CD A U.S. Environmental Protection Agency Industrial Environmental Research
f ^» Office of Research and Development Laboratory
Research Triangle Park, North Carolina 27711
EPA"600/7-78"018
SOURCE ASSESSMENT
SAMPLING SYSTEM:
DESIGN AND DEVELOPMENT
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum, interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-018
February 1978
SOURCE ASSESSMENT
SAMPLING SYSTEM:
DESIGN AND DEVELOPMENT
by
D.E. Blake
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
Contract No. 68-02-2153
Program Element No. INE623
EPA Project Officer: William B. Kuykendal
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
ACKNOWLEDGEMENT
The work described in this report was sponsored by the Process
Measurements Branch, IERL/RTP, U.S. Environmental Protection Agency. The
Technical Project Officer was Mr. William B. Kuykendal, who provided support
and guidance on numerous occasions during the conduct of this project. His
help is gratefully acknowledged.
Many other people were instrumental in the success of this program.
Mr. Jeff Kennedy and Dr. Michael Evans of Acurex contributed to the successful
completion of the project, Mr. Kennedy during SASS construction and checkout,
and Dr. Evans during the cyclone calibration efforts. Several other con-
tractors were generous with help and support, particularly Dr. Wallace Smith
and Mr. Kenneth Cushing of Southern Research Institute, and Drs. Philip Levins
and Judi Harris of Arthur D. Little, Inc.
iii
-------
TABLE OF CONTENTS
Section - Page
1 INTRODUCTION 1-1
2 PRELIMINARY DESIGN HISTORY 2-1
2.1 The High-Volume Stack Sampler 2-2
2.2 The Combustion Power Company Sampler 2-5
3 DESIGN PHILOSOPHY 3-1
3.1 Contract Requirements 3-1
3.2 EPA Guidelines 3-3
3.3 Component Design Philosophies 3-5
3.3.1 Incinerator Train Design and Construction 3-5
3.3.2 Individual Component Design 3-7
3.4 Conceptual Design Study 3-15
4 SASS DETAILED DESIGN AND CONSTRUCTION 4-1
4.1 Heated Probe 4-1
4.2 Particulate Collection System 4-6
4.2.1 The Cyclones 4-13
4.2.2 The Filter Holder 4-17
4.2.3 The Cyclone Oven 4-25
4.3 Organic Module 4-25
4.4 Impinger Assembly 4-40
4.5 Vacuum Pumps 4-46
4.6 Control Unit 4-51
4.7 System Specifications 4-59
5 CYCLONE CALIBRATIONS 5-1
5.1 Monodisperse Aerosol Cyclone Calibration Tests .... 5-3
5.1.1 Monodisperse Aerosol Cyclone Calibration Method . . . 5-4
5.1.2 Initial Monodisperse Aerosol Cyclone Calibrations . . 5-8
5.1.3 Final Monodisperse Aerosol Cyclone Calibrations . . . 5-9
5.2 Polydisperse Powder Cyclone Calibration Tests 5-11
-------
TABLE OF CONTENTS (concluded)
Section
5.2.1 Polydisperse Powder Cyclone Calibration Method. . . . 5-14
5.2.2 Initial Calibration Tests 5-ZO
5.2.3 Final SASS Cyclone Calibration Tests 5-Z7
SASS MODICICATIONS AND IMPROVEMENTS 6-1
6.1 First Field Tests 6-1
6.2 Completed Modifications 6-7
6.3 Potential SASS Modifications 6-13
APPENDIX A - CYCLONE CALIBRATION DATA ANALYSIS A-l
APPENDIX B - ISOKINETIC SASS TRAIN CONCEPTUAL DESIGN . . . . B-l
APPENDIX C - CONCEPTUAL DESIGN STUDY C-l
C. 1 Conceptual Designs C-3
C.I.I Cost Summary C-14
C.2 Schedule C-14
C.3 Discussion C-14
C.4 Recommendation C-21
APPENDIX D - CONVERSION FACTORS FOR NONMETRIC UNITS D-l
USED IN THIS REPORT
REFERENCES R-l
Vi
-------
LIST OF ILLUSTRATIONS
Figure Page
1 Acurex High-Volume Stack Sampler 2-3
2 Design schematic - SASS train 3-2
3 Conceptual cooler designs for organic module 3-10
4 Organic module functional schematic diagrams 3-14
5 Source Assessment Sampling System (SASS) 4-3
6 Schematic of SASS 4-5
7 Heated probe 4-7
8 Probe details 4-9
9 Cyclone train and filter with oven 4-11
10 SASS cyclone dimensions 4-14
11 Top view of SASS cyclones 4-15
12 10-ym cyclone assembly 4-19
13 Small and medium cyclone assembly 4-21
14 Filter housing assembly 4-23
15 Rotation of sampling probe 4-26
16 Organic module schematic 4-28
17 Organic module -exploded view 4-29
18 Organic module 4-31
19 Sorbent cartidge — original design 4-34
20 Sorbent cartridge assembly clamp 4-37
21 Redesigned sorbent cartridge 4-39
22 Impinger train out of case 4-41
vn
-------
LIST OF ILLUSTRATIONS (continued)
Figure
23 Comparison of standard impinger bottle with SASS 4-44
oversized bottle
24 Assembly drawing, impinger train 4-45
25 Impinger train 4-47
26 Vacuum pumps 4-49
27 Construction of Gast 1022 vacuum pump 4-52
28 Control unit 4-53
29 Control unit wiring schematic. 4-57
30 Typical cyclone fractional efficiency curve 5-2
31 Schematic representation of the vibrating orifice 5-5
aerosol generator
32 Collection efficiency at 400°F, 4 SCFM -SASS middle .... 5-12
cyclone. Turquoise dye particle density = 2.04 gm/cm3
33 Polydisperse powder cyclone calibration apparatus 5-15
schematic
34 Detail of polydisperse powder cyclone calibration 5-17
apparatus
35 Dust feeder. 5-18
36 Si02 test dust 5-23
37 Si02 - small dust cyclone cup catch 5-23
38 Si02 test dust - small cyclone filter catch 5-25
39 Results of tests of the constancy of the size. . . 5-29
distribution of aluminum powder passing through the
test apparatus
40 Aluminum test dust 5-31
viii
-------
LIST OF ILLUSTRATIONS (concluded)
Figure Page
41 Aluminum dust - small cyclone cup catch 5-31
42 Aluminum dust — large cyclone cup catch 5-33
43 Results of calibration of EPA SASS cyclones 5-35
44 SASS cyclone calibration data 5-38
45 First SASS test sampling location 6-2
46 Vacuum pump characteristic curves 6-9
47 Curve of distribution of feed and collected dust A-4
48 Curve of feed and collected dust (log - probability plot). . A-5
49 Comparison of cyclone collection efficiency calculated . . . A-7
from Equations (3) and (7).
50 In-stack velocity control system B-3
51 In-stack split stream probe B-5
52 Hot gas recycle B-7
53 Cool gas recycle B-10
54 Schematic — isokinetic module B-13
55 Design A system schematic C-4
56 Design B system scehmatic C-8
57 Design C system schematic C-10
58 Design D system schematic C-13
IX
-------
LIST OF TABLES
Table
1 Comparison of gas cooler concepts 3-12
2 SASS system specifications 4~60
3 SASS cyclone calibration initial tests 5-10
4 Summary of monodisperse aerosol calibrations of 5-13
middle cyclone
5 Summary of calibration results 5-36
6 Comparison -calibration results for two sets of SASS. . . . 5-39
cyclones
7 First SASS test - particulate catch 6-3
8 Particulate matter in lu cyclone - first SASS test 6-4
(concentration in ppm weight)
9 Organic fractions from Tenax extract - first SASS test . . . 6-6
10 Compound classes identified in Fraction 2 6-6
11 Size distribution of feed and collected dust A-3
12 Calculated cyclone efficiency A-8
13 Program for calculating cyclone efficiency A-10
14 Listing of input for the example problem in Figures 1. ... A-19
through 3
15 Program output for the sample problem A-20
16 Evaluation of concepts B-12
17 Design A summary C-5
18 Design B summary C-9
19 Design C summary C-ll
20 Design D summary C-15
-------
LIST OF TABLES (concluded)
Table Page
21 Summary of costs for SASS designs C-17
22 Costs to upgrade an existing HVSS to full SASS capability. . C-19
23 Total investment required to upgrade a HVSS to full SASS . . C-20
capability
XI
-------
SECTION 1
INTRODUCTION
The Process Measurements Branch, IERL/RTP, has developed a strategy
for sampling and analysis in Environmental Assessment Programs (References
1 and 2). Three levels of sampling/analysis detail are specified. Level 1
is a screening approach to measure organic and inorganic mass emissions
(within a factor of 2 to 3). Levels 2 and 3 provide quantitative and/or
continuous monitoring of specific pollutant species.
The Source Assessment Sampling System (SASS) is the primary sampling
tool for Level 1 gaseous and particulate emissions from ducted sources.
The SASS train performs the following functions:
• Extractive sampling of gaseous streams from ducts or stacks
t Measurement of particulate mass loading and size distribution
0 Collection of organic species for subsequent analysis
• Collection of vaporous trace elements for subsequent analysis
In addition to these functional requirements, the SASS train must be
portable, corrosion resistant, easily cleanable, reliable, and accurate. The
primary purpose of the project described in this report was to design and
construct a sampling train meeting all of these requirements.
During the project, there were several changes in design philosophy.
Some changes resulted from improvements of SASS train performance initiated
by project personnel, others resulted from changes in Level 1 procedures
1-1
-------
imposed by EPA. To make clear the progress of this project and the reasons
for the various changes, the report has been organized chronologically.
Section 2 describes sampler development efforts that preceded and
led to the SASS. Section 3 presents the initial SASS design philosophy
and the Conceptual Design Study that was made early in the project.
Section 4 reviews the design, construction, and initial testing of a full
SASS train based on the recommended conceptual design. Section 5 relates
the cyclone calibration history at Southern Research Institute and Acurex.
Section 6 discusses the problems reported by field test crews and recom-
mends solutions.
Other documents useful in understanding the design and operation of
the SASS include the Operating Manual (Reference 3) and the detail design
drawings*. The drawings, part of the contractually-required documentation
of the SASS development program, contain all required fabrication informa-
tion for the components designed.
*SASS detail design drawings are property of EPA/RTP, Industrial Environ-
mental Research Laboratories. Generated under EPA Contract No. 68-02-2154.
1-2
-------
SECTION 2
PRELIMINARY DESIGN HISTORY
To reach its present development, the SASS train has passed through
several evolutionary stages. The first stage was the High Volume Stack Sampler
(HVSS). This was the first Method 5 sampling tain that operated at high flowrate
and was compact, rugged, and easy to set up and use. The immediate precursor
of the SASS was the Combustion Power Company train (the CPC train), an
adaption of the HVSS to include particulate fractionation by series cyclones.
Although these earlier trains were designed and used for somewhat
different purposes than the SASS, they shared with the SASS operation at high
volumetric flowrate (that is, they were designed to operate at up to 4 scfm*,
compared to about 1 scfm for most sampling trains). In addition, many of
the components designed for these early sampling trains have been used with
only minor modification in the SASS.
In this section we review the design and method of operation of the
HVSS and CPC sampling trains, and discuss how they evolved into the SASS.
throughout this report, numerical quantities are presented in the units most
commonly employed by workers in the fields of sampling and analysis. A con-
version table is given in Appendix E to allow conversions of nonmetric units
to standard S.I. units.
2-1
-------
2.1 THE HIGH-VOLUME STACK SAMPLER
The first particulate sampling train developed by EPA personnel was
used during the middle 1960's to test municipal incinerator emissions. This
train sampled at the rate of 5 cfm. The Control Systems Laboratory (CSL)
of EPA used this type of sampling train for evaluating sources between about
1965 to 1972. During this period other EPA personnel needing to sample a
large number of sources for establishing emission levels developed a lighter,
more compact sampling train with a nominal flowrate of 0.75 cfm.
The original high volume sampling train was bulky, difficult to trans-
port and, in many respects, lacked the refinements of a comnercial product.
It was clear, however, that for certain applications -- particularly for evaluating
the effectiveness of particulate control devices — there was a need for a
high volume sampler for routine field use.
In 1972 Acurex Corporation began developing what came to be known as
the High Volume Stack Sampler (HVSS). The HVSS was to conform to the basic
requirements of Method 5, Standards of Performance for New Stationary Sources,
Federal Register Volume 36, No. 247. The HVSS is made up of a heated probe;
an absolute particulate filter housed in a temperature-controlled oven;
an impinger and gas drying module for moisture determination; a control module
for monitoring temperatures, pressures, and flowrates throughout the system;
a vacuum pump; and associated hoses and electrical lines connecting these
components. Figure 1 shows the HVSS as designed and constructed.
The specific features of the HVSS are listed below:
• Fiberglass, cushioned cases for carrying and shipping each piece
of equipment
• Oil-less vane pump modified for low leakage
2-2
-------
oo
^CONTROL UNIT
25-FOOT UMBILICAL LINE
OVEN WITH
CYCLONE
AND FILTER
IMPINGER TRAIN
AND ICE BATH
25-FOOT SAMPLE HOSE
10 CFM VACUUM PUMP
PUMP-CONTROL UNIT HOSE
Figure 1. Acurex high-volume stack sampler.
-------
t Two Magnehelic gages for accurate readout of the pitot tube range
of 0 to 4 inches of water
• Round probe body for ease in sealing the sampling port
• Probe can be rotated for sampling horizontal ducts located at same
level or below the sampling train
• Impinger train and ice bath separable from oven
t Stainless steel, Lexan, or Teflon used instead of glass -- all
glassware eliminated to avoid breakage problems
• Enlarged impinger bottles with demisters to prevent water carry-
over at flowrates up to 8 cfm
• Stable unirail traversing stand for guiding probe in horizontal
or vertical directions
• Circuit breakers instead of fuses
• Separate power lines for heaters and pump to assure availability
of required power
The control module, probe, and vacuum pump from the HVSS have been
used without modification on the SASS. The impinger assembly, particulate
collection system, and oven were extensively modified. Specific information
about SASS component design is given in Section 3. More information about
the HVSS can be found in Reference 4.
2.2 THE COMBUSTION POWER COMPANY SAMPLER
The next steps in the development of the SASS were taken in
November 1975, when Acurex began designing a special sampling train for
tests of the fluidized bed combustor at the Combustion Power Company in Menlo
Park, California. This train (which became known as the CPC train) was in
essence the "front half" of a SASS. It consisted of a probe, a three-stage
cyclone assembly and backup filter, an oven to maintain the cyclones above
2-5
-------
condensation temperatures, an impinger assembly, a vacuum pump and a control
module. The train was designed for operation at 4.0 acfm. Development of
the CPC train was funded by EPA.
Because the CPC train had to be produced on a tight schedule and limited
budget, all of the components were standard HVSS parts except the cyclone
assembly and oven. So that four discrete particulate size fractions would
be obtained, three series cyclones, with nominal DSQ cut sizes of 1, 3, and
10 ym, and a backup filter were desired.
Considerable previous work on cyclone design for sampling had been
conducted by Southern Research Institute, and their work had led to a design
for just such a cyclone train. The train had been built and some limited
testing conducted by TRW, Inc. in early 1975 (Reference 5). The cyclones had
been designed and constructed according to the standard Lapple desiqn methods
(Reference 6). The 1-um and 3-um cyclones appeared to be satisfactory for
incorporation into the CPC train, with only minor modifications. However,
the 10-um cut-size cyclone, 7 inches in diameter and 30 inches high, was much
too large to use in a portable sampling train. In addition, laboratory
experiments showed an excessive buildup of dust in the very long inlet to the
10-pm cyclone. Clearly, another cyclone design was required.
Acurex designed, constructed, and tested a "stub" cyclone of 10-ym
cut size, but much smaller dimensions, based on the work of Andrew McFarlane
at the University of Texas (Reference 7). The cyclone was 7 inches in diameter
and 10 inches high, including a large dust collection cup. It achieved the
10 urn cut size in such a small package by controlling the cyclone shape and
incorporating vanes to reduce the internal vortices.
This much smaller cyclone allowed the design of a portable three-stage
series cyclone assembly and backup filter in an oven of manageable size and
2-6
-------
weight. The cyclones and filter were closely packed together, and connecting
tubing was designed so that the cyclones and filter could be fitted into an
oven only about 50 percent larger than the standard HVSS oven. These
cyclones, filter, and oven became the basis for the CPC train design, and
ultimately, the SASS particulate handling system.
2-7
-------
SECTION 3
DESIGN PHILOSOPHY
The basic SASS design philosophy was established by EPA prior to the
initiation of our program on March 3, 1976. Figure 2 shows the design
schematic for the complete SASS train. This basic design concept has been
adhered to throughout the course of the SASS development program.
3.1 CONTRACT REQUIREMENTS
Acurex was awarded a contract to design, fabricate, and deliver three
sampling systems for use in environmental assessment studies of industrial
and energy processes. These systems were to be designed to meet the fol-
lowing requirements;
1. Extract a sample from a process stream at 3 to 5 cfm
2. Separate the particulate fraction of the sample into four size
ranges
a. >10 micrometers
b. 3 to 10 micrometers
c. 1 to 3 micrometers
d. <1 micrometer
3. The organic fraction shall be adsorbed on a Tenax porous polymer
adsorber. Requirements for the adsorber are as follows:
a. Operating temperature -- not to exceed 60° C
3-1
-------
ORGANIC
COLLECTION
MODULE
GO
ro
HEATED
PROBE
HEATED OVEN
WITH 3 CYCLONES
AND FILTER
i
TRACE
ELEMENT
COLLECTOR
GAS
PUMP
SYSTEM
CONTROL
MODULE
Figure 2. Design schematic - SASS train.
-------
b. Bed depth — 70 mm minimum
c. Face velocity -- not to exceed 1.65 ft sec
4. The inorganic fraction shall be collected in scrubbing impingers
3.2 EPA GUIDELINES
At the beginning of the development program EPA suggested guidelines
for the detailed design phase. These in included:
1. The cost of the completed SASS train was to be kept as low as
possible. A target figure of $17,000 for a barebones, but
complete, SASS train was suggested. An additional cost of not
more than $3000 was allowable for an automatic control feature in
the event that such a feature was judged desirable.
2. The SASS design was to be made as interchangeable as possible with
the High Volume Stack Sampler (HVSS). This was felt to be
desirable because many of the potential users of the SASS train
already owned an HVSS. In some cases it might be possible to
upgrade an HVSS to a SASS, cutting costs considerably to users.
3. Cyclones, rather than a stage impactor, were specified as the
method for determining particulate size distribution. The primary
reasons for choosing cyclones was the desire to collect large
particulate samples (~1 gram) for subsequent chemical and bio-
logical analysis, and the requirement for trouble-free field use.
It was considered necessary to maintain a constant sampling flow-
rate through the cyclone assembly, since cyclone size cut varies
with gas flowrate. However, if the cyclone flowrate is held
constant, then isokinetic flow generally cannot be maintained at
3-3
-------
the nozzle for the train configuration shown in Figure 2, unless
special measures are taken. (Several ways of simultaneously
obtaining constant cyclone flow and isokinetic nozzle flow are
discussed in Appendix B.)
4. The probe and oven were to be heated to eliminate the possibility
of 503 or organics condensing.
5. The organic sorbent material in the organic module was to be kept
dry and held at 60°C.
6. A flowrate of 4.0 acfm was to be maintained at the cyclones.
7. The only acceptable materials of construction for the SASS parts
that would contact the sample stream were to be Type 316 Stainless
steel, fully-fluorinated Teflon, or Pyrex glass.
