September 2004
Revised
Environmental Technology
Verification Report
Opsis AB
LD500 Continuous Emission
Monitor for Ammonia
Prepared by
Battelle
Battelle
The Business of Innovation
Under a cooperative agreement with
vvEPA U.S. Environmental Protection Agency
ElV etV etV
-------
September 2004
Revised
Environmental Technology Verification
Report
ETV Advanced Monitoring Systems Center
Opsis AB
LD500 Continuous Emission Monitor
for Ammonia
by
Ken Cowen
Ian MacGregor
Kelley Hand
Joseph Carvitti
Mike Rectanus
Thomas Kelly
Karen Riggs
Battelle
Columbus, Ohio 43201
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development, has financially supported and collaborated in the extramural program described
here. This document has been peer reviewed by the Agency. Mention of trade names or
commercial products does not constitute endorsement or recommendation by the EPA for use.
11
-------
Foreword
The U.S. Environmental Protection Agency (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 verification 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. In 1997, through a competitive cooperative agreement, Battelle was awarded EPA
funding and support 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.
Battelle conducted this verification under a follow-on agreement to the original cooperative
agreement. Information concerning this specific environmental technology area can be found on
the Internet at http://www.epa.gov/etv/centers/centerl.html.
111
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the
verification test, analyze the data, and prepare this report. We would like to thank Ernie
Bouffard, Connecticut Department of Environmental Protection; John Higuchi, and Glenn
Kasai, South Coast Air Quality Management District; and Tom Logan, U.S. Environmental
Protection Agency, for their careful review of this verification report.
iv
-------
Contents
Page
Notice ii
Foreword iii
Acknowledgments iv
List of Abbreviations viii
1 Background 1
2 Technology Description 2
3 Test Design and Procedures 5
3.1 Introduction 5
3.2 Test Design 6
3.3 Test Conditions 8
3.4 Test Procedure 8
3.4.1 Reference Method 8
3.4.2 Dynamic Spiking 9
3.5 Quality Assurance Procedure 11
3.5.1 Performance Evaluation Audit 11
3.5.2 Technical Systems Audit 12
3.6 Data Comparisons 12
4 Quality Assurance/Quality Control 14
4.1 Equipment Calibrations 14
4.1.1 Host Facility Equipment 14
4.1.2 Calibration Check/Dynamic Spiking Equipment 14
4.2 Audits 14
4.2.1 Performance Evaluation Audit 14
4.2.2 Technical Systems Audit 15
4.2.3 Audit of Data Quality 15
4.3 QA/QC Reporting 15
5 Statistical Methods and Reported Parameters 16
5.1 Agreement with Standards 16
5.2 Linearity 16
5.3 Precision 17
5.4 Calibration and Zero Drift 17
5.5 Response Time 17
v
-------
6 Test Results 18
6.1 Agreement with Standards 18
6.2 Linearity 18
6.3 Precision 19
6.4 Calibration and Zero Drift 21
6.5 Response Time 21
6.6 Ease of Use 22
6.6.1 Installation 22
6.6.2 Zero and Span Checks 23
6.6.3 Dynamic Spiking 23
6.6.4 Data Handling 25
6.7 Data Completeness 25
6.8 Cost 25
7 Performance Summary 26
8 References 27
Figures
Figure 2-1. Opsis LD500 Ammonia CEM 2
Figure 2-2. Schematic Diagram of the Opsis LD500 3
Figure 3-1. Schematic of CEM Locations in Ammonia CEM Verification 7
Figure 6-1. Linear Regression of the LD500 NH3 Response vs. Expected Response 20
Figure 6-2. Example LD500 Rise and Fall Time Plots 24
Tables
Table 3-1. Summary of Flue Gas Parameters or Constituent Concentrations
at AEP's Mountaineer Plant 8
Table 3-2. Summary of PE Audits 11
Table 3-3. Summary of Data Obtained in LD500 Verification Test 12
Table 6-1. Agreement of LD500 with Ammonia Gas Standards 19
Table 6-2. Precision (% RSD) of LD500 During Dynamic Spiking Periods 20
Table 6-3. Calibration and Zero Drift for LD500 During Weeks 1 and 5 22
vi
-------
Table 6-4. LD500 Rise and Fall Times 23
Table 7-1. Summary of LD500 Verification Results 26
vii
-------
List of Abbreviations
AEP
American Electric Power
A
agreement
AMS
Advanced Monitoring Systems
ASTM
American Society of Testing and Materials
CEM
continuous emission monitor
cm
centimeter
CTM
conditional test method
EPA
U.S. Environmental Protection Agency
ETV
Environmental Technology Verification
ft3
cubic foot
h20
water
I/O
input/output
IC
ion chromatography
ISE
ion selective electrode
L
liter
lb
pound
m
meter
mg
milligram
min
minute
nh3
ammonia
nh4
ammonium
NIST
National Institute of Standards and Technology
NOx
nitrogen oxide
PE
performance evaluation
ppmv
parts per million volume
ppmwv
parts per million, wet volume basis
QA
quality assurance
QC
quality control
QMP
quality management plan
RSD
relative standard deviation
SCR
selective catalytic reduction
TSA
technical systems audit
viii
-------
Chapter 1
Background
The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental tech-
nologies 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 tech-
nologies 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 EPA's National Exposure Research Laboratory and its verification organization partner,
Battelle, operate the Advanced Monitoring Systems (AMS) Center under ETV. The AMS Center
recently evaluated the performance of a continuous emission monitor (CEM) for ammonia
(NH3), the Opsis AB LD500 (LD500).
1
-------
Chapter 2
Technology Description
The objective of the ETV AMS Center is to verify the performance characteristics of environ-
mental monitoring technologies for air, water, and soil. This verification report provides results
for the verification testing of the Opsis LD500. Following is a description of the LD500, based
on information provided by the vendor. The information provided below was not subjected to
verification in this test.
The LD500 is an optical open-path monitoring system designed to measure ammonia, water
vapor, hydrochloric acid, hydrogen fluoride, oxygen, and temperature. The LD500 allows
multiplexing of monitoring paths and can be configured with up to eight individual paths.
