I
I
I
I
I
I
I
1
1
I
     &EPA
                  United States
                  Environmental Protection
                  Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/1-90-002
March 1990
                  Air/Superfund
                  AIR/SUPERFUND
                  NATIONAL TECHNICAL
                  GUIDANCE STUDY SERIES
                  Comparisons of Air Stripper
                  Simulations And Field
                  Performance Data

-------
 I
 I
 I
 I
 I
 I
                                   •COMPARISONS  OF  AIR  STRIPPER
                             SIMULATIONS  AND  FIELD PERFORMANCE DATA
 I
 I                                            by
                                      IPEI Associates,  Inc.
                                       11499  Chester Road
                                     Cincinnati, Ohio  45246
I
                                    Contract No. 68-02-4394
                                          \ssignment f
                                          PN 3759-25
I                                  Work Assignment No.  25
I
                            James Durham, Technical Representative


I                           U.S. ENVIRONMENTAL PROTECTION AGENCY
•                       OFFICE OF AIR QUALITY PLANNING AND STANDARDS
                         RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711
I
I
I
                                         February  1990

-------
                                 DISCLAIMER


     This report was prepared for the U.S. Environmental  Protection Agency by
PEI Associates, Inc., Cincinnati, Ohio,  under  Contract  No.  68-02-4394, Work
Assignment No. 25.  The contents are reproduced  herein  as received from the
contractor.  The opinions, findings, and conclusions  expressed are those of
the authors and not necessarily those of the U.S.  Environmental  Protection
Agency.

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                  CONTENTS


Figures 	      v
Tables	     vi
Acknowledgment	viii

     1.   Introduction	      1

               Air stripping principles 	      1
               Basic elements of ASPEN software 	      3
               User interface software	      4
               Input/output format	      4
               Comparison of ASPEN simulations	      5

     2.   Site Selection Criteria 	      6

               Pollutant type	      6
               Pollutant concentration	      8
               Water volume treated 	      9
               Air/water ratio	      9
               Control equipment	     10
               Data availability and quality	     12

     3.   Site Background Summaries	     13

               Tacoma Well 12A	     13
               Rockaway Township Site 	     18
               Brewster Well  Field No. 1	     23
               Verona Well Field	     26
               Western Processing Site	     33
               Hicksville MEK Spill  Site	     38
               Gilson Road (Sylvester's) Site	     42

     4.    ASPEN Simulations and Performance Comparisons 	     47

               General results of the ASPEN site comparisons	     47
               Site-specific  comparisons	     49
               Summary	     66

     5.    Emissions Tradeoffs From Air Strippers	     68

               Emissions estimate methodology 	     70
               Emissions comparisons for the control  options	     72
               Summary	     78

-------
                            CONTENTS (continued)
                                                                          Page
     6.   Conclusions and Recommendations 	    80
            Conclusions	    80
            Recommendations 	    82
References	    84
Appendices
     A.   Summary of Site Parameters	    85
                                       IV

-------
1
1
1
1

1
•


1
IB




1
•





I
•



1
1
1
1
1
1
1
1

•

FIGURES
Number
1 Diagram of the air-stripping process 	 ....

2 Process flow diagram for Tacoma Well 12A 	

3 Process flow diagram for Rockaway Township 	

4 Process flow diagram for Brewster Well Field 	

5 Process flow diagram for Verona Well Field . . 	

6 Process flow diagram for Verona Well Field during removal
action of highly contaminated zone 	

7 Process flow diagram of the CADRE system at Western
Processing 	

8 HTAS process flow diagram for Hicksville MEK Site 	

9 HTAS process flow diagram for Gil son Road Site 	 	

10 Process flow diagram for "cross-flow" air stripper for
handling reduced gas volumes ..... 	







V





Page
2

17

22

25

29

30


34

40

46


83










-------
TABLES
Number
1
2

3

4
5
6
7
8
9

10
11

12

13
14

15

16

17


Pollutants Identified at the Selected Simulation Sites . . .
Range of Concentrations of the Most Commonly Identified
Compounds 	
Currently Operating Facilities Using Air Emission Controls
on Air Strippers 	
Summary of Selected Sites 	 	 	
Summary of Selected Design Data for Tacoma Well 12A 	
Summary of Selected Design Data for Rockaway Township. . . .
Summary of Selected Design Data for Brewster Well Field. . .
Summary of Selected Design Data for Verona Well Field. . . .
Concentration Range for Selected VOC Contaminants at the
Western Processing Site 	
Summary of Selected Design Data for Western Processing Site.
Summary of Selected Design Data for Hicksville MEK Spill
Site 	
Concentrations of Organic Contaminants Found in the Gilson
Road Site Ground water 	
Summary of Selected Design Data for the Gilson Road Site . .
Comparison of ASPEN Simulations to Actual Performance
at the Tacoma Well 12A Site 	
Comparison of ASPEN Simulation to Actual Performance at the
Rockaway Township Site 	
Comparison of ASPEN Simulation to Actual Performance at the
Brewster Well Field Site 	
Comparison of ASPEN Simulation to Actual Performance at the
Verona Well Field Site 	
Page
7

8

11
14
16
20
24
28

32
35

41

43
44

51

53

55

56

-------
1
1
1

1





1

I
•
1
•

1






1
1
1
1
1
1
1

1


Number
18


19

20

21

22

23
24
25

26
27

28










TABLES (continued)

Summary of Inlet and Outlet Concentrations and Collection
Efficiencies for Selected Compounds, April 30, 1987. . . .

Comparison of ASPEN Simulation with Actual Performance at
the Western Processing Site 	
Comparison of ASPEN Simulation with Actual Performance at
the Hicksville MEK Spill Site 	
Comparison of ASPEN Simulation with Actual Performance at
the Sylvester's Gilson Road Site 	
Emission Factors Used by ASPEN Model for Fuel Combustion
Emissions 	

Emissions Comparison for the Tacoma Well 12A Site 	
Emissions Comparison for the Brewster Well Field Site. . . .
Emissions Comparison for the Verona Well Field Site 	

Emissions Comparison for the Western Processing Site ....
Emissions Comparison for the Hicksville MEK Site 	

Emissions Comparison for the Sylvester's Gilson Road Site. .







vii



Page

59


61

65

65
71
/ -L
73
73
75

75
77

77










-------
                               ACKNOWLEDGMENT


     This report was prepared for the  U.S.  Environmental  Protection Agency,
Office of Air Quality Planning and Standards,  by  PEI  Associates,  Inc.,
Cincinnati, Ohio.  The project was directed by Mr.  David  Dunbar and managed
by Mr. Gary Saunders.  The principal author was Mr. Gary  Saunders.   The
author would like to acknowledge Mr. James  Durham,  the  Agency's Technical
Representative, for his overall guidance and direction,  and Mr. Tony Rogers
and Dr. Ashok Damle from Research Triangle  Institute  (RTI)  for their work on
the ASPEN simulations.

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                  SECTION  1
                                INTRODUCTION

1.1  AIR STRIPPING PRINCIPLES
     One of the more common problems noted at  Superfund  sites is the
contamination of ground water by volatile  organic  compounds  (VOCs).   In some
cases, the contamination has been discovered because  the VOCs have
contaminated an aquifer used as a drinking water supply  for  a community.  In
other cases, the contamination is found in and near dumped or spilled
materials that threaten to contaminate available water resources.   One
remedial alternative that is used to reduce or remove the VOC contamination
from water is air stripping in a tower that uses either  packing media or
trays.
     Air stripping generally involves the  countercurrent contact between air
and contaminated water by use of a packing material or trays to provide a
large surface area for the transfer of VOCs from the  water to the air (Figure
1).  The ability to strip a compound from  the  water depends  on several
factors, including the air/water ratio, the packing or tray  type,  and the
Henry's Law value for the compounds of interest.   The objective is to remove
the VOCs from the water.
     When being considered for remediation purposes,  the air stripper design
should be evaluated for removal efficiency and cost of operation.   A design
evaluation may examine variations in water flow, chemical  composition,
contamination levels, and different stripper design considerations (air/water
ratios, packing types, packing height, etc.).   One approach  to this
evaluation is a computerized simulation of key design parameters.   Although
numerous program approaches are available, a computerized process simulator
(known as ASPEN) was used in this project  to simulate the stripping process
and to evaluate the capital and annual costs of stripper operations.
     The purpose of this project was to collect available design and
operating data on operating air strippers  and  to input the design and
operating parameters into the ASPEN simulator  through a  user interface

-------
Contaminated
Water
            Storage
              Tank
           (Optional)
                         Pump
                                                        voc
                                                      Control
                                                      (Optional)
                                           Packing
Air
                                               "Clean" Water
            Figure 1. Diagram of the air-stripping  process.

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
program.  The results from the ASPEN  simulator were compared  to  the operating
data gathered for the sites to determine  the  relative  accuracy of the ASPEN
model results when compared with the  actual performance data.  A wide range
of.design air/water ratios, pollutant concentrations,  and  pollutant types
were sought for comparison of actual  performance  versus ASPEN predictions.  A
total of seven sites were used for comparison purposes.

1.2  BASIC ELEMENTS OF ASPEN SOFTWARE
     The ASPEN process simulation software (the  acronym ASPEN represents
Advanced System for Process Engineering)  is designed  to aid in  the evaluation
of process unit operations, energy and material  balances,  and sizing and cost
of major pieces of process equipment.  One of the advantages  of  the ASPEN
simulation software is that a complicated process can be  defined by
simplified modules constructed in a flowsheet style.   In  addition, physical
and thermodynamic properties of chemicals can be accessed by  built-in
libraries and routines.  This minimizes the need to  obtain physical and
thermodynamic properties for specific compounds.   Using these features, an
ASPEN-based model of an air stripper has been developed that  includes unit
operations for the control of air emissions by  vapor-phase carbon adsorption
or catalytic oxidation.  The ASPEN air stripper module uses the Onda-
correlation method to estimate mass transfer  coefficients for liquid and gas
phases  in the calculation  of stripping efficiency.
     Although the programming concept of ASPEN  is relatively simple, the
actual  programming and data input in dimensional  units compatible with the
ASPEN language are somewhat complicated.  The development of the air stripper
module  eliminates the need to program in ASPEN,  but  data  input is still
required.  To this end, a  user-friendly-data  input program was developed to
generate the data input file needed to run the  ASPEN air  stripper module.
This user interface software allows the generation of the data input file
while the ASPEN air stripper module programming remains transparent to the
user, and no knowledge of  the ASPEN programming language  or file structure is
required.

-------
1.3  USER INTERFACE SOFTWARE
     The Research Triangle  Institute  (RTI) programmed and compiled  the user
interface software in an executable file  in the BASIC programming language.
The interface software allows  a  user  to choose between evaluating an existing
air stripper design or "creating"  an  air  stripper design to  achieve a desired
removal efficiency for specified VOCs.  The ASPEN simulation can  evaluate
performance for simultaneous  removal  of up to 20 VOCs from a library of
approximately 400 chemicals.   Default values are provided throughout the
interface software, and key chemical  parameters are automatically accessed by
the choice of VOCs.  For the  seven sites  used for the ASPEN  comparison the
model was run in "rating" mode to  evaluate existing stripper designs.
     The air stripper model is supplied with options for air emission
controls.  The user can select from the  following options:   1) no control, 2)
vapor-phase carbon adsorption, and 3) catalytic  oxidation.   He/she can also
select  for inclusion in system evaluation a  liquid-phase carbon adsorption
module  for final  "polishing" of the water exiting  the air stripper.

1.4   INPUT/OUTPUT  FORMAT
      Because of  its  complexity  and hardware requirements,  the ASPEN
simulation  software  is  maintained on a VAX mainframe computer and  cannot  be
run  on a personal  computer.  The  user interface software is designed  to
operate on  PC's  to generate the ASPEN input file required for proper
execution of the program.
      The ASPEN  output  format  that is  incorporated  into the  ASPEN simulation
 software provides results  in  a  form  that most people do not find very useful.
Therefore,  a customized output  report format had to be provided that  presents
 results in a format that the  user finds  both useful and readable.   The output
 report consists of three main sections:  background information  on ASPEN, a
 summary of input data,  and a  performance and economic analysis.  These
 sections provide the basis for a  short engineering-style summary report of an
 air stripper design.  The ASPEN simulation for the selected sites  uses this
 format for the output report.

-------
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
 I
I
I
1.5  COMPARISON OF ASPEN SIMULATIONS

     The project included a comparison  of  ASPEN  simulations  with the

performance reported for operating air  strippers with  regard to the accuracy

of • performance and cost predictions.  The  operating data  for seven air

stripper systems representing a variety of designs and chemical  species to be

stripped from ground water were selected to  achieve this  goal.   It was also

deemed desirable to evaluate the operation and costs of emission controls.

Of course such an evaluation was subject to  data availability and quality.

The results of these comparisons are discussed in the  following sections.

-------
                                  SECTION 2
                           SITE  SELECTION CRITERIA

     The sites selected for evaluation and comparison with ASPEN  simulation
results had to meet several criteria.  The seven sites  selected for
evaluation are all  included on the National  Priorities  List  (NPL)  as
Superfund sites.  The U.S.  Environmental Protection Agency  (EPA)  Regional
Offices, State programs,  and the potential sites themselves  were  contacted
concerning the availability of data.  Two reports were  also  used  as a
starting point for identifying existing  air  stripper systems.  '    Several  of
the sites identified in these reports as having air emission controls are not
yet operational, or no data were available.   Two important  criteria for the
ASPEN simulations were data availability and data quality.   The other
selection criteria will be discussed  in  the  following  subsections.

2.1  POLLUTANT TYPE
     Available information on Superfund  site contamination  indicates that
trichloroethylene is one of the  most  commonly identified  contaminants at
Superfund sites.  The chemical  characteristics of trichloroethylene, most
notably its Henry's Law constant, make  it  a  relatively easy compound to strip
out of water.  A diversity in pollutants to  be stripped from ground water
(representing a wide range of Henry's Law  constants) was  sought  to evaluate
the ability of the ASPEN simulation and  chemical library  to make  accurate
predictions of removal efficiencies for  the  various  compounds.
Trichloroethylene was found at  six of the  seven sites.   Other compounds
frequently identified included  tetrachloroethylene  (perchloroethylene),
1,1,1-trichloroethane, chloroform, dichloromethane  (methylene chloride),
1,1-dichloroethylene, 1,2-dichloroethylene (cis-  and trans- forms), and vinyl
chloride.  Table 1 presents a complete  listing of  the  25 compounds found  at
the seven selected sites.  The Henry's  Law constants for these compounds
range  from 10"2 to 10"6 atm-m3/mole,  which represents  the least difficult  and

-------
1
1

1







1
•H



1

•
*



1

•
•



1

1

1
1
1
1


TABLE 1. POLLUTANTS




Compound

Acetone
Acrylonitrile
Benzene
Carbon tetrachloride
Chloroform
1,2-Dichlorobenzene
1,1-Dichloroethane
1,2-Dichloroethane
1,1-Dichloroethylene
1 , 2-Di chl oroethyl ene
CIS
trans
Hexachloro-l,3-butadiene
Hexachl oroethane
Isobutanol
Dichloromethane
Methyl -tert-butyl ether
Nitrobenzene
Isopropyl alcohol
1,1,2 , 2-Tetrachl oroethane
Tetrachl oroethyl ene
Toluene
1 , 1 , 1-Tri chl oroethane
1,1,2-Trichloroethane
Trichloroethylene
Vinyl chloride
Surrogates
Ethyl -propyl ether





IDENTIFIED AT THE SELECTED




Number of sites

1
1
1
1
3
1
2
2
2
3


1
1
1
4
1
1
1
1
4
2
4
1
6
3




7


SIMULATION SITES .



Chemical abstract
service (CAS) -number

67-64-1
107-13-1
1076-43-3
56-23-5
865-49-6
95-50-1
75-34-3
107-06-2
75-35-4

156-59-2
156-60-5
87-68-3
67-72-1
78-83-1
75-09-2
1634-04-4
98-95-3
67-63-0
79-34-5
127-18-4
108-88-3
71-55-6
79-00-5
79-01-6
75-01-4

628-32-0




-------
the most difficult to strip  compounds, respectively.  Most  of the  sites
include several  compounds, and  their  complexity ranges  from removal  of a
single compound  to as many as  19  compounds.  Thus,  the  seven sites not only
provide the compound most commonly  found  in contaminated  ground water, but
also provide an  extensive range of  contaminants for evaluation.

2.2  POLLUTANT CONCENTRATION
     Contaminant levels ranging from  a few parts  per  billion (ppb) by weight
to the 100,000 ppb range were  desired.  At the seven  sites, the average total
VOC concentration levels typically  ranged from 10 to  350  ppb.  In  nearly all
cases, the actual contamination levels were below the initial design
concentrations based on monitoring  well data.  In several  cases, the actual
long-term influent level was 10 to  100 times less than  the design
concentration.  When pumping and  treatment began, the concentration declined
from the initial values found  in  monitoring wells to  a  substantially lower
value.  It is hypothesized  that the sampling and  monitoring wells  were
established under "static"  conditions,  and the dynamic  action of pumping
water established a new equilibrium.
     The decrease in VOC concentrations  under  these circumstances  result in
an overdesign of the air stripper to achieve the  target effluent levels.  The
range of pollutant concentrations at the  various  sites, however, allows for
the evaluation of several different removal  efficiencies.  Many of the  sites
with the more common VOCs are treating water to  attain levels of less  than  5
ppb of  specific compounds.   Table 2 shows the  range of influent
concentrations for the most common compounds.


  TABLE  2.  RANGE OF CONCENTRATIONS OF THE MOST COMMONLY IDENTIFIED COMPOUNDS
Compound
Chloroform
1,2-Dichloroethylene
Methyl ene chloride
Tetrachl oroethyl ene
1,1,1-Trichloroethane
TCE
Range,
1.3 -
12.5 -
31.4 -
5 -
8.2 -
1.4 -
ppb
781
100
8170
378
1440
8220
                                      8

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
2.3  WATER VOLUME TREATED
     The typical air stripper treats less than 500 gal/min.   The water volume
treated by the seven selected sites ranges from 100 to 3500  gal/min.  These
air strippers represent a wide range of designs.   In all  but one case, the
air stripper design involved a single stripping column.   In  the case of the
air stripper treating 3500 gal/min, five parallel  stripper columns are used
to treat the ground water to limit the height and  diameter of the column
required.

2.4  AIR/WATER RATIO
     Besides the physical characteristics (Henry's Law,  temperature) and
desired removal efficiency, the air/water ratio is an essential design
parameter in the sizing and performance of an air  stripper.   The sites
selected represent a wide range of design air/water ratios,  from a low of 20
to a high of 310 (volume basis).  In general, lower values result in taller
columns, depending on the compounds involved.  Lower air/water ratios can
also result in lower gas volumes to be treated by  additional gas emission
control systems and higher VOC concentrations in the uncontrolled gas stream.
In general, lower air/water ratios reduce VOC control  costs  and result in
more effective operation of VOC controls.  Actual  designs typically represent
a compromise between compounds to be treated, the  variability in the influent
concentrations, and total gas volume to be handled.  Most new designs use
air/water ratios of 125 or less.
     An additional  factor encountered during  the data acquisition was the use
of high-temperature air strippers (HTAS).  Two of  the systems use HTAS to
improve removal efficiency of hard-to-strip compounds [e.g., alcohols,
acetone, methyl ethyl ketone (MEK)] without excessively increasing the
air/water ratio.  The operation at elevated temperature (greater than 140°F)
modifies the Henry's Law constants for these  compounds.   High-temperature
strippers increase the energy costs for preheating the water.

