United States
Environment Protection
Agency
Environmental Scienoet Research
Laboratory
ne»earch Triangle Pant NC 27711
EPA-600/2-79-073
April 1979
Research and Development
Evaluation of
Particulate Mass
Monitors Using the
Beta Gauge
Technique
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield. Virginia 22161.
-------
EPA-600/2-79-073
April 1979
EVALUATION OF PARTICULATE MASS MONITORS USING
THE BETA GAUGE TECHNIQUE
by
Meryl R. Jackson
Particle Data Laboratories, Ltd.
Elmhurst, Illinois 60126
Contract No. 68-02-1216
Project Officer
John Nader
Emission Measurements and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.,
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
11
-------
ABSTRACT
A field study was conducted to evaluate two commercially avail-
able beta gauge instruments for measuring particulate mass
concentrations in stationary source emissions. Performance of
the instruments was compared with a manual method of measure-
ment at a ferrite rotary-kiln calciner, at a slurry cement kiln
with an electrostatic precipitator, and at an oil-fired boiler.
Tests were conducted over a 168-hour period to establish instru-
ment accuracy, calibration error, drift, and system reliability.
Descriptions of the instruments, test programs and test sites
are presented together with a detailed summary of the experi-
mental data.
The accuracy of the beta gauge instruments was strongly depend-
ent upon the sampling characteristics of the extractive probes.
The instruments tested were not capable of correctly measuring
the particulate concentration in the stack, nor of operating
continuously for a 168-hour period. In the case of the cement
kiln, the particle concentration measured by the beta gauge
instruments correlated well with the concentration determined
from the filter catch portion of the manual method, but not
with the fiber plus probe catch.
Particle deposition in the probe of the beta instruments was as
high as 86% (average) for the cement plant emissions.
111
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1
2. Conclusions 2
3. Recommendations 3
4. Site Descriptions 4
5. Instrument Descriptions 6
6. Text Program 8
7. Field Experience 12
8. Calibration Checks of Beta Gauges 16
9. Results 17
10. Discussion 36
References ' 39
Appendices 40
A. Instrument Design and Operation - Argos 1 41
B. Instrument Design and Operation - Model 2414 51
C. Description of the Particulate Reference Method 56
Using the EPA Sampling Train
D. Statistical Analysis 62
v
-------
FIGURES
Number Page
1 Point Location 10
2 Filter Catch Only as Reference, LSI 2!)
Instrument - Site 1
3 Filter Catch Plus Probe Washing As Reference, 2(>
LSI Instrument - Site 1
4 Filter Catch Only as Reference, LSI Instrument - 27
Site 2
5 Filter Catch Plus Probe Washing as Reference, 28
LSI Instrument - Site 2
6 Filter Catch Only as Reference, RAC Instrument - 29
Site 2
7 Filter Catch Plus Probe Washing, RAC 30
Instrument - Site 2
8 Filter Catch Only as Reference, LSI Instrument - 31
Site 3
9 Filter Catch Plus Probe Washing as Reference, 32:
LSI Instrument - Site 3
10 Filter Catch Only as Reference, RAC Instrument - 33
Site 3
11 Filter Catch Plus Probe Washing, RAC 34
Instrument - Site 3
VI
-------
TABLES
Number Page
1 Test Program - Site 1 9
2 Program Test Conditions - Site 1 11
3A Site 1 - Data for L.S.I. -
Observations with Filter Catch Only as Reference 18
3B Site 1 - Data for L.S.I. -
Observations Using Filter Catch Plus Probe
Washings as Reference 19
4A Site 2 - Data for RAC and L.S.I. -
Observations with Filter Catch Only as
Reference 20
4B Site 2 - Data for RAC and L.S.I. -
Observations with Filter Catch Plus Probe
Washings as Reference 21
5A Site 3 - Data for RAC and L.S.I. -
Observations with Filter Catch Only as Reference 22
5B Site 3 - Data for RAC and L.S.I. -
Observations Using Filter Catch Plus Probe
Washings as Reference 23
6 Regression Analysis of Data 24
7 Material Masses Deposited in the Probe System 35
8 Absolute Mean Value Plus 95% Confidence 37
9 Absolute Mean Value Plus 95% Confidence 37
10 Instantaneous Drift Using Linear Interpolation 38
VII
-------
ACKNOWLEDGMENTS
The support of Lear Siegler Inc. and Research Appliance Corpora-
tion in their extensive technical support of the testing program
is gratefully acknowledged. Our thanks are also due for the
services, assistance and hospitality extended to us at each of
the three testing sites. We hope the information gained by our
hosts to some extent off-sets their efforts on our behalf.
Vll 1
-------
SECTION 1
INTRODUCTION
Present methods recommended by EPA for measurement of stack
particulate loading are manual and provide information averaged
over a minimum period of two hours. Particulate code compliance
of any stationary emission source or performance evaluation of
any control equipment is judged upon the results of three tests
of between two to four hours duration performed within a one or
two day period. A need exists for an automatic and more or less
continuous measurement system of particulate emission rate from
stationary sources similar to those presently available for mea-
surement of sulfur dioxide and nitrogen oxides.
This report presents the result of an experimental study of the
performance of two commercially available automatic and more or
less continuous particulate mass monitoring systems. The major
criteria of performance was measurement accuracy relative to the
EPA manual test method, and reliability of unattended instrument
operation over a 168 hour period.
Three test sites were selected for simultaneous measurement of
particulate loading by the one manual and two automatic systems.
The sites were selected to provide different effluent character-
istics with respect to particle physical properties, effluent gas
concentration, and particulate emission rates.
The data are presented and evaluated to provide assistance in the
formulation of emission monitoring requirements, to furnish manu-
facturers with technical guidelines on the present and desired
performance characteristics, and to inform potential users of
automated particulate monitoring equipment capabilities.
-------
SECTION 2
CONCLUSIONS
Two commercially available beta gauge methods for measurement of
stack gas particulate loading concentrations were inaccurate in
their measurement of total dust loading concentrations.
Both instruments had repeated mechanical and electronic failures
indicating poor reliability and impacting on accuracy of measure-
ments .
The degree of inaccuracy was affected largely by removal of par-
ticulate by deposition in the extractive probes before reaching
the measurement system.
The investigation showed that the particulate concentrations as
measured by the beta gauges bear no significant statistical
correlation to the measurements obtained by the manual test
method for the ferrite calciner or the oil fired boiler.
In the case of the cement kiln effluent the beta gauge measure-
ments correlated well with the filter catch of the manual method.
Neither beta gauge instrument operated continuously for 168 hours
without failure at any of the three sites tested.
-------
SECTION 3
RECOMMENDATIONS
The physical principles involved in the measurement of mass by
beta attenuation are well established and do provide a potential
means of automatic mass concentration measurement. The instru-
ments tested do not perform as intended mainly due to deficiencies
in the interface between the stack gas and the filter tape. If
adequate transportation of particles from the stack gas to the
filter can be achieved, the system concept would appear to be
viable.
It is therefore recommended that research and development effort
be expended in correcting the deficiencies of the sampling inter-
face. It is suggested that an alternative approach to previous
effort in this area is to provide a filter transport system so
that the particulates are collected on the filter at the sampling
nozzle in the stack, and are transported to the measurement
system as a filter deposit rather than as an aerosol.
-------
SECTION 4
SITE DESCRIPTIONS
The following three sites were selected for performing the tests:
SITE 1
The stack selected for testing of instruments at the first site
was a natural draft stack venting a gas fired hard ferrite
rotary calcining kiln, approximately 9 meters long and 1.3
meters diameter. A raw material slurry mixture comprising a
finely divided metal oxide and other salts suspension in water
was sprayed into the upper end of the calciner. The calciner
operating temperature was near 1300°C. Convective cooling arid
dilution air reduced the gas temperatures in the stack generally
to the range 260°C to 370°C.
The stack effluent contained approximately 18% moisture and two
components of particulate material. One component appeared a.s a
submicron fumed material, while the other was apparently red
iron oxide in the 10 micrometer and larger size range. Gas a.nal-
ysis indicated that the C02 and 02 concentrations of 4.6% and
14.0% respectively were governed only by the natural gas combust-
ion heating in the calciner.
The stack comprised a nominal 60 cm diameter vertical steel tube
approximately 13.6 meters in height and capped with a rotating
cowl. The calciner breeching entered the stack at a height of
approximately 4.6 meters. The sampling plane was located at
approximately 6 meters above the breeching and 3 meters below
the top of the stack.
Under normal operation the calciner was fed with slurry 24 hours
a day, 5 days a week, alternately from one of two slurry tanks.
At the week-ends the slurry feed was stopped, and the calciner
temperature reduced to 870°C. At approximately four hour inter-
vals during operation, the natural gas and slurry feeds were
stopped while the fire ring was removed by rodding and the gas
flame photoelectric cell sight tube was cleaned. The slurry
feed was stopped for approximately 5 to 10 minutes during this
process.
-------
SITE 2
A wet slurry cement kiln exhaust was selected as the second
testing site. Sampling was performed in the near-vertical duct
between the electrostatic precipitator and the ID fan prior to
exhausting to the stack. At the testing location the gas was
flowing vertically downward through a rectangular cross section-
al duct 2.14 meters x 1.27 meters. The average gas temperature
was 175°C., gas moisture was approximately 33%, and CC>2 and C>2
concentrations were near 8% and 11% respectively. The average
particle size of the collected particles was approximately 25
micrometers, but the size distribution extended over the range
2 to 80 micrometers.
