PB85-121044
DEMONSTRATION OF A MAXIMUM RECYCLE, SIDESTREAM SOFTENING
SYSTEM AT A PETROCHEMICAL PLANT AND A PETROLEUM REFINERY
The University of Houston
Houston, TX
Oct 84
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
-------
EPA-600/2-84-176
October 1984
DEMONSTRATION OF A MAXIMUM RECYCLE, SIDESTREAM SOFTENING SYSTEM
AT A PETROCHEMICAL PLANT AND A PETROLEUM REFINERY
by
Jack V. Matson
Wendy Gardiner Mouche
Eric Rosenblum
Larry McGaughey
Environmental Engineering Program
Civil Engineering Department
The University of Houston
Houston, Texas 77004
Cooperative Agreement
CR-807419
Project Officer
Donald Kampbell
Robert S. Kerr Environmental Research Lab
Environmental Protection Agency
Ada, Oklahoma 74820
ROBERT S. KERR RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
i.
4.
7.
9.
REPORT NO.
EPA-600/2-84-176
2.
TITLE AMD SUBTITLE
Demonstration of a Maximum Recycle, Sidestream Soften-
ing System at a Petrochemical Plant and a Petroleum
Refinery
AUTHORIS)
J. V. Matson, W. G. Mouche
, E. Rosenblum, L. McGaughey
PERFORMING ORGANIZATION NAME AND ADDRESS
Civil Engineering Def>artment
University of Houston
Houston, TX 77004
12. SPONSORING AGENCY NAME AND ADDRESS
R. S. Kerr Environmental Research Laboratory
P. 0. Box 1198
Ada, OK 74820
15
16
17.
a.
13.
. SUPPLEMENTARY NOTES
•
3. RECIPIENT'S ACCESSION'NO.
PB35 121044
5. REPORT DATE
October 1984
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
CBGB1C
il.S«!X3W&SOX8R«!!tX»S. Coop. Agr.
CR807419
13. TYPE OF REPORT AND PERIOD COVERED
Final 5/80 - 11/83
14. SPONSORING AGENCY CODE
EPA/ 600/15
. ABSTRACT
New full-scale maximum recycle sidestream softening systems at USS Chemicals,
Houston, Texas and TOSCO refinery, Bakersfield, California were evaluated as a
technology to achieve vzero wastewater discharge.. Softener process efficiency
was optimum at a pH control range of 10.3 to 10. 5 "at 40 C and using a high mixing
intensity. A problem of heat exchanger biofouling from the high dissolved organ-
ics in recycle water was effectively controlled by using Bromocide with chlorine.
A total organic carbon balance over the cooling water system showed raw makeup
water and process water contribute 1/3 and 2/3- of the organics, respectively.
Major organic sinks were drift (60%), biodegradation (30%), and volatilization
(10%) . Softener sludge as analyzed for chromium by leachate tests was classi-
fied as nontoxic. Heat exchanger equipment. averaged . two mils /year internal
corrosion. External corrosion from drift aerosols was corrected by installation
of a ferrous sulfate reactor in the blow down system and improved drift elimi-
nators in cooling towers. The TOSCO water problem of high silica and low mag-
nesium was corrected by adding caustic and magnesium sulfate to the softener.
Both plants operated satisfactorily at near zero liquid discharge. Operating
costs and benefits are discussed.
DESCRIPTORS
Water treatment
Circulation
Petroleum refining
Industrial plants
Effluents
DISTRIBUTION STATEMENT
Release to public
KEY WORDS AND DOCUMENT ANALYSIS
b. IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Sidestream softening 13B
Recycle streams
Scale inhibitor
Biofouling
Organic sinks
Corrosion
19. SECURITY CLASS (This Report} 21. NO. OF PAGES
Unclassified 219
20. SECURITY CLASS (This page) 22. PRICE
Unclassified
.EPA Form 2220-1 (9-73)
-------
DISCLAIMER
Although the research described in this article has been funded wholly
or in part by the United States Environmental Protection Agency under assistance
agreement CR807419 to the University of Houston, it has not been subjected
to the Agency's peer and administrative review and therefore may not
necessarily reflect the views of the Agency and no official endorsement
should be inferred.
ii
-------
FOREWORD
EPA is charged by Congress to protect the Nation's land, air, and water
systems. Under a mandate of national environmental laws focused on air and
water quality, solid waste management and the control of toxic substances,
pesticides, noise, and radiation, the Agency strives to formulate and imple-
ment actions which lead to a compatible balance between human activities and
the ability of natural systems to support and nurture life.
The Robert S. Kerr Environmental Research Laboratory is the Agency's
center of expertise for investigation of the soil and subsurface environment.
Personnel at the Laboratory are responsible for management of research pro-
grams to: (a) determine the fate, transport and transformation rates of
pollutants in the soil, the unsaturated zone and the saturated zones of the
subsurface environment; (b) define the processes to be used in characterizing
the soil and subsurface environment as a receptor of pollutants; (c) develop
techniques for predicting the effect of pollutants on ground water, soil and
indigenous organisms; and (d) define and demonstrate the applicability and
limitations of using natural processes, indigenous to the soil and subsurface
environment, for the protection of this resource.
Zero discharge technology is the final goal for development and imple-
mentation of reuse/recycle water systems at industrial plants. Quality data
is needed to support regulation, enhance technology transfer, and encourage
acceptance of such feasible systems. The research report presents results of
a recycling cooling water system study for two different industrial processes.
Topics covered are (1) performance optimization, (2) impact of recycle streams,
(3) fate of organic contaminants, and (4) sludge toxicity.
Clinton W. Hall
Director
Robert S. Kerr Environmental
Research Laboratory .
iii
-------
ABSTRACT
The purpose of this project was Co document the performance of a maximum
recycle, sidestream softening system at the USS Chemicals Petrochemical Plant
in Houston, Texas. The concept was to reuse all effluent streams as makeup
to the cooling water system, with one exception—the demineralizer regenera-
tion water. A blowdown from the cooling water system was processed through
two precipitation softeners to remove the potential scale-forming constit-
uents and returned. Thus, the plant approached zero liquid discharge.
The system was started up in November 1979, concurrent with the demoth-
balling of the ethylene unit. The styrene unit had been in continuous opera-
tion. In May 1980, the research effort was commenced. There were many
problems with the new, innovative system. Consequently, much of the effort
involved direct interaction in improving the system.
The most significant efforts involved the control of biofilm formation
in an extremely high organic concentration cooling water; the enhanced per-
formance of the precipitation softeners; the evaluation of the toxicity of
the softener sludge (and its subsequent delisting as a toxic material); and
the delineation of an emergency blowdown system.
There are many things yet to be learned in this very applied area of
research. Hopefully, this document can be used as an example of the success-
ful application of prec'ipi tat ion softening to maximize water recycle; and
"'•at many of the advances made can be applied to new system.
-------
CONTENTS
Abstract ................... .......... iv
Figures ................. . ........... vii
Tables .............................. xi
Acknowledgement ...... . ........ . ......... xiii
Introduction . . < ^_ ....................... 1
General Conclusions ....................... 3
1. The Development of Zero Discharge Sidestream Softening . 6
Slowdown and Zero Discharge Systems ........ 6
Early Sidestream Softening Research ........ 7
Recent Developments ....... ......... 15
2. Sidestream Softener Design . . ...... ." ...... 24
Cooling Water Quality ...... ......... 27
Sidestream Flow Rate ...... . ......... 31
Softening Agents, pH Adjustment, and Additional
Treatment .................... 34
.3. Sidestream Softener Operation at USS Chemicals ..... 37
Operational Data . ..... . ......... ... 37
4. Sidestream Softener Performance . • ..... . ...... 68
Alkalinity Determination and Softener Efficiency . . 68
Removal of Silica by Sidestream Softening ..... 35
Effect of Mixing Speed on the Softening Process . .
5. Monitoring and Control of Biofouling in a Zero Discharge
Sidestream Softened Cooling System — USS Chemicals . . 120
6. Treatment of Chromate in Softener Sludge and Cooling
Tower Slowdown ..... ....... . ....... ^37
Evaluation and Handling of Chromate Leachate from
Softener Sludge .......... . ...... 137
Chromate Removal from Cooling Tower Slowdown .... 149
7. Total Organic Carbon Mass Balance ............ 166
8. Costs ...... ..... ............... 170
-------
9. Zero Discharge Sidestream Softening at TOSCO 175
Introduction 175
General Plant Description 176
Sidestream Softener Process Design, TOSCO Refinery .... 177
Preliminary Process Design 177
Mass Balance Equations 178
Chemical Dosage Calculations 181
Preliminary Design Results 182
Final Process Design 187
Performance of TOSCO Zero Slowdown System 191
Comparison of Predicted and Actual Performance 192
Actual Versus Design Operating Conditions 192
Actual Versus Estimated Water Quality Data 194
General Assessment of Sidestream Softener System 197
Recommendations and Conclusions 199
-------
FIGURES
Number Page
1-1 Typical induced draft open recirculating cooling tower 6
1-2 Sidestream softening system with optional demineralization .... 9
1-3 Sidestream softened water treatment plant (Rice) 14.
1-4 Pilot cooling tower with zero blowdown capability (Reed) ..... 14
1-5 Georgia-Pacific sidestream softening ... 17
1-6 View of the USS Chemicals ethylene and styrene petrochemical 19
plant
1-7 Side view of the two sidestream softeners ............ ^9
1-8 View of softeners and soda ash delivery system ..... 20
1-9 Control room and filters 20.
1-10 Side view of recarbonator 21
1-11 Top view of recarbonator 21
2-1 Schematic diagram of a sidestream softening system ........ 30.
3-1 Schematic diagram of USS Chemicals sidestream softening system . . 33
3-2 Flow diagram for water recycle system (USS Chemicals) 39
3-3 USS Chemicals softener effluent pH, 1980-81 43.
3-4 USS Chemicals softener effluent calcium hardness, 1980-81 .... 44-
3-5 USS Chemicals softener effluent alkalinity, 1980-81 45,
3-6 Ethylene and styrene unit cooling water pH, 1980-81 ....... 45
3-7 Ethylene and styrene unit cooling water calcium hardness, 1980- 47
81 ;
4a
3-8 Ethylene and styrene unit cooling water alkalinity, 1980-81 . . .
vii
-------
3-9 USS Chemicals recarbonator effluent pH, 1980-81 ......... 49
3-10 USS Chemicals recarbonator effluent calc-turn hardness, 1980-81 . . 50
3-11 USS Chemicals recarbonator effluent alkalinity, 1980-81 ..... 51
3-12 USS Chemicals softener ;inf lueiat_-and. effluent magnesium hard-
ness, 1981 ........................... 57
3-13 USS Chemicals softener influent and effluent silica, 1981 ... 58
3-14 Results of dye trace study of softener hydrodynamics ...... 63
3-15 Results of dye tracer study of recarbonator hydrodynamics. ... 64
4-1 Comparison of TOC and 1C in USS Chemicals cooling water before
and after softening ....................... 73
4-2 Comparison of various cooling water alkalinity measurements. . . 74
4-3 Efrect of cooling water alkalintiy measurement on soda ash
dosage ............................. 75
4-4 Effect of soda ash dosage on removal of calcium from cooling
water . . ........................... 77
4-5 Comparison of various softener effluent alkalintiy measurements- 78
4-6 Effect of softener effluent alkalinity measurement on soda ash
dosage ............................. 79
4-7 Effect of soda ash dosage on additional removal of calcium and
softener effluent ....................... _§0
4-8 Effect of soda ash dosage on cooling water calcium removal and
[Ca]/[C03] ratio ............. . .......... 81
4-9 Relation between [Ca]/[CO,] ratio and calcium removal from USS
Chemicals cooling water .................... 82
4-10 Relation between softening reaction pH and final [Ca]/[CO_] . . 83
4-11 Effect of lime added on soda ash dosage and calcium removal. . . 87
4-12 Effect of percent calcium removal on softener flow rate and
treatment cost ......................... 88
4-13 Effect of softener pH on silica removal ............. 91
4-14 Effect of softener sludge TSS on silica removal ........ 96
4-15 Effect of silica/magnesium ratio on effluent silica* ....... 97
viii
-------
4-16 Equilibrium isotherm for adsorption of silica onto magnesium
hydroxide floe (jar test) 101
4-17 Linearized Freudlich isotherm for silica adsorption (jar test). . . 102
4-18 Linearized Knight isotherm for silica adsorption (jar test) . . . 103
4-19 Equilibrium isotherm for silica adsorption (south softener) .... 104
4-20 Linearized Freudlich isotherm for silica adsorption (south softener) 105
4-21 Linearized Knight isotherm for silica adsorption (south softener) . 106
4-22 Silica adsorption isotherms obtained at various mixing speeds . . . 107
4-23 Schematic of sampling points in the north softener 108
4-24 Effect of mixing speed on calcium carbonate solubility 110
4-25 Effect of mixing speed on line dosage Ill
4-26 Effect of mixing speed on softener effluent turbidity 113
4-27 Effect of mixing speed on sludge recycle rates 114
4-28 Effect of mixing speed on calcium removal 116
4-29 Effect of mixing speed on softener sludge settling 117
5-1 Apparatus for monitoring biofouling in the cooling system , , . . 122
5-2 Detail of biofouling monitor apparatus showing aluminum heat
exchanger ........ ............. 123
5-3 Process diagram of biofouling monitor system .... 124
5-4 Variations in surface condensor vacuum of the biofouling moni-
tor apparatus 130
5-5 Variations in cooling water free halogen residual ........ 131
5-6 Variations in friction and heat transfer resistance ....... 133
5-7 Variations in cooling water TOG ....... ..... 134
5-8 Variations in viable microbial cell count in USS Chemicals
cooling water ............... 135
-------
6-1 Process diagram for leached chromate adsorption experiment. . . . 143
6-2 Equilibration of Texas Toxicity Test 144-
6-3 Solubility of Cr(OH)_ as a function of solution pH 145-
6-4 Equilibration of EP Toxicity Test 150
6-5 Isotherm for re-adsorption of leached chromate 151
6-6 Linearized Freundlich isotherm for chromate re-adsorption .... 152
6-7 Cr(VI) contained in chromium adsorbed onto softener sludge. . . . 153
6-8 Process schematic for reduction and removal of chromate from
cooling tower blowdown 154
6-9 Effect of ferrous sulfate dosage on chromate removal 158
6-10 Effect of reduction pH on chromate removal from deionized water . 160
6-11 Effect of reduction pH on chromate removal from USS Chemicals
cooling water 161
6-12 Effect of precipitation reaction pH on chromate removal from
USS Chemicals cooling water 162
6-13 Effect of reduction reaction pH on chromate removal at differ-
ent settling times 163
9-1 Sidestream Softening Schematic 201
9-2 Preliminary Design Sidestream Softening".~Y ." 202
9-3 Sidestream Softening System Schematic Diagram ... . 203
-------
TABLES
Number Page
1-1 Oak Ridge Water Quality Data . ......... . ........ 8
1-2 Feasibility Tests of a Sidestream Softener for Oak Ridge . . . , . ^Q
1-3 Recirculating Water Quality in a Pilot Plant Using C02 for pH
Control ...... . . . ...... ..............
1-4 Recirculating Water Quality at Georgia-Pacific Plant
1-5 Actual Versus Design Recirculating Water Quality Parameters at
Southern California Edison Coolwater Plant
3-1 Average Water System Sample Analysis ........ ....... ^
3-2 Cooling and Softener System Analysis Averages (1980-81) .....
-------
5-1 Output of Biofouling Monitor System . . . 125
5-2 Abbreviations and Symbols Used in Biofouling Experiment 126
6-1 Comparison of Leachate Extraction Procedures 140
6-2 Texas Toxicity Test Leachate Values , 142
6-3 EP Toxicity Test Leachate Values 147
6-4 Comparison of Chromium Concentration in Softener Sludge as
Determined by EP and Texas Toxicity Tests 148
7-1 TOG Results, WWTU 168
7-2 TOG Results, Makeup 169
8-1 Chemical Cost for Lime Softening System 171
8-2 Chemical Cost Per Unit Gallons Water Treated ........... . 172
8-3 Chemical Usage For Cooling Water System . . . 173
8-4 Chemical Cost Per Unit Gallons Recirculation Water ....... . 174
9-1 Water Quality Data Base 179
9-2 Softening Reactions 183
9-3 Projected Water Balance , , , 183
9-4 Values Assumed_for Preliminary Design Calculations . 184
9-5 Projected Water Quality Data , 186
9-6 Design Loading Factors ..,.....,,..,,.,.. 186
9-7 Comparison of Preliminary and Final Flow Data ........... . , 189
9-8 Comparison of Preliminary and Final Water Quality Estimates , . . 189
9-9 Comparison of Preliminary and Final Chemical Consumption 190
9-10 Comparison of Preliminary and Final Equipment Sizing .,..,,. 190
9-11 Comparison of Design Estimate and Actual Operating Data , , » , , 193
9-12 Comparison of Estimated and Actual Water Quality Data .. , , . . , 196
9-13 Estimated Annual Cost Data , , , . , 196
xii
-------
ACKNOWLEDGEMENT
As this final report represents the cumulative efforts of many people
over the past two years, the authors would like to acknowledge the following
researchers whose reports were incorporated into the text, and whose work
forms the substance of the present manuscript:
Section 1
Matson, Jack V. and Puckorius, Paul. "Status of Sidestream Soft-
ening," £ojjrjiaj^_cj:_the_j::oc^ 2(1), pp. 21-30.
Matson, J.V. "Cooling Water Recycle by Softening," Annual Report to
National Science Foundation (November 1977).
Section 2
Matson, Jack V. and Harris, Teague G. "Zero Discharge of Cooling
Water by Sidestreara Softening," Jouranl WPCF, 51(11), p. 2602
(1979).
Matson, Jack V.; Gardiner, Wendy M.; Harris, T.G.; Puckorius, P.R.
"Zero Discharge in Cooling Towers," Proceedings 4th Annual Confer-
ence, Industrial Energy Conservation Technology, 1980.
Section 3
Gardiner, Wendy and Matson, Jack V. "Demonstration of a Maximum
Recycle Sidestream Softening System at a Petrochemical Plant," First
Annual Progress Report, EPA Grant #CR 807419-01, August 11, 1981.
Velez, Fernando; Matson, Jack V.; and Amador-Pena, E. "Calculation
of Carbon Dioxide Stripped Out in a Sidestream Softening System."
Project Report to USS Chemicals, April 27, 1982.
Section 4
Alvarez, Hector R. "Calcium Carbonate Softener Efficiency Study,"
Seventh Quarterly Report to Environmental Protection Agency, Summer
1982.
Kill
-------
INTRODUCTION
Our national goal, as promulgated by Congress, is to eliminate pollution
from industrial facilities within the decade of the 1980s [1]. At first
glance, this "zero pollution" goal appears to be an expensive, hypothetical
ideal difficult to obtain. In reality, however, the technology to attain
this goal is already being developed and can be implemented.
At the forefront among techniques to eliminate wastewater pollution is
sidestream softening. Since cooling tower blowdown accounts for "the single
greatest source by volume of all types of industrial water discharges (over 1
trillion gallons per year)[2], application of sidestream softening to achieve
zero discharge constitutes a major reduction in industrial water pollution.
Sidestream softening treats a portion of the cooling water in a lime
softener, then returns the treated water to the cooling system. The softener
precipitates scale-forming materials from the cooling water as a sludge which
can be dewatered and landfilled. This allows cooling water to be recircu-
lated indefinitely without the normally necessary effluent discharge. Other
plant waste streams may also be added to the softener sidestream, resulting
in further reduction of aqueous discharge, Thus it is quite possible to
attain a true zero discharge operation with sidestream softening.
In addition, the chemicals involved in the softening reaction (lime, and
soda ash) are relatively inexpensive and the theory of softening is well
understood. Because of these benefits, at least 20 plants in the United
States currently use sidestream softening to attain zero discharge.
The incentives to industry for the construction and operation of a zero
blowdown or discharge sidestream softening system have been both regulatory
and economic in nature [3, 4]. Because of the EPA effluent limitations for
chromium and zinc, operational requirements are either blowdown treatment or
a switch to alternate corrosion inhibitors. Since chromate is widely recog-
nized as the best corrosion inhibitor for cooling water systems, zero
discharge sidestream softening is often the method of choice. It allows for
the continued use of chromate in higher concentrations .than permitted within
discharge limitations. Depending upon the costs of water, pretreatment, and
wastewater treatment .in a given industry, sidestream softening could provide
operating cost savings which yield a capital payback period of about five
years. In some arid regions of the country, this type of water reuse may
emerge as the only economical alternative.
Although the concept is simple, a cost-effective process design and
cooling water treatment program requires meticulous selection. High levels
of total dissolved solids (TDS) from increased cycles of concentration sig-
-------
nificantly affect Che chemistry of both the softener and the cooling water.
Also, interactions between chemicals added to the cooling and softening
systems can be detrimental to the proper function of each.
The practical design information obtained is documented as
follows:
1. Development of zero discharge sidestream softening technology;
2. Basic criteria used to determine the suitability of various side-
stream softening techniques;
3. Example of design application in a "zero" or minimum aqueous dis-
charge system at the USS Chemicals installation in Houston, Texas;
4. Improvements of the system based on experimental results during its
start-up and operation.
Section 1 of the report contains a history of the development of zero-
discharge sidesream softening, a description of the USS Chemicals plant, and
the sidestream softening system. Section 2 illustrates the basis for side-
stream softening design, and provides the background for the case studies
which make up the rest of the report.
Section 3 documents the zero-discharge sidestream softened cooling tower
system at USS Chemicals. Raw data assembled in tabular form in the Appendix
can be used as a basis for further research. Section 4 shows the results of
experiments concerning specific aspects of the USS Chemicals sidestream soft-
ening operation, designed to optimize the softening process.
Section 5 describes experiments with the biocides used to control bio-
logical fouling in the heat exchanger system. 'The need to' blowdown period-
ically is addressed in Section 6, with in the areas of chromate precipita-
tion; and the landfill disposal of chromate-bearing sludge. Section 7 deals
with accounting for the organic material entering and leaving the cooling
water system. Section 8 gives an evaluation of the costs of the zero-
discharge sidestream softening system.
-------
GENERAL CONCLUSIONS
The sidestream softening concept was particularly attractive to USS
Chemicals because the cost of their water was relatively high, the efflent
discharge regulations were tough, and they wanted to continue to use chrome
and zinc as corrosion inhibitors in the cooling water system. Prior to the
installation of a sidestream softening, USS Chemicals had only a rudimentary
(physical, chemical solids removal) wastewater treatment system.
The decision to minimize blowdown was made in the mid-1970^s. The
economics of sidestream softening were the most advantageous, especially when
viewed as a preinvestraent to facilitate compliance with, future regulations.
The fact that the plant approaches zero discharge would minimize expenses in
the future.
The major initial problems encountered at USS Chemicals were the microbio-
logical fouling in the cooling water system and high sludge disposal costs
because of the original classification as toxic. Both problems were resolved
satisfactorily during the course of this research.
The major problem that was unresolved during the study phase of the
project was drift loss from the cooling towers. The drift with the concen-
trated dissolved solids deposited on equipment and pipelines with a 50-foot
radius of the towers and created an external corrosion problem. At one
point, solids from the drift accumulated on a transformer and shorted it out.
Plant personnel then started using an improved coating on the equipment .to
mitigate the problem and the transformer is cleaned every six months.
Finally, in June 1981, USS Chemicals rebuilt the largest ethylene cool-
ing tower and installed high efficiency drift eliminators. As expected, the
TDS concentration in the cooling water system increased (roughly 50 percent).
In September 1982, the styrene plant was mothballed due to poor economic
conditions. This move further increased TDS levels in the ethylene cooling
tower, which approached 30,000 mg/L. In December 1982, USS Chemicals
installed the chrome destruction blowdown system researched in Section 6.
They are now blowing down at low flow (5 gpm) to decrease the TDS concentra-
tions in the cooling water to more reasonable levels.
The project demonstrated that a- cooling water system could be used .as a
sink for wastewaters, and that a softening system installed as a sidestream
could control scaling. To the many industrial plants which currently employ
cooling water systems as a basic utility, sidestream softening offers an
opportunity to maximize reuse of industrial cooling water.
•
In the TOSCO case the design, startup, and operation was smooth from the
-------
beginning. Microbiological problems were minimized because no recycle streams
with organics were introduced into the cooling water system.
TOSCO was unique in the use of caustic instead of lime in the softener;
and in the use of magnesium sulfate to remove silica from the cooling water.
Also, dissolved solids were in the low range, and no external corrosion due to
drift deposition was observed.
As a footnote, the TOSCO refinery was shut down in early 1984 for
economic reasons associated with the price of gasoline.
Recommendations for Further Study
This initial attempt to demonstrate the technology of a minimum blowdown
system also served to improve sidestream softening process performance.
Nevertheless, as with most scientific efforts, the research only serves to
multiply the questions. In some order of priority, the. following tasks need
to be performed:
1. Development of a design methodology that can take into account the
various process options;
2. Determination of the impact of the high dissolved solid level on
calcium carbonate, magnesium hydroxide, and. silica adsorption in the
softening reaction.
3. Investigations into the best types and optimum concentrations of
scale inhibitors in sidestram softening systems.
4. In-depth economic analyses of sidestream softening systems.
5. Evaluation of the various softening options.
In addition, much work could be done to quantify the environmental
effects of sidestream softening as a minimum discharge strategy.
-------
REFERENCES - INTRODUCTION
1. Codified in scattered sections of 12, 15, 31, 33 U.S.C. (1972) cited in
Environmental Quality 1981, 12th Annual Report of the Council on Environ-
mental Quality, U.S. Government Printing office.
2. Matson, J.V. and Harris, T.G., "Zero Discharge of Cooling Water by Side-
stream Softening," Journal WPCF, 51(11), pp. 2602-2614 (November,
1979).
3. Curtis, M. "Economic Attractiveness of Sidestream Softening," UH/NSF
Workshop on Zero Discharge of Cooling Water by Sides tream Softening,
University of Houston, Houston, Texas, June 1-2, 1981.
4. Lihach, N. "Coping with Zero Discharge," EPRI Journal (June, 1981).
-------
SECTION 1
THE DEVELOPMENT OF ZERO DISCHARGE SIDESTREAM SOFTENING
A common cooling water system consists of a cooling tower, conveyance
system, and heat exchangers. Heat is transferred through the metal tubes of
the heat exchanger into the water. The water is cooled in the cooling tower,
then returned to the exchangers. Heat exchange requirements determine the
makeup water flow rate.
SLOWDOWN AND ZERO-DISCHARGE SYSTEMS
A major problem in recirculating systems is that evaporation concen-
trates dissolved solids in the cooling water to levels at which scaling
occurs on heat exchange surfaces. Therefore, a purge stream (blowdown) is
continuously maintained to limit the maximum concentrations of certain dis-
solved species in the system. A cooling water system schematic is shown in
Figure 1-1. Source or "makeup" water is pumped into the cooling tower basin
to replenish losses from evaporation, blowdown, and drift.
Evaporation, Drift
f
Makeup
Process Heat
Exchange
Figure 1. Typical induced-draft open recirculating
cooling tower.
In addition, a variety of chemicals may be added to the cooling water to
prevent scaling and corrosion. Chromium (in hexavalent form) and zinc
inhibit corrosion. Polyphosphates and phosphonates and other microbiocides
minimize microbial fouling. When used, these chemical additives are also
discharged in the blowdown.
Some recirculating systems are closed to the atmosphere: cooling is by
-------
sensible heat transfer alone, through air fins analogous to an automobile's
radiator. Other cooling water systems, such as the existing once-through
systems in which no cooling tower is necessary. Many once-through systems
are now required by federal law to convert to recirculating systems in order
to reduce thermal discharges.
Some cooling water discharges from recirculating cooling towers contain
chrome and zinc, which can be toxic to aquatic life. Effluent standards
currently call for concentrations of hexavalent chrome in the range of 0.02
mg/L, and zinc below 0.5 mg/L. The phosphorus in blowdown is a potential
eutrophication agent. Constituents of chlorine and microbiocides may also be
toxic to the-environment. Heat and high dissolved solids levels enhance
additional environmental problems.
The zero-discharge recirculating cooling system solves the problem of
toxic discharge by continuously recirculating the cooling water with only
minimal discharge. The zero discharge system removes deleterious materials
by softening, demineralization, or some other form of advanced treatment. A
technology actively under consideration (including ion exchange, reverse
osmosis, and electrodialysis), lime/soda softening of a cooling water side-
stream is' economically attractive and the most highly developed.
The sidestream softening process (Figure 1-2) removes the principal
scale-forming agents of calcium, magnesium, and silica. Lime is added to
raise the pH to 10.5 - 11.0 where ionic bicarbonate is converted to ionic
carbonate, which reacts with ionic calcium to precipitate out as calcium
bicarbonate. Dissolved silica is removed by adsorption onto magnesium
hydroxide, which precipiates as floe and is removed in the softener settler.
The treated water is then clarified, filtered, and its pH is adjusted for
return to the cooling tower.
EARLY SIDESTREAM SOFTENING RESEARCH
The concept of integrating a softening unit into a cooling water system
to achieve zero discharge was originated by Fowlkes in the late 1960s at
Union Carbide's Oak Ridge Gaseous Diffusion Plant (Oak Ridge, Tennessee) [1],
Proposed regulations for effluent to the Clinch river severely restricted
chromate discharges. The corporation did not want to discontinue the use of
superior chromate-bearing corrosion inhibitors, or to treat the blowdown
prior to discharge.
Originally, blowdown was to be blended into two existing makeup water
softeners. The blowdown water could be successfully softened by laboratory
jar tests. A one-year operational test was conducted in 1969, then continued
for another 1 1/2 years. Corrosion rates were monitored with bare steel and
copper coupons. Rates were below 1 mpy (mill per year). They were compar-
able to previous rates in the system.
The Oak Ridge cooling water system processed about 12 MGD makeup water.
Fifteen percent of this makeup water flow was blown down and recycled to the
makeup water softeners. A bleed to the firewater system reduced the TDS
-------
TABLE 1-1. Oak Ridge Water Quality Data (1, 2, 3)
Parameter Makeup Operating Test Loop
Calcium Hardness as CaC03 72 179 959
Magnesium Hardness as CaC03 38 161 . 519
Sulfate as S04 62 487 2345
Dissolved Solids 170 1106 4869
Chlorides as Cl 59 647
Silica as Si02 23 111
-------
Hot
Reclrculating
Sidestream
Heat Exchange
Equipment
Lime or Caustic
Reuse Streams
Sidestream Return
Reuse Streams
Soda Ash
Magnesium
Reuse Streams
Cooling Towers
Acid or COo
pH Adjustment Filters
Sludge Tank
Disposal
Softener -Clarifler
Optional
D
Issolved So
ids
Removal by
Reverse Osmosis
I or |
Electrodyalysis
Figure 1-2. Sidestream softening system with optional demineralization
-------
TABLE 1-2. Feasibility Tests of a Sidestream Softener for Oak Ridge (2)
Parameter
pH
"M" Alkalinity (as CaC03) ppm
Total Hardness (as CaCC^) ppm
Calcium (as CaCO^) ppm
Sulfate (as SO^) ppm
Dissolved Solids ppm
Silica (as SiC^) ppm
Softening
6.55
7
566
364
737
1440
36
After 104°F (60°(
Softening
9.70
76
80
22
725
1474
9.5
Note: Feed rates: 150 ppm lime and 640 ppm soda ash
in
-------
level from a calculated 4000 mg/L to actually 1100 tng/L. Fowlkes set up a
test loop without a firewater bleed [2] to measure corrosion rates for the
true zero discharge situation. He found that corrosion rates were higher,
but tolerable. The qualities of the makeup water, actual cooling water, and
test loop water are shown in Table 1-1. Fowlkes also tested the concept of a
separate softener for blowdown treatment. This took advantage of a hot water
temperature and the high concentration gradients of the reactants [3]. The
tests results were encouraging (Table 1-2). The separate softener was
installed in 1974.
