Document text
UNCLASSIFIED
PREDICTING PRODUCT HATER DUAL I TV FROM THE
688-GALL0N-PER-H0UR REVERSE OSH CU > COLD REGIONS
RESEARCH AND ENGINEERING LAB HANOVER NH J R BOUZOUN
FEB 88 CRREL-SR-88-2 F/G 24/4
Special Report 88-2
February 1988
AD-A1S4 988
US Army Corps
of Engineers
Cold Regions Research &
Engineering Laboratory
Predicting product water quality from
the 600-gallon-per-hour reverse osmosis
water purification unit
Field water supply on the winter battlefield
John R. Bouzoun
OTIC
ELECTE
JUN 0 31988
H
Prepared for
OFFICE OF THE CHIEF OF ENGINEERS
Approved for public .elease; distribution Is unlimited.
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE
REPORT DOCUMENTATION PAGE
lb RESTRICTIVE MARKINGS
form Approved
OMB No 0704 0188
Exp Date tun 30. 1986
la REPORT SECURITY CLASSIFICATION
Unclassified
2a SECURITY CLASSIFICATION AUTHORITY
2b OECLASSIFICATION / DOWNGRADING SCHEDULE
4 PERFORMING ORGANIZATION REPORT NUMBER(S)
Special Report 88-2
6a NAME OF PERFORMING ORGANIZATION 6b OFFICE SYMBOL
U.S. Army Cold Regions Research (lf aPPhcable)
and Engineering Laboratory CECRL
6c. ADDRESS (City, State, and ZIP Code)
72 Lyme Road
Hanover, New Hampshire 03755-1290
8a. NAME OF FUNDING /SPONSORING
ORGANIZATION
8b OFFICE SYMBOL
(If applicable)
i DISTRIBUTION /AVAILABILITY OF REPORT
Approved for public release; distribution is
unlimited.
5 MONITORING ORGANIZATION REPORT NUMBER(S)
7a NAME OF MONITORING ORGANIZATION
Office of the Chief of Engineers
7b ADDRESS (City, State, and ZIP Code)
Washington, D.C. 20314
9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
8c. ADDRESS (City, State, and ZIP Code)
10. SOURCE OF FUNDING NUMBERS
PROGRAM
ELEMENT NO.
6.27. 30A
PROJECT
TASK
no 4A7627
NO
30AT42
CSS
1 1 . TITLE (Include Security Classification)
Predicting Product Water Quality from the 600-Gallon-Per-Hour Reverse Osmosis Water Purification
Unit. Field Water Supply on the Winter Battlefield. _
12 PERSONAL AUTHOR(S)
Bouzoun, John R. _
14 DATE OF REPORT {Year, Month, Day) 15 PAGE COUNT
February 1988 14
13b TIME COVERED
FROM TO
13a TYPE OF REPORT
17
COSATI CODES |
FIELD
GROUP
SUB-GROUP
- r“
18 SUBJECT TERMS {Continue on reverse if necessary and identify by block number)
Drinking water,, ^ — Water purification /
Reverse osmosis^;
Water purification /
Water quality .«*(
A preliminary equation for predicting the total dissolved solids (TDS) concentration in the product
water from the 600-gph ROWPU is presented. The equation requires the raw water temperature and
TDS concentration as input data. Both of these variables can be easily measured in the field. The equa¬
tion is presently limited to raw water TDS concentrations in the range of 800-900 mg/L. As data be¬
come available for a greater range of raw water TDS concentrations, including seawater, the equation
will be modified. The standard error of the estimate is 3.4 mg/L. ^ ^ , j. %
20 DISTRIBUTION /AVAILABILITY OF ABSTRACT
13 UNCLASSIFIED/UNLIMITED □ SAME AS RPT □ pnc USERS
22a NAME OF RESPONSIBLE INDIVIDUAL
John R. Bouzoun
DO FORM 1473, 84 MAR 83 APR edition may be used until exhausted
All other editions are obsolete
21 ABSTRACT SECURITY CLASSIFICATION
Unclassified _
22b TELEPHONE (Include Area Code) 22 c OFFICE SYMBOL
(603) 646-4100 _ CECRL-EA
il exhausted SECURITY CLASSIFICATION OF THIS PAGE
UNCLASSIFIED
PREFACE
This report was prepared by John Bouzoun, Research Environmental Engineer,
Applied Research Branch, Experimental Engineering Division, U.S. Army Cold
Regions Research and Engineering Laboratory. Funding for this report was pro¬
vided by DA Project 4A762730AT42, Design, Construction and Operations Tech¬
nology for Cold Regions, Mission Area, Combat Service Support, Work Unit 001,
Field Water Supply on the Winter Battlefield.
