EPA-600/2-77-035
February 1977
Environmental Protection Technology Series
'- ^'< * ฃ "^
,ป, , 0s*'
!. > <. , <*,. Jv
4 V -V-* - * v * '"'' *."'>"*'" "* '-* % -. -* '- "''''*' '/ '-' /' *"0 "' '<4H^*-^--V <'4-ซ<* A*^"^ * *4*"- M,, *4 '< t * * ****v, ,*" * ** ** ' */ * ** ** * * ,ซ *.x * s"1-* 4A ^ tl^
^ < ***** 4;"", !jv^ ^ *^*ฐ -*^x , > *<<,<' ฃ V < < * 3 s "" ^ * ^ < x%^ *4
^v^M&t^M***: :v^:>fcv* /: '^5
-------
RESEARCH REPORTING SERIES
Research reports of the Off ice of Research and Development, U.S Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental 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.
-------
FIELD EVALUATION OF AN
AUTOISOKINETIC STACK PARTICULATE SAMPLING SYSTEM
by
Walter S. Smith and Emil W. Stewart
Entropy Environmentalists, Inc.
Research Triangle Park, North Carolina 27709
Contract No. 68-02-2226
Project Officer
Thomas E. Ward
Emissions Measurement 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
ART
.,,,;ENTAL PROTECTION .^;ซr/
4. iป
-------
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
The performance of a prototype autoisokinetic stack par-
ticulate sampling system designed to maintain automatically iso-
kinetic sampling conditions was evaluated in field tests at
stationary sources. Tests were conducted to determine the oper-
ating limits and characteristics of the system. Preliminary
tests demonstrated the necessity for making several modifications
to the existing systems to improve the level of performance for
the field program. Improvements were made in the problem areas
of the performance of the mass flowmeter, flow totalizer, and
flow control valve systems and in the sampling nozzles.
The autoisokinetic sampler was tested at four field instal-
lations selected to provide a wide range of sampling conditions.
The results of the testing and analysis showed that the sampling
system maintained acceptable isokinetic sampling rates of 100%
ฑ 10 at only one of four sources. An inverse relationship be-
tween the percent of the isokinetic rate and the temperature of
the gas stream being sampled was found. The evaluation revealed
that the sampler will operate only in the narrow range of stack
gas static pressure of + 3 inches water column. The physical
hardware was found to be fragile and difficult to operate in the
field. Overall, the autoisokinetic sampling system failed to
meet the design goals.
111
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1
2. Summary of Results 3
3. Conclusion 6
4. Recommendations 8
5. Testing Protocol 9
6. Evaluation Testing 13
Preliminary testing 13
Field evaluation 36
7. Field Test Results 41
8. Analysis of Testing 48
v
-------
FIGURES
Number Page
1 Original test protocol diagram 10
2 Revised test protocol diagram 12
3 Sampling train flow diagram 15
4 Calibrations of flow meter and totalizer as 1?
originally furnished
5 Totalizer calibrations at four flow rates 18
6 Final calibrations, mass flow meter and totalizer 19
7 Typical mass flowmeter calibration (Hastings) 20
8 Pitot tube error 22
9 Performance of 1/4 inch nozzle, as received 23
10 Performance of 3/8 inch nozzle, as received 24
11 Steady-state calibration - 1/4 inch nozzle 25
12 Steady-state calibration - 3/8 inch nozzle 26
13 Performance of unmodified nozzles- sensor only 28
14 Performance of nozzle modification #2- sensor only 29
15 Performance of nozzle modification #3- sensor only 30
16 Performance of nozzle modification #4-sensor only 31
17 Performance of nozzle modification #5-sensor only 32
18 Performance of nozzle modification #6-sensor only 33
19 Performance of nozzle modification #7-sensor only 34
20 Performance of nozzle modification #7-sensor and 35
controller
21 Performance of final 1/4 inch nozzle, pre-field 39
and post-field testing
22 Performance of final 3/8 inch nozzle, pre-field 40
and post-field testing
23 Gas stream velocity versus percent isokinetic 42
24 Gas stream temperature versus percent isokinetic 43
25 Grain loading versus percent isokinetic 44
VI
-------
TABLES
Pa a
Number f-M
1 Summary oฃ Field Test Results 5
2 Results From the EPA Stationary Source
Simulator Facility Runs (Source #1) 45
3 Results From the Neutral Pressure Power Boiler
Stack, (Source #3) 46
4 Results From the High Temperature Wood-Fired
Boiler Stack, (Source #4) 47
VII
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge Thomas E. Ward of the
Environmental Sciences Research Laboratory for his assistance
and guidance on this project. The authors also wish to thank
Liberty Chair Corporation for their cooperation and assistance
in providing a valuable test site.
Vlll
-------
SECTION 1
INTRODUCTION
A prototype particulate sampling system capable of main-
taining isokinetic sampling conditions automatically was de-
signed, manufactured, and tested under EPA contract number
68-02-0546 and described in EPA report number EPA 650/2-74-029.
The purpose of this project was to field test the "autoisokinet-
ic" sampling system and define the operating limits and para-
meters and to make any recommendations toward improving the
sampling systems. The design goals are specified below:
Isokinetic sampling of gas streams having veloc-
ities in the range of 6 to 46 meters per second.
Isokinetic sampling of gas streams having
temperatures of -1ฐC to 535ฐC.
A sampling rate of 15 to 60 standard liters per
minute.
A response time capable of following flow fluc-
tuations of ฑ 10% with a frequency of 30 to 120
cycles per minute.
Automatic control of the sampling rate.
Provide a visual output reading.
Electrical outputs to accommodate continuous
recording of sampling rate.
The results of the evaluation tests are summarized in
Section 2. Section 3 presents the conclusions and demonstrates
the reasoning behind them. The recommendations for improving
the sampling system are documented in Section 4 and can be
separated into those 1) pertaining to the operating level and
2) concerning the handling characteristics of the equipment.
-------
Details on the testing protocol including sources tested
are presented in Section 5. Ten simultaneous tests using an EPA
Method 5 train and the autoisokinetic sampling system were
scheduled at each source.
Section 6 details the actual equipment evaluation testing
procedure and the required modifications. A description of the
testing at each of the four sources is also given. Results of
the tests are presented in Section 7, along with tables and
graphs demonstrating the results. The performance capabilities
and the procedure of isokinetic sampling with this system are
presented in Section 8.
-------
SECTION 2
SUMMARY OF RESULTS
The autoisokinetic sampling equipment did not function
satisfactorily. Modifications to the supplied equipment were
made during the preliminary testing period to improve the level
of performance.
The percent isokinetic sampling rate was determined from
data generated by a dry gas meter and S-type pitot tube which
were added to the original system. The carbon-vane pump of the
original system was replaced with a leakless pump to preserve
the integrity of the dry gas meter readings.
Other modifications were:
(1) The sensitivity of the bias adjustment valve was found to
be insufficient. The valve was replaced with a more
sensitive valve-control system.
