EPA-650/2-74-086-Q
SEPTEMBER 1974
Environmental Protection Technology Series
-------
EPA-650/2-74-086-Q
PROCEDURES FOR MEASUREMENT
IN STRATIFIED GASES
VOLUME I
by
A. Zakak, R. Siegel, J< McCoy, S. Arab-Ismali,
J. Porter, L. Harris, L. Forney, and R. Lisk
Walden Research Division of Abcor, Inc.
201 Vassar St.,
Cambridge, Massachusetts 02139
Contract No, 68-02-1306
Program Element No. 1AB013
ROAP No. 21ACX-092
EPA Project Officer: William B. Kuykendal
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
September 1974
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does mention
of trade names or commercial products constitute endorsement or recommen-
dation for use.
-------
TABLE OF CONTENTS
Section Ti tle Page
LIST OF FIGURES., v
LIST OF TABLES lx
I CONCLUSIONS 1
II RECOMMENDATIONS 2
III INTRODUCTION 3
IV PROGRAM 5
A. TASK I - LITERATURE AND FIELD SURVEY 5
1. LITERATURE SEARCH 5
a. Background and Summary 5
b. Sampling Systems 7
c. Gas Mixing Systems 13
d. Conclusions and Recommendations 13
2. FIELD SURVEY 16
a. Sampling Procedure 16
b. Results 17
c. Conclusion 30
3. GAS STRATIFICATION DOCUMENTATION 30
4. CONCEPTUAL OCCURRENCE OF GAS STRATIFICATION 32
B. TASK II - ANALYTICAL ACTIVITIES AND LABORATORY
EXPERIMENTS 38
1. ANALYTICAL ACTIVITIES 38
a. Analytical Simulation 38
b. Sampling Methodologies 123
c. Jet Mixing of Flue Gas Streams 124
2. LABORATORY EXPERIMENTS 139
a. Wind Tunnel and Test Set Up 139
b. Testing Program 142
c. Results and Conclusions 187
C. TASK III - FIELD DEMONSTRATION 191
1. SAMPLING SYSTEM AND CALCULATION PROCEDURES 191
a. Sampling System Arrangement 191
b. Calculation Procedures 199
iii
-------
TABLE OF CONTENTS (continued)
Section Title Page
2. PRELIMINARY SURVEY TESTS 199
a. Sampling Procedure 203
b. Results 203
c. Discussion of Results and Conclusions 214
3. DEMONSTRATION TEST 223
a. Sampling Procedure 225
b. Results 228
c. Discussion of Results and Conclusions 246
V DISCUSSION AND RECOMMENDED PROCEDURES 252
A. GENERAL..... 252
B. PROCEDURES FOR REPRESENTATIVE SAMPLING USING AN
INTRINSIC TRACER OR REFERENCE GAS 253
C. SAMPLING ARRAY PROCEDURE 255
1. PRE-SURVEY TO ASSESS DEGREE OF STRATIFICATION 255
2. RIGOROUS SURVEY AND SELECTION OF SAMPLING POINTS... 256
3. DESIGN, CONSTRUCT AND INSTALL AN AUTOMATIC ARRAY OF
PROPORTIONAL SAMPLERS 257
VI REFERENCES 262
APPENDIX A through J - See Volume II
-------
LIST OF FIGURES
Number Title Page
SECTION IV.A. —*-
1 A Null Probe 14
2 COz Concentrations Before Dust Collector 31
3 Plane A-A West Side 3-16-70 33
4 Plane A-A West Side 34
5 Plane A-A West Side 35
6 Plane A-A West Side 36
7 Conceptual Occurrence of Gas Stratification 37
SECTION IV.B
1 Probability of Obtaining An Accuracy Within 15% of 9-Point
Analysis for 02 In A Large Duct31 39
2 Schematic - Large Combustion Unit 43
3 Normalized C02 Traverse Data at Dust Collector of poal Fired
Power Plants , 44
4 Tangential Method for Duct Division31. 49
5 EPA Sampling Point Locations 50
6 Types of Asymmetric Velocity Distribution in Pipes 52
7 Traverse Plan for Rectangluar Duct 38 55
8 Error in 4-Point Averaging of Some Arbitrary Axially-
Symmetric Velocity Distributions 57
9a through 21-b- See Appendix G
22 Error in Emission for Rectangular and Circular Ducts as a
Function of Total Number of Probes 100
23 Error in Emission for Rectangular Ducts as a Function of
Strategy and Total Number of Probes 101
24 Error in Emission for Circular Ducts as a Function of
Strategy and Total Number of Probes 102
25 Emission Error vs. Number of Probes for Different Probe
Locations in Rectangular Ducts 103
26 Mean Emission Error vs. Number of Probes for Different
Probe Locations in Rectangular Ducts 104
27 Emission Error vs. Number of Probes for Log Linear Method
for Probe Locations in Circular Ducts 105
28 Emission Error vs. Number of Probes for the Tangential
Method for Probe Locations in Circular Ducts 106
29 Mean Emission Error vs. Number of Probes for Different
Probe Locations in Circular Ducts 107
30 Emission Error vs. Number of Probes for Different Probe
Locations in Rectangular Ducts 115
31 Mean Emission Error vs. Number of Probes for Different
Probe Locations in Rectangular Ducts 116
32 Error in Emission for Rectangular and Circular Duct as a
Function of Total Number of Probes 117
33 Error in Emission for Rectangular Ducts as a Function of
Strategy and Total Number of Probes 118
34 Gas Jet Mixer t" [ t\ 133
35 Passive Mixing Schemes !!!!!!!!!!!!!!!.'.'! 130
-------
LIST OF FIGURES (continued)
Number Ti 11e page
36 Ideal Fan Power for Jet vs. Average Flue Gas Velocity for
Different Flue Diameter at Diameter of Flue Duct/Dtameter
of Jet Orifice = 1.0 135
37 Ideal Fan Power for Jet vs. Average Flue Gas Velocity for
Different Flue Diameter at Diameter of Flue Duct/Diameter
of Jet Orifice = 1.0 136
38 Wind Tunnel Plan View (One Fan) 140
39 Aluminum Honeycomb Cells 141
40 Wind Tunnel Plan View (Two Fans) 143
41 Velocity Distribution in Round Section 144
42 Average Velocity Using the Annubar Element in Round Section. 145
43 Velocity Distribution in Square Section 146
44a Velocity and Concentration Distribution Data in Square
Section 148
44b Velocity Distribution in Square Section 149
44c Concentration Distribution of Ethane in Square Section. 150
45a Velocity and Concentration Distribution Data in Round
Section 151
45b Velocity Distribution in Round Section 152
45c Concentration Distribution of Ethane in Round Section 153
46a Average Concentration Using the "Annubar Element" 154
46b Annubar Flow Element 155
47a Experimental Set-up for Manually Adjusting Sampling Time At
Each Sampling Position , 158
47b Sampling in Square Section at Constant Flow Rate With
Manually Adjusting-Sampling Time for Each Position 159
47c Sampling in Square Section at Constant Flow Rate With
Manually Adjusting Sampling Time for Each Position 160
48a Velocity and Concentration Distribution Data in Square
Sect 1 on 161
48b Velocity and Concentration Distribution in Square Section... 162
49a Velocity and Concentration Distribution Data in Round
Section 164
49b Velocity Distribution in Round Section 165
49c Concentration Distribution of Ethane 1n Round Section. 166
50a Velocity and Concentration Distribution Data in Round
Section 167
50b Velocity Distribution in RoUnd Section 168
50c Concentration Distribution of Ethane in Round Section 169
51a Average Velocity Using the Annubar Element in Round
Section. 170
51b Average Velocity and Concentration Using an Annubar Element
for Sampling 171
52a Velocity and Concentration Distribution Data 1n Square
Section Using the Chebyshef Method for Sixteen Point
Traverse 173
52b Velocity Concentration Distribution in Square Section Using
the Chebyshef Method 174
vi
-------
LIST OF FIGURES (continued)
Number Title Page
53a Velocity and Concentration Distribution Data 1n Square
Section Using The Centroid of Equal Area Method for
Sixteen Points Traverse 175
53b Velocity Concentration Distribution in Square Section
Using the Centroid of Equal Area Method 176
54a Velocity and Concentration Distribution Data in Square
Section Using the Circular Analog Method for Sixteen Points
Traverse 177
54b Velocity Concentration Distribution in Square Section Using
the Circular Analog Method , 178
55a Velocity and Concentration Distribution Data in Square
Section Using the Chebyshef Method for Sixteen Point
Traverse 179
55b Velocity and Concentration Distribution Data in Square
Section Using the Centroid of Equal Area Method for Sixteen
Points Traverse 180
55c Velocity and Concentration Distribution Data in Square
Section Using the Circular Analog Method for Sixteen Points
Traverse 181
56a Velocity and Concentration Distribution Data in Square
Section Using the Chebyshef Method for Sixteen Point
Traverse 183
56b Velocity and Concentration Distribution Data in Square
Section Using the Centroid of Equal Area Method for a
Sixteen Point Traverse 184
56c Velocity and Concentration Distribution Data in Square
Section Using the Circular Analog Method for Sixteen Points
Traverse 185
57 Velocity and Concentration Distribution Data in Round
Section 186
SECTION IV.C. i
la Sampling Arrangement 193
Ib Humidity Test Arrangement 194
Ic Photographs 195
Id Photographs , 196
2 Emission Rate Calculation Procedure 200
3 Data Sheet 201
4 After Air Preheaters Ducts - Plan View 202
5 Results From Test Run at the Scrubber South Inlet Duct 204
6a Data Sheet Test 1 206
6b Data Sheet Test 2 207
7a After Air Preheater Duct (North Side) Velocity, SO*
Concentration and Temperature Traverse at - 130 MW Gross
Output , 208
7b After Air Preheater North Duct Velocity, S02 Concentration
and Temperature Profile at - 130 MW Gross Output 209
vil
-------
LIST OF FIGURES (continued)
Number Title Page
8a After A1r Preheater Duct (South Side) Velocity, S02 and C02
Concentration and Temperature Traverse at ~ 150 MW Gross
Output 210
8b After A1r Preheater South Duct Velocity, S02 and C02
Concentration and Temperature Profile at ~150 MW Gross
Output 211
8c After A1r Preheater Duce (South Side), Port No. 4 Traverse
With Fixed Reference Probe for ~ 150 MW Gross Output 212
8d After A1r Preheater Duct (South) Velocity and S02
Concentration Profiles at - 150 MW Gross Output 213
9a South Duct S02 Concentration at 150 MW for 1.9% Sulfur 011.. 219
9b South Duct Velocity Profile at 150 MW 220
9c South Duct Velocity Profile at 150 MW 221
10 South Duct Velocity Profile at 150 MW 224
11 After A1 r Preheater Sampl 1 ng Arrangement 226
12 C02 and 02 Concentrations at the Inlet South Duct to the
Scrubber Usi ng the Fyri tes 229
13a S02 Recorded Output Signal at 0.5 in/min - Response Curves.. 235
T3b C02 Recorded Output Signal at 0.5 in/m1n - Response Curves.. 240
14a Schematic of Sampling Plane Position Relative to Air
Preheater 249
14b Probable Flow Pattern at the After A1r Preheater Ducts 250
SECTION V.C.
1 For Non-Reverse Flow in Ducts 258
2 For Non-Reverse Fl ow i n Ducts 259
3 For Non-Reverse Flow in Ducts 261
viii
-------
LIST OF TABLES
Number Title
SECTION IV.A.
1 Coal-Fired "Outlet" 18
2 Coal-Fired "Inlet" 21
3 01l-F1red "West" 25
4 Oil-Fired "East" 27
5 Summary Table of Coefficient of Variation (%) for the
Sampl 1 ng PI anes 29
SECTION IV.B.
1 Observed Coefficient of Variation for C02 Traverse at
Various Sampling Locations 4$
2 Observed Coefficient of Variation for C02 Traverse for
Various Coal-Fired Plants 46
3 Location of Measuring Points for Log-Linear Method 51
4 Test Poi nts for Rectangul ar Ducts 54
5 Station Locations and Weights for Averaging 56
7 Velocity Traverse Points 1n Rectangular Ducts with
Perpendicular Ports 63
8 Test Results for a Rectangular Duct (51 x 10') with Velocity
and Concentration Profiles of the Form: 64
9-1 Test Results for Case II 67
9-2 Test Results for Case III 68
9-3 Test Results for Case IV 69
9-4 Test Results for Case V 71
9-5 Test Results for Case VI 73
9-6 Test Results for Case VII 74
9-7 Test Results for Case VIII 75
9-8 Test Results for Case IX 76
9-9 Test Results for Case X 77
9-10 Test Results for Case XI 79
9-11 Test Results for Case XII 81
9-12 Test Results for Case XIII-1 83
9-13 Test Results for Case XIII-2 84
9-14 Test Results for Case XIII-3 85
9-15 Test Results for Case XIII-4 86
10-1 Average Errors for Four Rectangular Duct Sample Cases 87
10-2 Average Error for Six Circular Ducts; Diameter Locations
Segregated 89
10-3 Average Errors for Six Circular Ducts Regardless of
Strategy and Diameter Location 90
10-4 Average Errors for Four Rectangular Ducts Regardless of
Strategy 91
10-5 Average Error for Ten Ducts Regardless of Strategy,
Geometry and Location 92
10-6 Average Error for Six Circular Ducts by Strategy and Probe
Number Regardless of Diameter Location 93
10-7 Average Error for Six Circular Ducts by Diameter Location
and Probe Number Regardless of Strategy 94
ix
-------
LIST OF TABLES (continued)
Number Title Page
11-1 Emission Error vs. Number of Probes Using Different Methods
for Traversing Rectangul ar Ducts 95
11-2 Percent Average Emission Error and Standard Deviation vs.
Number of Probes 96
11-3 Emission Error vs. Number of Probes Using Different Methods
for Traversing Circular Ducts 97
11-4 Percent Average Emission Error and Standard Deviation vs.
Number of Probes 99
12-1 Average Errors for Eight Rectangular Duct Sample Cases 108
12-2 Average Errors for Eight Rectangular Ducts Regardless of
Strategy 110
12-3 Average Percent Error for Fourteen Ducts Regardless of
Strategy, Geometry and Location Ill
12-4 Emission Error vs. Number of Probes Using Different Methods
for Traversing Rectangular Ducts 112
12-5 Percent Average Emission Error and Standard Deviation vs.
Number of Probes for Rectangular Ducts 114
13 Approaches Based on The General Equation for Finite Samples. 125
14-1 Ideal Fan Power for Flue Gas Jet at *150°C (300°F) and
101.32 N/m2 (1 Atm) 133
14-2 Ideal Fan Power for Flue Gas Jet at ~150°C (300°F) and
101.32 N/m2 (1 Atm) 134
15 Testi ng Program Resul ts 147
16 Testing Program Results ' 147
17 Variation of Percent Error in Emission Rate and Total Flow
as Calculated From Different Traversing Techniques Used In
the Square Section... 190
SECTION IV.C.
1 Mystic Station Duct Summary 192
2 Major Equipments Specification 198
3a Values for SOz Concentration and Velocity for the Actual
15 Points Traverse at 150 MW 216
3b Interpolated Values fo SOz Concentration and Velocity for 15
Probes Equal Area 217
3c Extrapolated and Interpolated Values for S02 Concentration
and Velocity for 9 Probes Equal Area 218
4 South Duct (Port 4, 5, 6) at 150 MW Comparing Different
Sampl 1 rig Methods 222
4a North Duct Data Reduction at -144 MW Gross Output 230
4b South Duct Data Reduction at -144 MW Gross Output 231
4c Average Concentrations From Both Ducts 232
5a Test No. 1 233
5b Test No. 2 233
6 Summary of the Demonstration Test Results from Both Ducts... 247
-------
SECTION I
CONCLUSIONS
Results from the literature and field surveys indicate that gas
stratification exists, but it is likely to be less general and less
severe than particulate stratification. For a given gas stream, it is
necessary to make a preliminary gas concentration survey to determine
the existence of spatial stratification.
Where stratification exists, we have concluded as a result of this
program, that there are two methods of obtaining representative gas
samples. Where conditions permit, we recommend a system of monitoring
the ratio of pollutants such as S02, NOX, etc. to C02 from a single
location. Then from the measured fuel flow and chemistry of the process,
the mass flow of C02 is the mass flow of the pollutant. Where conditions
do not permit such a system, we recommend a schedule of manual surveys
and installation of a multi-element proportional sampler and gas velocity
array.
We also examined the use of devices to mix stratified gas streams.
Our conclusion is that passive mixing devices are unlikely to be useful
mixing devices due to the practical problem of retrofitting high pressure
drop devices to existing boiler duct work. An active mixing device using
jet of flue gas was theoretically examined and appears promising; however,
definite conclusions on the effectiveness of this approach is withheld
in the observance of an experimental evaluation.
-------
SECTION II
RECOMMENDATIONS
In the course of this project, areas requiring further development
were Identified. These are as follows:
A. A program 1s needed to develop an automatic Instrumentation
system for extracting continuous representative gas samples from stratified
gas streams, for example, a multi-probe automatic proportional gas sampler
which would be practical in terms of cost and adaptability to various process
streams.
B. A program is needed to develop techniques for determining the
total gas flow profile or velocity vector 1n process streams. It 1s likely
that in practice a significant fraction of errors in emission measurements
are attributable to errors 1n gas velocity/flow determination. (It Is
understood that development in this area is 1n progress.)
C. Further documentation of the extent and frequency of gas strati-
fication 1n process streams together with statistical analyses of data
would be helpful in refining sampling methodologies. (It 1s understood
that a program to gather some of this data will be starting soon.)
D. An experimental program is required to assess fully the practicality!
of using jets to mix flue gases.
-------
SECTION III
INTRODUCTION
This final report marks the conclusion of a ten-month program, the
object of which was to develop methods for the continuous extraction of
a representative gas sample from stratified streams. A representative
sample is one which permits an accurate deduction of a gaseous pollutant
emission or mass flow through a test section. This program is the first
study known to us involving theoretical and experimental investigation into
the problem of gas sampling in stratified flow. However, the concept of
accounting for stratification in the determination of particulate emission
1s well known and appreciated. In the course of this program, we found
that there is little documentation on the extent of spatial variations of
gas concentrations in full-scale power plant effluent streams. Neverthe-
less, the available documentation^ well as our program data indicate that
such stratification exists, although it is unlikely that gas stratification
1s as widespread or as severe as particulate stratification. One of the
results of this program was the formulation of procedures for obtaining
representative gas samples in the presence of gas concentration stratifica-
tion.
The program was organized into three tasks, viz.:
Task I - Survey
Task II - Development
Task III - Demonstration
Task I activities were divided into literature and field survey sub-
tasks. The first sub-task involved a literature and personal contact
survey to obtain and evaluate: 1) documentation of gas concentration and
velocity profiles for process streams, 2) information on sampling procedures,
3) information on mixing process streams, and, 4) information on sampling
devices. The second sub-task involved the performance of field measurements
on four ducts of two power plants to obtain Information on gas stratification,
-------
Task II was divided Into an analytical development sub-task and an
experimental sub-task. In the analytical sub-task, sampling methodologies
were Investigated and a method of mixing flue gases using gas jets was
examined. In the experimental sub-task, laboratory experiments using
various measurement techniques were conducted in a wind tunnel. The results
of these sub-tasks Indicate several approaches to the problem of obtaining
emission measurements 1n stratified flow.
Task III was a demonstration task. Based on the results of Tasks I
and II, procedures for extracting rtpresentatlve gas samples from stratified
process streams were identified and the most potentially reliable of these
were further developed. Different steps utilized in these favored procedures
were then demonstrated on a full-scale power plant.
-------
SECTION IV
PROGRAM
A. TASK I - LITERATURE AND FIELD SURVEY
1. LITERATURE SEARCH
The literature survey Involved an examination of the open lit-
erature as well as a personal inquiry into the present knowledge of flue
gas stratification and the state of existing technology for sampling
stratified gases. Approaches to sampling flue gases are evaluated for
sampling in stratified gas streams.
a. Background and Summary
The problem of sampling stratified gas streams for gases
has not been as universally appreciated by practitioners as has the problem
of sampling for particulate matter. However, where investigators have been
concerned with gas stratification, the approach has been to mimic the spatial
methodology used for the sampling of particulate matter, viz., to sub-divide
the sampling plane of the test duct into a number of equal sub-areas and to
extract samples from the centroids of the sub-areas.
This search found documentation for only single arbitrary
point sampling and the centroid of equal approach mentioned above. No
recommended methodology was found such as those commonly reported for
velocity and particulate sampling.
It was found that people experienced in the sampling of efflu-
ent from boilers recognize a rule of thumb in regard to the composition of
gas streams. This rule is not documented but is in accord with intuitive
notions of air in-leakage, i.e., high gas velocity gives low Og concentration.
No instrumentation designed for obtaining a representative
sample of gases was found which samples gases representatively for emission,
although multipoint gas samplers have been reported which account for spatial
variation over the sampling plane. In the following section (Section IV, B.l.b.,
Sampling Methodologies), it is shown by mathematic development that to obtain
a representative gas sample, certain sampling parameters must be made proportional
-------
to the gas stream velocity (speed) at the sampling location. While no
Instrumentation Intended for proportional sampling of gas was found, there
are available a number of descriptions of actual and conceptual instruments
for the isold netic sampling of particulate matter. Since isokinetic sampling
is a special case of proportional sampling, these instruments are described
and evaluated in terms of applicability to gas sampling with consideration
for simplicity, hence economy of implementation.
Other schemes which obviate traversing techniques are evaluated.
A method using a diffusion tube(s) across the sampling plane was examined.
In principle, this device can obtain a spatial average of the concentration
along the tube and an emission average for the special case of a flat (constant
value) velocity profile. However, in practice, it is expected that a temper-
ature profile in the sampling plane would also effect the sample. For these
reasons an 1n-stack diffusion tube technique is not a promising area for
development.
Another technique examined was the use of a tracer gas to obtain
a representative sample from a single point stack gas extraction. Two general
approaches to this method are to introduce a special tracer gas into the gas
stream or to use an intrinsic gas in the effluent. The approach of using
special tracer gases is, in principle, an unreasonable approach. On the other
hand, the use of an intrinsic tracer, e.g., C02, has no obvious technological
problems and thus warrants further developmental work (see section IV, A. 2.,
Field Survey).
This search found reports of gas mixing using mixing orifices
or baffles (usually half area). The devices are for use on laboratory or
pilot plant scale experiments. This approach is not practical for retro-
fitting to full-scale systems because of the obvious adverse pressure drop
which would be produced in the flue gas stream.
-------
b. Sampling Systems
b.l. General
Given a suitable sampling methodology, certain resolutions
must be made as to which approach will be used to Implement sampling. As
shown 1n the analytical section for sampling methodologies, a sample must be
extracted which 1s volumetrically proportional to the stack gas velocity. The
first resolution to be made is whether to use a single traversing (moving)
sampling nozzle or a fixed array of sampling nozzles. A combination of these
approaches, I.e., a moving array, would compound the disadvantages of both
methods with little advantage over either method.
While the single traversing nozzle reduces the complica-
tions of building an array with a multitude of proportional sampling nozzles,
the expense 1s Incurred of building a complicated in-stack mechanism for
moving the nozzle about. Additionally, the single probe approach is Intrin-
sically unable to obtain a simultaneous measurement over the sampling plane.
Without a formal cost effectiveness analysis, 1t seems that a fixed array of
sampling probes would be the most satisfactory approach.
The practical choice of the type of proportional sampling
system is between sampling at a flow rate proportional to the local stack
gas velocity or at a fixed flow rate for all nozzles in the array but at
times proportional to the stack gas velocity. Both of these approaches, of
course, require a sample integrating scheme to operate. The approach in which
the sampling time is proportional to the stack gas velocity will not provide a
rigorous simultaneous measurement. However, the whole array could be sampled
1n a short time compared to process times, e.g., 1 minute. Therefore, if the
whole array is sampled cyclically, the measurement would be virtually simul-
taneous. This technique requires that the gas analyzer has a response time
much shorter than the scan time over the array.
-------
b.2. Proportional Samplers
Regardless of the choice {type of methodology employed) of
methodology (probe location and number), it is necessary that gas concentra-
tion measurements at a sampling point be weighted by the local gas velocity
and the area ascribed to that probe. In the common methodologies, where each
sample is taken to represent an equal area of the duct, it is required that
all samples extracted be volumetrically proportional to the local stack gas
velocity. The obvious ways to satisfy these conditions are the following:
(a) sample flow rate proportioned to the local
stack gas velocity for equal times
(b) sample equal flow rates for times proportioned
to the local stack gas velocity
The approach indicated by (a) is used in the isokinetic
sampling of particulate matter. A variety of devices has been proposed and
employed for the automatic collection of particulates. Before discussing these
devices, additional requirements for isokinetic sampling which do not apply
to proportional sampling should be Indicated. By definition isoklnetic sam-
pling denotes that the stack gas stream is not accelerated in the vicinity of
the Inlet to the nozzle. This requirement involves constraints applicable
only to ioskinetic sampling, viz.:
the axis of the sampling nozzle must be parallel
to the stack gas stream with the nozzle facing
into the stream
the gas velocity (vector) at the face of the
nozzle must be identical to the stack gas
velocity in the neighborhood of the nozzle
It is seen that when (b) above is satisfied, the require-
ments for proportional sampling are satisfied. The proportionality of the
sample flow to the stack gas velocity is the nozzle area.
8
-------
This search found no reports of automatic proportional
samplers; however, documentation of several isokinetic samplers was found.
Descriptions of particular automatic isokinetic instruments are presented
in Appendix A.
Three general approaches have been used in automatic
isokinetic samplers. These are the following:
(1) null balance probe
(2) computation and adjustment of sample flow rate
based on information from a velocity measurement
in the neighborhood of the sampling nozzle
(3) passive, gas stream driven sampler
In the null balance nozzle, the flow (velocity) of the
sample through the nozzle is matched to the stack gas flow (velocity) just
outside the nozzle. In this approach, the temperature and pressure of the
sample stream at the sample sensor are assumed to be (or made to be) identical
to the temperature and pressure of the stack gas stream at the reference
sensor. Sensors which have been or could be used include:
* static pressure probes
• temperature probes
• thermal anemometers
Examples of the null static pressure types are described in Volume II, on
pages 1 through 3 and Reference 50. An example of a null temperature probe
type is shown on page 4. Thermal anemometers located in the sampling nozzle
and just outside the nozzle are straightforward applications of a null approach.
An attractive feature of this approach is the simplicity with which an error
signal is obtained to operate a closed-loop servo-mechanism for flow control.
It 1s significant that the only reference to a system implementing automatic
multl-point Isokinetic sampling was of the null balance type (see page 3).
-------
Computation of the sample flow rate from a velocity
measurement made 1n the neighborhood of the sampling nozzle Involves the
automation of the usual manual procedure for 1sok1net1c sampling. Examples
of Instruments using this approach are shown In Volume II, on pages 5
through 9. Even to the casual glance this approach Is more complex than
the null balance approach, and hence, 1t 1s not attractive for a multl-point
sampling system.
The last approach for discussion 1s a passive device
which 1s operated, hence adjusted, by the gas flow. This 1s an attractive
approach since 1t eliminates the need for a servomechanlsm for control. A
particular example of this type of device Is shown on page 10 1n Volume II.
b.3. Arrays and Traversing Mechanisms
Probe arrays and mechanical traversing mechanisms may
be used to Implement sampling methodology. Shown In Appendix B are some
examples of these approaches. Generally, the arrays have been used to
obtain a spatial average of concentration over the sampling plane. The array
system (Volume II, P. 13) could have an interesting extension 1f the sample
time for each probe were adjusted to be proportional to the stack gas velocity.
This would of course require a companion array of velocity sensors and a
logic system for control.
It was suggested that a gas sample be extracted from an
2
Annubar Instrument (Volume II, p. 16). However, sampling from this Instrument
did not lead to a representative sample. Sampling conditions and results
are reported In the laboratory tests section.
Mechanical systems which move a single sensor over a
sampling plane have been constructed, but these devices can become very com-
18
plex. A simpler system which is mechanically driven, is commercially avail-
able for a particulate analyzer. This system is shown in Volume II, p. 17.
Inquiry Into the rationale for the trajectory of this system revealed that it
was designed principally for convenience and based on the assumption that it
was ah Improvement over a single point measurement.
10
-------
It 1s significant to note that 1n rectangular ducts,
averaging was recommended (Volume II, p. 18) along the diagonal axis of the
test section, including the corner points and sampling through a common
manifold. By this method, a true average can be obtained only when a
uniform flow exists across the duct cross section.
b.4. Diffusion Tube Samplers
The use of diffusion tubes located over the sampling
plane has been suggested as a method for obtaining a representative gas
sample from flue gas. It 1s shown in Appendix C that this approach will,
in principle, give a spatial sample average, but is unlikely to give a
representative sample for emissions except in the special but trivial case
of virtually constant gas velocity over the sampling plane. Diffusion
3
tubes have been used in stack sampling where the Intent has been only to
separate the sample gas from moisture and particulate matter. Diffusion
tube devices are temperature dependent; hence, in applications where a
spatial average 1s sought, an erroneous answer is obtained where gas
temperatures are stratified over the sampling plane.
A reference to a proposed method for controlling tem-
perature of a diffusion sampling tube using a heat pipe was found (see
Appendix D). However, the fact that diffusion tubes will only provide a
spatial average and not an emission average makes this approach unpromising
and would not warrant the temperature stabilizing development effort.
A sampling scheme using an out-of-stack diffusion tube
is outlined in Appendix C, Figures C-l and C-3. This conceptual design
eliminates the need for a servo-mechanism for proportional control and is
similar to the device shown in Appendix A, Figure A-10. Preliminary cal-
culations were made 1n order to estimate the length of diffusion tubes
required. The detailed calculations are shown in Appendix C. Results
show that the length of Teflon PTFE necessary to permeate 5,000 ppm at 100°C
11
-------
from a 50 x 1
-------
c. Gas Mixing Systems
A system which mixes the flue gases would reduce the sampling
procedure to a simple single point gas concentration measurement, providing
emission data when multiplied by the total flow. The flow measurement may
be made anywhere along the duct system. Baffles, usually one-half area
plates, have been used for mixing gas streams * . This approach is incom-
patible with retrofitting to full scale systems because intolerable pressure
drops are produced in the process stream.
The method of using jets to mix the gas streams was examined
analytically (see jet mixing section, Section IV, B.C.). Preliminary results
Indicated that the power required for the mixing jet can be modest as compared
to the power required to drive flue gas through the duct. However, scaled
laboratory experiments should be performed to test the analytical results
prior to any full-scale demonstration of this approach.
d. Conclusions and Recommendations
This search found that there is no generally recognized or
specifically reported sampling methodology (sampling strategy) for stratified
gases. When the problem of measuring stratified gases has occurred, 1t has
been handled on an ad hoc basis, frequently by the use of rakes or probe
arrays which account only for spatial variations of gas concentrations with
no consideration of velocity, hence emission. Development of a rational
sampling methodology is a promising analytical approach. Such an approach
was undertaken as a part of this study and 1s reported elsewhere in this
report.
No reports of automatic proportional gas samplers were found.
However, three general approaches to isoklnetic sampling have been identified,
which could form the conceptual basis for automatic proportional sampler
design. Of the three approaches, null balance and the gas driven approach
seem most promising for instrument development, since these devices appear
to offer some simplicity of Implementation. A conceptual example 1s given in
Figure 1. A null probe 1s shpwn using either a thermal anemometer or a
13
-------
Flue gas
transducer
FOR ONE SAMPLE LINE
Null or proportional
balance amplifier
motorized control valve
From mixing chamber
To analyzer
Filter unit
Sampling pump
Figure 1. A Null Probe
-------
thermocouple sensor approach7'8'9'10. The null probe method 1s a straight-
forward application of servo-mechanisms. For multl-point arrays the sensors
and servomechanisms would, of course, have to be repeated for each element
of the sample array. The development of this concept 1s promising but re-
quires an extensive program.
The conclusions reached from examining mechanical traversing
schemes 1s that a considerable effort with a large amount of complicated
machinery 1s necessary to achieve a rational traversing plan. From a common
sense point of view, mechanical traversing schemes appear complex compared
to the more promising fixed array of probes.
A possible area for development of the probe array approach
is to use both a sampling nozzle array system (conventional gas-extractive
probes) with an associated velocity array. The data from the velocity array
would be used to control the sampling time (the volume by design) sequentially
for each probe proportional to the local stack gas velocity. A logic controlled
system for this approach, though straightforward, 1s not trivial. It has been
reported by Grandville that at locations remote from bends or obstruction
and where the mean velocities are greater than about 20 ft. per sec., the
flow distribution patterns are often similar, although the total flow rate
may vary. Given these conditions, a simple control system using a timer
program could be implemented which would be dependent only on the velocity
profile (relative velocity) and not the value of total flow. On installation
of the system, the timer would be adjusted for the particular.flow profile.
The total flow may be determined by a single, independent velocity measure-
ment. This appears to be a very promising area for development; however,
more substantial evidence on the behavior of flow profile preservation under
fluctuations of total flow would be necessary before pursuing this approach.
The use of a tracer introduced to the gas stream is not a
promising approach. The use of an intrinsic tracer, e.g., COg, 1s a recom-
mended approach. The scheme of producing a constant gas concentration over the
sampling plane by use of an upstream mixing device 1s not promising for when
15
-------
devices such as baffles are used, Intolerable pressure drops are Introduced
Into the system. The approach which Involves using a jet mixing scheme 1s
promising. However, before attempting a full-scale application, scaled lab-
oratory experiments are recommended to validate the analytical results Indi-
cated 1n this report.
2. FIELD SURVEY
The results of the field survey are presented herein. Data were
taken from a total of four sampling planes using NDIR analyzers for S02 and
COg concentrations and an S-type pitot tube for velocity. Two of the sampling
planes were located across the ductwork of a coal-fired power plant (Bow,
New Hampshire), and two were located across the ductwork of an oil-fired
/
power plant (Weymouth, Massachusetts). The data from the coal-fired power
plant were taken from locations just before and just after the electrostatic
predpltator. The data from the o1l-f1red plant were taken from locations ,
on ductwork just before the stack breaching. In all cases, existing sampling
ports were used; hence, no special efforts (other than inspection) or costs
were made to obtain sampling planes in locations of known gas stratification.
a. Sampling Procedure
The data taken were simultaneous concentrations of SO* and CCL
by NDIR analysis and pitot tube heads. Attempts were made to collect infor-
mation on gas temperature and angle of attach but Instrumental problems pre-
cluded reliable data. A fine sampling mesh (2 Inches) was used near the walls
but somewhat further apart (on some locations) near the center (6-1nch mesh).
Because of the fine sampling mesh, several hours were necessary to traverse
a duct; hence, temporal variations in concentration may effect the data.
16
-------
b. Results
The data have been reduced and normalized to the following forms:
S00 concentration
(a) 2
/u\
* '
average SO,, concentration
S00 concentration
4
C0« concentration
average S0« concentration
(cj /TT" _ a velocity
average / A P average velocity
The set of points represented by (a) and (c) describe the S02
and velocity profiles over the sampling planes. The set of points represented
by (b) describe the differences between the SOg concentration profile and
the COg profile.