The Oak Ridge facility was still operational in June 1978 in the redun-
dant mode (i.e., with both front-end and sidestream softeners). A low pH
(6.4-6.5) has been maintained in the recirculating water system. The perfor-
mance of this softener system [4] has been a successful scaleup of the
earlier laboratory feasibility tests.
In 1973 Grits and Glover investigated sidestream treatment for a 50 MW
electric utility [5]. They recommended a higher pH control (7.5-8.5) for
their recirculating water system compared to Fowlkes. Control limits were
established for all the scaling and corrosive materials, which included
silica. Their sidestream softening process was boosted from five to 30 the
cycles of concentration of the cooling water. This was a remarkable increase
of 600 percent.
Grits and Glover later reported that the optimal sidestream configura-
tion to achieve zero discharge consisted of a softener followed by a reverse
osmosis module to control TDS not removed in the softener [6]. They reported
that such a system could be economically applied to any tower operation.
However, capital and operating costs of lime softening were much less than
for reverse osmosis and ion exchange processes. Suspended solids, silica,
and calcium/magnesium precipitation could limit recovery by the reverse
osmosis systenij and foul ion exchange resins.
In 1975, Matson and Perry performed a pilot study for sidestream soft-
ening of a petrochemical plant cooling water system [7, 8], Treated effluent
was added to the makeup of the cooling water system. Sidestream softener
then treated roughly half the makeup water flow. The pH in the system was
controlled by carbon dioxide added to the water. This eliminated sulfate
ions present from sulfuric acid treatment. Bicarbonate alkalinity was con-
served, thereby minimizing the soda ash requirement in the sidestream soft-
ener. Results from the lime softening loop pilot test (Table 1-3) indicated
that the water was supersaturated by carbon dioxide and held the high bicar-
bonate alkalinity at a reasonable pH [9j. The technical limitation for the
system was calcium carbonate scaling potential.
Some problems were detected in testing the feasibility of the system but
none insurmountable. Chemical additives such as phosphonates and polyphos-
phates interfered with the softening process, so use was discontinued.
Silica levels were effectively controlled with the addition of magnesium
oxide powder to the softening reactor. The plant at which the plot tests
were conducted was modified to achieve zero blowdown with lime softening, and
is the subject of the study [10].
11
-------
TABLE 1-3. Recirculating Water Quality in Pilot Plant
Using Carbon Dioxide for pH Control, (9)
River Recirculating
Parameter Water Cooling Water
Calcium hardness, in milligrams per liter as CaCC>3 95
Magnesium hardness, in milligrams per liter as CaCO^ 10
Total alkalinity, in milligrams per liter as CaCO} 105
Total dissolved solids, in milligrams per litera 240
Nonhardness TDS in milligrams per liter 135
Total suspended solids, in milligrams per liter 50
Silica, in miligrams per liter 9
Chrome (hexavalent, in milligrams per liter)
Zinc, in milligrams per liter
Total hardness exiting softening loop
290
51
247
4800
4520
70
20
29
2
155 mg/L as CaC03
aSum of hypothetical compounds in solution
°TDS, less calcium and magnesium
12
-------
A high-density solids contact unit for sidestream softening [11] was
developed by Frazer in 1975. The unit recycled precipitated solids to the
reaction zone in a relatively small volume of water. This allowed the reac-
tions to occur in the presence of 50 g/L of dry solids. Frazer also advo-
cated the use of a two-stage treatment at high TDS levels to remove magnesium
in the first stage, and remove calcium in the second stage.
In 1977, Darji reported on the advantages of the high density solid pro-
cess for sidestream softening [12]. His case studies involved electric power
generation plants with poor quality makeup water. Darji indicated that
makeup softeners were workable in conjunction with sidestream softeners if a
plant's service water also required softening. He recommended two-stage
softeners followed by a filter for suspended solids removal for the side-
stream configuration.
Webb in 1975 made an economic evaluation of sidestream softening to
achieve zero discharge at a coal burning electric power generating plant
[13]. He compared six different options, combining pretreatment of river
water versus no treatment; sidestream softening versus blowdown recovery by
brine concentration; and a sidestream treatment with and without recarbona-
tion. His least cost option was untreated river water makeup with sidestream
softening and recarbonation. In the analysis, the size of the evaporation
pond was the critical factor. It was smallest in the least cost case. Webb's
sidestream treatment had two stages: magnesium removal by addition of
caustic (NaOH) in the first stage and calcium removal by the addition of lime
(Ca(OH)2) in the second stage. Soda ash was not added in Webb's theoretical
model; therefore, a very high bicarbonate alkalinity was needed in the cool-
ing water. Conclusions were that sidestream softening treatment would
achieve zero discharge.
Rice summarized the variety of process schemes for eliminating blowdown.
They included reverse osmosis, vapor compression, and solar ponds in steam/
electric power plants [14]. He acknowledged that most of the schemes
included the use of a sidestream warm lime treatment. For systems that
required further TDS removal prior to sidestream recycling, effluent from the
lime-treatment process was the logical source of water supply. It was clari-
fied, free of colloidal silica, and greatly reduced in organic matter. In
the economic analysis, schematics included the sidestream lime softener as an
integral part of the process for achieving zero discharge (e.g., Figure 1-3).
Wirth and Westbrook also considered the zero discharge concept for the
electric power industry [15]. Their generalized system consisted of chemical
(lime-soda ash) softening, filtration, and brine desalting. They investi-
gated the optimum salinity for a given recirculating cooling water. All
their schemes included the use of a sidestream softener. Their economic
analysis showed that the combination of sidestream softening and electrodial-
ysis had the lowest overall capital requirements. They concluded that such a
combination was ideal in an optimum salinity region to achieve the necessary
calcium, silica, and overall salt removal from the system.
A pilot test program using sidestream softening as a means to achieve
13
-------
Figure 1-3. Sidestream water treatment plant (Rice).
W»l«f Olil/lbuUon
HttdM
Tow«f rut.
1 1
1 1
II
II
U
Coaling
Tovrar
*—
i^^fa.
\y^"
Chtm
F«td
Pump
c
mcJrcuUUng | .
•towoown
Figure 1-4. Pilot Cooling Tower with ZBD capability (Reed, et. al.)
14
-------
zero blowdown was reported by Reed et al. [16], A sophisticated pilot
cooling tower (Figure 1-4) was constructed and tested an ultra-high TDS
recirculating water level of the type expected by zero blowdown systems cur-
rently under design. Reported corrosion rates of stainless steel and copper
alloy specimens were low — less than 2 rapy. Specific detail of the test
conditions was not provided. The performance of the system depended on the
IDS level, the mechanical operations, and the various cooling water treat-
ments applied. They concluded that zero blowdown employing a sidestream lime
softening process would be workable for both new and existing cooling tower
systems.
RECENT DEVELOPMENTS
In 1977, Hennings et al reported on a sidestream softening system that
had been operating since 1976 at the Georgia-Pacific petrochemical plant in
Plaquemine, Louisiana [17]. A schematic of the system in shown in Figure 1-
5. As at Oak Ridge, the concept evolved from the successful recycling of
blowdown to makeup water softeners. Four independent cooling towers were
connected into a common circuit. A single sidestream softener served the
system.
Corrosion rates in the cooling water system were reported to be less
than 1 mpy. Typical cooling water quality is shown in Table 1-4. Microbio-
logical problems were minimal, probably because the sidestream softener acted
as a disinfection process and chlorine was conserved through the system.
Precipitation was not a problem. Hennings concluded that the sidestream
softening concept was the most economic choice for their plant to meet envi-
ronmental standards. Deutsch [10] indicated that the plant was successfully
operating in the zero discharge mode as of February 1979.
Recently, Curtis did an economic analysis of sidestream softening for a
petrochemical plant [18]. He compared its costs to expenditures actually
incurred in conventional (i.e., discharge) cooling water operations. Zero
discharge sidestream softening offered substantial savings by conserving
cooling water chemicals, eliminating blowdown treatment, and cutting makeup
water requirements by 20 percent. These savings equaled the total amortized
cost of the sidestream softening system. After-tax savings were estimated at
$100,000 using sidestream softening for the period 1980-2000, with a return
on investment of 12 percent. Curtis concluded that sidestream softening was
breakeven based on current economics, but became positively attractive when
environmental regulations and/or water conservation forced the issue.
In August 1978, Southern California Edison brought on-stream its Cool-
water electric power generation plant, with both front-end and sidestream
softeners designed into the cooling water system [19]. The makeup water was
from the underground Mojave River. EPA mandated that no aqueous effluents
were to be discharged. The actual makeup water quality was better than
expected. Also, the front-end softeners were not needed. In October 1978,
the cooling water system was running at about 3,000 mg/L TDS. This was far
below the expected steady state of 15,000 mg/L. Apparently, the heat load
was less than expected to evaporate enough water to concentrate the TDS to
15
-------
Che expected values. In Table 1-5, the water quality for the plant is shown.
A number of interesting features differentiate this Coolwater facility
from other sidestream softening systems. Instead of lime, NaOH was added to
the softening reactor for pH control. It was cheap, easy to store, and was
simpler to feed to the reactor. Powdered magnesium oxide was added to ensure
sufficient silica removal. Such operational control maximized bicarbonate
alkalinity in the recirculating water and allowed pH control close to the
calcium carbonate solubility limits. The soda ash requirement for the soft-
ening process was reduced about 100 mg/L.
Five plants utilizing the sidestream softening concept that have been
reported on in the literature are as follows:
1. Oak Ridge (DOE) (TN)
2. Georgia-Pacific (LA)
3. S.C. Edison, Coolwater (CA)
4. Martin Drake Power (CO)
5. USS Chemicals (TX)
The Martin Drake Municipal Power Plant has operated in the sidestream-soft-
ener mode since July 1978. USS Chemicals (formerly Arco Polymers, Inc.) went
onstream in the fourth quarter, 1979. It differs from other plants in that
carbon dioxide is used to control cooling tower pH. Also, cooling tower
makeup includes both process effluent and rainfall runoff. This enhances the
system's recycle capabilities. See Figures 1-6 through 1-11 for photographs
of the facility.
In May 1980, the U.S. Environmental Protection Agency funded the two-
year study of the USS Chemicals sidestream softening system which is this
final report. Operational data collected during the project period is dis-
cussed in this report. A number of experiments designed to optimize various
aspects of the zero discharge cooling system are also noted.
16
-------
ChmmoU
Zinc
Oiipcrtanl
CVomalt
Zinc
Zlne
Chroma)*
Zlne
MiU.ono*
JO.OCO GPU
Actual- le'AT
— 1
S/S
Kilnr
Phinol
To««r
18,000 GPM
Acluol-IS'ar
1
•
— n
5/S
Fill..
PVC
Tow«r
30,000 GI'M
Fillw
S/S
Fllln
— 1
(
ftnionic r./r,.,i
P(X»m»fCa(0l02
Llm*[caUH)2]
(Soda A«n (Na2COj)
Sludg. fV
,_/
/
~ To Undllll Vy "
71
ill
'L
Sod
To»«r 8
At fUqi.
Mainlau
1
ICO
in
_ o
l\
ir
^auslic/CMormt
18,000 GPM
Otugn-20'aT
Actual- 16* ar
lowdown
Ued la
n i CycUt
W4ll HjO
•gaiillir
Sludgi Zinc
Oliptrtanli
Recycle HjO
Figure 1-5. Georgia-Pacific sidestream softening.
17
-------
TABLE 1-4. Recirculating Water Quality at Georgia-Pacific (17)
Parameter Concentration (mg/L)
PHT 6.5
IDS 8000
Total Hardness 480
Calcium 141
Magnesium 31
Chloride 650
Sulate 4500
Sodium and Potassium 2648
Silica 40
Bicarbonate 56
Total Alkalinity 46
^Unitless
TABLE 1-5. Actual Versus Design Recirculating. Water Quality Parameters
for S.C. Edison Coolwater Plant
Concentration (mg/L)
Parameter Actual Design
Calcium as CaC03
Magnesium as CaCOo
Alkalinity as CaCC^
Sulfates
Chlorides
Silica
800
100
50
1300
300
50
300
300
150
t
t
150
Not given
18
-------
"^••••^^^•^
cr~4=sF»^r;-;£ i5
^^S^^ivSr^::5^ ' if»o;«£-TT
. ;•• ^^^^'j^ftgijMarTP1"* «•»-•=» ^Oi- -^r; - ^
^il^^^-j^^
i?!SfiSL, /'asfe^^iSfe r:^Sto?
B S^^^^^^^^fe^^^^^
FIGURE 1-6. View of the USS Chemicals ethylene and styrene petro-
chemical plant.
a IK i if v jgay *^A # K flr*^a_*t»«>^_
i,1*.* *• »-"'iBlf-*l&fl^«v»":S<»,.
FIGURE 1-7, Side view of the two sidestream softeners,
19
-------
£ r'"!::|l-?l":f
"~*-r*-:—-_ • 1 r _ » :, r
FIGURE 1-8. View of softeners and soda ash delivery system.
u. ••...
FIGURE 1-9. Control room and filters.
20
-------
FIGURE 1-10. Side view of recarbonator.
FIGURE 1-11. Top view of recarbonator.
21
-------
SECTION 1 REFERENCES
1. Fowlkes, C.C., "Softening of Cooling Tower Slowdown Water for Reuse,"
Cooling Tower Institute, Reprint TP-112A, (Oak Ridge Gaseous Diffusion
Plant Report K-P-4023), January 5, 1973.
2. Fowlkes, C.C., "Corrosion Rates to be Expected at Zero Slowdown of
Recirculating Water," Materials Protection, pp. 26-30, October, 1974.
i
3. Fowlkes, C.C., Personal Communication, August 9, 1974.
4. Kotesky, R., "Operation of Sidestream Softening Systems," Presented at
Cooling Water, Zero Discharges by Sidestream Softening Workshop, Univer-
sity of Houston, Houston, Texas, June 2, 1978.
5. Grits, C.J. and Glover, G., "Zero Slowdown from Cooling Towers, the
Problem and Some Answers," Proceedings, 34th International Water Confer-
ence, Pittsburgh, Pennsylvania, p. 13, October 30, 1973.
6. Grits, C.J. and Glover, G., "Cooling Slowdown in Cooling Towers," Water
and Wastes Engineering, pp. 45-42, April, 1975.
7. Matson, J.V. and Perry, M.I., "Complete Reuse of Cooling Tower Blow-
down," Cooling Tower Institute, Reprint 145A, January, 1975.
8. Matson, J.V. and Perry, M.I., "Lime Softening of Cooling Tower Slow-
down," Presented at the 79th National Meeting, AIChE, Houston, Texas,
March, 1975.
9. Matson, J.V., "Treatment of Cooling Tower Slowdown," Journal of the
Environmental Engineering Division, ASCE, Vol. 108, No. EEI, pp. 87-98,
February, 1977.
10. Deutsch, D.J., "Lime Softening Helps Cooling Tower Operators," Chemical
Engineering, p. 60., February 12, 1979.
11. Frazer, H.W., "Sidestream Treatment of Recirculating Cooling Water,"
Cooling Towers, AIChE, pp. 76-81, 1975.
12. Darji, J., "Reducing Slowdown from Cooling Towers by Sidestream Treat-
ment," Presented at W.W.E.M.A. Conference, Atlanta, Georgia, April,
1977.
22
-------
13. Webb, L.C., "Sidestream Treatment of Cooling Tower Systems—A Step
Toward Environmental Improvement," Proceedings of the Amerial Power
Conference, Vol. 37, pp. 832-841, 1975.
14. Rice, J.K., "Design Options—Evaluating Cooling Systems with Zero
Aqueous Discharge," Generation Planbook, pp. 81-66, 1976.
15. Wirth, L. and Westbrook, G., "Cooling Water Salinity and Brine Disposal
Optimized with Electrodialysis Water Recovery/Brine Concentration Sys-
tem," Combustion, pp. 33-37, May, 1977.
16. Reed, D.T.; Klen, E.F.; and Johnson, D.A., "Sidestream Softening as a
Means to Achieving Zero Slowdown from Evaporative Cooling Systems,"
Cooling Tower Institute, January 31, 1977.
17. Hennings, J.; Misenheimer, G.; and Templet, H., "Sidestream Softening of
Cooling Tower Slowdown," Cooing Tower Institue, January 31, 1977.
18. Curtis, M., "Economic Attractiveness of Sidesream Softening," Proceed-
ings of the 3rd Conference on Treatment and Disposal of Industrial
Wasteaters and Residues, houston, Texas, pp. 83-87, April, 1978.
19. Matson, J.V., "Report of Trip to the Southern California Edison Cool-
water Power Generation Plant at Daggett, California," October 5, 1978,
Memo to File, Houston, Texas, October 21, 1978.
23
-------
SECTION 2
SIDESTREAM SOFTENER DESIGN
The zero discharge sidestream softening system includes the following
processes:
(1) A softener (solids contact reactor/clarifier) to remove scale-
forming elements.
(2) A pH adjustment tank to acidify the treated water to the proper
cooling water pH.
(3) Gravity or sand filters to remove any suspended particles.
An optional unit process can be added after the filters to reduce the total
dissolved solids (TDS) concentration, e.g.,
(4) electrodialysis or reverse osmosis unit.
Such a unit may be warranted for the control of chloride which is corrosive
at very high concentrations (> 10,000 mg/L).
Scale-forming materials which are removed by the softening reaction are
calcium and silica. They precipitate as calcium carbonate and calcium
sulfate (CaCO-j and CaSO^) and silica dioxide (Si02) respectively. Calcium,
the most prevalent scale-forming element, is removed at the high pH in the
softener as calcium carbonate. Dissolved silica is removed by adsorption
onto magnesium hydroxide floe which precipitates at pH's > 9.0.
Lime (CaO) or caustic (NaOH) is added to the softener reaction zone to
raise the pH. The selection is based on cost: lime is usually cheaper, but
it produces more sludge than caustic. Soda ash is required when insufficient
carbonate (C0?~) is available to precipitate the calcium. If silica scale is
the controlling constituent, then magnesium can be added to remove it. Suit-
able sources of magnesium are magnesium oxide, magnesium sulfate, magnesium
chloride or dolomitic lime.
The objective of the preliminary design process is to specify the soft-
ener size, sidestream flow rate, and the rate of addition of chemicals neces-
sary to maintain the quality of the recirculating cooling water within the
required limits. The design engineer must accomplish the following:
(1) Determine the actual quality of the makeup water and set water
quality limits for the recirculating cooling water.
24
-------
(2) Measure the softener efficiency and calculate the rate of side-
stream flow by mass balance analysis.
(3) Select the feed rate of softening agents plus acid and additives to
remove scale in the softener, adjust the cooling water pH, and to
control fouling and corrosion in the cooling system.
This design process requires knowledge of both the applicable water
chemistry and the many interactions between the various treatments involved.
Results obtained during the actual operation of the softener and cooling
system may also suggest design improvements as described in this report.
COOLING WATER QUALITY
Cooling water standards define the limiting concentrations of scaling,
fouling, and corrosive compounds which contribute to the deterioration of the
cooling system (including pipes, filters, and heat exchangers). In side-
stream softening, the chromate-base corrosion inhibitors protect heat-
exchange equipment and are recycled with the cooling water. Likewise, oxi-
dizing agents used as biocides to control fouling pass through the softener
intact. Therefore, the softener design process is primarily concerned with
the scaling potential of the cooling water, and removal of scale-forming
elements, namely, calcium and silica. Water quality standards are determined
by the acidity of the cooling water and the solubility product constants of
calcium carbonate. If greater accuracy is required, the solubility product
can be corrected to account for the effects of temperature and activity.
Calcium
Calcium forms scale by precipitation as calcium carbonate and/or calcium
sulfate. Calcium precipitation is a function of both ion concentration and
the pH of the cooling water. Scale formation as calcium carbonate (CaCO^) is
shown by the solubility equation:
(Ca2+)(CO§~) =
clcdl~ KSP
where K is the solubility product for calcium carbonate. When the product
of the calcium and carbonate free ion concentrations exceeds the solubility
product, calcium carbonate will precipitate forming scale. At 25°C, where
the activity coefficient is unity, K = 4.82 x 10~9 (moles).
However, carbonate ion concentration is also a function of:
25
-------
[H*][CO?~] -n
- - - — — = K2 = 4 x 10 ij"
[HCO^l
The higher Che hydrogen ion concentration, the lower is the pH, and the
proportion of carbonate ion in the cooling water decreases. In fact, calcium
carbonate scale formation is usually prevented by controlling the pH of the
cooling water between 6 and 7. Carbonate concentrations are very low in this
pH range so that the little calcium carbonate present will remain soluble.
Calcium sulfate is controlled by limiting the concentration of calcium.
Klen [1] found in pilot tests that calcium sulfate scale did not form in
cooling water with calcium present in concentrations less than 700 mg/L as
, even when sulfate concentrations exceeded 50,000 mg/L.
Below 42°C, the hydrated form of calcium sulfate is most common
'2H2°) 5 ac higher temperatures, the anhydride (CaSO^) can be found.
Since solubility decreases with temperature, the hottest heat exchanger tem-
perature controls the precipitation reaction.
As a rule, calcium sulfate scale will form when the calcium concentra-
tion of cooling water exceeds 280 mg/L. Empirical measurements of calcium
sulfate scale formation also indicate that the "cycles of concentration" a
given cooling water can undergo without forming precipitate are limited by
the solubility product for CaSC^.
Silica
Silica scale occurs in the form of silica dioxide (SiC^), a polymeric
colloid which forms slowly when the silica solubility concentration is
exceeded for a given temperature. Determination of the maximum allowable
silica concentration is based upon the coldest heat exchanger temperature as
follows :
Si02 (mg/L) = 4.7T + 24
where T is the temperature of the coldest exchanger in °C.
Magnesium silicate or sipiolite (MgO-SSiC^'S.Sl^O) precipitate in the
cooling water system if the magnesium concentration in the cooling water is
sufficiently high. However, since both magnesium and silica are removed in
the softener (i.e., they are not conservative constituents), sipiolite scale
need not be considered as a parameter in preliminary softener design.
Other scale-forming compounds include calcium phosphate, iron oxide, and
barium sulfate [3]. Calcium phosphate can become a precipitation problem
when the makeup water contains sufficient phosphate or if phosphate-based
inhibitors are used to control corrosion. High concentrations of phosphate
26
-------
inhibitors interfere with the softening reaction, and should not be used with
a sidestream softening system. Makeup water phosphate concentrations are
normally quite low (1 ppm). Neither iron nor barium are likely to be present
in cooling tower makeup water in concentrations to be included in preliminary
design.
SIDESTREAM FLOW RATE
Once the actual makeup water quality has been determined, and the cool-
ing water quality limits have been established, the quality of the softener
effluent must be estimated. Softener flowrate can then be based on a mass
balance over the entire cooling water system. Calculation of chemical feed
rates to completes the preliminary design.
The water quality of a reuse stream will determine its input location.
Relatively clean water can be added directly to the cooling tower -basin (see
Figure 2-1). Reuse streams containing high suspended solids may be input
just before the filters in the sidestream treatment train. Other reuse
streams can be routed through the softener.
The total sidestream flow may therefore be a blend of plant waste plus a
recycle stream. Water quality variations will define a sidestream water
composition which should reflect the worst reasonble operating conditions.
Recirculating water quality must also be estimated and then checked after
sizing the softener flow rate.
Softener Efficiency
The procedure for estimating the efficiency of the softening process is
similar to that used to establish water quality standards. It describes the
softener reaction in terms of relevant chemical equations and adjusts equa-
tions for the effects of temperature and ionic strength. It defines the
chemical equations that the remove calcium and silica in the softener.
As reported by Matson and Harris [3], the concentration of calcium
remaining in solution in the effluent from a typical softener is much greater
than that predicted by the calcium carbonate solubility product constant.
The difference may be attributed to kinetic limitations, e.g., a supersatura-
tion of calcium ions in the softener which results in the formation of ion
pairs [4]. Based on a typical residual calcium concentration of 30 mg/L as
CaC03 at 20°C, I = 0.02, Matson developed an "apparent" solubility product
for calcium carbonate formed in the softener:
Ksp(app) = tCa2*] [CO2,'] = (30)x(30) = 900 mg/L as CaC03
where KSp(app) is the apparent solubility product constant.
This apparent constant can be further modified to accommodate variations
in temperature [6], such that
2'7
-------
K (T
Ksp(app, T°C) = Ksp(app, 20°C) x KS
sp
In addition, Che effects of ion activity may be expressed as [7]
^ = PP. T°C)
, corr) 2
d
where Yj is the divalent ion activity coefficient for a particular solution,
calculated from the Davies equation. The corrected apparent solubility pro-
duct constant, KS p(app Corr) may be assumed to aPPly to a solution in which
the calcium and carbonate concentrations are equal [8]; thus the effluent
calcium concentration is defined
CaHeff = Ksp(app, corr)
Determination of effluent silica concentration involves additional com-
putation, Silica is removed by adsorption onto magnesium hydroxide floe,
which precipitates at high pH. Adsorption is described by the Freundlich
equation, in which
Si02
=
Si02
where - — = silica adsorbed per mass of hydrogen floe precipitate (mg/mg)
Mg(OH)2
k, n = physical constants derived empirically from silica/magnesium
adsorption isotherms
Magnesium concentration in the softener effluent, on the other hand, is
calculated by the solubility product constant:
(Mg2+)(OH-)2 = Ksp = 10~U'6 (25°C, 1 atm)
The difference between the initial and effluent magnesium concentrations is
the magnesium precipitate which enters into the Freundlich equation above.
Thus, the effluent magnesium concentration is controlled by adjusting the
softener pH which also removes silica.
Final calculation of softener effluent silica concentration for design
2-8
-------
purposes requires additional mass balance analysis for the cooling system, as
follows.
Mass B-a lance Analysis
The mass balance analysis traces the fate of significant constituents
throughout the softening and cooling cycles' to determine the necessary soft-
ener flow rate and the characteristic concentrations of the recycled cooling
water.
The zero discharge sidestream softening system is essentially a closed
cooling system. Once materials enter the system, they will remain within the
system. They could be removed in the softener, the filters, or by drift
loss. the addition of lime, caustic, soda ash, and/or magnesium salts will
precipitate calcium carbonate, magnesium hydroxide, and silica. The high pH
and alkali in the softener will also precipitate iron, zinc, trivalent chrom-
ium, and phosphates.
Most suspended materials are removed, including biological organisms,
corrosion residue, and some pretroleum hydrocarbons. If any materials are
precipitated, from either added chemicals or reactions within the system,
these are removed in the softener and filters.
Some materials entering this system are assimilated. An example is
organics reacting with an oxidizing agent such as chlorine. Ammonia and some
light hydrocarbons are lost by air-stripping from the recirculating water in
the cooling tower. However, nonvolatile materials that remain soluble will
pass through the softener and filters to recycle within the system. These
are chlorides, sulfates, hexavalent chromates, bromides, and molybdates.
In zero blowdown systems, (Figure 2-1), the concentrations of these ions
are controlled by drift losses, such that (at steady state)
or
Qm
L^dJ
where C is the concentration of a given constituent, Q is the flow rate of
water in the system, and the subscripts m, w, and d indicate the location of
concentration and flow as that of the makeup water, cooling water, and drift,
respectively. (The drift rate, Qd, is often approximated on the basis of
cooling water evaporation rate).
The nonconservative dissolved solids removed in the softener are cal-
29
-------
Makeup
Softener
Recirculating
Water
*• Sludge
Schematic Diagram Of Sidestream Softening System
Figure 2-1.
-------
cium, magnesium, carbonate, and silica. A mass balance over Che cooling
tower is expressed by the steady state expression
Qmcm + °-sCs = °-dCw + QsCw
where Qs and Cs are the flow rate and constituent concentration going into
the softener loop.
The softener sidestream flow rate, Qg, is determined by the relative
concentrations of the controlling constituent.
Qmcm -
Q_ is determined by heat dissipation requirements, Cm is determined by
analysis of the raw water source. GW is defined according to the previous
section on cooling water quality. The only unknown on the right-hand side of
this mass balance equation is C . While Cg for calcium is calculated as
softener efficiency, determination of the silica concentration of the soft-
ener effluent requires several additional steps to compute. When the above
equation is solved for the cooling water constituent concentration, Cw, and
the expression for Cw is substituted into the Freundlich equation for silica
adsorption, the Qs term cancels out and the resultant equation yields
Cs(Si)
1/n
Qmcm(Si) * Qdcs(Si)
This equation may be solved for Cg to yield the softener effluent silica
concentration and can be used to solve for Q as described. Since calcium
and silica determine two independent sidestream flow rate values, the larger
value for the sidestream flow rate, Qs is necessary to attain the required
cooling wa-ter quality, and the determining constituent is said to control.
Complications arise if insufficient magnesium is present in the makeup
water to adsorb the silica. The solution to the above equation will, there-
fore, yield a Cs(g£) greater than the upper limit even if the pH is high. An
external source of magnesium must then be added.
SOFTENING AGENTS, pH ADJUSTMENT AND ADDITIONAL TREATMENT
Once the softener flow rate has been selected, the rate of addition of
softening agents (either lime or caustic) can be easily calculated according
31
-------
to the s toichiometric equivalents of the scale-forming elements removed.
Subsequent acidification (t^SC^ or C02) regulates the pH of the softener
effluent and the recirculating cooling water in order to prevent precipita-
tion of calicum carbonate in the cooling system. Additional treatment may be
required to prevent corrosion or fouling.
Softening Agents
Lime or caustic are added to raise the pH of the softener to the point
where calcium and magnesium will precipitate as calcium carbonate and magne-
sium hydroxide. Since bicarbonate converts to carbonate at pH higher than
10.4, and the pKg of magnesium hydroxide is 10.7, the pH of the softner is
generally raised to between 10.5 and 11.0, depending on temperature. Addi-
tion of lime or caustic is determined by the amount required to 1) neutralize
the alkalinity of the cooling water sidestream and convert it to the carbon-
ate form; 2) raise the softener pH to the proper level; and 3) precipitate
the magnesium as MgCOH^. This calculation is usually accomplished by
expressing all constituents in terms of parts per million of calcium carbon-
ate equivalents (ppm CaCO^), such that
[CaOH, NaOH] = [Alkalinity] + (Mg2"1"] + [OH] final
Also note that when alkalinity is measured at the methyl orange titration
point (4.5), the corresponding lime required must be doubled to insure con-
version of all the alkalinity to the carbonate form.
When lime is used for pH adjustment of the softener water, in some
cases, soda ash may be required to provide enough carbonate for complete
calcium precipitation. Soda ash requirements are calculated as the differ-
ence between calcium and alkalinity in the pH adjusted softener water; hence
[Na2C03] = [Ca2+] + [Ca2+]added ~ 2[Alkalinity]
The alkalinity concentration (in ppm CaCo^~) is usually doubled to account for
the conversion of monovalent bicarbonate to divalent carbonate.
pH Adjustment
Calcium carbonate scale is controlled by pH of the cooling water within
the limits of CaC03 saturation, usually in the range of pH 6 - 8. The pH may
be lowered by addition of sulfuric acid, or by carbon dioxide. Sulfuric acid
increases the TDS by adding sulfates to the cooling water. Carbon dioxide is
air-stripped in the cooling tower so that excess carbon dioxide must be
replinished. Furthermore, carbon dioxide cannot reduce alkalinity, so that in
certain circumstances, acid addition is preferred.
32
-------
Some cooling water system are being operated at the higher range of pH 7
- 8.5. The advantages for a sidestream-sof tened system are (1) less lime or
caustic is required to raise the pH of the softener sidestream; (2) less acid
is reuired to lower the pH of the softener effluent; and (3) reduced IDS
results in reduced corrosion potential. However, when cooling water acidity
is maintained so close to the pH of calcium carbonate saturation (pH ),
careful monitoring is necessary to prevent the accidental formation of scale.
When calculating the pH$ of the cooling water, allowance must be made
for the effects of ionic strength and ion pairing. According to McGaughey
and Matson [9], when the formation of calcium sulfate ion pairs (CaSO^) are
taken into account, saturation pH can be equated such that
Y?