The LD500 (Figure 2-1), housed in a 19-inch rack cabinet, is the central unit of a laser diode gas
monitoring system. Up to four laser diode heads can be installed, each one a complete laser
control and data sampling system monitoring a specific gas. The LD500's laser module emits
near-infrared light, operates continuously, and is tunable. The laser is scanned rapidly (in
kilohertz frequency range) over the absorption line of the gas to be measured for 10 to
30 seconds. An internal reference beam maintains the wavelength stability of the laser diode. At
the end of the measurement interval, the averaged spectmm is evaluated. The results are com-
pared through a least-squares fitting procedure with the known absorbance cross section of the
gas. The Beer-Lambert absorption law is used to determine the gas concentration from the
absorption measured in the monitoring path, using the known monitoring path length.
A schematic of the LD500 is shown in
Figure 2-2. The LD500 includes an emitter and a
receiver to be mounted on ports on the flue gas
duct. The laser signal is sent through a fiber
optic cable to the receiver where it is divided
into two fiber optic cables, one providing the
signal to the emitter and the second providing
the light signal for calibration. The emitter
projects the infrared energy across the stack or
duct. The receiver focuses the projected infrared
Figure 2-1. Opsis LD500 Ammonia CEM ener8y t0 a solid-state detector. The raw signal is
converted to a digital communication signal and
transmitted through a communications optic fiber back to the LD500. The LD500 processes the
final signal and presents a concentration. The receiver is equipped with a calibration/audit cell.
2
-------
©o
Calibration unil control signal
Figure 2-2. Schematic Diagram of the Opsis LD500
LD500 System Drawing Legend
Red
1
110 V AC power
Blue
2
DC power input cable
1,2,4
Purge air
3
110 V AC power
5
Zero air
4
DC power cable
6
NH, span gas
5
C alibration control cable
6
Spare control signal
Green
7
Communications cable for analog
1
Laser source energy fiber optic
mput/output
2
Detector signal fiber optic
8
Communications cable to data system
3
Laser source energy fiber optic from
9
Analog input/output signal
splitter
10
Temperature sensor signal
11
Calibration control signal
The calibration cell is 5.11 inches (130.0 mm) long andheated to a constant temperature of
150°F. A solenoid valve unit is connected to the LD500, providing daily automatic zero and
span calibration. In calibration mode, the gas is flushed at a low flow rate through the cell and
vented through a Vi-inch tube at a secure point.
For absolute zero and span calibrations, a flat mirror is folded in to deflect the calibration laser
beam through the calibration cell. At the beginning of the calibration cycle, the mirror is auto-
matically folded in; it is folded out upon completion of the calibration cycle. The sam e laser
source is used both for measurements and for calibration checks. An add-on spiking run is
performed with the mirror folded out, i.e., the measurement light beam is measured on the
detector. In this mode, a concentration entered in the calibration cell is added to the measured
stack concentration.
3
-------
A signal input and output unit is connected to the LD500. Signals include stack temperature
entering the system and analog output signals being delivered to an outside data system. The
LD500 stores all raw data, measurements, and logged data from the signal system on its internal
hard drive. For data presentation, a personal computer provides real-time graphics of monitored
results. In this verification test, the LD500 was set up to provide discrete readings of NH3
concentration every 10 seconds, without averaging or smoothing the data.
4
-------
Chapter 3
Test Design and Procedures
3.1 Introduction
The objective of this verification test of the LD500 was to evaluate its ability to determine
gaseous ammonia in flue gas under normal operating conditions in a full-scale coal-fired power
plant equipped with selective catalytic reduction (SCR) nitrogen oxide (NOx) control
technology.
This verification test was conducted according to procedures specified in the Test/QA Plan for
Verification of Continuous Emission Monitors for Ammonia at a Coal-Fired Facility(1) at
American Electric Power's (AEP's) Mountaineer Plant in New Haven, West Virginia, from
July 15 to August 15, 2003.
The performance parameters addressed by the test/QA plan included:
¦ Agreement with standards
¦ Relative accuracy
¦ Linearity
¦ Precision
¦ Calibration and zero drift
¦ Response time
¦ Ease of use
¦ Data completeness.
Agreement with standards was assessed for the LD500 based on the differences between LD500
readings and known concentrations of ammonia prepared from ammonia compressed gas
standards. Relative accuracy refers to the degree of agreement of LD500 readings with flue gas
ammonia measurements made by a reference method. Precision was assessed in terms of the
repeatability of the LD500 ammonia measurements with stable ammonia concentrations.
Linearity, calibration drift, zero drift, and response time were assessed using commercial
compressed gas standards of ammonia and high purity nitrogen zero gas. The effort spent in
installing and maintaining the LD500 was documented and used to assess ease of use. The
amount of time the LD500 was operational and the maintenance activities performed were
recorded to assess data completeness.
5
-------
3.2 Test Design
The LD500 was installed at AEP's Mountaineer Plant approximately two weeks prior to testing,
and a shakedown run was conducted before verification testing began. The LD500 was installed
between the exit of the SCR and the inlet of the air heater. Upstream of this location, the gas
flow exiting upward from the SCR catalyst beds underwent a 180° turn, to flow downward
through the duct where the CEM was installed. A port for reference method sampling was
located in the same duct with the LD500. The sampling ports were assigned so that the LD500
was unaffected by the operation of any other CEM or by the reference method sampling. The
LD500 was equipped with a heated calibration cell that was used during dynamic spiking and
for calibration.
Testing began on July 15, 2003, and continued until August 15, 2003. The boiler and SCR
operated continuously during the test period. During verification testing, the LD500
continuously monitored ammonia over the five-week test period. The LD500 provided discrete
integrated ammonia readings at 10-second intervals, with no smoothing or averaging of
successive 10-second readings. The discrete 10-second readings were used directly or averaged
over longer time periods to address each of the target performance parameters. The LD500
vendor provided a correction to the raw LD500 data that was applied to all LD500 data as
follows:
Corrected LD500 concentration = (rawLD500 concentration - 0.2) x 0.13/0.12
The first correction was needed to address an incorrect baseline (zero) concentration that was
part of the pre-test calibration. The factor of 0.13/0.12 was needed to adjust to the actual
calibration cell length of 13 centimeters (cm).
