AMD Feed Water
Elapsed Time Dissolved Oxygen
Clock Hours in Feed AMD mg/1
1189 0.6
1470 0.6
1565 0.9
1664 0.9
Note: All measurements made with a YSI (Yellow Springs)
Dissolved Oxygen Meter
Since a relatively long suction line for the AMD feed water (120') was
utilized, the dissolved oxygen level was checked at the AMD source and
after the feed pump. These analysis indicated no air leaks were present
in the suction piping.
46
-------
Raw water quality variation is presented in Figure 17. A reduction
in all values is evident at about a 1600 elapsed time clock reading.
This was due to a large amount of rain which fell on the area during
this period. This caused an increase in the water table and had the
effect of diluting the feed water concentrations. This sharp drop in
contaminent level was followed by a gradual increase until about 2900
hours when additional heavy rains were experienced which caused another
drop in concentration levels.
Operation of the Pretreatment System
The pretreatment (prior to RO) consisted of filtration followed by
ultraviolet light disinfection. Initially a sand filter followed by
5 micron cartridge filtration was utilized. The function of the sand
filter was to protect the cartridge filters and hence provide longer
runs. It was soon discovered that the sand filter was not removing
any particulate matter and use of the sand filter was discontinued
after about 100 hours of operation. The 5 micron polypropylene fil-
ters were manufactured by Pall-Trinity (Filter // MCY 1001 YCH2) . They
contained 3.7 sq ft of filter area. These filters (two in parallel)
had a life of approximately 24 hours or 7,500 gallons per filter.
Because of this relatively short life and the high costs involved,
a switch was made to 10 micron filters. These were also manufactured
by Pall-Trinity (Filter # MCY100 1EE) and contained 3.7 sq ft of
filter surface. The 10 micron filters (two in parallel) had an average
life of 108 hours which was equivalent to 32,400 gallons per filter.
The great difference in filter life indicated a significant number of
particles smaller than 10 micron and larger than 5 micron were present
in the mine discharge. Since there was no apparent effect on the RO
unit operation, 10 micron filters were used during the remainder of
the study.
The purpose of the ultraviolet light disinfection unit was to kill the
iron bacteria present in the mine drainage and, hence, prevent
bacterial oxidation of iron II to iron III. Iron oxidation studies
were performed to enable evaluation of the effectiveness of the UV
light. Figure 18 presents the data collected on the iron oxidation
rates of four different samples. The four samples were stored in
polyethylene bottles. Once each day the cap was removed and then
replaced co allow oxygen equilization and then shaken vigorously to
insure an oxygen saturated sample. The raw AMD sample only filtered
through a 10 micron filter exhibited a rapid and immediate reduction
in iron (II) level. On the other hand a filtered sample in which the
pH was adjusted to 2.5 experienced little iron (II) oxidation for a
period of 100 hours and then a rapid reduction at approximately the
same rate as the sample which had no pH adjustment. The sample which
had been exposed to ultraviolet radiation had little decrease in iron
II level until 100 hours and then a gradual decrease until no iron (II)
remained at 350 hours. The sample disinfected with formaldehyde had
little decrease in iron (II) level when the experiment ended at 676
hours (iron II was 42 mg/1 at this time).
47
-------
-p-
CO
120
All Analysis by
Atomic Absorption
1200
1600
Total Iron
2000
2400
2800
3200
Aluminum
3600
Elapsed Time Clock (Hrs)
FIGURE 17
RAW WATER QUALITY VARIATION
-------
Filtered AMD pH 3.35
Filtered AMD pH adjusted to 2.5
Filtered AMD UV Light pH 3.35
Filtered AMD 2 ml Formaldehyde added
NOTE: Total Iron =
54 mg/1 for all
samples. UV light
radiation time =
53 seconds
50
100 150 200 250
Elapsed Time (hrs)
FIGURE 18
300
350
IRON OXIDATION RATES
-------
At the completion of this oxidation study it was obvious that iron
oxidizing bacteria were responsible for the rapid decrease in iron
(II) levels since the oxidation rate exceeded the natural chemical
oxidation rate by many orders of magnitude (12). It was also apparent
that the oxidation was inhibited by merely lowering the pH. This is
not surprising, since this can upset biological oxidation until the
organisms acclimate to the new pH level. Once this occurred the
oxidation proceeded at the same rate experienced in the raw sample at
pH 3.35. It may be noted (Figure 18) that the UV light arrested the
oxidation for approximately 100 hours and a slower rate of oxidation
persisted for the next 250 hours. It was apparent that the UV light
did not effect a 100% kill, however the possibility of sample contam-
ination during the frequent analysis does exist. The sample disin-
fected with formaldehyde was the most stable as demonstrated by the
small decrease in iron (II) level. In any event addition of acid, UV
light disinfection, or formaldehyde will provide protection against
iron oxidation and subsequent fouling of the RO membrane-
Some mechanical difficulties were experienced with the UV light system.
After 626 hours of use the bulbs (Model P-247) failed due to destruc-
tion of the end caps by the UV light. These were replaced with bulbs
containing a foil wrap around the end caps. This wrap was apparently
not effective since the bulbs again failed after only 740 hours of
usage. At this point, a different model bulb was installed (P246) .
These bulbs lasted through the end of the study, an additional 900
hours without failure. Additional operating data is required to deter-
mine the life of this new model bulb. The manufacturer, however,
guarantees the P-246 bulb for 7000 operating hours.
Operation of the Tubular RO System
The tubular RO system was operated in five separate phases during
this study. Generally the phases were marked by different module
configuratonns or new modules. Figure 19 presents the module arrange-
ments utilized for the first four phases of the tubular study. Phase I
utilized 60 Type 310 modules in a 6-4-2 array with 5 modules in series
in each row. The last two modules in bank 2 and the last 3 modules
in bank 3 (Figure 19a) contained volume displacement rods (VDR) which
increased the brine velocity. The normal inside diameter in the
tubular system is 1/2 inch which corresponds to a linear brine veloc-
ity of 1.64 fps per gpm of brine flow. The volume displacement rods
effectively increase this velocity to 2.5 fps per gpm of brine flow.
When utilizing VDR's the headloss through the module increases
significantly. In a module without the VDR's the headloss at 1 gpm
brine flow is 5 psi per module, while with VDR's this increases to
22 psi per module.
The Phase II study utilized 35 type 310 modules in a 3-2-2 array
(Figure 19b). VDR's were utilized only in the last two modules of
bank 3. Phase III and IV utilized 15 type 310 modules in series with
50
-------
1
—
L
(
vlv
V
V
V
a) Phase I
V
V
V
i- V|V V -,
I_Q v v v J
Module Arrangement
V
V
(b) Phase II Module Arrangement
V
V
(c) Phase III & IV Module Arrangement
NOTE: Each square represents one module, modules
marked V contain turbulence promoter rods.
FIGURE 19
MODULE ARRANGEMENTS UTILIZED FOR TUBULAR 310 MODULES
51
-------
VDR's in the last two modules (Figure 19c).
The entire flux history for Phases I to IV is presented in Figure 20.
During Phase I the product water flux decreased steadily from 13 gsfd
to 8.5 gsfd in only 480 hours. It was originally assumed that the
system had been contaminated with iron oxidizing bacteria because of
trouble experienced with the UV light. At 160 hours the system was
disinfected with a quatenary ammonium compound (L-ll-X). This accounts
for the slight increase in flux noted at 160 hours, since the membranes
were relaxed (operated at low pressure) and this generally results in
a flux increase for a short period of time. Immediately after the
disinfection, the flux continued to decline rapidly. At this point
it was felt that the brine velocities might possibly be too low and
that concentration polarization effects were the cause of the rapid
fouling. It was therefore decided to increase the brine velocities.
This could not be accomplished with the configuration utilized in
Phase I because of the higa head losses experienced across the VDR
modules. To reduce this head loss and increase the velocities the
module arrangement was changed for Phase II (Figure 19b) to a 3-2-2
array. The minimum brine velocity was increased from 1.2 - 1.4 fps
to 2.0 - 2.2 fps. Prior to starting Phase II operation, the modules
were flushed with ar, ammoniated citric acid solution (1.5 wt % citric
acid - buffered to pH 4 with ammonia) in an attempt to remove any
iron fouling which may have occurred during Phase I. The system was
then put into operation for Phase II. It may be seen in Figure 20
that the flux initially declined and then stabilized at a. value of
about 7 gsfd. The initial high flux readings were probably due to
membrane relaxation, since it was later found the ammoniated citric
acid was not effective in removing iron fouling from AMD fouled RO
membranes. The stabilizing of the flux values during the latter part
of Phase II could have been a result of the increasing brine veloc-
ities, but a definite conclusion cannot be made without additional
study. It will also be noted in Figure 20 that the water recovery was
lowered to about 50% in Phase II, and this may have had a stabilizing
effect on the flux rates.
Phase III operation was identical to Phase II except the module config-
uration was changed to 15 modules in series (see Figure 19c). This
change was made in order to reduce the high pressure pump requirements.
As may be seen (Figure 20) the flux remained stable throughout the
entire Phase III. The recovery was also increased to about 60%
for the entire Phase III. At the end of Phase III, the back pressure
valve clogged and the system ran for 10 to 12 hours with no brine flow.
This resulted in the end modules becoming completely clogged with
03804. It was therefore necessary to switch to 15 new modules and
this marked the beginning of Phase IV as shown in Figure 20.
The initially high flux experienced in Phase IV was probably a result
of relaxation (0 pressure) while the modules were not in use. The
flux immediately began to decline to approximately the same levels
experienced in Phase III.
52
-------
o o
3 o
"O a;
o pi
w
PH
100
75
50
c/o
on o
co c
vo
X
3 I
O
00
15
10
Q
ii
60 modules in a
6-4-2 Array.
Minimum brine velocity
1.2 - 1.4 fps
35 modules in a
3-2-2 Array.
Minimum brine velocity
2.0 - 2.2 fps
III & IV
15 modules in series
Minimum brine velocity
2.0 - 2.2 fps
NOTE: See Figure 19
for details on
Module Arrangement
400 800 1200
Elapsed Time (Hours)
1600
2000
FIGURE 20
TUBULAR RO SYSTEM OPERATION WITH TYPE 310 MODULES
-------
At the end of Phase IV it was noted that the flux declines experienced
with the tubular system were not experienced with the hollow fine
fiber system or a spiral wound system also operating at the same site
(13). It was also noted that both the spiral wound system and the
hollow fiber system had considerably higher salt rejection as compared
to the tubular 'system. It was felt that this fact may have had some
influence on the flux declines experienced. To test this theory, five
high flux-high salt rejection modules (Type E610) were installed and
put into operation. Figure 21 presents the flux history for these
modules. An extremely high initial compaction set was experienced
during the first 40 hours of operation. The flux then stabilized at
about 15-16 gsfd until 240 hours elapsed time. A gradual decline
then occurred through about 440 hours, at which time the flux
stabilized at 12.5 gsfd for the remainder of the study. The flux
decline experienced from 240-440 hrs was also experienced on the hollow
fiber unit and was believed caused by a higher than nomial iron (III)
content in the AMD. The decline was entirely due to iron fouling,
since operation at 40-45% recovery was well below the 03864 fouling
range. The modules were flushed with a sodium-hydro-sulfite solution
(4 wt %) for one hour. This resulted in a dramatic increase in flux
as shown in Figure 21. Since additional operating time was not avail-
able, it is not known how much of this flux increase was due to
cleaning and how much was due to membrane relaxation. It is felt,
however, that a substantial gain was accomplished, since membrane
relaxation alone would not account for an increase in flux of about
33%.
The salt rejection properties for both the types 310 and 610 tubular
modules are shown in Table 12. The salt rejection was 'calculated
based on average brine concentration experienced on the membranes,
i.e. the average of feed and brine concentrations. This procedure
allows comparison of salt rejection, while operating the RO system
at different product water recoveries.
Generally salt rejection for the type 310 modules was in the range
of 98.5-99% for Ca, Mn, Fe, Al, and SO^. Silica rejection was
extremely low at an average of 46 percent. The type 610 modules had
significantly higher salt rejection in the range of 99.5-99.6% for Ca,
Mg, Mn, Fe, Al, and 504. Silica removal was also considerably higher
than the 310 modules at an average of 93.9 percent rejection. No
apparent changes occurred in the salt rejection throughout the
operational period.
The mechanical operation of the tubular system was excellent. No
module failures were experienced over the entire 2800 hour operation.
This reflects the improvements made in tube construction since the
last study (3). The problem experienced with the plugged back
pressure valve at the end of Phase III was a result of the extremely
low total brine flows during this time, and the fact that the high
pressure pump was feeding two separate RO systems resulting in less
than positive brine flow control. This problem is not anticipated in
full scale systems.
54
-------
l-i 0)
0 >
3 O
•a u
o a)
10
a.
