EPA-650/2-75-020

February 1975
Environmental Protection Technology Series
                  CONTINUOUS MEASUREMENT
                     OF  TOTAL GAS  FLOWRATE
                  FROM  STATIONARY SOURCES
                                OHice o( Research and Development

                                US Environmental P'oteclion Agency

                                     Washington. DC 20460

-------
                            EPA-650/2-75-020
CONTINUOUS  MEASUREMENT
  OF  TOTAL GAS  FLOWRATE
FROM  STATIONARY  SOURCES
                  by

          E. F. Brooks, E. C. Beder,
     C. A. Flegal, D. J. Luciani, and R. Williams

             TRW Systems Group
              One Space Park
         Redondo Beach, California 90278
           Contract No. 68-02-0636
             ROAP No. 21ACX-AE
          Program Element No. 1AB013
     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

               February 1975

-------
                        EPA REVIEW NOTICE

 This report has been reviewed by the National Environmental Research
 Center - Research Triangle Park. Office of Research and Development,
 EPA, and approved for publication.  Approval does not signify that the
 contents necessarily reflect the views and policies of the Environmental
 Protection Agency, nor does mention of trade names or commercial
 products constitute endorsement or recommendation for use.
                   RESEARCH REPORTING SERIES


Research  reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency , have been grouped into series.  These broad
categories were established to facilitate further development and applica-
tion of environmental technology.  Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:

          1.  ENVIRONMENTAL HEALTH EFFECTS RESEARCH

          2. ENVIRONMENTAL PROTECTION TECHNOLOGY

          3. ECOLOGICAL RESEARCH

          4. ENVIRONMENTAL MONITORING
          5. SOCIOECONOM1C ENVIRONMENTAL STUDIES

          6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
          9. MISCELLANEOUS

This report has  been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series.   This series describes research performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia  22161.
                              11

-------
                                ABSTRACT

    The program objective was to evaluate hardware and techniques for the
continuous measurement of the total  gas flowrate from stationary sources,
specifically in large or complex ducts where total flow metering devices
such as plate orifices are not practical.  Work consisted of formulation
of operating specifications, evaluation of commercially available velocity
sensors, development and evaluation  of flow mapping techniques, and
field demonstration of both hardware and technique.  Results showed that
total volumetric flowrate can be measured with accuracies consistently
better than 10% in either circular or rectangular ducts through proper
placement of from one to eight flow sensors, when standard traversal
techniques would require twenty to fifty traverse points.  The rectangular
duct mapping techniques developed during the program were found to have
optimum accuracy immediately downstream of an elbow.  Several  off-the-
shelf velocity sensors were found acceptable for use in the specified
stack-type environment.  The field demonstrations verified the acceptability
of both hardware and techniques.
    This report was submitted in fulfillment of Contract No. 68-02-0636,
by TRW Systems Group under the sponsorship of the Environmental Protection
Agency.  Work was completed as of March, 1975.
                                   iii

-------
                            CONTENTS
                                                           Page
Abstract                                                    iii
List of Figures                                              vi
List of Tables                                               xi
Acknowledgements                                            xiv
Sections
I     Conclusions                                             1
II    Recommendations                                         2
III   Introduction                                            3
IV    Task I - Operating Specifications                       6
      4.1  General                                            6
      4.2  Flow Environment                                   6
      4.3  External Environment                               8
      4.4  Instrument System Performance Specifications       8
V     Task II - Instrument Selection                         17
      5.1  Literature Bibliography                           17
      5.2  Instrument Survey                                 18
      5.3  Instrument Selection                              18
      5.4  Instrument Acquisition                            36
      5.5  Advanced Instrument Acoustic Velocimeter          36
VI    Task III - Mapping Technique Evaluation                50
      6.1  General                                           50
      6.2  Description of Facilities                         50
      6.3  Mapping Techniques                                55
      6.4  Mapping Test Results                              67
      6.5  Mapping Test Summary                             105
                               iv

-------
                      CONTENTS - continued

Sections                                                   Page
VII   Task IV - Laboratory Assessment                       112
      7.1  Facility Description and Scope of Testing        112
      7.2  Analytical  Instrument Description                116
      7.3  Basic Instrument Calibration                     121
      7.4  Stability                                        142
      7.5  Time Response                                    146
      7.6  Sensitivity to Orientation                       146
      7.7  Environmental Testing                            149
      7.8  Laboratory Test Summary and Final Evaluation     165
VIII  Task V - Field Demonstrations                         191
      8.1  General                                           191
      8.2  Facility Description                             191
      8.3  Test Conduct                                     194
      8.4  Flow Data Dorrelation                            197
      8.5  Test Results                                     198
      8.6  Summary  of Results                               213
IX    Discussion of Results                                 216
X     References                                            220
XI    Glossary                                              221
XII   Appendices                                            223

-------
                                 FIGURES

No.
 1.  Flow Measurement Control  Volume                               9
 2.  Duct Analogy for Transformation  of  Test  Velocity to
     Standard Velocity                                            12
 3.  Combined Reversed Pi tot  Tube                                  21
 4.  Flare Gas Flow Probe                                          22
 5.  Flare Gas Probe Specifications                                23
 6.  Drag Meter                                                   26
 7.  Thermo Systems Hot Film  Sensors  Tested                        28
 8.  TSI  Metal Clad Sensor and Anemometer                          29
 9.  TSI  Specification Sheet                                       30
10.  Elementary Fluidic Velocity Sensor                            32
11.  Annubar                                                      35
12.  Acoustic Measurement  of  Flow Velocity in Duct  (Schematic)     38
13.  Stack Flue Interior Noise,  Mohave Power  Plant  Unit 1,
     March 2, 1973, Operating at 760  ±5  Megawatts.   B&K 4136
     Microphone.   One-third Octave Spectrum                        42
14.  Stock Flue Interior Noise,  Mohave Power  Plant  Unit 1,
     March 2, 1973, Operating at 760  ±5  Megawatts.   B&K 4136
     Microphone.   Narrow,Band Spectrum                             43
15.  Stack Flue Interior Noise,  Mohave Power  Plant  Unit 1,
     March 2, 1973, .Operating at 760  ±5  Megawatts.   Statham
     PL 80TC-0.3-350 Pressure Transducer.  Narrow Band
     Spectrum                                                     44
16.  Stack Flue Interior Wall  Vibration, Mohave  Power Plant,
     March 2, 1973, Operating at 760  ±5  Megawatts.   B&K 4136
     Microphone.   One-third Octave Spectrum                        45
17.  Precipitator Inlet Flue  Interior Noise,  Mohave Power
     Plant Unit 1, March 3, 1973, Operating at 775  ±5 Mega-
     watts.  B&K 4136 Microphone.  One-third  Octave Spectrum       46
                                     vi.

-------
                            FIGURES - Continued
No.                                                               Page
18.  Precipitator Inlet Flue Interior Noise,  Mohave Power
     Plant Unit 1, March 3, 1973,  Operating  at 775 ±5 Mega-
     watts.   B&K 4136 Microphone.   Narrow Band Spectrum            47
19.  Precipitator Inlet Flue Interior Wall  Vibration, Mohave
     Power Plant Unit 1, March 3,  1975,  Operating at 775 ±5
     Megawatts.  Narrow Band Power Spectral  Density                48
20.  Circular Duct Mapping Test Configuration, 1973                51
21.  Pitot Traverse Installation -- Circular Duct                  52
22.  Rectangular Duct Mapping Test Configurations, 1973            53
23.  Schematic of Rectangular Duct Reference Traverse Map          54
24.  Annubar Location During Duct Mapping Test,  1973               56
25.  Schematic -- Top View of Mapping Test Facility, 1974          57
26.  Position and Orientation of Reference Probes                  58
27.  Position and Orientation of Test Probes for Circular
     Section                                                       59
28.  Rectangular Section Probe Locations                           60
29.  Probe Locations Relative to Elbows in 1974 Mapping Tests      61
30.  Probe Locations -- 1975 Testing                               62
31.  Four and Five Point Methods for Rectangular Duct Mapping      66
32.  Row Average Method                                            68
33.  Typical Flow Profiles from Circular Duct Mapping Test         69
34.  Circular Inlet Blockage -- 1974                               72
35.  Results of Four and Five Point Analysis for a Partially
     Developed Flow Run                                            80
36.  Rectangular  Inlet  Blockage -- 1974                            94
37.  Probe Placement for General Rectangular Duct Applications    109
38.  Probe Placement after a Rectangular Elbow                    110
                                    VII

-------
                           FIGURES - Continued

No.                                                                Page

39.  Low Speed Wind Tunnel                                           113

40.  TSI Wind Tunnel                                                114

41.  Coal Fired Combustion  Flue Gas Simulator                       115

42.  Manufacturer's Accuracy for Instruments Tested                 122

43.  S Probe Calibration Factor                                     124

44.  Schematic of Ramapo Mark V Flow Meter                          125

45.  Ramapo Mark V Calibration:  Free Stream Velocity Versus
     Ratio of Output Voltage to Input Voltage                       127

46.  Difference between Factory Calibration and Test Velocity
     in Percent Versus Test Velocity for Ramapo Mark V              128

47.  Diagrammatic Explanation of Ramapo Output at High Speed
     in Wind Tunnel Testing                                         129

48.  Results of Force Calibration of Ramapo Mark V as Percent
     Difference in Factory  and Test Velocity Versus Equivalent
     Test Velocity                                                  130

49.  Moment Applied at Strain Gauge Bridge of Ramapo Mark V         132

50.  Factory and Test Calibration Curves for Hastings-Raydist
     AFI-10K Probe as Probe Output Voltage Versus Standard
     Velocity                                                       134

51.  Absolute Difference in Percent between Velocity Data Points
     and Reference Curve Fit Velocity Versus Standard Velocity
     for Hastings Probe  .                                           135

52.  Comparison of Calibration Curves for Three Hastings
     Probes and Test Data for AFI-10K Probe                         136

53.  Calibration of TSI Metal Clad Sensor as Power Squared
     Versus Velocity                                                138

54.  Calibration of TSI Metal Backed Sensor as Power Squared
     Versus Velocity                                                139

55.  Calibration of TSI Wedge Sensor as Power Squared
     Versus Velocity                                                140

56..  Annub'ar Calffiratton Factor                                     143

57.  Bamapo probe Sta&lli;ty Dgta Reduction S&eet                    144

                                   vii i

-------
                           FIGURES - Continued
No.                                                                Page
58.  S Probe Orientation Angles                                     148
59.  S Pitot Probe Orientation Sensitivity Data                     150
60.  Ramapo Probe Orientation Sensitivity Data                       151
61.  Hastings-Raydist Probe Orientation Sensitivity  Data             152
62.  Annubar Orientation Sensitivity Data                           153
63.  TSI Metal Clad Sensor Orientation Sensitivity as
     Normalized Velocity Versus Orientation Angle                   154
64.  TSI Metal Backed Sensor Orientation Sensitivity as
     Normalized Velocity Versus Orientation Angle                   155
65.  TSI Wedge Sensor Orientation Sensitivity as Normalized
     Velocity Versus Orientation Angle                              156
66.  Accuracy of Hastings AFI-10K as Percent Difference Between
     Test Data Points and Calibration Curve Fit Versus Velocity      160
67.  Ramapo Purged Probe                                            161
68.  Pre-and Post-Environmental Calibrations of TSI  Metal  Clad
     Sensor as Sensor Power Squares Versus Sensitivity               162
69.  Pre-and Post-Environmental Calibrations of TSI  Metal  Backed
     Sensor as Sensor Power Squared Versus Velocity                  163
70.  Pre-and Post-Environmental Calibration of TSI Wedge
     Sensor as Sensor Power Squared Versus Velocity                  164
71.  Ramapo Mark VI                                                 170
72.  Ramapo Mark VI Specification Sheet                             171
73.  Hastings Stack Meter                                           174
74.  Hastings Stack Gas Meter Specification Sheet                   175
75.  TSI Total Vector Acemometer                                    178
76.  TSI Total Vector Anemometer Specifications                     179
77.  Approximate Pi tot-Static Probe Response at Large  Yaw Angles     186
78.  Schematic of Moapa Power Plant                                 192
                                   IX

-------
                            FIGURES  -  Continued
                                                                   Page
79.  Field Demonstration Rectangular Duct Geometry                   193
80.  Schematic of Traversing Mechanism for Ramapo Drag  Meter         195
81.  Stack Velocity Profiles                                        203

-------
                                  TABLES

No.                                                              Page
 1.  Key Word Classification for Flow Instruments                  19
 2.  Method for Determining Volumetric Flow from Point
     Measurements in a Circular Duct                              64
 3.  Circular Duct Mapping Test Results, 1973                     71
 4.  1974 Circular Duct Point Sensor Mapping Results               73
 5.  Annubar Calibration Factors for 1974 Calibration  Tests  --
     Configurations 3 and 4                                       74
 6.  Summary of 1974 Circular Duct Mapping Test Results            75
 7.  Results of Rectangular Duct Analysis by Four  and  Five
     Point Methods           .                                     81
 8.  Results of Rectangular Duct Analysis by Rows                  85
 9.  Examination of 1973 Row Average Data after a  Rectangular
     Elbow                                                        88
10.  Annubar Duct Measurement Results, 1973                       90
11.  Summary of Annubar Duct Measurement Results,  1973            92
12.  1974 Row Average Results after a Straight Inlet
     (Configuration 1)                                            95
13.  Annubar Calibration Factors for 1974 Calibration  Tests  —
     Configuration 1                                              96
14.  Acceptable Flow Mapping Region for Constant Inlet Geometry
     and Variable Flowrate for 1974 Mapping Test — Straight
     Rectangular  Inlet                                            98
15.  1974 Row Average Results after a Rectangular  Elbow
     (Configuration 2)                                           100
16.  Annubar Calibration Factors for 1974 Calibration  Tests  —
     Configuration 2                                             101
17.  Row Average  Results for 1975 Testing                        103
18.  Annubar Results for 1975 Testing                            104
                                    XI.

-------
                            TABLES  - Continued

No.                                                              Page
19.  Recommended Circular Duct Flow Measurement Techniques        107
20.  Recommended Rectangular Duct Flow Measurement Techniques     111
21.  TSI Hot Film Sensor Calibration Summary                      141
22.  Stability Test Results                                       145
23.  Response Time Test Results                                   147
24.  Orientation Sensitivity Test Results                         157
25.  Post Environmental Test Calibration Results                  159
26.  Ellison Instruments Combined Reversed ("S") Pitot Probe
     Calibration Results                                          166
27.  Ramapo Mark V Flow Meter Calibration Results                 173
28.  Haystings-Raydist AFI-10K Gas Flow Probe Calibration
     Results                                                      176
29.  Thermo Systems Hot Film Sensor Calibration Results           181
30.  Annubar Calibration Results                                  183
31.  Yaw Characteristics of Pitot-Static Probes                   185
32   Laboratory Test Summary                                      188
33.  Duct Traverse Summary, 1974                                  200
34.  Duct Traverse Summary, 1975                                  201
35.  Stack Traverse Summary, 1975                                 201
36.  Summary of Annubar Continuous Monitoring Data with Scrubber
     Off, 1974                                                    205
37.  Computation of Total  Flow from Coal Analysis and Measured
     C02 Concentration, 1974                                      206
38.  Average Daily Flow through Ducts,  1975                       208
39.  Average Daily Flow through Stack,  1975                       209
                                    xn

-------
                            TABLES  -  Continued


No.                                                                   Page

40.  Summary of Annubar and Ramapo  Short Term Monitoring Data,
     1974                                                             210

41.  Field Demonstration Pitot Traverse Data, 1974                    211

42.  Row and Column Analysis of Field Test Pitot-Static Traverse
     Data, 1974                                                       212

43.  Summary of Pitot Traverse Row Average Data, 1975                 214

-------
                 ACKNOWLEDGEMENTS

     Program personnel wish to thank the instrument
manufacturers whose probes were evaluated for their
help in answering questions which arose during the
testing.  Ellison Instruments in particular supplied
several probes and also on-site advice during the
first field demonstration.  TRW Systems Group is very
grateful to the Nevada Power Company and the Southern
California Edison Company for their assistance and use
of facilities for the field demonstrations.
                        xiv

-------
                      SECTION  I
                     CONCLUSIONS
Desired system accuracy specifications (5% to 10% of full
scale) can be met with hardware and measurement techniques
examined during the program.
Total flow measurement can normally be made with better
accuracy in circular ducts than in rectangular ducts, due
to the way in which the duct itself conditions the flow.
Several point mapping techniques, notably the Log Linear 4
method, are adequate for total  flow measurement in cir-
cular ducts, using eight or less sampling points.
The  row averaging technique, using eight or less velocity
sensors, is adequate for total  flow measurement in rectanq-
ular ducts when supported by in place calibration.

The region immediately downstream of an elbow has been
identified as the most desirable place to determine total
volumetric flow in a rectangular duct when line averaging
techniques are used.
The Ellison Annubar is adequate for measurement in circular
ducts, and also in rectangular  ducts when calibrated in
place.  A purge capability may  be required in moist, par-
ti cul ate laden flows.
The Ramapo Fluid Drag meter had the best characteristics
among the point sensors tested.
The S pi tot probe tested had several undesirable charac-
teristics, including shifts in  calibration as a function
of Reynolds number and high angular sensitivity.  For
normal manual traversing, it is believed that the
ellipsoidal nosed pitot-static  probe can be used to
obtain data several  percent more accurate than can be
obtained using an S probe.
The field demonstrations confirmed the basic adequacy
of the techniques and sensors tested.

-------
                       SECTION  II
                    RECOMMENDATIONS

Results of this program should be coordinated with gas
composition sampling technology to produce a system to
determine both total flow rate and gas composition.
Manufacturers of flow sensors which may be used in the
applications of interest during the program should
consider offering total systems to include measurement,
data processing and readout capabilities.  Also, velocity
sensors should have the capability to measure local
pressure and temperature, since these parameters are
always required for data reduction.
Inexpensive, reliable linear traversing mechanisms for
point sensors should be made available as a shelf item.
Consideration should be given  to replacing the centroid
of equal areas method by the Log Linear method for
circular duct mapping.
To optimize accuracy in EPA Method 2  (Reference 1), con-
sideration should be given to  converting velocity to
standard conditions before averaging, rather than after.
Work should be done to standardize design characteristics
of the S pitot probe to optimize the  accuracy.
Ellipsoidal and hemispherical  nosed pitot-static probes
were designed to minimize error with  respect to the
magnitude  of the total velocity vector.  For applications
of interest in this program,  it would be very desirable
to have a  probe which would minimize  error with respect
to the axial component of the  velocity vector.

-------
                              SECTION III
                              INTRODUCTION
     The program is concerned with the measurement of total  volumetric
flow on a continuous basis.  Volumetric  flow measurement is important as
an end in itself, as well as being a requirement in support of other
measurements.  Examples of the former are total  emissions measurement for
regulatory purposes, and monitoring of process stream devices such as
blowers for control and maintenance purposes.  Total  flow measurement is
required in support of pollutant concentration measurements to translate
a concentration reading into mass output per unit time.   Program specifications
are based on continuous monitoring requirements in full  scale coal fired
electric utility plants.
     The more often a monitoring measurement must be  performed, either for
regulatory or control purposes,  the more desirable it is to have a reliable,
in place system which can continually perform the measurement without
requiring an operator.  In-line  flow metering technology for small lines or
long straight pipes is capable of providing reliable, highly accurate
measurements.  This program is not concerned with these  applications but
rather with ones where the duct  size, shape, or configuration makes metering
of the total flow impractical.
     The class of sensors considered cons-ists of devices which are inserted
into the flow stream.  The total flow rate  is then inferred from the readings
of these instruments.  Thus the  primary  program objectives have been to
determine what available instrumentation is applicable to the problem,  and
to determine what techniques are required to correctly use the instruments.
     The general approach used is similar to that given  in the standard
EPA Methods 1 and 2 (Reference 1) -- taking velocity  measurements at a
number of discrete points in the flow and averaging the  data to obtain  a
total flow rate.  These standard methods are not generally practical for
continuous monitoring due to the large number of data points usually
required, so one program objective was to determine the  minimum number
and location of sampling points  required for accurate measurement.
                                  -3-

-------
     The program began in October 1972, and the technical  effort was
concluded in March 1975.   The task breakdown was as  follows:
     Task I  Operating Specifications.   This task involved the formulation
of environmental, instrument, and system specifications in order to identify
and bound the problem being considered.  The specifications were used in
part to evaluate accuracy, reliability, and survivability  of  candidate
sensors and measurement systems.
     Task II Instrument Selection.  This task consisted of a  survey of
available instrumentation and subsequent analysis of manufacturers' data
to select those instruments best qualified for further evaluation.   A
special effort was devoted to determine the feasibility of using acoustic
flowmeters.
     Task III  Mapping Technique Evaluation.  Task II results clearly
showed that most applicable velocity instruments were point sensors,  such
as pitot probes.  This meant that development and/or evaluation of flow
mapping techniques in both circular and non-circular ducts was required.
Task III involved development of test facilities and testing  to determine
optimum sensor deployment schemes.
     Task IV  Laboratory Assessment.  Instruments selected for further
evaluation in Task II were tested in a low speed wind tunnel  to determine
their accuracy characteristics.  Environmental tests were  also performed
to determine sensor survivability under conditions specified  in Task I.
Instruments best suited for field test operation were then selected.
     Task V  Field Demonstration.  Continuous flow monitoring systems
were successfully tested at the Nevada Power Company station  at Moapa,
Nevada.
     Success was achieved in the development of rectangular duct mapping
techniques which require as few as one velocity sensor regardless of duct
size.  Acceptability of mapping techniques such as the Log-Linear method
for use in circular ducts was verified.  Acceptable techniques involve
the use of between one and eight velocity sensing devices.  Instruments
found acceptable both from the standpoint of accuracy and survivability
include the Ellison Annubar, which was the only non-point sensor tested,
and the Ramapo Fluid Drag Meter.
                                   -4-

-------
     EPA Methods 1 and 2 were both examined as reference manual  techniques.
It is being recommended that the Log-Linear method be considered as an
improvement over the centroid of equal  areas mapping technique for measure-
ment in circular ducts.  In addition, the standard pitot-static  probe
should be used whenever possible in place of the S type probe in order
to improve system accuracy.
     Final results indicate that continuous flow measurement accuracies
of 5% to 10% can be readily attained in non-developed flows in circular
and rectangular ducts using state of the art instrumentation.
                                  -5-

-------
                               SECTION IV
                   TASK I  - OPERATING SPECIFICATIONS
4.1  GENERAL
     The following specifications served as a base for instrument and
technique selection and evaluation.   They were obtained in part from the
program RFP and work statement.  The flow environment data apply specif-
ically to coal fired electric generating plants.
4.2  FLOW ENVIRONMENT
     •   Gas Temperature - Range:-18 to 200°Cwith fluctuations of
         +17°C in a one hour period.
     •   Gas Composition
              Constituent               Concentration Range (Vol)
                 S02                          0-3000 ppm
                 S03                          0-100 ppm
                 CO                           0-50 ppm
                 C02                          12-18%
                 H20                          4-14%
                 NOX                          0-500 ppm
                 02                           1-5%
                 Hydrocarbon                  0-1500 ppm
                 Np                           Balance
     •   Particulate Loading - Range 0-7 grams per standard cubic me-
         ter,  with particle sizes ranges from submicron to 300y.
         Loadings may be exceeded on a temporary basis due to soot
         blowing.  The particulate material is frequently extremely
         adhesive due to the moisture and H2S04 mist content of the
         flue gases.  Insensitivity or resistance to both fouling
         and abrasion are therefore required.