Reference method sampling (see Section 3.4.1) was conducted on each weekday during the first
and fifth weeks. On each day of reference method sampling, duplicate reference method samples
were collected simultaneously using parallel sampling trains over each of three different
sampling periods. The 10-second LD500 readings during these periods were used to calculate
one-hour averages with the intent to compare them with the ammonia concentrations measured
by the reference method.
The rectangular duct at the test location was 20 feet 8 inches by 32 feet in cross section.
Figure 3-1 shows a schematic of the test configuration in the duct. The LD500 monitoring path
traversed the duct approximately five feet from the duct wall, parallel to the 20-foot side. The
reference method port was located in a corner of the duct, approximately eight feet from the
20-foot side. External access to the reference method port was severely restricted. Consequently,
it was not possible to use a probe long enough to penetrate well into the duct. In fact, reference
method samples could be collected only at a depth of less than one foot inside the inner wall of
the duct. Because of concern about the representativeness of the reference measurements, the
ammonia concentrations across the duct were mapped after the conclusion of the test period, to
assess ammonia uniformity in the duct. The results of that effort, and the limitations of the
reference measurements, are reported in Section 3.4.1.
6
-------
Reference sample port
a
nh3 cem
Light
Paths
8'
36" approx.
Facility NH3 CEM
r 0" OPL
NH3 CEM
NH3CEM
Figure 3-1. Schematic of CEM Locations in Ammonia CEM
Verification
During the third week of testing, the LD500 was challenged with a series of dynamic spikes of a
compressed ammonia gas standard and nitrogen zero gas using the calibration cell in line with
the cross-duct light path. The LD500 responses to the ammonia spikes were determined by
subtracting the average ammonia concentration observed without spiking from the average
ammonia concentrations observed during spiking. The results of these runs were used to assess
the agreement with standards, linearity, and precision of the LD500. Ten-second readings from
the dynamic spiking also were used to assess the response time of the LD500.
During each day of reference method sampling, zero and span checks were conducted by
challenging the LD500 with a commercial compressed ammonia gas standard and nitrogen zero
gas using the LD500 calibration cell. These zero/span checks were used to assess the zero and
calibration drift of the LD500 during the test period. During the second week of the test, the
LD500 operated continuously without any performance testing. The LD500 was not operational
during the fourth week of testing due to a hard drive failure.
Throughout the verification test, the LD500 was operated by the vendor's own staff or by
Battelle staff trained by the vendor. The intent of the testing was to operate the LD500
continuously in a manner simulating installed operation at a combustion facility. As a result,
once the verification test began, no adjustment or recalibration was performed other than that
which would be conducted automatically by the LD500 in normal unattended operation.
Maintenance procedures were carried out as needed, but testing was not interrupted in such
cases. Those maintenance procedures consisted of cleaning the optical windows in the duct on a
7
-------
few occasions during the test. This maintenance was necessitated by the lack of an adequate
supply of clean facility instrument air, and would not be needed in a normal permanent
installation.
3.3 Test Conditions
Table 3-1 shows the levels of ammonia and other constituents in the flue gas stream at AEP's
Mountaineer Plant. Some of the data in Table 3-1 were obtained during the reference method
sampling runs (see Section 3.4.1). Note that the percent moisture values in Table 3-1 vary
widely. This variability does not appear realistic, and may result from measurement error in the
reference sampling.
Table 3-1. Summary of Flue Gas Parameters or Constituent Concentrations at AEP's
Mountaineer Plant
Parameter/Constituent
Typical Concentration or Range
nh3
NOx
Sulfur dioxide
Oxygen
Dust loading
Moisture
Carbon dioxide
Temperature
0.8 to 2.5 parts per million on a wet volume basis (ppmwv)(a)
37 parts per million volume (ppmv)(b)
540 ppmv(b)
3.1 to 4.28%(c)
4.3 grains/dry standard cubic foot(b)
4.4 to 10.8%(c)
14.6 to 15.6%(c)
648 to 679°F(c)
Typical 15-minute values in ppmwv calculated from the LD500 10-second readings.
(b:i Typical values supplied by AEP.
(c:i As measured during reference method sampling.
3.4 Test Procedure
3.4.1 Reference Method
The test/QA plan(1) called for comparing the LD500 results with those from a time-integrated
measurement of ammonia in flue gas obtained using a modified EPA Conditional Test Method
(CTM027).(2) That conditional test method is similar to a draft American Society for Testing and
Materials (ASTM) method(3) for measuring ammonia. However, the draft ASTM method calls
for analysis by ion selective electrode (ISE) whereas EPA CTM027 calls for analysis by ion
chromatography (IC). The draft ASTM method also calls for a smaller volume of a more dilute
acid solution in the sampling impingers than does EPA CTM027. Since the dilute acid is more
appropriate for measuring low levels of ammonia, EPA CTM027 was modified to use the ASTM
acid volumes and concentrations for this verification test.
8
-------
During verification testing, reference sampling was conducted simultaneously with two
collocated trains, with each sampling run lasting 60 minutes. Thus, each of the three reference
sampling periods during a test day provided two reference ammonia samples for comparison
with the LD500 data. Field blank samples also were recovered from one blank sampling train on
each of three days during each week that reference method samples were collected. Additionally,
on each of three days during each week of reference sampling, one sample train was spiked with
ammonia solution to serve as a field spike sample. The spike was added as an aqueous standard
directly to the front impinger in the train.
Four reference method samples (two from each week of reference method sampling) were also
spiked with additional ammonium after analysis and then reanalyzed to establish the spike
recoveries. A performance evaluation audit of the reference method using National Institute of
Standards and Technology (NIST)-traceable ammonia standards was also conducted.
The reference method blank, spike, and audit sample results met all applicable criteria stated in
the test/QA plan,(1) indicating that the reference method sampling was properly carried out.