The process operated on a 24 hour per day basis at an approxi-
mately constant input feed rate and quality. The 4 precipitator
banks rapped sequentially on a 30-minute cycle.
SITE 3
Testing was performed at this site in the stack of an oil-fired
boiler used for heating purposes only. The rated capacity of
the boiler was 100 MBTU/hr. and was fired with #6 oil. Testing
was performed in the 1.55 meter diameter stack at a plane 9.15
meters above the boiler breeching and 1.55 meters below the top
of the stack. The average stack gas temperature was 250°C, the
moisture content was approximately 2% and the carbon dioxide and
oxygen compositions were approximately 15% and 4% respectively.
The boiler was equipped with forced draft combustion air fans
only resulting in stack gas velocities of near 4 meters/sec.
-------
SECTION 5
INSTRUMENT DESCRIPTIONS
BETA GAUGE INSTRUMENTS
General Principle of Operation
In the general case the mass of particulate material filtered
from a measured volume of stack gas is determined by the attenu-
ation of a beam of beta particles in passing through the sample.
The attenuation is related to the mass of material by the equa-
tion
N. = N e
1 O
where N is the rate at which beta particles pass through the
filter material before collection of the sample,
Nj_ is the rate at which beta particles pass through the
filter material and collected sample,
A is the area of the filter through which the beta parti-
cles pass to reach the detector,
m is the mass concentration of sample collected on the
filter, 4
a is the mass absorption coefficient of C beta particles
and is approximately constant for elements with atomic
number to mass ratios of 0.45 to 0.50.
Recasting equation (1) we obtain
In (N0)/(Ni) = Ama (2)
= Km
where K is a constant.
Since for a system of constant filter deposition and source-
detector geometry, m, the mass concentration is linearly related
to total mass of deposition, M, by the ratio of the deposit area
to the area through which the beta particle pass.
-------
Thus
M = K' In (No)/(Ni) (3)
where K1 is a constant.
If the volume of the gas passed through the filter under stated
conditions is V, the mass concentration of the gas under the same
conditions is given by
M/V = K'/V x In (No)/(Ni) (4)
From equation (4) , measurement of particulate concentration in a
gas stream is derived from measurement of the sampled gas volume
and the rate at which beta particles are received after passing
through the clean filter and the same filter with particulate
deposit. This relationship is the basis upon which the beta
gauges operate.
More detailed descriptions of the actual instruments manufactured
by Lear Siegler Inc. and Research Appliance Corporation used in
this program are presented in Appendices A and B, respectively.
MANUAL METHOD
The manual method of emissions testing was performed using a
Scientific Glass Blowing sampling train conforming to the require-
ments detailed in Method 5 of the Federal Register 36, #247
Part II (Dec. 23, 1971) (Ref. 1). In this technique particulate
material is withdrawn isokinetically from the stack through a
heated probe.
The particulate material is removed primarily by a cyclone and
back-up filter system. The particulate catch includes however
all particulate material retained on the walls of the probe and
glass-ware up to and including the filter holder. This fraction
of the particulate catch is collected by brushing and washing the
contaminated surfaces, and evaporating these washing to dryness.
Further details of the design and operation of this manual test
method are presented in Appendix C.
-------
SECTION 6
TEST PROGRAM
The purpose of the test program was to compare the particulate
emission rate of each beta gauge with the rate determined by the
manual method recommended by EPA and known as Method 5 as detail-
ed in Ref. 1. This method is the reference method.
During each test run the beta gauges were operated automatically
at the measurement rate determined optimum for each site. Test
runs were performed using the manual method as detailed before.
Comparison was made of the average beta gauge emission rate
measurement during the manual test run and the manual method
emissions rate measurement.
TEST PROGRAM, SITE 1
Originally the program was planned to compare the particulate
emission rate as measured by the two beta gauges and the manual
method, under the following conditions:
1. Single point beta gauge measurement with full traversing
of the manual method.
2. Single point beta gauge measurement with single point
measurement by the manual method.
However due to mechanical and electrical problems with the
Research Appliance Company instrument, tests were conducted to
compare only the Lear Siegler instrument with the manual method.
Table 1 details the program and identifies the test points shown
in Figure 1. Table 2 summarizes the test conditions.
-------
Table 1. TEST PROGRAM - SITE 1
Day
1
Date
1/17
2 1/18
3 ! 1/19
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1/20
1/21
1/22
1/23
1/24
1/25
1/26
1/27
1/28
1/29
1/30
1/31
2/ 1
2/11
2/12
2/25
2/26
3/11
3/12
3/13
3/14
3/15
3/16
3/18
Lear Siegler
Port Position
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
Centroid 3
Centroid 3
Centroid 4
Centroid 4
Av. Velocity
Av. Velocity
Av. Velocity
Av. Velocity
Av. Velocity
Av. Velocity
Av. Velocity
1
2
5
6
Centroid 3
Centroid 3
Centroid 4
Av. Velocity
Av. Velocity
1
2
5
6
Av. Velocity
Av. Velocity
Clean
Probes
X
X
X
X
X
EPA Trains
No. 1 Hour Runs
Travers- Single
ing Point
2
2
-
-
2
2
2
2
2
2
2
2
2
2
2
—
—
2
2
2
2
2
2
2
2
-------
PORT 2
PORT I
Figure 1. Point Location
10
-------
Table 2. PROGRAM TEST CONDITIONS - SITE 1
No. Days run
No. LSI test days
No. RAC test days
No. Manual tests total
Traversing
Single point
No. Probe cleaning
Days run between cleaning
Actual days between cleaning
26
26
0
42
20
22
4
1,4,6
1,4,6
,9
,41
TEST PROGRAM, SITE 2
Examination of the data obtained from site 1 showed little or no
valid statistical relationship between the emission rates meas-
ured by the beta gauge instrument and the manual method. It was
therefore decided that the tests performed at site 2 would be
conducted to provide more data under similar test conditions,
rather than to attempt investigation of many variables. Thus at
site 2 the nozzles of all three instruments were located within
the central velocity plateau of the gas flow. They were located
closely enough together to provide reasonably similar gas condi-
tions without creating nozzle induced disturbances.
26 one-hour test runs were made with the manual method with the
precipitator operating under normal conditions. 6 half-hour runs
were performed during which time certain fields of the precipita-
tor were switched off in order to create a higher emission dust
loading.
TEST PROGRAM, SITE 3
30 one-hour runs were performed with the manual method when all
three nozzles were located in the central velocity plateau region
of the stack.
11
-------
SECTION 7
FIELD EXPERIENCE
SITE 1
Stack Characteristics
The data developed during the tests showed the stack gas quality
to be extremely variable in temperature, and particulate loading
during normal calciner operation. Gas velocity and moisture con-
tent were largely constant during normal operation. During cal-
ciner idle operation when temperature only was maintained without
any material throughput, effluent particulate loadings were low
and exhibited small absolute variations in emission rate.
During normal operation the following ranges were experienced:
gas velocity mps 3.3-4.2
gas temperature °C 260-370
moisture % 15- 18
particulate loading mg/m^ 250-850
During idle operation the following ranges were experienced:
gas velocity mps 3.3-4.2
gas temperature °C 260-370
moisture % 5-6
particulate loading mg/m^ 15- 45
In the early stages of testing the calciner operated essentially
24 hours per day, 5 days per week. Later in the tests, plant
processing difficulties reduced operational time to essentially
16 hours per day, 5 days per week.
Argos I; Lear Siegler Trie.
The following problems were encountered in operation of the
Argos I and attempts were made by Lear Siegler to find solutions.
1. Condensation of water from the sampled gas in the pump and
immediate pump connections caused repeated damage to the rotary
vane type of pump.
12
-------
A condensate knock-out jar was added upstream of the pump, and
the air by-pass system, used to control sampling rate, was heated
to raise the dew point. These modifications did not entirely
correct the problem and a Roots blower was subsequently used.
Since this type of pump does not rely upon the free movement of
carbon or other vanes, and since a high torque electric motor
was used to drive the pump, it was hoped that the system would
have a greater tolerance to condensation. Despite this, some
fouling of the pump occurred which during shut-down conditions
caused severe sticking of the pump motor. Additional condensate
control is needed under cold ambient temperature and high stack-
gas moisture conditions.
2. Deposition in the probe was a recurring problem despite
efforts to heat the probe to avoid condensation. No cure was
found for this cause of loss of sample prior to collection on
the filter tape.
3. The probe solenoid valve was subject to particulate contami-
nation and consequent jamming. As in the case of probe deposi-
tion no permanent solution to the problem was found. Frequent
cleaning by-passed the problem.
4. Deposition occurred inside the tubes within the instrument.
Most deposit formed in the bell-housing and vertical tube above
the filter tape. However material collected in the orifice meter
downstream of the filter. This had the effect of modifying the
orifice calibration and thus changing the volume of gas sampled.
More importantly it indicated a leakage of particulate through
or around the filter. Since the material was red in appearance
the possibility was discounted that the build-up was of volatile
material cooling and condensing on expansion after the orifice.
No solution was found to this problem.
5. During operation of the instrument an electronic failure
occurred in the counting circuit. The failure was diagnosed by
Lear Siegler personnel and corrected.
With the exception of the above listed particulate fouling prob-
lem, the instrument operated well during the test period, seemed
to be unaffected by extremes of cold and other winter environ-
mental factors, and was easy to install and relocate at the site.
Stack Monitor Model 2412; Research Appliance Co.