PHS = pK2 - PK - log TCa - log Alk - log - i -
where ko = second acidity constant for carbonate system
KS = solubility product constant for calcium carbonate
TQ = total calcium, m/L
T
S04 = total sulfate, m/L
Alk = total alkalinity, m/L
Yj_ = activity coefficient (from Davies equation)
The saturation pH determined by the McGaughey method is higher than would be
without correcting for ion pairs.
Acid feed rate is calculated from stoichiometric equivalents required to
neutralize alkalinity and lower the pH to the predeterminied pHs. Acid addi-
tion is in the softened sidestream, and in the cooling tower basin. Sulfuric
acid dosage need be equal to the total alkalinity of the makeup water and the
softener effluent sidestream, plus the stoichiometric equivalents required to
lower the pH of the effluent to the saturation pH,
[H2S04] = [TAlk]m + [TAlk]g + log'1 (pHs)
While sulfuric acid is most commonly added to acidify the softener
sidestream, carbon dioxide has been suggested as an alternative, and is cur-
rently being used for pH adjustment at the USS Chemicals sidestream softening
system. Perhaps the main advantage to carbon dioxide as an alternative to
sulfuric aci is that it does not increase the level of sulfates in the
cooling water. Lower sulfate levels decrease the corrosivity of the cooling
water, as well as the TDS. Also, the dangers of acid overdose are minimized
with carbon dioxide since, unlike sulfuric acid, an excess of COo will never
33
-------
drop Che pH of the cooling water below 4.3 (i.e., the pH at which CQ^ ^s no
longer absorbed). Acidification with carbon dioxide maintains the alkalinity
of the cooling water and reduces the soda ash requirements. Since high
alkalinity can also increase calcium carbonate precipitation, one can add
some sulfuric acid to the softener sidesream effluent. This destroys
residual alkainity and precipitates ionic silica. In practice, carbon
dioxide can be added to the cooling tower basin to lower the pH of the
cooling water before it reaches the heat exchangers.
Additional Treatment
Compounds are added to cooling water to control scaling, corrosion,
fouling, and microbial growth. Many of these substances interfere with the
softening reaction and cannot be used in a zero-discharge system. Each addi-
tive must be evaluated individually to its compatibility with the sidestream
softening process.
Scale inhibitors, such as phosphate esters, phosphonates, and polyacry-
lates, are very effective even at relatively low dosages. They keep scaling
minerals in solution even in water with high scaling potential. They also
prevent precipitation in the softener. Their effect must be offset by the
addition of excess lime or caustic that results in higher effluent calcium
hardness. The same may be said of lignins, tannins, alginates, and starches,
all of which have been used to reduce scale. A recent development in scale
prevention involving modification of the scale crystal structure has been
found to have no deleterious effect on softener efficiency. It even produces
a better quality softener effluent when soda ash is used to supply alkalin-
ity. Most scale inhibitors are removed in the softener and need to be added
continuously.
Corrosion control is often achieved by.the addition of chromate-base
corrosion inhibitors. Hexavalent chrome is not removed in the softener. It
is recycled to the cooling tower system, as are concentrations of other
conservative ions (such as chloride and sulfate) that require higher levels
of corrosion inhibitor. Chromate levels in several systems have been
increased from 20 to 50 ppm to prevent pitting on the heat exchangers. Non-
chromate inhibitors include phosphates, polyphosphates, molybdates, sili-
cates, and nitrates. The polyphosphates cause "after-precipitation," and are
removed in the softener along with the silicates when sufficient magnesium is
present. Other phosphates, including orthophosphate, are removed in the
softener, but do not interfere with softener efficiency. Trivalent chromium,
precipitates out in the softener and is removed as a sludge.
Suspended solids or foulants from any number of sources are controlled
by the addition of dispersants or suspending agents (including polyacryl-
ates). Generally, they are all incompatible with the operation of the side-
stream softener. They prevent the softener sludge from settling. The
nonionic low-foaming surfactants used to control organic foulants (e.g.,
biological slime). They can disrupt the clarifier bed when they are added on
a slug basis. Some of these products are less disruptive than others so one
that causes minimal interference should be selected.
34"
-------
It is necessary before selecting a biocide for control of microbiolog-
ical growth, to determine the fate of the toxicant (1) at the high pH of the
softener, and (2) at the high concentration of organic potentially present in
the recirculating cooling water. Where the cooling system pH varies from
approximately 8 to 11, the biocides will increase in concentration, either as
added or in some degraded form. For example, chlorine is reduced to chloride
and increases the corrosivity of the recirculating water. Also, it is inef-
fective at high alkalinity. By comparison, chlorine dioxide has fewer chlo-
rides and still retains its effectiveness in the softener. The hypochlorite
ion will pass through the softener if the chlorine demand of the CaCC>3 slurry
is low. Other oxidizing biocides (bromine, ozone, 11207) also control micro-
organisms at relatively high pH without forming dissolved solids. Non-
oxidizing agents which contain dispersants are not suited for sides-tream
softening for the reasons given above.
35
-------
SECTION 2 - REFERENCES
1. Klen, E.F. et a I. "Calcium sulfate Solubility in Dynamic Cooling Tower
Systems." WWEMA Conference proceedings, 1976.
2. Matson, J.V. "Treatment of Cooling Tower Slowdown," Journal Environ-
mental Engineering Division, ASCE, 102(87), 1977.
3. Matson, J.V. and Harris, T. op. cit.
4. Cenada, F.C. et al. "The Calcium Carbonte ion Pair as a Limit to
Hardness Removal," J._ AWWA, 66(524), 1974,
5. Fowlkes, C.C. Softening of Cooling Tower Slowdown for Reuse," Cooling
Tower Institute, Houston, Texas, 1973.
6. Fowlkes, C.C. "Corrosion Rates to be Expected at Zero Slowdown of
Recirculating Water," Materials Performance, 13(26), 1974.
7. Nancollas, G.H. Interactions in Electrolyte Solutions, Addison Wesley
(New York), 1966.
8. Nakayama, F.S. and Ranick, B.A. "Calcium Electrode Method for Measuring
Dissociation and Solubility of Calcium Sulfate Dihydrate," Anal. Chem.,
39(1022), 1967.
9. McGaughey, L.M. and Matson, J.V. "Practical Applications of Ion Asso-
ciation Theory: Prediction of the Calcium Carbonate Saturation pH in
Cooling Water," Water Research, 14(12), pp. 1729-35, December 1980.
10. Matson, J.V. et al. "Energy (Cost) Savings by Zero Discharge in Cooling
Towers," Proceedings of the 4th Annual Industrial Energy Conservation
Technology Conference, April 1982, pp. 221-230.
36
-------
SECTION 3
SIDESTREAM SOFTENER OPERATION AT USS CHEMICALS
The USS Chemicals plant was construction in 1961 on a 60-acre plot in
Pasadena, Texas near Houston. The plant manufactures basic chemical feed-
stocks, ethylene and styrene). Capacity for the ethylene unit is 500 million
pounds per year and for the styrene unit, 120 million pounds per year.
In November 1979, the plant was converted to a minimum aqueous discharge
system with sidestream softening (Figure 3-1). A flow diagram for the water
recycle system is given in Figure 3.2. The plant has two cooling towers.
They operate at 110 to 170 cycles of concentration. A sidestream of hot
cooling water is pumped from each tower to a splitter box, which feeds the
lime softeners. Two softeners (referred to as "North" and "South") in paral-
lel remove high concentrations of calcium, magnesiim, and silica from the
cooling water. Lime is added as a slurry to raise the pH in the softener
above 10, and soda ash (Na2C03) is added to provide sufficient carbonate to
complete the softening reaction.
After softening, the sidestream is acidified by injection of carbon
dioxide (COo) in the recarbonator. The softened water is then filtered to
remove any small particulate matter which did not settle in the softener or
recarbonator units, and returned to the cooling tower.
The USS Chemicals water reuse system also incorporates other recycle
streams, some of which contain processed wastewater, characterized by high
TDS, TSS, and organic and oil contaminants. Average water quality data for
the USS Chemicals softener sidestream is presented in Table 3-1.
OPERATIONAL DATA
Since the start-up, the sidestream softening at USS Chemicals, much
effort has been inputted into careful monitoring of the water quality. A
sizeable record of information has been obtained concerning chemical consump-
tion rates, corrosion rates, and quality of both makeup water and softener
effluent. Together, they form an accurate profile of the startup and opera-
tion of a successful zero discharge sidestream softening system.
Hardness, Alkalinity, and pH
Among all water quality parameters, the most critical to the sidestream
softener operation are calcium hardness (CaH), total alkalinity (T-Alk), and
37
-------
CO
CD
STORMWATER
RIVEg
WATER
RIVER WATER
CLARIFIERS
SLUDGE
PROCESS
WASTEWATER
TREATMENT
DECOKING
WATER
SPENT REGENERANT
-H I-
COOLING
TOWERS
-j FILTERS |«»-JRECARBONATOR|^-{ SOFTENERS}
SLUDGE
SLUDGE
(To landfill)
USS CHEMICALS PROCESS FLOW SCHEMATIC OF THE MAXIMUM
RECYCLE, SIOESTREAM SOFTENING SYSTEM
Figure 3-1. Schematic diagram of USS Chemicals sidestream softening system.
-------
FLOW DIAGRAM FOR WATER RECYCLE SYSTEM
LO
VO
evopcxot'on' 'drill
STYRENE COOLING TOWER
SmmENE OOILER OD
bo*rr
Iced
denincralitttr woter
SVC
WATER
RUNOFF
RAINWATER
RUNOFF
cVro ID kmdni
ETHY BOILER BO
ETHYLENE COOLING TOWER
Figure 3-2.
-------
pH. Data for these variables collected from five separate sampling points
are presented in Figures 3-3 through 3-11. Since both the north and south
softeners received the same influent, the similarity of their effluent qual-
ity (Figures 3-3 through 3-5) was. expected. Furthermore, variations in
softener effluent quality can be seen to error, roughly, the quality of the
influent from the ethylene and styrene cooling towers (Figures 3-6 through 3-
8). Water quality in the recarbonator shows the effect of acidification with
carbon dioxide (Figures 3-9 through 3-11).
.The relation between calcium removal and softener pH can also be derived
from a comparison of Figures 3-3, 3-7, and 3-10. The difference between
influent CaH and calcium hardness in the recarbonator (i.e., calcium removed)
does not appear to increase with softener pH, suggesting that the addition of
more lime to increase softener pH does not serve to improve calcium removal
in the softener.
A "trend analysis" performed over the eight-month period from September
1, 1980 through April 30, 1981 revealed no continuous trends, either increas-
ing or decreasing. The pH, CaH, and T-Alk variables are presented appropri-
ately in terms of mean (average), standard deviation, and confidence interval
in Table 3-2. Data for both tabular and graphical formats were obtained by
plant operators every two hours at the north and south softeners, and every
four hours at the other sampling points. The data was averaged over ten-day
periods, as indicated on the time-axis of Figures 3-3 through 3-11; for the
purpose of trend analysis, data were analyzed by 21-day "moving" averages
over the entire 242-day period.
Magnesium Hardness and Silica
Softener influent and effluent concentrations of magnesium and silica
were also monitored concurrently. Figures 3-12 and 3-13 present water
quality data for these from-February to April, 1981. Silica concentration in
the softener effluent is a function of magnesium in the so-ftener influent.
Therefore, an increase in influent magnesium usually corresponds to a
decrease in effluent silica. Alternatively, where influent magnesium is
constant., changes in the levels of silica in the softener effluent parallel
changes in silica in the softener influent.
Water quality and Quality Assurance
Comprehensive analyses of different streams in the cooling water recycle
system where done on four separate occasions. The average of these analyses
was presented earlier (Table 3-1). Individual analyses are presented in
tabular form in Appendix B, along with ionic balance sheets. The accuracy of
these analyses was established on two separate occasions by independent
quality assurance tests.
Split samples were collected on February 17, 1981 from eight streams and
four sludge locations, one for the University of Houston laboratory and one
for the Robert S. Kerr Environmental Research Laboratory (EPA) in Ada, Okla-
40
-------
TABLE 3-1. Average Water System Sample Analysis*
Location
Treated Raw Water
Ethylene Influent
to Softener
Styrene Influent
to Softener
Combined Filter
Effluents
Guard Basin Effluent
WWTU Effluent
M-1008
No.
13
20
21
35
43
50
70
a*-
108
276
282
110
126
72
108
Mg2+t
15
30
35
9
15
8
15
K+
14
343
180
434
135
41
14
**
35
9,195
5,975
7,750
2,887
7,075
34
cr
46
11,100
6,444
8,442
2,971
1,099
45
T.Alk*
76
391
286
259
187.3
90
76.5
Si02
7
62
45
46
23
11
8
pH
7.74
7.40
7.52
8.32
8.90
7.51
7.5
TDS
210
26,000
13,700
24,200
5,140
3,620
354
Turbidity
(Nil)) TSS
2.5 85.5
42.8 256
12.0 141
21.1 182.5
44.4 143.0
41.6 91.5
6.0 85
•roc
7
548
328
476
214
121
8
*A11 values in mg/L unless otherwise noted
*Values in mg/L Ca003
-------
TABLE 3-2. COOLING AND SOFTENING SYSTEM AVERAGES
SEPTEH1JER 1980 THROUGH APRIL 1981
NJ
Ethylene
Cooling
Tower
Styrene
Cooling
Tower
North
Softener
South
Softener
Recarbon-
ator
Calcium Hardness
(ppo CaCO.j)
Standard Confidence
Mean Deviation Interval
250.5
268.5
120.8
120.2
101.6
47.9
33.2
44.1
52.9
31.2
247.4 -
253.7
266.3 -
270.6
110.9 -
122.7
117.7 -
122.8
- 99.5 -
103.6
Alkalinity
(ppm CaCOj)
Standard Confidence
Mean Deviation Interval
362.1
262.0
109.7
111. B
216.7
99.7
67.9
31.2
23.6
41.6
355.5 -
368. B
257.7 -
266.4
108.1 -
111.2
110.7 -
112.9
214.0 -
219.4
pli
Standard Confidence
Mean Deviation Interval
7.46
7.39
10.25
10.23
8.21
7.28 -
7.77
7.21 -
7.72
10.02 -
10.76
9.97 -
10.93
7.97
7.44 -
7. 47
7.38 -
7.41
10.24 -
10.27
10.21 -
10.24
8.17 -
9.26
-------
10.8
10.7
10.6
10.5
pH 10.4
10.3
10.2
10.1
100 L-^
O - O NORTH SOFTENER
SOUTH SOFTENER
10 15
PERIOD
20
OCT NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-3. Softener effluent pH at USS Chemicals, 1980-81,
-------
175 r
150
125
to
O
a 100
75
X
O
O
50
25
O O NORTH SOFTENER
SOUTH SOFTENER
10 15
PERIOD
20
OCT NOV DEC. JAN. FEB. MAR. APR.
Figure 3-4. Softener effluent calcium hardness, 1980-81.
-------
175 r
150
rr>
8 125
o
to
O
e 100
< 75
_j
<
50
25
O O NORTH SOFTENER
SOUTH SOFTENER
/~\ I i i i i i i i i i i i t i i
10 15
PERIOD
i i i i
20
OCT. NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-5. Softener effluent alkalinity, 1980-81.
45
-------
pH
8.0
7.9
7.8
7.7
7.6
75
7.4
7.3
7.2
7.1
7.0
O O ETHYLENE
STYRENE
10 15
PERIOD
20
OCT NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-6. Ethylene and styrene unit cooling water pH, 1980-81,
46
-------
340 r
300
260
o
O
O
o
CJ
220
x
o
O
180
140
100
O—O ETHYLENE
STYRENE
I I j j t i
10 15
PERIOD
20
OCT NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-7. Ethylene and styrene unit cooling water calcium
hardness, 1980-81.
47
-------
600 r
500
rO
o
o 400
o
en
O
300
< 200
100
O O ETHYLENE
STYRENE
i , , , , i , , . , i . , . . i . . , .
10 15
PERIOD
20
i I I i
I I
OCT. NOV. DEC. JAN. FEB MAR. APR.
Figure 3-8. Ethylene and styrene unit cooling water alkalinity,
1980-81.
48
-------
X
Q.
9.4
9.2
9.0
8.8
8.6
8.4
8.2
8.0
7.8
7.6
7.4
7.2
7.0
RECARBONATOR
10 15 20 25
PERIOD
OCI NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-9. Recarbonator pH, 1980-81.
49
-------
175
150
125
rO
100
o»
J 75
I
o
o
50
25
RECARBONATOR
10 15
PERIOD
20
OCT NOV. DEC. JAN. FEB. MAR. APR.
Figure 3-10. Recarbonator calcium hardness, 1980-81.
50
-------
400 r
350
300
rO
o
o
o
250
f 200
H
§ 150
100
50
I I L _J I I A i_. t I I I lit
10 15
PERIOD
_I | I
RECARBONATOR
i t it
20
j I
OCT NOV. DEC JAN. FEB. MAR. APR.
Figure 3-11. Recarbonator alkalinity, 1980-81.
51
-------
homa. The only difference in the two laboratory analyses was in the measure-
ment of alkalinity. An interference of CaCC>2 fines before filtration caused
an error in the total alkalinity measured. A quality assurance test was per-
formed by the Water and Wastewater Analysts Association.
Chemical Usage and Cost Rates
Concurrent with the analysis of water quality in the lime softening and
cooling water systems, chemical usage rates and chemical costs (cost per day
and cost per 1000 gallons) were also monitored. For two periods—from June
through September, 1980, and from March through May, 1981 — the four costs
basic to the lime softening system were measured. Lime, soda ash, CC^, and
sludge for both periods are shown in Tables 3-3 and 3-4, 'These parameters
were calculated by comparing measurements of amounts of chemicals used (or
disposed) with usage rates as determined by inventory check.
As regards the lime softening system, the lime calculations for the
second period were more precise since only then was the amount of dilution
water added to the lime slurry storage tank measured. Also, since most COo
consumption occurs in the ethylene and styrene cooling towers, the design
rate of 1000 Ibs/day to the recarbonator is assumed as the COo consumption
rate. (Consequently, the consumption rate of the cooling water is decreased
by 1000 Ibs CC^/day.) Best estimates divided the production of filter cake
evenly between the raw water clarifier and the lime softener.
Cooling water costs and chemical usage rates are listed in Table 3-6.
Since most of the cooling water chemicals are added on a batch basis, con-
sumption figures are usually derived from daily operator records. However,
chlorine and carbon dioxide were also monitored by pressure gauges.
Carbon Dioxide _
One major disadvantage in using carbon dioxide for control of pH is that
the cooling tower acts as a gas stripper that continually removes C02 from
the water. As a result, carbon dioxide must be continually replenished, and
roughly ten times as much C02 must be used, as compared to equivalent pH
control with l^SO^.. Furthermore, it is frequently difficult to measure
carbon dioxide in cooling water accurately, since CC>2 is often lost to the
atmosphere during sampling and transfer. Therefore, a method was developed
at the University of Houston to determine CC^ i-n USS Chemicals cooling water
as a function of alkalinity, and to calculate the extent of W^ stripped out
by the cooling tower.
The method is based upon ah equilibrium equation for carbon dioxide in
an aqueous system,
+) (HCOp
52
-------
TABLE 3-3. Chemical Usage Rates - Lime Softening System
CO
Sludge
Lime
Soda Ash
Dates
days
tons Ib/day
C02
i!/day
tf/day
6/16 - 7/4
7/5 - 7/30
7/31 - 8/18
8/19 - 9/7
9/8 - 9/27
20
26
19
19
20
166,972 8348.6
130,606 5023.3
64,021 3369.5
115,656 6087.2
127,515 6375.8
50.92 5092
37.476 2883
34.540 3636
34.998 3368
32.171 3217
11,350 568
20,350 783
16,100 847
21,950 1155
19,000 950
20,000 1,000
26,000 1,000
19,000 1,000
19,000 1,000
20,000 1,000
*Based upon sludge produced from lime softeners
+Based upon C0« added to recarbonator
-------
TABLE 3-4. CHEMICAL USAGE RATES FOR LIME SOFTENING SYSTEM
Dates
(1981)
3/5-1/25
3/26-4/15
A/16-5/6
5/7-5/27
Ho. Days
21
21
21
21
t wt.
Ave.
we/day
Lime
Dry Tons Tons/Day
13.971
10.667
49.765
42.677
177.080
2.094
1.937
1
2.370
2.032
2.108
Soda Ash
Ibs Ibs/day
7.700 366.67
15.600 742.86
19.750 940.48
16,300 776.19
59,350
706.55
CO/
Iba Ibs/day
21,000 1,000
21,000 1,000
21,000 1,000
21,000 1,000
84.000
1,000
Filter Cake*
(produced)
Ibs Ibs/day
146.260 6964.8
104.389 4970.9
178.560 8502.9
121,850 5802.4
551.059
6560.2
Estimate of dosage to recarbonator
Estimate of lime softener sludge produced
-------
TABLE 3-5. Average Qisnical Costs and Usage Races—Softener
Flow T.-ijT"3 Soda Ash 002 Sludge Total
Period 1000 gal/day $/# #/day $/# #/day $/# #/day 5/1 T/day $/1000 gal
6/16/80 -
9/27/80 871.1 .035 3639.2 .102 860.6 .019 1000 41.03 2.92 .402
3/5/81 -
5/27/81 801.4 .039 4216 .108 706.6 .026 1000 9.88 3.28 .377
55
-------
TABLE 3-6. Average Chemical Costs and Usage—Cooling Tower
Chemical
Calcium Dispersant
Zinc Inhibitor
Chromate Inhibitor
Non-Oxidizing Biocide
Chlorine (C^)
Bromine (B^)
Carbon Dioxide (CC^)
TOTAL
*/*
1.63
0.67
1.08
2.02
0.10
3.00
0.03
*/day
111.5
53.9
26.5
26.7
256.7
63.0
14,018.0
$/kgal
.041
.008
.006
.012
.006
.042
.094
.209
Period: 3/5/81 to 5/27/81
Flow: 4457 kgal/day
56
-------
Cn
-j
250
$
3200
in
o
-150
to
in
LU
z
a
I100
to
50
-O INFLUENT TO SOFTENER
O D NORTH SOFTENER REACTION ZONE
A A SOUTH SOFTENER REACTION ZONE
I I
8 12 16 20 2^ 28 4 8 12 16 20 24 28 I 6 9 13 17 21
FEBRUARY MARCH APRIL
Figure 3-12. Softener influent and effluent magnesium hardess, 1981.
-------
tn
00
100
80
60
< 40
o
20
o
-o
INFLUENT TO SOFTENER
O D NORTH SOFTENER REACTION ZONE
A A SOUTH SOFTENER REACTION ZONE
I I I
12 16 20 24 28 4 8 12 16 20 24 28 I 5 9 13 17 21
FEBRUARY MARCH APRIL
Figure 3-13. Softener influent and effluent silica, 1981.
-------
where (H2CO§) = (H2C03) + (C02)aq
Since the concentration of carbonic acid may be considered negligible with
respect to the concentration of dissolved carbon dioxide, the above equation
reduces to
In waters of low ionic strength, at 25°C, K^ = 6.35 [1]; additional
constants adjusted for temperature are available in most tables. However, in
cooling waters of high ionic strength, the acidity constant must be further
corrected for ionic strength, e.g., by the Guntelberg approximation of the
Debye-Huckel limiting law
pK; = pKl - 0.5 /i
(1 + 1.4
By substituting the adjusted activity constant into the above equation, the
amount of C02 stripped out in the cooling tower (and hence the amount of C02
required to replace it) may be calculated simply by measuring the cooling
water temperature, pH, and total alkalinity.
An example of this method of C02 determination can be seen in Tables 3-7
and 3-8. In Table 3-7, cooling water from the ethylene system was analyzed
(for Ca2+, Mg2 + , K+, Na + , Cl~, and SO2') and found to have an ionic strength
of 0.4986, or 0.50. Temperatures were 46°C and 32°C, and pH was 7.2 and 7.8
at the top and bottom of the cooling tower, respectively. Alkalinity was
measured as follows:
1. Samples of cooling water from the top and bottom of the tower were
each acidified to 4.5 with 0.02 N H2S02~ in order to determine total alkalin-
ity.
2. The samples were purged with air for one hour to strip out C02 and
destroy all carbonate alkalinity.
3. After raising the pH of the sample back to its original value with
0.02 N NaOH, the samples were again tested for alkalinity—in this case,
alkalinity due to organic acids, or some other non-carbonate source.
The difference between the first and second alkalinity tests yielded the
carbonate alkalinity used for determination of CCU in cooling water,
according to the procedure outlined below.
59
-------
(C02)
where (H+) = log -1 (-pH)
"5
(HCOj) = 10" [Alk]c03
KI = log"1 (-pK.{)
2. [C02]stripped "
Once the amount of carbon dioxide stripped out by the cooling tower is deter-
mined (in terms of mol/L), it is simple to convert this figure to percentage
CC>2 stripped, or to mg/L as CaCC^. Furthermore, the feed rate of carbon
dioxide required to maintain the cooling water pH may be determined by calcu-
lation of the total carbonate alkalinity removed between the top and the
bottom of the cooling tower. Hence:
= 10
where TC03 = [Alk]CO;3 + [C021
4. C02 Feed = [C021 consumed (mol/L)
x 44 .8 x 1 lb x 3.78 L x 50,000 gal x 1440 min
mol 'Zo5~g gal mm day
Table 3-8 presents similar data for cooling water for the styrene
system; all calculations are performed in the same manner as for the ethylene
unit. As a check on the method, the sum of the carbon dioxide consumed in
the ethylene cooling tower, styrene cooling tower, and recarbonator systems
was compared with the daily consumption as calculated by C02 inventory.
Since inventory consumption was 19,000 Ibs C02/day, and the calculated con-
sumption was 17,962 Ibs C02/day for a difference of 5.5 percent, the method
was considered acceptable.
Hydrodynamics
As a further investigation into the actual operation of a sidestream
softening system, a dye tracer study was performed on both softeners and the
60
-------
TABLE 3-7. Determination of Carbon Dioxide Consumption in
the Ethylene Cooling Tower System
Parameter Cooling Tower Top Cooling Tower Bottom
pH
T (°C)
Alkalinity (mg/L CaC03)
Non-Carbonate
Carbonate
P*l
pK;
[CO,] (mg/L CaCO-0
7.2
46
340.0
51.0
289.0
6.318
6.110
23.5
7.8
32
342.0
85.0
257.0
6.140
6.140
5.6
Total Carbon Dioxide Consumed = 13,163 Ibs
Ionic Strength =0.50
Flow Rate = 50,000 gpm
61
-------
TABLE 3-8. Determination of Carbon Dioxide Consumption in
the Styrene Cooling Tower System
Parameter
pH
T (°C)
Alkalinity (mg/L CaCO-j)
Non-Carbonate
Carbonate
P*l
PK;
[C02] (mg/L CaC03)
Total Carbon Dioxide Consumed =
Cooling Tower Top
7.3
37
290
40
250
6.306
6.15
7.8
3957 Ibs C02/day
Cooling Tower Bottom
7.8
25
292
so ;
212
6.358
6.20
2.3
Ionic Strength =0.32
Flow Rate = 15,000 gpm
62
-------
cr>
LO,
'max
1.2
I.I
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
O.I
0.0
SOUTH SOFTENER
A A NORTH SOFTENER
0 0.5
1.0
1.5 2.0 2.5 3.0 3.5 4.0 4.5
Figure 3-14. Results of dye tracer study of softener dynamics. (C = 24 ppb [North]
max
28 ppb [South] )
-------
'max
1.2
I.I
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
O.I
0.0
O RECARBONATOR
0 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50
t
Figure 3-15. Results of dye tracer study on recarbonator dynamics. (C =430 pph)
fflci A
-------
recarbonator to compare real and theoretical detention times. Rhodamine B
dye was injected into the north and south softeners in single slugs to 24 and
28 ppb, respectively (Figure 3-14). Monitored effluent revealed immediate
dispersion of the dye, with gradual elimination characteristic of a CFSTR.
Observed detention times of 97 and 111 minutes compared favorably with the
calculated softener detention time of 120 minutes.
Unlike the softeners, the recarbonator accepted a slug of 430 ppb which
did not appear at its maximum concentration until t/to= 0.5, for a detention
time of 51 minutes. The calculated detention time for the softener was 80
minutes; the profile of the recarbonator effluent (Figure 3-15) is similar to
plug flow with axial dispersion.
Corrosion
Corrosion rates were monitored with mild steel coupons placed in both
the etylene and styrene cooling systems. Each coupon had a surface area of
19.75 cm , and a density of 7.65 g/cm . Coupon exposure in the styrene unit
varied from 26 to 221 days, and was carried out at various times between
December 1979 and April 1981. On the average, penetration of the mild steel
coupons amounted to 0.52 mpy, as determined by the equation:
•o = k x W
A x D x T
where P = penetration (mpy)
W = coupon weight loss (g)
A = coupon surface area (cm )
D = coupon density (g/cm )
T = exposure time (days)
k = product of all conversion constants (e.g., 365 days/year, etc.)
Final formula after substitution of all constants and coupon values was
P = 951.1 W
T
There was a slight problem with debris setting in the small diameter
corrosion rack pipes, so the coupons were cleaned regularly with a dilute
solution of 1 + 1 HC1. Coupon exposure in the ethylene unit was similar to the
styrene unit. Average corrosion rate was 2.83 mpy. The greater penetration
65 "•
-------
was attributed to the inclusion of a wastewater treatment stream with high
IDS and high organic content. The increased conductivity, biological foul-
ing, and miscellaneous impurities of the wastewater stream did contribute to
the higher corrosion rate.
66
-------
SECTION 3 - REFERENCES
1. Stumm, W. and Morgan, J.J. "Dissolved Carbon Dioxide," Chapter 4 in
Aquatic Chemistry, 2nd ed. New York: Wiley-Interscience, 1981.
2. Gardiner, W. and Matson, J.V. "Demonstration of a Maximum Recycle Side-
stream Softening System at a Petrochemical Plant," First Annual Progress
Report, EPA Grant #CR 807419-01, August 11, 1981.
67
-------
SECTION 4
SIDESTREAM SOFTENER PERFORMANCE
Operation of the minimal aqueous discharge sidestream softened cooling
system at USS Chemicals provided numerous opportunities to study the zero
discharge softener design. This chapter presents the results of three sepa-
rate studies of softener performance.
The first study concerns the various measurements of alkalinity which
may be used to determine the soda ash (sodium carbonate) requirements of the
softener. The second study comprises investigations into the removal of
silica from cooling water, including the effects of temperature, ionic
strength, and magnesium concentration. The third study evaluates the effects
of mixing on the softening process.
Alkalinity Determination and Softener-Efficiency
This study was conducted to improve calcium removal by using a total
organic carbon (TOG) analyzer to determine the sodium carbonate (Na2C03)
requirements. Since soda ash (Na2C03) is added to the softener to make up
for any carbonate deficiency, measurement of available ionic bicarbonate
determines the amount of soda ash added. At the USS Chemicals plant, ionic
bicarbonate is measured by testing methyl orange alkalinity. The method used
calculates carbonate on the basis of inorganic carbon determined by a TOC
analyzer.
MATERIALS AND METHODS
The calcium removal in the softeners at USS Chemicals was simulated by
treating actual cooling water and softener effluent from the plant in jar
tests. Separate jar tests were performed with soda ash dosages based on
methyl orange alkalinity, and total inorganic carbon alkalinity. The rela-
tive efficiency of these methods was determined by comparing residual calcium
levels after a predetermined period of softening had occurred. An additional
series of jar tests was performed to determine the relationship between lime
and soda ash dosage, and the ratio of calcium to carbonate residual. A third
test concerned the relative contributions of the bicarbonate ion and organic
acids to total alkalinity. Finally, an economic evaluation was made to
analyze the cost of improving calcium removal.
Measurements included methyl orange alkalinity (M-Alk); phenolphthalein
alkalinity (P-Alk); inorganic carbon (1C); and total organic carbon (TOC);
calcium, silica and magnesium concentrations, and pH of the cooling water
samples. 1C and TOC were measured with a Beckman Model 915A TOC Analyzer;
all other tests were performed by wet chemical analysis as outlined by
Standard Methods [2].