Blank sample concentrations were less than 10% of any duct sample ammonia concentration,
and laboratory spike recovery was always within 10% of the expected value (and usually within
5%). Field spike recoveries were well within the 20% acceptance criterion, and the audit sample
results agreed within about 5%. However, due to the inability to extend the reference method
probe across the duct (see Section 3.2), concern arose that the reference samples might not
adequately represent the duct ammonia concentrations, for comparison with data from the CEMs
undergoing verification.
To address this concern, after the verification test was concluded, an ammonia mapping study
was conducted to assess the representativeness of the reference method sampling location
relative to the light path of the LD500. In this mapping study, reference method samples were
collected simultaneously at the reference method port and at four locations along the LD500
monitoring path (32, 50, 68, and 86 inches inside the inner duct wall) on August 20 and again
on August 21. The results of the ammonia mapping study showed that ammonia concentrations
at the reference sampling point were typically two to five times lower than those at points along
the LD500 light path. The difference between the reference point results and those from points
along the light path was generally greatest for the points on the light path that were farthest into
the duct. Based on these observations, the reference data were judged to be not representative of
the flue gas sampled by the LD500. Consequently, no quantitative assessment of the relative
accuracy of the LD500 and reference method results is made in this report.
3.4.2 Dynamic Spiking
During the third week of testing, the LD500 was challenged with a series of dynamic spiking
runs using the calibration cell in line with the cross-duct light path. During these runs, the
effective ammonia concentrations in the light path were increased by 2.14, 5.22, and
8.15 ppmwv above the flue gas concentration. At each of these spike concentrations, a series of
runs was conducted that produced 12 spiked and 12 unspiked sample measurements. The path
length of the flue gas duct was 6.05 meters (m), whereas the path length of the calibration cell
was 13 cm (i.e., 0.13 m). The internal volume of this cell was 0.5 liters (L). To perform a
9
-------
dynamic spike, this cell was purged with either a standard ammonia gas mixture or nitrogen zero
gas. The purge flow rate to the cell was 1.3 L/minute (min), which produced approximately 2.6
cell volume changes per minute. A five-minute purge was adequate to obtain a stable reading.
To obtain a dynamic spike observation, a standard ammonia gas mixture was introduced to the
calibration cell until a stable reading was observed. A two-minute period of readings was then
obtained, and the cell was allowed to purge for an additional two minutes before another two-
minute period of readings was obtained, thus providing two spiked measurements. The cell
purge gas was then changed to zero nitrogen, and the cycle was repeated to obtain two periods of
unspiked ammonia readings. In summary, the following procedure was used to obtain dynamic
spiking data:
1. Allow the calibration cell to purge for approximately five minutes with the ammonia
standard (this yields 13 cell volume changes).
2. Select the next two minutes of 10-second LD500 readings and calculate a two-minute
average value. This value is the first spiked sample measurement.
3. Allow the calibration cell to purge with the ammonia standard for an additional two minutes
(this yields 5.2 cell volume changes).
4. Select the next two minutes of 10-second LD500 readings and calculate a two-minute
average value. This value is the second spiked sample measurement.
5. Repeat steps 1 through 4 using zero nitrogen to purge the cell to obtain two unspiked sample
measurements.
This procedure for collecting the spiked and unspiked measurements was conducted a total of
six times at each of the three spike concentrations, to obtain 12 spiked and 12 unspiked
measurements at each concentration (36 spiked and 36 unspiked observations). A single average
unspiked reading was determined from all of the unspiked LD500 readings, and that unspiked
average was subtracted from each spiked measurement before comparisons were made to the gas
standard spike concentrations.
Linearity was evaluated using all 36 two-minute average spiked observations. The expected
LD500 response was calculated based on the concentration of the ammonia standard gas and a
0.13-m/6.05-m factor to correct for the difference between the 13-cm light path in the
calibration cell and the 6.05-m monitoring path in the flue gas duct. A temperature
compensation factor of 1.84 and a spectral effects correction factor of 0.665, both supplied by
the vendor, were applied to correct the expected LD500 response for the difference between the
calibration cell temperature (150°F) and the flue gas temperature (350°C). The actual LD500
spike response was calculated by subtracting the average reading when zero gas passed through
the calibration cell from the average reading when spike gas passed through the cell.
10
-------
3.5 Quality Assurance Procedures
QA/quality control (QC) procedures were performed in accordance with the quality management
plan (QMP) for the AMS Center(4) and the test/QA plan for this verification test.(1) These proce-
dures are briefly described in this section. Results of the QA/QC procedures are presented in
Section 4.
3.5.1 Performance Evaluation Audit
A performance evaluation (PE) audit was conducted to assess the quality of the measurements
made in this verification test. This audit addressed only measurements that factor into the data
used for verification, i.e., the LD500 and the staff operating the LD500 were not the subject of
the PE audit. This audit was performed once during the verification test by analyzing a standard
or comparing a reading with one that was independent of standards used during the testing.
Table 3-2 summarizes the approach and equipment used for the PE audits and shows the
expected agreement of audit results. These audits were the responsibility of Battelle staff and
were carried out with the cooperation of facility staff. Results of the PE audit are summarized in
Sections 3.4.1 and 4.2.1.
Table 3-2. Summary of PE Audits
Parameter
Audit Equipment/Approach
Expected Tolerance
Flue Gas
Independent pressure measurement
±0.5 inch of H20
Differential
(Magnehelic gauge, LN342539 )
Pressure
Mass (H20)
Calibrated weights
±1% or 0.5 gram, whichever is
larger
Ammonia (overall
Spike reference method trains
±20% bias in spike recovery
measurement)
Ammonia (ISE
Independent audit sample- NIST
±10% of standard
analysis)
solution
concentration
Ammonia (IC
Independent audit sample- NIST
±10%) of standard
analysis)
solution
concentration
Planned PE audits of flue gas temperature and barometric pressure were not performed. These
deviations from the test/QA plan were documented in the program files, but have minimal
impact on the results of this verification.
11
-------
3.5.2 Technical Systems Audit
Battelle's ETV Quality Manager performed a technical systems audit (TSA) on July 16, 2003.
