November 2010
Environmental Technology
Verification Report
GAS IMAGING TECHNOLOGY, LLC
SHERLOCK® VOC
Prepared by
Battelle
Batfeiie
The Business of Innovation
Under a cooperative agreement with
U.S. Environmental Protection Agency
-------
November 2010
Environmental Technology Verification
Report
ETV Advanced Monitoring Systems Center
GAS IMAGING TECHNOLOGY, LLC
SHERLOCK® VOC CAMERA
by
Brian Boczek and Amy Dindal, Battelle
John McKernan, U.S. EPA
-------
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development,
funded and managed, or partially funded and collaborated in, the research described herein. It
has been subjected to the Agency's peer and administrative review process. Any opinions
expressed in this report are those of the author (s) and do not necessarily reflect the views of the
Agency, therefore, no official endorsement should be inferred. Any mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
11
-------
Foreword
The EPA is charged by Congress with protecting the nation's air, water, and land resources.
Under a mandate of national environmental laws, the Agency strives to formulate and implement
actions leading to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, the EPA's Office of Research and
Development provides data and science support that can be used to solve environmental
problems and to build the scientific knowledge base needed to manage our ecological resources
wisely, to understand how pollutants affect our health, and to prevent or reduce environmental
risks.
The Environmental Technology Verification (ETV) Program has been established by the EPA to
verify the performance characteristics of innovative environmental technology across all media
and to report this objective information to permitters, buyers, and users of the technology, thus
substantially accelerating the entrance of new environmental technologies into the marketplace.
Verification organizations oversee and report verification activities based on testing and quality
assurance protocols developed with input from major stakeholders and customer groups
associated with the technology area. ETV consists of six environmental technology centers.
Information about each of these centers can be found on the Internet at http://www.epa.gov/etv/.
Effective verifications of monitoring technologies are needed to assess environmental quality
and to supply cost and performance data to select the most appropriate technology for that
assessment. Under a cooperative agreement, Battelle has received EPA funding to plan,
coordinate, and conduct such verification tests for "Advanced Monitoring Systems for Air,
Water, and Soil" and report the results to the community at large. Information concerning this
specific environmental technology area can be found on the Internet at
http ://www. epa.gov/etv/centers/centerl .html.
in
-------
Acknowledgments
The authors wish to acknowledge the contribution of the many individuals, without whom, this
verification testing would not have been possible. Quality assurance (QA) oversight was
provided by Michelle Henderson, U.S. EPA, and Zachary Willenberg, Battelle. We thank Mr.
David Fashimpaur of BP, for hosting the laboratory testing phase of this verification test at the
BP, Naperville, IL research complex. Also, we acknowledge the support of Mr. Jeffrey Panek
and Dr. Paul Drayton of Innovative Environmental Solutions, Inc. for operating the leak
generation equipment and performing data collection during the laboratory testing phase. We
thank Ms. Julie Woodard, Ms. Fran Quinlan Falcon, and Mr. Barry Kelley of the Dow Chemical
Company for providing the field test site and for supporting the verification test team during the
field testing phase. We gratefully acknowledge the support of the American Chemistry Council
(ACC) and the Texas Chemical Council (TCC) as collaborators to this verification test and
would like to specifically thank Mr. Jim Griffin (ACC) and Ms. Christina Wisdom (TCC) for
their personal dedication to this verification test. Finally, we want thank Mr. David Williams
and Mr. Eben Thoma of the U.S. EPA for their review of the test/QA plan and/or this verification
report.
IV
-------
Contents
Page
Foreword iii
Acknowledgments iv
List of Abbreviations viii
Chapter 1 Background 9
Chapter 2 Technology Description 10
Chapters Test Design and Procedures 11
3.1 Test Overview 11
3.2 Experimental Design 13
3.2.1 Detection of a Chemical Gas Leak Using the Sherlock VOC Imaging
Spectrometer 13
3.2.2 Method Detection Limit 13
3.2.3 Confounding Factors 14
3.2.4 Detection of a Chemical Gas Species Relative to a Portable Monitoring Device.. 16
3.2.5 Field Testing Procedures 17
3.3 Qualitative Evaluation Parameters 17
Chapter 4 Quality Assurance/Quality Control 18
4.1 Reference Method Quality Control 18
4.1.1 Bias and Accuracy of Sample Screening Measurements Using Portable Monitoring
Device 19
4.1.2 Confirmation of Detected Leaks 20
4.1.3 Bias and Accuracy of Enclosure Equilibration Gas 20
4.1.4 Bias and Accuracy of Bagging Procedure 22
4.1.5 Bias and Accuracy of Gas Chromatography Analytical Method 22
4.2 Audits 23
4.2.1 Technical Systems Audit 23
4.2.2 Data Quality Audit 25
Chapters Statistical Methods 26
5.1 Method Detection Limit 26
5.2 Percent Agreement 26
Chapter 6 Test Results 28
6.1 Method Detection Limit 28
6.2 Detection Agreement to a Portable Monitoring Device 28
6.2.1 Laboratory Testing 33
6.2.2 Field Testing 33
6.3 Confounding Factors 37
6.4 Operational Factors 38
Chapter 7 Performance Summary 39
Chapter 8 References 42
-------
Tables
Chemical Leaks Evaluated with the GIT Sherlock® VOC Imaging Spectrometer
During Laboratory Testing [[[ 14
Test Conditions Evaluated During Laboratory Testing ................................................. 15
TVA Calibration Responses [[[ 20
TVA Calibration Check Samples [[[ 21
Confirmation of Detected Leaks by TVA [[[ 21
Known Leak Rate Test Results [[[ 22
Summary of Positive Control Check Responses [[[ 24
Sherlock® VOC Method Detection Limits at 10 feet with Cement
Board Background [[[ 29
Sherlock® VOC Method Detection Limits at 30 feet with Cement
Board Background [[[ 30
. Sherlock® VOC Method Detection Limits at 10 feet with Gas
Cylinder Background [[[ 31
. Sherlock® VOC Method Detection Limits at 30 feet with Gas
-------
FIGURES
Figure 1. Sherlock® VOC Camera 10
vn
-------
List of Abbreviations
ACC
AMS
CH4
DQI
EPA
ETV
ft
f/
GC
GIT
g/hr
IR
kg/hr
LCD
LOD
Mph
NRMRL
PID
ppmv
QA
QC
QMP
Sherlock® VOC
TCC
TQAP
TVA
U.S.
VOC
American Chemistry Council
Advanced Monitoring Systems
Methane
Data Quality Indicator
Environmental Protection Agency
Environmental Technology Verification
Feet, foot
Focal ratio, F-number
Gas Chromatography
Gas Imaging Technology, LLC.
Grams per hour
Infrared
Kilogram per hour
Liquid crystal display
Limit of Detection
Miles per hour
National Risk Management Research Laboratory
photoionization
Parts per million by volume
Quality assurance
Quality control
Quality Management Plan
Sherlock® VOC imaging spectrometer
Texas Chemical Council
Test Quality Assurance Plan
Toxic Vapor Analyzer
United States
Volatile organic compounds
Degrees Fahrenheit
Vlll
-------
Chapter 1
Background
The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental
technologies through performance verification and dissemination of information. The goal of the
ETV Program is to further environmental protection by accelerating the acceptance and use of
improved and cost-effective technologies. ETV seeks to achieve this goal by providing high-
quality, peer-reviewed data on technology performance to those involved in the design,
distribution, financing, permitting, purchase, and use of environmental technologies.
ETV works in partnership with recognized testing organizations; with stakeholder groups
consisting of buyers, vendor organizations, and permitters; and with the full participation of
individual technology developers. The program evaluates the performance of innovative
technologies by developing test plans that are responsive to the needs of stakeholders,
conducting field or laboratory tests (as appropriate), collecting and analyzing data, and preparing
peer-reviewed reports. All evaluations are conducted in accordance with rigorous quality
assurance (QA) protocols to ensure that data of known and adequate quality are generated and
that the results are defensible. The definition of ETV verification is to establish the performance
of a technology under specific, pre-determined criteria or protocols and a strong quality
management system. High quality data are assured through the implementation of the ETV
Quality Management Plan. ETV does not endorse, certify, or approve technologies.
The EPA's National Risk Management Research Laboratory (NRMRL) and its verification
organization partner, Battelle, operate the Advanced Monitoring Systems (AMS) Center under
ETV. The AMS Center recently evaluated the performance of the Sherlock VOC imaging
spectrometer by Gas Imaging Technology, LLC. (GIT), a portable, passive infrared (IR) imaging
spectrometer operating in the spectral range of 3 to 5 micrometers
-------
Chapter 2
Technology Description
This verification report provides results for the verification testing of GIT's Sherlock® VOC
imaging spectrometer. Following is a description of the GIT Sherlock® VOC imaging
spectrometer technology (hereafter referred to as the Sherlock VOC), based on information
provided by the vendor. The information provided below was not verified in this test. Figure 1
shows the Sherlock® VOC imaging spectrometer.
The Sherlock® VOC is an infrared optical imaging
instrument for video imaging of gas leaks. The Sherlock®
VOC is man portable and battery operated. The Sherlock®
VOC is based on patented Image multi spectral Sensor
imaging technology.
The Sherlock VOC can be used for IR imaging purposes in
many types of industries: oil, gas, chemical, power
generation, mining, pulp and paper, to name a few.
The Sherlock VOC has a single focal ratio (F-number or f/)
of 2.5 in a 75-millimeter (mm) focal length lens that is
embedded in the body of the instrument. The horizontal
field of view is approximately seven degrees. The ™ t c, , t®\7f\/^
cu 1 i«vr»r« u • ju Figure 1. Sherlock VOC
Sherlock VOC can be carried by an operator using an P
optional EasyRig harness which enables pointing and
scanning while looking at the liquid crystal display (LCD). This design leaves the operator with
both eyes free to watch for safety hazards in the environment of a processing plant.
In addition to displaying the infrared image to the operator on a hooded LCD display the
Sherlock® VOC can store both 14-bit digital video clips on an embedded frame grabber or to an
external 8-bit digital video recorder.
10
-------
Chapter 3
Test Design and Procedures
3.1 Test Overview
This verification test was conducted according to procedures specified in the Test/QA Plan for
Verification of Leak Detection and Repair Technologies^\TQAP) and adhered to the quality
system defined in the ETV AMS Center Quality Management Plan (QMP).(2) Battelle conducted
this verification test with support from British Petroleum (BP), Innovative Environmental
Solutions, Inc., The Dow Chemical Company, Sage Environmental Consulting, and Enthalpy
Analytical, Inc.