03 o
tO vO
O
CO
50
30
20
15
10
5 Modules in Series
Without Turbulence Rods
Minimum Brine Velocity
1.2 - 1.5 fps
200
400 600 800
Elapsed Time (Hours)
1000
FIGURE 21
TUBULAR RO SYSTEM OPERATION WITH TYPE 610 MODULES
-------
TABLE 12
SALT REJECTION CHARACTERISTICS
TUBULAR RO SYSTEM
LABORATORY ANALYSIS
Type 310
Raw Water
Ion
Calcium
Magnesium
Manganese
Iron (Total)
Aluminum
Silica
Total Dissolved
Solids
Quality
.. mg/1
111 ±
83 ±
14 ±
70 ±
8 ± 0
11 ±
1319
6
4
0.6
6
.3
0.5
± 85
Brine
Quality
m.§/1
287 ±
220 ±
39 ±
180 ±
21 ±
16 ±
3523
98
82-
11
71
6
2
± 1100
Modules
Product
Quality
mg/1
2.8
1.6
0.31
1.42
0.2
7.2
53 ±
± 0.8
± 0.4
± 0.06
± 0.3
± 0.08
± 0.5
11
Salt
Rejection
" %
98.55 ±
98.89 ±
98.79 ±
98.81 ±
98.6 ±
46.4 ±
97.66 ±
0.32
0.17
0.19
0.21
0.53
5.0
9.82
Type
Brine
Quality
mg/1
168
132
22
102
14
22
2074
610 Modules
Product
Salt
Quality Rejection
mg/1 %
0.58
0.39
0.08
0.31
0.1
1.0
17
99.57
99.64
99.55
99.55
99.61
93.9
99.13
FIELD ANALYSIS
Calcium
Total Hardness
as (CaC03)
Iron Total
Iron (II)
Sulfate
pH
118 ±
602 ±
67 ±
64 ±
774 ±
3.38
6
42
5
4
45
± 0.06
373 ±
1939
226 ±
186 ±
2056
2.96
74
± 378
60
37
± 326
± 0.1
4.5
22 ±
1.5
1.4
19 ±
4.1
± 1.1
6
± 0.2
± 0.2
4
±0,2
98.24 ±
98.37 ±
98.94 ±
98.85 ±
98.73 ±
—
0.31
0.37
0.15
0,09
0.27
200
1500
200
160
1450
3.0
0.80
6.0
0.65
0.62
2.0
4.0
99.47
99.45
99.54
99.49
99.80
Notes: All ranges shown at 95% confidence level.
Salt Passage (%) = 200 (Product Water Quality ) / (Feed Quality + Brine Quality)
See Appendix for detailed data.
-------
In summary a number of points can be made regarding tubular system
operation. The use of volume displacement rods is definitely not
recommended. The price paid in headloss far exceeds the benefits
obtained. If higher velocities are required, it appears recircula-
tion of brine would be the preferred alternative. With regard to
required velocities, it appears that a minimum velocity of about 1.5
fps is desirable, since operation at this velocity with the high flux-
low salt passage modules was satisfactory. It should be noted,
however, that the recoveries during this phase were quite low
(45 - 70%), and that operation at higher recoveries may require higher
velocities to offset the concentration polarization effects due to
higher brine concentrations. Additional study at higher recoveries
is necessary to answer this question. It is not known if the high
initial flux losses experienced with the type 310 modules was
specific for the modules utilized or a result of the lower salt
rejection. In any event high salt rejection modules(greater than
99%) are definitely recommended for both flux and product water
quality considerations.
Comparing the flux history in Figure 20 (310 modules) to the flux
history from the previous field testing (3) significantly lower flux
declines were noted in the present study. For example greater than
80% of the original flux was lost in 400 hours in the previous study
compared to about 45% of the original flux in the present study.
This would Indicate that tbe pretreatment system did have some effect
on the tubular system operation.
Operation of the Hollow Fiber RO System
The operation of the hollow fiber RO system was accomplished concur-
rently with the tubular system using identical feed water. The
initial permeator received at the site had an abnormally high salt
passage (greater than 10%). This was a result of an improperly applied
corrosion coating on the aluminum permeator shell, which resulted in
poor brine flow distribution. This permeator was immediately replaced
with a 316 stainless steel shell permeator and this corrected the salt
passage problem.
The initial 838 hours of operation of the hollow fiber system were made
using one permeator operating at 75% nominal product water recovery.
Two additional permeators were then added to the system to form a 2-1
array (see Figure 14). This 2-1 array was operated for an additional
1832 hours. The flux history for the three hollow fiber modules is
presented in Figures 22 and 23. Figure 22 is the flux history for
permeator No. 6r>l. This permeator operated initially as a single unit
at 75% recovery and then as the final stage in a 2-1 array. Figure 23
is the flux history for permeators 1129 and 1131 which were operated
as the first stage of the 2-1 array. The permeators were operated in
the 2-1 array to allow high recovery experiments and confine the
expected CaSO^ fouling to a single permeator. This also allotted
57
-------
00
e
O-
ec
o t^
rH O
fc 00
VD
M
0) 1
•u
(0 M
O O
3 O
2.25
2.0
1.75
800
200
400 600
Elapsed Time (hours)
Operated as a Single Pertneator
Water Recovery - 74.3 ± 0.28%
Brine Flow - 0.54 ± 0.01 gpm
Log-Log Slope - 0.0306 ± 0.008 gpm
1200
1600 2000
Elapsed Time (hours)
2400
Operated as Final Stage
Permeator in a 2-1 Array
Overall Water Recovery
Percent
I - 76.2 ± 0.6
II - 84.4 ± 0.7
III - 78.5 ± 0.5
IV - 75.3 ± 0.5
2800
FIGURE 22
HOLLOW FIBER RO OPERATION PERMEATOR #691
-------
VD
Operated as 1st Stage Permeators in a 2-1 Array
2.75
2.5
2.25
D.
CC
O 00
,-1 vo
CO CO
^ 0.
^J Q
o c
3 -*
•O
O
p-1 -
Permeator 1131
Hour Clock F Elapsed Time + 2102 hr
400 800 1200
Elapsed Time (Hours)
2.75
2.5
2.25
Permeator 1129
Hour Clock « Elapsed Time + 2102 hr
0
400 800 1200
Elapsed Time (Hours)
Individual Permeator Recovery- 60.3±1.4%
Brine Flow - 1.20 ± 0.026 gpm
Log-Log Slope - 0.0298 ± .005
1600
2000
Individual Permeator Recovery- 60.3±1.4%
Brine Flow - 1.20 ± 0.026 gpm
Log-Log Slope 0.0289± .005
1600
2000
FIGURE 23
HOLLOW FIBER RO SYSTEM OPERATION
-------
collection of long term flux data on the first stage permeators. Flow
control orifices were utilized in the first stage permeators to insure
equal flow of feed water to each permeator. A discussion of these
orifices may be found in the Appendix on hollow fiber operational data.
It should be noted that the hollow fiber system is not normally operated
as a stage system, i.e., generally all permeators are operated in
parallel. It should also be noted that the fluxes for the hollow
fiber system were not listed as flow per unit membrane area, but as
flow per permeator (module). This was done since the exact membrane
area per permeator was not known.
The initial 838 hours of operation of the hollow fiber system (Figure 22)
resulted in a relatively low flux decline. The log-log slope for this
period of operation was 0.0306 ± 0.008 at the 95% confidence level.
It should be noted that this slope was obtained by a regression analysis
of 75 individual data points, all of which are not plotted in Figure 22.
The correlation coefficient was 0.85 which indicates that all the flux
decline is not due to membrane compaction. Membrane compaction is the
loss of flux due to plastic flow (compaction) of the membrane. Pure
compaction will plot a straight line on log-log paper and the correla-
tion coefficient should be in the range of 0.95 to 0.97. Additional
regression analysis on portions of the flux curve of Figure 22 indicate
an increase in the log-log flux slope between 200 to 300 and 500 to 600
operating hours. This indicated fouling was occurring. An increase
in the pressure drop across the permeator was also noted which would
indicate fouling. This fouling was mainly due to iron precipitation,
since calcium sulfate (CaSO^) fouling does not generally cause an
increase in pressure drop (see later discussion on CaS04 fouling),
across the permeator, but rather an increase in salt passage. The
water recovery during this period was 74.3 ± 0.28% and the brine flow
0.54 ± 0.01 gpm at the 95% confidence level.
After operating 838 hours with a single permeator, two additional
permeators were added and operated in a 2-1 array. The flux history
of these two new permeators is presented in Figure 23. The log-log
flux slopes were 0.0298 ± 0.005 and 0.0289 ± 0.005 at the 95% confidence
level. The regression analysis performed on this data also indicated
fouling was occurring. Additional regression analysis on portions of
the data indicated significant flux curve slope changes from 550 to
800 and 1300 to 1500 elapsed hours. Flux decline slopes outside of
these time intervals closely approached the theoretical levels (high
regression coefficients), indicating the fouling was occurring only
for limited time periods and not continuously. The pressure drop
across the bundle also increased during these periods of flux decline
which would indicate iron fouling. Since the first stage permeators
operated at about 60% water recovery calcium sulfate fouling was
unlikely.
In order to evaluate the effect of the ultraviolet light on system
operation, the light was turned off at 140 hours and remained off
until 595 hours (Figure 23). This time interval represented 978 hours
-------
to 1433 hours elapsed time on the last stage permeator (Figure 22).
As may be seen in these figures, the fact that the UV light was off
had little apparent effect on the flux rates. During this period,
however, the pressure drop across the bundle did increase about 12 psi
(from 18-30 psi). This fact would indicate some iron fouling was
occurring. It is not known how much of the pressure drop was associ-
ated with the feed flow distributor. It is also interesting to note
that during the time the UV light was off the second stage permeator
had essentially no flux loss and no increase in bundle pressure drop.
This would indicate that the first stage permeators were effectively
filtering out the iron bacteria or other substance which was causing
the increasing pressure drop across the first stage permeators.
When the two additional permeators were brought into operation, the
original permeator was used as the last stage in the 2-1 array. Even
though some iron fouling had occurred, the permeator (#691) was not
flushed with any cleaning solutions. The flux history for the second
stage operation of this permeator is presented in the lower curve of
Figure 22. From 838 to 1400 hours the overall system recovery was 76.1
± 0.6%. Essentially no flux was lost during this period (Phase I,
Figure 22). At 1224 hours the overall recovery was increased to 85%
and an immediate decline in flux was noted. The recovery was reduced
to 75% at 1269 hours and the flux recovered to the same value as it
was prior to the increase in recovery. The flux remained stable at
this level until 1430 hours. If the high recovery period is ignored
this period of time (838 to 1430 hours) represented a very stable
period of operation showing essentially no flux decline for the last
stage permeator. Brine flow rate during this period was 1.58 ±
0.03 gpm. During the same period of operation the flux slopes for the
two permeators in the first bank were also stable with brine flows of
1.20 ± 0.026 gpm. A comparison of brine flows and flux decline slopes
is shown below.
Case Brine Flow Flux Decline Slope Water Recovery
(gpm) (%)
1 0.54 ± 0.01 0.0306 ± 0.008 74.3
2 1.20 ± 0.026 0.012 ± 0.006 60.3
3 1.58 ± 0.03 0.011 ± 0.004 76.2
Based on the above comparison there appears to be a correlation between
flux decline and brine flow. It should be noted, however, that Case 1
was over a different time period than Cases 2 and 3, and an absolute
comparison cannot be made. In view of the large differences in decline
rates it is recommended that a minimum brine flow of about 1 gpm be
maintained.
At 1450 hours the overall system recovery was again raised to 85%.
The flux immediately began to decline at a rapid rate (Figure 22) in
the second stage permeator. The salt passage also increased, but the
61
-------
head loss across the bundle remained constant. This was a definite
indication of CaSO^ fouling. The CaSO^ precipitation occurs in the
outer most fibers, since the brine is most concentrated in this area.
A disruption in brine flow distribution results in some areas of the
fibers receiving no brine flow. This causes an increased salt passage.
The brine now short circuiting around these areas finds the path of
least resistance and this results in little change in head loss across
the fiber bundle. However, if the condition persists for a long enough
time an increase in headless would be noted as a greater and greater
portion of the brine flow area is plugged. During this same period
(595 - 668 hrs., Figure 23) the flux decline slope increased for the
two first stage permeators. The brine flows were also reduced to about
1.0 gpm during the period as compared to 1.2 gpm prior to increasing
the system recovery. This decrease may have initiated the flux decline
noted in the first stage permeators.
At 1537 hours (Figure 22), a fifteen minute high brine flow flush was
performed on the second stage permeator and the water recovery was
lowered to 80%. The flux decline rate decreased, but was still signif-
icantly higher than experienced at 75% water recovery. At 1700 hours
the last stage permeator was flushed in an attempt to remove some of
the CaSO^ precipitation which had occurred. A 50 gallon (2 wt %)
solution of ammoniated citric acid was recirculated through the
permeator for 2 hours at a pH of 4.0. A new solution was then made
and buffered to pH 8.3 and this solution was recirculated for 2 hours.
This flushing was only marginally successful and recovered only about
25% of the flux lost due to CaSO^ fouling. The system was put back
into operation and operated at 80% recovery for an additional 250
hours. The flux decline slope was approximately the same as before
the flush, and still significantly higher than the 75% recovery level.
At 2100 hours the last permeator was flushed with a 3.4 wt % EDTA and
1.7 wt % Na2S2C>4 solution for a period of two hours. This flush
recovered about 50% of the total flux lost due to CaSC>4 fouling. Next
a 2 wt % sulfamic acid flush was used, but this did not recover any
additional flux. At this point it was decided to wait until the end
of the study before attempting any additional cleaning on the second
stage permeator. The system was put back on line and operated another
600 hours at an overall recovery of 75.3 ± 0.5%. The flux over this
period was very stable as shown in Figure 22 from 2100 to 2700 hours.
The flux history at higher recoveries provided some valuable informa-
tion with regard to CaS04 fouling. Figure 24 presents a plot of the
CaSO^ solubility product experienced in the brine at various time
intervals and compares these values with the flux history over the
same period.