Tables 1 through 4 present the normalized data. The distance
values denote the Insertion depth of the probes. The standard deviation
(S.D.), the mean (average) and the coefficient of variation (CV) have been
calculated for both the set of points associated with a port and the set
of points associated with the whole sampling plane. This breakdown was done
because the total time to traverse the whole sampling plane took several hours
(approximately a work day 1n some cases) and no fixed-point reference probes
were used to gather correction data to account for temporal variation 1n the
process stream. Therefore, the set of points associated with a port are
more closely related 1n time than the set for a whole sampling plane.
A summary of the amount of gas stratification for all sam-
pling planes is shown in Table 5, along with the amount of velocity strat-
ification. The amount of gas stratification shown for the coal-fired power
plant data 1s so low that it probably reflects the precision of the analyzers
as much as the amount of gas stratification. The amount of S02 stratification
for the oil-fired power plant data is higher than the suspected precision of
the analyzers, but the magnitudes represent only a moderate amount of
stratification.
17
-------
TABLE 1
COAL-FIRED "OUTLET"
ySF
0.48
0.93
0.98
1.05
1.09
1.09
1.12
1.12
1.14
1.14
1.14
1.14
1.14
1.09
1.09
1.09
1.12
1.09
1.11
1.11
1.13
1.13
1.13
1.11
1.11
1.11
1.11
1.09
1.14
0.89
„
0.127
1.073,
11.84*
SOj/CC
^ so2
0.983
0.970
0.983
0.952
0.963
0.963
0.963
0.966
0.968
0.974
0.961
0.955
0.955
0.966
0.966
0.961
0.955
0.955
0.961
0.966
1.058
1.066
1.058
1.058
1.071
1.069
" 1.057
1.051
1.045
1.045
«,•
0.045
0.995
4.52*
SOg/SUj
ysoycOa
•flF//£
s^
1.016
1.029
1.016
1.029
1.041
1.041
1.041
1.054
1.066
1.073
1.079
1.073
1.079
1.092
1.092
1.065
1.079
1.079
1.085
1.092
1.079
1.073
1.079
1.079
1.092
1.098
1.085
1.079
1.079
1.079
* •
0.023
,1.068
2.15*
S.P.
0.0205
0.0263
0.1156
Distance
(Inch)
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
AVG
AVG.
1.000
1.005
0.991
•£"
0.47
0.95
0.93
0.9S
0.95
.02
.05
.02
.05
.02
.02
| .02
g, .02
» >05
~ .05
S .05
.02
Z 1.05
£ 1.09
^ 1.09
I i-°'
— 1.09
1.14
1.12
1.09
1.09
1.07
1.07
1.05
0.98
—
0.115
1.021
11.26*
SIX PORT
cy*
2.05*
2.62*
11.68*
50"2 S02
0.868
0.900
• 0.911
0.932
0.922
0.922
0.932
0.932
0.932
0.932
0.932
0.932
0.922
0.922
0.916
0.916
0.938
0.956
0.933
0.933
0.945
0.944
0.944
0.944
0.933
0.927
0.933
0.922
'0.933
0.911
~
0.016
.927
1.73*
so. c
1.016
1.054
1.079
1.104
1 .092
1.092
1.079
1.079
1.079
1.079
1.079
1.079
1.066
1.066
1.054
1.054
1.079
1.054
1.054
1.054
1.041
1.054
1.054
1.066
1.054
T.066
1,054
1.066
1.079
1.079
—
0.0176
1.0668
1.651
lstanc<
(Inch)
0
2
4
6
8
10
12
14
16
IB
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
st"
AVG
i *
•
1
P
J
£
y
?
f%
I
«#*
i
"^
* Coefficient of variation
18
-------
TABLE 1 CONT'
/SF s
0.00 (
0.70
0.81
0.93
0.93
0.93
0.93
0.93
0.91
0.91
0.86
0.91
0.86
0.86
0.86
0.91
0.91
0.91
0.93
0.91
0.95
0.93
0.93
0.95
0.95
0.93
0.93
0.93
0.93
0.93
0.86
0.051
0.904
5.64*
"2 S02
1.965
1.000
.023
.014
.021
.021
.058
.041
.047
.041
.047
.047
.063
.055
.055
.065 ,
.065
.072
.079
.072
.065
.093
.069
.072
.063
.079
.079
.072
.065
1.072
1.096
0.028
1.054
Z.66X
SO,
0.885
0.916
0.916
0.928
0.928
0.928
0.947
0.954
0.960
0.954
0.960
0.960
0.966
0.966
0.960
0.954
1
0.954
0.960
0.966
0.960
0.954
0.979
0.972
0.960
0.966
0.966
0.966
0.960
0.954
0.960
0.966
0.019
.952
1.99*
Distance
(Inch)
0 C
2 C
4 1
6 1
8 1
10 1
12 1
14 1
16 1
18 1
20 1
22 | 1
24 g> 1
26 £ 1
28 * 1
30 » 1
y
32 f 1
34 ~
36 f
38 -ti
*•*•
40 gi
42 -*
44
46
48
50
52
54
56
58
60
1
/E£ 5ff? S02
1.23 1.009
1.81 1.009
.04 1.022
.12 1.022
.12 1.022
.09 1.022
.09 1.035
.09 1.022
.07 1.028
.09 1.017
.07 1.022
.09 1.028
.19 1.035
.19 1.037
.14 1.035
.14 1.026
.16 1.028
.16 1.014
.19 1.014
.19 1.026
.19 1.026
.19 1.026
.16 0.984
.16 0.984
.16 0.997
.16 0.984
.14 0.984
.12 0.978
.12 0.978
9.98 0.991
—
3.179 0.018
1.088 1.013
5.45* 1.78*
!°i
0.966
0.966
0.979
0.979
0.979
0.979
0.991
0.979
0.991
0.960
0.979
0.985
0.991
0.979
0.991
1.004
0.991
0.991
0.991
1.004
1.004
.004
.004
.004
.016
.004
,004
0.997
0.997
1.010
--
0.013
.99
1.31*
Distance
(Inch)
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
. 56
58
60
J.D.
£VG
f V
f
2
r*
ft
3
3?
3
r»
^
|
2
i
"""
* Coefficient of variation
19
-------
TABLE 1 CO,T'
f i
0.47
0.74
0.86
0.91
0.91
0.91
0.91
0.93
0.95
0.93
0.93
0.93
0.93
0.93
0.95
0.95
0.93
0.91
0.98
1.02
1.02
1.02
0.98
0.98
1.02
0.98 *
0.98
0.98q
0.95
0.93
..
0.102
0.927
11.01
5&"2 so2
0.937
0.943
0.943
0.977
0.964
0.971
0.977
0.971
0.964
0.964
0.964
0.976
0.984
0.996
0.990
0.976
0.970
0.976
0.975
0.983
0.989
1.001
0.988
0.988
1.000
0.994
0.988
0.974
0.989
1.003
.-
0.016
0.977
1.64*
S02 Distance
^~ (Inch)
0.885
0.891
0.891
0.922
0.914
0.916
0.922
0.916
0.916
0.916
0.916
0.928
0.928
0.941
0.941
0.928
0.922
0.928
0.941
0.947
0.957
0.979
0.966
0.556
0.985
0.972
0.966
0.954
0.954
0.954
—
0.026
0.935
2.781
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
jgP
0.33
0.74
0.74
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.93
0.95
0.95
g 0.95
1 0.95
» 0.95
3 0.95
3 0.95
~ 0.93
§. 0.95
2 0.95
|_ 0.98
*" 1 02
1.02
1.05
1.02
1.02
1.02
0.98
1.02
0.95
0.129
0.929
13.881
WZ S02
eo-2 *$
0.976
0.990
0.990
.032
.032
.037
.024
.024
.024
.037
.065
.109
.107
.099
1.099
1.094
1.094
1.082
1.082
1.075
1.088
1.087
1.087
1.087
1.079
1.074
1.074
1.087
1.081
1.073
1.093
0.036
1.063
3.391
!i
0.928
0.941
0.941
0.960
0.972
0.979
0.966
0.966
0.966
0.979
0.991
1.016
1.023
1.023
1.023
1.010
1.010
0.991
0.991
0.985
0.997
1.004
1.004
1.004
1.004
0.991
0.991
1.004
0.997
0.997
1.016
0.024
0.989
2.43
Distance
(Inch)
0
?
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
4?
44
46
48
50
52
54
56
58
60
S.D.
AVG
C/
F
o
c
n
„
in
Q
8
2
I
Coefficient of variation
20
-------
TABLE 2
COAL-FIRED "INLET"
1
0.83
0.92
0.97
0.95
0.95
0.94
.05
.06
.06
• 08
.05
.05
.05
.02
.02
0.97
0.90
0.87
0.86
0.81
0.76
0.75
0.75
0.108
0.942
(0.1146)
11.46*
W2 C02
0.877
0.950
0.989
0.989
0.996
1.003
0.983
0,990
0.990
0.990
0.997
0.991
0.977
0.964
0.971
0.985
0.985
0.978
0.972
0.978
0.979
0.979
0.972
0.024
0.977
(0.0246)
2.46*
!4
0.860
0.939
0.978
0.978
0.984
0.984
0.965
0.958
0.952
0.952
0.965
0.939
0.939
0.926
0.932
0.932
0.932
0.913
0.900
0.906
0.900
0.900
0.893
0.032
0.935
(0.0342)
3.42*
Distance
(Inch)
0
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
63
66
69
70
SEP
-SF
• 0.94
0.98
1.05
1.08
.06
=L .10
n
** .06
g- .06
I -05
& .03
A .02
? .02
^ 0.97
g 0.95
- 0.97
^n
£. 0.98
0.98
0.98
0.97
0.95
0.95
0.97
0.98
0.76
0.067
0.993
(0.0674)
6.74X
§£
0.943
0.956
0.943
0.930
0.943
0.963
0.963
0.911
0.957
0.963
0.943
0.957
0.996
1.003
1.002
1.002
0.989
1.002
0.994
0.994
0.988
0.988
1.000
1.008
0.994
0.027
0.973
(0.0277)
2.77X
i|
0.978
0.991
0.965
0.952
0.952
0.965
0.952
0.913
0.939
0.965
0.926
0.939
0.978
0.991
1.004
1.004
1.004
0.991
1.017
1.017
1.017
1.017
1.030
1.030
1.017
0.033
0.982
(0.0336)
3.36X
Distance
(Inch)
0
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
2.
R
C
%
a.
m
2
?
J
1
_
[?
^
I
54 r-
57
60
63
66
69
70
$TD.D.
AVG.
C V*
AVG. OF SEVEN PORT
so2/W2
«yco2|so2/co2
JEF
t*Z
S.D.
0.0319
0.0247
0.1360
AVG
1.005
1.002
1.016
C V*
3.88X
2.47X
13.38X
* Coefficient of Variation
21
-------
TABLE 2 CONT'
£SE
.05
.10
.16
.16
.19
.19
.19
1.19
1.13
0.98
0.95
0.81
0.90
1.00
0.94
0.86
0.130
1.050
(0.0124)
12.41
SOL SO, SO, Distance
sF^J iff"
1
u>
TJ
O
ft
O
O
•n
I
* Coefficient of variation
22
-------
TABLE 2 CONT1
1 53? I
.08
.16
.14
.17
.19
.22
.21
.22
.02
0.98
0.86
0.81
0.83
0.83
0.81
0.78
.019
.031
.057
.057
.057
.042
.055
.058
.051
.076
.020
.034
.028
.015
.015
.028
.056
.069
.095
.095
.095
.095
.108
.082
.082
.108
.043
.043
.030
.017
.017
.030
0.173 0.018 0.032
1.019 1.040 1.066
0.1697) (0.0173) (D. 030)
6.97X 1.73X 3.0X
Distance
(Inch)
0
2
4
6
8
10
12
24
36
48
60
62
.64
66
68
70
5*
a
Southsldc
•V
$
J
f
3
/F
1.05
1.13
1.19
1.21
1.22
1.24
1.24
1.21
1.10
1.10
0.84
0.83
0.79
0.78
0.78
0.76
0.194
1.029
(0.1885)
18.85X
53!
0.984
1.005
1.032
1.065
1.042
1.037
1.029
1.046
1.005
1.029
1.032
1.017
1.032
1.047
1.087
1.041
0.024
1.033
(0.0232)
2.32X
$02
0.952
0.965
0.991
1.030
1.030
1.017
1.017
0.991
0.965
0.939
0.913
0.900
0.913
0.926
0.939
0.900
0.046
0.961
(0.0479)
4.79X
Distance
(Inch)
0
2
4
6
8
10
12
24
36
48
60
62
64
66
70
|.D.
£vg.
£ v*
^
5*
«•
I
3!
J
**
I
Coefficient of Variation
23
-------
TABUF 2 CONT
rt£
0.94
1.08
1.10
1.06
1.10
1.10
1.10
1.13
1,13
1.06
mm
0.75
0.83
0.81
0.75
0.68
0.194
0.997
0.195
99. 5S
Ek S02
0.969
1.007
1.007
1.015
1.016
1.023
1.018
1.034
1.034
1.034
—
1.045
1.045
1.016
1.032
1.061
0.021
1.023
0.02
21
5
0.991
1.017
1.017
1.017
1.004
1.004
0.978
0.96S
0.965
0.96S
—
0.939
0,939
0.913
0.913
0.939
0.036
0.971
0.037
3.7*
Distance
(Inch)
0
2
4
6
8
10
12
24
36
48
60
62
64
66
68
70
Std. Oevn.
Avg.
cv*
«
j1
••*
1
1
s
^
**
I
1
1
* Coefficient of Variation
24
-------
TABLE 3
OIL-FIRED "WEST"
/AP
0.86
0.84
0.84
0.84
0.86
0.84
0.93
.02
.09
.05
,14
.19
.19
.28
.35
.37
.37
.37
1.39
1.37
.35
.37
.35
.35
.39
.39
.39
.42
.35
0.219
1.191
18.38S
VV
*
SBg S02
0.993
0.993
1.027
1.027
1.084
1.084
1.123
1.162
1.162
1.141
1.141
1.123
1.123
1.121
1.063
1.041
1.063
1.082
1.082
1.041
1.018
1.059
1.059
1.018
1.018
1.018
1.039
1.039
1.059
0.049
1.069
4.58X
SOg/SOg
5oyto~2
W//ZZ
SO,
1,065
1.065
1.065
1.065
1.104
1.104
1.143
1.143
1.143
1.143
1.143
1.143
1.143
1.104
1.065
1.025
1.065
1.065
1.065
1.025
0.986
1.025
1.025
0.986
0.986
0.986
0.986
0.986
1.025
0,057
1.064
5.36%
S.D.
0.0622
0.0530
0.2731
Distance
(Inch)
0
2
4
6
8
10
12
PI
16 f
M
in
20 £
"22 f
24 £
30 S
36 f?
42 £
48 £
50 I
52 ^
54
56
58 •
60
62
64
66
68
70
71
Avg. We
Avq.
1.000
1.001
1.01
1
0.65
0.60
0.60
0.60
0.56
0.56
0.56
0.60
0.65
0.65
0.65
0.81
0.84
0.93
0.93
0.95
0.98
.07
.16
.19
.21
.28
.32
.30
.40
1.44
1.30
1.30
1.14
0.305
0.938
32. SIX
st Ports
C V*
6.22%
5.30%
27.04%
i$
0.788
0.938
1.U02
1.002
1.002
1.041
1.023
1.059
1.018
1,059
1.041
1.059
1.018
1.018
1.059
1.018
1.002
1.059
1.034
1.009
1.005
1.050
1.004
1.025
1.000
1.000
1.000
0.979
0.979
0.051
1.010
5.05X
SO,
0.789
0.907
0.986
0.986
0.986
1.025
1.025
1.025
0.986
1.025
1.025
1.025
0.986
0.986
1.025
0.986
0.986
1.025
0.946
0,907
0.907
0.907
0.867
0.867
0.828
0.828
0.828
0.828
0.828
0.079
0.942
8. 38X
Distance
(Inch)
0
2
4
6
8
10
12
14
m
16 £•
18 *
20 £
22 |
24 *
30 £
36 ?
42 ^
48 xj
so'?
52 ~
54
56
58
60
62
64
66
68
70
71
S.D.
Avg.
C V*
Coefficient of Variation
25
-------
TABLE 3 CONT'
jffi
0.56
0.58
0.60
0.60
0.60
0.60
0.60
0.65
0.70
0.74
0.77
0.74
0.77
0.77
0.86
0.95
.05
.07
.07
.07
.09
.09
.14
.19
.23
1.30
1.35 .
1.44
1.32
0.276
0.913
30.23X
5>£
0.740
0.817
0.856
0.870
0.906
0.906
0.906
0.927
0.943
0.943
0.943
0.979
0.963
1.000
0.927
0.842
0.870
0.906
0.957
0.957
0.952
0.945
0.929
0.961
0.929
0.973
0.957
1.007
0.988
0.058
0.924
6.28t
!^2
0.867
0.946
0.946
0.946
0.986
0.986
0.986
1.025
1.025
1.025
.025
.065
.065
.104
.025
0.946
0.946
0.986
1.025
1.025
0.986
0.946
0.946
0.946
0.946
1.025
1.025
1.025
1.025
0.049
0.993
4.93*
Dlstanc
(Inch)
0
2
4
6
8
10
12
14
16
18
20
22
24
30
36
42
48
50
52
54
56
58
60
62
64
66
68
70
71
5.0.
fV3*
B
-fM
f
n
r»
§
f
r*
£
**
j?
j
T
»*
* Coefficient of variation
-------
TABLE 4
OIL-FIRED "EAST"
fS. i S°2
0.76 0.718
0.89 0.881
0.94 0.988
0.96 1.050
0.98
0.98
0.98
.00
.02
.00
.02
.00
.00
0.98
0.98
0.94
0.96
0.94
0.94
0.94
0.94
0.89
0.89
0.91
0.89
0.94
.050
.082
.117
.084
.103
.103
.121
.103
.121
.121
.121
.087
.107
.126
.091
.091
.111
.097
.Q97
.060
.082
.082
0.94 1.082
0.94 1.060
0.94 1.025
0.052 0.083
.947 1.066
S.5X 7.79X
SO./SB,
so2
5*2
0.796
0.961
1.061
1.127
1.127
1.160
1.160
1.127
.127
.127
.127
,127
.127
.127
.127
.094
.094
,094
.061
1.061
1.061
1.028
1.028
0.995
0.995
0.995
0.995
0.995
0.961
0.080
1.064
7.5X
S.D..
0.0730
Distance
(Inch)
0
2
4
6
8
10
12
14
16
18
20
22 2
24 1
30 £
36 £
42 S
m
48 S
50 o
52 fl
54 5
56 g,
58 A
60
62
64
66
68
70
71
Avg.
Avg.
1.000
./SF OT2 S02
^ ™Z ^
0.30
0.59
0.64
0.64
Q. 62
0.59
0.66
0.68
0.70
0.76
0.83
0.85
0.87
0.87
0.85
0.83
0.89
0.91
0.96
0.98
.02
.08
.08
.08
.13
.15
1.21
1.21
1.19
0.226
0.867
26.06X
East Ports
CJL*
7.3X
0.749
0.932
1.000
1.029
1.056
1.099
1.148
1.173
1.159
1.159
1.181
1.146
1.191
1.146
1.111
1.111
1.126
1.111
1.111
1.132
1.154
1.175
1.154
1.111
1.132
1.060
1.078
1.082
1.082
0.088
1.099
8.0X
SOg Distance
0.995
1.127
1.194
1.227
1.260
1.293
1.194
1.160
1.127.
1.127
1.127
1.094
1.127
1.094
1.061
1.061
1.094
1.061
1.061
1.061
1.061
1.061
1.061
1.061
1.061
0.995
1.028
0.995
0.995
0.077
1.098
7. OX
0
2
4
6
8
10
12
14
16
18
20
22
24
30
36
42
48
50
52
54
56
58
60
62
64
66
68
70
71
J.D
Avg
£
rt
§
M
ft
R
£
*
0.0762
1.000 7.62X
0.1250 ,999 12.51*
* Coefficient of Variation
27
-------
TABLE 4 CONT1
/SP
'T^ff
Ur
1.02
1.13
1.13
1.13
1.11
1.11
1.08
1.11
1.11
1.11
1.13
1.15
1.15
1.17
1.13
1.19
1.21
1.23
1.21
1.21
1.21
1.21
1.25
1.25
1.25
1.30
1.30
1.30
1.21
0.070
1.175
5.961
SO, S02
0.617
0.753
0.803
0.846
0.879
6.893
0.879
0.893
0.908
0.875
0.875
0.903
0.848
0.848
0.819
0.805
0.792
0.805
0.792
0.805
0.805
0.805
0.805
0.819
0.819
0.854
0.872
0.887
0.903
0.059
0.834
7.07X
sir
0.663
0.796
0.862
0.895
0.928
0.928
0.928
0.928
0.928
0.895
0.895
0.862
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.796
0.829
0.862
0.862
0.862
0.062
0.836
7.41X
Distance
(Inch)
0
2
4
6
8
10
12
14
16
18
20
22
24
30
36
42
48
50
52
54
56
58
60
62
64
66
68
70
71
$.0.
Jvg.
JV*
5
«o
B>
3
r»
1
ft
r*
ft
S
rt
3
* Coefficient of variation
28
-------
TABLE 5
SUMMARY TABLE OF COEFFICIENT OF VARIATION" (%) FOR THE SAMPLING PLANES
Coal -Fired
Inlet
Outlet
01l-F1red **
West
East
so2/^2
3.2
2.1
6.2
7.3
(so2/co2) / ($o2/co2)
2.5
2.6
5.3
7.6
rST ISTT
13
12
27
13
* NOTE: As a rule of thumb the range of a set of numbers 1s about 3 to 6 times
the standard deviation. Therefore, the range of stratification about
the average value, represented In percent, Is ± 3 to 6 times the C.V.
with ± 4-1/2 times as typical. For example, in the case of the coal-
fired Inlet duct, one would expect extreme deviation for the S02 con-
centration values to be 1n the neighborhood of ± 14% for the mean.
** This oil fired plant 1s Edgar Station. The field demonstration, TASK III,
was performed at Mystic Station, also an oil fired plant.
29
-------
The reason that the concentration ratios of S02 to C02 vary
about the same amount as the S02 concentration 1s not obvious. Clearly,
this result 1s not in accord with the hypothesis which predicts that S02
and C02 are stratified in the same way, i.e., ratios are preserved through
dilution, etc. One could look to the effect of the analysis technique
itself for a possible explanation. In any case, further experiments described
under Task II are in accord with the above-stated hypothesis.
c. Conclusion
If one assumes that the precision of the analyzer is ± 2 to
± 5 percent, the amount of gas stratification for the coal-fired power plant
is almost negligible (range for all S02 data points 1s + 7% and - 6% from
the average of the reading for respective parts).
The data for the oil-fired plant do show noticeable stratifi-
cation with a range for SCL from - 25% to + 18% (from the average for a port)
for all S0« concentrations.
The ratio data of S02/C02 are not consistent with the theory
which predicts that this ratio should be a constant over the sampling plane.
It is expected that this Inconsistency is the result of the sampling technique
used, since subsequent measurements during Task III show a constant ratio.
3. GAS STRATIFICATION DOCUMENTATION
During the literature search task of this program, data were sought
on the existence of stratification in gas streams for both gas concentration
and the associated gas velocity. It was discovered that this type of data are
scarce; however, a few examples were found and are presented below.
No data with both velocity and concentration were obtained from the
literature search. However, some data showing gas concentration stratification
were found. Data supplied by the Bureau of Mines are shown in Figure 2. The
COconcentration ranges approximately 40% from the mean at a location at the
30
-------
DUCT BEFORE OUST COLLECTOR
— 4
9.5
3 c;
llJO
—2 G
9.5
— 1 (•
10.7
A
12-4
,;
13.8
G
14.1
G
13.5
B
c
14.1
14-1
G
14.1
•-
IZ.9
U
c
c
14.,
0
i3.e
0
136
0
13.6
u
0
C
12,9
G
13.8
©
13.8
©
13.2
CT
E
0
G
12.1
G
u.e
©
11.8
U '
F
FIGURE 2
CO. CONCENTRATIONS BEFORE DUST COLLECTOR
-------
inlet to a dust collector. Data supplied by TVA are shown in Figures 3
through 6 and show various degrees of SOp stratification at a test section
following a furnace of a coal-fired power plant (Shawnee, Kentucky). Addi-
tional data on C02 and S02 stratification were obtained from the field
survey and the final field demonstration tasks, which were collected from
oil-fired power plants. These data are documented below. Evidence of gas
stratification was also presented on Page 44 of Wai den Research's final
report entitled "Improved Chemical Methods for Sampling and Analysis of
Gaseous Pollutants from the Combustion of Fossil Fuels.
4. CONCEPTUAL OCCURENCE OF GAS STRATIFICATION
As an aid in the investigation of the problem of conducting emission
measurements on process streams, some conceptual processes are presented which
could lead to gas stratification. These are presented not as locations of
assured stratification, but as possible locations where stratification can be
expected.
Various conceptual designs forecasting the probable occurrence of
gas stratification are shown in Figures 7, a,b,c, and d. Figure 7a shows
two different streams joining a common stream. The stratified plane can occur
downstream of the junction. In Figure 7b, two identical streams branch out
from the main then meet again in one stream. One branch has air in leakage
causing a stratified plant to occur downstream of the junction. Figure 7c
depicts the same condition as Figure 7b, except a different degree of air in
leakage occurs in the two identical streams. A stratified plane can occur
after the air in leakage location. Figure 7d shows the condition of a by-
pass branching that follows a different process than the main stream. In
this case the stratified plane can occur after the junction.
-------
PLflNE fl-fl WEST SIDE 3-IB -10
Figures (TVA) 46 feet x 24 feet
502 DISTRIBUTION - INT =50 PPM
-------
PLflNE fl-fl WEST SIDE
1311
U)
SQ2 DISTRIBUTION - INT.=50 PPM
Figure 4.
-------
PLflNE fl-fl WEST SIDE
<*>
en
502 DISTRIBUTION - INT.=50 PPM
Figure 5.
-------
PLflNE fl-fl WEST SIDE
CO
O»
S02 DISTRIBUTION
Figure 6.
- INT.=50 PPM
-------
(a)
Stratified Plane
Air i
Leakage
Stratified
Plane
(b)
Stratified Plane
Air in Leakage
(O
Figure 7. Conceptual Occurrence of Gas Stratification
Stratified Plane
37
-------
B. TASK II - ANALYTICAL ACTIVITIES AND LABORATORY EXPERIMENTS
1. ANALYTICAL ACTIVITIES
a. Analytical Simulation
The results of the analytical simulation sub-task are pre-
sented below. The purpose of this sub-task was to develop procedures for
emission measurements for flow of stratified gases in rectangular and
circular ducts.
a.l Background
(a) Representative Stack Sampling for Gases
The classical problem of stack sampling for gases Is to
associate the (average) samples taken generally at different times and
positions with either an average concentration or, with greater difficulty,
an average throughput. The collection of a gas sample which accurately rep-
resents the flow at a given point is far simpler than the complexities of
representative particulate sampling. Isokinetlc sampling rates are not a
requirement since the (thermal and turbulent) eddy d1ffus1v1t1es are far
too large to permit significant moelcular fractionation in the very weak
centrifugal fields imposed by any velocity mismatch attainable in sampling.
The precision and accuracy of the determination are a function of the
method(s) employed, the experimental design, and the variability of the
source. There is very little information extant on sampling methodologies
for continuously extracting a representative gas sample from a non-uniform
stream. There 1s, however, considerable background on sampling with manual
analytic methods.
38
-------
O>
o
01
Q.
.O
-------
Literature 1s scarce on sampling locations for obtaining
representative average concentrations of gases with ducts. Although this
subject 1s frequently discussed 1n the literature on partlculate sampling,
the discussions are concerned with Inertial separation of particles from
the gas stream. Since inertlal separation of pollutant gases from carrier
gases does not occur, the particulate sampling discussions are not generally
applicable (i.e., isokinetic sampling is not necessary and stratification
due to duct bends, etc., is not present).
It is a commonly held belief that gas stratification is
not present in ducts with turbulent gas flow29. However, eddy diffusion
studies show that a straight duct length of the order of 100 duct diameters
is required for good mixing of a highly stratified gas. It follows that in
many large-scale combustion systems where infiltration air is known to occur,
there is no direct location where the gas is well mixed. Luxl carried out a
total of 792 Orsat oxygen traverses in ten different ducts of fossil fuel
combination sources. He found (Figure 1) that stratification is generally
present and that single point samples are usually nonrepresentative in large
ducts.
32
ASME specifies a multi-point sampling system for Orsat
analysis of flue gases and a single point sampling system for S03 and S02, but
there 1s no explanation of this apparent inconsistency. The multi-point loca-
tions specified are also selected by ASME for velocity traverse. It may be
argued that velocity traverse schemes can give representative average velocities
in a duct where the velocity profile is not generally flat; hence the same
techniques should give representative values of emissions. However, until a
careful study is made, the accuracy of this technique cannot be established.
The emission of material from a combustion source is
described by the general equation:
40
-------
Ea « f Ca v • n dA (1)
a JA a
where Ea 1s the emission of material (a), C_ 1s the concentration of (a),
-»• a
v the flue gas velocity along the duct, A 1s the cross sectional area of the
duct, and n the unit vector normal to A. It follows from Equation (1) that
Ca and v are coupled if neither 1s constant across the duct; hence, they
a
should be measured together.
A traverse using a continuous oxygen or carbon dioxide
analyzer can quickly determine the extent of Infiltration air stratification.
If the gases are not significantly stratified, the pollutant gases may be
assumed to be well mixed and the concentration can be determined from a
single sampling point. However, it Is more prudent to sample at more than
one point as a check on the mixing of the pollutant gases.
For rigorous measurements, a set of replicate samples
should be taken at traverse points and a statistical analysis performed In
order to establish the flow pattern and an estimate of residual error. For
subsequent measurements in this duct, fewer sampling points can be used and
the accuracy predicted .
(b) Variance of Concentration in Large Ducts
A major problem 1n the high precision determination of
pollutant emissions 1s the variation 1n species concentration which may
exist 1n a large duct as a result of air Infiltration and poor mixing
(stratification). In the course of NAPCA's (now EPA) extensive studies of
coal-fired power plant effluents34, many C02 concentration profiles were
obtained by traverse of large ducts at different sampling locations. The
following discussion, based on a random selection of this test data and,
therefore, Incomplete, outlines the magnitude of the problem and the In-
fluence of both sampling location and equipment type on the results obtained.
41
-------
Typical sampling locations are Illustrated in the power
plant schematic (Figure 2). The most common locations are at the entrance
and exit of the duct collection equipment. The statistical method adopted
was to calculate the mean C0« concentration for each traverse plane, the
standard deviation from the mean, and the coefficient of variation (CV*=
100 a/mean). Two examples, Illustrating relatively homogeneous and strat-
ified flows, are given 1n Figure 3. The observed coefficient of variation
of the C02 concentration is given in Table 1 as a function of sampling
location. The relatively low value of CV*at the outlet of the dust collec-
tor, presumably the result of good mixing, suggests this to be the location
of choice for both simplicity and high precision. The relatively high value
of CV*at the inlet to the dust collector may be associated with air infil-
tration at the air preheater, which is a common occurrence.
The most frequently used sampling locations for emissions
determinations are the inlet and outlet of the dust collectors. The coefficient
of variation of the CCU concentration at these locations 1s given in Table 2
as a function of equipment type. With the exception of Plant No. 3, CV*at
the inlet is relatively large compared to the outlet values.
Conclusions which may be drawn from these data are:
C0« (or 0«) traverses are extremely valuable for selection of sampling locations
and determination of the number of samples required for a high precision
emissions determination.
(c) Flue Gas Flow Measurement Methods
Although this report 1s directed to the problems of gas
sampling, an understanding of velocity and total flow measurements is essential
to the development and analysis of representative sampling systems. The
discussion that follows summarizes the general methodology in practice today,
which 1s used to infer velocity profiles and total flow from a finite number
of point measurements.
* Coefficient of variation
42
-------
fly ash,
so2,so3
conbustlon
fuel I *•
C.M.J
1
pulverizer bottoms
infiltration
air
3
1 fuel
air
Infiltration air
1
C, H, S "
• 1 1
1
from
2
economizer
air
pre-
heat
j
J
i
CO
F
ion
nech.
ector
\y
\
s. ./
EC
.0.
DDt
/
l«
e
i
Stack
combustion air
bottom ash
collected ash
occasionally the E.S. precipitator may be upstream of the air-preheater
A Plant input data
j 1 j Flue gas sampling points
Figure 2. Schematic - Large Combustion Unit.
-------
'HOMOGENEOUS1
1.03 1.01 0.98 1.00 1.01
1.00 1.01 0.98 1.01 1.01
3"3"
i
Ave C02 - 11. 7%
CV*- 1.5*
"STRATIFIED"
4 .92 1.0 1.0 .97 .95
2 1.01 1.02 1.01 1.00 .90
5 1.05 .99 1.02 1.01 .90
2 1.00 .93 .93 1.01 .86
Uiin'1 . *-
t
4'8"
I
Ave C02 « 12.6
CV*- 9.3%
* Coefficient of variation
Random selection from six plants.
Figure 3. Normalized C02 Traverse Data at Dust Collector of Coal
Fired Power Plants.
44
-------
OBSERVED COEFFICIENT OF VARIATION FOR COg TRAVERSE
AT VARIOUS SAMPLING LOCATIONS
Sampling Position
it
Furnace (1)
Economizer Inlet (2)
Economizer Outlet (3)
(Dust Collector Inlet)
Outlet of Dust Collec-
tor (4)
T(F)
2400
860
360
350
No. of Traverse Points
12
12
8
8
24
24
18
18
co2 (cv), %
4.0 .» ,.
5.2 4*b
3.8 >3 8
4.8
1:1 >6-2
O « C > A O
32
(2) Sampling should be conducted at the outlet of dust collectors in the
absence of other information.
(3) For simplified methods, single point sampling at the dust collector
outlet appear to be feasible (Coefficient of Variation < 555).
See Figure 2.
45
-------
TABLE 2
OBSERVED COEFFICIENT OF VARIATION FOR C02 TRAVERSE
FOR VARIOUS COAL-FIRED PLANTS
Plant
No. Type of Boiler Firing -
1 Horizontally Opposed
2 Cyclone
3 Spreader Stoker
4 Corner
5 Vertical
C * cyclone
E • electrostatic precipitator
I * dust collector inlet
0 «. dust collector outlet
Dust Collection
Equipment
C
E
C
C,E
C,E
Sampling
Location
I
0
I
0
I
0
I
0
I
0
No. of
Traverse Points
24
12
24
24
18
9
18
12
24
12
co2(cv), *
9.3
2.3,1.4
4.6
3.2
1.5
1.02
8.8
0.97
7.1
3.2
-------
Total flow (Q) across a cross section of a duct can be
described by an equation similar to Equation (1) above, as an integral over
the area:
Q = f v • dA (2)
where v 1s the velocity normal to a differential area, dA.