(2) The protective shroud surrounding the nozzle and velocity
sensors was removed to eliminate interference with pitot
measurements.
(3) The existing nozzles were found to be defective. New
nozzles were fabricated and used in place of the defective
nozzles.
A summary of results for the field testing program is pre-
sented in Table 1. The twenty-seven test average percent iso-
kinetic for the autoisokinetic sampler was 86.2% with a standard
deviation of ฑ 15.8%. While the particulate concentration re-
sults of the autoisokinetic samples showed an average positive
bias of only 2.65% for the twenty-seven tests, the standard
deviation was + 23.6%.
-------
The Federal Register, Vol. 36, No. 247, December 23, 1971,
Method 5, states that the percent isokinetic of the sampling
rate must be between 90% and 110% for particulate sampling re-
sults to be considered acceptable. Nine out of the ten runs
were within the acceptable limits in source #1 in which the
effluent consisted of flyash redispersed in air. No acceptable
isokinetic values were obtained in tests at source #2 because
high negative stack gas static pressure caused the control valve
to lock full open, thus making the sampling system operate at the
highest sampling rate regardless of the effluent velocity. The
design stack gas static pressure limitations were determined to
be ฑ 3 inches water column. For source #3, the neutral pressure
boiler stack, only one of the seven tests conducted yielded an
acceptable isokinetic sampling rate. The other tests at this
source were below the acceptable limit. The percent isokinetic
was consistently low on all runs in source #4, the high stack
gas temperature wood fired boiler.
-------
CO
H
.-4
i i
i >
CO
w
cฃ
CO
H
CO
W
Q
J
i i
PL,
Pi
O
>"
OS
CO
.
rH
PH
_1
PQ
<
H
W
U Pl
H -H
P T3
CD Ctf
PI 0
H _j
_v>
r-^
0 Pl
(/) *H
H nj
O M
P O
3
to
CD r-~ cr> o
Q ...
r-~ rH O
i 1 to
T3 +1 +1 +1
4-*
CO
+ i
+1
to o to
P! ...
cd *3- to rg
CD rH rH rH
S ' +' +'
>
0)
Q VO LO I I
t^ LO OO
T) -53- CNl rH
P
CO
P! LO
O
,J n l
r~i r^
-t-> 0
cti X
H 4.)
>
CO
t>
H 0
H
0 4-)
H (p
*-ป P!
CD -H
PJ -X
H o
*v; f/\
r^H m
0 -H
W o
rH +->
3
=\ฐ <;
CD
o
t_j
H
3
O
CO
0) O OO OO
Q VO rH
O
<* it 01
T3
P -h +i +i
CO
^j- VO LO
+1 ...
to cxi oo
p| o f- r~ป
Oj rH
CD
s
rH tO *
VO
IO
CXI
+1
LO
vO
.
CXI
+
oo
rH
*3-
+ 1
rH
O
0
rH
OO
LO
rH
+1
CXI
vO
oo
-------
SECTION 3
CONCLUSIONS
The autoisokinetic sampler failed to sample isokinetically
over the tested operating range. Of the twenty seven tests per-
formed using the autoisokinetic sampler, ten tests were within
the acceptable limits for isokinetic sampling. The percent iso-
kinetic ranged from a low of 54% to a high of 112$. By way of
comparison, all twenty seven EPA Method 5 tests, which were con-
ducted simultaneously with the autoisokinetic tests, were within
acceptable isokinetic limits. The auto sampler performed accept-
ably on source #1 with the .277 inch nozzle. The autoisokinetic
sampler did not exhibit the minimum acceptable performance level
for the neutral pressure boiler stack or for the wood fired boil-
er stack.
The results indicate that the percent isokinetic is inverse-
ly related to the temperature of the gas stream sampled. Re-
lationships between the operating percent isokinetic for the auto-
isokinetic sampler and gas stream velocity, temperature and
particulate loading were sought. The particulate loading and
gas stream velocity had no noticeable effect on the percent iso-
kinetic. However, a correlation was observed between the gas
stream temperature and the percent isokinetic. For the sources
tested, the temperature ranged from 74ฐF to 659ฐF and the percent
isokinetic ranged correspondingly from 112% to 61%.
The mass flowmeter and totalizer gas measurement systems of
the sampler are inadequate for source sampling tests. In the
sample gas flow range of 0.5 to 2.0 SCFM the mass flowmeter
accuracy ranges from +6 to +1% of actual flow, and the totalizer
accuracy from +6.5% to +1%. At flows of 0.3 SCFM or less, the
errors of both devices are greater than 10%. Calibration curves
for both devices are non-linear, further complicating corrections
of the output.
The autoisokinetic sampling equipment was designed to oper-
ate in stack gas static gage pressures within the range of ฑ 3
inches of water column. This restricts its use to sampling sites
within this range. Sources with stack gas static pressures out-
side this range are common.
-------
The following general conclusions are based on observations
made during operation of the autoisokinetic sampling equipment
in the field. Arrangement and location of the probe pitot con-
nections and the sensor line connections make rotation of the
probe extremely difficult. Sensor lines attached to the nozzle
are fragile in field sampling situations. The weight of the
probe, sample box, and umbilical card make remote sampling sites
difficult to set up and make sampling traverses difficult and
physically demanding to perform. The stiff umbilical makes
kinking of the sampling lines a major problem. System leak tests
require an open control valve, which causes the following
problems. The bias pressure setting must be adjusted far from
its normal sampling position to hold the control valve open.
The subsequent adjustment back to the proper position for source
sampling is tedious and time consuming. Also, care must be taken
to insure that no silica gel is carried from the fourth impinger
into the control valve because the silica gel can cause the con-
trol valve to malfunction. There is no efficient way to clean
the nozzle because of the sensor line in its center. The most
important failure is reflected in the erratic isokinetic sampling
rates observed over the operating range. Even with the modified
systems, the apparatus is fragile, cumbersome, and complicated
to operate in the field.
-------
SECTION 4
RECOMMENDATIONS
In order to improve the existing system, certain changes
should be made. These changes can be classified as affecting one
of the two general areas of performance or handling characteris-
tics. The following suggestions are made to improve performance:
replace totalizer and mass flowmeter with a more accurate
means of measuring the totaling flow rates
.increase the nozzle's gas flow sensor sleeve diameter
.provide a visual display of instantaneous percent isokinetic
sampling rate
.provide a better method of nozzle-probe connection to reduce
time required for connection and chance of damage while
making the connection
.fix sample box heating compartment
.provide more detailed operating instructions
The following suggestions are made to improve handling char-
acteristics :
move the fluidic converter and controller valve to console
'or other location to give quicker access for adjustment or
repair
.increase sensor and pitot lines length
.replace umbilical cord with a lighter more flexible one
.extend the stainless steel tubing the full length of the
probe
.change manner of sensor line connections to reduce time and
chance of damage
-------
SECTION 5
TESTING PROTOCOL
The original schedule called for an equipment check out and
familiarization period to determine if the euqipment was working
as described by its accompanying manuals and to allow the testing
personnel to learn how to operate the equipment as per instruct-
ions (See Figure 1). From that point, the field evaluation test-
ing was scheduled to begin. A rigorous program of ten runs each
at a variety of sources was proposed. Each test site was chosen
for conditions (i.e., gas stream velocity, temperature, particu-
late loading, entrained moisture, etc.) which would establish the
operating limits and parameters of the autoisokinetic sampling
train. Testing at each site would follow standard EPA Method 5
procedures. The four types of sources selected are given below.