-------
2.5  CONTROL EQUIPMENT
     The two control  equipment  options  for the ASPEN air  stripper simulation
modules made it desirable to obtain  data  for comparable systems  from actual
operating strippers.   The air stripper  report produced by Radian indicated
the existence of such air strippers, most of which operate without VOC
controls.1  Four of the 12 sites  used an  incineration technique  that was not
compatible with the ASPEN module  (direct  incineration) for VOC controls.  The
remaining sites were either not operating or no data were available for use
in this study.  Table 3 includes  all sites currently operating with VOC
controls.  Three of the sites shown  in  this table are included  in the
comparison study.  Several Superfund sites also propose controls for future
air stripping operations.
     Three of the seven sites selected  for the ASPEN performance comparison
used some kind of VOC control.  The  Verona Well Field Site uses  vapor-phase
carbon adsorption.  The carbon, however,  is not regenerated on  site; it is
changed every 6 to 12 months, as  necessary.  The Western  Processing Site also
uses vapor-phase carbon adsorption to  control VOC emissions.   This system
differs in that it is regenerated by gases from a direct-fired  incinerator
(CADRE), which, in turn, destroys the  VOCs released  from  the carbon beds
during regeneration.  The Gilson  Road  Site uses the  boiler that  preheats the
stripper water to its operating temperature as  an  incinerator to destroy VOCs
liberated by the air stripper.
     In the future, many more air strippers will probably include VOC
controls.  For this project, however,  only a limited number of sites could
provide useful data.  No data were available from  sites  using catalytic
oxidation; however, this is still considered a  possible  control  technology
for air strippers.  Catalytic oxidation systems  are  also  being considered and
proposed for  in situ soil vapor extraction, which  has  similar gas-stream
characteristics.  Direct incineration  or hybrid  systems  such as CADRE  are not
currently included as ASPEN options.
                                      10

-------
           TABLE  3.   CURRENTLY OPERATING FACILITIES USING AIR EMISSION CONTROLS ON AIR STRIPPERS
Facility name
BKK Landfill
City of Plainfield
Hughes Aircraft
Lowry Landfill I
Verona Well Field
McClellan AFB
Gil son Road
Chem Central
Western Processing
Unifirst
Chem-Dyne
Palos Verdes
Landfill
City
West Covina
Plainfield
Tuscon
Denver
Battle Creek
Sacramento
Nashua
Grand Rapids
Kent
Williamstown
Hamilton
Palos Verdes
State
CA
NO
AZ
CO
MI
CA
NH
MI
WA
VT
OH
CA
Type3
LF
DW
GW
LF
GW
GW
GW
GW
GW
GW
GW
LF
Major b
contaminants
Landfill leachate
PCE, TCE
TCE, DCA, TCA
1,1 -DCA, 1,2-DCA
1,1,1-TCA, 1,1-DCA,
PCE, TCE
MEK, acetone,
various VOC
MeOH, EtOH, acetone
MEK, Tol, others

1,2 DCA, Tol, PCE,
TCE, TCA, chloro-
form, methyl ene
chloride, others
PCE

Landfill leachate
Typec
Flare
GAC
GAC
GAC
GAC
INCIN
INCIN
GAC
GAC
GAC
GAC
Flare
Control
required,
Y/N


Y
Y
Y
Y
Y

Y
Y


aLF = landfill, DW = drinking water,  and GW = ground water.
bPCE = Tetrachloroethylene; TCE = trichloroethylene; TCA = 1,1,1-trichloroethane; 1,1-DCA =
 1,1-dichloroethane; 1,2-DCA = 1,2-dichloroethane;  MEK = methyl ethyl  ketone;  MeOH = methandl;  EtOH =
 ethanol;  Tol = toluene.
CINCIN = incineration;  GAC = granulated activated carbon.

-------
2.6  DATA AVAILABILITY AND QUALITY
     A major limitation in obtaining  data was the assurance  that  available
data were of good quality.  Some  sites were eliminated because  the  number of
contaminants were limited to one  or two components already well represented
(e.g., trichloroethylene).  In  most cases, however, complete or nearly
complete design or operating data were not readily available from the  sites.
For example, in some cases,  initial contacts with Agency or  site  personnel
indicated that complete design  information was not available or that
operating data were limited. The lack of performance data for  VOC  control
systems was particularly a problem  at sites that use emission controls.   In
other cases, there was a reluctance to provide information,  either  because of
legal or political situations or  because of the effort required to  provide
the information.  Where data were not forthcoming, the sites were dropped
from further consideration in this  project.
     For the seven sites selected,  the design and performance data  are
relatively complete and are of  sufficient quality to provide reasonable
accuracy.  One area in which data completeness was a problem involved  costs.
In general, the basis of the cost values provided was difficult to  ascertain.
The costs for the Brewster Well Field Site and the Hicksville MEK Site were
the exception.  The allocation  of costs was well defined for these  sites.
For the remaining sites, only general costs were provided, and  documentation
of what was included in the direct  and indirect costs was limited.
                                    12

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                  SECTION  3                   -  -     -
                          SITE BACKGROUND  SUMMARIES

     As discussed in Section 2, the seven  sites  selected  represent a
substantial diversity in design,  contaminants, concentrations,  and VOC
controls.  Although none of the sites is entirely representative of the
majority of operating air strippers,  together they provide a wide range of
parameters for the ASPEN simulation process  to use for comparison of actual
and simulation values.  The air stripper designs at  these sites also should
be within the range of designs and concentrations expected in future site
remediations.  This section presents background  information on  each of the
selected sites.  Table 4 summarizes pertinent information, and  Appendix A
provides detailed design and operating information.

3.1  TACOMA WELL 12A
     The Tacoma Well 12A Site is located  in  the  southern  part of Tacoma,
Pierce County, Washington.  During the summer months,  Well 12A supplies the
water-processing plant with the higher daily average and  peak flows
associated with summer water usage.  Contamination of the well  was first
discovered in 1981, at which time the well was shut  down.  No alternative
supply of water could be developed to replace the lost production of this
well (3500 gal/min).  In the meantime, the site  was  placed on the NPL and
became eligible for Superfund monies.
     During the time the source and extent of aquifer contamination were
being determined, an interim remedial action was planned  to return the well
to service while the VOCs were being removed from the well water.  The plan
called for the use of a pilot-scale stripper to  document  the feasibility of
air stripping as a remedial alternative and  the  subsequent scale-up of the
stripper if the pilot-scale treatment proved successful.   The air stripper
would then be installed and operated to remove VOCs  from  the water.  A short
time frame of 11 months was planned for implementation of this  plan.
                                      13

-------
                     TABLE 4.  SUMMARY OF SELECTED SITES

Site
num-
ber Site name
1 Tacoma Well 12A
2 Rockaway Township
Location
Tacoma, WA
Rockaway
Township, NJ
Primary
contaminants
Tetrachloroethane
Trichloroethylene
Dichloroethylene
Trichloroethene
Methyl -tert-butyl
ether
VOC
con-
trols
No
No
Comments
5 towers to
handle 3500
gal/mi n
Other com-
pounds
identified
 3   Brewster Well
      Field
Brewster,  NY
Diisopropyl ether

Tetrachloroethylene
Trichloroethylene
Dichloroethylene
No
4





5





6

7



Verona Well Field





Western
Processing




Hicksville MEK
Spill
Gil son Road
(Sylvester's)
Site

Battle Creek,
MI




Kent, WA





Hicksville,
NY
Nashua, NH



1,1,1-Trichloro- Yes
ethane
Trichloroethylene
Tetrachl oroethyl ene
Dichloroethylene
Other VOCs
Dichloromethane Yes
Trichloroethylene
1,1,1-Trichloro-
ethane
Methyl ethyl ketone
Other VOCs
Methyl ethyl ketone No

Isopropanol Yes
Acetone
Toluene
Dichloromethane
Nonregener-
able vapor-
phase GAC



Vapor-phase
GAC with
incinerator
used during
regenera-
tion
HTAS

HTAS boiler
used for
incinera-
tion

     The remedial investigation determined potential  areas and sources of

aquifer contamination.  The VOC concentration in the  water varied because

seasonal startup of the well drew in essentially uncontaminated or only

slightly contaminated water before drawing in the contaminated plume

resulting from the reversal of the normal  aquifer flow.
                                      14

-------
I
I
I
I
I
I
I
1
I
I
I
I
I
I
I
I
I
I
I
     Four chlorinated organic solvents  were  identified  in  the well  water.
These were 1,1,2,2-tetrachloroethane, trans-l,2-dichloroethylene,
trichloroethylene, and tetrachloroethylene.  The pilot-scale testing included
various air/water ratios and column packings at VOC concentrations  at or near
the maximum levels encountered during normal seasonal operations.   Successful
operation of the pilot-scale unit indicated  that air  stripping  could be used
to remove VOCs from the water.  Of the  four  compounds,  1,1,2,2-tetrachloro-
ethane was the most difficult to remove and  had the highest design
concentration (300 ppb).
     The scale-up of the air stripper design called for five parallel towers
packed with 1-inch saddles.  The use of five towers allowed for shutdown of
one tower for maintenance while maintaining  the overall  removal  efficiency.
Each tower handles 700 gal/min.  The design  air/water ratio is  310  ft
air/ft3 water, which represents the highest  value  of  air/water  ratio in the
seven selected sites.  The Tacoma Well  12A Site also  represents the largest
water treatment volume of the seven sites.   As  is  typical  with  many sites,
actual operation usually results in influent levels substantially  lower than
the initial design value.  The air stripper, however, must be designed tc
provide a given effluent level for the  highest  concentrations of the various
contaminants.  The compound that is most difficult to strip may also be a
controlling factor.  In this case, the  1,1,2,2-tetrachloroethane fit both
criteria—the highest influent concentration and the  most  difficult to strip.
In actual operation, the other compounds are undetectable  in the effluent.
     The high air/water ratio results  in lower  packing  heights  and  reasonably
sized multiple stripping towers; however, substantial energy  is required to
move 29,000 ft3/min through each tower.  These  high gas volumes and the
extremely dilute VOC concentration in  the gas  stream  make  selection of any
VOC option difficult.  The system, however,  has worked  well.  Table 5
summarizes the key parameters, and Figure 2  is  a process flow diagram.  Given
these tradeoffs, the design selected today might differ from  that  selected in
1982.
                                      15

-------
        TABLE 5.  SUMMARY OF SELECTED DESIGN DATA FOR TACOMA WELL 12A



Number of stripper columns:   5
Water volume treated:    3500 gal/min (700 gal/min per column)
Air volume:    145,000 cfm (29,500 cfm per column)

Tower height - 32 ft
Tower diameter - 12 ft
Packing height - 23 ft
Packing type - 1 in. plastic saddles

Air/water ratio - 310 (volume basis)
Air mass velocity - 1.56 kg/mz-s
Water mass velocity - 4.16 kg/m2-s


CONTAMINANT LEVELS

        Chemical                     Design, ppb               Actual, ppb

1,1,2,2-Tetrachloroethane                300                       40.9

trans-1,2-Dichloroethylene               100                       14.3

Trichloroethylene                        130                       44.6

Tetrachloroethylene                        5                        0.9

Designed for 89 percent removal of 1,1,2,2-Tetrachloroethane

No emission controls

Date of initial operation:  July 1983
                                      16

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
     29,000ft /min
      Air
Fan
                        4 k To Atmosphere
t
                                      700  gal/min
                                         -Q-
                                         Pump
                                                   Well
                                       To  Water Treatment
                                          and  Distribution
                                Pump
      Figure 2.  Process flow diagram for Tacoma Well 12A.
                           17

-------
     Further remedial actions involving  the  suspected  source  of contamination
              '*
are being planned for this site.   These  include  in  situ  vapor extraction of
the contaminated soil and the pumping and  liquid-phase carbon adsorption
treatment of the highly contaminated aquifer layer  directly under and
surrounding the source of contamination.   These  actions  should reduce the
availability of contaminants and  substantially reduce  the  extent of the
contaminant plume, which would reduce the  long-term remediation time required
to achieve cleanup.

3.2  ROCKAWAY TOWNSHIP SITE
     The Rockaway Township Site is a well  field  that supplies drinking water
to this township in northern New Jersey.   Maximum pumping  capacity is 1900
gal/min and the nominal flow from three  wells is 1400  gal/min.  In late 1979
[prior to the Comprehensive Environmental  Response  Compensation and Liability
Act (CERCLA)], trichloroethylene was detected in two of  the three wells at
levels ranging from 50 to 220 ppb.  Initially, these two wells, which were
nearest to the suspected source of the contamination,  were removed from
potable water service.  One of the two contaminated wells  was operated as a
"blocking" well to protect the remaining uncontaminated  well, and the
contaminated water was pumped directly into  a small stream.  Alternatives
were investigated to replace or remediate  the lost  production from these
wells while the blocking well was being  operated.
     In October 1980, the remaining well  was found  to be contaminated with
two compounds:  diisopropyl ether and methyl-tert-butyl  ether from a
different suspected source of contamination.  The alternatives available to
the township were to obtain water from surrounding  municipalities, to develop
a new well field, or to remediate the ground water.  In  the interim, the site
was placed on the NPL pending selection and  implementation of remedial
alternatives.
     Initially, a liquid-phase granulated activated carbon (GAC) adsorption
system with an expected carbon life of 6 to  8 months before regeneration was
necessary was selected to remove VOCs from the ground water.   Actual
operation, however,  showed that the operating time before breakthrough of the
two ether compounds was between 4 and 6 weeks, which presented an
                                      18

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
unsatisfactory operating condition and cost  (approximately  $32,000 per carbon
change-out).
     Pilot-scale tests with well  water were  begun  to  investigate the
feasibility and design requirements for an air stripper.  The two ether
compounds are more difficult to strip than trichloroethylene because their
Henry's Law constants are nearly an order of magnitude  lower.  The two ether
compounds were producing odor and taste problems  in the water and were
determined to have an extremely low taste threshold.   It  was also determined
that diisopropyl ether was the major compound associated  with taste and odor
problems.  Consequently, any design necessary to  remove the ether compounds
was found to remove trichloroethylene to satisfactory levels during the
pilot-scale testing.
     Scale-up of the pilot test air stripper resulted in  the design
summarized in Table 6.  The design basis selected  was 99.9  percent removal of
diisopropyl ether, based on an influent concentration of  4000 ppb.  This
influent level was based on hydrogeological  studies  and the estimated spill
quantity.  The target value of 4 ppb for diisopropyl  ether  was considered low
enough to avoid taste and odor problems. The air stripper  was placed in
service in February 1982.  The liquid-phase  GAC system has  been held in
reserve should maintenance be required on the stripper or should final
polishing of the treated water be required.
     In actual operation, the levels of diisopropyl  ether and methyl-tert-
butyl ether only increased to levels of 50  to 60  ppb.  The  trichloroethylene
levels were initially in the range of 200 to 300  ppb, but they fluctuated
greatly.  The diisopropyl ether and methyl-tert-butyl ether levels remained
relatively constant from the last half of 1981 through October 1982, at which
time they gradually decreased.  Diisopropyl  ether has remained undetectable
since February 1982, and methyl-tert-butyl  ether levels have very slowly
declined over the years.  The concentration  of trichloroethylene in the water
has also decreased to relatively low levels  over the years, but samples taken
every 2 weeks still show substantial variation in the influent concentration.
Benzene and toluene, which were detected in  monitoring wells, have not yet
been detected in the stripper influent.  Several  other compounds, however,
have been discovered since the system has been in operation, including
1,1-dichloroethylene, cis-l,2-dichloroethylene, chloroform,
                                      19

-------
       TABLE 6.  SUMMARY OF SELECTED DESIGN DATA FOR ROCKAWAY TOWNSHIP
Number of stripper columns:   1
Water volume treated:    1400 gal/min
Air volume:    37,500 cfm

Tower height - 35 ft
Tower diameter - 9 ft
Packing height - 25 ft
Packing type - 3 in. Tellerettes

Air/water ratio - 200 (volume basis)
Air mass velocity - 3.60 kg/m2-s
Water mass velocity - 14.91 kg/m2-s
CONTAMINANT LEVELS

        Chemical

Trichloroethylene
Diisopropyl ether

Methyl-tert-butyl ether

Tetrachloroethylene
Design,  ppb

    300

  4,000
   Actual,  ppb

       28.3

No longer detected

        3.2

        0.9
Contaminant levels for other compounds not specified in the design.
efficiency specified:  99.9 percent diisopropyl  ether

No emission controls

Date of initial operation:  July 1982
                               Removal
                                      20

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
1,1,1-trichloroethane, and (occasionally)  dichloromethane  and
1,1-dichloroethane.  Some doubt exists as  to  whether  the last  two compounds
are actually present or are a result of laboratory  problems,  as they are
observed intermittently and at very low concentrations.  The  possible source
of these additional compounds had not been determined;  however, the two
dichloroethylene compounds could be decomposition products of the original
trichloroethylene spills as the trichloroethylene slowly breaks down in the
aquifer.  The other compounds may indicate that  some  other spill  has occurred
recently or that contamination from a more distant  source  has  finally
traveled through the aquifer.  In any case, the  air stripper  has sufficient
design capacity to remove these other compounds.
     The design liquid volume of 1400 gal/min (peak at  1900)  and a nominal
air/water ratio of 200:1 place this system in the upper range for both design
parameters.  The design concentrations were relatively  high,  but current
operation and influent conditions place this  system on  the lower end of the
contaminant concentration scale.  This system is not  equipped with any VOC
controls for air emissions.  Table 6 summarizes  the design and operating
parameters.  Figure 3 is a process flow diagram.
     A point of interest not modeled by ASPEN, but  one  that has an impact on
water quality, is the effect of the air stripper operation on other water
quality parameters.  The influent water is slightly acidic, as it contains a
small quantity of C02 dissolved as carbonic acid  in the water.  The water
also contains a small quantity of dissolved iron  (FeO).  As the water passes
through the air stripper and comes in intimate contact  with air,  the CC^ is
stripped from the water, which alters the  pH.  This also oxidizes the iron to
Fe90,, which is much less soluble in a neutral or alkaline condition than in
  C. 0
an acidic environment.  Whereas such a change in equilibrium  has not affected
operation of the air stripper, it has affected the  final water quality and
the potential for "scale" within the distribution system and  at the point of
use.  The township currently plans to use  a sequestering agent to combine
with the iron in the water to prevent further problems  with iron.
                                      21

-------
                        t k To Atmosphere
From
Well  field
1400  gal/min
 37500
 O / UUU y-v

ft  /min/""\
Fan
                                                  Liquid
                                                  Phase
                                                   GAC
             Figure 3.   Process flow diagram for Rockaway  Township.
                                    22

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
3.3  BREWSTER WELL FIELD NO. 1
     The Brewster Well Field Area No.  1  is composed  of nine wells,  and when
combined with the output of Well  Field Area No.  2, supplies drinking water to
the Village of Brewster in Putnam County,  New York.   Sampling conducted in
August 1978 indicated that tetrachloroethylene and trichloroethylene were
detected at levels of 166 and 5 ppb, respectively, or an overall value of 171
ppb.  Testing the following month showed levels at  174 ppb and 220 ppb for
total VOCs, which prompted shutting down the well  field.  The need to resume
production to maintain an adequate water supply resulted in pilot-scale
testing and full-scale operation of an air stripper  to treat the contaminated
ground water.  This air stripper was placed in operation in October 1984.
Subsequently, contamination of Brewster Well Field  Area No. 2 and a deep well
(designated DW-2) were observed and are the subject  of a recent Remedial
Investigation/Feasibility Study (RI/FS) and Record  of Decision  (ROD)
indicating another air stripper with VOC emission controls as the preferred
treatment technology.
     The compounds of primary interest at Brewster Well Field No. 1 were
tetrachloroethylene, trichloroethylene, cis- and trans-l,2-dichloroethylene,
and  vinyl chloride.  Occasionally a variety of trihalomethanes  were found in
the  well water,  but they were believed to be caused by chlorination of the
well water and not directly related to a source of contamination.   In
addition to a full-scale stripper,  several pilot-scale strippers were also
tested  for removal efficiency with  various packings and air/water ratios.
The  test data showed  that at  the influent concentrations measured,  a minimum
air/water ratio  of 20:1  (volume basis) was needed to meet Maximum Contaminant
Levels  (MCLs) specified  by  the Safe Water Drinking Act.  The  full-scale
column  typically operates at  an air/water ratio of 50:1.
     The water volume handled by this air stripper is 300 gal/min,  which  is
considered to be in a range typical of most air strippers.  The air/water
ratio of 50:1 is also considered typical for this application,  as are the
concentrations of the pollutants in the water.  Table 7 summarizes  the
stripper design  parameters, and  Figure 4  is a process flow diagram  for
Brewster Well Field.
     The estimated cost  of  construction and the assumptions used in the
development of operating and  maintenance costs for this system  appeared  to  be
                                       23

-------
      TABLE 7.  SUMMARY OF SELECTED DESIGN DATA FOR BREWSTER WELL FIELD
Number of stripper columns:   1
Water volume treated:    300 gal/min
Air volume:    2000 cfm

Tower height - 27 ft
Tower diameter - 4.75 ft
Packing height - 17.75 ft
Packing type - 1 in. Saddles

Air/water ratio - 50
Air mass velocity - 0.28 kg/m2-s
Water mass velocity - 11.41 kg/m2-s


CONTAMINANT LEVELS

        Chemical                      Design,  ppb             Actual,  ppb

Tetrachloroethylene                       215                      200

Trichloroethylene                          77                       30

1,2-Dichloroethylene                       68                       38

Vinyl chloride                              2                       ND


Designed for removal down to 5 ppb for all  compounds  except vinyl  chloride
(nondetectable levels).