Measuring gas stream velocities at various points across
a flue or duct is a velocity traverse. From the geometry of the ducts and
the geometry of the traverse point, the gas velocities may be used to calculate
the volumetric flow. This calculation 1s the evaluation of the Integral In
Equation (2). The integral may be evaluated graphically; however, the most
common technique 1s to divide the test section into a number of equal area
zones and determine the mean velocity in each zone. The velocities for each
zone are averaged and the volumetric flow is given by:
V * vA (3)
where V is the average velocity and A Is the area of the test section.
Discussion of techniques for dividing test sections into
equal area zones will be limited to ducts of circular and rectangular cross
sections, since sampling in ducts of other forms 1s rare. For ducts of unusual
shape, the volumetric flow should be determined by graphic integration for
accurate results.
The number of test zones into which a flue is divided will
depend upon the uniformity of velocity distribution and the accuracy desired, and
not upon the size of the duct, since for any two similar ducts (different only
in size) with similar velocity distributions, an equal number of velocity read-
ings will be required to determine the average velocities with equal accuracy.
However, in practice, the size of the pitot tube will limit the number of
velocity measurements in small ducts.
-------
The tangential method divides a duct of circular cross
section Into n equal zones, a circular central zone and (n-1) annular zones
(Figure 4). Each zone is divided into two equal area annular parts and the
velocity measurement is made at the radius of the boundary between the equal
area parts. The mathematical derivation of the division of a circular cross
35
section by this technique 1s found In Ower . The method of dividing a
circular duct by this technique is shown in Figure 4. The EPA sampling pro-
36
cedure uses this technique , as shown in Figure 5.
The log-linear method is an alternative which gives
higher accuracy. The circular cross section is again divided into equal
area annular zones, but the velocity is not arbitrarily measured at the
center of area of each zone. Instead, the measurement points are calculated
on the basis of an empirical analysis of the flow through circular pipes.
The development of this method, Including the determination of the measure-
ment poii
Table 3.
oc
ment points, may be found 1n Ower . The results are summarized in
37
For fully developed flow, Winternitz found that a
four point log-linear traverse gave an error of less than 0.556; whereas
the ten point tangential method overestimated the mean velocity by about U.
For nonfully developed flow, the ten point tangential technique was somewhat
better than the four point log-linear, but an eight point log-linear method
was superior to the ten point tangential method. The six point log-linear
method will give results with an error of less than 1% in flow distributions
as asymmetric as that shown 1n curve A of Figure 6.
Because the log-linear method provides better accuracy
than the tangential method for an equal number of measurements, 1t 1s rec-
ommended for velocity traverses in ducts of circular cross section. At the
present time, the log-linear method is in general use 1n the United
Kingdom33*38.
-------
Formula for determining location of point* >'n timihr duct
norti • r«eti/t*tu
of totaHon of
1r —
where r,= distance from center of duct to point p
will be in tame unitt a» R.
Example: Duct radius = R; 20 points total.
Distance to point 3 = r*.
_ /2*'(2-3-l)
*~\ *
H 09 0.707/1
cross section of circular duct
Figure 4. Tangential Method for Duct Division31
49
-------
Location of traverse points in circular stacks
fPercent of stack diameter frsi 1»;fde vail to traverse point)
Cross section Of circular stac'k divided into 12 equal
areas, shewing location of traverse points at centroid of each area.
8
o
O
" ** ^* " ^ ~
O
1 1
1
1
0 1 *
1
, i _.
r I
0 t 0
I
1
p r 1
1
o i e
i
e
o
O
Cross section of rectangular stack divided into 12 equal
areas, with traverse points at centrpid of each area.
Traverse
point
timber
on a
diameter
1
2
Number of traverse points on a diameter
2
11 fi
85.4
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
a
22
23
24
4
6.7
25.0
75.0
93.3
6
4.4
14.7
29.5
70.5
85.3
95.6
8
3.3
10.5
19.4
32.3
67.7
80.6
89.5
96.7
13
2.5
8.2
14.6
22.6
34.2
65.8
77.4
85.4
91.8
97.5
12 ! 14
2.1 1.8
6.7 5.7
11.8
17.7
25.0
9.9
14.6
20.1
35.5 '25.9
64.5
65.0
82.3
88.2
93.3
97.9
36.6
63.4
73.1
79.9
85.4
90.1
94.3
98.2
16
T.6
4.9
8.5
12.5
16.9
22.0
28.3
37.5
62.5
71.7
78.0
33.1
37.5
91.5
95.1.
93.4
18
4.4
7.5
10.9
14.6
18.3
23.6
29.6
38.2
61.8
70.4
76.4
31.2
20 | 22 1
M
3.9
6.7
9.7
12.9
16.5
20.4
25.0
30.6
38.8
61.2
69.4
75.0
85.4 79.6
S9.1
92.5
95.6
98.6
33.5
87.1
SO. 3
93.3
96.1
98.7
n
3.5
6.0
8.7
11.6
14.6
18.0
21.8
26.1
31.5
39.3
60.7
68.5
73.9
78.2
82.0
8S.4
83.4
91.3
94.0
96.5
98.9
24
1 1
3.2
5.5
7.9
10.5
13.2
16.1
19.4
23.0
27.2
32.3
39.8
60.2
67.7
72.8
77.0
80.6
83.9
86.8
89.S
92.1
94.5
96.8
98.9
Figure 5. EPA Sampling Point Locations.
-------
TABLE 3
LOCATION OF MEASURING POINTS FOR LOG-LINEAR METHOD 3?
No. of Measuring
Points Per Diameter Distance from Wall in Pipe Diameters
4 0-043, 0-290, 0-710, 0-957
6. 0-032, 0-135, 0-321, 0-679, 0-865,
0-968
8 0-021, 0-117, 0-184, 0-345
0-655, 0-816, 0-883, 0-979
10 0-019, 0-076, 0-1.53, 0-217, 0-361
0*639, 0-783, 0-847, 0-924, 0-981
51
-------
djLi o-« 0-6^ o-e
Rotio of well <*»lonc« to dtonttMr, i/
-------
The technique used for dividing a rectangular duct
Into equal area zones 1s to divide the section Into a number of geo-
metrically similar rectangular zones and to measure the velocity at the
centroid of each zone. The rules regarding the number of zones are more
arbitrary than for circular ducts. As 1n the case of circular ducts, the
accuracy of the traverse will depend on the uniformity of the flow and the
number of velocity measurements. However, it is a convention to increase
the number of sampling points with duct size. The recommended number of
test points in two different procedures are given in Table 4 as a function
of the cross section area of the duct. The EPA method contains a similar
formulation .
39
British Standards 1042 recommends a division into at
least 16 zones, with five velocity measurements in each corner zone, and
three velocity measurements on each wall zone. The velocity 1n each zone 1s
averaged before averaging the velocities over the duct. No zone should be
•)
greater than 36 inches , which means that more than 16 zones are necessary
for ducts with cross
1s shown in Figure 7.
2
for ducts with cross sectional areas over 4 feet . The BSI Traverse Plan
The National Engineering Laboratory (U.K.) has found that
errors of 2% or more can occur for certain types of asymmetric flow distribu-
07
tlons when the 16 part, 48 point traverse is used .
A variety of sample point location methods based on
42
other mathematical formulae have been presented by ASME . A table of these
methods is shown in Table 5. Figure 8 gives a comparison of these methods
when applied to some hypothetical flow profiles.
(d) Summary
The main point of this discussion relevant to the de-
velopment of procedures to obtain representative samples for gas flow 1s
the Intrinsic coupling of velocity and concentration in the area integral
(Equation 1) describing the total emission In a general non-uniform system.
If we are to obtain a representative value for concentration, it must be
53
-------
TABLE 4
TEST POINTS FOR RECTANGULAR DUCTS
A. Haaland40
Cross Section Area
Square Feet
Less than 2
2 to 25
Greater than 25
Number of Test
4
12
20 or
Points
more
B. ASTM^l
Inside Cross Sectional
Area of Flue, ft2
Minimum Number of Test Points
r to
2< to 12
4
6-24
More than 24
54
-------
-t- i-
-4- +
B/24 \
Arrangement of pftot tub?
positions for each corner
panel of the airway
Figure 7. Traverse Plan for Rectangular Duct
38
55
-------
TABLE 5.
STATION LOCATIONS AND WEIGHTS POM AVERAGING
Averaa Inf for linear Interval 0 £ x < 1
Averaging In • circular duct, In Interval 0 £ r 1 1
:ft..i>a>
OF
STATIONS
n
2
3
L
S
6
7
8
9
10
n
METHOD
CEITROIU or luvA A,6
.935li
0.3162
.51.77
.7071
.83*7
.9187
0.2687
.5000
.«»55
.7638
.B(VO
.9571
0.2673
.li',29
.5976
.7071
.8018
,086Ii
.9636
0.2500
.I.331
.5590
.*«11.
.7500
.8292
.9011i
.9682
0.2357
.Ii062
.5270
.6236
.7071
.7817
.81498
.9129
.9718
0.2S36
.3873
.5100
.W16
.6701
.7M6
,110/2
,w*i
.vzro
.?71.7
r
W
i/*
1/3
iA
1/5
1/6
1/7
1/8
1/9
1/10
V
KMTImll. W tU'AL AhFX.
NIWTON-COTt3
X
(b)
0
1
7
.1667
W
0.1250
.3750
.3750
.1250
0.0778
.3556
.1333
.3556
.0778
0.0660
.2601.
.1736
.1736
.2601,
.06W
0.01.83
.2571
.0321
.3238
.0321
.2571
.01.88
o.3ii35
.2070
.0766
.1730
.1730
.0766
.2070
.Oli35
0.03li9
.2077
-.0327
.3702
-.1601
.3702
-.0327
.2077
.031.9
0.0319
.1757
.0121
.2159
.O6ii5
.061$
.2159
.0121
.1757
.0319
W
NFWTOti-iOTES
CHtBTSHCT
•J.
X
(•)
0.2113
.7887
0.1U&
.$000
.8536
0.1027
.Ii072
.5928
.8973
0.0838
.3127
.Sooo
.6873
.9162
0.0669
.2887
.3667
.6333
.7113
.9331
0.0581
.2352
.3381
.Sooo
.6619
.7614
.9U9
(f)
0.0513
.2036
.2961.
.1)1.87
.5513
.7036
.7961!
.9li87
O.oJJd
.1995
.2356
.U160
.5000
.58UO
.76U.
.8005
.9558
(f)
O.OU9
.1561.
.2500
.3'i36
.iiSSl
.51.19
.6561,
.7500
.8Ji36
.9581
X
r
(•)
O.LS97
.8881
0.3827
.7071
.9239
0.3203
.6382
.7699
.91.73
0.2891
.5592
.7071
.8290
.9572
0.3586
.5373
.6057
.7958
.81.31
.9660
0.2140
Ml>9
.5811
.7071
.8136
.871.5
.9705
U)
0.??66
.1.513
.51M
.6698
.7L25
.8388
.8921.
.9710
0.2103
.U66
.1.8514
.61.50
.7071
.761*
.871.3
.891(7
.9776
(E)
0.201.6
.3951.
.5000
.5862
.6768
.7361
.8102
.86fV)
.9185
.9788
r
V
(•)
1/2
V3
1A
1/5
V6
1/7
(O
1/8
V9
(s)
1/10
W
CHUTSHEF
X
(t)
0.2113
.7887
0.1127
.5000
.8873
0.0691.
.3300
.6700
.9306
0.01.69
.2308
.5000
.7692
.9531
(•) All >
aqua]
(b) Tr.pt
(c) Ptr»t
(StMf
(d) Three
(a) 0.2 .
(t ) Tiro I
inter
(t) TwS
inter
OAUSS
r
d)
O.t597
.8881
0.3357
.7071
.91.20
0.2635
.S7W
.8185
.96i7
0.2166
.UK*
.7071
.8771
.9763
V
(•)
1/2
0.2778
.1.1.1*
.2778
0.1739
.3261
.3261
.1739
0.1185
.2393
.2fll.lt
.2393
.11S5
MOTE3
i««iur*w»rrti of
L weight
letdil rule
wile rule
>«on'» rule)
-ei£hthi rule
0.8 rule
-etatlon
•*•!•
-•tit ten
T«H
56
-------
6
.6
RADIUS, r
(o) VELOCITY DISTRIBUTIONS
1.0
CENTROID OF EQUAL AREAS
NEWTON-COTES
CHEBYSHEF
GAUSS
DISTRIBUTION
I
-0.26
-000,
-000,
-000,
II
•H.I6
-020
+002
-000,
III
40.77
-008
+001
-OOOj
(b) ERROR, %. IN 4-POINT APPROXIMATION
ERROR IN 4-POINT AVERAGING OF SOME
ARBITRARY AXIALLY-SYMMETRIC VELOCITY
DISTRIBUTIONS.
Figure 8.
57
-------
weighted by the velocity at the sampling point, unless either the velocity
or the concentration is uniform or can be made uniform by the sampling
system.
It is interesting to note that the conditions which
often lead to stratification of particulate matter are often the most
favorable for gas sampling. Components such as fans impart high shears and
mix the gases passing through, tending to reduce stratification and non-
uniform profiles. Particles which are subject to non-uniform inertia forces
in the same processes are, in contrast, often segregated and consequently,
large concentration gradients develop.
a. 2 Simulation
The objective of the analytical simulation was to develop
a measure of accuracy for the extraction of representative samples for
several measurement strategies as a function of hypothetical and actual
flow and concentration profiles, and specific system parameters, such as
number and location of sampling points. Based on these results, we have
attempted to develop a set of statistics that could be used to establish
generalized sampling strategies.
To predict concentration and velocity profiles in ducts
would require enormous effort, considering that the errors generated by
the computational techniques and the unknown initial condition at the up-
stream end of the duct system would render the model virtually useless in
predicting stack gas effluent rates. To apply this approach, some of the
problems that would have to be overcome are the following:
(a) Numerical solution methods must be developed for the
simultaneous solution of the three-dimensional forms of:
(1) the continuity equation
(2) the equations of motion (Navier-Stokes Equations)
(3) energy equation
(4) equations of state for the fluid phase
(5) viscosity dependence on temperature
58
-------
(b) The geometric configuration of the stack and the con-
ditions existing at the duct walls must be specified.
(c) The Initial condition 1n the fire box must be specified.
The boundary conditions on the duct walls depend, in general, on prevailing
weather conditions such as wind and humidity. The initial condition 1n the
fire box depends on the reaction rate in coal burning, and a variety of other
conditions which affect coal combustion. Since these effects are not gen-
erally known, 1t would be futile to attempt to simulate the flow, heat and
mass transport through the network of ducts and equipment leading to the stack.
Therefore, in this study we adopted an approach such that one
specifies duct geometry, actual concentration and velocity profiles in the
duct, and the number of probes and location and area associated with each
probe. The model, which has been programmed for digital solution, will then
compute the actual effluent rates, viz.:
JA CVdA JA VdA
t
/A "A /A dA
The error in the measured and actual effluent rate 1s then calculated as:
CVdA '
c - -A
I/A
f CVdA
JA
E2
|/A
JA
VdA
VdA
59
-------
By calculating a series of such cases, graphs have been pro-
duced showing the error 1n stack gas measurement (I.e., average velocity
effluent emission rate) as a function of number of probes, probe location,
strategy, and specified concentration and velocity profiles. These graphs
serve as guidelines for determining the best number, location, and strategy
for placing probes in a specific duct if some a priori estimate for measure-
ment (survey) of the duct profiles 1s available.
a.3 Simulation Results
Sixteen fictitious and actual stratification profile sets
were Identified and prepared for sampling strategy evaluation. (See
Appendix G for figures 9-a through 21-b and Table 6.) These cases Included
the analytic functions:
CASE,
u(x,y)
cU.J
1 -
for a rectangular duct where a and b are ducts dimensions and Q 1s the
volumetric flow rate; and
CASE II
U(r,6) • -Sfc [l - f]
?ra L J
c(r,8) - C0 s1n6
for a circular duct of radius a.
60
-------
Case III (Figures 9-a and 9-b) consists of normalized
velocity and concentration (CO^) profiles derived from exhaust emission
measurements of a T-53 aircraft gas turbine combustor at Idle setting
(circular duct). Case IV (Figures 10-a and 10-b) consists of velocity and
temperature profiles taken from an EPA test of a coal fired boiler sampled
at a precipltator inlet (rectangular duct). For test purposes, the tem-
perature profile in degrees Fahrenheit has been taken as the concentration
profile for the duct. Case V, also a rectangular duct and shown in Figures
11-a and 11-b, illustrates actual velocity and S02 stratification immediately
downstream of a coal-fired TVA boiler. Cases VI - IX, shown in Figures 12-a
and 12-b , 13-a and 13-b, 14-a and 14-b and 15-a and 15-b, are hypothetical
concentration and velocity profiles for circular ducts as defined by Walden
staff. Cases X and XI, shown in Figures 16-a and 16-b and 17-a and 17-b,
respectively, are also hypothetical profiles but for rectangular ducts.
Cases XIII-1 to XIII-4, shown in Figures 18-a,b to 21-a,b are actual con-
centration velocity data obtained from a TVA duct immediately downstream
of a coal-fired power plant furnace. Case XII, shown in Table 6, 1s for
an experimental wind tunnel test conducted at Maiden.
Strategies selected for testing of the rectangular duct
cases Include: equal area (1, 9, 16, 49, 100 probes); British Standard 1042
(48 probes); Newton Cotes* (9, 16, 49, 100 probes); Gauss* (9, 16 probes);
Chebyshef (9, 16, 49, 100 probes); equal area with square area segments (50
probes); and an equal area circular ring analog (8, 16, 24 probes) defined
by the sketch below (4 zones, 16 probes).
* points weighed equally not. appropriate weighting factors
61
-------
This latter procedure 1s applicable when two ports are located
perpendicular to one another at the center points of adjacent sides in a
rectangular duct. Traverse positions for the zones are listed In Table 7.
Note that probes located according to the Chebyshef procedure for zones of
this type were analyzed for Case XII.
Strategies selected for testing of the circular duct cases
Include: the centreid of the duct (1 probe); tangential (2, 4, 6 points per
diameter; 2 diameters* and 4 diameters**); and log linear (4, 6, 8 points
per diameter; 2 diameters* and 4 diameters**). In addition, both the log-
linear and tangential strategies have been investigated with orthogonal
probes, containing 6 probes per diameter with the diameters rotated so that
additional probe sampling locations of 30° and 120°; 45° and 135°; and 60° -
and 150° can be simulated.
Test results for the 16 sample cases are shown in Tables 8
and 9-1 thru 9-15 and are summarized in Tables 10-1 thru 11-4, Figures 22 thru
29, Tables 12-1 thruJ2-5 and Figures 30 thru 33. In Tables 10-1 thru 10-7
the left most entry for the average error in emission or Velocity has been
* Probes oriented at 0° and 90°
** Probes oriented at 9°, 4$°, 90°, and 135°
62
-------
TABLE 7
VELOCITY TRAVERSE POINTS IN RECTANGULAR DUCTS
WITH PERPENDICULAR PORTS
PERCENT OF DISTANCE ACROSS DUCT
Traverse
Point 2
1 7.3
2 17.7
3 82.3
4 92.7
5
6
7
8
9
10
11
12
Number
3
4.6
15.2
35.6
64.4
84.8
95.4
--
—
--
« •»
--
.-
of Zones
4
3.4
10.7
19.8
36.0
64.0
80.2
89.3
96.6
—
«
--
_-
Across Duct
5
2.6
8.3
14.8
23.1
38.8
61.2
76.9
85.2
91.7
97.4
...
--
6
2.2
6.8
12.0
17.8
25.4
39.8
60.2
74,6
82.2
88.0
93.2
97.8
63
-------
TABLE 8
TEST RESULTS FOR A RECTANGULAR DUCT (5' x 10')
WITH VELOCITY AND CONCENTRATION PROFILES OF THE FORM;
Probes
50
I Probes
9
16
49
100
C(x,y) = C
1 + — 11 +
Strategy: Equal Area with Square Equal Area Segments
Error (e) Computed by Program
'emission
+ 2.52%
'ave. velocity
+ 2,52%
Strategy: Newton-Cotes*
Error (e) Computed by Program
eenvission
- 74.9%
- 55.5%
- 30.6%
- 21.0*
'ave. velocity
- 75.0%
- 55.6%
- 30.6%
- 21.0%
Error (e) Hand Calculated
£emission £ave. velocity
n.c.** + 2.5%
Error (e) Hand Calculated
'emission
75.0%
55.5%
n.c.
n.c.
'ave. velocity
- 75.0%
- 55.5%
- 30.5%
- 21.0%
-------
TABLE 8 (continued)
Strategy: Chebyshef
I Probes
Error (e) Computed by Program
"emission
'ave. velocity
Error (c) Hand Calculated
9
16
49
100
- .012%
+ .124%
+ .020%
+ .007%
- .013%
+ .124%
+ .021%
+ .008%
WIN 1 J J 1 WII
0%
n.c.
n.c.
n.c.
ave. vc i ui
- .034%
+ .117%
+ .013%
+ .0004%
Strategy: British Standards 1042
Probes
48
Error (e) Computed by Program
emission
+ 3.85%
'ave. velocity
+ 3.85%
Error (e) Hand Calculated
'emission
n.c.
'ave. velocity
3.97%
Strategy: Gauss*
I Probes
9
16
Error (e) Computed by Program
'emission
- 19.0%
- 26.5%
'ave. velocity
- 19.0%
- 26.5%
Error
'emission
- 20.4%
n.c.
Hand Calcul ated
"ave. velocity
- 19.0%
- 26.6?
-------
TABLE 8 (continued)
Strategy: Equal Area
Probes
Error (e) Computed by Program
"emission
ave. velocity
1
9
16
49
100
+ 124.9%
+ 11.4%
+ 6.36%
+ 2.04%
+ 1.00%
+ 125.0%
* 11.4%
+ 6.36%
+ 2.04%
+ 1.00%
Error (e) Hand Calculated
'emission
n.c.
n.c.
n.c.
n.c.
n.c.
'ave. velocity
125.0%
11.4%
6.34%
2.03%
0.99%
Strategy: Circular Analog-Equal Area
Probes
8
16
24
Error (e) Computed by Program
emission
- 3.98%
+ 16.5%
+ 14.9%
'ave. velocity
- 3.98%
+ 16.5%
+ 14.9%
Error (e) Hand Calculated
"emission
n.c.
n.c.
n.c.
'ave. velocity
- 4.0%
+ 16.5%
+ 14.9%
* Points weighted equally not by appropriate weighting factors
** n.c. means not calculated
-------
Test Results for Case II
Strategy: Centrold
Probes
1
Error (e) Computed by Program
eemlsslon eave. velocity
- 51.3 % + 99.4 %
Strategy: Log Linear
Error (c^ Computed by Program
8 Probes
4 probes/61aneter~2 diameters*
4 probes/d1a*eter-4 diameters**
6 probes/41ameter-2 dlamews*
6 probes/d1ameter-4 diameters**
8 probes/dlameter-2 diameters*
8 probes/d1ameter-4 diameters**
6 probes/d1aroeter-2 diameters(30° and 120°)
6 probes/d1ameter-2 diameters(45° and 135°)
6 probes/diameter-2 diameters(60* and ISO9}
eenrisslon
- 23.1 at
- 11.1 %
- 18.4 *
- 7.9 %
- 15.9 %
- 5.2 %
+ 3.8 %
+ 2.5 %
- 21 .0 %.
eave. velocity
- 1.19*
- 1.20*
- 2.5 %
- 2.5 *
- .001*
- .0043!
- 2.5 %
- 2.5 %
- 2.5 X.
Strategy: Tangential
Error (e) Computed by Program
I Probes
2 probes/dlameter-2 diameter*
2 probes/dlameter-4 diameter?**
4 probes/dlameter-2 diameters*
4 probes/d1ameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/diameter-4 diameters**
6 probes/dlameter-2 diameters(30° and 120°)
6 probes/41 arneter-2 diameters(45° and 135°)
6 probes/d1amster-2 diameters(60° and 150°)
* 0° and 90°
** 0°, 45°, 90°, 135°
67
eemlsston
+ 26.1 %
+ 31.3 %
- 1.24*
+ 8.2 %
- 7.2 %
+ 2.8 %
+ 15.5 %
+ 12.8 *
- 15.2 %
eave. velocity
- 0.22%
- 0.22%
+ 0.005*
*- 0.005*
+ '0.119*
+ 0.117*
+ 0.116*
* 0.115*
+ 0.115*
-------
TABLE 9-2
Test Results for Case III
Strategy; Centrold
I Probes
/I
Strategy; Loo Linear
I Probes
4 probes/dlameter-2
4 probes/dlameter-4
6 probes/dlameter-2
6 probes/dlameter-4
8 probes/dlameter-2
8 probes/dlameter-4
6 probes/dlameter-2
6 probes/dlameter-2
6 probes/dlameter-2
diameters*
diameters**
dlamerers*
diameters**
diameters*
diameters**
diameters(30° and 120°)
d1ameters(45° and 135°)
diameters(60° and 150°)
Error (e) Computed by Program
remission cave, velocity
+ 8.6 % - 10.0 %
Error (e) Computed by Program
eemlsslon
-
•
m
-
•f
•f
+
•f
-
1.60X
0.91%
1.621
.74X
.145X
1.08X
.48X
.13755
1.17X
eave. velocity
- 1.98%
- 1.85X
• 2.041
• 1.9 X
- .191X
.059?
- 1.77X
- 1.78X
- 1.81X
Strategy: Tangential
I Probes
2 probes/dlameter-2 diameter*
2 probes/d1ameter-4 diameters**
4 probes/dlameter-2 diameters*
4 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters(30° and 120°)
6 probes/dlameter-2 diameters(45° and 135°)
6 probes/dlameter-2 d1ameters(60° and 150*)
Error (c) Computed by Program
* 0° and 90°
0°. 45°, 90°, 135°
68
eemlsslon
+ 10.1 X
+ 12.0 X
+ 4.2 X
+ 5.4 X
+ 2.7 X
+ 3.8 X
+ 5.2 X
+ 4.8 X
+ 3.5 X
cave, velocity
+ 10.5 X
+ 10.7 X
+ 4.2 X
+ 4.4 X
+ 2.5 X
+ 2.7 X
+ 2.9 X
* 2.9 X
+ 2.8 X
-------
TABLE 9-3
Test Results for Case IV
Strategy: British Standards 1042
# Probes
48
Error (e) Computed by Program
eemlsslon save, velocity
+ 5.4% + 5.4 %
Strategy: Equal Area
Error (e) Computed by Program
# Probes
1
9
16
49
eemlsslon
+ 37.7 %
+ 8.4 %
+ 13.4 %
+ 7.7 %
eave. velocity
+ 36.0 %
+ 8.2 *
+ 13.4 %
+ 7.8 %
Strategy: Circular Analog-Equal Area
t Probes
8
16
24
Error (e) Computed by Program
eemlsslon
- 2.8 *
- 2.6 %
- 3.6 %
eave. velocity
- 2.9 %
- 3.0 %
- 4.0 %
Strategy: Newton-Cotes
'i Probes
9
16
49
Error (e) Computed by Program
eemlsslon
- 84.2 %
- 70.9 %
-41.5 %
eave. velocity
- 84.3 %
- 70.9 %
-41.6 %
Points weighted equally not by appropriate weighting factors
69
-------
TABLE 9-3(cont.)
*
Strategy: Gauss
Error (e) Computed by Program
# Probes eemission eave. velocity
9 + 7.2 % + 7.1 *
16 + 3.3 % + 3.9 «..
Strategy: Chebyshef
Error (e) Computed by Program
# Probes eemission eave. velocity
9 + 8.0 % + 7.8 *
16 + 11.4 % + 11.2 X
49 + 4.0 % + 4.1 %
*
Points weighted equally not by appropriate weighting factors
70
-------
TABLE 9-4
Test Results for Case V
Strategy: British Standards 1042
# Probes
48
Error (e) Computed by Program
eemission eave. vel ocHy
+ 1.15% + 3.54%
Strategy: Equal Area
# Probes
1
9
16
49
Error (e) Computed by Program
eemission
+ 3.13%
+ 21.8 %
+ 11.4 %
+ 6.0 %
eave. velocity
+ 7.44%
+ 22.9 %
+ 14.7 %
+ 5.3*%
Strategy: Circular Analog-Equal Area
# Probes
8
16
24
Error (e) Computed by Program
eemission
.72%
- 7.8 %
- 8.4 %
eave. velocity
+ 7.5 %
+ 0.81%
.088%
Strategy: Newton-Cotes
§ Probes
9
16
49
Error (e) Computed by Program
eemission eave. velocity
- 88.1 % - 87.3 %
- 66.1 % - 68.1 %
- 37.9 % - 36.4 %
Points weighted equally not by appropriate weighting factors
71
-------
TABLE 9-4(cont.)
Strategy: Gauss
Error (e) Computed by Program
# Probes eemission eave. velocity
9 - .685% + 6.1 %
16 + 2.2 % - 7.0 %
Strategy: Chebyshef
Error (e) Computed by Program
# Probes eemission eave. velocity
9+ 16.9 % + 18.5 %
16 + .63% + 5.3 %
49 + 6.3 % + .012%
* Points weighted equally not by appropriate weighting factors
-------
TABLE 9-5
Test Results for Case VI
Strategy; CentreId
I Probes
1
Strategy; Log Linear
I Probes
4 probes/dlameter-2
4 probes/d1ameter-4
6 probes/d1ameter-2
6 probes/dlameter-4
8 probes/d1ameter-2
8 probes/dlameter-4
6 probes/d1ameter-2
6 probes/dlameter-2
6 probes/dlameter-2
diameters*
diameters**
dlamerers*
diameters**
diameters*
diameters**
d1ameters(30e and 120°)
d1ameters(45° and 135°)
d1ameters(60° and 150°)
Strategy: Tangential
I Probes
2 probes/dlameter-2 diameter*
2 probes/dlameter-4 diameters**
4 probes/dlameter-2 diameters*
4 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/dfameter-4 diameters**
6 probes/dlameter-2 d1ameters(30° and 120°)
6 probes/dlameter-2 diameters(45° and 135°)
6 probes/dlameter-2 diameters(60* and ISO9)
Error (e) Computed by Program
eemlsslon eave. velocity
+ 115.5 % + 68.6 %
Error (e) Computed by Program
eemlsslon
-
-
m
-
-
-
-
-
+
24
18
15
9
9
2
4
2
1
.5
.3
.9
.1
.5
.3
.7
.3
X
X
X
X
X
X
X
X
.66X
eave.
-
-
-
*»
-
m
+
+
+
13.
8.
9.
4.
6.
1.
»
1.
4.
velocity
2 X
OX
3 X
1 X
9 *
66X
29X
m
6 X
Error (e) Computed by Program
eemlsslon
+ 24.9 X
+ 35.4 X
- 6.3 X
+ 1.68X
- 7.9 X
.27X
+ 4.7 X
+ 7.4 X
+ 11.8 X
eave. velocity
+ 14.9 X
+ 21.5 X
- 1.14X
+ 4.7 X
- 3.5 X
+ 2.1 X
+ 6.9 X
+ 7.7 X
* 11.7 X
0* and 90°
** r. 45% 90°, 135°
73
-------
TABLE 9-6
Test Results for Case VII
Strategy; CentreId
I Probes
1
Error (e) Computed by Program
eemlsslon eave. velocity
+ 129.6 % 4 68.6 X
Strategy: Log Linear
Error (e) Computed by Program
I Probes
4 probes/dlameter-2 diameters*
4 probes/d1ameter-4 diameters**
6 probes/d1ameter~2 dlamerers*
6 probes/d1amtter-4 diameters**
8 probes/dlameter-2 diameters*
8 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters(30° and 120°)
6 probes/dlameter-2 diameters(45° and 135%)
6 probes/dlameter-2 diameters(60* and 150°)
eemlsslon
- 11.1 X
- 6.6 X
- 9.9 X
- 5.5 X
- 5.8 X
- 1.26X
- 2.7 X
- 1.10X
+ 4.3 X
eave. velocity
- 13.2 X
- 8.0 X
- 9.3 X
- 4.1 X
- 6.9 X
- 1.66X
+ 0.29X
+ 1.11*
+ 4.6 X
Strategy; Tangential
Error (c) Computed by Program
I Probes
2 probes/dlameter-2 diameter*
2 probes/d1ameter-4 diameters**
4 probes/dlameter-2 diameters*
4 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/dlameter-4 diameters**
6 probes/dlameter-2 d1ameter,s(30° and 120°)
6 probes/dlameter-2 diameters(45° and 135°)
6 probes/dlameter-2 diameters(60° and ISO*)
* 0* and 90e
** 0°. 45°, 90°, 135°
eemlsslon
+ 10.1 X
* 15.2 X
- 1.82X
+ 3.1 X
- 3.7 X
+ 1.11X
+ 4.1 X
+ 5.9 X
* 11.8 %
eave. velocity
* 14.9 X
* 21.5 X
- 1.14X
+ 4.7 X
- 3.5 X
+ 2.1 X
+ 6.9 X
+• 7.7 X
+ 11.7 X
74
-------
TABLE 9*
Test Results for Case VIII
Strategy: Centrold
I Probes
1
Strategy: log linear
I Probes
4 probes/dlameter-2
4 probes/d1ameter-4
6 probes/dtameter-2
6 probes/dlameter-4
8 probes/dlameter-2
8 probes/dlameter-4
6 probes/dlameter-2
6 probes/dlameter-2
6 probes/dtamet«r-2
diameters*
diameters**
dfamerers*
diameters**
diameters*
diameters**
diameters(30° and 120"}
diameters(45* and 135*)
diameters(60* and 150°)
Error (e) Computed by Program
eemlsslon cave, velocity
+ 1078.7 X • + 429 %
Error (e) Computed b Proram
eemlsslon
- 13,0 X
+ 12.1 X
- 23.0 X
- 1.24X
- 19.4 X
+ 2.7 X
+ 17.4 X
+ 20.5 X
+ 16.7 X
cave, velocity
- 22.5 X
- 8.5 X
- 18.8 X
- 7.8 X
- 15.4 X
- 4.8 X
+ 1.95X
* 3.2 X
- 6.9 X
Strategy; Tangential
I Probes
2 probes/diameter
2 probes/diameter
4 probes/diameter
4 probes/diameter
6 probes/diameter
6 probes/diameter
6 probes/diameter
6 probes/diameter
6 probes/dtameter-
•2 diameter*
•4 diameters**
•2 diameters*
•4 dlamettrs**
2 diameters*
•4 diameters**
2 diameters(30» and 120*)
2 dlmtarkfcS* and 135*)
d1ameter*'(60* and 150*)
Error (e) Computed by Program
eemlsslon eave. velocity
- 36.4 X + 2.6 X
- 19.6 X 4 5.7 X
- 19.3 X - 12.1 X
+ 3.5 X - 1.92X
- 20.1 X - 13.8 X
+ 2.4 X - 3.2 X
+ 22.4 X + 6.5 X
+ 25.0 X +• 7.3 t
* 23.5 X - 3.5 X
* 0° and 90"
0',
45°,
90'.