Source Simulator -- Controlled medium loading, ambient
temperatures and moisture content less than 3%, in addition to
controlled gas stream velocities, were characteristics of this
effluent. This was a control situation for the field sampling.
The EPA Source Simulator was selected for these runs. Simultan-
eous runs with a standard EPA Method 5 train accompanied various
runs. Other runs established sampling probe directional param-
eters and other limitations of the equipment.
Asphalt Plant With Water Scrubber -- Typical stack effluent
from this type of source presents a high concentration of water
vapor with possible entrained moisture, low gas stream velocities,
low temperatures, moderate to heavy particulate loading and low
C02 concentration. Particular attention to the functioning of
the static balance probe, pitot tube, and gas totaling device
was scheduled here.
Coal-Fired Power Plant-- Temperatures and stack gas moisture
here were in the middle range, particulate loading heavy, and the
C02 concentration high. This type of operation was assumed to be
a major application for the autoisokinetic sampler.
The sampling probe of the autoisokinetic sampler has an
aerodynamic sensor in the center; therefore, a comparison of the
particulate collection of this sampling probe design to that of
an EPA type train was made. Simultaneous EPA Method 5 samples
were run at this site.
-------
58S
ฃ
o
&
n! in
H
OJ
3g
o in
'S
%*
n) o
in >
o
(U (fl
Cl
-I r-t
l-l ,0
rtg
rt o
H rt
in
o-'d
CD (1)
!"ง
s^
in in
S'-^
i-S
<1) 4->
o a>
^H
p, ซ
S o
II
CO
8
-2
KI
H
O iH
O 3in
rt
P< .
CS
4->
CO
e
ed
!-.
bO
(8
O
U
o
4->
O
^i
PH
C/)
(D
3
00
H
10
-------
Glass Manufacturing Plant-- This source is characterized by
high gas stream temperatures and velocities in combination with
low moisture and particulate loading. Tests here were scheduled
to establish the upper stack gas temperature limits of usefulness
of the automatic sampler.
Upon receipt of the contract and the autoisokinetic equip-
ment and after a close inspection of the equipment, the test pro-
tocol was revised (See Figure 2). The preliminary inspection and
testing programs were expanded to include additional EPA source
simulator air testing of the supplied nozzles, to measure the
shroud interference with the gas stream flow around the nozzle,
and to include the manufacturing and preliminary testing of
nozzles to replace the originally defective ones.
New source sites were also chosen, due to the unanticipated
closure and lack of availability of two of the proposed sites.
The scheduled asphalt plant tests were substituted by a coal-
fired boiler having a high negative static pressure. A high
temperature wood-fired boiler was exhanged for the high temp-
erature glass manufacturing plant. The original coal fired
power plant boiler and the ambient air room exhaust sites re-
mained in the new protocol.
Each series of tests was conducted with an EPA Method 5
particulate sampling train sampling within four to six inches
of the autoisokinetic train sampling point.
11
-------
d>
+_>
Clj
H {/)
SH O
O*'H
f-4 IT,
fX
CL. Ui
03 C
H
c ป->
o to
o>
13 O
4=
u
4~>
c +-*
0) 03
e
ft cij
H i 1
3 3
(U (I)
T3 O
C to
oj
45
t/> to
0 -H
(H r-l
3 43
t/}
S 13
4-J
G> H>
CH E
t I
1 0
03 I 1
s]
I-J N
3 O
O C
X %
M (1)
O C
)S
I/;
.. 4)
+J 4-1
C
0) 13
e c
ft o3
3 to
cr c
be
H
>, u
r~4 Cj
r-l O ft
ซ 1 > (/)
rt c
H (/)
OJ Oj
rH I/I
S -i o
Q ftU
1
H
tM
13
0
f>
^ ' ' *
T '/)
ft- 'r-<
O
0) rH
4_J -rH
H 0
C/j ซ
r-f
03
^-i
0)
z; , .
-^ o
to
d> i/^
-I-J QJ
rH IH
CO ft
j_
QJ
H
Q JH
13 to
O O
H ft
r-H U
CO CT3
* B^
n>
r-Nl >
41. -H
Hi cs
*J Wi
H O
C/3 2^
1 i
!/l I/,
3 C
9 ฃ
w i/)
3
ts O
O
f,
o
c
+J
o
u.
CZ
t:
* 'C *
+-*
,,
6
H
4-J
2
r-H
^
> *->
W SH
v^5 P-
'~" ' f^ ^
t/4 , i
d 13
o c
*-* U-
CD
r_j
O
I
I
|
1
i
S-J i
1 '
K
1^1 l
-------
SECTION 6
EVALUATION TESTING
PRELIMINARY TESTING
Soon after the autoisokinetic equipment was received, tests
were made to verify that the equipment was in working order and
that all of the necessary equipment was at hand. This exercise
revealed limitations of the accompanying manual, EPA Report No.
650/2-74-029, regarding assembly of the system by persons unfa-
miliar with the specific hardware. For example, the functions
of the three two-position toggle switches located on the control
console, between the temperature readout dial and the sample
box heater control, are never mentioned. Color and letter-coding
of pneumatic lines proved quite helpful during setup.
During system assembly, the following points were noted:
Unions from pump-to-mass flowmeter and pump-to-knockout
jar missing.
Glass probe liner broken. (Removal of broken glass from
probe liner proved quite difficult).
Probe heater wires (nichrome) severed in two places, and
exposed in numerous others.
Vacuum line through umbilical kinked at quick-connects
on both ends and where line exits umbilical wrapping.
Coil springs around vacuum line at these points ineffec-
tive due to the weight of umbilical.
These problems were temporarily circumvented to allow im-
mediate shop testing of the system. The broken glass probe liner
was by-passed. This arrangement removed the filter and impingers
from the system, yet allowed testing of the automatic controller,
the most important facet of the equipment.
With the system so assembled and a gas cylinder providing
the pneumatic air supply, power was supplied to the control
console where all instrumentation appeared to be operational.
The fluidic converter was preset for the existing conditions
(3/8" nozzle, and 0.1 to 0.2 inches of H^O pitot readings).
13
-------
A fan supplied the gas stream and the controller was observed
to automatically adjust the sampling flow rate to follow
generally the induced fluctuations in the gas stream.
One criterion for evaluating the autoisokinetic equipment
is to calculate the percent isokinetic of the sampling rate.
To determine the actual percent isokinetic, the gas stream ve-
locity and the volume of gases sampled by the autoisokinetic
train must be ascertained. Therefore a S-type pitot tube and a
dry gas meter were added to the existing autoisokinetic sampling
train in order to gather the needed data.