Dr. C. James Martel and Robert Sletten of CRREL technically reviewed the
manuscript of this report.
The contents of this report are not to be used for advertising or promotional pur¬
poses. Citation of brand names does not constitute an official endorsement or ap¬
proval of the use of such commercial products.
ii
CONTENTS
Page
,1
,1
Abstract . i
Preface . i i
Conversion factors . iv
Introduction . 1
Total dissolved solids flux . 2
Salt concentration equation . 3
The proportionality constant . 3
Product water flow . 5
Overall product water TDS concentration . 5
Conclusions . 6
Literature cited . 7
ILLUSTRATIONS
Figure
1. Principle of reverse osmosis . 2
2. Proportionality constants as a function of raw water temperature . 4
TABLES
Table
1. Water quality and quantity data . 3
2. Salt flux rates . 3
3. Calculated and predicted proportionality constant . 4
4. Measured and predicted TDS concentrations . 6
iii
Accession For
/
NT IS CF.A&I
w
PT1 ' TAB
n
Uu.ituiounced
□
Ju..t : r : o;it i on _
—
By - - _
Plstrlbutl on/
Aval lability t‘ >
■:e9
tAvnll tux'/-
. r
list 1 recoin!
CONVERSION FACTORS: U.S. CUSTOMARY TO
METRIC (SI) UNITS OF MEASUREMENT
These conversion factors include all the significant digits given in the conversion
tables in the ASTM Metric Practice Guide (E 380), which has been approved for use
by the Department of Defense. Converted values should be rounded to have the
same precision as the original (see E380).
Multiply
By
To obtain
degrees Fahrenheit
toc=(t«F-32)/1.8
degrees Celsius
feet
0.3048*
meters
gallons
0.003785412
cubic meters
pounds
0.4535924
kilograms
♦Exact
Predicting Product Water Quality from the 600-Gallon-Per-Hour
Reverse Osmosis Water Purification Unit
Field Water Supply on the Winter Battlefield
JOHN R. BOUZOUN
INTRODUCTION
In March 1974 the Army officially stated its need for a water purification unit that was capable
of purifying most of the raw waters encountered in the field, including seawater. As a result, the
600 gallon-per-hour (gph) Reverse Osmosis Water Purification Unit (ROWPU) was accepted for
use by the Army in June 1979.
The 600-gph ROWPU is a trailer-mounted mobile water purification unit Power for the unit is
supplied by a 30-kW diesel generator that is also mounted on the trailer. The entire unit is
approximately 19 feet long and 8 feet wide and high, and weighs 17,000 pounds. The ROWPU has
three water treatment processes in series that remove the suspended and dissolved solids from the
raw water. They are the multimedia filter, the cartridge filter, and the reverse osmosis vessels.
The multimedia filter consists of six layers of synthetic pellets, anthracite, sand and gravel
contained in a steel tank. A polymer is added to the raw water to cause small suspended particles
to flocculate together into larger particles, which are removed by the multimedia filter. The car¬
tridge filter consists of eight woven polypropylene filters with 5-pm openings. The cartridge fil¬
ter removes the very small particles that are left in the water coming from the multimedia filter.
Then the water passes through the four reverse osmosis pressure vessels, each with two spiral-
wound polyether-urea reverse osmosis elements. Each of these eight elements has 159 ft2 of sur¬
face area, giving a total surface area of 1272 ft2. As the water passes through these elements, the
dissolved solids are removed by a membrane separation process.