Date of initial operation:   October 1984
                                      24

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                            'k To Atmosphere
          300 gal/min
From
Well Field
            Pump
2000
ft3/min
                                                 Fan
                                         To Water Distribution
      Figure 4.  Process  flow diagram for Brewster Well  Field.
                              25

-------
well documented.  Costs of major components  (stripper  column,  piping,
instrumentation, etc.) were outlined to provide  total  capital  costs instead
of a single reported capital cost value without  documentation.

3.4  VERONA WELL FIELD
     The Verona Well Field Site is located near  Battle Creek,  Michigan,  in
Calhoun County.  This well field is composed of  approximately  30 wells that
supply potable drinking water to the Battle  Creek,  Michigan,  area.   In 1981,
detectable levels of VOCs were found during  routine testing.   The
contaminated wells were identified and taken out of service,  and the pumping
load was shifted to other wells and away from the contaminant  plume in the
aquifer.  By 1984, however, it was apparent  that the summer maximum-day
demand could not be met and other alternatives needed  to  be considered.
These included development of new wells and  the  treatment of the contaminated
ground water.
     The five contaminated wells were to be  placed  back into operation to act
as blocking wells to protect the rest of the well field from the spread of
contamination.  An air stripper was recommended  as  the most cost-effective
method of treatment.  Under interim removal  action  authority,  however,  v
operation of the five wells began before the air stripper was  installed.
Temporary treatment with a liquid-phase GAC  adsorption unit was applied.
This provided a unique opportunity to compare the installation  and  operating
costs of both liquid-phase GAC and air stripping.
     Michigan regulations require the use of best available control
technology (BACT) for control of air emissions where any  new source of
carcinogens or suspected carcinogens are involved.   In this case, vapor-phase
activated carbon adsorption was the selected control technology.  The cost of
this system was included in the initial feasibility and cost-effectiveness
study.
     The compounds of concern found in monitoring wells were
1,1-dichloroethane,  1,2-dichloroethane, 1,1,1-trichloroethane,
1,1-dichloroethylene, cis-l,2-dichloroethylene,  trichloroethylene,  and
tetrachloroethylene.  It should be noted that concentrations from monitoring
well data prior to startup were 15 times higher  than those found at the
                                      26

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
influent of the liquid-phase GAC adsorption  system  or  the  air stripper once
operation began.
     The air stripper was designed to handle a nominal  water flow of 1950
gal/min (maximum of 2500 gal/min).  This places the design capacity of the
column on the high-end of stripper designs for water volume treated.  This
design, however, also incorporates a minimal  air flow  design to reduce the
gas volume handled by the vapor-phase GAC.   Thus, the  operating air/water
ratio is 20:1, the lowest within the group of the seven selected sites.  This
design results in a very tall stripping column with a  packing height of 40
feet.  Table 8 summarizes the design data.   Figure  5 is a  process flow
diagram for the site.
     The air passes through the air stripper, an induced-draft fan, and an
indirect-fired natural gas heater and into two vapor-phase carbon adsorbers
to adsorb VOCs from the gas stream.  The natural-gas-fired heater is used to
lower the relative humidity of the gas stream to less  than 50 percent by
providing a 30° to 35"F temperature increase to improve vapor-phase GAC
performance.  The vapor-phase GAC system does not have any provision for
onsite regeneration.  Based on the design parameters,  change-out of carbon
(to be regenerated offsite) was to be required approximately once a year.
     The operation of the liquid-phase GAC adsorption  system continued for 17
weeks.  It was terminated and removed once the air  stripper system became
operational.  The performance of the liquid-phase system indicated that
liquid-phase carbon would have to be replaced at least every 6 months.  When
the system was shut down, it was discovered  that breakthrough had occurred
for the 1,2-dichloroethylene, 1,1,1-trichloroethane, and 1,1-dichloroethane.
     Since its operation began in 1984, occasional  tests were made of the
vapor-phase GAC system, but the daily levels were not  routinely monitored
during the tests.  Initial test results for  the air stripper and vapor-phase
GAC indicated high removal efficiencies of VOC from the water and air.  When
VOC concentrations at the outlet of the vapor-phase GAC system increased, the
carbon was replaced.  The concentration of VOCs in  the water influent
remained relatively constant.  The potential  source of the contamination was
identified and in 1987 and early 1988, a removal action was implemented to
reduce the areas of highest contamination.   Part of this removal action
included the use of the stripper's "extra VOC stripping capacity," which was
                                      27

-------
       TABLE 8.  SUMMARY OF SELECTED DESIGN DATA FOR VERONA WELL FIELD
Number of stripper columns:   1
Water volume treated:    1950 gal/min
Air volume:    5000 cfm

Tower height - 65 ft
Tower diameter - 10 ft
Packing height - 40 ft
Packing type - 3.5 in. polypropylene  pall  rings

Air/water ratio - 20 (volume basis)
Air mass velocity - 0.39 kg/m2-s
Water mass velocity - 16.85 kg/m2-s


CONTAMINANT LEVELS

        Chemical                      Design,  ppb              Actual, ppb

1,1-Dichloroethane                         34                        5.7

1,2-Dichloroethylene                        8                       ND

1,1,1-Trichloroethane                     150                       12.1

cis-l,2-Dichloroethylene                  229                       11.1

1,1-Dichloroethylene                       11                       ND

Trichloroethylene                          62                        1.1

Tetrachloroethylene                        94                        9.2


Designed to remove contaminants to less than 5 ppb.

Vapor-phase GAC adsorption system for VOC  emissions  control
Number of beds:  2
Bed diameter:  10 ft
Bed depth:  4 ft
Carbon weight:  9,500 Ib
Carbon type:  Calgon BPL, 4x6 mesh
Air preheater included (indirect, natural  gas-fired)
Nonregenerable system, change once/year
Removal efficiency:  90 percent

Date of initial operation:  September 1984
                                      28

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
            Water
          1950
          gal/min
From
Well
Field
           Vapor
           Phase
           GAG
                                                  Atmosphere

                                                   t_     t
          Heater
         (Natura
           Gas)
                                    Fan
 Air
5000 ft3/min
                                        Clean  Water
          Figure 5.  Process flow diagram for Verona Well Field.
                                29

-------
From Highly
Concentrated
Zone
From
Well
Field
       »|Tank|—I
         Vapor
         Phase
          GAC
                                Lgl
         Heater
                                  Fan
Air
                                                Atmosphere
                                                 L     t
                                      Clean Water
   Figure 6.  Process flow diagram for Verona Well Field during
            removal action of highly contaminated zone.
                              30

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
not being used because influent VOC levels  were  substantially below design
levels.  Ground water was pumped from the aquifer  at  and  near the source of
the contamination and pretreated by passing it through temporary liquid-phase
GAC adsorbers to remove the bulk of VOC contaminants.   This arrangement is
shown in Figure 6.  Additional  compounds not normally found in the air
stripper influent include dichloromethane and vinyl chloride.  These data
were included in the ASPEN comparison data  set because they represented
elevated levels of VOC being sent to the stripper. All other parameters
(water and air flow) remained constant.  The data  provided also included test
data from the vapor-phase GAC adsorption system.  This removal action has
been completed and the system has been returned  to its original design.  No
other major problems have been noted for this system.
     Site installation costs, for the air stripper system and estimated
operation and maintenance costs were provided.   A  complete breakdown of all
the capital costs was not available.  Costs for  operation and maintenance,
however, were provided for energy and annual vapor-phase GAC replacement.

3.5  WESTERN PROCESSING SITE
     The Western Processing Site is located in  Kent,  Washington (near
Seattle) in King County.  The 13-acre site  was  originally a hazardous waste
treatment, storage, and disposal facility (TSDF) under the Resource
Conservation and Recovery Act (RCRA) regulations.   Contamination of surface
and ground water was discovered, and a RCRA Corrective Action Order was
issued for the site in 1980.  The site continued to operate until it filed
for bankruptcy because it was unable to comply with the corrective action
order.  The site was placed on the NPL in 1982.
     A series of removal actions included the drums dumped, stacked, or
buried at the site and the contaminated surface  soils.  The site is located
next to a tributary of the Green River, and the  water table depth ranges
between 3 and 12 feet with an average depth of 6 feet to the top of the water
table.  This shallow aquifer extends downward some 75 feet.  Extensive
sampling of this aquifer revealed 87 priority pollutants and 12 other
hazardous pollutants.  Forty-nine of the compounds are either known
or suspected carcinogens.  Twenty-nine samples wewre present in
concentrations above 1000 ppb.
                                      31

-------
         TABLE 9.  CONCENTRATION RANGE FOR SELECTED VOC CONTAMINANTS
                       AT THE WESTERN PROCESSING SITE
    	Compound	Range,  fig/liter
     Benzene                                      77 to 2,200
     1,2-Dichloroethane                           16,000
     1,1,1-Trichloroethane                        100 to  340,000
     1,1-Dichloroethane                           320 to  33,000
     Chloroform                                   130 to  27,000
     1,1-Dichloroethylene                         87a
     trans-1,2-Di chloroethylene                   390,000a
     Ethyl benzene                                 32a
     Dichloromethane                              1,200 to 720,000
     Fluorotrichloromethane                       920a
     Tetrachloroethylene                          37 to 50
     Toluene                                      110 to  22,000
     Trichloroethylene                            830 to  210,000
     Vinyl chloride                               360a
aNoted in only one sample.
     The site is extremely heterogeneous and varies  widely in both types of
contaminants and their concentrations.   Table 9 presents  a partial listing of
the range of concentrations found in seven monitoring wells.   Of the 14
compounds listed in Table 9 chloroform,  dichloromethane,  1,1,1-trichloro-
ethane, and trichloroethylene appeared  to be the most widely  distributed.  In
addition to these VOCs, many semi volatile and nonvolatile organics, were
found as were a number of inorganic species.
     The extreme variability in the concentrations made it rather difficult
to select a design that would meet Safe Water Drinking Act Maximum
                                      32

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Contaminant Levels (MCL) for these VOCs.   The  entire  treatment system also
needed to address inorganic materials as  well  as  the  semi-  and nonvolatile
organic compounds.  Air stripping was the selected  remedial  alternative for
VOCs and is the first of many steps in the water-treatment  process.
     Two stripper towers are used at this site.   One  tower,  the larger of the
two, is designated as the "extraction" tower.  This tower treats ground water
from a larger area of the site.   The smaller of the two towers, designated as
the "trans-tower" for removal of trans-l,2-dichloroethylene,  receives its
water from three contaminated wells located across  the  tributary from the
main site.  Both air strippers vent into  an induced-draft fan (one for each
tower), and the gases are then combined and heated  to reduce relative
humidity.
     Emission controls were required on this system and a vapor-phase GAC
adsorption system is also used.   The system is a  two-bed CADRE system.  The
CADRE system differs from a conventional  steam-regenerated  GAC system in that
it uses an incinerator to generate heat for desorption  and  regeneration of
the carbon beds.  This same incinerator also combusts the VOCs as they are
desorbed from the GAC.  The principal advantage to  this system is that the
incinerator treats a much smaller gas volume with much  higher VOC
concentrations than normally exits the strippers, which lowers the energy
costs.  It also eliminates the handling of the liquid wastes that would be
generated by conventional steam regeneration.  Figure 7 is  a schematic of
this system.
     The two air strippers are designed around "primary" and "secondary" air
and water flow rate values.  The primary  values represent the initial lower
volume operating values for the site, when the contaminant  concentrations
would be expected to be the highest.  The secondary values  represent the
higher air and water flow rates that occurred  when  remediation was well under
way and contaminant concentrations were somewhat  reduced.
     As mentioned earlier, the extreme variability  in concentration and the
wide range in contaminants present some difficulty  in the design of an air
stripper.  The average design concentration of the  various  compounds was
292,980 ppb (15 compounds), and the maximum expected  level  was five times
greater than the average.  Design removal  efficiency  for the extraction air
stripper was 96.59 percent.  The extreme  variability  also introduces some
difficulty in the design of a vapor-phase GAC  absorber  system.  The
                                      33

-------
CO
-ti
    100
    gal/min
    From
    Extraction
    Wells
        Pump
                                   ff
                                    Fan
Extraction
Tower
         45
         gal/min
 Air
From
"Trans"
Wells _
2150
ft3/min
    Pump
                                                        Fan
                    Trans"
                    Tower
Air
500
ft3/min




Scrubber

t.
i
Incinerator -
used during
Regeneration
only
f
Vapor
> Phase
GAG
t...

_J
Atm.
t i

                   Figure 7.  Process flow diagram of  the CADRE system at Western Processing.

-------
I

I

I

I

I

I

I

I

I

I
   TABLE 10.  SUMMARY OF SELECTED DESIGN DATA FOR WESTERN PROCESSING SITE


Number of stripper columns:   2a

EXTRACTION TOWER
Water volume treated:    100 gal/min (initial),  200 gal/min (maximum)
Air volume:    2150 cfm (initial), 2670 cfm (maximum)

Tower height - 26 ft
Tower diameter -40 ft
Packing height - 40.5 ft
Packing type - 2 in. Jaeger Tripack

Air/water ratio - 160 (initial), 100 (maximum) (volume basis)
Air mass velocity - 1.04 kg/mz-s (initial), 1.30 kg/m2-s (maximum)
Water mass velocity -  5.36 kg/m2-s (initial), 10.72 kg/mz-s (maximum)


CONTAMINANT LEVELS

     Compound                                     Range, ppb

     Benzene                                      77 to 2,200
     1,2-Dichloroethane                           16,000
     1,1,1-Trichloroethane                        100 to 340,000
     1,1-Dichloroethane                           320 to 33,000
     Chloroform                                   130 to 27,000
     1,1-Dichloroethylene                         87
     trans-1,2-Di chloroethyl ene                   390,000
     Ethyl benzene                                 32a
     Dichloromethane                              1,200 to 720,000
     Fluorotrichloromethane                       920
     Tetrachloroethylene                          37 to 50
     Toluene                                      110 to 22,000
     Trichloroethylene                            830 to 210,000
     Vinyl chloride                               360


TRANS TOWER

Tower height - 28 ft
Tower diameter -2 ft
Packing height - 22.5 ft
Packing type - 2 in. Jaeger Tripack
Water volume treated:  45 gal/min (initial),  60  gal/min (maximum)
Air volume:  500 ft3/min

Air/water ratio - 83.1 (initial), 62.3 (maximum)  (volume basis)
Air mass velocity - 0.97 kg/m2-s
Water mass velocity -  9.65 kg/m2-s (initial), 12.87 kg/m2-s (maximum)

(continued)
                                      35

-------
TABLE 10 (continued)
CONTAMINANT LEVELS

  -   Chemical

trans-1,2-Dichloroethylene
Trichloroethylene
Vinyl chloride
Dichloromethane
Design, DDD

   4,000
      73
     270
     140
      Actual," ppb

           NO
            9
          140
          100
Design Removal Efficiencies
     Chemical

trans-1,2-Di chloroethylene
Tri chloroethylene
Vinyl chloride
Dichloromethane
   Removal
efficiency. %

    98.12
    58.90
    88.89
    28.57
     Effluent
concentration, ppb

        75
        30
        30
       100
VOC Emission Controls
     Vapor-phase GAC adsorption with integral incinerator for carbon
regeneration and VOC destruction (Calgon CADRE).

Number of beds:  2
Bed diameter:  6 ft
Bed depth:  2 ft
Carbon weight:  7,200 Ib
Carbon type:  BPL
Air preheater included (indirect, natural gas-fired)
Removal efficiency:  95 percent

Date of initial operation:  September 1988


aTwo separate columns with different flow and pollutant characteristics.   Air
 is combined with a common GAC system.  Designed as extraction and
 trans-towers at two designated flow rates.
                                      36

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
"trans-tower" air stripper was expected to treat water contaminated primarily
with trans-l,2-dichloroethylene and containing smaller concentrations of
trichloroethylene, vinyl chloride, and dichloromethane.  Table 10 summarizes
the design parameters for this system.
     During actual operations, the contaminant concentration has been
approximately 10 percent of the design levels for the 15 design compounds for
the extraction tower.  Weekly sampling data on the trans-tower indicate that
trans-l,2-dichloroethylene has virtually disappeared.  Only two compounds
(vinyl chloride and tetrachloroethylene) are at or higher than their design
levels.  Concentrations of vinyl chloride have consistently doubled their
design concentration.  This is not meant to imply that many compounds are
still found at substantial levels (dichloromethane and trichloroethylene are
found at concentrations greater than 8000 ppb each);  instead, it merely
indicates that the design basis does not match the actual  operation at this
time.  A bentonite slurry wall around the site has apparently reduced, but
has not stopped contaminant migration in the ground water.   Water from the
treatment facility can be discharged to the local  publicly owned treatment
works (POTW) or returned to the site for soil flushing to assist in VOC
removal.
     At the time of this report, the largest loading  of VOCs to the
vapor-phase GAC system comes from the extraction tower; the trans-tower
provides only minor amounts to the overall  loading (mostly vinyl  chloride).
The primary water flow rate of 100 gal/min puts the extraction tower air
stripper at the low end of the scale for water volume handled.  The primary
air/water ratio of 160:1 places the design in the  upper range for air/water
ratios.  The level of contamination encountered, however,  is the highest of
all the sites included in this comparison and the  high air/water ratio
appears proper within this context.
     The vapor-phase GAC system has posed some problems.   During a
regeneration cycle in January 1989 the carbon in one  of the beds caught fire.
The precise cause of this fire was not provided.  The bed  was repaired and
placed back in service.   Modifications made to the system  at that time should
prevent any recurrence of this problem.   Although  emission  testing of the
CADRE system may still be required,  the performance of the  vapor-phase GAC is
monitored frequently by testing inlet and outlet concentrations with an
                                      37

-------
ultraviolet phptoionization detector.   The photoionization detector is
calibrated before each inlet/outlet test by using 1,2-dichloroethylene.
     Cost data were requested for this site because it represents a
relatively recent installation.  Cost  data, however,  were not available
either in the files of the EPA Regional  Offices or at the site itself.  Thus,
no comparison of cost data is possible.