 Below 230°C, S03 may be present as H2SO» mist, at a loading of 0-0.1
 grams/son.

                                    -6-

-------
     •    Water Mist  Loading  -  Normally  below  0.3 grams/son in  process
         equipment demister  effluent  gases.   Note:  The mists  present
         in  process  equipment  effluent  gases  can, depending upon the
         abatement process,  be contaminated with limestone, calcium
         sulfate and sulfite,  magnesia, magnesium sulfate  and  sulfite,
         sulfuric acid,  sodium sulfate  and sulfite, and ammonium
         sulfite-bisulfite.  Resistance to the  effects of  these
         materials is,  therefore,  required.
     •    Pressure -  Range 600  - 790 torr, with  27 torr  (mm Hg)
         vacuum to 36 torr positive pressure departure
         from ambient,  with  pulsations  at 30-120 cycles per minute.
     •    Velocity -  Magnitude  1.5-38  meters  per second, with fluctua-
         tions of j^6 m/s,  and flow periodicity of 30-120  cycles
         per minute; direction primarily  parallel to  duct  axis
         but may be  random at  any  point,  including reverse flow.
     The following are  not specifications but are general  application
descriptions.
     •    Vibration - Maximum frequency  1.0 HZ
                     Maximum amplitude  1.0cm   displacement
     •    Access - 1-inch pipe  nipple  to 12-inch x 12-inch  inspection
         plate
     •    Disturbances - Eddy currents and static zones due to  duct
         changes in  size and direction  both  upstream  and downstream
         of the probe/sensing  element zone.
     •    States of Motion (Flow Types)  -  Both laminar and  turbulent,
         with changes between  flow types  occurring as a function of
         changes in  plant load.
     •    Duct Sizes  - Pilot scale, range  7-70 cm.  diameter.  Full
         scale 1-12  meter diameter.
     •    Duct Shapes -  Circular, rectangular, and irregular.
     •    Duct Axis Orientations -  Vertical,  horizontal, and at any
         angle between  vertical and horizontal.

                                   -7-

-------
     •   Length of Straight Run (Ducting) - 0-10 equivalent duct
         diameters between direction or cross-sectional  shape
         changes
4.3  EXTERNAL ENVIRONMENT
     The following conditions are applicable to parts of the system not
in the flow.
     •   Temperature - Range -29°C to 55°C, with direct sunlight on exposed
         components
     •   Pressure -  Range 600 to 790 torr, with fluctuations of  +25 torr

     •   Relative Humidity - 0-100%, with conditions such that
         condensation forms water or frost
     •   Moisture - Rainfall, snow, and melting snow
     •   Vibration - Amplitudes as high as  1.0 cm displacement,
         and frequencies as high as 1.0 KHZ
     t   Electrical Power - 110V or 220V AC, 50-60 HZ, with +20%
         fluctuations
4.4  INSTRUMENT SYSTEM PERFORMANCE SPECIFICATIONS
4.4.1  Definition of Parameter to Be Measured
     The term  "total volumetric flow" must be properly defined in order
to have unambiguous physical significance.  From a fluid mechanics
standpoint,  it is simpler to work in terms of total mass flow rate, and
then define  volumetric flow rate in terms of the mass flow.  Consider the
duct shown  in  Figure 1.  The four sides and the entry and exit planes form
a control volume.  By definition, the sides are solid, so that all fluid
must enter  through the left plane and leave through the right plane.  For
simplicity  (which does not compromise accuracy), assume that the flow
rate into the  control volume is always exactly the same as the flow rate
out of  the  control volume.  This relation becomes true as the size of
the control  volume approaches zero.  The flow through the control volume
may then be given as the  flow through the exit plane of the control volume:
                                    -8-

-------
                Mean  Flow
                Direction
Flow enters from the left and exits
to the right.

Velocity at a point in the exit plane is given
by
     tT = u T  +  v ~j + w k
where
     T, J, "k* are unit vectors in the directions
     shown, forming an orthogonal coordinate system
and
     u, v, w are the scalar components of u in

     the ~t, *^ and Tc directions, respectively

The vector~n is the unit vector normal to the
exit plane, so that
    —»  . .A
     n = i

and the net flow component out of the duct at the point
shown is

     TT . ~n  =  u
     Figure 1.   Flow measurement control  volume
                         -9-

-------
                   m = //ptT • rT dA = pu A                            (1)
                       A
where
                   •
                   m = mass flow rate, gm/sec
                                                 •3
                   p  = local fluid density, gm/cm  (gas phase only)
                   ~u = local velocity (vector), m/sec
                   "n~ = unit vector normal to exit plane, dimensionless
                                         2
                   A = exit plane area, m
                   u = tT • "n", scalar velocity component normal to exit
                       plane, m/sec
This equation and its application in terms of hardware became the base on
which much of the program was built.  The average of the product of the
normal velocity component and the fluid density is exactly defined by
this equation, and the proper definition of the terms "average velocity"
and "volumetric flow" must in turn follow from the definition of p"u~.
     It is common practice to evaluate equation (1) in a large pipe or
duct by performing a velocity traverse, for example using EPA methods 1
and 2.  The integration in equation 1 is then approximated by the
summation
                         N
                   m;  !>nVAn                                   (2)
                        n=l
for N area segments.   Usually the segments are of equal area, so that
                          N
                   "• = &  X>nun
                          n=l
     In terms of commonly measured parameters in a gas, this becomes
                          n=l
where
                   p = gas static pressure, torr
                                   -10-

-------
                                                2
                   R = gas constant, e.g.  in
                                             sec °K

                   T = gas static temperature, °K

The flow through any one area segment is then assumed to be

                   \ - fT  R^ un ' VVA

In order to switch from mass flow rate to volume flow rate, the concept
of "standard conditions" needs to be introduced.  This is illustrated in
Figure 2.  For this control volume, the gas enters at arbitrary pressure
and temperature.  It then undergoes whatever changes are required in
order to emerge uniformly with a static temperature of 20°C and at an
absolute static pressure of 760 torr, which have been defined as standard
conditions for this program.  Also by definition, the gas is considered
to be chemically frozen and there are no phase changes.   This is important
from a practical measurement standpoint.  It means that  the mass flow rate
being considered is that of the gaseous components only, since common
velocity sensors for operation in a gas stream respond to gas flow, and
not to liquids or solids which may be entrained in the flow.  Therefore
defining a gas flow in terms of standard rather than actual conditions
implies changes in pressure and temperature only -- not  composition and
phase.  Thus for any one area segment where the velocity and density are
considered to be uniform,

                   m = pUAA = p u AA                                 (5)
where
                   ( )  = value at standard atmospheric conditions.
Given this background, it is now possible for us to define volumetric
flow rate and average velocity.
     For any one area segment, the volumetric flow at standard conditions
is given as
                   V,  = a. AA                                       (6)
                    sn     n
                                    -11-

-------
//









J
X
X
X
X
X



TEMPERATURE
PRESSURE
DENSITY
VELOCITY

AREA
/ /








ml_
1
X
X
X
X
s

,
my
-- ~
7
x
X
X
X
X
STATION STATION
1 2
(TEST) (STANDARD)
Ti TS
Pi Ps
Pl ps
V] Vc
s
A A
.Xi










>
/
/
X









FROM CONSERVATION OF MASS,
                               = m
SO
AND
                        PI  vi A = PS  vs A
                               PI
   Figure 2.   Duct analogy  for  transformation of test
              velocity to standard velocity
                           -12-

-------
where
                    V   = volumetric flow rate at standard conditions
                      n    3
                          m /sec
and volumetric flow at actual conditions is given as

                    V  = u M                                          m
                     n    n
where
                    Vn = volumetric flow rate at actual conditions,
                          3
                         m /sec
For the term V to have an unambiguous meaning, the actual flow temperature
and pressure must always be given and must be constant for the region  of
interest.  In the term Vg, the pressure and temperature are given by
definition.  For the entire duct as shown in Figure 1, we have
where V  is the total volumetric flow rate at standard conditions.  There
       ^                                  •
is no directly corresponding equation for V except for the special case
of zero density stratification.  This is explored in more detail in
Appendix A.  All volumetric flow data presented in this report are given
as volumetric flow rate at standard conditions unless otherwise noted.
     In summary, the following relations are those used in the program:

                    ms = pu~ A  Total mass flow                        (1)

                   Vs  = uj A  Total volumetric flow at               (8)
                               standard conditions
and by definition are related as follows:

                    m  = puA = P~~ u~ A = P~~ V                         (9)
                     b          bo      b  5
4.4.2  Preliminary System Error Analysis
     Recall from equation (4) that the mass flow rate through any one
area segment during a traverse is given by
                    •  _ A   p
                    01  " N  ET u
                                    -13-

-------
The uncertainty in this single point measurement may then be given by
(see Reference 2 and Appendix A):
         02   =(Mr  02  + fam r   2 .  [am I2   2
          mn    \9A/

                 +

where

          o   = standard deviation of quantity x
           /\
If a measurement system is postulated which determines each of the varia-
bles independently of the others, the above equation can be written
                    222222
                  O'    0«    0_     On     OT     0.,
                   m_A_£      R      T      u                /,,x
                   To" ~   o"   ";> T    5"    "~o       ?"               VI I /
                   nr    A     p      R£    r      u

Under the best of circumstances, the following assumptions may be made
for normal field measurements:

                   1.  OA = .01 A

     This is based upon the assumption of a 2a error of 1% in the duct
linear dimensions.
                   o   o    Op   aT
                       -f-T'-T- -005
These are based on 2o accuracy of 1% for each case.
Thus we have
                     2                        2
                   o-          o          o   a
                    ?5-= (.Ol)2 + 3(.005)2 + —SJ-
                    m                         u
                   a'2              2
                       = .000175 + a u
                    .2             — p"
                    m
                                    -14-

-------
     The equation can now be examined in terms of velocity uncertainty,
which is of central importance to the program.  For zero uncertainty in
velocity, we get

                   2am = ± .026 m

i.e. a 2.6% uncertainty for a 95% confidence level.  For 2a  = .Olu, .05u,
and .10u, respectively, we obtain 2am   = .028m , .056m ,  and .103m .   This
shows that under optimum circumstances, a local accuracy of about 3% can
be attained when the accuracy of each individual  measurement is 1%.  It
also shows that if the velocity accuracy is considerably worse than the
others, the system uncertainty is only slightly greater than the velocity
uncertainty.
     The above analysis is for a single point measurement only, and does
not consider errors resulting from differences in point measurement and
true average flow through the area segment, time lag error, and other
error sources, which add to the minimum 3% error.   The analysis shows  the
parameters which must be involved for any orthodox measurement system,
and an estimate of the best results to be reasonably expected.  Within
this scope,the following specifications are recommended for individual
components and the combined system.
4.4.3  Component Specifications
     •   Range - any component such as a thermocouple, pressure
         transducer, etc., must be compatible with the range
         given in Section 4.2.
     0   Accuracy - Within +3% of full scale reading for velocity
         instrument subsystems (such as pitot probe-pressure
         transducer combinations) in a uniform flow field; within
         +2% of full scale reading for support measurements such
         as absolute temperature and pressure.
     •   Velocity probe angular response - For a velocity probe
         aligned with the duct axis, error in measurement of
         axial velocity component should be less than +5% for
         off-axis flow angles up to 10°.
                                   -15-

-------
4.4.4  System Specifications
     •   Accuracy - Within +_5% of values  obtained  by pitot trans-
         versal  in circular pipes with straight runs greater than
         ten pipe diameters; within +_10%  of values obtained by
         pitot traversal  in circular pipes with less than  10 pipe
         diameters of straight run or in  noncircular ducts.
     •   Desired repeatability =  +_2% of full  scale
     •   Desired zero and span drift - Zero drift  <+1.5%/24 hour;
         full scale drift <+_1.5%/24 hour  (with standard mainten-
         ance) ;  and <+3% week.
     •   Desired response time -  <30 seconds  for 95% of full  scale
         flow change.
     •   Calibration - maximum of once per week, by pitot  tube
         traverse.  Note:  Mass balance methods may be used as
         alternative reference techniques where data of sufficient
         accuracy is available.
     •   Maintainability - Routine maintenance < one-half  hour
         per eight hour shift (probe installed and available for
         use).  Refurbishment  ^12 man hours  per month (probe
         installed and available  for use).
     •   Readout - Readout in plant control room with continuous
         printout on hard copy as volumetric  gas flow per  unit
         of time versus calendar  (or chart) time.
                                    -16-

-------
                                SECTION V
                     TASK II  -  INSTRUMENT SELECTION

5.1  LITERATURE BIBLIOGRAPHY
     A literature search was  performed to gather data in  the following
general  areas related to the  program:
     •   Instrumentation (Vendor catalogues)
     •   Velocity Measurement Techniques
     •   Instrument Survivability Problems
     •   Stationary Source Air Pollution
     t   Gas Sampling Techniques
     •   Thermodynamic Parameter Measurements
     •   Power Utility Characteristics
     t   Error Propogation
Reports were categorized, numbered and cross  referenced by computer.
Vendor data was the most used in order to produce the instrument survey
described in the following section.
     Three reference books should be singled  out due to their importance
during the program.  The first is "The Measurement of Air Flow," by
Ower and Pankhurst (Reference 3).  The book deals in detail  with many
common velocity measurement devices, including the pitot  probe and its
variations.  It also examines flow in pipes and associated measurement
techniques.  It was the best single source of information found during
the program.  More technically detailed information for some specific
problems was obtained from "Boundary Layer Theory," by Schlichting
(Reference 4), including Reynolds principle of similarity, turbulent pipe
flow, and flow through pipes of non-circular cross-section.   Error
analysis techniques were adapted from "The Analysis of Physical Measure-
ments," by Pugh and Winslow (Reference 2), which provides an excellent
insight into system error propagation.
                                    -17-

-------
5.2  INSTRUMENT SURVEY
     A greater variety of instruments is available for flow measurement
than for the measurement of practically any other engineering parameter.
A literature and vendor search turned up the list of "key words"  shown
in Table I to categorize various types of flow instruments.  The  general
industry survey results are presented in Appendix B, which lists  manu-
facturers, medium (gas, liquid or solid) for instrument use, general
instrument classification as given in Table I, and the appropriate TRW
file document number.  This list was compiled in late 1972 and therefore
does not contain any new instruments produced since then.
5.3  INSTRUMENT SELECTION
     Two simple elimination steps were used to narrow the field of
instruments to those most applicable to this program.  The first was
procedural, and that was to consider only gas flow measurement instruments,
which is in line with program specifications.  The second decision was to
eliminate from consideration instruments which are designed solely for use
in either small or long, straight, circular pipes, such as orifice or
venturi meters or other metering systems which process the total  flow.
This program  is concerned with applications where the duct may not be
round,  straight runs  are very short, and characteristic dimensions of the
ducting make  an instrument  to process the total  flow  impractical.
     The  remaining instruments can be characterized most generally as
those instruments which  react to localized  conditions in the flow, usually
by being  immersed in  it, in a way which allows estimation  of the total
flow  by inferring a  relationship between the  local  and total flow.  For
example,  in  a pitot  traverse, the total flow  is  determined  by assuming
that  the  average flow at the  discrete  sampling points is equivalent to
the  total  flow.  The  instrument may  make measurements at a  single point,
such  as a stationary  pitot  probe; at a  number of points, such as in a
pitot traverse; along a  line, such as  an acoustic velocimeter, or in  a
plane,  such  as a drag meter with a  large target.
      After the initial  list was  narrowed to fit  the scope  of the program,
a final selection of probes for  test was performed.   Selection was based
 largely on user recommendations  when available.   The instruments chosen

                                    -18-

-------
Table 1.  KEY WORD CLASSIFICATION FOR FLOW INSTRUMENTS
Number
        Type
   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
electromagnetic
flow tubes
laminar
mass
non-obstructing
nozzles
open channel
orifice meter
pi tot-type
positive displacement
turbine
variable area rotometer
Venturis
thermal anemometer
mechanical anemometer
acoustic
vortex shedding
other
                          •19-

-------
and the reasons for their selection are given below,  the reasons based
on information available at the time, late 1972.   The state of the art
of acoustic velocimeters was specifically evaluated  as a separate sub-
task, and is discussed below in Section 5.5,
5.3.1  S Type (Stauscheibe) Pitot Probe
     Since this is the most commonly used field instrument for point
velocity measurement, it was determined that  its  properties should be
examined.  The specific probe used was provided by the Ellison Instrument
Division of the Dieterich Standard Corporation.  The  probe and its
specifications are shown in Figure 3.
5.3.2  Hastings-Raydist Flare Gas Flow Probe  (Figures, 4, 51
     The probe is actually a type of pitot probe that is constructed
with two openings at the probe tip.  These openings are connected together
by an  internal stainless steel tube.  A portion of this tube is heated and
thermo-electric sensors measure temperature gradients along the wall of
the  tube, external to the  flow stream.  Purge gas is injected into the
tubing  in an  arrangement which forms a pneumatic bridge in a manner similar
to that  of  a  gas  density balance widely used as a detector in gas chroma-
tography.   At zero line velocity,  the bridge is balanced and purge gas
exhausts out  both openings at the  probe tip equally.
     As  flow  across  the tip occurs,  a differential pressure is developed
as with  any pitot type probe, unbalancing the bridge.  Purge gas still
exhausts out  both openings, but now  they are slightly unequal.  The thermo-
electric sensors measure the shift in temperature gradients along the
heated  portion of the tube which are related to the main gas flow creating
the  differential  pressure  at the tip.
     Because  the  purge gas is continuously exhausted into the flowing gas
whose  velocity is being measured,  corrosive or particulate laden gas is
prevented from entering the probe  to produce fouling.  For this reason,
the  Hastings  Mass Flow Probe has found wide use in refineries for measuring
the  velocity  of gases in flares contaminated with tar-like material.
     The Hastings Mass Flow Probe  is normally  calibrated for air in a
wind tunnel using either air or nitrogen as the purge gas.  A calibration
                                    -20-

-------
                                                                              2000 SERIES —TECHNICAL DATA ON
                                                                              COMBINED  REVERSED  PITOT TUBES
DIGTEHICH STANDARD CORP • DRAWER M • BOULDER COLORADO 3O3DB USA • TEL 3O3 m«7-1OOO • TLX OB-BO3 • CABLB OIBBTAN ODO
       COMBINED  REVERSED PITOT  TUBE

       The  Ellison  Combined Reversed pilot tube is  designed to the
       Stauscheibe concept. This  pilot tube has numerous design fea-
       tures. It can be  used  to measure many liquids, clean or  dust-
       laden air, and steam. It also produces a higher velocity pressure
       or differential gage reading for a given flow rate which allows
       measuring of lower flow rates and produces a more  accurate
       measure of all flow rates. If the large sensing tubes need clean-
       ing after use, they can be  easily cleaned without special tools.
       A  special water-cooled version of the Combined Reversed type
       is available for use with temperatures up to +2000*F. These im-
       portant features  have  made Ellison's Combined Reversed  pitot
       tube the choice of many flow specialists around the world.
                                    Combined Reversed Type
       INSTALLATION  PROCEDURE

       As with all pitot tube installations,  it is of  primary importance
       that  the impact tube be pointed directly upstream. This means
       that  the opening of one of the bent  tubes on the Combined Re-
       versed type points directly upstream, and the other points down-
       stream. To insure a uniform flow and accurate measurement, it is
       desirable to have as much unobstructed pipe or duct as possible
       upstream  from  a  pitot  tube as well as a reasonable  distance
       downstream. Unobstructed  pipe or duct means no bends, tees,
       valves, or changes in diameter. For air or gas, 15 to 25 diameters
       of the pipe or duct, depending on velocity are sufficient for the
       upstream  side.  For  liquids and  steam 20 to  50 diameters are
       desired. Downstream lengths should be Mi to Vz of  the upstream
       length. If  the above suggested lengths are not possible, a com-
       plete traverse is recommended. It  is very important that the rate
       of flow be stabilized during the time a complete manual traverse
       is being made. Any change in velocity will greatly affect the
       accuracy of the calculated average velocity obtained from the
       traverse.
        FORMULAS FOR VELOCITY AND  VOLUME

           Velocity —Combined Reversed Type Only

            Gas  Flow                      V = 903.4

            Air Flow*                      V = 3300

            Liquids —Wet Seal, using Hg     V = 3008

            Water" —Dry Seal              V = 107.5
                                           V = 395.9

            Water'* —Wet Seal, using Hg     V .- 381.1
SELECTING  A MEASURING
INSTRUMENT

The  pitot tube may be  used  with a wide variety of measuring,
recording, or controlling instruments. It is most commonly used
with an Inclined manometer or a Vertical manometer.  Ask for a
copy of Ellison's new General Catalog. To determine  the maxi-
mum scale range of the  instrument you will need, use the follow-
ing formulas:
      hw = (V/3300)1 for Air.

      hw = d(V/903.4)' for Gas.

      hm = (V/381.1)' for Water-Wet Seal to  Mercury  (only)

      hw = (V/107)' for Water -Dry Seal

      hm = (V/395.9)1 for Water —Dry Seal
                                                                      Terms—
       V = velocity in feet per minute

       d = density in pounds per cubic foot

       hw - differential pressure in inches of water

       hm = differential pressure in inches of Mercury

       A = inside area of pipe in square feet

       a ~ inside area of pipe in square inches
 CORRECTIONS FOR  NON-STANDARD CONDITIONS
 —ALL PITOT TUBES

 AIR—For calculating Air Flow at temperatures and pressure other
 than at standard conditions, use Gas Flow formula. Air Density or

        d = 1 325 x Inches of Mercury (Absolute) in pipe
                       459.6 + Temperature in  *F.