The purpose of this TSA was to ensure that the verification test was being performed in accord-
ance with the test/QA plan(1) and that all QA/QC procedures were implemented. As part of the
audit, Battelle's ETV Quality Manager reviewed the reference sampling and analysis methods
used, compared actual test procedures with those specified in the test/QA plan, and reviewed
data acquisition and handling procedures. An independent EPA audit was conducted by the EPA
Quality Manager at the same time as the Battelle audit.
3.6 Data Comparisons
Table 3-3 summarizes the data used for the verification of the various performance parameters
Chapter 5 presents the statistical procedures used to make these comparisons. Because of the
limitations of the reference data (Section 3.4.1), relative accuracy is not listed in Table 3-3 or
discussed in the subsequent sections of this report.
Table 3-3. Summary of Data Obtained in LD500 Verification Test
Performance
Parameter
Objective
Comparison Based On
Total Number
of Data Points
for
Verification
Agreement
with
Standards
Determine degree of
quantitative agreement with
compressed gas standard
Dynamic spiking with
NH3 gas standards
36
Linearity
Determine linearity of
response over a range of
ammonia concentrations
Dynamic spiking with
NH3 gas standards
36
Precision
Determine repeatability of
successive measurements at
stable ammonia levels
Repetitive measurements
during each dynamic
spiking run
72
Cal/Zero Drift
Determine stability of zero gas
and span gas response over
successive days
Zero gas and NH3 gas
standard analyses
14
Response
Time
Determine rise and fall times
Recording successive
readings in dynamic
spiking runs
11
The results of the dynamic spiking were used to assess the agreement of the LD500 results with
respect to calculated ammonia concentrations determined from the spike gas concentration. For
each spiking run, the difference between the ammonia concentration measured by the LD500
12
-------
and the calculated ammonia concentration from spiking was determined. A total of 36 spike
results were obtained. The differences were then used to assess the agreement of the LD500
results with the ammonia standard concentrations as described in Section 5.1.
Linearity of the LD500 response was assessed by linear regression of the two-minute average
data from the dynamic spiking runs, as described in Section 5.2. The measured ammonia
concentrations and the calculated ammonia concentrations were used to assess linearity over the
range from 2.14 to 8.15 ppmwv above background. A total of 36 data points (12 two-minute
averages each at 2.14, 5.22, and 8.15 ppmwv above background) were used for this assessment.
Precision of the LD500 was assessed based on the average percent relative standard deviation
(% RSD) of the 10-second readings over the duration of each dynamic spiking period, as
described in Section 5.3. An average % RSD was determined at each of the three spiking
concentrations.
Calibration and zero drift were verified by repeatedly challenging the LD500 with an ammonia
compressed gas standard and a nitrogen zero gas, respectively, during the first and fifth weeks of
the test, as described in Section 5.4. In this procedure the LD500 light path passed through the
calibration cell only (i.e., not across the duct). Consequently, the LD500 readings had to be
adjusted (scaled up) to correct for the difference between the 0.13-m cell length and the 6.05-m
path length programmed into the LD500 for cross-stack measurements. Seven data points were
used to assess zero drift, and seven were used to assess calibration drift.
LD500 response time was assessed in the third week of the test based on the successive
10-second readings during the dynamic spiking runs, as described in Section 5.5. The data from
the dynamic spiking run at the highest concentration (8.15 ppmwv above background) were
used to provide the clearest indication of response time. Five measures of rise time and six of
fall time were used in this evaluation.
No additional test activities were required to determine the data completeness achieved by the
LD500. Data completeness was assessed by comparing the data recovered from the LD500 with
the maximum amount of data recoverable upon completion of all portions of these test
procedures. The test was conducted over a period spanning approximately 746 hours.
Setup and maintenance needs were documented qualitatively, both through observation and
through communication with the vendors and trained facility staff during the test. Factors
included frequency of scheduled maintenance activities, downtime of the LD500, and number of
staff needed to operate or maintain it during the verification test. The approximate purchase cost
of the LD500 was also determined based on information provided by the vendor.
13
-------
Chapter 4
Quality Assurance/Quality Control
This section summarizes the results of QA/QC efforts in this verification. Because the CTM027
reference data were not used, for the reasons described in Section 3.4.1, the QA/QC results for
the reference method are not included here.
4.1 Equipment Calibrations
4.1.1 Host Facility Equipment
Monitoring devices in place at AEP's Mountaineer Plant, including an ammonia CEM of a type
not verified in this test, were calibrated according to normal facility procedures. All calibration
results were documented according to facility procedures and are available as supporting
documentation for this test.
4.1.2 Calibration Check/Dynamic Spiking Equipment
The accuracy of the dry gas meter used for measuring the spike gas flow rate during the
calibration checks and the dynamic spiking activities was confirmed by Battelle by comparison
against an electronic bubble flow meter (M30 Mini-Buck Calibrator, A. P. Buck, Inc.). This
calibrator has a flow rate range of 01 to 30 L/min. The range of flows confirmed with this
calibrator was approximately 5 to 10 L/min. The M30 Mini-Buck was itself calibrated by the
manufacturer against a NIST-traceable flow standard.
4.2 Audits
4.2.1 Performance Evaluation Audit
The PE audits of differential pressure and mass measurements showed results within the
expected tolerances in Table 3-2. As noted in Section 3.4.1, the PE audit results of the reference
method analyses were also within the tolerances in Table 3-2.
14
-------
4.2.2 Technical Systems Audit
Observations and findings from this audit were documented and submitted to the Battelle
Verification Test Coordinator for response. No major findings were noted. All minor findings
were documented, and all required corrective actions were taken. The records concerning the
TSA are permanently stored with the Battelle Quality Manager.
4.2.3 Audit of Data Quality
At least 10% of the data acquired during the verification test were audited. Battelle's Quality
Manager traced the data 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 during the technical review process.
4.3 QA/QC Reporting
Each audit was documented in accordance with Sections 3.3.4 and 3.3.5 of the QMP for the
ETV AMS Center.(4) Once the audit report was prepared, the Battelle Verification Test
Coordinator ensured that a response was provided for each adverse finding or potential problem
and implemented any necessary follow-up corrective action. The Battelle Quality Manager
ensured that follow-up corrective action was taken. The results of the TSA were sent to the EPA.
