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
------- |