8. The SASS was to be designed for ease of sample recovery and post-
test cleanup.
There were several problems with these guidelines. Item 1 (low
cost) conflicts directly with Items 2 (SASS to be interchangeable with HVSS)
and 8 (limiting the permissible materials of construction). To balance these
and other conflicting requirements, a Conceptual Design Study was conducted
in which four possible SASS designs were compared for technical feasibility
and cost. This design study is discussed in Section 3.4.
Two of the guidelines were changed later in the SASS development
program. The sorbent material tempertature (Item 5) was lowered and the
sorbent was allowed to contact the condensate. Also, the flowrate at the
cyclones (Item 6) was later changed from 4.0 acfm to 4.0 scfm. The reasons
for these and several other system changes are discussed in Section 5.
3-4
-------
3.3 COMPONENT DESIGN PHILOSOPHIES
The initial design philosophy for each SASS system component was
established at a meeting at Acurex on March 10, 1976. Representatives of
Acurex, EPA, and TRW, Inc. (the inorganic analysis contractor for the
task level of effort program) attended. The meeting was to establish the
SASS basic design parameters and to specify the requirements for two
"incinerator ship" sampling trains.
3.3.1 Incinerator Train Design and Construction
The incinerator ship trains were to be partial SASS trains for TRW,
Inc. to use in assessing the combustion efficiency of the incinerator ship
Vulcanus. Because of the inflexible sailing date of the ship, it was con-
sidered essential that the trains be delivered to TRW no later than June 1,
1976.
Since the ship was to be burning a liquid chlorocarbon, it was
anticipated that particulate would be almost nonexistant, and that organics
would constitute the primary pollutant of interest. Accordingly, the organic
module would have to be designed, constructed, and tested, but the cyclone
assembly would not be required. The remaining components -- impinger
assembly, control module, vacuum pump, umbilical, hoses — would duplicate
standard HVSS components to the greatest extent possible. TRW was to supply
a special water-cooled probe to withstand the anticipated high incineration
temperatures.
The primary task was the design and construction of the organic
module. Several conceptual designs were considered, and a thin-film heat
exchanger concept was chosen for development. The sample gas was to be
cooled to 55°C, condensate removed, then the gas was to be reheated to 60°C
3-5
-------
and passed through a porous polymer sorbent bed. The detailed design of an
organic module was essentially completed on March 25. At that time, EPA
directed that the operating temperature of the sorbent be changed to 20°C
and any condensate formed allowed to pass through the sorbent. These changes
were required due to information on organics retention developed by Arthur
D. Little, Inc.
Based on these new requirements, a modified organic module was designed,
and two complete incinerator ship trains were constructed and assembled at
Acurex in April and May 1976. The incinerator ship trains were delivered
to TRW on June 1, 1976, as originally scheduled. The actual sampling of the
incinerator ship Vulcanus took place on March 5 to 13, 1977. The materials
being burned -- waste and byproducts from the manufacture of alkylchlorides,
dichloroethane, vinyl chloride, and epichlorohydrin -- could be classified
as primary low molecular weight chlorinated aliphatics. Sampling was successfully
accomplished (Reference 8), although the high hydrochloric acid content of
the sampled gases caused severe corrosion of some SASS components. Corrosion
was observed when temperatures were reduced to a point where condensation of
water vapor occurred. The resulting hydrochloric acid produced etching and
pitting of the Type 316 stainless steel in the organic module and in the
impinger assembly.
Because of the very demanding schedule required to design, construct,
test, and ship the incinerator trains in 2 1/2 months, the initial component
designs had to be carried through to fabrication. There was simply no time
for design changes or exploration of parallel design concepts. In general,
this approach produced adequate component designs that became part of the
present SASS train. However, in a few cases (discussed in Section 6), some
later modifications were made.
3-6
-------
3-3.2 Individual Component Design
The individual component design philosophies, established at the March
1976 meeting, are described below.
Probe
A standard Acurex heated probe was selected for the SASS train.
This probe was designed for the HVSS system, and had demonstrated good
performance at high (2 to 8 acfm) volumetric flowrates. It was decided to
use a 316 stainless steel liner rather than glass or quartz, primarily on the
grounds of simpler construction and greatly increased durability and
reliability.
Oven
Because of the size of the cyclones and filter, the oven is the
largest single SASS train component. It is essential, therefore, to minimize
its size and weight. The oven previously designed for the CPC train was as
small as practicable, given the fixed cyclone size. For the SASS train, the
weight was reduced by making the outer shell of the oven out of aluminum. The
inner shell was made of stainless steel for corrosion resistance.
Cyclones and Filter
The cyclone assembly designed for the CPC trains was basically
satisfactory, although some modifications were required. The CPC cyclones
were fitted together with screwed flanges; it was deemed desirable to replace
these with quick disconnect aircraft clamps so the cyclone assembly could be
rapidly disassembled and cleaned in the field. Also, the CPC cyclones were
interconnected in a very convoluted way, making assembly and disassembly
difficult. Because some sharp bends in the interconnecting tubing were
3-7
-------
points of possible erosion or dust buildup, it was felt to be desirable to
reconfigure the interconnecting tubing to eliminate these problems.
The filter holder used in both the HVSS and CPC trains had performed
satisfactorily in previous field tests. It was used in the SASS with only
minimal changes: the materials of construction were changed from 304 to 316
stainless steel, a Teflon 0-ring was used rather than Viton, and the Teflon
coating was eliminated. The Teflon coating was considered unacceptable
because so-called Teflon coatings usually contain impurities to improve
adhesion and film-forming.
Organic Module
The organic module — the part of the SASS train for capturing
the organic pollutant species -- was a completely new component with no prior
design or field history. Accordingly, there was considerable discussion of
its form and function. The organics were to be captured by cooling the hot
gas stream from the oven and passing the gas through a dry porous polymer bed
of Tenax gas-chromatographic resin (although other sorbent materials were
under consideration). The organic module was to:
• Cool the gas from the oven (at 205°C to 55°C)
• Collect any condensate formed
• Reheat the gas back to 60°C and pass ft through the sorbent bed
• Maintain the sorbent bed at 60°C + 2°C
• Be easy to disassemble, clean, and reassemble
t Have a low gas pressure drop
• Be constructed entirely of 316 stainless steel, Teflon, or Pyrex
These requirements dictated a two part design: the first part would
cool the gas and collect any condensate; the second part would reheat the gas
3-8
-------
slightly and then pass it through the sorbent bed. Designing the first part
-- the gas cooling and condensate collection system ~ appeared to be more
difficult.
Three separate conceptual designs were considered for the cooling
section: an externally-cooled coil, a parallel tube heat exchanger (with the
sample gas passing down the inside of the tubes and cooling fluid outside),
and a thin film heat exchanger in which a thin film of the sample gas passes
between cooled walls. As each design was considered in detail, it became
clear that the two most difficult constraints were the requirement to cool
the gas from 205°C to 55°C, and the requirement for easy disassembly,
cleaning, and reassembly. Both of these constraints favored a design with a
large surface area composed of smooth, easily-accessed surfaces.- The thin
film heat exchanger concept was judged clearly superior in these areas.
Figure 3 shows schematically each of the gas cooler concepts; Table 1
lists the advantages and disadvantages of each.
The second part of the organic module is the section where organics
are sorbed on a porous polymer sorbent. In setting the design philosophy, it
was agreed to be essential to package the sorbent in a replaceable cartridge
which could be removed for analysis, while testing proceeded with a clean
replacement cartridge. The conceptual design of the second part of the
organic module will not be discussed in detail, because about 6 weeks
after the start of the project, the operating parameters for the organic
module were changed by EPA. The key changes were a reduction in sorbent
operating temperature from 60°C to 20°C, and a requirement that condensate
pass through the sorbent bed along with the gas. (These changes are
discussed in more detail in Section 6.) The effect on the organic
3-9
-------
COOLING
WATER IN
COOLING
WATER OUT
COOLED
GAS
a. Coiled - tube cooler
HOT
GAS
COOLING
WATER IN
COOLING
WATER OUT
COOLED
GAS
b. Parallel - tube cooler
Figure 3. Conceptual cooler designs for organic module,
3-10
-------
HOT
GAS
IN
COOLING
WATER
IN
COOL
GAS
OUT
COOLING
WATER
OUT
c. Thin - film cooler.
Figure 3. Concluded.
3-11
-------
TABLE 1. COMPARISON OF GAS COOLER CONCEPTS
Design
Coil
Parallel
Tubes
Thin Film
Advantages
Easy to construct. Relatively
inexpensive. Low gas pressure
drop. Capable of adequate
cooling with metallic
construction, but not glass.
Low gas pressure drop.
Capable of adequate cooling
with metallic construction,
but not with glass. Compact.
Cleaning very easy. All surfaces
accessible. Low pressure drop.
Small, compact. Capable of
adequate cooling with either
metallic or glass construction.
Disadvantages
Very difficult to
clean. Must be large
and bulky to achieve
necessary heat
transistor.
Moderately difficult
to clean. Complex
design. Relatively
costly.
Somewhat more expensive
than the coil. Cheaper
than the parallel tube
design.
3-12
-------
module design was to allow a combination of the gas cooling and sorption
functions into a simpler unit. Figure 4 shows schematically the organic
module functions at the start of the project, and after the change.
Impinger Assembly
The purpose of the impinger assembly in the SASS train is to collect
trace elements remaining in the gas stream. A second purpose is to dry the
sample gas to avoid damaging the vacuum pumps and control module. The
standard HVSS four-bottle impinger trains provide the basis for the SASS
impinger design, with one important modification. Because of the relatively
high pressure drop in several of the SASS components, the impinger assembly
operates at a substantial vacuum (10 to 20 inches Hg). The actual
volumetric flowrate in the impingers is as high as 12 acfm, leading to
excessive splashing of the impinger solutions and possible solution carryover.
In order to eliminate this problem, special oversized glass impinger bottles
are required.
Gas Flow and Temperature Monitoring and Control
The standard HVSS vacuum pump and control module were used for the
SASS. The HVSS control module, being designed for the requirements of Method
5, is more complex than is required for the SASS. For example, the dry gas
meter is not required. Also, because of the semiquantitative nature of the
mechanical SASS, mechanical dial thermometers might have been used instead
of the thermocouple readouts present on the HVSS.
An analysis of the cost of the existing HVSS control module
compared to the cost of redesigning a simpler control system was needed to
make a final decision, along with consideration of such intangible factors
as convenience and the desirability of HVSS-SASS component interchangeability.
3-13
-------
HOT
GAS 205°C
FROM
OVEN
GAS
COOLER
1
55°C
CONDENSATE
COLLECTION
ORGANIC
SORPTION
GAS
REHEAT
COOL
GAS
'TO
IMPINGERS
a. Original organic module function.
GAS 205°c
FROM
OVEN
GAS
COOLER
20° C
ORGANIC
SORPTION
20° C
COOL
GAS TO
IMPINGER
CONDENSATE
COLLECTION
b. Revised organic module function.
Figure 4. Organic module functional schematic diagrams,
3-14
-------
3.4 CONCEPTUAL DESIGN STUDY
A wide range of hardware configurations could be provided within the
constraints imposed by the basic design philosophy. An early task in the
SASS development program was the preparation of a Conceptual Design Study.
The purpose of the study was to postulate several SASS designs covering a
range of costs and capabilities, and then do a detailed cost-benefit analysis
on each. One of the designs was then recommended to EPA, who made the final
decision.
The Conceptual Design Study was submitted to EPA as an internal study
report in April 1976. Since the study was not available in the open
literature, it has been reproduced in Appendix C. It should be remembered
that costs are in early 1976 dollars.
On the basis of the Conceptual Design Study, EPA concurred with the
recommendation and accepted Option B as the basis for SASS train development.
Upon receipt of EPA approval, work was immediately begun on the formal design,
construction, and testing of the SASS.
3-15
-------
SECTION 4
SASS DETAILED DESIGN AND CONSTRUCTION
Official notice was received from EPA on July 19, 1976, to proceed
with the detailed design, construction, and testing of three SASS units based
on Design B of the Conceptual Design Study. A design effort was immediately
begun which led to construction of the first SASS trains. Although some
changes and modifications were made subsequently, the first trains are
substantially the same as current models.
Figure 5 shows a photograph of a SASS, and Figure 6 presents a
schematic diagram. These figures show a current (mid-1977) SASS, but could
equally well represent the first SASS trains, although the first trains
had one rather than two vacuum pumps (SASS changes and modifications are
discussed in Section 6). The operation of the SASS is shown in Figure 6.
This figure shows the progression of the sample gas through the extractive
probe, the cyclones and filter, the organic module, the impinger/trace element
collector, the vacuum pump(s), and the control module.
In this section we discuss the design of each SASS component.
4.1 HEATED PROBE
The SASS probe extracts gas/particulate samples from the source being
tested, monitors the temperature and gas velocity of the source, and maintains
sample temperatures above the condensation point of water/S03 mixtures.
Since the SASS is designed to operate at a sampling rate four to five
4-1
-------
Control Module
Trace Element
Impingers
Oven With Cyclones
Gas Cooler &
Organic Module
Sampling Probe
Compressors
Figure 5. Source assessment sampling system (SASS)
-------
til
Convection
oven
Filter
Stack T.C.
^ 4-UJ
Gas cooler
S-type pi tot
Gas
;emperature
T.C.
Stack velocity (AP)
magnehelic gauges
Sorbent
cartridge-
Gas
meter
Condensate
collector
Imp/cooler
trace element
collector
Coarse adjustment
0' Fine valve
s adjustment
[<( val ve £p>'
Orifice AH1
magnehelic gauge
-------
times higher than conventional Method 5 sampling equipment, the SASS probe
must be adaptable to this high flowrate.
The HVSS Probe met all of these requirements and was selected for
the SASS. Figure 7 shows the probe disassembled, while Figure 8 shows the
internal arrangement of the assembled probe. The important features of the
probe are the Type 316 stainless steel sampling tube; the fiberglass-insulated
strip heater (incorporating a thermocouple for feedback temperature control)
wrapped around the sampling tube; a round probe body to allow sealing of
the sampling port and rotation of the probe as necessary; strain relief for
all electrical, thermocouple, and pitot line connections; a calibrated S-type
pi tot; and easily interchangeable probe tips with diameters from 1/4 to 3/4
inch in 1/16-inch increments as standard equipment. The probe is designed
to withstand duct temperatures of up to 600°F. Standard probe lengths are
3, 5, and 10 feet; longer or intermediate lengths can be fabricated as required,
Water cooled probes compatible with the SASS have been constructed and tested.
They should be used for stack temperatures exceeding about 500°F.
4.2 PARTICULATE COLLECTION SYSTEM
The design, construction, testing, and calibration of the particulate
collection system (the three cyclones, the filter holder, and the cyclone
oven) has been a significant part of the SASS development program. The pur-
pose of the particulate collection system is to maintain the sample gas
stream at 400°F while collecting the particulate in three cyclones and a
backup absolute filter. Figure 9 shows the particle collection system
built for the first SASS unit.
4-6
-------
Rear Cap
1
Front Cap
Temp.
^_ -_
Probe
Liner
- _ *,, „
,. , ._ _ •_
Heater Power Plug
Sampling ]
Nozzle-1
-Heating Tape
Stack Temperature
T.C.
S-Type Pitot Tube-
Figure 7. Heated probe.
-------
HEATING TAPE
PROBE
LINER
PITOT TUBE
CONNECTIONS
PROBE TEMP
T.C. PLUG
(BLUE)
HEATER
POWER
PLUG
INTAKE
NOZZLE
L- STACK TEMP T.C.
STACK TEMPERATURE END VIEW
T.C. PLUG (RED)
Figure 8. Probe details.
-------
11
Figure 9. Cyclone trail, and filter with oven.
-------
4-2.1 The Cyclones
The SASS cyclone assembly is patterned closely on the cyclone design
developed by Aerotherm and Southern Research Institute for the Combustion
Power Sampler (see Section 2). Some changes from that design were made to
increase the ease of working with the cyclones in the field, but the size
and performance of the cyclones were essentially unchanged. Figure 10
schematically illustrates the three SASS cyclones and shows key dimensions
of each.
In keeping with the basic design philosophy, the cyclone assembly was
to be designed entirely of 316 stainless steel and Teflon. It was to be
lightweight and compact for easy field use, and easily assembled, disassembled,
and cleaned in the field. These requirements imposed limitations on the
physical arrangement of the cyclones and their method of construction.
After considering several possible cyclone fabrication methods -- rolling
and welding of flat platestock, spinning maching from billets, and deep-
drawing from sheet stock -- spinning was chosen. Spinning is the best way
to produce cyclones with thin walls for light weight with good dimensional
tolerance and durability. The cyclone bodies are manufactured from Type
316 tubestock by rapidly spinning the tube axially while forcing the tube
into a mold using a stationary hot mandrel. Top and bottom flanges and
inlet and outlet tubes are machined and welded to the spun bodies. All
welding is performed in an inert helium atmosphere with a 316 welding rod,
and welded parts are later vacuum annealed to eliminate retained strain
in the weld areas.
To reduce the overall size of the cyclone assembly, the cyclones and
backup filter must be closely packed. Figure 11 shows a top view of the
4-13
-------
Nominal cut size 10 vm
Dimensions
Cyclone
10 (im
3fjm
1 fim
Total Height
A, cm
25.0
27.1
16.5
Body Height
B, cm
13.2
15.2
7.6
Body Diameter
C, cm
15.2
8.1
3.7
Inlet ID
D, cm
5.1
1.6
0.9
Figure 10. SASS cyclone dimensions.
-------
I
en
Figure 11. Top view of SASS cyclones.
-------
first SASS cyclone assembly. Considerable care was taken to arrange the
connecting tubing so the cyclones were nestled together closely, yet with
minimum lengths of connecting tubing between cyclones, and no sharp bends
in the tubing. Minimizing tubing length and bending reduces particulate
deposition in the tubes.
Figure 12 and 13 show details of the individual cyclones. Several
points may be noted. Comparing Figures 11 and 12, it is seen that the
method of attaching the top flange to the body of the large cyclone has been
changed. In Figure 11, the top flange is attached by screws, while in
Figure 12, the flange is attached with an aircraft clamp. The screw attach-
ment method was used only in the first prototype cyclone assembly. Clamps
were used in subsequent SASS units and have proven to be convenient and
reliable for field use.
The sealing gasket and 0-ring shown in Figures 12 and 13 are white
Teflon. These Teflon seals proved to be troublesome because of their noncompliance
and tendancy to cold-flow, leading to difficulty in getting satisfactory leak
tests. As described in Section 6, Viton seals were specified for the SASS
in June 1977.
4.2.2 The Filter Holder
The filter holder houses and supports an absolute backup filter for
the series cyclones. Since the filter is likely to require frequent replace-
ment when sampling in streams with high particulate loading (1 grain/standard
cubic foot or higher), it is desirable that the filter be easily changed
during a test. The HVSS filter holder met these requirements, except
for the materials for construction (it is made of Type 304 stainless
steel). Figure 14 shows a SASS filter housing assembly; it is dimensionally
4-17
-------
10-um cyclone
10-um cyclone breakdown
Figure 12. 10-um cyclone assembly.
-------
3-ym cyclone assembly
1-ym cyclone assembly
Figure 13. Small and medium cyclone assembly.
-------
Figure 14. Filter housing assembly.