This verification test simulated gas leaks of various chemicals in a controlled laboratory
environment. The ability of the Sherlock® VOC imaging spectrometer to qualitatively detect gas
leaks of select chemicals species by visual images under controlled environmental conditions -
including varied stand-off distances, wind speeds, and background materials - was verified and
the method detection limits under each test condition were determined. This passive IR camera
has not been evaluated under the ETV Program for other compounds or species other than those
tested under this verification test. The potential exists for the identification of other species that
have an IR absorbance feature(s) in this spectral range under ideal test conditions.
Additionally during laboratory testing, the ability of the Sherlock VOC imaging spectrometer to
qualitatively detect the gas leak by visual images relative to a quantitative concentration
measurement made by a portable monitoring device acceptable under U.S. EPA Method 21 -
Determination of Volatile Organic Compound (VOC) Leaks^ for the determination of VOC
leaks from process equipment was verified for each chemical at each test condition during
laboratory testing. During laboratory testing, "acceptable under U.S. EPA Method 21" meant
that the portable monitoring device used met all of the performance requirements of Section 6 in
U.S. EPA Method 21 with the exception of those requirements related to a specific leak
definition concentration specified in any applicable regulation. A specific leak definition
concentration was not used to qualify leaks during laboratory testing in a regulatory sense.
This verification test also verified the ability of the Sherlock® VOC imaging spectrometer to
detect gas leaks of various chemicals relative to a portable monitoring device acceptable under
U.S. EPA Method 21 under "real world" conditions at a chemical plant in Freeport, TX. During
field testing, "acceptable under U.S. EPA Method 21" meant that the portable monitoring device
used met all of the performance requirements of Section 6 in U.S. EPA Method 21; a specific
leak definition concentration of 500 parts per million by volume (ppmv) was utilized. Reference
sampling was conducted to determine the mass rate of specific chemical species emitted from
each leaking component observed with the Sherlock VOC imaging spectrometer and with the
portable monitoring device acceptable under U.S. EPA Method 21.
11
-------
This verification test of the Sherlock VOC imaging spectrometer was conducted October 20
through October 24, 2008 at the BP research complex in Naperville, Illinois (laboratory testing)
and December 1 through December 5, 2008 at the Dow Chemical Company plants (field testing)
in Freeport, TX in compliance with the data quality requirements in the AMS Center Quality
Management Plan (QMP). The TQAP for this verification test indicated that field testing would
be conducted at two field sites. Due to production scheduling issues, a second field site could
not be obtained in a timely manner and this verification test was completed using only one field
test location. Confirmation from a second field site was obtained during the writing of these
reports and field testing occurred outside of this verification test in March 2010. The reader is
encouraged to contact either Gas Imaging Technology or the Texas Chemical Council (TCC) to
obtain the results of testing completed at the second field site. As indicated in the test/QA plan,
the testing conducted satisfied EPA QA Category III requirements. The test/QA plan, the
verification statement, and this verification report were reviewed by the following experts.
• Dave Fashimpaur, BP
• Julie Woodward, Dow Chemical
• Jim Griffin, American Chemistry Council
• Christina Wisdom, Texas Chemical Council
• EbenThoma, U.S. EPA.
One technical expert came to the laboratory testing, and one technical expert came to the field
site to observe testing. Verification testing was conducted by appropriately trained personnel
following the safety and health guidelines for BP and Dow's facilities.
The Sherlock VOC imaging spectrometer was verified by evaluating the following parameters:
• Method detection limit - The minimum mass leak rate that three separate individuals can
observe using the imaging spectrometer under controlled laboratory conditions. This
parameter was not evaluated during the field testing phase.
• Detection of chemical gas species relative to a portable monitoring device- The ability of
the imaging spectrometer to qualitatively detect a gas leak by visual images relative to a
quantitative concentration measurement made by a portable monitoring device acceptable
under U.S. EPA Method 21. This parameter was evaluated in both the laboratory and
field testing phases.
• Confounding factors effect - Background materials, wind speed, and stand-off distance
were carefully controlled during laboratory testing to observe their effects on the method
detection limit. During field testing, these variables as well as meteorological conditions
were recorded.
• Operational factors - Factors such as ease of use, technology cost, user-friendliness of
vendor software, and troubleshooting/downtime were evaluated.
Due to unavailability of a second Sherlock® VOC imaging spectrometer during the laboratory
and field testing portions of this verification test, inter-unit comparability could not be completed
during laboratory and field testing.
12
-------
Prior to the start of the verification test, personnel from Industrial Scientific Corporation (a
vendor licensed to distribute the Sherlock® VOC imaging spectrometer) setup the imaging
spectrometer according to their recommended configuration for optimal performance.
3.2 Experimental Design
®
3.2.1 Detection of a Chemical Gas Leak Using the Sherlock VOC Imaging Spectrometer
During both the laboratory testing and field testing, the Sherlock® VOC imaging spectrometer
was operated by a representative of Industrial Scientific Corporation. This verification test used
two additional confirming individuals beyond the camera operator to confirm the observation of
a leak in an effort to eliminate potential operator bias. The two additional confirming individuals
were the Battelle verification test coordinator and an additional verification test team member.
The use of three individuals to observe a chemical leak with the imaging spectrometer is not
standard practice when using the imaging spectrometer; typical operation relies on a single
camera operator to observe the presence of a chemical gas leak.
The detection of a chemical gas leak in either the laboratory or field setting was determined by
the camera operator, as well as, two confirming individuals who reported the results qualitatively
as either "detect" or "non-detect" observation. All three individuals must have agreed on the
results for the observation to be considered a "detect." When all three individuals did not agree
on a detection, the observation was reported as a "non-detect." A non-detect was also recorded if
the camera operator did not observe a detection (i.e., no confirmation of a non-detect was
performed). Each observation was conducted using the viewing screen of the imaging
spectrometer.
The TQAP for this verification test required that camera observers have five seconds to identify
the origin of the leak or be able to track the plume back to the leaking component when
observing chemical gas leaks (i.e., identify the source of the leak). However, during laboratory
and field testing the observers were allowed two minutes. This change was made during
laboratory testing to account for system hysteresis and upon discovering that several liquid
compounds at very low flow rates did not generate a continuous plume. Rather, the leaks were
observable as intermittent "puffs" of chemicals emanating from the valve at a frequency on the
order of 10 seconds to two minutes. This lag resulted at lower syringe pump feed rate settings
and the reduced hot nitrogen carrier gas volume flow rates.
3. 2. 2 Method Detection Limit
Method detection limits were determined only in the laboratory portion of this verification test.
To determine the method detection limit, a known mass leak rate from the packing of a 1-inch
valve attached to certified gas cylinders and calibrated flow meters was set at a nominally
detectable level either specified by the vendor's limit of detection (LOD) for a particular test
condition, or based on previous literature by Panek et al.(4) When all three observers identified
the leak, the leak rate was reduced by the testing staff using calibrated flow meters. Once a leak
rate that was not identifiable by all three people was reached, the mass emission rate was again
increased using the calibrated flow meters to the level where all three could again identify the
leak using the passive infrared imager. This rate was the method detection limit for the
Sherlock® VOC imaging spectrometer under the tested conditions. This process was completed
for every testing trial identified in Section 3.2.3. Table 1 identifies the type of chemical leaks
evaluated with the Sherlock® VOC imaging spectrometer during laboratory testing.
13
-------
The TQAP for this verification test stated that 1,2-dichloropropane (i.e., propylene dichloride)
and hydrochloric acid also would be used during laboratory testing . The stock solution of
propylene dichloride was suspected by laboratory personnel of having been cross-contaminated
by a different chemical compound. A second stock solution of propylene dichloride could not be
obtained from a chemical vendor before the conclusion of laboratory testing. Thus, propylene
dichloride was not used during laboratory testing. The laboratory staff also expressed concerns
of causing damage to the delivery syringe in the chemical delivery system with the use of
hydrochloric acid. Because hydrochloric acid could not be delivered through the chemical
delivery system without causing damage to the system, a known leak rate could not be generated
during laboratory analysis, therefore hydrochloric acid was not evaluated.
Table 1. Chemical Leaks Evaluated with the GIT Sherlock® VOC Imaging Spectrometer
During Laboratory Testing
Chemical Chemical Group
1,3-butadiene Olefm
Acetic acid Acetate
Acrylic acid Acid
Benzene Aromatic
Dichloromethane Chlorinated
(methylene chloride)
Ethylene Olefm
Methane Alkane
Methanol Alcohol
Pentane Alkane
Propane Alkane
Styrene Aromatic
3.2.3 Confounding Factors
Because passive IR imagers rely on the physical characteristics of the environment and the
molecules being imaged to create an image viewed by the operator (via temperature/emissivity
differences between naturally occurring ambient IR radiation and the thermal emission or
absorption of the leaking gas), environmental characteristics may confound the measurement.
For example, if there is not sufficient thermal emission or absorption by the leaking gas, the
passive IR imager may not be able to detect a leak against an ambient thermal background.
During laboratory testing, experimental factors of background materials, wind speed, and stand-
off distance were altered for each chemical tested. These experimental factors were chosen,
because the performance of passive imagers is dependent on physical characteristics of the leak,
atmospheric conditions, and background materials. The change of background material
demonstrates the ability of the Sherlock® VOC imaging spectrometer to passively detect the leak
with a background scene similar to petrochemical process piping and vessels (curved metal gas
cylinders) and with a background that is different than the leaking component and more uniform
in nature (cement board - representing control buildings, sidewalks, and other uniform flat
background surfaces). The wind speed variations and the stand-off distances inform on the
atmospheric and optical pathway effects on the method detection limit, and in turn on real-world
limitations. Table 2 presents the specific test conditions evaluated during laboratory testing.
14
-------
It was originally intended that all test conditions would be completed for all chemicals; however,
it was not possible for 1,3-butadiene, acrylic acid, methylene chloride, methane, and styrene for
the following reasons.
Previous testing of the Sherlock® VOC imaging spectrometer using methane had been
completed by the laboratory facility outside of the verification test. Consequently, methane was
used during test equipment setup to confirm that the equipment produced method detection limits
for methane that were consistent with those produced during previous testing by the laboratory.