The peaks of the CaSO^ curve correlate well with the high flux decline
periods. The curves in Figure 24 indicate a molar solubility product
of about 25 to 35 x 10~5 is all which can be attained without CaS04
fouling. Above this range CaSC>4 fouling will cause rapid flux declines,
62
-------
1200
1600
2000
2400
2800
FIGURE 24
COMPARISON OF CaSO/ MOLAR SOLUBILITY PRODUCT AND .PRODUCT WATER FLOW
-------
The conclusion can then be made that CaSO^ levels in the raw feed water
will determine the maximum level of product water recovery.
Operation of the hollow fiber unit was terminated at total operating
hours of 2670. At this time the entire unit was flushed with a 4 wt %
solution of ffla2520^' This caused a flux increase in all permeators
as shown in Figures 22 and 23, but did not restore the bundle pressure
drop to the valuas experienced at the start of the study. This indi-
cates that the flushing was not completely successful in removing the
iron fouling. The units were then filled with product water and
disinfected with formaldehyde.
A summary of the chemical analysis for the hollow fiber unit is
presented in Table 13. Average brine and product water quality and
the range at the 95% confidence levels are shown. Salt rejection
percentage is presented for all three permeators. In general, salt
rejection was in the range of 99,2 to 99.7% with the exception of
silica which was 94-98%. Total dissolved solids rejection was about
99%. It can be noted that the product water produced even at these
high salt rejections does not meet USPHS standards with regard to iron
and manganese which are 0.3 mg/1 and 0.05 mg/1 respectively. The pH
is also too low at 4.2. The product water would therefore require
additional treatment. If the product water were blended with an
existing drinking water supply it would probably meet the requirements.
If it is desired to utilize the product water directly the iron and
manganese levels will have to be reduced and sufficient buffering
capacity added to provide a stabilized water. This buffering capacity
is extremely important since the water is almost equal to distilled
water and could cause corrosion problems,
A single set of analyses was performed to determine the carbon dioxide
levels in the feed, product and brine flows of the hollow fiber system.
These analysis indicated the raw feed water had a C02 level of 6.4 mg/1.
The product water contained 6.9 mg/1 while the brine only contained
2.9 mg/1. Mass balances across the system were correct within 1.5%.
This means that essentially all the CC>2 was passing through the
membrane i.e. no rejection of C02 was being observed. The test
prodedure consisted of purging the CC>2 from the mine water by air
stripping. The COj laden air was bubbled through a potassium chromate-
sulfuric acid solution, through an iodine solution and finally through
barium hydroxide. The C02 caused a precipitate of barium carbonate
to form and the CC>2 concentration was determined by back titration of
the barium hydroxide solution.
In summary the operation of the hollow fiber system was satisfactory
since only a slight amount of iron fouling was experienced. This
fouling can be controlled by periodic flushing of the membranes or by
addition of acid to lower the feed pH as suggested by others (8) .
Acid addition will eliminate iron fouling by keeping all ferric iron
in the soluble state. Calcium sulfate fouling can be controlled by
keeping the recovery at the proper levels. It may be concluded that
64
-------
TABLE 13
SALT REJECTION CHARACTERISTICS
HOLLOW FIBER RO SYSTEM
LABORATORY ANALYSIS
Ion
Calcium
Magnesium
Manganese
Iron (Total)
Aluminum
Silica
Total Dissolved
Solids
Calcium
Total Hardness
As (CaC03)
Iron (Total)
Iron (II)
Sulf ate
pH
Raw Water
Quality
mg/1
111 ±
83 ±
14 ±
70 ±
8 ± 0
11 ±
1319
118 ±
602 ±
67 ±
64 ±
774 ±
3.38
6
4
0.6
6
.3
0.5
± 85
6
42
5
4
45
± 0.06
Brine Product Water
Quality Quality
mg/1 mg/1
487
381
57 ±
308
35 ±
49 ±
5809
523
2656
309
280
3146
2.8
± 69
± 56
10
± 60
5
6
± 862
FIELD
± 53
± 274
± 40
± 37
± 381
± 0.02
0.76 ± 0.2
0.66 ± 0.1
0.12 ± 0.03
0.60 ± 0.2
0.21 ± 0.09
0.83 ± 0.3
25 ± 9
ANALYSIS
1.04 ± 0.2
5.0 ± 1
0.60 ± 0.06
0.57 ± 0.06
3.2 ± 1.0
4.2 ± 0.1
Salt Rejection %
Module
1129
99.67
99.67
99.66
99.60
99.34
94.40
99.10
99.74
99.77
99.78
99.77
99.90
—
± 0.05
± 0.06
± 0.16
± 0.16
± 0.05
± 1.9
± 0.4
± 0.11
±0.06
± 0.03
± 0.03
± 0.06
1131
99.66 ±
99.64 ±
99.61 ±
99.43 ±
99.67 ±
94.20 ±
98.95 ±
99.76 ±
99.77 ±
99.76 ±
99.75 ±
99.90 ±
—
691
0.07
0.08
0.08
0.26
0.08
1.9
0.7
0.11
0.71
0.04
0.04
0.04
99.69
99.65
99.61
99.62
99.30
97.66
99.35
99.63
99.59
99.63
99.62
99.80
—
± 0.09
± 0.07
± 0.12
± 0.11
± 0.3
± 0.6
± 0.3
± 0.08
± 0.10
± 0.04
± 0.05
± 0.06
NOTES: All ranges at 95% confidence level
Salt Passage (%) = 200 (product water quality) /(feed water quality + brine quality)
See Appendix for detailed data
-------
the feasibility of acid mine treatment by RO has been established.
The mechanical operation of the hollow fiber system was excellent.
No failures were experienced. With regard to the remainder of the RO
components, the only item which required maintenance was the RO high
pressure pump. Frequent greasing was required for the pump packing.
No shear pan failures were experienced in the entire 3800 hour run
since the pins were replaced four times during the run. The system
was never shut down longer than 15 minutes for pump maintenance.
Operation of the Brine Treatment System
The brine exiting from the RO units contains all the impurities orig-
inally present in the raw waste. The main constituents include calcium,
magnesium, manganese, iron, aluminum, silica, and sulfate. The concen-
tration of these elements will be related to the feedwater quality
and the water recovery of the RO unit (see Table 13). In any brine
treatment system, the objective is removal of these pollutants. In
the case of iron and aluminum this is easily accomplished, since these
metals are quite insoluble in certain pH ranges and can be precipi-
tated as the metal hydroxides. Magnesium on the other hand is quite
soluble and difficult to remove. Manganese when oxidixed to the
travelent state will also form insoluble hydroxides, but complete
removal can be obtained only at a pH above about 9.5. Silica will
also complex with calcium at this high pH and be removed from the liquid
phase. Calcium sulfate will precipitate to a limited extent in
accordance with its solubility, however, the overflow from a neutra-
ization system will be saturated with respect to CaSO^. Since many
of the impurities can be removed by simple neutralization, the brine
treatment system consisted of neutralization followed by aeration to
oxidize the iron and possibly manganese, and then sedimentation.
The brine treatment system was operated on either the tubular RO brine
or the hollow fiber brine, but not a mixture. Bench scale tests were
run in conjunction with the full scale system to adequately define
the operating criteria which were utilized.
Bench Scale Testing
Iron oxidation tests were performed at pH 6.6 and 7.7 to determine the
iron oxidation rates. Figure 25 presents the data collected. It is
obvious that a pH of about 7.7 is required for rapid and complete
oxidation of the ferrous iron. The source of water for this test was
RO brine from the hollow fiber unit operating at 75% recovery.
Bench scale settling rate tests were performed at various oxidation
times. The results of these tests are plotted in Figure 26. It may
be seen that the settling rate increases with aeration time at a con-
stant pH of 7.7. It is not known if higher settling rates would be
produced at lower oxidation times by raising the pH. This is an area
which required additional study. In any event the settling rates at
66
-------
oc
ra
-------
1000
CO
Qj
U
i-l
4)
3
er
05
-a
•H
>=(
o
CO
c
o
f-i
ts
o
0,
750,
Aerated 12 min.
Aerated 30 min.
Aerated 60 min
Aerated 120 min
500
250
Test Run in a
1000 ml Graduated
Cvlinder
Tests Performed on Hollow Fiber
RO Brine at 75% Recovery, pH = 7.7
Hydrated lime dosage = 1000 mg/1
For analysis see Figure 25
20 30
Settling Time (Min)
FIGURE- 26
SETTLING RATE TESTS
-------
60 and 120 minutes of aeration provide sufficient settling velocity
for reasonable clarifier overflow rates.
A test was also run where the iron was not oxidized. The pH was
raised to 7.7 and the waste flocculated and settled. This procedure
produced a settling rate comparable to the curve for 60 minutes
aeration (Figure 26), however, the effluent contained 70 mg/1 of
soluble iron (II). It was, therefore, concluded that this mode of
treatment was not feasible.
In order to evaluate neutralization on a bench scale, two experiments
were run to evaluate CaS04 precipitation, which might occur in the
system. The results of these two tests are shown in Table 14, Both
tests were run in a 1000 ml graduated cylinder. Test I utilized 75%
RO brine. The lime (Ca(OH)2) dosage was 1288 mg/1 which was equivalent
to 695 mg/1 of calcium. As may be seen in Table I, the effluent
contained 800 mg/1 Ca after 22 hours of sedimentation. This was 360
mg/1 higher than the influent brine. Considering the calcium added
when neutralizing the brine, a net of 335 mg/1 of calcium was precip-
itated. The sulfate analysis, however, remained constant. The calcium
could have been precipitated as a calcium silicate complex and as
calcium carbonate (from the C02). It is also possible that all the
lime did not dissolve. It was obvious however, that neutralization
and settling produced a water higher in calcium when compared to the
RO brine.
Bench scale Test II (Table 14) was run on 85% recovery RO brine. Lime
(Ca(OH)2) dosage was 1800 mg/1 or 97<+ mg/1 calcium. This test lasted
38 hours. During this time a steady decrease in calcium was found in
the effluent. Sulfate also decreased indicating CaSO^ was precipitating
Total calcium in the feed considering the lime added was 1656 mg/i
which means 816 mg/1 precipitated. The sulfate precipitated was
2200 mg/1. This would account for an equivalent of 915 mg/1 calcium.
The possible error in the sulfate analysis (±10%) could account for
this difference. In any event, the calcium in the effluent was still
higher than the influent brine.
Full Scale System Operation
The full scale neutralization system was operated on the RO brines
for four runs. Tables 15 and 16 summarize the operation of the
neutralization system. The results confirmed the bench scale tests.
Good removals of iron, and aluminum were achieved (95 to 99%). Iron
was removed as iron hydroxide and aluminum as aluminum hydroxide.
Silica removals were also high (95%) and removal was most probably via
a calcium silicate complex. Manganese removals were lower in the range
of 55 to 60% as manganese hydroxide. Calcium on the other hand
increased as previously noted in the bench scale studies (due to
lime addition for neutralization). Manganese was not removed to any
extent and in fact exhibited an increase in one case. It may be
69
-------
TABLE 14
BENCH SCALE NEUTRALIZATION TESTS
Test I
Feed
Effluent
296
280
2950
2600
440
30 Min
5.8
0
3000
3200
800
22 Hrs
0.8
0
3000
3200
800
Total Iron, mg/1
Iron (II), mg/1
Sulfate, mg/1
Total Hardness (CaCO-^) mg/1
Calcium, mg/1
Test Specifications
Lime Dosage - 1288 mg/1 Ca(OH)2 = 695 mg/1 Ca
pH - 7.9
Brine from Hollow Fiber RO Unit - 75% Recovery
Temperature 73° F
Test in 1000 ml graduate
Test II Feed
Total Iron, mg/1 ' 505
Iron (II) 450
Sulfate, mg/1 6250
Total Hardness CaC03, mg/1 3900
Calcium, mg/1 682
Effluent
30 Min
4.3
0
5000
5100
1180
12 Hrs
0
0
5000
4900
1180
21 Hrs
4750
4650
1040
38 Hrs
4050
4150
840
Test Specifications
Lime Dosage - 1800 mg/1 Ca(OH)2
pH - 7.9
Brine from Hollow Fiber RO Unit
Temperature 74° F
Test in 1000 ml graduate
974 mg/1 Ca
85% Recovery
70
-------
TABLE 15
Run //
Brine Flow Rate, gpm
Duration of Run, Hrs.
Aeration Time, Hrs.
Settling Time, Hrs.
Hydrated Lime Dosage, mg/1
(Ca(OH)2
INFLUENT WATER QUALITY
pH
Iron, mg/1
Calcium, mg/1
Total Hardness, mg/1
as CaC03
Sulfate, mg/1
EFFLUENT WATER QUALITY
pH
Iron
Calcium, mg/1
Total Hardness as CaCC>3, mg/1
Sulfate, mg/1
RALIZATION SYSTEM OPERATION
1
2.5
73
1.95
11.5
505
2.9
108
368
1580
1400
6.5-7.0
6.0
340
L 1430
1250
2
1.7
118
2.86
17.0
1215
2.8
287
440
2950
2500
7.8
4.5
800
3350
2650
3
0.88
37
7.2
32.8
1720
2.8
475
720
4000
6500
7.6
0.9
1080
5000
4750
4
1.7
46
2.86
11.5
1215
2.9
340
560
2600
3100
7.7
5.5
800
3700
3500
71
-------
concluded that the quality of the effluent from the neutralization
system operating at a pH of 7.6-7.8 eliminated the possibility of
recycling this water to the RO system (in the light of the high
calcium content) as suggested by Hill et al. (9). It is possible
that operation at a lower pH would allow recycling, but clarity of
the settling tank overflow fould be adversely effected causing addi-
tional treatment problems. It is also possible that other treatment
systems could be utilized to render the sedimentation tank overflow
amenable to recycling back through the RO unit and hence eliminate
this liquid waste stream.