3.6  HICKSVILLE MEK SPILL SITE
     The Hicksville MEK Spill Site represents  one of  the simpler of the seven
selected sites because it involves only one contaminant,  methyl  ethyl  ketone.
The air stripper at this site, however,  was a  different technology from that
used at the other sites.  It is High-Temperature Air  Stripper (HTAS),  in
which the contaminated water is brought up to  an elevated temperature before
being contacted with air as in a normal  stripper.   The advantage of this type
of system is that compounds such as some alcohols and ketones,  which have
relatively high vapor pressures and low Henry's Law constants,  can be
stripped from the water with a much smaller air stripper.   In fact,  tests
conducted at this site indicated that  a conventional  column with five times
the air volume of the HTAS or with substantially more packing would  be  *+
required to achieve the same degree of removal  as the HTAS.
     The Hicksville MEK Spill Site is  located  in Hicksville,  New York, on
Long Island in Nassau County.  An overturned tank truck resulted in  a spill
of approximately 4800 gallons of MEK.   The cleanup action  was initiated as an
emergency removal action, limited duration ($1  million and 6 months).   The
total operating time for this air stripper was  125 days before  the removal
action was completed.  The water was treated in a batch-type operation
(untreated water was stored in one 6000-gal  tank,  and the  treated water was
stored in two 50000-gal  tanks until  tests verified that the required degree
of treatment was achieved); however, the stripper ran continuously during
each batch.
     The HTAS concept uses a small  package boiler to  provide the steam
required to heat the water to an elevated temperature.   The incoming water is
preheated by passing it through an indirect-steam heat exchanger to  bring it
up to operating temperature.   The typical  water temperature was  then 180° to
195°F when it entered the stripper.  The air entering the  stripper was not
                                      38

-------
I
I
I
I
1
I
I
I
I
I
I
I
I
I
I
I
I
I
I
preheated.  The contact between  air  and water  and the  heat  losses  that
occurred through the stripper walls  decreased  the water  temperature as it
passed through the stripper.   The elevated temperature also results in
greater water vapor losses than  encountered  in traditional  strippers.
     One option available (but not used in this  case)  is the ability to use
the package boiler as a VOC control  device by  venting  the air exiting the
stripper through the boiler via  the  primary  air  supply and  using the boiler
as a direct-flame incinerator.  The  addition of  a high-moisture air stream
such as this would decrease the  boiler efficiency somewhat, would  require the
use of more fuel to provide the  necessary heat input  (the heat available from
methyl ethyl ketone combustion would be negligible),  and could reduce the
peak flame temperature.  In this case, however,  the air  was vented from the
top of the air stripper to the atmosphere without controls.  Figure 8 is a
process flow diagram for the HTAS system.
     The air stripper performed  well during  this removal action.  In general,
target removal levels were achieved  without  having to  rerun batches of water.
The typical operation entailed running a batch through the  stripper twice.
The first pass was designed to lower the VOC concentration  from an estimated
15,000 ppb influent to 250 ppb effluent.  The  water was  reheated for the
second pass and the VOC concentration was lowered to  less than 50  ppb.  The
nominal air/water ratio was approximately 200:1  for the  two passes combined.
The ASPEN simulation used data from  only one of  the two  passes.  The effect
of the two passes was essentially the same as  passing  the water through the
two separate air stripper columns.  Table 11 summarizes  the key parameters.
     The water treatment rate of 100 gal/min places this air stripper on the
low end of the range for air strippers, whereas  the actual  air/water ratio of
120 places it in the mid-range.   The MEK concentration is high (15,000 ppb)
and represents a highly contaminated water stream.  The  HTAS is a unique
system for removing a compound that  is usually most difficult to air-strip.
Extreme fouling difficulties, however, are an  example  of a  problem not easily
modeled nor even considered by ASPEN and reflect the  "other" considerations
that are important to an air stripper design.
     Scaling in the preheater heat exchanger due to iron oxide was a serious
problem.  The water contained significant quantities  of  iron dissolved as
FeO.  Passage through the air stripper at an elevated  temperature  caused the
                                      39

-------
                                                Atmosphere
From Well  or
50,000  Gallon
Tanks
                                                     Air
                                                         -2
1600
ft3/min
                                                  Boiler
                                                  Combustion
                                                  Emissions
           Fan
                          Pump
                                                                       (2) 50,000
                                                                       Gallon Tanks
                 Figure 8.  HTAS process flow diagram for Hicksville MEK Site.

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
  TABLE 11.  SUMMARY OF SELECTED DESIGN DATA FOR HICKSVILLE MEK SPILL SITE



Number of stripper columns:   1
Water volume treated:    100 gal/min
Air volume:    1600 cfm

Tower height - 24 ft
Tower diameter - 3.6 ft
Packing height - 15 ft
Packing type - 2 in. Jaeger Tripack

Air/water ratio - 120 (volume basis)
Air mass velocity - 0.97 kg/m2-s
Water mass velocity - 6.71 kg/mz-s


CONTAMINANT LEVELS

        Chemical                      Design, ppb             Actual. ppb

          MEK                           15,000                   15,000


Designed for 98.33 percent removal of methyl ethyl ketone.

No emissions control

Date of initial operation:  June3 1984

Note:  High-Temperature Air Stripper (HTAS)


aOperated for 3 months for removal action.
 iron to oxidize to Fe203 and to precipitate.  This problem was so severe

 that, without some form of treatment, the heat exchanger would plug within  2

 days of operation.  The problem was solved by adding hydrochloric acid  to
 lower the pH to 4.0, running the water through the HTAS, and  (after achieving

 the desired VOC removal level) raising the pH by the addition of a sodium

 hydroxide solution.  This substantially reduced the need to clean the heat

 exchanger.  The Fe203, however, precipitated upon causticizing and caused

 problems with the reinjection of the water into wells.
                                      41

-------
3.7  GILSON ROAD (SYLVESTER'S)  SITE
     The seventh site included  in  the  ASPEN  performance  evaluation was the
Gil son Road Site (also known as Sylvester's)  in  Nashua,  New Hampshire, in
southern Hillsborough County.   This  site  uses  an HTAS  to remediate
contaminated ground water.
     The 6-acre site was originally  a  sand quarry (or  sand  borrow pit) and
then later used to dispose  of both household  and hazardous  waste.  This
unapproved disposal operation began  in the late  1960's.   The materials
included drums of waste as  well as hazardous  liquid  chemicals and sludges
that were allowed to percolate  into  the ground.   The contamination spread
through the aquifer.  The total quantity  of  materials  present could not be
determined, but EPA determined  from  available  information that more than
800,000 gallons were disposed of at  the site  in  1979.  The  site was placed on
the NPL in October 1981.
     Contamination in the ground water included  VOCs,  heavy metals, and semi-
and nonvolatile organics.  A single  remediation  technology  could not handle
all of these contaminants,  so a treatment train  was  set  up.  The first step
of the treatment removes the inorganics (metals) to  prevent fouling of other
equipment.  The next step is the HTAS  for removal  of VOCs.   The last step
involves biological treatment.   Extremely high concentrations of VOCs were
found in the ground water samples  (Table  12).  Several of these compounds,
most notably the alcohols,  ketones,  and tetrahydrofuran,  are difficult to
strip at normal temperatures.   At  operating  temperatures of 175"F, however,
the Henry's Law constants are sufficiently high  to make  air stripping
practical with reasonable column heights  and  air/water ratios.
     To minimize the migration  of contaminated ground  water, a bentonite
slurry wall was installed around the site.   As with  many sites, after the
initial operation, ground water contamination  levels drop to a fraction of
the design value.  With the exception  of  1,1,1-trichloroethane and
trichloroethylene, all actual operating contaminant  levels  are below design
values.  The air stripper design was based on  seven  compounds, and 75 percent
removal of isopropanol is the controlling compound in  the design.  Table 13
summarizes the design parameters for the  HTAS.
                                      42

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
    TABLE 12.  CONCENTRATIONS  OF  ORGANIC CONTAMINANTS  FOUND IN
                 THE GILSON  ROAD  SITE GROUND WATER
     Compound
Concentration, ppb
Acetone
Benzene
Isobutanol
Chlorobenzene
Chloroform
1,1-Dichloroethane
1,2-Dichloroethane
Diethyl ether
Dimethyl sulfide
Ethyl benzene
Ethyl chloride
Ethylene chloride
Isopropanol
Methyl acetate
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Dichloromethane
Tetrachloroethylene
Tetrachloroform
1,1,1-Trichloroethane
1,1,2-Trichloroethane
Trichloroethyl ene
Toluene
Vinyl chloride
Xylenes
     310,000
       3,400
       3,560
       1,100
      31,000
          15
      18,000
      20,000
       3,500
       1,200
         320
      73,000
      26,000
       2,400
      80,000
      21,000
       3,500
     122,500
         570
   1,500,000
       2,000
          17
      15,000
      29,000
         950
      10,000
                                 43

-------
     TABLE 13.  SUMMARY OF SELECTED DESIGN DATA FOR THE GILSON. ROAD SITE



Number of stripper columns:   1
Water volume treated:    300 gal/min
Air volume:    2080 cfm

Tower height - 33 ft
Tower diameter - 4 ft
Packing height - 16 ft
Packing type - 16 KOCH Type Trays @ 1 ft intervals

Air/water ratio - 51.4 (volume basis)
Air mass velocity - 1.01 kg/m2-s
Water mass velocity - 16.08 kg/m2-s


CONTAMINANT LEVELS

        Chemical                      Design,  ppb            Actual. ppb

Isopropyl alcohol                       36,000                   532.0
Acetone                                 36,000                   472.7

Toluene                                 22,000                14,884

Dichloromethane                          8,300                 2,365

1,1,1-Trichloroethane                      430                 1,340

Trichloroethylene                          740                 1,017

Chloroform                               1,200                   469


Designed for  75 percent removal of  isopropyl alcohol.

Date of  initial operation:  June 1986

Note:  HTAS design.  Boiler used to  heat water is also used as a direct-fired
       incinerator.  VOC  removal reported to be 99.95 percent.
                                      44

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
     Both air and water are heated in  this  HTAS.  A  heat  exchanger/economizer

uses the water exiting the air stripper  to  preheat the  incoming water,  the

water is then heated to its operating  temperature by steam provided from the

boiler/fume incinerator.  After the preheated water  passes through the

stripper, it flows through the enconomizer,  and  the  air is sent to the

boiler/fume incinerator.  The boiler not only provides  the steam for the HTAS

and other operations around the site,  but also provides a method of VOC

destruction.  The overall  average VOC  destruction is reported to be 99.95

percent, with 99.99 percent removal  of tetrahydrofuran  (note this is not

included as one of the monitored pollutants).  Emissions  are exhausted from

the stack along with other combustion  emissions.  Figure  9 is a process flow

diagram for this system.
     The design air/water ratio of approximately 50:1  (volume basis) is a low

to moderate design value.   The design  water treatment  rate of 300 gal/min

places the stripper in the moderate size category.   Unlike the other air

strippers included in this study, this system uses a tray design as opposed

to "conventional" packing.  The use of the  boiler as a  fume incinerator

represents a type of control  not considered by the ASPEN  simulation.  The

relatively high contamination levels and moderate air/water ratio, however,

result in a relatively contaminated air  stream compared with that of the

other strippers in this study, and VOC controls  are  required.
                                      45

-------
O-l
From
Inorganic
Treatment
                               Atmosphere
                                Condensate
                            Water
                          Preheating
                          Economizer
                                                                          Air
                                                                        2080
                                                                        ft3/min
                                                                             Fan
                                                   Pump
                                                                            To Treatment
                        Figure 9.  HTAS process flow diagram for Gil son Road Site.

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                  SECTION  4
                ASPEN SIMULATIONS AND PERFORMANCE  COMPARISONS

     The parameters for each of the seven  sites  were  collected and input into
the ASPEN model through the user-interface software.   The design and
operating parameters for each air stripper included column dimensions,
packing height and types, water and air volumes  treated,  control device
parameters, and species and influent concentrations of contaminants.  Cost
data were also input when available; if not,  default  values were used.  The
ASPEN air stripping program was then run in  a "rating" mode to compare
ASPEN's predicted results with actual performance  data.   Equipment and
operating costs were also compared when data were  available.

4.1  GENERAL RESULTS OF THE ASPEN SITE COMPARISONS
     In general, ASPEN performance predictions compared quite favorably with
actual performance predictions for both individual compounds at a site and
with predictions of overall performance.  For approximately half of the
individual estimates of chemical removal,  the predicted performance levels
were within 1 percent of the actual performance  data.   In addition, the
general tendency for compounds outside the 1  percent  relative error band at
any given site was for all the compounds either  to be overpredicted or
underpredicted.  In other words, when the  ASPEN  simulations underpredicted
performance by more than 1 percent relative  to the actual performance data,
it generally underpredicted for all compounds.  Some  notable exceptions to
this observation are discussed in Section  4.2.  Much  of this tendency to
overpredict or underpredict performance appears  to be due to measurement or
estimation inaccuracies involved with such elements as air flow measurements
and temperatures or to nonideal conditions within  the stripper, such as
channeling (for overprediction of performance).  At six of the seven sites,
overall estimated performance was within 2 percent of the actual performance
level, and performance was underpredicted  slightly at five of these sites.
Thus, it may be concluded that the ASPEN simulation tends to slightly under
                                      47

-------
 predict  performance.  This would result in a somewhat conservative design  if
 ASPEN were  run  in  the "design mode" to assist in the minimum design needed
 for  an air  stripper.
   ,.  Although predictions were well correlated for most cases, some compounds
 and  designs presented a challenge to model using the ASPEN software.  For
 example,  some compounds known to be difficult to strip (e.g., acetone and
 methyl ethyl ketone) did not correlate well with actual performance values.
 Differences between actual and predicted values for another compound,
 dichloromethane, were also significant, with the ASPEN model overpredicting
 removal  by a substantial margin.  The cause of this large relative error
 could be  due to sampling or analytical contamination,  as dichloromethane
 (methlyene chloride) is a commonly used laboratory material.
     A substantial relative error (compared with actual performance) occurred
 in the area of surrogate compounds.  Although the ASPEN library of chemical
 properties includes nearly 400 compounds,  several very important compounds,
 most notably cis-l,2-dichloroethylene, trans-l,2-dichloroethylene,
 1,1,1-trichloroethane, and the tetrachloroethanes,  are not currently
 included.  Modeling these compounds in the simulation  required the selection
 of a chemically similar compound.   The user input program allows for the
 modification of the Henry's Law value as needed if the default value provided
 with the  compound needs to be modified.  For compounds such as
 cis-l,2-dichloroethylene or trans-l,2-dichloroethylene,  for which no Henry's
 Law or other physical  data are present in  the ASPEN library, the compound
 1,2-dichloroethane was selected to be the  surrogate and Henry's Law is
 modified  to reflect the dichloroethylene compound,  not the dichloroethane
 compound.  Although the two compounds are  similar,  the presence of
double-bonded carbon atoms in the  dichloroethylene  compounds alter the
 physical   behavior of the compounds in water.   In general,  this introduced a
 slightly greater relative error than  would be found for the remaining
compounds in the simulation at a specific  site.   Also,  the error could be in
the opposite direction of the rest of the  compounds on the list.   (These
relative errors  are discussed in Section 4.2  for the individual  sites.)   The
use of surrogates appears to introduce a larger relative  error than using the
actual  compound's properties when  predicted performance is compared to actual
reported values.   The  effect of any individual  error on the overall
                                      48

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
performance, however, depends on the relative  error  and  the relative
contribution of each compound to the overall  influent  concentration.  Another
problem associated with the use of surrogate  compounds is  that if the
selected surrogate compound is also represented  in the influent,  it cannot be
run during the simulation because only one  Henry's Law value can  be selected
for a compound during the run.  For those sites  where  both 1,2-dichloro-
ethylene and 1,2-dichloroethane were found, only one or  the other could be
modeled per run because only the surrogate  or  the actual compound could be
selected—not both.
     Cost comparisons were available for five  systems  but  cost data were
relatively incomplete.  In general, the predicted capital  costs of air
strippers appeared to be reasonably close to  the actual  capital costs when
site-specific factors were considered.  Operating costs  reported  by the
sites, however, were generally lower than those  predicted  by ASPEN, probably
because of differences in costs included (e.g.,  overhead and taxes) by the
ASPEN cost subroutines that were not included  by the various sites in their
operating cost estimates.  The sites generally based operating costs on
capital recovery and energy-related costs.
     Control equipment options at two of the  sites were compared.  Each site
was run through the ASPEN simulation to determine the  size and cost
requirements for each control option (catalytic  oxidation  and vapor-phase GAC
adsorption).  The results of these simulations are discussed in Section 4.2
for each individual site.

4.2  SITE-SPECIFIC COMPARISONS
     This section discusses site-specific results of the ASPEN simulations
and the comparison of the ASPEN data with available  actual site data.
Performance ratios relative to ASPEN predictions are reported for individual
compounds and overall performance.  Ratios  greater than  1.00 represent an
under prediction by the ASPEN software when compared with  actual  site data.
Ratios less than 1.00 represent an overprediction of performance  by the ASPEN
software when compared with actual site data.  Where applicable,  the use of
surrogates is noted for individual sites, along  with other assumptions or
default values.
                                      49

-------
4.2.1  Tacoma Well 12A
     The ASPEN library of compounds contained  three  of the four compounds of
concern at this site.  A surrogate compound  was  selected  for the compound
1,1,2,2-tetrachloroethane and the Henry's  Law  constant was appropriately
modified.  The compound selected as a surrogate  for  this  simulation was
1,1,2-trichloroethane.  The presence of tetrachloroethylene as a compound of
concern precluded its use as a surrogate although  it would have been the most
desirable choice.  In addition,  1,2-dichloroethane was selected as a
surrogate for the trans-l,2-dichloroethylene found in the contaminated water.
Table 14 compares the results of the simulation  with the  actual performance
data.
     As presented in Table 14, the results for the surrogate compound for
1,1,2,2-tetrachloroethylene demonstrated a significant difference between
actual and reported results.  All other values were  underpredicted slightly
when compared with performance data.  The  overall  results showed an
approximate 2 percent difference between observed  performance and predicted
performance.  The surrogate compound for trans-l,2-dichloroethylene provided
accurate results in this case.
     Because the stripper design represented five  identical, parallel
stripping columns, the input was set up to calculate the  removal efficiency
and costs for one column.  The performance aspect  of the  model produced
reasonable estimates of overall  performance  at the high air/water ratio.  The
cost comparison between the reported capital costs and the predicted costs,
however, showed significant differences.  The  reported cost of the project
was $750,000 in 1983.  The estimated cost  of one module (purchased equipment
cost) was $236,600 (January 1986 dollars)  or a total of $1,183,000 for the
five modules.  The total estimated capital cost  for  this  system was
$1,904,650.  Even when considering the period  between 1983 and 1986 during
which time the inflation rate was low, the difference cannot be accounted for
in the cost.  The $750,000 figure was supposed to  represent the total cost,
including installation, engineering, and equipment cost.   It actually
appears, however, to be closer to the base equipment costs rather than the
total system cost.  The discrepancy between  the  two  costs cannot be resolved
at this time.
                                      50

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
             TABLE  14.  COMPARISON OF ASPEN SIMULATION TO ACTUAL
                   PERFORMANCE AT THE TACOMA WELL 12A SITE
Chemical
contaminant
1,1,2,2-Tetra-
chloroethane
trans-l,2-Di-
chloroethylene
Trichloro-
ethylene
Tetrachloro-
ethylene
Total VOC
Surrogate
contaminant
1,1,2-Tri-
chloroethane
1,2-Dichloro-
ethane



Influent
concentra-
tion, ppb
40.9
14.3
44.6
0.9
100.7
Observed
removal
effi-
ciency,
%
95.00
99.99
99.99
99.99
97.50
Predicted
removal
effi-
ciency,
%
89.17
99.98
99.98
99.98
95.59
Comparison
ratio
actual/
predicted
1.07
1.00
1.00
1.00
1.02
4.2.2  Rockawav Township Site

     All but three of the seven chemical  contaminants  currently found at the

site are included in the ASPEN library of compounds.   The three compounds not

included are 1,1,1-trichloroethane,  cis-l,2-dichloroethylene,  and

methyl-tert-butyl ether.  The surrogate compounds  selected to  represent these

compounds during the ASPEN simulation were 1,1,2-trichloroethane,

1,2-dichloroethane, and ethyl-propyl-ether,  respectively.  The choice of
1,1,2-trichloroethane as a surrogate for 1,1,1-trichloroethane is a good one

because the two chemicals are very similar in  structure and molecular weight,

the Henry's Law constant has been appropriately adjusted. Ethyl-propyl ether

is a less ideal choice, although it is similar in  molecular weight and of the

same chemical family structure.  The differences in the physical

characteristics are great enough that, even with the adjustment to its

Henry's Law constant, some difference between  actual  and predicted

performance would be expected.  The choice of  1,2-dichloroethane as a
                                      51

-------
surrogate is a good one because its  chemical  structure  is  reasonably similar
to that of cis-l,2-dichloroethylene.   No  1,1-dichloroethylene was  included in
the run because its logical  surrogate is  1,1-dichloroethane,  which was
already included in the chemical  list.
     Table 15 shows the results of the ASPEN  simulation.   The predicted
performance and actual performance are in good  agreement  for  most  of the
compounds.  The predicted removal  efficiency  for methyl-tert-butyl ether is
lower than the observed value by approximately  9 percent.   The significance
of this difference on overall performance is  negligible because of the small
contribution of methyl-tert-butyl  ether to the  overall  VOC loading.  Also,
the difference in the predicted versus actual removal could be accounted for
by the detection limits of the monitoring methodology.  Overall, the ASPEN
predictions differ slightly from predicted performance  when compared with
actual data.  These differences are considered  insignificant, however, when
compared with the monitoring methodology.  It should be noted that this is a
high air/water ratio design  (approximately 200:1).
     The estimated cost for  installing the air  stripper system was $375,000
versus the ASPEN-predicted cost of $269,240.  Extensive site  preparation
costs and pilot study work may account for a  portion of this  difference.  The
utility costs for the operation of the column agree quite well; the reported
estimated costs were approximately $57,000, compared with the predicted
$48,000.  Differences between these two values, when compared on the basis of
same unit cost for electricity, can be accounted for by changes in air
temperature throughout the year, which changes  actual  energy requirements.
The values  input into this ASPEN simulation represent only a limited time
frame, which expands to a longer time frame (i.e.,  one  year)  for operation
costs.  In  addition, other costs, such as the use of an iron sequestering
agent, are  not included in the cost of operating the air  stripper; however,
they represent a real cost because the air stripper operation has  a negative
effect on the dissolved iron  in the ground water.
                                      52

-------
1
1
1
1
•

1



1
••
1
1

1




1

1

1
•••
1
1
1
1
1




TABLE 15. COMPARISON OF ASPEN SIMULATION TO ACTUAL





Chemical
contaminant

Trichloro-
ethylene
Methyl -tert-
butyl ether
1,1-Dichloro-
ethylene
cis-l,2-Di-
chloroethy-
lene
Chloroform

1,1,1-Tri-
chloroethane
1,1-Dichloro-
ethane
Total VOC



•»
Not modeled.