Table 2. Test Conditions Evaluated During Laboratory Testing
Chemical Species
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methane
Methanol
Pentane
Propane
Styrene
Laboratory Test Conditions
y
1
—
^g
43
a
g
-off distance; 0-
ard background
" O
g 42
•§ -g
4= |
o S
'
^
^
^
^
^
^
^
^
^
'
y
ITS
•S "I
•5 0
r-] 'vj
a y
SŁ
-off distance; 0-
tal gas cylinder
^ GJ
^ ^
•§ T3
4n ^
o id
^H O
^
^
'
^
^
^
^
^
^
^
^
1
T3
g
'^
^
43
&
s
1
•n
-off distance; 2.
ard background
" O
a f,
•§ -g
S-H G
o S
^H O
'
^
^
^
^
^
^
^
1
"
.S c
> s
43 2
a ^
g o
•A 4D
-off distance; 2.
tal gas cylinder
^ ^
•§ T3
4n ^
O S
^H O
^
^
^
'
^
^
^
y
1
—
f^
•r;
43
a
S
-off distance; 5-
ard background
CH o
•g g
^H g
O bfi
o O
s j§
-off distance; 5-
tal gas cylinder
^ ^
•§ T3
j^
o s
^H O
^
^
'
^
^
y
1
—
f^
•r;
43
a
S
-off distance; 0-
ard background
CH o
•g g
^H g
o S
m o
^
^
'
^
^
^
^
^
^
^
y
ITS
•S "I
•5 0
^r-] _ 'vj
QH ^
S 42
-off distance; 0-
tal gas cylinder
^ ^
tn "«
4n ^
0 S
r*~i o
^
^
'
^
^
^
^
^
^
^
1
T3
g
'^
^
4^
&
S
1
•n
-off distance; 2.
ard background
CH o
•g g
4= |
O CL>
m o
^
^
^
'
^
^
1
&
?H rt
'fe ^
2
a ^
g o
>A 4D
-off distance; 2.
tal gas cylinder
^ ^
•§ T3
4n ^
0 S
m o
'
^
^
^
^
^
y
1
—
^c
43
a
g
-off distance; 5-
ard background
CH o
•g g
4= |
O CL>
m o
^
^
^
'
^
y
ITS
•S "i
« s
'S 2
^ 60
a ^
S ^
-off distance; 5-
tal gas cylinder
T3 (D
a g
•2 ^
4n f*
o id
m o
^
^
^
^
^
The 2.5 and 5-mph wind speed test conditions were not completed for acrylic acid. After
completion of the 0-mph wind speed test condition, laboratory personnel indicated that the
acrylic acid was dissolving the rubber plunger gasket in the liquid delivery syringe in the vapor
generator system. Laboratory personnel indicated that the rubber seemed to be "dissolving"
inside the syringe and the syringe was no longer providing a steady flow of acrylic acid into the
chemical delivery system. Additional testing using this compound was abandoned due to safety
and chemical handling concerns.
15
-------
The 2.5 and 5-mph wind speed test conditions were not completed for 1,3-butadiene and styrene
due to safety and potential exposure concerns. During laboratory setup the week prior to
verification testing, the exhaust of the test apparatus, which feeds into the general laboratory
building exhaust, was balanced and smoke tested to ensure that compounds leaking from the
system were captured in either the vertical hood canopy mounted over the leaking component or
the downwind hood mounted adjacent to the test system. Unbeknownst to laboratory personnel,
the building general exhaust system was operating at a lower setting during air balancing and
smoke testing due to decreased occupancy in the building. During the week of the test, the
general building exhaust was increased due to the presence of the test compounds entering the
exhaust system. The change in building exhaust flows caused the capture of the chemical
compound by the overhead hood and the hood mounted next to the test system to decrease.
Neither the vendor nor laboratory personnel had documentation of respirator fit-testing, so
respirators could not be used during testing.
To address this problem, the leaking valve was placed next to the side hood during wind speed
testing and testing of those chemical compounds which are liquids at standard conditions
commenced in order of increasing boiling point. Upon completion of wind speed testing for
acetic acid, the laboratory had a slight odor of acetic acid. This indicated to laboratory personnel
that locating the leaking valve next to the side hood during wind speed testing did not adequately
capture all of the chemical compounds exhausting from the test system. At this point, wind
speed testing of 1,3-butadiene and styrene was abandoned because these compounds have higher
chemical toxicity and exposure by the verification test team, vendor, and laboratory staff to these
compounds would have occurred during wind speed testing.
During benzene, methylene chloride, and ethylene testing, several of the wind speed tests and
background tests were not conducted because the method detection limit for lower wind speed
(or background) conditions exceeded the highest reliable flow rate capable of being provided by
the chemical leak delivery system at test conditions which were expected to produce a lower
method detection limit (refer to Section 6.3 for discussion of the observed influence of
confounding factors). For example, a 5-mph wind speed test was not conducted at a 10 ft stand-
off distance with a cement board background with benzene because the method detection limit
exceeded the highest reliable flow rate of the chemical delivery system for the 10 ft stand-off
distance, cement board background, and 2.5-mph wind speed.
3.2.4 Detection of a Chemical Gas Species Relative to a Portable Monitoring Device
The detection of a single chemical gas leak in either the laboratory or field environments was
determined by the operator as well as two confirming individuals as previously described in
Section 3.2.1 and reported qualitatively as either "detect" or "non-detect."
During laboratory testing a portable monitoring device acceptable under U.S. EPA Method 21
sampled the leak after the method detection limit was determined for the specified test
conditions. During laboratory testing, "acceptable under U.S. EPA Method 21" meant that the
portable monitoring device used met all of the performance requirements of Section 6 in U.S.
EPA Method 21 with the exception of those requirements related to a specific leak definition
concentration specified in any applicable regulation. A specific leak definition concentration
was not used to qualify leaks during laboratory testing in a regulatory sense. The portable
monitoring device used during laboratory testing was an Industrial Scientific IBRTD MX6 with
photoionization (PID) sensor and SP6 motorized sampling pump which was supplied calibrated
from the instrument supplier; no additional calibrations were performed during laboratory
16
-------
testing. During field testing a portable monitoring device acceptable under U.S. EPA Method 21
was used to screen each leaking component as part of the bagging reference method used.
During field testing, "acceptable under U.S. EPA Method 21" meant that the portable monitoring
device used met all of the performance requirements of Section 6 in U.S. EPA Method 21; a
specific leak concentration of 500 ppmv was utilized. The portable monitoring device used
during field testing was a Thermo-Environmental Toxic Vapor Analyzer (TVA).
3.2.5 Field Testing Procedures
Field testing was conducted to allow for performance evaluation under "real world" conditions.
Chemicals that were tested in the laboratory were targeted for evaluation at the field sites. The
flow rates of field leaks were quantitatively determined by a reference method called EPA
Protocol for Equipment Leak Emission Estimates^, which is also referred to as the "bagging
method." Specific details and procedures for this reference method are provided in the TQAP
for this verification test. This method involves completely enclosing the leak with non-
permeable material, collecting the leak with ambient air entering the bag, and performing mass
measurement of the bagged leak by an analytical method. Only those leaks above the field test-
assigned 500 ppmv leak definition concentration, as measured by the Thermo-Environmental
TVA, were observed with the passive infrared imagers and collected as reference samples under
this verification test.
The verification test team moved through the plant screening for possible leaking components
using a Thermo-Environmental TVA, a portable monitoring device acceptable under U.S. EPA
Method 21. Once a leak was detected with the portable monitoring device, leak characteristics
and environmental factors (i.e., type of component, background material, data and time,
temperature, etc.) were recorded qualitatively. Meteorological data was retrieved from the
nearest meteorological data station, the Dow Chemical Company's on-site weather station. As
space permitted, the camera operator took readings at three stand-off distances (10, 30, and
greater than 30 ft if possible). Every reading was verified by an additional two confirming
individuals and recorded as either "detect" or "non-detect" as specified in Section 3.2.1. Once
all cameras had scanned the leak, the bagging team members (Sage Environmental Consulting)
commenced collecting duplicate reference samples of the leak into evacuated SUMMA canisters.
Reference sampling concluded with a final screening by the portable monitoring device
acceptable under U.S. EPA Method 21 to verify that the leak concentration had not changed from
the beginning to the end of testing the component. Only those leaks which showed less than a
20% difference between the pre- and post-screening with the portable monitoring device were
considered consistent enough to report in the results without a data qualifier. The concentration
of the collected reference samples was determined according to the analytical method in U.S.
EPA Method 18 -Measurement of Gaseous Organic Compound Emissions by Gas
Chromatography.^ Upon conclusion of the five days of field testing, all reference samples were
shipped to Enthalpy Analytical, Inc. for U.S. EPA Method 18 analysis.
3.3 Qualitative Evaluation Parameters
Operational factors such as maintenance needs, ease of use, data output, and software
requirements were documented based on observations by Battelle.
17
-------
Chapter 4
Quality Assurance/Quality Control
QA/quality control (QC) procedures were performed in accordance with the QMP for the AMS
Center and the TQAP for this verification test. As noted throughout Chapter 3, there were
deviations from the TQAP, but the work was performed as described in the previous sections.
None of the deviations from the test/QA plan resulted in any adverse impacts on the quality of
the data produced by this verification test. QA/QC procedures and results are described in the
following subchapters.
4.1 Reference Method Quality Control
Laboratory testing did not use a specified reference method for determining the leak rate of the
test conditions. Rather, certified gas cylinders and laboratory grade liquid compounds were used
with calibrated flow meters and a calibrated syringe pump to generate a known leak rate in terms
of mass per unit time from the leaking valve. As a laboratory QC measure, laboratory personnel,
randomly and without the knowledge of the camera operator of the additional confirming
individuals, increased or decreased the mass leak rate to reduce the opportunity to predetermine
an outcome. In addition, laboratory blanks (i.e., pure nitrogen gas) and replicate tests were used
to reduce uncertainties and verify method detection limits established in prior tests.
The field testing portion of this verification test used accepted methods to generate reference
samples. Reference samples were collected using EPA Protocol for Equipment Leak Emission
Estimates and the concentrations of compounds in the collected reference samples were
determined according to the analytical method in U.S. EPA Method 18 Measurement of Gaseous
Organic Compound Emissions by Gas Chromatography
The quality of the reference measurements collected during field testing was assured by
adherence to the requirements of the data quality indicators (DQIs) and criteria for the reference
collection and analytical method critical measurements, including requirements to perform initial
calibrations and calibration checks of the portable monitoring device acceptable under U.S. EPA
Method 21, confirming the leak rates changed less than 20% before and after bagging, assessing
the bias and accuracy of the bagging procedure, and assessing the bias and accuracy of the gas
chromatography (GC) laboratory analysis by developing calibration curves traceable to certified
gas standards, and performing positive and negative control checks. The following sections
present key data quality results from these methods.
18
-------
4.1.1 Bias and Accuracy of Sample Screening Measurements Using Portable Monitoring
Device
A DQI is established in the TQAP for this verification test for the bias and accuracy of sample
screening measurements using a portable monitoring device. This DQI is assessed by
performing calibrations of the Thermo-Environmental TVA used to screen leaking components
during the field portion of the verification test and analyzing calibration check samples. During
laboratory testing the portable monitoring device was an Industrial Scientific IBRID MX6 with
PID sensor and SP6 motorized sampling pump which was supplied calibrated from the
instrument supplier; per the TQAP for this verification test, no additional calibrations were
performed during laboratory testing.