15
-------
Chapter 5
Statistical Methods and Reported Parameters
The statistical methods presented in this chapter were used to verify the performance parameters
listed in Section 3.1.
5.1 Agreement with Standards
The agreement (A) of the LD500 with respect to the ammonia gas standards was assessed using
Equation 1:
c
I | f(X d
A = 100% ^
x
14
where d refers to the difference between the expected ammonia concentration from the dynamic
spiking and the two-minute average LD500 ammonia readings (corrected for the average
background concentration) during the spiking period, and x corresponds to the expected
ammonia concentration. Sd denotes the sample standard deviation of the differences, while t('n_, is
the t value for the 100(1 - a)th percentile of the distribution with n-1 degrees of freedom. The
agreement was determined for an a value of 0.025 (i.e., 97.5% confidence level, one-tailed). The
A value calculated in this way can be interpreted as an upper confidence bound for the relative
d
bias of the LD500, i.e., — , where the superscript bar indicates the average value of the
x
differences or of the reference values. The agreement with standards was calculated separately at
each of the spiking levels, using the 12 spike results at each level. The three most outlying
results (i.e., the three largest d values) were excluded in the calculation, i.e., the agreement was
calculated with nine data points at each spike level.
5.2 Linearity
Linearity was assessed by a linear regression analysis of the two-minute averages from the
dynamic spiking runs using the calculated ammonia concentrations as the independent variable
and the LD500 results as the dependent variable. Linearity is expressed in terms of slope,
intercept, and coefficient of determination (r2).
16
-------
5.3 Precision
Precision was calculated in terms of the average percent RSD of the LD500 readings over the
duration of each of the 12 spike and 12 zero two-minute periods during each dynamic spiking
run. For each two-minute period during each dynamic spiking run, all 10-second readings from
the LD500 were recorded, and the mean and standard deviation of those readings were
calculated. Precision (P) was then determined as:
where SD is the standard deviation of the LD500 readings and X is the mean of the LD500
readings in each period, and the overbar in Equation 2 indicates an average over all 12 periods.
Precision was determined with both ammonia and zero gas provided to the cell. Note that the
calculated precision is subject not only to the LD500 variability, but also to the variability of the
flue gas ammonia background and the dynamic spiking procedure. The precision observed with
zero gas in the calibration cell indicates the variability due to the flue gas background.
5.4 Calibration and Zero Drift
Calibration and zero drift are reported in terms of the mean, RSD, and range (maximum and
minimum) of the stable readings obtained from the LD500 in daily sampling of the same
ammonia standard gas and zero gas. These readings were obtained with the calibration cell
isolated from the cross-duct light path, i.e., the flue gas ammonia background was not a factor in
these tests. The readings were adjusted to account for the difference between the calibration cell
length and the cross-duct path length programmed into the LD500. Seven ammonia standard
readings and seven zero readings were used for this calculation. This calculation, along with the
range of the data, indicates the day-to-day variation in zero and standard gas readings.
5.5 Response Time
Response time was assessed in terms of both the rise and fall times of the LD500 in the dynamic
spiking runs. Rise time (i.e., 0% to 95% response time) was determined based on the 10-second
LD500 readings as the gas supplied to the calibration cell was switched from zero gas to the
ammonia standard during the dynamic spiking run. Once a stable response was achieved with
the gas standard, the fall time (i.e., the 100% to 5% response time) was determined based on the
LD500 readings as the gas supplied was switched from the ammonia standard back to zero gas
during the dynamic spiking run. The observed rise and fall times are highly dependent on the
replacement time of the gas standard or zero gas in the calibration cell. Rise and fall times were
determined for the LD500 using the data from the dynamic spiking run at 8.15 ppmwv above
background. A total of 11 data points were obtained relevant to response time for the LD500.
(2)
17
-------
Chapter 6
Test Results
The results of the verification test of the LD500 are presented in this section. The LD500
outputs ammonia concentrations without correction for flue gas conditions. Therefore, the
concentrations are on a wet volume basis, i.e., ppmwv. Note that all test results originate from
discrete 10-second readings reported by the LD500 without smoothing or averaging, as
described in Section 3.2
6.1 Agreement with Standards
Table 6-1 presents the data and resulting percent agreement of the LD500 with respect to each of
three ammonia gas standards used for the dynamic spiking runs during Week 3 of the
verification test. Shown in Table 6-1 are the two-minute average background-corrected LD500
readings, the expected ammonia concentrations, the resulting differences, and the overall A
values at each of the three spike concentrations calculated using Equation 1 in Section 5.1. The
calculated A was 10.4% at a 2.14-ppmwv spike concentration,7.9% at a 5.22-ppmwv spike
concentration, and 14.3% at an 8.15-ppmwv spike concentration. Note that these A values arise
from relatively small differences between the LD500 and standard results. For example, the
median of the differences listed in Table 6-1 is 0.43 ppmwv. Most often, the LD500 readings
were higher than the expected spike concentrations. In addition, since one average background
concentration was used for the duration of the spiking runs at each concentration, normal
variation in flue gas ammonia concentrations may have contributed to the differences between
expected and observed concentrations.
6.2 Linearity
Figure 6-1 presents the linear regression analysis of the LD500 response based on the two one-
minute averages obtained during the dynamic spiking runs versus the expected ammonia
response. The linear regression equation is shown in the figure and includes the 95% confidence
intervals of the slope and intercept in parentheses. This linear regression shows a slope of 1.198
(± 0.036), an intercept near -0.52 (± 0.21) ppmwv, and a coefficient of determination (r2) of
0.970.