135*
75
-------
TABLE 9-8
Test Results for Case IX
Strategy; CentreId
I Probes
-1
Strategy; Log LInear
I Probes
4 probes/dlameter-2 diameters*
4 probes/d1ameter-4 diameters**
6 probes/dlameter-2 dlamerers*
6 probes/d1ameter-4 diameters**
8 probes/dlameter-2 diameters*
8 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters(30° and 120°}
6 probes/d1ameter-2 d1ameters(45° and 135°)
6 probes/dlameter-2 diameters(60° and ISO*)
Error (e) Computed by Program
eemlsslon eave. velocity
•*76.5 % +429.2 X
Error (e) Computed by Program
eemlsslon
- 18.3 %
- 2.1 X
- 20.7 %
- 7.9 X
- 18.8 %
- 6.5 X
+ l.OOX
+ 4.8 X
* .52X
cave, velocity
- 22.5 X
• 8.5 X
- 18.8 X
- 7.8 X
- 15.4 X
- 4.8 X
+ 1.95X
+ 3.2 X
- $.9 X
Strategy: Tangential
Error (c) Computed by Program
I Probes
2 probes/dlameter-2 diameter*
2 probes/dlameter-4 diameters**
4 probes/dlameter-2 diameters*
4 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/dlameter-4 diameters**
6 probes/dlameter-2 diameters(30* and 120°)
6 probes/dlameter-2 diameters(45° and 135°)
6 probes/dlameter-2 diameters(60° and 150°)
eemlsslon
- 19.7 X
- 15.2 X
- 17.2 X
-. 4-.7 X
- 18.0 X
- 5.5 X
+ 3.5 X
+ 7.0 X
+ 2.6 X
eave. velocity
+ 2.6 X
•* 5.7 X
- 12.1 X
- 1.92X
- 13.8 X
- 3.2 X
+ 6.5 X
+' 7.3 X
- 3.5 X
0° and 90°
** 0°, 45% 90°, 135e
76
-------
TABLE 9-9
Test Results for Case X
Strategy: British Standards 1042
# Probes
48
Error (e) Computed by Program
cemission eave. velocity
- .40% - .42%
Strategy: Equal Area
# Probes
1
9
16
49
Error (e) Computed by Program
eemission eave. velocity
+140.7 %
+ 14.3 %
+ 6.2 %
+ 1.33%
+108.8 %
+ 12.9 %
+ 7.2 %
+ 2.3 %
Strategy: Circular Analog-Equal Area
I-Probes
8
16
24
Error (e) Computed by Program
eemisslon save, velocity
- 3.1 % - 2.3 %
+ 19.8 % + 14.4 %
+ 18.4 % + 12.9 %
Strategy: Newton-Cotes
# Probes
9
16
49
Error (e) Computed by Program
eemisslon
- 73.3 %
- 54.6 %
- 29.0 %
eave. velocity
- 77.3 %
- 57.2 %
- 31.2 %
Points weighted equally not by appropriate weighting factors
77
-------
TABLE 9-9(cont.)
*
Strategy: Gauss
Error (e) Computed by Program
Probes eemission eave. velocity
9 - 19.2 % - 17.0 %
16 - 53.3 % - 43.1 %
Strategy; Chebyshef
Error (e) Computed by Program
# Probes eemtssion save, velocity
9 + 1.80% + 1.84%
16 + .190% + 0.101%
49 - 1.32% + .019%
Points weighted equally not by appropriate weighting factors
78
-------
TABLE 9-10
Test Results for Case xi
Strategy: British Standards 1042
# Probes
48
Error (e) Computed by Program
remission eave. velocity
+ 0.20% + 0.24%
Strategy: Equal Area
if Probes
1
9
16
49
Error (e) Computed by Program
semission eave. velocity
+ 34.5 %
+ 17,9 %
+ 55.6 %
+ 20.7 %
+ 10.3 %
+ 2.3 %
+ 10.2 %
+ 3.4 %
Strategy; Circular Analog-Equal Area
# Probes
8
16
24
Error (e) Computed by Program
eemlssion eave. velocity
+ 9.7 % + 8.4 %
'+ 12.6 % + 7.9 %
+ 11.7 % + 6.9 %
Strategy: Newton-Cotes
# Probes
9
16
49
Error (c) Computed by Program
eemission
- 79.0 %
- 58.4 %
- 29.6 %
eave. velocity
- 82.3 %
- 60.9 %
- 32.2 %
Points weighted equally not by appropriate weighting factors
79
-------
TABLE 9-10 (cont.)
Strategy: Gauss*
Error (e) Computed by Program
# Probes eemission eave. velocity
9 - 10.0 % - 8.43%
16 - 35.8 % - 25.8 %
Strategy: Chebyshef
Error (e) Computed by Program
# Probes eemisslon save, velocity
9 + 10.4 % + 9.1 %
16 0.074% 0.005%
49 - 1.8 % 0.014%
* Points weighted equally not by appropriate weighting factors
80
-------
TABLE 9-11
Test Results for Case XII
Strategy: British Standards 1042
Error (E) Computed by Program
$ Probes eemission cave, velocity
48
Strategy: Equal Area
Error (e) Computed by Program
$ Probes eemission eave. velocity
1 + 172.0 % + 27A %
9 + 58.9 % + 15.5 X
16 + 68.6 % + 16.1 %
25 + 64.5% + 11. H%
Strategy: Circular Analog-Equal Area
Error (e) Computed by Program
f Probes eemission eqve. velocity
8 + 78.4 % -f 16.7 %
16 + 113.0 % + 8.2 %
Strategy: Circular Analog-Equal.Area-Chebyshef Locations
Error (e) Computed by Program
# Probes eemission eave. velocity
e """ +100.0 % + 24.2 %
16 + 123.0 % + 13.8X
81
-------
TABLE 9-11 (cont.)
Strategy: Chebyshef
Error (e) Computed by Program
# Probes eemisslon eave. velocity
9 + 59.6% + 16.1 %
16 + 65,7%. + 14.7 %
25 + 61.4% + 9.2 %
82
-------
# Probes
48
# Probes
1
9
16
49
# Probes
8
16
24
TABLE 9-12
Test Results for Case XIII-1
Strategy: British Standards 1042
Error (e) Computed by Program
eemlssion eave. velocity
+ 2.2 % t 4.0 %
Strategy: Equal Area
Error (e) Computed by Program
eemission eave. velocity
+ 5.7 % + 7.4 %
+ 9.9 % + 13.5 %
+ 11.2 % + 14.8 %
+ 7.9 % + 8.2 %
Strategy; Circular Analog-Equal Area
Error (e) Computed by Program
eetmssion
+ 3.5 %
+ 5.3 %
+ 6.3 %
save, velocity
+ 5.9 %
+ 2.8 %
+ 3.6 %
# Probes
9
16
49
Strategy: Chebyshef
Error (e) Computed by Program
eemission eave. velocity
+ 9.9 % + 13.6 %
+ 8.7 % + 12.7 %
•f 6.6 56 + 3.5 %
83
-------
Probes
48
# Probes
1
9
16
49
TABLE 9-13
Test Results for Case XIII-2
Strategy: British Standards 1042
Error (e) Computed by Program
cemission eave. velocity
+ 4.6 % + 4.0 %
Strategy: Equal Area
Error (e) Computed by Program
eemission
+ 11.1 %
+ 16.5 %
+ 18.6 %
+ 10.2 %
eave. velocity
+ 7.4 %
+ 13.5 %
+ 14.8 %
+ 8.2 %
Strategy; Circular Analog-Equal Area
# Probes
8
16
24
Error (e) Computed by Program
eeimssion
+ 4.1 %
- 6.6 %
- 7.4 %
eave. velocity
+ 5.9 %
- 2.8 %
- 3.6 %
Probes
9
16
49
'Strategy; Chebyshef
Error (e) Computed by Program
eemission
+ 16.8 %
+ 16.4 %
+ 2.8 %.
eave. velocity
+ 13.6 %
+ 12.7 %
+ 3.5 %
84
-------
TABLE 9-14
Test Results for Case XIII-3
Strategy: British Standards 1042
Probes
48
Error (e) Computed by Program
eemission ea ye. veloc1ty
+ 2.5 % + 3.5 %
Strategy: Equal Area
Probes
1
9
16
49
Error (e) Computed by Program
eemission eave. velocity
+ 2.4 % + 7.1 %
+ 15.5 % + 17.9 %
+ 12.8 % + 16.0 %
+ 6,4 % + 7.4 X.
Probes
8
16
24
Strategy; Circular Analog-Equal Area
Error (e) Computed by Program
eemission
+ 10.3 %
+ 1.2 %
+ 0.33%
eave. velocity
+ 11.0 %
+ 1.4 %
+ 0.60 %
Probes
9
16
49
Strategy: Chebyshef
Error (e) Computed by Program
eemission eave. velocity
+ 14.1 % + 17.5 %
+ 7.5 % + 11.6 %
+ 2.8 % + 1.7 %
85
-------
Probes
48
TABLE 9-15
Te,st Results for Case XI11-4
Strategy: British Standards 1042
Error (e) Computed by Program
eemission eave. velocity
+ 4.1 % + 3.5 %
Probes
1
9
16
49
# Probes
8
16
24
Strategy: Equal Area
Error (e) Computed by Program
eemission
+ 3.0 %
+ 16.5 %
+ 15.1 %
+ 7.0 %
eave. velocity
+ 7.1 %
17. 9 %
16.0 %
+ 7.2 %
Strategy: Circular Analog-Equal Area
Error (e) Computed by Program
eemission
+ 11.4 %
+ 2.1 %
+ 1.1 %
eave. velocity
+ 11.0 %
+ 1.4 %
+ 0.6 %
Probes
9
16
49
Strategy; Chebyshef
Error (e) Computed by Program
eemission
+ 15.4 %
+ 10.1 %
+ 1.2%
eave. velocity
+ 17.5 %
+ 11.6 %
+ 1.7 %
86
-------
# Probes
48
I Probes
1
9
16
49
# Probes
8
16
24
f Probes.
9
16
49
TABLE 10-1
Average Errors for Four Rectangular Duct Sample Cases
Strategy: British Standards 1042
Error (e) Computed by Program
eemission eave. velocity
2.2
1.6 1.8
Strategy; Equal Area
2.4
Error (e) Computed by Program
eemission
59,3 59.3
16.3 16.3
10.3 10.3
4.3 4.3
eave. velocity
46.7 46.7
15.5 15.5
11.4 11.4
4.7 4.7
Strategy: Circular Analog-Equal Area
Error (e) Computed by Program
eemission eave. velocity
0.8 4.1 2.7 5.3
5.5 10.7 5.0 6.5
4.5 10.5 3.9 6.
Strategy: Nev/ton-Cotes*
Error (c) Computed by Program
eemission cave, velocity
-81J 81.1 -82.8 '82.8
-62.5 62.5 -64.2 64.2
-33.8 33.8 -35.3. 35.3
* Points weighted equally not by appropriate weighting factors
87
-------
TABLE 10-1. (cont.)
Strategy: Gauss*
Error (e) Computed by Program
# Probes eemlsslon cave, velocity
9 -5.7 9.3 -3. 9.6
16 -20.9 23.6 -18. 19.9
Strategy: Chebyshef
Error (E) Computed by Program
# Probes eemlssion cave, velocity
9 9.3 9.3 9.3 9.3
16 3.1 3.1 4.1 4.1
49 1.8 3.3 1.0 1.0
* Points weighted equally not by appropriate weighting factors
-------
.TABLE io-2
Average Error for Six Circular Ducts; Diameter Locations Segregated
Strategy: Centrold
I Probes
1
Error (c) Computed by Program
cenrlsslon cave, velocity
180.8
326.3 343.4
184.1
Strategy; Log Linear
Error {c) Computed by Program
I Probes^
4 probes/dlameter-2 diameters*
4 probes/d1ameter-4 diameters**
6 probes/dlameter-2 diameters*
6 probes/diameter-.4 diameters**
8 probes/d1ameter-2 diameters'*
8 probes/dlameter-4 diameters**
6 probes/dlamcter-2 diameters (30° and 120')
6 probes/d1ameter-.2 diameters (45* and 135°)
6 probes/dlatneter-2 diameters (60* and 150*)
eemisslon
-15.3
- 4.5
-14.9
*• 5.4
-1K5
- 1.9
2.5
4.1
0.5
15.3
8.5
14.9
5.4
11.6
3.2
5.
5.2
7.9
eave. velocity
-12.4
-6.
-10.1
-4,7
-7.5
•2.1
0.03
0.7
-1,5
12.4
6.
10.1
4.7
7.5
2.2
1.4
2.1
4.5
Strategy: Tangential
Error (e) Computed by Program
I Probes
2 p.robes/dlameter-2 diameters*
2 probes/d1ameter-4 diamet&rs**
4 probcs/diameter-2 diameters* •
4 probes/d1ameter-4 diameters**
6 probes/tilamcter-2 diameters*
6 probes/d1a»>eter-4 diameters**
6 probes/diamcter-2 d1ameters(30' and 120«)
6 prober./d1ameter-2 diameters(45* and 135°)
6 probcs/diameter-2 diameters(60° and 150°)
eemisslon
cave, velocity
2.5
9.8
-6.9
2.9
-9.
0.8
9.2
10.5
6.3
21.2
21.4
8.3
4.4
9.9
2.6
9.2
10.5
11.4
7.5
10.8
-3.7
.1-7
-5.3
0.1
5.
5.5
3.2
t
7.6
10.9
5.1
2.9
6.2
2.2
5.
5.5
5.5
and 90°
**
45°,
90% 135*
89
-------
TABLE 10-3
AVERAGE ERRORS FOR SIX CIRCULAR DUCTS
REGARDLESS OF STRATEGY AND DIAMETER LOCATION
I PROBES
1
4
8
12
16
24
32
cEMISSION
326.3
2.5
-4.1
1.1
-4.4
-2.3
-1.9
343.4
21.2
15.0
9.3
8.2
4.0
3.2
eVELOCITY
180.8
7.5
-1.8
-0.3
-3.9
-2.3
-2.1
184.1
7.6
9.5
5.1
5.5
3.5
2.2
90
-------
TABLE 10-4
AVERAGE ERRORS FOR FOUR RECTANGULAR DUCTS
REGARDLESS OF STRATEGY
I PROBES
1
8
9
16
24
48
49
cEMISSION
59.3
0.8
-15.3
-12.9
4.5
1.6
-9.3
59.3
4.1
2.9
22.
10.5
1.8
13.8
eVELOCITY
46.7
2.7
-15.3
-12.3
3.9
2.2
-9.9
46.7
5.3
29.3
21.2
6.
2.4
13.7
91
-------
TABLE 10-5
AVERAGE ERROR FOR TEN DUCTS
REGARDLESS OF STRATEGY, GEOMETRY AND LOCATION
# PROBES
1
4
8
9
12
16
24
32
48
49
EMISSION
219.5
2.5
-3.2
-15.3
1.1
-8.9
-0.6
-1.9
1.6
-9.3
229.7
21.2
13.0
29.
9.3
15.5
5.6
3.2
1.8
13.8
eVELOCITY
127.1
7.5
-1.
-15.3
-0.3
-8.3
-.7
-2.1
2.2
-9.9
129.1
7.6
8.7
29.3
5.1
13.8
4.1
2.2
2.4
13.7
92
-------
TABLE 10-6
AVERAGE ERROR FOR SIX CIRCULAR DUCTS
BY STRATEGY AND PROBE NUMBER
REGARDLESS OF DIAMETER LOCATION
I PROBES
GENTROJD
1
LOG LINEAR
8
12
16
24
32
TANGENTIAL
4
8
12
16
24
eEMISSION
326.3
-15.3
-1.9
-8.
-5.4
-1.9
2.5
1.4
4.2
2.9
0.8
343.4
15.3
8.3
10.
5.4
3.2
21.2
14.9
10.3
4.4
2.6
eVELOCITY
180.8
-12.4
-2.7
-6.7
-4.7
-2.1
7.5
3.5
2.1
1.7
0.1
184.1
12.4
4.6
6.8
4.7
2.2
7.6
8.
5.6
2.9
2.2
93
-------
TABLE 10-7
AVERAGE ERROR FOR SIX CIRCULAR DUCTS
BY DIAMETER LOCATION AND PROBE NUMBER
REGARDLESS OF STRATEGY
# PROBES
4
8
12
16
12
12
12
8
16
24
32
(0°,90°) 2 diameter
(0°,90°) 2 diameter
(0°,90°) 2 diameter
(0°-90°) 2 diameter
(30° ,120°) 2 diameter
(45°, 135°) 2 diameter
(60° ,150°) 2 diameter
(0°f45°,90M350) 4 diameter
(0°,45%90°,135°) 4 diameter
(0° ,45°, 90° ,135°) 4 diameter
(0°t45°,90M350) 4 diameter
eEMISSION
2.5
-11.1
-12.
-11.5
5.9
7.3
3.4
9.8
-.8
-2.3
-1.9
21.2
11.8
12.4
11.6
7.1
7.8
9.6
21.4
6.5
4.0
3.2
eVELOCITY
7.5
-8.1
-7.7
-7.5
2.5
3.1
0.9
10.8
-2.2
-2.3
-2.1
7.6
8.8
8.2
7.5
3.2
3.8
5.0
10.9
4.5
3.5
2.2
94
-------
TABLE 11-1
EMISSION ERROR VS. NUMBER OF PROBES
USING DIFFERENT METHODS FOR TRAVERSING RECTANGULAR DUCTS
Number
of
Probes
8, 9
16
24
48, ,49
Strategy % Error
Case
Number
IV
V
X
XI
IV
V
X
XI
IV
V
X
XI
IV
V
X
XI
Equal
Area
8.4
21.8
14,3
20.7
13.4
11.4
6.2
10.3
7.7
6.0
1.33
2.3
Circular
Analog
- 2.8
- 0.72
- 3.1
9.7
- 2.6
- 7.8
19.8
12.6
- 3.6
- 8.4
18.4
11.7
Chebyshef
8.0
16.9
1.8
10.4
11.4
0.63
0.19
0.074
4.0
6.3
- 1.32
- 1.8
British
Standard
5.4
1.15
- 0.4
0.2
95
-------
TABLE 11-2
PERCENT AVERAGE EMISSION ERROR
AND STANDARD DEVIATION
VS. NUMBER OF PROBES
From A111 Strategies
Probes % Average Emission Standard Deviation
8, 9 8.78 8.67
16 6.30 8.13
24 4.52 12.61
48, 49 2.57 3.22
96
-------
TABLE 11-3
EMISSION ERROR VS UMBER OF PROBES USING DIFFERENT METHODS FOR
TRAVERSING CIRCULAR DUCTS
Total
•Muter
of Case
Probes Nwber
4 II
III
VI
VII
VIII
IX
•8 II
III
VI
VII
VIII
IX
12 II
III
VI
VII
VIII
IX
16 II
III
VI
VII
VIII
IX
2 Diameter
(1)
- 23.1
- 1.6
-24.5
- 11.1
- 13.0
-18.3
- 18.4
- 1.62
- 15.9
- 9.9
• 23.0
- 20.7
• 15.9
- 0.145
- 9.5
- 5.8
- 19.4
- 18.8
4 Diameter
(2)
- 11.1
- 0.91
- W.3
- 6.6
12.1
- 2.1
Loo Linear
2 Dlawter 2 Dlawter
(3) (4)
3.8 2.5
0.48 0.137
- 4.7 ' - 2.3
• 2.7 - 1.1
17.4 20.5
1.0 4.8
STRATEGY X ERROR
2 Dlawter
(5)
-21.0
- 1.17
* 1.66
4.3
18.7
0.52
2 Dlawter
(1)
26.1
10.1
24.9
10.1
- 36.4
- 19.7
- 1.24
4.2
- 6.3
- 1.82
- 19.3
-17.2
- 7.2
2.7
- 7.9
- 3.7
- 20.1
- 18.0
Tangential
4 Dlawter 2 Dlawter 2 Dlawter 2 Dlawter
(2) (3) (4) (S)
31.3
12.0
35.4
15.2
- 19.6
- 15.2
8.2
5.4
1.68
3.1
3.5
4.7
15.5 12.8 - 15.2
5.2 4.8 3.5
4.7 7.4 11.8
4.1 5.9 11.8
22.4 25.0 23.5
3.5 7.0 2.6
-------
TABLE 11-3 (Cent.)
EMISSION ERROR VS NUMBER OF PROBES USING DIFFERENT METHODS FOR
TRAVERSING CIRCULAR DUCTS
'ot«J
lumber
if
robes
M
32
1) 0%
2) 0'.
3) 30*
4) 45*
5) 60s
C*se 2 Dimeter
Number 0)
II
III
VI
VII
VIII
IX
II
III
VI
VII
VIII
IX
90*
45*. 90«, 135*
. izcr
. 115*
. 150*
Log Linear
4 Diameter 2 Diameter 2 Dian
(2) (3) (4)
- 7.9
- 0.74
- 9.1
- 5.5
- 1.24
- 7.9
- 5.2
1.08
- 2.3
- 1.26
2.7
- 6.5
STRATEGY X ERROR
Tangential
eter Z Diameter 2 Diameter 4 Diameter 2 Diameter Z diameter 2 Dlaneter
(5) (1) (2) (3) (4) (5)
2.8
3.8
0.27
1.11
2.4
- 5.5
-------
TABLE 11-4
PERCENT AVERAGE EMISSION ERROR
AND
STANDARD DEVIATION VS. NUMBER OF PROBES
Num£er From All Strategies
of "—^ '—
Probes % Average Emission Error Standard Deviation
4 2.52 25.22
8 - 4.12 17.84
12 1.15 12.13
16 - 4.39 9.57
24 - 2.29 4.64
32 - 1.91 3.54
99
-------
90
80
70
3 60
£
£ 50
8
40
30
20
10
Q Rectangular
O Circular
• Both (average)
Figure 22. Error 1n Emission for Rectangular and
Circular Ducts as a Function of Total
Number of Probes
10
20 30 40
Number of Probes
50
60
100
-------
90
80
70
«? 60
o
I/)
M
50
£ 40
0)
Q.
Figure 23. Error 1n Emission for Rectangular
Ducts as a Function of Strategy
and Total Number of Probes
30
20
10
Circular Analog - Equal Area
Standards
10
20 30
Number of Probes
40
50
60
70
101
-------
V
•f
«J
£ 30
VI
s. 20
10
Figure 24. Error 1n Emission for Circular Ducts as
a Function of Strategy and Total Number
of Probes.
Curve 1 Log Linear A
Curve 2 Tangential Q
Curve 3 2 dlam. (0°, 90°) •
Point 4 1 dlam. (30°, 120°)
Point 5 2 dlam. (45°. 135°)
Point 6 2 dlam. (60°, 150e) •
Curve 7 4 dlam. (0°, 45°. 90*» 135°) O
>C£T
10 20 30 40 SO 60 70
Number of Probes
102
-------
40
30
Q Equal Area
• Circular Analog
D British Standard
A Chebyshef
20
10
0
60
lU
t:
o
-10
-20
-30
NUMBL'R OF PROUHS
Figure 25. Imission I rror vs. Nuirl-er of Probc»s for Different Probe
Locations in Rectangular Ducts
103
-------
40
Mean Emission Error and
Standard Deviation
30
20
10
i.
o
-10
60
l/»
V
e
-2C
-30
Nun.ber of Probes
Figure 26. Mean Emission Error vs. Number of Probes for Different
Probe Locations in Rectangular Ducts
104
-------
40
30
20
10
\
I
O 2 0
•400°, 45°, 90°, 135°
O 2 D 30°, 120°
• 2 D 45°, 135°
A 2 D 60°, 150°
LU | ^
c
o
•r-
V)
•| -10
UJ
-20
0
O
o
Q
i
-------
e
40
30
20
8
10
• o
** n
o 0
o
-10
I
O 2 D
• 4 D, 0°, 45°, 90°, 135°
D 2 D, 30°, 120°
• 2 D, 45°, 135°
A 2 D, 60", 150°
20
30
40
50
60
-20
- o
o
o
-30
Number of Probes
Figure 28. Emission Error vs. Number of Probes For the Tangential Method
For Probe Locations In Circular Ducts
106
-------
Mean Emission Error
and Standard
-30
Figure 29. Mean Emission Error vs. Number of Probes For Different
Probe locations In Circular Ducts
107
-------
TABLE 12-1
AVERAGE ERRORS FOR EJGHT RECTANGULAR DUCT SAMPLE CASES
4 Probes
48
Strategy: British Standards 1042
Error (c) Computed by Program
eemission eave. velocity
2.5 2.6 3.0 3.1
Strategy: Equal Area
Error (e) Computed by Program
# Probes
1
9
16
49
eemission
32.4
15.4
12.4
6.1
32.4
15.4
12.4
6.1
eave. velocity
27.0
15.6
13.4
6.2
27.0
15.6
13.4
6.2
# Probes
8
16
24
Strategy: Circular Analog-Equal Area
Error (e) Computed by Program
eemission
4.1
3.0
2.3
5.7
7.2
7.1
Strategy; Newton-Cotes*
eave. velocity
5.6 6.9
2.8 4.3
2.1 4.0
# Probes
9
16
49
Error (e) Computed by Program
eemission eave. velocity
-81.1 81.1 -82.8 82.8
-62.5 62.5 -64.2 64.2
-33.8 33.8 -35.3 35.3
Points weighted equally not by appropriate weighting factors.
108
-------
TABLE 12-1. (cont.)
Strategy: Gauss*
Error (e) Computed by Program
# Probes eemlssion eave. velocity
g- 5.7 9.3 - 3.0 9.6
16 -20.9 23.6 -18.0 19,9
Strategy: Chebyshef
Probes
Error (e) Computed by Program
eemission save, velocity
11.7 11.7 12.4 12.4
*6 6.9 6.9 8.1 8.1
4g 2.6 2.6 1.8 1.8
Points weighted equally not by appropriate weighting factors
109
-------
TABLE 12-2
AVERAGE ERRORS FOR EIGHT RECTANGULAR DUCTS
REGARDLESS OF STRATEGY
Number
of
Probes
eemlsslon
eveloclty
1
8
9
16
24
48
49
32.4 32.4
4.1 5.7
7.4 22.3
4.9 17.4
2.3 7.1
2.5 2.6
3.3 10.2
27.0
5.6
7.
3.9
2.1
3.0
3.9
27.
6.9
22.7
17.0
4.0
3.0
10.3
110
-------
TABLE 12-3
AVERAGE PERCENT ERROR FOR FOURTEEN DUCTS
REGARDLESS OF STRATEGY, GEOMETRY AND LOCATION
Number
pf
Probes
I
4
8
9
12
16
24
32
48
49
e emission
158.3
2.5
- 1.6
- 7.4
1,1
- 1.9
- 0.46
- 1.9
2.5
- 3.3
165.7
21.2
12.1
22.3
9.3
13.4
5.2
3.2
2.6
10.2
evelocity
92.9
7.5
- 1.2
- 7.
- 0.3
- 3.9
- 5.4
- 2.1
3.
- 3.9
94.3
7.6
6.6
22.7
5.1
2.3
3.7
2.2
3.
10.3
in
-------
TABLE 12-4
EMISSION ERROR VS. NUMBER OF PROBES
USING DIFFERENT METHODS FOR TRAVERSING RECTANGULAR DUCTS
Number
of
Probes
8, 9
16
24, 25
Case
Number
IV
V
X
XI
XII
XIII-1
XIII-2
XIII-3
XIII-4
IV
V
X
XI
XII
XIII-1
XIII-2
XIII-3
XIII-4
IV
V
X
XI
XII
XIII-1
XIII-2
XIII-3
XIII-4
Equal Area
8.4
2.1.8
14.3
20.7
58.9
16.5
16.5
15.5
9.9
13.4
11.4
6.2
10.3
68.6
18.6
15.1
12.8
11.2
64.5
Strategy
Circular Analog
- 2.8
- 0.72
- 3.1
9.7
78.4
4.1
11.4
10.3
3.5
- 2.6
- 7.8
19.8
12.6
113.0
- 6.6
2-1
1.2
5.3
- 3.6
- 8.4
18.4
11.7
- 7.4
1.1
0.33
6.3
percent error
Chebyshef British Standard
8.0
16.9
1.8
10.4
59.6
16.8
15.4
14.1
9.9
11.4
0.63
0.19
0.074
65.7
16.4
10.1
7.5
8.7
61.4
112
-------
TABLE 12-4 (Cont.)
Number
of
Probes
Case
Number
Strategy percent error
Equal Area
Circular Analog
Cnebysnef
British Standard
48, 49
IV
V
X
XI
XII
XIII-1
XIII-2
XIII-3
XIII-4
7.7
6.0
1.33
2.3
10.2
7.0
6.4
7.9
1.2
2.8
4.0
6.3
- 1.32
- 1.8
2.8
4.1
2,5
6.6
5.4
1.15
- 0.4
0.2
4,6
2.2
113
-------
TABLE 12-5
PERCENT AVERAGE EMISSION ERROR
AND STANDARD DEVIATION VS. NUMBER OF PROBES FOR RECTANGULAR DUCTS
Number
of From All Strategies
Probes % Average Emission rateStandard Deviation
8, 9 16.5 19.1
16 15.7 26.1
24 14.4 26.9
48, 49 3,7 3.1
114
-------
40
30
O Equal Area
• Circular Analog
O British Standard
& Chebyshef
20
10
8
8
40
60
UJ
O
8 -
•20
-30
Number of Probes
Figure 30. Emission Error vs. Number of Probes for Different
Probe Locations 1n Rectangular Ducts
115
-------
40
30
20
10
J 0
-10
to
All Cases
Mean Emission
Error and Standard
Deviation
Cases IV, V, X,
_XJ, only
50
60
-20
-30
Number of Probes
Figure 31. Mean Emission Error vs Number of Probes for
Different Probe Locations 1n Rectangular Ducts
116
-------
100
96
92
88
84
80
76
72
68
64
60
56
52
48
44
40
36
32
28
24
20
16
12
8
4
• Both
O Rectangular
O Circular
Figure 32. Error 1n Emission For Rectangular
And Circular Duct As a Function
Of Total Number of Probes
1 1
4 8 12 16 20 24 28 32 36 40 44 48 52 56 60
Number of Probes
117
-------
u
Q
U
ft)
o
c
o
o
88 p
84 -
80 -
76 -
72
68
64 -
60
56
52
48
44
40
36
32
28
24
20
16
12
8
Figure 33. Error in Emission for Rectangular
Ducts as a Function of Strategy and
Total Number of Probes
Standard
4 8 12 16 20 24 28 32 36 40 44 48 52 56
Number of Probes
118
-------
computed by summing actual errors; the rightmost entry for the average
error in emission or velocity has been computed by summing the absolute
values of the errors. In Tables 11-1 thru 11-4 all of the entries for
average error are calculated by a summation of actual errors, i.e. allowing
the positive and negative errors to cancel in the summation. Figures 22
thru 29 are derived from Tables 10-1 thru 11-4 while Figures 30 thru 33
are derived from Tables 10-1 thru 12-5.
Our evaluation of these results follows. Note that these
results are based on very limited data (9 rectangular and 6 circular ducts).
A principal finding of our study is that many more test cases are required
before reliable conclusions can be reached. All conclusions are therefore
based on the assumption that the 16 cases analyzed thus far represent good
samples of the broad range of actual industrial velocity-concentration
stratification profiles.
a.4 Conclusions and Recommendations
In general, increasing the number of probes for a given
strategy would increase the probability of determining average velocity or
emissions. Figures 22, 23, 24 and 26, and Figures 31, 32, 33 indicate that
this is the case for all strategies except the modified Gauss and the Circular
Analog-Equal Area strategies for the first four rectangular ducts that were
studied. These two exceptions are probably due to the small number of cases
studied. We therefore postulate that the error limits in sampling under
stratified conditions decrease with an increasing in the number of probes.
However, the magnitude of error for a particular case may increase by in-
creasing the number of sample probes. For an extreme example, consider the
case where a single probe yields a zero percent error while four sample probes
yield results that are 1Q% in error. Obviously, one does not conclude from
this that using one probe is a better general sampling procedure than using
four probes, but only that it is possible for the correct result to be ob-
tained with one probe. Therefore, intuition indicates that, in general, the
119
-------
magnitude of probable error will decrease with an increase in the number
of samples. Furthermore, it is postulated that the magnitude of probable
error will decrease with an increase in the number of samples at a faster
rate for some methodologies than for others.
For circular ducts, the log-linear technique tends to under-
predict average emissions while the tangential strategy tends to overpredict.
For the same number of probes and diameters the tangential strategy is
slightly better than the log-linear technique in error magnitude; however,
as the probes are oriented through an axis of flow symmetry the log-linear
technique appears to be somewhat superior. The high mean error of the cen-
troid strategy (1 probe in the center of the duct) indicates how poor this
method is for emission measurements in stratified gases.
Generally speaking, for circular ducts, for the same number
of probes, the number of diameters and diameter locations may have a sig-
nificant effect on the emissions measurement. For example, consider the
case where C = f(r,9) while the concentration dependency is "greater" on
r than 9. Here 8 probes on 2 diameters will likely be as, or more, effective
a strategy than 8 probes on 4 diameters. However, if the dependency were
reversed, i.e., the concentration were more dependent on 6 than on r, 8
probes on 4 diameters would be more effective than 8 probes on 2 diameters.
Clearly, it would seem that an initial detailed characterization of the shape
of the concentration and velocity profiles is necessary to determine the most
effective measurement strategy for a given duct sampling location. Without
this a priori knowledge, the tangential strategy with 16 probes should
provide a good indication of mean emissions from a circular duct (see
Table 11-4) which indicates a mean error of 4.4% for the cases analyzed.
In contrast to the strategies that have been evaluated for
circular ducts, different strategies give substantially different results
for rectangular ducts. For the first 4 cases analyzed, the modified Newton
Cotes technique gives the highest mean error while the .Chebyshef technique
gives the lowest. Note, however, that the British Standards technique using
120
-------
48 probes has a lower mean error (1.8%) than the 49 point Chebyshef tech-
nique (3.3%), even though the size of the area segments are substantially
2
greater than the recommended 36 in . Preliminary results show that 16
probes located in accord with the Chebyshef strategy should provide a
satisfactory indication of mean emissions from a rectangular duct. (See
Figure 23 which indicates a mean error of 3.1% for the cases analyzed.)
For rectangular ducts, the "circular analog" technique also
appears to be a rather good strategy; however, its effectiveness clearly
depends on the existence of flow symmetry within the duct.
As was the case with the circular ducts, a single probe
located at the center of a rectangular duct yields a higher average error
in predicted emissions (59.3%). Since in practical cases it would be
highly desirable to measure average emissions with a single probe, it would
be useful to develop a technique with which to determine the focus of points
within ducts that would lead to "zero error" in emission prediction. By
applying such a technique to a large group of sample concentration-velocity
profiles, one could develop sets of curves, the study of which could lead to
the development of an acceptable single point measurement strategy. Such an
approach could be developed through modification of the existing analysis
computer program.