 WATER — For calculating Water Flow  at temperatures other than
 60*F., use Liquid  Flow formula.  Water densities, d, are  shown
 below for various temperatures.
Tamp.
+ 32'F.
40
50
60
70
80
90
100
110
120
Deniitxld]
62.42
62.43
62.41
62.37
62.30
62.22
62.11
61.99
61.86
61.71
Temp.
4-130'F
140
150
160
170
160
190
200
210

Oeniity Id)
61.55
61.38
61.20
61.00
60.80
60.58
60.36
60.12
59.88

        -At standard conditions (d = .07495). +70'F. 29.92" Barometer, see correc-
        tion formulas for other conditions.
"At 60*F. with d =62.37. Dry seal is when manometer is calibrated dry
and water does not contact indicating liquid during use. wet seal is when
manometer is calibrated dry, and water does contact indicating liquid
during use.
                                 Figure  3.    Combined  reversed  pitot  tube
                                                               --21-

-------
             HASTINGS-RAYDIST
                                                   Specification Sheet No.  513A
HASTINGS  GAS  FLOW  PROBE    MODEL  API   SERIES
              FOR MEASURING VELOCITY OF WET AND DIRTY GASES
            v.        RANGE:   0-1000 fpm   OR  0-6000 fpm    of*  o-/0,000
                                A DEPENDABLE, NON-CLOGGING FLOWMETER
                                     FOR CONTAMINATED GAS LINES.
FEATURES
• CONTINUOUS PURGE PRINCIPLE
• NO EXPOSED SENSORS 01 WIRES
•EASILY INSTALLED or MOVED
•EXPLOSION-PROOF TYPE HOUSING
• 0-5 VOLT D-C OUTPUT SIGNAL
• PROVIDES LONG LINE TRANSMISSION
  CAPABILITY
•REMOTE: RECORDING, CONTROL, ALARM,
  INDICATION
•PURGE WITH AIR, N? OR PROCESS GAS
• CHOICE OF TWO RANGES: 0-1000 or 0-6000 fpm

GENERAL
  The Hastings Gas Flow Probe is the result of
nearly two decades of experience in dealing with
difficult-to-measure, corrosive, or inflammable
gases.  Using  a unique Hastings  thermal  prin-
ciple with continuous purging, the measured gas
does not come in contact with the internal parts
of the  probe. Thus  plugging,  fouling,  con-
densation, corrosion, etc., are no longer prob-
lems.
  The Probe is constructed entirely of stainless
steel  for  all  parts (internal  and  external)
through which any gas flows. Solid-state circuits
are built into the explosion-proof  type housing
and require only connection toa 24 volt d-c power
source and remote read out. The output signal of
0-5 volts may be connected to any remote data
logging device,  meter,  recorder  or readout
desired. The purge gas required  is quite small
and normally is less than 30 cu.ft. / hr. The probe
may also be used without the purge system in
relatively clean and dry lines where plugging is
not a problem.

PRINCIPLE OF OPERATION

CONTINUOUS PURGE MODE
  The probe is constructed with two openings at
the probe  tip. These openings are connected
together by an internal stainless steel tube. A
portion of this tube is heated and thermo-electric
sensors measure temperature gradients along
the wall of the tube, external to the flow stream.
  Purge  gas is injected into the tubing in an
arrangement which forms a  pneumatic  bridge.
At zero line velocity, the bridge is balanced so
that no flow occurs through the sensing portion of
the tube.  The purge gas exhausts  out  both
openings  equally  at the  probe tip.
  As flow across  the tip occurs, a  differential
pressure  is developed,  unbalancing the bridge
and causing a small amount of purge gas to flow
through  the  sensing section.  Purge gas still
exhausts  out both openings,  but now they  are
slightly unequal.  The thermo-electric sensors
measure  the  shift  in  temperature gradients
along the heated  portion of the tube which  are
related to the main gas flow creating the  dif-
ferential  pressure at the tip.
  Since the purge  gas is continuously exhausting
into the main line, it prevents the main line  gas
from entering the probe and prevents fouling.
         	1,1
p= nc\ ="t lS?i -v
s
;
[

tuGlNG
E 14PS
OMPOtSSlO*
"^ *
:
NSOB
P^>
'-^ t
5 ; 1
]

djRCE GAS
'LOW MOC
MAMlFOLO
AT ZERO Fl'OW
AP • &P?
AP. ' fiP4
NO TjO* "^S1 SErrtO^i
APL> ZERO
flT MAIN LINE FLOW
APL IS DEVELOPED. SO

AT ALL T.MES RJ»0£
C4St«HAUSIS iNIO
PWOBE TiPQFlNNGS
0, «NO 02

MAIN |
— LINE /
   CONTINUOUS PURGING PRINCIPLE ot OPERATION
                          ( Manufactured under one or more U.S. Patents and Pat. Pending)
                      Figure  4.   Flare  gas  flow probe

                                          -22-

-------
PURGE GAS REQUIREMENTS
  Regulated, clean dry air, nitrogen or process
gas may be used as the purge gas. Consumption
is less than 30 standard cu. ft. / hr. at a pressure
approximately 15 psi above the static pressure of
the line  being  measured.  Since  such a small
amount is flowing into relatively large lines, the
percentage of  air added to  the whole is in-
significant. If inert gases are desired, nitrogen is
recommended.

PURGE MANIFOLD
  Included with  the probe  is  a  purge  gas
balancing manifold  built into a weatherproof
conduit type  enclosure. This may  be mounted
anywhere near the probe. It includes two valves
tor balancing the bridge and zeroing the probe.

POWER CONVERTER
  An optional 24 volt d-c converter is offered for
those installations  where 24 vdc is not available.
Built into a Grouse Hinds explosion-proof type
housing and rated  at Vj amp, it is available for
use from either 115 volt or 230 volt a-c lines.

CALIBRATION
  Calibration is related to the gas density and the
velocity  profile.
Velocity = KrK?-V1NO
where
   K,   =Velocity Profile Factor; typically  .8
   K,  = Density Factor;   V .075/ gas density
   VINO = Velocity from probe calibration curve

Example: What is  the full scale (5 volt) range of
an AFI-6K when measuring stack gas having a
density of .092 Ibs./ft.?
   Velocity = K, • K?^V|ND
           = (.8)  ( V  .0757 .092 ) (6000)
           = 4333 fpm

  TYPICAL  CALIBRATION  CURVE
  SELECTION CHART
          Model                Air Range
          AFI-1K               0-1000 fpm
          AFI-6K               0-6000 fpm
   Above includes:   Probe, Purge Minifold, and curve ot
                Air Velocity versus 0-5 volts.
  ACCESSORIES
    Model AOC-2     115 v. i-c/24 v. d-c power converter •
    Model ADC-3     230 v. i-c/24 v. d-c power converter
    Meter 24-1-419   0-5 volt meter for remote reidout
    Meter 24-1-420   0-5 volt meter relay single point control
    Meter 24-1-421   fj-5 «olt meter relay double point control
SPECIFICATIONS
POWER : J« V. a-c ( ' 4 V.) (" JIO ma
OUTPUT:  0-S v. d-t. ft 4 ma (m»).
PURGE GAS: API.IK 5 clh
          AFI-4K 30 clh
          <> IS psig regulated
DIMENSIONS:  ?robe—W . «" « «" Overall
           with U" • u «" o.O. Wand .
MANIFOLD: «" I »" > IV," Weatherproof Type Bo«
MATERIALS: 104 and lit Stainless Steel lor
          all parls in contact with gas
HOUSING: Croine Hindi Type
COMPRESSION SEAL FOR PROBE: Male l»" NPT Threaded
                         Connection
MANIFOLD CONNECTIONS: Vi" O.O. Tubing to Probe, H" NPT to
                   Purge Gas Supply

INSTALLATION

  The probe is supplied with a compression seal
fitting for easy installation. The fitting has a liv
NPT male thread that will easily connect to  a
I'/i"  female threaded  gate  valve.  It may be
mounted in any position  if used  in  the purge
mode.
  The purge manifold is connected to a regulated
gas supply at 10-30 psig. The manifold should be
connected  to the  probe by means of V&" O.D.
tubing.
  Electrically,  2  wires from  the  probe  are
required for the 24 v. d-c input power, and 2 wires
for  connection to the remote read-out.
9
0)4
23
z
r-2
2
3 1
o
0
(
AIF



/
y
>
I VELI
IT

,/
/



/



•^









^.^^—




2 3 4 5 (
DCITY IN THOUSAND OF FPI
HASTINCS-RAYDIST
A TELEDYNE COMPANY
\r1
-.-.- J:
OUTLINE
. .
1 CCMLtT • "' — C
t / , f 1
" "' .'i/" - '"'
DIMENSIONS API SERIES
Literiture lUiilible upon request:
Hjslingi Vicuum Saugn Citaloj No. 3000
Hastings Mcleod Gauge Spec. Sheet No 3408
Hastings Gauge Tube Accessories spec. Sheet No. 35!
Hasting;, Vicuum Gauge Reference Tubes Spec. Sheet No. 351*
Hastings Air-Meiers Calalog No. 4000
Hillings Mass Flcmmeters lor (Uses Calalog No. 5000
Hastings Cilibnted Gis leiis Spec. Sheet No. 9040

SPECIFICATION SHEET NO. 513A PHONE 703-723-6531
HAMPTON, VIRGINIA 23361 TWX: 710-882-0085
PRINTED IN U.S.A. COPYRIGHT O 8-72
                      Figure  5.   Flare  gas  probe specifications

                                               -23-

-------
curve is obtained for air velocity vs output in voltage.   Smaller changes
in voltage will be noted as the velocity increases.
     While air is usually the most convenient purge gas to use, other
gases may be used to obtain different sensitivities.  Nitrogen is used to
replace air with no major change in calibration.   Most other gases such as
methane, propane, or natural gas may also be used but the original air
calibration curve is no longer applicable.   It is not possible to multiply
the air purge curve by a constant factor to obtain the purge curve for
some other gas.  Changing the purge gas only changes the  relationship
between output and voltages and the indicated velocity.  The sensitivity
of this relationship is affected by the thermal conductivity of the purge
gas with methane producing a greater sensitivity than air.
     The Hastings Probe is capable of measuring the velocity of stack
gases from 0 to 30 m/sec, but the accuracy becomes poorer above about
20 m/sec.  The Hastings probe is designed to operate at a temperature
range from -34°C to +315°C with the electrical  zero shifting approximately
5% over this range.  In addition, as with most flow measuring devices, the
gas density must be determined to correct the indicated velocity measure-
ments.  The variations expected in gas composition present no survivability
problems with the Hastings Probe since the probe is constructed of stainless
steel and the purge gas continuously exhausting into the  stack prevents
the stack gases from entering the probe and fouling the system.  Because
the measured gas does not come into the internal  parts of the probe and
the probe is continuously purged, water mists,  limestone, calcium sulfate
and sulfite, sulfuric acid mists, sodium sulfate and sulfite, and ammonia
sulfite-bisulfite resistance appears assured.  One user reported use in
the presence of zinc oxide dust with oxides of sulfur and moisture being
present in the measured gas stream.
     Positive and negative factors identified for these sensors are as
fol1ows:
Positive Factors
     0   Probe can be used in extremely wet, corrosive and particulate
         laden gas streams; Q-J grams of parttculate per cubtc meter
         of gas is acceptable

                                   -24-

-------
     •   Calibrations should remain constant for six months
     •   The probe is very sensitive to changes in flow rate at low
         flow rates
     •   Accuracy of full scale voltage readout is reported by the vendor
         to be +2% of full scale
Negative Factors
     •   The velocity vs voltage output curve is non-linear; changes
         in velocity at the low end of the scale cause the greatest
         change in voltage
     •   The high sensitivity of the probe at low velocities tends to
         accentuate small zero drifts.  One user found probe unsatis-
         factory because of zero shifts
     t   The electrical zero will shift 5% over the operating tem-
         perature range
     •   The probe must be isolated from vibrations
     t   Users contacted indicated initial setup and adjustment
         problems
     •   Gas velocity range between 0 and 20 m/sec can be measured
         with good accuracy; between 20 and 30 m/sec measurements
         have less accuracy
  ^
 5.3.3  Ramapo Mark V Flowmeter (Figure 6)
     Ramapo manufactures the Mark V Flowmeter Probe that measures flow in
 terms of dynamic forces acting on a fixed body in the flow stream.  Bonded
 strain gauges, in a bridge circuit outside the fluid stream and shielded
 by stainless steel, translate this force into an electrical output pro-
 portional to the flow rate squared.  The electronics used are identical
 to those used for pressure transducers.  Gas velocities of 1.5 to 38 meters
 per second can be measured, but the probe has a turn down ratio of 10:1
 which means that two different probes will be required to cover the entire
 range.
     Gas temperatures from 0 to 260°C are acceptable with the Ramapo
 probe.  Corrections to the indicated velocity must be made for changes in

-------
Figure 6.   Drag meter
        -26-

-------
density.  Particulate loadings of 0 to 15 grains per standard cubic foot,
with particle size ranging from submicron to 300 micron are not expected
to be a problem unless the sensing head becomes excessively loaded.
     Positive and negative factors identified for these sensors are
as follows:
Positive Factors
     •   Particulate loadings of 0-7 gm/scm will be acceptable
         if the disc in the probe is coated with Kel-F, or if
         provisions for steam cleaning are included
     •   Resistant to chemical attack (SO^, NO  , H^O, etc.) if purge
         is used in internal cavity
     •   Kennecott Copper of Salt Lake City rated probe as their
         primary selection for use in smelter stacks
     t   TRW has a good service record with device in measuring the
         velocity of liquids
Negative Factors
     t   A purge may be required to protect parts from chemical
         attack
     t   One probe will not cover the velocity range of 1.5 to
         38 m/sec, as identified in the specifications
     •   Limited previous applications for stacks
5.3.4  Thermo-Systerns Incorporated Hot Film Sensors (Figures 7-9)
     Thermo-Systerns manufactures a velocity transducer that measures
velocity, based on calibration at a given temperature and pressure, by
sensing the cooling effect of a moving flow stream over a heated sensor
surface.  The sensor is heated electrically by current from a control
amplifier.  The sensor is one leg of a bridge which is continuously bal-
anced.  Another leg of the bridge serves as a temperature sensor to com-
pensate for temperature changes.  The output signal is proportional to
the square root of velocity.  Standard probes are subject to calibration
shifts due to particulate buildup on the probe.
                                    -27-

-------
                                        NEW

                                 METAL BACKED
                                 HOT FILM SENSOR
                                 - LARGE, RUGGED,
                                   FILM SURFACE
                                 - HEAVY QUARTZ
                                   COATING
        NEW

METAL CLAD SENSOR
- RUGGED
- METAL ENCASED
      SENSOR
  A.  METAL CLAD SENSOR
                                  B.  METAL BACKED SENSOR
                                  WEDGE HOT FILM
                                        OUARTZ R
                                             GOLD FILM ELECTRICAL LEAD*
                                   -OUARTZ COATED PLATINUM
                                   FILM 10.004*1 0040" EACH SIDE)
                                     C.   WEDGE  SENSOR
        Figure 7.   Thermo systems hot film sensors tested
                             -28-

-------
                             Electronic Control Circuit
                            " and Bridge Built Into Probe
                                           All Stainless Steel,
                                           High-Reliability Parti
                                                           Sensor
   VT 161 FEATURES
      Self-Contained — Bridge and
      Electronic Control Integral
      with Probe
      Sensitivity from a few ft/min.
      to over 200 ft/sec.
      New RDM' Bridge System Gives
      High Stability -  No Controls or
      Adjustments Needed
      Operates on Batteries or
      115/230V AC
      Linear Output Available
      Temperature Compensated
      Calibration Included
      Cable  Length Not Critical
   THE VT 161  IS AVAILABLE
   AS FOLLOWS:
   1. Probe Package for 12V DC
     Power - Non-Linear Output -
     No Readout
   2. Probe Package for 115/230V AC
     Power Using Model 1605 Power
     Supply — Non-Linear Output -
     No Readout
   3. Linear Signal Conditioner with
     Probe Package - Analog Readout
   4. Linear Signal Conditioner with
     Probe Package — Digital Readout


   •Patented (Pulse Duration Modulation!
         NEW

METAL CLAD SENSOR
- RUGGED
- METAL ENCASED
       SENSOR
The first really rugged thermal sensor-
takes  advantage of the wide range.
stability and fast response  of thermal
techniques without fragility.
WIDE APPLICATION
    - TESTING
    - RESEARCH
    - CONTROL
       Ducts, Pipes, etc.
       Towers
       Field Studies, Remote Points
       Mobile Equipment
       Multi-Point Studies
                                         VELOCITY
                                      TRANSDUCER


                                      SERIES  VT 161

                                              AIR
                                       AND OTHER GASES
                                          0 - 200 ft/sec
 THERMO-SYSTEMS Inc.
2500 Cleveland Ave.North
 St. Paul, Minnesota  55113
      (612) 633-0550
     Telex (297—482)
Form No. S161-172
                                                                                         Printed in US.A.
                    Figure  8.   TSI  metal  clad  sensor and  anemometer

                                                 -29-

-------
                               MODEL VT  161  VELOCITY TRANSDUCER
          PROBE PACKAGE FOR 12V DC POWER SOURCE - NON-LINEAR - NO READOUT
PRINCIPLE OF OPERATION

The VT 161 measures velocity (actually "standard"
velocity based on calibration at 70 F and 1 atmosphere
or mass flow) by sensing the cooling effect of a moving
flowstream over the heated sensor surface. The sensor
is heated electrically  by  current from the control
amplifier. The sensor is one leg of a bridge which is
continuously balanced  by the unique, POM* amplifier
system, requiring no adjustments. Another leg of the
bridge serves as a temperature sensor  to compensate
for temperature changes. The output signal is propor-
tional to the electrical power dissipated in the sensor
or about the square root of velocity. If desired a linear
output is provided  by a  very accurate curve-fitting
amplifier that  is part of the Signal Conditioner pack-
ages shown below.
                                                                        SPECIFICATIONS
                                                         Velocity Range: 0 to 200 ft/sec (600 mpsl. (Dete n furnished to
                                                           give good accuracy throughout the above range. Any specific
                                                           smaller range can be utilized. Calibration is at standard condi-
                                                        '   tions of temperature and pressure.)
                                                         Output: Approx. 0.5 to 3V DC. Impedance - 100 ohms
                                                         Accuracy: ±0.1% of FS and ±3% of Reading
                                                         Repeatability: leu then 0.1% (repeated conditions)
                                                         Response Time: 50 milliseconds
                                                         Environmental Conditions (Senior Section): Temperature -20°C to
                                                           100°C, Pressure 0.01 to 500 psia
                                                         Cable: 15' Standard (specify otherwise)
                                                         Power Supply: 12V *?'2 V DC (+15V DC optional). 0.5 Amp. Max.

     VCLOCITV (FT/MCI
TYPICAL CALIBRATION
CURVE-VT161
                         MODEL 1605 POWER SUPPLY
                         Converts  116/230V  DC
                         power to 12V DC for op-
                         eration of VT  160 Proem.
                         Convenient terminals serve
                         as junction point for signal
                         leads
        MODEL VT 1612  PROBE WITH LINEAR SIGNAL CONDITIONER-ANALOG  READOUT
                                                                         SPECIFICATIONS
                                                         (Signal conditioner includes power supplies, linearizing circuit*,
                                                         output amplifiers and analog meter.)
                                                         Analog Output: Rear Terminals; 50 ohm impedance; 0 — 6V DC
                                                            full scale range; 50 millisecond response time; and total linear
                                                            accuracy ±5% of reading in 50 to 1 range.
                                                         Analog Meter: 0 - 100% and 0 - 10% of full scale. ±2% fS.
                                                         Cabinet: Aluminum, portable, with stand — 8.5 x 3.5 x 11
                                                         Power Required:  115/230V AC
                                                         Probe Specifications: Same as VT 161 except for linear output and
                                                            AC power. The VT 161 Probe can be retrofitted later to make a
                                                            VT 1612 or VT 1613 system.
                                                                   Model No.
                                                                   VT 1612-1
                                                                          -2
                                                                          3
                                                                          -4
                                                                          •5
                                            Range
                                            0- 1000 fpm
                                            0- 10mpi
                                            0 - lOOfps
                                            0 - 10,000 fpm
                                            0- 100 mps
                                                                   (other ranges may be specified as needed)
           MODEL VT 1613  PROBE WITH LINEAR SIGNAL  CONDITIONER-DIGITAL READOUT
                                                                         SPECIFICATIONS
                                                         (Same specifications for Probe,  Lmeanzer and Analog output as
                                                         above.)
                                                         Digital Display: 3'/. digit, neon; 100% over range; 2 second response
                                                            time, 0.2 second/reading display, and il digit resolution. BCD
                                                            output optional.
                                                         Ranges:     Modal No.             Range (direct reading)
                                                                    VT 1613-1             0-2000 fpm
                                                                          •2             0 - 20.00 mps
                                                                          -3             0 - 200.0 fpi
                                                                          4             0 - 100.0 mps

                                                                    (other ranges may be specified as needed)
         tsij   THERMO-SYSTEMS INC.
                       3000MOUTH CLEVELAND * VC.
                       ST. PAUL. MINNESOTA  SSI 13
612633 0660
                             Figure  9.    TSI  specification  sheet

                                                      -30-

-------
     Positive and negative factors identified  for these sensors  are
   follows:
Positive Factory
     •   Can measure velocities of 1.5 to 38 m/sec
     •   Changes in pressure make only small changes in measured
         velocity
     t   TRW has had experience with advanced  designed transducers
         that operate at high temperatures.  The experience indicates
         the transducers may operate in a stack environment
     •   Transducers can operate above 200°C
Negative Factors
     •   Output in volts vs velocity is not inherently linear, sensi-
         tivity decreases at higher velocities
     •   Sensors are subject to fouling by particulate
     •   Gas viscosity and thermal conductivity must be known or
         sensors must be calibrated in place
5.3.5   FluiDynamics Devices Unlimited Sensors

     FluiDynamics manufactures two flowmeters  considered for testing -
a cross-flowing sensor and a co-flowing sensor.  The fluidic cross-flow
velocity sensor employs a free jet normal to the measured flow,  two
receivers or total head tubes, and a differential pressure gauge, trans-
ducer or manometer to indicate the pressure from the receivers,  as shown
in  Figure 10.  The supply gas can be nitrogen, air, natural gas or other
gases at a  standard supply pressure of 50 psi.  Sensor operation depends
on  the  jet  entraining and mixing with the surrounding fluid.  This causes
the jet to  spread out with distance from the nozzle, and the jet velocity
to  decrease.  Consequently the total head profiles  becomes flatter with
distance from the nozzle.  Entrainment of the cross flow causes jet
deflection.
      Characteristics of the  total head profiles are used to measure  the
jet deflection  by means of two total head tubes that are appropriately

                                   -31-

-------
           ..frfft*
     NOZZLE




    JfcT
VlfASOftKD
          MfAt> TUBES

                             ^r-iuK±.:>jf*^*a
  Figure  10.   Elementary fluidic velocity sensor
                        -32-

-------
located with respect to the nozzle.   The undeflected total  head profile
imposes equal pressure on the two total  head tubes that is  proportional
to the cross-flow velocity, within the operating range.  The relationship
of differential pressure with velocity is linear for the cross-flow
fluidic sensor up to 18 m/sec.
     The cross-flow fluidic sensor has been used in flare gas stack appli-
cations by several oil companies in the United States.   Since flare stack
gas streams contain adhesive particles which cause fouling of the receiver
ports, these users have found it necessary to back purge the receiver
ports.  The users indicate that this purging nearly eliminates fouling
except in extremely dirty flare stacks.
     The FluiDynamic co-flowing fluidic velocity sensor is a device some-
what similar in principle to the cross-flow fluidic velocity sensor but
employing a jet of supply fluid parallel to and in the same direction as
the ambient fluid stream whose velocity is being measured.  The co-
flowing sensor is complementary to the fluidic cross-flow velocity sensor
in that it can be used to measure much higher velocities, i.e., up to 150
or 180 m/sec in air but at the same time overlaps the velocity of the
cross-flow fluidic velocity sensor although not being capable of measuring
the extremely  low velocities that can be measured with the latter.
     In order  to measure the full velocity range, a cross-flow fluidic
sensor  (.3 to  18 m/sec) and a co-flowing sensor (15 to 150 m/sec) must
be used simultaneously.  Both instruments have a linear AP readout over
the indicated  ranges and should be able to use the same supply gas source.
With the proper differential pressure readout device, the required flow
fluctuations should be detectable.  The fluidic sensors measure gas
velocity at  one station in the flow stream.