This beta gauge failed to operate satisfactorily during the
entire test period. Problems were encountered in the computer
control system, the gas flow measurement and moisture control
unit. Prior to termination of the test period the instrument
was returned to the manufacturers for modification and check-
out.
13
-------
Valid data was not obtained with this instrument and thus
neither field data nor its analysis is presented for the RAC
instrument at this site.
Manual Method
The manual_method was performed as detailed in Reference 1 with
the exception that single point measurements were made as indica-
ted in Table 1. Duplicate one-hour tests were performed each day,
Operation parameters such as nozzle size and nomograph settings
are shown on the copies of typical raw data sheets presented in
Appendix D.
SITE 2
Stack Characteristics
The cement calcining kiln operated at near constant conditions
throughout the test period. Typical parameter values were:
gas velocity mps 13.7
gas temperature °C 182
moisture % 29- 34
particulate loading
mg/m3 20- 1009
The high emission rates observed were due to the deliberate re-
duction of precipitator efficiencies by sequentially switching
off selected fields. Under normal conditions the particulate
loading ranged from 20 to 200 mg/m3.
Argus I: Lear Siegler Inc.
There were no significant mechanical problems with this instru-
ment. However, the high moisture content of the stack gas
showed problems in filter tape moisture pick up. This was evi-
denced by tape breakage and visible wet spots.
Additional tape heaters were installed and appeared to correct
the condensation problem.
Stack Monitor Model 2412: Research Appliance Co.
There were no significant mechanical problems with this instru-
ment. However, the computer control program occasionally failed
and automatically shut off further automatic instrument opera-
tion. It was not determined whether brief line interrupts were
the cause of this.
14
-------
Manual Method
No unusual problems were encountered in using the manual instru-
mentation.
SITE 3
Stack Characteristics
The oil fired boiler was used for heating purposes only. Since
more than one boiler was needed to supply sufficient stream for
the total load, it was possible to operate this boiler under
near constant conditions.
Typical parameters were:
gas velocity mps 3.8-5.0
gas temperature °C 227-271
moisture % 2
particulate loading mg/m 61-150
Argos I; Lear Siegler Inc.
Mechanical and electrical problems were encountered which re-
quired constant attention during testing of this site. These
problems centered around failure of the tape transport and by-
pass valve solenoid control systems and electric failure of the
counting circuitry. The instrument failed to operate for 168
hours.
Stack Monitor Model 2414: Research Appliance Co.
Dilution air-line freeze-ups caused some problems which were
overcome by use of heating tapes and moisture knock-out jars.
Problems were also encountered by chattering of the dryer drain
solenoid valve. The cause of this malfunction was not estab-
lished.
Manual Method
No unusual problems were encountered in using the manual method.
15
-------
SECTION 8
CALIBRATION CHECKS OF BETA GAUGES
At sites 1, 2 and 3, tests were performed to deposit stack parti-
culates on preweighed pieces of filter. The mass collected was
monitored by the beta instruments and the filter pieces were re-
turned to the laboratory for gravimetric analysis.
At sites 2 and 3 attempts were made to introduce known masses of
particulate material into the probe nozzles of the beta instru-
ments. Comparison was attempted of the known mass introduced
and the mass as measured by the beta gauges. Both coarse and
fine powders were used and were introduced as aerosols from a
squeeze bottle and as bulk powders by pouring.
At site 3 a check was made of the quantity of material passing
through one thickness of filter paper by inserting up to 3 layers
in the sampling stage of the Lear Siegler instrument.
Instrument drift of the Lear Siegler was observed using the
built-in zero and reference calibrate conditions. The RAC in-
strument does not have these automatic calibrate-check capabili-
ties.
16
-------
SECTION 9
RESULTS
Typical field data sheets and data tabulations are presented for
sites 1, 2 and 3 in Appendices D, E and F. Details of the sta-
tistical analysis are presented in Appendix G.
Since considerable deposition of particulate material was observ-
ed in the probe system of the Argos I, analysis has been perform-
ed using both the total particulate and filter catch only of the
manual method as reference. The occurrence of particle deposi-
tion in the RAC instrument is masked by the back-purge system,
and no direct estimate of the extent of deposition could be
determined.
Tables 3, 4 and 5 summarize the data obtained for sites 1, 2 and
3 respectively and show the filter-only and filter-plus-probe-
washings base for comparison.
Regression analysis was performed and is summarized in Table 6.
The spread of the data and the regression lines are shown in
Figures 2 through 11.
The quantities of the particulate material removed from the
probe and pre-filter pipe-work on the beta gauge are shown in
Table 7.
Alumina powder in the 10 micrometers size range and sodium
bicarbonate in the size range up to 150 micrometers were used for
these tests. Deposition of both these powders in the probe sys-
tem generally resulted in near zero mass readings by the beta
gauges. Examination of and cleaning of the probe systems after
these test confirmed that the majority of the powders introduced
into the probe tip failed to reach the filter tapes.
Up to 3 layers of filter tape were placed simultaneously in ser-
ies in the filter holder of both beta gauge instruments at site
3. Visual examination of the three layers of filter tapes showed
discoloration of all three, indicating passage of material through
the filter medium. Gravimetric analysis showed that approximate-
ly 2% of the sample was passing the first filter and a gravime-
trically undetectable quantity passed the second filter.
17
-------
Table 3A. SITE 1 - DATA FOR L.S.I.
OBSERVATIONS WITH FILTER CATCH ONLY AS REFERENCE
Reference
Observation
Mq/M3
170.9
120.4
127.2
39.5
12.5
112.8
222.4
50.2
201.3
200.6
183.1
210.7
10.8
254.3
43.0
245.3
145.9
62.9
222.2
212.0
202.8
131.2
63.4
217.9
216.2
247.0
247.9
253.1
227.5
97.0
257.3
93.4
LSI
Observation
Ma/M3
473.5
358.3
359.5
102.8
49.4
406.5
393.6
381.6
339.0
392.2
379.9
357.7
72.9
183.7
95.5
92.9
59.2
304.7
349.2
263.9
302.1
267.4
312.2
303.9
432.7
240.9
236.2
279.9
263.3
345.5
422.2
328.0
LSI%
Change from
Reference
177.1
197.6
182.6
160.1
296.0
260.4
77.0
660.2
68.4
95.5
64.5
69.8
57.5
- 27.8
122.0
- 62.1
- 59.5
384.3
57.2
24.5
49.0
103.9
392.4
39.5
105.9
- 2.5
- 4.7
10.6
15.7
256
107.5
251
Average = 145.2 Std. Dev. = 170.4
Note: if the reference value was under 10 Mg/M3, it was omitted
from this calculation due to the sensitivity of this type of
measurement to such readings.
18
-------
Table 3B. SITE 1 - DATA FOR L.S.I.
OBSERVATIONS USING FILTER CATCH PLUS PROBE
WASHINGS AS REFERENCE
Reference
Observation
Ma/M3
265.6
260.1
235.5
48.0
16.6
213.4
286.5
56.5
473.3
219.6
334.2
237
287.0
75.3
28.5
292.3
147.7
297.5
281.7
214.5
347.1
255.9
232.9
192.0
139.6
489.9
546.8
457.8
440.6
567.8
408.2
252.6
343.2
Average = 40.1
LSI
Observation
Mcr/M3
473.5
358.3
359.5
102.8
49.4
406.5
393.6
381.6
339.0
392.2
422.2
379.9
357.7
82.5
72.9
183.7
95.5
92.9
59.2
304.7
349.2
263.9
302.1
267.4
312.2
303.9
432.7
240.9
236.2
278.9
263.3
345.5
328.0
LSI%
Change from
Reference
78.3
37.8
52.7
114.0
198.4
90.5
37.4
575.9
- 28.4
78.6
26.6
60.3
24.6
9.5
155.8
- 37.2
- 35.3
- 68.8
- 79.0
42.1
.6
3.1
29.7
39.3
123.6
- 37.9
- 20.9
- 47.4
- 46.4
- 50.7
- 35.5
368
- 4.4
Std. Dev. = 116.2
19
-------
Table 4A. SITE 2 DATA FOR RAC AND LSI -
OBSERVATIONS WITH FILTER CATCH ONLY AS REFERENCE
Reference
Observation
Mg/M3
11.04
10.25
18.38
25.17
21.19
28.39
27.26
13.82
12.15
38.52
44.86
69.64
93.98
188.40
228.80
185.20
221.30
18.48
12.19
RAC
Observation
Mg/M3
12.82
4.36
18.74
18.61
18.74
24.58
19.87
1.65
.18
32.16
27.10
42.34
58.48
139.60
142.30
129.20
99.97
2.49
1.53
LSI
Observation
Mg/M3
15.31
12.74
30.90
28.50
27.52
30.83
27.21
16.36
14.49
45.93
46.89
79.28
139.30
184.40
220.40
183.50
176.90
31.32
18.68
RAC %
Change from
Reference
+ 16.2
- 57.5
+ 2.0
- 26.1
- 11.6
- 13.3
- 26.2
- 88.4
- 98.5
- 16.5
- 39.6
- 39.2
- 37.8
- 25.9
- 37.8
- 30.2
- 54.8
- 86.5
- 87.4
LSI %
Change from
Reference
+ 38.7
+ 24.3
+ 68.1
+ 13.2
+ 29.9
+ 8.6
.2
+ 18.4
+ 19.3
+ 19.2
-I- 4.5
+ 13.8
+ 48.2
- 2.1
- 3.7
- 0.9
- 20.1
+ 69.4
+ 53.2
Average= - 40.0 21.2
Std. Dev= 31.9 24.8
Note: If the reference value was under 10 Mg/M3, it was omitted
from this calculation due to the sensitivity of this type of
measurement to such readings.