The results "from the experimental wind tunnel tests shown
in Table 9-11 indicate an extermely poor measurement accuracy for all strategies
using less than 25 probes. However, experimental results have not duplicated
the error magnitudes suggested by the computer analysis. It is our judgment
that these differences may be associated with a difficulty in fitting the
usually irregular profile data of Table 6. That is, the computer fit may
tend to mask some of the extreme variations in the raw data which would, of
course, not happen in experimental tests. This suggests that extremely
stratified flows must be measured with a large number of probes (-50) to
assure reasonable measurement accuracy. Also, the simulation program should
be modified to allow extreme profiles to be adequately represented.
121
-------
The addition of the four TVA ducts (Cases XIII-1 to XI11-4)
increased the data base for our earlier conclusions and added considerable
weight to their reliability. We still believe that many more cases should
be analyzed to insure that our selected profiles are indeed representative
of the broad range of actual velocity-concentration stratification profiles.
These results have confirmed our earlier assessment
that the circular-analog-equal area strategy is extremely effective for
cases exhibiting flow symmetry. The additional data have flattened the
summary curve for this strategy (Figure 33) so that increasing the number
of probes now appears to yield a reduction in sampling error. In fact, the
equal area strategy now appears to be the most effective strategy analyzed
to date for rectangular ducts.
Figure 33 indicates that, in general, irrespective of strategy
(random probe location), there appears to be no significant difference in
measurement accuracy between the circular and rectangular strategies that
have been evaluated in this program.
These new results do not differ substantially from earlier
findings except in regard to estimated error magnitude. For example, the
error in emission measurement associated with a single probe located at the
center of a rectangular duct has been reduced in our calculations from 59.3%
to 32.4%. However, considering the range of values making up this average,
it must be concluded that a single point strategy is a poor choice unless
one has prior knowledge of the flow characteristics.
Note also that the 16 point Chebyshef procedure now'indicates
a mean error of 6.9% rather than the 3.1% reported earlier. However, the
circular analog strategy now appears to be superior with a 16 probe mean
error of 3.0% and an 8 probe mean error of 4.1%.
122
-------
b. Sampling Methodologies
In heterogeneous sampling, it is generally not possible to
obtain a representative sample from a single arbitrary sampling location.
In order to study the different strategies for obtaining a representative
sample from stack gases, a mathematical analysis of the emission rate
general equation was performed and different sampling methodologies were
derived.
is equal to:
Analysis: The emission of material from a combustion source
f Ca v
JA
ndA
(1)
where: Ca = concentration of species (a)
v = flue gas velocity along the duct
ii = unit vector normal to A
A = cross sectional area of a duct
Ea = emission rate of specie
a
Assuming the velocity vector is perpendicular to area (A), we can also
write for species (a)
f -*-
h It
v dA
(2)
where:
w
t
q
mass of species (a) measured during time (t)
sampling time
volumetric sampling rate
We can approximate (2) by writing the integral as a finite sum, viz.,
123
-------
(3)
where: n = number of sampling locations
i=n
Z= A
1=1
v. = average velocity in sampling area A.
Therefore:
Ww A wufl utuA
l w i r\i «/5»o"o "«v «r\-.
F = _U_l+_2_2-l+ + JQJUL (4)
qiti q2*2 Vn l '
Equation (4) is the general equation that can be used to determine the
emission rate from a finite number of measurements. It is obvious that
in the limit of an infinite number of sampling points, an accurate answer
is obtained.
Table 13 shows that by placing special conditions or con-
straints on the parameters of equation (3) a simplified expression is derived
and a sampling method(s) suggested. The different methods presented in
Table 13 make up the group of sampling strategies which may be implemented
to obtain a representative gas sample from a process stream.
c. Jet Mixing of Flue Gas Streams
From information obtained in the literature search, we con-
cluded that in-the-stack mixing devices are not suitable for practical reasons,
i.e., for mechanical installation (retrofit) and/or adverse pressure drops in
the process streams. A scheme which may avoid both of these probelems involves
the use of gas jets to mix the process stream (See Figure 34). in this scheme,
a slip-stream from the flue gas is withdrawn from the duct and then pumped
back into the duct as a mixing jet.
124
-------
TABLE 13
APPROACHES BASED ON THE GENERAL EQUATION
FOR FINITE SAMPLES
Special Cases
and Conditions
Applicable Equation
Sampling Methods
For Steady State
1.
An •
- V*
v t ... v
E •
A1 hv1 . W2V2 . Vnl
^^" I " —" * "" '*•' ^IL • • • •»~"^ I
t Ml q2 qn I
L J
Concentration, velocity and
sampling rate must be known at
each sampling position
a. Automatically traversing
with one probe
b. Using several probes with
sequential sample analysis
2.
v, v?
tj- ... tn
v
•" qT
where K 1s a known
constant
3. A, • A2
... v
t21 .,.. tn
E • A1 K Cav
The proportional constant K must
be known and the average concen-
tration of the mixed samples,
also the total sampled volume
per unit time
W.V.
n
Concentration, velocity,
sampling time and flow rate for
each sample location must be
known
a. Automatically traversing
and automatically adjusting
the sampling flow rate
b. Use one probe or several
probes with one mixed
sample analysis
c. For v/q to equal a constant
either the velocity at the
entrance of the sampling
nozzle* 1s adjusted equally
or proportionally to the
velocity of the flue gas at
the sampling location
a. Automatically traversing
with one probe
b. Using several probes with
sequential sample analysis
125
-------
TABLE 13. (Cont.)
Special Casts
and Conditions
Applicable Equation
Sampling Methods
4. A •
where Kj 1s a known
constant
5. A. • Ag • . ..
vl #v2
6. AI » A-
v, -x2
*!"*£" —
...
"
•" *
E • A
Cav
1-1
The proportional constant 1C, must
be known and the average concen-
tration of the mixed samples, also
the sampling time for each sample
location
n n
Concentration and velocity must
be known at each sample position
The velocity being uniform over
the cross section of the duct,
only one measurement 1s needed;
the average concentration of the
mixed samples must be known
or- nA^Cvav
The concentration being uniform
over the cross section of the
duct, only one measurement 1s
needed; the velocity of each
sampling position must be
known
a. Automatically traversing
and sampling at constant
flow rate only automati-
cally adjusting sampling
time for each sampling
position
b. Use one probe or several
probes with one mixed
sample analysis
a. Automatically traversing
and sampling at constant
flow rate
b. Use one probe or several
probes with sequential
sample analysis
a. Automatically traversing
with one probe or using
several probes with one
mixed sample analysis
Automatically traversing
with one probe or using
several probes for velocity
measurement, with one
sampling position
126
-------
v, » v» « ... v.
'1 "V2
TABLE 13 (Cent.)
Special Cases
and Conditions
Applicable Equation
Sampling Methods
a. A.
"• i.
L • rwi^v
The concentration and velocity
being uniform over the cross
section of the duct
a. Sampling with one probe
for sample analysis
9.
E « nA1 Cav Vav
The concentration and velocity
averages must be known
a. Sampling with many probes
averaging
127
-------
Counter Rotating Fans
a. Fan or Turbine Mixer
C
b. Gas Dynamics Mixer
Figure 34. Gas Jet Mixer
128
-------
The problem of 1n-the-stack mixing was formulated analytically.
The formulation for the two dimensional case (a tractable example) 1s presented
1n Appendix H-l. A less rigorous analysis was also performed and 1s presented
1n Appendix H-2. Calculations based on this analysis were performed on the
problem of mixing flue gas by jets, and are presented 1n a following section,
(See c.2).
c.l Background
In principle, with the flattening of the concentration profile,
the problem of extracting a representative sample becomes trivial, since a
single point sample may be taken at any sampling rate; i.e., non-proportional
sampling. Thus, a method which flattens (equal to a constant) the concen-
tration profile 1s a very powerful method, and any reasonable scheme to
achieve this should be fully Investigated.
Concentration profiles can be flattened by mixing the gas
stream. In any gas stream, dlffusional mechanisms tend to reduce small-scale
stratification; thus the role of any mixing approach should be to promote
large-scale (relative to duct size) mixing. Schemes to promote mixing fall
Into two classes, viz.:
• Passive
• Active
Passive mixing schemes extract energy from the gas stream to
promote mixing and, hence, necessarily introduce a pressure drop in the system.
While it is theoretically attractive to think of using this pressure drop as
a flowmeter, passive mixing elements are Impractical for retrofit to full-scale
systems for the reasons of excessive pressure drop. However, for new plant
designs, passive mixing methods might be considered. Some examples of passive
mixing schemes are:
• mixing orifices (Figure 35-a)
• mixing disks or plates (Figure 35-b)
• swirl vanes (Figure 35-c)
129
-------
o
a.
A
2
•€>
Swirl Vane
Figure 35. Passive Mixing Schemes
130
-------
Active mixing elements take a variety of forms. These
do not necessarily cause a pressure drop 1n the system; In principle,
they can cause a pressure Increase 1n the system. For these reasons,
active mixing Is a viable approach.
Some active mixing schemes are as follows:
• fan or turbine mixers (Figure 34-a)
• gas jet mixer (Figure 34-b)
c.2 Calculations and Results
Ideal fan power for mixing flue gas by a jet was calculated
for different duct sizes. Duct diameters ranged from 2.4 m (8) to 9.1 m
(30); the flue gases were assumed to be at 150°C (300°F) and one atmosphere.
The Ideal fan power for the jet was calculated by using the following equa-
tion:
P - WQH
where:
P - theoretical power, watts
W « specific weight of fluid, N/M3
Q * volumetric flow rate, m /s at flue condition
H « developed head, m
The developed head, of the jet was assumed to equal the velocity
v2
pressure head, I.e.* fa FHctlonal and other losses were not accounted for.
It was also assumed that good mixing In the flue would be obtained when the
following Is met (See Appendix H-2 for the theoretical development of this
analysis):
VD
UI
1
131
-------
where:
v = velocity of the flue gases 1n the flue (m/s)
D = diameter of flue (m)
U = velocity of the jet stream at the orifice (m/s)
d = diameter of jet duct (m)
Tables 14-1 and 14-2 and Figures 36 and 37 show results
for cases in which the jet inlet orifice is 1/10 the duct diameter.
In order to appreciate the power required under actual
power plant conditions, the following examples are given:
Example 1 Potter, P.J., "Power Plant Theory and Design",
p. 322, The Roland Press Co., N. Y., 2nd Edition,
1959.
o
Given: an induced draft fan designed at 193.5 m/s (410,000
ACFM). When the gas flow is equal to 89.7 m3/s (190,000 ACFM) the static
3 2
pressure 1s equal to 3.14 x 10 N/m (12.6 in. of water). If an inlet
damper control is used to control the gas flow, 72.5% of the outlet damper
power is used. The efficiency of the fan is equal to 0.7, and the shaft
hoursepower becomes:
for outlet control:
Shaft hp « 89'7 X 4 X ]° = 402 x 1(}3 W (539 Hp)
for inlet control:
Shaft hp * 402 x 103 x .725 = 291 x 103 W (391 HP),
and the developed pressure will equal 7 x 102 N/m2 (2.8 in., varies as the
square of the flow). Therefore, the theoretical fan power is equal to
291 x 103 x 0.7 « 205 x 'iO3 W (275
132
-------
TABLE 14-1
IDEAL ft* POWER FOR FLUE 6AS JET AT -1WC (300*F) AND 101.32 N/m2
(1 ATM)
Diameter ATM of
Of Flux Flue
2.4 4.67
(8')
3.6 10.52
<12')
j*» 6.1 29.18
" (20')
7.6 45.61
<25')
9.1 65.68
(30')
Velocity
In Flue
(•/$)
3.0
6.1
12.2
18.3
24.4
30.5
3.0
6.1
12.2
15.2
3.0
6.1
9*
.1
3.0
6.1
3.0
6.1
FloH Rate Ratio of
In Flux Flue to Jet
(B'/S) Dlaaeter
14.01 10
28.49
56.97
85.46
113.95
142.43
31.56
64.17
128.34
159.90
87.54
178.0
265.54
136.83
278.22
197.05
400.65
Diameter
Of Jet
0.24
0.36-
0.61
0.76
0.91
Velocity
In Jet
(«/$)
30
61
122
183
244
305
30
61
122
152
30
61
91
30
61
30
61
Jet
Area
<•*)
0.0467
0.1052
0.2918
0.4S6I
0.6568
Flow Rate
In Jet
1.401
2.849
5.697
8.546
11.395
14.243
3.156
6.417
12.834
15.990
8.754
17.800
2.655
13.683
27.822
19.705
40.065
Developed
Head-u*/29
47.24
189.28
757.12
4703.8
3029.10
4732.93
47.24
189.28
757.1}
1183.29
47.24
189.28
425.81
47.24
189.28
47.24
189.28
Power * Qj x H x w
H • 8.48 R/a*
W
•560 ( *0.7S)
4.474
35.868 46.1
121,176 162.5
287.169 85.1
560.915 (752.2
1.268 ( 1.7
10.067 13.5
80.759 108.3
157.715 (211.5
3,505 ( 4.7]
28.038 { 37.6
94.629 (126.9
i
'
5.444 ( 7.3)
43.847 ( W.ei
7.904 { 10.6)
63.086 ( 84.6}
Specific wight at 1508C and 101.32 N/n2x 103
-------
CO
TABLE 14-2
IDEAL FAN POWER FOR FLUE GAS JET AT -IStTC (300*F) and 101.32 H/m2
(1 ATH)
Diameter Area of
Of Flux Flue
(») <«')
2.4 4.67
(8')
3.6 10.52
02')
6.1 29.18
(20')
7.6 45.61
(25' )
9.1 65.68
(30')
Specific weight
** 3
ftj/«1n.
Velocity
In Flue
Otfs)
5.1
20.2
30.3
45.5
2.2
8.9
13.4
20. 2
0.8
3.2
4n
.8
7.2
0.5
2.0
3.1
4.6
0.3
1.4
2.1
3.2
at 150-C and
Flow Rate Ratio of
In Flux . Flue to Jet
(•'/$) Diameter
23.6 10
(50.000)**
94.4
(200.000)
146.6
(300.000)
212.4
(450.000)
23.6
94.4
141.6
212.4
23.6
94.4
141.6
212.4
23.6
94.4
141.6
212.4
23.6
94.4
141.6
212.4
101.32 N/b'x-103
Diameter Velocity Jet
Of Jet In Jet Area
(it) (n/s) (•*)
0.24 51 0.0467
202
303
455
0.36 22 0.1052
89
134
202
0.61 8 0.2918
32
48
72
0.76 5 0.4561
20
31
46
4.91 3 0.8568
14
21
32
Flow Rate
In Jet
(«'/$)
2-36
9.44
14.16
21.24
2.36
9.44
14.16
21.24
2.36
9.44
14.16
21.24
2.36
9.44
14.16
21.24
2.36
9.44
14.16
21.24
Developed
flead -u*/2g
131
2080
4685
10539
25
409
916
2080
3
53
120
268
1
21
48
109
0.6
10
23
53
Power • Qj x H x w
w « 8.48 */•*
(W)
(H.P.)
2.535 ( 3.4)
163.532 (219.3)
552.340 (740.7)
1.863,504 (2499)
500
32.140
107.977
367,854
62
4.175
14.094
47.426
22
1.640
5,667
.19,313
11 <
820
2.759
9.396
0.67)
43.1)
144.8)
493.3)
0.084)
5.6)
w • w •
18.9)
63.6)
0.03)
2.2)
7.6J
25.9)
0.015)
l.ll
3.7}
12.6)
-------
746
(1000)
671
(900)
597
(800)
522
(700)
ex
I
CO
o
447
(600)
373
(500)
296
(400)
•; 224
-g (300)
149
(200)
74
(100)
Figure 36. Ideal Fan Power for Jet vs Average Flue Gas* Velocity
For Different Flue Diameter At Diameter of Flue Duct/
Diameter of Jet Orifice » 1.0
* Flue gas assumed to be at 150°C
and 101.32 N/m2 x 193
6.1
(20)
12.2
(40)
18.3
(60)
24.4
(80)
30.5
(100)
\ ^ ^ t * r
Flue gas velocity, Vf m/s (ft/sec)
36.6
(120)
42.7
(140)
135
\
-------
74.6
(100)
67.1
(90)
59.7
(80)
Q
v, 52.2
° (70)
44.7
(60)
o>
1 1
ex
37.3
(50)
29.8
(40)
-O
22.4
(30)
14.9
(20)
7.4
(10)
Figure 37. Ideal Fan Power for Jet vs Average Flue Gas* Velocity
For Different Flue Diameter at Diameter of Flue Duct/
Diameter of Jet Orifice - 1.0
Flue gas assumed
to be 150°C and
101.32 N/m2 * 10
7.6 9.1 10,7
(25) (3C) (35)
tn/s (ft/sec)
136
-------
A jet used to mix 89.7 m3/s (190,000 ACFM) flowing through
a 3.0 m (12 ft.) diameter flue at 9 m/s (30 ft/sec) would need an Ideal
fan power for jet mixing of about 37 x 103 W (50 HP) (See Figure 36).
This 1s approximately equal to a fifth of the power used to drive the flue
gases. A quick estimate of the electrical cost per year can be obtained
1f we assume 4000 hrs/yr operation at 2tf/kw hour, giving a cost of electricity
equal to:
37 x IP3 x TO"3 x 4000 x 2 *
Example 2 Boston Edison Mystic River Power Plant,
from personal communication.
A flow of 141.6 m3/s (300,000 ACFM) 1s driven by a fan* with
o
a motor of maximum rating equal to 932 x 10 W (1250 HP), at an Inlet pressure
of 35 x 10 N/m (14" W.6.) vacuum and with a delivery pressure of about
?
101.33 N/m (1 atm). If the temperature of the flue gases 1s assumed to be
equal to 150°C (300°F) the ideal fan power becomes
Q x P - 141.6 x 35 x 102 = 495 x 103 W
where
p
P • developed pressure, N/nr
3
Q * actual flow rate m/s
If an overall efficiency** of 0.6 1s used, the actual horsepower equals
820 x 103 W (1100 HP) which 1s very close to the fan motor rating.
* the booster pump 1s not considered
** fan n « 0.7 and motor coupling n * 0.85
137
-------
If one considers jet mixing in the precipitator outlet duct,
7.3 m x 4.9 m (241 x 16'), the equivalent diameter is calculated to be
close to 6.1 m (20(). At 92 m3/s (195,000 SCFM) 141.6 m3/s (-300,000 ACFM
at ~150°C and 101.33 N/m ), the Ideal fan power for the jet is equal to
14.2 x 103 W (19 HP) (See Figure 36). The jet opening is equal to 0.61 M
(21) with a jet velocity of 48.8 m/s (160 ft/sec).
In the case of jet mixing in the after preheater section
with an equivalent diameter approximately equal to 3.6 M (121), the ideal
fan power for the jet is equal to 108 x 103 W (145 HP) (See Figure 36).
The jet opening is equal to 0.37 m (1.21) with a jet velocity of 134 m/s
(440 ft/sec).
From this it is concluded that, according to the jet mixing
site, the power needed can range from as low as 1/20 of the total ideal fan
power used in the main flue to as high as 1/5 of the total Ideal fan power.
It is significant to note that the fan power varies inversely
with the absolute temperature (speed and capacity being constant). In the
last example the flue gases are assumed to be at ~150°C (300°F); if lower
temperatures are used, more power will be required for the same actual flow
rate.
The results of this analysis imply that the power requirements
for the mixing jet can be modest compared to the power required to drive flue
gas through the duct. While this approach cannot be considered a general
solution for the problem of obtaining representative samples from all strati-
fied gas streams, it is likely to be applicable to some particular stratified
stream. However, scaled laboratory experiments should be performed to test
the results indicated herein before any full-scale application is attempted.
In summary, this approach now appears to be more promising than at the beginning
of this program.
138
-------
2. LABORATORY EXPERIMENTS
Tests with velocity and concentration traveses in square and
round sections of a wind tunnel were conducted and the results are presented
below. Different sampling methods are also discussed.
a. Wind Tunnel and Test Set Up
The installation of the wind tunnel equipment was first com-
pleted as shown in Figure 38. Ports and probe holders were installed at
two test sections, a circular section and a square section. The plan view
of the Installation shows the approximate locations of the planned test
sections. An additional 0.61 m (21) length of circular duct was Introduced
for the Annubar element, following the pltot tube traversing port. Aluminum
honeycomb cores (See Figure 39) were used in the round section following the
90° bend to produce a more uniform flow.
Provisions were made for metering ethane from a 1-A cylinder
Into the tunnel as the test gas. Analysis of ethane from gas extracted
from the test sections was by FID. The average concentration of ethane was
held below the lower limit of flamability in air (LEL in air 3%). Gas
velocity* was measured by a standard pitot static tube using a MKS, Inc.
Baratron instrument. Control of the velocity stratification and test gas
stratification was by use of a damper following the fan and a baffle28
introduced at the entrance of the wind tunnel and by position of gas in-
jection respectively. The referee method of measurement for the emission
of ethane through the test section was by rotameter reading of the ethane
supply.
* when air is used at temperature TS (K°) and 101.33 x 103 N/m2 the
velocity Vs(m/s) Is given by:
Ah x T.
Vs = 0.87
1.333 x 10*
where Ah = differential pressure in N/m2 (1 mm Hg = 1.333 x 102 N/m2)
139
-------
Tracer Gas Inlet
Honeycomb cell
To Exhaust
Line
0 62
(34-1/2")
variable length
Round Section
Air Flow
0.83
(32-3/4")
Figure 38. Wind Tunnel Plan View (One Fan)
140
Tracer Gas Cylinder
Tracer Gas
Detector
Experimental Ducts Showino
Tracer Gas Set Up
Scale: ~1:40 Dims. Meters
Sections
Square
Round
Plywood
Galvanized Steel
-------
Figure 39. Aluminum Honeycomb Cells
141
-------
In the most recent experiments, the wind tunnel was mod-
ified to Increase the maximum flow rate. A fan the same size as the
original fan was Installed 1n series (see Figure 40). The blades were
Installed to operate in counter-rotation to the original fan 1n order
to Increase efficiency. This modification provided a 2.9% Increase in
maximum flow or a maximum average velocity of 8.3 m/s (27.3 ft/sec) in
the circular section and 3.5 m/s (11.6 ft/sec) 1n the square section.
b. Testing Program
A series of tests (Test No. 2 through Test No. 13) were
conducted with different velocities and concentration profiles. Several
sampling methods were evaluated. In all cases, the tracer gas was Intro-
duced at about 16.2 x 10"6 m3/s (-1 I1ter/m1n).
Test No. 2 (see Figures 41, 42, and 43) is an evaluation
of the air flow through the wind tunnel measured by traversing the square
as well as the round section using the pitot tube and also using the Annubar
element 1n the round section. The calculated flow was surprisingly very
close In all cases, as shown in Table 15.
Test Nos. 3 and 5 were conducted on the square and the
round sections of the wind tunnel, respectively, In both cases almost all
conditions were kept the same. The calculated flow from both traverses
agreed within 3.B%
The total emission as calculated from the round section was
about 6$ higher than expected and from the square section about 10% lower
than expected. This discrepancy may be due to the values of the velocities
obtained when reverse conditions predominated. Results are shown in
Figures 44-a, b, c and 45-a, b, c and Table 16.
Table 16 compares the value of the total emission obtained
from Test No. 6 by using the 'ANNUBAR1 sampling port (see Figure 46 a and b)
for sampling vs. the ethane Introduced. An error of about -34% was observed.
142
-------
To Exhaust
Honeycomb stralrhteners
Figure 40. Wind Tunnel Plan View
(Two Fans)
143
Scale ~1:40
Section Description
A
AN
B
E
F
FI
H
I
P
R
S
Air tank
Annubar Element
Baratron
Ethane Tank
Fan
FID
Hydrogen Tank
Ethane Injection Port
Sampling Pump
Rotameter
Sampling Ports
-------
TEST NO. 2
Conditions:
Baffle: Inclined -22.5° to horizontal
Flow Area: 0.073 m2 (0.7854 ft2)
Temperature: 18°C
*
A
B
C
1
1.5
(4.86)
1.3
(4.21)
1.3
(4.35)
2
3.1
(10.31)
3.9
(12.86)
3.1
(10.31)
3
4.1
(13.31)
4.8
(15.75)
4.4
(14.58)
4
2.3
(7.68)
2.9
(9.41)
2.6
(8.42)
Port
Port
* Velocity 1n m/s
Total flow: 0.21 m3/s (455.73 ft3/m1n) at 18°C
or. 0.2 m3/s (427.57 ft3/min) at 0°C
Figure 41. Velocity Distribution In Round Section
144
-------
TEST NO. 2
CONDITIONS: Baffle: Inclined -22.5° to horizontal
Flow area: 0.7854 (ft2)
Temperature: 18°C
Annubar reading
*P (mrcHg) 0.085
Temperature T
°R 524.4
0.0765 x 520
T 0.07585
Element Constant S 0.68
Formula Used
Qn = 585.24 x S x
S
2UU!2f. m*TZ
427.5 at 32°F
Figure 42. Average Velocity Using the Annubar Element In Round Section
145
-------
TEST NO. 2
CONDITIONS:
Baffle: Inclined 522.6° to horizontal
Flow Area: 0.18 nr (280 1n2)
Temperature: 20°C
*
A
B
C
D
E
F
G
1
1.5
1.8
1.8
1.6
1.3
~0
-5.9
2
1.7
1.8
2.0
2.0
1.8
1.1
-1.5
3
1.8
1.9
2.0
2.2
2.1
1.4
1.3
4
2.0
2.1
2.3
2.3
2.1
1.8
1.5
5
2.2
2.2
2.5
2.1
1.4
1.2
0.8
6
2.0
2.1
2.3
1.6
0.8
~0
-1.0
7
2.0
2,2
2.2
1.8
~o
-2.8
-5.1
* Velocity 1n m/s
Total flow 0.217 m3/s (459.10 ft3/m1n) at 20°C
or 0.20 m3/s (427.80 ft3/m1n) at 0°C
Figure 43. Velocity Distribution In Square Section
146
-------
TABLE 15
Traverse A1r Flow Reduced To
Section 0°C m'/s (32°F ftVmin)
Remarks
Square
0.20
(428)
P1tot tube traverse, reverse flow
was accounted for by subtracting
Circular
Circular
0.20
0.20
(428)
(428)
P1tot tube traverse
Average velocity 1s obtained
using the ANNUBAR element
TABLE 16
Traverse
Section
Square
Circular
Circular
A1r Flow Reduced To
0°C m3/s (32°F ftVmin)
0.197 (418)
0.205 (435)
0.202 (427)
Ethane Introduced
kg/s
20.370 x 10'6
20.370 x 10"6
20.370 x 10'6
Emission Rates
Calculated kg/s
18.240 x 10"6
18.640 x 10"6
18.517 x 10"6
using
"ANNUBAR" element*
Sampling from high pressure side tube or sampling from sampling tube -
See Figure 9-b for details.
147
-------
TEST NO. 3
CONDITIONS:
Baffle: Inclined 22.5° to horizontal g
Ethane Flow rate: 0.97 l/m1n » 20.3 x 10 kg/s
Injected after baffle in square section
Temperature: 20°C 2
Flow Area: 0.18 mz (280 1n )
*
AV*
M c**
B V
B C
'I
°l
*l
'I.
«Vc
1
1.4
288
1.7
293
1.8
268
1.7
218
1.3
158
0
143
-5.1
183
2
1.6
290
1.6
280
1.8
265
1.8
205
1.8
155
1.1
135
-1.5
145
3
1.8
305
1.8
265
2.0
230
2.1
165
2.0
120
1.4
120
1.3
140
4
1.9
295
2.1
250
2.3
180
2.2
120
1.8
95
1.4
100
1.1
115
5
2.1
330
2.3
275
2.4
210
2.2
130
1.5
105
1.4
110
1.1
.130
6
1.9
338
2.2
198
2.3
208
1.7
128
0.8
113
0
113
-1.1
148
7
1.9
248
2.2
148
2.0
138
1.5
128
0
128
-2.3
168
-4.9
203
* Velocity In m/s
** concentration must be multiplied by 0.3347 to convert to ppm ethane
T6tal Flow = 0.212 m3/s (448.67 ft3/min) or 0.197 m3/s (418.07 ft3/m1n) at 0°C
Total Emission 18.240 x 10"6 kg/s
Figure 44a. Velocity and Concentration Distribution Data 1n Square Section
148
-------
TEST CONDITIONS: Baffle: Inclined 22.5°
Total Flow: 0.21 (-450 ft3/m1n) at 20°C
Flow Area: 0.18 m2 (280 1n2)
6 7
Flow Direction
Reverse Flow
Scale: 1cm = 0.61 m/s
(2 ft/sec)
Figure 44b. Velocity Distribution in Square Section
149
-------
TEST CONDITIONS:
Flow
Direction
Baffle: Inclined 22.5°
Total Flow: 0.2 m3/s (-450 ft3/m1n) at 20°C
Flow Area: 0.18 m2 (280 1n2)
Ethane Flow Rate: 16.2 x 10"6 m3/s (0.97 l/m1n)
Duct Wall
/ . . 1 • •
f
, *
Scale: 1cm - 35 ppm
Figure 44c. Concentration Distribution of Ethane In Square Section
150
-------
TEST NO. 5
Conditions:
Baffle: Inclined 22.5° to horizontal fi
Ethane Flow rate = 0.97Vm1n = 20.37 x 10"° kg/s
injected after honeycomb in
round section
Temperature = 18°C 9
Flow Area = 0.073 m3 (.07854 ft*)
n V*
« C**-
B V
B c
c v
L c
1
1.4
540
1.2
740
1.2
790
2
3.3
90
3.9
45
3.4
40
3
3.9
130
4.8
125
4.6
no
4
2.7
340
2.8
470
2.7
500
Port
Port
* Velocity in m/s
** Concentration must be multiplied by 0.3347 to convert to ppm ethane
Total Flow = (464.40 ft3/min) or (435.20 ft3/m1n) at 0°C
Total Emission = 21.64 kg/s
Figure 45a. Velocity and Concentration Distribution Data in Round Section
151
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.219 m3/s (465 ft3/min)
Flow Area: 0.073 m2 (0.7854 ft2)
Flow
Direction
Scale: 1 cm
0.61 m/s
(2 ft/sec)
Figure 45b. Velocity Distribution in Round Section
152
-------
TEST CONDITIONS:
0.219 m3/s (465 ft3/m1n)
Baffle: Inclined 22.5°
Total Flow:
Flow Area: 0.073 m2 (0.7854 ft2)
Ethane Flow Rate: 16.2 x 10"6 m3/s (0.97 1/min)
Flow
Direction
Wall
Scale: 1 cm • ~ 35 ppm.
Figure 45c. Concentration Distribution of Ethane in Round Section
153
-------
TFST NO. 6
Conditions:
Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate = 0.97£/min =20.370 x 10
Temperature 18°C «
Flow area = 0.073 m2 (0.7854 fir)
Flow rate = 0.21 m3/s (454 ft3/min)
06
kg/s
Special Condition
Well mixed sample
Sampling from cap
Sampling from H.
**
Sampling position
(No. 1)
press
No. 2
2r3
3
4
5
6
7
8
9
10
11
12
Concentration*
220
150
150
1
160
220
290
370
330
340
310
340
330
300
155
145
* Concentration must be multiplied by 0.3347 to convert to ppm ethane
** Sampling position No. 1 is holes facing stream flow as in Figure
other positions, holes facing stream at an angle
Figure 46a. Average Concentration Using the "Annubar Element"
154
-------
Down Stream
Pressure Tube
Interpolating High
Pressure Tube
Tygon Sampling
Tube
Cap With
Sampling Hole
0.305 m (12") Circular Section
0.025 m (r) Dia. Tube .188" Wall
Low Pressure
High Pressure
Tygon Sampling Tube
Steel Frame Stand
Figure 46b. Annubar Flow Element
-------
In Test No. 6, ethane was well mixed by introducing it into
the square section port and sampling in the round section after the fan
using the 'ANNUBAR1 element. Since no concentration stratification existed,
the correct emission was obtained as expected.
The tracer gas was then introduced after the honeycomb in the
round section and the 'ANNUBAR1 element was used as a sampling probe to
obtain an average sample at different positions. The first position of the
element was with the holes facing the flow; other positions were with the
element at different angles to the flow obtained by rotating the element
360° about the longitudinal axis. It was found that one position gave the
correct average concentration, i.e., equal to the mixed sample concentration.
This particular position was No. 2-3, where the holes faced the stream at
approximately a 45° angle.
These tests show that for this particular velocity and con-
centration profile in the duct, a specific position of the Annubar was
found to give the average concentration. However, this position is not ex-
pected to provide the average concentration for another set of profiles
As discussed in SectlonlV .B.l.b. of this report, it is gen-
erally necessary to use proportional samplers. That is, gas concentration mea-
surements at a sampling point must be weighted by the local gas velocity
and the area ascribed to that probe. One method of satisfying this con-
dition is to sample at equal flow rates for times proportioned to the local
stack gas velocity. This method reduces the emission rate equation to:
1=n
2*i
1=1
156
-------
where:
E = emission rate
Aj = area ascribed equally to all probe's location
V, V, tf
K, • proportional constant equal to -r- = T=- • "* • —
1 'i *2 tn
C - average concentration of the mixed samples
t.j = sampling time for each sample location
Test Nos. 7 and 8 were conducted to demonstrate this method.
The sampling set up 1s sketched in Figure 47-a. Sampling from all locations
was done sequentially at an equal rate. A stop watch was used to measure
the sampling time which had been calculated previously from velocity profile
data. The average concentration was obtained by collecting the total sample
In a mylar bag and then sampling the mixed contents with a F.I.D.
Test results are given in Figures 47-b and 47-c. Both tests
were performed at the same locations across a traversing port of the square
section of the wind tunnel; the only difference was the proportionality
constant K. Results from Test Nos. 7 and 8 show a +2% and -6% discrepancy
in emission rate respectively, from that calculated from each sampling loca-
tion velocity and concentration traverse.
Test No. 9-1 was run to determine the velocity and concentration
Profiles generated in the square section using the high speed wind tunnel as
shown 1n Figure 40. Results of the test are shown in Figures 48-a,b. A 91
Point traverse* was performed and the average velocity was 3.5 m/s (11.6 ft/sec
ft/sec) with a total flow of 0.64 m3/s (1356 ft3/min.). The calculated
total emission was 1.935 higher than the Injected value.
•^""•"^•«**»IB»
Numbers 1,2,3,...7 are the original 49 centroid of equal area sampling
locations. Number 1-2, 3-3, .. 6-7 are located respectively mid-way
between 1 and 2, 2 and 3 .. 6 and 7.
157
-------
Mylar Bag
Wind Tunnel '
Sampling Probe
Diaphragm Pump
Stop Watch
Figure 47a. Experimental Set-up for Manually Adjusting Sampling Time
At Each Sampling Position
158
-------
TEST NO. 7
Conditions:
Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate: 16.2 x 10'6 m3/s (0.97*/m1n)
Temperature: 16°C
Flow Area 0.18 m2 (280 1n2}
va
A *b
R V
B C
« c
*l
' I
'}
r V
6 C
y.1-7
[ ^1-1
prom bag
Cau
K
Ed.