TAKLE_16
SUMMARY OF ATOMIC ABSORPTION ANALYSIS OF NEUTRALIZATION SYSTEM
Neutralization Run
Influent Effluent Influent Effluent
Calcium, mg/1 715 1030 500 800
Magnesium, mg/1 572 990 376 361
Manganese, mg/1 86.4 38.8 55.9 29.9
Iron, mg/1 493 0.3 312 1.4
Aluminum, mg/1 54 0.1 25 0.1
Silica, mg/1 76 4 45 3
72
-------
SECTION VIII
GENERAL DISCUSSION
Discussion of Flushing Techniques
Two sources of fouling were experienced during the field operation,
i.e. iron and CaSO^. Flushing methods were evaluated for effectiveness
in removing the precipitates from the RO modules and the effects of
the flushing solution on the RO membranes.
For removal of iron fouling, a two-weight percent product water solution
of citric acid adjusted to a pH of 4 with ammonia was evaluated. This
solution did not effectively dissolve the iron from the membrane, as
little flux changes were experienced. To further evaluated this
solution some precipitated iron was scraped from a raw AMD storage tank
and put into the ammoniated citric acid. The sample was then mixed on
a magnetic stirrer for 2 hours. Visual inspection revealed little if
any iron precipitate had dissolved. This same experiment was performed
using sodium hydrosulfite (1^23204) in a 4 weight percent solution and
the results were dramatic. All traces of iron precipitate were dissolved
within 15 minutes. This is consistent with results from a previous
study (3). The next step was to use sodium hydrosulfite on an actual
RO module. The 610 tubular modules which had been fouled with iron
were flushed with a 4 weight percent solution for 1.5 hours. This
resulted in an increase in flux from 12 gsfd to 17 gsfcL After operating
the system for about 2 hours the salt passage was at the same level as
prior to the flush. The hollow fiber modules were then flushed with
sodium hydrosulfite; flux increases were also experienced from 2.26 to
2.42 gpm per module for the first stage modules. Salt passage returned
to normal after 20 hours of operation. Based on the above testing it
appears that sodium hydrosulfite is an effective method for cleaning
AMD-iron fouled RO membranes.
For removal of CaSO, fouling the following solutions were evaluated:
a 2 wt % solution of citric acid buffered to pH 8 with ammonia; a 3.4
wt % solution of EDTA - 1.7 wt % solution of Na2S204; and a 2 wt %
solution of sulfamic acid. The only module fouled with CaSO^ was the
hollow fiber module No. 0691. (This was the only module operated at
high recovery.)
Flushing in the field was first accomplished utilizing the ammoniated
citric solution at pH 8 for 2 hours. This resulted in restoration
of about 25% of the flux which had been lost due to CaSO^ fouling.
Analysis of the flushing solution indicated an increase of calcium
levels from 19 tag/I to 870 mg/1 indicating calcium was being brought
into solution. The module was next flushed using the EDTA - Na2S204
solution. This resulted in restoration of flux from 1.4 to 1.6 gpm,
while the flux prior to CaS04 fouling was 1.8 gpm. The final flushing
73
-------
in the field was accomplished using the sulfamic acid solution for a
period of two hours. This resulted in no appreciable increase in flux.
No additional field flushing was attempted. At the completion of the
study, permeator 691 was returned to the laboratory for additional
flushing. Based on flush water analysis the best solution appeared
to be ammoniated citric acid at pH 8. It was felt that the time of
flushing was important and that the limited success in the field with
this solution was a result of insufficient flushing time. Prior to
flushing, the module was recharacterized and then the bundle was removed
for inspection. A solid ring of CaS04 precipitate was found near the
brine exit end of the module. The bundle was reinstalled and the
flushing precedure along with performance results are presented in
Table 17. As may be seen the precipitate was difficult to remove,
however, after 20 hours of flushing the performance was restored to the
same levels that existed in the field prior to the high recovery run and
subsequent CaSO^ fouling. Inspection of the bundle indicated no traces
of the calcium sulfate deposits previously noted. The fiber bundle was
then unrolled and all fibers inspected. No pockets of precipitates
were found in the entire bundle. Tests were then run on the fiber to
determine if any damage had occurred during 2670 hours of operation and
the many chemical flushes which were performed. Fiber strength, and
elasticity indicated absolutely no damage had occurred to the fibers
as all tests were comparable to new fiber (14). It was concluded that
CaSO^ fouling could be removed utilizing ammoniated citric acid at pH 8
and that no membrane damage was observed from utilization of the various
flushing solutions or operation on AMD for 2670 hours.
Economic Consideration :for RO - AMD Operation
Based on the results of the field evaluation phase, estimates of the
costs associated with treatment of AMD via RO were prepared. The flow
sheet utilized is shown in Figure 27. The following assumptions were
made to arrive at the costs shown.
1. Hollow fiber RO modules were utilized.
2, RO product water capacity was 750,000 gpd.
3. Chemical additive costs were based on field testing results.
4. Diatomaceous earth filtration was utilized.
5. No costs for buildings or land were included.
6. The product water from the plant meets USPHS standards.
7. No costs were included for disposal of residuals.
8. Operating manpower included a plant manager and a crew of 3.
Total salary and administrative costs - $50,000 per year.
74
-------
TABLE 17
LABORATORY RO MODULE CLEANING RESULTS
Test Sequence - Description
Original performance data
prior to field operation
Performance upon returning
from field
Product Water
Flow Water Salt
400 psi-68°F Recovery Passage Bundle
gpm % % AP
2.17
1.64
76.4
2.7
74.8 11.1
Performance after 3 hours
shell feed flush
1.82
74.7 10.7
15
Performance after 3 hours
shell feed - 3 hours
distributor feed flushes
Performances after 5 hours
distributor feed flush -
14 hour tap water flush
6 hour shell feed flush
1.98
1.85
75
75.6
9.4
3.5
3.5
NOTES: Flushing solution ammoniated citric acid pH 8.
Test solution 1500 mg/1 NaCl
75
-------
AMD
Feed Water
i
i
I
Pretreatment I
1,000,000
. Filtration f — ~~T
. pH adjustment! gp
3, Disinfection j
/
Recycled i
RO
Hollow Fiber
Brine
Product Water 790,000 gpd
750,000 gpd |
250,000 god
Neutralization
& Oxidation
V
\
\
Variable Depending \^ ^/
upon
l
3flw W3f,«=r
^— , ___ ___— • -^
Quality
pH
Adjustment
f\ f -\ ^
Overflow S
\
'
X
l
To Disposal
Variable Depending
Upon Raw Water
Quality
^^
L—-- ^
Sedimentation
^~~^^^
*
^
Sludges To
Disposal
37,500 gpd
•v
\ Overflow
j 40,000 gpd
/
Mn & Fe
Removal
Chlorination
\
1
Drinking
Water
790,000 gpd
FIGURE 27
FLOW SHEET USED FOR COST ESTIMATES
-------
9. Power costs at 1.0
10. Chemical additives include acid, diatomaceous earth, lime,
chlorine, flushing chemicals for RO membranes, potassium
permanganate.
11. RO module life 4 years - replacement cost 28c/gpd capacity.
12. Brine treatment system of concrete construction with high
speed floating aerators.
13. Product water treatment system utilizes a portion of the sedi-
mentation tank overflow for neutralization and potassium
permanganate for manganese oxidation - followed by filtration
and chlorination.
Shown below are the major cost items for the treatment system of Figure 27.
All cost estimates are based on vendor quotations or recent purchase prices
I. Capital Costs
A. Pretreatment
Filtration (diatomaceous earth)
pH control
Disinfection $ 29,000
B. RO System
Modules
Pumps and plumbing
Instrumentation. $385,000
C. Brine Treatment System
Aeration unit (high speed surface aerator)
Sedimentation unit
Chemical feeders and controls $ 58,000
D. RO Product Water Treatment
Iron and Mn removal
Final filtration
Chlorination . $ 31,000
TOTAL CAPITAL COST $503,000
Amortized Ld 6% - 20 yr = 15C/1000 gal Product Water
77
-------
II. Operating Costs in c/1000 gal Product Water
A. Chemical additives ....... 4.8
B. RO Modules 17.4
C. Power . 7.0
D. Maintenance - Materials 2.0
E. Operating Manpower 17.3
TOTAL... 48.5
The costs presented herein are estimated based on present day prices.
Advancement in RO hardware will undoubtedly bring price reductions in
the RO equipment. Also refinement of the flow sheet may also result
in more economical operation. One must also consider that two tasks
are being performed i.e. waste treatment and production of potable
water.
78
-------
SECTION IX
ACKNOWLEDGEMENTS
Many people in the Rex Ecology Division contributed to the success
of this project. Design of the demonstration system was made by
F. Toman with assistance from J. E. Milanowski. Operation of the
unit was performed by F. Toman, M. K. Gupta, and D. G. Mason.
Laboratory analyses were provided by the Ecology Division Analytical
Laboratory headed by R. E. Wullschleger. Bench scale experiments
were conducted by M. K. Gupta. The report was written by D. G. Mason
and M. K. Gupta.
Assistance from the staff at the F.PA Acid Mine test site, Norton,
West Virginia is appreciated. Review of RO operations with Robert
Scott and Roger Wilmoth provided valuable information. Alvin Irons
and Randolph Lipscomb provided expert assistance in field operation
of the RO unit.
Guidance from the EPA project officer, Ronald Hill, and key State of
Pennsylvania personnel, David Maneval and John Buscavage, is also
appreciated.
79
-------
SECTION X
REFERENCES
(1) EPA-ORM Mine Drainage Pollution Control Projects PPBS Code 14010,
updated periodicallv.
(2) EPA-ORM Mine Drainage Pollution Control Reports, Cincinnati, Ohio
April 1, 1970.
(3) Mason, D.G., Ecology Division, Rex Chainbelt Inc., Treatment of Acid
Mire Drainage by Reverse Osmosis, EPA Progress Report System No.
14010 DYK 03/71
(4) Kreinen, S.S. et al, The Reclamation of Acid Mine Water by Reverse
Osmosis, Third Symposium on Coal Mine Drainage Research, Mellon
Institute, May 19-20, 1970.
(5) Riedinger, A, and Schultz, J., A.cid Mine Water Reverse Osmosis
Test at Kittaning, Pennsylvania, Research and Development Progress
Report No. 217, Office of Saline Water, Washington, B.C., 1966.
(6) Furukawa, D., Flushing Techniques to Restore Flux in Reverse
Osmosis Plants, Division of Research, U.S. Bureau of Reclamation,
December 16, 1968.
(7) Xusbaum, I., et_ al, Reverse Osmosis Membrane Module, Research and
Development Report No. 338, Office of Saline Water, Washington,
D.C., March 1968.
(8) Acid Mine Waste Treatment Using Reverse Osmosis, by Gulf Environ-
mental Systems, EPA Progress Report System No. 14010 DYG 08/71.
(9) Hill, R.D., Wilmoth, R.C, and Scott, R.B., Neutralization Treatment
of Acid Mine Drainage, 26th Purdue Industrial Waste Conference,
Lafayette, Indiana, May 4-6, 1971.
(10) "Standard Methods for Examination of Water and Waste Water", APHA
AWWA WPCF. American Public Health Association, New York, New
York, Thirteenth Edition, 1971.
(11) Bennett, C. and Franklin, N., "Statistical Analysis in Che-istry
and the Chemical Industry", John Wiley and Sons, New York, 1954.
(12) Singer, P. and Stumm, W. , Kinetics of the Oxidation ol__Ferrous
Iron. Presented Second Symposium on Coal Mine Drainage Research
MeTIon Institute, Pittsburgh, Pennsylvania, May 14-15, 1968.
(13) Wilmoth, R., Private communication EPA Acid Mine Drainage
Treatment Lab, Norton, West Virginia, May 1971,
81
-------
(14) Tomsic, V., Private Communication, DuPont Co., Wilmington,
Delaware, November 1971.
82
-------
SECTION XI
PUBLICATIONS
Portions of the work described herein will be utilized in a technical
paper to be presented at the 4th Symposium on Acid Mine Drainage
Mellon Insitute, Pittsburgh, Pennsylvania in April 1972.
83
-------
SECTION XII
GLOSSARY OF TERMS
AMD - Acid Mine Drainage
Brine - The waste stream exiting a Reverse Osmosis unit, also can be
called concentrate
Concentrate - The waste stream exiting a Reverse Osmosis unit, can also
be called brine
Flux - Rate of water passage through a Reverse Osmosis membrane usually
expressed as gallons per sq ft of membrane per day (gsfd)
Permeate - Water which has passed through a Reverse Osmosis membrane -
could be termed product water
Permeator - A hollow fiber Reverse Osmosis module, trade mark of DuPont
Product Water - Water which has passed through a Reverse Osmosis membrane -
could be termed permeate
RO - Reverse Osmosis
Salt Rejection - Measure of the amount of salts not passing through
the membrane
Salt Passage - Measure of the amount of salts passing through the membrane.