A leakless self-oiling pump of the type used in a typical
EPA Method 5 particulate sampling train meter box was substituted
for the carbon-vane pump and the dry gas meter was placed in-
line after the leak-free pump. The mass flowmeter was retained
in the system, remaining between the umbilical and the vacuum
pump as shown in Figure 3.
With this modified autoisokinetic sampling train, a one hour
test run was conducted, with an air flow provided by a fan di-
rected at the 3/8" nozzle which produced pitot readings in the
range of 0.5 to 2.5 inches of water. Approximately 15 changes
in the gas flow were induced during the course of the test by
manual manipulation of the fan. The sampling rate indicated by
the Hastings mass flowmeter fluctuated correspondingly, indi-
cating that the system consisting of the null-balance, control
valve, and fluidic converter was responding to the fluctuations
as described in the manual.
During the test setup and throughout the one hour of testing
several potential problem areas were noted: supply air usage
rate, low sensitivity of bias adjustment valve, a discrepancy
between the digital flow totalizer and the mass flowmeter, and
the level of operation appeared to be below acceptable isokinetic
conditions.
Pneumatic supply pressure was supplied via a gas cylinder
at a rate of approximately 0.75 cfm. Discussions with a General
Electric engineer familiar with the system confirmed .75 cfm as
a normal usage rate. Over a period of several months of field
testing, the cost of compressed air cylinders supplying 0.75 cfm
would have become prohibitive, so a compressor was rented to
provide the pneumatic air supply.
The bias adjustment valve sensitivity was too low to make
the required adjustment to reach the bias setting specified by
the instruction manual. This valve was replaced with a more
sensitive valve control system.
14
-------
.- f,
H
-
f
*
LJL
C_
t
u
1
a
c
i
ซ/
<
>
a
c.
1
a.
2:
CL.
i
U
*2
CO 3
CO C
;^ 1 1
1
U.
_J
CL
d
cr
/
U.
cc.
c
ex.
a.
j
> en.
1 LU
- h-
< LU
:s:
sS
- LU
: 2:
\
I
7
I
ac
e
s:
i
X
0
CO
t
1
1
1
1
1
is
1 m 1
1 a:
1 ฃ '
I LU 1
* !
LU ,
Q CO 1
i <^) )
1 =C I
1 h- Q
1 lil | i
I S LU
1 > I
i - LU
g^
a
CL
0.
i 1
ซ
^2=LU
co o p.
A
1 c-i
< C
1 ฐ
1 -H
1 -P
i 2
' 5
1 ,-H
l S
1
1 x
_.
' m 6
. 03
1 o: JH
i it! ^
1 t rt
i ^ ^
1 Is
] S o
1 g ฐ
. ^ MH
1 LU
1 S -
i 5 nj
i w i
|
1 .S
1 M
1 e"
i OT
i
i
t^>
CD
H
15
-------
Calibration checks were made on the digital flow totalizer
and mass flowmeter with the results as shown in Figure 4. The
mass flowmeter was corrected somewhat by adjustment of the cali-
bration potentiometers labeled "meter" and "gain" but the flow-
meter was still out of calibration. The totalizer could be
calibrated for a given constant flow, but when the flow was
decreased, the totalizer read lower than the actual value, and
when the flow was increased over the calibrated flow setting, the
totalizer read higher than the actual value. The totalizer would
not respond to flows less than 0.2 - 0.3 cfm. Calibrations for
flow rates are presented in Figure 5. The mass flowmeter was
taken to a Hastings-Raydist office for a manufacturer's calibra-
tion. The flowmeter was designed to operate at flows up to 5
cfm. Specifications for the flowmeter are ฑ 1% of full scale.
At a flow rate of 0.5 cfm, therefore, an error of 101 is allowed
by the specifications. The Hastings-Raydist engineers calibrated
both the flowmeter and the totalizer. The final calibration ad-
justment still yielded a positive 5% bias for both meters. The
results are given in Figure 6. Figure 7 gives a typical mass
flowmeter calibration curve.
16
-------
JH
O
S-.
tn
W
0
- 2
- 4
- 6
- 8
-10
-12
-14
-16
-18
-20
-22
& Mass Flow Meter
ฉ Digital Flow Totalizer
0.5
1.0
1.5
2.0
2.5
Figure 4.
Flow Rate, scfm
(Timed Dry Gas Meter Readings)
Calibrations of flow meter and totalizer
as originally furnished
17
-------
?H
o
rt
-P
O
ฃ-
+ 12
+ 10
+ 8
+ 6
+ 4
+ 2
- 2
- 4
- 6
- 8
-10
-12
14
-16
cfm
cfin
I
i i i
i _ i _ i i _ i i
0.5
2,0
2.5
1.0 1.5
Flow Rate, scfm
(from timed dry gas meter readings)
Figure 5. Totalizer calibrations at four flow rates
18
-------
rt
w
18
16
14
12
10
-2
-4
-6
-8
-10
-12
ฉ TOTALIZER FINAL CALIBRATION
A FLOW METER FINAL CALIBRATION
1 . I
2 -4 -6 -8 1.0 1.2 1.4 1.6 1.8 3.0 2.2 2.4
Figure 6.
Flow Rate, scfm
(Timed Dry Gas Meter Readings)
Final calibrations, mass flow meter and totalizer
19
-------
CFM
READING
Figure 7. Typical mass flowmeter calibration
20
-------
Because the proximity of the built-in pitot tube to the noz-
zle shroud was 1/4 inch of free space, experiments were con-
ducted to determine the accuracy of the pitot tube. The tests
determined that one inch of free space was the closest that a
pitot tube could be installed without interference by the shroud.
The built-in pitot tube, at best, had an error of -10% at zero
rotation. Figure 8 shows the data plotted as error versus angle
of rotation. Included is typical data from a S-type pitot alone.
Experiments that were designed to demonstrate the perfor-
mance level of the autoisokinetic controls were then conducted
in the EPA Source Simulator Facility (SSF). The nozzle diameters
were measured and the results are presented below:
Nominal Size Actual Nozzle I.D. Nozzle Static I.D. of Area Equivalent
Probe, O.D. to Annular Area
Inches Inches Inches Inches
0.25 0.282 .082 0.269
0.375 0.355 0.128 0.332
To verify that the null-balance method was working, an ex-
periment that removed the fluidic controller and allowed for man-
ually reading a null-balance condition between the sensors was
designed and implemented. In addition, tests were made with the
fluidic controller operating the null-balance system. The re-
sults for each nozzle are presented in Figures 9 and 10. When
compared with the data presented in Figures 11 and 12 which are
taken from the manual accompanying the autoisokinetic equipment,
a definite deterioration in the performance level of the systems
was noted.
With the system operating at the above level, no meaning-
ful field data would be collected; therefore, methods of im-
proving the operating level of the autoisokinetic equipment were
sought.