-------
equivalent to the HVSS filter housing, but is machined from type 316 stain-
less steel. The figure shows Teflon seals; Viton-A seals were later substituted,
4.2.3 The Cyclone Oven
The oven provides a constant temperature environment for the cyclones
and filter, as well as mechanical protection. It also supports the probe
by means of a collar attached to the side of the oven, which securely clamps
the probe. The probe and oven collar are so designed that the probe can be
rotated to any angular position* as shown in Figure 15.
The SASS cyclone oven is a sturdy, double-walled box with a stainless
steel liner, 2 inches of fiberglass insulation and a lightweight aluminum
outer shell. The exterior of the oven is safe to touch even when the interior
has been heated to 400°F. Figures 5 and 9 show the configuration of the
cyclone oven.
The maximum power dissipation of the single-sheathed heating element
is 1200 watts, sufficient to bring the oven to 400°F in less than 15 minutes.
A fan circulates the heat to the cyclones and filter, so that these too are
heated within 15 minutes. The fan may also be used to hasten cooling when the
heater is off and the oven door is left open. The temperature in the oven
during a test is normally maintained at 400° +_ 5°F by means of a temperature
controller located in the control module. A temperature of 500°F can easily
be achieved but, to protect the silicone rubber gaskets on the door jamb,
/
it is recommended that the oven not be operated beyond 450°F. Level 1 proce-
dure requires the oven temperature to be maintained at 400°F.
4.3 ORGANIC MODULE
The sample gas leaves the filter holder at 400°F, cleaned of particulate
but still containing any organic vapors or trace element vapors that are
present. The organic module cools the gas stream, and passes the cooled gas
4-25
-------
1N3
CTl
PROBE
ROTATABLE
THROUGH
360°
Figure 15. Rotation of sampling probe.
-------
and any condensate through an absorbent bed. The bed will collect organic
species of less than some characteristic volatility, as well as some fraction
of the metallic trace elements present.
As discussed in Section 3, several design concepts were considered
for the organic module. The concept finally chosen uses a thin-film heat
exchanger, vertically oriented, with the sorbent bed located at the base
of the cooling section. Precise temperature control of the sorbent bed was
deemed essential for reproducible organic species collection. Accordingly,
the organic module temperature control system is designed with the ability
to either heat or cool the sample gas stream to maintain the sorbent bed
at its operating temperature of 20°C +_ 1°C.
Figure 16 is a schematic illustration of the organic module. The
left half of the apparatus shown in Figure 16 is the operating part, with
the right side devoted to temperature maintenance and control. Figure 17
shows an exploded view of the left (active) side of organic module. Figure
18 shows a photograph of an organic module. Referring to Figure 17,
the hot sample gas from the cyclone oven enters the organic module at top
left. The hot gas passes down the thin (about 0.040-inch gap) annular space
between the gas cooler inner wall and the gas cooler outer wall. Both the
inner and outer walls are actively cooled (the cutaway view in Figure 16
clearly shows the coolant liquid flow passages).
The inner wall of the gas cooler section is cooled by the continuous
flow of cold (~40°F) water from a pump located in the ice water-filled impinger
assembly case. The inner wall provides about 70 percent of the cooling necessary
to reduce the sample gas stream from 400°F to 68°F (20°C). The remaining
cooling is provided the water-jacketed outer cooler wall. This outer water
4-27
-------
I
ro
CO
Return Flow
Impinger Bath
Cold Water From
Impinger Bath
Hot Gas
From Oven
Liquid Passage
Gas Passage
Gas Cooler
Sorbent
Cartridge
Section
Condensate
Reservoir
Section
3-Way Solenoid Valve
To Heat Exchanger
In Impinger Case
From Heat Exchanger
In Impinger Case
Cooling Fluid
Reservoir
Immersion
Heater
Liquid Pump
Temperature
Controller
Figure 16. Organic module schematic.
-------
Ice Water In
Ice Water Return
Hot Gas In
Gas Cooler
Inner Wall
Coolant Fluid Out
Coolant Fluid In
Sorbent Cartridge
-Sorbent Cartridge Holder
=3—^Cool Gas Out
3
t—^Condensate Out
Figure 17. Organic module -exploded view.
4-29
-------
Figure 18. Organic module.
4-31
-------
jacket is directly connected by means of a small pump to the cooling fluid
reservoir on the right side of the organic module (see Figure 16). The
temperature of the cooling fluid reservoir can be adjusted either up or down
to control the temperature of the gas cooler water jacket, and thus the sample
gas. A dual-set-point temperature controller can energize a heating element
to raise the temperature of the cooling fluid reservoir, or can switch the
three-way solenoid valve at the top of the cooling fluid reservoir to dump
heat to a heat exchanger in the impinger case (when the water jacket
temperature must be reduced). This temperature control system can maintain
the sample gas temperature just upstream of the sorbent cartridge at any
temperature between about 15°C and 80°C. As previously mentioned, the stan-
dard operating temperature of the sorbent bed for Level 1 sampling is 20°C.
As the sample gas is cooled, water or acid will usually condense.
The organic module is designed so that the cooled gas leaving the gas cooler
section, along with any condensate formed, passes through the sorbent bed.
The sorbent is typically a porous polymer gas chromatographic bed packing
material (Rohm and Haas XAD-2 resin is currently specified for Level 1 sampling)
It is contained in a sorbent cartridge, which is designed to be easily removed
and replaced with an identical clean cartridge when several SASS tests are
to be made sequentially.
Two types of sorbent cartridge are in use. Figure 19 shows the
initial design, the one most commonly used. Two circular swatches of 316
stainless steel mesh are stretched and held over the ends of an open 316
stainless steel tube by two crimp rings, also of 316 stainless. The upper
crimp ring incorporates a flange to support the sorbent cartridge and provides
a seal so that condensate and gas pass through the sorbent material inside
the cylinder.
4-33
-------
80 MESH 316SS
GO
-P-
UPPER CRIMP RING
AND SEALING FLANGE
316SS TUBE
LOWER CRIMP RING
Figure 19. Sorbent cartridge - original design.
-------
Although this sorbent cartridge design was functionally satisfactory,
some problems were encountered with removing and replacing the crimp rings in
the field. A special clamp (Figure 20) was provided to remove and reinstall
the crimp rings; however, use of the clamp was sometimes troublesome under
the difficult conditions often prevailing at the sampling site. Several
instances of broken or bent crimp rings were reported. Therefore, a modified
sorbent cartridge was made available in mid-1977. Figure 21 shows the new
design. The lower crimp ring has been replaced with a threaded cap nut that
holds the lower screen in place. This makes it simple to remove and replace
the sorbent by unscrewing the knurled cap. The top screen is still held in
place with a crimp ring, since it should not be necessary to remove the top
screen in the field. To make it easier to remove and replace the top crimp
ring, the body of the sorbent cartridge has been made of thicker material,
allowing a taper to be cut on the upper crimp ring mating surface. This
allows the upper crimp ring to be removed and reinstalled with a soft hammer.
After the gas and condensate exit the lower screen of the sorbent
cartridge, they pass into the condensate reservoir section (see Figure 16).
Condensate collects in the bottom of the reservoir and is periodically pumped
to a storage bottle; gas exits a tube at the top of the condensate reservoir
section and passes to the impinger/trace element assembly. The organic
module has been designed to be as simple and easy to use as possible. The
active part of the organic module has been constructed in three separate
sections, held together with aircraft snap-clamps. These clamps permit easy
disassembly of the unit for sample recovery, cleanup, and ready access to
any particular part of the unit. All of the parts of the organic module that
4-35
-------
JS.
I
CO
--J
Figure 20. Sorbent cartridge assembly clamp,
-------
80 Mesh 316SS
CO
70 MM
Upper Crimp Ring
And Sealing Flange
316SS Tube
Threaded Cap Nut
Figure 21. Redesigned sorbent cartridge.
-------
contact the gas stream are smooth and easily accessible for cleaning -- a
major factor in choosing the thin-film gas cooler design over other designs
such as a coiled-tube.
The basic structure of each section of the organic module is a tube
with flanges at the ends. Where available, standard sizes of 316 stainless
steel tubing have been used to minimize fabrication costs. In the gas cooler
section, the close clearance desired between the inner and outer cooling
surfaces dictated nonstandard dimensions. These surfaces are made by rolling
and welding flat sheetstock. Throughout the organic module (and the entire
SASS) tungsten-inert gas (TIG) welds using 316 welding rod are specified.
Vacuum annealing and dye penetrant testing of all welds is standard procedure.
4.4 IMPINGER ASSEMBLY
The impinger assembly immediately follows the organic module. It
collects any remaining trace elements for subsequent analysis and dries the
sample gas stream to avoid damaging the gas pumps and flow monitoring
instrumentation. The impinger assembly is pictured in Figure 22. Four
heavy wall glass bottles contain chemical solutions or moisture sorbent; 316
stainless steel and Teflon tubing directs gas flow. The first impinger bottle
(on the right in Figure 22 contains an oxidative solution of hydrogen peroxide
to collect sulfur oxides. The next two bottles contain a solution of 0.2 molar
ammonium persulfate with 0.02 molar silver nitrate to collect trace elements.
In each of these three liquid-containing bottles, a straight section of
tubing ducts the sample gas below the liquid level. The sample gas bubbles
through the liquid, allowing the various pollutant species to be scrubbed
out. The fourth impinger bottle contains granular silica gel to dry the gas.
In this bottle also, the gas is ducted to the bottom of the bottle by a
stainless steel tube and flows upward through the silica gel granules.
4-40
-------
Figure 22. Impinger train out of case.
4-41
-------
Several characteristics of the SASS led to design modifications from
a standard high-volume impinger assembly. The most important characteristic
was the large bottle volume required because of the high sample gas flowrate
and the high vacuums that can exist at the impinger. Both of these factors
tend to produce considerable splashing of the reagent liquids, with the
possibility of liquid carryover to adjacent bottles. Figure 23 compares a
standard impinger bottle (as used in the HVSS high volume Method 5
train) with the special bottle designed for the SASS. The SASS bottle,
substantially larger, is designed with a wide base and relatively narrow
neck to reduce the possibility of splashover. Also, the SASS bottle is con-
structed of heavy glass stock (about 1/4 inch thick) that is resistant to
breakage.
The Teflon caps that seal the top of the glass bottles are standard
HVSS parts, as are the gas tubes that convey the sample gas to the bottom of
the bottles. The connecting tubes between the bottles, however, are special-
ly designed for the SASS. In the initial SASS trains, the connecting tubing
was of formed (rigid) 316 stainless steel tubing. Considerable difficulty
was experienced with rigid connectors. When they were tightened down to be
leak tight, the entire impinger assembly was held in a fixed position; any
jostling or moving of the bottles tended to rotate a connector fitting, and
caused leaks.
To eliminate this problem, and to make assembly and handling of the
impinger assembly easier, flexible bellows-type connectors were installed in
most SASS trains. Figure 24 shows an assembly drawing of the impinger
train and includes some detail of the bellows connectors. Recently there
have been reports of corrosion of the stainless steel bellows with sampling
4-43
-------
3"
3"
1
14"
I
V
12"
I
1
OVERSIZED IMPINGER BOTTLE STANDARD IMPINGER BOTTLE
Figure 23. Comparison of standard impinger bottle with
SASS oversized bottle.
4-44
-------
£. ITEM -10 USSO Ok) ITCI-t -3 OKJ\_Y
4=»
on
IMPIN&EK
D 50726 7233-O65
Figure 24. Assembly drawing, impinger train.
-------
streams high in chlorides or fluorides. A Teflon bellows connector for elim-
inating any possibility of corrosion is now standard. The remaining com-
ponents of the impinger assembly (shown in Figure 22) include a thermocouple
to monitor temperature of the gas exiting the silica gel, a small pump to agi-
tate the ice/water slurry surrounding the bottles, a carrying tray so the
entire impinger assembly can be lifted out of its ice bath when required, and
a coiled tube that acts as a heat exchange surface to cool the organic module.
Figure 25 shows the assembled impinger train in its fiberglass case. The
fiberglass case serves two purposes: it acts as a shipping case for the impinger
assembly, and it holds the ice and water coolant during sampling operations.
4.5 VACUUM PUMPS
Two vacuum pumps connected in series are used with the SASS. The pumps
are identical to the pumps used with the HVSS train. These carbon vane-type
pumps (Gast Model 1022) are modified by Acurex with a special shaft seal
to reduce the leak rate to better than Method 5 standards. Each pump has
a 3/4-hp motor, a flowrate of 10 acfm at zero pressure drop, and weighs-59
Ibs including all fittings. Each pump requires 10 amps/115 VAC. Figure 26
illustrates the pumps and associated fittings, gages, valves, and hoses.
Pump features include:
• Smooth, pulse-free flow
• High vacuum capacity
• Self-lubricating carbon vanes
• Special shaft seal
• Coarse and fine flow control valves located on pump
• Carrying handle
• Operates in open air for good cooling
4-46
-------
Figure 25. Impinger train.
4-47
-------
Figure 26. Vacuum pumps
-------
• Aluminum filter and muffler jars
t Vacuum gage to indicate filter condition
• Quick disconnect fittings
Figure 27 illustrates the construction of the Gast Model 1022 vacuum pump.
4.6 CONTROL UNIT
The control unit contains all of the instruments for measuring stack
velocity, sampling flowrate and cumulative flow, and temperatures at various
points in the sampling system (Figure 28). All of the controls for the
sampling system are located in the control unit except the valves for
controlling sample flowrate. The valves are mounted on the vacuum pump,
which is placed adjacent to the control unit when using the sampling system.
Thus all of the controls and measurement displays are centered about the
control unit.
The SASS control unit is identical with the HVSS control unit, even
through some HVSS control unit components such as the dry gas meter, the
interchangeable orifice plates, and the timer are not necessary for Level
1 sampling. There are three reasons for not modifying the HVSS control unit.
First, it was desired that the SASS and HVSS sampling trains should share as
many interchangeable parts as feasible, since many users will own both trains
(discussed in Appendix C). Second, even though the cost of a SASS control unit
without unnecessary components would be lowered, the design costs for the new
control unit would largely offset any savings. And third, in the future it
may be desirable to use the SASS for some Level 2 procedures for which accurate
gas flowrate measurement would be required.
4-51
-------
Figure 27. Construction of Gast 1022 vacuum pump.
4-52
-------
c_n
OJ
115V 15 AMP
POWER INPU
POWER TO
PROBE AND
OVEN HEATERS
OVEN FAN
THERMO-
COUPLE
INPUTS
MAGNEHELICS
SHOWING AP
FOR PITOT
TUBES
PITOT INLETS
SAMPLE INLET
SAMPLE
EXHAUST
MAGNEHELIC SHOWING
AP FOR ORIFICE
DRY GAS METER DIAL
ELAPSED TIME INDICATOR
SELECTOR SWITCH KEY
AEROTHERM
ACUREX Corporation
MAIN POWER
ON/OFF SWITCH
MAIN POWER
INDICATOR LIGHT
DIGITAL TEM-
PERATURE
INDICATOR
PROBE HEATER
INDICATOR
LIGHT
PROBE HEATER
ON/OFF SWITCH
PROBE HEATER
CONTROLS
OVEN HEATER
INDICATOR LIGHT
OVEN HEATER
ON/OFF SWITCH
OVEN HEATER
CONTROLS
FAN ON/OFF SWITCH
Figure 28. Control unit.
-------
The various switches, gages, and connections seen on the face of the
control unit are described below:
Switches
There are five electrical switches with the following functions:
• Main power (with pilot light and 3-ampere, 115-VAC circuit
breaker)
• Probe heater (with pilot light and 15-ampere, 115-VAC circuit
breaker)
• Oven heater (with pilot light and 15-ampere, 115-VAC circuit
breaker)
• Fan power
• Elapsed time indicator start/stop switch
Circuit breakers are used instead of fuses to avoid the problem of
running short of fuses. The oven circulation fan is connected so that
during heating, the fan is in operation regardless of the position of
the fan control switch. When the oven heater is "off," the fan may be turned
"on" with the oven door open to hasten cooling of the oven, cyclones, and
fi1ter.
Elapsed Time Indicator
An elapsed time indicator is used to determine when to move from one
traverse point to the next, when performing Method 5 sampling. It is also
useful for SASS sampling to monitor impinger solution change intervals, data
logging intervals, and total sampling time. The indicator has a resolution
of 1/10 of a minute. The indicator can be reset to zero, and started or
stopped with a pushbutton located near the indicator.
4-55
-------
Oven and Probe Heater Temperature Controls
Power to the oven and probe heating elements is modulated with adjust-
able temperature controllers. These controllers use chrome 1-alumel
thermocouples for temperature sensing. Each controller has the following
features:
0 Actual temperature continuously displayed
• Maximum set-point is limited to 500°F by a mechanical stop
0 Power cycling is indicated by red and green lights
The oven and probe heater controls are located on the control unit
because under some conditions the oven is inaccessible for adjustment. For
example, while sampling large stacks, the oven may be located beyond the edge
of the sampling platform. Feedback control of the temperature is used
because the ambient conditions under which stack sampling is performed are
highly variable. The concern for accurate temperature control is based on
the fact that many of the effluents sampled have condensible components.
These components, such as, water and sulfur oxides, must be maintained as vapor
prior to filtering out the particulates. Also, the cyclones require constant
temperature operation.
The actual switching of power to the heating elements in the probe and
oven is done by heavy duty relays (lower right corner, Figure 29). This
greatly increases the capacity for the system to use longer probes which
require greater power.
Temperature Display
A digital temperature indicator is used together with an eight-point
selector switch. The selector switch permits monitoring the temperature at
each of the following locations:
4-56
-------
Jii
cn
m
VAC
»
115 »
VAC »__,_
»
PROBE POWER SW ISA
<•
o—cr\jo-
OVEN POWER SW ISA
I
-crxo—or\p-
POWER SWITCH
3A
POWER
LIGHT
T
O
TIMER
(^) POWER
\£/ LIGHT
'
TEMP 3
CONT -^r-
OVEN J)
O
O
5
POWER
LIGHT
(E
OVEN
RELAY
TEMP
CONT
PROBE
OVEN
HEAT
r\
PROBE
RELAY
^c<
PROBE
HEAT
' I
I
j
FAN
SWITCH
A
r
O
FAN
t
Figure 29. Control unit wiring schematic.
-------
t Stack
• Probe
• Oven
• Inipinger train outlet
a Gas meter inlet
« Gas meter outlet
e Two "spare" locations
The temperature range is 0°F to 1500°F with an accuracy of +4°F.
Gas Flow
The cumulative sample gas flow is measured by a Rockwell Model 415 gas
meter, a high accuracy meter used for testing. The measurement is displayed
by a digital counter and pointer with a resolution of 0.005 cu. ft.
Pressure Gages
Three Magnehelic pressure gages can be seen on the face of the control
unit. One is used for monitoring the pressure drop across the orifice meter
(see following discussion on orifice meter). The other two gages are connected
in parallel and indicate the pressure differential of the pitot tube used for
measuring stack velocity. One of the gages has a range from 0 to 0.5 inches
of water; the other, usually 0 to 4 inches of water. Thus the pitot tube
pressure differential can be determined with high precision over the full
range.
Umbilical Line Connections
The umbilical line between the control unit, oven, and probe makes the
connections with the control unit as follows:
• Multipoint connector with AC power leads to oven, fan, and probe
• Dual-pin thermocouple connectors for the stack, probe, and
impinger thermocouples.
4-58
-------
The separate 25-foot sample hose connects to the vacuum pumps. The
exhaust hose of the pump is conected to the "inlet" fitting located on the
control unit. The sample gas then passes through the gas and orifice meters
in the manner of the typical Method 5 sampling train.