20
-------
Table 4B. SITE 2 DATA FOR RAC AND LSI -
OBSERVATIONS WITH FILTER CATCH PLUS PROBE WASHINGS AS REFERENCE
Reference
Observation
Mg/M3
119.20
127.80
87.49
98.57
86.14
98.21
62.59
216.80
90.94
142.50
158.90
137.60
26.90
20.10
19.19
92.22
101.30
157.30
222.30
755.90
902.60
1009.00
709.40
181.60
134.40
RAC
Observation
Mg/M3
23.89
12.82
4.36
3.72
3.22
4.59
18.61
18.87
18.61
18.74
24.58
19.87
1.65
.18
.14
32.16
27.10
42.30
58.48
139.60
142.30
129.20
99.97
2.49
1.53
LSI
Observation
Mg/M3
28.45
15.31
12.74
15.37
15.87
16.51
20.28
30.90
28.50
27.52
30.83
27.21
16.36
14.49
9.76
45.93
46.89
79.28
139.30
184.40
220.40
183.50
176.90
31.32
18.68
RAC %
Change from
Reference
- 80.0
- 90.0
- 95.0
- 96.2
- 96.3
- 95.3
- 70.3
- 91.3
- 79.5
- 86.8
- 84.5
- 85.6
- 93.9
- 99.1
- 99.3
- 65.1
- 73.2
- 73.1
- 73.7
- 81.5
- 84.2
- 87.2
- 85.9
- 98.6
- 98.9
LSI %
Change from
Reference
- 76.1
- 88.0
- 85.4
- 84.4
- 81.6
- 83.2
- 67.6
- 85.8
- 68.7
- 80.7
- 80.6
- 80.2
- 39.2
- 27.9
- 49.1
- 50.2
- 53.7
- 49.6
- 37.3
- 75.6
- 75.6
- 81.8
- 75.1
- 82.8
- 86.1
Average = - 86.6 - 69.8
Std. Dev. = 10.0 17.9
Note: If the reference value was under 10 Mg/M3, it was omitted
from this calculation due to the sensitivity of this type of
measurement to such readings.
21
-------
Table 5A - SITE 3 DATA FOR RAC AND LSI -
OBSERVATIONS WITH FILTER CATCH ONLY AS REFERENCE
Reference
Observation
Mg/M3
72.25
71.79
59.14
56.14
74.84
68.98
75.07
70.21
64.86
52.00
44.00
42.89
57.93
56.69
56.87
56.72
52.73
56.98
59.36
60.13
58.55a
79.64
62.79
65.67
51.75
60.59
54.10
50.57
49.26
51.95
RAC
Observation
Mg/M3
33.95
36.29
37.21
34.48
28.56
25.21
24.64
19.92
21.02
18.22
18.20
16.54
22.71
17.15
152.40
22.12
16.16
15.91
25.91
22.24
27.08
26.34
28.25
27.19
LSI
Observation
Mg/M3
61.28
60.76
43.98
44.65
45.25
29.10
44.78
33.34
40.43
34.89
40.32
35.43
38.43
31.55
30.95
28.70
20.74
25.48
22.51
19.52
16.95
21.56
16.11
22.34
17.63
11.67
27.28
18.95
12.91
RAC %
Change from
Reference
- 53.0
- 49.4
- 37.1
- 38.6
- 44.0
- 42.7
- 42.6
- 64.9
- 63.0
- 67.9
- 65.5
- 70.9
- 61.7
- 71.5
+160.3
- 72.2
- 74.3
- 75.8
- 49.9
- 63.3
- 49.9
- 47.9
- 42.7
- 48.4
LSI %
Change from
Reference
- 15.2
- 15.4
- 25.6
- 20.5
- 39.5
- 57.8
- 40.3
- 52.5
- 37.6
- 31.6
- 6.0
- 38.8
- 32.2
- 44.5
- 45.4
- 45.6
- 63.6
- 57.1
- 62.6
- 66.7
- 78.7
- 71.1
- 75.5
- 56.8
- 70.9
- 78.4
- 46.1
- 61.5
- 75.1
Average = - 56.4 - 48.0
Std. Dev. = 12.5 20.3
Rejected as rogue.
22
-------
Table 5B. SITE 3 DATA FOR RAC AND LSI -
OBSERVATIONS USING FILTER CATCH PLUS PROBE WASHINGS AS REFERENCE
Reference
Observation
Mg/M"3
91.97
88.19
72.96
69.83
99.26
101.20
130.70
146.10
150.90
77.61
103.20
103.10
69.81
75.52
74.98
96.53
68.83
72.26
88.45
97.38
114.10
139.90
79.90
99.80
80.65
104.50
79.75
75.34
61.03
62.24
RAC
Observation
Mg/M
33.95
36.29
37.21
34.48
28.56
25.21
24.64
19.92
21.02
18.22
18.20
16.54
22.71
17.15
15.24
22.12
16.16
15.91
25.91
22.24
27.08
26.34
28.25
27.19
LSI
Observation
Mg/M
61.28
60.76
43.98
44.65
45.25
29.10
44.28
33.34
40.43
34.89
40.32
35.43
38.43
31.55
30.95
28.70
20.74
25.48
22.51
19.52
16.95
21.56
16.11
22.34
17.63
11.67
27.28
18.95
12.91
RAC %
Change from
Reference
-63.1
-58.9
-49.0
-50.6
-63.2
-75.6
-76.1
-73.6
-72.0
-81.1
-73.6
-77.1
-74.3
-82.4
-33.6
-84.2
-79.8
-84.1
-67.9
-78.7
-66.4
-66.0
-53.7
-56.3
LSI %
Change from
Reference
-33.4
-31.1
-39.7
-36.2
-54.4
-71.2
-66.1
-77.2
-73.2
-55.0
-60.9
-49.2
-49.1
-57.9
-67.9
-58.3
-71.3
-71.2
-76.9
-82.9
-87.9
-73.0
-83.9
-72.3
-83.1
-85.4
-63.8
-68.9
-79.3
Average = -65.5 -64.8
Std. Dev. = 23.6 16.1
23
-------
Table 6. REGRESSION ANALYSIS OF DATA
SITE 1
LSI FC 4- PW
LSI FCO
SITE 2
RAC FC + £W
RAC FCO
LSI FC + PW
LSI FCO
SITE 3
RAC FC + PW
RAC FCO
LSI FC + PW
LSI FCO
Regression Equation
y'
Y1
y1
y1
y1
y1
y1
y1
y1
y'
= 279.8 + .323 (x - 271.1)
= 279.8 + .5588 (x - 154.7)
= 33.9 + .15 (x - 230.4)
= 41.8 + .608 (x - 66.8)
= 57.5 + .216 (x - 230)
= 70 + .9 (x - 66.8)
= 29.9 + .38 (x - 86.6)
= 24 4- .03 (x - 57.6)
= 30.9 + .05 (x - 92.1)
= 31.4 + .47 (x - 60.4)
Standard Error
of the Estimate Se
135.1
114.8
13.9
12.3
24 . 8
14.3
26.5
6.6
13.2
12.6
Correlation
Coefficient R
.381
.378
.952
.969
.929
.980
.260
.040
.098
.320
ro
FC = Filter catch
PW = Probe washings
FCO = Filter catch only
-------
'500-
Figure 2. FILTER CATCH ONLY AS REFERENCE
LSI INSTRUMENT
SITE 1
0
100 200 '300 400
REFERENCE READING M9/M3
500
25
-------
Figure 3. FILTER CATCH PLUS PROBE WASHING
AS REFERENCE
LSI INSTRUMENT
SITE 1
o
100 BOO 300
REFERENCE READING
400
500.
26
-------
Figure 4. FILTER CATCH ONLY AS REFERENCE
LSI INSTRUMENT
SITE 2
210 —
0
30
60
90
120
REFERENCE READING M9/M
ISO
3
210
27
-------
Figure 5. FILTER CATCH PLUS PROBE WASHING
AS REFERENCE
LSI INSTRUMENT
SITE 2
TOO
600
500
400
O
2!
Q
<
LJ
a:
300
800
.100
: 100 200 300 100 500
REFERENCE READING M9/M3
600
TOO
28
-------
"210 -
Figure 6. FILTER CATCH AS REFERENCE
RAC INSTRUMENT
SITE 2
0
30
fcO
120 »50 ISO 2IO
REFERENCE READING M9/M:
29
-------
Figure 7. FILTER CATCH PLUS PROBE WASHING
AS REFERENCE
RAC INSTRUMENT
SITE 2
0
100 200 300 400
REFERENCE READING
"500
,3
600
700
30
-------
Figure 8. FILTER CATCH ONLY AS REFERENCE
LSI INSTRUMENT
SITE 3
0
IO 80 30 40 5O 60
REFERENCE READING M9/M3
TO
31
-------
Figure 9. FILTER CATCH PLUS PROBE WASHING
AS REFERENCE
LSI INSTRUMENT
SITE 3
RE&RESSIONUNi
_J
'20
60
80
100
teo
REFERENCE READING-
32
-------
Figure 10. FILTER CATCH ONLY AS REFERENCE
RAC INSTRUMENT
SITE 3
60
«*» 50
Q
LjJ
o:
o 30
a:
10
REGRESSION/LINE
10 20 30 40 50
REFERENCE READING M9/M'
60
70
33
-------
Figure 11. FILTER CATCH PLUS PROBE WASHING
AS REFERENCE
RAC INSTRUMENT
SITE 3
140
120
CD
21 80
Q
UJ
O 60
<
cc
I
2O 40 60 80 WO
REFERENCE READING M9/M3
120
34
-------
Table 7. MATERIAL MASSES DEPOSITED IN THE PROBE SYSTEM
Site
1
2
3
Date
1/22/74
1/28/74
1/29/74
3/12/74
3/18/74
LI/21/74
12/11/74
Days from
Previous
Washing
7
6
1
8
6
7
7
Mass material collected grams
Lear Siegler
10.70
5.27
5.98
28.17
25.52
2.67
0.54
Research Appliance Co.