Calibration of the TVA was conducted using various levels of certified methane (CH/^-in-air gas
standards. The TQAP for this verification test required the use of five calibration points (an un-
spiked gas standard plus four additional concentrations); however, only three additional gas
standard concentrations were obtained. Because component leaks were only bagged as reference
samples if their concentration was greater than 500 ppmv and because the calibration response of
the TVA was evaluated using an un-spiked gas standard (0 ppmv) and three additional
concentrations of gas standards (500, 1000, and 9600 ppmv) thereby bounding the 500 ppmv
reference sample bagging threshold, there was no effect on data quality.
The calibration response of the TVA was analyzed at the start and end of each verification test
day or if the overall TVA sensitivity changed by greater than 10% (based on the calibration
check data, which are presented in Table 5). The minimum acceptance criterion for this
reference method DQI was that the TVA calibration response must agree within 10% of the
concentration of each gas standard. Table 3 presents the results of all TVA calibration responses
collected during this verification test. Inspection of the data present in Table 3 shows that all
calibration response measurements were confirmed to be within 10% of the calibration gas
standard concentration.
The TQAP for this verification test required that a calibration check sample be analyzed using
one concentration of the calibration gas standards at a minimum frequency of 5% of all bagged
reference samples collected. Sixteen calibration check samples were analyzed with the TVA
during the course of field testing and nine duplicate reference samples were collected resulting in
a calibration check sample frequency of 178% of all bagged reference samples collected (i.e., 16
calibration check samples completed during the collection of nine duplicate reference samples).
These checks were performed more frequently to ensure no drifting of the instrument occurred
during downtimes to ensure optimum performance. The minimum acceptance criterion specified
in the TQAP for this verification test is that the check standard must be within less than or equal
to a 10% change in response from the previous calibration of the TVA. If the calibration check
sample showed a change in response greater than 10%, then recalibration of the TVA was
performed and any affected reference sample components collected would be rescreened.
During this verification test, calibration check samples were performed using a certified 500
ppmv CH/j-in-air gas standard. Table 4 presents the results of all calibration check standards
performed during verification testing. Inspection of the data presented in Table 4 indicate that
reference samples 08 A and 08B should have been rescreened after recalibration of the TVA and,
therefore, are considered suspect data and reported with a data qualifier.
19
-------
Table 3. TVA Calibration Responses
Date [Time]
12/1/2008 [13:33](a)
12/2/2008 [09:01]
12/2/2008 [14:08]
12/2/2008 [16:05]
12/3/2008 [08:41]
12/3/2008 [09:30]
12/3/2008 [10: 12]
12/3/2008 [17:06]
12/4/2008 [10:04]
12/4/2008 [13:20]
12/4/2008 [16: 12]
12/4/2008 [17:23]
12/5/2008 [08:59]
12/5/2008 [11:20]
12/5/2008 [14:01]
Calibration
0
TVA Output
Concentration
(ppmv CH4)(b)
0.70
0.40
1.0
1.0
0.80
0.70
0.80
0.60
0.60
ND
0.60
0.20
0.60
1.2
0.20
Gas Standard Concentration (ppmv
500
TVA
0.40
-0.80
1.2
5.6
-1.4
-0.60
-1.2
-7.2
-0.60
ND
-0.80
-1.4
ND
4.0
3.4
1000
Calibration Response (as %
-1.3
-0.10
11
4.2
ND
-4.4
-0.60
-8.2
-0.30
-0.10
-1.5
-1.7
-0.70
3.0
3.3
CH4)
9600
Error)(c)
-0.80
-0.60
2.1
4.2
-0.70
-4.9
0.10
-8.0
-1.0
-0.30
-1.0
-1.1
-0.70
-8.3
-3.1
(a) An end-of-day TVA response was not collected on 12/1/2008. Data for leak location 1 is included but flagged
because there are acceptable reference and bagging measurements.
(b) Concentration data presented for un-spiked gas standard, since % error calculation is not possible. This point is
used in calibrating the Thermo-Environmental TVA.
(c) Percent error is calculated as [(TVA calibration response, ppmv CH4 - Calibration Gas Standard Concentration,
ppmv CH4)/ Calibration Gas Standard Concentration, ppmv CH4] x 100%.
ND - Not detected
4.1.2 Confirmation of Detected Leaks
A DQI is established in the TQAP for this verification test for the confirmation of detected leaks.
This DQI is assessed by analyzing the concentration of a leaking component before and after
bagging the component. These measurements were completed for all leaking components which
were bagged and collected as reference samples. The acceptance criterion for this DQI is that
the pre and post screening measurements collected with the TVA agree within 20%. Table 5
presents the results of all pre- and post-bagging measurements completed during the collection of
reference samples.
4.1.3 Bias and Accuracy of Enclosure Equilibration Gas
A DQI is established in the TQAP for this verification test for bias and accuracy of the enclosure
equilibration gas. This DQI requires that if the blow-through bagging procedure is used to
collect reference samples, then the equilibration gas in the bag is collected and analyzed for
contamination prior to collection of reference samples. During the verification testing, reference
samples were collected using the vacuum-method which does not require the use of an
equilibration gas; therefore, this DQI was not applicable.
20
-------
Table 4. TVA Calibration Check Samples
Date [Time]
Calibration Check
Response
(as % Error)(a)
Comments
12/2/2008 [11:17]
12/2/2008 [12:15]
12/2/2008 [14:05]
12/2/2008 [14:08]
12/2/2008 [15:10]
12/2/2008 [15:43]
12/3/2008 [9:23]
12/3/2008
12/3/2008
12/3/2008
12/3/2008
12/4/2008
12/4/2008
[10:30]
[11:32]
[13:57]
[15:45]
[11:43]
[13:23]
0.40
-5.2
-16
1.2
1.4
2.0
64
0.80
-0.60
0.60
0.60
1.6
-17
Recalibration only. No rescreening necessary because no
reference samples had been collected between this
calibration check sample and TVA calibration.
Found leak; recalibrated only. No rescreening necessary
because reference samples had yet to be collected this
day.
Recalibration only. No rescreening necessary because no
reference samples had been collected between this
calibration check sample and the previous check.
12/4/2008 [15:30]
12/4/2008 [17:25]
12/5/2008 [10:38]
24
-1.4
-3.0
Recalibration only. Reference samples 08A and 08B
were inadvertently not rescreened and are therefore
considered suspect and results reported with qualifier.
(a) Percent (%) error is calculated as [(TVA calibration check response, ppmv CH4 - Calibration Gas Standard
Concentration, 500 ppmv CH4)/ Calibration Gas Standard Concentration, 500 ppmv CH4] x 100%.
Table 5. Confirmation of Detected Leaks by TVA
Reference
Sample
Numbers
01C, 01D
02A, 02B
03A, 03B
05A, 05B
06A, 06B
07A, 07B
08A, 08B
09A, 09B
10A, 10B
Concentration
Pre-bagging
>100,000(a)
20,500
>100,000(a)
>100,000(a)
18,000
18,000
8,000
800
>100,000(a)
Measured by TVA
Post-bagging
>100,000(a)
20,500
>100,000(a)
>100,000(a)
23,000
17,000
8,000
870
>100,000(a)
(ppmv CH4)
Relative %
Difference^
0%
0%
0%
0%
24%
5.7%
0%
8.4%
0%
Comments
Data is considered suspect and
results reported with qualifier.
(a) The concentration of the leak at the component was high enough to cause the TVA to flameout. Concentration
estimated as greater than 100,000 ppmv CH4.
(b) Relative percent (%) difference calculated using the following calculation:
21
-------
2 X ?re — bagging concentration — Post bagging concentration
Relative % difference = ; ; ; ^ ; X 10096
?re — bagging concentration + Post bagging concentration
4.1.4 Bias and Accuracy of Bagging Procedure
A DQI is established in the TQAP for this verification test for the bias and accuracy of the
bagging procedure. This DQI is assessed by bagging an artificial leak at a known rate in the
middle of the analytical calibration curve. The procedure followed is that specified in U.S. EPA
Protocol for Equipment Leak Emission Estimates using certified CH/j-in-air gas standards and
calibrated flow meters. This DQI indicator was assessed at the beginning and end of the week of
field sampling. An acceptance criterion of 80 to 120% recovery is required for the bagging
equipment to pass the known leak rate test. Table 6 presents the results of the known leak rate
test. As shown in Table 6, this DQI was met before and after reference sampling.
Table 6. Known Leak Rate Test Results
Date [Time]
Leak Rate
Level
Emission Rate
(kilogram per hour [kg/hr] CH4)
Theoretical
Measured
% Recovery(a)
P re- Test
11/28/2008 [12:45]
11/28/2008 [12:20]
Low
High
4.31xlO'4
1.75X10'3
4.23x 10'4
1.60X10'3
98.1%
91.4%
Post-Test
12/5/2008 [14:35]
12/5/2008 [14:43]
Low
High
1.25X10'3
2.43 x 10'3
1.32 xlO'3
2.50 xlO'3
106%
103%
(a) Percent Recovery is calculated as (measured emission rate, kg/hr CH4) / (theoretical emission rate, kg/hr CH4) x
100%
4.1.5 Bias and Accuracy of Gas Chromatography Analytical Method
A DQI is established in the TQAP for this verification test for the bias and accuracy of the GC
analytical method used to quantify the concentration of leaks collected during reference
sampling. This DQI was assessed through initial calibration, and by performing positive and
negative control samples. These assessments are discussed in the following paragraphs.
Initial Calibration. Initial calibration of the GC was conducted by using various levels of
certified calibration gases starting with an un-spiked gas standard and then a minimum of four
additional concentrations of gas standards. The TQAP for this verification test required that the
initial calibration be performed at the start and end of every analytical sequence or if overall
instrument sensitivity changed by greater than 10%. To ensure accuracy of the initial calibration,
the instrument must be calibrated using certified gas standards. The minimum acceptance
criteria specified for this assessment is that all gas standards must be within 2% of their certified
value.
The analytical laboratory that performed the GC analytical method (Enthalpy Analytical, Inc.)
purchased gas standards with certification accuracies of ± 2%, as specified by the gas supplier.
In addition, the GC analytical laboratory produced diluted gas standards from these purchased
standards using a gas dilution system compliant with U.S. EPA Method 205^ which specifies
gas dilution systems must produce calibration gases whose measured values are within ± 2% of
the predicted levels from a certified gas standard.
22
-------
Positive Control Checks. The TQAP for this verification test required that positive control
checks be performed at a minimum frequency of 10% of all samples tested using one
concentration of calibration gas standard. The minimum acceptance criteria for positive control
checks is that the positive control check response is less than or equal to a 10% change in
response from the initial calibration after adjustment of the overall instrument sensitivity. Forty
sample measurements were conducted by the GC analytical laboratory using triplicate injections
and 19 positive control checks were performed exceeding the minimum frequency of 10% of
samples tested. The results of the positive control checks are provided in Table 7. As
demonstrated by Table 7, all positive control checks met this acceptance criterion.