      The variations expected in the gas composition and temperature will
 present no problems with the FluiDynamic fluidic sensor with respect to
 chemical  resistance since the fluidic sensor is fabricated from 316
 stainless steel.  The sensors are reported by users to have very good
 resistance to particulate fouling.  Positive and negative factors identi-
 fied for these sensors are as follows:
                                  -33-

-------
Positive Factors
     •    Gas temperature range of 0 to 200°C is acceptable
     t    Sensor and probe are fabricated of stainless steel,  therefore,
          resistant to chemical corrosion
     •    Has been used successfully in heavily particulate-laden
          flare stack applications
     •    Has been used successfully in near-saturated and water mist
          laden gas streams
     •    Has been used successfully with trace concentrations of
          H2S04
     •    Excellent maintenance and recalibration record reported by
          users
     •    Rugged construction suitable for field use
     t    Accuracy and repeatability of 3% of reading is expected
     •    Device has been used with continuous readouts on strip
          charts in control house
     •    Excellent sensitivity at low flow rates
     •    Small pressure drop  in gas flow caused by instrument
     •    No flow reference required
Negative Factors
     •    Two sensors required to cover required velocity range, a
          standard Fluidic sensor for  .3-18 m/sec and a co-flowing
          sensor for 15-150 m/sec
     •    Receiver ports may require a gas purge to keep particulate
          matter from clogging up the ports
5.3.6   Ellison  Annubar  (Ellison Instrument Division, Dieterich Standard
        Corporation)  (Figure 11)
     The Annubar is distinct from the other sensors considered in that it
is  not  a point  sensor — it obtains a type of average velocity along a
line which  spans the pipe or duct in which it is placed.  The instrument

                                   -34-

-------
 I
CO
U1
 I
                                      ASME PAR UW 16 DRILL OR BURN
                                      ONE 1-3/8" DIAMETER HOLE AND
                                      PROVIDE 1/4" MINIMUM WELD BE AD
                                      ALIGN TO PIPE AXIS (TYPICAL! ~
                                                                            ASTM SPEC. A18VGRADE No. 2
                                                                            2/3000 LBS. 1"NPT FORGED
                                                                            STEEL WELD COUPLING IS
                                                                            SUPPLIED.
                         4- 3/8"  DIAMETER SENSING PORTS
                         ARE LOCATED IN APPROPRIATE
                         CENTERS OF CONCENTRIC ANNUL).
                                                                        FLANGE INCLUDED
                                                                        TO CUSTOMER'S
                                                                        SPECIFICATIONS
                                                                                         ARMCO 17-4 PH
                                                                                         HARDENED STAINLESS
                                                                                         COMPRESSION FERRULE
                                                                        FACTORY TIG WELDED
                                                                        UNLESS OTHERWISE
                                                                        SPECIFIED
                                                                                                                                      PERMANENT
                                                                                                                                      METAL TAG
                                                                                                                                      WITH 3" CHAIN
                                                                                                                                                   STAINLESS 1/2". 1/4" or
                                                                                                                                                   1/8" NPT.or 1/4" TUBE
                                                                                                                                                   COMPRESSION. OR
                                                                                                                                                   BRASS SHUT-OFF VALVE
                                                                                                                                                   WITH  1/4"SAE FLARE
                                                                                                                                                   CONNECTION
                                                                        1/8" NPTF CLEAN-OUT PLUGS
                                                                                                                              DIETERICH STANDARD CORPORATION

                                                                                                                              DRAWER M BOULDER COLORADO OO3OB UOA
                                                                                                                         751 TO 755 ANNUBAR  FLOW ELEMENTS
                                                                                                                              ; 9-23-70: ADDED CLEAN-OUTS; WAS 27/64
    o    o 1  &  o    o  }••-
                                           .^
         FLANGE MOUNT OPTION
MATERIAL: 303 STAINLESS. OTHER MATERIALS  OPTIONAL.
/»3TO4 PIPE DIAMETERS  IS RECOMMENDED FOR  DOWNSTREAM SIDE.
V<6OR  MORE PIPE DIAMETERS IS RECOMMENDED  FOR UPSTREAM SIDE
W AFTER  VALVES. ELBOWS AND ETC. SEE  CALCULATION REPORT E - 79
   ©PERMANENT TAG SHOWING MIN.. NORM. & MAX. DESIGNED FLOWS. METER
   READINGS FOR DESIGNED FLOWS. TAG NO.. LINE SIZE. SER. NO. S MET6RED FLUID.
   »EC.FPV PIPE" D'
   «-HFrnil F na 11 „
   SCHEDULE OR I.D.
                                                                        • 'ELEMENT MADE TO TH.S D.MENSION
                                                                        UNLESS OTHERWISE SPECIFIED. PARTS
                                                                        |N TH,S REQ|ON g, GASKET NOT SUPPLIED
                                                                        EXCEPT WELD COUPLING
                        1968 - OIETERICH STANDARD CORP.. BOULDER. COLO.
                                                                              Figure  11.   Annubar

-------
output is a differential pressure.   The high pressure component is
obtained from having four upstream  facing holes open into the probe body.
The resultant pressure is then sensed by a single tube in the center of
the probe.  The low pressure component is obtained from a single orifice
facing downstream and also located  at the center of the probe.   The four
forward facing holes are located according to a four point velocity
averaging technique for flow in circular ducts.  The probe output must
then be related to the total flowrate by means of an empirical  calibration
factor.  A probe must be manufactured specifically for a given  installation,
since it completely crosses the pipe or duct.
Positive Factors
     •    The instrument is inherently an averaging one, which is
          desirable for this program
     •    Good results have been reported for stack applications
     t    Corrosion and abrasion effects are considered to be negligible
Negative Factors
     •    The probe appears to be suited only for measurements in cir-
          cular ducts
     •    Factory calibration factors may be in error if it is necessary
          to use the instrument near a large flow disturbance
5.4  INSTRUMENT ACQUISITION
     A Hastings Flare Gas Flow Probe, a Ramapo Fluid Drag Meter, and six
Thermo Systems hot film sensors were purchased for evaluation.  An order
was placed with Fluidynamics but eventually had to be cancelled due to
delivery date slippage.  Ellison Instruments provided free of charge
an S type pi tot probe and an eventual total of seven Annubars during
the program.  Of these, a 150" Annubar and two 65" Annubars were left
permanently installed at the Nevada Power Company Moapa station at the
conclusion of the second field demonstration.
5.5  ADVANCED INSTRUMENTATION - ACOUSTIC VELOCIMETER

     The intention at the start of the program for this subtask was to
build  (or have built) and test an acoustic velocimeter for the application
                                   -36-

-------
of interest.  This type of device takes a line average across the duct,
which laboratory testing proved to be a good way to determine total  flow.
Other desirable aspects of such systems, which are currently used to
monitor liquid and high pressure gas flows, are relative ease of instal-
lation and removal, and the fact that no system elements are immersed in
the flow.
      Very early in the effort it became apparent that no hardware was
available for applications in a gas at one atmosphere in large ducts.
The main effort was then devoted to obtaining baseline information which
would have to be considered in any system design.
5.5.1  Technical Discussion
      Figure 12 shows the basic arrangement of the acoustic flowmeter of
interest.  Sound beams are transmitted both up- and downstream.  The speed
of propagation  is  the vector sum of the local sound velocity and the flow
velocity.  The  parameter measured is the time of flight of a sound pulse,
or the change in phase of a CW beam.   In steady flow, an upstream and a
downstream measurement traversing identical paths, combine to yield the
flow  speed without an explicit determination of the sound speed.

        It is easy to  show that the  greatest accuracy  results when  the
  sound path and the  duct axis  are 45°  apart.
        The nature of  the gas  flow and  the flue place  a  number  of inherent
  constraints on the  choice of  transducer.   These constraints are discussed
  below.
  5.5.1.1   Losses in  Acoustic  Intensity in the Flue Duct -
           Acoustic Impedance  Mismatch
        A basic constraint is  the poor  efficiency for  acoustic  energy
  transfer between a  solid (transducer  face) and a low pressure gas.   The
  energy transfer for a 1 -dimension normal incidence model  would be:
  where n is the ratio of specific acoustic impedance of the gas and solid
  or (pc) /(pc)  ; p  is density  and c  is  speed of  sound.

                                    -37-

-------
                       \v
                                                                                         T
I
CO
00
I
          r
\

                                                                \
                                                                   \

                     Figure  12.  Acoustic measurement of flow velocity in duct (schematic)

-------
      The transmission coefficient is independent of the direction of
sound propagation, i.e., from gas to solid or from solid to gas.  Values
of pc's are roughly:
           Solids (metals and dielectrics)  * 10* to 107 MKS Rayls
           Liquids (non-metals)             * 106 to 3.106
           Gases (0°C. 100 atmos)           * 101* to 5.101*
           Gases (0°C, 1 atmos)             * 102 to 5.102
Therefore energy transmission coefficients and the resulting sound inten-
sity level dB losses are approximately:
           Solids •*-» liquids:                      10"1 to 1
                                                   (-10 to 0 dB)
           Solids *-»- gases at 0°C and 100 atmos:   10"3 to 2.10"2
                                                   (-30 to -17 dB)
           Solids *-+ gases at 0°C and 1 atmos;     10"5 to 2.10"4
                                                   (-50 to -37 dB)
           Quartz +-+ air at 20°C, 1 atmos:         10"1*
                                                   (-40 dB)

The  above sound  power losses are  incurred both at the  transmitter and at
the  receiver.  Therefore, for the subject gas flue, the total loss to be
expected from acoustic  impedance mismatch is of  the order of 80 dB.  For
100  atmos gases, on the other hand, this loss would be between 35 and
60 dB.  We have  demonstrated a major problem for acoustic flowmeters
when employed in low pressure gas flows.  Innovations  in transducer design
may  reduce these losses but not sufficiently to  alter  the main conclusion.
       Spherical  Divergence Loss  (Figure 12)
       The spherical divergence of the sound beam results in a 6 dB loss
for  each doubling of distance from  the source.
       The source radius can be taken as follows.  If the transducer disc
has  a  radius "a", then  an equivalent source sphere is  defined to be
ira2  = 4ir r 2, or r  = i a.  The  subject application permits large
transducers; assume a   to be 10 cm, and therefore r  = 5 cm.  With a
path length of L = D/cos 45° = 1420 cm, the number of  radial doublings,

                                  -39-

-------
n % 8 is obtained from 2nrQ = L.   The resulting intensity loss is
% 8 x 6 = 48 dB.  By contrast, a  one meter duct would have approximately
5 doublings or a loss of 5 x 6 =  30 dB.
      Molecular Absorption of Sound Energy
      As the sound beam propagates ^ 14  M across the exhaust flue,  it
loses energy to the flue gas by molecular absorption.  The energy  decay
is exponential I(x) = I  exp {-2Ax} where A is the absorption coefficient.
Rewriting we get
          ALT = Ll(x) - Lj  =10 log]0 I(x)/I0 = -8.7 Ax (dB)

For air at 27°C and 37% relative humidity, Kinsler & Frey (Reference 5)
give the following values for (8.7A) dB/M:
                 f(KHz)    8.74 (dB/M)   8.7A (14 M) dB
                   20          0.65            9.1
                   40          1.1            15.5
                   100          4.2            59.0
These values are indicative of losses which should occur in the flue gas.
Evidently, it  is advisable to use frequencies below about 40 KHz.

 5.5.1.2   Summary of  Inherent  Acoustic Intensity Losses
          in  the Subject Flues -
          From  the  above three sources of  intensity  loss, we  get the
 total:
          Acoustic  Impedance  Mismatch               ^ 80  dB
          Spherical Divergence  (20  cm Transducer)   ^ 48  dB
          Molecular Absorption Loss  (f <40 KHz)      * 15  dB
                                    Total          ~ 143  dB

 These  losses do not include  additional  losses  that  could be  produced
 from (1) reflections in a gas flow  with  temperature,  density  and
 therefore, specific acoustic impedance  gradients  as encountered in
 the subject flues, (2)  scattering by turbulent eddies.
                                 -40-

-------
      Molecular absorption of sound energy shows a strong dependence
on frequency, and limits the proposed acoustic flowmeter to 40 KHz
or less.
5.5.1.3  Background Noise and Vibration Levels -
      Background noise and vibration measurements were made by TRW on
the exhaust flues of the SCE Mojave Power Plant,  in Clark County,
Nevada, March 2 and 3, 1973.   The objective was to determine acoustic
instrument operating requirements for the subject application.  The
power plant has two 750 MW units which were operating near capacity at
the time of the measurements.
      Results of the tests are shown in Figures 13 through 16, and
Figures 17 through 19.   From the noise tests,  the noise sound pressure
level  (or intensity level) is quite high  (100  to 70 dB) up to about 20 KHz;
at this frequency, the intensity drops very sharply.  It appears that
the level should be below 50 dB for frequencies above 30 KHz.
      Wall vibration measurements show 3u displacements not exceeding
0.2 cm  for frequencies up to ~ 500 Hz.  Above  500 Hz, displacements are
much  smaller.
       These  tests  demonstrate that the proposed acoustic flowmeter will
have  to operate above about 30 KHz to avoid the high background noise
levels  at  lower frequencies.
  5.5.2   Conclusions
        A supplier  survey has  shown  that  an  off-the-shelf  (with minor
  modifications)  acoustic flowmeter  for  the  subject  applications is  not
  available.
        A feasible  study has  shown  that  two  loss  factors  place large  power
  demands  on  the transducers  for  the  subject application.   One is  the
  extreme acoustic  impedance mismatch  between  (solid) transducer and  a
  one-atmospheric  pressure gas;  the  other is the  spherical  divergence  loss,
  caused by the very  long sound  path  length.
        The frequency range most suitable for  the subject application
  should be between 30-40 KHz  to avoid the high background  noise  levels

                                 -41-

-------
Figure 13.   Stack  flue  interior noise, Mohave power plant unit 1, March 2, 1973,  operating  at
            760  ±5 megawatts.  B&K 4136 microphone.  One-third octave spectrum.

-------
CO
                                              v\ \   v '•
                                              \\\ /> i T~A  ,1

                                              \HAft  Ui—<
                                                                                            JO
                                                                                           to

                                                                                       5oo


       Figure  14.  Stack flue interior noise, Mohave power plant  unit 1, March 2, 1973,  operating at

                  760 ±5 megawatts.  B&K 4136 microphone.  Narrow band spectrum.

-------
I
-fc>
-fa
                                                                                    ±00
         Figure 15.   Stack  flue  interior  noise, Mohave power plant unit 1, March 2, 1973, operating at

                     760  ±5 megawatts.  Statham PL  80TC-0.3-350  pressure  transducer.   Narrow  band  spectrum.

-------
en
i
                                                                                        500
            Figure 16.  Stack flue interior wall vibration, Mohave power plant, March 2,  1973,  operating at
                        760 ±5 megawatts.  B&K 4136 microphone.  One-third octave spectrum

-------
en
i
                                    (00
(ooo
                                                                            (O Ooo     40 ooo
                                                    i
         Figure 17.  Precipitator  inlet flue interior  noise, Mohave power plant unit 1, March 3,  1973,

                    operating  at  775 ±5 megawatts.  B&K 4136 microphone.  One-third octave spectrum

-------
Figure 18.   Precipitator  inlet flue interior noise, Mohave power plant unit 1, March 3,  1973
            operating  at  775 ±5 megawatts.  B&K 4136 microphone.  Narrow band spectrum

-------
00
I
                             J_
          Figure 19.   Precipitator  inlet flue interior wall vibration, Mohave power plant unit 1,
                      March  3,  1973, operating at 775 ±5 megawatts.  Narrow band power spectral  density

-------
in the low ultrasonic and audible ranges,  and to avoid high molecular
absorption losses by humid flue gases above * 40 KHz.
      For an assumed transducer of 20 cm diameter,  the beam major lobe
widths are 3.8° for 30 KHz and 3.0° for 40 KHz; these beam widths appear
suitable for the subject application.
      Pulsed sound waves are likely to produce inaccurate measurements
because very short duration pulses (less than a millisec) are needed for
accurate times of arrival; these short duration pulses will have a power
spectrum bandwidth greater than 1 KHz and therefore, a relatively low
signal-to-noise ratio.
      Continuous single frequency or narrowband, measuring phase dif-
ferences in the 30-40 KHz range, can operate with all  power in a low
noise level region, and may therefore be expected to obtain more accurate
velocity data than with pulsed waves.
      The present feasibility study would indicate the acoustic method
to be marginal at best for the subject application.  A more refined
assessment of transducer performance is needed, however, to show what
improvements in efficiency are available.  In any case, however, this
study has defined the difficulties which have thus far prevented the
development of an acoustic flowmeter by commercial suppliers, for appli-
cations similar to that discussed here.
                                  -49-

-------
                              SECTION VI
                TASK III - MAPPING TECHNIQUE EVALUATION
6.1  GENERAL
     With the exception of the Annubar,  all  instruments chosen  for evalua-
tion were point sensors, and the Annubar had factory calibration factors
only for use in circular ducts.   This made it clear that sensor evaluation
alone would not accomplish program objectives -- it would also  be necessary
to find ways in which to deploy the sensors.  It was therefore  decided to
perform laboratory flow mapping tests in order determine acceptable sensor
placement techniques.  During the course of the program, mapping tests
were performed for many configurations in two different facilities.  It
was arbitrarily decided early in the effort that only techniques which
would require not more than eight velocity sensors would be considered,
since use of a large number of sensors would be expensive and complicated.
     Two kinds of probes were utilized for mapping work:   the  hemispherical
nosed pitot-static probe and the Annubar.  Pitot-static probes  were used
for all point measurements, and the results are considered to apply to
point sensors in general so long as the sensor is small with respect to
the size of the duct.  The Annubar had to be evaluated directly since it
is not a point sensor.  Most of the effort was devoted to flows in
rectangular ducts, which traditionally have not been as well characterized
as the flows in circular ducts.
6.2  DESCRIPTION OF FACILITIES
     Initial mapping tests were performed during 1973.  Test configurations
are shown in Figures 20 - 22.  For the circular test section configurations,
pitot-static probe data were taken along four diameters, as shown in
Figure 21.  For both configurations, the fan inlet was open to  the
atmosphere, resulting in negligible temperature and density stratification
in the flow.  The purpose of the testing was to investigate velocity
stratification.  A sixty-four point traverse was taken in the rectangular
test section, as shown in Figure 23.  Resulting data were then  analyzed
in accordance with the techniques of interest.  For both circular and
rectangular duct data, the reference flow rate was computed from traversal
data in the test plane.  An Annubar was installed during most of the rectang-
ular mapping tests, and was located downstream of the test section, as

                                  -50-

-------
 INTAKE
BLOWER
                         HONEYCOMB FLOW
                         STRAIGHTENERS
                            4.88
                                          SAMPLING
                                          STATION
                                           1.22
        .36  DIAMETER DUCT


     CASE A   PARTIALLY DEVELOPED FLOW
                                                                  .30 x .41  DUCT
BLOWER
                     MOVABLE
                     VANE
                                 t
                                .76
                      3.05
.30  x  .41  DUCT
                                                .20
                           1.52
                                                             .SAMPLING
                                                             STATION
  — .15

•36 DIAMETER    -30 *  -41
DUCT           DUCT
                       CASE  B   FLOW AROUND A CORNER
 All Dimensions  in Meters
    Figure  20.  Circular  duct mapping test configurations,  1973
                                     -51-

-------
TRAVERSING
MECHANISM
                                                              DUCT
                        PITOT-STATIC PROBE

                              TRAVERSE PANEL
                                                                  INSERT
                                                                 DUCT
                                                               REMOVABLE PANEL
                                                               (3)
      Figure  21.   Pi tot traverse installation - circular duct
                                 -52-

-------
                7
    BLOWER
                         HONEYCOMB FLOW
                         STRAIGHTENERS
                                                          SAMPLING
                                                            STATION
   .30 x .41 DUCT


CASE A  PARTIALLY DEVELOPED FLOW
                                                        SAMPLING STATION
                           MOVABLE
                           VANE
                                               .20
                          .30 x .41  DUCT
                            CASE B  TWO STREAM DUCT
                                            .36
                                t
                               .76
                               J_
                                                           SAMPLING
                                                            STATION
               .30 x .41   DUCT

                          CASE  C  FLOW AROUND A CORNER

All  Dimensions in Meters

   Figure  22.   Rectangular  duct  mapping  test  configurations, 1973

                                    -53-

-------
1 | 9
	 J 	 -J
1
1
4

1

i j
\- ' ' 1
i
i
i
!
I

	 	 	 i 	 	 ....
...... 1 	 . . j
8 1
1



	 1 	 1 	 1 	 ' ' !
I'll
L. •• ! l
i 	 f - ~ 1 	 • 	 7~ "- -
1 1 ' 1
1 : 1 !
' ! ' '
i ' | ! 1
	 .. 	 - h 	 ! 	
1 ' l
1 1 i
|_ 4 L L i. - -
r " ; ~ | ^ !
1 ' ' 1
---•J- — -+---J- — -[----
1 ' ' 1
! i 1
! ' j. 1 .
' ' I
1 , i '
1 , i 1