18
-------
Table 6-1. Agreement of LD500 with Ammonia Gas Standards
Zero-
Expected
Zero-
Expected
Zero-
Expected
adjusted
adjusted
adjusted
adjusted
adjusted
adjusted
LD500
LD500
Differ-
LD500
LD500
Differ-
LD500
LD500
Differ-
response'3'
response®
ence
response'3'
response®
ence
response'3
response®
ence
(ppmwv)
(ppmwv)
(ppmwv)
(ppmwv)
(ppmwv)
(ppmwv)
' (ppmwv)
(ppmwv)
(ppmwv)
Spike Concentration 1
(c)
Spike Concentration 2
-------
12
10
y = 1.198 (+/- 0.036)x - 0.521 (+/- 0.209)
r2 = 0.9695
8
6
4
2
0
0
1
2
3
4
5
6
7
8
9
Expected NH3 Response (ppm)
Figure 6-1. Linear Regression of the LD500 NH3 Response vs. Expected Response
Table 6-2. Precision (% RSD) of LD500 During Dynamic Spiking Periods
Period
Average Reading
(ppmwv)
Average SD
(ppmwv)
Average RSD (%)
Spike l(a)
2.82
1.37
48.6
Spike 2(b)
6.26
1.49
24.1
Spike 3(c)
10.12
1.59
15.8
Zero
0.69
1.00
-
Zero
0.61
0.96
-
Zero
0.68
1.07
-
131 Using a spike gas with an ammonia concentration of 122 ppmwv.
lbl Using a spike gas with an ammonia concentration of 297 ppmwv.
lcl Using a spike gas with an ammonia concentration of464 ppmwv.
20
-------
6.4 Calibration and Zero Drift
Span and zero checks were conducted during Weeks 1 and 5 of the verification test. These checks
were conducted by flowing either a zero gas or a standard gas through the calibration cell, with
that cell isolated from the cross-duct light path. The 10-second LD500 readings were then
adjusted to account for the fact that the cell path length was only 0.13 m, rather than the 6.05-m
cross-duct path length used in the LD500 software. Table 6-3 presents the results of these checks,
showing the NH3 concentration in the cell, the LD500 readings, and the average, standard
deviation, %RSD, maximum, and minimum of those readings. Zero values determined during the
test show individual results from -0.46 to 14.0 ppmwv, though most results were zero. The span
values show standard deviations of 11.4 and 4.66 ppmwv in the two weeks, with no significant
trend with time. These standard deviations result in RSD values of 2.43% in Week 1 and 0.96%
in Week 5.
6.5 Response Time
LD500 response time was estimated using the 10-second readings generated during the dynamic
spiking runs. Because these checks were performed in an external calibration cell with a light
path through both the duct and the calibration cell, the LD500 readings were subject not only to
the LD500's time response, but also to the adsorptive nature of ammonia and the physical
changeover due to gas replacement in the calibration cell. The internal volume of the calibration
cell was 0.5 L, and the gas flow rate through the cell was 1.3 L/min. The LD500 would be
expected to display the final concentration only after the reference cell had changed volumes
several times. Therefore, the final concentration would not have been displayed until the standard
gas or zero gas had been flowing through the calibration cell for at least one minute.
Consequently, the response times indicated below should be taken as procedural changeover
times and not as the instrument response times of the LD500. Vendor-supplied information
indicates that true LD500 response times are generally a few seconds.
Table 6-4 presents the CEM rise and fall times during the dynamic spiking in Week 3. Figure 6-2
presents examples of the Week 3 rise and fall curves. Table 6-4 shows that the rise times
observed with the LD500 averaged 53 seconds. The observed fall times averaged 57 seconds. The
data leading to both of these average response times varied considerably. Because these
measurements were recorded with a light path through both the calibration cell and the flue gas,
these readings reflect both the variability of flue gas concentrations and the time needed to
replace the gas in the calibration cell. The observed response times are consistent with the
concentration profiles expected based on the cell volume and gas flow rate.
21
-------
Table 6-3. Calibration and Zero Drift for LD500 During Weeks 1 and 5
Gas Standard
Concentration
(ppmwv)
LD500
Reading
(ppmwv)
LD500 Standard Relative Standard
Average Deviation Deviation
(ppmwv) (ppmwv) (%)
LD500 LD500
Minimum Maximum
(ppmwv) (ppmwv)
Weeks 1 and 5 Zeros
0
0.0
0
-0.01 (-0.46)(a)
0
0.0
0
0.0
0
0.3 (14.0)
0
-0.01 (-0.46)
0
0.0
0.04 (1.86) 0.115 (5.36) —
-0.01 (-0.46) 0.3 (14.0)
Week 1 Spans
479
10.4 (484.6)
479
10.1 (470.6)
479
10.1 (470.6)
479
9.8 (456.6)
10.1 (470.6) 0.245 (11.4) 2.43
9.8 (456.6) 10.4 (484.6)
Week 5 Spans
479
10.4 (484.6)
479
10.3 (479.9)
479
10.5 (489.3)
10.4 (484.6) 0.100(4.66) 0.96
10.3 (479.9) 10.5 (489.3)
(a) First value shown is actual reading of the LD500; value in parentheses is corrected to adjust for the fact that the
light path passed through only the 0.13m calibration cell, and not the 6.05 m path programmed in the LD500.
6.6 Ease of Use
The LD500 has some features that make it easy to use. Other features add complexity to its use.
Once the LD500 was set up and calibrated, it required very little maintenance. Zero and span
checks (Section 6.4) revealed that the LD500 did display a slight variation around zero but
maintained its calibration setting very well. Other specific aspects of installation and operation
are discussed below.
6.6.1 Installation
Installation and LD500 setup at the site required two Opsis engineers. The engineers set up the
LD500 central unit in the instrument trailer along with a desktop computer, bundled cable to
reach from the analyzer to the duct (100 m), mounted and strung cable connecting the emitter/
receiver modules at the duct, aligned the optics, plumbed in purge air, and verified LD500
performance.
22
-------
Table 6-4. LD500 Rise and Fall Times
Week 3
Rise/Fall Time
Rise/Fall(a)
Time
(seconds)
Rise
12:29:09
93
Rise
12:54:26
66
Rise
13:22:19
31
Rise
14:04:18
42
Rise
14:39:52
32
Average Rise
53
Fall
12:10:39
54
Fall
12:40:16
28
Fall
13:07:53
73
Fall
13:49:17
47
Fall
14:22:06
106
Fall
14:57:19
31
Average Fall
57
(a) Flue gas background concentration approximately 1 ppmwv; dynamic spike concentration 8.15 ppmwv above
background.