K+zVcau
or zvjCj
Port No.
5
2.10
330
2.26
280
2.44
205
2.10
135
1.49
115
1.41
120
1.15
130
205
1/4
.*
*,c
27.60
29.60
32.00
27.60
19.50
18.50
15.10
169.10
2654
V1C1
693
632
500
283
171
168
150
2597
2597
Footnotes
a velocity m/s
b concentration must be
multlpHled by 0.337 to
convert to ppm ethane
c time in seconds
d emission rate (not
converted
Discrepancy + 2%
Figure 47b. Sampling 1n Square Section at Constant Flow Rate With
Manually Adjusting-Sampling Time for Each Position
159
-------
TEST NO. 8
Conditions: B Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate: 16.2 x 10"6 m/s
Temperature: 16°C
Flow Area: 0.18 m2 (280 in2)
Va
A *b
B V
B C
'S
»?
'cv
'I
r v
G C
W
1*1
from bag
Cau
K
Ed =
IWEtjXC^
or_ Zv^
Port No.
5
2.10
335
2.25
280
2.44
2.5
2.10
135
1.49
115
1.41
125
1.15
135
190
1/8
t c
M
55.1
59.2
64.1
55.1
39.0
37.0
30.2
339.8
2460
V1C1
703
630
500
284
171
176
155
2619
Footnotes
a velocity m/s
b concentration
c time in seconds
d emission rate (not
converted)
Discrepancy - 6%
Figure 47c.
Sampling 1n Square Section at Constant Flow Rate With
Manually Adjusting Sampling Time For Each Position
160
-------
Figure 433. Velocity and Concentration Distribution Data in Square Section
TEST NO. 9
A Vl
C2
B V
C
C V
C
D V
C
£ V
C
F V
C
G V
C
1
2.1
85
2.6
95
2.8
105
2.3
120
2.3
100
2.6
80
2.9
60
1-2
2.1
90
2.1
95
2.8
115
2.6
115
2.3
105
3.0
80
3.5
55
2
2.1
85
2.8
106
2.6
125
2.6
125
2.8
110
3.2
85
4.0
55
2-3
3.0
95
3.0
115
3.1
125
3.0
125
3.0
120
3.3
90
4.0
60
3
4.1
105
3.1
120
3.1
130
3.3
130
3.3
115
3.6
90
4.0
60
3-4
3.1
no
3.6
120
3.8
130
3.9
125
3.4
115
3.4
100
4.0
65
4
3.5
no
3.7
120
4.0
130
4.2
120
3.8
100
3.8
105
4.0
70
4-5
3.5
115
3.8
115
3.9
120
4.4
105
3.9
95
3.9
90
3.9
70
5
3.5
120
3.6
110
4.1
115
4.5
105
4.4
85
4.1
70
3.6
60
5-6
3.0
95
3.5
90
4.2
90
4.9
80
4.6
65
3.9
60
3.6
50
6
2.8
100
3.2
85
4.1
90
4.9
70
5.1
50
4.6
45
3.8
45
6-7
2.3
85
3.0
65
4.3
65
4.7
60
5.0
35
4.9
30
3.8
30
7
2.3
95
3.0
75
4.0
65
4.7
55
4.7
35
4.6
25
3.8
25
CONDITIONS: Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate: 20.23 x ID"6 kg/s
Temperature: 20°C
Flow Area: 0.18m2 (280m2)
Number of Fans: 2 fans in series
velocity m m/s
concentration must be multiplied by 0.298 to convert to ppm ethane: Average Velocity * 3.5 m/s (11.6 ft/sec)
Total Flow = 0.64 m3/s (1356 ft3/min) at 20°C Total Emission = 20.62 x 10"6 kg/s
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.64 m3/s (1355 ft3/min)
Flow Area: 0.18 m2 (280 in2)
Average Velocity: 3.5 m/s (11.6 ft/sec)
ro
Scale:
-9
1 cm - 45.05 x 10 ' kg/s
Flow
Direction
Figure 4"8b. Velocity and Concentration Distribution In Square Section
-------
Test No. 9-2 was run on the circular section of the wind
tunnel using the traversing ports and locations* as shown 1n Figure 49-a.
The velocity and concentration distributions are shown In Figures 49-b,c
respectively. The calculated average velocity was 7.9 m/s (26 ft/sec),
giving a total flow of 0.575 m3/s (1218 ft3/m1n) which 1s approximately
10% less than the calculated flow from the square section. The total
emission rate v,
the real value.
emission rate was 23.66 x 10 kg/s (0.02366 gin/sec), about 17% higher than
Because of the considerable disagreement between these and
the square section results, Test No. 9-3 was run using the traversing ports
rotated at a 45° angle as shown 1n Figure 50-a using the same radial loca-
tions. The velocity and concentration distrubutlons are shown 1n Figures
50-b,c, respectively. The calculated average velocity was 8.2 m/s
(27 ft/sec), giving a total flow of 0.606 m3/s (1284 ft3/m1n), which 1s
about 5% less than the calculated flow from the square section. The total
emission rate was 25.06 x 10"6 kg/s (0.02506 gm/sec.), about 24% higher
than the real value.
The results of Test No. 9-4 using the Annubar element are
shown 1n Figures 51-a,b, The total flow was equal to 0.515 m3/s
(1092 ft3/m1n), approximately 19% lower than the value obtained from the
square section. During this test, the tracer gas was Injected before the
fan location to obtain a mixed sample. After a concentration profile was
generated, the tracer gas was injected after the fan location. Sampling
from the Annubar element was done through the tap opposite the pressure
taps at a rate of about 50 x 10"6 m3/s (3 I1ters/m1n).
* Rad11 to stations 1n zone A, b and C respectively are 0,063 m (2.5"),
0.108 m (4.25"), and 0.140 m (5.5"). Stations locations 1n zone g-A,
A-B and B-C are located mis-way between g and A, A and B, and, B and C.
163
-------
TEST NO. 9-2
Conditions:
Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate = 0.97 fc/min = 20.23 x 10
Temperatures 20°C
Flow Area
No. of Fans 1n Series • 2
-6
kg/s
0.073 m3 (0.7854 ft2)
/
s*l**
*Vc
*'*l
B V
c
B-CJ!
r V
C C
1
6.1
195
4.9
160
4.3
20
4.2
20
4.2
10
4.1
5
2
10.0
5
10.7
0
11.1
0
11.6
0
11.1
0
9.5
0
3
9.0
15
9.2
5
8.3
0
7.7
0
7.3
0
6.7
4
4
c 8.1-8.3
£50-85
8.2
260
8.4
460
8.2
560
8.1
425
7.8
260
Port 1
•rt 2
* Velocity 1n m/s
** Concentration must be multiplied by 0.298 to convert to ppm ethane
1 Dlsreguard significant figures in table
Total flow = 0.515 m3/s (1218 ft3/min)
Total emission = 23.66 x 10 kg/s
Figure 49a. Velocity and Concentration Distribution Data 1n Round Section1
164
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.57 m3/s (1218 ft3/m1n)
Flow Area: 0.073 m2 (0.7854 ft2)
No. of Fans in Series: 2
Flow
Direction
Scale: 1 cm = 1.5 m/s (5 ft/sec)
Figure 49b. Velocity Distribution In Round Section
165
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.575 m3/s (1218 ft3/min)
Flow Area: 0.073 m2 (0.7854 ft2)
Ethane Flow Rate: 16.2 x 10"6 m3/s (0.97 ft/min)
No. of Fans In Series: 2
Flow
Direction
Duct Wall
Scale: 1 cm = - 35 ppm
Figure 49c. Concentration Distribution of Ethane in Round Section
166
-------
TEST NO. 9-3
Conditions:
Baffle: Inclined 22.5° to horizontal
Ethane Flow Rate =0.97 i/m1n = 20.23 x 10
Temperature 20°C
Flow Area 0.073 m2 (0.7854 ft2)
No. of Fans in Series = 2
-6
kg/s
y*
SAC**
*cV
*-BVc
•!
B-cc
c v
c c
1
7.5
50
7.4
30
7.7
7
8.3
2
8.1
0
7.6
0
2
9.8
3
10.7
0
11.1
0
11.1
0
11.3
0
10.6
0
3
8.6
25
8.6
10
8.1
2
7.9
2
8.0
1
8.2
0
4
£8.1-7.1
£50-270
6.6
750
6.3
870
6.2
650
6.3
370
6.2
150
* Velocity in m/s
** Concentration must be multiplied by 0.298 to convert to ppm ethane
1 Disregard significant figures in table
Total flow * 0.606 m3/s (1284 ft3/min) using 25 points traverse
Total emission = 26.06 kg/s
Figure 50a. Velocity and Concentration Distribution Data 1n Round Section1
167
-------
TEST TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.606 m3/s (1284 ft3/min)
Flow Area: 0.073 m2 (0.7854 ft2)
No. of Fans in Series: 2
Flow
Direction
Scale: 1 cm = 1.52 m/s (5 ft/sec)
Figure 50b. Velocity Distribution in Round Section
168
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow:
Flow Area:
Ethane Flow Rate: 16.2 x 10"6 m3/s (0.97 1/rnln)
No. of Fans 1n Series: 2
0.606 m3/s (1284 ft3/min)
0.073 m2 (0.7854 ft2)
Flow
Direction
Scale: 1cm tt - 35 ppm
Figure 50c. Concentration Distribution Of Ethane In Round Section
169
-------
TEST NO. 9-4
Conditions: Baffle: Inclined 22.5° to horizontal
Flow Area: 0.073 m2 (0.7854 ft2)
Temperature: 20°C
No. of Fans in Series = 2
Annubar reading
AP (mm Hg)
Temperature TQ
K
Iff
0,0765 x 520
T
Element Constant S
Formula Used*
Total Flow at T°
Qn, ft3/m1n
m3/s
* On = 7.897 x S x 0.7576
0.49
528
0.07534
0.58
nn . 585.24 v ,
\^ll •"••••••^^^ ^ ^
1092
0.515
x Ml.994)2 x 1.0 x
. 1 i| AP
j_j»i I 1 .8663
/TT f
1
i n „ 1 ..if AP(mmHg)
1.866
Figure 51 a. Average Velocity Using the Annubar Element 1n Round Section
170
-------
TEST NO. 9-4.
Conditions:
Baffle: Inclined 22.5° to horizontal
Flow Area: 0.073 m2 (0.7854 ft2)
Temperature: 20°C
No. of Fans 1n Series: 2
No.
of
Fans
1
2
Sample*
Mixed
Not Mixed
Not Mixed
Not Sampling
Mixed
Not Mixed
Not Sampling
Not Mixed
Not Sampling
Velocity press, head
N/m2 (mm Hg)
28.67 (0.215)
28.67 (0.215)
7.33 (0.055)
8.00 (0.060
62.67 (0.47)
62.67 (0.47)
65.33 (0.49)
32.00 (0.24)
33.33 (0.25)
Cone, of
Tracer
Gas**
no
10
no
85
10
-
85
Approx. angle of
rotation of element
to obtain Av. Cone.
0°
0°
55° upward
55° upward
0°
0°
0°
45° upward
45° upward
* Sampling rate - 50 x 10"6 m3/s ( 3 liters/mln)
** To convert to ppm multiply by 0.298
Figure 51b. Average Velocity and Concentration Using an Annubar Element
For Sampling
171
-------
As shown in Figure 14b, the Annubar element was also used
for sampling. When the element was not rotated, the total emission rate
was about 88% lower than the real value. By rotating the element, the
average concentration was obtained at an angle of approximately 45° upward,
while using the two fans in series. It is significant that the value of the
velocity head (1n mm Hg) measured from the Annubar element was virtually
unaffected by simultaneously sampling from the reference tap. This indicates
that for special cases of flow and concentration profiles, this instrument
can be used to obtain a representative gas sample and total flow simultaneously.
However, since in all cases the Annubar was rotated to obtain the correct
average concentration value (as determined by total gas flow and tracer
injection rate), this approach Is applicable only in cases where the profiles
remain constant and the apparatus can be aligned on the basis of an Indepen-
dent knowledge of the correct average concentration.
Results of Test Nos. 10-1 through 10-3 are shown in Figures 52-a
through Figure 54-b. A 16 point traverse was done on the square section
using 3 different sampling methods. The Chebyshef method gave an average
velocity approximately 2% higher than the 91 point traverse method, and an
emission rate of 3% less. The Centroid of equal area method gave an average
velocity of 2% less and an emission rate of 6% less. The Circular analog
method gave a much higher average velocity as well as emission rate, 17% and
2U higher, respectively, than the 91 point traverse method.
Three sets of tests were run for different velocity and con-
centration profiles. These profiles were generated by operating one or
both fans and/or changing the wind tunnel damper position. In two sets of
tests, three methods of sampling were Investigated in the square section:
the Chebyshef method, the Centroid of Equal Area method and the Circular
Analog method, for a 16 point traverse. The data from the first set of
tests, Test No. 11, are shown in Figures 55a, b, and c. The emission rate
as calculated from the Chebyshef method for the 16 point traverse was about
8% less than the Injected amount, while the average velocity was about
172
-------
TEST NO. 10-1
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: fully open
Temperature: 22°C
Flow Area: ,0.18 m2 (280 1nz)
Ethane Flow Rate: 0.97 £/m1n = 20.096 x 10"6 ka/s
No. of Fans in Series: 2
Average Velocity: 3.6 m/s [11.8 ft/sec)
Emission Rate: 19.49 x 10'6 kg/s
1234
ft v*
* c**
R V
Bc
r v
L C
n V
°C
1
3.3
100
3.4
120
2.2
90
1.1
75
2
3.3
125
3.9
130
3.8
110
4.2
60
3
3.9
90
4.4
105
4.4
90
4.2
60
4
2.1
60
4.5
55
4.8
35
4.2
15
* Velocity in ft/sec
** Concentration must be multiplied by 0.298
to convert to ppm
1 Disregard significant figures In Table
B
Percent
Distance
10.27
40.72
59.28
89.73
Distance (m\ -
0.044
0.173
0.252
0.382
(1.72")
1 ' • r fe /
(6.82")
(9.92")
(15.03")
Figure 523. Velocity and Concentration Distribution Data 1n Square Section
Using the Chebyshef Method for Sfxteen Point Traverse'
173
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.65 m3/s (1380 ft3/min)
Flow Area: 0.18 m2 (280 1n2)
Average Velocity: 3.6 m/s (11.8 ft/sec)
Scale:
-9
I on - 45.05 x 10 * kg/s
B
Flow Direction
Figure 52b. Velocity * Concentration Distribution in Square Section Using the Chebyshef Method
-------
TEST NO. 10-2
Conditions: Baffle: Inclined 22.5° to horizontal
Damper: Fully open
Temperature: 22°C
Flow Area: 0.18 m? (280 ft2)
Ethane Flow Rate: 0.97 £/m1n = 20.096 x 10"6 koA
No. of Fans In Series: 2 y
Average Velocity: 3.5 m/s ill.40 ft/sec)
Emission Rate: 18.85 x 10"° kg/s
a v*
M c**
B V
0 C
p V
L c
D V
C
1
3.2
95
3,2
125
1.5
85
0.7
70
2
3.4
130
3.5
135
3.7
100
4.1
65
3
3.5
90
4.1
100
4.7
80
4.3
70
4
2.1
70
4.1
55
4.9
45
4.6
20
1234
A
B
*
•
•
•
*
4
•
•
•
•
•
•
•
.
•
.
•
* Velocity 1n ft/sec
** Concentration must be multiplied by 0.298
to convert to ppm
1 Disregard significant figures In Table
Figure 53a. Velocity and Concentration Distribution Data In Square Section
Using the Centrold of Equal Area Method for Sixteen Points
Traverse1
175
-------
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.63 m3/s (1335 ft3/min)
Flow Area: 0.18 m2 (280 1n2)
Average Velocity: 3.5 m/s (11.40 ft/sec)
Scale:
-9
1 cm - 45.05 x 10 * kg/s
B
Flow
Direction
Figure 53b. Velocity * Concentration Distribution in Square Section Using the Centroid of
Equal Area Method
-------
TEST NO. 10-3
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: Fully open
Temperature: 22° C0 «
Flow Area: 0.18 m2 (280 ft*)
Ethane Flow Rate: 0.97 *-/m1n.= 20.096 x 10
No. of Fans 1n Series: 2
Average Velocity: - 4 m/s [13.00 ft/sec)
Emission Rate: 24.15 x 10'6 kg/s
kg/s
AV*
M c**
B V
B C
1
3.2
115
3.3
100
2
3.3
120
3.1
110
3
4.1
120
3.2
115
4
3.9
120
3.8
120
5
4.1
105
4.5
100
6
4.2
85
4.9
75
7
4.5
65
4.9
60
8
3.8
60
4.5
30
* Velocity 1n m/s
** Concentration must be multiplied by 0.298 to
convert to ppm ethane
1 Disreguard significant figures 1n table
1
•— '• «^
2
b «_
1
.. .j
ak
r-jH
.1
2
3
h?-
5
6
7
8
6
• • •—
7
• » •
8
. V-M
% Distance
Across Duct
Distance (m)
3.4
10.7
19.8
36.0
64.0
80.2
89.3
96.6
0.014
0.045
0.084
0.914
0.272
0.340
0.380
0.411
(0.57")
(1.80")
(3.30")
(6.00")
(10.70")
(13.40")
(14.96")
(16.20")
Figure 54a. Velocity and Concentration Distribution Data in Square Section
Using the Circular Analog Method for Sixteen Points Traverse
177
-------
CD
TEST CONDITIONS:
Baffle: Inclined 22.5°
Total Flow: 0.717 m3/s (1520 ft3/min)
Flow Area: 0.18 m2 (280 in2)
Average Velocity: - 4 m/s (13.0 ft/sec)
Scale:
-9
1 cm - 45.05 x 10 * kg/s
Flow
Direction
Figure 54b. Velocity * Concentration Distribution in Square Section Using The Circular Analog Method
-------
TEST NO. 11-1
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: Fully open
Temperature: 20°C,) ?
Flow Area: 0.18 nr (280 In ) fi
Ethane Flow Rate: 0.97 */m1n = 20.23 x 10"b kg/s
No. of operating fans: 1
Average velocity: 2.6 m/s (8.4 ft/sec)
Emission rate: 18.57 x 10'6 kg/s
1234
AV*
« c**
•2
r V
U C
°t
1
-8.3
115
2.3
160
2.7
170
3.0
no
2
0.0
145
4.1
155
3.8
130
4.1
105
3
2.9
120
4.5
120
4.5
85
2.1
85
4
3.3
105
4.3
75
4.9
40
2.6
90
/*""
A
B
>
P
D
.
•
.
•
,
.
.
•
.
.
.
•
,
.
.
.
ercent D
1 stance Distance (m)
10.27 0.044 |
40.72 0.173
59.28 0.252
89.73 0.382 <
;i.72")
6.82")
9.92")
15.03")
* Velocity 1n m/s
** Concentration must be multiplied by ~0.31 to convert to ppm
Note: expected average velocity from average concentration measurement
of a mixed sample equal to 2.6 m/s (8.5 ft/sec)
Figure 55a. Velocity and Concentration Distribution Data in Square Section
Using the Chebyshef Method for Sixteen Point Traverse
179
-------
TEST NO. 11-2
Conditions
Baffle: Inclined 22.5° to horizontal
Damper: Fully open
Temperature: 20°C 9
Flow Area: 0.18 m2 {280 In*)
Ethane Flow Rate: 0.97 Jl/mln = 20.23 x 10
No. of Operating Fans: 1
Average Velocity: 2.6 m/s (8.5 ft/sec)
Emission Rate: 19280 x 10"° gm/sec
kg/s
1234
.
AV*
rt c**
B V
8 C
r v
1 C
r
i
-7.4
120
2.0
155
2.6
165
2.9
120
2
1.1
150
3.9
150
3.5
140
4.1
115
3
3.3
120
4.5
105
4.3
80
2.1
85
4
3.6
100
3.9
95
4.9
45
2.1
100
A
B
* Velocity 1n m/s
** Concentration must be multiplied by -.0.3 to convert to ppm
Note: expected average velocity from average concentration measurement of
a mixed sample equal to 8.5 ft/sec.
Figure 55b. Velocity and Concentration Distribution Data 1n Square Section
Using the Centroid of Equal Area Method for Sixteen Points
Traverse
180
-------
TEST NO. 11-3
Conditions: Baffle: Inclined 22.5° to horizontal
Damper: Fully open
Temperature: 20°C« «
Flow Area: 0.18 nr (280 in*) fi
Ethane Flow Rate: 0.97 Jl/min = 20.23 x 10"° kg/s
No. of Operating Fans: 1
Average Velocity: 3.7 m/s (12.2 ft/sec)
Emission Rate: 27.03 x 10'6 kg/s
A V*
M c**
B V
B C
1
0.7
135
2.4
170
2
2.6
140
2.7
170
3
3.8
140
3.0
165
4
4.5
;145
3.8
.150
5
4.1
100
4.9
90
6
4.2
100
5.0
65
7
4.1
90
4.9
50
8
4.2
75
4.5
35
* Velocity 1n m/s
** Concentration must be multiplied by -0.30
to convert to ppm.
Note: expected average velocity from average
concentration measurement of a mixed
sample equal to 2.6 m/s (8.5 ft/sec)
B-E
%—
2
— .—
J
f
3
" •""
4
• 9*mm
-|
1
2
3
4 5
"^w»- •*
5
6
6
•••«•
7
7
f •••
8
8
•» • •
% Distance
Across Duct
3.4
10.7
19.8
36.0
64.0
80.2
89.3
96.6
Distance
in Inches
0.014 (0.57"
0.046 (1.30"
0.084 (3.30"
0.914 (6.00"
0.272
0.34
0.38
0.411
10.7"
13.4"
14.96")
^16. 2")
Figure 55c. Velocity and Concentration Distribution Data in Square Section
Using the Circular Analog Method for Sixteen Points Traverse
181
-------
the same as that calculated from the known value for the total flow
through the square section. The Centrold of equal area method gave an
emission rate of 5% less than the Injected amount, while the average
velocity was practically the same as in the previous method. By using
the Circular analog method for 16 point traverse, the emission rate was
found to be higher than the known value by approximately 34% and the
average velocity was over-estimated by about 43%. This considerable
agreement 1s due partially to the inadequacy of the Circular analog
method in accounting for the reverse flow occurring near the corner of
the square duct.
The data summarized for the second set of tests, Test No. 12,
are shown in Figures 56a, b, and c. The emission rate calculated from the
Chebyshef, the Centroid of equal area and the Circular analog methods dif-
fered by -42%, -32%, and +28%, respectively, with the known value of the
Injected tracer gas. The average velocity calculated from the Chebyshef
traverse method was 65% less than the actual average velocity and 50% less
when the Centroid of equal area method was used, but exceeded by 18% the
actual value when the Circular analog method was used. This large discrepancy
is probably due to highly reversed flow conditions in the duct.
Test No. 13 was run on the circular section with the damper
partially open. A 25 point traverse, on two diameters Including the center
line value, was sampled, and results are given in Figure 57. The calculated
emission rate was about the same as the known injection value, although the
average velocity was found to be 25% less than the expected average value.
182
-------
TEST NO. 12-1
Conditions: Baffle: Inclined 22.5° to horizontal
Damper: Partially open
Temperature: 20°C
Flow Area: 0.18 mz
Ethane Flow Rate: 0.97 fc/nrin = 20.23 kg/s
No. of Operating Fans: 2
Average Velocity: 1.1 m/s (3.6 ft/sec)
Emission Rate: 11.7 x 10'6 kg/s
AV*
* c**
*t
r V
U C
°l
1
-8.2
100
3.6
115
-0.0
105
•8.9
85
2
2.1
115
4.2
140
3.9
130
4.5
100
3
3.8
150
4.1
140
4.1
110
3.9
90
4
0.7
115
1.6
80
1.1
40
-2.6
35
* Velocity 1n m/s
** Concentration must be multiplied by -0.3 to convert to ppm
Note: expected average velocity from average concentration measurement of
a mixed sample equal to 7.6 m/s (25 ft/sec)
Figure 56a. Velocity and Concentration Distribution Data in Square Section
Using the Chebyshef Method for Sixteen Point Traverse
183
-------
TEST NO. 12-2
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: Partially open
Temperature: 20°C «
Flow Area: 0.18 mz (280 in*) 6
Ethane Flow Rate: 0.97 fc/m1n = 20.23 x 10 ° kg/s
No. of Operating Fans: 2
Average Velocity: 1.6 ms/ (5.3 ft/sec)
Emission Rate: 13.78 x 10'6 kg/s
AV*
H c**
*l
r V
C C
°Vc
1
-5.8
100
3.9
115
0.0
105
-5.9
85
2
2.1
115
4.1
140
3.8
125
4.6
85
3
3.6
155
4.1
140
4.1
100
3.6
95
4
0.9
130
1.6
70
1.1
40
0.0
35
* Velocity 1n m/s
** Concentration must be multiplied by - 0.3 to convert to ppm
Note: expected average velocity from average concentration measurement
of a mixed sample equal to 7.6 m/s (25 ft/sec)
Figure 56b. Velocity and Concentration Distribution Data 1n Square Section
Using the Centroid of Equal Area Method For a Sixteen Point
Traverse
184
-------
TEST NO. 12-3
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: Partially open
Temperature: 20°C9 «
Flow Area: 0.18 n£ (280 in*)
Ethane Flow Rate: 0.97 Vm1n H 20.23 x 10"
No. of Operating Fans: 1
Average Velocity: 3.6 m/s (11.8 ft/sec)
Emission Rate: 25.92 x 10"6 kg/s
kg/s
A V*
« c**
B V
c
1
3.5
150
0.4
115
2
3.3
120
2.3
no
3
3.9
140
3.6
120
4
4.3
145
4.1
135
5
4.0
120
4.3
120
6
3.9
no
4.3
85
7
3.9
105
3.9
40
8
4.1
80
3.6
35
* Velocity 1n m/s
** Concentration must be multiplied by - 0.3 to convert to ppm
Note: expected average velocity from average concentration measurement
of a mixed sample equal to 7.6 m/s (25 ft/sec)
Figure 56c. Velocity and Concentration Distribution Data 1n Square Section
Using the Circular Analog Method for Sixteen Points Traverse
185
-------
TEST NO. 13
Conditions:
Baffle: Inclined 22.5° to horizontal
Damper: Partially open
Temperature 20°C 9 9
Flow Area: 0.073 nT (0.7854 fr) fi
Ethane Flow Rate8: 0.97 £/min =20.23 x 10"° kg/s
No. of Operating Fans: 2
Average Velocity: 5.9 m/s (19.4 ft/sec)
Total Emission: 20.46 x 10-6 kg/s
V*
£-A c**
A?
A-B I
B V
B C
B-C I
C V
L C
1
g4.9-4.1
£235-225
3.6
210
4.1
195
4.7
165
5.1
150
5.2
135
2
7.0
210
8.0
120
8.9
55
9.7
15
9.4
5
8.3
5
3
5.8
235
6.6
210
7.3
145
7.4
85
7.4
40
7.0
20
4
4.3
220
3.8
190
3.8
155
3.8
130
3.8
115
3.6
105
Port 1
ort 2
a Injection port was changed to previous the honeycomb section
* Velocity 1n m/s
** Concentration must be multiplied by 0.32 to convert to ppm ethane
Figure 57. Velocity and Concentration Distribution Data In Round Section
186
-------
c. Results and Conclusions
Different sampling methods were Investigated in Test Nos. 2
through 13. These included: sampling from the round and the square section
and the annubar at low airflow rate; sampling with proportional time;
directional sampling in the round section; directional sampling from the
Annubar element; and sampling from the square section using the 16 point
traverse by the use of either the Chebyshef method, the Centrold of equal
area method, or the Circular analog method. The calculated flow from
traversing both the square and circular sections as well as from using the
Annubar element closely agreed when the average velocity was low. A good
estimate of the emission rate was also obtained from both round and square
sections at lower flow rates. When sampling time from different locations
was proportioned to the local air velocity in Test Nos. 7 and 8, a good
agreement In emission rate was obtained compared to the calculated emission
rate using velocity and concentration measurement from each location.
The proportional time method has potential application in
non-reverse stratified gas flow when a true total emission rate Is desired.
The application of a continuous automatic method for proportional sampling
Is also possible by automating the above procedure.
The most recent results (at higher air flow rate), obtained
from traversing the square and circular sections, showed a good agreement
between the calculated air flow from each section only when the traversing
ports of the round section were rotated by an angle of about 45°. These
results demonstrate the inability of the perpendicular (2 diameters) traverse
method to account for certain velocity profiles with azimuthal dependence.
By rotating the traversing scheme, a more accurate flow rate may be obtained.
This Indicates that a knowledge of the flow and emission profiles 1s helpful
in selecting the optimum orientation of the sampling diameters.
187
-------
The total emission rate obtained from the square section
using the 91 point traverse was very close (-1%) to the real value (no
reverse flow existed 1n this section). This 1s 1n accord with the Intuitive
notion that the error should decrease with an increase 1n the number of
samples.
The emission rate calculated from the circular section 1n
Test No. 9 differed substantially from the known injected value. Because
the average velocity was Increased about three fold using the two fans 1n
series, the lateral diffusion of the tracer gas was limited by reduced
residence time and only a high concentration spike was obtained. This
spike-like profile slipped through the sampling mesh; hence, traversing
methods did not lead to the correct answer. It Is significant to note that
by comparing the velocity and concentration profiles one could see a greater
diffusion of tracer gas in areas where the velocity was lower. When the
tracer gas injection port and the injection tube 1n the circular section
were modified (Test No. 14), less stratification occurred and the emission
rate agreed closely with the known value. However, in this test, the
calculated air flow differed considerably (-25%) from the known value.
This Implies that the traversing procedure was adequate only for the
combined velocity and concentration profiles. Thus, without a previous
knowledge of the velocity and concentration profiles, the actual value of
the emission rate as well as total flow can only be assured by a minimum
of traverses along 4 diameters, 45° apart, with at least 8 sampling
points/diameter.
Results from the 'Annubar Element' tests showed a substantial
error 1n the flow rate measurement in one test and good agreement 1n
another test, but in all cases showed a much larger error than that observed
in the circular section traverse runs. The Annubar element was observed
to give the average concentration only when rotated, which shows the in-
ability of this element to measure the flow and average concentration when
interpolation occurs in a line where no average velocity or concentration
188
-------
exists. It 1s suggested that perpendicularly positioned 'Annubars1 could
be used for better results, which would Increase the probability of ob-
taining a reliable answer. Also, in high air flows, I.e., 3.3 m/s (11 ft/sec),
the Annubar could be used simultaneously for sample extraction and as a flow
sensor.
The average velocity and total emission rate, calculated from
the Chebyshef and Centroid of equal area methods for two sets of tests using
only 16 points for traversing in the square section, were very close to the
values obtained from the 91 point traverse method. When a highly reversed
flow was generated (Test No. 12), both methods failed to give the actual
value of the emission rate or the total air flow; instead, a much lower
value was given. In contrast, the Circular analog method for the square
section using 16 traverse points gave a much higher average velocity and
emission rate In all cases. The latter method 1s expected to give larger
errors in cases where no average emission rate exists in the 2 perpendicular
traversing planes. Therefore, 1t is recommended that one use either the
Chebyshef or the Centroid of equal area method for a 16 point traverse in
cases where highly reversed flow is not present. For reverse flow conditions,
a much higher number of sampling points is recommended (see Table 17).
189
-------
TABLE 17.
VARIATION OF. PERCENT ERROR IN EMISSION RATE AND TOTAL FLOW
AS CALCULATED FROM DIFFERENT TRAVERSING TECHNIQUES USED IN THE SQUARE SECTION
Test
No.
9-1
9-1
10
11
1212
No. of
Fans In
Operation
2
2
2
1
2
Damper
Position
fully open
fully open
fully open
fully open
partially
open
No. of
Traversing
Points
91
49
16
16
16
Centroid
Emission
1.9
<1
-6
-5
-32
of Equal Area
Av. Velocity
<1
-2
<1
-50
* ERROR*
Chebyshef
Emission Av. Velocity
-3 2
-8 <1
-42 -65
Circular Analog
Emission
21
34
28
Av. Velocity
17
43
18
Flow
Condition
no reverse
no reverse
no reverse
reverse
highly
reverse
within ±31 reproduceablHty
TEST CONDITIONS:
Baffle:
Inclined 22.5° horizontal
Ethane Flow Rate: 0.97 1/m1n
Flow area: 0.18 m2 (280 1n2)
Temperature: 20*C
2023 x 10"6 kg/s
-------
C. TASK III - FIELD DEMONSTRATION
Boston Edison's Mystic Station power plant was the site of field experi-
ments during the final task of this program. Three locations in the plant
duct work were first considered 1n an attempt to find gas stratification:
the after air preheater ducts, the after predpitator duct, and the Inlet
ducts to the scrubber of Unit No. 6. Table 1 summarizes the conditions
of these locations.
The air preheater outlet ducts and the south inlet duct to the scrubber
were surveyed. Because air 1n leakage was likely to occur 1n the air pre-
heater s, gas stratification was found in a preliminary survey at the air
preheater outlet ducts. The demonstration test was therefore run at this
location.
The sampling train, sampling procedures and results for both the pre-
liminary survey and final demonstration test are discussed below.
1. SAMPLING SYSTEM AND CALCULATION PROCEDURES
a. Sampling System Arrangement
The sampling system was prepared according to the schematic
arrangement shown 1n Figure 1-a through 1-d.
S-type pitot tubes heads (about 0.3 m (T) long), sampling
probes and dust 'filters were constructed of 316 stainless steel. Extensions
were provided so that probes and pitot tube lengths could be varied from
2.13 m (71) to 3.66 m (12'} according to the sample location requirement.
Because of subsequent delays in delivery of stainless steel tubes for
S-Pitot tubes, extensions were made of hard copper 0.057 m O.D. (3/16" O.D.).
The connections of the sample line to the probe tube as well as manometer
outlet lines to the S-p1tot tube were all Swagelock quick disconnects. This
arrangement simplified and expedited the sampling procedure from different
probe sets since only one sampling train were available.
191
-------
TABLE 1
No.