85
-------
SECTION XIV
APPEND 1C I ES
I. Operating Data Hollow Fiber RO System
This Appendix contains all operating data for the hollow fiber RO system,
The normalized flux values were calculated using the temperature and
pressure correction equations shown below:
For Temperature - Q68 = QT x (1.0166768"1)
where: Q68 = Product water flow at 68°F
QT = Product water flow at observed temperature °F
T = Temperature °F
For Pressure - Qk00 = Qo [800/(P1 + P2)]
where: QI+QO = Product water flow at 400 psig
Qo = Observed product water flow
P1 = Pressure into the module, psi
?2 = Pressure out of the module, psi
When the system was operated as a 2-1 array flow control orifices were
utilized to insure equal flow distribution to the two first stage
permeators. The orifice headloss equation is:
AP = QB x 22.2293
where AP = the head loss across the orifice in psi
Q = the flow through the orifice
The orifices were located in the brine line ahead of the pressure gauge,
hence, the interstage pressure shown in Table 1-2 includes the orifice
pressure drop.
87
-------
TABLE 1-1
HOLLOW FIBER OPERATING DATA
FOR SINGLE PERMEATOR
Elapsed
Time
(hrs)
0
3
9
19
21
25
25
32
43
49
52
68
73
81
94
95
101
114
118
119
125
137
146
151
163
188
198
211
216
222
235
240
243
245
247
257
263
284
289
305
309
314
Pressure
In
Out
(jpsi)
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
380
400
398
400
400
380
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
390
390
390
390
390
390
389
390
390
390
400
400
400
380
400
398
400
400
380
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
Temp.
°F
53.0
53.2
52.3
52.3
53.2
54.0
54.0
53.8
53.0
53.0
52.0
52.9
54,0
53.0
53.0
53.0
53.0
52.0
53.0
53.0
53.0
52.9
53.0
53.0
53.0
53.0
53.0
53.0
53.0
53.0
53.0
54.0
54.0
53.0
53.0
53.0
53.5
5.3
53.0
53.0
54.0
53.5
1 Normalized to 400
Brine
Flow
gpm
0.50
0.51
0,51
0.51
0.51
0.55
0.54
0.43
0.54
0.53
0.54
0.53
0.53
0.54
0.54
0.49
0.59
0.58
0.58
0.60
0.59
0.59
0.58
0.58
0.58
0.59
0.59
0.58
0.58
0.58
0.58
0.58
0.59
0.58
0.59
0.60
0.52
0.52
0.53
0.53
0.53
0.53
psi -
Product
Flow
sec /gal
35.50
35.60
36.20
36.40
36.10
36.60
36.50
37.00
37.40
37.30
37.30
37.60
37.40
37.60
38.10
37.20
37.40
39.40
38.90
37.20
38.00
36,90
36.70
37.80
38.10
38.00
38.10
38.50
38.75
38.90
38.60
38.60
38.35
38.85
38.10
38.50
38.45
39.00
39.00
39.00
38.65
38.75
68°F
1
Normalized
Product
Flow gpm
2.166
2.153
2.149
2.137
2.123
2.092
2.098
2.077
2.082
2.087
2.087
2.077
2.048
2.071
2.044
2.067
2.056
1.955
2.081
2.061
2.034
2.087
2.095
2,141
2.018
2.023
2.018
1.997
1.984
1.977
1.992
1.959
1.972
1.979
2.018
1,997
1.983
4.338
1.971
1.971
1.957
1.968
Product
Water
Recovery - %
77.2
76.9
76.5
76.4
76.5
74.9
75.2
74.9
75.0
75.1
75.0
74.9
75.0
74.8
74.4
73.2
73.1
72.5
72.5
73.0
72,7
73.3
73.7
73.1
73.0
72.9
72.9
72.8
72.7
72.6
72.9
72.9
72.7
72.5
72.7
72.2
74.9
74.4
74.5
74.5
74.7
74.6
88
-------
TABLE 1-1 (Continued)
HOLLOW FIBER OPERATING DATA
FOR SINGLE PERMEATOR
Elapsed
Time
(hrs)
329
337
355
377
378
381
394
405
419
422
432
443
449
478
478
491
523
527
539
546
563
575
587
595
615
636
641
660
685
693
698
705
718
718
740
764
772
789
812
838
Pressure
In
Out
(psi)
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
405
405
400
390
405
405
405
405
415
405
390
402
410
400
400
400
400
400
400
400
4 jG
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
390
385
400
400
398
400
400
390
385
390
390
385
387
390
383
383
380
380
380
383
380
380
Temp .
°F
53.0
54.0
54.0
52.5
54.0
53.0
52.5
52.5
52.5
53.0
52.0
52.5
52.5
53.0
53.0
53.0
53.0
53.0
53.0
53.5
52.9
53.0
53.0
54.0
53. C
54.0
54.0
53.5
53.5
54.2
55.0
54.5
54.0
54.0
54.0
54.0
55.0
54.8
55.0
54.8
Brine
Flow
gPm
0.53
0.53
0.52
0.52
0.55
0.55
0.55
0.54
0.54
0.53
0.53
0.53
0.52
0.49
0.54
0.54
0.53
0.53
0.51
0.50
0,51
0.52
0.51
0.52
0.52
0.51
0.50
0.51
0.50
0.52
0.51
0.47
0.46
0.47
0.52
0.51
0.48
0.48
0.46
0,47
Product
Flow
sec/gal
39.50
39.0
39.40
40.10
37.50
38.50
39.00
38.50
38.50
38.50
38.90
38.90
38.40
39 . 20
39,20
38.50
39.60
38,90
40.30
41.00
39.50
39.65
39.70
39.60
39.50
40.40
41.50
40.10
41.00
39.30
39.90
40.10
40.35
40.40
41.40
41.10
40.85
40.20
43.90
42.00
Normalized
Product
Flow gpm
1.956
1.939
1.919
1.933
2.017
1.997
1.988
2.014
2.01^
1.007
/ . 0 !". -1
1.993
1,986
1.961
1.961
1.997
1.930
1.964
1.932
1.920
1.938
1.927
1.929
1.898
1.911
1.884
1.881
1.921
1.860
1.955
1.895
1.894
1.915
1.913
1.874
1.887
1.868
1.897
1.738
1.823
Product
Water
Recovery - %
74.3
74.5
74.4
74.2
74.3
73.8
73.7
74, J
7 4 .3
74.5
74.5
74.6
74.9
74.7
73.8
74.2
73.9
64.6
74.5
74.6
74.7
74.6
74.7
74.6
74.6
74.4
74.4
74.6
74.4
74.7
74.7
76.1
76,3
75.8
73.4
74.2
75.5
75.6
74.6
75.1
1 Normalized to 400 psi - 68°F
89
-------
TABLE 1-2
Elapsed
Time
hrs.
1
9
22
70
97
116
140
140
145
163
187
191
212
235
259
284
308
332
340
355
359
380
386
390
404
414
426
431
432
447
452
471
476
484
497
516
540
562
568
HOLLOW
Pressure,
In
400
400
400
400
400
410
410
410
410
410
410
410
410
410
410
410
410
420
410
410
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
410
400
2nd
stage
345
345
350
350
345
345
349
340
347
342
340
345
348
348
340
342
347
345
345
345
333
330
350
347
347
346
346
330
327
325
325
300
325
325
330
330
330
340
328
FIBER
psi
Out
315
315
315
315
310
305
305
300
304
300
300
304
305
303
302
300
305
302
308
301
293
290
320
320
317
316
316
292
289
285
285
290
285
285
290
290
290
300
290
OPERATING DATA 2-1 ARRAY
Feed
Temp
OF
54.6
53.0
54.5
54.5
54.0
53.0
53.0
53.0
54.0
53.0
54.0
54.7
54.0
53.5
54.0
54.0
54.0
55.0
54.0
54.0
54.5
54.0
54.5
53.8
53.8
54.0
53.5
54.0
54.5
54.5
54.5
55.0
55.0
54.0
54.5
54.0
54.0
54.0
54.5
Brine
Flow
sec/gal
39.6
39.6
40.7
42.7
38.4
36.8
37.3
35.5
35.5
35.7
35.3
35.0
36.0
36.0
35.4
35.5
35.5
35.6
36.1
35.7
37.5
37.8
69.0
68.3
68.1
68.4
68.6
40.5
39.0
39.3
38.7
40.8
40.2
40.8
52.2
40.5
40.1
40.3
41.9
Product Water Flow (sec/gal
1129
28.4
28.7
28.6
28.7
29.5
29.8
29.9
30.2
29.7
30.0
30.2
29.5
29.8
29.9
30.0
30.1
29.9
29.8
29.8
30.2
30.5
31.2
30.5
31.1
31.0
31.1
31.6
31.5
31.0
30.9
30.7
30.9
30.7
30.9
30.6
30.5
30.5
30.6
31.4
Module
1131
28.9
29.5
29.2
29.1
30.1
30.3
30.5
30.7
30.2
30.4
30.9
29.8
30.3
30.4
30.7
30.6
30.3
30.4
30.5
30.8
31.3
32.0
31.2
31.8
31.4
32.0
32.1
32.0
32.2
31.5
31.5
31.5
31.6
31.6
30.7
30.8
30.7
30.6
31.9
691
50.0
51.0
49.9
49.1
51.3
52.5
52.0
53.0
52.5
52.5
52.6
51.2
51.5
52.0
51.6
52.1
51.3
51.5
51.1
51.8
52.6
53.5
49.5
50.3
50 ,,6
51.1
52.3
55.0
55.9
55.4
54.8
54.2
54.8
55.0
54.0
54.2
53.3
55.0
55.6
90
-------
TABLE 1-2 (Continued)
Elapsed
Time
hrs.
578
595
499
512
620
636
646
661
667
669
684
692
706
713
729
741
756
760
778
802
829
852
906
922
947
971
975
978
987
1011
1016
1019
1023
1033
1059
1087
1111
1131
1150
1171
1195
1221
HOLLOW
Pressure,
In
400
400
400
400
405
400
400
400
303
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
410
401
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
2nd
stage
340
338
431
340
342
340
338
342
343
334
334
334
331
322
321
325
325
335
335
332
332
330
330
333
335
343
330
325
328
328
328
320
321
322
326
326
325
325
325
325
325
325
FIBER
psi
Out
320
312
319
318
320
315
313
318
319
309
309
309
301
290
290
295
292
305
303
303
302
302
300
302
302
310
310
310
310
320
320
301
309
309
310
310
310
310
310
310
310
310
OPERATING DATA 2-1 ARRAY
Feed Brine Product Water Flow (sec/gal)
Temp
°F
54.0
54.0
54.0
54.0
54.0
54.0
54.0
54.5
54.0
54.0
54.5
54.0
54.0
54.0
53.0
54.0
54.0
54.2
54.0
54.0
54.0
54.8
54.8
54.0
54.0
54.2
55.0
54.0
54.0
54.5
55.0
55.0
55.0
54.3
56.0
55.0
55.0
55.0
55.0
55.0
55.0
55.0
Flow
sec/fial
28.0
57.6
67.7
67.9
69.4
69.5
70.2
70.2
69.8
52.6
52.6
52.5
52.5
41.4
41.8
42.6
42.6
49.0
49.7
51.4
51.7
52.0
52.5
53.5
53.4
53.0
47.5
46.9
46.3
46.1
45.8
46.3
46.4
46.4
40.9
45.8
46.3
46.3
46.0
46.6
46.5
46.5
1129
31.3
31.9
31.9
32.0
31.8
32.2
32.7
32.1
32.4
32.4
32.5
32.4
32.8
33.1
33.4
33.3
33.0
32.6
32.9
33.1
32.9
33.1
32.9
33.2
32.8
32.7
32.1
32.6
32.7
32.6
32.7
32.5
32.7
32.5
32.2
32.4
32.8
32.6
32.8
32.6
32.9
32.7
Module
1131
31.7
32.7
32.7
32.5
32.5
32.9
33.3
32.7
32.8
33.0
33.3
33.1
33.4
33.8
34.1
33.6
33.7
33.1
33.2
33.5
33.4
33.4
33.5
33.7
33.4
32.9
32.4
33.7
33.1
33.7
33.4
33.4
33.4
33.0
32.9
33.6
33.4
33.4
33.6
33.4
33.5
33.4
691
49.5
53.3
53.1
53.6
54.0
55.2
56.9
56.8
57.9
58.5
59.9
59.9
60.8
63.2
69.9
62.4
62.6
60.3
61.1
62.1
62.5
63.2
64.0
65.4
66.2
65.4
61.6
61.3
61.6
62.5
62.5
61.9
62.5
62.5
61.3
63.0
63.3
63.3
63.1
63.6
64.1
64.2
91
-------
TABLE 1-2 (Continued)
HOLLOW
Elapsed Pressure,
Time
Mrs.