PERFORMANCE AT THE ROCKAWAY




Influent
Surrogate concentra-
contaminant tion, ppb


28.3
Ethyl -propyl-
ether 3.2
4.0
1,2-Dichloro-
ethane
6.4
1.3

1,1,2-Trichloro-
ethane 20.0

2.0
65.2
(61.2)a



TOWNSHIP SITE


Observed Predicted
removal removal
effi- effi-
ciency, ciency,


99.99 99.96
99.99 91.56
99.99

99.99 99.98
99.99 99.96


99.99 99.25

99.99 99.98
99.99
99.29







Comparison
ratio
actual/
predicted


1.00
1.09


1.00
1.00


1.01

1.00

1.01



Predicted results do not include this compound.



53












-------
4.2.3  Brewster Well field Site
     The Brewster Well field Site was modeled for three compounds.   A
surrogate compound for 1,2-dichloroethylene (1,2-dichloroethane) was selected
for the simulation, and the Henry's Law constant was  modified appropriately
in the data input.  Table 16 shows the results  for the three compounds.
     The most obvious result is that performance is overpredicted somewhat
for each compound.  The results were closest for tetrachloroethylene and the
farthest apart for trichloroethylene.   The maximum relative error, however,
was approximately 5 percent.  Potential  causes  for these differences include
the channeling of air or water through the air  stripper,  inaccuracies in
estimating water temperature,  inaccurate estimates of air flow through the
stripper, and the sampling methodology used. Also the air/water ratio at
this site is 50:1, which is the second-lowest value of the selected sites.
The hypothesis is that the lower the air/water  ratio,  the greater the
potential for nonideal effects such as channeling to  occur within the
stripper column.  Overall, however, the predictions are within 2 percent of
the overall observed VOC removal  efficiency, and based on the other potential
sources of error, the predicted results appear  to correlate well with
observed performance.
     The reported installed cost for the air stripper was $138,000 versus a
predicted cost of $100,000.  Some of the costs  attributed to the air stripper
project at the Brewster Well Field are estimated from a total  water project
cost (i.e., a percentage of site preparation costs, piping costs,  etc.).
Changes in these factors would change  the estimated cost.   The assumptions
used to generate the estimated cost are well documented.
     The estimated annual  costs were predicted  to be  $52,500/year versus the
actual  site-estimated costs of $26,138/year, or approximately one-half of the
ASPEN-predicted cost.  The largest portion of this difference is due to the
length of time used for the capital cost recovery period.   The data reported
by the site represent a 20-year period,  whereas the ASPEN prediction was
based on 10 years.  When the ASPEN estimate is  based  on a 20-year period, the
predicted annual cost is $20,720,  which is somewhat lower than actual
estimates for the site.  This  variation can be  attributed to the difference
in initial capital costs estimated for the site.
                                      54

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
             TABLE 16.  COMPARISON OF ASPEN SIMULATION TO ACTUAL
                 PERFORMANCE AT THE BREWSTER WELL FIELD SITE

- -

Chemical
contaminant
Tetrachloro-
ethylene
Trichloro-
ethylene


Influent
Surrogate concentra-
contaminant tion, ppb

200

30
Observed
removal
effi-
ciency,
%

98.50

93.33
Predicted
removal
effi-
ciency,
%

99.42

98.67

Comparison
ratio
actual/
predicted

0.99

0.95
1,2-Dichloro-    1,2-Dichloro-
ethyl ene ethane
Total VOC
38
268
95.59
97.01
99.67
99.37
0.96
0.98

     In summary, the ASPEN model  overpredicted performance slightly and

underestimated equipment and operating costs.   Much  of the difference in the

values, however, can be accounted for in the  assumptions  inherent to data

gathering; thus, for this site,  the predictions appear to represent

satisfactory performance and costs estimates.

4.2.4  Verona Well Field Site

     The Verona Well Field Site  included nine  contaminants in the influent

water to the air stripper.  Only seven of the  nine compounds were modeled

because two of the compounds were not included in the  ASPEN library.  The two

compounds not modeled were cis-l,2-dichloroethylene  and l,l-dichloroethyl£ne.

Surrogate compounds could not be selected for  these  two because the compounds

that would normally be selected  as surrogates  were already included in the

compound list.  The surrogate compound selected for  1,1,1-trichloroethane was

1,1,2-trichloroethane.  These data were input  into ASPEN,  and Table 17

presents the performance results.
                                      55

-------
             TABLE 17.   COMPARISON OF ASPEN SIMULATION TO ACTUAL
                  PERFORMANCE AT THE VERONA WELL FIELD SITE

Chemical
contaminant
1,1-Dichloro-
ethane
1,2-Dichloro-
ethane
1,1,1-Tri-
chloroethane
cis-l,2-Di-
chloro-
ethylene
1,1-Dichlgro-
ethylene
Trichloro-
ethylene
Tetrachloro-
ethylene
Dichloro-
methane
Vinyl
chloride
Total VOC

Influent
Surrogate concentra-
contaminant tion, ppb
6.6
4.1
1,1,2-Tri-
chloroethane 10.3
15.3
1.0

2.1

16.4

41.2
34.0
131.0
114. 7a
Observed
removal
effi-
ciency,
%
98.53
29.27
76.70
83.01
99.99

99.99

99.99

66.26
99.99
82.90
83.32a
Predicted
removal
effi-
ciency,
%
100.0
88.85
83.06



99.98

100.00

99.14
100.00

97.79a
Comparison
ratio
actual/
predicted
0.99
0.33
0.92



1.00

1.00

0.67
1.00

0.85a

Compounds not modeled.   Results reflect performance excluding noted
 compounds.
                                     56

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
     The results indicate that ASPEN overpredicted  performance for this case
by approximately 15 percent.   The compounds  that  account  for this
overprediction are 1,2-dichloroethane,  dichloromethane, and the surrogate for
1,1,1-trichloroethane.  The difference  between  the  observed arid predicted
values for the surrogate is approximately 8  percent.   The relative error for
1,2-dichloroethane and dichloromethane  is substantial.  The cause of this
error is not readily apparent but the fact that only  two  compounds were
affected suggests that it may be due to a sampling  and measurement error
(i.e., contamination of the sample)  rather than a substantial  problem with
the ASPEN simulation.  For example,  the observed  removal  efficiency is
significantly lower than would be expected for  both compounds, especially
1,2-dichloroethane.  Dichloromethane (methylene chloride) is a common
laboratory solvent and could be a source of  sample  contamination.  As the
air/water ratio decreases the various nonideal  factors (such as channeling or
poor water distribution) may be more important  in actual  performance than can
be modeled.  This possible source of overprediction would be less likely to
cause a problem at higher air/water  ratios.   The  Verona Well Field air
stripper represents the lowest air/water ratio  (20:1) of  the seven selected
sites.
     The air emissions from the stripper pass through a vapor-phase GAC
adsorber system for control of VOCs. This was  included in the cost of the
total system, which and was estimated to be  $675,000.  The ASPEN prediction
for the air stripper/carbon adsorber system  cost  was  $514,000, which is lower
than the reported value by 24 percent.   In addition,  the  vapor-phase carbon
adsorption system actually operated  as  a nonregenerative  system rather than a
steam-regenerable system.  This, of  course,  would affect  the cost of
operation.  The reported cost of operation and  maintenance was $223,000
including capital-recovery costs. The  ASPEN predictions  were much lower
($185,300), partially because of the use of  steam regeneration rather than
carbon replacement and offsite regeneration.
     Initial testing data from the Verona Well  Field  indicated that
performance of the vapor-phase GAC system was quite good, despite the
extremely dilute concentrations of the  contaminants in the gas stream.  As
would be expected, the lighter-molecular-weight compounds broke through the
carbon first because the adsorption  capacity varies with  concentration and
                                      57

-------
molecular weight.  Based on this  initial  evaluation,  it  appeared that a
carbon change-out every year (12  months),  as  initially planned,  was going to
be normal routine.  When the additional  contamination of the highly
contaminated zone was added to the normal  well  field  contaminants,  however,
"normal" contaminant concentrations doubled (sometimes tripled).  In
addition, two compounds not normally seen at  the air  stripper,  vinyl chloride
and dichloromethane, were added in substantial  quantities,  as the adsorption
capacity of the liquid-phase treatment system used before the air stripper
was limited for these compounds.
     The inlet and outlet concentrations of the vapor-phase GAC system were
not routinely checked, however, under the interim removal action, the system
was tested on a more frequent basis to determine if changeout was needed.
During a test conducted in April  1987, both inlet and outlet concentrations
were measured.  Some pollutant concentrations fell below the detection limits
of the test method.  Table 18 summarizes the data on  the water and air
concentrations for this system.  Several items from these test data are worth
noting with regard to vapor-phase GAC performance.  First,  despite the
increased water concentrations, the test method determined that inlet VOC
concentrations in the gas were very dilute.  Second,  the overall removal
efficiency is quite low  (12.8%) because the two most prevalent compounds at
the inlet are also found at the outlet.  Finally, the outlet concentration of
dichloromethane  is much greater than the inlet concentration, which suggests
that  the vapor-phase carbon was saturated with VOCs.   The data suggest that
vinyl  chloride was being controlled only slightly and that dichloromethane
was being desorbed at a  rate  nearly equal to the  inlet,  which resulted in  an
emission rate higher than  if  it were not controlled at all.  Such  situations
must  be  accounted for in the  design of carbon adsorbers, because competitive
adsorption/desorption will  occur when multiple compounds are involved.   Other
data  provided from  a test  conducted 2 months later suggest  the  performance
had  improved, but some  of  the test data were inconsistent  in that  the  water
and  air  sample VOC  concentrations did not match well.
      The estimated  once-a-year carbon change-outs were  substantially
 increased to  once every 4  to  6 months,  usually  as a  result  of testing.   Under
normal  circumstances, when concentrations  of various  contaminants  are
relatively  steady,  the  operating  characteristics  and  time  before VOC
                                       58

-------
             TABLE  18. SUMMARY OF INLET AND OUTLET CONCENTRATIONS AND COLLECTION EFFICIENCIES
                                 FOR SELECTED COMPOUNDS, APRIL 30, 1987

Air stripper
Chemical
1,1-Dichloroethane
1,2-Dichloroethane
1,1-Dichloroethene
cis-l,2-Dichl oroethene
Dichloromethane
Tetrachl oroethene
1,1,1 -Tri chl oroethane
Trichl oroethene
Vinyl chloride
Total
Influent,
/ig/liter
6.6
4.1
1.0
15.3
41.2
16.4
10.3
2.1
34.0
131.0
Effluent,
/zg/liter
0.5
2.9
ND
2.6
13.9
ND
2.4
ND
ND
22.4
Removal
efficiency,
92.42
29.27
99.99
83.01
66.26
99.99
76.70
99.99
99.99
82.90
GAC adsorption system
Inlet,
/jg/m3
101
8.7
BDLb
BDL
197
37.4
BDL
BDL
162
516.1
Outlet
east,
Mg/m3
NDa
ND


330
ND


131
461
Removal
effi-
ciency, %
99.99
99.99


-67.51
99.99


19.14
10.68
Outlet
west,
/zg/m3
ND
ND


323
ND


116
439
Removal
effi-
ciency, %
99.99
99.99


-63.96



28.40
14.94

 None detected.
3Below detection limit.

-------
breakthrough may be well defined and require  little  if any monitoring.  When
              \
variations from the normal  condition occur and  extreme variability is the
norm, routine and frequent  monitoring should  be required.
  _  One option available for evaluating  GAC  designs is to determine if
breakthrough would be allowable for certain difficult-to-adsorb compounds and
to run the simulation without including them  in the  chemical  compound list:
This would determine how much of a change occurs in  VOC control equipment
size and cost occurs as a result of allowing  certain compounds to pass
through uncontrolled.  This was not done  for  the performance  comparison, but
the results could be used to evaluate different designs.
4.2.5  Western Processing Site
     The Western Processing Site was the  most complex site to evaluate
because it involved numerous compounds.   Of the 22 VOCs tested for in the
ground water, 18 had measured concentrations  above the detection limits.
Fourteen of these 18 compounds were modeled by  the ASPEN model.  The four
compounds not modeled were  hexachlorobutadiene,  hexachloroethane,
nitrobenzene, and 1,1,2-trichloroethane.   With  the exception  of
1,1,2-trichloroethane, none of these compounds  is currently included in the
ASPEN library.  The concentrations of these compounds,  however, were minor
compared with concentrations of other compounds.  Only one surrogate compound
was needed for this simulation (1,1,2-trichloroethane was  substituted for
1,1,1-trichloroethane).
     Table 19 presents a comparison of ASPEN  simulations and  actual
performance results at the  Western Processing Site.   With  several notable
exceptions, the predicted removal  efficiency  was greater than the observed
removal efficiency.  In most instances, the difference between the actual and
predicted values was within a 2 percent relative error.  The  difference for
benzene was greater, approximately a 7 percent  relative error.  Two other
compounds, 1,2-dichloroethane and 1,2-dichlorobenzene,  displayed large
differences between the observed and the  predicted values.  No reason for
these discrepancies was apparent but the  values suggest either a sample
testing problem or the need to select a smaller Henry's Law constant.
     Two compounds, carbon  tetrachloride  and  isobutanol, showed a ratio of
zero between the ASPEN run  and the actual  performance.   In the case of carbon
tetrachloride, the test method indicated  an inlet concentration of 5 ppb
                                    60

-------
1
1
1
•
1



1



1
•
1



1

1




1

1
•
1
1
1
1
TABLE 19. COMPARISON OF ASPEN SIMULATION WITH ACTUAL
PERFORMANCE AT THE WESTERN PROCESSING SITE


Chemical
contaminant


Benzene
Carbon tetra-
chloride
Chloroform
1,2-Dichloro-
ethane
1,1-Dichloro-
ethylene
1,1,1-Tri-
chloroethane
Trichloro-
ethylene
Vinyl chloride
Dichloro-
methane
Tetrachloro-
ethylene
Toluene

1,2-Dichloro-
benzene
Isobutyl-
alcohol
Methyl ethyl
ketone
Total VOC





Influent
Surrogate concentra-
contaminant tion, ppb


73
5
781

22

89
1,1,2-Tri-
chloroethane 1,440

8,220
159

8,170
378
551


11

10

1,480
21,389


61

Observed
removal
effi-
ciency,
%


93.15

99.36

77.27

94.38
99.65

99.94
99.37

99.63
98.68
99.09


54.55

0.00

70.27
97.52




Predicted
removal
effi-
ciency,
%


99.95
99.97
99.95

99.63

99.98
98.12

99.96
99.99

99.97
99.96
99.93


99.52

3.25

48.17
96.21





Comparison
ratio
actual/
predicted


0.93
0.00
0.99

0.78

0.94
1.02

1.00
0.99

1.00
0.99
0.99


0.55

0.00

1.46
1.01





-------
and an outlet concentration of less than 5 ppb without specifying the
concentration.  Carbon tetrachloride would be expected to be stripped from
the water at high efficiency, and this result appears to be caused by the
testing methodology.  Its effect on the overall results is very small.  For
isobutanol, the difference between the observed removal efficiency of
essentially zero percent and the predicted efficiency of 3.25 percent is not
significant and can be considered to be accurate.
     The removal efficiency for two of the compounds, 1,1,1-trichloroethane
(using the surrogate) and methyl ethyl ketone, was underpredicted by ASPEN.
The error in the use of the surrogate produced only a 2 percent relative
error for 1,1,1-trichloroethane and, therefore, is considered to correlate
well with actual performance.  Methyl ethyl  ketone removal efficiency was
significantly underpredicted by ASPEN (by approximately 46%).  Actual removal
efficiency for methyl ethyl ketone was reported to be 70.27 percent versus
the 48.17 percent predicted by ASPEN.  Although methyl  ethyl  ketone is a
particularly difficult compound to remove (as evidenced by its low Henry's
Law constant), there are no apparent reasons for this value to be
significantly different from the predicted value,  as sample contamination
would not normally cause this type of discrepancy.  For the moment, this
variation remains an unexplained anomaly.  The error is significant because
methyl ethyl ketone represents an important contributor to overall
contaminant levels.  When combined with the effects of the surrogate used for
1,1,1-trichloroethane, they offset the slight overpredictions for the
remaining compounds.
     As was the case at the Verona Well Field Site, VOC emissions are
controlled with a vapor-phase GAC adsorption system.  The system is not
steam-regenerated, but uses an incinerator to provide hot gas for
regeneration.  Capital cost data for this system and its associated operating
costs were not available for comparison with ASPEN simulation values.  These
costs may not be directly comparable in this case  because of regeneration
differences.
     PEI reviewed the data from both the inlet and outlet tests conducted at
the vapor-phase GAC adsorption system.  These inlet and outlet tests were
conducted with a photoionization type detector calibrated by using a known
concentration of 1,2-dichloroethylene in a carrier gas.  Such checks are
                                      62