No.
of Ducts
of Ports/Duct
A1r Preheater
Outlet
Duct
2
3
Electrostatic
PredpUator
Outlet Duct
1
12
Scrubber
Inlet
Duct
2
5
Diameter of Port
Ports Location
Duct Dimensions
[width x helghth]
Ports Elevation From
Ground Level
Duct Location
Pressure
0.076 m (3")
nearly centrold
of equal area
locations
3.1m x 3.3m
(10'4" x IT)
3m (10')
Indoors
Negative
0.076 m (3")
symmetrically
located on both
sides of the duct
14.8m x 5.1m
(48'6" x 16'8")
13.1m (43')
Outdoors
Negative
0.12 m (4")
centrold of
equal area
locations
2.1m x 3.7m
(7' x 12')
24m (-80')
Outdoors
Positive
192
-------
ascarlte
drlerlte
OJ
Nomenclature
Room Air
Ta
Tf
Tw
Td
H
ts
BP
q
cso,
pro •
MO
ambient temp.
flue gas temp.
wet bulb temp.
dry bulb temp
velocity pressure head
flue static pressure head
barometric pressure
sample flow rate
concentration
dry gas molecular
weight
2 way valve
Flue
Dust Filter and
Sampling probe
>2 Span Gases
;0 Press Gas
2 Mixture
to vent
From
Reference
Probe
Sampling
il / Probe
/ set
Instrument
Unit
nil
Rue
gases
kS-tube and
thermocouple
Dust filter
Water Trap
water Well
Inclined
'Gauge
and
Pyromerer
, _ -1 I1ter/m1n
/ Manometer
Flow rate Indicator
Thermometer
Intertech. S02 analyser
Recorder
Beckman 400 C02 analyzer
Recorder
leather
Station
Tw.Td
8P
Orsat
Analysis
T
MO
Figure la. Sampling Arrangement
cco
-------
Flue
Dry Bulb Thermometer
Wet Bulb Thermometer
W1ck
Water Well
Balston Filter
Pump (-1.5 fr/m1n)
Figure Ib. Humidity Test Arrangement
194
-------
Figure 1c.
195
-------
Figure 1-d
196
-------
All thermocouples were made of 1ron/constant1n wires. To
keep the wires resistant to a maximum of 2.2 ohms (maximum calibration
setting for pyrometer), 6 m (-20'} of 14 gage Insulated wire was used to
fabricate each thermocouple. For each thermocouple, the hot junction was
silver soldered, then covered lightly with Teflon tape and finally shielded
with stainless steel tube (about 0.3 m (T) of the thermocouple wire was
shielded). An Alnor pyrometer was used to read the flue gas temperature.
o o
A refrigerator of 0.043 m capacity (-1.5 ft ) was used to
cool the sample prior to analysis. Two small holes were drilled 1n the
refrigerator door for the Inlet and outlet flue gas lines. The flue gas
sampling line in the refrigerator was made of approximately 1.8 m (61)
of stainless steel (1/4" tube) followed by 6 m (-20') of poly-flo tube
(1/4"). This line was shaped 1n a large coll form and was reinforced in
position with 2 steel Oexxlon angle Irons. The water vapor and other
6 3
condensables were trapped 1n a 250 x 10"° m (250 cc) Pyrex vial introduced
Into the refrigerator. The temperature inside the refrigerator was kept
near freezing (~3°C). The freezer section was used to remove most of the
remaining moisture content of the sampled flue gas stream 1n a plastic
trap. All connections were perfectly sealed.
Specifications for the major pieces of equipment are given
in Table 2.
The percent moisture was calculated from wet and dry bulb
temperatures measured by a separate sampling train, as shown in Figure 1-b.
The pressure in the line near the flow meter and the mea-
suring instruments for C02 and S02 gases were kept nearly atmospheric by
using a short vent line opened to atmosphere (the C02 analyzer was followed
by the S02 analyzer).
197
-------
TABLE 2
MAJOR EQUIPMENTS SPECIFICATION
Equipment
Specification and Manufacturer
Dwyer gage
SO 2 Gas Analyzer
C0« Gas Analyzer
02 Fyrite
Refrigerator
Pyrometer
Pump
Dust Filter
Filter Element
Model 400, Dwyers
Infra Red, Uras 2, Intertech Co.
Infra Red, Beekman Model No. 864, Beckman Instruments,
Inc.
Bachareh Industrial Instrument Co., Pittsburgh, Pa.
(1.5 ft3) Sears Cat. No. 3467370N.
Sears Roebuck and Co.
Alnor No. 1500, Alnor Instruments
Flue Gas Pump No. T-5 (prototype), Metal Bellows Corp.,
Sharon, Mass.
Balston Filter 95-S, Balston Inc., Lexington, Mass.
Filter tube Grade A, Balston Inc., Lexington, Mass.
198
-------
Sufficient sample line was used from the refrigerator to the
analyzers, and the temperature of the sample gas equilibrated to room tem-
perature. In all sampling cases, the same flow rate (-2 Uters/min) was
-6 3
maintained; about 16.6 x 10" m /s (1 I1ter/m1n) only went to the analyzers;
the remainder was purged. The purge line was provided to increase the
response time of the system by rapidly flushing the lines ahead of the pump.
It was also used for the Fyrite analysis of Og in the gas sample.
Chart recorders (Texas Instruments Inc.) were provided for
each analyzer for recording data on a continuous basis.
b. Calculation Procedures
The method used for calculating the emission rates from
traverses 1s outlined in the flow sheet shown in Figure 2. The data sheet
used in data collection 1s shown In Figure 3.
2. PRELIMINARY SURVEY TESTS
The following 1s a discussion of preliminary survey tests conducted
on the exhaust ducts of Unit No. 6 at the Mystic Station. During the pre-
liminary survey, the outlet ducts of the air preheaters (See Figure 4) and
the south Inlet duct to the scrubber were examined for gas stratification.
Oxygen and carbon dioxide were sampled at different locations In the ducts
and analyzed by Fyrltes.
Oxygen and carbon dioxide stratification was found in the air pre-
heater outlet duct (north side), while no stratification was found at the
scrubber's Inlet duct. On the basis of this evidence, it was decided to
survey both outlet air preheater ducts for S02, C02 (NDIR) and velocity.
199
-------
. eo /::•!
-
K • -5- • propertlcMllty eanitinl
K Convert 1on f«ct»r at
M
of trtwntng potfti
A. tfftctlw flu* tre«. •*
Flijur* Z. Cnlitlon Kiti C«UuUtle«
200
-------
Figure 3
TEST UNIT
STATION
TEST NUMBER
TEST LOAD
DATE
Position
PORT
T °C
Vh
Vel
FR
02
•
C02
S02
Time
PORT
T °C
Vh
Vel
FR
02
C02
S02
Time
WET BULB TEMP
AMBIENT AIR TEMP
AVERAGE F.G. TEMP
*TATIC PRESS-
AVFRARF VELOCITY
RABOMFTRTC PRESS.
DRY BULB TEMP
IN. WATER
F.P.S.
IN. H9
201
-------
r\»
o
Reference
0.076 m (3")
Ports
Port 6-a
Port 6
Port 5 Port 4
3.1 m
SOUTH DUCT
ill"
ue gas
— (b- — <^ —
O.C76 m
(3") Ports
~~Y~
Reference
Port 1-a
Port la
Port 3
T
Z-
PortJL
3.1 m
NORTH OUCT
Figure 4. After Air Preheaters Ducts - Plan View
-------
a. Sampling Procedure
The sampling train described 1n the previous section was
used In these tests.
The after air preheater ducts were surveyed using one sampling
probe set which kept the reference probe set at a constant position. Because
the location was "Indoors" a heated sampling tube was not required to prevent
freezing.
Since the 0.076 m (3") dia. ports were less than 2.7 m (9 feet)
above floor level, some difficulty was encountered 1n introducing a longer
length probe. A 0.9 m (3 foot) extension was used with a 2.7 m (9 foot)
probe to sample at locations 2.4 m (8 feet) and 3.0 m (10 feet) 1nsl1de the
ducts. This extra length was connected after the 2.7 (9 foot) probe was Intro-
duced about 1.2 m (4 feet) Inside the duct.
To assure a vertically straight probe assembly 1n the duct,
the sampling tube, the S-tube and the thermocouple were secured firmly to
a Dexxlon No. 225-S angle Iron. The whole assembly was Introduced Into the
port and fixed 1n position by a specially fabricated stand, as shown 1n
Figure 1-c. A tight seal was obtained by closing the port with asbestos
tape, wrapped in aluminum foil.
When sampling at the Inlet duct to the scrubber, a 2.7 m
(9 foot) stainless steel tube (1/4" O.D.) was used to collect the gas sample.
A Dexxlon steel angle iron was used to give horizontal support to the sampling
tube. A 2.7 m (9 foot) extension was provided for both the sampling tube
and the steel support and an asbestos rag sealed the port. 02 and C02
analyses of the gas sample were performed by using the Pyrites.
b. Results
Test results from surveying the south Inlet duct to the
scrubber did not show any gas stratification (see Figure 5).
203
-------
North
3.6 m x 2.1 m duct
Sampling Ports
Flue Gas Flow
2.1 m
Test Run Using Pyrites For C02 and 02 Content
Port 4
(m) x 10'1
6
12
18
24
30
% co2
11.0. 11.5
10.5, n.o
11.0. 11.5
n.o. 11.5
11.0
% Q2
6.0
6.5
6.0
6.0. 6.0
6.0, 6.0
Port 3
(m) x 10"1
6
12
18
% co2
11, 11.5
11.0, 11.5
11.0. 11.5
•
% o2
6.5, 6.0
6.5, 6.0
6.5
* On April 16, 1974 a constant load on the scrubber was kept as well as constant
gross MW output during the testing period.
Figure 5. Results From Test Run* At The Scrubber South Inlet Duct
204
-------
Port 4 was surveyed at five locations, from 0.6 m to 3 m 1n
the duct, and Port 3 was surveyed for three locations, from 0.6 m to 1.8 m
1n the duct. In all cases, the percent C02 was equal -11% ± 0.556 and the
percent 02 was equal -655 + 0.5%. Since gas concentration stratification was
not observed, further tests at this location were not attempted.
Test results from the preliminary survey showed gas stratifi-
cation In the ductwork after the air preheaters In all cases. Results from
the Fyrlte tests run on the north duct are presented in Figures 6-a and 6-b.
At Port No. 3 the oxygen content varied from 4.5% at 0.6 m (2 feet) to about
7.5% at 2.9 m (9.5 feet), and the C02 content decreased from 10.5% at 0.6 m
(2 feet) to 8.5% at 2.9 m (9.5 feet).
Results from Test Nos. 3 and 4, using the NDIR analyzers for
flue gas analysis, are presented 1n Figures 7 a through 8 d. The after air
preheater duct (north side) was surveyed at 130 mw gross output. For all
locations In the duct, the S0« concentration was sampled as well as the SO
concentration from a reference point. The velocity and temperature at the
reference and sampling points were also measured. The reference S09 concen-
tration varied ±11 ppm around a mean value of 749 ppm, while the values at
the different sampling locations varied ±47 ppm around a mean value of SO
concentration equal to 680 ppm. The velocity 1n the duct ranged from as low
as 8 m/s (28 ft/sec) to as high as 15 m/s (48 ft/sec). Lower values were
generally obtained near the top of the duct at a probe Insertion of 2.4 m
(8 feet) and 3.0 m (10 feet). The S02 concentration times velocity term
(which 1s proportional to the emission rate) varied ±26% about the mean value,
with a lower value of 35% below the mean and the highest 43% above the mean.
Temperature stratification was also observed to vary from 120°C to 155°c near
the center of the duct.
205
-------
Figure 6a
TEST UNIT No. 6, After Air Preheater
STATION
TEST NUMBER
TEST LOAD
1
DATE 3/18 and 3/19. 1974
"150 MW
Position
0.6m (2')
1.2m (4')
1.8m (6()
2.0m (63/4'
2.4m (73/4'
Q,6m (2')
1.2m (4')
lifim («'.)
2.4m (81)
2.9m 19.5'1
PORT
T °C
130
145
140
) 135
145
Vh
1.0
1.0
0.9
0.6
0.7
Vel
FR
1
02
7.5
7.0
6.5
7.0
6.5
6.0
6.5
C02
9.0
9.5
10.0
9.5
9.5
9.5
9.5
10.0
S02
Time
i
4
PORT
T °C
145
145
145
140
135
Vh
1.7
1.7
1.5
1.4
1.0
Vel
FR
02
7.0
6.0
5.0
R.O
R,R
7.0
6.5
6.5
7.0
C02
10.0
10.5
11.0
10.5
10. n
10.0
S02
Time
WET BULB TEMP
AMBIENT AIR TEMP
AVERAGE F.G. TEMP
AVERAGE STATIC PRESS.
AVERAGE VELOCITY'
BAROMETRIC PRESS.
DRY BULB TEMP
IN. WATER
F.P.S,
IN. HQ
206
-------
TEST UNIT
• Figure 6b
After Air preheater
STATION Mystic
TEST NUMBER 2
DATE 3/19/74 P ~3:30
TEST LOAD Peak Load ~154 MW
Position
0.6m (21)
1.2m (44)
1.8m (6')
2.4m (8')
2.9m(9.5'
REPEAT (A
1.2m (41)
PORT
T °C
130
150
150
145
140
F 4:45
150
Vh
2.0
2.0
1.6
0.9
0.6
\
?.o
Vel
FR
02
4.5
•4.5
5.0
5.5
5.5
5.5
6.5
7.0
7.0
7.5
5.5
5.5
C02
10.5
11.0
10.0
8.5
8.5
8.5
10.5
_
SOp
_ . __
Time
PORT
T»C
Vh
Vel
FR
02
C02
S02
-
Time
WET BULB TEMP
AMBIENT AIR TFMP
AVERAGE F.G- TEMH
AYFRAGE STATIC PRFSS.
AVFRAfiE Vfil,OCTTY
BAROHFTRIC PRESa.
DRY BULB TEMP
°F
°F
IN. WATER
F.P.S.
IN. HO
207
-------
TEST m. 3
Ffflurt 7*. After Air Prebeater Duct (North Side) Velocity. SO, Concentration and Tetpentur* Traverse
•t - 130 MH Gross Output*
^
S62 co2
(PP«) (t)
PORT NO. 1
0.6n
1.2H
l.Bfe
2.4»
3.0m
2'
4'
6'
8'
10
70S
720
630
630
') 645
M PORT NO. 2
8 0.6m
1.2m
2!4«
3.0*
2'
4'
i:
10
735
760
730
675
} 645
PORT NO. 3
0.6»
2.*4*
3.0H
21;
4'
6'
8'
10'
70S
705
690
615
) 615
a std. dev. 47
7*r°rT.«n "°
Cp"
^
-------
Figure 7b. After A1r Preheater North Duct Velocity,
S02 Concentration and Temperature Profile*
At - 130 MW Gross Output
/
3m
9 m/s
+ 602 ppm
^ 135°C
8 m/s
+ 602 ppm
140°C
12 m/s
+ 675 ppm
145°C
14 m/s
+ 718 ppm
140°C
14 m/s
+ 718 ppm
yls 125°C
0.6m
10 m/s
+ 644 ppm
140°C
12 m/s
4- 674 ppm
145°C
12 m/s
+ 719 ppm
150°C
15 m/s
+ 774 ppm
155°C
15 m/s
+ 739 ppm
150°C
10 m/s
+ 653 ppm
135°C
9 m/s
+ 633 ppm
140°C
10 m/s
+ 629 ppm
140°C
11 m/s
+ 719 ppm
145CC
11 m/s
+ 709 ppm
120°C
T — 1
Port 3
Port 2
Port V
3.3m
3.1m
* from preliminary survey
209
-------
Flfurt 8«. After Air Preheater Duct (South Side) Velocity. SO. and CO, Concentration and Tovperature Traverse
At . 150 MM Gross Output* ^ z
TEST NO. 4
PORT NO
0.6m (2
i-si!
2.4 («'
3.0. (1
PORTNG
O.ta (2
1.2> 4
1.8i 6
2.4« «
3.0» (1
PORT NO
0.6. (2
1.2* 4
1.8* 6
2.4* 8
3.0* (1
oStd.
?.C-or
C.
S2
(PP»)
. 4
'} 900
1 850
•) 825
) 810
0') 810
. 5
') 910
915
855
') 855
O1)
. 6
') 840
910
885
') 855
0') 860
Oev. 36
T 863
W
12.
11.
11.
10.
10.
12.
12.
11.7
11.3
11.5
12.2
12.1
12.0
12.0
0.6
11.8
(
soz
(PP«)
0.6 •
855
870
840
870
870
855
865
870
870
870
860
860
870
860
9
863
.-
W
(2') Inside
11.3
11.3
11.4
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
11.2
11.3
0.0
11.3
V$
(pp.)
908
843
848
803
803
918
913
848
848
833
913
888
848
863
39
863
!«v
(*)
12.3
11.8
11.4
10.8
10.8
12.6
12.6
11.7
11.3
11.5
12.2
12.1
12.1
12.0
0.6
11.8
Ratio
Rm
-------
Figure 8b. After Air Preheater South Duct Velocity,
S0? and CO? Concentration and Temperature Profile*
At -150 MW Gross Output
/
3m
12 m/s
+ 863
140°C
10 m/s
+ 848
140°C
11 m/s
+ 888
140°C
14 m/s
+ 913
130°C
13 m/s
+833
/K 120°C
0.6m
+
15 m/s
+ 848
130°C
17 m/s
+ 848
130°C
17 m/s
+ 913
140°C
17 m/s
+ 918
130°C
11 m/s
+ 803
130°C
13 m/s
+ 803
130°C
14 m/s
+ 848
130°C
15 m/Si
+ 843
125°C
15 m/s
+ 908
125°C
Port 6
Port 51-1 Port 4
3.1m
3.3m
* from preliminary survey
211
-------
1000
ro
CO
900
80
60
§ 40
c
o
CsJ
O
20
800
0.6 m
(2')
Reference Probe Port No. 6
(Position, 0.6 m Inside)
Sampling Probe Port No. 4
(Variable Position)
* I.....L
11:00 a.m. 12:00 1:00 2:00
Time, Hours
3:00
Figure 8c. After A1r Preheater Duct (South Side), Port No. 4 Traverse With Fixed Reference Probe
Fcr ~ 150 MW Gross Output
-------
Velocity Profile
S02 Concentration Profile
ro
CO
Velocity Scale :
1 cm = 3m/sec
S02 Concentration Scale:
1 cm = 200 ppm
V"-- Ouct
;0*V Wall \
Figure 8d. After Air Preheater Duct (South) Velocity and S02 Concentration Profiles
At - 150 MW Gross Output
-------
Results from Test No. 4 were obtained from testing at the
south duct after the air preheater at -150 mw gross output. S02 and C02
concentrations were reported for both the sampling probe and the reference
probe gas samples. The reference S02 concentration varied ±9 ppm about a
mean value of 863 ppm, while the reference CO* concentration held steady at
11.3% CO*. The sample concentration values at the different locations in
the duct varied for both S02 and COg. A standard deviation of ±36 ppm about
a mean of 863 was obtained for S02 concentrations, and a variation of ±0.6%
about a mean of 11.8% was obtained for C02 concentrations. The ratio of
SOo/CO, concentrations had a mean value of 73.1 x 10 with a standard
-4
deviation of ±1.4 x 10 .
The reference sample held a reasonably steady value for S02
and C02 concentrations in the flue gas during the traversing 1n Port No. 4.
(See Figure 4-c). The velocity In the duct varied from as low as 110 m/s
(34 ft/sec) to a high of 17 m/s (57 ft/sec) near the center of the duct.
The S02 concentration times velocity term varied ±19% about the mean value,
with a lower value of 28% below the mean and the highest 43% above the mean.
Temperature 1n the duct varied from 120°C (248°F) to 140°C (284°F). The
velocity and S02 concentration profiles are both displayed in Figure 8-d.
c. Discussion of Results and Conclusions
The north and south ducts following the air preheater outlet
were surveyed for gas stratification, using three ports equally separated on
the width of the duct and at locations Inside the duct ranging from 2 feet
to 10 feet. The gas samples were first analyzed for oxygen and carbon
dioxide using the Pyrites. Stratification of both 02 and C02 was observed
1n the north duct for about 150 mw gross load; this was confirmed at 130 mw
gross output using the analyzer to detect S02 concentrations. A lower degree
of S02 and C02 stratification was observed at the south duct at 150 mw,
Indicating that the conditions 1n the two ducts are not similar. The c.v. for
S02 concentrations In the ducts is less than 10%, on an average, but the c.v,
1n the velocity time S02 concentration term 1s about 25% at -130 mw gross output-
214
-------
These results show a greater emission rate stratification than 502 concentra-
tion stratification, which is due to the low velocity value observed at the
same location where a low S02 concentration value was found. Confidence in
these results depend upon the knowledge that during sampling, the reference
samples maintained a nearly steady value for S02 and C02 concentrations.
Also, the concentration ratios of S02 and C02, for the points associated
with the whole sampling plane, did not vary more than -2%, which is consistent
with the expected precision qf the analyzers, i.e. ±1%. This result is in
accord with the hypothesis which predicts that S02 and C02 are stratified in
the same way.
Based on the stratification data obtained initially, the south
duct test results were further analyzed. Interpolated and extrapolated values
for velocity and S02 concentrations, for 9 to 15 probes equal area strategy,
were taken from actual data profiles and the CV term calculated (See Tat?le 3-a
to 3-c and Figures 9-a thru 9-c).
The average values for velocity, S02 concentration and the CV
term, for the 9 and 15 equal area strategy and for the actual data, are
tabulated in Table 4 with the standard deviation and the coefficient of
variation. Practically speaking, the CV term (concentration x velocity) can
be considered to be the same for all cases and therefore the 9 probe equal
area strategy Is comparable to the 16 probe equal area strategy.
If 9 probes were used, the average concentration times velocity
term is equal to 12100 (ppm x m/sec). One could attempt to compare the
measured emission value to the actual emission value based on 9989 x 10"6 m3/s*
of residual fuel oil used at 1.951 S*. The specific gravity of the oil may be
taken as 0.96.
* From Boston Edison operating data, expected fuel flow rate at 150 mw equal
9500 gallons/hr.
215
-------
TABLE 3-a
VALUES FOR SOg CONCENTRATION AND VELOCITY FOR
THE ACTUAL 15 POINTS TRAVERSE*
AT 150 MW
PORT
PORT
PORT
v, r
o
CV**
*
At
4 4-1
4-2
4-3
4-4
4-5
5 5-1
5-2
5-3
5-4
5-5
6 6-1
6-2
6-3
6-4
6-5
or CY
(*)
(2(), (41), (6'), (8')
C
ppm SO 2
908
843
848
803
803
918
913
848
848
833
913
888
848
863
863
39
4.5
and (101
V
m/s
15
15
14
13
11
17
17
17
15
13
14
11
10
12
14
2.4
17.4
) Inside the duct
CV
ppm x m/s
13,620
12,650
11,870
10,440
8,830
15,610
15,520
14,420
12,720
10,830
12,780
9,770
8,480
10,360
11,990
2,300
19,2
Coefficient of variation
216
-------
TABLE 3-b
INTERPOLATED VALUES FOR S02 CONCENTRATION AND VELOCITY
FOR 15 PROBES EQUAL AREA
PORT 4
PORT 5
PORT 6
V", t~, or
a
CV* (%)
4-1
4-2
4-3
4-4
4-5
5-1
5-2
5-3
5-4
5-5
6-1
6-2
6-3
6-4
6-5
CT
C .
ppm SO 2
908
865 x
850
810
803
918
915
865
848
848
833
880
900
850
860
863
35
4.0
V
m/s
14
15
14
13
11
15
17
17
16
13
11
13
12
10
12
14
2.1
15,2
CV
ppm x m/s
12,450
13,440
12,170
10,860
9,050
13,710
15,340
14,760
13,440
10,850
9,140
11,800
10,970
8,810
10,480
11,800
2,000
17.3
* Coefficient of variation
217
-------
TABLE 3-c
EXTRAPOLATED AND INTERPOLATED VALUES
FOR S02 CONCENTRATION AND VELOCITY
FOR 9 PROBES EQUAL AREA
PORT 4
PORT 5
PORT 6
7, T, CT
a
CV* (%)
4-1
4-2
4-3
5-1
5-2
5-3
6-1
6-2
6-3
C
ppm S02
908
850
803
918
865
848
833
900
855
865
38
4.0
V
m/s
15
14
12
16
17
14
12
12
11
14
2.1
15.2
CV
ppm x m/s
14,110
12,170
9,540
15,390
14,760
11.890
10,410
10,970
9,900
12,100
2,160
17.8
* Coefficient of variation
218
-------
Duct Wall
o Port 4
5
APort 6
800 900 1000
SOConcentration, ppm
1100
Duct Wall
Figure 9a. South Duct S02 Concentration at 150 MW for 1.955 Sulfur Oil
219
-------
Duct Wall
6 9 12 15 18 21
Velocity, m/sec
Duct Hall
Figure 9b. South Duct Velocity Profile at ISO MU
220
-------
O)
A Position 1
O Position 2
Position 3
0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7
Location of Probes in Duct Width, m
3.0
Figure 9c, South Duct Velocity Profile at 150 MW
221
-------
TABLE 4
SOUTH DUCT (PORT 4, 5, 6) AT 150 MM*
COMPARING DIFFERENT SAMPLING METHODS
Method and
No. of Probes
r
ppm S02
C.V*
7
m/s
C.V** CT
ppm x m/s a
C.V.
(«)
15 probes ***
(real data)
863
39 4.5
14
2.4 17.4 11,990
2,300 19.2
15 probes
(equal area)
863
35 4.0
14
2.1 15.2 11,800
2,000 17.3
ro
r\>
ro
9 probes
(equal area) 865 38 4.0
* percent sulfur In fuel = 1.9%
** Coefficient of variation
^Average temperature ~130°C
% moisture =9.5% (measured)
14
2.1 15.2 12,100
2,160 17.8
-------
The expected emission of SOg (assuming 3% of the sulfur 1n
the oil converts to S03) based on the sulfur content 1n the oil 1s equal to:
1 x IP3 x 9989 x 10"6 x 0.96 x 1.9 x 64 x 0.97 _ 355 x 1Q-3 kg/s
32 x 100
For the actual data, if one assumes that the north duct will behave similarly,
we obtain:
Area x C x V x
. 2x10.5x12100
64
34 x 10"
34 x 10
"J
64 x (1-0.095), 430 x 1Q-3
k /s
The percent difference between the actual and the expected
emission 1s equal to +2U. The difference may be due to difficulties en-
countered in measuring the pulsating flow in the duct and/or failure to con-
sider the boundary layer effort on the flue gases flow rate. It 1s also
possible that this difference Is based on the reported value of the fuel
flow rate which has not been verified.
By attempting to extrapolate more points near the boundary
(see Figure 10), the mean value for the velocity was found to be equal to
13 m/sec, about 8% lower than the measured value. It is therefore expected
that the calculated percent error In emission rate will Include at least
error due to boundary layer effect.
Based on these results and analyses, a nine probe equal area
strategy was planned for the final demonstration test.
3. DEMONSTRATION TEST
i
Based on the stratification data collected in the preliminary survey,
plans were made for the demonstration test at the locations after the air pre-
heaters for a 9 equal area strategy for each duct, I.e., a total of 18 sampling
223
-------
Position 3
Position 2
Position 1
' 6 10
6 to
6 10
Port 6
10
12
11
12 13
11
10
10
11
12
13 14
12
13
13
13
12
13 14
10
7
13 13
16 16
16 16
•
Port 5
11
13
14
14 14
15
15
16
17
17
17 17
17
17
17
17
16
16 16
15
10
13 13
16.2 15
16 16
J L.
Port 4
10
11
12
10 5
12
13
13
14
14
12 6
14
15
15
15
15
14 7
14
8
_i
Mean Value of All Points « 13 m/sec
* Dimension 1n m/sec
Figure 10. South Duct Velocity Profile at 150 MW
224
-------
points for both ducts. The set up Included 6 sets of probes, one for each
sampling port and 2 sets of probes for the two reference ports (one ref-
erence probe for each duct). At each location S02, C02 (NDIR) and (L (Fyrite)
concentrations plus temperature and velocity were measured. Methods of data
collection and test results are discussed below.
a. Sampling Procedure
The sampling train arrangement was as described in Section IV,c.l.a.,
shown in Figures 1-a through 1-d.
A schematic diagram of the set up of probes and stands is
shown in Figure 11. The reference probe assembly was maintained at 10.6 m
(21) inside the duct, and each duct has one reference assembly. The probe
tip assembly at each port was placed Into the duct in three different positions,
I.e., about 0.6 m 1.7 m and 2.7 m into the duct (duct dimensions are 3.1 m
wide and 3.3 m high).
A gauge was used to monitor the high sulfur fuel oil tank
during the test period. The test ran from about 11 o'clock in the morning
until 4 o'clock 1n the afternoon. During the test, a constant load (-144 mw
gross output) was maintained on the boiler unit No. 6 as well as the scrubber
unit. The latter was checked by running Pyrites tests on both C02 and 02
analyses at the inlet south duct to the scrubber.
Sampling proceeded as follows:
1. Calibration of analyzers using both zero and span gases
2. Sampling from Reference la at the north duct
3. Sampling from Port No. 1 at the first position
4. Sampling from Reference la
5. Sampling from Port No. 2 at the first position
225
-------
SIDE VIEW
ELEVATION
South Duct
Port Port Port
6 5 4
Probe
Assembly
Cerrent 0.6
Elocks (21)
Dexxion
Stand
T
2.7 m
(91)
North Duct
Port Port Port
3 1
Figure 11. After Air Preheater Sanplino Arrannenent
226
-------
This sequence was followed for all points at the first position. The probe
assembly was adjusted for the second position as soon as sampling from the
first position was completed. The same sampling procedure was then repeated
for the second position and the third position. At the termination of a run
the analyzers were calibrated.
About 36 samples were analyzed, involving 18 sampling points,
centroid of equal areas, and 18 reference points. At each location S02> C02
(NDIR) and 0« (Fyrite) concentrations plus temperature and velocity were
measured. S02 and C02 data were recorded continuously.
When 02 analysis was performed (by Fyrite), the sample was
taken from the purge line while the flue gas stream leading to the analyzers
was closed. This was done because hand pumping to the Fyrite causes pressure
fluctuations which effect the analyzers output signal, specifically the
Intertech S02 analyzer.
The purge stream as well as the analyzer flue gas stream were
adjusted (same setting for each sampling point) by checking the flow measuring
elements (rotameters).
The sampled gas temperature (equal to room temperature) and
humidity (controlled by refrigerator temperature) were also kept constant
for all sampling points. Therefore, equal volumes per unit time were sampled
from each location.
Humidity tests were run separately, twice for each duct. The
sampling arrangement shown previously 1n Figure 1-b was used.
Sampling at the Inlet south duct to the scrubber was performed
at a constant location, about 1.2 m Inside the duct at Port No. 4, using the
same arrangement described 1n the preliminary survey section (see Section IV .C.l.b.).
227
-------
b. Results
Test results from surveying the south and north ducts after
the air preheaters showed CCL, CL, and SCL gas stratification while the
test load was kept constant at 144 mw gross output from Unit No. 6. The
reference sample from each duct held a reasonably steady value for S02
and C02 concentrations 1n the flue gas during the whole traversing period.
The C02 and 0« concentrations, at the Inlet south duct to the scrubber,
also held a steady value during the test period (See Figure 12). Test
results are presented 1n Tables 4-a thru 5-b and Figures 12 and 13. All
calculations made are Included In Appendix I.
The reference $02 concentration at the north duct varied
±12 ppm from a mean value of 943 ppm; the CO,, concentration varied ±0.2%
from a mean value of 11.9%, while the 02 concentration did not show any
variation. The normalized values at the different sampling locations varied
for S02 concentration by ±52 ppm from a mean value of 917 ppm with a low
value of 840 ppm and a high value of 990 ppm; for C02 concentration by
±0.9% from a mean value of 11.6% with a low value of 10.4% and a high
value of 12.8%; and for 02 concentration by ±0.4% from a mean value near
6.5% with a low value of 6% and a high value of 7%. The velocity in the
duct ranged from as low as 10 m/s to as high as 18 m/s with an average value
of 15 m/s and a standard deviation of ±2.5 m/s. Temperature stratification
was also observed varying from 130°C to 150°C.
The coefficient of variation of tfee concentration times
velocity term (which is proportional to the emiss\ftn rate) was ±22%, ±24%,
±16% for S02, C02 and 02> respectively; the lowest and highest deviation
from the mean was -37% and + 24% for S02, -39{( and +35% for C02, -24% and
and + 27% for 02. The average S02/C02 concentration ratio was equal to
79.1 x 10"4 with a coefficient of variation of 2.1%.
The reference S02 concentration at the south duct varied
±8 ppm from a mean value of 1042 ppm. The variation of C02 concentration
was negligible, while the 02 concentration varied ±0.2% from a mean value
of 4%.
228
-------
North
t
t
f
Wet
Scrubber
;
t
t
i
.3.6 m x 2.1 m Duct
Sampling Ports
2.1 m
Position
Port No. 4
1.2 m
Indlde
% co2
10.5
10.0
10.0
10.0
10.0
10.0
% o2
6.0
6.0
6.0
6.0
5.5
6.0
Time
11:05
11:35
12:05
.12:45
1:30
2:15
* On April 24, 1974 from 11:00 a.m. to 2:30 p.m.
Figure 12. C02 And 02 Concentrations At The Inlet South Duct To the
Scrubber Using the Pyrites*
229
-------
TA8U 4-a
NORTH OUa DATA KEDUCTI0* AT -144 MM GROSS OUTPUT
IN*
<*>
O
««*, £>
PORT NO. 1
1-1 408
1-2 413
1-3 403
PORT NO. 2
2-1 423
2-2 408
2-3 408
PORT NO. 3
3-1 418
3-2 413
3-3 408
o 6.0
Vorf 411
«**{*) 1.5
CONDITIONS: B.P.
V
TSAV
Gd
V
•Average value of
** Coefficient of
V
H x 102 C
(H/m2)
1.37
0.87
0.62
1.99
1.37
1.25
1.87
1.37
1.25
100.9 N/*2
97.2 N/m2
411*IC
0.99
0.835
.ft.P.xHxT
,-0.19
15
12
10
18
15
14
18
IS
14
2.5
15
17
(29.8" Ho) x 103
(28.7" Hg)x 10?
s
I2
6.0
6.0
7.0
6.5
6.0
7.0
6.0
6.0
6.5
v
X
11.8
11.6
10.6
12.5
11.6
10.9
12.5
12.0
10.9
ppn
945
915
855
9(0
915
870
975
930
870
5*
6.C
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
0.0
6.0
0.0
c,
X
11.6
12.1
12.1
11. «
12.0
12.0
11.6
12.0
12.1
0.2
11.9
1.8
ppra
-
930
-960
960
930
945
945
930
945
945
12.0
943
1.2
V
x
6.C
6.0
7.0
6.5
6.0
7.0
6.0
6.0
6.5
0.4
6.3
6.8
^R
X
12.1
V..4
10.4
12.8
11.5
10.6
12.8
11.9
10.7
0.9
11.6
7.5
pptn
960
900
840
975
915
870
990
930
870
52
917
5.6
CN V (pp» x »/s) CTSO?
°2
(x 104)
90
72 '
70
117
90
98
108
90
91
15
92
16.3
U?-)
181
137
104
230
172
151
230
178
ISO
41
170
24.1
-.
14,400
10,800
8,400
17,550
13.725
12,180
17.820
13.950
12.180
3,000
13.445
22.4
x 10f*
79.3
78.9
80.8
76.1
79.5
80.5
77.3
78.1
81.3
1.7
79.1
2.1
S-tube calibration factor
variation
-------
TABLE 4-b
SOUTH DOCT DATA REDUCTION AT -144 MU GROSS OUTPUT
Ts
Position fK)
PORT NO. 4
4-1 403
4-2 403
4-3 393
PORT NO. 5
5-1 403
5-2 403
5-3 393
PORT NO. 6
6-1 413
6-2 408
6-3 403
o 6.3
7 or C~ 403
CV** (t) 4.9
commons: B.P.