WIDTH
8
.' 57
i— 	


i
. —



— 	 	 —


r
64
















i
\
HEIGH!
8

OBTAIN  MEASUREMENT AT  CENTER OF EACH SUBSECTION

     Figure 23.   Schematic of rectangular duct
                  reference traverse map
                       -54-

-------
shown in Figure 24.  No Annubar was available during the circular
duct mapping tests.
     A new mapping facility was constructed specifically for this
program during 1974.  The configurations used are shown in Figure 25,
and probe locations for 1974 testing are shown in Figures 26 - 29.  Probe
locations for testing during 1975 are shown in Figure 30.  The new
facility was built for several reasons.   Having the fan at the duct
outlet rather than at the inlet allowed greater freedom of variation
for inlet conditions.  The straight, circular reference section down-
stream of a honeycomb flow straightener was used to provide a total
flow measurement in an area free from significant flow angularities.
It also eliminated the need for a full traverse of the test section
for each run, so that the number of runs could be maximized.
6.3  MAPPING TECHNIQUES
6.3.1  Circular Duct Techniques
     Many empirical techniques are available for determining volumetric
flow in circular ducts through a number of point measurements.  Perhaps
the most traditional method is the centroid of equal areas technique,
as given in EPA Method 1 (Reference 1).   This is the best technique to
use if nothing is known about the flow profile shape.  It has long been
recognized that flow in a straight circular pipe eventually assumes a
parabolic type profile as a steady state condition.  This is called
fully developed flow.  This characteristic of pipe flow was used by
Winternitz and Fischl in developing the Log-Linear Mapping Technique
(Reference 6).  It is based on the assumption that pipe flow profiles
can be generally characterized by the equation
          v(r)  =  A + B log                                         (12)
where
             v  =  magnitude of velocity, m/sec
                                  -55-

-------
                                                  ANNUBAR
i
in
             FLOW

           DIRECTION
                             PLANE OF

                          PITOT TRAVERSE



                     Dimensions in Meters
                                                                                                     -O—
                            Figure  24.   Annubar  location during  duct mapping  test, 1973

-------
I
tn
           CONFIGURATION
            IN ORDER OF
              TEST
                                                    MEASUREMENT
                                                       PLANE
                                                                             .69



H
MEASUREMENT
PLANE

FAN
H
                                                                           .41
                                                                    .91
                                                                                   FLOW
                                                                                   DIRECTION
                                                                                              c>
                                                                                                 Dimensions in Meters
                                 INLETS           TEST SECTIONS
                                 H DENOTES  LOCATION OF HONEYCOMB PANEL
                                                                              REFERENCE SECTIONS
                            Figure  25.  Schematic  -- top  view of mapping  test facility, 1974

-------
                       VERTICAL
01
ca
i
                                     ANNUBAR
— * 6 1n to

d "7
00 . -M.
PROBE
1

j







PROBE
2

.- 20


^
                                               HONEYCOMB
ANNUBAR
HONEYCOMB
                                                                                                   Dimensions in Meters
                                  Figure 26.   Position and  orientation of reference  probes

-------
en

-------
Dimensions in Meters

'

.91
1
1

.4
i



6


;
^

^
l
-o-
+
PROBE 5


	 0 	 0 0

                         ANNUBAR
                                                                     l.i
                                                 .04
                                                                        -is
()
                                                                    ANNUBAR
                      Figure 28.   Rectangular section probe locations

-------
                AIIHUBAR
                LOCATION
            T1
             • 4)
       ANNUBAR
       LOCATION
          PITOT PROBE
          LOCATION
                                          .90
                                     1.09
                                                 .61
                                                          -.41
                          A.  PROBE LOCATIONS AFTER RECTANGULAR ELBOW
PITOT PROBE
LOCATION
                 i      i
                                      .7] DIA
                                 .97
                                1.22
                                      .76
                                                DUCT  HEIGHT
                                                .91
                                                                     Dimensions  in Meters
                             B.  PROBE LOCATIONS AFTER CIRCULAR ELBOW
Figure 29.   Probe  locations  relative  to elbows  in  1974 mapping tests

                                      -61-

-------
PITOT TRAVERSE /
LOCATIONS
t
W
I

Location
A
B
C
D
E
F
G
H
I
\ B C

I





E







G



Distance From Elbow
Pi tot
Traverse
2.20W
1.36W
.92W
.80M
.61W
.42W
.23W
--
--
Annubar
2.91W
2.06W
1.62W
1.50W
1.31W
1.12W
.94W
.75W
.56W


^
Flow
Direction



W = .406 M




                        D    F   H
Annubar
Locations
t
1


A




B C














E




G












:

^


•
Flow
Direction
Figure 30.   Probe locations --  1975 testing
                  -62-

-------
             A, B  = constants, m/sec
                y .= distance from pipe wall toward center, m
                d  = pipe diameter, m
As the authors state, "The basis for the proposed integration rules was
the selection of metering position in each annulus such that the exact
mean velocity for a particular cross-section would occur at the selected
gauging points, provided the chosen logarithmic functions are an adequate
representation of the considered pipe-flow velocity profiles."
     One of the conclusions in Reference 6 is that for a given number of
measurement points, the Log-Linear technique gives more accurate results
than the centroid of equal areas technique (also known as the tangential
method).  This conclusion was accepted as valid and it was decided to use
the Log-Linear technique for reference measurements during the test,
as well as evaluating it for general use with a minimum number of points.
All methods tested are shown in Table 2.   Each method involved measurement
of velocity at specific points along two orthogonal diameters.
     The Annubar, as described in Section 5, was designed around a four
point per diameter mapping technique.  Its used in circular ducts is
straightforward:  it is installed according to factory specifications
and a factory calibration factor is used to correct the output.  The
computed flow for each run was then compared to the reference flow measure-
ment to determine instrument accuracy.
6.3.2  Rectangular Duct Techniques
     A literature search revealed no rectangular duct mapping techniques
involving eight or less sampling points.  Consequently, new methods had to
be developed.  In order to do so, properties of flows in rectangular
ducts were considered.  Some of the words done in this area by Nfkuradse
is documented in Chapter XX of Reference 4, where it is observed that
secondary flows occur "in all straight pipes of non-circular cross-section."
In a rectangular duct, the secondary flows tend to generate high velocities
at the corners.  This tendency means that flow angularities with respect
to the duct axis should always be expected regardless of the duct length.
                                   -63-

-------
            Table 2.  METHOD FOR DETERMINING VOLUMETRIC  FLOW FROM POINT MEASUREMENTS IN A CIRCULAR DUCT
RULE
Gauss-
Sherwood
Gauss-
Sherwood
Aichelen
New Empirical
Log-Linear
Log-Linear
Log-Linear
Log-Linear
i
NUMBER OF POINTS
PER DIAMETER
3
Weights*
4
Weights*
2
3
Weights*
2
2 + 4
4
10
DISTANCE FROM WALL
IN PIPE DIAMETERS
.060 .500 .940
585
.036 .208 .792 .964
.1739 .3261 .3361 .1739
.119 .881
.081 .500 .919
2 1 2
.112 .888
.112 .888 (One Dia.) Q .
.043 .290 .710 .957 ( „<)
U i a
.043 .290 .710 .957
.019 .076 .153 .217 .361
.639 .783 .847 .924 .981
TOTAL NUMBER
OF POINTS
(ON TWO ORTHOGONAL DIAMETERS)
5
8
4
5
4
6
8
20
I
cr>
             ^Multiply measured value by weight,  sum, divide by  sum of weights.

-------
     Another practical problem associated with flow measurements in
rectangular ducts in process streams is that the ducts are normally very
short.  In a power plant, for example, it would not be expected to find
any runs of rectangular ducting comparable in length to that of the
circular exhaust stack.  Circular duct mapping techniques are derived
from knowledge of fully developed flow characteristics, and then applied
to non-fully developed flows.  There has been no directly workable
analogue of this for the rectangular case.
     Common circular duct mapping techniques may be termed "universal"
in that they are designed to apply in any circular duct without a need
for in-place calibration.  Clearly it would be desirable to have such
techniques available for continuous monitoring in rectangular ducts.
Continuing this process, it is possible to define various types of
techniques in order of acceptability as follows:
     1.  Universal - sensors can be installed at pre-selected locations
         to give accurate total flow measurement without in-place calibra-
         tion or prior knowledge of flow conditions.
     2.  Single calibration - installed system needs to be calibrated
         in place at one flow condition only to determine a constant
         calibration factor.
     3.  Full calibration - installed system needs to be calibrated
         in place over the full range of flow conditions to produce
         a curve of calibration factor vs. flow rate.
Techniques 2 and 3 may also require some prior knowledge of the flow  to
determine sensor placement.  The goal of producing a Universal  technique
was not fully realized, although results indicate that methods  developed
are definitely in the Single Calibration category and may be universal
immediately downstream of an elbow.  A broader data base is needed to
verify the latter.
     The first techniques attempted were the four and five point methods
shown in Figure 31.   The objective of the investigation was to  determine
if a single location on the diagonals for each point would result in
a correct computation of the total flow rate.  If these points  existed,

                                   -65-

-------
en

-------
then the methods would be directly analogous to circular duct mapping
techniques.  As is shown below, these proposed methods failed.
     Existing 1973 test data were then re-analyzed to determine other
possible techniques, and the Row Average Method was discovered.  This
method involves determining the average velocity along a line spanning
the duct and assuming that it is the average velocity for the entire
duct.  Methodology involves determining which two walls to take the
line average between, and the proper location of the line.  This is
illustrated for a fictitious case in Figure 32.  By definition, rows
are taken along lines of maximum velocity stratification.  Proper location
with respect to the side walls was determined during testing by comparing
local row averages to the total flow average.  This method proved to be
successful, especially in the wake of an elbow, and came to be the
recommended method for use with point sensors.
     There was initially no expectation that the Annubar would work well
in rectangular ducts, since its hole pattern was based on a circular duct
technique.  An Annubar was installed along the center row for most of
the 1973 rectangular duct testing, and the results showed a surprisingly
repeatable calibration factor.  As a result, an Annubar was used during all
subsequent rectangular duct mapping tests.  It eventually became clear
that the Annubar and Row Average methods are really very similar, since
they both involve line averaging.  This led to the further realization
that both techniques work best directly downstream of an elbow.
6.4  MAPPING TEST RESULTS
6.4.1  Circular Duct Techniques
6.4.1.1  1973 Testing -
     A total of eight runs was performed for the two configurations shown
in Figure 20.  Typical velocity profiles from each configuration are
shown in Figure 33 from which it is clear that the flows were undeveloped,
i.e. non-axisymmetric and not parabolic.  For each run, data were taken
along two sets of orthogonal diameters (Figure 21) thus giving two sets
of data for each method per run, or a total of sixteen sets of data for
each method.

                                   -67-

-------
     Tabular Entries  Are  Velocity At  Traverse  Points
     	*-Direction Of  High Gradients
Direction
   Of
  Low
Gradients
2
1
0
0
0
1
4
3
4
0
3
3
6
5
4
3
5
7
8
7
6
6
7
9
10
9
9
9
10
10
12
11
10
9
11
12
Row Average
  7.00
  6.00
  5.50
  4.50
  6.00
  7.00
                                     Overall Average =6.00
Rows are taken in the direction of highest velocity gradients.
Row averages for rows 2 and 5 agree with overall average, so
either of these rows would be recommended for sensor placement
                 Figure  32.   Row  average method
                             -68-

-------
                 DISTANCE FROM WALL

                    DUCT DIAMETER
10
i
        1.00
          .50
    AIR SPEED,  METERS/SEC


 25              30 '


\	\
                                                                                       PARTIALLY

                                                                                       DEVELOPED FLOW
                              Figure 33.   Typical  flow profiles from circular duct mapping test

-------
     Results are shown in Table 3.  For each run the average velocity
was determined using the Log-Linear 10 method along all four diameters
for a total of forty points.  Each method was then evaluated with respect
to the reference velocity.  Since two sets of orthogonal diameters were
used, each method produced two sets of data per run except for the
Log-Linear 2+4 method, which had four possible combinations.  The average
error listed is the average of the absolute errors.
     The results show that the Aichelen and Gauss-Sherwood 3 methods were
consistently less accurate than the others.  The results for the other
five methods are excellent in terms of program requirements—worst case
errors were always under 5 percent, and average errors always under
2 percent.
     As mentioned above, the reason for this part of the testing was to
confirm the adequacy of previously documented methods for circular duct
flow determination.  The results in Table 3 show that five of these
methods indeed gave very good results in nondeveloped flows.
6.4.1.2  1974 Testing -
     As shown in Figure 25, two basic configurations were used -- a
straight inlet and a tiitered elbow inlet.   Variation of the inlet flow
was accomplished by blocking off various sections of the inlet with
4" x 8" aluminum plates.  Three measurement techniques were evaluated --
the Log-Linear 4 method (8 points) and the New Empirical method (5 points),
for which data were obtained by two pitot static probes, and the Annubar.
Reference data were provided by two pitot static probes in the reference
section, utilizing the Log-Linear 4 method, and another Annubar.  The
Log-Linear 4 data in the reference section was taken to be the reference
flow data as a result of the 1973 tests.  Accuracy of this reference
measurement was estimated to be not worse than +2%.
     Inlet blockage configurations for the test are shown in Figure 34.
These were chosen at random to create a variety of flow patterns at the
test section.   For each blockage condition, data were taken at four different
flow rates.  A total of 24 runs was performed for the elbow case, and 32
runs were performed for the straight inlet case.   Results for the two
configurations are given in Tables 4 and 5 and a summary is given in Table 6.

                                  -70-

-------
                 Table 3.  CIRCULAR DUCT MAPPING  TEST  RESULTS,  1973
CONFIG. .
A
B
RUN
NO.
1
2
3
1
2
3
4
5
AVERAGE ERROR
WORST CASE ERROR
AVFRflRF
VELOCITY
M/SEC
21.75
17.53
8.11
23.03
17.68
18.87
19.46
14.44


ACCURACY OF METHOD, %
L.L. 2
(4)*
- .5
+ .4
0
-1.4
-2.3
- .8
+ .1
-4.3
-3.3
-2.4
-3.6
-4.0
-1.0
+ 1.9
-1.7
-1.7
1.8
4.3
I
L.L. 4
(8)
+ .4
+ .1
0
- .5
+ .8
- .8
+3.1
+ .1
+ .8
- .3
+ 1.8
+ .5
+ 1.0
+4.6
+ 1.7
+1.9
1.2
4.6
L.L. 2+4
(6)
+ .1, + .6
- .4, + .2
+ .2, - .2
- .7, -1.2
- .8, - .8
-1.1, 0
+1.1, -2.7
+2.1, -1.5
+ .4, -2.3
-2.9, - .4
+ .2, -2.4
-2.0, -1.2
+1.0, +2.4
-1.0, +4.1
+1.2, - .2
1.1, + .4
1.2
4.1
A.
(4)
+ .2
+1.2
- .5
- .7
- .8
+ .4
-3.1
+4.4
-3.2
+7.3
-3.5
+5.7
+2.3
+5.5
-2.7
+7.5
3.1
7.5
N.E.
(5)
- .5
- .1
- .2
+ .2
-1.1
- .8
+ .9
-2.2
+1.0
- .8
+1.5
-1.1
- .7
+2.4
+1.8
- .4
1.0
2.4
G.S. 3
(5)
+ 1.5
+1.8
+ 1.9
+1.9
+ 1.1
+ 1.5
+1.9
+3.7
- .1
+4.7
+2.8
+3.9
+6.3
+1.7
+1.6
+5.4
2.6
5.4
G.S. 4
(8)
+ .1
+ .6
+ .2
- .2
+ .8
0
-1.2
-2.7
-1.2
-1.7
- .6
-4.0
+ .7
+3.4
- .6
-2.5
1.3
4.0
Total  number of data points required

-------
                 Shaded portion of inlet blocked  off  for  runs  shown.
                 Each set of four runs was  performed  at  four different
                 flow rates.
            Runs  89-92,
            113-116,  117-120
                     93-96,  121-124
97-100, 125-128
101-104, 129-132
                                                              105-108,  133-136
            109-112, 137-140
                         141-144
                       Figure  34.  Circular inlet blockage - 1974

                                        -72-

-------
    Table 4.   1974 CIRCULAR DUCT
   Conf. 4:   Runs 89-112
    POINT SENSOR MAPPING RESULTS
         Config. 3:  113-114
Run
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
VREF
rn3/sec
6.2
4.8
4.3
3.1
4.9
4.3
4.0
3.1
5.3
4.4
4.1
3.2
4.6
4.3
3.8
3.1
5.0
4.4
4.0
3.2
5.2
4.4
4.2
3.2
6.8
5.0
4.6
3.3
L.L.
1.020
.995
1.043
1.032
1.052
1.035
1.005
1.084
.954
.994
1.051
1.127
1.034
1.045
1.103
1.065
1.004
.979
1.043
1.037
1.031
1.065
1.105
.960
1.033
1.026
1.011
1.043
N.E.
1.097
1.096
1.140
1.169
1.076
1.086
1.164
1.127
1.016
1.072
1.083
1.108
1.144
1.067
1.213
1.071
1.082
1.149
1.060
1.121
1.141
1.125
1.120
h072
.953
.950
.977
.961
Run
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
126
127
133
139
140
141
142
143
144
VREF
3,
ra /sec
6.7
5.1
4.6
3.4
5.7
4.6
4.3
3.3
5.6
4.6
4.2
3.2
5.2
4.4
4.1
3.0
5.5
4.7
4.6
3.2
5.6
4.7
4.2
3.2
6.1
4.8
4.3
3.2
L.L.
1.022
.978
.992
1.005
1.029
1.075
1.026
1.026
1.052
1.043
1.028
.972
1.032
1.051
1.046
1.058
1.012
1.000
1.031
.984
1.039
1.024
1.059
1.048
1.064
1.025
1.077
1.031
N.E.
.973
.950
.980
1.021
.981
1.099
1.095
1.022
1.001
1.066
1.018
.985
1.061
1.072
1.061
1.093
.810
.898
.868
.809
.982
.990
1.009
1.006
1.029
.950
1.024
1.023
I  I   -  Log-Linear 4

           VREF
N.E. =
_  New Empirical

       VREF

-------
       Table 5.  ANNUBAR CALIBRATION FACTORS FOR 1974
                 CALIBRATION TESTS - CONFIGURATIONS 3 AND 4
Conftg. 4:  Runs 89-112
                   Conftg. 3;  Runs 113-144
Run
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
V
m /sec
6.2
4.8
4.3
3.1
4.9
4.3
4.0
3.1
5.3
4.4
4.1
3.2
4.6
4.3
3.8
3.1
5.0
4.4
4.0
3.2
5.2
4.4
4.2
3.2
6.8
5.0
4.6
3.3
%
REF
.738
.691
.692
.665
.705
.673
.684
.658
.742
.672
.693
.658
.700
.723
.694
.729
.720
.719
.681
.702
.732
.772
.712
.703
.684
.683
.688
.675
*r
TEST
.711
.669
.670
.624
.687
.636
.689
.640
.695
.668
.681
.644
.704
.734
.718
.705
.650
.729
.680
.677
.733
.757
.734
.692
.663
.663
.667
.666
Run
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
126
127
133
139
140
141
142
143
144
V
ro' /sec
6.7
5.1
4.6
3.4
5.7
4.6
4.3
3.3
5.6
4.6
4.2
3.2
5.2
4.9
4.1
3.0
5.5
4.7
4.6
3.2
5.6
4.7
4.2
3.2
6.1
4.8
4.3
3.2
%
REF
.667
.695
.689
.709
.696
.698
.694
.694
.687
.680
.673
.680
.714
.690
.698
.672
.746
.713
.689
.670
.727
.745
.726
.729
.677
.672
.672
.683
TEST
.662
.679
.685
.698
.689
.695
.699
.702
.671
.683
.669
.675
.716
.687
.689
.659
.743
.695
.679
.652
.733
.734
.735
.722
.662
.659
.645
.644
   - -698
o$ =
S  = . 687
+_ .025 (+3. 7%)
+ .030 (+4.4%)
Factory Value:

   S = .667
                              -74-

-------
Table 6.   SUMMARY OF 1974 CIRCULAR DUCT MAPPING TEST RESULTS
Method
Annubar(Test
Section)
Annubar
(Reference
Section)
Log-Linear 4
New Empirical

Average Error, Percent, and Standard
Deviation, Percent
Conf. 3
.2.6 +_ 3.9

+2.5 +_ 3.1
+2.9 +_ 2.5
-0.9 + 7.1
+1.2 + 4.5*
Conf. 4
-3.3 + 5.0

+5.3 +_ 4.0
+3.6 + 4.1
+10.8 + 3.8

Total
-3.0 +_4.4

+4.6 + 3.7
+3.2 + 3.3
+4.1 +_8.0
+5.6 +6.2*
  Excludes runs 133-136
                              .75-

-------
Annubar  results are  presented  in  terms of  the  standard calibration
factor,  S,  as  used by  the  factory:

          V S ^s
            s      bA

      where
          V = true  volumetric  flow  at standard  conditions
          S = Annubar calibration factor

          S,, = uncorrected volumetric flow at  standard conditions
               calculated  from  Annubar output.
For  the  runs performed,  the average  S value  from the  factory was S =.667.