-
-
-
-
-
.54
.48
35
-------
SECTION 10
DISCUSSION
In the analysis of the performance of the particle measuring
devices, the performance criteria used were relative accuracy,,
calibration error, instantaneous drift and long term drift.
Relative accuracy is defined as the difference between the mea-
surement system and the reference system outputs expressed-as a
percentage of the reference value. Calibration error is defined
as the difference between the measurement system reading and the
exact concentration of material on a calibration filter, express-
ed as a percentage of the measurement system reading. Instant-
aneous drift is defined as the drift in the system between
successive readings taken with no time lag between readings. It
is expressed in terms of 95% confidence intervals in full scale
deflection. Long term drift is defined as the drift in the
system between readings of the system taken 24 hours apart. It
is expressed in terms of 95% confidence intervals of full scale
deflection.
The data of Table 6 clearly shows that the standard error of the
estimate for site 1 is sufficiently high to exclude data from
this site from further consideration. The data shown in Figures
2 and 3 clearly demonstrate the lack of significant relationship
between the particulate emission rates measured by the Lear
Siegler and manual reference method.
The flowrate control valve in the sampling system on the beta
gauge monitor at site 1 was found to be malfunctioning due to
contamination. Higher volumes of air were sampled than was
recorded or, in effect, the calibration for flowrate was off.
The result was higher mass concentrations with the beta gauge
monitor compared to the reference method.
For sites 2 and 3 it will be seen that the standard error of the
estimate is not too unrealistic, except perhaps for the values
of 24.8 and 26.5. However, the correlation coefficient for site
3 data is unacceptable. The correlation coefficient data for
site 2 is good and shows the improvement in correlation for both
beta gauges if the filter catch only is used as a basis for
comparison.
36
-------
It is to be concluded therefore that the beta instruments pro-
vide valid predictive data for site 2 only and do not provide
accurate data at any site. The improvement in correlation coef-
ficient and standard error of the estimate for the filter-catch-
only data, indicates that the complex mechanism governing deposi-
tion of particulates in the sampling probes is a major factor
determining the accuracy of the instruments.
The data shown in Table 7 clearly shows the extent of the line
loss problem. The lower value for the RAC instrument is due to
the back-purge sequence in the normal operation of the instru-
ment.
Accuracy for sites 2 and 3 using the two reference systems is
given in Table 8. Accuracy for site 1 is not suitable for pre-
sentation because the data was invalid. Table 8 shows that the
best accuracy is given for site 2 when using filter catch only
as the basis for comparison. The data clearly show the impact of
the line loss of particulates on the accuracy of the measurement.
The calibration errors shown in Table 9 demonstrate the inherent
absolute inaccuracy of the beta systems. The RAC unit shows
better calibration accuracy than the LSI system, which is to be
contrasted with the relative accuracy shown in Table 8.
TABLE 8. ERROR IN RELATIVE ACCURACY
EXPRESSED AS ABSOLUTE MEAN VALUE
PLUS 95% CONFIDENCE INTERVAL
Site 2
Site 3
FC + PW
FCO
% Error LSI
105.6
70.7
RAC
106.6
104.7
LSI
97.0
88.6
RAC
112.7
81.4
TABLE 9. CALIBRATION ERROR
EXPRESSED AS ABSOLUTE MEAN VALUE
PLUS 95% CONFIDENCE INTERVAL
Site 2
Site 3
Error in %
LSI
256.6
RAC
211.9
LSI
380.5
RAC
74.6
37
-------
Table 10 presents the instantaneous drift of the LSI instrument
at sites 2 and 3. The values are in terms of divisions. Simi-
lar data cannot be obtained for the RAC instrument since there
is no automatic zero check capability.
TABLE 10. INSTANTANEOUS DRIFT
EXPRESSED AS ABSOLUTE MEAN VALUE
PLUS 95% CONFIDENCE INTERVAL,
BOTH FOR DIVISIONS AND IN TERMS OF FULL SCALE READINGS
Site 2 Site 3
Drift Divs.
% of Full Scale
5.9
15
2.72
7.5
The 24 hour drift data was combined for sites 2 and 3 because of
the small amount available. Expressed as absolute mean value
plus 95% confidence interval, the 24 hr. drift is 5.8 divisions
or in percentage of full scale reading it is 17.4%.
Instantaneous and long term drift data for the LSI instrument
shows a maximum error of 17.4% of full scale. The absolute
magnitude of this error depends upon the calibrated full scales
range of the instruments.
38
-------
SECTION 11
REFERENCES
1. Federal Register, Vol. 36 #247 Part II,
December 23, 1971.
39
-------
APPENDICES
A. Instrument Design and Operation -
Argos 1
B. Instrument Design and Operation -
Model 2414
C. Description of the Particulate Ref-
erence Method Using the EPA Sampling
Train
D. Statistical Analysis
40
-------
APPENDIX A
INSTRUMENT DESIGN AND OPERATION
ARGOS 1, TRANSPORTABLE MODEL, LEAR SIEGLER INC.
GENERAL DESCRIPTION
The instrument is shown diagramatically in Figure 1 to comprise
four main component modules.
Module 1 is the sampling system and comprises a heated stainless
steel probe and nozzle. The nozzle was located on a single point
while each sample was withdrawn. A motorized ball valve was
located in the probe to isolate the instrument from the stack
during periods when sampling was not being performed.
A motorized probe designed to sweep across the stack to obtain a
representative sample of the entire cross-section of the stack
is available for use with the Argos 1, but was not used in the
tests reported here.
Module 2 is the Radiometric Measuring Unit and contains the
fiberglass filter tape transport, particle filtering stage, beta
source and detector systems.. The filter tape transport system
moves the tape in a precise manner using a metering roll so that
sequentially the same area of tape is presented—
(a) to the counting stage where transmitted beta particles
are counted over a one minute time period
(b) to the filtering stage where a pre-selected volume of
gas is passed through the filter
(c) to the counting stage where transmitted beta particles
are counted over a one minute time period
An adjacent clean section of filter tape is then subjected to the
same sequence.
The upper fitting of the filter holder is mounted rigidly to the
sample transport piping. The lower fitting is moveable by a
motor operated cam and connects to the remainder of the sample
piping by way of a teflon bellows. The lower portion of the
filter housing is pressed up against the filter to provide an air
tight seal during a sample cycle and is separated from the filter
41
-------
Module 1
M
Gn
D I.J
3i/
Module 2
H
D:
Module 3
LJ U
^\ I Jfi
Module 4
Figure A-l. Argos 1 Mass Analyser
42
-------
and upper portion of the housing during tape movement„ The tape
proceeds sample by sample from the storage reel to the takeup
reel. The instrument is equipped with an optical tape fault
control, that automatically turns the instrument off if the tape
tears, or the roll end is reached. This prevents the instrument
from taking in unfiltered gas which may cause dust deposition in
the transport tubes, orifice and pump. To prevent condensation
within the instrument, heating jackets are used over th© length
of the sample transport pipe,, The temperature of the gas is
measured with a sensor located just down stream from the filter
holder and displayed on a temperature indicator. The tempera-
ture is controlled by a thermostat which controls the current to
the heating jacket elements„
Module 3 is the Gas Measurement Module and contains a venturi
orifice meter and differential pressure regulator that controls
the flow through the venturi by activating a motorized vacuum-
pump bypass valve. A vacuum gauge and temperature sensor are
provided together with the venturi calibration curve, to permit
calculation of the sample flow rate,, In use, the limit switches
on the differential pressure regulator are set above and below a
venturi pressure drop value calculated to provide near isokinetic
sampling under average conditions. Flow rates of up to 10 M^/hr
are utilized in sampling, and are obtained using either a rotat-
ing vane pump or Roots blower. While the sturdy Roots blower is
more suitable for permanent operation, it is suggested by Lear
Siegler Inc. that the lower weight and easier maintenance rotat-
ing vane pump is more suitable for short-time measurements with
the transportable system used in this study. Experience however
showed the necessity of using the Roots blower under the test
conditions.
Module 4 comprises the Computer and Control Unit which controls
all the mechanical functions of the instrument, computes the
mass concentration as given in Equation 4 and provides an output
signal suitable for recording. The entire mechanical sequencing
is controlled by a cam timer; the period of sampling is controll-
ed by an adjustable clock timer.
Computation of the ratio of the beta counts is performed by
moving voltage potentiometers according to the number of beta
impulses received by the G. Mo tube. Two sealers are used to
suitably match the high rate beta pulses to the low rate poten-
tiometer stepping motors. The ratio of the potentiometer outputs
at the conclusion of counting is displayed on a strip chart
recorder. The recorder output is manually converted by a chart
to mass concentration emission rate.
The control module contains a function switch to permit either a
zero check in which the tape is transported and counted without
collection of a filtered deposit, or an upscale reference check
in which a deposit is simulated by interposing a partially
43
-------
obscuring plastic plate between the source and detector during
the "count sample" event in the sequence.