V
a"
V
* Average value
V«
H X 10* C1
1.62
1.37
0.99
1.62
1.49
1.18
1.37
0.99
0.62 •
100.9 N/B2
97.6 N/*2
403*K
1.01
0.835
jp.P.xHiT^
•0.19
On/-;)
16
15
12
16
15
14
15
13
10
2.0
14
14.3
(29.80 Ho)
(28.84 Hg)
°2
t
7.0
5.5
6.0
6.0
5.0
5.5
4.5
4.0
4.5
Si
C02
X
12.8
12.5
11.8
13.9
13.4
12.3
13.5
1-.4
13.0
ppra
960
990
930
1050
975
975
1050
1020
1005
CR
°2
X
4.0
4.0
4.0
4.0
4.0
4.0
4.5
4.0
4.5
0.2
4.1
5.3
X
13.5
13.4
13.5
13.5
13.5
13.5
13.5
13.5
13.5
-0
13.5
.0
!>02
ppm
1035
1050
1035
1035
1035
1035
1050
1050
1050
8
-1042
0.7
V
X
7.0
5.5
6.0
6.0
5.0
5.5
4.0
4.0
4.0
1.1
5.2
21.1
CR
io2
X
12.8
12.6
11.8
13.9
13.4
12.3
13.5
13.4
13.0
0.7
13.0
5.3
CN V (pom x m/s) CJiS02
ppm
965
980
935
1055
980
980
1040
1010
995
37
993
3.7
U2
(x 10*)
112
82
72
96
75
77
60
52
40
22
74
30
(x 10«)
205
189
142
222
201
172
202
174
130
30
182
16
*>2
15.440
14.700
11,220
16,880
14,700
13,720
15,600
13,130
9.950
2.200
13.900
16
xlO!«
75.4
77.8
79.2
75.9
73.1
79.7
77.0
75.4
76.5
2.0
76.7
2.6
for S-tube Calibration factor
*• Coefficient of variation
-------
TABLE 4-c
AVERAGE CONCENTRATIONS FROM BOTH DUCTS
a
*
mean
C.V.**
o
mean
C.V.**
°2
(%)
1.0
5.8
17%
°2
(x 104)
20
83
25%
co2
/ Qj \
\ " 1
1.0
12.3
8%
co2
(x 104)
36
176
20%
so2 so2/co2
-4
ppm x 10
59 2.2
955 77.9
6% 2.8%
so2
ppm
2,600
13,600
19%
1 • 18
W v/
Snean was a^so ^OLtnc^ to be e(?ual to
(ALL DRY BASIS)
** Coefficient of variation
232
-------
TABLE 5
ro
CO
CO
TD('C)
54.4
57.2
57.8
58.9
60.0
61.1
62.2
63.3
64.4
65.6
66.7
NORTH DUCT
TW(°C)
47.8
47.8
48.3
48.9
48.9
48.9
48.9
49.4
49.4
49.4
50.0
5-a TEST NO. 1
T (°C
ID1 I
48.9
50.6
52.2
53.3
54.4
56.1
57.2
59.4
SOUTH DUCT
:) TW(*C)
44.4
45.6
46.1
46.7
46.7
47.2
47.2
47.8
TD(°C)
53.9
57.2
59.4
61.1
63.3
64.4
65.5
66.6
67.8
68.9
NORTH DUCT
TW(°C)
47.2
47.8
48.3
48.9
48.9
48.9
48.9
48.9
49.4
49.4
5-b TEST NO.
TD<°
46.1
48.9
51.1
52.2
54.4
56.1
57.2
58.9
60.0
62.2
2
SOUTH DUCT
C) TW(°C)
44.4
45.0
46.1
46.1
46.7
47.2
47.8
47.8
47.8
47.8
-------
Figure 13
Recorder Output
13-a. S02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
13-b. CORecorded Output Signal at 0.5 1n/m1n - Response Curves
234
-------
S02 SPAN
GAS IN
Figure 13*. S02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
235
-------
t/1
z
o
o
l/l UJ
—J l/>
2 0
•-, LJ
3
O c
ss
1.0 .9
- PROBE (P5SITIC
PQ
RT 1
L- REF I
N 2)
- REF 2,
.7
\i ;
.- PROBE! (POSITION |]
PORT fr
f
REF 2
i
I
•2
I]
1
4
'
t
iqure 13*. S02 Recorded Output Signal at 0.5 In/min - Response Curves
236
-------
2:
o
LJ
*
o
0 £
o >
Ul C
o. -
8s
UJ t-
oe 4
ru
O
1.0
.e
1 -
I
.7
REF 1
PORT
--f -4-
.6 .5
08E (POSI T I
RT I
.3
T
-
ON 2)
Figure 13a. *502 Record d Output Signal at 0.5 1n/n.1n - Response Curves
237
-------
PROBE (POSITION $
PORT I
.5
.4
i
• 3
.2
y«EFi
r "'
~l_
PROBE CPOS
PORT 6
fTlON ,2)
Figure 13a. S02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
238
-------
o
o
u
0 i 1.0
O N
UJ C
o -
cc *">
u
U» K
a <
IM
O
.9
REF 2
r -
"PORT 3
-i
.5 .4
•' i
I
REP
.3
1
.2
PORT 2
i
i--- - -
-4
H
Figure 13«. S02 Recorded Output Signal at 0.5 In/mln - Response Curves
-------
£
X.
C
s
5
Z Ul
o >
C02 GAS
1.0
.8
.7
.4
.3
.2 .1 |
o?
o u>
UJ u>
o or
S '
o
Ul
ct
-------
-Z
g
i-
u
o ±
o^
Ul C
o -
tt^
8°
OJ
f?
,j
I
m—* —
J
i
1
t
„*
DRT
;
*
;
,
l
/ A€F ,
<
- PCfcT 2
f
i
[
-
1
1
'
|
1.0
.6 .5 .A
.2 .1
FYR
ITE INTRODUCE )
IN LINE
PROeE(P04lT|aN)
PORT.I
OU:T
Figure 13b. C02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
241
-------
* 1.0
t/>
°I
o ^
uj r
o "
SS
u
(X <
rg
O
u
«T 3
1
.7 .6 .5
"1
FEREiNCE I
I
I
Figure 13b. C02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
242
-------
t
z
• -J
I/I
•
i
t
i
1
t
PO
'WE
^
T 2
• 1
1
i
PROBE
-
—
0>OS<
. . . ,
- -—
I
TION
— —
!
-
- —
o x
UJ r
O -
* *l
o o
u
(VJ
O
u
3 .2 .1 0
. _ - -
TfON
2)
i
— —
"
i
i
Figure 13b. C02 Recorded Output Signal at 0.5 1n/m1n - Response Curves
243
-------
CO- RECORDED OUTPU T SIGNAL
AT
ct>
joO
-t(F
H
O
IV)
rti
ro-
<•»
in
1
O
•
w
-------
The normalized values at the different sampling locations
varied for S02 concentration by ±37 ppm from a mean value of 993 ppm
(with a lowest value of 935 ppm and a highest value of 1055 ppm); for C0?
concentration by ±0.7% from a mean value of 13.0% (with a lowest value of
11.8% and a highest value of 13.9%); and for 02 concentration by ±1.1%
from a mean value of 5.2% (with a lowest value of 4% and a highest value
of 7%).
The velocity measured In the duct ranged from as low as
10 m/s to as high as 16 m/s with an average value of 14 m/s and a -standard
deviation of ±2.0 m/s. Temperature stratification was also observed varying
from 120°C to 140°C.
The coefficient of variation of the concentration times
velocity term was ±16%, ±16%, ±30% for S02, COg, and 02, respectively;
the lowest and highest deviation from the mean was -28% and +21% for S02,
-28% and + 21% for C02, and -45% and +51% for Og. The average S0g/C02 con-
centration ratio was equal to 76.7 x 10" with a coefficient of variation of
26%.
A few'sections of the recorder outputs are displayed In
Figures 13-a and 13-b. Stratification for both S02 and C02 was recorded.
The response of the C02 analyzer was faster than that of the S02 analyzer,
as shown In the recorded calibration curves. It took approximately 3 minutes
for the C02 analyzer to approach a 100% value and 6 minutes for the S02
analyzer to approach 100% value.
The average concentrations and CV term for both ducts are
given 1n Table 4-c. S02, 02 and C02 were respectively, 955 ppm, 5.8%, 12.3%
(dry basis). The coefficients of variation for the CxV term for SQ2, 02,
and C02 were, respectively, ±19%, ±25% and ±20%. The average S02/C02 concen-
tration ratio was equal to 77.9 x 10"4 with a coefficient of variation of 2.8%.
245
-------
Humidity test results are shown in Table 5-a. The percent
moisture calculation, based on average values for wet and dry bulb temper-
atures from both tests, is given in Appendix I, Calculation A. The percent
moisture in the north and south ducts was equal to 11% and 10%, respectively.
The molecular weight of the flue gas stream is calculated in Appendix I,
Calculation B. The relative gas density for the north and south ducts was
equal to 0.99 and 1.01 respectively. Emission calculations and total
flue gas flow rate are given in Appendix I, Calculation C. The estimated
q
S02 emission from both north and south ducts was equal to 475 x 10 kg/s,
17% higher than the calculated amount from fuel oil analysis (see Appendix I,
Residual Fuel Oil Analysis). The C0« emission from both ducts was equal to
42 kg/s, 35% higher than the calculated amount from fuel oil analysis. The
3
total flow measured was equal to 176 m /s at standard conditions (273°K and
101.33 N/m2), which is equivalent to 373,000 SCFM (dry basis) or 573,000 ACFM
(dry basis). An analysis of these results is presented in the following section'
c. Discussion of Results and Conclusions
Actual concentration and velocity data, describing stratifi-
cation in a flue gas stream have been recorded at Boston Edison's Mystic Station
for the air preheater outlet ducts for the higher sulfur oil fired Unit No. 6,
immediately downstream of a 90° bend. A 9 probe equal area strategy was used.
A summary of the results of the demonstration test is given in Table 6.
The expected emissions of S02 and CO-, based on the residual
fuel oil analysis (as shown in Appendix I), were equal to 405 x 10~3 kg/s
and 31.0 kg/s, respectively. The total S02 and C02 emissions from the south
and north ducts as calculated in Appendix I, Calculation C are equal to
475 x 10"3 kg/s and 42 kg/s, respectively. This implies a +17% error on S02
emission and a +35% error on C02 emission. The total measured flow rate was
equal to 176.0 m3/s (dry basis) at STP*.
273°K and 101.33 N/m2
246
-------
TABLE 6
SUMMARY OF THE DEMONSTRATION TEST
RESULTS FROM BOTH DUCTS
Test Load 144 ™ 9ross output
Fuel Consumption 9.92 kg/s (9820 gallons/hr, sp. gr. 0.96)
Fuel 011 Analysis 2.1% S, 85.36% C, 11.45% H
(North Duct) (South Duct)
Average S02 Concentration 917 ppm 993 ppm
Average C0« Concentration 11.6% 13.0%
Average 0« Concentration 6.3% 5.2%
Average S02/C02 Concentration 79.1 x 10" 76.7 x 10"4
Average Velocity 15 m/s 14 m/s
Average % Moisture 11X 10*
Average Temperature 138°C 130°C
Average Static Pressure 97.2 N/m x 10 97.6 N/m2 x 103
Effective Area of Duct 10.5 m 10.5 m
S02 Emission 229 * 10 W* 246 x ™-3 kg/s
C02 Emission 19-9 k9/s 22.1 kg/s
Flow Rate (DRY BASIS) At
273°K and 101.33 N/m2 x 103 89.5 m /s 86.5 m /s
247
-------
The system was further analyzed 1n an attempt to discover the
source of the error. Theoretical combustion based on 18% excess air (Cal-
3
culatlon D) forcasted 12.08 m of dry gases/kg of oil at standard conditions.*
The estimated sulfur dioxide concentration was equal to 1180 ppm (dry basis)
and the estimated carbon dioxide concentration was equal to 13.2% (dry basis).
Because air in leakage occurred at the air preheater, the excess dilution
caused by air was calculated from an oxygen balance (see Calculation E,
Appenldx I); the expected S02 and COg concentrations after dilution were
found to be equal to 1055 ppm and 11.8% respectively. The average SO^ con-
centration from measured values was equal to 955 ppm, I.e., 9% lower than
expected. The average C02 concentration from measured values was equal to
12.3%, I.e., 4% higher than expected.
An analysis of the dry flue gas volumetric rate was performed
(Appendix I, Calculation E). Based on the assumption that a +26% error
originates from the flow rate measurements (+17% - (-9%)), air 1n leakage was
estimated to b6 equal to 19.9 m /s (42,000 SCFM). As a result of this analysis,
the oxygen concentration balance closed to within 9% and the C0« emission
error to within 5%. Therefore, it is concluded that the large error in mea-
suring the emission rate originated from an Inadequate flow measurement.
55
It was previously stated that "an accurate determination of
the quality of gas flow 1s a prerequisite to a realistic evaluation of the
total pollution effluent". The flow conditions at the after air preheater
ducts were severe. The ducts followed a 90° bend, and a pulsating flow was
encountered. The probable flow pattern is shown In Figure 14a and 14b. By
using the S-pitot tube as a measuring device in a pulsating flow and not
attempting to measure the velocity near the walls, additive errors were generated'
The boundary layer error had already been estimated from preliminary test results
to be near +8%; however, larger errors can also be expected . Consequently,
the true flow is less than the measured flow, as was discovered in this
experimental test.
273°K and 101.33 N/m2
248
-------
air preheater
reference ' '
port
plan
Indoors
-------
From A1r Preheater
Reverse Flow Zone
Dust Accumulation
INJ
O1
o
90° Turn Without
Corrective Devices
Figure 14b. Probable Flow Pattern At fhe After A1r Preheater Ducts
-------
To Improve the accuracy and reliability of flow measurements
1n severe flow conditions (reverse flow, pulsating flow, etc.), new flow
measuring devices have to be developed to handle all these flow conditions.
Also, the boundary layer must be elevated 1n detail or Its error accounted
for 1n the analysis.
Because of the errors involved In the instrumental measure-
ment and sampling, the S02/C02 concentration ratio technique can not be
evaluated to better than experimental errors, i.e., analyzer accuracies, boiler
feed rate, fuel composition, etc., (see estimations in Appendix I).
251
-------
V. DISCUSSION AND RECOMMENDED PROCEDURES
A. GENERAL
At the conclusion of the three tasks in this program, two
approaches can be identified which may be applied as the present state-
of-the-art for extracting a representative gas sample from stratified
process streams. These approaches are:
1. Single point sample extraction and analysis for the ratio
of the test gas species, e.g., S02, to a reference species which is an
intrinsic tracer of the process, e.g., C02.
2. Multi-point sample extraction of sample gas.
In our opinion, approach 1 is generally preferred over approach 2,
on the basis of instrumental simplicity and absence of possible flow mea-
surement problems.
At this time the exact accuracy of each method can not be specified
in general terms. In the demonstration task (Task III), we found that it
was not possible to close the material balance using approach 2 because of
suspected error in the flow/velocity measurement, which was estimated at
about 26% from the stoichiometry of the process. (This problem is explained
in greater detail in the previous section and Appendix I.) The results
from the intrinsic tracer or reference gas technique were encouraging, al-
though there are still uncertainties associated with the gas analyzer mea-
surement and the measurements for process parameters such as fuel rate.
The instrumental advantage of this procedure is that a single extraction
probe may be used and proportional sampling is not required.
252
-------
B. PROCEDURES FOR REPRESENTATIVE SAMPLING USING AN INTRINSIC
TRACER OR REFERENCE GAS
A reference tracer gas, produced during the process which
produces the pollutant gas, may be used with process rate data to determine
the mass flow or emission of the pollutant species. For example, in a com-
bustion system such as described in Task III of this program, the mass flow
of S0« was determined from the ratio of the concentration of S02 to the con-
centration of C02 in the flue gas, while the mass flow of C02 was calculated
for fuel firing rate and fuel analysis, viz.:
p
M = §, M'
where: M 1s the pollutant gas mass flow
M1 1s the reference gas mass flow
C 1s the pollutant concentration
C' Is the reference concentration.
Before applying this procedure, the test plane should be checked
for gas stratification by surveying with a gas analyzer for the pollutant
species desired or by Orsat or Fyrlte for C02 or Og. As a rule of thumb,
about 16 samples well dispersed over the test plane should be taken. If
stratification 1s not present, a stratified gas procedure 1s not necessary.
If stratification 1s present, the intrinsic tracer/reference gas procedure
may be applicable. In order to determine whether or not this procedure is
applicable, a survey should be made by traversing the sample probe from the
sampling system over the test plane, Incorporating the two analyzers. This
check is necessary to assure that the test gases and reference gases are
stratified 1n the same manner. If it is known that the process stream is
a mixture from two or more units manifolded together and the fuel in each
unit 1s not Identical, this method will not be applicable.
253
-------
The survey data Indicate that the set of ratios of test gas con-
centration to reference gas concentration should be within ±10% of the mean
value for approximately 16 measurements. If this condition is met, the
probe should be fixed for operation at or near a location giving the mean
value.* An example of a typical sampling system for a test gas and a
reference gas is given in Section IV under the equipment description for
Task III.
At the same time that the gas concentration ratio is monitored,
it is necessary to obtain process data on the fuel rate and the composition
of the fuel, e.g., carbon content. From this data the mass flow of the
reference gas can be calculated.
This approach is particularly interesting for the special applica-
tion of determining the pollutant removal efficiency of a gas control system
when the inlet and/or outlet gas streams are stratified. The fraction passed
through the control device may be represented in the form of the previous
equation by the following:
i =
M
1
VCo'
W
H'
where the subscripts o and 1 denote outlet and inlet, respectively. To be
valid, the reference^gas species must not have been removed by the control
device, I.e., M'Q * M'r The fractional efficiency of the control system is
given by the following:
These specifications are arbitrary; however, they are in accord with the
procedures used in the field demonstration program. In our demonstration,
this procedure predicted the S02 emission at 9% less than the value pre-
dicted stolchlometrlcally.
254
-------
This technique requires only a single point gas extraction 1n
the ductwork before and after the control device. The sampling system
should be of the type described for S02 and C02 1n Section IV, Task III
of this report. If two sets of analyzers are used to determine the ratio
of Cj/C.1 and Co/Co1 (inlet and outlet), real-time efficiency measurements
can be determined for the control device, thus allowing the device to be
adjusted for optimum operation. This is similar to the approach one would
use if the gases were unstratifled; of course, only the pollutant species
would be measured in this case.
This special procedure obviates the problems of obtaining fuel
rate and fuel analysis data as well as total gas flow Information. For
these reasons, 1t is a very attractive procedure for measuring removal
efficiencies of gas scrubbers.
C. SAMPLING ARRAY PROCEDURE
Sampling array concepts have been discussed earlier in Section IV.
The sampling array procedure 1s generally applicable to all stratified
streams. However, in some cases, the effort required to Implement this
procedure can be unattractively high. In addition, as discussed 1n
Section IV, Task HI, problems associated with measuring gas flow/velocity
appear to adversely effect the measurement of the pollutant mass flow or
emission. For this reason, sampling arrays are considered to be a "second
choice" method to be used when the intrinsic tracer method Is not applicable.
1. Pre-Survey to Assess Degree of Stratification
A preliminary Investigation of the degree of gas and velocity
stratification Is necessary. The gas stream is surveyed at different locations
1n one plane of the duct, using Orsat or Fyrltes for analysis of 02 and C02
and S-p1tot tubes for measurement of velocity. The survey is valid only if
a constant process load is monitored. In this case, spatial variations are
Independent of time variations (unsteady state operations), thus allowing
255
-------
the spatial variations to be fully evaluated. To check the stability of
the operation during testing, a reference probe must be kept at a constant
location 1n the duct while gas 1s withdrawn for analysis. Sequential
analysis of gas and velocity measurements from a reference probe and a
sampling point probe 1s recommended. When gas stratification 1s identified,
a more rigorous survey must follow, If insignificant stratification is
observed, 1t is not necessary to use a stratified gas procedure.
2. Rigorous Survey and Selection of Sampling Points
In order to obtain the complete velocity and concentration
profiles of a gas stream section, the manual methods for traversing rec-
tangular and circular ducts are recommended.
The 48 point equal area method or the 49 British Standard
Methods may be used for rectangular duct traverses. For circular ducts, the
cross section may be divided into 32 equal areas; this method traverses 4
diameters with 8 sampling points per diameter. As shown 1n Section IV, this
number of samples with these locations provides a high probability of estima-
ting the true emission.
When the velocity and concentration profiles are established
for a given process load, these data can be used to determine an appropriate
simplified sampling method which may be employed to obtain the emission rate
to within an acceptable deviation from the real value. These data should be
used in order to select a reduced number of sampling points from the rigorous
survey data which give results that are equal or nearly equal to the results
obtained from the rigorous survey. The possible methods may be selected from
the following:
a. Use the computer program procedures detailed in Appendix J.
256
-------
b. Generate expected values for a selected number of
points, by manual interpolation between the data points, and then compare
the mass flow from the summation of the generated points to the results
from the rigorous survey. When an acceptable estimation, e.g. 5% is
achieved with a minimum number of points, this number can be Implemented
in the measurement array.
It is likely that 9 or 16 points will be required for the
equal area strategy for rectangular ducts. For circular ducts, the probable
requirement will be 16 points obtained from 4 diameters, i.e. 4 points per
diameter for high azimuthal stratification, or 16 points on two diameters
for high radial stratification. If a reverse flow is encountered, a larger
number of sampling points is recommended. When the appropriate number of
sampling points is determined, based on the knowledge of the flow and con-
centration conditions, one can then proceed to design the appropriate
sampling system and array.
3. Design. Construct and Install an Automatic Array of
Proportional Samplers"
The design of a sampling system is dependent mainly on the
nature of the flow and concentration conditions. For non-reverse flow con-
ditions and highly stratified concentration conditions, two alternative
methods may be used to adjust sampling flow in proportion to gas velocity
at the sampling points. One method would Involve sampling from each point
with a flow rate proportional to the velocity at the sampling point. A
schematic diagram of this sampling system is shown In Figure 1. An
alternative method would be to sample sequentially from each point for
times proportional to the velocity at that point. The schematic diagram
of this system is shown in Figure 2.
257
-------
to stack
Ea = As x K x Cav x q
* A • area of sampling nozzle
Figure 1- For Non-Reverse Flow 1n Ducts
258
-------
T ? ? ? 7 ?
As
?i •?!?!?
|i I i i i
M |l j > I
il i III
I I II I
• a. * «. It f
?!
i I!
ii
i M
in
Control Box
Sequential
Operation
• M M I I i
i I M I M i
Figure 2. For Non-Reverse Flow In Ducts
259
-------
When spatial concentration stratification 1s negligible
and/or velocity stratification (C.V.* <^ 10%), one can obtain an average
concentration value by directly sampling with equal flow rates from each
sampling point and mixing the sampling streams. If the value of the total
flow 1s known, a reliable estimate of the emission rate may be obtained. The
schematic diagram of this system 1s shown 1n Figure 3.
When reverse flow exists in the duct, the concentration
and directional velocity for each sampling point must be recorded separately
for each point. The concentration times velocity values are then algebraically
added to account for the direction of the flow (reverse flow is negative
velocity).
Coefficient of variation
260
-------
to
cr>
E = CVA
To
Comnon
Sample
Analyzer
Equal Flow Rate
Figure 3 For Non-Reverse Flow In Ducts
-------
SECTION VI
REFERENCES
1. Personal communication with R,W. Robinson, Manager, Field Testing
and Performance Results, Combustion Engineering, Inc., Windsor,
Connecticut.
2 Personal communication with K.O. Plache, Ellison Instrument D1v. of
Dieter1ch Standard Corp., Boulder, Colorado.
3. Spence, R.D., e_t al, and C.E. Rodes, "A Polymeric Interface for
Monitoring SC>2 Emission from Stationary Sources", Paper No. 42C
presented at the 74th National Meeting of the AICHE, New Orleans,
Louisiana, March 1973, pp. 11-15.
4. Nieuwenhulzen, O.K. and H. Posthumus, "An Accurate Simple Method for
Air-Flow Measurement in Field Tests on Air-Cooled Heat Transfer
Equipment", J_. Inst. Fuel, 4p_, 313 (1967), pp. 45-47.
5. Nonhebel, G., ed., Gas Purification Processes for A1r Pollution
Control. London.
6. Stalrmand, C.J., "The Sampling of Dust Laden Gases", Trans.Inst.
Chem. Engrs., 29 (1951), pp. 15-44.
7. Cooper, H.B.H., Jr., and A.T. Rossano, Jr., "Source Testing for Air
Pollution Control", Environmental Science Services, 24 Danbury Rd.,
Wilton, Conn. 06897, Div. of E.R.A., pp. 63-66.
8. Hyde, P., "Partlculate Sampling of Wlgman Burners", Forest Research
Laboratory, Oregon State University, CorvalUs, Oregon, October,
1968.
9. Rosinskl, J., and A. Lleberman, "An Automatic Isokinetlc Sampling",
Appl. Sci. Res., 6_, Sec. A (1956), pp. 92-96.
10. Coenen, W., "A Simple Flow Sensitive Arrangement With a Thermistor
and its Technical Application to Dust Measurement", Staub-Reinhalt
Luft, 29_, No. 11 (1969), pp. 16-22.
11. Grlndell, D.H., "Monitoring Smoke and Flue Dust Emission", AEI Eng.,
2., No. 5 (1962), pp. 229-235.
12. Gilmore, J.S., et al, "State of the Art: 1971 Instrumentation for
Measurement of "PartTculate Emissions from Combustion Sources, Volume
II: Partlculate Mass - Detail Report", Thermo-Systems, Inc., St. Paul,
Minnesota (PB-202-666), April 1971.
13. Granville, R.A. and W.G. Jaffrey, "Dust and Grit in Flue Gases",
Engineering, Feb. 27, 1959, p. 285.
262
-------
14. Wilson, K.D. and D.A. Falgout, "A New Approach to Isoklnetic Null
Probe Design", Environmental Engineering, Inc., Gainsville, Florida,
Paper No. 72-32.
15. Edouard, L., "La Mesure des Concentrations en Poussiere des Gaz a
I'Entre'e des Chemine"es de la Centrale Thermlque de Crell, Le G$n1e
Civil, Physique IndustHelle, 1961, p.270.
16. Donoso, J.J., "An Automatic Multiple Smoke Sampler", AIME TRANS.,
P_- Metals, 188 (March 1950), pp. 610-612.
17. Hyde, P., "Particulate Sampling of Wigman Burners", Forest Research
Laboratory, Oregon State University, Corvallis, Oregon, October 1968.
18. Moore, A.S., "Sampling Dust in the Bureau of Mines Coal-Fired Gas
Turbine", Combustion. (October 1973), pp. 28-30.
19. Jackson, M.R., et al_, "Prototype Fly Ash Monitor for Municipal Incin-
erator Stacks",Troc. National Incinerator Conf. (1970), p. 182.
20. "Development of an Automatic Fly Ash Monitor", report prepared by'IIT
Research Institute under contract to APWA Research Foundation, Chicago,
Illinois.
21. Personal communication with Dr. Winston, IKOR Inc., 2nd Avenue,
Burlington, Mass. 01803.
22 Bosch J , "Device for the Continuous Determination of the Dust Flow
1n Flowing Gases", Staub-Relnhalt Luft, 32, No. 11 (1972), pp. 8-14.
23 Ounsted, D., "A Rapid, Multipoint, Oxygen Analyzer for Power Station
Flue Gases", J_. Inst. Fuel, 4£ (1969), pp. 408-411.
24 Williamson, R.C. and J.A. Russell, "On-L1ne Gas Analysis of Jet Engine
Exhaust", Society of Automotive Engineers (SAE), Combined Fuels and
Lubricants, Power Plant and Transportation Meetings, Pittsburgh, Pa.,
October 30 - November 3, 1967. Paper No. 670945.
25 Fair J.R., B.B. Crocker, and H.R. Null, "Sampling and Analyzing",
Chem. Eng., (Sept. 18, 1972), p. 146.
26 Dresia H and F. Spohr, "Experience With the RadlometHc Dust
Measuring'Unit Beta Staubmeter \Staub-Re1nhalt Luft, 31_, No. 6 (1971),
p. 19. (English).
27. Benedict, R.P., Fundamentals of Temperature, Pressure and Flow
Measurements,. John Wiley & Sons, inc. 1969. p. 129.
263
-------
28. Fitton, A. and C. P. Sayles, "The Collection of a Representative
Flue Dust Sample", Engineering, p. 229, Feb. 22, 1952.
29. Personal communication with Prof. Paul Glever, University of
Michigan, (1969).
30. Sherwood and Plgford, Absorption and Extraction. McGraw H111 Book Co.,
New York (1952).
31. Luxl, F. C., "Analyzing and Control of Oxygen in Boiler Flue Gas",
Paper No. 61-WA-340, ASME Annual Meeting (1961).
32. ASME, "Flue and Exhaust Gas Analysis", Report No. PTC 19.10 (1968).
33. Hawks!ey, P. G. W., et.al., "Measurement of Solids in Flue Gases",
British Coal Utilization Research Association, Leatherhead (1961).
34. Personal communication with R. Larkin, NAPCA, Cincinnati, Ohio,
35i 9Yn]-rAE' and R< C* Pankhurst» Measurement of Air Flow. Pergamon Press
(1966). ——————___-___™__
36. Federal Register, Vol. 36, No. 247, EPA, Standards of Performance
for New Stationary Sources, (December 23, 1971).
37. "Determining Dust Concentration in a Gas Stream", PTC 27-1957, The
Amer. Soc. of Mech. Engineers, New York (1957).
38. "Methods of Testing Fans for General Purposes", Part I, B S 848
British Standards Institution, London, (1963).
39. "Flow Measurement", B. S. 1042: 1943, British Standards Institution,
London (1951).
40. Haaland, H. H., Editor, "Methods for Determination of Velocity, Volume,
Dust and Mist Contents of Gases", Bulletin WP-50, 7th Ed., Western
Precipitation D1v., Joy Manufacturing Co., Los Angeles (1968).
41' {WO)"2* Subcomm1ttee VI' Tenta^ve Standard Method for Sampling Stacks,
42. ASME, Fluid Velocity Measurement. PTC 19.5.3-1965.
43. Keffer, J. F. and Baines, W. D., "The Round Turbulent Jet in A Cross
Wind", Journal of Fluid Mechanics. Vol. 15, Pt. 4, pp. 481-496, (1963).
44. Platten, J. L. and Keffer, J, F., "Deflected Turbulent Jet Flows",
Journal of Applied Mechanics. Vol. 38, No. 4, pp. 756-758, (1971).
264
-------
45. Pratte, B. D. and Balnes, W. D., "Profiles of the Round Turbulent Jet
In A Cross Flow", Journal of the Hydraulics Division. ASCE, Vol. 92,
No. HY6, pp. 53-64, (1967).
46. Wu, J., "Near-Field Trajectory of Turbulent Jets Discharged at Various
Inclinations Into a Uniform Cross Flow", AIAA Journal, Vol. 11, No 11
pp. 1579-1581, (1973).
47. Chi 1 ton, T. H., Genereaux, R. P., "The Mixing of Gases for Reaction",
AIChE Trans., 25, (1930).
48. Hoult, D. P., Weil, J. C., "Turbulent Plume in a Laminar Cross Flow",
Atmospheric Environment, 6, (1972).
49. Deutsche, N., DIN 4702 Blattz, Seite 16, Beuthvertrieb GmbH, Berlin
30, Koln, (DecemEer 1967).
50. B. P. Research Centre, Sunbury on Thames, Middlesex, R. W. Butcher,
"Continuous Combustion Equipment", Tech Memo No. 120496 Project No.
106, 15.1.68.
51. B. P. Research Centre, Petroleum Div. Sunbury on Thames, "The Sampling
of Gases from Ducts, Design of Multihole Sampling Probes'1, Tech Memo
No. 110030, Project No. 110, 7.3.60.
52. Personal communication with Dr. A. Orlng, Pittsburgh Energy Research
Center, Pennsylvania.
53. Personal communication with Falkenberry, TVA, Chatanooga, Tennessee. • '
54. Prepared for Environmental Protection Agency under Contract No.
22-69-95.
55. Burton, C. L., "Quantitation of Stack Gas Flow", Journal of the Air
Pollution Control Association, Vol. 29, No. 8, p. 631, (1972).
265
-------
TECHNICAL REPORT DATA
(Fteate read Iminictloni on the reverie be fort committing)
\. REPORT NO.
EPA-650/2-74-
-a and b
2,
4. TITLE AND SUBTITLE
Procedures for Measurement in Stratified Gases,
Volumes I and II (Appendices)
3, RECIPIENT'S ACCESSION-NO.
B. REPORT DATE
September 1974
S, PERFORMING ORGANIZATION CODE
r.AUTHOR^AZakak> R.siegel, J.McCoy, S.Arab-Ismali
J. Porter, L. Harris, L. Forney, and R. Lisk
8. PERFORMING ORGANIZATION REPORT NO
9 PERFORMING OPO \NIZATION NAME AND ADDRESS
Walden Research Division of Abcor, Inc.
201 Vassar Street
Cambridge, Mass. 02139
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21ACX-092
i. 6SNf RACt/GRANt NO.
68-02-1306
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 6/73-5/74
14. SPONSORING AGENCY CODE
IS. BUTFttMfNTARY NOTES
is. AMTRACT The reporf giveg results of a program to develop methods for the continuous
extraction of representative gas samples from gas streams that exhibit compositional
stratification. The program considered available data in the literature, as well as
field data generated during the program. Wind tunnel tests and mathematical model-
ing were used to develop sampling methodologies which are recommended. Data from
the literature, as well as program data, indicate that stratification exists, although
it is unlikely that gas stratification is as widespread or as severe as particulate
stratification. Depending on conditions, two different methods are recommended.
The first method involves monitoring the ratio of SO2, NOx. etc. to CO2 at a single
location. Then, from the measured fuel flow and chemistry of the process. the mass
flow of CO2 can be predicted. The product of the measured ratio and the predicted
mass flow of CO2 is the mass flow of the pollutant. Where conditions do not permit
using this method, it is recommended that a schedule of manual surveys be conducted
followed by installation of a multi-element proportional sampler and gas velocity
array.
17.
Kf V WORD* AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFHRS/OPIN ENDED TlRMS
e. COSATI F
Air Pollution
Measurement
Gases
Sampling
Stratification
Wind Tunnels
Mathematical Models
Data
Carbon Dioxide
Air Pollution Control
Stationary Sources
Proportional Sampler
Gas Velocity Array
13B , 12A
14B
07D, 07B
IS. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThliRtpon)
Unclassified
21.N0.6MAdE*
276
20. SECURITY CLASS (ThltptftJ
Unclassified
22. PRICl
«M (••rm ssao-t
266
------- |