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
typically conducted every 4 hours.   PEI  examined daily  average  inlet  and
outlet concentrations for a 26 day  period  in  the spring of 1989.   Daily
average removal efficiencies ranged from a low  of  17.4  percent  to a high of
7&.S percent.  The average for this period was  50.2  percent removal,  which is
far below the design value of 95 percent.
     The photoionization detector cannot determine which compounds are
present in the gas stream.  The ultraviolet lamp used must be of  sufficient
strength to photoionize all of the  VOCs  of interest; otherwise, VOCs  could be
present but not detected.  For example,  detection  of dichloromethane  requires
a very-high-intensity lamp.  Further,  it cannot be assumed that the
fractional compositions of the chemicals in the air  stream are  identical
between the inlet and outlet.  Most likely they are  not, as was demonstrated
at the Verona Well Field Site.  One problem noted  with  the testing of the
vapor-phase 6AC system is the missing  mass of VOCs not  found during testing
of the inlet.  Although considerable variability was observed in  the readings
that made up the daily averages, they  were always  considerably  less than what
would have been expected from a mass balance.  The photoionization detector
may not have the proper bulb installed,  which would  make it impossible for
the detector to "see" compounds such as  dichloromethane.  Also, some dilution
of the gas stream might be occurring as  a  result of  the "trans-"  stripper.
In any event, limited test data from the site seems  to  indicate that
performance is significantly below design  values.
4.2.6  Hicksville MEK Soil! Site
     The use of a high temperature  to remove  methyl  ethyl ketone  presents a
unique problem to the air stripper model.   The  ASPEN software automatically
adjusts the Henry's Law constant for water temperature  for the  compounds;
thus, increases or decreases are automatically  computed for most  design
situations.  The Onda-correlation,  however, has a  temperature limit of 45'C
in its estimation of mass transfer coefficients.   The  program assumes that
any temperature above this limit is 45'C.   The  "true"  mass transfer
coefficient is likely to be different from that predicted by the  Onda-
correlation, and the only method of compensating  for this limitation is to
alter the apparent Henry's Law constant  for the compound.
     The presence of only one compound in  the water  allowed for the altering
of the Henry's Law constant to match the performance observed for this
                                      63

-------
stripper design.  A good match between  predicted  and  actual  performance was
obtained by changing the constant  from  a  literature value of 4.35x10"
atm-m3/mole to 1.25xlO~4 atm-m /mole.   At this  Henry's  Law value,  the
predicted removal efficiency was a 96.79  percent  (Table 20)  versus an
observed value of 98.41 percent.
     The application of HTAS would probably  be  limited  to cases where
difficult-to-strip compounds are encountered in significant  quantities.  For
compounds with Henry's Law constants greater than 10"  atm-m /mole, a
correction to the Henry's Law constant  is probably not  necessary to
compensate for the temperature limitation of the  calculation of mass transfer
coefficients.
     The predicted installed cost  of the  air stripper was $212,000 versus the
reported cost of $323,000.  The reported  total  cost,  however, also included
an auxiliary package boiler and tank arrangement  that could  not be included
in the ASPEN simulation.  When this is  considered, the  ASPEN cost predictions
should compare well with the actual cost  of  the air stripper for this site.
Much of the operating cost associated with this stripper involved the
combustion of fuel oil to heat the water  for stripper operation.
     The ASPEN model could not predict  the problems associated with iron
oxide in the water or the costs associated with correcting these problems
(acidification and neutralization).  Therefore  although the  model  is an
important tool for predicting the  removal efficiency  for VOCs, other
important water-quality parameters cannot be ignored.
4.2.7  Sylvester's Gil son Road Site
     The Gil son Road Site uses a HTAS to  remove isopropyl alcohol, acetone,
and other compounds from the contaminated ground  water.  As  at previous sites
1,1,2-trichloroethane was selected as a surrogate for 1,1,1-trichloroethane,
and an appropriate adjustment was  made  to the Henry's Law constant of the
surrogate.  All other compounds were selected with the  default values for
Henry's Law constant for each compound.
     Table 21 summarizes the observed and predicted performance for each
compound.  For all compounds except acetone, the  model  slightly over-
predicted results when compared with actual  performance data.  The model
predicted nearly complete removal, whereas the observed performance suggested
                                      64

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
TABLE 20.  COMPARISON OF ASPEN  SIMULATION  WITH ACTUAL
  * PERFORMANCE AT THE HICKSVILLE  MEK SPILL  SITE
Influent
Chemical Surrogate concentra-
contaminant contaminant tion, ppb
Methyl ethyl
ketone
TABLE 21. COMPARISON
PERFORMANCE AT THE
Chemical Surrogate
contaminant contaminant
Isopropyl
alcohol
Acetone
Toluene
Dichloro-
methane
1,1,1-Tri- 1,1,2-Tri-
chloroethane chloroethane
Trichloro-
ethylene
Chloroform
Total VOC
15,000
OF ASPEN
SYLVESTER
Influent
concentra-
tion, ppb
532
473
14,884
236
1,340
1,017
469
18,951
Observed
removal
effi-
ciency,
%
98.41
Predicted
removal
effi-
ciency,
%
96.79
Comparison
ratio
actual/
predicted
1.02
SIMULATION WITH ACTUAL
'S GILSON ROAD SITE
Observed
removal
effi-
ciency,
%
95.30
91.93
99.87
93.79
99.45
99.71
99.06
99.41
Predicted
removal
effi-
ciency,
%
99.08
59.99
100.00
100.00
100.00
100.00
100.00
99.00
Comparison
ratio
actual/
predicted
0.96
1.53
1.00
0.94
0.99
1.00
0.99
1.00

                          65

-------
a slightly lower removal  efficiency.   The  overpredictions  of performance were
greatest for isopropyl  alcohol  and  dichloromethane.   Some  of this
overprediction may be the result  of nonideal  conditions  not  fully accounted
for by the model.  This stripper  differs from those  at other sites,  not only
because of its operating temperature,  but  also because it  uses trays instead
of packing to obtain the mass transfer from  liquid to gas.
     The performance of the stripper with  respect to acetone indicated a
large difference between the actual and predicted removal  efficiencies;
actual removal efficiency was underpredicted by approximately 56 percent.  Of
the compounds on the list for this  site, acetone was the most difficult to
strip, with a Henry's Law constant  of  2.50 x 10  atm-m  /mole.  As was the
case for the Hicksville MEK Spill site, the  use of the HTAS  presents some
computational difficulties for the  ASPEN model because of  the temperature
limitations on the Onda-correlation method of calculating  mass transfer
coefficients.  It may be prudent  to increase the effective Henry's Law
constant by approximately one-half  to  one  order of magnitude for this
compound and other compounds with values less than  10"   atm-m /mole to
overcome the temperature limitations  imposed by the  model.
     The cost of the stripper reported by  the operation  was  $45,000.  This
represents base equipment cost, not installed capital cost.   This correlates
well with the value of $50,600 predicted by  ASPEN.   Because  this stripper
also  includes a boiler to provide steam for  the stripper and other remedial
processes on site, the installed cost  would  be substantially more than
predicted by ASPEN.  The actual annual cost  would  also  be  greater.

4.3   SUMMARY
      The ASPEN performance predictions generally correlated well with
observed performance and provided estimates  that were within 2 percent of  the
observed removal efficiency.  The use of surrogate  compounds generally
introduced a larger error than was observed  for the actual compounds at  a
given  site.  This  is believed to be due to the fact that although chemically
similar, the physical data from the ASPEN  library that was used  in computing
mass  transfer coefficients were sufficiently different to introduce larger
errors in the predicted  efficiency.  The effect of these errors  depends  on
the fractional composition of the  total VOC loading of the surrogate
                                      66

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
compound.  With the exception of the surrogate compounds  and some isolated
chemicals, the ASPEN predictions tended either to  over- or underpredict for
all compounds at a particular site.
     Two compounds (1,2-dichloroethane and dichloromethane) presented a
problem with regard to calculating removal efficiency.  Part of the problem
with dichloromethane could be explained by sampling errors; however,
1,2-dichloroethane seemed to present a problem whether  it was a surrogate or
not.  Ketones, notably methyl ethyl  ketone and acetone, also may present a
problem, but these compounds were encountered in only one conventional air
stripper, and a great discrepancy occurred in predicted versus observed
performance.  The use of a HTAS represents a special  case in which
temperature limitations in the calculation of mass transfer coefficients may
require a change in the Henry's Law constant used  to estimate removal
efficiency.
     Based on this limited data set, it appears that at lower air/water
ratios the ASPEN model may slightly overpredict performance, and at higher
air/water ratios it may slightly underpredict performance.  Although data are
limited, this result is not unexpected because nonideal factors may have a
more detrimental effect on performance at lower air/water ratios than at
higher ratios (e.g., greater than 50:1).
     In both cases where vapor-phase GAC adsorption was used, the ASPEN model
predicted that a much greater quantity of carbon was needed to ensure good
performance.  Test data available from both sites  indicated that actual
performance was much lower than design performance.  The  presence of
compounds such as vinyl chloride and dichloromethane greatly increases the
carbon requirements of the adsorber, which suggests that  another alternative
may be appropriate for application where these compounds  are found.
     Cost comparisons were limited by available site data.  In general,
site-specific factors and lack of itemized costs from the sites limited the
ability to compare cost data.  In some cases, the  predicted costs appeared to
be reasonably close to reported costs.  More information  is needed to make
more definitive comparisons.
                                      67

-------
                                  SECTION  5
                   EMISSIONS TRADEOFFS  FROM AIR  STRIPPERS

     Air stripping, by design,  transfers VOC's from water  to  air according to
physical laws defined by equilibrium relationships, diffusion,  and mass
transfer.  Often, the focus is  on the removal of the  VOC from the water and
the achievement of target concentrations after the water passes through the
stripper.  This focus tends to  overshadow  other  aspects of the  stripper
operation.
     Four of the seven sites included in the performance comparison are not
equipped with VOC emission controls.  In the future,  more  air strippers will
likely be equipped with some form of VOC emission control.  These may be
required by air toxic regulations that  are Applicable or Relevant and
Appropriate Requirements (ARARs)  for Superfund sites  or under other
guidelines or directives.  The  application of VOC emissions control,  however,
is not without its costs or impacts.  Discussions regarding the application
of a system to remove or control  VOC emissions tend to focus  on the cost of
the equipment and its performance.
     The specification and installation of VOC controls will  probably change
the design of air strippers.  For example, early designs of air strippers
tended to rely on large air/water ratios to remove VOCs from  contaminated
water.  This usually meant that large quantities of air had to  be moved
through the system to produce very dilute  VOC-bearing gas  streams.
Application of VOC controls to  these systems tended to require  large and
expensive control systems because much  of  the sizing  of equipment has a
direct relationship to gas volume handled. The  sizing of  catalytic oxidizers
is also directly influenced by  the amount  of gas handled because this
influences both incinerator size and fuel  requirements.  Vapor-phase GAC
adsorption systems have limits  on the velocity and pressure drop through a
carbon bed that need to be considered in addition to  the amount of carbon
required for the quantity and concentration of the VOCs  in the  gas stream.
In either case, a reduction in  gas volume  through a reduction of the
                                      68

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
air/water ratio will  generally result in  a  more  cost-efficient control
system.
     The operation of any air-stripping system generates  the potential  for
air emissions.  Most  apparent are the uncontrolled  VOC  emissions from the
stripper itself.  The operation of the stripper,  however,  requires the use of
pumps and fans to move water and air and  energy  (electricity) to operate this
equipment.  Although  electricity may not  be generated on  site, some
incremental increase  in electrical generation  and the amount of fuel
combusted at a generating station is required  to produce  this energy, and
this results in some  incremental increase in emissions  of NOX, SC^, CO,
particulates (PM1Q),  and nonmethane hydrocarbons.  Generally, the quantity of
electrical energy used is small compared  with  the output  of a single
generating station.  Also, the impact of  this  generation  remote from the site
of the air stripper is not considered. Although this impact is real, it is
incrementally small.
     A more direct local impact should be considered when VOC controls are
applied to air strippers.  Both control options  (catalytic oxidation and
vapor-phase GAC adsorption) require the combustion  of fuel, which usually
results in a local impact of greater concern than electrical consumption.
Catalytic oxidation,  for example, requires  fuel  combustion to establish and
maintain the incinerator temperature for  VOC destruction.   This fuel
combustion produces S02, NOX, CO, and nonmethane hydrocarbons.  In addition,
the presence of halogenated VOCs produces halogenated acids (HX) upon
combustion, which also must be considered.   Vapor-phase GAC adsorption also
generates emissions.   The model assumes steam  regeneration of the carbon and
the use of a small boiler to produce the  needed  steam.  The combustion of No.
2 fuel oil to generate this steam produces  the same types of pollutants as
the catalytic oxidizer with one exception.   The  organic constituents captured
by the carbon are assumed to be recovered,  stored,  and  shipped offsite for
further processing (e.g., solvent recovery).  The aqueous material from steam
regeneration would be contaminated with the same VOCs,  and this material is
assumed to be processed by the air stripper prior to its  discharge.  No HX
are assumed to be formed because this option does not involve the combustion
of halegenated VOC streams.
                                      69

-------
     Offsite regeneration does not mean that  no emissions are created; it
only means that they are not created on site.   Depending on the regeneration
method used, the emissions could be nearly identical  in  either option
(incineration or steam regeneration).   For the purpose of this evaluation,
offsite regeneration is assumed to be equivalent to onsite steam
regeneration.
     Both options also contribute to the production of carbon dioxide (C02)
as a result of fuel combustion.  Although not considered a pollutant, C02 is
becoming more important from a global  warming perspective.  The quantity of
C02 is displayed for information purposes in  the ASPEN simulation results.

5.1  EMISSIONS ESTIMATE METHODOLOGY
     The estimation of the uncontrolled VOC emissions rate is a straight-
forward material balance.  The quantity of VOCs in the influent water is a
known quantity that must be input as part of  the initial input data.  After
calculating removal efficiency, the ASPEN model assumes  that the VOCs
stripped from the water enter the air stream.   The quantity of each compound
stripped from the water is summed to give the overall uncontrolled VOC
emission rate.
     Estimating air emissions from the control options is somewhat more
complex because estimates have to be made for the average fuel consumption
(106 Btu/h) required to operate the process.   Emission factors from AP-42
are then used to convert fuel use to emission rates for  each of the fuel
combustion pollutants.  The emission factors  used are shown in Table 22.
     The catalytic oxidation option estimates the amount of fuel required to
heat and maintain the air temperature at approximately 700"F.  The fuel is
assumed to be natural gas.  The heat requirement from this estimate is used
to estimate the quantity of fuel combustion products and pollutants.  The
quantity of products of the combustion from the VOCs exiting the stripper is
calculated and added to the fuel combustion products.  Halogenated compounds
are assumed to convert to HX and are summed for all halogenated compounds.
The total quantity of pollutants generated from the catalytic oxidizer
(except C02) are summed and then compared with the uncontrolled emission
rates to determine if a net increase or decrease in pollutants occurred as  a
result of using the control equipment.
                                      70

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
               TABLE 22.   EMISSION FACTORS USED BY ASPEN  MODEL
                       FOR FUEL COMBUSTION EMISSIONS3
Pall utant
so2
CO
VOC (nonmethane)
co2
Natural
lb/106 ft3
0.6
100.0
20.0
5.3
116,596.0
aas
lb/106 Btub
6.00 x 10"4
0.100
0.020
5.30 x 10"3
116.36
Distil
lb/103 gal
142(S)
20
5
0.34
22,747.0
late oil
lb/106 Btuc
1.022(S)
0.144
0.036
0.002
163.65
aFrom AP-42, Supplement 13.

bAssumes natural gas heat content of 998 Btu/ft3.

cAssumes distillate oil heat content of 139,000  Btu/gal.
     The vapor-phase GAC adsorption option  uses  a  similar approach to that of

catalytic oxidation because fuel  combustion emissions  are estimated from heat

input estimates.  To estimate the heat input for the regeneration of the

carbon, either the actual carbon  used (rating mode) or the carbon required

(design mode) is used to estimate steam requirements,  given the adsorption

cycle time.  This, in turn, gives the number of  regeneration cycles and an

assumed value of 3.5 Ib steam/1b  carbon required to regenerate the carbon.

The heat input is then estimated  from the total  quantity of steam required

per year and averaged to a 10  Btu/h heat input  value.   The emission factors

from Table 22 were then applied and summed  for comparison in a manner similar

to that for catalytic oxidation.

     The comparison between uncontrolled VOC emissions and emissions produced

as a result of using a control  option resembles  a  comparison between

dissimilar pollutants with different air toxics  implications.   The intent of

the comparison,  however, is to  demonstrate  that  the use of air emission

controls may have other impacts.   In fact,  the air emissions from the

vapor-phase carbon adsorption may underestimate  the true magnitude of
                                      71

-------
short-term impacts.  Because the boiler operation may be intermittent, the
values shown represent long-term averages,  not short-term peak values.

5.2  EMISSIONS COMPARISONS FOR THE CONTROL  OPTIONS
     The data from the seven selected sites were used as a basis for emission
estimates for three scenarios:  no controls,  vapor-phase GAC adsorption, and
catalytic oxidation.  Each site was run in  the design mode to determine what
emissions would occur given the air stripper performance and allowing the
ASPEN software to estimate control  equipment sizing  for the two control
options.  The emission rates for each site  are based on this design mode
value.
5.2.1  Tacoma Well 12A
     The selection of either option for this site results in a net emissions
increase from the operation of the control  equipment (Table 23).  The
increase in emissions resulting from the vapor-phase GAC system, however, is
much smaller than that resulting from the catalytic  oxidizer.  This
difference is caused by the large gas flow  from this stripper design, because
a large quantity of heat is required to bring the gas stream up to
temperature.   The vapor-phase option has a  much lower overall impact because
each of the compounds included in the contaminant list is very easy to
adsorb, and the limiting factor for the GAC design is gas velocity through
the beds, not a large carbon requirement.
5.2.2  Rockawav Township
     No comparison could be performed for this site  because of missing data
in the ASPEN library for the surrogate compound ethyl-propyl  ether.  Computer
runs gave invalid results, which are not included in the simulation.
5.2.3  Brewster Well field
     The selection of a vapor-phase GAC adsorption system resulted in a much
higher net emission increase when compared  with emissions from catalytic
oxidation (Table 24).  Although the air/water ratio  for this stripper is
relatively low, the presence of two difficult-to-adsorb compounds (vinyl
chloride and the surrogate 1,2-dichloroethane) increases the carbon
requirements substantially.  This directly  affects the steam and heat input
                                      72

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
        TABLE 23.  EMISSIONS COMPARISON FOR THE TACOMA WELL 12A SITE
No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 0.0765
HX
SO,
NO^
COX
VOC (nonmethane)
C02
Emissions from control
option (excluding C021
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h



0.3785
0.0533
0.0133
0.0007
60.60
0.4459

-0.3694
Catalytic oxida-
tion,.kg/h


0.0651
0.8905
6.050
1.216
0.3159
7035
8.5350

-8.4600

aNegative values indicate an overall  emissions increase over the uncontrolled
 option.  Positive values indicate a  net decrease.
      TABLE 24.  EMISSIONS COMPARISON FOR THE BREWSTER WELL FIELD SITE

No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 0.0183
HX
S09
NO*
COX
VOC (nonmethane)
co2
Emissions from control
option (excluding CO^l
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h



31.79
4.479
1.120
0.062
5090

37.45

-37.43
Catalytic oxida-
tion, kg/h


0.016
0.005
0.033
0.007
0.002
38.46

0.062

-0.044

aNegative values indicate an overall  emissions  increase  over the uncontrolled
 option.  Positive values indicate a  net  decrease.
                                      73

-------
requirements, .which are reflected by the pollutant  emission rate.   According
to the ASPEN prediction, selection of the vapor-phase GAC system would
significantly increase S02 and NOX emissions.   The  GAC system boiler would be
classified as a major SC^ source under Prevention of Significant
Deterioration (PSD) review.
5.2.4  Verona Well Field
     Table 25 shows the results of emissions estimated for the Verona Well
Field.  Although this site has a vapor-phase GAC system installed,  the data
available on its performance suggested that  its performance was poor, given
the compounds it was trying to control.   The ASPEN  design suggested a much
larger GAC system would be required, this larger system is reflected in Table
26.  The presence of compounds such as 1,2-dichloroethane, dichloromethane,
and vinyl chloride greatly increased carbon  requirements  for proper
operation.  Thus, although the low air/water ratio  helps  in the design of the
catalytic oxidizer, it does little to help reduce the estimated emissions
from the GAC system because of the poor adsorbability of  the compounds
mentioned in the preceding sentences.   For such a small removal rate, the use
of GAC extracts a significant emissions penalty in  this case.   In  addition,
it would be subject to PSD review because it would  be classified as a major
SO- source.
5.2.5  Western Processing Site
     The emission rate comparisons for this  site show some astounding numbers
(Table 26).  The emission rate estimates are orders of magnitude greater than
all other estimates.  In addition, the heat  input rates required to produce
these emission rates are extremely high (to  put this into context,  the
necessary heat input rate would represent a  substantial portion,
approximately 2.38 million MW, of the U.S. steam-generating capacity for
generating electricity).  This example demonstrates that  one should carefully
examine the results.
     In this case, the extremely large numbers  are  caused by the presence of
large quantities of difficult-to-absorb compounds.   The two compounds that
have the greatest effect on the report values are vinyl chloride and
dichloromethane, both of which require extremely large quantities  of
activated carbon to control.  This further suggests that  vapor-phase GAC
                                      74