Negative Control Checks. The TQAP for this verification test required that negative control
checks be performed at a minimum frequency of one out of every 10 samples tested. The
minimum acceptance criterion for this assessment is that all negative control responses must
remain lower than the lowest calibration standard for the chemical analyzed. Forty sample
measurements were conducted by the GC analytical laboratory using triplicate injections and
four negative control checks were performed meeting the minimum frequency of one negative
control check per 10 samples analyzed. All negative control checks performed were non-detect
for the compounds analyzed indicating an analytical result below the method detection limit for
the compound. The method detection limit for methane, ethylene, styrene, benzene,
1,3-butadiene, methylene chloride, and propylene dichloride was 1.00 ppmv for each compound.
4.2 Audits
Two types of audits were performed during the verification test, a technical systems audit (TSA)
of the verification test procedures, and a data quality audit. Because of the nature of bagging
reference method, a performance evaluation audit, as is usually performed to confirm the
accuracy of the reference method, was not applicable for this verification test. Audit procedures
for the TSA and the data quality audit are described further below.
4.2.1 Technical Systems Audit
The Battelle AMS Center Quality Manager performed a TSA during the both the laboratory and
field testing portions of this verification test to ensure that the verification test was performed in
accordance with the QMP for the AMS Center and the test/QA plan.
The TSA of the laboratory portion of the verification test was performed on October 22, 2008.
During this TSA, the Battelle AMS Center Quality Manager observed the test procedures used to
determine method detection limits and the response of the Industrial Scientific IBRID MX6 with
PID sensor and SP6 motorized sampling pump at the each method detection limit. These
procedures were observed during some of the testing conducted with acrylic acid, benzene,
methylene chloride, and styrene.
The TSA of the field testing portion of the verification test was performed on December 3,
2008. During this TSA, the Battelle AMS Center Quality Manager observed the procedures of
the bagging reference method, including the confirmation of the detected leaks by means of pre-
and post-bagging screening of the leaking component with the Thermo-Environmental TVA,
construction of the bagging enclosure, and duplicate reference sample collection, as well as the
audited the observations of the leak component with camera. In addition, the Battelle AMS
23
-------
Center Quality Manager observed both the performance of a calibration drift check and
recalibration as well as an end-of-day calibration response check of the Thermo-Environmental
TVA.
Table 7. Summary of Positive Control Check Responses
Positive Control
Check Sample ID
GC100pgl67#2
GC100pgl67#2
GC100pfl69F#4
GC100pfl69F#4
GC100pfl69F#4
GC100pfl69F#4
GC100pfl69F#4
GC100pfl69F#4
GC102pg44 #3
GC102pg44 #3
GC100pgl69 #2
GC100pgl69 #2
GC100pgl69 #3
GC100pgl69#4R
GC100pgl69#4R
GC102pg52 #4
GC102pg52 #4
GC102pg52 #4
GC102pg52 #4
Compounds
Measured by GC
Method
Benzene
Benzene
Ethylene
1,3 -butadiene
Ethylene
1,3 -butadiene
Ethylene
1,3 -butadiene
Ethylene
1,3 -butadiene
Ethylene
1,3 -butadiene
Ethylene
1,3 -butadiene
Methane
Methane
Methane
Methane
Methane
Methane
Methane
Pentane
Methylene chloride
Benzene
Propylene dichloride
Styrene
Pentane
Methylene chloride
Benzene
Propylene dichloride
Styrene
Pentane
Methylene chloride
Benzene
Propylene dichloride
Styrene
Pentane
Methylene chloride
Benzene
Propylene dichloride
Styrene
Expected
Response
(Picoampere
Second)
39.8
39.8
13.7
27.3
13.7
27.3
13.7
27.3
13.7
27.3
13.7
27.3
13.7
27.3
22.4
22.4
7.10
7.10
15.7
15.9
15.9
122
17.6
148
36.1
31.9
122
17.6
148
36.1
31.9
67.7
10.2
82.0
21.1
17.8
67.7
10.2
82.0
21.1
17.8
Actual Response
(Picoampere
Second)
39.3
39.0
13.8
26.9
13.7
26.7
13.5
26.3
13.4
25.7
13.7
26.9
13.8
27.2
22.8
22.7
6.95
6.73
15.3
15.5
15.8
127
17.7
150
35.4
34.0
125
17.3
147
34.4
32.7
67.5
9.86
79.5
20.5
18.4
70.3
10.2
82.3
21.2
18.6
Percent Error(a)
-1.1%
-1.9%
+0.39%
-1.6%
-0.61%
-2.4%
-1.6%
-3.9%
-2.4%
-5.8%
-0.44%
-1.5%
+0.39%
-0.43%
+1.6%
+1.3%
-2.1%
-5.3%
-3.3%
-2.5%
-0.39%
+4.2%
+0.60%
+1.1%
-2.1%
+6.7%
+2.7%
-1.9%
-0.75%
-4.6%
+2.4%
-0.35%
-3.4%
-3.1%
-2.9%
+3.8%
+3.7%
+0.16%
+0.35%
+0.49%
+4.5%
(a) Percent error is calculated as [(Actual Peak Response, peak area - Expected Response, peak area)/ Actual Peak
Response, peak area] x 100%.
The TSA of both the laboratory and field testing portions resulted in one finding and one
observation. The finding identified that only one field test (at a chemical plant) has been
24
-------
conducted as part of this verification test as opposed to the two field sites (one a chemical plant
and the other a petrochemical plant) identified in the TQAP for this verification test. The
observation noted documentation errors and improvements to the manner in which data were
recorded were discussed on-site with the verification test leader; immediate changes based on the
discussed improvements were implemented.
A ISA report was prepared, and a copy was distributed to the EPA AMS Center Quality
Manager.
4.2.2 Data Quality Audit
Records generated in the verification test received a one-over-one review before these records
were used to calculate, evaluate, or report verification results. Data were reviewed by a Battelle
technical staff member involved in the verification test. The person performing the review added
his/her initials and the date to a hard copy of the record being reviewed.
100% of the verification test data was reviewed for quality by the Verification Test Coordinator,
and at least 10% of the data acquired during the verification test were audited. The data were
traced from the initial acquisition, through reduction and statistical analysis, to final reporting to
ensure the integrity of the reported results. All calculations performed on the data undergoing
the audit were checked.
The data quality audit resulted in four findings (on three separate topics) that were addressed
related to the documentation of the number of confirming individuals at the method detection
limits in the laboratory phase raw data, exclusion from the verification report of concentration
measurements made by the PID sensor for dichloromethane (methylene chloride), methanol, and
propane during the laboratory phase of this verification test, and data transcription errors.
A data audit report was prepared, and a copy was distributed to the EPA AMS Center Quality
Manager.
25
-------
Chapter 5
Statistical Methods
The statistical methods used to evaluate the quantitative performance factors listed in Section 3.2
are presented in this chapter. Qualitative observations were also used to evaluate verification test
data.
5.1 Method Detection Limit
The method detection limit was assessed using the procedures described in Section 3.2.2 and the
TQAP for this verification test. The overall detection limit variation was calculated as the
standard deviation of the method detection limits determined under all the conditions tested for
each chemical of interest. The equation for standard deviation is as follows:
5 =
(i)
where Sx is the standard deviation of all method detection limits determined for chemical x, n is
the number of replicate samples, Ck is the leak rate measured for the Mi sample, and c is the
average leak rate of the replicate samples. If the sample sizes were small (n < 10), standard
deviations provide a biased estimate of variability. Therefore the range is provided when there
were fewer than 10 samples collected.
5.2 Percent Agreement
Percent agreement was used to assess the agreement between the Sherlock VOC imaging
spectrometer and the monitoring device acceptable under U.S. EPA Method 21 in the laboratory
for each compound tested. The inverse of the percent agreement is the percentage of the results that
the technology would detect a leak when U.S. EPA Method 21 would not. The equation for percent
agreement is as follows:
,4
Percent Agreement = — x 100%
where A the number of tests that both units agree and Tis the total number of tests. To determine if
both the monitoring device acceptable under U.S. EPA Method 21 and the Sherlock® VOC
imaging spectrometer agreed, the method detection limits at each test condition were first
reviewed. If the method detection limit of the Sherlock® VOC imaging spectrometer was below
26
-------
the highest reliable flow rate of the chemical delivery system (reported as <), then the Sherlock
VOC imaging spectrometer was noted as being able to detect the chemical gas leak under those
®
specified test conditions. Similarly, if the method detection limit of the Sherlock VOC imaging
spectrometer was above the highest reliable flow rate of the chemical delivery system (reported
as >), then the Sherlock® VOC imaging spectrometer w
chemical gas leak under those specified test conditions.
as >), then the Sherlock® VOC imaging spectrometer was noted as not being able to detect the
Next, the response of the monitoring device acceptable under U.S. EPA Method 21 was
reviewed for the same test conditions. If the monitoring device acceptable under U.S. EPA
Method 21 produced a response greater than zero, the monitoring device was considered capable
of detecting the chemical gas leak. Similarly, if the monitoring device acceptable under U.S.
EPA Method 21 produced a response equal to zero, the monitoring device was considered
incapable of detecting the chemical gas leak.
The responses of both the Sherlock VOC imaging spectrometer and the monitoring device
acceptable under U.S. EPA Method 21 under the same test conditions were compared. If both
the Sherlock® VOC imaging spectrometer and the monitoring device acceptable under U.S. EPA
Method 21 proved capable of detecting the chemical gas leak, then both units were considered to
have agreed under the specific test condition. Likewise, if either the Sherlock VOC imaging
spectrometer or the monitoring device acceptable under U.S. EPA Method 21 proved incapable
of detecting the chemical gas leak under the specified test conditions, then the units were
considered to have disagreed. Test conditions, under which a response from the either the
Sherlock® VOC imaging spectrometer or the monitoring device acceptable under U.S. EPA
Method 21 were not obtained, were excluded from the comparison.
27
-------
Chapter 6
Test Results
As mentioned previously, this verification test included both quantitative and qualitative
evaluations. The quantitative evaluation was conducted to assess the method detection limits of
the Sherlock® VOC imaging spectrometer, the detection of chemical gas species relative to a
portable monitoring device acceptable under U.S. EPA Method 21, as well as, by testing the
influence of confounding factors. The qualitative evaluation was performed to document the
operational aspects of imaging spectrometer when they were used during verification testing.
The following sections provide the results of the quantitative and qualitative evaluations.