      Several important features are shown in the summary.  The first is
that the new Empirical method did not provide good results for the elbow
case and for one configuration  in the straight inlet case.  In both
instances,  a systematic shift was observed.  For the elbow case, the
average for the New Empirical method was  10.8% above the reference value,
and for runs 133-136 for the straight inlet case, the average was about
15% low.  For the other 28 straight inlet  runs,  the average was 1.2% high,
which is very acceptable.  It is significant to note that for the elbow
case, the shift was very systematic -- the standard deviation for those
runs was comparable to that for the other methods.
      Annubar results and the Log-Linear  4 results are all comparable,
and the Annubar results are reasonably close to the factory predictions.
It is noteworthy that the Log-Linear results in the test section were
systematically high,  and that the Annubar output was systematically low,
which resulted in a higher than predicted calibration coefficient.   It is
believed that the shift for the Log-Linear results was due to flow
angularity  in the test section.  As has  been mentioned,  the desired
parameter is the axial velocity component.   In non-aligned flows, as will
be shown more completely in Section 7, a  hemispherical pitot-static
probe tends to indicate too high an axial component.  On that basis, it
would be expected that a particular method, such as the  Log-Linear, would

-------
 indicate a  higher velocity in a region of significant flow angularity, such
 as  the  test section, than in a more one-dimensional area, such as the
 reference section.  This angular response problem was, as previously
 mentioned,  one of the reasons for constructing the reference section.
      It is  believed very likely that the Annubar results varied from the
 factory calibration due to the presence of the rods used to support the
 duct  itself and the pitot probes.  Two half inch diameter steel rods were
 placed  upstream of each Annubar to stiffen the duct and hold it in a cir-
 cular shape.  In addition, the pitot-static probes slid along these rods.
 The presence of the rods inevitably created a low pressure (and hence
 low velocity) region along the axis of the duct, which is where the
 Annubar rear static port was located.  This localized low pressure area
 would result in lower than normal Annubar output, which in turn resulted
 in a  higher than predicted calibration factor.  This explanation is given
 added credence by the fact that the reference Annubar calibration factor
 was higher  than that for the one in the test section, since the flow was
 better  aligned axially in the reference section so that the reference
 Annubar static port would lie more directly in the wake of the support rods
 than would  the static port of the test section Annubar.
      In summary, it is felt that the Log Linear results would have been
 better  if an ellipsoidal nosed pitot probe had been used (further explana-
 tion  is given in Section 7) and the Annubar results would have been better
 if the  instruments had not been directly in the wake of the support rods.
 Even with these discrepancies, however, data for these methods is quite
 acceptable.   Results for both Annubars and the Log Linear 4 method averaged
 better than 5% accuracy and less than 5% standard deviation, which is well
 within program objectives.   The New Empirical  method was accurate only
 for some configurations, however, although errors were very systematic
 and could be accounted for in actual  conditions by an in-place calibration.
     The tabulated data show one practical consideration very clearly,
 and that is  that the Annubar, from which one measurement is required,
 gave results comparable in  accuracy and consistency to the Log Linear
method,  which required eight individual  measurements, and gave better
                                  -77-

-------
results than the New Empirical method, which required five individual
measurements.  The cost difference between one channel  of instrumentation
for the Annubar and eight for a fixed Log Linear array of point sensors is
obvious.  From an analytical standpoint, test data reveal no significant
difference between the Log-Linear method and the Annubar.  From a hardware
cost and system complexity standpoint, the Annubar is economically
more desirable.
6.4.2  Rectangular Duct Techniques
6.4.2.1  1973 Testing -
     A total of forty runs was performed in the configurations shown in
Figure 22.  The breakdown was as follows:
          Case A - Partially developed flow:   5 runs
          Case B - Two stream flow:           19 runs
          Case C - Flow around a corner:      16 runs
An Annubar was installed for Cases B and C but was not available for
Case A.  In Case A, only flow rate was varied.  In Cases B and C, both
flow rate and  the position of the moveable vane shown in Figure 22
were varied.
     The first techniques examined were  the four and five point techniques
discussed in Section  6.3.  The analysis  proceeded as follows:
     1.  Compute average velocity by averaging 64 data points.
     2.  Mathematically curve fit the data points along  the diagonals.
     3.  Specify a desired  accuracy  (e.g., 1%, 5%).
     4.  Begin at the center  of  the  duct and  proceed in  1 percent length
         increments along each half  diagonal, computing  a four or five
         point average  (using the center velocity as constant  in the
         five  point method) at each  increment until  the  computed velocity
         agrees with  the reference velocity within the specified accuracy.
     5.  Repeat the procedure starting  at  the corners and moving toward
          the middle.
                                    -78-

-------
This analysis was used to determine the tolerance band for the accuracy
specified, so long as the velocity is only increasing or only decreasing
within the band, which would generally be the case.   Results for a typical
run are shown in Figure 35 .  The results  at that  point seemed
highly encouraging.  They showed a wide bandwidth -- 5 to 10 percent or
more, for the location of the sampling points to obtain a mapping accuracy
of 2 to 5 percent, which is within acceptable limits.  The wide bandwidth
was encouraging since it tended to show that a wide variety of flow con-
ditions could have overlapping bandwidths.  The overlaps can be tabulated
to determine an optimum sensor location to give acceptable accuracy over
the full range of flow conditions.  This in turn means that the four
and/or five point methods would in fact be universal within certain
accuracy limitations.
     More complete data analysis quickly dampened the initial enthusiasm
for the four and five point methods.  Overall results for the three test
configurations are shown in Table  7  .  The five point method worked quite
well for the partially developed flows and for the two stream flows with
vane settings 1 and 2.  For these 12 runs, sensors in the region  .75 to
.79 would give accuracies of at least +5%.  Results for the four point
method were not so good -- there was no region which was acceptable
for all 12 runs.  For case C, flow around a corner, the region .42 to
.50 was acceptable for 5% accuracy using the five point method for 13
of the 16 runs, while the four point method was acceptable in the region
.35 to .40 for 13 of the 16 runs.  The disappointment here is that while
results were very consistent for this case, the regions were not close
to those for cases A and B.  This showed rather conclusively that neither
method will be as universal in scope as the corresponding circular duct
methods such  as  the  Log-Linear  technique.

     Failure  of the four and five point methods led  to a search  for other
usable techniques, and the Row  Average Method was eventually discovered.
The method was described in  Section 6.3.2.   For each of the forty completed
runs, row averages were determined from the 8x8 reference matrix, and
compared to the total  flow average.  The row  orientation can usually be
                                   -79-

-------
          DATE:  3-13-73

          RUN NO:  2

          DUCT CONFIGURATION:  Partially Developed, 1/4 Flow

          AVERAGE VELOCITY:    8.56 m/sec
SPECIFIED
ACCURACY
.5
2
5
FRACTIONAL TOLERANCE BAND*
.68 -
.66 -
.62 -
t
.68
.70
.73
r.
.74 -
.71 -
.67 -
>
.74
.76
.80
         *
          r given in fraction of half diagonal
Figure 35.  Results of four and five point analysis for a partially
            developed flow run
                                -80-

-------
                                  Table 7.   RESULTS OF  RECTANGULAR  DUCT ANALYSIS
                                            BY  FOUR AND FIVE  POINT  METHODS
            ENTRIES  INDICATE  PROBE PLACEMENT  REGIONS ALONG  HALF  DIAGONAL  LENGTH  IN PERCENT TO ACHIEVE


            VELOCITY MEASUREMENT ACCURACY AT  TOP OF COLUMN.   SEE  FIGURE 2




            CASE  A.   PARTIALLY  DEVELOPED FLOW

INLET '
1
3/4
1/2
1/4
1/8
METHOD:
VELOCITY
ACCURACY:





NET ACCEPTABLE
REGION
5 POINT
0.5%
70 - 71
70 - 71
77 - 77
X
73 - 73
X
2%
67 - 74
66 - 74
75 - 78
71 - 76
70 - 75
X
5%
59 - 80
59 - 80
71 - 81
67 - 80
66 - 79
71 - 79
4 POINT
0.5%
63 - 64
64 - 65
71 - 71
68 - 68
68 - 68
X
2%
60 - 67
61 - 67
69 - 73
66 - 70
66 - 70
X
5%
54 - 73
54 - 73
66 - 75
62 - 73
62 - 73
66 - 73
I
oo

-------
                            Table 7  (continued).
RESULTS OF RECTANGULAR DUCT ANALYSIS

BY FOUR AND FIVE POINT METHODS
              CASE  B.   TWO  STREAM  FLOW

INLET
1
1
1
1/16
1
3/4
1/2

VANE
1
1
1
1
2
2
2
METHOD :
VELOCITY
ACCURACY :







NET ACCEPTABLE
REGION
1
3/4
5/8
1/2
1/8
1
3/4
1/2
1
3/4
1/2
V2
2.5
2.5
2.5
2.5
2.5
3
3
3
4
4
4
5












NET ACCEPTABLE

REGION
5 POINT
0.5%
76 - 76
76 - 77
74 - 75
73 - 74
77 - 77
80 - 81
82 - 83
X
45 - 45
49 - 49
X
X
X
86 - 86
86 - 86
89 - 89
49 - 49
53 - 53
51 - 51
X



2%
73 - 80
74 - 79
72 - 77
71 - 76
75 - 78
79 - 82
79 - 85
X
44 - 46
48 - 50
43 - 43
48 - 49
50 - 51
85 - 87
86 - 87
88 - 90
49 - 50
52 - 54
50 - 52
61 - 62



5%
67 - 82
68 - 82
65 - 81
65 - 80
71 - 81
75 - 85
71 - 89
75 - 80
43 - 48
47 - 52
42 - 43
47 - 50
49 - 52
83 - 89
83 - 89
85 - 92
47 - 52
51 - 56
49 - 53
59 - 64



4 POINT
0.5%
78 - 78
79 - 79
78 - 78
76 - 76
80 - 80
84 - 84
87 - 88
X
X
X
37 - 37
41 - 41
X
31 - 31
X
95 - 95
X
43 - 43
41 - 41
52 - 52



2%
77 -880
77 - 81
76 - 79
74 - 78
79 - 81
83 - 85
86 - 89
X
36 - 37
39 - 39
37 - 37
41 - 41
43 - 43
31 - 31
32 - 33
94 - 95
39 - 39
43 - 43
41 - 41
52 - 52



5%
73 - 83
74 - 83
72 - 82
70 - 81
76 - 83
81 - 87
81 - 92
81 - 81
35 - 38
38 - 40
36 - 38
40 - 42
43 - 44
29 - 32
31 - 34
93 - 97
38 - 41
42 - 44
40 - 42
51 - 54



co
ro
i

-------
                             Table  7  (continued)
               CASE  C.   FLOW AROUND A CORNER
RESULTS OF RECTANGULAR DUCT ANALYSIS

BY FOUR AND FIVE POINT METHODS
T Ml r~T
INLET
1
3/4
5/8
1/2
1
3/4
1/2
A I ft
3/4
1/2
1
3/4
1/2
1
3/4
1/2
\M n it i T~
VANE
0
0
0
0
1R
1R
1R
1L
1L
1L
2R
2R
2R
2L
2L
2L
METHOD:
VELOCITY
ACCURACY :
















NET ACCEPTABLE
REGION
5 POINT
0.5%
43 - 44
44 - 45
44 - 44
53 - 55
45 - 45
42 - 42
46 - 46
41 - 42
35 - 35
49 - 50
43 - 44
42 - 43
28 - 28
52 - 53
29 - 32
51 - 53

X
2%
40 - 46
42 - 47
42 - 46
48 - 61
42 - 48
40 - 45
43 - 48
38 - 45
33 - 36
45 - 53
40 - 46
39 - 46
26 - 30
49 - 55
25 - 36
45 - 56

X
5%
35 - 53
37 - 53
39 - 51
42 - 74
37 - 54
35 - 51
39 - 53
33 - 53
30 - 40
40 - 62
35 - 52
34 - 53
25 - 33
41 - 60
25 - 52
25 - 62

42 - 51*
4 POINT
0.5%
36 - 36
37 - 37
38 - 38
38 - 38
37 - 38
35 - 35
39 - 39
34 - 34
31 - 31
40 - 40
36 - 37
35 - 35
X
51 - 52
30 - 32
52 - 53

X
2%
34 - 38
35 - 39
36 - 40
36 - 40
36 - 40
33 - 36
37 - 41
32 - 36
29 - 32
38 - 43
35 - 39
33 - 37
31 - 34
49 - 54
25 - 35
48 - 56

X
5%
30 - 42
35 - 42
35 - 43
35 - 45
35 - 44
30 - 40
34 - 44
29 - 41
27 - 34
33 - 48
35 - 43
30 - 41
28 - 36
43 - 58
25 - 40
25 - 61

35 - 40*
t
co
CO
I
          *0mits 3 runs

-------
determined by inspection of the duct work.   For example,  for Case C
the rows were always being taken parallel  to the plane of the elbow
rather than normal to the plane of the elbow.   Results are tabulated
in Table 8.  Overall results were very encouraging,  but those for
Case C, the elbow, were especially good.   These results were further
analyzed, and results are given in Table  9.   For these 16 runs,  a very
consistent pattern of higher row averages  toward the walls and lower
row averages toward the center was observed, and the total row average
variation was only about 14%.  This shows  that for the specific  con-
figuration used, a row average taken practically anywhere across the
duct would agree fairly closely with the  overall average  velocity.
     As mentioned above, an Annubar was installed for Cases B and C
as shown in Figure 24.  Since it was horizontal, as  was the elbow in
Case C, it was in effect installed along  a  row.  A reading from the Annubar
was taken at the conclusion of each pi tot traverse for a  total of 35
runs.  The data reduction scheme was as follows:
               V =  average velocity

               R =  gas  constant

              T  =  static temperature

              p^ =  static pressure

                 =  pitot probe differential  pressure
                    12  R T
              V.  = velocity from Annubar  reading

               k  = calibration  factor

                 = Annubar differential pressure



                                 -84-

-------
        Table 8.  RESULTS OF RECTANGULAR DUCT ANALYSIS BY ROWS
                     X INDICATES .95
                                        row
1.05
CASE A.   PARTIALLY DEVELOPED FLOW
INLET

1
3/4
1/2
1/4
V8
AVERAGE (%)
ROW NUMBER
1


X
X
X
60
2
X
X
X


60
3
X
X

X
X
80
4
X

X
X
X
80
5


X
X
X
60
6


X
X
X
60
7
X
X
X
X
X
100
8

X


X
40
                                   -85-

-------
 Table 8 (continued).  RESULTS OF RECTANGULAR DUCT ANALYSIS BY ROWS
CASE B.   TWO STREAM FLOW
INLET

1
.1
1
1/16
1
3/4
1/2
1
3/4
5/8
1/2
1/8
1
3/4
1/2
1
3/4
1/2
1/2
VANE
POSITION

1
1
1
1
2
2
2
2.5
2.5
2.5
2.5
2.5
3
3
3
4
4
4
5
AVERAGE (%)
ROW NUMBER
1
X
X
X
X















21
2
X
X
X
X


X
X
X
X
X
X
X
X

X

X
X
79
3



X
X
X













16
4










X





X
X

16
5









X
X
X




X
X

26
6






X
X

X
X
X





X

32
7




X
X
X
X
X
X
X
X
X
X
X
X

X

68
8



X















5
                                 -86-

-------
Table 8 (continued).  RESULTS OF RECTANGULAR DUCT ANALYSIS BY ROWS
CASE C.   FLOW AROUND A CORNER
INLET

1
3/4
5/8
1/2
1
3/4
1/2
1
3/4
1/2
1
3/4
1/2
1
3/4
1/2
VANE
POSITION

0
0
0
0
1R
1R
1R
1L
1L
1L
2R
2R
2R
2L
2L
2L
AVERAGE (%)
OVERALL
AVERAGE (%)
ROW NUMBER
1
X












X


12
22
2
X
X


X
X
X
X
X
X
X
X


X

69
72
3
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
100
58
4
X

X
X
X
X

X

X
X




X
56
40
5
X
X

X
X
X
X
X

X
X





56
42
6
X
X
X
X
X
X
X
X
X
X
X
X
X
X

. X
94
60
7
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
100
85
8


X









X
X
X

25
18
                               -87-

-------
        Table 9.   EXAMINATION  OF  1973  ROW  AVERAGE  DATA AFTER
                  A RECTANGULAR ELBOW
Date/Run

4-5 1
4-6 2
3
4
5
6
7
8
4-9 1
2
3
4
5
6
7
8
Average
0, %
Location, Fraction of Duct Width
.062
1.051
1.071
1.065
1.055
1.051
1.085
1.102
1.133
1.075
1.076
1.089
1.124
1.095
1.072
1.081
1.095
1.082
+2.1
.188
1.005
.997
.990
1.000
1.065
1.026
1.052
1.086
.996
1.007
1.008
1.048
1.070
.995
1.012
1.014
1.023
+3.0
.312
.991
.948
.964
.967
1.030
.994
1.015
.912
.964
.960
.965
1.016
1.000
.962
.986
.932
.975
+3.3
.438
.969
.929
.956
.967
.935
.923
.956
1.002
.957
.928
.931
.915
.951
.952
.945
.929
.947
+2.4
.438
.963
.951
.962
.968
.927
.963
.944
.980
.953
.933
.940
.922
.892
.967
.950
.929
.946
+2.3
.312
.972
.997
.979
.979
.976
.954
.950
.956
.976
.982
.956
.931
.959
.974
.962
1.008
.969
+1.9
.188
.994
1.031
.1.021
1.002
1.004
.980
.967
.993
.996
1.031
1.010
1.007
.978
1.002
.993
1.041
1.003
+2.0
.062
1.064
1.080
1.065
1.058
1.010
1.074
1.019
.934
1.077
1.088
1.102
1.037
1.057
1.072
1.076
1.049
1.054
+3.7
Tabular values = Row avera9e at Iocat1on noted
                 Average from complete traverse
                                -88-

-------
For accurate operation,  V =  V»,  so that
                                2 R T
The Annubar may be defined as acceptable for use in a particular
installation if the calibration factor k is a known function of the total
volumetric flow.  Preferably, k should be constant for a particular
installation.  The instrument would be considered universal  if k were
constant for all installations, i.e., the instrument could be used
without any in-place calibration.  The term  k is  being used  to avoid  confusion
with the factory calibration factor S for use in circular ducts.

     The rectangular duct data were analyzed to determine k for each run
for the two general test configurations.  Results are shown in Table 10.
Two runs were omitted from the first group for calculation of mean value
and standard deviation under the assumption of operator error.  This left
a  net  17 runs in the first group and 16 in the second group for consideration.
For the 17 runs in the first group, the maximum deviation from the mean
value  for k was less than 5 percent, and less than 6 percent for the second
group.  The largest error based on the mean value for all 33 runs was
8.1 percent.  The data summary presented in Table 11 is encouraging.  It
shows  the mean  value of k to be quite constant for each of the two
configurations:  for the first group, use of the mean value of k would
result in an average velocity error of less than 4.6 percent, 95 percent
of the time, and an error of less  than 6 percent for the second case.
Using  the overall mean value would result in velocity errors of less than
7.2 percent, 95 percent of the time.

      1973 results were encouraging for the Row Average and Annubar techniques.
The purpose of  the 1974 testing was then to broaden the data base to get a
better estimate of the universality of these methods.
                                  -89-

-------
Table 10.  ANNUBAR DUCT MEASUREMENT RESULTS,  1973

PUN
r\Uli
NO.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

v
M/sec

2.2
10.3
17.0
17.3
15.6
14.2
12.4
13.3
12.6
16.6
14.5
18.8
20.0
20.0
20.5
14.5
5.5
8.2
10.9


k

.7085
.6554
.6393
.6824
.6809
.6767
.6846
.6896
.6924
.6799
.6395
.6776
.6533
.7192
.6677
.6684
.6685
.6709
.6906
% ERROR FOR k

BASED ON AVERAGE k
FOR TWO-STREAM FLOW
+5.494
-2.412
-4.809
+1.608
+1.385
+0.759
+1.936
+2.680
+3.097
+1.236
-4.780
+0.893
-2.725
+7.087
-0.581
-0.476
-0.462
-0.104
+2.829
BASED ON OVERALL
AVERAGE FOR K
+8.008
-0.087
-2.541
+4.029
+3.800
+3.160
+4.365
+5.127
+5.554
+3.648
-2.511
+3.297
-0.407
+9.639
+1.788
+1.895
+1.910
+2.276
+5.279
    A.   RESULTS FOR TWO STREAM FLOW CASE
    Calibration factor k calculated from
                   2 R T
                     -90-

-------
Table 10 (continued).   ANNUBAR DUCT MEASUREMENTS  RESULTS,  1973

RUN
NO.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

V
M/sec

19.9
20.1
20.6
15.0
16.0
16.5
15.0
12.4
16.3
13.6
15.8
13.3
11.5
12.4
12.9
11.3


k

.6520
.6616
.6722
.6551
.6400
.6365
.6557
.6026
.6431
.6215
.6333
.6510
.6494
.6082
.6301
.6169
% ERROR FOR k

BASED ON AVERAGE k
FOR CORNER FLOW
1.987
3.488
5.146
2.471
0.109
1- .438
2.565
-5.741
.594
-2.784
- .939
1.830
1.580
-4.865
1.439
-3.503
BASED ON OVERALL
AVERAGE FOR k
- .605
.858
2.474
- .133
-2.435
-2.968
- .041
-8.136
-1.962
-5.255
-3.456
- .757
-1.001
-7.282
-3.944
-5.956
         B.  RESULTS FOR FLOW AROUND A CORNER CASE
           Calibration factor k calculated  from
                          V2 R T

                          	-
                            D
                            -91-

-------
Table 11.   SUMMARY  OF  ANNUBAR  DUCT MEASUREMENT RESULTS,  1973


A
B
OVERALL
ave

.6716
.6393
.6560
STANDARD DEVIATION a
ACTUAL
±.015
±.019
±.024
I
±2.3
±3.0
±3.6
2a
%
±4.6
±6.0
±7.2
  Explanation of standard  deviation (a):
  For a normal  distribution,  there is a 68 percent
  chance that the measurement error will be less than
  la, and a 95 percent  chance that the error will  be
  less than 2a.
  Runs 1  and 14 from  group A were omitted from
  calculations  of  k     and a
   E
•V?
               (k -
                              n = number of runs
                         -92-

-------
6.4.2.2  1974 Testing-
     Inlet configurations used for the rectangular mapping were analogous
to those for the circular case:  a straight inlet and a mitered elbow
inlet as shown in Figures25 and 29.  Emphasis was placed on testing
behind an elbow for two reasons:   good results were obtained there for
1973 testing, and an elbow is the most common flow disturbance in
rectangular duct systems.  The reference total flow measurements were
taken in the reference section in the same manner as for circular duct
testing, so that complete traverses of the test section were not required.
Two techniques were used to obtain data:  the row averaging technique
developed in 1973 and the Annubar.  The Annubar, as before, was placed
along the center "row."  In this  case, the rows were parallel  to the
short sides unlike in 1973 when they were parallel to the long sides.
Recall that the rows are defined  aa being in the plane of the  elbow for
that configuration.  For the straight inlet, blockage with the 4" x 8"
aluminum plates was generally applied in a manner consistent with the
definition of rows being taken in the plane of maximum velocity strati-
fication.  A total of 48 runs was performed for the straight inlet case,
and a total of 40 runs was performed for the rectangular inlet case.
Blockage configurations are shown in Figure 36.  In most cases, row
averaging data was taken at two locations:  18.9% and 13.9% of the
duct height from the top of the duct.  The 18.9% position was  chosen
since the most consistently accurate data in the 1973 testing  was
obtained at 18.8% of the duct width.