In order to improve the performance of the system, the noz-
zle protective shroud was removed. Large open threaded holes in
the sides of each nozzle were discovered upon removal of the
shroud. The most inexpensive method to make repairs was to fa-
bricate new nozzles which were constructed using thinner wall
tubing and the static pressure sensors from the original nozzles,
Also, the new nozzle tips were shortened so that the nozzle and
pitot tube were on the same plane. The protective shrouds were
left off.
21
-------
/
Added Pitot Tube
^ 1" free space
Built in Pitot Tube
1/4" free space
Typical Pitot Tube
X readings without probe
-90 -80 -70 -60 -50 -40 -
20 30 40V,50 60 70 80 90
X
X
X
ป ซ X
Figure 8. Pitot tube error
22
-------
120
C!
oo
H
0)
C
H
O
tn
o
ซปp
110
a>
i-H
N
M
O
(3
o 100
(U
90
80
70
60
50
20
O Sensor only (measured by mass
meter)
H Sensor only (measured by Dry Gas
meter)
0
30 40 50 60
Stack Gas Velocity, feet/sec.
70
80
Figure 9. Performance of the 1/4 inch nozzle, as received
23
-------
120
110
100
r-4
N
N
o
03
I/I
90
80
.o 70
+J
S
LLJ
o 60
50
10
0 Sensor only (measured by mass meter)
El Sensor and Controller (measured by
mass Meter)
O Sensor only (measure^ with a dry gas
meter)
A Sensor and Controller (measured with a
dry gas meter)
_L
I
30
20 30 40
Stack Gas Velocity, feet/sec.
50
_L I
60
Figure 10. Performance of the 3/8 inch nozzle as received
24
-------
tS
oi
4J
H
o
o
3
05
-------
1.10
1.05
1.0
.95
.90
J5 .85 -
S .80 _
Adjustable Range
H Sensor only
ฉ Sensor plus Controller
J I L_ L
J L
10 20
30 40
Gas Velocity, feet/sec.
J L
50 60
Figure 12. Steady-state calibration 3/8 inch nozzle (taken from manual)
26
-------
With the new nozzles, the null balance system operation was
rechecked, again by-passing the controller, with the results as
shown in Figure 13. Both nozzles nulled at 90% isokinetic when
operated manually. Past history indicated a drop in performance
when the controller operated the system. Therefore, further noz-
zle modifications were fabricated and tested to achieve an oper-
ating level of at least 110% isokinetic when the controller was
bypassed. Figures 14 through 19 show the sensor modifications
tested and the results therof. Modification #7, consisting of
the pressure holes, was finally selected for the new nozzles.
The results of the ''sensor and controller" tests on the final
modified nozzles are presented in Figure 20.
With the new nozzle configuration and with the previously
mentioned problem areas corrected (new glass probes, rewiring
heater, etc.), the field evaluation program was initiated.
-------
14U
135
130
125
120
115
110
105
100
3 95
^-*
o>
c
ฃ 90
o
i_4
* 85
80
75
70
Unmodified nozzle
ฉ 3/8 inch nominal
R 1/4 inch nominal
o p
-^^""^ Sensor only
- ^ >=
\ V
-
rn ป, ^r-x f *
o d3 o Qj3_ ^o _?
__n ฐ .... yu CT
--0-^ฎ-0 GQ ^ Q E
- G
1 1 1 1 -
15 20 25 30 35 40 45 50 55 60 65 70
Velocity, feet/sec.
Figure 13. Performance of unmodified nozzles
28
-------
145
140
135
130
125
120
115
5/8" O.D. Tubing
Nozzle Modification #2
0 3/8" Nominal Nozzle
Q 1/4" Nominal Nozzle
Sensor only
i
110
105
100
95
90
85
80
75
0
70
I I
I i
20 25 30 35 40 45 50 55 60 65 70 75 80
Velocity, feet/sec.
Figure 14. Performance of nozzle modification #2, sensor only
29
-------
135
130
125
120
115
110
105
100
95
90
0
i i
g
ง 85
U4
8
HH
.*> 80
75
70
3/8" O.D. Tubing Nozzle Modification #3
3/8" Nominal Nozzle
(Due l-o Time fnnfii-Ka i nts
___ J p and Low 3/8" Results,
sf^~^~~ 1 ^ 1/4" Not Tested)
| Sensor Onlv
-4^7^ ^^
\ c
\ \
\ A
-
-
-
_> Q ฎ Q
o __ฎ -
___ -S- ~ "
_^_ - ^r~ ^)
ฉ
-
-
-
1 1 1 1 1 1 1 I 1 1 1
15 20 25 30 35 40 45 50 55 60 65 70
Velocity, feet/sec.
Figure 15. Performance of nozzle modification #3, sensor only
-------
o
H-1
H
140
135
130
125
120
115
110
105
100
95
90
85
80 _
75 _
701
G
3/8" O.D. Tubing
Nozzle Modification 14
3/8" Nominal Nozzle
(Due to time constraints
and low 3/8" results,
1/4" not tested)
Sensor only
O
.O
15 20
25
30 35 40 45 50 55 60 65
Velocity, feet/sec.
Figure 16. Performance of nozzle modification #4, sensor only
31
-------
u
140
135
130
125
120
115
110
105
100
95
90
85
80 _
75 _
70 _
0
Rivet
Nozzle Modification #5
1/4" Nominal Nozzle
(Due to time constraints
and low 1/4" results, 3/8"
not tested)
Sensor only
20 25 30 35 40 45 50 55 60 65 70 75
Velocity, feet/sec.
Figure 17. Performance of nozzle modification #5, sensor only
32
-------
u
160
155
150
145
140
135
130
125
120
115
5? 110
105
100
95 U
90 U
A I | i
Nozzle Modification #6
ฉ 1/4" Nominal
Sensor only
J I I L
15 20 25
30 35 40 45
Velocity, feet/sec.
j L
50 55 60 65 70
Figure 18. Performance of nozzle modification #6, sensor only
33
-------
U
ti
H
140
135
130
125
120
115
110
105
100
95
90
85
80
75
El
Sleeve
Nozzle Modification #7
15 20 25 30 35 40 45 50 55 60 65
Velocity, feet/sec.
Figure 19. Performance of nozzle modification #7, sensor only
70
34
-------
J.4U
135
130
125
120
115
110
105
100
U
i i
g 95
ง 90
ซป>
85
80
75
70
Nozzle Modification #7
- Sensor and Controller
ฉ 3/8" Nominal
D 1/4" Nominal
-
ฉ
ฉ
ฉ
00
ฎ ฉ G
- a
ฉ ฉ ฉQ qjpj m
ฉ 0 ja
ฉ ฉ
s ฉ
-
ฉEl Q
a
i i i i i i i i i i i
Q
15 20 25 30 35 40 45 50 55 60 65 7
Velocity, feet/sec.