6.1 Method Detection Limit
The method detection limit of each chemical compound was determined according to the
procedures discussed in Section 3.2.2. Table 8 through Table 11 present the method detection
limits of each chemical compound determined during laboratory testing. Table 8 through Table
11 identify each test condition evaluated (i.e., stand-off distance, background material, and wind
speed), the temperatures of the laboratory and of the chemical leak, the response of the portable
monitoring device acceptable under U.S. EPA Method 21, and, the method detection limits for
each test condition. Table 12 summarizes the range of method detection limits in units of grams
per hour (g/hr) found during laboratory testing as well as presents the overall detection limit
variation for each compound. The overall detection limit variation presented in Table 12 was
calculated using Equation 1 in Chapter 5.
6.2 Detection Agreement to a Portable Monitoring Device
The detection of a single chemical gas leak in either the laboratory or field environments was
determined by the operator as well as two confirming individuals as discussed in Section 3.2.1. The
leak rate was known from certified gas cylinders and calibrated flow meters in the laboratory, or was
determined through the bagging method during field testing. During both the laboratory and field
tests, a portable monitoring device acceptable under U.S. EPA Method 21 was used to sample the
leaks. The following section presents results on the ability of the Sherlock® VOC imaging
spectrometer to detect a chemical gas species relative to a portable monitoring device acceptable
under U.S. EPA Method 21.
28
-------
Table 8. Sherlock® VOC Method Detection Limits at 10 Feet with Cement Board
Background
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Wind
Speed
(mph)
0
0
2.5
5
0
0
2.5
0
2.5
0
2.5
5
0
2.5
5
0
2.5
5
0
2.5
5
0
Ambient
Temp.
70.4
72.2
74.9
74.9
71.0
73.0
72.3
70.9
72.3
71.2
71.3
71.5
71.5
70.0
70.0
71.9
71.3
70.9
70.5
71.8
71.1
71.5
Leak
Temp.
71.0
77.6
84.6
79.4
95.7
80.3
81.5
74.8
78.4
71.3
72.2
72.2
76.6
91.5
83.3
79.3
82.6
77.9
70.4
71.9
71.6
80.5
M21 Device
Cone.
(ppmv)
> 2,000
25
511
675
43
877
> 2,000
N.A.(b)
N.A.(b)
No data(c)
842
382
N.A.(b)
N.A.(b)
N.A.(b)
0.70
190
164
N.A.(b)
N.A.(b)
N.A.(b)
>2000
Method
Detection Limit
(g/hr)
8.1
1.7
9.2
23
7.4
3.2
> 70 (a)
>70(a)
>70(a)
3.3
250
> 275 (a)
2.1
29
69
0.83
55
> 55 (a)
0.88
9.8
13
15
(a) The leak could not be detected below the highest reliable flow rate supplied by the delivery system.
(b) N.A. - not applicable. The ionization potential of this compound is higher than is capable of
detection by the device used. Therefore, any raw data measured with this device is not reported in
this table.
(c) No data - the leak concentration was inadvertently not collected by laboratory personnel using the
M21 device.
29
-------
Table 9. Sherlock® VOC Method Detection Limits at 30 Feet with Cement Board
Background
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Wind
Speed
(mph)
0
0
2.5
5
0
0
2.5
5
0
0
2.5
0
2.5
5
0
2.5
5
0
2.5
5
0
Ambient
Temp.
71.5
71.7
74.8
74.7
71.7
70.0
74.8
74.6
69.5
70.5
71.3
71.7
72.2
70.0
72.0
70.3
70.3
70.8
71.8
72.3
71.0
Leak
Temp.
72.2
76.5
87.4
81.5
92.0
77.8
83.0
79.2
74.3
70.7
72.2
83.3
92.2
82.1
76.6
85.8
80.1
70.7
71.7
72.0
80.0
M21 Device
Cone.
(ppmv)
> 2,000
44
39
516
0.80
> 2,000
> 2,000
> 2,000
N.A.(b)
No data(c)
863
N.A.(b)
N.A.(b)
N.A.(b)
168
36
111
N.A.
N.A.
N.A.
>2000
Method
Detection Limit
(g/hr)
27
1.9
32
81
0.92
14
> 70 (a)
> 70 (a)
>70(a)
17
278
3.5
66
>69(a)
2.5
39
> 55 (a)
1.3
22
82
25
(a) The leak could not be detected below the highest reliable flow rate supplied by the delivery system.
(b) N.A. - not applicable. The ionization potential of this compound is higher than is capable of
detection by the device used. Therefore, any raw data measured with this device is not reported in
this table.
(c) No data - the leak concentration was inadvertently not collected by laboratory personnel using the
M21 device.
30
-------
Table 10. Sherlock® VOC Method Detection Limits at 10 Feet with Gas Cylinder
Background
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Wind
Speed
(mph)
0
0
2.5
5
0
0
2.5
0
2.5
0
2.5
5
0
2.5
5
0
2.5
5
0
2.5
5
0
Ambient
Temp.
70.5
71.8
74.0
74.7
71.7
72.6
74.3
70.9
74.1
71.6
71.1
71.5
70.9
70.0
70.2
70.6
72.4
72.0
71.2
71.8
70.9
72.4
Leak
Temp.
71.5
80.5
89.4
78.9
97.0
92.3
83.7
80.9
81.3
71.6
72.0
71.9
83.3
91.2
83.4
79.5
85.0
79.6
71.0
71.8
71.5
85.6
M21 Device
Cone.
(ppmv)
> 2,000
No data(a)
0.40
377
26
> 2,000
1022
N.A.(C)
N.A.(C)
No data(a)
1,790
547
N.A.(C)
N.A.(C)
N.A.(C)
12
185
190
N.A.(C)
N.A.(C)
N.A.(C)
>2000
Method
Detection Limit
(g/hr)
13
1.7
9.2
23
2.3
7.0
> 70 (b)
>70(b)
> 70 (b)
22
> 209 (b)
> 278 (b)
2.1
29
69
1.9
19
47
1.1
9.8
14
18
(a) No data - the leak concentration was inadvertently not collected by laboratory personnel using the
M21 device.
(b) The leak could not be detected below the highest reliable flow rate supplied by the delivery system.
(c) N.A. - not applicable. The ionization potential of this compound is higher than is capable of
detection by the device used. Therefore, any raw data measured with this device is not reported in
this table.
-------
Table 11. Sherlock® VOC Method Detection Limits at 30 Feet with Gas Cylinder
Background
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Wind
Speed
(mph)
0
0
2.5
5
0
0
2.5
5
0
0
2.5
0
2.5
5
0
2.5
5
0
2.5
5
0
Ambient
Temp.
71.2
71.5
74.6
74.4
71.4
71.8
75.0
74.6
69.8
71.5
71.0
71.7
72.2
70.1
71.9
75.1
72.0
70.6
71.6
71.0
73.0
Leak
Temp.
72.0
78.4
88.7
79.2
88.4
77.1
83.8
79.4
77.2
72.0
72.2
80.1
93.7
84.0
76.5
88.8
81.7
70.3
71.5
71.8
85.4
M21 Device
Cone.
(ppmv)
> 2,000
39
111
560
3.7
> 2,000
> 2,000
> 2,000
N.A.(b)
No data(c)
1402
N.A.(b)
N.A.(b)
N.A.(b)
228
443
113
N.A.(b)
N.A.(b)
N.A.(b)
>2000
Method
Detection Limit
(g/hr)
27
1.9
32
81
3.2
14
> 70 (a)
> 70 (a)
>70(a)
32
> 278 (a)
3.5
66
>69(a)
2.8
> 55 (a)
> 55 (a)
1.3
22
235
25
(a) The leak could not be detected below the highest reliable flow rate supplied by the delivery system.
(b) N.A. - not applicable. The ionization potential of this compound is higher than is capable of
detection by the device used. Therefore, any raw data measured with this device is not reported in
this table.
(c) No data - the leak concentration was inadvertently not collected by laboratory personnel using the
M21 device
32
-------
Table 12. Sherlock® VOC Range of Method Detection Limits and Overall Method
Detection Limit Variation (g/hr)^
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Minimum
8.0
1.7
0.92
3.2
> 70 (c)
3.3
2.1
0.83
0.88
15
Maximum
27
81
7.4
>70(c),(d)
> 70 (c)
> 278 (c)
> 69 (d)
> 55 0), (d)
235 (c)
25
Overall Variation^
29
31
127
31
25
67
(a) Minimum and maximum values shown were measured at a 0-mph wind speed unless otherwise noted.
(b) When sample sizes are small (N < 10), standard deviations provide a biased estimate of the variability, therefore
only the range is provided when there were fewer than 10 method detection limits were determined.
(c) Measured at a 2.5-mph wind speed condition.
(d) Measured at a 5-mph wind condition.
6.2.1 Laboratory Testing
Table 13 presents the percent agreement between the ability of the Sherlock® VOC imaging
spectrometer and of a portable monitoring device acceptable under U.S. EPA Method 21 to
detect a chemical gas leak under the conditions tested. Percent agreement was calculated
according to Equation 2 in Chapter 5. The calculation of percent agreement excludes those
laboratory test measurements for which a response was not collected using a portable monitoring
device acceptable under U.S. EPA Method 21. In addition, percent agreement was not evaluated
for methylene chloride, methane, methanol, and propane because these compounds have an
ionization potential greater than that which could be supplied by the Industrial Scientific IBRID
MX6 with PID sensor (i.e., the device is incapable of detecting these compounds).
(S)
Table 13. Summary of Detection Agreement Between Sherlock VOC and a Method 21
Portable Monitoring Device
Gas Tested
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Ethylene
Pentane
Styrene
No. of Tests in
which Agreed
4
11
4
4
2
8
4
Total No. of
Tests
Completed
4
11
4
10
6
12
4
Percent
Agreement
100%
100%
100%
40%
33%
75%
100%
6.2.2 Field Testing
® -
During field testing, nine leaking components were viewed using the Sherlock VOC imaging
spectrometer using the procedures described in Section 3.2.1. Table 14 identifies whether each
chemical species gas leak was observed by the camera and the concentration of the leak as
33
-------
determined by a portable monitoring device acceptable under U.S. EPA Method 21. In addition,
this table identifies the type of component that was leaking, the average chemical-specific mass
leak rate from the component as determined by reference sampling, the distance the leak was
observed and the wind speed. Daily meteorological conditions were obtained from the Dow
Chemical Company's on-site weather station. Although the wind speed and daily maximum and
minimum temperatures were obtained from this meteorological tower, the actual wind speed and
ambient and background temperatures at each leak location at the time of observation are
unknown. Additional discussions describing each leak location are provided in the following
sections.
Table 14. Summary of Field Testing Results Using the Sherlock15 VOC
Leak
Location
1
2
3
5
6
7
8
9
10
Leaking
Component
Type
3 -inch Plug
'/4-inch Tube
'/2-inch
Connector
6-inch Block
Valve
V CUV ^f
8-inch Block
T r0l Ł-
Valve
Control Valve
Flange
2-inch Block
T r0l Ł-
Valve
1 -inch Valve
Plug
6-inch
Pressure
Relief Valve
Wind
Speed
(mph)
8
21
21
21
21
18
18
18
5
Stand-off
Distance
(ft)
12
10
30
10
10
10
10
10
10
10
M21 Device
Screening
Cone.