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
      TABLE 25.  EMISSIONS COMPARISON FOR THE VERONA WELL FIELD SITE
No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 0.050
HX
SO,,
NO
CO
VOC (nonmethane)
C0«
Emissions from control
option (excluding C021
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h



146.10
20.59
5.147
0.286
23,400

172.1

-172.1
Catalytic oxida-
tion, kg/h


0.038
0.030
0.195
0.039
0.010
226.5

0.312

-0.262
Negative values  indicate  an overall emissions increase over the uncontrolled
 option.   Positive  values  indicate a net decrease.
       TABLE 26.   EMISSIONS  COMPARISON FOR THE WESTERN PROCESSING SITE
No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 0.355
HX
SO,
N(r
CO
VOC (nonmethane)
C02
Emissions from control
option (excluding CO^l
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h


7
1.160 x 10;
1.634 x 10?
4.085 x 10?
2.269 x 10n
1.857 x 10y
7
1.366 x 10'
7
-1.366 x 10'
Catalytic oxida-
tion, kg/h


0.280
0.012
0.083
0.167
0.004
96.72

0.396

-0.041
Negative values indicate an  overall  emissions increase over the  uncontrolled
 option.  Positive values indicate  a  net decrease.
                                      75

-------
adsorption is an inappropriate VOC control  method,  at  least for these
compounds.
     By comparison, catalytic oxidation  looks  very  attractive on an overall
emissions basis.  A slight emission increase of approximately 11.5 percent
over the uncontrolled levels is predicted  for  this  option.   The low gas
volumes help to reduce the net emission  increase.
5.2.6  Hicksville MEK Spill  Site
     The Hicksville MEK Site represented one of two cases where a net
reduction of emissions resulted from the addition of controls.   In this case,
a net reduction occurred with the use of catalytic  oxidation.   This reduction
was due in part to low gas volumes and high gas-stream temperatures.   The
carbon adsorber option indicated a net increase in  emissions,  although only
one compound was present.  Table 27 shows  the  results  of the comparison.
     Methyl ethyl ketone is a relatively easy  but dangerous compound  to
adsorb.  The heat release rates from the adsorption process can be so high
that fires can occur within the carbon bed if  special  precautions are not
taken.  This generally involves humidification of the  gas stream to carry
away the heat of adsorption.  The model  does not consider this  problem in its
evaluation of adsorber designs.
     The use of HTAS involves the use of a boiler to increase water
temperature.  The fuel required would likely cause  a further net increase in
emissions for both options.   This design,  however,  also offers  the
opportunity to use the boiler as an incinerator, which reduces  the
supplementary fuel requirements for a separate catalytic oxidizer.
5.2.7  Sylvester's Gilson Road Site
     This site was the only other site of  the  seven where a net reduction of
emissions was predicted for one of the control options. As at  the Hicksville
MEK Spill Site, the catalytic oxidation  option provided a net reduction in
emissions due, in part, to the operation of the HTAS.   A substantial  increase
in emissions was predicted with the use  of a vapor-phase GAC system compared
with uncontrolled values.  This increase was due to the presence of several
compounds that are difficult to adsorb.  These compounds increase the carbon
requirement and hence the emission rate.  Table 28  compares the results.
                                      76

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
         TABLE  27.  EMISSIONS COMPARISON FOR THE HICKSVILLE MEK SITE
No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 0.318
HX
S0«
NO
CO
VOC (nonmethane)
rn
LU«
Emissions from control
option (excluding C021
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h


2.438
0.343
0.086
0.0048
390.3

2.872

-2.553
Catalytic oxida-
- tion, -kg/h

0.00
0.014
0.050
0.010
0.0026
58.52

.074

0.244
Negative values indicate  an overall emissions increase over the uncontrolled
 option.   Positive values  indicate a net decrease.
    TABLE 28.   EMISSIONS COMPARISON FOR THE SYLVESTER'S GILSON ROAD SITE
No controls,
Pollutant kg/h
Uncontrolled VOC
emissions 1.278
HX
S00
NO
COX
VOC (nonmethane)
C02
Emissions from control
option (excluding C021
Net emissions decrease
(excluding C02)
Vapor-phase
GAC, kg/h



3905
550.1
137.5
7.64 ,
6.252 x 10°

4600

-4599
Catalytic oxida-
tion, kg/h


0.176
0.025
0.114
0.023
0.006
136.4

0.344

0.934
Negative values indicate an overall  emissions  increase over  the  uncontrolled
 option.  Positive values indicate  a  net decrease.
                                      77

-------
     Under this scenario,  the  boiler would have  to be very  large  and would be
subject to PSD review.  Under  normal circumstances,  little  or  no
consideration would be given to  attempting to capture isopropyl alcohol.
Acetone can also present special  problems in terms of the use  of  a
vapor-phase GAC system.  Because this  site uses  its  boiler  as  an  incinerator,
the net emission reduction predicted to  result from  the  use of catalytic
oxidation may be similar to actual  operation at  this site.   Using the boiler
to heat the water would tend to  increase the net emissions  resulting from the
vapor-phase GAC system option.

5.3  SUMMARY
     In each case the selection  of vapor-phase GAC  adsorption  as  a control
option results in a net increase in the  overall  emissions.   In cases where
large quantities of carbon would be required, the net emission increases
could be substantial and may necessitate a rethinking of the need or desire
to control such compounds.  Such results would also  indicate another
alternative would be more appropriate.
     A smaller net emissions increase  was generally  predicted  for the
catalytic oxidation option than  for the  GAC  option  because  the heat input and
emissions are controlled more by gas volume  than by the chemical  composition
of the gas stream.  The Tacoma Well 12A  Site was the only one where the
emissions from the operation of catalytic oxidation  were much larger than
those from carbon adsorption.   These larger  emissions were  due to the high
gas volumes that would have to be handled  by the oxidizer.   A net emissions
decrease was predicted for both HTAS because of  the reasonable gas volumes
and the high operating temperatures.  Actual  net emission reductions would
probably be less because of the boiler operation for heating the water going
to the stripper.
     With the focus on air-stripper performance, these  net emissions
increases or decreases tend to be overshadowed or not even considered.   The
ASPEN model provides  a method for at least  comparing net emissions  increases
                                      78

-------
the vapor-phase carbon adsorption option.
I
I
         or decreases for evaluation of existing  designs  or  potential  control
•       alternatives.  This could also be used to  indicate  the  need for other reviews
         such as a PSD review for new major sources,  as was  shown for four sites with
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                      79

-------
                                  SECTION  6
                       CONCLUSIONS AND  RECOMMENDATIONS
6.1  CONCLUSIONS
     The first major objective of the  project  was  to  provide a straight-
forward method of inputing data for the  ASPEN  air  stripper and control  option
modules while keeping the programming  requirement  transparent to the user.
The user-interface software provides this  by allowing the  user to input data
through a series of menus with default values  available  to evaluate existing
designs and new design options for sizing  and  cost purposes.  A user's guide
and documentation have been developed  and  are  supplied as  a separate
document.
     The second major goal of this project was to  develop  an output report
format that provided the most information  in a usable format.  The default
printing characteristics are inherently  cryptic and are  displayed in less
familiar units than most people are used to reading.   The  report format
developed under the project converts these ASPEN software  outputs to more
conventional and familiar units (e.g., percentage  removal, kg/h, etc).
     The third major goal of this project  was  to compare the predicted
performance with actual performance for  a  sampling of air  strippers operating
under a variety of conditions and treating several different chemicals.
Seven strippers were selected for comparison.   Twenty-five different
compounds were evaluated for strippers with air/water ratios ranging from
20:1 to 300:1.  In general, performance  predicted  by the ASPEN model matched
the actual performance within 1 percent  relative accuracy.  In some isolated
instances the predicted performance did  not match  the observed performance,
generally for compounds such as methyl ethyl  ketone and  dichloromethane.   In
some cases, the cause of the discrepancy could not be determined.
     A small number of compounds were  not included in the ASPEN library of
chemicals.  For these compounds (e.g., 1,1-dichloroethylene, 1,2-dichloro-
ethylene, and 1,1,1-trichloroethane),  the selection of a surrogate compound
                                      80

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
that was chemically similar (i.e., similar in structure and molecular weight)
was required.  A greater relative error was noted for most simulations when
surrogate chemicals were used.  Again,  the error was generally within 4
percent of the actual value.
     One tentative conclusion derived from these simulations is that the
ASPEN model tends to overpredict performance slightly at lower air/water
ratios (less than 50:1) and to underpredict slightly at higher air/water
ratios.  More study of this situation would require many more stripper
designs to be evaluated.  This seems plausible,  however, because nonideal
effects (channeling) would tend to occur more often at lower air/water ratios
than at higher ones.
     One area in which data were lacking concerned cost comparisons.  Costs
were often quoted with only limited or no supporting data or assumptions
available from the site.  Therefore, only limited cost comparisons were
possible, and it was sometimes difficult to tell  which items were included in
site costs estimates so that a meaningful comparison could be made.
     Limited control option comparisons could be made for sites where
controls were used.  Both sites using vapor-phase GAC control did not match
the ASPEN model design exactly because the ASPEN model assumes steam
regeneration.  The ASPEN model could be used, however, for evaluation of the
carbon requirements and a comparison of existing designs.   The output
suggested that the two designs evaluated were inadequate to provide  a high
degree of VOC removal.  Limited site data confirmed this,  as VOC removal
efficiency was very low at these two sites and substantially below design
values.  Results in these two cases suggest that some other control  option
would be appropriate and/or that better monitoring of performance was needed.
     The fourth major goal of this project was to evaluate the emissions
tradeoffs that occur when controls are  applied to air strippers.   Each of the
seven selected sites was run on the ASPEN simulation to develop control
equipment designs for the air stripper  off-gas.   In general,  the application
of emission controls resulted in a net  increase  in total  emissions although
VOCs were controlled.   These conclusions were based on energy-usage
calculations and emission factors.   It  was also  predicted  that, except for
very large air flows,  the use of vapor-phase GAC  adsorption with steam
regeneration results in a larger net emissions increase than does catalytic
                                     81

-------
oxidation.  In addition, for some systems,  adequate control  of certain
compounds (such as vinyl chloride and dichloromethane)  may be impractical.
The ASPEN software does not attempt to weigh  the toxics effects of the
uncontrolled VOC emissions against other emissions.  It simply indicates how
much of particular pollutants will be generated.

6.2  RECOMMENDATIONS
     The modules assembled for this version of the ASPEN software represent
the most common stripper design and most probable emissions  control options.
An additional option of nonregenerable vapor-phase GAC  would probably be
useful in the future.  The program has shown  itself to  be versatile enough to
simulate HTAS.  To date, however, only limited application of this technique
has been seen at actual Superfund sites.
     An alternative that may see more use in  the future is cross-flow
stripping.  This technique can be used in situations where relatively high
                                                                       _2
concentrations of easily stripped compounds (Henry's Law constants = 10   to
10   atm.m3/rool) can be stripped with low air/water ratios,  which results in
low air volumes requiring control.  More-difficult-to-strip  compounds
                                o            c
(Henry's Law constants = low 10   to high 10    atm.m3/mol) can then be
stripped in a second column operating at much higher air/water ratios.  An
example of such a system is shown in Figure 10.   This may become a viable
system in the future to reduce control costs  and may compete with traditional
systems.  If this system appears to be a viable  approach, an additional
module for ASPEN should be considered, as such a simulation  now could only be
achieved by running the ASPEN simulation at least twice and  using the results
of one run as the input for the second.
                                      82

-------
00
CO
                                 Emissions
                                  Control
Water w/  Low
Henry's  Law
Constant
Compounds
                         Atmosphere or
                         Emissions  Control
                         (Separate From Low
                           A/W Tower System)
                                                                    Air
                    Low  Air/Water
                     Ratio Tower
                High  Air/Water
                  Ratio  Tower
                   Figure  10.  Process flow diagram for "cross-flow" air stripper
                              for  handling reduced gas volumes.

-------
                                 REFERENCES


1.   Radian Corporation.  Air Stripping of Contaminated Water Sources Air
     Emissions and Controls.   Prepared for the U.S.  Environmental  Protection
     Agency.  July 1987.

2.   U.S. Environmental Protection Agency.   Compilation of Air Pollutant
     Emission Sources.  AP-42,  Fourth  Edition.   September 1985.
                                     84

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
                                           APPENDIX A
                                   SUMMARY OF SITE PARAMETERS
85

-------
                                TABLE A-l.  AIR STRIPPER LOCATIONS AND CONTRACTS

Site
number
1
Site name
and
location
Tacoma Well 12A
State/local
contact and
phone
Ken Merry
EPA/Superfund
contact and
phone
Kevin Rochlon
Startup
date
July 1983
Operating
hours per
year
2500
Type of
emissions
control
None
00
en
            Tacoma, WA
            Pierce County
            Rockaway Township
            Rockaway Township, NJ
            Morris County
                          Tacoma Public
                           Utilities
                          (206) 593-8210

                          Steve Levinson
                          Rockaway Dept.
                           of Health
                          (201) 627-7200

                          Mary Lou Parm
                          NJ DEP
                          (609) 292-5383
Region X
(206) 442-2106
                February
                 1982
               8760
               None
Brewster Well Field
Village of Brewster, NY
Putnum County

Verona Well Field
Battle Creek, MI
Calhoun County
            Western Processing
            Kent, WA
            King County
            Hicksvillle MEK Spill
             Site
            Hicksville, NY
            Nassau County
Robert Wing
Region II
(212) 264-8670
                                      Pat McKay
                                      Michigan Dept. of
                                      Natural Resources
                                      (517) 373-8448
                                            Loren
                                             McPhillips
                                            Region X
                                            (206) 442-4903
October
 1984
                September
                 1984
                September
                 1988
                                            Robert Cobiella June
                                            Region II
                                            (201) 321-6646
                     1984
8760
               8760
             GAC (nonre-
             generable)
            Not
             established
             with fume
             incinerator

            3 months un-
             der removal
             action
None
            Vapor-phase
            Vapor-phase
            GAC (CADRE)
               None
   (continued)

-------
  TABLE  A-l  (Continued)
Site
number
7
Site name
and
location
Gil son Road Site
(Sylvester's)
Nashua, NH
Hillsborough County
State/local
contact and
phone
Robert Ostrofsky
NH Dept. of En-
vironmental
Services
(603) 882-3631
EPA/Superfund
contact and
phone
Chet Janowski
Region I
(617) 573-9623
Startup
date
June 1986
Operating
hours per
year
8760
Type of
emissions
control
Boiler/In-
cinerator
00

-------
                                  TABLE A-2.  AIR STRIPPER AIR AND WATER FLOWS
   Site
   num-
   ber
       Water flow
                                      Dry air flow
Water vapor   Air and water vapor
'C    kg/h  kmol/h  m3/h
                                      kg/h  kmol/h  m3/h     kg/h   kmol/h   m3/h    kg/h   kmol/h   m3/h
00
00
1   10.0 788,720 43,818 794.9
2   11.8 315,488 17,527 318.0
3   12.8  67,605  3,756  68.1
4   12.8 439,430 24,413 442.9
5    7.2  22,535  1,252  22.7
6   88.0  21,938  1,218  22.7
7   79.4  68,101  3,780  68.1
                          10.0 295,974 10,241  246,357
                          20.0  76,545  2,649   63,713
                          15.6   4,093  141.6    3,407
                          20.0  10,206  353.1    8,495
                          11.4   4,389  151.8    3,653
                          29.0   3,266  113.0    2,718
                          43.3   4,205  145.5    3,500
                                                                3,848.7 213.8  2,332.7  299,823  10,455 248,691
                                                                  491.9   27.3    645.5   77,037  2,676  64,359
10.6   0.6
66.3
28.5
21.2
27.5
3.7
1.6
1.2
1.5
13.9    4,104  142.2   3,421
87.0   10,272  356.8   8,582
37.4    4,417  153.4   3,690
27.7    3,287  114.2   2,746
35.9    4,232  146.9   3,536

-------
                                  TABLE A-3.  ORGANIC CONTENT OF WATER STREAMS
00
IO

Site Chemical Design,
number contaminant ppb
1 1,1,2,2-Tetra-
chl oroethane 300
trans-l,2-Di-
chloroethylene 100
Trichloro-
ethylene 130
Tetrachloro-
ethylene 5
Total VOC 535
2 Trichloro-
ethylene
Diisopropyl
ether 4,000
Methyl -tert-
butyl ether
1,1-Dichloro-
ethylene
cis-l,2-Di-
chloro-
ethylene
Chloroform
Water entering
ppb

40.

14.

44.

0.
100.

28.

ND

3.

4.


6.
1.