A quick-disconnect fitting is provided at the sample "exhaust" outlet
on the control unit. A length of tubing can be connected at this point for
leading toxic sample gases away from the control unit area.
The control unit has a removable back cover to provide ease of access
for repairs and to set the three-position orifice meter (right side,
Figure 29). Three orifices are used for high accuracy measurements over
the following flow ranges:
t 0.3 to 1.3 cfm
• 1.0 to 3.0 cfm
• 2.0 to 6.0 cfm
For Level 1 sampling, the largest orifice (highest flowrate) will always
be used.
4.7 SYSTEM SPECIFICATIONS
Table 2 shows the size, weight, and electrical specifications for
the SASS components. Four 20-amp circuits or three 30-amp circuits are
recommended when using the SASS. Ice consumption for a 5-hour Level 1 test
will vary somewhat with ambient temperature, but should not exceed 175 Ib.
Average ice use is 100-150 Ib.
4-59
-------
TABLE 2. SASS SYSTEM SPECIFICATIONS
Component
Probe
Oven
Cyclones/ Filter
Organic Module
Impinger Assy
Pumps (each)
Control Module
Size
5 ft x 3 in.
22 x 15 x 22 in.
21 x 16 x 15 in.
20 x 25 x 9 in.
19 x 20 x 10 in.
16 x 18 x 9 in.
26 x 20 x 15 in.
Weight
15 Ib.
30
18
40
35
60
70
Electrical
Voltage
115
115
—
115
115
115
115
Maximum
Amperage
5.5
10
—
6
1
10
1
4-60
-------
SECTION 5
CYCLONE CALIBRATION
The characterization of the SASS cyclones has been underway almost
continuously since the development of the SASS. Initial efforts were
conducted by Southern Research Institute using a Vibrating Orifice Aerosol
Generator. Later calibration tests were performed by Acurex using
a different method involving dispersions of polydisperse aluminum spheres. At
the time of writing, results have been obtained with both methods that are
reasonably consistant and are believed to represent the actual performance of
the cyclones.
The object of the various cyclone calibration tasks ultimately is to
determine the cyclone efficiency curve; from the curve can be obtained a
commonly used figure-of-merit for the cyclone called the D50 cut diameter.
Figure 30 illustrates these concepts. The efficiency of particle collection
is plotted against the particle diameter. For each particle diameter,
therefore, the effectiveness of the cyclone is determined. For example,
Figure 30 shows that for this particular (ficticious) device, if a large
number of 2.5-um diameter particles are introduced, 17.5 percent will be
collected and 82.5 percent will pass through uncollected. The particle
diameter at which half of the particles are collected is the D5Q cut
5-1
-------
CJ1
IN3
O
z
UJ
O
HI
z
O
UJ
O
O
UJ
Z
O
O
>-
O
1.0 r
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
6.1
1
L
2 3 4 56789 10
PARTICLE AERODYNAMIC DIAMETER, nm
Figure 30. Typical cyclone fractional efficiency curve.
-------
diameter; Figure 30 shows the DSQ cut diameter of that device to be 3.0 \m.
The DSQ cut diameter, often abbreviated to "cut size," is commonly used as a
rough indication of the collection cut-off of a cyclone.
Note that Figure 30 expresses particle diameters as aerodynamic
particle diameters. It is important to distinguish aerodynamic diameters
from physical diameters. The physical diameter is the dimension of the
particle obtained by physical measurement, for example with a microscope and
reticle. For nonsymmetrical particles, the physical diameter of a given
particle may have several different values, depending on the measurement axis
chosen. The aerodynamic diameter (sometimes called the Stokes diameter) is
defined as the diameter of the equivalent spherical particle of unit specific
gravity having the same terminal settling velocity as the particle in
question. The advantages of using the aerodynamic diameter to characterize
the particles used for cyclone calibration are twofold. First, each particle
is uniquely characterized, independent of any choice of physical dimension.
Second, and most important, since the basic cyclone separation mechanism
depends on Stoke's law, measuring particle diameter in terms of Stoke's law
behavior assures that calibration data will be valid over wide ranges of
particle size, shape, and density.
5.1 MONODISPERSE AEROSOL CYCLONE CALIBRATION TESTS
The Southern Research Institute (SoRI) calibration efforts using
monodisperse aerosol occurred in two phases. Initial calibrations were made
at room temperature, with the intent of calculating cyclone D5Q cut diameters
at 400°F by use of accepted design equations. Based on these initial calcu-
lated values, certain modifications were made to the cyclones to shift their
cut points closer to the 1, 3, and 10 ym values desired. It was discovered
5-3
-------
subsequently that the design equation used to adjust the cyclone D5Q's to
400°F was inapplicable to small cyclones such as the SASS. The second phase
of calibrations at SoRI involved actual calibrations at 400°F. These efforts
resulted in calibration data (for the middle cyclone only) that are believed
to be accurate.
5.1.1 Monodisperse Aerosol Cyclone Calibration Method
The same basic procedure has been used in all of the SoRI calibration
work (Reference 9). The SASS train cyclones were calibrated using a Vibrating
Orifice Aerosol Generator (VOA6). The VOAG generated monodisperse ammonium
fluorescein particles and turquoise dye particles with diameters from 2 micro-
meters to 7 micrometers. The VOAG used in this study was designed and built
at SoRI. However, similar devices have been reported by several authors
previously, and a commercial unit is available from Thermo Systems, Inc.
Figure 31 is a schematic diagram showing the operating principle of
the VOAG. A solution of known concentration (in this case, a solution of
fluorescein (£20^12^5) ^n ®'^ NfyOH or a solution of turquoise dye in water)
is forced through a small orifice (5-, 10-, 15-, or 20-pm diameter). The
orifice plate is attached to a piezoelectric ceramic which, under electrical
stimulation, will vibrate at a known frequency. This vibration imposes
periodic perturbations on the liquid jet causing it to break into uniformly
sized droplets. Knowing the liquid flowrate and the perturbation frequency,
the droplet size can be readily calculated. The solvent is evaporated from
the droplets leaving the nonvolatile solute as a spherical residue. The
final dry particle size can be calculated from the droplet size through the
known concentration of the liquid solution.
5-4
-------
CHAMBER
CHARGE NEUTRALIZER
>
\SSDRYING
•»^^^
VIBRATING
ORIFICE -—^____^
FLOW
METERS -i^
•
h
S^
CONTROL ^*^ I
VALVES ^^
f |
\v.
»;
V
L ,
y
i
| X
/ /
;/
ir
i
1,
^
.
k
i
>
»
r
r
h
^
SIGNAL GENERATOR
^X^
^
\/ MEMBRANE
^C *" FILTER
^^^ SYRINGE
PUMP
X ABSOLUTE
FILTER
V-
(/r\> DRY AIR
Figure 31. Schematic representation of the
vibrating orifice aerosol generator.
5-5
-------
To calculate the dry particle size, the expression
dp = (QCV/10TTF) 1/3
where
Cv is the solution concentration or volume of solute/volume of solution
Q is the solution flowrate (cm^/min), and
F is the perturbation frequency (Hz)
By using smaller orifices, one can obtain much higher operating
frequencies. This in turn yields higher particle number concentrations
and allows a shorter running time to collect the same mass per stage. The
running time must be sufficiently long, however, to allow accurate determi-
nation of the cyclone collection efficiency. It was found that the 20-um
orifice was consistently easier to use in particle generation, primarily
because of fewer clogging problems.
Prior to particle generation, the orifice plates were washed in deter-
gent with ultrasonic agitation and then rinsed several times in distilled
water, also with ultrasonic agitation. After the filter and liquid handling
system was flushed several times with aerosol solution to be used, an orifice
plate was placed, still wet with distilled water or blown dry, into the crystal
holder and the syringe pump was turned on. A jet of air was played over the
orifice plate to keep the surface clean until enough pressure was built up
behind the orifice to form a jet.
After a stream of particles was generated, a determination of mono-
dispersity had to be made. Two methods were used to accomplish this. By
using a small, well-defined air jet to deflect the stream of particles, it
was possible to tell when the aerosol was mono or polydisperse. Depending
on particle size, the stream was deflected by the air at different angles.
If the aerosol was polydisperse, several streams could be seen at one
5-6
-------
time. By varying the crystal oscillation frequency, the system could be
fine-tuned to give only a single deflected particle stream, indicating
monodispersity. The particle stream was then dispersed and diluted in a
plexiglass cylinder 6 inches in diameter and 18 inches high through which
dry, clean vair flowed at 0.83 cfm. Polonium 210 alpha particle sources were
placed around the dispersed particle stream at the orifice and at the outlet
of the plexiglass cylinder around the exiting aerosol stream to act as charge
neutralizes to reduce agglomeration and loss of particles due to
electrostatic forces. A second check of monodispersity and particle size was
optical microscope examination of a filter on which particles were collected
as they exited from the outlet of the plexiglass cylinder. In general, it
was found that about 4 percent to 8 percent by mass of the particles were
twice the volume (1.26 x diameter) of the primary particles. Microscopic
photography was used as a validation of monodispersity and particle size.
When it had been determined that particles of the correct size were
being generated, a pump was started which was preset to give the correct
actual flowrate through the cyclone. The aerosol stream flowed at the rate
of 0.83 cfm from the outlet of the plexiglass mixing column, and to this heated
room air was added at a rate and temperature to give the final indicated
flowrate and temperature. The heated aerosol stream then entered the cy-
clone, which was kept in an oven and at or near the aerosol stream temper-
ature. A glass fiber filter caught all the particulate that passed the cyclone.
On several occasions, the aerosol tended to drift from monodispersity.
In order to avoid sampling a polydisperse aerosol, periodic filter samples of
the heated aerosol stream were taken and checked by optical microscopy. This
also provided a good check on the sphericity of the aerosol because the final
particles instead of the primary liquid droplets were investigated.
5-7
-------
Throughout the testing, close watch was kept on the temperature and
flowrate of the aerosol stream. Any discrepancies were quickly corrected,
and readings of all temperatures were recorded periodically to insure
repeatability of the tests and test results.
After each test, the cyclone and filter substrate were washed to
dissolve and rinse off all the aerosol particles. The wash solutions used
were 0.1N NH4OH for ammonium fluorescein and distilled water for turquoise
dye.
A Bausch and Lomb Spectronic 88 Spectrophotometer, calibrated with
solutions of known concentration of the aerosol solute (turquoise dye or
ammonium fluorescein), was used to measure the absorbance of both the wash
from the cyclone and the filter from which the concentration was determined.
From knowledge of the amount of wash solution, the dilution factor, if any,
and the absolute concentration, the mass of particles in the cyclone and on
s
the filter was calculated. With these two masses known, the collection
efficiency of the cyclone for that particular particle size was calculated.
5.1.2 Initial Monodisperse Aerosol Cyclone Calibrations
As previously described, the initial cyclone calibration tests on the
SASS cyclone were conducted at ambient temperature to calculate performance
at the 400°F SASS operating temperature. The Lapple equation (Reference 10)
predicts the variation in D,-Q cut diameter with changes in other variables:
D50 = (9n B
where
050 = cyclone 059 cut diameter
Vi = gas viscosity
p = density of particles
5-8
-------
Nc - number of turns made by the gas stream in the cyclone body
and core
Bc = width of the cyclone inlet
Vc = inlet air velocity
Table 3 shows the results of the initial SASS cyclone calibra-
tions. The D50 values are calculated at 4.0 scfm (6.5 acfin) flowrate, 400°F
temperature, and 1.0 specific gravity, although the actual measurements were
made at 4.0 acfm, 75°F, and 1.35 g/cm3 assuming the applicability of the
/
Lapple equation. In fact, during later attempts to calibrate the cyclones at
400°F, it was found that the equation does not accurately predict parametric
variation in small cyclones such as the SASS. Accordingly, it was concluded
that the calibration values shown in Table 3 are incorrect, and that
calibrations should be performed at the operating conditions of the SASS
train (4.0 scfm and 400°F). This was done during the final calibration tests
at SoRI.
5.1.3 Final Monodisperse Aerosol Cyclone Calibrations
Having decided to perform the SASS cyclone calibrations at 400°F, the
major problem to be overcome was the thermal instability of the ammonium
fluorescein dye used for previous room-temperature work. The first method
attempted was to calibrate the cyclones at 70°F, 200°F, and 350°F and then to
extrapolate the results to 400°F. This method proved to be unsatisfactory
when it was found that — contrary to expectations -- ammonium fluorescein
particles smaller than 4 un in diameter were unstable at 350°F. Attempts to
alleviate this problem were largely unsuccessful. After some futher work, it
was determined that ammonium fluorescein was not suitable for calibrations at
elevated temperature. Attention was turned to finding a different dye
material.
5-9
-------
TABLE 3. SASS CYCLONE CALIBRATION INITIAL TESTS
Cyclone
Calculated D5Q Values
at 400°F, 4 scfm, 1.0 g/cm-3
Large
Middle
Small
10.2
3.2
0.80
Several strict requirements must be met by a satisfactory dye candidate,
Most or all of the following characteristics must exist:
• Nontoxic
• Stable at temperatures up to 500°F or above
• Soluble in water or other nontoxic, nonresidue forming solvent
• Amporphous -- dries to form solid, homogeneous spheres when
disperesed in solution from a VOAG
• Known or easily measured density
• Has a definite, distinct absorption spectrum peak for absorption
spectroscopy measurement between 400 NM and 900 NM
Of several samples from three chemical companies, du Pont's
"Pontamine" Fast Turquoise 8 GLP dye was the first found to satisfactorily
meet the requirements listed above. A spectral analysis performed on a
dilute water solution of this dye indicated a distinct absorption peak at 622
namometers. Measurements with a Helium-Air pycnometer gave a density of 2.04
gm/cm3. The sample seemed pure and its stability at 400°F was excellent.
The expansion problems encountered with small diameter ammonium fluorescein
particles were absent. Aerosol particles made from a solution of the dye in
distilled water were very nearly, if not perfectly, round.
5-10
-------
When the high temperature calibrations technique was fully developed,
primary interest was in the performance of the middle cyclone. Accordingly,
only the middle cyclone was calibrated. As the first calibration tests
indicated that the D50 cut diameter was somewhat larger than the 3.0 pm
desired value, the middle cyclone was also calibrated with the vortex buster
removed. The vortex buster is a sheet metal cross that is normally placed in
the dust collection cup at the bottom of the cyclone. It breaks up the
normal swirl pattern in the cyclone, thus reducing collection efficiency and
increasing 059 cut diameter. The swirl buster had been added earlier, based
on (erroneous) estimates that the 059 was too low.
The results of the high-temperature calibration of the middle cyclone
are shown in Figure 32. Note that the physical particle diameter is used.
The 059 physical cut diamters are 3.4 and 2.5 TJTI with and without the swirl
buster. The 059 aerodynamic cut diameters are 4.9 and 3.5 ym respectively.
These are obtained by using the Lapple equation to correct to unit particle
density.
Table 4 summarizes all of the monodisperse aerosol cyclone calibration
tests of the middle cyclone. The "correct" values of the D^Q cut diameter are
the Turquoise Dye aerodynamic values, that is, 3.5 ym without the vortex buster,
and 4.9 ym with the vortex buster.
5.2 POLYDISPERSE POWDER CYCLONE CALIBRATION TESTS
The Acurex SASS cyclone calibration test series was
conducted by an entirely different method than was used by SoRI. The Acurex
method involves the dispersion of polydisperse particles, at concentrations
representative of actual field conditions. It was desirable to do a second
series of calibration tests for several reasons:
5-11
-------
100
80
u
UJ
o
LL
Ul
O
O
o
o
60
40
20
7
O/ £
TURQUOISE DYE
400°F, 4 SCFM
O
OWITHOUT VORTEX BUSTER
A WITH VORTEX BUSTER
f I I Jill
2 3 45678 910
PHYSICAL PARTICLE DIAMETER, micrometers
Figure 32. Collection efficiency at 400°F, 4 SCFM
SASS middle cyclone. Turquoise dye
particle density =2.04 gm/cm3.
5-12
-------
TABLE 4. SUMMARY OF MONODISPERSE AEROSOL CALIBRATIONS OF
MIDDLE CYCLONE
Material
Ammonium
Fluorescein
Ammonium
Fluorescein
Ammonium
Fluorescein
Turquoise
Dye
Turquoise
Dye
Turquoise
Dye
Vortex
Buster
i
IN
IN
IN
IN
OUT
IN
Temperature
Ambient
200QF
350°F
Ambient
400°F
400°F
Flow Rate
ft.Vmin
Actual/Standard
5.37/~- -
5. 4 I/
5.46/ —
5.42/
6.50/4.00
6.50/4.00
D so. Physical
Micrometers
2.8
3.5
4.2
2.2
2.5
3.4
D so Aerodynamic
Micrometers
3.3
4.0
4.9
3.1
3.5
4.9
en
i
-------
• A separate series of calibrations using a different method
would -- if the results agreed -- greatly increase confidence in
the correctness of the calibrations
• The extremely low particle mass concentrations with the SoRI cal-
ibration method required confirmation at more realistic loadings
• The physical state of the SoRI dye particles was unknown. There
was a feeling that any stickiness caused by hydroscopic absorption
might bias the results
0 At the time the A/A tests began, problems with the SoRI method at
elevated temperature made ultimate success with that method
uncertain
The Acurex calibration program was not without its share of problems
and false starts. Several different test dusts and data reduction methods
were tried before successful calibration data were obtained. Throughout the
program, however, the apparatus and basic method of conducting the calibra-
tions was unchanged.
5.2.1 Polydisperse Powder Cyclone Calibration Method
Figure 33 shows the test apparatus in schematic form. Metered
amounts of air and test dust are combined in a powder feeder. The powder
feeder is so designed that the dust particles are deagglomerated and
suspended in the air. The dust cloud then enters a heater, where its tem-
perature is raised to the desired level. The hot dust cloud then enters the
cyclone being evaluated. Each particle is either captured by the cyclone
(ending up in the cyclone cup) or exits the cyclone and is captured on
an absolute backup filter. Clean air is exhausted from the filter holder
to the room.
5-14
-------
en
on
Metered
air
Mete red
dust
Powder
feeder
Uoat-Qy.
neater
'1
k /
F
Filter
Cyclone
Figure 33. Polydisperse powder cyclone calibration apparatus schematic,
-------
Figure 34 shows the apparatus in more detail, while Figure 35 shows
a cutaway view of the dust feeder, the heart of the apparatus. As shown in
Figure 34, compressed air is filtered, regulated, and then passed through
a square-edged critical flow orifice. The flowrate through this type of orifice
depends only on upstream pressure, so long as the temperature remains constant
and the downstream absolute pressure is less than about half of the upstream
pressure. In the calibration apparatus these conditions are easily met, so
a constant volumetric flowrate of 4.00 scfm was held simply by maintaining
a constant reading on the upstream pressure gauge.
The carrier air -- clean, dry, and at constant flowrate -- now enters
the dust feeder. Figure 35 shows the operation of the feeder. The dust
metering element is the feeder disk, which has a small groove milled
coaxially on one face. The feeder motor rotates the disk continually. Test
dust is stored in the dust hopper. The hopper also contains a rotating
stirring shaft that assures that the groove in the disk (which passes
directly under the hopper) is filled with test dust. A wiper arrangement on
the outside of the dust hopper rides against the disk, and assures that the
hopper dust is contained and that the surface of the disk — except for the
groove -- is free of dust.