(ppmv)
> 100,000
20,500
> 100,000
> 100,000
20,500
17,500
8,000^
835
> 100,000
Leak
Detected by
Camera?
No
Yes
No
No
No
No
No
No
No
No
Bagging Results:
Average Leak Rate
(g/hr)
8.79 (methane)
4.31 (ethylene)
0.951 (ethylene)
2.32 x 10"3 (ethylene)
7.78 (methane)
5.24 x 10"2 (ethylene)
8.68 x 10"3 (styrene)
0.077 (benzene)
3.44(a) (benzene)
1.95 x 10"3 (ethylene)
0.282 (benzene)
1.92(b) (1,3-butadiene)
0.350 (methylene
chloride)
6.78
(propylene dichloride)
(a) As reported in Table 5, the pre- and post-bagging leak concentrations, as measured by the TVA, differed by
24.4%. This exceeds the minimum acceptance criterion of 20% for the DQI for the confirmation of detected
leaks. Thus, this data is considered suspect and reported with this data qualifier.
(b) As reported in Table 4, the calibration check response for the TVA, conducted after screening this component,
resulted in a 24% difference. This exceeded the minimum acceptance criterion of 10% for the DQI for the bias
and accuracy of sample screening measurements using a portable monitoring device. After recalibration of the
TVA, the leak concentration from this component was not reconfirmed with the TVA. Thus, this data is
considered suspect and reported with this data qualifier.
Leak Location 1. A leak was identified originating from a 3-inch plug in service with a process
stream containing ethane, ethylene, methane, and propane. Screening of the component with the
TVA caused an over range reading (estimated as > 100,000 ppmv). The leak was viewed but not
detected with the Sherlock® VOC imaging spectrometer at a stand-off distance of 12 ft with the
sun at the observers back. The leak was bagged and a duplicate reference sample was collected
into two evacuated SUMMA canisters. The SUMMA canisters were shipped to the off-site GC
laboratory and analyzed for ethylene and methane concentrations. Daily weather conditions, as
34
-------
reported by the on-site weather station, were clear conditions, a daily minimum and maximum
temperature of 41 and 61 degrees Fahrenheit (°F), respectively, with wind out of the east at 8
mph.
The average mass leak rate of ethylene measured at this leak location was 4.31 g/hr. This value
is lower than the lowest method detection limit measured for this compound during the
laboratory phase of this verification test.
Leak Location 2. A leak was identified originating from a H-inch tube in service with a process
stream containing ethane and ethylene. Screening of the component with the TVA resulted in a
concentration reading of 20,500 ppmv. The leak was viewed with the Sherlock VOC imaging
spectrometer at stand-off distances of 10 and 30 ft with the sun to the left of the observer. The
leak was detected by the imaging spectrometer at the 10 ft stand-off distance but was not
detected at the 30 ft stand-off distance. Wind direction at the location was noted as originating
from behind the observer and the site was shaded by piping and other equipment. The leak was
bagged and a duplicate reference sample was collected into two evacuated SUMMA canisters.
The SUMMA canisters were shipped to the off-site GC laboratory and analyzed for ethylene
concentration. Daily weather conditions, as reported by the on-site weather station, were clear
conditions, a daily minimum and maximum temperature of 42 and 70 °F with wind out of the
south southeast at 21 mph.
The average mass leak rate of ethylene measured at this leak location was 0.95 1 g/hr. This value
is lower than the lowest method detection limit measured for this compound during the
laboratory phase of this verification test.
Leak Location 3. A leak was identified originating from a %-inch connector in service with a
process stream containing acetylene, ethane, ethylene, methane, propane, and propylene.
Screening of the component with the TVA caused an over range reading (estimated as > 100,000
ppmv). The leak was viewed with the Sherlock® VOC at a stand-off distance of 10 ft with the
sun to the right of the observer. The leak was not detected by the imaging spectrometer at this
stand-off distance. Wind direction at the location was noted as originating from the right of the
observer and the site was shaded by piping and other equipment. The leak was bagged and a
duplicate reference sample was collected into two evacuated SUMMA canisters. The SUMMA
canisters were shipped to the off-site GC laboratory and analyzed for ethylene and methane
concentrations. Daily weather conditions, as reported by, were clear conditions, a daily
minimum and maximum temperature of 42 and 70 °F with wind out of the south southeast at 21
mph.
The average mass leak rate of ethylene measured at this leak location was 2.32 x 10"3 g/hr. This
value is lower than the lowest method detection limit measured with the Sherlock® VOC for this
compound during the laboratory phase of this verification test.
Leak Location 4. Leak location 4 contained a leaking component that was misidentified as
being in service with styrene. This sample location was confirmed to be in ethylbenzene service
and thus no analytical results are reported for this leak location. The Sherlock VOC imaging
spectrometer was not able to detect this leak.
Leak Location 5. A leak was identified originating from a 6-inch block valve in service with a
process stream containing benzene, ethane, ethylene, ethylbenzene, styrene, and toluene.
Screening of the component with the TVA caused an over range reading (estimated as
35
-------
> 100,000 ppmv). The leak was viewed with the Sherlock VOC at stand-off distance of 10 ft;
the leak could not be detected at this distance. The site was shaded and the viewing background
was concrete. The leak was bagged and a duplicate reference sample was collected into two
evacuated SUMMA canisters. The SUMMA canisters were shipped to the off-site GC
laboratory and analyzed for benzene, ethylene, and styrene concentrations. Daily weather
conditions, as reported by the on-site weather station, were clear conditions, a daily minimum
and maximum temperature of 48 and 79 °F with wind out of the north at 21 mph.
The average mass leak rates of ethylene, styrene, and benzene measured at this leak location
were 5.24 x 10"2, 8.68 x 10"3, and 0.077 g/hr, respectively. These values are all below the lowest
method detection limits measured with the Sherlock® VOC for these compounds during the
laboratory phase of this verification test.
Leak Location 6. A leak was identified originating from an 8-inch block valve in service with a
process stream containing benzene, toluene, hexane, and other aromatic hydrocarbons.
Screening of the component with the TVA resulted in a concentration reading of 20,500 ppmv.
The leak was viewed with the Sherlock VOC at a stand-off distance of 10 ft with the sun to the
right of the camera observer; the leak could not be detected at this distance. The site was an
exterior location and weather conditions were noted as slightly overcast with moderate wind
originating from the right of the observer. The leak was bagged and a duplicate reference sample
was collected into two evacuated SUMMA canisters. The SUMMA canisters were shipped to
the off-site GC laboratory and analyzed for benzene concentration. Daily weather conditions, as
reported by the on-site weather station, were clear conditions, a daily minimum and maximum
temperature of 48 and 79 °F with wind out of the north at 21 mph.
Leak Location 7. A leak was identified originating from a control valve flange in service with a
process stream containing benzene, butane, butylbenzene, all isomers of diethylbenzene, ethane,
ethylbenzene, ethylene, hexane, toluene, and other aromatic hydrocarbons. Screening of the
component with the TVA resulted in a concentration reading of 17,500 ppmv. The leak was
viewed with the Sherlock® VOC at a stand-off distance of 10 ft with the sun behind the camera
observer; the leak could not be detected at this distance. The site was located on the second deck
of the chemical plant and weather conditions were qualitatively noted as very windy. The
viewing background was other plant piping and equipment. The leak was bagged and a duplicate
reference sample was collected into two evacuated SUMMA canisters. The SUMMA canisters
were shipped to the off-site GC laboratory and analyzed for benzene and ethylene
concentrations. Daily weather conditions, as reported by the on-site weather station, were partly
cloudy conditions, a daily minimum and maximum temperature of 43 and 65 °F with wind out of
the north at 18 mph.
The average mass leak rates of ethylene and benzene measured at this leak location were 1.95 x
10"3 and 0.282 g/hr, respectively. These values are all below the lowest method detection limits
measured with the Sherlock® VOC for these compounds during the laboratory phase of this
verification test.
Leak Location 8. A leak was identified originating from a 2-inch block valve in service with a
process stream containing 1,3-butadiene. Screening of the component with the TVA resulted in
a concentration reading of 8,000 ppmv. The leak was viewed with the Sherlock® VOC at a
stand-off distance of 10 ft; the leak could not be detected at this distance. The site was an
exterior location on a marine vapor recovery line at a marine vapor recovery system and weather
conditions were qualitatively noted to be very windy and overcast. The leak was bagged and a
36
-------
duplicate reference sample was collected into two evacuated SUMMA canisters. The SUMMA
canisters were shipped to the off-site GC laboratory and analyzed for 1,3-butadiene
concentration. Daily weather conditions, as reported by the on-site weather station, were partly
cloudy conditions, a daily minimum and maximum temperature of 43 and 65 °F with wind out of
the north at 18 mph.
The average mass leak rate of 1,3-butadiene measured at this leak location was 1.92 g/hr. This
value is lower than the lowest method detection limit measured with the Sherlock® VOC for thi
compound during the laboratory phase of this verification test.
Leak Location 9. A leak was identified originating from a 1-inch valve plug in service with a
process stream containing dichloromethane (methyl ene chloride). Screening of the component
with the TVA resulted in a concentration reading of 835 ppmv. The leak was viewed with the
Sherlock® VOC at a stand-off distance of 10 ft; the leak could not be detected at this distance.
The site was an exterior location and weather conditions were qualitatively noted as overcast
with calm winds. The viewing background was concrete ground and a few metal pipe supports.
The leak was bagged and a duplicate reference sample was collected into two evacuated
SUMMA canisters. The SUMMA canisters were snipped to the off-site GC laboratory and
analyzed for dichloromethane concentration. Daily weather conditions, as reported by the on-
site weather station, were partly cloudy conditions, a daily minimum and maximum temperature
of 43 and 65 °F with wind out of the north at 18 mph.
The average mass leak rate of dichloromethane (methylene chloride) measured at this leak
location was 0.350 g/hr. This value is lower than the lowest method detection limit measured
with the Sherlock® VOC for this compound during the laboratory phase of this verification test.
Leak Location 10. A leak was identified originating from a 6-inch pressure relief valve in
service with a process stream containing 1,2,3-trichloropropane, 2,3-dichloropropanol, 2-methyl-
2-pentenal, l-chloro-2,3-epoxypropane, and 1,2-dichloropropane (propyl ene di chloride).