1226
1243
1256
1249
1273
1289
1315
1341
1366
1386
1395
1418
1442
1462
1484
1532
1561
1580
1603
1628
1642
1689
1718
1738
1764
1786
1814
1815
1825
In
400
410
400
400
405
405
400
400
402
400
402
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
2nd
stage
331
335
320
320
333
325
320
320
320
315
320
320
318
318
318
315
315
315
315
316
315
315
308
310
310
308
305
315
320
FIBER
psi
Out
311
315
305
305
320
310
300
302
305
300
305
305
300
300
300
298
298
298
298
297
298
298
295
296
295
292
290
298
300
OPERATING DATA 2-1 ARRAY
Feed Brine Product Water Flow (sec/gal)
Temp
°F
55.1
54.5
54.5
54.0
54.0
52.5
53.5
54.8
55.1
54.0
55.0
55.0
55.0
54.0
54.0
54.0
55.0
54.0
54.0
54.0
54.5
54.5
55.0
54.5
55.0
55.0
54.0
54.0
54.0
Flow
sec/gal
46.8
47.0
40.4
40.6
51.4
39.2
39.8
39.5
39.8
40.1
40.3
43.2
41.8
41.6
40.0
40.4
40.7
40.5
40.6
40.8
41.2
41.4
41.2
41.1
41.1
41.6
41.3
41.0
40.8
1129
32.4
32.7
33.4
33.5
32.6
33.6
33.9
33.8
33.6
34.6
33.7
34.1
34.4
34.9
34.7
35.5
35.6
35.5
35.3
35.5
34.5
34.7
34.9
35.1
35.1
35.4
35.5
33.2
33.3
Module
1131
33.2
33.7
33.5
33.9
33.4
34.0
34.2
34.2
34.2
35.0
34.2
34.6
34.8
35.5
35.3
35.7
35.9
35.7
35.8
35.8
34.9
35.1
35.3
35.8
35.7
35.7
35.8
34.0
33.6
691
63.8
64.2
59.6
60.5
56.5
60.5
61.1
60.4
60.1
61.7
60.5
60.8
61.2
62.8
62.2
63.0
63.2
63.6
63.4
63.7
61.8
62.3
63.0
63.8
' 62.8
63.0
63.4
59.6
59.0
92
-------
TABLE 1-3
OPERATIONAL DATA HOLLOW FIBER RO SYSTEM
(2-1 ARRAY FIRST STAGE MODULES)
Elapsed
Time
(hrs)
1
9
22
70
97
116
140
140
145
163
187
191
212
235
259
284
308
332
340
355
359
380
386
390
404
414
426
43:
43
-------
TABLE 1-3 (Continued)
OPERATIONAL DATA HOLLOW FIBER RO SYSTEM
(2-1 ARRAY FIRST STAGE MODULES)
Elapsed
Time
(hrs)
578
595
599
612
620
636
646
661
667
669
684
692
706
713
729
741
756
760
778
802
829
852
906
922
947
971
975
978
987
1011
1016
1019
1023
1033
1059
1087
1111
1131
1150
1171
1195
1221
Normalized
Water Flow
Module
1129
2.408
2.482
2.222
2.480
2.471
2.469
2,440
2.448
2.435
2.457
2.431
2.455
2.439
2.424
2.455
2.403
2.425
2.426
2.414
2.414
2.429
2.390
2.407
2.410
2.434
2.375
2.443
2.464
2.445
2.430
2.405
2.446
2.429
2.469
2.394
2.434
2.409
2.424
2.409
2.425
2.404
2.418
Product
gpm
1131
2.378
2.421
2.168
2.442
2.422
2.416
2.396
2,403
2.405
2.412
2.372
2.406
2.392
2.374
2.401
2.381
2.374
2.390
2.393
2.385
2.393
2.369
2.364
2.374
2.390
2.361
2.417
2.383
2.416
2.354
2.351
2.381
2.378
2.432
2.343
2.348
2.366
2.366
2.451
2.367
2.360
2.368
Brine Flow
gpm Each
Module
1.68
1.08
1.01
1.00
0.99
0.98
9.95
0.96
0.95
1.08
1.07
1.07
1.06
1.20
1.15
1.18
1.18
1.11
1.09
1.07
1.06
1.05
1.04
1.02
1.01
1.02
1.12
1.13
1.13
1.13
1.14
1.13
.1.13
1.13
1.22
1.13
1.12
1.12
1.13
1.12
1.11
1.11
Water
Recovery
% Each
Module
53.1
6.31
64.8
65.0
65.3
65.4
65.5
65.9
66.0
62.8
63.0
63.0
62.9
59.9
60.7
60.2
60.3
62.2
62.3
63.8
63.0
63.1
63.4
63.7
64.1
64.0
62.4
61.5
61.6
61.5
61.5
61.6
61.9
60.1
61.6
61.7
61.8
61.5
61.9
61.8
62.0
Bundle
Pressure
Drop
psi
35
35
35
37
41
38
41
37
41
39
40
40
43
46
49
43
43
37
38
42
43
45
45
43
42
43
43
46
43
43
43
51
49
40
45
47
47
46
47
47
47
94
-------
TABLE 1-3 (Continued)
OPERATIONAL DATA HOLLOW FIBER RO SYSTEM
(2-1 ARRAY FIRST STAGE MODULES)
Elapsed
Time
(hrs)
1226
12A3
1256
1249
1273
1289
1315
1341
1366
1386
1395
1418
1442
1462
1484
1532
1561
1580
1603
1628
1642
1689
1718
1738
1764
1786
1814
1815
1825
1832
Normalized Product
Water Flow
Module
1129
2.418
2.376
2.381
2.396
2.424
2.412
2.385
2.339
2.336
2.335
2.335
2.323
2.305
2.312
2.319
2.278
2.236
2.279
2.292
2.277
2.326
2.314
2.303
2.304
2.284
2.272
2.312
2.433
2.408
2.396
gpm
1131
2.359
2.306
2.370
2.367
2.366
2.383
2.361
2.312
2.295
2.308
2.301
2.289
2.279
2.272
2.280
2.266
2.217
2.267
2.260
2.258
2.299
2.287
2.277
2.259
2.245
2.253
2.293
2.376
2.387
2.353
Brine Flow
gpm Each
Module
1.11
1.11
1.25
1.23
1.11
1.26
1.24
1.26
1.25
1.23
1.24
1.19
1.21
1.20
1.23
1.22
1.21
1.21
1.21
1.21
1.21
1.21
1.20
1.20
1.21
1.20
1.20
1.24
1.24
1.24
Water
Recovery
% Each
Module
62.2
62.0
59.0
59,0
61.9
58.4
58.5
58.4
58.5
58.2
58.7
59.5
58.9
58.7
58.1
58.0
58.0
58.1
58.2
58.2
58.7
58.7
58.6
58.5
58.3
58.4
58.3
59.1
59.0
59.0
Bundle
Pressure
Drop
psi
41
47
45
46
44
44
45
44
47
51
47
48
49
50
48
52
52
52
52
51
52
52
59
58
57
60
63
51
45
45
95
-------
TABLE 1-4
Elapsed
Time
839
847
860
908
955
954
978
979
983
1002
1026
1030
1050
1073
1098
1122
1146
1170
1178
1193
1197
1218
1224
1228
1242
1252
1264
1269
1271
1285
1290
1309
1315
1322
1335
1355
1378
1401
1406
1416
HOLLOW FIBER
(2-1 ARRAY,
Normalized
Product Water
Flow gpm
Module 691
1.815
1.827
1,308
1,838
1.801
1.802
1.809
1,813
1.770
1,825
1.797
1.797
1.799
1.802
1.826
1.809
1.809
1.786
1.813
1.808
1.822
1.824
1.809
1.809
1.807
1.788
1.762
1.769
1.743
1.776
1.795
1.861
1,780
1.803
1.792
1,. 800
1.831
1.719
1.745
SYSTEM i
2ND STAG
Brine
Flow
Rate
gpm
1.51
1,52
1.47
1.41
1.56
1.63
1.61
1.69
1.69
1.68
1.70
1.71
1.67
1.67
1.69
1.69
1.69
1.69
1.66
1.68
1.60
1.59
0.87
0.88
0.88
0.88
0.87
1.48
1.54
1.53
1.55
1.47
1.49
1.47
1.15
1.48
1.50
1.49
1.43
1.852
Brine
Flow
Rate
gpm
1.51
1.52
1.47
1.41
1.56
1.63
1.61
1.69
1.69
1.68
1.70
1.71
1.67
1.67
1.69
1.69
1.69
1.69
1.66
1.68
1.60
1.59
0.87
0.88
0.88
0.88
0.87
1.48
1.54
1.53
1.55
1.47
1.49
1.47
1.15
1.48
1.50
1.49
1.43
2.14
Overall
Syscem
Recovery
%
78.1
77.8
78.4
79.3
76.9
75.9
76.1
75.0
75.3
75.3
74,9
75.3
75.6
75,5
75. 1
75.1
75.3
75.4
75.6
75.2
75.8
75.6
85.4
85.1
85.1
85.0
84.9
76.7
75.9
76.4
76.2
77.1
76.8
77.0
81,4
77.2
77.1
77.1
77.3
70.1
Bundle
Pressure
Drop
psi
30
30
35
35
35
40
44
40
43
42
40
40
43
45
38
42
42
43
37
44
40
40
30
27
30
30
30
38
38
40
40
10
40
40
40
40
40
40
38
30
96
-------
TABLE 1-4 (Continued)
HOLLOW FIBER SYSTEM OPERATION
(2-1 ARRAY, 2ND STAGE MODULE)
Elapsed
Time
(Hrs)
1434
1437
1450
1459
1474
1484
1499
1506
1507
1522
1530
1545
1551
1567
1580
1594
1599
1616
1640
1667
1691
1744
1761
1785
1810
1813
1816
1825
1850
1855
1857
1862
1872
1897
1925
1949
1969
1988
2010
2034
Normalized
Product Water
Flow gpm
Module 691
1.746
1.518
1.715
1.692
1.673
1.633
1.601
1.578
1.608
1.558
1.571
1.574
1.564
1.440
1.564
1.566
1.563
1.552
1.534
1.527
1.495
1.481
1.457
1.435
1.412
1.509
1.554
1.539
1.482
1.469
1.548
1.511
1.527
1.501
1.485
1.480
1.480
1.485
1.474
1.462
Brine
Flow
Rate
gpm
1.04
0.89
0.88
0.86
0.86
0.85
0.85
0.86
1.14
1.14
1.14
1.14
1.45
1.44
1.41
1.41
1.22
1.21
1.17
1.16
1.15
1.14
1.12
1.12
1.13
1.26
1.28
1.30
1.30
1.31
1.29
1.29
1.29
1.47
1.31
1.30
1.30
1.30
1.29
1.29
Overall
System
Recovery
%
82.3
84.5
84.6
84.8
84.7
84.6
84.8
84.6
80.5
80.3
80.3
80.1
75.8
75.5
76.4
76.4
79.1
79.3
79.7
79.8
79.8
79.9
80.1
80.1
80.2
78.8
78.2
78.1
77.9
77.8
78.1
78.0
78.1
76.1
77.8
77.9
78.0
77.8
78.1
77.9
Bundle
Pressure
Drop
psi
26
11
22
22
25
25
24
24
25
25
25
30
32
31
30
33
30
32
29
30
28
30
31
33
33
20
15
18
8
8
19
12
13
16
16
15
15
15
15
15
97
-------
TABLE 1-4 (Continued)
Elapsed
Time
(Hrs)
2059
2065
2081
2094
2087
2112
2128
2153
2179
2204
2225
2233
2256
2280
2300
2322
2371
2399
2418
2442
2466
2480
2527
2556
2576
2602
2624
2652
2653
2663
2670
HOLLOW FIBER
(2-1 ARRAY,
Normalized
Product Water
Flow gpm
Module 691
1.460
1.450
1.438
1.609
1.600
1.640
1.614
1.610
1.589
1.582
1.594
1.574
1.566
1.573
1.559
1.574
1.567
1.536
1.552
1.557
1.549
1.584
1.571
1.566
1.552
1.566
1.574
1.604
1.656
1.654
1.646
SYSTEM OPERATION
2ND STAGE
Brine
Flow
Rate
gpm
1.29
1.28
1.28
1.49
1.48
1.17
1.53
1.51
1.52
1.51
1.50
1.49
1.39
1.44
1.44
1.50
1.49
1.47
1.48
1.48
1.47
1.46
1.45
1.46
1.46
1.46
1.44
1.45
1.46
1.47
1.46
MODULE)
Overall
System
Recovery
%
78.0
78.2
78.1
75.6
75.5
80.1
74.8
74.9
74.9
75.1
74.7
75.2
76.3
75.6
75.2
74.5
74.4
74.5
74.4
74.5
74.6
75.3
75.2
75.0
74.8
74.9
75.0
74.8
75.8
75.8
75.9
Bundle
Pressure
Drop
psi
15
20
20
15
15
13
15
20
18
15
15
15
15
18
18
18
17
17
17
17
19
17
17
13
14
15
16
15
17
20
18
98
-------
TABLE 1-5
Hour
Clock
RDG
Hrs.