If the dust feeder is operating properly, the test dust will be firmly
packed into the groove. Since the disk rotates at a constant angular
velocity, the amount of dust per unit time that leaves the dust hopper will
be constant. The rate of delivery of dust thus depends on the size of the
groove and its speed of rotation. For all of the SASS cyclone calibration
tests, the groove size and rotational speed were set to deliver 0.26 grams/
minute, which corresponds to a dust cloud loading of 1 grain/standard cubic
foot.
5-16
-------
en
i
AM
MPSIO—»f—
MIN
FILTER
HOLDCft
CVCLttM
COLLECTION
COP
\
\
PRESSURE
REGULATOR
v THERMOCOUPLE
TEE
\
CONTROL PANEL
FILTER
Figure 34. Detail of polydisperse powder cyclone calibration apparatus.
-------
HOPPER MOTOR
01
CO
HOPPER
DRIVE
COUPLING
3
3 DUST HOPPER
HOPPER COVER / FEEDER AIR INLET
FEEDER DUST CLOUD
OUTLET TUBE
TRANSPARENT
FEEDER
FEEDER
HOUSING
FEEDER DISC
COUPLING
Figure 35. Dust feeder.
-------
Approximately 180° away from the dust hopper, and centered over the
groove, is the feeder outlet tube (see Figure 35). Since the feeder is
sealed, all of the air that enters through the air inlet tube must exit
through the outlet. The end of the outlet tube is positioned just above the
groove in the disk. The outlet tube is sized so the air velocity at the
entrance is about 800 ft/sec. This high velocity air cleans all of the dust
out of the groove and conveys it out of the dust feeder. The cleaned groove
then rotates back to the dust hopper, where it is refilled.
In practice, the operation of the dust feeder is quite simple.
Operator attention is required only periodically to see that the groove is
being properly filled and to refill the dust hopper when needed. The high-
velocity air at the feeder outlet tube does have a tendency to clean out the
groove about a 1/4-inch ahead of the tube and some slugging (breaking
away of the dust in the groove in chunks) is evident.. Thus there are
undoubtedly momentarily higher dust loadings than the 1 gr/scf average value.
The dust cloud velocity in the outlet tube is deliberately held at
near sonic conditions in order to assure maximum dispersion of the dust
particles. At the point the outlet tube enters the heater (see Figure 34),
a step enlargement in the diameter of the tube reduces the gas velocity
through the heater to about 400 ft/sec. The heater itself is simply a
stainless steel tube wrapped with a heating tape. The wall temperature of the
tube is maintained at about 600°F, allowing the dust cloud to reach 400°F by
the time it exits the 20-inch long heater. Feedback temperature control of
the exit dust cloud temperature is maintained by a thermocouple and
temperature controller.
The cyclone being calibrated is attached to the outlet of the heater,
and is wrapped with thermal insulation during calibration tests so the 400°F
5-19
-------
exit temperature is maintained. A standard SASS filter holder supports
a glass fiber absolute filter.
Conduct of a calibration test is straightforward. The dust hopper
is loaded with the desired test dust and a filter paper is preweighed and
installed in the filter holder. The air and the heater are turned on.
After temperatures have equilibrated, disk rotation is begun, starting the
flow of test dust. The test continues for long enough (1 to 5 hours) to
collect sufficient dust in the cyclone and on the filter for subsequent
analysis. The disk rotation is then stopped, the heater is turned off, and
after the system has cooled (about 30 minutes) the air is turned off and the
samples are removed. Dust found inside the cyclones is brushed down into the
collection cup. Dust found inside the cyclone outlet tube is added to the
filter catch.
5.2.2 Initial Calibration Tests
In the first series of experiments undertaken, a highly classified
silica dust material was used. The dust (AC Fine test dust) was classified
by the Donaldson Company into nine size fractions. This particular material
was chosen because of its availability, its wide use as a test dust in the
HVAC industry, and the familiarity of Donaldson Company with the behavior of
the material in their classifiers.
The approximate performance of each of the SASS cyclones was believed
known from previous SoRI tests. Accordingly, three to six of the classified
dusts were chosen for calibrating each cyclone, depending on the mean
particle size of the dust and the anticipated cyclone cut size. The size
distribution of the test dust and the dust collected in the cyclone cup was
measured using the X-ray sedograph at EPA/RTP.
5-20
-------
From the size distribution data, it should have been possible to
construct a cyclone efficiency versus particle size curve for the particle
size range of the test dust. When this was attempted, it became apparent
that the experimental results were inconsistent, and in some cases,
contradictory. For several experiments, for example, the mass median size of
the cyclone cup catch was smaller than the feed material; the filter catch
mass median diameter was even smaller. This result is clearly impossible
unless the size distribution measurement method is faulty or unless the test
dust is changing its characteristics during the test.
There is some evidence that the latter explanation is the cause of the
unexpected test results. Figures 36, 37, and 38 are scanning electron
micrographs of the feed, cyclone cup, and filter fractions, respectively,
from a calibration run with the small cyclone. The magnification is 3000X.
It is qualitatively apparent that the cyclone cup fraction is smaller than
the feed fraction, as indicated by the X-ray Sedograph measurements. The
most interesting point, however, is the appearance of the particles. The
test dust particles (Figure 36) are generally smooth and show cleavage
planes. The particles collected by the cyclone (Figure 37), however, are
very rough and pitted, and seem to be rounded off. The filter fraction
largely consists of very small particles that are not evident in the test
dust. All of this seems to indicate that the test dust has been eroded and
reduced in average size somewhere in the calibration apparatus. As
velocities in the dust cloud outlet tube and heater are deliberately kept
high (near sonic) to avoid reagglomeration of the dust, it is suspected that
particle-particle contact in this region is causing the erosion. The
hardness and frangibility of the test dust undoubtedly is also a major
factor.
5-21
-------
Figure 36. SiCL test dust.
Figure 37. SiCL -- small dust cyclone cup catch.
5-23
-------
Figure 38. S10? test dust -- small cyclone filter catch
5-25
-------
It was clear that to overcome the problem of test dust degradation,
either the apparatus or the test dust would have to be changed. The easiest
solution appeared to be to use a test dust not subject to erosion. After
considering several alternatives, a spherical aluminum powder was chosen for
the next series of tests.
5.2.3 Final SASS Cyclone Calibration Tests
The aluminum powder chosen is produced by several manufacturers by
atomization from the melt. Individual particles are spheroidal, a large
majority having the appearance of perfect spheres. Because the particles are
malleable, erosion or other damage during the calibration operation was
considered unlikely.
The data reduction method originally planned for use with the aluminum
powder depended on measuring the size distribution of the material entering
the cyclone, the amounts of material collected in cyclone dust cup and on the
filter, and the size distribution of either the dust cup or the filter
material. Knowing this information allowed calculation of the cyclone
efficiency at each differential particle size. Coulter counter measurement
was chosen for determining the distribution, since for the particle sizes of
interest (1 un to 20 \m) it was known to give reliable, reproducible, and
relatively inexpensive results.
It quickly became apparent that the apparent size distribution of the
test dust entering the cyclone was being changed during passage through the
dust feeder and heater section of the test apparatus. Six tests were made in
which the test apparatus was operated normally except no cyclone was attached
— all of the output of the apparatus was collected on an absolute filter.
When the size distribution of the collected aluminum powder was measured, it
varied noticeably from test to test. Aluminum powder as suppled by the
5-27
-------
manufacturer was very constant in size distribution, Figure 39 shows the
results of these tests. As the aluminum test dust passed through the test
apparatus, the size distribution gets larger — agglomeration is occurring,
small particles are somehow being removed, or both.
To overcome the problem of test-to-test variations in the test dust
entering the cyclone being tested, a different data reduction strategy was
used. The size distribution and quantity of dust collected in both the
cyclone dust cup and on the filter is measured. From this information a
simple material balance on each differential element of particle size allows
reconstruction of the distribution of the test dust entering the cyclone, no
matter how it may have been changed during passage through the dust feeder
and heater. This method of analysis was used for the final, and successful,
series of cyclone calibrations. Details of the data reduction method are
given in Appendix A.
Two complete sets of SASS cyclones were calibrated. Each set consis-
ted of three cyclones -- one small, one medium, and one large. The first set
calibrated was a part of an EPA-owned SASS train. The large and medium
cyclones were calibrated both with and without their swirl busters; thus a
total of five complete cyclone calibrations were obtained. Figures 40 through
42 show scanning electron photomicrographs of the test dust and represen-
tative samples of dust collected by the cyclones. The spheroidal nature of
the particles is easily seen.
Figure 43 shows the calibration results. Note that cyclone
efficiency is plotted against physical particle diameter; converting to
aerodynamic diameter requires correcting for the nonunit density of the
aluminum particles. This conversion has been done in Table 5, which shows
the DSQ cut diameters for each of the five cyclone configurations expressed as
5-28
-------
E
="•
DC
LU
s
<
0
O
90
80
70
60
50
40
30
20
10
9
8
RANGE OF DISTRIBUTION VALUES
MEASURED AFTER PASSING
THROUGH FEEDER AND HEATER
LOWER LINE = DISTRIBUTION
OF ALUMINUM POWDER AS MANUFACTURED
10 15 20 30 40 50 60 70 80 85
PERCENTAGE UNDERSIZE
90
95
98
Figure 39. Results of tests of the constancy of the
size distribution of aluminum powder passing
through the test apparatus=
5-29
-------
Figure 40. Aluminum test dust.
Figure 41. Aluminum dust -- small cyclone cup catch
5-31
-------
Figure 42. Aluminum dust -- large cyclone cup catch.
5-33
-------
CO
tn
•o
9)
4->
U
O)
-------
TABLE 5. SUMMARY OF CALIBRATION RESULTS
DgQ Cut Diameter, ym
Cyclone
Large
(with SB)
Large
(w/o SB)
Medi urn
(with SB)
Medi urn
(w/o SB)
Small
Aerotherm
Physical
6.6
6.2
3.65
2.18
1.05
Aerodynami c
10.8
10.2
6.0
3.6
1.55
SoRI
Aerodynami c
4.9
3.5
5-36
-------
both physical and aerodynamic diameters. The SoRI 059 values for the medium
cyclone are shown for comparison. The agreement between the two methods is
good.
A second set of SASS cyclones has recently been calibrated by this
method. The cyclones tested are a part of a SASS train owned by KVB, Inc.
These cyclones were calibrated with the swirl busters removed from the large
and middle cyclones. Figure 44 compares the calibration results for the
KVB and EPA SASS cyclones. The two cyclone sets compare quite well for the
large and medium cyclones, and reasonably well for the small cyclone. This
is to be expected, as all of the uncertainty factors in the calibration method
— particle deagglomeration, Coulter counter accuracy, particle uniformity
—are more significant for the smaller particle sizes.
Note that Figure 44 correlates cyclone efficiency with physical
particle diameter, as measured with the Coulter counter. Table 6 shows the
physical and aerodynamic 059 cut diameters of the KVB and EPA cyclones. It
is interesting to note that the DgQ aerodynamic cut diameters of the two SASS
cyclone sets calibrated are reasonably close -- averaging 9.7, 3.7, and 1.5
ym — to the desired cut diameters of 10, 3, and 1 ym that were established
at the start of the SASS development program.
5-37
-------
KVB cyclones
•-—- EPA cyclones
tn
k
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Test dust - aluminum
Test dust density • 2.7 g/cm3
Calibration temp. - 400°F
Calibration flowrate - 4.0 scfm
Particle diameter measured
by Coulter counter
Large
cyclone
(No S.B.)
Middle
cyclone
(No S.B.)
.3 .4 .5 .6 .7 .8.91 2 3
Physical particle diameter, urn
5 6 7 8 910
15 20
Figure 44. SASS cyclone calibration data.
-------
TABLE 6. COMPARISON -- CALIBRATION RESULTS FOR TWO SETS OF SASS CYCLONES
D50 Cut Diameters, ym
KVB EPA
Cyc 1 one
Large3
Med i uma
Small
Physical Aerodynamic Physical Aerodynamic
5.61 9.2 6.20 10.2
2^.30 3.8 2.18 3.6
0.81 1.3 1.05 1.7
aSwirl busters removed
5-39
-------
SECTION 6
SASS MODIFICATIONS AND IMPROVEMENTS
As with any complex instrument, field use has demonstrated the need
for changes to ths SASS. Some changes have been accomplished, and some have
been recommended and are being studied now. In this section each actual or
potential SASS modification is described and discussed.
6.1 FIRST FIELD TESTS
Most of the modifications were the result of comments from the field
crews for the first few field tests, which are briefly described below. In
all of the field tests, the problems noted were overcome in the field by the
test crews, and the desired data was obtained.
Pulverized Coal Furnance Test
This was the first SASS test. It was conducted on the stack of
Acurex's pulverized coal-fired furnace before the baghouse on May 24 to 28,
1976. The train used was made by combining one of the incinerator trains with
the cyclone/oven section from the CPC train. This produced a SASS nearly
identical in design, and functionally equivalent, to current SASS trains. The
test was conducted without incident, and cleanup and sample recovery were
straight-forward. Figure 45 shows the sampling location. Table 7 shows the
particulate collected during the 5-hour test, while Table 8 is the
concentration of various elements in the 1-pm cyclone catch, as measured
by spark source mass spectroscopy. The fractional organics catch, obtained
6-1
-------
TO BAGHOUSE
SASS TRAIN
* * * * * ' ' *
•**'*' '*'*'' *'*''' *
oooooo
oooooo
HEAT EXCHANGERS
OOOOOO
oooooo
oooooo
oooooo
o o
o o
AEROTHERM MULTIFUEL FURNACE
(BURNING WESTERN KENTUCKY COAL)
Figure 45. First SASS test sampling location.
6-2
-------
TABLE 7. FIRST SASS TEST -- PARTICULATE CATCH
Fraction
Probe + Nozzle
lOy Cyclone
3y Cyclone
ly Cyclone
Filter
Grams
0.35
14-95
15.47
10.15
2.16
Percent
0.8
34.7
35.9
t
23.6
5.0
100.0%
6-3
-------
TABLE 8. PARTICULATE MATTER IN ly CYCLONE -- FIRST SASS TEST
(Concentration in ppm weight)
Element
Uranium
Thorium
Jismuth
Lead
Thallium
Mercury
Gold
Platinum
Iridium
Osmi urn
Rhenium
Tungsten
Tantalum
Hafnium
Lutetium
Ytterbium
"huli urn
Erbium
Hoi mi urn
Dysprosium
Cone.
45
12
0.3
95
20
NR
1
<0.3
0.7
0.3
3
0.4
2
3
5
Element
Terbium
Gadolinium
Europium
Samarium
Neodymium
Praseodymi
Cerium
Lanthanum
Barium
Ces i urn
I od i ne
Tellurium
Antimony
Tin
Indium
Cadmi urn
Silver
Palladium
Rhodium
Cone.
2
3
2
5
58
urn 14
120
45
310
25
0.1
<0.2
10
8
STD
3
0.5
Element
Ruthenium
Molybdenum
Niobium
Zirconium
Yttrium
Strontium
Rubidium
Bromi ne
Selenium
Arsenic
Germanium
Gallium
Zinc
Copper
Nickel
Colbalt
Iron
Manganese
Chromium
Cone.
65
30
420
320
310
560
0.2
3
36
97
170
MC
150
220
160
MC
620
400
Element
Vanadium
Titanium
Scandium
Calcium
Potassium
Chlorine
Sulfur
Phosphorus
Silicon
Aluminum
Magnesium
Sodium
Fluorine
Oxygen
Nitrogen
Carbon
Boron
Beryl 1 i urn
Lithium
Hydrogen
Cone
MC
MC
68
MC
MC
14
MC
MC
MC
MC
MC
MC
-100
NR
NR
NR
270
14
36
NR
6-4
-------
by extracting the sorbent per Level 1 procedures, is shown in Table 9.
Table 10 shows GCMS analysis of Fraction 2.
KVB Boiler Test
The second SASS test was on the stack of the KVB, Inc. experimental
coal-fired boiler. The test was conducted on July 21 to 22, 1976, using the
same train used for the Acurex test. This train used a single vacuum pump
and the test was conducted at 4.0 acfm at the cyclones, the standard flowrate
at that time. Problems noted included difficulty in disassembling the
Swagelock fittings in the cyclone assembly when hot, some dust deposition in
an inaccessible spot in the middle cyclone, insufficient strength in the
welding of the inlet tube on the small cyclone, some corrosion in the organic
module, and bypass of gas due to backflow holes in the impinger stems.
Comanche Power Plant
The stack of the Comanche Power Plant, Colorado Public Service Company,
Pueblo, Colorado was tested by Radian Corporation on September 2 to 10, 1976.
This SASS test was successfully completed, except for an inability to achieve
the desired flowrate through the SASS. The decision was made to supply
Radian with another pump and required hoses and valving to allow them to com-
plete their sampling program at 4.0 scfm (6.49 acfm at 400°F).
Radian also reported some difficulty in leak-testing the cyclone
assembly and the organic module, and a problem with breakage of the impinger
bottles due to interference with the top cap lockdown nuts.
Exxon Mini pi ant
Battelle Columbus Laboratories used a SASS to sample the Exxon pres-
surized fluidized bed combustor during the week of February 7, 1977 and again
in April 1977. Battelle reported some difficulty in sealing the SASS using
6-5
-------
TABLE 9. ORGANIC FRACTIONS FROM TENAX EXTRACT - FIRST SASS TEST
Fraction Weight
Fraction 1 28.0 mg
Fraction 2 42.0 mg
Fraction 3 84.0 mg
Fraction 4 72.0 mg
Fraction 5 70.0 mg
Fraction 6 24.0 mg
Fraction 7 <1 mg
Fraction 8 <1 mg
TABLE 10. COMPOUND CLASSES IDENTIFIED IN FRACTION 2
m/e Compound Class
129 Tetrahydroacenaphthenes
149 Phthalate esters (plasticizers)
153 Acenaphthenes, biphenyls
165 Fluorenes
177 Phenanthrenes
181 Tetrahydrophenanthrenes
205 Tetrahydrofluoranthrenes
227 Chrysenes
251 Benzopyrenes
-------
the Teflon gaskets and 0-rings supplied. They also noted considerable evidence
of corrosion in the organic module, with a considerable amount of green con-
densate collected. The condensate was very high in nickel and chromium ion.
6.2 COMPLETED MODIFICATIONS
The various modifications and improvements to the SASS that have been
implemented are listed here in approximately chronological order.
Lower Sorbent Temperature
In the original design philosophy for the organic module, the sorbent
temperature was to be maintained at 60°C and condensate was to be collected
before it reached the sorbent (see Sections 2 and 4.3). Midway through the
initial design cycle, EPA directed that the sorbent temperature was to be
held at 20°C, and any condensate formed upstream of the sorbent was to be
allowed to pass through the sorbent bed. These changes were accomplished
before the construction of the first organic module.
Increase Train Flowrate
During the Comanche Power Plant SASS test. Radian Corporation reported
they were unable to achieve the desired flowrate of 4.0 scfm with the single-
pump train they were using. Communications between Acurex, Radian, and
EPA revealed a misunderstanding as to the desired SASS flowrate. Acurex
and Southern Research Institute (who were then calibrating the cyclones) had
understood the desired flowrate to be 4.0 afm at the cyclones, while Radian
had been informed by EPA that the operating flowrate was to be 4.0 scfm. In
order to collect a sufficiently large sample within a reasonable sampling
time, the SASS flowrate was established at 4.0 scfm. Two major SASS changes
were required to increase the train flowrate by some 62 percent. The cyclones,
which had been designed for the lower flowrate, would have to be recalibrated
6-7
-------
and possibly modified. And, because of the increased pressure drop at the
higher flowrate, two vacuum pumps operating in parallel would replace the
single pump previously used. (The cyclone redesign and calibration efforts
are discussed in Section 5.)