Screening of the component with the TVA caused an over range reading (estimated as >
100,000 ppmv). The leak was viewed with the Sherlock® VOC at a stand-off distance of 10 ft;
the leak could not be detected at this distance. The site was an exterior location (on top of a
storage tank platform) and weather conditions were qualitatively noted as overcast, breezy, and
cold. The leak was bagged and a duplicate reference sample was collected into two evacuated
SUMMA canisters. The SUMMA canisters were shipped to the off-site GC laboratory and
analyzed for 1,2-dichloropropane concentration. Daily weather conditions, as reported by the
on-site weather station, were partly cloudy conditions, a daily minimum and maximum
temperature of 41 and 50 °F with wind out of the north at 5 mph.
6.3 Confounding Factors
The method detection limits generated during laboratory testing presented in Table 8 through
Table 11 were inspected to identify general trends that the confounding factors of stand-off
distance, wind speed, and background materials impart on the method detection limits for the
gaseous chemical species leaks observed using the Sherlock® VOC imaging spectrometer. The
following general trends were noted when using the imaging spectrometer.
• Stand-off Distance - Method detection limits generally increased as the viewing distance
increased. Two exceptions to this general observation were found. The first occurred
37
-------
when viewing a pentane leak at a 2.5-mph wind speed and a cement board background.
The second occurred when viewing an acrylic acid leak with no wind and a cement board
background.
• Wind Speed - Method detection limits generally increased with increased wind speed;
• Background Materials - Method detection limits were generally lower when viewed
against the cement board background. A single exception to this observation was noted
when viewing an acrylic acid leak at a 0-mph wind speed and at a 10 ft viewing distance.
During field testing, confounding factors were recorded either quantitatively or qualitatively and
are reported in Table 14. A rigid analysis of the influence of confounding factors was not
undertaken using field testing data, however, it is generally noted that because the optical
imaging instrument detected only a few of the chemical leaks in the field, the confounding
factors of wind speed, stand-off distance, and background materials affected the detection
capability of the cameras.
6.4 Operational Factors
The Sherlock® VOC imaging spectrometer was found to be easily setup in a small, three ft by
three ft area and deployed within approximately 10 minutes for portable gas leak observations.
In terms of field portability, the imaging spectrometer was moderate in weight (weighing
approximately 19 pounds with battery), easily carried by one person and was provided with a
rugged shipping case for transportation.
The Sherlock® VOC imaging spectrometer may be powered with either an Anton Bauer Titan
Power Supply/Charger, a 70 watt unit that accepts 90 to 265 volt (alternating current) at 50 to 60
Hz, for stationary applications or with an Anton Bauer Compatible Digital Hytron Nickel Metal-
Hydride battery for mobile field observations. The battery for each instrument was used and
held its charge when performing visual screening of leaking components. The camera observer
sees the infrared image through a standard, mounted 3.5-inch on diagonal liquid crystal display
viewing screen when using the imaging spectrometer; these images may be recorded to either an
internal CompactFlash card (up to 2 gigabytes) or to a Sony DV-1000 Digital Video Recorder
system or any other digital recording system.
Ease of use was not investigated with a newly trained operator, as the personnel from Industrial
Scientific Corporation operated the Sherlock® VOC during both laboratory and field testing.
Verification test team members, however, did observe that the instrument was used by the
camera operator with relative ease. The Sherlock® VOC imaging spectrometer is not
intrinsically safe, and cannot be used in explosive atmospheres or environments.
During this verification test, all chemical leaks were required to be observed by the instrument
operator and two additional confirming individuals to be considered as "detected" by the optical
imaging device. During verification testing, there were instances where either one or two of the
three observers were able to observe the chemical leak. This indicates that the ability of the
operator using the instrument to positively identify the chemical leak may have an influence on
the operation of the system.
The cost of the Sherlock® VOC imaging spectrometer is $89,000 and includes the LCD video
display, a Pelican shipping case, a battery and battery charger, personal computer, HYP AT
software, and all necessary cables.
38
-------
Chapter 7
Performance Summary
Method Detection Limits. Method detection limits were determined during the laboratory
testing. Table 15 summarizes the minimum and maximum method detection limit obtained
during laboratory testing using the Sherlock® VOC imaging spectrometer. Specific details,
including the test conditions at which these method detection limits were obtained, are provided
in Table 8 through Table 11 in Chapter 6. The overall detection limit variation for each chemical
obtained using the imaging spectrometer are presented in Table 12 in Chapter 6.
Detection of Chemical Gas Species Relative to a Portable Monitoring Device. The ability of
the Sherlock® VOC imaging spectrometer to detect a gaseous leak of a chemical relative to a
portable monitoring device acceptable under U.S. EPA Method 21 was assessed during both
laboratory and field testing. During laboratory testing, after the method detection limit had been
reached for a particular chemical under the specified test conditions, the leak was sampled by the
portable monitoring device. Table 15 presents the percent agreement between the ability of the
Sherlock VOC imaging spectrometer and of a portable monitoring device acceptable under U. S.
EPA Method 21 to detect a chemical gas leak under the conditions tested in the laboratory.
During field testing a portable monitoring device acceptable under U.S. EPA Method 21 was
used to screen each leaking component as part of the bagging reference method used. Table 16
reports the responses of the portable screening device when screening leaking components,
identifies whether the Sherlock® VOC imaging spectrometer was able to detect the chemical leak
from the leaking component, and reports the chemical-specific mass rate of emissions from the
leaking component as obtained through reference sampling.
39
-------
®
Table 15. Summary of Sherlock VOC Method Detection Limits^ and Percent Agreement
with a Method 21 Monitoring Device During Laboratory Testing
Compound
1,3 -butadiene
Acetic acid
Acrylic acid
Benzene
Methylene chloride
Ethylene
Methanol
Pentane
Propane
Styrene
Method Detection
Minimum
8.0
1.7
0.92
3.2
> 70 (b)
3.3
2.1
0.83
0.88
15
Limit (g/hr)
Maximum
27
81
7.4
>70(b),(o)
> 70 (b)
> 278 (b)
> 69 (c)
> 55 (b), (c)
235 (b)
25
Agreement with Method 21
Monitoring Device
Total No. of Percent
Tests Performed Agreement
4 100%
11 100%
4 100%
4 40%
No data(d)
2 33%
No data(d)
8 75%
No data(d)
4 100%
(a) Minimum and maximum method detection limits were measured at a 0-mph wind speed unless otherwise noted.
(b) Measured at a 2.5-mph wind speed.
(c) Measured at a 5-mph wind speed.
(d) Percent agreement was not evaluated for methylene chloride, methanol, and propane because these compounds
have an ionization potential greater than the energy which could be supplied by the Industrial Scientific IBRID
MX6 with PID sensor.
Confounding Factors. Stand-off distance, wind speed, and background materials generally
impacted the performance of the Sherlock VOC imaging spectrometer (e.g., increasing the
viewing distance from the leak increased the method detection limits). Details of the effects of
confounding factors can be found in Section 6.3.
Operational Factors. The Sherlock VOC imaging spectrometer was found to be easily set up
and ready to deploy in 10 minutes. The camera is moderate in weight (19 pounds with battery)
and operated on batteries when performing visual screening of leaking components. Because the
cameras were operated by Industrial Scientific personnel and there were some disagreements on
detections with the two confirming individuals, the ability of the operator may influence the
operation of the camera. The Sherlock® VOC imaging spectrometer is not intrinsically safe, and
cannot be used in explosive atmospheres or environments.
The cost of the Sherlock® VOC imaging spectrometer is $89,000 and includes the LCD video
display, a Pelican shipping case, a battery and battery charger, personal computer, HYP AT
software, and all necessary cables.
40
-------
®
Table 16. Summary of Field Testing Results of the Sherlock VOC
Leaking
Leak Component
Location Type
1 3 -inch Plug
2 !/4-inch Tube
, '/2-inch
Connector
6-inch Block
Valve
8-inch Block
Tr 1
Valve
7 Control Valve
Flange
2-inch Block
Tr 1
Valve
1 -inch Valve
Plug
6-inch
10 Pressure
Relief Valve
Wind
Speed
(mph)
8
21
21
21
21
18
18
18
5
Stand-off
Distance
(ft)
12
10
30
10
10
10
10
10
10
10
M21 Device
Screening
Cone.
(ppmv)
>100,000
20,500
>100,000
>100,000
20,500
17,500
8,000^
835
>100,000
Leak
Detected by
Camera?
No
Yes
No
No
No
No
No
No
No
No
Bagging Results:
Average Leak Rate
(g/hr)
8.79 (methane)
4.31 (ethylene)
0.951 (ethylene)
2.32 x 10"3 (ethylene)
7.78 (methane)
5.24 x 10"2 (ethylene)
8.68 x 10"3 (styrene)
0.077 (benzene)
3.44(a) (benzene)
1.95 x 10'3 (ethylene)
0.282 (benzene)
1.92(b) (1,3-butadiene)
0.350 (methylene
chloride)
6.78
(propylene dichloride)
(a) As reported in Table 5, the pre- and post-bagging leak concentrations, as measured by the TVA, differed by
24.4 %. This exceeds the minimum acceptance criterion of 20% for the DQI for the confirmation of detected
leaks. Thus, this data is considered suspect and reported with this data qualifier.
(b) As reported in Table 4, the calibration check response for the TVA, conducted after screening this component,
resulted in a 24% difference. This exceeded the minimum acceptance criterion of 10% for the DQI for the bias
and accuracy of sample screening measurements using a portable monitoring device. After recalibration of the
TVA, the leak concentration from this component was not reconfirmed with the TVA. Thus, this data is
considered suspect and reported with this data qualifier.
41
-------
Chapter 8
References
1. Test/QA Plan for Verification of Leak Detection and Repair Technologies, Battelle,
Columbus, Ohio, September 18, 2008.
2. Quality Management Plan for the ETV Advanced Monitoring Systems Center, Version 7.0,
U.S. EPA Environmental Technology Verification Program, Battelle, Columbus, Ohio,
November, 2008
3. EPA Method 21- Detection of Volatile Organic Compound Leaks, EP A-600/2-18-110; U. S.
EPA, September 1981.
4. Panek, J., P. Drayton, and D. Fashimpaur. Controlled Laboratory Sensitivities and
Performance Evaluation of Optical Leak Imaging Infrared Cameras for Identifying Alkane,
Alkene, and Aromatic Compounds, Proceedings of the 99* Annual Conference and
Exposition of the Air and Waste Management Association, New Orleans, June 20-23, 2006,
Manuscript number 06-A-159-AWMA, Curran Associates, Inc., Red Hook, New York,
March 2007.
5. EPA Protocol for Equipment Leak Emissions Estimates, EPA-453/R-95-017; U.S. EPA:
Research Triangle Park, NC, November 1995.
6. EPA Method 18 - Measurement of Gaseous Organic Compound Emissions by Gas
Chromatography, 40 CFR, Part 60, Appendix A; April, 1994.
7. EPA Method 205 - Verification of Gas Dilution Systems for Field Instrument Calibrations,
40 CFR, Part 51, Appendix M, September, 1996.
42
------- |