68
-------
Experiment #1: Soda Ash Dosage and Calcium Removal
The first series of tests consisted of softening cooling water and soft-
ener effluent ("pre-softened") samples from USS Chemicals with soda ash
dosages determined according to several different methods. In all cases, the
softening procedure was as follows:
(1) Add lime to pH 10.5 and soda ash as determined.
(2) Stir mixture for at least 45 minutes to simulate full-scale soften-
ing process.
(3) Settle for 10 minutes.
(4) Filter 100 ml through 0.45um paper filter.
Prior to softening, all samples were analyzed for calcium, alkalinity, 1C and
TOC, pH, and silica. After softening, the samples were analyzed for calcium
alone.
In the cooling water experiments, alkalinity (P-Alk and M-Alk) and 1C
measurements were used to determine soda ash requirements according to the
following formulae (all concentrations in ppm
Alkalinity
[Na2C03] = [Ca2+]cw + [Ca2*]lime - [M-Alk]
2(IC) _______
[Na2C03] = [Ca2+]cw + [Ca2+]lime - 2[lC-Alk]
1C
[Na2C03] = [Ca2+]cw + [Ca2+]Ume - [iC-Alk]
where the subscripts cw and lime with reference to calcium indicate the cal-
cium concentration originally present in the cooling water (as determined by
analysis) and the calcium added in the form of lime, respectively. Inorganic
carbon (1C) determined by TOC analysis in terms of mg C/L was converted to
IC-Alk according to the following conversion formula:
llC-Alk]
12 mg 1C mnol C
It may be noticed that the expression mmol HCO^/mmol C is equivalent to the
proportionality term oiHCO^; hence the above formula can be abbreviated
69
-------
[IC-AlkKppra CaC03) = m?LIC x 4.17
No lime was added to the softener effluent, which entered the laboratory
at a sufficiently high pH. Therefore, soda ash dosages for "pre-softened"
samples were determined as follows:
[Na2C03] = [Ca2+]s - U
and
[Na2C03] = [Ca
2+
s
where [Ca2+]s indicates the calcium concentration of the softener effluent
before the second softening reaction.
Experiment #2: Calcium/Carbonate Ratio
The second set of tests involved softening USS Chemicals cooling water
with varying amounts of lime and soda ash so as to obtain different ratios of
calcium to carbonate in the softener filtrate. in the first experiment, 400
mg/L lime (As Ca(OH)2) raised the pH of the cooling water to 9.4-9.7, and
soda ash was added from 50 - 175 mg/L (as Na2C03). In the second experiment,
the lime dosage was increased to 420 mg/L, which raised the pH to 10.3-10.5
while the same range of soda ash dosages were obtained. The final experiment
involved various lime dosages from 450-650 mg/L, with soda ash added from
219-505 mg/L.
The softening procedure in the second set of tests was identical to that
outlined above, except that in addition to effluent calcium, silica and pH
were also analyzed, and the alkalinity of the filtrate was measured. Thus,
carbonate after softening was determined (in ppm CaC03) as
[CO2;'] = [P-Alk] - [OH~] - [Si02]
The ion activity product and saturation index were calculated for each fil-
trate, according to the method discussed earlier.
The results of the second set of tests were also used to estimate the
contribution of organic acids to alkalinity (OA-Alk). After calculating
carbonate (as above) and bicarbonate where
r ..
[HC03]
70
-------
the organic acid alkalinity was calculated with respect to the total alka-
linity such that (in ppm CaCOo):
[OA-Alk] = [P-Alk] + [M-Alk] - [HCO^] - 2[COj~] - [Si02] - [OH"]
In addition, graphs derived from these results were used to obtain projec-
tions concerning the impact of lime and soda ash dosage on softener economy.
RESULTS AND DISCUSSION
TOG analyses performed over several months indicated relatively constant
values of total carbon (TC), inorganic carbon (1C), and total organic carbon
(TOC) for both the cooling water samples and the softener water samples,
carbon ranged from 8.0 to 24.5 percent by weight (34.5-91.0 mg/L), maintain-
ing an average of 17.8 percent 1C. In the softener effluent, inorganic
carbon varied between 1.8 and 7.2 percent, an average of 4.9 percent. TOC
averaged 302 mg TOC/L in the cooling water, and 246 mg TOC/L for the softener
effluent. As shown in Figure 4-1, nearly all of the 1C present in the
cooling water was removed during softening down to a level of around 10 mg
IC/L. On the other hand, TOC appeared to maintain approximately the same
level both before and after softening (Table 4-1).
Effect of Alkalinity Measurement on Calcium Removal
Figure 4-2 illustrates that IC-Alk is significantly lower than M-Alk,
and that P-Alk is lower still. As a result,, soda ash dosage (Figure 4-3) for
softening the -cooling water is much higher when calculated by the value of
2(IC-Alk) rather than M-Alk, and the subsequent removal of calcium (Figure 4-
3) is more complete. When soda ash dosage is calculated by the still lower
level of IC-Alk even more soda ash added, and more calcium is removed.
The same patterns are observed for analyses and jar tests of softener
effluent samples (Figs. 4-4 through 4-7). Figure 4-6 illustrates the soda
ash dosage as determined by IC-Alk/2, and the still smaller value, IC-Alk/4.
As in the cooling water jar tests, the smaller alkalinity value resulted in
higher soda ash dosage, and more complete calcium removal (Figure 4-7). When
the initial calcium concentration in the softener effluent was low ( 65 ppm
as CaCO^), further softening did not effect significant calcium removal.
Furthermore, when initial calcium concentration was greater than 90 ppm,
further softening only reduced the residual to a final concentration in the
range of 60 ppm.
Calcium/Carbonate Ratio in Softener Effluent
Results from the second series of experiments (in terms of percent
calcium removal and final calcium/carbonate ratio) are presented in Table 4-
2. The calcium/carbonate ratio was deemed significant since an excess of
either calcium or carbonate in the softener filtrate was considered an indi-
cation of inefficient softening. Both calcium removal and calcium/carbonate
71
-------
TABLE 4-1. TOG Analyses of Cooling Water Before and After Softening
Before Softening
Run # TOG
1 247
2 27
3 281
4 264
5 279
1C
68
69
91
84
76
TOC/TC
78.2
76.7
75.5
75.9
78.4
After Softening
TOG
244
235
292
254
276
1C
10
11
10
10
8
TOC/TC
96.1
95.5
96.7
96.2
97.0
72
-------
400
TC-
1C
TOC
350
300
250
u
E
200
150
100
50
Figure 4-1.
RUN
Comparison of total organic carbon (TOC) and inorganic
carbon (1C) in USS Chemicals cooling water before and
after softening (a = before softening; b = after softening)
73
-------
to
O
O
o
O
E
o.
o.
400 -
300 -
200 -
100 -
-
mm
•
_
M-ALK
IC-ALK
P-ALK
\
\
\
\
\
\
\
\
\
\
\
\
\
\
•••
1
••*
\
\
\
\
X
\
\
\
\
\
\
\
\
iw
v^1
\X
\\
\\
\\
\\
\X
x\
f^
X
\
\
X
X
X
X
X
\
\
\
v
s
\
\
\
\
\
-r
^>-
\
\
\
\
\
\
\
\
\
\
\
\
N.
\
\
\
\
\
\
^M
\
\ V
x
\ V
\ N
\ V
\ N
\ V
\ N
\ V
\ N
\ N
\ Vl
\ N
\
\ S
\S
\
PI
•••
\
V
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
^^
\
\
\
\
\
\
X
\
X
X
X
X
X
X
\
\
\
\
\
\
\
\
\
••
••
23456 78
RUNS
Figure 4-2. Comparison of various alkalinity measurements of
USS Chemicals cooling water.
74
-------
680
640
600
560
520
0 480
o 440
z
I 400
Q.
Q.
~ 360
UJ
8 320
Q
ro 280
O
^ 240
0
Z
200
160
120
80
40
-
-
-
-
-
-
ml
_
-
No2C03 BY IC-ALK
HI No2C03 BY 2 (
\s|
\\1
'XN!
* \\1
-SS|
r~i
:
\
\
\
\
\
\
\
\
\
\
\
C-ALK)
No2C03BY M-ALK
'•:
*-n
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
~
:•:• r
:j:::
•'•' ! s
S'^
"
^^*
__
L
;"H
234
RUNS
8
Figure 4-3. Effect of cooling water alkalinity on soda ash dosage.
75
-------
TABLE 4-2. Results of [Ca2+]/[C02~] Experiments
Test #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Dosage
Ca(OH)2
Added
400
400
400
400
400
420
420
420
420
420
420
450
500
550
600
650
(mg/L)
Na2C03
Added
50
100
125
150
175
50
75
100
125
150
175
219
290
362
433
505
([C*2+]e)
[co32-]e
6.92
7.68
4.70
4.31
2.76
2.84
2.37
2.25
1.54
1.27
1.13
0.90
0.60
0.43
0.40
0.30
Percent
Removal
41.4
49.7
52.8
57.5
63.2
47.9
50.0
55.5
63.0
69.9
71.9
77.2
81.5
85.8
87.0 .
88.3
SI
1.09
1.05.
1.07
1.02
1.09
1.13
1.18
1.12
1.10
1.03
1.00
1.01
1.00
0.92
0.87
0.90
Organic
Acids
(% T.Alk)
10.2
20.8
10.4
5.4
18.4
19.0
15.8
16.3
12.3
13.5
16.3
5.7
4.0
7.8
18.3
11.5
76
-------
J
320
300
280
-« 260
o
S 240
CJ
E 220
Q.
Q.
- 200
z
9. 180
^_
5
cr 160
^™
z
£ 140
I 120
:o 100
(j
80
60
40
20
k
_
-
.
_
-
-
-
-
-
-
-
-
-
-
-
1 INITIAL [Ca2*]
[Ca2*] AFTER TREATMENT
BASED ON [M ALK]
i. [Co2*] AFTER TREATMENT
;:n BASED ON 2[ic ALK!
- \ " • J
: v':
[Ca2*] AFTER TREATMENT
BASED ON [lC ALK]
: q
) X
;i^
j;
:|
•jq
•! \
1 2
T
•• -,
:• I
-,
•'. .- »
:• , • .
'.- f •
ri
!• 'X «
-; / »
34567
RUNS
\
•
ll
8
Figure 4-4. Effect of soda ash dosage on removal of calcium from
USS Chemicals cooling water.
77
-------
M-AIK
IC-AIK
P-AIK
150
140
130
5* 120
0
o 110
E
a 100
K__
£ 90
^^
z
U 80
y*1
-J 70
<
60
50
40
30
-
-
-
-
•»
- r.
k
k
l\
7
i
\
\
\
\
\
\
\
\
\
\
\
\
s
•7
;';
:-i:
^
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
-
f
1 2 3
\
\
\
\
\
\
\
x^
\
\
™
:'
'•:•
|-
i:
-
V
\
\
\
\
\
\
\
\
\
\
\
™
i
4 5
T
\
\
\
\
\
\
\
\
N
\
;.;
i
6
r
s
\
\
\
\
\
\
;•;
\-
'.•'.
i_
\
\
\
\
\
\
\
\
\
\
\
s
\
\
\
\
\
X
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
'•:
f
-
^"
\
\
\
\
\
\
••
•
i:
-
T
\
\
\
\
\
\
\
\
\
s
\
\
\
\
\
\
\
s
\
\
\
\
\
\
\
\
\
\
\
\
\
•
1
i
\
V
\
\
\
X^
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
-r
•I;
F;
78 9 10 II
RUNS
Figure 4-5. Comparison of various alkalinity measurements of
USS Chemicals softener effluent.
78
-------
No2C03 BY
[IC-ALK]
Na2C03 BY
[IC-ALK]
1
55
50
45
3* 40
o
CM
i 35
E
1 30
to
Hi 25
o
2
20
15
10
5
-
-
-
-
-
-
-
-
~
81
~
T!T
I
i
—
— i
I 2 34 5 6 789 10 II
RUNS
Figure 4-6. Effect of softener effluent alkalinity on soda ash
dosage.
79
-------
INITIAL [Ca2+]
r i fIC A|K1
[Co2*] AFTER TREATMENT BASED ON - — - - -
"
r T
[Ca2+J AFTER TREATMENT BASED ON
AIK
1
130
120
§110
o
O
EIOO
a.
a.
^ 90
5 80
cr
o
70
i?^
o 60
•4-
3 50
40 -
30 -
20 -
10 -
567
RUNS
8
10 II
Figure 4-7. Effect of soda ash dosage on additional removal of
calcium from softener effluent.
80
-------
100
90
80
J 70
O 60
£50
40
30
20
10
400mg/L LIME
O - O % REMOVAL
D - D
42O mg/L LIME
O ---- -0 % REMOVAL
•— • -B [Co2+]/[c02~]
OPTIMUM
EFFICIENCY
LINE
10.0
9.0
8.0
70
6.0,
5.0
o
o
*0p
3.0
20
1.0
25 50 100 150 200 250 300 350
Na2C03 ADDED
(mg/L)
Figure 4-8. Effect of soda ash dosage on cooling water calcium removal and
[Ca]/[CO ] ratio.
-------
o
2
LJ
100
90
80
70
60
50
40
LIME DOSAGE
O—Q 400 mg/L
420 mg/L
A 450 mg/L
A 500 mg/L
V 550 mg/L
600 mg/L
mg/L
30
20
10
1.0 2.0 3.0 4.0
[Co2*] / [CO2"]
5.0
6.0
7.0
Figure 4-9. Relation between [Ca]/[CO,] ratio and calcium removal from
USS Chemicals cooling water.
82
-------
5.0r
4.0
3.0
1.0
LIME DOSAGE
400 mg/L
420 mg/L
450 mg/L
500 mg/L
550 mg/L
600 mg/L
650 mg/L
9.0
10.0
pHe
10
12.0
Figure 4-10. Relation between softening reaction pH and final
[Ca]/[CO.] ratio.
83
-------
ratio are plotted as functions of soda ash dosage in Figure 4-8. Increased
lime dosage resulted in increased calcium removal for all levels of soda ash
concentration. At both 400 mg/L and 420 mg/L lime, removal also improved
with increased soda ash dose. Conversely, as more soda ash was added, the
ratio of calcium to carbonate in the softener filtrate decreased and
approached unity. This trend is also pictured in Figure 4-9, which shows
calcium removal improving with decreasing calcium/carbonate ratio in the
filtrate. Interestingly, despite wide variation in lime and soda ash dosage,
calcium removal appears to approximate 75 percent as the filtrate calcium/
carbonate ratio approaches 1.0.
Figure 4-10 plots calcium/carbonate filtrate ratios against pH, illus-
trating that when the softener pH reaches 10.4 - 10.5, the calcium/carbonate
ratio most nearly approaches unity. Also, the solubility product constant
^SDCaoo corr)^ and saturation index (SI) calculated for the softener fil-
trate at 'each pH indicate that the softened cooling water was well within the
stability limits, and was not prone to precipitation. Calcium of free
calcium (which determines, in part, the saturation index) was based on the
average sulfate concentration and ionic strength of cooling water. The free
calcium concentration averaged 45 percent of the total calcium concentration
of the filtrate. It is probable that a portion of the organic compounds
exist as acids which contribute to the total alkalinity of the water, and
must be accounted or in alkalinity analyses in order to arrive at accurate
measurements of carbonate and bicarbonate.
Calcium Removal and Treatment Costs
The effect of calcium removal on the economics .of sidestream softening
operation can be seen by relating calcium removal to softener effluent
quality and sidestream flow rate for a given cooling water. The relation
between softener effluent quality and softener flow rate is derived from a
mass balance around the cooling tower (Figure 2-1). As explained in Section
2, this results in the equation
Qsc = QdCw + QsC
or
ss dw sw
Cw -
Hence, for a given makeup water quality and flow, the flow from the cooling
tower to the softener decreases as removal of calcium increases.
Operating costs (i.e., chemical costs) were calculated for various soft-
ener flow rates based on maintaining a cooling water calcium concentration of
267 mg/L. Constant flows and calcium concentrations were assumed as listed
in Figure 4-11. Results of these computations are presented in Table 4-3.
Lime and carbonate dosages were presumed to result in the same rates of
removal as in jar tests 12 - 16, where initial calcium was somewhat higher
(324 mg/L as CaC03).
84
-------
On the basis of those assumptions, costs were calculated with calcium at
as Ca(OH)2, and soda ash at llilv as Na2C03 and plotted as a function of
percent calcium removal in Figure 4-12. Since an increase in calcium removal
rate results in a decrease in softener flow rate, a minimum cost was obtained
for lime addition at the point where further addition of lime yields only
minimum reductions in softener flow. On the other hand, soda ash dosage
increases proportionately with calcium dosage, so no minimum was discovered.
CONCLUSIONS
Softener efficiency can be improved by basing the soda ash dosage on
inorganic carbon alkalinity, as determined by the TOC analyzer. Inorganic
carbon represents only 20 to 25 percent of the total carbon in cooling water,
hence inorganic carbon alkalinity values are generally lower than methyl
alkalinity values. As a result, the soda ash dosages calculated on the basis
of IC-Alk are higher, and the calcium removal is improved.
From an operating cost standpoint, cost is proportional to the rate of
calcium removal. When a given system was theoretically operated at the opti-
mum pH for calcium removal (pH 10.4 - 10.5), operating costs escalated with
improved calcium removal, despite the smaller softener flows which resulted
from better quality softener effluent. This was chiefly due to the increased
cost of soda ash which accompanied the addition of greater quntities of lime,
since an optimum lime dosage was obtained when the softener efficiency
reached 75 percent.
The most convenient parameter for estimating softener efficiency is the
ratio of calcium to carbonate in the softener effluent, which approaches
unity under conditions of maximum efficiency.
Removal of Silica by Sidestream Softening
Silica (silica dioxide, Si02) is one of the scale-forming constituents
in cooling water. Research has been performed to explain the mechanism of
its removal by softening, and several models have been proposed to account
for silica adsorption onto magnesium hydroxide floe. In this study, the
effects of different process variables upon silica removal are examined,
specifically:
1) Source of magnesium precipitate
2) Proportion of silica to magnesium in the softener
3) Softener pH
4) Softener temperature
5) Concentration of total suspended solids (TSS in the softener)
In addition, results of laboratory studies of USS Chemicals cooling
water and actual full-scale softener data were compared with both the Freund-
lich and Knight ("modified Freundlich") models for silica removal by adsorp-
tion onto magnesium hydroxide floe.
85
-------
TABLE 4-3. Cost of limp and Soda Ash Based on Reduced Flow to Softener
t
Percent
Removal
50
55
60
65
70
77
80
85
88
[Ca2+]e
138
124.20
110.4
96.6
82.8
63.43
55.2
41.4
33.12
GPM
682.61
620.55
568.85
525.08
486.57
443.25
426.63
401.53
387.85
Line
(ppm
€3(01)2)
390
398
407
420
436
465
485
580
680
Line Cost
0.067
0.061
0.058
0.055
0.053
0.051
0.052
0.055
0.063
Soda Cone.
(ppm
Na2C03)
75
95
117
145
175
235
271
360
485
Soda Cost^
0.047
0.054
0.061
0.070
0.078
0.096
0.106
0.132
0.173
Total Cost ^
0.114
0.115
0.119
0.125
0.131
0.147
0.158
0.187
0.236
t 7+
Mass balance around cooling tower was computed by maintaining Ca concentration at
276mg/L as
Costs in $/gal cooling water treated
86
-------
500r
oo
A Na2C03ADDED
100 20O 300 400 500 600
LIME ADDED (mg/L as CaC03)
700
Figure 4-11. Effect of linie added on soda ash dosage and percent calcium removal,
-------
25.0
20.0
15.0
10.0
5.0
SOFTENER FLOW RATE (GPM)
TOTAL CHEMICAL COST (c/min.!
A LIME COST U/min.) .
• — -A SODA ASH COST (t/min.)
I I I
700
600
500
400
300
200
100
10 20 30 40 50 60 70 80 90 100
% REMOVAL
Figure 4-12. Effect of percent calcium removal on sidestream
softener flow rate and treatment cost.
88
-------
THEORY
The dissolution and disposition of silica in water involves hydration
and dehydration reactions, as described below [4]:
hydration
(Si02)x + 2H20 -H
dehydration
Above pH 9, Si(OH)4 first ionizes to form Si(OH)30~, then (at higher pH) it
forms Si(OH)202 . The relevant equilibrium constants (at 25°C) are as
follows:
[H+] Si(OH)oO~ ,n
K, = 3 = 1.58 x 10"10 PK, = 9.8
1 [Si(OH)4J F l
17
K2 = _ _ 2 = 1.58 x 10 1Z pK2 = H.8
[Si(OH)30"]
Silica scale differs from other precipitates (such as CaCOo and CaS04)
in that its solubility increases with increasing temperature. As mentioned
earlier (Section 2) relation is described by the equation
Si02 (mg/L) = 4.7 T + 24
where Si02 indicates the maximum concentration of silica which will remain in
solution at a given temperature T (°C) [5].
Silica scale is formed by polymerization of dissolved silica which
results in an amorphous colloid. This colloid has a high affinity for di-
and trivalent cations, which are nearly always found in scale samples. the
lime softening process takes advantage of this affinity by exposing the
silica to ionic magnesium (Mg^+) at high pH. As a result, one reaction which
has been proposed to describe the mechanism of silica removal is as follows:
Si(OH)4 + (OH") t Si(OH)30~ + H20
Si(OH)30" + Mg(OH)2 t Mg(OH)2 Si(OH)30~
Previous studies also indicate that the removal of silica by magnesium
hydroxide floe is better characterized as adsorption, rather than a stoichio-
raetric chemical reaction. As such, it is often described by the Freundlich
isotherm model
89
-------
q = £ = kC1/n
m
where k and n are constants, q is the adsorption of silica onto magnesium
hydroxide floe on a weight/weight basis, and C is the final concentration of
silica in solution.
Empirical data of silica adsorption applied to the Freundlich model is
usually analyzed according to the linearity of a plot of log (x/m) versus log
C. According to the linearized Freundlich equation log, ((x/m) = log K +
1/n) log C. However, Knight [4] has proposed a modification of this model,
replacing the final concentration of silica (C) with the "log mean concentra-
tion" term (LMC) such that
(i) = k1 (LMC)3'
m
where
ci - cf
LMC = —
Cf
^ = initial silica concentration
If = final silica concentration
Knight's model is particularly useful, in that it accounts for any variation
of adsorption intensity which may occur as the initial silica concentration
is depleted during the course of the adsorption process. Its main drawback
is that when the equation is used to predict the extent of adsorption, the
final silica concentration must be estimated in advance.
In Figure 4-13 [7], shows silica concentration in cooling water effluent
as a function of magnesium concentration prior to reaction.
MATERIALS AND METHODS
Four different kinds of jar tests were designed to study different
aspects of silica removal. The first two tests involved laboratory-prepared
solutions of silica and magnesium in deionized water (DI H20), while the
second two tests were conducted with actual cooling water obtained from USS
Chemicals. In addition, full-scale tests were performed by varying condi-
tions at the softeners of USS Chemicals.
Experiments with Deionized Water
The first jar test was designed to determine the relative effectiveness
90
-------
O
§
I
40
38
36
34
30
28
26
24
22
20
o
38.6 %
O
9.0 92 9.4 9.6 9.8 10.0 102 10.4 10.6 10.8
PH
Figure 4-13. Effect of softener pH on silica removal.
91
-------
of different compounds as sources of magnesium for silica removal. In this
study, separate solutions of 60 mg SiOo/L (from sodium metasilicate) were
treated with 50 mg/L magnesium from MgO and Mg(OH)2» respectively. The lime
softening process was simulated as follows:
1) 500-ml aliquots of silica solution were placed in 2-L beakers, to
which were added magnesium from either compound.
2) Calcium was added to raise the pH to 10.4; soda ash requirement was
calculated according to the method described earlier.
3) The solutions were stirred continuously for 45 minutes, after which
they were allowed to settle for 15 minutes.
4) The flasks were decanted and filtered through a 0.45 m Millipore
membrane filter, and the filtrate was analyzed for relevant constit-
uents, principally silica and magnesium.
Initial and final silica concentrations were determined by the molybdo-
silicate blue method; analyses were performed with a B&L Spectronic 20 Spec-
trophotometer. Initial and final magnesium concentrations were determined by
the EDTA method.
Experiments with USS Chemicals Cooling Water
The second set of jar tests was designed to develop isotherms for silica
adsorption onto magnesium hydroxide floe. Silica removal followed the same
procedure as the previous test, except that only Mg(OH>2 was used as a source
of magnesium, and the magnesium concentration was maintained at 16.75 mg
Mg/L. Initial silica concentration varied from 40 to 120 mg
In the third series of jar tests (performed with USS Chemicals cooling
water), lime and soda ash dosages were varied in order to determine the
effect of different final pH on silica removal. In this case, the source of
magnesium was the softener sludge which had an average magnesium content 12.7
mg/L. After addition of lime and soda ash, the procedure followed was ident-
ical to that followed above. An average analysis of USS Chemicals cooling
water used in the jar tests is presented in Table 4-4.
Finally, the fourth test was designed to determine the equilibrium iso-
therms for silica adsorption from cooling water, and involved spiking the
cooling water with different amounts of silica to an initial concentration
varying from 43.8 to 163.8 mg SiC^/L.
It should be noted that the initial laboratory tests which were per-
formed with USS Chemicals cooling water yielded results which were not easily
reproduced, due to extensive spectrophotometric interference by the intense
yellow color of the water. Subsequent analysis indicated that the interfer-
ing color was produced by an average of 25 mg CrO|~/L and organic acids equal
to about 15 percent of total alkalinity. In order to eliminate this inter-
fer-ence, all cooling water used in laboratory tests was treated with ferrous
sulfate (FeSO^*7H20) and powdered activated carbon (PAC), prior to silica
removal.
In addition to the laboratory-scale tests, full-scale studies were per-
92
-------
TABLE 4-4. Average Initial Characteristics of
the USS Chemicals Water for Jar Tests
pH: 8.1
Alkalinity
P = 0
M = 377 ppm as CaCO-j
Hardness
Total 357 mg/L as CaC03
Ca2+ 305 mg/L as CaC03
Mg2+ 52 mg/L as CaC03
Silica (Si02)
44 mg/L
93
-------
formed by measuring influent and effluent silica and magnesium at the North
and South Softeners of the USS Chemicals plant under a variety of conditions.
In the first study, silica removal was determined in the presence of various
concentrations of suspended solids—from 22,000 to 59,000 mg/L—in order to
observe the effect of TSS on the efficiency of silica adsorption. Also,
influent and effluent concentrations of both silica and magnesium were mea-
sured under stable conditions to determine the optimum ratio of silica to
magnesium with respect to final silica concentration. Finally, variable
speed mixers were installed in the North Softener, and the effects of various
mixing speeds were examined over a period of four months.
Determination of the effect of temperature on silica removal was
obtained by a comparison of results reported in the literature by various
researchers, whose referenced works should be consulted for procedural
details.
RESULTS AND DISCUSSION
Effects of Different Magnesium Sources on Removal of Silica
Results of the silica removal jar tests performed with different magne-
sium compounds are summarized in Table 4-5. As can be seen, no real differ-
ence between Mg(OH)2 and dolomitic lime (Ca(OH)^'MgO) was observed with
respect to removal of silica. However, when magnesium must be added to cool-
ing water to remove silica, Mg(OH)2 may be preferred to dolomitic lime, since
the latter introduces excessive quantities of calcium. This additional
calcium may in turn require additional soda ash for effective softening,
thereby increasing alkalinity. Furthermore, as discussed above, when the pH
is raised above the optimum for silica removal, SiC>2 may redissolve, increas-
ing the silica residual.
Effect of pH on Removal of Silica
Since the solubility of silica increases with pH, it was suspected that
there should be an optimum pH at which the highest percent removal of silica
occurs. Jar tests conducted at 25°C with 20,000 mg/L suspended solids from
softener sludge as a source of magnesium (Figure 4-15) indicate that pH 10.45
was the optimum pH for removal of silica from USS Chemicals cooling water.
Effect of Total Suspended Solids on Silica Removal
A four-month study of full-scale sidestream softening operations at USS
Chemicals included continual monitoring of the softener suspended solids
(TSS), as well as influent and effluent concentrations of silica and magne-
sium. Results of this study, as regards the effect of suspended solids on
silica removal, are summarized in Figure 4-16. As expected, increasing the
TSS resulted in increased removal of silica, since the suspended solids con-
tained approximately 15 percent magnesium (by weight).
However, while full-scale removal improved as the TSS concentration
increased to nearly 60,000 mg/L, the percent removal obtained in the labora-
tory at 20,000 mg TSS/L was roughly double that obtained in the full-scale
plant at the same level of suspended solids.
94
-------
TABLE 4-5. Comparison of Magnesium Source for Silica Removal
Magnesium
Compound Used
Initial
Mg
ppm as
Final
Mg
ppm as
Initial Final
Si02
ppm
Si02
ppm
Percent
Si02
Removed
MgO
Mg(OH)2
50
50
0
0
60
60
29
31
52
48
95
-------
cr>
Q
O
UJ
a:
CD
O
100
90
80
70
60
50
40
30
20
10
5 10 15 20 25 30 35 40 45 50 55 60 65
SLUDGE CONCENTRATION ~- x |0B
Figure 4-14. Effect of softener sludge TSS on silica removal.
-------
70
65
60
55
.^ 50
o>
~ 45
_ <»
*& 40
(
35
30
25
20
15
10
5
o
o
y = 12,14 X
1^=0.88
CL8S
I I I I I
O
0.5 1.0 1.5 2.0 25 3.0 3.5 4.0 4.5 5.0 5.5 6.0
LMg J
Figure 4-15. Effect of silica/magnesium ratio on silica removal.
Mgi = 12.5 mg/L T = 112 °C pH - 10.4
97
-------
Effect of SiQ9/Mg Ratio on Silica Removal
In Figure 4-17, effluent silica concentration is plotted as a function
of the ratio of silica (as rag SiC>2/L) to magnesium (as mg Mg/L) in the
softener influent at USS Chemicals. As can be seen, as the SiC^/Mg ratio
increases beyond 2.5, more and more silica remains dissolved in the softener
effluent. Final selection of an optimum SiCU/Mg ratio requires both the
determination of maximum effluent silica concentration allowable and an
economic analysis of relevant costs.
Equilibrium Isotherms and Models for Silica Adsorption
Equilibrium isotherms, relating final silica concentration (mg SiOo/L)
to adsorption (mg Si02/mi Mg), are presented for laboratory jar tests with
both deionized water and USS Chemicals cooling water in Figure 4-18. As can
be seen, silica appears to be better adsorbed by magnesium in 'the USS Chem-
icals cooling water than in the deionized water samples, at least at higher
concentrations.
This increased adsorption is also reflected in the adsorption equations
derived from linearized plots of both Freundlich and Knight (modified Freund-
lich) models (Figures 4-19 and 4-20). Parameters of the derived equations
are presented in Table 4-6, where the higher 1/n and (1/n)1 values correspond
to an increased intensity of silica adsorption. A comparison of the correla-
tion coefficients for the different models also reveals a generally improved
fit of the Knight model to the empirical data.
A further comparison can be made between removal of silica from USS
Chemicals' cooling water in the laboratory, and adsorption obtained when
silica was removed in the full-scale softener at the USS Chemicals plant
(Figures 4-21 through 4-23). The exponential terms in both the Freundli-ch-
and Knight equations are greater for the adsorption isotherms derived from
full-scale operations, indicating that laboratory data may tend to underesti-
mate the adsorption that may actually be obtained in the field. One reason
for improved adsorption by full-scale methods may be the continuity of the
process in contrast to the non-equilibrium behavior of the batch tests.
Effect of Mixing on Silica Removal
At the beginning of 1982, two variable speed mixers were installed in
the North Softener at USS Chemicals, and softener performance was monitored
at different mixing speeds over a four-month period. While a full report of
variables investigated are presented below, equilibrium isotherms developed
for silica removal at various mixing speeds is presented in Figure 4-24.