1234
1302
1395
1590
1899
2051
2314
2528
2904
3074
3162
3489
ANALYSIS DATA HOLLOW
Ca
mg/1
115
115
115
97
97
102
107
112
120
125
123
103
Mg
mg/1
85
86
87
74
72
77
77
78
89
94
90
82
Mn
mg/1
FEED
13.8
14.3
14.5
12.4
12.7
13.2
13.5
14.8
14.6
15.0
14.9
13.8
FIBER SYSTEM (AA
Total
Iron
mg/1
WATER
71
74
75
54
57
59
69
74
80
84
80
64
Silica
mg/1
11
10
10
10
10
11
11
11
11
12
11
12
DATA)
Al
mg/1
8.1
7.8
8.0
7.4
8.3
7.9
8.5
8.5
8.4
9.0
8.6
9.0
TDS
mg/1
1281
1340
1337
1044
-
1248
1278
1277
1498
1510
1387
1313
FINAL BRINE
1234
1302
1395
1590
1899
2051
2314
2528
2904
3074
3162
3489
480
452
422
376
384
404
433
715
576
624
568
407
378
360
335
312
280
307
320
572
436
495
438
337
58.3
54.6
52.0
47.2
25.2
51.0
53.3
86.4
67.0
71.7
67.2
50.0
315
298
284
212
223
241
292
493
386
426
376
254
45
42
40
42
43
49
43
76
53
57
51
49
31.2
26.0
31.0
29.0
32.0
35.4
34.0
54.0
39.0
40.0
35.0
29.0
5567
5165
4791
4153
-
4862
5235
8197
7154
7394
6389
4994
99
-------
TABLE 1-6
Hour
Clock
Rdg
Hrs
ANALYSIS DATA HOLLOW FIBER SYSTEM (AA DATA)
Ca
mg/1
Mg
mg/1
Mn
mg/1
Total
Iron
mg/1
Silica
mg/1
FIRST STAGE BRINE
2314
2528
2904
3074
3162
3489
2314
2528
2904
3074
3162
3489
2314
2528
2904
3074
3162
3489
1234
1302
1395
1590
1899
2051
2314
2528
2904
3074
3162
3489
257
316
319
339
328
253
0.71
0.55
0.73
0.73
0.78
0.58
0.85
0.70
0.77
0.66
0.69
0.54
0.40
0.52
0.51
0.60
0.56
0.74
1.06
1.09
2.21
2.68
2.09
1.48
184
225
233
268
242
201
PRODUCT
0.54
0.44
0.49
0.55
0.58
0.43
PRODUCT
0.66
0.56
0.51
0.52
0.55
0.44
PRODUCT
0.44
0.92
0.72
0.50
0.40
0.56
0.77
1.23
1.47
2.07
1.48
1.18
32.4
43.2
36.8
41.3
38.8
31.8
WATER
0.10
0.04
0.08
0.14
0.09
0.09
WATER
0.12
0.06
0.09
0.13
0.09
0.09
WATER
0.05
0.15
0.10
0.05
0.08
0.08
0.15
0.20
0.25
0.42
0.25
0.22
167
212
205
236
211
154
PERMEATOR
0.79
0,76
0.43
0.45
0.51
0.34
PERMEATOR
1.04
0.96
0.43
0.42
0.47
0.36
PERMEATOR
0.27
0.70
0.49
0.40
0.32
0.44
1.19
1.82
0.27
2.15
1.42
0.99
26
31
30
31
29
27
#1129
0.8
1.0
1.1
2.0
1.0
1.0
#1131
0.8
1.0
1.3
2.0
1.0
1.0
#691
0.50
0.40
0.75
0.50
0.40
0.40
0.8
1.0
1.3
2.0
1.0
1.0
Al
mg/1
20
23
22
24
22
21
0.1
0.1
0.4
0.4
0.1
0.1
0.1
0.1
0.4
0.3
0.3
0.1
0.10
0.40
0.18
0.20
0.04
0.15
0.10
0.10
0.40
0.50
0.40
0.10
IDS
mg/1
3104
3618
4035
4188
3704
3157
24
14
28
-
16
27
26
11
27
-
41
31
14
4
10
33
-
45
34
9
40
—
29
30
100
-------
TABLE 1-7
FIELD ANALYSIS DATA HOLLOW FIBER SYSTEM
Total Calcium
Hour
Clock
Rdg
1253
1282
1301
1328
1348
1380
1445
1632
1663
1682
1778
1826
1938
1956
2051
2314
2366
2528
2717
2810
3031
3051
3114
3370
Total
Iron
mg/1
74
67
74
75
65
71
76
54
50
56
55
50
55
56
56
57
68
70
100
88
66
70
75
78
Iron II
mg/1
66
72
67
66
65
70
70
49
50
47
51
50
51
56
56
57
65
69
76
78
73
76
75
78
C> f\
so4
mg/1
RAW
890
800
730
860
890
725
760
710
580
620
590
700
890
720
690
780
760
663
710
920
780
950
890
940
Hardness
(CaC03)
mg/1
AMD
560
580
728
716
684
600
640
480
500
500
500
350
465
588
550
560
580
650
650
650
720
700
800
680
Hardness
(CaC03)
mg/1
320
328
324
340
308
380
280
240
225
250
275
340
305
255
300
290
300
300
250
300
290
310
250
320
pH
Units
3.3
3.5
3.4
_
3.6
3.6
3.3
3.6
3.5
3.3
3.5
3.4
-
-
-
-
-
3.3
3.4
3.4
-
-
-
—
Meter
TDS
mg/1
_
_
_
1300
_
1275
1200
-
-
1300
-
-
1250
-
1100
1100
1100
1200
1200
1400
1350
1250
1400
INTER-STAGE BRINE
2314
2366
2482
2528
2717
2810
3031
3051
3114
3370
150
149
155
190
240
210
186
235
195
190
142
118
150
190
185
190
210
185
198
181
1300
1500
1700
1950
1700
2400
2600
2350
2350
2000
1400
1650
1400
1600
1500
1600
1680
1800
2200
1700
600
650
600
800
750
800
860
680
1000
700
_
-
3.0
2.9
3.1
3.2
—
-
-
—
2500
2300
2500
2650
2900
2900
3450
3400
3100
3200
101
-------
TABLE 1-8
Hour
Clock
RDG
Mrs,
1253
1282
1301
1328
1348
1380
1445
1632
1.663
1682
1778
1826
1938
1956
2051
2314
2366
2482
2528
2717
2810
3031
3051
3114
3370
FIELD ANALYSIS DATA HOLLOW FIBER SYSTEM
Total Calcium
Total
Iron
mg/1
304
275
285
388
300
315
255
189
189
199
200
150
220
220
230
260
305
280
498
540
450
370
450
390
340
Iron
mg/1
303
270
300
225
260
290
208
182
183
182
175
167
260
200
230
270
290
280
464
510
395
415
255
345
328
II S04
mg/1
3600
2800
2700
2500
2800
3100
3150
2500
2450
2500
2500
2450
2600
3000
2040
2600
2500
2800
6000
5000
4500
3800
3800
3850
3100
Hardness
Ca CO-j
mg/1
BRINE
3040
2000
2420
3100
2520
2300
2400
1900
2000
2100
2070
2070
2200
2300
2250
2200
2200
2600
4100
4000
3250
3500
3600
3100
2600
PRODUCT WATER PERMEATOR
2314
2366
2482
2528
2717
2810
3031
3051
3114
3370
0.47
0,48
0.45
0,45
0.51
0.58
0.49
0.48
0.50
0.55
0.42
0.41
0.45
0.40
0.50
0.57
0.49
0,48
0.49
0.52
2
5
2
2
1
1
4
2
2
2
5.0
5.5
3.0
4.0
4.0
4.0
3.8
4.0
5.0
3.0
Hardness
Ca C03
mg/1
1400
1740
1440
1640
1200
1300
1300
900
850
1000
920
940
1020
1010
1990
1000
1100
1100
1700
1500
1500
1850
1500
1400
1400
1129
5.0
1.0
1.0
2.0
2.0
2.0
3.0
3.6
2.0
1.0
PH
Units
2.8
3.0
2.8
-
-
2.9
3.1
3.0
2.6
2.7
2.8
2.9
-
-
-
-
-
2.9
2.7
2.9
2.9
_
~
-
-
_
-
3.9
4.3 .
4.4
4.5
-
_
-
-
Meter
TDS
mg/1
4750
4350
4600
4300
3900
4100
3950
3900
3900
3900
3950
3850
3800
3750
3800
4000
3600
4000
7800
9000
4800
-
8500
4950
4900
12
12
11
13
14
12
10
11
12
11
102
-------
TABLE 1-9
Hour
Clock
RDG
Hrs.
1253
1282
1301
1328
1348
1380
1445
1632
1663
1682
1778
1826
1938
1956
2051
2314
2366
2482
2528
2717
2810
3031
3051
3114
3370
2314
2366
2482
2528
2717
2810
3031
3051
3114
3370
Total
Iron
mg/1
0.81
0.77
0.731
0.79
0.70
0.60
0.62
0.41
0.38
0.35
0.36
0.33
0.59
0.46
0.45
0.55
0.61
0.60
1.10
1.34
1.20
1.59
1.61
1.40
1.55
0.56
0.59
0.50
0.52
0.52
0.55
0.50
0.52
0.53
0.49
FIELD
Iron
mg/1
0.76
0.77
0.70
0.79
0.65
0.58
0.52
0.38
0.34
0.32
0.33
0.32
0.54
0.45
0.42
0.53
0.51
0.59
1.10
1.31
1.20
1.60
1.68
1.36
1.52
0.52
0.55
0.50
0.50
0.50
0.55
0.47
0.50
0.51
0.48
ANALYSIS DATA
II S04
mg/1
PRODUCT WATER
3
4
3
5
3
1
1
1
1
1
1
2
2
3
2
4
3
.3
3
8
10
17
19
12
10
PRODUCT WATER
2
3
2
1
I
1
4
3
1
3
HOLLOW FIBER SYSTEM
Total Calcium
Hardness
Ca CO 3
mg/1
PERMEATOR
_
_
_
_
2.5
4.0
2.5
2.0
2.0
10.8
0.7
5.5
5,5
5.5
13.0
13.0
5.0
8.0
10.0
10.0
12.6
14.0
12.0
10.0
PERMEATOR
5.0
5.5
3.0
4.0
4.0
4.0
3.2
3.0
7.0
3.0
Hardness
Ca C03
mg/1
691
_
_
_
_
2.5
2.0
2.5
1.0
1.0
2.0
0.2
4.5
4.5
4.0
1.5
5.0
2.0
3.0
3.0
4.0
5.8
6.8
4.7
4.0
1131
3.0
4.0
1.0
2.0
2.0
2.0
1.2
1.0
2.0
1.0
PH
Units
4.5
4.5
4.4
_
4.6
4.8
4.6
4.8
4.4
4.4
4.5
4.5
-
-
-
-
-
3.6
4.0
4.1
4.3
-
-
-
-
_
-
3.7
4.2
4.4
4.5
-
-
-
-
Meter
TDS
mg/1
18
18
16
17
14
14
14
10
10
11
12
12
11
11
13
17
- 16
17
25
26
25
28
29
28
26
12
11
11
13
15
12
10
11
12
11
103
-------
II. Tubular RO System Data
This section contains the data collected on the tubular RO system.
The following equations were utilized for normalizing the flux data.