Figure 20. Performance of nozzle modification #7, sensor and controller
35
-------
FIELD EVALUATION
SOURCE 1 -- The field evaluation period was started with ten
simultaneous tests made in the EPA Stationary Source Simulator
Facility (SSSF) with moderately heavy particulate loading, med-
ium velocities, and low temperatures. Five tests were made with
the 3/8" nozzle and five with the 1/4" nozzle. The probes and
nozzles of the two sampling trains were positioned at a proximity
of two inches. Test readings were taken at five minutes inter-
vals during the thirty minute runs. The sampling location simu-
lated ideal field conditions: proximity to a well equipped shop,
spacious setup and operating areas, the autoisokinetic sampling
box and probe mounted on a rolling table, complete control of the
source, and room temperature climate conditions. Under these
test conditions, the difficulties of obtaining good leak tests
were the only drawbacks.
SOURCE 2 -- A coal-fired power plant boiler with a large
negative stack gas static pressure was the source for the next
series of tests. At this test, the weight of the autoisokinetic
sampling box and probe required a table and three people to be
properly moved from point to point and from port to port. Two
people handled the table-box-probe unit while the third person
guided and supported the heavy probe to prevent it from sagging
and breaking at the box-probe fastening point.
On each of the four sets of test conducted, the controller
valve locked open as soon as the autoisokinetic probe was intro-
duced into the stack. When the probe was removed from the stack,
the system would appear most of the time to work as designed. It
was determined, after a conference with the designer, that the
autoisokinetic equipment was designed to operate only within
the stack gas static pressure range of + three inches of water
column. Further testing at this source ( - 13 inches of water]
was terminated due to the insignificance of any data generated
with the controller valve continually forced open by the source.
SOURCE 3 -- A neutral pressure coal-fired power boiler was
tested, providing medium temperatures, high velocities, and heavy
particulate loading as testing parameters. For these tests, ex-
tensive site preparation was required including access ladders to
the testing platform and planking for the platform. An identical
setup involving the EPA 5 equipment and the autoisokinetic e-
quipment as required at the previous source was needed here,
including the table support. The following handling problems
36
-------
for the autoisokinetic sampler are mentioned in detail because
no significant problems were encountered with the simultaneous
Method 5 tests. The ambient temperature remained below freezing
and the 15-20 mph winds caused a considerable chill factor. The
handling of the equipment became more awkward and clumsy: the
1/4 inch nozzle was damaged and repaired with new tubing to re-
place the sensor lines. All of the movement that precipitated
the damaging of the nozzle was occasioned by the malfunctioning
of the controller valve which was first assumed to be caused by
the freezing temperatures but was later determined to be caused
by silica gel granules. After the repair of the 1/4 inch nozzle,
the scheduled ten tests were completed. Due to the malfunctioning
and unpredictability of the boiler, the tests were ended at this
source.
During the course of making the seven tests, several recur-
ring problems were encountered: in separate incidents the con-
trol valve stuck open and closed, two of four holes in a sensor
line were plugged by the particulate matter in the effluent, the
heater in the sample box worked erratically and eventually failed
(the wire broke but was repaired in the field), and finally the
sensor lines on the 1/4 inch nozzle broke and were repaired again.
All of these problems created delay while "identify and repair"
operations were made.
The control valve malfunction was determined to be caused
by silica gel granules which were carried through the check valve
into the control valve. This was remedied by removing the check
valve and by placing glass wool in the last impinger to catch any
silica gel that might be forced into the controller during op-
eration of the sampler.
A major problem not previously encountered was in properly
deciding whether to use a 1/4 inch or 3/8 inch nozzle, as the
boiler load changed drastically throughout the testing period.
For example, during the time required to correct the previously
mentioned problems, the boiler loading would deviate from the
preliminary pretest data, with the magnitude of the boiler load-
ing change depending on the weather and on other plant operations.
If the original nozzle size became inappropriate, changing to
the other nozzle required at least an hour,by which time the
boiler operation would normally change again. Also, as the load-
ing changed, the static pressure ranged in value from 0-1.5
inches of water. This change required tedious, time consuming
adjustment of the instrument controls.
SOURCE 4 -- The last source tested was the wood-fired boiler
that provided operating parameters of high temperatures, low
velocities, and light particulate loading. Again, ten simul-
taneous tests were scheduled. The EPA Method 5 nozzle was placed
37
-------
in one port within four inches of the autoisokinetic nozzle which
was positioned through the other port.
Several new problems occurred, one of which involved the
autoisokinetic probe assembly. The one-inch stainless steel
sheath for the glass liner does not extend the length of the lin-
er but is instead butt-welded to the larger two-inch diameter
outside sheath, which also encompasses the pitot tubes. The
butt-weld failed, breaking the glass liner, requiring a new glass
liner and repair by using another butt-weld. Another problem
was caused by the testing site: a fire during the night burned
the compressor air supply line. These were the only time consum-
ing problems encountered and all ten simultaneous tests were
completed.
With the completion of the last field test, the equipment
was again setup at the SSSF in order to compare the post- test
operation with the pre-test operation. The results are shown in
Figures 21 and 22. A slight drop in performance seems to have
occurred during field tests.
38
-------
1 A f\ ,^_^_________^_^_____^___ ____^_____ . ~____ _ _
xiu i "
I Ambient air tests with
135 i- controller operating nominal
; nozzle size, 1/4"
130 r
! ฉ December
125 h 0 March
120 -
115 |_
i
110
105
95
0
90
85
80
75
ฉ
m
70
ฉ
ฐง
100 .
ฉ
ฉ
ฉ ฐ &
ฉ 0 ฉ
m a E
0
B E ra
m
ฉ 0
ฉ
ฉ 15 20 25 30 35 40 45 50 55 60 65 70
Velocity, feet/sec.
Figure 21. Performance of final 1/4 inch nozzle, pre-field and post-
field testing
39
-------
140
~
130
Ambient air tests with
controller operating
nominal nozzle size 3/8"
125 I
120
115
110
105 h
100 h
95 -
85
80
75
70
0
ฉ
ฉ
0
ฉ
a
0
>ฉ
0 ฉ
ฉ
ฎ H Q
D
ฉ
O December
Q March
J I
m
10 15 20 25 30 35 40 45 50 55 60 65
Velocity, feet/sec.
Figure 22. Performance of final 3/8 inch nozzle, pre-field and post-
field testing
40
-------
SECTION 7
FIELD TEST RESULTS
This section presents the results of the field tests. Com-
ments on the design and functionalism of the equipment, and the
preliminary testing are covered in Section 8 "Analysis of Test-
ing System."
Ranges for each parameter tested are given in the chart be-
low, along with the source tested which best exemplified the
particular range of each parameter.
PARAMETER RANGES
Temperature
Velocity
Particulate
Loading
Static
Pressure
Low
Source 1
20 - 250ฐ
Source 4
20 - 43 fps
Source 4
0-0.35 gr/dscf
Source 1
0-1 in. H20
Medium
Sources 3 ง
250 - 500
Source 1
43 - 67
Source 1
0.35 - 0.63
Source 3
1 - 5
High
Source 4
500 ง greater
Source 3
67 + greater
Source 3
0.63 greater
Source 2
5 ง greater
The results from each test in the specified parameter ranges
are plotted versus percent isokinetic in Figures 23,24, and 25.