While these curves depart from the standard isotherm forms derived
earlier in the laboratory and in the South Softener, it appears that greater
adsorption was obtained at higher rpm. However, further research in this
area is required in order to develop more consistent data.
98
-------
TABLE 4-6. Freundlich and Knight Isotherm Parameters
for Adsorption of Silica onto Magnesium Hydroxide Floe
Freundlich: qe = K(Ce)1/n
,(l/n)' Ci - Cp
Knight: qe = K'(LMC) where LMC = - e
In Ui)
C
Freundlich Knight
Water K 1/n r2 K (1/n)' r2
DI H20 .688 .255 .8843 .52 .305 .9051
USS Chemicals
Cooling Water '°98 «788 -9754 -091 -772 '5817
South
Softener -084 -898 -6994 -049 1-°° -7877
99
-------
CONCLUSIONS
In summary, experiments in the laboratory indicated that silica may be
removed equally effectively with either Mg(OH)2 or M§° as tne source of
magnesium. Silica is removed most efficiently when the pH in the softener is
raised to 10.4 - 10.5. Furthermore, removal of silica improves with
increased contact time and higher temperature. Studies performed with full-
scale softener operation show that silica removal is improved at higher
concentrations of suspended soliids which contain magnesium (i.e., softener
sludge). Also increased mixing speeds may tend to improve silica adsorption,
but this area must be studied further.
Isotherms developed from both laboratory-scale and full-scale experi-
ments suggest that, in general, adsorption in full-scale, continuous process
softeners exceeds that of laboratory batch reactors (i.e., jar tests). This
tendency should be considered whenever laboratory isotherms are applied to
full-scale processes for design purposes.
The Effect of Mixing Speed on the Softening Process
The zero-discharge sidestream softening system currently operating at
the USS Chemicals plant uses two Infilco softeners, called to as the North
and South Softeners. In this study, the effect of variable mixing in the
first chamber of the reaction zone was investigated by means of two turbine
mixers (Eurodrive Inc. Type REGO Dll BD1 80 N4C) installed in the North
Softener. As mentioned earlier, South Softener remained unmodified, as a
control, for the purpose of comparison.
MATERIALS AND METHODS
To determine the effects of mixing on the North Softener (as compared
with the slow-mixed, unmodified South Softener), grab samples were collected
from different sampling locations in both softeners on a regular basis
(Figure 4-25). All water samples were analyzed in the Environmental Engi-
neering Laboratories at the University of Houston according to Standard
Methods, 15th Edition.
Determination of the effect of mixing on softening efficiency was
divided into three different parts, and the corresponding work was performed
during three periods. During the first period (eight weeks), the various
parameters were measured every other day in both softeners and the plant
operators performed alkalinity, pH, and calcium hardness measurements four
times a day. The data collected were used as the baseline of comparison
between the modified and unmodified softeners. In the second period, the two
mixers were installed in the North Softener, and the plant operators were
trained to adjust the softeners so that both softeners received the same flow
rate, and dosages of lime and soda ash. During the third period, the effi-
ciency of the modified softener was tested at various first chamber mixing
rates. In these tests, the modified softener turbine mixer speed was varied
while the flow rate, and lime and soda ash dosages, recirculation rate, and
sludge blowdown for both softeners were kept the same. Four different mixing
speeds were tested—30, 40, 50, and 70 rpm. Each mixing speed was maintained
until steady state condition was reached, except for 30 rpm, since it was
found that at 30 rpm mixing was inadequate for a satisfactory equipment
response.
100
-------
6.0
5.0
4.0
(22.)
qe W
3.0
2D
1.0
0.5
O O Dl H20
Q—D uss CHEMICALS
; »l6.75mg/L
i i i
10 20 30 40 50 60 70 80 90 100
[Si]f (mg/L)
Figure 4-16. Equilibrium isotherm for adsorption of silica onto
magnesium hydroxide floe.
101
-------
0.70
0.65
0.60
0.55
0.50
0.45
0.40
lo Qe
0.30
0.25
0.20
0.15
0.10
0.05
D USS CHEMICAL COOLING WATER
O - O DEIONIZEDH20
0.20.40.6 0.8 1.001.2 1.4 1.6 1.82.002.2
log [Si02]e(mg/U
Figure 4-17. Linearized Freundlich isotherm for silica adsorption
(Jar Tests).
102
-------
log Qe
(mg/mg)
0.70
0.65
0.60
0.55
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
D USS CHEMICAL COOLING WATER
DEIONIZED WATER
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
log LMC
Figure 4-18. Linearized Knight isotherm for silica adsorption
(Jar Tests).
103
-------
O SOUTH SOFTENER
0.25
10 20 30 40 50 60 70 80 90 100
[Si]f (mg/L)
Figure 4-19. Equilibrium isotherm for adsorption of silica
onto magnesium hydroxide floe (USS Chemicals
south softener).
104
-------
0.7 r
0.6
0.5
0.4
logQe
(mg/mg)
0.3
0.2
O.I
o
o
o
I I I I I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
log[Si02](mg/L)
Figure 4-20". Linearized Freundlich isotherm for silica
adsorption (south softener).
105
-------
0.7
0.6
0.5
0.4
log Qe
(mg/mg)
0.3
0.2
O.lh
I _ I
A
A
I I i i
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
log LMC
Figure 4-2 L Linearized Knight isotherm for silica adsorption
(south softener).
106
-------
7.0
6.0
5.0
4.0
(mg/mg)
3.0
2.0
1.0 I-
O O 40 RPM
A- A 50 RPM
n 7°
10 20 30 40 50 60 70 80 90 100
[Si02]e(mg/l)
Figure 4-22. Silica adsorption iostherms obtained at various
mixing speeds (north softener).
107
-------
o
CO
Sludge
Effluent
oo
oo
Mixing Zone
Reaction
Zone
Influent
Schematic of Snmpllng Points In North Softener.
Figure 4-23 .
-------
RESULTS AND DISCUSSION
In order to analyze the efficiency of the better mixed North Softener,
the data collected during the third period of the work plan was used to plot
the effect of different mixing speeds on various operational parameters. For
most of the cases, the shapes of the graphs were determined by the averages
of several data points, since data was generally well-scattered. However,
for the sake of accuracy, all data points are plotted on certain figures, and
the averaged values are distinguished from the raw data presented.
Effect of Mixing Speed on Solubility
Product of Calcium Carbonate
The solubility product of calcium carbonate, Kg was calculated for
each grab sample according to the following calculations:
P.Alk = 1/2 (COp + (Si02) * + (OH")** all terms as mg/L CaC03
[0)3]* = 2(P.Alk - Si02 * |2-) * 10~5 moles/liter CO 3
[Ca2+] = Ca2"1" mg/L CaC03 * 10~5 moles/liter Ca2"1"
Ksp
**[OH~] was considered negligible at operating pH = 10.4
When the product of calcium and carbonate concentrations in solution exceed
the KS CaC03 will begin to precipitate; hence the lower the K value in
the softener, the greater is the efficiency of the softening reaction.
Solubility products of calcium carbonate in solutions mixed at different
rpm are plotted in Figure 4-26. Since the solubility product obtained for
solutions mixed at speeds from 30 to 70 rpm were higher than the typical
solubility product of calcium carbonate (Kg = 2.8*10"'), results indicate
that in all cases solutions were supersaturated with CaCOo. It is observed
that the lowest Kgp values were obtained at 30 rpm (1.67*10"°) and 70 rpm
(1.71*10"°). However, since mixing at 30 rpm was judged inadequate, 70 rpm
may be considered the optimum mixing speed at which the lowest KS_ will
occur. By comparison, the K value of the South Softener in which the only
motion in the reaction zone is given by the slow velocity paddles (2 rpm),
the K is equal to 2.18x10, indicating a substantially greater tendency
for calcium and carbonate to remain in solution.
Effects of Mixing Speed on Lime Requirements
Figure 4-27 is a plot of the North Softener mxing speed versus the South
109
-------
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
1.9
10
1 '••
1" 1.7
1.6
1.5
1.4
1.3
1.2
I.I
10
-
-
- .
!-
^ ^ _ _ j. t_ r* - f ±
Xboutn Softener
s 2 18 Paddles 2 rpm
s = o'.5. •
:/^\sS4ifxio-7
- s = 4.69x|0~y \
A
1 1 1 1 1 1 1 i i i »-
10 20 30 40 50 60 70 80 90 100
Mixing Speed, rpm.
North Softener K versus Mixing Speed.
Figure 4-24. Effect of mixing speed on calcium carbonate
solubility in USS Chemicals north softener.
110
-------
o
tr
CO
CO
CO
2.5
2.4
2.3
2.2
2.1
2.0
1.9
1.8
1.7
1.6
1.5
1.0
1.3
1.2
x = 2.47
s= 0.65
s = 0.48
7 = 1.85
7 = 1.65
s=0.35
x = l.67
s=0.44
10 20 30 40 50 60 70 80 90 100
Mixing Speed, rpm.
South Softener/North Softener Lime Ratio versus Mixing
Speed in North Softener.
Figure 4-25. Effect of mixing speed on softener lime dosage.
Ill
-------
Softener/North Softener lime ratio. This ratio compares lime requirements
for the South Softener to that of the faster mixed North Softener; a higher
SS/NS lime ratio indicates that less lime dosage was added to the North
Softener to obtain similar softening. Accordingly, 70 rpm is the optimum
mixing speed at which lime requirement in the North Softener was reduced to
the least amount. This phenomena is due to the fact that high mixing helps
to dissolve the lime particles, improving the softening process. By compari-
son, the minimum mixing in the South Softener does not dissolve the lime in
the water very well.
It has been estimated that the North Softener operated at 70 rpm
requires 45 percent less lime to accomplish softening comparable to the South
Softener, which had a minimum amount of mixing.
Effect of Mixing Speed on Effluent Turbidity
Figure 4-28 shows the average effluent turbidity in NTU units of
effluent from the North Softener operated at 30, 40, 50, and 70 rpm. the
results, which indicated minimum turbidity at 40 rpm, may be explained as
follows: (1) at low mixing speed (between 30 and 40 rpm), the process of
nucleation or formation of new crystals occurs; (2) at 40 rpm, a hindered
settling process takes place, and due to the settling of crystals, the tur-
bidity of the effluent decreases; (3) when the mixing speed exceeds 40 rpm,
the higher speeds tend to break up the crystals and create tiny particles
which do not settle well, and which are not efficiently removed by the
filters.
According to the phenomena observed in Figure 4-28, a mixing speed of 40
rpm is the most adequate to obtain less turbidity in the effluent.
Effect of Mixing .Speed on Sludge Recycle
Figure 4-29 shows the data for the South Softener/North Softener sludge
recycle rates as a function of mixing speed (rpm). Similar to the previous
SS/NS lime ratio, the sludge recycle ratio in the softeners also depends upon
the intensity of mixing speed. Although Figure 4-29 indicates that 30 rpm
results in the highest SS/NS sludge recycle rate, it was not considered the
optimum point in this particular experiment because, as was stated earlier,
mixing speed is inadequate for the equipment operation. Consequently, 40 to
50 rpm may be considered the optimum mixing speed to accomplish minimum
sludge recycle.
Effect of Mixing Speed on Percentage of Calcium Removal
The percentage of calcium removal is one of the most significant indi-
cators of softener efficiency. Figure 4-30 shows a plot of percentage of
calcium removal versus mixing speed. While it can be observed that the per-
cent calcium removal was significantly improved at higher mixing speeds (as
compared with the South Softener), there was no appreciable difference in
removal rates among the higher speeds.
112
-------
55
50
P 45
I 35
i_
^ 30
"c
§ 25
uj 20
15
10
5
0
I I I L
10 20 30 40 50 60 70 80 90 100
Mixing Speed, rpm.
North Softener Effluent Turbidity, NTU, versus Mixing Speed,
rpm.
Figure 4-26. Effect of mixing speed on softener effluent
turbidity.
113
-------
1.5
COICO
T '.4
a
tr
03
o
1.3
cn
r I.I
5? 1.0
0.9
0.8
0.7
0.6
10 20 30 40 50 60 70 80 90 100
Mixing Speed, rpm.
South Softener/North Softener Sludge Recycle Ratio versus/
Mixing Speed in North Softener.
Figure 4-27. Effect of mixing speed on softener sludge recycle
rate.
114
-------
Effect of Mixing Speed on Sludge Setting in Mixing Zone
Figure 4-31 is a plot of sludge settling in the mixing zone versus
mixing speed. As can be seen from that plot, the sludge settling increased
when the mixing speed increased from 30 to 50 rpm; however, less sludge
setting occurred when mixing speed increased from 50 to 70 rpm. These
results are comparable to those obtained for turbidity, since increased
settling results in decreased turbidity. Consequently, it seems that a
mixing speed of 70 rpm is preferable to producing the least sludge settling
in the mixing zone.
Effect of Mixing Speed on Silica Removal
The limited effects of mixing speed on silica adsorption were reported
above. As stated earlier, while no quantifiable trend emerged, increased
mixing seemed to increase maximum attainable rate of silica adsorption.
CONCLUSIONS
Mixing in the softener reaction zone is an important aspect of the soft-
ening process. A substantial improvement in softener performance was
obtained when the North Softener was fitted with turbine mixers capable of
maintaining faster mixing speeds than the South Softener. This was espe-
cially true of savings in lime required for the softening reaction.
It was also determined that once enough mixing was provided for good
process control (e.g., 40 rpm), no significant difference in softening effi-
ciency was obtained when higher mixing speeds (50 and 70 rpm) were applied to
the system. The results are summarized in Table 4-7.
115
-------
80
70
60
_ 50
o
o
0>
O
O
30
20
10
20
O
O
30
40
50
60
70
Mixing , Speed, rpm.
North Softener '/. Calcium Removal versus Mixing Speed, rpm.
Figure 4-28- Effect of mixing speed on calcium removal,
116
-------
40
0)
c
o
N
I 30
CO
20
i i
J I I I
10 20 30 40 50 60 70 80 90 100
Mixing Speed, rpm.
Sludge Settling (mixing zone, ml/L) versus Mixing
Speed (rpm).
Figure A-29- Effect of mixing speed on sludge settling in
the mixing zone.
117
-------
TABLE 4-7. Effect of Mixing Speed on Softener Operations
RPM
2
30
40
50
70
K
(xlO6)
2.18
1.67
1.92
2.02
1.71
Lime Ratio'
(SS/NS)
1.00
1.65
1.85
1.67
2.47
Turbidity
(NTU)
53
32
45
55
Sludge
Recycle^
(SS/NS)
1.0
1.3
1.0
1.02
0.90
Percent
Ca
Removed
60.9
68.0
63.4
65.4
63.5
Sludge
Settling
(mg/L)
27
27
28
31
24
^Indicates the ratio of the SS (South Softener) to the NS (North Softener).
For example, a ratio of 1.65 means that 69 percent more lime was used in the
South Softener.
118
-------
SECTION 4 - REFERENCES
1. Butler, James N. Carbon Dioxide Equilibria and their Applications.
Harvard university, Addison-Wesley Publishing Company, 1982.
2. McGaughey, L.M. and Matson, J.V. "Prediction of the Calcium Carbonate
Saturation pH in Cooling Water." Univrsity of Houston, Department of
Civil Engineering, April 1980.
3. Standard Methods for the Examination of Water and Wastewater, American
Public Health Association (Washington, B.C.),1980.
4. Iller, R.K. The Chemistry of_ Silica. Wiley and Sons (New York), 1979.
5. Matson, J.V. and Harris, T. "Zero Discharge of Cooling Water by Side-
stream Softening," Journal WPCF, 51 (11), pp. 2602-2614 (November 1979).
6. Knight, J.T. "Chemistry of Sidestream Softening and Silica Reduction,"
Journal Cooling Tower Institute, 2(2), p. 45 (1981).
7. Betz, L.D. Handbook of Water Treatment.
8. Betz, L.D.; Noll, C.A.; and Maguire, J.J. "Adsorption of Soluble Silica
from Water,"
119
-------
SECTION 5
MONITORING AND CONTROL OF BIOFOULING IN A
ZERO DISCHARGE SIDESTREAM SOFTENED COOLING SYSTEM
Recirculating cooling tower systems (RCT) that contain high concentra-
tions of organic matter present significant challenges to control of biofoul-
ing. Organic concentrations in cooling water are increasing because environ-
mental regulations and higher water costs are stimulating recycle schemes
involving cooling water systems. An RCT in a plant that reuses much of its
process water as makeup, and also treats and reuses its blowdown will have
high organic concentrations that will stimulate biofouling.
USS Chemicals, Houston, Texas, began operating a recycle/reuse RCT
system in 1979. Effluent streams, including stormwater, process wastewater,
and boiler blowdown, were added as makeup to the RCT system. The only
effluent streams not reused were the demineralizer regeneration and rinse
waters. Blowdown was treated in a lime softener for calcium, magnesium, and
silica removal and recycled to the RCT system.
However, biofouling was such a problem that by January, 1981, production
was limited by the resulting loss of heat transfer. Organic levels, in terms
of total organic carbon (TOG), had increased to approximately 500 rng/L. In
contrast, most RCT systems rarely rise above 50 mg TOC/L. Conventional bio-
fouling control methods were not effective so a different approach was
attempted in which a brominated biocide ws used in conjunction with a more
traditional chlorinated control.
MATERIALS AND METHODS
To determine the effectiveness of the various biocidal treatments, a
fouling monitor system was installed on a slipstream from the cooling water
return line to monitor biofouling. Its purpose was to simulate conditions in
the main system and provide a real time readout of the extent of biofouling
as measured by reduction in heat transfer.
In the plant control room, surface condenser vacuum was monitored.
Although influenced by a variety of factors, the vacuum reading was sensitive
to biofouling, and rapid decreases in vacuum (in the absence of other causes)
were directly attributable to biofouling.
The total halogen residual in the cooling water was measured six times
per day by the DPD colorimetric method [3]. The TOC was analyzed daily.
Microorganism counts in the cooling water were also conducted on a daily
basis.
120
-------
The Fouling Monitor System
The fouling monitor system consisted of two major components:
1. A 0.0127 m (0.5 in) I.D. carbon steel tube containing ports (for
pressure drop measurement) and a heat transfer section (consisting
of an electrically heated block which was clamped around the tube).
(Figures 5-1 and 5-2)
2. A microcomputer to calculate frictional resistance and overall heat
transfer coefficient.
Temperature probes inserted in the heated block determined the radial
temperature profile; the system also included a flow meter and bulk water
temperature probes. Both the pressure drop and entrance length sections were
heated to match surface temperature conditions in the heat transfer section.
Output from the microcomputer was displayed on a television monitor and
included all pertinent measurements and calculations. A cassette recorder
stored the data. The system is diagrammed in Figure 5-3.
To simulate a composite heat exchanger in the plant, the fouling monitor
was operated at the following conditions.
1) Flow rate was maintained at 0.90 m s~* (3 fps) by constant head feed
tank with a manually controlled valve;
2) Feed water temperature was maintained at 40°C by a manually con-
trolled heat exchanger; and
3) Electric power to the block was maintained at 100 watts
Data was averaged over each hour of operation and recorded on an hourly
basis. Output consisted of the parameters listed in Table 5-1.
Data and calculations were recorded on cassette tape and transcribed at
a later date. Manual determinations of temperatures and pressure drop were
conducted approximately every three days as a check of computer results and
to anticipate a computer failure. Bulk water and block temperature were
measured manually using a Yellow Springs Instrument Tele-Thermometer and
pressure drop was measured with an inclined mercury manometer.
Overall Heat Transfer Coefficient
Overall heat transfer coefficient is defined by
U = 9
2Lrii (T2 ' TB)
*W.G. Characklis and Associates, Consulting Engineers, 516 West Cleveland,
Bozeman, Montana 59715
121
-------
N)
*?
i —'
„§
o 3
•o
1.27cm
-i^-nv
;
TLMPF.RATUHE PROBES
7, - -JJUINUM BufiCK AT RADIUS = l.5Ocm
',-.- ALUMINUM ULOCK AT RADIUS = 7.37cm
ifi, c)l.'LK WATER TEMPERATURE AT INLET
re-.- ij-ji.K WAIER TEMPERATURE AT OUTLET
"»;(' .Milt lli!AjlOtl('ll
cui;>on sii'o! oij.e
tdic. * I 57 cm)
- pressure poil 1
fi '
i——-^ni/—T
h
U1.8 cni-
-J I ru2
I j ^pressure porl 2
12.7cm
dirbcliun o) How
TUBULAR FOULING MONITOR SYSTEM
Figure 5-1. Apparatus for monitoring biofouling in the USS Chemicals cooling system.
-------
NJ
CO
CTXJ
m m
\
\
6.35c
m]
IG.IBcn
R, = C.95 cm
R2~- 8.09cm
Rj = 1.50cm
R = 7. 37 err.
L- -12.70cm -i i
i I i
ALUMINUM TEST HEAT EXCHANGER
Figure 5-2. Detail of biofouling monitor apparatus showing aluminum heat exchanger.
-------
FOULING MONITOR
MICROCOMPUTER
OUTPUT
N>
BLOCK
If-
BULK WATER
TEMPERATURE
iciwr
HEATER.
C.I
^
X
$
\
\
\
x
x
x
\
s
s,
1 Ml U
^i
1 1
• .
1 •
VW^>S
FLOW METER/1
a V
ni
!
"l
U
**
~4
w^
>,
^
J
\
X
X
X
X
X
X
^5^
PRESSURE
DROP
rBULK WATER
TEMPERATURE
FLOW CONTROL!
i
\
CALCULATE.
FRICTIONAL RESISTANCE
OVERALL HEAT TRANSFER
RESISTANCE
CONVECTIVE
CONDUCTIVE
CONTROL:
FLOW RATE OR PRESSURE
DROP
HEAT FLUX OR SURFACE
TEMPERATURE
Fouling Monitor System
Figure 5-3. Process diagram of biofouling monitor system.
CASSETTE STORAGE
-------
TABLE 5-1. The Output of Fouling Monitor/Microcomputer System
FOULING MONITOR MEASUREMENTS
-Bulk water temperature at the tube
inlet, TB1
-Bulk water temperature at the tube
outlet, TB2
-Block temperature at two radii,
Tl, T2
-Flow rate, F -
-Pressure drop, Ap
MICROCOMPUTER CALCULATIONS
-Overall heat transfer coefficient, U,
using a heat flux calculated from the
thermal conductivity of the block and
the difference between T^ and T2
-Friction factor, f, from pressure drop
and flow rate data
125
-------
TABLE 5-2. Abbreviations and Symbols
NOTATION
d = Cube diameter
f a friction factor
F = flow rate
g = gravity
k 3 thermal conductivity of block
L =» length of heat transfer section
L = length between pressure ports
q = applied heat
r. = block radius at 0.0150 m
i
r.. = block radius at 0.0737 m
11
T = block temperature at r.
T = block tempertaure at r..
TB =* average bulk water temperature
TB 3 bulk water temperature at the tube inlet
TB = bulk water temperature at the tube outlet
(J = overall heat transfer coefficient
U = overall heat transfer resistance
v a mean fluid velocity
Ap = pressure drop
pf = fluid density (water)
PH = density of mercury
126
UNITS
(m)
(dimensionless)
Cms"1)
Cms"2)
(m)
(m)
(W)
(m)
(m)
(W
(w"1m2°C)
(m s"1)
(m Hg)
(kg m"3)
(kg m"3)
-------
Overall heat transfer resistance, U~ , is the inverse of U. The heat flux,
q, was maintained constant and was calculated from 1.2 anc* ^1 and tne thermal
conductivity of the heated block.
_ 2 L k (T2 - TL)
In (rii/ri)
Frictional Resistance
Frictional resistance is characterized by the dimensionless friction
factor.
2 d
Hg
L PC v2
p f vm
where o^g is the specific gravity of mercury, P£ is the fluid specific
gravity, I, is the pressure drop length, and vm is linear mean velocity (see
Table 5-2 for an explanation of symbols and
Sampling Method and Analytical Techniques for Deposit Characterization
Two 0.05 m length sections of tubing, in series with and identical to
the heat transer section tubing, were removed with a pipe cutter; one tube
was preserved in a 100 mg/L HgCl2 solution to prevent biological growth and
the other stored in a dry container.
The preserved sample tube was analyzed for total mass by removing the
attached mass and drying for three hours at 103°C. The sample was cut in
longitudinal sections to observe pitting and corrosion and stored, with
dessicant, for comparison with later samples.
The bacterial count was determined by removing the biomass with a rubber
policeman, diluting to 25 mL and then performing a standard plate count [1].
The areal bacterial density was determined by dividing the total number of
bacteria on the sample tube by the inside area of the tube. The standard
plate count does not permit development of the more fastidious aerobes or
obligated anaerobes and other potentially important aquatic bacteria.
Identification of bacteria was performed by observing morphological
characteristics of different colonies (i.e., gram stain, cell size, motility,
and endospores) and by observing cultural characteristics of macroscopic
appearance on growth media (i.e., amount of growth, color, opacity, and
form). Further identification was made using indole, oxidase, citrate,
glucose, an4 catalase tests. The standard MPN procedure [1] was used to
determine presence of coliform bacteria.
127
-------
Control Strategy
The chemicals originally used for biofouling control were chlorine, a
surfactant, and a non-oxidizing biocide. Chlorine was added continuously to
the RCT by bubbling the gas through the water in the cooling tower basin.
Free chlorine residual levels were maintained in the 0.05 - 0.2 mg/L range.
The surfactant (chlorine helper) was also added continuously. The non-
oxidizing bLocide (a quaternary amine) was batch fed to the system three
times per week. In response to excursions in which heat transfer was
reduced, additive rates were increased to maximum levels.
The alternative control strategy was the use of Bromocide®, (bromo-
chlorodimethylhydantoin, or BCDMH) manufactured by Great Lakes Chemicals, as
a complement to gaseous chlorine. Gaseous chlorine mixes with water to
produce hypochlorous acid according to the following reaction:
C12 + H20 £ HOC1 + H"1"
The product, hypochlorous acid, in turn establishes an equilibrium with the
hypochlorite ions
HOC1 t OC1~ + H+
as described by the equation
K = [H+HOCI-]
[HOC1]
The pK of the reaction, which equals 7.5 at low ionic strength and
temperatures of 20°C, decreases at higher ionic strengths and temperatures.
In the cooling water at USS Chemicals, the pK approaches 7.0 at a temperature
of 30°C and ionic strength of 0.5. Thus, the predominant species is the
hypochlorite ion.
However, the hypochlorite ion is a much stronger oxidizing agent than
hypochlorous acid, and dissipates quickly with organic material. More impor-
tantly, the hypochlorite ion is a less effective biocide because the micro-
organisms can more easily repel the charged ions. Conversely, the uncharged
hypochlorous acid can diffuse readily into a microbial cell [2].
Bromocide* was chosen as an alternative to gaseous chlorine since it is
a more effective biocide in the pH range encountered at USS Chemicals. (7.6
- 7.8). Bromine chemistry is analogous to chlorine chemistry in that the
bromine will form hypobromous acid and hypobromite ion. Moreover, the pK is
approximately one pH unit higher than chlorine, i.e., pK^ = 8.63 [3]. There-
fore, the predominant species of bromine in the USS Chemicals cooling water
is hypobroraous acid.
128
-------
RESULTS AND DISCUSSION
Debugging of the biofouling monitor system was completed in late 1980.
In January, 1981, the comprehensive monitoring system commenced operation,
including analysis of halogen residual, viable cell counts, and measurement
of condenser vacuum was in operation.
Before Bromination
During the first thirty days, the condition of the cooling water dete-
riorated as shown in Figure 5-4, which indicates that the surface condenser
vacuum decreased from a high of 25 inches of mercury to a rough average of
21. (A decrease in vacuum indicated intensified biofouling.) In an attempt
to counteract the biofouling, increased amounts of chlorine were added to the
cooling water system. As shown in Figure 5-5, the free residual rose to an
average of 0.5, but the biofouling condition persisted.
Section of heat exchanger tubing was wthdrawn on January 24 (Figure 6).
The fouling deposit was concentrated on the bottom of the tube. Corrosion
and pitting was evident on the tube surface after the deposit was removed.
Dry mass of the fouling deposit (103°C for 3 hours) was 32.38 g cm . A
bacteria count revealed approximately 380 organisms per cm^ of the tube wall
surface. The bacterial count was conducted several days after the sample was
removed and should be considered as an estimate.
Preliminary investigation of types of bacteria revealed the presence of
Actinomycetes, Bacillus, Flavobacterium, Moraxella, and Alcaligenes. No
coliform organisms were present.
After Bromination
On day thirty, the chlorinators were shut down and the BCDMH was added
to the system at a rate of 100-150 Ibs/day. The results of the BCDMH addi-
tion are shown in Figure 5-6, which indicates the reduction of fouling by the
change in heat transfer resistance and frictional resistance, respectively,
in the fouling monitor. Heat transfer resistance increases steadily due to
fouling, up to approximately day ten (actually day 30 of the test). After
that, heat transfer resistance decreases to approximately clean conditions.
The decrease in heat transfer resistance begins immediately after the change
to BCDMH treatment.
Figure 5-6 also indicates the progression of friction factor. The
results are almost identical to heat transfer resistance. Friction factor
increased steadily until BCDMH was added, then the friction factor decreased
to clean conditions.
For the next thirty-day period, BCDMH was added with the surfactant and
non-oxidizing biocide. As Figure 5-4 shows, the condenser vacuum gradually
increased into the range of 23 to 24 inches of mercury; there was also visual
129
-------
U)
o
•5
E
§
26
25
23
22
21
20
19
30
60 9O 120
Time (days)
Surface Condenser Vacuum as a Function of Time
150
Figure 5-4. Variations in surface condenser vacuum of the biofouling monitor apparatus.
-------
2.5
~ 2.0
o>
c
0>
Oi
o
o
I
1.0
05
0
30
120
Figure 5-5.
6O 90
Time (days)
Free Halogen Residual for Recirculating Cooling Water as a Function of Time
Variations in cooling water free halogen residual.
150
-------
evidence of biofilm reduction. Stringers on the cooling tower slats gradu-
ally disappeared. The TOG and viable cell count measurements were not indi-
cative of biofouling conditions in the cooling water system. The TOG actu-
ally increased during the month of February as shown in Figure 5-7. However,
the increase may have been an artifact of the cleanup operation rather than
evidence of higher organic loadings to the system. The viable cell counts
showed no sensitivity to biofouling or the resultant cleanup operation, as
shown in Figure 5-8.
Starting at day ninety, BCDMH was blended with chlorine on a 1:10 weight
ratio basis which was roughly 30 lbs:300 Ibs/day to decrease the cost of
biofouling control. The cooling water system was considered to be in good
condition. Over the next ninety-day period, an operational strategy evolved
in which the free halogen level was gradually increased by step changes from
0.25 to 0.45 at 0.05 mg/L intervals. The goal was to establish minimum
residual level at which biofuling incidences were negligible. At 0.45 mg/L,
this level was achieved. The condenser vacuum gradually increased and sta-
bilized at an acceptable level.
CONCLUSIONS
The bromine compound BCDMH was effective in the control of biofouling in
a high pH, high organic content cooling water. In conjunction with chlorine,
it produced a free halogen residual which proved an effective strategy in
biofouling control.
Furthermore, the biofouling monitor was a sensitive, real time indicator
of biofouling. While condenser vacuum was a key plant parameter for the
detection of biofouling, viable cell counts and total organic carbon did not
correlate with biofouling conditions.
132
-------
-J 1=
< z
SS
r- O
DC
5 K
LU g
DC
r-
DC
iii
LU , .
LL UJ
co o
z z
DC H
L± CO
p~ il*
H W
< m
LU OC
I
Z
O
1-
o
DC
LL
10'
£ 5
* lS
^^ 3
10°
1 V^
26
LU 24
1-
DC 22
20
0.025
X"N
J^
>
*"- 0.015
0
CVJ
^E
0.005
(E °-°15
o
H 0.01C
0
if 0.05C
i i i i i
^^"^» ^BW ^^ ^^^ '^ ^^^P 1 /
o before biocide
1111 &
• '•I i
0
•«• •»•***
*\ •*
V***x
1 1 1 1 1
1 v I ' ' 1
•
* ^^
— ^
1
/\ ' ' '•-^•*
)- •" •• *
II 1 I 1
0 10 20 30 40 50 60
«- new biocide
TIME (days)
Heat Transfer Resistance and Friction
Factor as a Function of Time
Figure 5-6. Variations in friction and heat transfer resistance
as indications of fluctuations in biofouling
133
-------
OJ
1000
900
800
700
5 600
o»
J 500
o 400
P 300
200
100
0
30
60 90
Time (days)
120
150
Figure 5-7,
Cooling Water Total Organic Carbon as a Function of Time
Variations in cooling water TOG.