In this instrument the volume sampled is determined as the pro-
duct of sampling time and volume flow rate. The sampling time
is determined from a timer while the volume flow rate is deter-
mined using the orifice meter and differential manometer system.
Since volumes measured by venturi meters are dependent upon the
temperature and pressure of the gas at the venturi, it is necess-
ary to select the instrument operating parameters to provide a
measure of volume flow. Further since both gas temperature and
pressure may change during each sampling cycle, it is necessary
to establish average parameter values by a reiterative experimen-
tal program to establish a reasonable measure of volume flow rate.
Pre-operation Set-up
The instrument was assembled as shown in Figure 1. The section
of the probe outside the stack was wrapped with a heating tape
and insulated. The sample tubes within the instrument were main-
tained at a temperature of between 150 and 170 C. The measuring
range of the instrument is designed to be in the range 0-50 mg
collected sample. The sample flow rate and total sampling'time
must be selected prior to.normal operation to give the appropri-
ate concentration range. Figures 2 and 3, relating nozzle size
to sampling rate and the venturi flow meter calibration curve are
used in the following reiterative manner to establish the required
sampling conditions.
1. The maximum anticipated stack gas velocity (V max), static
pressure (Ps), temperature (Ts), and dust loading were estab-
lished by preliminary measurement.
2. Figure 2 was used to select the correct nozzle size ba.sed
upon V max, to give an actual volume flow rate of between 5 and
9 M3/hr. The flow rate was determined using Figure 2. Lear
Siegler Inc. recommended that this indicated flow rate be in-
creased by 10% to 20% to ensure that the gas sample is always
withdrawn from the stack under super isokinetic conditions.
3. The indicated actual sample flow rate was corrected to
standard conditions.
4. The desired sample concentration upper limit was estab-
lished and used to calculate the total sample volume required to
equate this concentration to the 50 mg upper design detection
limit of the instrument.
5. The desired sampling time was established using the
volume flow rate determined in step 2 and the total volume estab-
lished in step 4 above.
44
-------
Figure A-2
RELATIONSHIP BETWEEN NOZZLE SIZE & SAMPLING RATE
. FT/SEC
J3I.2 -I
120
no H
IOO
90 ~
80 -,
3 *>
LJ
> 50
30
20-
:0-
5 -
Q
M/SEC
40—1
lOcnm
20—I
10
0
12mm
I I I I I I I I
l ' I ' I
2345&789IOIII2
-nM3/HR
i r i i i i i
0 I £ 3 4 5 6
SAMPLING RATE
rFT3/MIN
7.06
45
-------
APfmmWATER)
300 -
2OO
IOO -
PI
O.lkp cm2
0.2 "
0.3 "
0.4 "
0.5 "
£(t=50°C,
0.905
0.855
0.802
0.746
0.684
C(t>*i'C)
0.842
0.796
0.747
0.694
0.637
cfctso'C)
0.791
0.747
0.701
0.652
0.598
I ATMOSPHERE= I.G33Kp/CmE
c =
(L033-p,)293
1.033 fe? 3 4-T)
o
7
r I ' I ' I r I
•3 ; o : i 12
r^r^n r T r^ f^r ^ i ' i • i T i T i •' i ' i
O ,589 f.766 2.943 4.120 5.297 6.475
1.177 2.354 3.531 4.7O9 5.886 7.06
(MS/HR) AT 20°c AND i ATMOSPHERE
Figure A-3. Venturi flow meter calibration curve & correction factors
-------
6. The sample volume was calculated under estimated average
venturi temperature and pressure conditions; Figure 3 was used
to establish the required venturi pressure drop. The limit
switches of the differential pressure controller were set a
+ 20 mm from the established pressure drop.
7. A sampling sequence was performed and the initial and
final gas temperatures and pressures were noted. The averaged
values were used to re-calculate the required venturi pressure
drop.
8. This calculation process was repeated until the pressure
and temperatures at the venturi used in the calculations agreed
with those obtained in the actual sampling cycle.
Output Format
The theoretical output is logarithmically related to particulate
concentration. However the instrument output signal is not
logarithmically transposed by the computer, making a manual
transposition necessary. This is performed using Table 1 which
relates instrument output to percentage of measuring range.
Thus the concentration at a particular output is calculated as
C = XM eqn. 5
where C is dust concentrations in mg/M^ at
1 atmosphere and 20°C
X is percentage of measuring range obtained from
Table 1
M is measuring range in mg/M^
Operation
In operation the instrument is fully automatic performing the
following sequence of operations.
1. lower filter stage bellows, advance tape forward to
locate, clean area between detector and source, close bellows.
2. for a period of 1 minute count, scale and record on one
potentiometer the number of beta particles received by the
detector. Nominal count rates in the range 15,000-20,000 cpm
are used.
3. lower bellows, move tape back to locate same filter area
in the filtration stage, close bellows.
47
-------
•TABLE A-l
% OF MEASURING RANGE FOR INDIVIDUAL RECORDER OUTPUTS
•(LOGARITHMIC CONVERSION)
Recorder Recorder
Output % of Measuring Range Output % of Measuring Range
10 0.0
11 0.5
12 1.1
13 1.6
14 2.2
15 2.7
16 3.3
17 3.9
18 4.4
19 5.0
20 5.6
21 6.2
22 6.8
23 7.4
24 8.0
25 8.6
26 9.2
27 9.8
28 10.5
29 11.1
30 11.8
31 12.4
32 13.1
33 13.8
34 14.4
35 15.1
36 15.8
37 16.5
38 17.2
39 17.9
40 18.7
41 19.4
42 20.2
43 20.9
44 21.7
45 22.5
46 23.3
47 24.0
48 24.9
49 25.7
50 26.5
51 27.4
52 28.2
53 29.1
54 30.0
55 30.9
48
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
31.8
32.7
33.6
34.6
35.6
36.5
37.5
38.6
39.6
40.7
41.7
42.8
43.9
45.1
46.2
47.4
48.6
49.9
51.1
52.4
53.7
55.1
56.5
57.9
59.3
60.8
62.3
63.9
65.5
67.2
68.9
70.6
72.4
74.3
76.2
78.2
80.3
82.4
84.6
86.9
89.3
91.8
94.4
97.1
100.0
-------
4. open sample probe ball valve.
5. stop cam timer, operate pump and clock timer simultane-
ously. The differential pressure controller will operate the
pump by-pass valve to maintain the venturi pressure within the
set limits of +_ 20 mm about desired and preset value.
6. the clock timer stops the pump and re-starts the cam
timer.
7. the pump by-pass valve is returned to the fully open end-
stop position, the sample probe valve is closed, the bellows are
lowered and the tape advanced to bring the same filter area be-
tween the source and detector, the bellows are closed.
8. for a period of 1 minute, count, scale and record on the
second potentiometer the beta particles received by the detector.
9. at the completion of counting, the ratio of the outputs
of the two potentiometers is displayed on a strip chart recorder
for a period of approximately 10 seconds.
10. the two potentiometers are re-set to zero and the sequence
restarted as 1 above.
Calibration
Zero Check and Span
A zero check mode may be selected by switch function. In this
mode the complete cycle of operation is performed except that the
pump operation step is omitted, thus providing count data to the
two potentiometers from the same area of clean filter. In the
zero calibrate mode the output should be 10% of full scale. A
maximum variation of +_ 4% of full scale is allowed for the mean
of ten sequential measurements provided that each individual
measurement lies within + 670 full scale of the mean value.
In the span check mode the complete cycle of operation is per-
formed except that the pump operation step is omitted, and for
the second beta count an absorber is physically interposed into
a reproducible position between the source and detector. The
absorber simulates a filter deposit and provides an arbitrary
up-scale reading. The span check is used only as a check on
operation and not as an upper range or linearity calibrate check.
The instrument should always indicate the same value for the
reference calibrate output with a maximum allowable variation of
+ 470 of full scale for the mean of ten sequential measurements.
Each individual measurement should lie within +_ 6% of full scale
of the mean value.
49
-------
Gravimetric Calibration
Prime calibration is performed by comparison of beta and gravi-
metric measurements of collected filter deposits. Lengths of
filter tape, cut to fit the calibrate filter holder, are dried
and weighed. Deposits are collected on these filters by manually
following the instrument sequence. The collected deposits and
filters are re-dried and weighed. The weight gain measured
gravimetrically is compared with that indicated by the instrument
to provide a calibration of the beta system. It is stated by
Lear Siegler Inc. that to obtain a calibration of the whole sys-
tem, standard particulate sampling methods must be employed such
as EPA Test Method 5, ASAE Performance Test Code 27 or ASTA
D-2928-71. Such comparisons will allow the percent error intro-
duced by moisture in the gas stream, sample collection, deposi-
tion in the probe and gas volume measurement inaccuracies to be
evaluated with respect to the various reference techniques.
50
-------
APPENDIX B
INSTRUMENT DESIGN AND OPERATION
MODEL 2414, AUTOMATIC STACK MONITOR
RESEARCH APPLIANCE CO.
GENERAL DESCRIPTION
The monitor system comprises nine major subsystems: (1) probe,
(2) boundary layer diluter, (3) sample line, (4) sampling module,
(5) heat exchanger, (6) dehydration module, (7) metering module,
(8) control console, and (9) computer.
Probe
The probe is permanently installed in the stack location deter-
mined as optimum for representative sampling of the specific
stack.
Boundary Layer Diluter
The boundary layer diluter is attached to the probe when a stack
involves high levels of temperature, humidity, or particulate
concentrations. This optional component conditions the effluent
sample stream with dry instrument air (-40°F dew point) before it
enters the sampling module.