Note that some sources fit into two different parameter ranges.
Tables 2,3, and 4 give the results of tests at each source
for each nozzle except for the high vacuum source, source #2. No
results are tabulated for source #2 since the control valve was
always locked full open, thereby causing the sampler to sample
at the highest rate regardless of the stack velocity.
41
-------
120
110 -
o
H
fri
M
B
H
tf
O
100 -
90 -
80
70
60 -
50
G
O
0
ฉ
O
O
O
0
J_
_L
O Low
H Medium
A High
0 0
(D
JL
J L
J L
_L
_L
J_
15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95
Velocity, feet/sec.
Figure 23. Gas stream velocity versus percent isokinetic
42
-------
0
H
W
Z
i i
0
tt>
t 1
125
120
115
_
110 -
105
100
95
90
85
80
75
70
65
60
55
50
_
0
O
ฉ
0 0
o Low
Q Medium
A High
0 0ฐ
0 0
-
-
-
-
-
-
-
-
-
70
1 1 1 1 1 1 '
80 90 100 HO '20 130
El
f
H
ED
t i
i i
, 280 790
Q
Q
El a
A
A
A
A
1 i
{,,,{,*
440 450 460470 480 490 500 1, S80 $90 600 ^ 660670
f 1
1 1
Temperature ฐF
Figure 24. Gas stream temperature versus percent isokinetic
43
-------
o
% ISOKINET]
125
120
115
110
105
100
95
eo
35
80
75
70
65
60
O Low
Q Medium
A High
E
E
B
ฉ
a ^
s m
-
ฉ
ฉ
ฉ
- ฐ 0
0
0 A
A
ฉ m
a B A
ฉ
0
i
65
cc
A
I I i i i I 1.1.1 I i
.10 .20 .30 .40 .50 .60 .70 .80 .90 1.0 1.10 1.20
Grain loading, gr/scf
Figure 25. Grain loading versus percent isokinetic
44
-------
_ N
r
LU
o:
Z3
o
CO
CO
o:
I
1 1
_i
>-H
<_)
ซcC
1 1
CsL
O
I
i
et
_J
ZD
2:
CO
LU
CJ
1
o
CO
o:
z.
o
1 1
h-
*
1
CO
^f
Q.
LU
LU
:n
i
i
s:
o
fys
1 1
CO
1
1
ZD
CO
LU
.
CM
LU
I
CO
cu
^ฃ
a.
LU
E
r
fO
i-
h-
r-
cu
c
r
O
in
t t
i
O
4_>
^
OJ
4->
03
15
O
i~* ^S
+-> "
(O O
d) *^
^~ i %
1 !
-^ T3
O C
l/> O
11 <_>
0
4-> LU
o
CO rv
O CJ3
C
ri
O 10
r- C
-!-> 0
C -M
r ปr
-^ x>
0 C
in o
i-i 0
0
r-
4-> U_
(O O
S- CO
I-5 Q
c -^
ai o;
O CD
c
0
o
N
^^
QJ
N
O
"Z.
i CT>
CO CM LO "^" OCOO i O i CTl COCOI
y^ is^ ^. (y, crir^-cvj r^ CM ^j- i < CM
1 1 1 1 1 1 +1 r CMCMCMCMCM-H
1 III II
O LO
LO o vo i r^^j-co o^ co i ซd- Or -h~.
r CM CM CO CM CO LO C"* ^~ VO O^ C3 VO CO
CT) O CTi O OCT^+lCTi CT> CT> CM OCM+I
r r~- r~ r~~
iDOLor~-cr> ^-r^coซ*CM
LO O vo r*^. oo r*^ vo r*^ LO CM
LOLOCMLOLO LOLOIDLOLO
OOOCDO OCDOOO
^f* r**^
csj vo CD co LO oo r*^ ^o r~~ r^*- o^ oo ^3 c\j
OO O CO CD i Oi OO C\J LO CO CTi^OLO
crปo o o OO+ICT>I oo oo+i
r r-* r r~ r r~" r^ r-~ r~" r
.
a> a>
Q 0
O^COCOVO-'VOCMVOVOr
CM VO LO CM COO1T3 r-~^J- LO i i CT)"D
LO ^J" CM ID LO > +* ^J* ^" ^" ^" ^" > 4-*
ซcC CO o
i- CO
cu a
o -a
CO to
t\ T"3
s- c
O !->
co co
45
-------
- -
oo
III
o
~D
O
oo
o
1
oo
ce:
LI i
_i
i i
0
co
o;
LU
3:
o
rt
UJ
Qi
ID
oo
oo
LU
Q-
1
1
1
LU
z.
LU
nr
I
i
^
O
cc
Lu
OO
\
1
OO
LU
ca
*
oo
UJ
_J
ca
ซ=c
1
r^
O
r
-P
C
in
1 (
1
o
4J
<:
LO
-o
o
J=
CD
s:
^^
Q-
LU
C
to
i.
1
o
Ol
_c
^
o
I/)
1 1
1
o
3
CU
M
to
O
+J
to to
D- -r-
CO
C
I
O3
S-
I
f^
ซซ
O in
r- C
C -4J
-S -o
o c:
in o
i i o
C
o
1
[ * 1 1
03 C_>
s- oo
-P 0
cu D:
O C3
Q
o
tfl
O tn
r- C
+-> 0
C 4->
-^ T3
O C
O
I " C_)
c
o
r
P LL.
03 C_>
S- OO
+-> O
CO rv"
O CI3
C
o
CU
CU M
r- CO
O O
CU i
M- N
LU O
~^"
oo ^^* 10 o^ r^^ i-D i i vo i i i f*~^
iocr>i <5a-Lf)vooOLnvDcncM
+ + l++lr r OOl r
+ + + + + +1
LO 1-^
t^ooiotncMCMLOi i CMI
OOOO+lOOOOO+l
ซ* i CM CO OO OO "3-
en ^J- oo o P>. CM rซ.
LO OO OO VO ซd- LO ซJD
o o o o o o o*
r*~ t"Vi. C\J OO ^D OO C^ Lf^ ^~" ^J* C?
VปO C3 i i O^ CD CsJ OO t^j **i" r*""* O^
i-.r>>crir^rj-i^r^r^Loio+i
. t
"^ ^
CU O>
Q O
.
oo r-. CM o r o r- .
v^ oo oo 01 T3 r*** LO vo en en *o
> 4J . . . > 4J
ooo<:oooooo 00
CM O
r~ i
+i
l^-
q_
* 0
C
01 c 0
> O -r-
cf. <- 4->
-t-> O3
103
1X5 > Q-
S- CO O
CO O d.
O TD CO
i. r
CO 03 Q.