-------
OJ
t_n
o
o>
o
§ 3
is
30
120
150
Figure 5-8.
60 90
Time (days)
Cooling Water Viable Microbial Cell Counts as a Function of Time
Variations in viable microbial cell counts in USS Chemicals cooling water.
-------
SECTION 5 - REFERENCES
1. Standard Methods for the Examination of Water and Wastewater, 15th Ed.,
American Public Health Association (Washington, D.C.), 1980.
2. Rice, J.K. Drew Principles of Industrial Water Treatment, Drew Chemical
Corporation (Boonton, New Jersey), p. 100 (1977).
3. Smith, R.M. and Martell, A.E. Inorganic Complexes, Vol. 4 of Critical
Stability Constants, Plenum Press (New York), p. 134 (1981).
136
-------
SECTION 6
TREATMENT OF CHROMATE IN SOFTENER SLUDGE AND COOLING TOWER SLOWDOWN
Of all the substances used routinely in the sidestream-softened zero-
discharge cooling system, chromate is the most toxic and the most problematic
in terms of its ultimate disposal. Chromate is added to cooling water in the
hexavalent form (Cr(VI)) which is also the most toxic. Although Cr(VI) is
not removed by the softening reaction, some portion of soluble chromate
(perhaps as Cr(lII)) does in fact precipitate into the softener sludge with
calcium carbonate and magnesium hydroxide. Careful evaluation of softener
sludge for chromium leachate potential can determine the extent of landfill
containment required for proper disposal of chromate-bearing sludge. Fur-
thermore, in the event that blowdown of chromate-treated cooling water is
required, chromate may be successfully removed by precipitation with ferrous
sulfate (FeSO^) prior to discharge.
EVALUATION AND HANDLING OF CHROMATE LEACHATE FROM SOFTENER SLUDGE
Disposal of chromate-bearing sludge from the sidestream softener is
regulated by state and federal waste disposal authorities, including the US
Environmental Protection Agency (EPA) and, in the state of Texas, the Texas
Department of Water Resources (TDWR). This report deals primarily with the
toxicity tests specified by those agencies, which determine both the extent
to which chromate-bearing sludge is hazardous, and whether it must be placed
in a permitted, secure, hazardous landfill , or a non-hazardous landfill.
In 1976, the Resource Conservation and Recovery Act (RCRA) was passed by
Congress, and in May, 1980, Subtitle C of the Solid Waste Act was amended by
RCRA "...to promulgate regulations to protect human health and the environ-
ment from improper management of hazardous waste..." Subtitle C of RCRA also
established a federal program which, when fully implemented, will provide
"cradle-to-grave" regulation of hazardous waste. Although parts 261-265 of
RCRA collectively contain the first phase of EPA's regulations carrying out
these directives, this report concerns only subpart C ("Characteristics of
Hazardous Waste") Section 261.24 ("Characteristics of EP Toxicity"), and
Appendix II ("EP Toxicity Test Procedure"). The current maximum concentra-
tion of chromium allowable in non-hazardous leachate is one hundred times the
National Interim Primary Drinking Water Standard of 0.05 mg/L, or 5.0 mg Cr/L
leachate by the EP toxicity test.
Also in 1976, TDWR issued a series of nine technical guides dealing with
industrial solid waste management. These technical guides defined wastes
according to their physical and chemical characteristics, and offered
137
-------
criteria for Che selection and siting of treatment alternatives, including
ponds and lagoons, landfarms, and landfills. In addition, they specified the
necessary monitoring and recordkeeping systems required by state law for
continuous handling of hazardous waste. The current TDWR maximum chromium
concentration allowable in leachate is the same as the federal standard, 5.0
mg Cr/L leachate, as determined, however, by the Texas toxicity test.
The EPA regulations, RCRA, and the TDWR guidelines were promulgated in
response to the alarming proliferation of non-authorized and unsuitable
landfills across the country, and were designed to stem the serious health
hazards they pose. In contrast, the new federal and state regulations
require compacted clay liners for many types of landfills, and outline proper
handling procedures (including complete manifests) for all hazardous waste.
Also, these regulations specify tests to determine the hazard posed by a
given material, and its suitability for disposal in a properly maintained
landfill.
MATERIALS AND METHODS
One such candidate for landfill disposal is the sludge from the softener
at the USS Chemicals plant. The thickened sludge is composed primarily of
precipitated CaCO^, Mg(OH)2 and some CrCOH)^, and is vacuum-filtered to a
final thickness of 75-80 percent solids. In appearance, it is a firm, light
brown cake which dries to a fine powder.
The lime-softened sludge used in testing was collected in a plastic bag
which was placed directly underneath the vacuum filter. A representative
sample was collected from the entire width of the filter by moving the bag
back and forth underneath the point of discharge. The bag was then sealed so
as to prevent the passage of moisture in or out of the sludge. A sample of
the sludge was weighed in the laboratory, dried at 105°C for at least 24
hours, then weighed again to determine the percent solids. The sludge was
divided into two parts, one part to be used for each of the leachate tests
investigated.
EP Toxicity Test
The EP toxicity test was performed as follows:
1. 100 g sludge (wet weight) of particle diameter 9.5 mm was placed
in a stainless steel extractor (Associated American instruments)
along with deionized water equal to sixteen times the sludge weight,
or 1600 mL.
2. The extractor was agitated with a two-bladed impeller, and the
initial pH was measured.
3. Since pH was greater than 5.0 (t 0.2), it was adjusted with 0.5 N
acetic acid at a maximum rate of 4.0 mL/g sludge, or 400 mL
138
-------
4. The extractor was agitated for a specified period of time, after
which the mixture was raised to a final volume equal to twenty times
the initial sludge weight, according to the formula
V = 20w - (16w + A)
where
V = volume of deionized water
w = weight of initial sludge sample
16w = volume of deionized water previously added in step (1)
A = volume of acid added in step (5)
After the sludge mixture was brought to the proper final volume, it was
decanted and filtered through a 0.45 urn Millipore filter, and the filtrate
was analyzed for chromium. As a regulatory test, the EP toxicity test is run
for 24 hours, but for experimental purposes, it was performed six different
times from 3 to 30 hours in duration, with duplicate samples run for each
time period tested.
Texas Toxicity Test
The Texas toxicity test, based on the dry weight of the sludge sample,
was performed as follows:
1. After percent solids measurements were made, a wet sludge sample
with an equivalent dry weight of 250 g was placed in a 2000 mL
Erlenmeyer flask with 1 L deionized water.
2. The solution was stirred for five minutes, stoppered, and allowed to
settle for a specified period of time.
Following the settling period, the solution was decanted and filtered through
a 0.45 m Millipore filter, and the filtrate was analyzed for chromium. As a
regulatory test, the Texas toxicity test is run for seven days, but for
experimental purposes, it was performed for five different times from one to
nine days, with duplicate samples for each duration. EP and Texas toxicity
tests are compared in Table 6-1.
In addition to the toxicity tests performed, further investiations were
made into the rate of chromium adsorption onto the previously leached sludge.
These tests were performed with the sludge sample leached for seven days
according to the Texas toxicity test method. This sludge sample was divided
into three portions of 20 g each which were in turn exposed for one hour to
150 mL of deionized water containing 20, 40, and 60 mg Cr/L, as described in
Figure 6-1. Sludge samples containing reabsorbed chromium were then leached
139
-------
TABLE 6-1. Comparison of Extraction Procedures
TDWR
Distilled Water Leaching Medium
7 days
250 g sample + 1000 L DI water
Represents a more realistic
eluant to be used in tests to
simulate mono-landfi11 ing
operations
Based on dry weight
EPA
Acetic Acid Leaching Medium
24 hours
Wet sample + 16x its weight in DI water
Lowering pH to 5 t 0.2 with acetic acid
Increases solubility of certain trace
elements while decreasing others
Differing amounts of acid used in each
case some solids will dissolve more
than others
Based on wet weight
140
-------
again, according to the Texas toxicity test methods, and both the leachate
and the resultant sludge were tested for chromium.
Determinations of chromium in both sludge and filtrate were made accord-
ing to Standard Methods. Total chromium was determined by graphite furnace
atomic absorption spectrophotometry, while hexavalent chromium was determined
by the colorimetric method. However, the colorimetric method was not used to
determine hexavalent chromium in the EP toxicity test leachae, since addition
of the diphenylcarbazide to that solution produced a white, milky precipitate
instead of the normal color development. Instead, the EP leachate was raised
to pH 8.0 to precipitate Cr(OH)-j, and the remaining solution was filtered
through a 0.45 m Millipore filter, and analyzed for chromium by atomic
absorption spectrophotometry. Hexavalent chromium was thus determined to the
be total chromium left in solution after Cr(OH)3 precipitation.
RESULTS AND DISCUSSION
Results of the Texas toxicity test, showing leachate concentration
(total Cr and Cr(VI)) at various testing periods, are presented in Table 6-2,
and summarized in Figure 6-2. As can be seen, Figure 6-2 indicates that
relatively high concentrations of chromium are leached from the.softener
sludge in the first day of testing, from which point on the chromium appears
to be readsorbed onto the sludge from the leachate. From an initially high
leachate concentration of 1.25 mg total Cr/L, the leachate chromium concen-
tration drops off by 40 percent over the next three days to 0.77 mg total
Cr/L by the fifth day. Between days five and seven, leachate chromium con-
centration decreases an additional 29 percent to 0.55 mg total Cr/L, with yet
another 29 percent decrease to 0.38 mg total Cr/L by the ninth day. Thus,
while the leachate concentration decreases more slowly over time, leachate
chromium does not appear to reach equilibrium with the sludge during the
period of the experiment.
The concentration of Cr°+ in the sludge leachate also decreased over
time. In 'all, the leachate Cr concentration decreased about 60 percent in
nine days, as compared to a decrease in total leachate chromium concentration
of 70 percent during the same period of time, however, since the initial
Cr&+ concentration was relatively small (0.5 mg Cr°+/L), the nominal concen-
tration of hexavalent chromium in the softener sludge leachate changed very
little, indicating that in nine days the sludge/leachate mixture had more
nearly reached equilibrium with respect to Cr""1". Furthermore, the total
chromium leachate concentration gradually approached the Cr^+ leachate con-
centration, suggesting that Cr^+ was removed from the leachate by readsorp-
tion onto the sludge, while the Cr"* remained in solution. Thus, it might be
expected that at the time of true equilibrium, the total chromate concentra-
tion in the softener sludge leachate would consist entirely of hexavalent
chromium.
The significance of Cr^+ readsorption is also supported by the observa-
tion that total chromium in the leachate decreases as the pH drops below 8.0.
According to Thomas and Theis [4], the normal range of minimum solubility for
Cr(OH)3 (Figure 6-3), pH 8-10, is shifted to a more acidic range in solutions
141
-------
TABLE 6-2. Texas Test Leachate Values
(expressed as parts per million)
Days Run Total Cr
ppm
1 1.25
3 0.770
5 0.785
7 0.530
9 0.380
ppm
0.50
0.50
0.465
0.29
0.205
ppm
0.75
0.27
0.29
0.50
0.19
pH
(Ave.)
7.9
7.8
7.8
7.1
7.0
142
-------
7-DAY SLUDGE
7
20g | 20g ! 20g
i i
150-ml
20 mgCr/L
150-ml
40mgCr/L
150-ml
60mgCr/L
STIR I HR. / FILTER / DISCARD FILTRATE
80-ml
DI-H20
80-ml
DI-H20
LEACH SEVEN DAYS / FILTER
SLUDGE
FILTRATE
SLUDGE
FILTRATE
DISSOLVE IN HCI
DISSOLVE IN HCI
80-ml
DI-H20
SLUDGEl
FILTRATE
DISSOLVE IN HCI
ANALYZE SLUDGE AND FILTRATE FOR Cr BY ATOMIC ABSORPTION
Figure 6-1. Process diagram for leached chromate adsorption
study.
143
-------
j=
i_
o
1.4
1.3
1.2
I.I
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
O.I
oTOTAL Cr
nCr64"
ApH
a
567
TIME (DAYS)
8
10
14.0
13.0
12.0
11.0
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
Figure .6-2. Equilibration of Texas Toxicity Test to determine chromate in softener
sludge leachate.
-------
10 12
pH
Concentration distribution of
various Cr(III) species as governed by
the solubility of solid Cr(OH).-nH?0.
Figure 6-3. Solubility of Cr(OH) as a function
of solution pH.
145
-------
with high concentrations of carbonate. Since, at pH less than 8, most
carbonate is converted to bicarbonate, it is not unreasonable to suppose that
at that point also the Cr^+ in the sludge leachate becomes increasingly
insoluble, causing the total dissolved chromium concentration in the leachate
to decrease.
The EP toxicity test results (Figure 6-3) indicate a steep increase in
leachate chromium during the first nine hours of exposure. However, once a
high point was reached (0.38 mg total Cr/L), the chromium concentration in
the leachate also began to decline, until it apparently reached an equiibrium
concentration at 18 hours, which was 52 percent lower (0.185 mg total Cr/L).
In contrast to this dramatic behavior, Cr concentration appeared relatively
stable, at around 0.04 mg/L, varying little throughout the duration of these
test, suggesting that differences in total chromium concentration in the
leachate were caused by varying concentrations of chromium(lll).
The most likely explanation for the change in Cr^"1" found in the leachate
is that Cr(OH).j in the softener sludge dissolved under the acidic conditions
specified in the EP toxicity test (Figure 6-3), increasing the leachate chro-
mium concentration. However, the dissolution of Cr(OH)-j also served to raise
the pH of the solution, causing the Cr^+ to precipitate. Thus it can be seen
in Figure 6-4 that the steepest decrease in leachate chromium concentration
corresponded to the steepest increase in pH (from 3.8 to 6.3), which in turn
corresponds to the decreasing solubility of chromium(IIl). The variation of
Cr concentration with pH would therefore account for both the surge and the
decline of total chromium in the leachate. The slight increase in hexavalent
chromium during the early hours of the test may also result from the release
of chromate in the sludge upon acidification.
A comparison of the EP and Texas toxicity tests is presented in Table 6-
4, for the specified duration of the tests, i.e., 24 hours for the EP test,
and seven days for the Texas test. Results indicate that the EP toxicity
test, which comes to equilibrium quickly under acidic conditions, shows
chromium to be less easily leached from softener sludge than the more "con-
servative" Texas toxicity test. The Texas toxicity test, in which Cr"* is
determined by the diphenylcarbazide colorimetric method also indicates a
greater amount of hexavalent chromium in the sludge leachate than does the EP
toxicity test, which Cr(OH)3 precipitation and atomic absorption for deter-
mination of hexavalent chromium.
Studies of adsorption of chromium onto softener sludge indicated a
generally favorable adsorption isotherm (Figure 6-5) plotted according to the
linearized Freundlich model in Figure 6-6. As mentioned earlier, adsorbed
chromium was considered to be that portion which did not leach out of the
sludge when exposed to deionized water for seven days; hence, the values are
somewhat lower than those for total chromium adsorbed by the sludge after one
hour of exposure to the various chromium solutions. In all cases, Cr^* com-
prised a relatively constant, small percent of adsorbed chromium, on the
order of 1 percent (Figure 6-7).
146
-------
TABLE 6-3. EPA Leachate Values
(expressed as parts per million)
Hours Duplicates Total Cr Cr6+ Cr3"1" pH
3 1 0.210 0.0250 0.185 4.4
2 0.240 0.0185 0.222 4.0
6 1 0.460 0.0295 0.431 4.0
2 0.250 0.0250 0.225 3.5
12 1 0.330 0.0370 0.293 5.4
2 0.360 0.0335 0.327 7.0
18 1 0.260 0.0350 0.225
2 0.110 0.0610 0.049
24 1 0.160 0.0355 0.125 7.4
2 0.240 0.0400 5.3
30 1 0.180 0.0335 0.147 7.5
2
147
-------
TABLE 6-4. Chromium Concentrations in Softener Sludge Leachate by
the EP and Texas Toxicity Tests (mg/L)
Test Total Cr Cr6"1" Cr3+
EP 0.20 0.038 0.125
Texas 0.53 0.29 0.5
148
-------
CONCLUSIONS
The Texas toxicity test results in a higher chromium concentration in
the softener sludge leachate, but extrapolation of the leachate chromium
concentration past the seven-day test period suggests that a longer leaching
time might more nearly approach equilibrium, yielding a lower leachate chro-
mium concentration close to that obtained with the EP toxicity test. Fur-
thermore, test results do not seem to be affected by the fact that the Texas
test attempts to simulate a "mono-disposal" situation, in which the chromium
waste is stored separately, while the acidic EP test approximates conditions
of "co-disposal." Nevertheless, as it now stands, the Texas toxicity test is
a more conservative gauge of a waste's leaching potential than the EP
toxicity test—at least with respect to softener sludge.
As far as chromium adsorption is concerned, chromium is favorably
adsorbed onto softener sludge at chromium concentrations between 20 and 60 mg
Cr/L. further research should involve not only higher concentrations, but'
also mixtures of chromium and zinc, similar to those commonly found in
chrome-zinc corrosion inhibitors.
CHROMATE REMOVAL FROM COOLING WATER SLOWDOWN
Even in the most efficient zero discharge systems, occasions may arise
in which it is preferable to perform an "emergency blowdown"—for instance,
to dilute high levels of chloride ( 10,000 ppm) or TOC ( 600 ppm). In
these instances, it is advisable to have a stand-by chromate removal treat-
ment scheme for the blowdown sream to insure that standards are maintained.
Experiments with cooling water from the zero-discharge sidestream softening
system installed at the USS Chemicals Plant, Pasadena, Texas indicate that
chromate removal from high ionic strength water may be accomplished more
efficiently and more economically than by the treatment scheme (see Figure
6.8) normally used in continuous discharge systems.
Cr(VI) can be removed from cooling water blowdown by several processes,
including (1) chemical reduction to Cr(III) and subsequent precipitation, (2)
electrochemical reduction and precipitation, (3) ion exchange, and (4)
reverse osmosis. The standard method of hexavalent chromium removal from
cooling tower discharge is by chemical reduction and precipitation. This
method can be divided into two steps: (1) reduction of Cr(VI) to trivalent
chromium (Cr(III)) using reducing agents such as sulfur dioxide, sodium
bisulfite, metabisulfite or ferrous sulfate; and (2) precipitation of Cr(lII)
as CrCOH)^ by pH adjustment to 8.5 with lime or caustic soda. The precipi-
tated hydroxide is then separated from the blowdown stream. The first step,
chemical reduction, requires only a reaction vessel and equipment for adding
the necessary reagents (reducing agent, alkali, and acid).
CHROMATE REMOVAL BY REDUCTION WITH FERROUS SULFATE
For years, chromium(VI) has been reduced by addition of ferrous sulfate
to remove chromate from blowdown of continuous discharge cooling systems.
149
-------
Ln
O
_J
\
J1
z
o
H
QL
UJ
O
2
O
O
1_
O
0.4
0.3
0.2
O.I
-D-
-Q.
12 18
TIME (HOURS)
Cr'
24
o
-D-"
I
10.0
9.O
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
30
Figure 6-4. Equilibration of EP Toxicity Test for determination of chroraate in
softener sludge.
X
ex
-------
0.01
Figure 6-5.
Ce (mg/L)
Isotherm for re-adsorption of leached chromate
onto softener sludge.
151
-------
0
Inq
(mg/g)
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
InC (mg/L)
-2
Figure 6-6. Linearized Freundlich isotherm for re-adsorption
of leached chromate onto softener sludge.
152
-------
AVERAGE (RUN NO. 182)
3
0.04 0.08 0.12 0.16
Cr6+(mg/L)
0.20
Figure 6-7. Cr(VI) contained in chromium adsorbed onto
softener sludge.
153
-------
Acid
Reducing
Agent
Alkali
Slowdown
Cn
-P-
1
Lagoon
Clarilier
Filler
Sludge
Sludge
Drying Beds
Centrifuge
Vacuum Filter
Incineration
Discharge
Chromate Chemical Reduction System
Figure 6-8. Process schematic for reduction and removal of chromate from
cooling tower blowdown.
-------
However, most procedure specifications do not take into account the effects
of ionic interaction which occurs in the high TDS cooling waters obtained in
zero discharge systems. To make up for this lack of information, chemical
removal of chromate by ferrous sulfate reduction in these studies was inves-
tigated in two different waters: distilled water, and a high ionic strength
cooling water from the USS Chemicals plant. Ferrous sulfate was chosen over
other reducing agents (sulfur dioxide, sodium sulfite, bisulfite, or metabi-
sulfite) in order to eliminate the need to lower the cooling water pH with
acid prior to reduction.
Removal of hexavalent chromium from cooling water by chemical reduction
typically requires four steps:
1. Cooling water blowdown is lowered to between pH 3 - 4 with sulfuric
acid to increase the reaction rate of the reduction process;
2. A reducing agent is added to convert Cr(VI) to Cr(III);
3. Cr(OH)3 is precipitated by raising the pH to 8.5 with either lime or
caustic soda; and
4. The effluent is clarified to remove the chromium hydroxide floe.
When the reducing agent FeS04 is used, the following reactions occur:
Reduction
2H2Cr04 + 6FeS04 »• Cr2(S04)3 + 3Fe2(S04)3 + 8H20
Precipitation
Cr2(S04)3 + 6NaOH * 3Na2S04 + 2Cr(OH)3
The stoichiometric amount of FeS04 required for Cr(VI) reduction (calcu-
lated from the reaction above) is 3.93 mg FeS04/mg Cr04. Ferrous sulfate
consumption in the reduction reaction is closer to the calculated stoichiome-
tric amount in the pH range of 2 to 5, due to the increased oxidation of the
ferrous ion by dissolved oxygen at pH higher than 5. Solubility of the
precipitate, Cr(OH)3 is dependent on pH as described in Figure 6-5, with
minimum solubility at pH 8.
MATERIALS AND METHODS
The effects on chromate(VI) reduction and removal investigated in this
study included ferrous sulfate dosage, pH, and rection and detention times.
All experiments were conducted in batch reactors consisting of 250 ml Pyrex
beakers containing 100 mL of chromate solution.
155
-------
The experiments were performed with two different chromate soltuions:
an "ideal" Low IDS solution prepared from a stock solution of 1000 mg/L
potassium chromate G^CrO^) diluted with distilled water to a concentration
of 25 mg Cr/L (measured as CrO^), and an actual cooling water from the USS
Chemicals plant with an average concentration of 25 mg Cr/L. Trivalent chro-
mate solutions were prepared from reagent grade chromium chloride
(CrCl3*6H20).
The pH was adjusted to various values with 0.1 N l^SO^. An Orion
Digital Ion Analyzer (Model 501) pH meter with hollow body calomel probe was
used to measure the various pH parameters of the experiments. Total initial
and residual chromium concentrations were determined using a Perkin-Elmer 372
atomic absorption spectrophotometer fitted with a Model HGA-2200 graphite
furnace.
Ferrous Sulfate Dose Determination
Procedure to determine the effect of- ferrous sulfate dose on chromium
(VI) reduction was as follows:
1. Without prior pH adjustment, Cr(VI) was reduced to Cr(III) by addi-
tion of varying amounts of ferrous sulfate (100-200 mg FeSO^/L).
2. After a mixing interval of two minutes, 50 mL of the reduced solu-
tion was pipetted from the reactor, to which aliquot was added
sufficient 0.1 N NaOH to raise the pH to 8.5 and precipitate Cr(III)
as Cr(OH)3.
3. The resultant mixture was then filtered through a 0.45 ym membrane
filter to remove suspended solids not readily removed by simple
sedimentation.
Initial and residual chromium concentrations were measured, with the residual
chromium after filtration considered to be unreduced and unprecipitated
Cr(VI).
Determination of pH Effects
The three different pH parameters for chromium removal include (1)
initial pH of the chromate solution; (2) pH of the solution during reduction
(i.e., after addition of FeSO^); and (3) pH of solution for precipitation of
Cr(OH)^. For removal of 25 mg Cr/L, a stoichiometric amount of 100 mg
FeSO^/L was added to the chromate solution. Procedure for determination of
the effect of pH was the same as described above, with the exception that the
pH of the solution during the reduction reaction was adjusted to pH 2.5, 4.0,
5.5, and 6.5.
Also, a difference in removal was detected when ferrous sulfate was
dissolved at pH 2 or 4 prior to addition to the chromate solution.
156
-------
Effect of Reaction Time on Chromium Removal
To determine the effect of reduction reaction time on Cr(VI) removal,
chromate solution was allowed to contact stoichiometric equivalents of FeSO^
for various times between two minutes and 45 minutes. Following reduction,
aliquots of solution were neutralized to pH 8.5, settled, filtered, and
analyzed as described above.
Precipitation of Trivalent Chromium
To insure the complete precipitation of Cr(III) from reduced chromate
solutions, a solution of trivalent chromium was prepared and precipitated
with caustic (NaOH). Precipitation was allowed to proceed for different
lengths of time, after which the solution was decanted and filtered and
analyzed for chromium. Complete Cr(III) removal was observed at all reaction
times greater than two minutes.
Effect of Detention Time on Chromium Removal
After following the Cr(Vl) reduction and removal procedure described
above with a stoichiometric dose of FeSO^, supernatant solution was allowed
to stand for various times from 20 minutes to 48 hours before being analyzed
for total chromium. After 24. hours, additional light precipitate was
observed; 24 and 48 hour samples were therefore refiltered prior to analysis.
RESULTS AND DISCUSSION
Effect of FeSO/t Dosage on Chromate Removal
The decrease in chromate remaining after increasing dosages of FeSO^ is
plotted in Figure 6-8. While, in general, chromate removal improves with
increased FeSO^ doses, it is clear that dosages beyond 200 mg FeSO^/L result
in increasingly smaller reductions in residual chromate. As a result, the
most efficient dosage of FeSO^ appears to be within the range of 100-200
percent of the stoichiometric dose, or (for the given water) 100-200 mg
FeS04/L.
Effect of pH on Chromate Reduction and Removal
Figure 6-9 shows the residual chromate obtained by reduction with FeS04
at various pH. As can be seen, the greatest chromate removal was achieved
when chromate was reduced in deionized water to pH 4.0. This was true
regardless of the reaction time allowed for reduction, which had little
effect: on chromate removal at the optimum pH* However, when reduction took
place at pH 2.5, a change in reduction time from two to 45 minutes did seem
to have a significant effect on chromate removal.
157
-------
2.5
2.0
o>
J 1.5
o,
1.0
0.5
I
50 100 150 200
FeS04 (mg/L)
Figure 6-9. Effect of ferrous sulfate dosage on chromate
removal from USS Chemicals cooling water.
Cr. = 24.3 mg CrCL/L
158
-------
The effect of reduction pH on removal of chromate from USS Chemicals
cooling water is illustrated in Figure 6-10. Unlike the deionized water,
cooling water chromate appeared to be more effectively removed when the
reduction reaction occurred at pH 6 - 8, with the best removal achieved by
addition of an FeSO^ solution acidified to 4.0, which dropped the cooling
water pH to 6.9.
The effect of precipitation pH on chromate removal is plotted in Figure
6-11. As can be seen, the lowest chromate residual was obtained when precip-
itation was carried out at 7.0. This result is contrary to accepted prac-
tice, based on the minimal solubility of Cr(OH)3 at 8.5, and suggests that
Cr(OH)-j solubility is minimized in the acidic region in cooling waters with
high ionic strength.
Effect of Reaction Time on Chromate Removal
The minimal effect of increased reduction time (longer than two minutes)
on chromate removal from deionized water has already been inferred from
Figure 6-9. Since increased detention time increases the size of the reactor
vessel required to accomodate a given flow, it seems clear that a detention
time longer than four minutes should not be necessary when the reduction
reaction occurs at or near the optimum pH of 4.0. Detention time of the
reduction reaction greater than two minutes appeared to have no effect on
chromate removal from USS Chemicals' cooling water.
The effect of the time allowed for the precipitation reaction on chro-
mate removal from USS Chemicals cooling water is shown in Figure 6-12.
Curiously, the pH of the reduction reaction seems to have had an influence on
the speed of the precipitation reaction. When the reduction took place at pH
less than 6.0, there was a substantial increase in chromate removal obtained
by increasing the precipitation time from 1 to 24 hours. However, when the
pH of the reduction reaction was maintained at 6.3, maximum removal during
chromate precipitation occurred almost immediately (as shown by the intersec-
tion of the residual chromate lines at this pH), and did not seem to improve
with time.
CONCLUSIONS
The major finding of this investigation was that chromate can be satis-
factorily removed from high ionic strength cooling water without prior acidi-
fication by addition of stoichiometric amounts of ferrous sulfate. This
modification of usual chromate removal practice would result in substantial
savings since no sulfuric acid would be required to lower the pH of the
cooling water prior to chromate reduction, and therefore, less caustic would
be required to raise the cooling water to pH 8.5 for precipitation and
removal of Cr(OH>3.
For USS Chemicals' cooling water, the following scheme appears to result
in optimal chromate removal:
159
-------
1.75
ell* 1.5
o
O>
1.25
_J
<
Q
=> 1.0
- 0.75
0.5
0.25
O—-
o-
Reaction Time
I I 2 minutes
—A 10 minutes
2 5 minutes
O 4 5 minutes
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
Reduction pH
Figure 6-10.
Effect of reduction pH on chromate removal.
from deionized water.
Cr. = 25 mg Cr04/L FeS04 = 10° m§/L
160
-------
r;5.0r
o
o
4.0
o>
E
< 3.0
a
CO
LLJ
IT
o
2.0
1.0
5.0 6.0 7.0
pH
8.0
9.0
Figure 6-11.
Effect of reduction pH on chromate removal
from USS Chemicals cooling water.
Cr. =28 mg CrO^/L
161
-------
28
26
~ 24
ii
€ 22
o
S 2°
5 18
1.16
O
? 14
4,2
O
3
^
*s>
0)
cr
10
8
6
4
2
0
FeS04pH = 2.0
FeS04pH = 4.0
23456789
Reaction pH
10
Figure 6-12. Effect of precipitation reaction pH on
chromate removal from cooling water.
Cr = 28 mg CrO~/L
162
-------
30
28
26
24
II
c? 22
o
(f>
o
O
O
T3
'35
20
18
16
14
12
10
8
6
4
2
Detention Time
I hour
24 hours
345678
Reduction Reaction pH
9 10
Figure 6-13. Effect of reduction reaction pH on chromate
removal from USS Chemicals cooling water at
separate settling times.
Cr. = 26.8 mg CrO~/L
163
-------
1. A stoichiometric amount of FeSO^ (3.93 mg FeSO^/mg Cro£~) is acidi-
fied to pH 4.0 and added to the cooling water
2. The cooling water is'then mixed continuously for two minutes.
3. The pH of the cooling water is raised to 8,50 with caustic soda to
precipitate CrCOH)^, and mixed for an additional two minutes.
4. The effluent is clarified to remove chromium hydroxide floe (a
settling time of 30 minutes is recommended).
164
-------
SECTION 6 - REFERENCES
1, EPA Federal Register—Hazardous Waste and Consolidated Permit Regula-
tions, May 19, 1980.
2. Standard Methods for the Examination of Water and Wastewater, 15th ed.,
APHA (Washington, D.C.), 1980.
3. Saltzman, B.E., "Microdetermination of Chromium with Diphenylcarbazide
by Permanganate Oxidation," Analytical Chemistry, Vol. 24, No. 6, June
1952.
4, Weber, W.J., Physicochemical Processes for Wastewater Treatment. Wiley
Interscience (1971).