A series of laboratory tests at Acurex led to the decision to use
two vacuum pumps. The tests demonstrated the capability of a single pump to
achieve 4.0 scfm with a clean system, but as the filter loaded up, the
increasing pressure drop quickly reduced the flow capacity of the system
below 4.0 scfm. The tests were conducted as follows:
A SASS train was assembled in field configuration, including clean
fiberglass filter, water in two impinger bottles, but without desiccant.
The probe and overn were maintained at 400°F. Separate tests were run with
one pump and with two pumps.
In each case, after the operating temperatures were reached, the
throttling valve at the pump(s) was opened in steps and the system flowrate
noted. When the pump was wide open, maximum system flowrate was achieved.
Then in order to simulate an increase in the filter pressure drop, the valve
between oven and organic module was closed in steps. This of course reduces
system flowrate and increases system Ap. Figure 46 shows the results of these
tests. All flowrate data in Figure 46 are corrected to standard conditions
(29.92 in. Hg., 70°F). As can be seen, the maximum flowrate attainable is
the operating zone. With a single pump, 4.0 scfm can be maintained between
9 in. Hg. (the Ap for a clean system) and 15 in. Hg. With two pumps, the
operating zone is between 9 and 21.5 in. Hg.
In subsequent field tests, two vacuum pumps have provided sufficient
capacity for accomplishing Level 1 sampling.
6-8
-------
en
X
o
OPERATING ZONE WITH 1 PUMP
OPERATING ZONE WITH 2 PUMPS
4 5 6 7 8 9 10
@400°F IN OVEN & 70"F @GAS METER
11
12 13 14 15
SYSTEM AP, IN. H6.
17 18 19 20 21 22 23
Figure 46. Vacuum pump characteristic curves.
-------
Replace Component Connectors
The original SASS design used Swage!oc connectors for the cyclone and
impinger assemblies. Several reports of difficulty with these connectors,
particularly in the cyclone assembly, led to a change to ball-cone type union
fittings. The new fittings have worked well, and allow disassembly and
reassembly of the cyclones and filter when hot, which was not possible with
the Swageloc fittings.
Replace Ball Valve \
The ball valve at the entrance to the organic module was found by
Radian (during the Comanche Power Plant test) to leak after some time at
elevated temperature. A valve from a different manufacturer was installed,
which has since given good service.
Modify Middle Cyclones!
The top flange of the middle cyclone was modified to eliminate a recess
that tended to collect dust and was difficult to clean.
Modify Impinger Stems
The HVSS impinger train is designed so that the impinger stems (the
straight sections of tubing that carry the sample gas from the top cap to
below the liquid level) have several small holes just above the liquid
surface. These holes help avoid backflush of the solutions when the train
pressure gradient is reversed. Concern that some sample gas was bypassing
the solutions led to a design change eliminating the small holes.
Modify Impinger Cap Nutl
Several field crews reported chipping of the top of the impinger
bottles due to interference with the locknuts retaining the top fittings. A
modified nut was designed that eliminated the problem.
6-10
-------
Change Seal Material
There were many reports of problems in getting a satisfactory leak
check using the Teflon gaskets and 0-rings originally designed for the SASS.
The Teflon seals seemed to work satisfactorily for experienced crews using
new seals. With the first use, however, the Teflon acquires a permanent set
that makes reuse problematical. It was recognized that a seal material other
than Teflon would best solve the problem. The basic prohibition of any
materials except 316 stainless steel, Teflon, and Pyrex glass for use in the
SASS required thorough testing of alternate seal material candidates.
After testing at EPA, Acurex, and Arthur D. Little, Viton A was
approved for SASS use, although Teflon is still the seal material of choice
when it can be used. Official approval was received from EPA on May 27, 1977.
Viton has worked well to solve the leak problems, but it seems to stick to
train surfaces when heated making disassembly difficult. Teflon-coated Viton
seals are now being evaluated.
Oven Wiring Change
Some problems with shorts in the oven electrical system were traced to
faulty installation of electrical wires in the oven base. Assembly
procedures were changed to eliminate the problem, and existing SASS owners
were notified of a simple field fix.
Modified Sorbent Cartridge
The original sorbent cartridge, which was designed to be assembled
with force-fit crimp rings, was redesigned with a screw-thread fitting for
easier assembly/disassembly. This change is discussed in Section 5.2..
Impinger Connector Tube Clamps
There are three 1/2-inch tubes that interconnect the impinger bottles,
and carry sample gas from one bottle to the next. In the original SASS
6-11
-------
design, rigid stainless steel tubes, bent into a U-shape, were used.
These performed adequately, but their rigidity tended to cause leaks at the
impinger top cap fittings if the impinger bottles were turned or twisted.
Also, with the rigid tubes the entire impinger assembly was fixed and move-
ment at any single part was generally not possible. For example, in order
to change the desiccant, the entire impinger assembly had to be removed from
the ice bath, worked on, and then be carefully leak tested before resuming
operation.
To overcome these problems, flexible metal bellows were substituted
in December 1976. These made operation of the train much easier. Subsequent
field experience showed evidence of corrosion in the bellows at locations
where the bellows construction left crevices where oxygen depletion could
occur. Also, some difficulty was experienced in cleaning the bellows.
In September 1977, an impinger assembly using flexible corrugated
Teflon tubing was constructed by Acurex and field-tested by TRW. They
reported excellent field performance, with very easy post test cleanup.
These Teflon connectors have been approved for Level 1 SASS use by EPA, and
are now being made available to SASS owners.
Oversized Filter Holder,
When sampling a source with a high particulate loading, it may be
necessary to replace the filter two or more times during a 5-hour Level 1
SASS test. Since the filter holder is hot (400°F), changing the filter is a
difficult and time-wasting procedure. Acurex designed an oversized filter
holder with about 2-1/2 times the surface area of the standard filter holder.
The oversized filter holder is compatible with existing SASS trains, and will
eliminate filter changes during most Level 1 tests. The oversized filter!
6-12
-------
holders were evalutated by TRW and SoRI with field tests; performance was
excellent. The oversized filter holders have been made available to SASS
owners.
Eliminate Spun Flanges on Medium Cyclone
The original cyclone design included a spun flange on the top of the
3-vim cyclone. This flange was thin and subject to bending. A welded-on
machined flange has been substituted.
Eliminate Crevice in Organic Module
Corrosion was noted in a crevice at the bottom of the condensate
collection section of the organic module. The welding technique was modified
to eliminate the crevice.
6.3 POTENTIAL SASS MODIFICATIONS
Several recommendations for SASS changes have been made that have not
yet been implemented. All suggestions received to date are listed below, in
no particular order:
• The cyclones are difficult to get at in the oven. Replumb so
fittings can be more easily reached. Consider moving inlet and
outlet holes in the oven walls so cyclone assembly is reversed
and the small cyclone is in front
• Consider an alternative to the aircraft-type clamps to eliminate
cut fingers by sampling crew members, and to improve sealing
• Use thicker metal in the cyclones so they are more sturdy
• Put index marks on the cyclones to allow easier assembly
• Design a jig to hold the cyclones outside the oven to allow
assembly and leak checking. The cyclone assembly could then be
inserted into the oven and connected using just the probe and
filter fittings.
6-13
-------
• Battalia reports erosion at the center of the filter paper.
Consider a design with a tapering inlet to reduce the entering
velocity.
t Consider rhodium plating the inside of the organic module to
eliminate corrosive attack. Will require concentric (non
scrapping) construction.
• Consider ways to improve the concentricity of the gas cooler
section of the organic module. Eliminate welds by boring heavy
wall tubing or spinning small tubing up to size.
• The nut at the gas outlet tube of the organic module interferes
with the Harmon clamp. Fix by moving or lengthening tube.
• Design a glass/teflon organic module to eliminate corrosion. Keep
existing temperature control section.
• Design a combined XAD-2 and condensate reservoir section
(eliminate the lowest clamp on the organic module)
• Increase the size of the impinger case. Insulate the case to
reduce ice use.
• Provide a draincock at the bottom of the impinger case
• Design heavier cyclone collection cups that are less likely to be
bent
• Provide an extra thermocouple to monitor the ice bath temperature
• Redesign the condensate bottle and adaptor for less fragility and
easier use
• Provide a level gage on the condensate reservoir
6-14
-------
• If the pressure drop can be tolerated, install check valves in
the impinger connector tubes to eliminate backflow
• Provide isokinetic sampling capability by adding an "Isokinetic
Module," described in Appendix B
6-15
-------
APPENDIX A
CYCLONE CALIBRATION DATA ANALYSIS
For a cyclone, the collection efficiency as a function of particle size
(n.(r)) can be calculated from
Mc(r)
n(r) -— - (1)
Mf(r)
where Mc(r) is the mass of particles of size r collected by the cyclone
during some time interval and Mp(r) is the total mass of particles of size r
introduced into the cyclone during the same time interval. Thus a series of
tests with different sized monodisperse dust could be used to establish the
collection efficiency of a cyclone. It is more common, however, to test the
cyclone with a polydisperse dust. In this case the available information is
the size distribution of the collected and input dust. This is commonly pre-
sented as the mass fraction of the dust with size less than (or greater than)
size r. If f (r) represents the size distribution, then M(r) is given by:
M(r) = M dr (2)
dr
where M is the total mass of the sample. Equation (1) then becomes:
dfc(r) / dfF(r)
n(r) = R - / - (3)
dr / dr
A-l
-------
where R is the ratio of the mass collected to the total feed mass. There-
fore, if the size distribution of the feed and the collected dust are avail-
able at discrete values of r, Equation (3) can be evaluated by numerical
differentiation of the two size distributions.
The accuracy of numerical differentiation is sensitive to the degree
of curvature of the function, the number of discrete points used to represent
the function and the technique used to calculate the derivative. These three
features are discussed in the following paragraphs.
Table|11 and Figures 47 and 48 show typical size distribution for
feed and collected dusts. Figure 47 is on a standard linear scale; whereas
Figure'48 is on log-probability paper. The importance of this type of plot
is that a normal (Gaussian) distribution will appear as a straight line.
Many dust samples are almost normally distributed. In general it is easier
to obtain accurate numerical deviation from near-linear curves then from
highly curved ones. In this case both figures exhibit a nearly linear nature
except near the ends of the curve in Figure 47. Therefore deviations
calculated near the ends of the curve in Figure 47 would be more sensitive
to the number of data points and the calculated procedure than those of
Figure 48. !
Transformation to the log-probit coordinates of Figure 48 offers
improvement to the accuracy of the numerical derivatives.
This transformation is given by:
y = probit (f)
x = An r
where probit (f) is defined by:
A-2
-------
TABLE 11. SIZE DISTRIBUTION OF FEED AND COLLECTED DUST
Particle size
(microns)
0.794
1.0
1.26
1.587
2.0
2.52000
3.17500
4.00000
5.04000
6.35000
8.00000
10.07900
12.69900
16.00000
20.15900
25.39800
fc
0
0
0
0
0
0.00500
0.01600
0.04700
0.10800
0.30900
0.55700
0.76100
0.87900
0.94100
0.97900
0.99100
f Filter
0.003
0.0
0.02
0.041
0.078
0.13900
0.2270
0.39400
0.64300
8.7200
0.97500
0.99500
1.00000
1.00000
1.00000
1.00000
fF
0.00169
0.00562
0.0112
0.023
0.0438
0.08031
0.13458
0.24201
0.40867
0.62541
0.79192
0.89251
0.94700
0.97416
0.99080
0.99606
Definitions:
fc - mass fraction of collected dust with diameter less than r
fFilter - mass fraction of dust with size less than r that is collected
on a filter at the outlet of the cyclone.
fp - mass fraction of dust with size less than r in the feed. Calculated
from mass conservation principles: fp(r) = Rfc(r) + (I-R) fpi
where R = total mass fraction of dust collected.
A-3
-------
1.0
0.8
0.6
f (r)
0.4
0.2
— FEED DUST
Q — COLLECTED DUST
i I L I
0 2 4 6 8 10
12 14
r (Mm)
16 18 20 22 24
Figure 47. Curve of distribution of feed and collected dust.
-------
100
90
eo
70
50
40
30
20
10
9
8
7
6
5
- FEED DUST
- COLLECTED DUST
0.01
0.1 0.2 0.5 1 2 5 10
20 30 40 50 60 70 80
f (r) x 100
90 95 98 99 99.8 99.9 99.99
Figure 48. Curve of feed and collected dust
(log — probability plot).
A-5
-------
probit(f)
VliT
F'
L ( e-l/2 t2 dt. (5)
» J
Then (df/dr) is given by:
df df dy dx 1 2/2 dy ,fi.
— = s —=5 e •> ' — vDy
dr dy dx dr rVzir dx
and Equation (3) becomes:
n(r) =
dyj
dx
where — c (= - — ) and — are numerically evaluated from the data
dx d An r dx
and yc (= Probit (fc)) and Yj are obtained from standard functions available
on most computers or from tables (like trigonometric or exponential functions).
Both methods for calculating the cyclone efficiency (Equations (3) and
(7)) were examined and, except in a few cases, yielded equivalent results.
Figure, 49 and Table 12 show the results of calculations for which the two
methods are noticeably different. Equation (7) yields much smoother results,
as expected.
The number of data points available to define the size distribution
depends on the type and resolution of the equipment used. For this program
a Coulter counter was used. Measurements were obtained at approximately
equal intervals of log r. Table 11 presents the distribution data for i
curves in Figure 47 and 48. Enough measurements must be taken to define
the linear mid-range (see Figure i47) and the curved regions near the ends.
If few data points are available to define the curved region, the transformation
to the more linear form (log-Probit) becomes more necessary.
A-6
-------
1.0
0.8
0.6
0.4
0.2
EQUATION (7)
Q — EQUATION (3)
3 4 5 6 7 8 9 10 12 14 16 18 20
r(ym)
Figure 49. Comparison of cyclone collection efficiency calculated from Equations (3) and (7),
-------
TABLE 12. CALCULATED CYCLONE EFFICIENCY
Particle size
(urn)
2
3
4
5
6
8
10
12
16
20
25
.52000
.17500
.00000
.04000
.35000
.00000
.07900
.69900
.00000
.15900
.39800
Collector
Equation (7)
0.06593
0.10162
0.14050
0.22840
0.53988
0.79362
0.93936
1.00642
1.05227
1.02801
1.00333
Efficiency
Equation (3)
0.05305
0.10985
0.13635
0.33593
0.45654
0.83964
0.94108
0.97848
1.00207
1.00000
1.00000
A-8
-------
There are numerous ways to compute numerical derivatives. Several are
listed below:
• Linear fit between points; compute derivates at a point from the
average of slopes to the adjacent points
• Quadratic curve fit through the three nearest points; compute
derivatives at a point from the slope of the quadratic through the
point and the two adjacent points
• Weighted quadratic curve fit if using the points on either side
and a centered quadratic
• Higher order polynomial curve fits
t Transcendental function curve fits
t Weighted linear curve fits using four or five adjacent points
All of these offer advantages or disadvantages for particular shapes of
curves. In the present analysis, the first three methods were considered.
This choice was based on the degree of linearity of the data. Method 2 gave
the most consistent results. This is probably a result of the "smoothing" of
data inhererent in the method. |
A listing of the computer program written to process the dust
distribution data and calculate cyclone efficiencies is given in Table 13.
Tables|14 and 15 show a listing of the input data and the output
corresponding to the example given in the Figures 47, 48, and 49.
A-9
-------
>
o
TABLE 13. PROGRAM FOR CALCULATING CYCLONE EFFICIENCY
11/1M/77 0-:<42:uu r>TO iU'^AA3;+3 Oli02ny S29 20
PE I LOr-'Krj P<0<;
2 pARAfLlrK P=50
5 C'JMMOtV /!. Im/AK/
f * H F ( P 5 , FU,P ( P ) » XH ( V ) , XCUP ( P ) « SF ( P ) , SCUP ( P ) t EFF ( P ) » R ( P ) i EF ( P > »
t) * OF'i (F') tDf-C (P)
b C0h;«0!' /CONST/
7 * '\URATIOi I SL(iP
-------
TABLE 13. Continued
1* P'lCLlJOfc. CO WON
2* C *****
3* C
4* CALL INPUT
5* C
6* C *****
7* C
8* C FIND RLCPt'IS FOR H VS F
9* C
10* IF
-------
TABLE 13. Continued
33* DO 20 1 = 1,IM
i )**?)/2
35* £FF{i)=KAT10*SF (D/SCUP( !)*£
36* 20 Ct)NTTMUt'
37* C *****
38* C
39* CALL OUTPUT
40* C
* *****
-------
TABLE 13. Continued
1* S'JPKO-Jf I;v.l. I'KPUT
2 + I sICU.H'E COP; (P.*
3 *• 1)1 H T ' v S 1 0 ; \! K £ C 0 l< D < "i 6 )
t* READ (b,l) (HECOHUd ) » 1 = 1 » IB)
b* 1 H-JRf'/M (1BA4)
7* C
6* C IM=i\!0. OF PTS., IbT = l FOR f-.C7T.R, 1FILT = 1 FOR FILTER
-J+ C
ID* RLAU (5.10) :UIGTF,1RTC,IFILT,ISLUP
11* 10 F OK MAT (1215)
12* WKIT^ (^,^0) N»1GT> iTGTCt iFlLT.ISLOP
13^ 2 a F URN. AT (/PX.?H!Ni=»IS»2X.5HlGTF=»I3«2X»5HIGTC=t 13 »2X»6HIFlLT= « I3-i2X
Ib* C
lb* C HrtTlU =
17* C
1 8 * M t; A U ( b » S (.1 ) »- A f 1 0
lf^t* 3H FOR?': AT
-------
TABLE 13. Continued
20* WHITt (b,40) RATIO
21* 40 FORMAT (/2X,22HMASS CUP / MASS FEED =,F10.5/)
22* IF(IFJLT.rQ.l) WRITE (6,45)
23* 45 FJRfiAT (/,2X,5H*****//,2X«6HFILTER//«2X,5H*****/)
24* IF(IGTT-.EO.l) WRITL (6,50)
35* 50 FURMAT (/ , 2X,49HINPUT PERCF.NTS HAVE BEEN CHANGED TO % LESS THAN R/
26* *)
27* MKITE (t, ,60)
28* 60 FORMAT ( MX , I.HI , 2X , lOHK ( MICRONS ) , 2X , 5HFF ( I) t 3X .7HFCUP ( I ) / )
29* C
50*' C T>-JPUT « IN HICRONS
31* C
32* DO 65 1=1tN
33* READ (5,30) R(I),FF(I I,FCyP(I>
3f* IF {IGTF.E'Q.D FF (1 ) = 100 .-FF (I )
3b* IF (IGTC.E«i.l) FCUP(I)=100.-FCUP(I)
36% FF(I)=FF(I)*.01
37* FCUR(J)=FCUP«I)*.01
58* WRITE. (6,70) I ,R (I ) »FF ( I ) , FCUP ( I )
59* 65 CONTINUE
<4Q* 70 FORMAT (2X,I3,7FlO»5)
41* IF (IFILT.E.Q.O) GO TO 90
WHITE (6,75)
75 FORMAT ( /£X , 5H*****// * 2X , "+HFEED// « 2X, 5H*****/ )
WHITE (6,60)
45* C
46* C CHANGE FILTER TO INPUT DUST
47* C
48* 00 60 1 = 1, IM
49* FF(I)=RATIO*FCUP(I)+(1.-RATIO)*FF(I)
50* WRITL (6,70) 1,R(I),FF(I),FCUP(I)
51* 80 CONTINUE
52* 90 CONTINUE
53* RETURN ^
54* EiMO
-------
TABLE 13. Continued
1* SUHROt.iTlNE OUTPUT
2* I'NtCLUHE CDHOM
5* wmrr (61 ID)
*+* til Fi.)h!V1AT < /3^X»25H*************************//(*2Xt 6HOUTPUT«//3**Xt25H*
^j* ********** * ********* *****/)
b* W'UTE (S,20) RATIO
7* 2U FORMAT ( /2x «£2HMASS CUP / MASS FEED =,F10.5/)
s* IP ( I SLOP. Ea. o) GU To 2y
10* 25 FOk^AT(/2X,3aHCOLLECTlOM LFFICIEMCY BASED ON OFC/OFI ,/3X . 28HI D
11* *FC Df I EFF/)
12* UU 28 1 = 1, i-J
15* WHITE (6,60) 1 ,OFC ( I ) ,OF H I ) , EF ( I )
j, 141. 2«
L, IS* 29
01 Ib* UNITE
-------
TABLE 13. Continued
23*
24*
25*
1 * S-. IBROU1 1 r«f. SG E. f R ( L « X » Y « EM )
2* IH^ENSIO'V X(lU)»YfiO) ,t
'6* J
-------
TABLE 13. Continued
2K* IPU-L + 1) 20,7,19
29* 20 A4 = X( I + 2)~XU + 1)
30*
31* 5
33* 0-1 = 2
34* f-jjj TO y
35* 6 COMTIfvUt-:
36* GO JO 16
37* 7 A4=A3
38* AX=A3
40*
41*
42*
43*
M-4*
45*
46*
47*
48*
49*
50*
51*
52*
53*
54*
55*
56*
57*
58*
59*
60*
19
9
10
11
12
13
14
15
16
H4=B3+Da
B1=84+DB
CTJ TO 5
A4 = A1
B4=B1
RO TO 5
IMS23) ll,lfl,M
Ti\i = R2/A;>
GO TO 17
IP(A6S(S12)-I-ABS(S13). ) 10, 10,
IP(A.RS(S24)+A6S(S54M 10,10,
IF(Aes(Sl2)-t-ABS(S34) ) 15,15,
T
-------
TABLE 13. Concluded
62* 25 A = 39
fa 5* 1 4 = A/f *P.b
6
-------
APPENDIX B
ISOKINETIC SASS TRAIN CONCEPTUAL DESIGN
Introduction
The SASS, as now designed, maintains a constant flowrate at the
cyclones to assure a constant D50 cut diameter. As a result, it is not
possible to continually vary the train flowrate to maintain isokinetic
sampling at the nozzle, as is done with Method 5 particulate sampling trains.