The reverse osmosis process is illustrated in Figure 1. Figure la shows a container divided in
half by a semipermeable membrane, with a known volume of clean water in the left half and the
same volume of a salt water solution in the right half. The clean water from the left half will flow
through the membrane into the right half, attempting to equalize the salt concentration, while the
membrane prevents the flow of salt in the opposite direction, as shown in Figure lb. If pressure is
applied to the right side (Fig. lc), the flow of water from the left side can be reduced and even
stopped. The pressure at which the flow is stopped is called the osmotic pressure of the salt
solution. If the pressure is increased above the osmotic pressure of the salt solution (Fig. Id), the
water flows into the left side of the container, leaving the salts in the right half of the container.
This is called reverse osmosis.
The purpose of this report is to present the analytical procedure used to develop a preliminary
equation to predict the total dissolved solids (TDS) concentration in the product water from the
Army’s 600-gph ROWPU. The equation presented here is based on a narrow range of raw water
TDS concentrations (800-900 mg/L) and therefore should be considered preliminary. After
additional data have been collected and analyzed, the equation will be modified. The equation is
simple to use and only requires the raw water temperature and the raw water TDS as input data.
b. Osmosis.
a. Equal volume of clean
and salty water.
c. Osmotic pressure. d. Reverse osmosis.
Figure 1. Principle of reverse osmosis.
Both of these parameters can be measured by the water purification specialist in the field using
the TDS meter and the thermometer that come with the ROWPU.
The final result of this project, after further testing and modification of the equation, will be a
simple table of product water TDS concentrations at different raw water temperatures and TDS
concentrations. This table will be submitted as a suggested change to the 600-gph ROWPU
Technical Manual (U.S. Army 1982) and to the Army Field Manual on Field Water Supply (U.S.
Army 1985). This information will be of value to water supply specialists, commanders and staff
officers, particularly in selecting raw water sources if more than one exists in their area of
operations.
TOTAL DISSOLVED SOLIDS FLUX
The flux of salt (TDS) through a reverse osmosis element is directly proportional to the differ¬
ence between the raw water and the product water salt concentrations. It can be expressed mathe¬
matically by the following equation (Lindsten 1986)
S = E2(Cr-Cp) (1)
where
S = salt flux (g/hr per ft2 of membrane)
Ct = salt concentration in the raw water (mg/L)
Cp = salt concentration in the product water (mg/L)
Ki - proportionality constant (L/hr per ft2 of membrane).
Another equation that expresses S as a function of product water TDS concentration is
*S - QpfCp
(2)
r
where
Q pf is the product water flux (L/hr per ft2 of membrane).
SALT CONCENTRATION EQUATION
Both eq 1 and 2 give the salt flux rate. Because they both equal S,they may be set equal to each
other:
K2(CT-Cp) = QpfCp. (3)
Equation 3 can be solved for Cp to give the following equation:
Cp = K2CT/(Qp{ + K2). (4)
Equation 4 gives the TDS concentration of the product water from the 600-gph ROWPU in terms
of three parameters. The TDS concentration of the raw water, CT, can be measured in the field us¬
ing the TDS meter issued with the ROWPU. The proportionality constant K2 and the product water
flow rate Qp remain to be calculated.
THE PROPORTIONALITY CONSTANT
If eq 1 is solved for K2 the proportionality constant, we get
K2 = S/(Cr-Cp). (5)
This means that we have to know the salt flux rate S, the TDS concentration in the raw water Cr,
and the TDS concentration in the product water Cp. It would also be beneficial if these data were
available at different raw water temperatures so its effect on the proportionality constant could be
determined. Fortunately these data do exist from previous work to determine the effects of raw
water temperature on the production rates of the ROWPU (Bouzoun et al. 1 986). Table 1 gives these
data.
Table 1. Water quality and quantity data.
Table 2. Salt flux rates.