5. Sorg, T.J., "Treatment Technology to Meet the Interim Primary Drinking
Water Regulations for Organics," Part 4, Journal American Water Works
Association, August 1979.
6. Grove, J.H. and Ellis, G.B., "Extractable Cr as Related to Soil pH and
Applied Cr." Soil Sci. Am. J., Vol. 44, 1980.
7. Donohue, J.M., "Treatment of Cooling Tower Slowdown," Industrial Water
Engineering, July/August 1978.
8. Kunz, R.G. et al. "Kinetic Model for Chromate Reduction in Cooling
Tower blowdown," Journal WPCF, Vol. 52, No. 9, September 1980.
9. Ibid, p. 2328.
10. Patterson, J.W. Wastewater Treatment Technology. Ann Arbor, Michigan:
Ann Arbor Science, 1975.
11. Marin, S. et al. "Methods for Neutralizing Toxic Electroplating Rinse-
water, Part I," Metal Finishing Journal, Vol. 18, 1972, p. 274.
12. Bennett, J.R. "The Treatment of Effluents from Metal Cleaning and
Finishing Processes," Metal Finishing Journal, Vol. 18, 1972, p. 274.
13. Donohue, J.M. Op. cit., p. 8.
14. Mracek, W.A. and Greenberg, L. "Control and Automation of Chromate
Waste Reduction Plants," Proc. Intl. Water Conference, 30, 91, Engineers
Soc. Western Pennsylvania, Pittsburgh, Pa., 1969, cited by Kunz, R.G.;
Hess, T.C.; and Yen, A.F. "Kinetic Model for Chromate Reduction in
Cooling Tower Blowdown," Journal WPCF, Vol. 52, No. 9, Sept. 1980, p.
2328.
165
-------
SECTION 7
TOTAL ORGANIC CARBON MASS BALANCE
The objective was to determine the fate of the organic material in the
cooling water system. There are basically two sources of organicsfor the
cooling water: wastewater treatment unit (WWTU) effluent and treated Trinity
River makeup water. Outputs for the above system are: drift, volatiles
stripped (from the cooling water) and evaporation (no organics). There is
also a biodegradable fraction of organics.
In order to complete the mass balance, experiments were done on the
makeup and WWTU water. The purpose was to determine volatile and biodegrad-
able fractions of both influents. Samples from the WWTU (unit) were taken
and seeded with microorganisms. Nutrients ^HPO^, NH4.C1, 1.0 g/2 L were
added; and the sample was aerated for 12 days. TC and 1C were measured and
the TOG was determined. The same procedure was uded for samples of the WWTU
water but using organisms from the ethylene cooing water and microorganisms
from an industrial waste treatment plant, the Washburn Tunnel facility. The
TOG results are tabulated in Table 7.1,
A similar method was used for the makeup water. The only difference was
the addition of ferrous sulfate in order to avoid incidence of residual C12
on the microorganisms. TOG was measured and the results are tabulated in
Table 7-2. Volatile organics fraction jar tests were also done for both
samples. For this test, the sample was filtered through a 0.24 ym filter in
order to eliminate possible effects of biodegradation by the microorganisms.
The samples were bled with air, and initial and final TOG after 24 hours
were taken to determine fraction of volatile organics. Results are tabulated
in Table 7-1 for WWTU and Table 7-2 for makeup.
To solve the mass balance equation:
f2 Q2C2 + f3 Qlcl H
C±: Initial TOG of makeup (18.5)
C2: Initial TOG of WWTU (63)
Q^: Flow of makeup water to ethylene cooling tower (635 gpra)
Qo: Flow of WWTU water to ethylene cooling tower (169 gpm)
fj_: Fraction WWTU volatile (0.1)
166
-------
f2: Fraction WWTU biodegradable (0.34)
f3: Fraction WWTU makeup biodegradable (0.5)
Q^: Drift loss
C^: TOG ethylene cooling tower (267.5 ppm)
Solving for drift loss:
Q4 - 45 GPM
The drift loss flow was calculated by difference to be 45 GPM. This
number contrasts with an estimate of 35 GPM by tracer analyses done prior to
this research, and 10 GPM by rough TDS mass balances over the system. If the
actual drift loss rate is in the 45 GPM range, that means the organics losses
are proportioned between biodegradation (31 percent), drift (63 percent), and
volatilization (6 percent). With lower estimates of drift loss, biodegrada-
tion would have to account for the organics losses.
167
-------
TABLE 7-1. TOC Results, WWTU
A. Biodegradable Fraction
Time
(days)
0
1
2
3
4
5
6
7
8
9
10
11
12
WWTU
(Microorganism Growth)
(ppm C)
56
55.2
47.0
-
39.5
37.5
37.5
37.5
-
-
-
37.5 '
-
WWTU
(Ethylene)
(ppm C)
56
53
49.5
-
42.0
40.0
38.5
37.0
37.0
-
-
-
37.0
WWTU
(Washburn Tunnel)
(ppm C)
56
55.2
49.5
-
43.5
41.5
41.0
38.5
38.5
-
-
-
38.5
% Biodegradable: 33.93%
B. Volatile Fraction
Time TOC
(days) (Volatile)
0 50
1 45
% Volatile: 10%
168
-------
TABLE 7-2. TOG Results Makeup
Time
(days)
0
1
2
3
4
5
6
7
8
TOG
(Ferrous Sulfate Added)
ppm C
11
9
8.3
7.5
5.5
5.6
5.5
TOG
(No Ferrous Sulfate Added)
ppm C
11
9.5
9.0
6.8
6.0
6.0
6.0
% Biodegradable: 50%
Volatile Fraction:.
Time TOG
• (days) (Volatile)
0 11
1 11
% Volatile: 0.0%
169
-------
SECTION 8
COSTS
The capitol cost of the sidestream softening system is difficult to
separate from the total cost of the project. The overall project included a
raw water treatment plant to remove suspended solids from Trinity River
water. The process equipment included two clarifiers with the chemical feed
systems, two multimedia filters, and a pump tank. Common facilities shared
with the sidestream softening system were a thickener, rotary vacuum filters,
and the control room.
The total project cost based on 1977 dollars was $3.5 million. A 50
percent allocation to the sidestream softening system results in a cost of
$1.75 million. This calculates to a cost factor of $2/gallon of flow through
the softeners, in terms of capitol cost.
The costs for chemicals in the softening system and for the cooling
water system were closely followed. In Table 8-1, the chemical costs for the
lime softening system on an average cost per day basis are shown. In Table
8-2, the costs are on the basis of cost per thousand gallons. For the latter
case, the costs averaged $.37/1000 gallons for the time period studied. The
number compares favorably with costs greater than $1/1000 gallons for pro-
cesses such as ion exchange, and reverse osmosis.
For the cooling water system, the chemical usage is indicated in Table
8-3. The chemicals include the dispersants, inhibitors, biocides, and carbon
dioxide. The chemical cost per unit gallons recirculation water is shown in
Table 8-4. The total cost for the period studied was $0.16/1000 gallons,
which is comparable to the costs in a normal recirculating water system if
sulfuric acid costs were substituted for carbon dioxide costs as it would be
in a typical cooling water system.
170
-------
Figure 8-1
Chemical Cost for Lime Softening System
Daces
1/5-3/25
3/26-4/15
W16-5/6
5/7-5/27
No. Days
21
21
21
21
I
Ave Cost
Day
Lime
$79.l5/Ton
Cost
Cose Day
3. 480. 30 165.73
3.218.79 153.28
3,938.90 187.57
3.377.88 160.85
14.015.87
166.86
Soda Ash
$0.1075/0
Coat
Cost Day
827.75 39.42
1.677.00 79.86
2.123.12 101.10
1,752.25 83.44
6.380.12
75.95
co2
$53.00/Ton
Cost
Cost Day
556.50 26.50
556.50 26.50
556.50 26.50
556.50 26.50
2.226.00
26.50
*
Sludge
59.88/Ton
Cose
Cost Day
722.52 34.41
515.68 24.56
882.09 42.00
601.94 28.66
2.722.23
32.41
C Coat
5.587.07
5, 967. 47
7,500.61
6,288.57
25.344.22
I Cost
Day
266.05
284.17
357.17
. 299.46
1.206.87
301.72
: 8.
iR
%S.
S--
m :p
oi
TJ
Based on average weight of 12.88 cona lime sludge/load and $127.00/loaJ, cost la approximated us
$9.88/tons. This checks within 5Z of cost by summation of loads disposed.
-------
Figure 8-2
Chemical Cost Per Unit Gallons Water Treated
Date
1/5-3/25
1/26-4/15
4/16-5/6
5/7-5/27
No. Days
21
21
21
21
Ave Cose
1000 Cal
1000 Cal
Day
615.14
817.90
792.70
758.61
Boi.pf
Lime
Cost
1000 Cal
0.2011
0.1815
0.2109
0.2129
0.2072
Soda Ash
Co tit
1000 Cal
0.0481
0.0945
0.1245
0.1104
0.0944
co2
Cose
1000 Cal
0.0125
0.0114
0.0126
0.0151
0.0129
Sludge
Coat
1000 Cal
0.0422
0.0291
0.0517
0.0180
0.0401
t Cose
1000 Cal
0.1261
0.1165
0.4196
0.1964
0.1747
-------
Figure 8-3
Chemical Usage For Cooling Water System
Dates
3/5-3/25
3/26-4/15
4/16-5/6
5/7-5/27
Ho. Days
21
21
21
21
Ave Ibs
day
Ca
Dispersancs
Ibs
70.26
57.17
158.05
83.41
77.09
-
Zinc
Inhibitor
Ibs
773.62
1.365.29
1.282.38
1.105.50
53.89
Cliroia.it e
Inhibitor
Iba
168.04
520.3)
717.04
621.06
26.51
Nun-
Ox Id Iz lug
Bloclde
Iba
415.45
1,163.26
664.72
NIL
26.71
Chloride
Iba
320.ua
6.651.96
7,438.41
7.152.60
256.71
Bromine
Iba
1.064.0
1.280.0
1.967.0
981.0
63.00
CO
Ibs
225,260
304,720
349.250
298.300
14.018.21
Split poundage due to addition of new calcium dlspersanc to cooling water.
-------
Figure 8-4
Chemical Cost Per Unit Gallons Recirculation Water
Period
I
II
III
IV
Dates
3/5-3/26
3/26-4/15
4/16-5/6
5/7-5/27
I
Ave Cuac
1000 Gal
Ca
Dlupersant
Cose
1000 Gal
0.0013
0.0011
0.0029
0.0025
0.0078
Zinc
Inhibitor
Cost
1000 Cal
0.0055
0.0098
0.0092
0.0079
0.0324
Chronate
Inhibitor
Coat
1000 Cal
0.0042
0.0060
0.0083
0.0072
0.0257
Non-
Oxldlzlng
Bloclde
Coat
1000 Cal
0.0090
0.0251
0.0143
0.0
0.0484
Chlorine
Coat
1000 Cal
0.0005
0.0103
0.0115
0.0111
0.0334
Bromine
Cost
1000 Cal
0.0341
0.0410
0.0630
0.0314
0.1395
C02
Cost
1000 Cal
0.0638
0.0863
0.0989
0.0845
0.3335
£ Coat
1000 Cal
0.1184
0.1795
0.2082
0.1446
0.6507
Figures based upon a 21-day period and 93.600 1000 gal rectrculatlng/duy.
-------
SECTION 9
ZERO DISCHARGE SIDESTREAM SOFTENING AT TOSCO
INTRODUCTION
In response Co needs for water conservation and environmental protec-
tion, the TOSCO Corporation elected to install a side stream softening
system at their Bakersfield, California petroleum refinery. This plant site
is typical of small to medium sized refineries except that it is located in
a semi-arid region. Local and federal governments had imposed significant
incentives to reduce freshwater consumption and to protect groundwater
resources. This environment led TOSCO to initiate engineering studies for
water conservation in 1975. Since blowdown from the open cooling water
system comprised a major portion of the plant wastewater, these studies
included a review of existing technology for the recovery and recycle of the
blowdown. In early 1980, a conceptual process design was prepared for a
sidestream softening process to enable the recycling of cooling tower, a
particulate scrubber and boiler blowdown streams. This design was unique in
that it combined treatment and recycle of waste streams other than cooling
tower blowdown into a single treatment system. TOSCO accepted the concep-
tual process and proceeded with final design and construction. The waste-
water recycling facilities, based upon a caustic softening process, were
started up in early 1982.
The objective of this report is to compare the predicted performance of
the conceptual process design to the actual performance of the completed
plant. The TOSCO sidestream softening system is well suited for this pur-
pose since only very minor changes were made in the process design during
final engineering and construction. The general features of the conceptual
design are reviewed. The design methodology is summarized, with emphasis on
the techniques used for chemistry calculations and prediction of process
performance. The various process alternatives that were investigated are
contrasted in this report, with discussion of potential advantages and dis-
advantages. The final process design is also briefly reviewed. Process
changes made during final engineering can have a significant impact upon
actual performance of the softening system. A comparison between predicted
and actual performance is also presented. Relevant factors contributing to
these comparisons are identified and discussed. This comparison provides a
convenient basis for the formulation of recommendations concerning the
design procedure used for caustic softening and recyclingof blowdown streams
to an open, recirculating cooling water system. These are presented in the
final section.
175
-------
GENERAL PLANT DESCRIPTION
TOSCO Corporation is a. medium sized independent oil refining and petro-
chemical company. A small refinery near Bakersfield, California was owned
and operated by TOSCO until adverse economic conditions forced a shutdown in
1983. The plant incorporated the conventional processes found in most crude
oil refining plants, including distillation, catalytic cracking, and coking
processes plus steam and cooling water utility systems. The Bakersfield
plant is located in an agricultural area with annual rainfalls averaging
less than 12 inches annually. Petroleum refining is also of major economic
importance. Thus, the Bakersfield area is in a semi-arid climate with a
significant economic dependence upon both irrigation and groundwater. With
the scarcity of surface water and the importance of the agricultural indus-
try, very stringent requirements are placed on discharges to receiving
streams. Similarly, dependence upon groundwater for potable supplies and
irrigation requires strict regulation of waste holding ponds and disposal
wells. The superfund taxes, based upon volume of waste injected into dis-
posal wells, represented a significant wastewater disposal cost. Prior to
the startup of the sidestream softening system, most process wastewater and
blowdown streams were disposed of by deep well injection.
The major wastewater streams at the TOSCO refinery included cooling
tower blowdown, boiler blowdown, thermal catalytic cracking (TCC) unit
scrubber blowdown, coker scrubber blowdown, and sour water stripper bottoms.
Of these streams, coker scrubber blowdown could not be considered for
recycle due to the presence of very fine solids, dissolved organics, and
reduced chemicals. This stream was disposed of by clarification, filtra-
tion, and deep well injection. Flow and water quality data for the remain-
ing streams are presented in Table 9.1. Although the sour water stripper
bottoms represented a large waste load and was of high quality in many
respects, it was decided to exclude this stream from the wastewater recycl-
ing project. At times the ammonia and sulfide concentrations would become
very high. The presence of these materials within a cooling tower system
could present unique corrosion problems. Earlier studies conducted at the
TOSCO plant indicated that the quality of the remaining waste streams was
acceptable for recycling following suitable physicochemical treatment.
The cooling water system used at the TOSCO plant in Bakersfield pre-
sented a unique situation also. The plant operated with a total of five
cooling towers. Four of these towers (Nos. 1-4) operated from a common
basin. Blowdown from cooling tower No. 5 discharged directly into this com-
mon basin. The blowdown stream from the No. 5 tower therefore served as a
feed or makeup stream for towers No. 1-4. With this flow scheme it was
possible to treat the four cooling towers as a single unit. The combined
recirculation rate of the five towers is about 53,000 gpm. A relatively
hard wellwater was used to provide makeup water for the cooling tower
system, as noted in Table 9.1. The heat exchange system served by the cool-
ing towers was mostly constructed of mild steel. This situation made the
continued use of chromate based corrosion inhibitors a major benefit for
selection of the sidestream softening system. A chemical softening process
treating a sidestream from the cooling towers would permit recycle of chrom-
ates, eliminating the environmental impact of a toxic chemical.
176
-------
Before the wastewater recycle plant was completed, wastewater at the
TOSCO plant was discharged either to an evaporation pond or disposal well.
Local and state agencies had imposed very strict standards for operation of
waste ponds and disposal wells. These governments were also exerting pres-
sure for industry to reduce its reliance upon both surface and groundwater
resources. The superfund tax burden for deep well injection of cooling
tower and boiler blowdown was in excess of $350,000 annually. In addition,
the cost for operation and maintenance of the injection well was strongly
influenced by the volume of wastes handled. These incentives were in con-
cert with improved corrosion protection that would be obtained by recycling
of chromates. For these reasons TOSCO decided to proceed with design and
construction of a sidestream softening system for the recovery and reuse of
several wastewater streams.
SIDESTREAM SOFTENER PROCESS DESIGN, TOSCO REFINERY
The process design for the zero blowdown sidestream softening system
for the TOSCO refinery will be reviewed in this section. The methods and
assumptions used in the preliminary process design will be discussed and
related to predicted performance. The TOSCO design provides a very inter-
esting example for application of techniques used for softener design. The
process used was unique in combining scrubber blowdown with sidestream soft-
ening. The cooling water system used at the Bakersfield refinery presents
an interesting flow pattern. Design equations were developed for this pro-
cess. The final design will also be briefly reviewed in this section.
Changes in preliminary design inevitably occur as final design and construc-
tion are completed. These differences will serve as a comparison of prelim-
inary and final designs.
PRELIMINARY PROCESS DESIGN
It was noted above that the cooling water system used at the TOSCO
refinery was unusual. The centralized basin and cooling towers 1-4 can be
treated schematically as a single cooling tower system. This system, how-
ever, receives two feed streams rather than the conventional single feed.
Fresh wellwater is added to replace evaporation, drift, and blowdown. The
blowdown stream from cooling tower 5 is also added to the central basin.
This is a low quality water that represents a significant load of scale
forming salts for the combined cooling water system. It was also decided to
include boiler blowdown and TCC scrubber blowdown in the sidestream softener
design. No changes in operation were anticipated for cooling tower 5. This
multiple cooling tower system is much more complex than systems typically
described in the literature. Unique mass balance relationships were derived
for the TOSCO system. Using these design models, equilibrium chemical cal-
culations, and assumptions based upon previous experience, a detailed pre-
diction of final cooling water quality was performed for the process recom-
mended to TOSCO. These calculations will be reviewed in this section.
The sidestream softening process recommended for the above cooling
water system is represented schematically in Figure 9.1. Definitions for
the symbols used in this schematic are included in this figure. This sche-
177
-------
matic shows the No. 5 cooling tower blowdown (QcT^ being added to the
central basin while boiler blowdown (Qfc) and TCC scrubber blowdown (QiCC^
are directed to the sidestream softening system. The boiler blowdown pro-
vides a significant source of both silica and alkalinity for the softener,
as shown in Table 9.1. The alkalinity would reduce requirements for soda
ash within the softener and silica will be removed during the softening pr-
ocess. The TCC scrubber blowdown can contain high suspended solids levels
but otherwise provides a good quality makeup source. These solids can be
removed within the softening system. In this process the softener removes
scale components (i.e., Ca, Mg, and SiC^) from cooling water and permits
recycle of two wastewater streams.
The sidestream softening system described by Figure 9.1. is unique to
the TOSCO plant. A steady state mass balance model is required to size the
softener and clarifier. Based on these mass balance equations, cooling
water quality and chemical dosages can be estimated. The scale constitu-
ents) that determine sidestream flow to the softener can also be identi-
fied, as discussed in a previous chapter. Steady state mass balances must
be developed for water, non-conserved soluble species, and conserved species
passing through the sidestream softener. Effects of all chemical additives
must be considered when designing a sidestream softener since all salts will
become highly concentrated in a zero blowdown system. Additives such as
H2S04 can affect scale formation, e.g., sulfuric acid may cause CaSO^ scale
to form within heat exchangers. The development of these mass balance equa-
tions and their use of estimating softener size and cooling water quality
will be considered in subsequent sections of this chapter.
MASS BALANCE EQUATIONS
The schematic diagram of the cooling water system suggests that mass
balances may be performed for the cooling tower/ sidestream softener as a
single system. Mass balances may also be performed for subsystems consist-
ing of the cooling tower and softener separately. The proper system or sub-
system for analysis must be selected in order to obtain desired information
concerning process design. The first step is usually to quantify flowrates
of all input and output streams by performing a water balance. Since the
flow rate to and from the softener is unknown, the entire system should be
selected for the water balance. Based upon Figure 9.1 this water balance is
given below:
All flow terms in Equation 9.1 with the exception of makeup, QM, softener
chemical additions, QQ, and waste softener sludge, Qvj, are known from plant
operating experience. These known flowrates are included in Table 9.1.
Reasonable estimates for QA, QQ, and Q^, based on prior experience, can be
used to quantify these terms. Using these assumptions, Equation 9-1 can be
used to estimate the makeup flowrate, QJJ.
The next objective is to estimate the required flowrate discharged to
178
-------
TAULE 9-1. WAT til IJUAMTV DATA BASE
CT70
a a
-.n
5T"
Water Source ^iu Ca* Me* M-Alk.* S iO2 Cl SO^ Na pll TSS** Nll-j |I2S
No. 8 Well - 100 33 106 17 36 72 45 8.4 -
No. 5 CT 55 440 120 24 140 330 783 342 7 .0 -
blowdown
Holler blowdown 50 - - 550 123 363 480 718 LI. I
Sour water
stripper
bottoms*** 130 61 - 3-.__.
TCC Scrubber
blowdown 25 50 24 100 21 24 60 56 8.4 6,700
Concentrations In ppm CaCO^. All other concentrations are nig/L
Total suspended solids.
h **
Ca, Mg, and Si02 per I). Miller, 3/26/80, except Nllj and II2S.
°- Na concentrations calculated for electroneutra lity
(D ^T*
I
o :
•o
-------
the sidestream softener, Qg. This flowrate, in conjunction with estimated
water quality, determines the size of the softener. The magnitude of Q3 is
estimated by performing a soluble species balance around the cooling tower
subsystem. This mass balance equation appears as follows:
Vm * QACA ^CTCCT + Vl ' Vw + Vw (9'2)
It should be noted that Equation 9-2 is not valid for species which can be
strongly influenced by aqueous equilibria such as CC>2, HCO^ and CO^. All
flow terms in Equation 9-2 are known except for Qs and Q^. These flowrates
can be related by a water balance around the softener subsystem:
QT = Qs+Qb+QTCc+Qc-Qw (9~3)
Combining Equation 9-2 and 9-3 and solving for Qs yields the following
design equation:
w r
For Equation 9-4 it was assumed that C^=Cjj for important scale forming
species. Equation 9.4 can be used to estimate the flowrate to the softener,
Qs, by using known flowrate and concentration data in addition to estimated
concentrations for the softener effluent. Application of Equation 9-4 to
each scale forming species that may limit cooling water concentration, i.e.,
CaC03, CaSO^, and Si02, must be performed to determine the required flowrate
to the softner. Maximum concentrations within the cooling water, Cw, can be
determined from prior experience with actual cooling water systems. Concen-
trations for softener effluent can be obtained either from jar testing on
simulated cooling water or from prior experience with sidestream softening.
For the TOSCO design, the estimated softener effluent quality was based upon
jar testing conducted by the NUS Corporation.
The concentration for both conserved and non-conserved soluble species
within the cooling water must also be estimated during process design. The
equation used for non-conserved species (i.e., Ca*^ Mg+^, Si02, Alkalinity)
can be obtained directly from a mass balance around the cooling tower sub-
system, Equation 9-2:
(9-5)
The desired Cw for the limiting scale species is used to estimate Q3 and Qf>
Values of Cw for remaining non-conserved species are estimated using Equa-
tion 9-5. Conserved species in the cooling water (e.g., Cl~, 50 = Na^,
etc.) can be influenced by chemical additions such as Na2C03, l^SO^, or
chromate salts. It is therefore desirable to include each input stream
individually in the equation used for conserved species. This equation can
be obtained by performing a mass balance around the entire system shown in
Figure 9-1. Performing this balance and solving for cooling water concen-
180
-------
tration, Cw, yields the following:
Cw = Q QT
» » — /% /•» ( f\ f*
(9-6)
where
Ql = Qd+Qs (9-7)
Q2 = QT+QW (9-8)
Equation 9-6 was used to estimate cooling water quality for all conserved
species. Prior to application of Equation 9-6 for conserved species it is
necessary to estimate required dosages of softening, neutralization, and
cooling water chemicals. These calculations will be discussed in the next
subsection.
CHEMICAL DOSAGE CALCULATIONS
A sidestream softener for a zero blowdown cooling water system must
remove both calcium and silica to prevent scale formation. Calcium is pre-
cipitated as CaC03 while silica (Si02) is removed with Mg(OH)2 that forms as
magnesium precipitates. The important softening reactions are shown in
Table 9-2. A base, usually lime, is added to raise pH and initiate precipi-
tation. Soda ash (^2^3) is added to balance calcium hardness and improve
calcium removal. Magnesium is often added, either as dolomitic lime or as a
soluble salt (e.g., MgSO^) to improve silica removal in the softener. After
softening reactions have been completed, the water must be neutralized prior
to returning to the cooling tower basin by addition of an acid. Sulfuric
acid is usually used for neutralization, but care must be exercised to avoid
conditions leading to calcium sulfate scale formation. The estimation of
these dosages will be described in this subsection.
The pH of the cooling tower can be raised using either lime (CaO) or
caustic soda (NaOH). The selection of either chemical must be based upon
the overall economics of the process. While lime is generally cheaper than
NaOH, it increases the soda ash requirement, produces more sludge, and
requires more complex handling equipment. The high cost of soda ash,
sludge disposal, and maintenance cost may lead to choice of NaOH rather than
lime. This choice is site specific, however. The required dosage of base
is estimated using the method of Matson (1,3). Matson recommends the fol-
lowing equation:
Base Dosage = C00 + HCO, + MgH + Si00 + AOH ^9~9^
(ppm CaC03) 23s 2
181
-------
In Equation 9-9, the quantities on the right side represent cooling water
values expressed in ppm CaCC^. The value of OH is obtained from the change
in pH from cooling water to softener values. Assumptions must be made at
this point concerning cooling water pH, alkalinity, and softener pH. Equa-
tion 9-9 may be used to estimate the dosage of NaOH or CaO.
The addition of soda ash (^2003) is used to improve removal of calcium
hardness. The usual practice, as recommended by Matson (1,3) is to obtain a
balance between calcium hardness and available carbonate. Using this as the
objective, the following equation is used:
Na2C°3 °OSe = CaH + Lime Dose - CO - 2(HCO ) - C0~ (9-10)
(ppm CaC03)
Again, all values to the right of the equal sign are in ppm €3003. Equation
9.10 shows the relationship between soda ash dose and lime dose. If NaOH
is used to raise cooling water pH, the second term on the right side of
Equation 9-10 is zero.
Estimation of magnesium requirements must consider the degree of
removal for silica. Matson (1) has shown that Si02 removal in sidestream
softeners follows a. Freundlich isotherm:
kCn (9-11)
where qe is the mass ration of silica removal to magnesium removal, C is the
Si02 concentration in the softener effluent, and both k and n are empirical
constants. Values of k and n are dependent upon temperature, and were dis-
cussed previously in this report. For any softener flowrate, Qs, there will
be some maximum value for SiC>2 in the softener effluent that must be main-
tained to prevent Si02 scaling. Using this value on available Mg and iso-
therm data are compared. If the available qe is greater than the isotherm
q , Mg must be added. The dosage of Mg can easily be calculated from the
two values of q . A soluble magnesium salt, MgSO^, was considered for the
TOSCO design since improved SiO£ removal has been found for this form of
magnesium (1). It should be noted that addition of a magnesium salt would
increase the required dosage of base, as shown by Equation 9-9.
As a final check of accuracy for the above water quality and dosage
calculations, check is performed for electroneutrality. Total positive and
negative equivalent concentrations are calculated for the estimated cooling
water quality. If these values are sufficiently close, the calculations are
accepted. If not, the entire process is repeated, using the results of the
previous iteration where necessary.
PRELIMINARY DESIGN RESULTS
The calculation methods described above were applied to the TOSCO
182
-------
TABLE 9-2. SOFTENING REACTIONS
Carbonic Acid System:
H2C03 + OH" H20 -t-
HCOj + OH' H20 + C03
Precipitation:
Ca2+ + CO* CaC03(s)
Mg2+ + 20H" Mg(OH)2(3)
Adsorption/Coprecipication:
Si02 + Mg(OH)2(s) Si02'Mg(OH)(s)
TABLE 9-3. PROJECTED WATER BALANCE
Item
Makeup Water*
No. 5 CT BLowdown
Boiler Slowdown
TCC Scrubber Slowdown
Evaporation (30° F T)
Drift
Treated Water Return
Slowdown, CT Nos. 1-4
Softener Flow
Sludge Flow
AVG
1518 gpm
55
50
25
1590
53
235
165
245
10
MAX
1513
55
50
25
1590
53
400
330
410
10
gpm
Includes 1 gpm flow for cooling tower additives
Includes softening chemicals water, 5 gpm
183
-------
TABLE 9-4. VALUES ASSUMED FOR PRELIMINARY DESIGN CALCULATIONS
Cooling water pH 7.0
Cooling water total alkalinity = 75 ppm
QA = 1 gpm
Qc = 5 gpm
Qw = 10 gpm
Cw(Si02) = 200 ng/L
CT(Si02) = 40 mg/L
Cp(Ca) = 50 ppm CaCOj
CT(Mg) = 20 ppm CaC03
CT(M.Alk) = 25 ppm CaC03
184
-------
system at Bakersfield. A total of four alternative softening processes were
evaluated using lime or caustic soda to raise pH and I^SCfy or C02 to neu-
tralize softener effluent. The scale component limiting softener flow with
sulfuric acid in use was Si02- When uing C02 for neutralization CaC03
limited softener flow, requiring use of much large softeners. Based upon a
consideration of operating and capital costs, the option using NaOH and
H2SC>4 was found to be best for TOSCO. Only those calculations for the NaOH-
H2S04 case will be considered in this report.
The values calculated for preliminary design of the TOSCO plant are
summarized in Tables 9-3 through 9-5. The water balance obtained for the
TOSCO plant is shown in Table 9-1. Values that were assumed for these cal-
culations are shown in Table 9-4. The blowdown from cooling towers 1-4 is
the flow to the sidestream softener. As shown in Table 9-3, a blowdown
flowrate of 165 gpm was required to maintain silica below 200 ppm in the
cooling water. An effluent silica concentration of 40 ppm was estimated for
the 40-C softener based upon prior experience. The total softener flowrate
was 240 gpm. Estimated water quality for cooling water and softener
effluent are shown in Table 9-5. These data explicitly show the level of
removal expected in the softener for scale forming chemicals. Data for con-
sumption of softening chemicals and sludge production are presented in Table
9-5. These data show that it was found necessary to add MgSO^o aid in
removal of Si02« If additional magnesium had been present in the well water
this may not have been required.
A schematic diagram of the process design for the sidestream softeners
is shown in Figure 9-2. This figure also shows the sizes recommended for
major process components. The loading factors used for design of the
settlers and filters are shown in Table 9-6. The overflow rate used for
clarification, 0.9 gpm/ft^ (1300 gpd/ft^)is within the range normally
encountered for lime softening systems. Actual settling data collected from
earlier studies were used to estimate the overflow rate. A solids flux
technique, using data supplied from previous studies, indicated that under-
flow should be in the range of 6% solids for the 20 ft. diameter reactor
clarifiers recommended. Two parallel softeners were recommended to provide
flexibility in operations. Each softener was designed to handle the full
softener flowrate of 245 gpm. A depth of 15 ft. was recommended, based upon
previous experience. The filter loading of about 6.0 gpm/ft^ was selected
as being reasonable for mixed media pressure filters. This value is well
within the range typically used for waters of fairly low turbidity. Again,
two parallel filters were recommended to permit continuous process operation
during backwash cycles.