For Level 1 sampling with the SASS, the usual procedure is to choose a nozzle
size from those available that gives the closest approximation to isokinetic
sampling. An "expected average" duct velocity is used. Isokinetic velocity
mismatch can usually be held to less than 20 percent when duct velocity
is constant during the test. If duct velocity varies, the mismatch can
be greater.
For particles less than about 5-jjn aerodynamic diameter, the levels
of isokinetic mismatch experienced with the SASS are probably not significant.
For larger particles, or for widely varying duct velocities, nonisokinetic
sampling may introduce significant errors.
Recognizing this potential problem, EPA/RTP authorized Acurex
to briefly investigate ways to modify the SASS to make it capable of true
isokinetic sampling, while maintaining constant flow at the cyclones.
Only concepts involving flowrate variation were considered. (A report was
prepared and submitted to EPA in June 1977.)
B-l
-------
Possible Approaches
The basic problem is to maintain a fixed flowrate through the SASS
train cyclones (4.0 scfm = 6.49 acfm at 400°F), while at the same time
varying the sampling velocity at the nozzle to adjust to varying stack gas
velocities. Four different ways of doing this were considered; these are
described below.
In-Stack Velocity Control
In this concept, the flowrate through the entire train is kept
constant at 4.0 scfm, while the velocity of the stack gas near the nozzle is
adjusted. Figure 50 illustrates the concept. The sampling nozzle is
surrounded by a large shroud that sucks in stack gas at a constant flowrate
so the nozzle is sampling isokinetically. Variations in the stack gas
velocity are automatically corrected for by the design of the system.
Figure 50 shows a representative set of flow streamlines, assuming
the velocity through the shroud is substantially higher than the stack
velocity. The flow streamlines are badly distorted near the periphery of the
shroud, but are nearly straight at the center of the shroud (where the
sampling nozzle is located). Thus, the stack gas entering the nozzle
contains an undistorted distribution of stack gas particulate. A similar
situation occurs when the stack gas velocity is higher than the shroud
velocity. In effect, the sampling system simply ignores changes in stack gas
velocity because the shroud provides a constant-velocity environment for the
sampling nozzle.
The main advantage of this concept is its simplicity and low cost.
All of the components of the SASS train, except the probe, are unchanged.
The new probe, additional ducting, and exhaust blower would be relatively
B-2
-------
Stack wall
DO
u>
Standard
samplin
nozzle
exhaust
to cyclone oven
Flow Streamlines
Figure 50. In-stack velocity control system.
-------
inexpensive. There are several potential drawbacks. The most serious is the
basic question "will it work?" Clearly, a substantial development/
calibration effort would be required to prove the concept, although a
preliminary evaluation indicates that the concept is valid over relatively
wide stack/shroud velocity mismatches (3:1 and more) for particles of 30 ]xn
diameter or less. Another potential drawback is the relatively large size
of the probe (a 4- to 6-inch port would probably be required).
In-Stack Split Stream Probe
In this concept, stack gas is filtered and mixed with sample gas in
varying ratio in order to allow a variation in nozzle sampling velocity.
This is another concept in which only the probe need be modified. Figure 51
shows a possible embodiment. A constant flowrate of 4.0 scfm is pulled
through the nozzle. In order to accommodate changing stack gas velocities,
a dilution side stream is provided. The system would work as follows:
• Gas flow through the train is set at 4.0 scfm
• The control valve position is adjusted so that about 3 scfm passes
through the nozzle and 1 scfm passes through the dilution stream
• The nozzle is sized so that at 3 scfm passing through the nozzle,
nozzle velocity matches stack gas velocity
0 The probe is inserted into the stack
• Stack gas velocity (i.e., pitot AP), temperature, pressure, and
dilution-stream flowrate are continuously monitored by a "black
box" that outputs a signal to the control valve
t In response to the signal, the control valve adjusts the flowrate
of gas through the nozzle so isokinetic conditions are maintained.
Since constant flowrate through the train is maintained by the
B-4
-------
Pi tot
Stack gas temperature'
Stack gas pressure •
Stack gas velocity •
Flow meter signal
Black
Box
/ \
T.C.
Control
valve
signal
Standard nozzle
Stack wall
To cyclone
oven
Figure 51. In-stack split stream probe.
B-5
-------
vacuum pumps, any increase (or decrease) in nozzle flowrate auto-
matically causes an exactly corresponding decrease (or increase)
in dilution-stream flow.
The primary advantage of this concept is relative simplicity -- the
SASS train need not be modified, except for the probe. The problem areas are
in developing the black box and the control valve. The black box would
be a microcomputer, the programming of which would be straightforward but
expensive. The control valve design would have to be evaluated to assure
that excessive dust buildup on and near the valve does not occur. Finally,
a direct reading of volume of gas sampled is not obtained with this concept.
The integrated dilution-stream flowmeter gas volume would have to be
subtracted from the dry gas meter reading at the end of the sampling period.
This concept would be quite attractive if relatively large numbers ~
several hundred ~ were to be built. For the much smaller quantities we are
considering, high development costs probably rule it out.
Hot Gas Recycl e ',
A different way to allow isokinetic sampling while maintaining
constant flow in the cyclones is to continually vary the amount of gas pulled
through the train (to adjust the nozzle velocity) while simultaneously changing
the flowrate in a feedback loop around the cyclones. Figure 52 illustrates
the simplest way to do this.
The SASS train is basically unchanged from its original configuration,
except for the addition of the recycle loop. Isokinetic sampling is achieved
by first choosing a nozzle size such that at 4.0 scfm through the nozzle, and
for the maximum stack gas velocity expected during the test, stack gas and
nozzle velocities are matched. The probe is then inserted into the stack,
B-6
-------
I
--J
Heated Probe
Differential
pressure
controller
Organic
module
Remote
actuated
valve
Vacuum
pumps
Exhaust
Figure 52. Hot gas recycle.
-------
and the pitot reading is used to set the control-module flowrate at a value
that gives isokinetic sampling at the nozzle. This flowrate will in general
be less than 4.0 scfm, and will vary as the stack gas velocity varies.
Throughout the duration of the test, the flowrate through the train (as
measured at the control module) will be changed frequently. In effect, we
will be operating the SASS train in the same way a Method 5 train is
operated, using a nomograph or programmed HP-65 to calculate flowrate from
stack gas velocity, temperature, and pressure.
The purpose of the recycle loop shown on Figure 52 is to automati-
cally maintain a constant 4.0-scfm flowrate at the cyclones while the train
flowrate varies. The small cyclone is used as differential pressure
flowmeter, in conjunction with a photohelic DP controller driving a remote
actuated valve. A gas pump moves the recycle gas stream from the low-
pressure point after the filter to the high-pressure point upstream of the
first cyclone. The recycle loop is fully automatic in that any deviation
from a flowrate of 4.0 scfm at the small cyclone immediately actuates the
recycle stream valve to correct the flowrate to that value.
In this concept, we have chosen to recycle hot gas from immediately
behind the filter. This eliminates the need for a heater on the recycle
stream, but requires a gas pump capable of continuous operation at 400°F.
The parts of the pump that contact the gas stream must, of course, be
constructed of 316 sst, or Teflon. A metal bellows pump meeting the
material, temperature, and flowrate requirements is available, but would be
extremely costly. We have been quoted a price of $2900 with 26-week delivery
in lots of 10. This high cost and long delivery time combine to make
unattractive any concept utilizing a hot gas pump. Another drawback is the
B-8
-------
possibility of condensing organics or trace elements if any part of the
recycle stream falls below 400°F. Also, a certain fraction of the vaporous
pollutants in the sample gas will be recycled many times before being
collected in the organic module and impinger train. The increased residence
time in the sampling system increases the probability of catalytic or other
reactive changes that could distort later pollutant analysis.
Cool Gas Recycle
One of the most serious problems with the hot gas recycle concept was
the very expensive gas pump made necessary by the system material restrictions
and the high temperature. In the cool gas recycle concept, this problem is
overcome by recycling cool, cleaned gas fron a downstream point in the train,
as shown schematically in Figure 53. Its operation is similar to that
described in the previous section for the hot gas recycle concept. The
primary difference is in the recycle loop, where the gas pump now operates at
ambient temperature, and a gas heater is required to raise the recycle loop
temperature to 400°F before reinjection upstream of the cyclones.
Since the gas pump need not operate above room temperature, a wider
choice of pumps is available. We have chosen a diaphragm pump as best
meeting the system flowrate/pressure drop/materials requirements. In the
particular pump chosen, all parts contacting the gas are 316 sst. except for
a Teflon-coated neoprene diaphragm and Teflon check valves. The pump
manufacturer has assured us that he has hundreds of such pumps in the field
that have operated for several thousand hours with no cracking or other
deterioration of the Teflon/neoprene diaphragm.
The reason we have chosen to use the small cyclone as a flow-
monitoring DP device requires some explanation. First, the small cyclone is
B-9
-------
CO
o
Heat probe
Thermocouple
Temperature
controller
Organic
module
Gas
Heater
Remote
actuated
valve
Vacuum
pumps
Figure 53. Cool gas recycle.
-------
the highest AP component in the train with a nonvariable AP (this gives
greatest sensitivity in the recycle flow control loop). Second, it is
already in the train, so pressure taps may be added with a minimun of change
and expense. Most importantly, there seems to be no good place to add
another DP device such as a venturi or an orifice. Upstream of the filters
dust buildup may be a problem. Downstream of the filter, the absolute
pressure varies considerably during a run, so DP device characteristics will
not be constant. For these reasons, the small cyclone seems now to be the
best choice. Its use, of course, depends on the constancy of AP as dust
builds up in the cup. This would have to be verified by calibration tests.
In both this concept and the previous one, less than 4 scfm of sample
gas may be pulled through the train. For example, if the stack velocity is
expected to vary by a factor of 50 percent during a sampling run (say between
67 and 100 ft/sec), then 4.0 scfm of stack gas will be sampled at maximum
stack velocity, and 2.7 scfm will be sampled at the minimum velocity. The
average flowrate during the sampling run will lie somewhere between, but is
not predictable at the start of the test. The net effect is to lengthen the
{
required sampling time to obtain a given total volume of sampled gas.
Comparison of Concepts
Table 16 shows an overview of the advantages and disadvantages of the
various concepts considered. Taking all factors into account, concept 4,
i,
cool gas recycle, is reconmended.
Figure 54 shows how the cool gas recycle concept might be implemented
in one possible layout for an "Isokinetic Module." The isokinetic module
will be packaged in a fiberglass case of dimensions 16 x 20 x 8 inches. It
will weigh between 40 and 45 pounds, and will draw a maximum of 900 watts of
B-ll
-------
TABLE 16. EVALUATION OF CONCEPTS
Approach
Advantages
Disadvantages
1. In-Stack t Low cost
Velocity t Simple — no modification
Control to existing trains except
to probe
• Measures sample gas flow-
rate through trains
• Pulls a full 4 scfm of
sample gas
t Hard to get into stack
• Does it really provide
isokinetic sampling
at the nozzle?
2. In-Stack t All in-stack
Split • Modification only to probe
Stream • Fully automatic isokinetic
Probe sampling
Less than 4 scfm sample
gas
Doesn't measure sample
gas flowrate directly
Design of control
may be difficult
High cost in small lots
3. Hot Gas
Recycle
t Simple and cheap — few
components
• Automatic flowrate control
for cyclones
• Totalizes sample gas volume
• "Dirty" gas is recycled
• Air pump runs hot at
high AP ~ may not
exi st
• Manual control for
isokinetic sampling
• Less than 4 scfm sample
gas
Cool Gas • Clean gas is recycled
Recycle • Uses cool air pump
• Automatic flowrate control
for cyclones
• Totalizes actual sample
gas volume
• Air pump must generate
small AP only
More components than
design 3 — somewhat
more expensive
Manual control for
isokinetic sampling
Less than 4 scfm
sample gas
B-12
-------
oa
l-O
relief
valve
T.C. at
cyclone
oven inlet
Photohelic DP
controller
o
Temperature
control ler
From upstream
of vacuum
pump
Remote actuated valve
Figure 54. Schematic — isokinetic module.
-------
power (assuming maximum gas flowrate through the recycle loop is 2.0 scfm).
The module can be easily attached to an existing SASS train for users who
require it. It is entirely self-contained, except for electric power. All
components that contact the gas stream are of 316 sst. or Teflon. Since
only cleaned gas is passed through, it should require only moderate between-
test cleaning. It is simple, so cost should be low (approximately $5000)
and reliability high.
Conclusions
Concept 4, Cool Gas Recycle, is the most satisfactory way to add
isokinetic sampling capability to the SASS train. This concept lends itself
to incorporating all new parts into an "Isokinetic Module." The module can
be provided as part of new SASS trains or can be easily added to existing
trains by connecting tees between the probe and large cyclone, between the
*
impinger assembly and the vacuum pumps, and on both sides of the small
cyclone. Adding the isokinetic feature only requires two important changes
in SASS Level 1 operating procedures: (1) the operator must regularly
adjust system flowrate to account for stack gas velocity changes, and (2) a
slightly longer sampling time is required to collect a given volume of sample
gas.
B-14
-------
APPENDIX C
CONCEPTUAL DESIGN STUDY
As a part of a program to assess the emissions from a wide variety of
stationary sources, the Environmental Protection Agency has established three
source assessment levels. Each level of assessment serves a different
purpose and may require different sampling and analytical instrumentation.
Level 1 will provide semiquantitative screening of pollutant emission
sources. The Level 1 Source Assessment Stack Sampler -- the end-product of
this program — will provide particulate, organic, and trace element samples
for subsequent analysis. The particulate samples will be classified by
particulate size; some of these sized fractions will be subject to toxicity,
cytotoxicity, and trace element concentration analysis. The organic
sample will be extracted from its adsorbent and classified into eight
different classes of organic materials, on a semiquantitative basis. Similarly,
the impinger solutions sample will be analyzed to determine the presence of metallic
X
and inorganic pollutants.
The nature of Level 2 and Level 3 source assessment is less well
defined, although the philosophy has been established. Level 2 assessment is
indicated where Level 1 screening has shown the presence of a particular
pollutant in significant amount or concentration. The source will be
resampled — either with the Source Assessment Stack Sampler or perhaps with
a special sampling train — and the collected sample carefully quantitatively
C-l
-------
analyzed for the significant pollutant. Level 3 assessment constitutes
applying the current state of the art to detailed analysis of one or more
pollutants. The most sophisticated techniques will be applied to elucidate
every aspect of the nature of the pollutant.
The purpose of this Conceptual Design Study is to prepare a report
describing several conceptual Level 1 Source Assessment Stack Sampler (SASS)
trains. The goal is to determine if a SASS train costing about $15,000, and
including all required features, can be built, and to predict the incremental
cost required to add various control and convenience features. Both design
and construction costs are important factors and are explicitly considered.
Some of the ground rules that Acurex and the EPA have agreed as a basis
for this design study are:
t The SASS train must perform the functions of sampling, determin-
ing particle size distributions into four size ranges, organic
material collection, and trace element collection
• The cyclone assembly (designed on a previous program) and the
organic collection module (designed as Task 1 of the present
program) will not be modified (they may in fact be modified
later, but for "this report will be considered to have an established
price)
• A wide range of possible SASS designs will be considered, from the
simplest manual train to a fully instrumented train with
semiautomatic flow control
• Manufacturing costs will be based on production runs of five units
• The cost of the SASS trains, at all functional levels considered,
must be kept as low as possible. We have chosen four possible SASS
designs for consideration. In the next section, each possible
C-2
-------
design is considered in detail. The costs and operational
attributes of each design are compared to Design B, which we
recommend as the best compromise of price and performance. It
should be noted that all of the price estimates made in this
report are for comparison purposes only, and should not be
interpreted as firm prices suitable for procurement purposes.
C.I CONCEPTUAL DESIGNS
Design A offers the greatest number of features of the designs
considered, and has the highest cost. It basically corresponds to the SASS
train described in RFP No. DU-75-A303 and addressed in our Proposal No.
2146-75-B. Figure 55 shows a schematic diagram of the design. The key
features are the semiautomatic flow controller (which maintains a constant
flowrate through the cyclones despite changes in System P), the dual impinger
train (to allow additional trace element collection solutions to be used),
and the complete, centralized instrumentation readout capabilities. Two
vacuum pumps are provided to allow use of very high pressure drop filter
media. Table 17 shows a summary of the design features and costs.