      Data from the straight inlet configuration, as shown in  Tables 12
and 13  generally agreed with the 1973 results.  The average Annubar cali-
bration factor in the test section was .646 for 46 of the 48 runs, compared
with .656 for the 1973 testing.  The standard deviation, however, was up
significantly, from 3.6% to 9.0%.  It is apparent that this increase in
the data spread was due to a much greater variety of inlet conditions for
the 1974 tests.  The deviation for this case was also significantly higher
than for either circular duct case discussed in the last section.  It
would seem apparent, then, that the Annubar cannot be expected to give
as consistent results in a variety of rectangular duct applications as in
a variety of circular duct applications.  The data do tend to  indicate,

-------
          Shaded portion of inlet blocked  off.   Each set of four runs
     1    was performed at four different  flow rates    .,
                                   2
Runs 1-4,
33-36, 49-52
5-8, 69-72           9-12,  57-60         13-16, 61-64
                                                                    8
17-20,  77-80          21_24j 73_76        25-28, 81-84         29-32, 65-68
                          10
                                                                    12
    37-40
   41.44          45-48, 85-88
53-56
            Figure 36.   Rectangular inlet blockage - 1974

                                   -94-

-------
   Table 12.   1974 ROW AVERAGE RESULTS AFTER A STRAIGHT INLET
              (CONFIGURATION 1)
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
M3/sec
6.72
4.76
3.97
2.62
5.74
4.45
3.94
2.64
6.62
4.80
4.07
2.74
4.71
4.07
3.61
2.62
4.48
4.09
3.73
2.25
4.36
3.90
3.61
2.60
xc
.159



.192
.134
.122
.128
.130
.158
.144
.133
.131
.129
.124
.150
.142
.171
.177

.177
.179
.176
.178
Location,
Fraction of
Duct Height
.189
1.077
1.051
1.039
1.068
.998
1.029

1.111
1.056
1.073
1.077
1.160



1.225
1.089
1.030
1.029
1.263
1.064
1.084
1.060
1.091
.139
.949



.962
1.003
1.028
1.020
1.009
.957
.991
1.017
1.047
1.045
1.106
.940
.995
.948
.909
1.151
.791
.655
.839
.675
Run
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
3 V
.IT/sec
5.35
4.40
4.03
2.79
5.25
4.40
4.10
3.00
6.86
5.05
4.22
2.91
4.90
4.19
3.94
2.84
6.20
4.65
4.11
2.89
5.33
4.60
4.07
2.81
XC


.119

.195
.166
.172
.236
.158
.153
.146
.172
.187
.168
.171
.159
.147
.133
.113
.136
.182
.232
.190
.227
Location,
Fraction of
Duct Height
.189
1.112
1.124
1.095
1.156
.986
1.097
1.066
.921
1.052
1.051
1.071
1.052
1.005
1.052
1.086
1.062
1.147
1.193
1.186
1.278
1.034
.865
.994
.886
.139


1.028

.872
.886
.878
.838
.969
.980
.988
.900
.890
.930
.850
.960
.973
1.020
1.064
1.015
.780
.709
.758
.734
XC = Interpolated location, in fraction of duct height, where row
     average equals V
 Other tabular data = Row averac*e at 1ocation ™ted
                                  V
 V obtained in reference section
                                -95-

-------
       Table 13. ANNUBAR CALIBRATION FACTORS FOR 1974

                 CALIBRATION TESTS - CONFIGURATION  1
Run
' 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
3 V
MJ/sec
6.7
4.8
8.0
2.6
5.7
4.5
3.9
2.6
6.6
4.8
4.1
2.7
4.7
4.1
3.6
2.6
4.5
4.1
3.7
2.3
4.4
3.9
3.6
2.6
S
REF
.697
.670
.664
.629
.694
.669
.663
.592
.695
.687
.676
.644
.732
.669
.634
.655
.677
.650
.665
.535
.669
.685
.657
.639
k
TEST
.708
.683
.676
.639
.677
.646
.629
.599
.684
.690
.677
.659
.670
.670
.610
.563
.633
.633
.633
.504
.568
.563
.567
.526
Run
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
V
M3/sec
5.4
4.4
4.0
2.8
5.2
4.4
4.1
3.0
6.9
5.0
4.2
' 2.9
4.9
4.2
3.9
2.8
6.2
4.7
4.1
2.9
5.3
4.6
4.1
2.8
S
REF
.709
.700
.719
.676
.675
-
.644
.490
.689
.680
.681
.665
.652
.637
.650
.652
.695
.668
.683
.649
.675
.666
.662
.612
k
TEST
.758
.739
.761
.746
.672
-
.614
.481
.709
.692
.690
.647
.603
.593
.595
.560
.695
.700
.715
.662
.570
.608
.569
.559
 S  =  .668,  ac = +  .026   (+3.9%)
           o   —
T  =  -646  a.   =  .058     (+9.0%)
        *  K
                                     Runs 20 and '32 not

                                     included

-------
 however,  that  the  data  spread  as a  function of flowrate for a specific
 configuration  will  be acceptably small.  Data for the sets of four runs
 which  correspond to specific inlet  conditions show much greater consistency
 than is  indicated  overall.  The result would be a suggested value of  .65
 for the Annubar calibration factor  for general use in rectangular ducts,
 subject  to  revision through in place calibration.  Another possibility is
 that the  accumulated test data can  be used to predict a calibration factor
 for a  specific installation.  This  was done to predict a calibration
 coefficient for field test use, and results are described in Section 1 ,
       The row mapping data for the  straight inlet configuration agreed
 reasonably  well with 1973 results,  but did show some undesirable features.
 As mentioned above,  the 1973 data showed the best results at 18.8% of the
 duct width  from ore wall.  Since the rows were 12.5% of the duct width part,
 however,  resolution of  the most desirable location was low.  The bulk of
 the 1974  data was  taken at two locations only 5% of the duct height apart,
 so that resolution was  improved considerably.   Data at this interval
 allowed interpolation to better locate the exact row location for agree-
 ment with the overall average,  and  this location turned out to be 16.1%
 of the duct width  from one wall, which agrees  with 1973 results.
 The disturbing part of the results  is that the data scatter for  this
 location and for the averages  obtained for the two test rows  is  fairly
 large,  as shown in  Table 12 .   What is even more  disturbing  is  that there
was an  average velocity  gradient of 14% between  the two rows,  which were
 only 5% of the duct width apart.   Such a  high  gradient  in the  vicinity of
 the measurement is clearly undesirable, since it would tend to lead to
 high amounts of data scatter.  For  several cases, though, the scatter was
 a  function  of  the  inlet configuration changes rather than for flowrate
 changes  for a  specific  inlet configuration, just as was the case for the
 Annubar.  As shown in Table 14, for several inlet conditions there was
 a  net  region where the  row average  agreed with the overall  average within
 5% for all  four flowrates, which is encouraging.
                                   -97-

-------
Table 14.   ACCEPTABLE FLOW MAPPING REGION FOR CONSTANT INLET
           GEOMETRY AND VARIABLE FLOWRATE FOR 1974 MAPPING
           TEST-STRAIGHT RECTANGULAR INLET
RUNS
1-4
5-8
9-12
13-16
17-20
21-24
25-28
29-32
33-36
37-40
41-44
45-48
NET ACCEPTABLE ROW MAPPING REGION
FOR 5% ACCURACY FOR FOUR RUNS, % OF DUCT HEIGHT
Insufficient Data
12.2 - 15.4
13.6 - 15.0
None
None
17.3-18.4
None
None
15.6-17.6
16.5-18.2
13.3-13.5
None
                              -98-

-------
      At first glance, the rectangular elbow data would appear to be
disastrous, but was in fact very instructive.  The test section measure-
ments were taken further downstream from the elbow than in 1973 to
evaluate the effect of that parameter.  The results revealed that probe
placement relative to an elbow is very important.  Results for the 40 runs
are shown in Tables 15 and 16,  and indicate virtually no agreement with
1973 results.  The greater distance from the elbow is clearly the reason
why, and it gave the following insight into flow around a rectangular elbow.
      Flow around a mitered rectangular elbow creates a high velocity
region toward the outside of the elbow and a low velocity recirculation
region toward the inside.  The presence of the elbow results in a somewhat
two dimensional flow immediately downstream of the elbow, where "uniform
two dimensional flow"  is  explained as follows:   Consider a coordinate
system  in  the duct with  the x axis parallel  to the duct axis, the y axis
in the  plane of the elbow, and the z  axis  normal to  the plane of the elbow.
The flow would be  two  dimensional  if  the velocity  vector at  any point  had
components  only in  the x and  y directions.   It would also be uniform if the
profiles  in the x  - y  plane were the  same  for all  values of  z.  This
would cl early. 6e  tb7e  tdeal situation  for use of  the  Row Average Method,
since all  row averages would  be  identically  equal  to the total average.
1973  results  indicated that an elbow  tends to produce  this  ideal  uniform,
two-dimensional flow.  The 1973  data  showed  that the highest row  averages
occurred toward the outside of  the duct,  leaving a lower  average  in  the
center.   The 1974 data conclusively  showed that  this condition worsens as
the  flow proceeds downstream:  the recirculating region  recedes  from the
ends  of the duct, creating  higher  velocities there,  and  grows  in  the
center   creating  lower velocities.   Thus  the relatively  flat profile
 becomes very three dimensional,  with  high  velocities on  three  sides  of
 the duct, and a low velocity  region  in the center, all  of which  is
 reflected in the  data.  The Annubar  readings were consistently low,
 resulting in a very high calibration factor, .853, which was 31% higher
 than the .65 obtained for other tests.   A systematic increase in the
 row averages obtained at the  same locations as  for the straight inlet  also
 occurred, as well as a change in the direction of the gradient between
 the two rows.  The one desirable result is that the magnitude of the gradient
 decreased  by a factor of three.
                                   -99,

-------
 Table  15.  1974  ROW AVERAGE  RESULTS
            (CONFIGURATION 2)
AFTER A RECTANGULAR ELBOW
Run
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
V
m /sec
' 5.6
4.6
4.0
2.9 ,
3.3
3.1
3.1
2.4
5.4
4.5
4.0
3.0
4.7
4.3
4.0
3.1
4.7
4.2
3.9
3.1
Location,
Fraction of
Duct Height
.239

1.006
1.171
1.090
1.012
1.090
1.047
.951
1.056
1.063
1.069
1.037
1.042
1.070
1.133





.189
1.284
1.180
1.181
1.266
1.221
1.265
1.218
1.212
1.061
.999
1.093
1.101
1.060
1.086
1.153
1.151
1.051
1.125
1.080
1.006
.139
1.317





1.426




1.143

1.176
1.222
1.209
1.025
1.105
1.083
1.000
Run
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
V
m /sec
5.3
4.6
4.1
3.2
4.0
3.8
3.4
2.8
4.5
4.3
3.9
3.1
5.1
4.5
4.2
3.2
5.4
4.7
4.3
3.2
Location,
Fraction of
Duct Height
.239












1.060
1.091


1.244



.189
1.190
1.191
1.245
1.229
1.264
1.138
1.177
1.252
1.099
1.055
1.189
1.151
1.050
1.028
1.053
1.074
1.259
1.250
1.140
1.215
.139
1.236
1.198
1.275
1.279
1.210
1.190
1.196
1.192
1.180
1.194
1.225
1.320


1.051
1.068
1.268
1.276
1.214
1.247
Tabular values = Row avera^e at 1ocation noted
                              V
V obtained in reference section
                             -100-

-------
      Table 16.  ANNUBAR CALIBRATION FACTORS FOR 1974
                 CALIBRATION  TESTS  -  CONFIGURATION  2
Run
49

50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
V
m /sec
5.6

4.6
4.0
2.9
3.3
3.1
3.1
2.4
5.4
4.5
4.0
3.0
4.7
4.3
4.0
3.1
4.7
4.2
3.9
3.1
S
REF
.729

.734
.726
.692
.650
.684
.685
.674
.717
.787
.697
.706
.773
.746
.751
.724
.734
.744
.765
.729
K
TEST
.769

.759
.788
.792
.737
.808
.829
.736
.741
.803
.740
.736
.896
.831
.865
.870
.912
.891
.848
.858
Run
69

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
V
m /sec
5.3

4.6
4.1
3.2
4.0
3.8
3.4
2.8
4.5
4.3
3.9
3.1
5.1
4.5
4.2
3.2
5.4
4.7
4.3
3.2
S
REF
.779

.750
.731
.714
.746
.729
.705
.755
.766
.738
.752
.722
.704
.717
.719
.687
.763
.751
.764
.801
K
TEST
.824

.906
.936
.906
.974
.959
1.058
1.055
.819
.863
.803
.833
.818
.766
.745
.721
1.046
.970
.920
1.003
 S =  .7315, os =+ .032 (+4.4%)


~K~=  .853,  0((  = + .093 (+10.9%)
                            -101-

-------
      The Annubar in the reference section also apparently showed the effects
of  the  elbow.   For  the  straight  inlet configuration, the  average Annubar calibra-
tion factor was  .668, lower than the .698 for the circular case.  It is
felt that this  lower value reflects the transition from a rectangular to
a circular duct  at  the entrance to the reference section.  The narrow
rectangular duct would tend to create a high velocity region along the
duct center, which would affect the Annubar reading more than it would
the Log Linear  reference traverse, since the latter uses no center!ine
data.  The higher local velocity would increase the Annubar reading, thus
lowering the calibration factor from the circular case.  For the elbow
inlet case, the  low velocity region created apparently persisted into the
reference section.  Again, since the Annubar responds more to the center-
line flow conditions than does the Log Linear method, the local  low
velocity region  resulted in a relatively low Annubar reading, and thus a
higher calibration  factor of .7315.
6,4.2.3  1975 Testing -
     This testing was performed solely to determine the optimum location
downstream of a  rectangular elbow for Row Average and Annubar measurements.
The mapping facility was set up as in configuration 2 for the 1974 testing
(Figure 25).  The test section was initially located as for the  1974 tests,
then moved forward toward the elbow.  Locations for the pi tot probe and
Annubar for each case are shown in Figure 30.   Note that the locations
are given in terms of the local duct width instead of the effective duct
diameter.  The flow properties of interest vary as a function of the duct
width.   Six row averages were taken over half the duct for each  run, along
with the normal  Annubar data.   Reference measurements were taken in the
reference section as usual.   At each location,  runs were performed at
maximum and minimum flow rates for inlet blockage conditions 1,  2 and 6
as shown in Figure 36.   A total  of 48 runs was  performed for row averaging,
and a total  of 60 for the Annubar.   Results are summarized in  Tables 17 and
18.
                                   -1Q2-

-------
             Table 17.  ROW AVERAGE RESULTS FOR 1975 TESTING
TEST
SECTION
LOCA-
TION
A
Bl
B2
C
D
E
F
G
ROW AVERAGE LOCATION, FRACTION OF
DUCT HEIGHT
.045
1.305
1.106
1.119
1.112
1.099
1.084
1.083
1.067
.136
1.093
1.026
1.022
1.026
1.045
1.032
1.042
1.009
.227
1.005
.987
.976
.966
.968
.957
.946
.963
.318
.908
.965
.970
.957
.964
.979
.960
.989
.409
.824
.958
.949
.963
.962
.977
.981
.985
.500
.730
.914
.927
.942
.928
.942
.978
.974 j
i
Tabular values are the mean ratio of the row average velocity to
the total average velocity for the six runs performed at each test
section location.
                                 -103-

-------
Table 18.  ANNUBAR RESULTS FOR 1975 TESTING
TEST
SECTION
LOCATION
i
i
A
Bl
B2
C
D
E
F
G
H
I
. DISTANCE BETWEEN
ANNUBAR AND ELBOW
AS FRACTION OF
DUCT WIDTH
2.906
2.062
2.062
1.625
1.500
1.3125
1.125
.9376
.75
.562
AVERAGE ANNUBAR
CALIBRATION FACTOR
FOR SIX RUNS
.843
.700
.703
.688
.649
.617
.617
.581
.571
.578

-------
     To obtain Table 17, row average velocity data at each row location
for each run were divided by the total  average velocity for that run.
These six ratios at each row location were then averaged to obtain the
tabular entries shown.  Results confirmed the belief that an elbow
tends to produce a uniform, two-dimensional  flow in its immediate
vicinity, and that the flow pattern changes  radically further downstream.
Consequently, the results showed that for the positions closest to the
elbow, a row average velocity taken between  about 14% and 40% of the duct
width for one wall would agree with the overall average velocity within
±5%.
     Data in Table 18 for the Annubar show that a "steady state" calibration
factor as a function of distance from the elbow was not reached, as had
been hoped.  At location D, which corresponded to the 1973 location, the
calibration factor was  .649, compared with .6393 for 1973 testing behind
an elbow.  This agreement is good, but the relatively large changes in
calibration factor with distance from the elbow make it appear unlikely
that good repeatability of the calibration factor would be attained in a
large number of different facilities.
6.5  MAPPING TEST SUMMARY
6.5.1  Circular Duct Techniques
     EPA Method 1 flow mapping techniques are effectively identical  for
both circular and rectangular ducts -- in each case the duct is divided
into a number of equal area segments and velocities are measured at the
center of each segment.  No knowledge of flow profiles is assumed.   Mapping
techniques such as the Log-Linear method retain the equal area segment aspect,
but go a step further in probe placement by recognizing that fully developed
pipe flows are axisymmetric and that the flow profile is roughly parabolic
in shape.  Symmetry is the most important point, since it means that an
average taken along one diameter is the same as that taken along any other
diameter.  This reduces the mapping problem to one of determining a proper
technique to obtain a correct line average along any one diameter.   Thus
circular duct techniques such as the Log-Linear method are developed
for the special case of fully developed pipe flow, and then their
                                  -105-

-------
 accuracy  in  non-developed  flows must be  empirically determined.
 Design of the Annubar followed this same route.
     Testing in 1973 downstream of  an  elbow, testing in 1974 downstream
 of  an elbow, and testing in 1974 downstream of a straight inlet with
 partial blockage too"k place at 1, 1.4, and 2.2 duct diameters downstream
 for a disturbance, respectively, for pitot traverse testing, and slightly
 further downstream for 1974 testing with the Annubar.   Three techniques
 were evaluated extensively:  the Log Linear 4, which requires a total
 of  eight points, the new empirical, which requires five, and the Annubar.
     On the basis of the test data, it is being concluded that the Annubar
 and Log Linear 4 methods qualify as being "Universal", and should be
 expected to have accuracies on the order of + 5% without the need for in
 place calibration.  The New Empirical method was found to be acceptable
 as  a "Single Calibration" method -- establishment of a calibration factor
 at  one flow rate should be adequate to insure accuracy over the entire
 range of flows at one location.  Testing showed that good accuracy can be
 achieved in applications where EPA Method 1  would require in excess of
 forty.   The Annubar, New Empirical, and Log  Linear 4 techniques require
 one, five, and eight measurements, respectively.   Results are summarized
 in Table 19.
 6.5.2  Rectangular Duct Techniques
     The flow profile symmetry which is the  basis for  circular duct
mapping techniques does not exist in fully developed flows  in rectangular
 ducts due to three dimensional  effects.  Work during the program did,
 however, identify a common situation in which flow profiles  are very
 similar in magnitude and shape across the duct.   This  situation occurs
immediately downstream of an elbow, which is  the  most  common flow
disturbance in rectangular duct systems in process streams.   The row
average and Annubar methods which were  successfully used during the
program achieved success by averaging flow along  a line  between the
duct walls in a consistent manner for a variety  of flow  conditions.
The critical  point is that the flow properties  in the  duct were such
that the average velocity along the line chosen was representative
of the  average velocity for the entire  duct.  Thus when  success  was
                                  -106-

-------
      Table 19.   RECOMMENDED CIRCULAR DUCT FLOW MEASUREMENT TECHNIQUES

A.  Single Probe Techniques
       Ellison Annubar - Install  and use according to factory
         instructions
                                              *
B.  Point Sensor Array or Traversal  Techniques
    Preferred:
    Log Linear 4
    Acceptable1:
    New Empirical
    Gauss Sherwood 4
    Log-Linear 2+4
          *
           Methods shown in Table 2
           In place calibration check considered a requirement;  in
           place calibration recommended for Log Linear 4 and Annubar

-------
achieved in a rectangular duct,  it was  for the  same  reason  as  in  a
circular duct -- similarity of flow profiles.
     Testing showed that accurate measurements  could be  taken  downstream
of an elbow with the Row Average and Annubar methods because  the  elbow
causes a violent disruption of the flow,  which  accomplishes two important
results:   1.  The disruption tends to  damp out influences  of disturbances
further upstream, resulting in consistent conditions at  the test  area.
2.  The elbow tends to condition the flow to a  uniform,  two-dimensional
shape immediately downstream of the bend.  Testing with  the straight
inlet in 1974 showed that both methods  work reasonably well for that
general case, but are more consistent downstream of  an elbow.
     The recommended circular mapping techniques have been  in existence
for several years, so a body of data exists for them outside  of this
program.  It was partly on the strength of this additional  work that
the Annubar and Log Linear techniques were recommended as being "Universal".
Since a body of outside data does not exist for Annubar  use in rectangular
ducts or the Row Average method, it is  felt that it  would be  premature  at
best to turn either technique "Universal".  Test data indicate that the
Annubar will probably always need in place calibration at one flow rate,
preferably near the mean flow rate, for use in  rectangular ducts, but
that this calibration should result in continuous monitoring  accuracies
on the order of +5%.  There is good reason to believe that the Row Average
technique is in fact "Universal" downstream of an elbow, but  more data
should be accumulated as a final judgement is made.   Recommended  installations
for the Row Average and Annubar techniques are given in  Figures 37 and  38.
Test conclusions are given in Table 20.
                                 -108-

-------
1
o
     ROW AVERAGE:
       REFINE ESTIMATED CALIBRATION
       FACTOR OF 1.00 BY IN-PLACE
       CALIBRATION

     ANNUBAR:
       REFINE ESTIMATED CALIBRATION
       FACTOR OF 0.65 BY IN-PLACE
       CALIBRATION
                        MEASUREMENT
                         PLANE
                        I
                       0>
                       FOR  VELOCITY  PROFILES AS  SHOWN  IN  Y
                       AND  Z  DIRECTIONS,  ROWS ARE  DEFINED
                       AS BEING  IN Z DIRECTION
               VELOCITY
ROW AVERAGE
 LINE (CHOOSE ONE)*
                                                     ANNUBAR
                                         VELOCITY
,19W
\_
^


t


• • • oi


i
I
t

	 • JW


t
.19W
                                                              *ROW AVERAGE MAY BE TAKEN
                                                               ANYWHERE IF SUPPORTED BY
                                                               ADEQUATE CALIBRATION DATA.
                            Figure  37.   Probe  placement for general  rectangular duct applications