Sample Line
This line connects between the Boundary Layer Diluter (if used)
or probe and the sample inlet motorized ball valve in the samp-
ing module. It is composed of stainless steel tubing with elec-
trical heating element adjacent, surrounded with insulation and
a weather jacket.
Sampling Module
This sealed, pressurized, stack-mounted unit includes a precision
sampling nozzle (heated to 250°F), punched filter paper tape,
beta radiation gauge, tape drive and photoelectric indexing
mechanism, and storage provision for replacement tapes. It also
has an integral heater that maintains temperature of 120°C in
effluent sample inlet tube and nozzle to prevent condensation of
entrained moisture. The sampling module case is pressurized
51
-------
continuously with dry instrument air to maintain a controlled
environment and assure optimum sampling accuracy.
The sampling nozzle shrouds and clamps the filter tape to produce
uniform sample spots 1" in diameter. Each 100' length of tape
will hold approx 600 individual samples. Replacement tapes are
stored in the sampling case for proper conditioning prior to use.
Heat Exchanger
This unit, an aircooled fin-type heat exchanger, reduces tempera-
ture of effluent sample to approx 38°C or less - and partially
condenses entrained moisture before the sample stream enters
dehydration module.
Dehydration Module
This stack-mounted component, a refrigerated condenser/dryer,
conditions the effluent sample stream to a 2°C dew point before
it enters the metering module.
Metering Module
This subsystem contains 2 dry gas meters, 3 adjustable flow-
measuring rotameters, 4 flow-control valves, a 14M^/m free-flow
vacuum pump, and a dryer for the instrument air used within the
system.
One dry gas meter measures the volume of effluent sample that
passes through system; the other measures volume of instrument
air delivered to boundary layer diluter. The rotameters measure
flow rates of (1) effluent sample stream, (2) dry air to diluter,
and (3) dry air to purge or back-flush probe and diluter assembly,
Flow rates are set by manually operated valves.
System design includes provision for periodic, automatic purging/
back-flushing of effluent sample line in addition to probe assem-
bly.
Control Console and Computer
The computer console controls all system sequences as well as
individual component functions are regulated automatically by
these two modules. The control console receives and assimilates
input signals from the other sub-systems, then feeds them to the
mini-computer for storage and subsequent calculation and print-
out. Visual readouts are provided on master control's front
panel for various relevant operating parameters, including indi-
cation of sequential phases. Instrumentation includes an elec-
tronic timer that controls timing intervals for beta gauging, a
digital beta impulse counter, and an add-subtract volume counter.
52
-------
The add-subtract volume counter adds sample volume flow, on a
cumulative basis, and automatically subtracts from it the volume
of dilution air introduced into the sample stream.
The mini-computer is operated by magnetic punched-card program,
Pre-Operational Set-Up
The total sample volume is determined as the difference in
readings of the two gas meters. One measures dilution air while
the second measures total sample plus dilution air under dry
conditions. The average gas meter temperatures and stack gas
temperature, humidity and velocity must be measured prior to
using the equipment for dust loading measurements. The nozzle
size must be selected together with the sample volume flow rate
to provide near isokinetic sampling under most stack gas condi-
tions. Since the filter impedance will increase as a deposit
is accumulated, the sample flow rate will decrease necessitating
selection of an average sampling rate. The dilutor air flow
rate should be adjusted to approximately 50% of this average
sample flow rate and the sample time selected to give a beta
count of the filter plus deposit approximately 50% of that on
the clean filter.
The computer must be programmed using the magnetic program cards
provided. The following six factors must be manually loaded
into the program.
1. set decimal point to 2
2. set the constant to provide the output dust loading in
the required units.
3. set sampling time
4. set counting time
5. set the time delay between sampling cycles
6. set the alarm level
Output Format
The computer system provides all time, function and calculation
capabilities. The output is printed on tape as the following
sequence of numbers for each sample
Number of beta particles counted on clean filter
Volume of gas sampled, cubic feet
Number of beta particles counted on filter with deposit
Calculated particulate emission rate mg/M^
A preset alarm level may be included which, if exceeded, causes
the printed emission rate to be preceeded and followed by a row
of dots.
53
-------
Operation
The entire operation of the monitor is controlled by the program
sequencer.
In step one the sequence programmer sends a signal to the tape
drive right motor. The motor is controlled by a triac. As soon
as the index hole on the filter tape passes away from the opti-
cal index reader the programmer is placed in a stop mode. The
programmer will remain in step 1 until the next index hole
passes into the optical index reader.
At this point the programmer will immediately jump to step two.
In step two the Beta Count Timer is triggered on. The Beta
Count Timer starts the Beta Counter and also places the sequence
programmer in the stop mode for the duration of time set on the
Beta Count Timer. At the end of this counting time the sequence
programmer will jump to step three thereby stopping further beta
counting.
Steps 3, 4, 5, 6, 7 are data enter steps. During these steps;
the beta count on the Beta Counter is entered into the computer.
Step 8 resets the Beta Counter to zero and resumes the calcula-
tion in the computer.
Step 9 is a drive tape left command, a reset command for the
sample volume counter, and a command to activate the tape out:
alarm system. If in this step the tape does not index properly
(the index hole on the filter tape does not move out of the
optical index reader) the control console will go into "tape out
alarm". In this mode a sharp, high pitch sound will be heard,
the "Tape Out" lamp will be lighted. To reset the system, place
computer on "stand by", place tape load switch, in Sampling
Module, in the load position (this lifts nozzle) and load fresh
tape (or repair break in the tape) made sure tape hole is und.er
optical index reader. After tape is loaded, place tape load
switch in normal position and move computer switch from stand by
to "on" position, push "Tape Out" indicator in and light will go
out. Depress start switch on computer, sampling sequence will
now automatically begin.
Step 10 is the sampling mode for the instrument and the sampling
pump should be one. Sampling will occur for the preselected
time. At this point the programmer will jump to step 11.
Steps 11 - 18 are a repeat of steps 1-8.
Steps 19 - 20 are not used on a single stack monitor. On a two
stack monitor step 19 is used to switch the system from one
stack to the other.
54
-------
Calibration
Prime calibration is performed by comparison of beta and gravi-
metric measurements of collected filter deposits. Lengths of
filter tape are dried and weighed. Deposits are collected on
these filters by manually following the instrument sequence.
The collected deposits and filters are re-dried and weighed.
The weight gain measured gravimetrically is compared with that
indicated by the instrument to provide a calibration of the
system.
55
-------
APPENDIX C
DESCRIPTION OF THE PARTICIPATE REFERENCE METHOD
USING THE EPA SAMPLING TRAIN
GENERAL PRINCIPLE OF OPERATION
The instrument is designed to permit isokinetic extraction of a
stack-gas sample, to filter the particulate material, to condense
and remove the moisture and to measure the total gas flow. The
instrumentation and its use is described in detail in Reference
1.
Instrument Design and Operation
The instrument is shown schematically in Figure 1-2 to comprise
the following modules.
1. Probe assembly including the sample nozzle and heated
probe, S type pitot tube, stack gas temperature measuring system
and the sample probe temperature measuring system.
2. Particle catchment comprising a cyclone and fiber glass
filter system in an oven capable of being heated above the stack
gas dew point.
3. Impinger train comprising a modified Greenberg-Smith
impinger containing water, a Greenberg-Smith impinger containing
water, an empty modified Greenberg-Smith and one containing
silica gel. All impingers are immersed in an ice bath.
4. Control box connected to the sampling box by flexible
connections. The control box contains a dry gas meter, orifice
meter, pump, manometers for monitoring pitot tube and orifice
meter differential pressures, and general control functions.
In operation particulate samples are withdrawn isokinetically
from the stack at preselected points. Isokinetic sampling velo-
city is obtained by matching the stack gas velocity as measured
by pitot tube and the volume flow rate as measured using the
orifice meter. Suitable adjustments are made for stack gas
humidity and composition, nozzle area and stack gas and orifice
temperature and pressure. These procedures are performed using
a standard nomograph relating operational parameters to the
pressure drop needed to be developed across the orifice meter
for isokinetic sampling.
Samples were collected in the heated cyclone and on the fiber
glass filter for gravimetric measurement. Material collected
on the probe walls was removed by washing and brushing and was
collected and weighed. Moisture condensed in the impingers was
56
-------
COARSE VALVE
AIR-TIGHT
PUMP
CHECK VALVE
VACUUM LINE
PARTICULATE SAMPUN.OJBAIN.>-^
Figure A-4 . Particulate sampling train
57
-------
determined by volumetric measurement, while moisture absorbed
in the silica gel was determined gravimetrically.
A typical data sheet and summary of calculations are shown in
Tables 1 and 2.
Calibration
The pitot tube correction factor was measured by comparison with
a standard pitot tube in a laminar air flow stream. Typical
data are shown in Table 3.
The dry gas meter was calibrated by comparison with a wet test
meter. Typical data are shown in Table 4.
Orifice meter calibration was performed internally using the
dry gas meter in the control box to determine the dependence of
flow rate upon pressure differential across the orifice. The
pressure differential corresponding to 0.75 scfm was used in
setting the nomograph for isokinetic adjustment.
58
-------
SCHEMATIC OF STACK
TIp
V^/
_|— .1
_L
PARTI£ULAT£
PLANT
DATE
L
ilte=r
V %• I
LOCATION .
OPERATOR.
STACK NO. .