The design costs for all four of the designs are based on the
assumption that much of the component design will be completed on the present
project. In general, design time required to build the incinerator trains
and then to upgrade those trains to full SASS capability (by addition of
probe, oven, and cyclones) will be charged to the present project; design
time required to eliminate known problems will be charged to the current
program. Redesign of problem areas (replacing stainless steel with
aluminum), ovens (eliminating fan and remote temperature controller), and
control modules (eliminating dry gas meters and orifice plates) for the
C-3
-------
o
i
Stack T.C. probe
i.e. r
Gas conditioner &
moisture collector
Pi tot AP
Gage
Sensor
umbilical
AP, TQ
'Porous polymer
adsorber
Cooled
section
Parallel
Imp/cooler'
trace element
collectors
Dry Gas Meter
Orifice Meter
Centralized Temperature
and Pressure ReaJ. jt
Control Module
_rv
Semiautomatic
control
module
Vacuum,
gage'
10 cfm vacuum pump
Figure 55. Design A system schematic.
-------
TABLE 17. DESIGN A SUMMARY
Component
Probe
Stack thermocouple
S-type pltot
SS 316 nozzles (8)
2-1/2" SS sheath
Vycor liner
Temp, controlled at control module
Oven
Large oven
Circulation fan
Temp, controlled at control module
Cyclones
3 standard cyclones
AP instrumentation
Filter Holder
No teflon coat
Special gasket
Gas Cooler & Tenax Module
Standard design
Impinger System
Parallel impinger trains
Large glass bottles
Extra instrumentation
Semiautomatic Controller
Compared to Design B
Design
Hours
40
1
60
180
Advantage
Closer to isokinetic sampling,
Noncontaminating at high
temperature. ' , , .
Rapid warm-up.
Allows flow measurement
by AP across small cyclone.
No chance of contaminating
analysis with teflon flakes.
Allows use of thin filters.
Allows extended analysis
of trace elements
Reduces carryover
Controls, monitors split
flows to parallel impingers.
Completely untended opera-
tion.
Disadvantage
Redundant with dry gas meter.
Hard to clean.
Component
Cost Change*
+265
+200
+10
+1230
+600
+2500
o
I
01
-------
TABLE 17. Concluded
Component
Vacuum Pump
Double vacuum pumps
Extra connecting lines
Umbilical Lines
Standard umbilical
Flow control umbilical
Control and Metering Module
Pressure readouts
Temp, readouts
Temp, controller
Orifice meter
Dry gas meter
Timer
Fan control
Power control
Un1ra1l
*Compared to Design B
Design
Hours
16
12
—
Advantage
Allows high AP filters.
Necessary with semiautomatic
flow controller.
Allows positioning of probe
for traversing.
Disadvantage
Cost, weight.
Component
Cost Change*
+924
+250
+150
+801
CT1
-------
purpose of reducing manufactured cost will be allocated to Design C or Design
D as appropriate.
Design B, shown schematically in Figure 56, is the simplest SASS
design that meets all of the functional requirements and is compatible with
existing HVSS components. The semiautomatic flow controller of Design A has
been eliminated, as have the duplicate impingers and vacuum pumps. The
temperature and pressure readout and control functions have been chosen so a
standard HVSS control module can be used. Table 18 summarizes the features
and costs of Design B.
Design C is shown schematically in Figure 57. This design is
a minimum interpretation of the basic SASS design. Numerous changes were
made to reduce production cost. These changes include eliminating flow
measurement devices from the control module (system flow is monitored by
means of the cyclone pressure drop); providing aluminum, rather than
stainless steel, sheathing for the probe; simplifying the umbilicals by
provision of single power, thermocouple, and pitot connectors; and
simplifying the oven by eliminating the convection fan and providing oven-
mounted temperature control. These changes significantly reduce the cost of
the Design C train; however, the ability to interchange components with
existing HVSS trains is largely lost.
The features and costs of Design C are summarized in Table 19.
The final conceptual design considered, Design D, is an all-out
attempt to achieve the lowest cost train possible. In this design (see
Figure 58), the control module has been eliminated completely, and each
component is locally controlled and monitored. The advantage of this concept
is cost — a sale price has been significantly reduced compared to Designs A,
B, and C. The disadvantages are threefold: in general, the oration of the
C-7
-------
o
i
00
Stack T.C.
•^ TT,
SS probe *L—i
Convection
/~ oven
Probe £.
T.C. r
Filter
I
I T^T
Dry gas meter
Orifice meter
Centralized temperature
and pressure readout
Control module
/
~l
Ovon
T.C.
Cyclone
AP gage
Gas Conditioning &
'moisture collector
r
Porous polymer
adsorber
Sampl e
T.C.'s
.1 0 cfm vacuum pump
Imp/ cooler
trace elanent.
collector
Impinger
T.C.
Figure 56. Design B system schematic.
-------
TABLE 18. DESIGN B SUMMARY
Component
Probe
Stack thermocouple
S-type pltot
SS 316 nozzles (3)
2-1/2" SS sheath
SS 316 liner
Temp, controlled at
control module
Oven
Large oven
Circulation fan
Temp, controlled at
control module
Cyclones /
3 standard cyclones
AP instrumentation
Filter Holder
Standard filter holder,
no teflon coat
Gas Cooler and Tenax Module
Standard design.
Impinger System
Large glass bottles
Standard case
Vacuum Pump
Single standard pump
Umbilical Lines
Standard umbilical
Control and Metering
Module
Pressure readouts
Temp, readouts
Temp, controllers
Orifice meter
Dry gas meter
Timer
Fan control
Power control
Design
Hours
-
Advantage
Allows high sampling temps.
Convenience, accuracy.
Rapid warmup.
Convenience.
Measures AP across small
cyclone to monitor flow.
No chance for contamina-
tion.
Eliminate splash and carry-
over, reduce contamination
of solutions. Cost less
than metal .
Standard control module.
Disadvantage
More expensive than aluminum.
Adds cost, weight.
Requires extra controls.
Redundent with dry gas
meter.
Harder to clean filter
holder.
C-9
-------
Convection
o
o
Gas Conditioning &
moisture collector
/"oven ''"•"'
—| Alum, probe
Porous polymer
adsorber
I _:
Imp/cooler
trace element
collector
Impinger
T.C.
Central lized
temperature
and pressure
readout
Control
module
10 cfm vacuum pump
Figure 57. Design C system schematic.
-------
TABLE 19. DESIGN C SUMMARY
Component
Probe
Static thermocouple
S-type pi tot
SS 316 nozzles (3)
2-1/2" aluminum sheath
SS 316 Uner
Local temp, control
Oven
Large oven
No circulation fan
Local temp, control
Cyclones
3 standard cyclones
AP Instrumentation
Filter Holder0
Standard filter, no Teflon
coat
Gas Cooler and Tenax Module
Standard design
Implnger System
Large glass bottles
Standard case
Vacuum Pump
Single standard pump
Umbilical Lines
Single power, thermocouple,
and pi tot connectors
*Compared to Design B
Design
Hours
4
28
24
60
Advantage
Lower cost than SS.
Eliminates remote controller
and connections.
Lower cost, more room in oven.
Eliminates remote controller
and connections.
Fewer lines needed on umbili-
cal, lower cost, simpler con-
nectors.
Disadvantage
Lower operating temperature.
Not as corrosion resistant.
Less convenient, less accu-
rate.
Longer warm-up.
Less convenient, less accu-
rate.
Interchangeability with stan- .
dard HVSS system is lost.
Component
Cost Change*
-44
-100
-20
-100
-90
o
-------
TABLE 19. Concluded
Component
Control and Metering Module
No dry gas meter
No orifice plates & meter
Timer
Pressure readouts
Temperature readouts
Power control
Design
Hours
180
Advantage
Significant simplification of
module.
Disadvantage
Measurements less accurate,
less central control, some
sacrifice In Interchanged-
bllity.
Component
Cost Change*
-1600
o
_J
rss
-------
Convection
oven
Gas Conditioning &
moisture col lector
/^—j Alum, probe
Porous polymer
adsorber
o
CO
Oven
thermometer
Cyclone
AP gage
Imp/cooler
trace elanent
col lector
Impinger
thermo,:ieter
t
f- I n rfin
I 0 cfin vacuum pump
Figure 58. Design D system schematic.
-------
train is less convenient and less accurate than Designs B or C; nearly every
component will require redesign (with its attendant costs); and the train
requires auxiliary support equipment (thermocouples and pitot tubes) to be
operated correctly.
Table 20 shows a summary of the design and cost features of Design D.
C.I.I Cost Summary
Table 21 presents a direct comparison of the design and construction
costs for each of the four designs, broken down by component. The basis for
preparing this table is an assumed production run of five units. So that the
actual total costs of the various designs may be compared, the bottom line of
Table 21 is a net cost per unit, obtained by adding the discounted component
cost and the pro rata design cost.
C.2 SCHEDULE
It is estimated that the design effort required to implement Designs
A, C, or D would require about 2 months. Construction and checkout time
would require approximately an additional 2-1/2 months, for a total program
length of about 4-1/2 months. Design B would require no additional time for
design and would require about 3 months for construction and checkout. (The
additional 1/2 month is required because the design and construction tasks
overlap for Designs A, C, and D.)
C.3 DISCUSSION
For each of the four designs considered, we have come up with a net
cost estimate that can be compared with the performance features of the
design, so that a cost/benefit analysis can be made. There are, however,
several less obvious factors that also should be considered.
C-14
-------
TABLE 20. DESIGN D SUMMARY
Component
Probe
No stack thermocouple
No pi tot tube
SS 316 nozzles (3)
Heat- traced SS 316 liner
Local temp, control
Oven
Large oven
No circulation fan
Local temp, control
Thermometers to measure oven,
gas temps.
Cyclones
3 standard cyclones
AP Instrumentation
Filter Holder
Standard filter, no Teflon coat
Gas Cooler and Tenax Module
Standard design except thermom-
eters except T.C. 's.
Umbilical
Power line only
Impinqer System
Large glass bottles
Standard case
Vacuum Pump
Single standard pump
Sample hose
^Compared to Design B
Design
Hours
130
24
16
-
24
16
-
-
Advantage
Considerably simpler, cheaper
than standard probe.
Low cost.
Low cost.
Disadvantage
Less accurate, less con-
venient.
Less accurate, less con-
venient.
Less accurate, less con-
venient.
Component
Cost Change*
-543
-120
-180
-400
o
-------
TABLE 20. Concluded
Component
Control and Metering Body
None
Timer
Separate timer box
Design
Hours
24
Advantage
Reduced cost.
Disadvantage
Reduced accuracy and con-
venience, local control of
all components required.
Component
Cost Change*
-4020
+40
o
_J
en
-------
TABLE 21. SUMMARY OF COSTS FOR SASS DESIGNS
Component
Nozzles
5' Probe
Oven
Cyclones
Filter Holder
Gas Cooler & Tenax Module
Impinger Trains
Umbilical
25' Sample Hose
Pump(s)
Control Module
Timer Unit
Unirail
Semiautomatic Controller
Subtotal
less 12X System Discount
Total Component Cost
Design Hours Required
Design Cost 0 $28/hr.
Net Cost Per Unit*
Unit Cost, Lots of 5
Design A
$ 265
1384
1500
6070
395
4000
3060
490
95
2098
4020
801
2500
26678
3201
23477
297
8316
25140
Design B
$ 159
919
1500
6070
385
4000
1230
490
95
924
4020
—
:_
19792
2375
17417
0
0
17417
_
Design C
$ 159
775
1250
6070
385
4000
1230
400
95
924
2420
—
IM
17798
2136
15662
296
8288
17320
_^^— — •— ^-^—
Design D
$ 159
376
1300
6070
385
4000
1230
90
95
924
—
40
—
____
14669
1760
12909
234
6552
14219
*0ne fifth the design cost plus the discounted component cost.
C-17
-------
In many cases, the organization that would purchase a SASS train would
have a continuing need for both SASS and HVSS train capabilities. Often the
potential SASS buyer will already have a HVSS train. It is clearly a desirable
feature that as many components as possible in the HVSS be compatible with
the SASS. This allows the user maximum flexibility in his sampling program,
and -- if separate HVSS and SASS trains are procured --allows one train to
be used as a backup and spare parts reserved for the other.
In many cases, the owner of HVSS train who desires to obtain a SASS
will wish to buy only enough SASS components to upgrade his HVSS. In this
case, the total cash outlay required to upgrade a HVSS is of interest. Table
22 shows the components that would have to be purchased to upgrade a HVSS to
each of the SASS designs. Design D, because it is a radically different
design than the HVSS, requires that an entirely new train be purchased -- the
upgraded cost is equivalent to the discounted SASS cost (Table 21). Design
C, which differs in several important ways from the HVSS, is also very costly
to upgrade; the upgrade cost of $14,194 is only moderately lower than the
SASS outright purchase price of $17,320.
For Design B, which was designed with HVSS compatibility specifically
in mind, the upgrade cost of $11,570 is significantly lower than the SASS
train cost of $17,417. Design A is also less costly to upgrade than to
purchase a complete SASS train, although -- because of the many convenience
features of Design A -- the upgrading cost is quite high. Table 23 shows the
total cost to an organization that already has an HVSS to upgrade it to a
SASS. It can be seen that the total cost for Design B is significantly lower
than for any other train, even the stripped-down Design D.
C-18
-------
TABLE 22. COSTS TO UPGRADE AN EXISTING HVSS
TO FULL SASS CAPABILITY
Components to be Purchased
Vycor Probe Liner
Additional Nozzles
5' Probe
Oven
Cyclones
Gas Cooler and Tenax Module
Impinger System
Semiautomatic Controller
Vacuum Pump
Umbilical Lines
Control and Metering Module
Umbilical Adaptors
Timer Unit
Total Upgrade Cost
Cost
Design A
424*
106
1500
6070
4000
2166*
3508*
1264*
19038
Design B
1500
6070
4000
11570
Design C
954*
1384*
6070
4000
736*
1050*t
14194
Design D
1104*
1524*
6070
4000
140
180*
174*
13192
*Includes pro rata design cost.
tAdapts modified umbilical to standard control module. Two different
adaptors required.
C-19
-------
TABLE 23. TOTAL INVESTMENT REQUIRED TO UPGRADE
A HVSS TO FULL SASS CAPABILITY
Design
Type
A
B
C
D
HVSS
Cost
9040
9040
9040
9040
Upgrade
Cost
19,038
11,570
14,194
13,192
Total
Cost
28,078
20,610
23,234
22,232
C-20
-------
Another factor that should be considered when deciding which design is
cost-effective is whether the SASS train will be used for some Level 2 source
assessment. The system requirements for Level 2 sampling are not as well
defined as for Level 1, but it seems reasonable to assume that measurement
and control of sample flowrates will be more stringent. This is particularly
significant when considering Designs C and D. In these designs, accurate
flow measurement has been eliminated, making these trains of doubtful utility
for Level 2 work. Designs A and B contain accurate flow monitoring
components.
C.4 RECOMMENDATION
After considering all factors -- design cost, component cost,
schedule, HVSS compatibility, versatility, and accuracy — we believe that
Design B is clearly superior to the others considered. We recommend that it
be the basis for SASS train development.
C-21
-------
APPENDIX D
CONVERSION FACTORS FOR NONMETRIC UNITS
USED IN THIS REPORT
cubic foot
day (mean
Fahrenheit
foot
gallon
grain
grain/
horsepower
hour
inch
To Convert From
:tual cubic feet/minute)
tish thermal unit)
)0t
in solar)
sit
ibic foot
ter
:an solar)
mercury (60°F)
water (60°F)
nd force)
iund mass
/inch2)
to
meter3/sec
joule
meter3
second
Celsius
meter
meter3
kilogram
kilogram/meter3
watt
second
meter
newton/meter2
newton/meter2
newton
kilogram
newton/meter2
Multiply by
4.719 x ID"4
1.0544 x 103
2.8317 x 10-2
8640
C = (5/9)(F-32)
0.3048
3.785 x ID'3
6.4799 x 10'5
2.2884 x ID"3
745.70
3600
2.54 x ID'2
3376.85
248.84
4.4482
0.4536
6894.76
scfm (standard cubic foot/
minute)
meter3/second
4.719 x 10~4
D-l
-------
REFERENCES
1. Dorsey, J. A., Johnson, L. D. Statnick, R. M. and Lochmuller, C. H.,
"Environmental Assessment Sampling and Analysis Phased Approach and
Techniques for Level 1," EPA-600/2-77-115, PB 268-563/AS, June 1977.
2. Hamersma, J. W., Reynolds, S. L., and Maddalone, R. F., "IERL-RTP
Procedures Manual: Level 1 Environmental Assessment," EPA-600/2-76-
160a, PB 257-850/AS, June 1976.
3. Blake, D., "Operating and Service Manual -- Source Assessment Sampling
System," Acurex/Aerotherm Report No. UM-77-80, EPA Contract No. 68-02
2154, February 1977.
4. Lapson, W. F., et. al., "The Development and Application of a High
Volume Sampling Train for Particulate Measurements of Stationary
Sources," presented at the 67th Annual meeting of the Air Pollution
Control Assoc., Denver, Colorado, June 9-13, 1974.
5. "Interim Report for Fabrication and Calibration of Series Cyclone
Sampling Train," TRW 24916-6019-TU-OO, EPA Contract No. 68-02-1412,
April 1975.
6. Perry, J. H. Chemical Engineers' Handbook. Third Ed., McGraw-Hill,
Inc., p. 1029, 1950.
7. Ancel, J. E., "Development of Cyclone for In-Stack Particulate Testing,"
M. S. Thesis #46556, University of Notre Dame, August 1973.
8. Clausen, J., et. al., "At Sea Incineration of Organic Chlorine Waste
on Board the M/T Vulcanus," EPA-600/2-77-196, September 1977.
R-l
-------
REFERENCES (Continued)
9. Gushing, K. M., et. al., "Particulate Sampling and Support," EPA-IERL
Monthly Technical Progress Narrative, No. 15, Contract 68-06-2131,
February 1977.
10. Perry, R. H. et. al., Chemical Engineers' Handbook. Fourth Ed., McGraw-
Hill, Inc., 1963.
R-2
-------
.„ TECHNICAL REPORT DATA
fftcase read Instructions on the reverse before completing)
EPA-600/7-78-Q18
2.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
ource Assessment Sampling System: Design and
Development
5. REPORT DATE
February 1978
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
D.E. Blake
8. PERFORMING ORGANIZATION REPORT NO.
.PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Corporation/Aerotherm Division
485 Clyde Avenue
Mountain View, California 94042
10. PROGRAM ELEMENT NO.
INE623
11. CONTRACT/GRANT NO.
68-02-2153
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 3/76-12/77
14. SPONSORING AGENCY CODE
EPA/600/13
15.SUPPLEMENTARY NOTES IERL-RTP project officer is William B. Kuykendal, Mail Drop 62,
ily/ 041—a uD i .
16. ABSTRACT
report chronologically describes the design and development of the
Source Assessment Sampling System (SASS). The SASS train is the principal sampling
element for ducted sources when performing EPA's Level 1 environmental assessment
studies. As such, it samples process streams and separates the samples into filter-
able particulate (four size fractions), organic vapors, and inorganic vapors. The
design concept and philosophy are discussed, as well as the evolutionary development
of the system. Developmental testing, problem areas, and subsequent system chan-
ges are described in detail. The report also includes a complete description of the
calibration procedures and system used to determine the size cut points of the parti-
culate fractionating cyclones used.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Sampling
Assessments
Industrial Processes
Dust
Vapors
Organic Compounds
Inorganic Com-
pounds
Calibrating
Cyclone Separa-
tors
Pollution Control
Stationary Sources
Source Assessment Sam-
pling System (SASS)
Environmental Assess-
ment
Particulate
13 B
14B
13H
11G
07D
Q7C
07B
07A
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
!1. NO. OF PAGES
221
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
R-3
------- |