As noted previously, the concentration of silica in the cooling deter-
mines softener design for the chemical treatment selected, i.e., NaOH and
I^SO^. Based on heat exchanger temperatures at the TOSCO refinery it was
recommended that the Si02 concentration in cooling water be maintained at a
maximum of 200 mg/L. A Si02 concentration of 40 mg/L in the softener
effluent was considered feasible, based on prior operating experience with
sidestream softening systems. Since MgSO^ addition was recommended, the
dosage could be increased to improve Si02 removal. The calculated detention
time in the recommended solids contact clarifiers was about 2.5 hr. This
185
-------
TABLE 9-5. PROJECTED WATER QUALITY DATA
I tera
Temp. Ca* MG* M.Alk.* SiCU Cl S04 Na pH TDS
CooLing Water 50° 360 230 . 75 200 L524 6465 3536 7.0 12,200
Softener Effluent 50° 50 20 25 40 1100 4690 2909 6.2 3,300
Concentration units of ppm CaCO^. All other values are in mg/L.
TABLE 9-6. DESIGN LOADING FACTORS
Item
Value
Softener Overflow Rate
Filter Hydraulic Loading
0.9 gpm/ft2
6.3 gpm/ft2
186
-------
should provide sufficient capacity for precipitation, clarification, and
sludge thickening. A sludge recycle rate of 20% was also recommended. This
high recycle rate was recommended due to the presence of "inert" solids in
the softener. Use of the softeners to clarify TCC scrubber blowdown intro-
duces these "foreign" solids that may be less effective for softening than
solids produced through precipitation. The preliminary process design, as
shown in Figure 9-2, was accepted by TOSCO management and plans were made
for final design and construction. The final design will be described in
the next section.
FINAL PROCESS DESIGN
Final plans and specifications were developed for the TOSCO sidestream
softener. These plans were based upon preliminary design calculations and
recommendations. During the final design process, preliminary calculations
are first double checked and corrected, if necessary. Vendor recommenda-
tions are also included for key process steps such as the solids contact
clarifiers and mixed media pressure filters. These steps commonly require
changes in flow, composition, and/or equipment siting during final design.
During preliminary design, a process schematic is typically used with no
concern for actual plant site layout. For final design, the physical layout
of the entire plant site is reviewed and the most economical routing of new
pipelines and new equipment layout should be determined. This review would
consider aspects such as available pipe racks, physical location of process
vessels, and available pressure and/or elevation for each flow. Such a
design review may lead to changes in the process flow pattern anticipated
during preliminary design. The results of the final process design will be
described in this section. A comparison between preliminary and final
designs will also be presented. Because the final design accurately
describes the actual TOSCO plant, this comparison will be important when
considering actual versus predicted process performance.
A schematic diagram for the final process design is included as Figure
9-3. This diagram is very similar to the preliminary process schematic,
Figure 9-2. For the final design, boiler blowdown was added to the filter
backwash tank rather than to the splitter box as shown in Figure 9-2. This
change has no effect on softener performance or water quality. The chemical
feed point for MgSO/^ was changed from the splitter box as originally pro-
posed to the softener feed well. This change is an improvement since it
increases the flexibility of operatic, i.e., magnesium feed can be adjusted
invididually for each softener. Another change to be noted is the change in
solids recycle rate for the solids contact clarifiers. The final design
recommends a recycle ratio of 10% whereas a 20% ratio was recommended
earlier. The maximum recycle ratio for final design is 16.7% of design
flow. While solids recycle is below initially proposed levels, the rate
used for final design was probably recommended by the supplier of the solids
contact clarifiers, Infilco-Degremont. Their extensive experience in water
treatment would ensure the final design values are reasonable. The final
design also added supernatant return from the softener sludge pond. The
only other change noted in the final design concerns the neutralization and
filter feed tanks. These tanks were combined in a single vessel with a
187
-------
separating weir in the final design. The obvious reason for this change is
a reduction in capital cost. None of the changes made in the final process
flow scheme would change the accuracy of the preliminary design calcula-
tions. The reduced sludge recycle rate would limit the maximum solids con-
centration attainable in the softening reactors. This may reduce precipita-
tion rates, leading to a less stable effluent, but this potential problem
should be very insignificant.
In Table 9-7 the flow data used for preliminary and final design are
compared. These data show that the major influent streams are the same for
preliminary and final design. For the final design, provision was made for
return of sludge lagoon supernatant and filter backwash. The flowrate
needed for treatment chemicals in final design was 3 gpm less than that
assumed during preliminary design. The net result of these changes is that
flowrate through the softener for final design is 2 gpm higher than in the
preliminary design. Sludge recycle is also lower in final design, as dis-
cussed previously. The estimated waste sludge flowrate was reduced from 10
gpm in preliminary design to 6 gpm for the final design. Considering all
changes in flow for the final softener design, the end result is that
treated water returned to the cooling system is 6 gpm higher than prelim-
inary estimates. These changes are not great enough to cause a major loss
of accuracy in preliminary calculations.
Water quality estimates are compared for preliminary and final design
in Table 9-8. These data show that the preliminary water quality estimates
described earlier were sufficient for final design purposes. A more
detailed description of projected water quality data is presented in Table
9-5. The preliminary calculations for cooling water quality were also used
for initial selection of treatment chemicals for the cooling system. Recov-
ery of cooling tower blowdown enabled the chromate-zinc corrosion inhibitor
system to be increased to 30-40 ppm CrO^ and 3-5 ppm zinc. It was also
recommended that cooling water pH be maintained in the range of 6.8 to 7.3
rather than 6.5-7.0 as had been done previously. A polymaleic acid (PMA)
scale inhibitor was recommended for use in the cooling water. Experience
with USS Chemicals has shown that PMA does not significantly interfere with
sidestream softener operations. A wetting agent was also recommended for
use in the cooling tower. Biological treatment recommended for the TOSCO
plant was continuous chlorination with a 0.2-0.5 ppm free chlorine residual.
The softening chemical requirements and sludge production rates are compared
in Table 9-9. Again, preliminary calculations were directly used for final
design.
The data of Table 9-10 compare preliminary and final design data for
softening equipment. As shown in Table 9-10, preliminary estimates for many
of the less important items were omitted. The only significant changes in
equipment recommendations were the modest size increases for the solids con-
tact clarifier and mixed media pressure filters. The reactor-clarifier
diameter was increased from 20 ft. to 22.5 ft., while the depth was
decreased from 15 ft. to 14 ft. for the final design. These changes may
have been made to accomodate standard production models from the supplier.
However, this increase changes design total hydraulic residence time from
about 2.3 hr. to near 2.7 hr., an increase of 17%. This change should
188
-------
TABLE 9-7. COMPARISON OF PRELIMINARY AND FINAL FLOW DATA
Item
Boiler Slowdown
TCC Scrubber Slowdown
Filter Bakcwash Return
Sludge Lagoon Supernatant
Slowdown, CT Nos . 1-4
Treatment Chemicals, Total
Softener Flowrate
Sludge Recycle Flowrate
Waste Sludge Flowrata
Treated Water Flowrate
Preliminary
Design
50 gpm
25
3
2
165
2
247
24.5
6
241
Final
Design
50 gpm
25
0
0
165
5
245 .
49
10
235
TABLE 9-3. COMPARISON OF PRELIMINARY AND FINAL WATER QUALITY ESTIMATES
Preliminary Final
Item Design Design
Cooling Water pH 7.0 ' 7.0
Cooling Water Temperature 50° C 50° C
Cooling Water TDS 12,200 mg/L 12,200 mg/L
Treated Effluent pH 8.2 8.2
Treated Effluent Temperature
Treated Effluent TDS 8,800 mg/L 3,300 mg/L
Treated Effluent TSS - 2 mg/L
TDS = Total Dissolved Solids
TSS = Total Suspended Solids
189
-------
TABLE 9-9. COMPARISON OF PRELIMINARY AND FINAL CHEMICAL CONSUMPTION
I tern
Anionic Polymer
MgSO^
Na2C03
NaOH
H2504 for
Softener
Sludge Production
All values are based
Preliminary
Dosage
(ppm)
-
260
380
340
49
-
on average
Design
Consumption
(Ib/day)
-
750
1,100
1,000
144
4,240
flow conditions.
Final
Dosage
(ppm)
0.5
50
380
340
49
-
Design
Consumption
(Ib/day)
1.5
150
1,100
1,000
144'
4,240
TABLE 9-10. COMPARISON OF PRELIMINARY AND FINAL EQUIPMENT SIZING
Preliminary
Item Design
Splitter Box 1000 gal
Softener Diameter 20 ft
Softener Depth 15 ft
Waste Sludge Sump
Neutralization/Filter Feed Tank
Filter Backwash Holding Tank
Filter Backwash Clear Well
Filter Diameter 7 ft
Final
Design
500
22
14
3,300
7,125
12,000
8,000
7
gal
.5 ft
ft
gal
gal
gal
gal
.5 ft
190
-------
increase both reaction time and clarification capacity, providing an efflu-
ent of higher quality. The overflow rate for the clarifier was decreased
from about 0.89 gpm/ft2 (1150 gpd/ft2) to about 0.65 gpm/ft2 (936 gpd/ft2 or
about a 19% increase in area. The data comparing filter diameters also
shows that the final design recommended larger filters. The filter loading
used for preliminary design was 6.1 gpm/ft2 while the loading for final
design was 5.4 gpm/ft2. This represents a decrease in filter loading of
about 11%. As expected, the final design is more conservative than the pre-
liminary design. The increases to major equipment siting appear to be in
the range of 10 to 20%. However, these increases are moderate and appear
reasonable. Increased size for softeners and filters should improve process
performance, capacity, and flexibility of operation.
The final design used for the wastewater recycle/sidestream softening
system was nearly identical to the preliminary design. No significant
changes were made to the process flow scheme. All chemical calculations for
the preliminary design were used for final process design and to select a
suitable cooling water treatment system. Modest increases were made in the
size of the softening reactors and final filters. These increases should
not significantly affect water quality estimates. The changes made to the
preliminary design should only serve to improve overall process performance
and reliability.
As a final note to the process design section, the "degree of conserva-
tiveness" applied will be briefly considered. During design of a process
such as sidestream softening, several opportunities are available for engi-
neers to use conservative estimates of important parameters. This is true
for plant engineering staff providing input design data and for the design
engineer estimating unknown parameters. It is also common practice for
design engineers to use a safety factor as a multiplier for initial size
estimates for process equipment. These factors may be as large as 1.5.
Furthermore, when used for clarifier design some engineers may apply the
safety factor to area while others may apply it to diameter. Unless an
effort is made to control these effects it is relatively easy for a process
design to become excessively conservative. For the TOSCO design, it was
endeavored to limit the tendency to "play it safe", since an economical
design was explicitly requested. Judgments concerning process sensitivity
and uncertainty were used as a guide for estimation of unknown quantities
required for design. It was also endeavored to provide broader process
flexibility to compensate for uncertainty in process efficiencies. For
example, a broad range of feed rates for MgSO^ can be used to offset a low
SiC>2 removal efficiency. The process design recommended to TOSCO was not
considered a conservative one. The provisions for process flexibility and
redundancy of key process units were included as alternatives to strongly
conservative assumptions and design.
PERFORMANCE OF TOSCO ZERO SLOWDOWN SYSTEM
Shortly after the final process design was completed and approved, a
contract was let to begin construction. The sidestream softening/wastewater
recycle plant was completed early in 1982. Operation of the plant commenced
191
-------
March 18, 1982. Startup of the softening process went very smoothly. All
softening chemicals recommended in the preliminary design were used during
softener startup. After about one week, the newly constructed softening
plant was lined out and operating well. During startup, TCC scrubber and
boiler blowdown streams were being fed to the softener system. In this sec-
tion the actual performance of the TOSCO softening system will be reviewed.
Measured performance will be compared to performance predicted during pre-
liminary design. These data will provide a convenient case for evaluation
of the techniques used for softener design. Finally, the economic benefits
accrued from wastewater recycle will be briefly reviewed.
COMPARISON OF PREDICTED AND ACTUAL PERFORMANCE
In the previous section, it was shown that no significant process
changes were made during final design. Thus, the basic premises used for
performance calculations should be applicable to the actual plant. However,
during process design it was endeavored to provide sufficient operating
flexibility in the softening plant to enable different modes of operation.
In this regard, each softener unit was sized to handle the design flowrate
and provisions were made to allow a fairly wide range of chemical feed
rates. These provisions allow actual influent flowrates to be varied when
both softeners are in operation. Also, a. variety of influent chemical com-
positions could be handled by varying individual chemical feed rates. Oper-
ating flexibility must be included in process design because actual water
quality is rarely identical to the quality used for design purposes. Fur-
thermore, chemical calculations presume equilibrium conditions which cannot
be attained in actual operation. Thus, actual process operations can be
quite different from conditions assumed during process design. Such differ-
ences can lead to a large discrepancy between actual and predicted perform-
ance. For these reasons the actual operating conditions will first be com-
pared with conditions used for design. These factors will be considered
when predicted versus actual performance data are contrasted.
ACTUAL VERSUS DESIGN OPERATING CONDITIONS
Several factors may cause a process operator to alter conditions
assumed to exist during process design. Plant expansions or modifications
can alter conditions known at the time process design was undertaken. Eco-
nomic conditions can have a strong influence over refinery operations, as
can the nature of feedstocks. Changes in refinery operations may alter heat
exchanges temperatures. These temperatures are of primary importance for
setting cooling water limits for scale forming species. In plants like
TOSCO using groundwater supplies, the quality of the we11water is also
likely to change with time. Changes in quality of cooling water makeup can
cause very significant changes in softener flowrate and chemical dosage
rates.
The data of Table 9-11 present a comparison between operating condi-
tions actually in use at the TOSCO plant and those used for design. Perhaps
the most important items listed in Table 3.1 are the flowrates for the cool-
192
-------
TABLE 9-L1. COMPARISON OF DESIGN ESTIMATE AND ACTUAL OPERATING DATA
Design
I ten Estimate Actual
pH 10.4 10.1 - 10.3
Qs 165 gpm 215 gpm
NaOH 340 mg/L 130 mg/L
Na2C03 380 mg/L 125 mg/L
MgS04 260 mg/L 230 - 260 mg/L
Softener Feed 245 gpm 130 - 1500 gpm (Avg 305 gpm)
All dosages are in mass units for chemical noted, 100% purity.
193
-------
ing tower sidestream, Qg, and softener feed. The actual flowrates are gen-
erally higher than values used for design. The sidestream flowrate from the
cooling tower system was increased by an average of about 50 gpm (30% of
design). This was done to reduce concentrations of scale component within
the cooling water. Actual softener feed rates encompass a rather wide
range, both above and below the design rate of 245 gpm. The upper limit on
feed rate of 1500 gpm was not used routinely, but was done during upset
conditions within the cooling towers. As usual, blowdown is used to control
cooling water quality. The design included two parallel process trains with
a stated capacity of 490 gpm when both trains were in use. The anticipated
effect of the increased flowrate through the softeners is to decrease the
quality of softener effluent and to maintain lower levels of scale chemicals
in the cooling water.
The softener pH data of Table 9-11 shows that actual practice was based
upon design recommendations. Control systems for chemical feed typically
operate over a band of concentrations. The design pH of 10.4 is near the
center of the actual pH range. However, the actual dosage for NaOH and
Na2C03 are well below design dosages. The reduced dosage reported for NaOH
is primarily due to maintenance of lower alkalinity in cooling water, as
will be discussed in a later section of this chapter. However, actual soda
ash dosage is well below the design level. The recommended soda ash dosage
was based upon maintaining equal molar concentratoins of calcium and total
carbonate species in the softener, assuming all influent alkalinity would be
converted to carbonate. The TOSCO plant was not operated in this manner.
Because soda ash is a relatively expensive chemical, its dosage was reduced
to improve process economy. Actual operation does not provide stoichiome-
tric carbonate concentrations. Since there is an excess of calcium hardness
in actual softener operations, actual effluent hardness should be higher
than levels anticipated during process design.
The data of Table 9-11 show that the estimated dosage for MgSO^ is very
close to the dosage used in actual operation. However, hardness levels in
the wellwater used for cooling tower makeup were lower than used for prelim-
inary design estimates. Wellwater hardness in 1982 had fallen to 42 ppm
calcium hardness (ppm CaC03) and 4 ppm magnesium hardness (ppm CaC03). As
shown in Table 9-1, during preliminary design it was assumed that calcium
hardness was 100 ppm CaC03 and magnesium hardness was 33 ppm CaC03. Based
upon current makeup water quality, the estimated MgSO^ dosage would be
higher than actual dosage. This comparison shows that the design approach
leads to a conservative estimate of MgSO^ addition for removal of silica in
the softener. However, the data also show that the method for dosage for
can be calculated with acceptable accuracy.
ACTUAL VERSUS ESTIMATED WATER QUALITY DATA
The previous discussions have provided the background for a comparison
of predicted and actual water quality data. These data are presented in
Table 9-12. Comparisons are presented for both softener effluent and
cooling water composition. The accuracy of water quality estimates for each
stream will be discussed. When necessary, information previously discussed
194
-------
will be used to better understand these comparisons.
The data for effluent from the softener are compared in Table 9-11.
Except for the differences noted for both Ca and Mg hardness, the agreement
between estimated and actual data is excellent. As shown by these data,
actual removal of Ga and Mg is significantly less than anticipated. It was
noted earlier that softener effluent hardness estimates were based upon jar
tests conducted on a synthetic water. The values used, 50 ppm CaC03 Ca
hardness and 20 ppm CaCC>3 rag hardness, were deemed optimistic but possible
based upon prior experience with similar softening systems. Calcium removal
is directly related to the dosage of soda ash. It was also noted that soda
ash addition was reduced to improve operating costs for the softener system.
With soda ash addition less than the stoichiometric amount effluent Ca hard-
ness would be expected to increase. Furthermore, higher hardness levels in
softener effluent would also cause the blowdown rate from the cooling water
system to be higher than predicted. This would result in reduced hydraulic
residence time in the reaction zone and resultant poorer removal efficiency.
It should also be noted that a polymaleic anhydride precipitation inhibitor
was added to the cooling water to control scale formation. Although this
inihibitor has been found to provide minimal interference with precipitation
softening operations, its presence may reduce removal efficiency for Ca and
Mg. Actual Mg removal for softener effluent was closer to the predicted
value than Ca removal.
A comparison of the remaining data for softener effluent shows excel-
lent agreement for the remaining components. Silica removal was very close
to predicted values. The design value of 40 mg/L Si02 was based upon prior
experience. This verifies the effectiveness of magnesium salt addition for
silica removal. The actual effluent pH is higher than anticipated, contrib-
uting to the discrepancy noted for alkalinity and sulfate concentrations.
The agreement for the components other than Ca and Mg are well within an
accuracy limit of 10%.
Data showing estimated and actual cooling water quality are included in
Table 9-12. These data show that predicted quality was reasonably close to
actual quality, but agreement was not as good as for softener effluent data.
All estimated data were higher than actual. The primary reason for this
discrepancy was the decision to maintain lower concentrations of scale form-
ing species in the cooling water. As noted previously, this was accom-
plished by increasing the rate of blowdown to the softener system, resulting
in a higher than predicted overall removal rate for Ca, Mg, and SiC^. The
primary reason for this was to reduce Ca concentrations to avoid potential
problems with gypsum (CaSO^) scale formation. Insufficient information on
actual operation was available to permit calculation of actual cycles of
concentration. The design concentration factor was about 42, however.
Detailed analytical data could not be obtained for wellwater and the
other feed streams, TCC scrubber blowdown and boiler blowdown. Thus, it is
difficult to ascertain the actual reasons for the discrepancies between
actual and predicted cooling water quality. The reduced alkalinity in
actual cooling water is due to the slightly reduced pH used for actual oper-
ation. This reduced alkalinity also decreased actual NaOH dosage, contrib-
uting to lower Na levels in the cooling water. The worst prediction was for
195
-------
TABLE 9-L2. COMPARISON OF ESTIMATED AND ACTUAL WATER QUALITY DATA
pH
Ca Hardness
(ppm CaCO})
Mg Hardness
(ppm CaCO^)
Na* (mg/L)
M.O. Alkalinity
(ppra CaC03)
el' (ng/L)
SO^ (mg/L)
Si02 (mg/L)
TDS (mg/L)
Na"1" concentration
8.2
50
20
2,900
25
1,100
4,700
40
8,300
was estimated
8.5
230-400
16-120
2,793*
35
1,200
4,500
42
3,700
to provide
7.0
360
230
3,540
75
1,525
6,470
200
12,250
electroneu trali
6.2-6.7
520-700
170-200
2,900-3,200
20-40
1,200-1,500
4,650-5,500
130-150
10,400
ty.
TABLE 9-13. ESTIMATED ANNUAL COST DATA
I tern Annual Savings Annual Cost
Wellwater Consumption
(300,000 gal/day @ 33c/1000 gal) $ 36,000
Disposal Well Operating Cost
(maintenance, etc.) 300,000
Reduced Chroma te Consumption 70,000
Disposal Well Tax Reduction 365 ,000
$771 ,000
Softening Chemicals 110,000
Operating Labor
(3 men per day @ $30/hr) 263,000
Softener Maintenance 100 ,000
$473,000
Estimated Net Cost Reduction $298,000
196
-------
sulfate, which was about 1400 mg/L, too high. The reason for this is
unknown, but may be due to actual wellwater alkalinity being lower than
anticipated. It was noted previously that makeup water quality had changed
for both Ca and Mg hardness. The chloride ion prediction appears to be
quite good, but preliminary estimates did not include contributions from
cooling water chlorination. Inclusion of chlorine feed in design calcula-
tions would have increased the predicted value for Cl. However, it is also
possible that actual drift losses were greater than the preliminary design
estimate of 53 gpm.
Overall, design predictions were sufficiently close to actual data for
both softener effluent and cooling water quality to permit a realistic
assessment of softener performance and need for chemical treatment of
cooling water. Predicted data were generally higher than actual data. This
would lead to conservative estimates of the impacts of softener operation on
the cooling water system. While excessive conservatism is undesirable, the
data of Table 9-11 show a moderate level of conservatism for most param-
eters. The data of Tables 9-11 and 9-12 show that careful application of
design techniques described in this report provide excellent estimates of
actual operating and performance data for sidestream softening systems.
GENERAL ASSESSMENT OF SIDESTREAM SOFTENER SYSTEM
The total cost for design and construction of the TOSCO sidestream
softening system in 1982 was $3,500,000. The TOSCO system provided for
recovery and reuse of cooling tower blowdown, boiler blowdown, and TCC
scrubber blowdown. Recovery of these wastewater streams significantly
reduced wastewater discharges at the Bakersfield refinery. Operating and
management personnel were pleased with system design and performance. Care-
ful planning and attention to detail during design and construction provided
for an easy startup. Since actual performance was fairly close to predicted
performance, there were no major "surprises" encountered during initial
operation. The accuracy of predicted performance enabled adequate prepara-
tion for cooling water treatment and sludge disposal. Operating costs, due
primarily to chemical additions and labor, were generally lower than antici-
pated. Actual economic benefits were accrued through operation of the
softener/wastewater recycle system that reduced overall refinery operating
costs. Cost reductions were achieved for wellwater supply, cooling water
chemical usage, maintenance costs, and superfund taxes.
The primary objective of the sidestream softener system at TOSCO was to
reduce wastewater production. This also leads to reduced demand for fresh-
water supply. Based upon actual operating data, wellwater consumption at
the Bakersfield refinery was reduced by a minimum of 300,000 gal/day after
the softener system was placed on-line. This led to estimated annual
savings of about $36,000, as shown in Table 9-13. The very high unit cost
of $0.33/1,000 gal includes power costs, well maintenance, and taxes on
groundwater usage. As noted previously, local and state government used
these taxes to increase incentives to reduce fresh water consumption. This
situation is unique because of the semi-arid climate at the Bakersfield
plant site.
197
-------
Recovery and recycle of cooling tower blowdown also had an impact on
consumption of cooling water treatment chemicals. The major changes noted
in cooling water treatment were increased concentration for both chromate
and Zn corrosion inhibitors. Chromate concentrations were nearly doubled,
going from about 20 mg/L to 30-40 mg/L Cr04 after startup of the softener.
Zinc concentrations were approximately 2 mg/L prior to startup and 3-5 mg/L
afterwards. The higher chromate concentration did not increase consumption
of chromic acid since chromate is not removed during softener operation. The
chromate is returned to the cooling tower with the treated effluent, effec-
tively reducing chroraate additions by recovery of cooling tower blowdown.
However, at the alkaline conditions maintained in the softeners, zinc is
removed as Zn(OH)2 with the sludge. Consequently, addition of Zn to cooling
water was increased by softener operation. These changes led to net savings
for cooling water treatment chemicals of approximately $70,000 per year, as
noted in Table 9-13. A potentially important cost increase could result
from increased corrosion rates within the cooling water system. Generally,
as the ionic strength of water increases the water becomes more corrosive.
The high levels used for both chromate and zinc were intended to offset this
effect. Measured corrosion rates at the TOSCO refinery using the standard
coupon test increased from about 0.4 mils per year (mpy) prior to wastewater
recycle to 0.7 mpy after recycling was implemented. Corrosion rates less
than 1 mpy are generally considered acceptable for well operated open
cooling tower systems (4,5,6). Thus, increased corrosion rates were not
considered to represent an additional cost burden.
Prior to softener operation, blowdown streams from the cooling tower
system, boilers, and TCC scrubber were disposed by deep well injection.
Operation of their disposal well represented a very significant cost for the
Bakersfield refinery. These costs include electrical power, maintenance,
and taxes. Power consumption in disposal well operations is strongly influ-
enced by the volume of wastewater injected. The injected wastewater must be
forced through a porous sand in the disposal zone. This zone is subject to
pluggage by solids remaining in the wastewater after filtration. Higher
injection well flowrates increase both the pressure drop required to main-
tain well flow and the rate of buildup of pressure drop due to plugging.
Both factors will have an impact on pumping efficiency for the injection
well system. Higher rates of plugging also increase the frequency of well
shutdown for cleanup operations designed to remove deposited solids. These
factors show that injection well operating and maintenance costs do not
increase linearly with wastewater flowrate; the incremental cost per unit
volume of wastewater increases as flow increases. While it is difficult to
isolate these costs, TOSCO personnel estimate that approximately $300,000
per year will be saved by recovery and re-use of wastewater previously
injected. These data are included in Table 9-13.
The superfund taxes levied on disposal well operation represent a sig-
nificant cost to the TOSCO refinery. These taxes are based on the volume of
wastewater injected via disposal well. The TOSCO sidestream softening
system reduced flow to the disposal well by about 170,000 gal/day. As shown
in Table 9-13, TOSCO estimates that superfund taxes would be reduced by
$365,000 per year due to recycle of wastewater streams. This represents an
average tax rate of about $5.89 per 1,000 gal of wastewater. This presents
198
-------
a very strong incentive for reduction of wastewater generation rates. The
data of Table 9.13 show that superfund taxes represent the single largest
savings for softener operation. Disposal well cost reductions, consisting
of maintenance and taxes, represent over 85% of total savings estimated.
The increased plant costs for softener operation consist of softener
chemical costs, operating labor, and estimated maintenance costs, as noted
in Table 9-13. Actual softener chemical costs of about $110,000 per year
were estimated, based upon initial softener operations. Chemical costs
estimated during process design were $135,000 per year. Softener operation
required addition of one operator per shift, leading to an estimated annual
cost of about $263,000. Although long term softener maintenance cost data
were unavailable, TOSCO estimates annual maintenance costs at $100,000
resulting in a total cost increase of about $474,000 annually. Data con-
cerning additional costs and benefits such as interest, property taxes,
income taxes, and depreciation were not available to the authors. However,
it is likely that these factors would increase the economic benefits attrib-
utable to softener operation.
The data of Table 9-13 show that the estimated total cost increase for
softener operation was $473,000 annually while estimated annual cost reduc-
tion was $771,000. The net effect on refinery operating cost was a decrease
of almost $300,000 annually. This net savings represent 8.6% of the total
cost of plant design and construction. The TOSCO design provides an eco-
nomic incentive for wastewater recycle in addition to the more general
incentives for environmental protection and water conservation. Incentives
for social concerns such as the two noted above are difficult to quantify in
economic terms. The TOSCO case is unique in that both economic and social
incentives favored construction and operation of wastewater recovery facili-
ties. The data of Table 9-13 show that the economic incentive is mostly
attributable to reduced tax burdens, however. In the absence of superfund
taxes, annual refinery operating costs would have increased, but the
increased cost would be minor.
RECOMMENDATIONS AND CONCLUSIONS
The design used for sidestream softening and wastewater recycle at the
TOSCO refinery in Bakersfield, California, provided a reliable and cost
effective system. The design technique used for preliminary process design
presented an accurate, well defined picture of actual process performance.
This design technique is obviously well suited for sidestream softener pro-
cess calculations when applied in a conscientious manner. Preliminary cal-
culations can be used to assess potential impacts of softener operations on
the overall plant with a high degree of confidence. The design technique
can also be used to investigate different treatment chemical combinations
with the objective of selecting the lowest cost alternative. Such a review
for the TOSCO plant led to selection of caustic soda rather than lime to
elevate pH in the softeners. Each plant site is unique, however, and dif-
ferent conditions may result in a different recommended chemical treatment
scheme.
199
-------
Further work is necessary to improve predictions for removal of hard-
ness in sidestream softeners. Prediction of effluent hardness was quite
poor for the TOSCO design. This work should include consideration of
effects of scale inhibitors on softener performance. Because several scale
inhibitors can strongly influence precipitation kinetics, their presence can
degrade softener performance. The situation is further complicated by the
multiplicity of scale inhibitors available for use in cooling waters. The
isotherm method used to predict silica removal is well suited for design of
sidestream softeners, however.
200
-------
,.c
'
QT,CT
^
*5 w
H SO
IT,
t
COOLINC
TOWER
SYSTEM
SIDESTREAM
SOFTENER
1 1 1
VCs QC>CC QTCC,
1
vs
Qs'Cw
"TCC
LEGEND
Q = Evaporation Loss
Q^ = Drift Loss
Q = Make-up Flow Rate
Q™ - Additive Flow Rate
Q = No. 5 Cooling Tower Slowdown
Q_ = No. 1-4 CT Slowdown
Q, = Boiler Slowdown
Q - TCC Scrubber Slowdown
Q = Softener Chemical Feed Rate
Q~ = Waste Sludge Flow Rate
0^, = Treated Water Flow Rate
FIGURE 9-1 SIDESTREAM SOFTENING SYSTEM SCHEMATIC DIAGRAM
201
-------
TCC Scrubber Slowdown
(25 gpm)
o
K>
Boiler
Slowdown
Solids
Contactor
Filter Feed .fJlkJackwash Line
Tank X"^>v
Filters
(7 ft)
Mixed
Media
Makeup t Additives
(1518 gpm)
Treated Water (235 gpm)
FIGURE 9-2 PRELIMINARY DESIGN SIDESTRKAM SOFTENING PROCESS SCHEMATIC
-------
Boiler
N>
O
TCC Scrubber Blowdoun
(25 gpm)
Solids
Contactor
(D=22.5 ft
11=14 ft)
Filters
(7.5 ft)
Mixed
Media
Mnkeup + Additives
(1518 gpm)
No. 5 CT
Bloudoun
(55 gpm)
Treated Water (235 gpm)
FIGURE 9-3 FINAL DESIGN SIDESTREAM SOFTENING PROCESS SCHEMATIC
-------
References
1. Matson, J.V., "Cooling Water Recycle by Softening," National Science
Foundation, ENV 77-06504, 1979.
2. Davies, C.W., Ion Association, Butterworths, London, 1962.
3. Matson, J.V., "Zero Discharge of Cooling Water by Sidestream Soft-
ening," Journal of the Water Pollution Control Federation, Vol. 51, No.
11, 1979, pp. 2602-2614.
4. Betz Handbook of Industrial Water Conditon, Betz Labortories, Trevose,
Pa., 1982.
5. Nalco Water Handbook, McGraw-Hill, 1980.
6. Drew Priciples of Industrial Water Treatment, Drew Chemical Company,
Boonton, N.J., 1977.
204
------- |