For Temperature:
Q68 = QQ * (1.01667(68~T))
where: Qes = Product flow @ 68°F (gpm)
Q = Product flow at observed temperature (gpm)
T = Temperature observed
For Pressure:
Qeoo = Q0 * (12QO/P1 + P2)
where: Qfion = Product flow at 600 psi
Q = Observed flow gpm
o
Pj = Pressure in
P2 = Pressure out
Therefore Normalized Flux is:
(68-T)
F600-68 = Q0(120°/pl + V x (1-01667 ) x (1440/N x 16.9)
where: ^600-68 = Flux at 60° Psi ~ 68°F in Safd
N = Number of modules utilized
NOTE: 1440 converts gpm to gpd & 16.9 sq ft membrane/module
104
-------
TABLE II - 1
TUBULAR RO SYSTEM OPERATIONAL DATA 310 MODULES
Phase 1-60 Modules
Elapsed
Time
(hrs)
23
41
64
80
96
112
125
145
161
174
188
194
207
218
240
255
280
304
328
336
381
407
430
448
470
492
503
517
529
541
546
575
589
620
625
661
1
Pressure
In
psi
600
600
600
605
600
602
600
600
601
600
600
600
610
600
600
600
600
600
600
600
605
620
620
625
630
700
700
700
700
700
700
700
700
700
700
700
Out
psi
410
400
425
420
422
422
432
320
432
389
415
411
430
420
420
400
400
395
355
370
330
339
300
302
305
485
470
470
455
440
375
440
440
432
432
459
Normalized to
Temp
°F
52.7
31.8
51.8
52.5
53.2
52.3
53.8
53.0
52.9
53.0
53.0
52.0
52.9
52.9
53.5
53.0
53.0
53.0
53.0
54.0
53.0
53.5
54.0
54.0
52.5
Phase
52.5
52.5
52.5
52.0
52.5
53.5
53.0
53.0
53.0
53.0
52.9
68°F -
Brine
Rate
sec/gal
31.7
26.8
33.3
31.3
33,2
32.2
35.2
33.4
33.9
27.9
30.8
21.9
33.3
31.9
31.0
30.4
28.3
25.4
23.4
24.5
21.3
20.9
18.8
18.64
18.4
Product
Rate
sec/5 gal
50.5
53.7
53.9
53.4
55.6
57.7
58.8
60.6
60.9
63.6
59.5
57.7
57.0
58.9
59.0
60.2
61.5
64.0
65.6
63.7
70.9
72.0
77.46
77.40
81.50
•9
Flux
gsf d
12.9
12.5
12.1
12.1
11.5
lj.2
10.7
10.5
10.4
10.5
10.9
11.1
11.1
10.9
10.8
10.9
10.7
10.4
10.5
10.4
9.9
9.4
9.2
9.0
8.7
Product
Water
Recovery
%
75.8
71.4
75.5
74.4
74.9
73.6
74,9
73.4
73,6
68,3
72.1
74.0
74.5
73.0
72.4
61.6
69,7
66.5
64.1
65.7
60.0
59.0
55.0
54.6
53.0
II - 35 Modules
32.5
28.8
28.2
26.1
24.8
20.9
24.7
24.3
23.6
23.3
25.3
600 psi
92.7
98.4
104.3
108.1
113.0
116.2
118.5
120.3
127.8
129.2
131.3
10.3
9.8
9.3
9.1
8.8
8.9
8.3
8.2
7.8
7.7
7.4
63.8
59.4
57.5
54.7
52.4
47.4
51.0
50.2
48.0
47.4
49.0
105
-------
TABLE II - 1 (Continued)
TUBULAR RO SYSTEM OPERATIONAL DATA 310 MODULES
Phase II - 35 Modules
Elapsed
Time
(hrs)
673
685
712
734
757
782
803
815
838
849
862
887
935
Pressure
In
psi
705
700
698
700
700
700
700
700
700
700
700
699
702
Out
psi
443
447
400
450
500
445
455
455
462
440
470
438
420
Temp
OF
53.0
53.0
53.0
54.0
53.5
53.5
54.5
54.0
54.0
53.5
54.0
54.8
54.8
Phase III
961
971
983
1153
1173
1196
1221
1245
1279
1293
1365
1387
1392
1539
1549
1573
1597
1622
1645
1668
1690
1718
1740
1
570
570
575
580
580
580
579
575
585
585
602
565
570
570
590
598
604
595
591
591
569
575
580
452
460
462
470
470
470
460
462
470
470
405
385
395
405
430
440
440
380
411
410
400
408
410
Normalized to
53.0
53.0
54.5
54.7
54.0
53.5
54.0
54.0
54.1
55.0
54.0
53.5
54.0
54.0
54.5
54.0
54.0
54.5
54.5
54.0
53.0
54.0
54.0
600 psi
Brine
Rate
sec/gal
23.2
24.1
21.44
24.5
29.2
24.0
25.1
25.0
24.6
23.7
25.3
23.5
22.1
& IV - 15
99.5
98.3
101.6
104.5
103.8
101.2
98.5
100.8
99.2
99.2
58.5
64.1
67.4
63.1
97.4
93.3
88.6
65.6
75.8
75.1
77.2
78.5
76.4
- 68°F
Product
Rate
sec/5 gal
133.5
134.4
142.5 ,
137.9
134.8
141.5
121.5
128.1
128.4
130.0
133.4
133.9
138.7
Modules
(sec /gal)
71.1
71.5
63.8
71.2
69.8
63.2
71.3
70.7
60.8
67.9
67.5
82.5
81.0
44.5
44.8
46.6
49.1
53.2
49.9
53.6
54.9
63.2
68.4
i
Flux
Ssfd
7.2
7.3
7.2
7.0
6.9
6.9
7.8
7.5
7.4
7.5
7.1
7.2
7.0
7.2
7.1
7.7
6.8
7.0
7.8
7.0
7.0
8.0
7.1
7.5
6.6
6.6
11.9
11.2
10.7
10.1
9.9
10.2
9.6
9.9
8.3
7.6
Product
Water
Recovery
%
46.5
47.3
42.9
47.0
52.0
45.9
50.8
49.4
48.9
47.7
48.6
46.7
44.3
58.3
57.9
61.4
59.5
59.8
61.6
58.0
58.8
62.0
59.4
46.4
43.7
45.4
58.6
68.5
66.7
64.3
55.2
60.3
58.4
58.4
58.4
52.7
106
-------
TABLE II -
1 (Continued)
TUBULAR RO SYSTEM OPERATIONAL DATA 310 MODULES
Phase III & IV - 15 Modules (continued)
Elapsed
Time
(Hrs)
1763
1790
1814
1868
1884
1908
1932
1949
Pressure
In Out
psi
592
603
604
600
600
604
610
590
psi
410
412
418
419
420
423
438
450
Temp
°F
54.0
54.0
54.8
54.8
54.0
54.0
54.2
54.0
Rate
sec/gal
72.6
72.0
72.4
72.9
74.6
73.9
78.7
95.7
Product
Rate
sec/gal
69.9
63.8
71.3
50.1
49.7
50.5
55.5
54.5
Flux1
gsfd
7.4
8.0
7.0
10.0
IiM
9.9
8.8
o _ i
Produc1,
Wata,'
Recovery
%
50.-'
53.0
50.3
59,3
60 0
59,4
58 „ 6
6j,7
1 Normalized to 600 psi - 68°F
107
-------
TABLE II - 2
TUBULAR RO OPERATIONAL DATA 610 MODULES
Elapsed Pressure
Time In Out
Hrs.
&
1 r".
41
f, 9
93
113
132
154
178
203
226
239
249
271
307
377
400
424
444
466
515
543
562
586
610
624
671
700
720
746
768
794
797
807
810
E£i
600
600
600
610
620
620
610
610
610
620
620
595
595
600
600
612
642
681
630
612
625
625
625
625
620
620
625
625
615
624
630
625
605
605
615
psi
520
525
530
540
550
550
54C
540
540
550
560
535
540
545
535
560
578
575
575
562
565
565
565
565
565
565
565
570
560
570
570
570
535
535
570
Temp
OF
55.0
54.3
56.0
55.0
55.0
55.0
55.0
55.0
55.0
55.0
54,5
54,5
54.0
52.5
53.5
55.0
55,0
55.0
54,0
54.0
54.0
55.0
54.0
54.0
54.0
54.0
54.5
55.0
54.5
55.0
55.0
54.0
54.0
54.0
54.0
Brine
Rate
sec/gal
63.2
66.4
69.9
71.5
66,5
65.5
67.2
70.2
71.0
69.5
79.2
73.2
69.2
81.2
70,, 3
80.5
68.4
72.6
75.8
63,7
69.5
68,1
72.0
72.4
67.2
73.8
74,1
76.3
72.2
74.0
71.3
70.5
59.0
66.0
66.0
Product
rate
sec/gal
63.5
84.4
84.1
83,3
82.8
84.6
83.6
84.8
88.4
87.3
87.0
89.9
92.3
94.6
94.9
96.7
94.4
95.6
101.4
107*6
103.5
102 « 4
HK.9
105.2
106.8
9f.. '-
>! / M '-J1
102.0
105.2
103,2
102.0
105., 0
.106.8
106.8
76.2
Fluxl
gsfd
21.4
16.8
15.7
15.9
15.7
15.4
15.8
15.6
15.0
14.9
14.9
15.1
14.8
14.6
14.5
13.4
13.2
13.1
12.7
12.2
12.6
12.5
12.4
12.4
12.2
13.5
13.2
12.5
12.4
12.3
12.4
12.3
12.7
12.7
17,1
Product
Water
Recovery
%
49.9
44.9
45.4
46.2
44.5
43.7
44.6
45.3
44.5
44.3
47.7
44.9
42.8
46.2
42.6
45.4
42.0
43.2
42.8
37.2
40.2
39.9
40.7
40.8
38.6
43.3
43.2
42.8
40.7
41.8
41.1
40.2
35.6
38.2
46.4
1 Normalized to 600 psi - 68°F
108
-------
TABLE II - 3
Hour
Clock
Reading
Mrs.
1253
1282
1301
1328
1348
1380
1445
1496
1632
1682
1778
1826
1938
1956
2003
2051
2717
31401
FIELD ANALYSIS DATA TUBULAR
Total
Iron
Total
mg/1
360
300
405
308
540
310
215
172
145
199
112
82
103
109
110
102
230
200
II
mg/1
270
260
305
186
310
270
210
160
130
182
95
88
109
100
102
99
200
160
SO
mg/1
BRINE
1900
3000
2700
1900
2800
2300
1800
1800
_
2500
1180
1360
1440
1560
1400
1240
1500
1450
Hardness
Ca C03
m8/1
ANALYSIS
2300
3400
2380
2180
2720
2400
1800
-
1400
2100
1000
1010
1170
1000
1170
1150
1500
1500
RO SYSTEM
Calcium
Hardness
Ca C03
mg/1
1340
1360
1380
980
1460
1300
800
_
700
1000
530
540
500
500
510
520
750
500
pll
Units
2.9
2.9
2.8
_
2.9
3.1
3.0
2.8
3.0
2.9
3.1
3.1
—
-
-
-
3.1
-
Meter
TDS
mg/1
3750
4350
4400
3400
4100
3950
3400
2425
3000
2250
1950
1950
2000
1900
2100
1950
2900
2000
PRODUCT WATER ANALYSIS
1253
1282
1301
1328
1348
1380
1445
1496
1632
1682
1778
1826
1938
1956
2003
2051
2717
31401
Note
1
1.92
1.92
1.96
1.78
1.94
1.76
1.71
1.59
0.85
0.35
0.89
0.90
1.05
1.00
1.04
1.08
1.39
0.65
: See
1.72
1.82
1.78
1.76
1.76
1.67
1.62
1.48
0.70
0.32
0.83
0.88
0.98
0.99
0.88
1.05
1.35
0.62
Hollow
23
23
23
28
27
21
14
14
5
1
9
12
9
11
13
12
17
2
20
29
31
35
60
18
16
16
5
2
8
16
13
11
13
13
14
6
Fiber Data Appendix
15
19
22
15
20
13
10
7
3
1
6
9
7
8
5
7
7
2
I for Raw
This Data for 610 Modules, all other Data
4.4
4.5
4.3
-
4.3
4.7
4.6
4.4
4.3
4.3
4.6
4.5
—
~
— •
— •
4.3
—
38
39
38
35
41
36
32
33
20
20
22
21
21
21
23
22
38
10
AMD Analysis
for 310
Modules
109
-------
TABLE II - 4
Hour
Clock
Reading
(hrs)
1229
1301
1396
1591
1898
2051
2528
31621
34891
1229
1301
1396
1591
1898
2051
2528
31621
34891
LABORATORY ANALYSIS DATA TUBULAR RO
Ca
mg/1
383
430
384
212
205
198
194
232
168
3.2
4.5
1.9
2.7
2.2
2.2
2.7
0.7
0.5
Mg
mg/1
295
337
310
167
149
147
138
170
132
1.9
2.1
1.9
1.7
1.1
1.2
1.5
0.5
0.4
Mn
Brine
44.8
52.0
48.0
27.6
47.4
25.0
25.3
28.2
21.7
Product Water
0.36
0.39
0.36
0.29
0.24
0.25
0.28
0.08
0.08
Total
Fe
mg/1
238
288
256
123
115
114
129
152
102
1.7
1.8
1.6
1.3
0.9
1.1
1.6
0.5
0.3
UNIT
Si
mg/1
18
19
19
15
15
15
14
20
22
7.0
7.7
6.0
7.0
—
7.1
8.0
1.0
1.0
Al
mg/1
25.5
25.0
27.0
17.0
17.0
16.1
15.0
15.9
14.0
0.2
0.3
0.3
0.2
—
0.2
0.1
0.2
0.1
TDS
mg/1
4224
4753
4307
2299
—
2334
3226
2616
2074
56
49
63
43
—
68
41
33
17
NOTE: See hollow fiber data Appendix I for raw AMD analysis
1 TU.!,, data for 610 modules, all other data for 310 modules
4U.S. GOVERNMENT PRINTING OFFICE: 197Z 484-485/214
110
-------
1
5
6
A cce.s.-, ion Number
2 .Snbjd t Fivld &. Group
05 D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization „ - ,
Ecology Division
Rex Chainbelt Inc.
Milwaukee, Wisconsin
Title
T?F.UraR.c;F DCTMnciTc: T^FMTMITDAT T^ATTDM m? Ar>-rr> HTM-P r>T> A T vr * n T?
10
22
Au thor(s)
Donald G. Mason
Mahendra K. Gupta
1A Project Designation
Project 14010 FQR
2] /Vofe
Citation
WofoT- Prv~l "ln-M nn Hnnt T-nl Re.QOaT'^Vl Sp-r-i PS 1 /, f>1 O W)R ^ /79 TTmr-i T-nnmeTrt-.al
Protection Agency, Washington, D. C.
Descriptors (Starred First)
Acid Mine Drainage*, Reverse Osmosis*, Iron*, Coal*, Neutralization,
Ultraviolet Light
25
Identifiers (Starred First)
Pennsylvania*
27
Abstract
The objective of this study was to determine the operational methods and procedures
necessary to successfully demineralize acid mine drainage utilizing reverse osmosis (RO).
Phase I consisted of laboratory bench scale investigations to determine methods for
controlling iron fouling and to select a process flow sheet. Phase II was field opera-
tion based on the flow sheet selected in Phase I.
The field test site was located in Mocanaqua, Pennsylvania. The source of acid mine
drainage was the discharge from an abandoned underground anthracite coal mine. Treat-
ment prior to RO consisted of filtration (10p) followed by ultraviolet light disinfection.
The brine from the RO unit was treated by neutralization, oxidation and settling. The
field test phase spanned a four month period. Frequent samples were analyzed to
characterize the operation of the system.
The results obtained indicated that it was feasible to demineralize acid mine drainage
by reverse osmosis. Membrane fouling due to iron was satisfactorily controlled. The
recovery of product water was limited to about 75% due to calcium sulfate fouling.
Product water was of potable quality in all respects except for iron, manganese, and pH.
Neutralization, oxidation and filtration would be required to meet potable standards.
Abstractor
Donald G. Mason
Institution _
Ecology Division. Rex Chainbelt Inc.
WR;I02 (REV JULY 1969}
WRS1 C
SEND TO; WATER RESOURCES SCIENTIFIC INFORMA1
U.S DEPARTMENT OF THE INTERIOR
WASHINGTON, D C 20240
* GPO: 196 9- 35 9-339
------- |