O T3 E
i- C 03
3 03 00
O +*
OO OO -X
46
-------
, .
h-
CU &ง
CU C
+J -I- ซ
ro Jxi C
O -r-
3 t/1 itj
O i < S-
- 1 h-
+-> 0
S- !->
n3 3
f^ ^t
co co ซ^ CM **D co cn CM co en
..........
=3-LT>cnCMCOCr>VQCOOOOO
1 IX5CS1COCM + CM + COCM
+ + + + + II
^o
IT)
T3
0
-C
CU
21
ej;
D-
Ul
o^~
ซt
O t/>
r- C
-P O
CU >-
C -P
I !
j<: "a
0 C
o
HI 0
oo<-oซd"i ^cMcnvocnioco
cnr-~ooencMi CMOCO
encnooenoe^oocn
i i i i r r
E
O
i
+-> U-
rtJ O
s- co
P Q
E -^
cu D;
O CD
E
O
O
ocMcncoiocOisJ-i ."vf
r^, i . o en vo co r^^ en CM o
i i r OOOOOCMr-
..........
i OOOOOOOOO
c
1
(O
s_
I
u
r
4->
CU
E
1^
O
Irt
i i
1
O
-p
3
^t-^CM^DCOCOCO^f^O
1 i CM^r^-COOOLOCOVO
^^ ^o oo oo co r*** co co r*** ^o
E
O
r-
P U.
fO 0
s- co
-P Q
E >ป.
cu a:
O CD
E
O
CJ>
r CMi r ^j-enr^o^J"
CMcococococncncn i
.
co cn
r-~ +i
CD E
> 0
eC *r
-P
r- (O
to >
i- CU
CU Q
>
O "O
s-
a> 03
O "O
S- E
3
-------
SECTION 8
ANALYSIS OF TESTING SYSTEM
This analysis is of three areas of the autoisokinetic samp-
ling system: testing procedures, equipment, and performance.
Problems and difficulties encountered in each area are discussed
and commented upon.
Procedural problems were caused mainly by inadequacies and
omissions in the manual accompanying the autoisokinetic equip-
ment EPA Report No. 650/2-74-029. Discrepencies and inconsistent
labeling were found in the set-up instructions. In section 3.5,
"System Set-Up and Operation", page 3-32, paragraph 2, two
"left hand knobs" are referenced when there are actually only
two knobs total. Figure 3.18 on page 3-29 shows the fluidic
controller and labels the left knob "Gain changer" and the right
"Gain Adjust", while Table 1, page 3-33, calls the right knob
"Variable Gain" and the left "Fixed Gain." No instructions on
how to make a leak test were given or specified. In the "Oper-
ation and Maintenance Manual" accompanying the equipment, all
adjustments are made with the vacuum pump off and out of the
stack. Instructions are to turn on the pump when the probe is
placed into the stack. This disagrees with the subsequent pro-
cedure used in the field evaluation testing, which was verified
as proper by the manufacturers of the equipment via a telecon con-
sultation. The new procedure requires that the vacuum pump be on
before the probe is placed in the stack. A further point of o-
mission was the fact that the three toggle switches on the con-
trol console were not mentioned by either of the supplied man-
uals and were not labeled on the control console in any way.
Concerning the equipment, discussions will start with the
mass flowmeter and the digital flow totalizer. The flowmeter
measurement range is so large that at a flow of 0.5 to 1.0 cfm,
the meter manufacturer's specifications of ฑ 1% full scale yield
potential errors of 10 to 5%. The problem is that these flow
rates are the rates at which most source sampling is performed.
The totalizer, calibrated as received, was 10 to 20% low in its
readings. After recalibration, the totalizer was 51 high which,
for calculating particulate loading, is still inadequate and un-
acceptable according to EPA requirements. The umbilical line is
much too heavy and stiff for sampling purposes, especially when
48
-------
connected to the sampling box which is also too heavy. The con-
nections for the umbilical are unsuited for quickly setting up
or disconnecting. The sample box set-up, besides being unwieldy,
has poor access to the controller valve or the back of fluidic
converter. In fact, the space is so tight that when the fluidic
converter was initially removed for checking purposes, it could
not be reinstalled completely because of pressure against one of
the lead-lag capacitors, which caused the bias pressure to change
It is assumed that the fluidic converter was originally installed
with pressure exerting against the capacitor from the controller
valve.
The probe is too heavy and cumbersome. The one inch stain-
less steel sheath does not extend the full length of the glass
liner. This creates a flexing condition which resulted in three
broken glass liners and in the sheath itself breaking a stainless
steel weld. The rotational ability of the probe is severely cur-
tailed due to the pitot connector and the sensor lines. The mode
of fastening the nozzle to the probe end is inconvenient due to
the hindering sensor tubing and pitot lines. The original noz-
zles had holes which would have prevented proper equipment op-
eration and source sampling. The sensor lines were very fragile
and broke three times through normal sampling circumstances.
Only the low temperature tests were valid. Temperature
seemed to affect the performance. As the temperature increased,
the percent isokinetic dropped. Stack gas static pressures in
excess of 3" water column made the sampler inoperative. The
design limits of stack gas static pressure were determined to be
ฑ 3" water column. Surprisingly, the particulate loading and
velocity appeared to have no noticeable effect.
Overall, the autoisokinetic sampling train does not fulfill
the design criteria and,at its present stage of development,
provides no advantage over the established methods of isokinetic
sampling.
49
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
2.
EPA-600/2-77-035
4. TITLE AND SUBTITLE
FIELD EVALUATION OF AN AUTOISOKINETIC STACK
PARTICULATE SAMPLING SYSTEM
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
3 RECIPIENT'S ACCESSION-NO.
1Q77
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
W. A. Smith and E. W. Stewart
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Entropy Environmentalists, Inc.
Research Triangle Park, N.C. 27709
10. PROGRAM ELEMENT NO.
1AD712
11. CONTRACT/GRANT NO.
68-02-2226
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 6/75 - 6/76
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The performance of a prototype autoisokinetic stack participate sampling system,
designed to maintain automatically isokinetic sampling conditions, was evaluated in
field tests at stationary sources. Tests were conducted to determine the operating
limits and characteristics of the system. Preliminary tests demonstrated the neces-
sity of making several modifications to the existing system to improve the level of
performance for the field evaluation program. Improvements were made in the problem
areas of the performance of the mass flowmeter, flow totalizer, and flow control
valve systems and in the sampling nozzles.
The autoisokinetic sampler was tested at four field installations selected to
provide a wide range of sampling conditions. The results of the testing and analy-
sis showed that the sampling system maintained acceptable isokinetic sampling rates
of 100% ฑ10 at only one of four sources. An inverse relationship between the percent
of isokinetic rate and the temperature of the gas stream being sampled was found.
The evaluation revealed that the sampler will operate only in the narrow range of
stack gas static pressures of ฑ3 inches water column. The physical hardware was
found to be fragile and difficult to operate in the field. Overall, the autoisoki-
netic sampling system failed to meet the design goals.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
*Air pollution
*Particles
*Sampling
*Systems
*Evaluation
*Field tests
13B
14B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
50
------- |