&£&!££ AMBIENT TEMPERATURE -3&
.BAROMETRIC PRESSURE •*?r
S+O
STATIC
PRESSURE •
^/S.
/3_ .
./.."?
,/3
./a
^/3
^/3
./•J
^/?
/"?
./3
,/3i^
STACK
TEMPERATURE
66O
tffO
6>#O
&&O
6X0
620
f &f)
S &f)
£,#0
^.ffO
t#o
&8O
AVERAGE (e%O
VELOCITT
• HEAD
UP,) Iv/SP^l
<93
^3
/>3
TT.JO/
.2 f^ ?^/
GAS SAMPLE T
AT DRV G>
INLET
(Tm.J.'F
• 4/£>
*/3-
1/.3.
t/3
4/4/
EMPERATURE
IS METER
OUTLET •
nr-w».~F
j2^
/27
3ff
£1
3.8
At
Zif~
&//~>
1Q+i~~
XCO
^oo
TEMPERATURE
OP GAS
LEAVING
CONDENSER OR
LAST IMPINGER
3jf
I*/
.13-
^f
-?/
Itf
&r
16, ~*
1C* !
^?7
^if '
.Iff
1
PUMP
VACUUM
In. H,
VELOCIT)
#
E
!
1
1
^73 '''V ! ! X3.«JY
VOLUME OF LIQUID
WATER COLLECTED
FINAL
OtmAL
LIQUID COUeCTE)
(HPINCER
VOLUME
/X9 //£..1
ftc no •£>
* B? /2."5
SILICA OEL
WEIGHT.
Zieffrf
JLtO
8.4-
O^SAT MEASUREMENT
1
2
ITIMEJ
CO,
o,
CO
N,
COMHENTSi
-------
SOURCE; Site 1.
1EST NUMBER; 1/24/74 Run 2
PPL BKXJECT WO;
ENGLISH UNITS CONVERSION METRIC UNITS
(29.92in. 70°F) FRCTOR (760mn 20°C)
Volure of sample at standard conditions,dry basis
instd
Il7.71 °R [ V
I in.Bgl m
p A H
bar + 13.6
= 24.46 cu.ft. x 0.02821 .6901
Volume water vapor in sanple at standard oonditions
V =|0.0474 cu.ft. |V
wstd |_ na J 1C
MDisture content in stack gas
B = Vwstd
» 5.328cu.ft. x 0.02821 .1503 M3
wo V V
instd + wstd
Particle concentration in stack gas on dry basis
c's • [o".01543 grip ^"|
L "iJtmstdJ
- 2.205 10~6 Mn
instd
17.89%
•1247 grains/scf x 2297
^6
17.89%
17.82xlO'
Ibs/scf
Stack gas volume flow rate on dry basis
.8211- r _ p
Qs = 3600 (I-3**)) Vs A| Tstd/s
4 Tstd.Ps~|
[jTgJavg^stdj
JtTgJavg^stdl
(A- 3.021 sq.ft. Vs»13.81fp8)
Process rate or BTU rating Pw
Emission Rate
(1.667 min. Ts
sec
(0.00267 in.Hg.cu.fto vlc + % &** * AH \
ml OR T™ 1576
sq.ft.)
- 92.59 %
92.59
60
ng/M3
aS.lSlxlO4 scfh xO.02821 1641 M3/hr
1.03.7 Ubeyhr. x 0.4536 .4703 kg/hr
Jba/ x 0.4536 kg/
-------
TABLE A-3
EQUIPMENT CALIBRATION DATA
PITOT TUBE
STANDARD TYPE PITOT TUBE S TYPE PITOT TUBE
Assume CF = 1.0
Air temperature = 68°F
P Vel:fps P CF:Calculated
0.33 38.3 0.50 0.8128
0.10 21.1 0.15 0.8176
Av. 0.8152
Calibration date: 12/18/73
TABLE A-4
EQUIPMENT CALIBRATION DATA
DRY GAS METER
Final Initial Difference P Temp Final Initial Differ- H °F
"WG ^F ence
2.100
3.212
4.229
5.491
6.571
7.565
0
2
3
4
5
6
.900
.100
.212
.425
.491
.571
Av.
1.100
1.110
1.017
1.066
1.080
0.994
6.367
0.55
0.50
1.1
1.5
2.0
2.6
77.9
78.0
77.8
77.8
77.5
77.5
21
23
24
25
26
27
.122
.262
.291
.386
.486
.503
20.00
22.122
23.262
24.300
25.386
26.486
Av.
1.122
1.140
1.029
1.086
1.100
1.017
6.494
0.
0.
0.
0.
0.
1.
1 74
1 74
3 74
5 75
75 76
00 76
61
-------
APPENDIX D
Statistical Analysis
The following statistical analysis was performed to evaluate the
data for Site 1. Identical analytical procedures were followed
for Sites 2 and 3.
Let P(t) be the particle loading of the stack at time t at the
orifice expressed in gm/M-* dry basis. Let R(t) be the flow rate
expressed in M-^/sec dry basis. Then P(t)r(t) is the particulate
concentration expressed in gm/sec at the orifice. The manual
instrument measurement at time to is x(t0) expressed in gm/M^.
It is given by
t + 1 hr
0 P(t) r(t) dt
x(to)
=' o
ft + 1 hr
0 r(t) dt
"o
The Argos 1 Beta Gauge takes its sample for only 43 seconds but
takes a similar measurement. We assume that r(t) the emission
rate of the stack, over the one hour span is given by
r(t) =
where ro is a fixed rate and^(t) is a random variable such that
for each t, the mean is 0 and the variance is
The beta gauge takes measurements at four minute intervals. We
denote these by
t + 43 seconds
0 P(t) r(t) dt
Y(t0) ='
'o + 43 seconds
r(t) dt
62
-------
To make this data compatible with that of the manual instrument,
we can write
15
t
1=1 J
15
k=l
t, + 4 min
K P(t) r (t)
dt
t, + 4 min
K r(t) dt
Where t^ is the kth four minute interval during the hour. The
estimate, which was used, for any k of
't, + 4 min
K P(t)r(t)dt
is
t, + 43 seconds
P(t)r(t)dt. Hence, we estimated
the numerator by
15
k=l
t, + 43 seconds
k P(t) r(t) dt
In addition from our assumed model, the expected value for any k
of
t, + 43 seconds
k r(t)dt = 1/83.7(rQ)
Using these we have
t + 1 hr
° P(t) r(t)
Co
t + 1 hr
0 r(t) dt
o
15 ft.
r
, , k=i J fck
15
dt £
= k=l .
15
k=l
t, + 4 min
K P(t) r(t) dt
' t, + 4 min
R r(t) dt
tk
i
+ 43 seconds
P(t) r(t) dt
15(1/15 ro)
63
-------
1
T5
15
k=l
t, + 43 seconds
k P(t) r(t) dt
ro/83.7
1
T5
15
£
k=l
f
t, + 43 seconds
k P(t) r(t) dt
fck
t, + 43 seconds
,.* r(t) dt
ck
j
This latter is the average of the 15 observations from the beta
gauge taken at four minute intervals. Performance criteria used
were relative accuracy, calibration error, instantaneous drift:
and long term drift.
Relative accuracy is defined as the difference between the mea-
surement system and the reference system outputs expressed as a
percentage of the reference value. Calibration error is defined
as the difference between the measurement system reading and the
exact concentration of material on a calibration filter, expressed
as a percentage of the measurement system reading. Instantaneous
drift is defined as the drift in the system between successive
readings taken with no time lag between readings. It is expressed
in terms of 95% confidence intervals in full scale deflection.
Long term drift is defined as the drift in the system between
readings of the system taken 24 hours apart. It is expressed in
terms of 95% confidence intervals of full scale detection.
64
-------
TECHNICAL REPORT DATA
(I'lease read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-073
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
EVALUATION OF PARTICULATE MASS MONITORS
USING THE BETA GAUGE TECHNIQUE
5. REPORT DATE
April 197Q
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
Meryl R. Jackson
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Particle Data Laboratories, Ltd.
Elmhurst, Illinois 60126
10. PROGRAM ELEMENT NO.
1A1010 26AAM64 (FY-73)
11. CONTRACT/GRANT NO.
68-02-1216
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences. Research Laboratory - RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 6/73 - 2/75
14. SPONSORING AGENCY CODE
EPA/600/9
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A field study was conducted to evaluate two commercially available .beta gauge instru-
ments for measuring particulate mass concentrations in stationary source emissions.
Performance of the instruments was compared with a manual method of measurement at a
ferrite rotary-kiln calciner, at a slurry cement kiln with an electrostatic precipi-
tator, and at an oil-fired boiler.
Tests were conducted over a 168-hour period to establish instrument accuracy, calibra-
tion error, drift and system reliability. Descriptions of the instruments, test pro-
grams and test sites are presented together with a detailed summary of the experimental
data.
The accuracy of the beta gauge instruments was strongly dependent upon the sampling
characteristics of the extractive probes. The instruments tested were not capable of
correctly measuring the particulate concentration in the stack, nor of operating con-
tinuously for a 168-hour period. In the case of the cement kiln, the particle concen-
tration measured by the beta gauge instruments correlated well with the concentration
determined from the filter catch portion of the manual method, but not with the fiber
plus probe catch.
Particle deposition in the probe of the beta instruments was as high as 86% (average)
for thp r.pmpnt. plant. pmi«;«;inn«;-
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
cos AT I Field/Group
* Air pollution
* Particles
* Mass
* Measurement
* Beta particles
Evaluation
Field tests
13B
20H
14B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
73
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
65
------- |