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

-------