Additional support from plant personnel was required to provide power at the duct for the
emitter/receiver, drop the cabling from the duct to the trailer, install a support bracket provided
by the vendor for the large receiver module, and mount the lenses on the flanges at the duct.
Teardown/demobilization required only one engineer and proceeded much more quickly.
6.6.2 Zero and Span Checks
As noted above, the emitter/receiver module of the LD500 has a built-in, temperature-controlled,
calibration gas cell (0.5 L volume, 13 cm length). A useful feature of the LD500 is that the laser
light can be directed along any of the following three paths: 1) through the duct alone, 2) through
the calibration cell alone, or 3) through both the duct and calibration cell. Such functionality in
combination with electronically operated solenoid valves allows for unattended zero/span checks.
The small size of the gas cell ensured that a stable response was achieved in a relatively short
time after the ammonia or zero nitrogen gas was introduced.
6.6.3 Dynamic Spiking
The laser light paths discussed in Section 6.7.2 also are useful when conducting dynamic spiking
studies. The variable laser light path makes it possible to perform dynamic spiking without
further LD500 modification. In this verification test, spiking was conducted manually, requiring
that trained personnel operate the LD500 data acquisition software in manual mode. However,
the LD500 system can be set up to carry out such spiking (known as an "add on" calibration)
automatically. The data adjustment after collection during dynamic spiking runs was somewhat
23
-------
July 28, 2003, Rise
15
13
11
9
7
5
3
1
¦1
Time (hh:mm:ss)
July 28, 2003, Fall
Time (hh:mm:ss)
Figure 6-2. Example LD500 Rise and Fall Time Plots
24
-------
cumbersome and required a correction for temperature to be recalculated at every duct
temperature observed during the dynamic spiking period. This adjustment can be done in the
LD500 data software; the vendor indicates that this capability will be in all future LD500
systems.
6.6.4 Data Handling
Data files were extracted from the data logging program by Opsis staff and copied onto floppy
discs for transfer to a Battelle laptop. The tab-delimited ASCII files were suitable for importing
directly to spreadsheet software.
6.7 Data Completeness
Data completeness was approximately 70% over the five-week test period. The LD500 required
intervention by vendor engineers in two instances to rectify errors that caused the data system to
lock up and, on one occasion, also caused the loss of data.
The first instance occurred when the LD500's autocalibration sequence failed. A feature of the
LD500 is its ability to conduct, at user-determined times, automatic zero/span checks using
compressed calibration gases. These calibration checks are typically scheduled during off-peak
hours as the LD500 cannot simultaneously measure duct ammonia concentrations and calibration
gases. On Monday morning of Week 1, the first scheduled zero/span failed and the LD500 froze,
thereby halting all data collection. After consultation with Opsis engineers, Battelle staff rebooted
the LD500 and successfully restarted data collection. The problem recurred later Monday after-
noon and early Tuesday morning. By later Tuesday morning, the Opsis engineer had arrived on
site, remedied the problem, and run a zero/span. No data correlating to reference sampling runs
were lost, but one zero/span was lost.
The second instance occurred on Monday of Week 4, when the hard disk in the LD500 failed.
Because the site was unattended, this failure was not detected until an Opsis staff member arrived
on site on Monday of Week 5, at which time he diagnosed the problem and ordered replacement
parts. The LD500 functioned intermittently throughout the day Tuesday, but it was not until early
Wednesday morning of Week 5 that the LD500 system was running normally. Approximately
nine days of data were lost, in addition to two scheduled daily zero/span checks.
6.8 Cost
The vendor indicated that the purchase cost of the complete LD500 system, as implemented for
this verification test, was approximately $55,000.
25
-------
Chapter 7
Performance Summary
Table 7-1 summarizes the results for each of the LD500 performance parameters. Note that all
quantitative results originate from discrete 10-second readings reported by the LD500 without
smoothing or averaging, as described in Section 3.2.
Table 7-1. Summary of LD500 Verification Results
Parameter
Performance Results
Comments
Agreement with
Standards
Relative
Accuracy
Linearity
Precision
Calibration and
Zero Drift
Response Time
Ease of Use
Completeness
10.4% at 2.14 ppmwv
7.9% at 5.22 ppmwv
14.3% at 8.15 ppmwv
Not calculated
Regression line = 1.198 (± 0.036)x
- 0.521 (±0.209) ppmwv, r2 = 0.970
48.6%) RSD at 2.82 ppmwv
24. P/o RSD at 6.26 ppmwv
15.8%o RSD at 10.12 ppmwv
Zero drift averaged 1.86 ppmwv
Span RSD values = 0.96 to 2.43%
Rise times average 53 seconds
Fall times average 57 seconds
Generally easy to use
70%o data capture
Results of three concentration
levels with 12 data points each;
nine data points used in each
calculation; median difference from
expected value = 0.43 ppmwv
Reference sampling location
unrepresentative of duct ammonia
concentrations^
Calculated over range of 2.14 to
8.15 ppmwv, 36 total data points
Discrete 10-second data, no
smoothing; variability due partly to
the variability of background
ammonia concentration in the duct
Minimal span drift over the five-
week test
Observed response times largely
due to concentration changeover in
the test cell
Missing data due to data system
lock up and hard drive failure
^ Reference sampling port was improperly located and did not allow sampling across width of duct. Mapping of
ammonia concentrations at points along the CEM light path confirmed that sampling at reference port could not
adequately determine duct ammonia concentrations.
26
-------
Chapter 8
References
1. Test/QA Plan for Verification of Continuous Emission Monitors for Ammonia at a Coal-
Fired Facility, Battelle, Columbus, Ohio, June 2003.
2. Procedure for the Collection and Analysis of Ammonia in Stationary Sources, Conditional
Test Method 027, U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, August 1997.
3. Standard Specification for Collection and Analysis of Ammonia Nitrogen in Flue Gas Using
Wet Chemical Sampling and Specific Ion Analysis, Draft Standard, ASTM, West
Conshohocken, Pennsylvania, October 2000.
4. Quality Management Plan (QMP) for the ETV Advanced Monitoring Systems Center, U. S.
EPA Environmental Technology Verification Program, prepared by Battelle, Columbus,
Ohio, Version 4.0, December 2002.
27
------- |