Average
raw
water temp
(°F)
Measured
product
water flow
(gaL / hr)
TDS (mg/L)
Raw Product
water water
Average
raw
water temp
(°F)
Salt flux
(g/fPhr)
33.7
348
800
42
33.7
0.0435
37.9
407
800
43
37.9
0.0521
42.1
430
800
41
42.1
0.0525
46.8
444
900
43
46.8
0.0568
51.5
459
800
47
51.5
0.0642
57.8
600
800
53
57.8
0.0946
68.3
687
900
60
68.3
0.1227
3
These data can be entered into eq 2 to determine the salt flux rate in the product water at the
various raw water temperatures. For example, at 33.7°F, eq 2 would give the following
5 = (42 mg/L) x (3.785 L/gal.) x (g/1000 mg) x (348 gal./hr) x (1/1272 ft2)
= 0.0435 g/ft2 hr.
The 1272 ft2 in this calculation is the area of the reverse osmosis element in the 600-gph ROWPU.
Table 2 gives the calculated salt flux rates at the various raw water temperatures.
Now that the salt flux rates have been calculated, the concentration of salts in the raw water
and in the product water from Table 1 can be used in eq 5 to calculate K 2 at the various raw water
temperatures. For example at 33.7°F, we get
K2 = (0.0435 g/ft2 hr x 103 mg/g) / (800 mg/L - 42 mg/L)
= 0.0574 L/ft2 hr.
Now that we have the proportionality constants at seven raw water temperatures Table 3), we
need to develop an equation that will predict the proportionality constant at any temperature
within the range of the temperatures of the experimental data. This is done by performing a
regression analysis of the proportionality constant as a function of their respective raw water
temperatures. In this case, if we do the regression analysis of the natural logarithms of the
proportionality constants as a function of the raw water temperatures, we get the following equa¬
tion:
K2 = 0.021 6e00281T r = 0.95 (6)
where e is the base of the natural logarithms (2.7183) and T is the raw water temperature (°F).
Figure 2 is a plot of the calculated proportionality constants and the line of best fit as given by
eq 6. Table 3 gives both the calculated and the predicted proportionality constants at the different
raw water temperatures.
Figure 2. Proportionality constants as a
function of raw water temperature.
Table 3. Calculated and predicted pro¬
portionality constants.
Average
raw Proportionality
water temp
(°F)
constant (LHP hr)
Calculated
Predicted
33.7
0.0574
0.0555
37.9
0.0688
0.0625
42.1
0.0692
0.0703
46.8
0.0663
0.0802
51.5
0.0853
0.0915
57.8
0.1266
0.1092
68.3
0.1461
0.1466
In addition to visually comparing the calculated and the predicted values of the proportionality
constants, we can determine how valid the correlation coefficient r is under different conditions.
m
i
i
4
We know that a value of r close to one (1) indicates a high positive correlation between the two sets
of data; a value of r close to zero (0) indicates that the two sets of data are not related. Because our
regression has resulted in an r value of 0.95, we can say that there is a high positive correlation
between the proportionality constant and the raw water temperature. (However, this does not im¬
ply a cause-and-effect relationship between them.) To determine how valid the correlation is be¬
tween raw water temperature and the proportionality constant, we test the null hypothesis, Hq -.t-
0, against the alternative hypothesis, H\ : r * 0. To do this we calculate a t-statistic using the for¬
mula
= Vor 2/(1 -r2)
where v is the degrees of freedom ( n - 2) and n is the number of data pairs. Then
t = 45i0.95?Kl -0.952) = ± 6.803.
The 95% confidence interval for the t-statistic with five degrees of freedom is ± 2.571. Since the /-
statistic w e calculated (± 6.803) falls outside of this range, we reject the null hypothesis that r = 0,
accept the alternative hypothesis that r * 0, and conclude that r is highly significant.
PRODUCT WATER FLOW
In addition to the proportionality constant and the raw water TDS concentration, it is neces¬
sary to know the product water flow rate before we can use eq 5 to predict the TDS concentration in
the product water. Bouzoun et al. (1986) developed an equation to predict product water flow as a
function of raw water temperature:
Qp = 188.3e0019T
where Qp is product water flow (gal./hr) and T is raw water temperature (°F).