9

3

6

9
7

3



2

0


4
3

0

0

0

0
0

0



0

0


0
0
kg/h

.033

.011

.035

.001
.080

.0090



.0010

.0013


.0020
.0004


1

1

2

4
5





1

1


2
3
stripper
kmol/h
•
.935 x 10"^


•
.706 x 10"^
C.
.31 x 10"6
.86 x IO"4

0.0001


c
.15 x 10°
I*
.31 x 10"&

f
.10 x 10°
.47 x IO"6
Water leaving stripper Weight remov-
ppb kg/h

2.05 0.002

ND

ND

ND
2.05 0.002

ND

ND

ND

ND


ND
ND
	 ai ci i IV.ICH
kmol/h cy, %

9.70 x IO"6 95.0

99.9

99.9

99.9
9.70 x IO"6 97.5

99.99



99.99

99.99


99.99
99.99
   (continued)

-------
    TABLE A-3 (Continued)

-------
TABLE A-3 (Continued)

Site
number








5

Chemical
contaminant
1,1,1-Tri-
chloro-
ethane
cis-l,2-Di-
chloro-
ethylene
1,1,-Di-
chloro-
ethylene
Trichloro-
ethylene
Tetrachloro-
ethylene
Dichloro-
methane
Vinyl chloride
Total VOC
Benzene
Carbon tetra-
chloride
Chloroform
Design,
ppb
150
229
11
62
94



2,000
700
20,000
Water
ppb
10.3
15.3
1.0
2.1
16.4
41.2
34.0
131.0
73
5
781
entering
kg/h
0.0046
0.0068
0.0004
0.0009
0.0073
0.0182
0.0151
0.0580
0.0017
0.0001
0.0177
stripper
kmol/h
3.46 x 10"5
0.0001
4.12 x 10"6
6.87 x 10"6
4.40 x 10"5
0.0002
0.0002
0.0007
2.18 x 10"5
6.49 x 10"7
0.0001
Water
ppb
2.4
2.6
ND
ND
ND
13.9
ND
22.4
5
5
5
leaving
kg/h
0.0011
0.0012



0.0062

0.0099
0.0001
0.0001
0.0001
stripper
kmol/h
8.27 x 10"6
1.24 x 10"5
99.99
99.99
99.99
0.0001
99.99
0.0001
1.28 x 10"6
6.49 x 10"7
8.40 x 10"7
Weight remov-
al efficien-
cy, %
76.70
83.01



66.26

82.90
93.15
0
99.36
(continued)

-------
TABLE A-3 (Continued)

Site Chemical
number contaminant
1,1-Dichloro-
ethane
1,2-Dichloro-
ethane
1,1-Dichloro-
ethylene
Fluorotri-
chloro-
methane
1,1,1-Tri-
chloro-
ethane
Trichloro-
ethylene
Vinyl
chloride
Dichloro-
methane
trans-1,2-
Dichloro-
ethylene
Chloro-
methane
Design,
ppb
17,000
8,000
500
500
300,000
200,000
300
700,000
200,000
100
Water entering stripper Water leaving stripper Weight remov-
ppb kg/h kmol/h ppb kg/h kmol/h cy, %
ND
22 0.0005 5.05 x 10"5 5 0.0001 1.01 x 10"6 77.27
89 0.0020 2.06 x 10"5 5 0.0001 1.03 X 10"6 94.38
ND
1,440 0.0327 0.0002 5 0.0001 7.52 x 10"7 99.65
8,220 0.1867 0.0014 5 0.0001 7.63 x 10"7 99.94
159 0.0036 0.0001 1 2.27 x 10"5 3.60 x 10"7 99.37
8,170 0.1856 0.0022 30 0.0007 8.24 x 10"6 99.63
ND
ND
(continued)

-------
   TABLE A-3 (Continued)
GO

Site Chemical Desigr
number contaminant ppb
Tetrachloro-
ethylene 1,800
Toluene 14,000
• 1,2-Di-
chloro-
benzene
Hexachloro-
butadiene
Hexachloro-
ethane
Isobutanol
Methyl ethyl
ketone
Nitrobenzene
1,1,2-Tri-
chloro-
ethane
Total VOC
Water entering
j
ppb
378
551
11
250
250
10
1,480
250
15
22,154
kg/h
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0086
0125
0002
0057
0057
0002
0336
0057
0003
5032
stripper
kmol/h
0.0001
0.0001
1.36 x 10"6
2.18 x 10'5
2.41 x 10"5
2.70 x 10"6
0.0005
4.63 x 10"5
2.26 x 10'6
0.0049
Water leaving
ppb
5
5
5
10
10
10
440
10
5
566

0
0
0
0
0
0
0
0
0
0
kg/h
.0001
.0001
.0001
.0002
.0002
.0002
.0100
.0002
.0001
.0126
stripper We
al
kmol/h
6
1
6
7
3
2

1
7
1
.02
.09
.80
.66
.35
.70
0.
.63
.32
.43
x 10"7
x 10'6
x 10"7
x 10"7
x 10'6
x 10'6
0100
x 10"6
x 10~7
x 10"3
sight remov-
efficien-
cy, %
98.68
99.09
54.55
96.00
96.00
0
70.27
96.00
66.67
,97.45
             Methyl  ethyl
               ketone
15,000  15,000   0.3407
0.0047    239
0.0049
0.0001
98.41
   (continued)

-------
TABLE A-3 (Continued)

Site Chemical
number contaminant
7 Isopropyl
alcohol
Acetone
Toluene
Dichloro-
methane
1,1,1-Tri-
chloro-
ethane
Trichloro-
ethylene
Chloroform
Total VOC
Design
ppb

36,000
36,000
22,000

8,300


430

740
1,200

Water
ppb

532
472.7
14,884

236.5


1,340

1,017
469
18,951.2
entering stripper
kg/h

0.
0.
1.

0.


0.

0.
0.
1.

0362
0322
0140

0161


0913

0693
0320
2911
kmol/h

0
0
0

0


0

0
0
0

.0006
.0006
.0110

.0002


.0007

.0006
.0003
.0138
Water leaving stripper W<
,1
ppb

26.8
42.7
20.7

16.0


7.8

3.5
4.9
122.4
kg/h

0.0017
0.0026
0.0013

0.0010


0.0005

0.0002
0.0003
0.0076
kmol/h

2.83 x
4.48 x
1.41 x

1.18 x


0.37 x

0.15 x
0.25 x
1.07 x
-R
10 b
io-5
io-5
c
10 5

c
10'5
c
10 5
io-5
io-4
sight remov-
1 efficien-
cy, %

95.30
91.93
99.87

93.79


99.45

99.71
99.06
99.41

-------
                                TABLE A-4.   COMPOSITION OF AIR LEAVING STRIPPER
Site number
1


Chemical compound
1,1,2 , 2-Tetrachl oroethane
trans-l,2-Dichloroethylene
Trichloroethylene
Tetrachl oroethyl ene
ppmv
0.018
0.011
0.026
0.001
kq/h
0.031
0.011
0.035
0.001
kmol/h
1.84 x 10~J
1.17 x 10":
2.71 x 10";
4.31 x 10"5
/jg/m3
125.8
44.6
142.1
4.1
                 Total  VOC
                                   0.056
               0.078
             5.77 x 10
                                                                        -4
                                                                                                 316.6
tn
Trichloroethylene
Diisopropyl ether
Methyl-tert-butyl ether
1,1-Di chloroethylene
cis-l,2-Dichloroethylene
Chloroform
1,1,1-Trichloroethane
Dichloromethane
1,1-Dichloroethane

Total VOC
0.026
 ND
0.004
0.005
0.008
0.001
0.018
 ND
0.002

0.039
0.0090

0.0010
0.0013
0.0020
0.0004
0.0064

0.0006

0.0207
                                                                                  0.0001

                                                                                1.15 x 10
                                                                                1.31 x 10
                 141.3
                                                                                  10 x 10
                                                                                  47 x 10
4.78 x 10

6.42 x 10

   0.0002
-5
-5
-5
-6
-5

-6
                  15,
                  20,
                  31,
  6.3
100.4

  9.4

324.9
                 Tetrachloroethylene
                 Trichloroethylene
                 1,2-Dichloroethylene
                 Vinyl  chloride

                 Total  VOC
                                   0.58
                                   0.10
                                   0.18
                                    ND

                                   0.86
               0.0134
               0.0019
               0.0024
               0.0177
                0.0821
                0.0152
                0.0268
                0.1241
               3,999.8
                 557.7
                 704.4
               5,261.9
                 1,1-Dichloroethane
                 1,2-Dichloroethane
                 1,1,1-Trichloroethane
                 cis-1,2-Dichloroethylene
                 1,1-Dichloroethylene
                 Trichloroethylene
                 Tetrachloroethylene
                                   0.08
                                   0.01
                                   0.07
                                   0.16
                                   0.01
                                   0.02
                                   0.12
               0.0027
               0.0005
               0.0035
               0.0056
               0.0004
               0.0009
               0.0073
               73 x 10
               10 x 10
               63 x 10
                0.0001
             4.12 x
             6.87 x
       10
       10
             4.40 x 10
-5
-6
-5

-6
-6
-5
317-. 8.
 58.9
412.0
659.2
 47.1
105.9
859.3
  (continued)

-------
  TABLE A-4  (Continued)
<£>
cn
Site number Chemical compound
Methyl ene chloride
Vinyl chloride
Total VOC
5 Benzene
Carbon tetrachloride
Chloroform
1,1-Dichloroethane
1,2-Dichloroethane
1,1 -Di chl oroethyl ene
Fl uorotri chl oromethane
1,1,1-Trichloroethane
Trichloroethylene
Vinyl chloride
Di chl oromethane
trans- 1 , 2-Di chl oroethyl ene
Chl oromethane
Tetrachl oroethyl ene
Toluene
1, 2-Di chl orobenzene
Hexachlorobutadiene
Hexachloroethane
Isobutanol
Methyl ethyl ketone
Nitrobenzene
1,1,2-Trichloroethane
Total VOC
ppmv
0.40
0.68
1.56
0.14
0
0.97

0.03
0.13
1.61
9.37
0.38
14.32

0.34
0.89 ,
4.48 x 10"3
0.14
0.15
0
2.16
0.29
0.01
30.92
kq/h
0.0120
0.0151
0.0480
0.0016
0
0.0176

0.0004
0.0019
0.0326
0.1866
0.0036
0.1849

0.0085
0.0124
0.0001
0.0055
0.0055
0
0.0236
0.0055
0.0002
0.4841
kmol/h
0.0001
0.0002
0.0005
2.05 x 10"5
0 *
1.48 x 10"H
c
4.04 x 10"c
1.96 x 10
2.45 x 10"!
1.42 x 10"J
5.71 x 10",
2.18 x 10"-3

5.12 x 10"5
1.35 x 10"5
6.80 x 10"c
2.11 x 10"c
2.32 x 10"°
0 .4
3.28 x 10 c
4.47 x 10"X
1.50 x 10"D
4.70 x 10"3
yq/m3
1,412.6
1,777.5
5,650.3
438.0
0
4,818.0

109.5
520.1
8,924.2
51,081.3
985.5
50,615.9

2,326.8
3,394.4
27.4
1,505.6
1,505.6
0
6,460.4
1,505.6
54.7
134,273,0
                  Methyl ethyl  ketone
41.25
0.3358
0.0047
123,546.7
   (continued)

-------
  TABLE A-4 (Continued)

Site number
7
Chemical compound
Isopropyl alcohol
Acetone
Toluene
Methylene chloride
1,1,1 -Tri chl oroethane
Trichloroethylene
Chloroform
ppmv
3.95
3.51
75.61
1.22
4.69
3.62
1.83
kq/h
0.0345
0.0296
1.0127
0.0151
0.0908
0.0691
0.0317
kmol/h
0.0006
0.0005
0.0110
0.0002
0.0007
0.0005
0.0003
yq/m'
9,857.1
8,457.1
289,342.9
4,314.3
25,942.9
19,742.9
9,057.1
                 Total VOC
94.43
1.2835
0.0138
366,714.3
VO

-------
                                       TABLE A-5.  AIR STRIPPER PARAMETERS
ID
O>

Air/water
Site
number
1
2
3
4
5
kg/kg
0.375
0.243
0.061
0.023
0.195
m3/m
309.
200.
50.
19.
160.
3
9
4
0
2
9
ratio
kmol/
kmol
0.234
0.151
0.038
0.014
0.121
Stripping tower design specifications
Packing
height, m
7.01
7.60
5.41
12.19
6.25
Number
of
trays
NA
NA
NA
NA
NA
Tower
diam-
eter, m
3.66
2.70
1.45
3.05
1.22

1
3
1
3
2
Description
of packing
or trays
inch Saddles
inch Tellerettes
inch Saddles
.5 inch Pall Rings
.0 inch Jaeger
Water mass
velocity,
kg/m2-s
4.165
15.306
11.404
16.729
5.362
Air mass
velocity,
kg/mz-s
1.563
3.714
0.690
0.389
1.044
            0.149   119.7
        0.093
          4.57
                               Tri-Pack

            NA      1.09     2.0 inch Jaeger
                               Tri-Pack
                                        6.531
                                            0.972
            0.062
51.4
0.038
4.87
16
1.22     Koch Type 3
16.182
0.999

-------
                                     TABLE  A-6.   SUMMARY OF ASPEN COMPARISON
VD
UD

Actual site data and performance
Site
number Chemical contaminant
1 1,1,2,2-Tetrachloroethane
trans- 1 , 2-Di chl oroethyl ene
Trichloroethylene
Tetrachl oroethyl ene
Total VOC
2 Trichloroethylene
Diisopropyl ether
Methyl -tert-butyl ether
1 , 1 -Di chl oroethyl ene
cis-1, 2-Di chl oroethyl ene
Chloroform
1,1,1 -Trichl oroethane
Dichloromethane
1,1-Dichloroethane
Total VOC
3 Tetrachl oroethyl ene
Trichloroethylene
1 , 2-Di chl oroethyl ene
Vinyl chloride
Total VOC
4 1,1-Dichloroethane
1,2-Dichloroethane
1,1, 1-Trichloroethane
ppb
40.9
14.3
44.6
0.9
100.7
28.3
ND
3.2
4.0
6.4
1.3
20.0
ND
2.0
65.2
200
30
38
ND
268
6.6
4.1
10.3
kg/h
0.033
0.011
0.035
0.001
0.080
0.0090

0.0010
0.0013
0.0020
0.0004
0.0064

0.0006
0.0207
0.0136
0.0020
0.0026

0.0182
0.0029
0.0018
0.0046
kmol/h Removal , %
1.94 x 10-4
M
1.17 x 10"T
2.71 x 10 "J
4.31 x 10"°
5.86 x 10"4
0.0001
c
1.15 x 10"i
1.31 x 10"?
2.10 x 10 "2
3.47 x 10"?
4.78 x 10"b
/"
6.42 x 10"°
0.0002
0.0821
0.0152
0.268

0.1241
2.93 x 10"r
1.82 x lO'J
3.46 x 10"D
95.0
99.99
99.99
99.99
97.50
99.99

99.99
99.99
99.99
99.99
99.99

99.99
99.99
98.50
93.33
95.59

96.01
98.53
29.27
76.70
ASPEN
predictions
Removal , %
89.17
99.98
99.98
99.98
95.59
99.96

91.56

99.98
99.96
99.25

99.98
99.29
99.42
98.67
99.67
99.88
99.37
100.00
88.85
83.06
ASPEN compari-
son ratio
Actual /ASPEN
removal
1.07
1.00
1.00
1.00
1.02
1.00

1.09

1.00
1.00
1.01

1.00
1.01
0.99
0.95
0.96

0.98
0.99
0.33
0.92
   (continued)

-------
  TABLE A-6  (Continued)
o
o
Actual site data and performance
Site
number Chemical contaminant
cis-l,2-Di chloroethane
1, 1-Dichloroethylene
Trichloroethylene
Tetrachl oroethyl ene
Dichl oromethane
Vinyl chloride
Total VOC
5 Benzene
Carbon tetrachloride
Chloroform
1,1-Dichloroethane
1,2-Dichloroethane
1 , 1 -Di chl oroethyl ene
Fl uorotri chl oromethane
1,1,1 -Tri chl oroethane
Trichloroethylene
Vinyl chloride
Dichl oromethane
trans -1,2-Di chl oroethyl ene
Chl oromethane
Tetrachl oroethyl ene
Toluene
1,2-Dichlorobenzene
Hexachlorobutadiene
Hexachloroethane
Isobutanol
Methyl ethyl ketone
ppb
15.3
1.0
2.1
16.4
41.2
34.0
131.0
73
5
781
ND
22
89
ND
1,440
8,220
159
8,170
ND
ND
378
551
11
250
250
10
1,480

0
0
0
0
0
0
0
0
0
0

0
0

0
0
0
0


0
0
0
0
0
0
0
kq/h
.0068
.0004
.0009
.0073
.0182
.0151
.0580
.0017
.0001
.0177

.0005
.0020

.0327
.1867
.0036
.1856


.0086
.0125
.0002
.0057
.0057
.0002
.0336
kmol/h Removal, %

4.
6.
4.



2.
6.


5.
2.









1.
2.
2.
2.

0.0001 ,
12 x 10"?
87 x 10"?
40 x 10"1*
0.0002
0.0002
0.0007
18 x 10"5
49 x 10"'
0.0001
£
05 x 10"?
06 x 10"°

0.0002
0.0014
0.0001
0.0022


0.0001
0.0001 f-
36 x 10"?
18 x 10"?
41 x 10"£
70 x 10"b
0.0005
83
99
99
99
66
99
82
93
0
99

77
94

99
99
99
99


98
99
54
96
96
0
70
.01
.99
.99
.99
.26
.99
.90
.15

.36

.27
.38

.65
.94
.37
.63


.68
.09
.55
.00
.00
.00
.27
ASPEN
predictions
Removal, %


99.
100.
99.
100.
97.
99.
99.
99.

99.
99.

98.
99.
99.
99.


99.
99.
99.


3.
48.


98
00
14
00
79
95
97
95

63
98

12
96
99
97


96
93
52


25
17
ASPEN
son
compari-
ratio
Actual/ASPEN
removal


1
1
0
1
0
0
0
0

0
0

1
1
0
1


0
0
0


0
1 1


.00
.00
.67
.00
.85
.93
.00
.99

.78
.99

.02
.00
.99
.00


.99
.99
.55


.00
.46
  (continued)

-------
TABLE A-6 (Continued)

Actual site data and performance
Site
number Chemical contaminant



6
7







Nitrobenzene
1 , 1 , 2-Tri chl oroethane
Total VOC
Methyl ethyl ketone
Isopropyl alcohol
Acetone
Toluene
Dichloromethane
1,1,1-Tri chl oroethane
Trichloroethylene
Chloroform
Total VOC
ppb
250
15
22,154
15,000
532
473
14,884
236
1,340
1,017
469
18,951
kg/h
0.0057
0.0003
0.5032
0.3407
0.0362
0.0322
1.0140
0.0161
0.0913
0.0693
0.0320
1.2911
kmol/h
4.63 x 10'!?
2.26 x 10"°
0.0049
0.0047
0.0006
0.0006
0.0110
0.0002
0.0007
0.0006
0.0003
0.0138
Removal , %
96.00
66.67
97.45
98.41
95.30
91.93
99.87
93.79
99.45
99.71
99.06
99.41
ASPEN
predictions
Removal , %


96.21
96.79
99.08
59.99
100.00
100.00
100.00
100.00
100.00
99.00
	 , :
ASPEN compari-
son ratio
Actual/ASPEN
removal


1.01
1.02
0.96
1.53
1.00
0.94
0.94
1.00
0.99
1.00

-------
I
I
I
I
I
I
I
I
I
I
I
                                    TECHNICAL REPORT DATA
                             (Please read Instructions on the reverse before completing)
 1. REPORT NO.
       EPA-450/1-9-002
2.
                              3. RECIPIENT'S ACCESSION NO.
4. TITLE ANO SUBTITLE
       Air/Superfund National Technical  Guidance Series
       Comparisons of Air Stripper Simulations  and Field
	Performance Data	
7. AUTHOR(S)           ~~"
                              B. REPORT DATE
                                                February 1990
                              6. PERFORMING ORGANIZATION CODE
        Gary L. Saunders

9- PERFORMING ORGANIZATION NAME AND ADDRESS
                              8. PERFORMING ORGANIZATION REPORT NO.

                                   DON 90-203-080-61-02
       PE-I Associates,  Inc.
       11499 Chester Road
       Cincinnati,  Ohio  45246
                                                             10. PROGRAM ELEMENT NO.

                                                                      61
                              11. CONTRACT/GRANT NO.

                                 68-02-4394
12. SPONSORING AGENCY NAME AND ADDRESS
       U.S. Environmental  Protection Agency
       Office of Air  Quality Planning and Standards
       Research Triangle  Park,  N.C.   27711
                              13. TYPE OF REPORT AND PERIOD COVERED
                                 Final
                              14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT

      One of the more common problems noted  at  Superfund sites is the contamination of
 ground water by volatile organic compounds  (VOCs).   One remedial alternative that is
 used to reduce or remove the VOC contamination  from  water is air stripping in a tower
 that uses either packing media or trays.  The  ability  to strip a compound from the
 water depends on several factors, including the  air/water ratio, the packing or tray
 type, and the Henry's Law value for the compounds of interest.   The objective is" to
 remove the VOCs from the water.  When being considered for remediation purposes, the
 air stripper design should be evaluated for removal  efficiency and cost of operation.
 One approach to tnis evaluation is a computerized simulation of key design parameters.
 Although numerous program approaches are available,  a  computerized process simulator
 (known as ASPEN) was used in this project to simulate  the stripping process and to
 evaluate the capital and annual costs of stripper operations.   The purpose of this
 project was to collect available design and operating  data on operating air strippers
 and to input the design and operating parameters into  the ASPEN simulator through a
 user interface program.  The results from the ASPEN  simulator were compared to the'
 operating data gathered for the sites to determine the relative accuracy of the ASPEN
 model  results when compared with the actual performance data.
J17.
KEY WORDS AND DOCUMENT ANALYSIS
^*- DESCRIPTORS
3
•
]

Air Strippers
Air Pollution
Superfund
ASPEN

r
b. IDENTIFIERS/OPEN ENDED TERMS
Air Strippers



19. SECURITY CLASS (This Report)
20. SECURITY CLASS (This page!
c. COSATI Field/Group




21. NO. OF PAGES
22. PRICE
Form 2220.1 (R«». 4-77) PMKVIOU* COITION is OMOUKTE

-------