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

-------