Equation 7 gives the product water flow in terms of gallons per hour. To convert the product
water flow to liters per hour, we multiply it by 3.785, the number of liters per gallon. Then, using
the identity Qp f = Q^A, where A is the total area of reverse osmosis membrane in the 600-gph
ROWPU (1272 ft2), we divide eq 7 by 1272 to get
Qpf = 0.560e°°197’.
OVERALL PRODUCT WATER TDS CONCENTRATION
Now we have equations to predict the proportionality constant and the product water flow, both
as functions of the raw water temperature. The only other input parameter required by eq 4 is the
raw water TDS concentration, which can be measured in the field using the TDS meter that
comes with the ROWPU.
If we substitute eq 6 and 8 for their appropriate symbols in eq 4, we get the following equation:
0.0216 eoo28l7’Cr/(0.560 e0019T+ 0.0216 e0028ir)
where
tVir -v<*
Cp = TDS concentration in the product water (mg/L)
Cr = TDS concentration in the raw water (mg/L)
T = raw water temperature (°F)
e = base of the natural logarithms (2.7183).
Now we can use eq 9 and the raw water temperature and TDS concentration data from Table 1
to predict the product water TDS concentration. Table 4 gives the predicted and measured product
water TDS concentrations at the various raw water temperatures.
Table 4. Measured and predicted TDS con
centrations.
Average
raw
water temp
(°F)
Raw
water
TDS cone
(mg/L)
Product water
TDS cone (mg/L)
Measured Predicted
33.7
800
42
39.8
37.9
800
43
41.3
42.1
800
41
42.8
46.8
900
43
50.2
51.5
800
47
46.4
57.8
800
53
49.0
68.3
900
60
60.3
The standard deviation of the predicted product water TDS concentrations (more formally
called the standard error of the estimate) is 3.4 mg/L. Two standard deviations on either side of
the mean encompass 95.4% of the observations in a normal frequency distribution, so we can say
that about 95% of the time the actual TDS concentration of the product water will be within 6.8
mg/L of the predicted concentration. For example, for a raw water temperature of 33.7°F and a
raw water TDS concentration of 800 mg/L, Table 4 shows a predicted product water concentration
of 39.8 mg/L. Adding and subtracting 6.8 mg/L gives a range of 33.0 - 46.6 mg/L. This means
that 95% of the time the actual TDS product water concentration from the ROWPU will be between
33.0 and 46.6 mg/L if the raw water temperature is 33.7°F and the raw water TDS concentration is
800 mg/L.
CONCLUSIONS
The equation developed in this report to predict TDS concentration in the product water from the
600-gph ROWPU is only preliminary. It was developed using a very limited range of raw water
TDS concentrations (800-900 mg/L). Therefore it should not be used for raw water TDS concen¬
trations outside of this range. Future work with the ROWPU will include using raw water with a
high TDS concentration (such as seawater and brackish water) at different temperatures to
determine product water flow rates and proportionality factors.
Once an equation is developed to accurately predict the TDS in the product water from the
ROWPU over a broad range of raw water TDS concentrations and temperatures, studies will be
conducted to determine the feasibility of developing similar equations for other dissolved chemi-
cals in the raw water. In this case particular attention will be paid to those chemicals that are in¬
cluded in the quality standards for treated water, such as magnesium and sulfate.
LITERATURE CITED
Bouzoun, JJt., S.C. Reed and C J. Diener (1986) An initial assessment of the 600-gallon-per-hour
reverse osmosis water purification unit; Field water supply on the winter battlefield. USA Cold
Regions Research and Engineering Laboratory, Special Report 86-20.
Lindsten, D. (1986) Development of US Army reverse osmosis water purification equipment.
USA Belvoir Research and Development Center, Report 2418.
U.S. Army (1982) Operator’s Manual, Water Purification Unit, Reverse Osmosis, 600 gph. Tech¬
nical Manual 5-4610-215-10.
U.S. Army (1985) Field Water Supply. Field Manual 10-52.
-i U.S GOVERNMENT PRINTING OFFICE. 1988 - 500 050/82004