-------
CD
I
 ROW  AVERAGE:
   REFINE  ESTIMATED  CALIBRATION
   FACTOR  OF  1.00  BY IN-PLACE
   CALIBRATION
ANNUBAR:
  REFINE ESTIMATED CALIBRATION
  FACTOR OF 0.65 BY IN-PLACE
  CALIBRATION

                MEASUREMENT
                PLANE
                              FLOW
                              DIRECTION
              ANNUBAR:    X= 1.50^  (NOMINAL)
              ROW AVERAGE:    X=
                                      +.40W
                                           'i
                                                                              ROWS ARE DEFINED TO BE IN
                                                                              THE PLANE OF THE ELBOW
                                                                              ANNUBAR
                                                                                      ROW AVERAGE LINE
                                                                                      (CHOOSE ONE)*
                                                            -.04 W
                                                                                       *ROW AVERAGE MAY BE TAKEN
                                                                                        ANYWHERE IF SUPPORTED BY
                                                                                        ADEQUATE CALIBRATION DATA.
                                                                                               +.25  W0
                                                                                        -.19W2         2
                                                                                             *•  -.04  W2

                                                                                        .5Wo
                                      -.60W
                                           1
                            Figure 38.   Probe placement  after a  rectangular elbow

-------
   Table  20.  RECOMMENDED RECTANGULAR DUCT FLOW MEASUREMENT TECHNIQUES


A.  Single Probe Techniques
       Ellison Annubar - Install as shown in Figures 37-38.
       In-place calibration required to modify estimated correction
       factor of .65

B.  Point Sensor Array or Traversal Techniques
       Row Average  Method  -  Install  as  shown  in  Figures  33-34.   For
       array,  use of eight sensors  recommended.   In  place  calibration
       required  to  minimize  systematic  error.  A row is  defined  as
       being  in  the direction  of maximum velocity stratification.
                                 -Ill-

-------
                               SECTION VII
                    TASK IV - LABORATORY ASSESSMENT

7.1   FACILITY DESCRIPTION AND SCOPE OF TESTING
    Two types of tests were performed during this task -- calibration
and environmental.  The bulk of the calibration work was performed in the
Fluid Mechanics Laboratory Low Speed Wind Tunnel, shown in Figure 39.
The wind tunnel had a closed rectangular test section 28.6 cm square
during the testing period.  The nominal speed capability was 3-30 m/sec.
Reference velocity measurements were provided by standard hemispherical
nosed pitot static probes (calibration factor 1.00) connected to a cali-
brated Baratron pressure transducer.  Temperature measurement was provided
by a calibrated chromel-alumel thermocouple immersed in the flow.  Addi-
tional calibration work was performed on the Ramapo Drag meter using
calibrated weights, and on the TSI hot film sensors using a standard
Thermo Systems Calibration Wind Tunnel (Figure 40)•
    Environmental testing was primarily performed in the TRW Process
Simulator (Figure 41) to determine sensor resistance to temperature and
humidity.  The facility also had a particulate generation capability,
but use of particulate was prohibited by local Air Pollution Control
District officials unless a scrubber was also operating.  Since a scrubbing
capability was not available at the time, it was necessary to evaluate
particulate effects through use of a portable sandblaster, which proved
to be adequate for test purposes.
    Calibration  tests were selected to characterize test sensors as
fully as possible within  the program scope.  The primary calibration was
a basic velocity  calibration in a uniform laminar flow with the test sensor
directly aligned with the flow.  Time  response tests were performed to
verify that the  sensors could adequately follow flow variations.  Stability
testing was performed to  determine  instrument drift characteristics.
Finally, orientation  sensitivity tests were performed to characterize  the
ability of the sensors to resolve the  required axial component of the  flow.
This  testing  is  important because in  the applications of interest,  it

-------
                                                   EXHAUST
LIQOJD I*; SUPPLY
FOR THERMAL CONTROL
        DIFFUSER
   TRANSVERSWG
   MECHANISM
   AND PROBE
                                                          SUPPLV SECTION FOR
                                                          RECfRCULATEDGAS
WATER SPRAY
SECTION


 TEST SECTION
                      Figure 39.  Low speed wind tunnel
                                  -113-

-------
                                 THERMAL ANEMOMETER  PROBES
                         CALIBRATORS AND  CALIBRATION
          CALIBRATION APPARATUS • MANOMETERS • CALIBRATION SERVICE
                                    CALIBRATORS
                                      'Air
                                       Water
                                      • Other Fluids
MODEL 1125 CALIBRATORS  FOR AIR
• HIGH ACCURACY
• CALIBRATE ANY PROBES FROM 0.1 FT/SEC TO MACH # 1
• EASY TO USE
• CAN BE USED WITH MOST GASES

The Model 1125 Calibrator is designed to calibrate hot wire and hot film
anemometer probes over a wide velocity range. Reference calibration
data is furnished giving velocity vs. pressure drop for each range.

SPECIFICATIONS:
Velocity Range: Probes can be calibrated from 0.1 ft/sec, to
             MACH #1 13 cm/sec to M #1) by appropriate
             use of interior chamber, interior nozzle and
             exterior nozzle.
Accuracy: t 1% above 10 fps; + 2% 0.5 to 10 fps; ^ 10% below
             0.6 fps
Background Turbulence Intensity: <0.1% at 100 fps
Nozzle Sizes:    2.83" (72mm) Dia. Interior Chamber;
             0.65 (16.5mm) Dia. Interior Nozzle: 0.15" (3.8mm)
             Dia. Exterior Nozzle.
Temperature Limits: 0 - 65°C (32 - 150°F)
Pressure Limit: 7 ATM Internal (20 ATM Max. Into Regulator)
Power Required: 115V AC 60 Hz or 230V AC 50 cos (specify)
Size: 20" High x 18" Wide x 10" deep
Weight: 15* (7 kg)
                                                                                   "•-.*.
ComprMMd Air
supply ix Gn
                                           MODEL 1125 BLOCK DIAGRAM
MODEL 1127  CALIBRATOR FOR WATER
The Model 1127 Calibrator is set up to utilize tap water for calibration of water probes.

SPECIFICATIONS
 Range: Three Calibrate Positions:
  0.02 to 0.2 fps (.6 to 6 cps)
  0.15 to 1.5 fps 14.5 to 45 cpsl
  1.5 to 30 fps 1.45 to lOmps)

Line Water Pressure: Must be at Least 10 psig (50 cm Hg) (If tine                 ,
pressure varies too much a constant head tank and sump pump can be
used pumping from an over flow lank to recirculate water]

Equipment Furnished: Tank, course regulator, fine regulator 1 ^ filter,
complete calibrator, extra larger nozzle, finings, valves and hoses.
(Manometer not furnished - Model 10134 or equivalent recommended)

MODEL 1128 CALIBRATOR (Not Shown)

The Model 1128 Calibrator is a combination of Model 1125 and 1127       i
for those who wish to calibrate in both air and water. Model 1128 comes
as one complete Model 1125 plus the tank and accessories necessary
to convert to the Model 1127 Water Calibrator.
                                             CALIBRATION
                                              CHAMBER
                                             L.ki Mod«i 11 25
40
                                THERMO-SYSTEMS Inc.
        00NORTH CLEVELAND AVE
         PAUL. MINNESOTA  55 MJ °
                                                                                    SI7 633 00,0
                                 Figure  40.    TSI  wind  tunnel

                                                 -114-

-------
 FLOWMETER


 FLOW CONTROL VALVE


 PRESSURE REGULATOR
"   POLLUTANT GAS
    MIXER SYSTEM

  AIR BLOWER
 (1000SCFMTO
  .1000 SCFM)
                                        WATER
                                      SPRAY INLET
                               DRY POLLUTANT
                               MATERIAL FEEDER
                             (0.1 TO 10 GRAINS FTJ)

               AIR
              DAMPER
\
NATURAL
  GAS
 INLET
  L    -r
  »— [X] - 1
      ^
                     NATURAL GAS BURNER
                        (770FTO400°F)
                                                                                           SCRUBBER
                                                                                            INLET
                                                                   SECTION A-A
INSTRUMENTATION
  PROBE PORTS
                                                                                O
                                                                                O
                                                                                O
                                                                                O
                                                       MATERIAL MIXING
                                                            VANES
                            -PROCESS SIMULATOR SECT ION-
                                                                            PROCESS
                                                                           SIMULATOR
                                                                          MONITORING
                                                                            SECTION
                                                                                                    SCRUBBER
                                                                                                     OUTLET
                                                                                                                              CHARGED
                                                                                                                              DROPLET
                                                                                                                              SCRUBBER
                                                                                                               VIEWB
                                                                                                      CHARGED DROPLET SCRUBBER
                                                                                                                                           EFFLUENT
                                                                                                                                         MONITORIN
                                                                                                                                           SECTION
                CONFIGURATION I
                 (ONE MODULE)
                                                                                                                 CONFIGURATION II
                                                                                                                  (THREE MODULES)
                                                                                                    .CHARGED DROPLET SCRUBBER.
                                                                                                           TEST SECTION
                    Figure 41.   Coal fired combustion flue gas simulator

-------
must be expected that the local flow will not always be parallel to
the duct walls.
    Testing of the S type pi tot probe revealed that the probe tested had
a non-constant calibration factor and relatively high sensitivity to small
changes in flow orientation.  These findings eventually led to a comparison
between the S type probe and the standard pitot-static probe and the sug-
gestion that the latter be used whenever possible.   Pitot-static probe
characteristics have been heavily studied and documented, so no testing
was specifically performed on these probes.  Probe characteristics given
were obtained from Reference 3.
7.2  ANALYTICAL INSTRUMENT DESCRIPTION
    The instruments tested were described and pictured in Section 5.3.  To
understand their calibration characteristics, it was necessary to determine
their analytical properties so that appropriate calibration parameters
could be isolated.  The most straightforward way to determine analytical
properties is to construct the desired flow output equation for each
instrument.  To do this, consider the flow shown in Figure 2.  The flow
cross-sectional area, A, is constant.  Recall that the three commonly
desired outputs are volumetric flow at standard conditions, Vg; volumetric
flow at actual conditions, V, and total mass flow, m.  Figure 2 illustrates
the relationship among the three as described in Section 4.  At each
station, each of the flow parameters is constant.  Actual conditions are
given at station 1, where we have

         V = uA                                                       04)
and      m = puA                                                      O5)

At station 2,  the  static temperature and static pressure have changed so
that they are  standard, which has been defined as 20°C and 760 torr.  The
flow is considered to be chemically frozen so that no change in molecular
weight occurs.  At station 2, we have

          vs = USA                                                    (is)
and       m  = p u A                                                  (17)
           J    JO

                                 -116-

-------
Since mass must be conserved in the system, we must have

          «   •
          m = ms


so that   PSU$A = puA


or        PSUS = pu                                                  (20)


also      V  _  u                                                    /0,N
          ~	77~                                                   \c-\l
          vs    us


Instrument properties can be evaluated from these basic relationships.

7.2.1  Pi tot-Static Probe

     A properly made pitot-static probe senses the stream dynamic pres-

sure, such that the output, Ap, is. given by

          AP = (l/2)Pu2                                              (22)
so that    u = \^                                                  (23)


Since      p =  ^                                                   (24)
we have    u = V-2^                                               (25)



and
                                         X/2RTAP                     ,oc.
                                         V  p                       (26)
                                                                     (27)

so that for the flow quantities we have
                V -  A                                               (28)
               Vs =  %;    >T             	                 (29)

                     r>       /9RTxn       .   J?nAn
                                                                     (30)
                                 -117-

-------
 7.2.2  "S" Probe and Annubar
      Each of these probes' output in a uniform stream can be given by

                   = k(l/2)Pu2                                        (31)

 where k is an appropriate calibration factor.  Performing similar
derivations,  we obtain
                                                                       (33)


                                                                       (34)

                                                                       (35)

                                                                       (36)

  The value of k can be deduced from factory data for the  "S"  probe  and is
  considered to be constant.   The calibration factor for the Annubar in  a
  uniform stream is not of great importance since the instrument  is  not a
  point sensor.  The major calibration evaluation of the Annubar  took place
  during the mapping tests.
  7.2.3  Ramapo Drag Meter
       The Ramapo output, fl,  is actually a voltage ratio,  the  ratio  of the
  strain gage bridge output to the bridge excitation voltage,  and is
  directly proportional to the drag force on the  instrument:
                    v
                                                                       (37)
  where for the flow range of interest,
                D = CD(l/2ku2S                                        (38)

                                    -118-

-------
where Vgout = bridge voltage out
      VBin  = bnd9e inPut (excitation) voltage
      k     = calibration constant (known)
      D     = drag force on sensor
      Cpj    = disc drag coefficient
      p     = gas density
      u     = gas velocity
       S     = disc frontal area
     The factory calibration for the probe lists the factor k as
- c.
" 5>
                        m     f   millivolts   \
                       V kg    (volt •  kilogram )
where the term kilogram is understood to mean the downward force exerted
by a one kilogram mass acted on by the acceleration of gravity at sea
level .   Thus we have
              E = kCD(l/2)Pu2S                                       (39)
                                                                     (40)

                                                                     (41)

                                                                     (42)

                                                                     (43)

                                                                     (44)
     The voltage output  of the  unit  tested  is a nonlinear function of
velocity.   The probe responds to  the stream dynamic pressure, so that for
general  stream conditions, we have
                                 -119-
u
us
if
Vs
m
Hastings
_ W ^n
' kCp,pS
T

PS ^kCDTS
- A J 2RT*
>kCDpS
_ . Ts J2RpJ?
PS ^kCDTS
_ « J 2pR
A 'RTkCDS
Raydist Flare Gas
J2RT/?
Flow Probe

-------
              Vout =  kAp = kO/2)pu2                                 (45)
                        V2V
                      —£r                                        06)
                 u=
                        2 RT "out                                   (47)
                           TCJ
                     T
                us '  j£   1   kl""                               08)
                V=  fl    *   kp                                    («>
               VS -  * £  Kl^                             (»>
                         s
                m=  fl   >-kRf^                                 <51>
7.2.5  Thermo Systems Hot Film Sensors
     The basic operation of a hot film sensor is as  follows:   a  small
sensing element is  electrically heated (i.e., resistance  heating) to a
temperature above the ambient stream temperature.  The power  required to
maintain this temperature overheat is monitored.  The power required is
related to the stream velocity by means of a calibration  curve.  For the
general case, the thermodynamic relationship is given by

              Nu -  a + b Ren                                         (52)
where
              Nu =  Musselt Number (heat transfer parameter)
              Re =  Reynolds number based on sensor diameter
         a, b, n =  Calibration factors (n^ 0.5)
An important point  to note for this type of application is that  the
Musselt  number and Reynolds number are inversely proportional to the
fluid thermal conductivity and viscosity, respectively.  For  the type of
anemometer proposed, the output parameter, which is  the bridge voltage
VR, may approximately be given as

                                 -120-

-------
                                  1/2
                                                                      (53)
 where
                p = ambient pressure
                u - freestream velocity
           CpC2 = calibration coefficients

 so  that
                m
                m -
                                                                      (55)

                                                                      (56)


                                                                      (57)
                     RT   ~~Cr-                                     (58)
7.3  BASIC INSTRUMENT CALIBRATION
    A minimum of three calibration runs was performed for each  sensor
prior to environmental testing.  Estimated accuracies from the  manufac-
turers as a function of velocity are shown in Figure 42.   Reference
measurement accuracies were as follows:  pressure:   0.5%  of reading  or
better; absolute temperature:   0.3% of reading;  accuracy  of pitot-static
probe calibration factor: 1% over range of test.
 7.3.1   "S"  Probe Calibration
      For each  run,  ten data points were taken with  the velocity increasing,
 and  ten with  the velocity  decreasing.  The differential  pressures from
 the  S  probe and  reference  pitot  probe  were then ratioed  to obtain the
 calibration  factor  k$J  since
                                  -121-

-------
ro
ro
             ACCURACY OF VELOCITY
             MEASUREMENT, PERCENT

                 16 r
                 14 -
                 12 -
                 10 -
                  6  -
                  4  -
                  2  -
                                       10
                                                                                               HASTINGS-RAYDIST
                                                                                               AFI-10K
                           THERMO-SYSTEMS VT1613
                           ELLISON "S" PITOT PROBE
                                                                 ELLISON ANNUBAR (CIRCULAR DUCT ONLY)
                  RAMAPO MARK V
                  UNITED SENSOR  PITOT-STATIC PROBE
15
                                                                    20
                                                                                   25
                                                                                                 30
35
                                                           VELOCITY, METERS/SECOND

                                    Figure  42.  Manufacturer's  accuracy  for instruments tested

-------
       upitot-static _ v>/p.s.  _' s'p.s.  _  v"'Vs._TEST= -852
and
               VREF = '823
7.3.2  Ramapo Calibration
     The Mark V Drag Meter is shown schematically in  Figure  44.  The
principle of operation may be recalled as follows:  a disc is  inserted
into the stream, at the end of a level arm.   The drag force  on  the disc
and on the exposed section of the arm is sensed by a  strain  gauge bridge

                                 -123-

-------
i
ro
i
         l.<
         1.42
         1.40
         1.38
         1.36
         1.34
         1.32
         1.30
              CALIBRATION FACTOR,
  p  V
£_ ^00  00
                           I=ks
                                 -A_
                                 _?~
                                   *.
                                   /v-
                     _Q
CALIBRATION RUN


   A  2
   D  3
TAG ON LEFT - V
               o
TAG ON RIGHT
                                                                        ^  INCREASING
                                                                        V   DECREASING
                                    MANUFACTURER:   ks    =  1.477

                                            TEST:   kc     = 1.376
                                                     test
                                                                     -0
                                                                     FREE STREAM VELOCITY, V^, METERS/SEC
                                            10              15             20

                                        Figure 43.  S probe calibration factor
                                                                     25
                           30

-------
            DISC
                                 LEVER
                                   ARM
                              STRAIN
                                GAGE
                              BRIDGE
         EFFECTIVE
         LEVER ARM
         LENGTH
±
Figure 44.   Schematic of Ramapo Mark V  flow meter
                        -125-

-------
mounted on the enclosed part of the arm.   The bridge input and output
voltages are used to determine the force  on the arm, and finally the
stream velocity, as given in Section 7.2.3.  A typical  calibration run
is shown in Figure 45-  The line is the reference equation and the points
are test data points.  Results are shown  more illustratively in Figure 46
for the three standard calibration runs.   It should first be mentioned
that the instrument accuracy is better than it would appear in Figure 46,
where it is clear that the Ramapo output  became consistently high with
increasing speed.  It is believed that the anomaly was  due to net down-
ward motion of the disc with increasing speed, as shown in Figure 47.
The probe caused considerable blockage, and thus higher local  velocities,
at the disc plane in the wind tunnel.   The disc face was initially aligned
with the static ports on the reference pi tot probe.  As the wind tunnel
speed increased, the disc moved with respect to the pitot probe, thus
putting the pitot probe into a region  of  lower and lower relative velocity
with respect to the disc.  Consequently,  it is important to look at the
data spread in Figure 46 as a key to instrument accuracy.  A nominal
accuracy of  1.5%  on the reference measurement and a 1% instrument accuracy
are consistent with the observed spread.
     Figure 46does show an apparent hysteresis with a magnitude of about
±1/2%—the values for the decreasing side of the velocity calibration are
typically 1% higher than the ascending values.  Although hysteresis loops
are not uncommon in cyclic handling of metal rods, it did seem unusual
to see this trend since the disc and arm vibrate quite perceptibly in the
airstream, meaning the loading and unloading are not very smooth, and
since the factory claims that the instrument is hysteresis free.  Since  the
observed level is ±1/2%, the effect does  not degrade the instrument accuracy
of ±1%, and so is not critical in the final evaluation.
     Since disc motion as discussed above caused discrepancies in the normal
calibration technique, independent weight testing was done to verify the
calibration accuracy.  This testing was done to verify  the accuracy of the
calibration factor k.  Results are shown  in Figure 48. These data also
constitute  the post-environmental test for the probe,  discussed below.
The factory stated flow range for the instrument is 3.66 to 36.58 m/sec, so the
larger errors for equivalent velocities less than 3.048 m/sec. should not  be

                                 -126-

-------
100
 z
 CO
 o
>no
 I
ro
 i
o
o
                                                                        FACTORY CALIBRATION

                                                                        TEST DATA POINTS
                                                                        FLAG ON LEFT - VELOCITY  INCREASING
                                                                        FLAG ON RIGHT - VELOCITY DECREASING
              I
                                                           I
I
I  I  I I
I
I
I
I   I   I   I  I  I
   .01                                  .1                                   1
                                                 OUTPUT VOLTAGE   mv
                                                 INPUT VOLTAGE  ' V
           Figure 45.   Ramapo Mark V calibration:   Free stream velocity versus ratio of
                         output voltage to  input voltage
                                                                                                                        10

-------
CO
i
              0
              o
                  -1
                  -2
                                    I


                                   5  /
                                    ' '
10
     15             20

VELOCITY, METERS/SEC
                                                                                             25
                                                          J
                                                          30
                                       V         V

                                  INCREASING DECEASING
                                      _<>
         DATE




         6-20

         6-20
                               	   	   AVERAGE

                      Figure  46.   Difference between  factory calibration and  test  velocity in

                                   percent versus test velocity for Ramapo Mark  V

-------
                                    oo
LOW SPEED:
DISC ALIGNED WITH
STATIC PORTS ON PITOT
PROBE - DIFFERENCE IN
INDICATED VELOCITY
SMALL
                                    oo
HIGH SPEED:

DISC DISPLACED BY
LARGER DRAG FORCE -
NOT ALIGNED WITH
PITOT PROBE - DISC
"SEES" HIGHER WIND TUNNEL
BLOCKAGE AND INDICATES
HIGHER VELOCITY THAN
PITOT PROBE
Figure 47.   Diagrammatic explanation of Ramapo  output at high  speed
            in wind tunnel  testing
                               -129-

-------
              +4,-
                 VELOCITY ERROR, %
              +2
CO
o
I
              -1
              -2
              -3
             -4
                     O A
  EQUIVALENT VELOCITY, METERS/SECOND

J	      I	I
                                          1Q
 15
                                       CD
                                                     0
-n-
 O

 A

 O
20
                           25
                                                                                               30
                                         35
                          CD    Q)
         O

         A
      V INCREASING

      V DECREASING
                       Figure 48.   Results of  force calibration  of Ramapo  Mark V as  percent difference
                                    in factory  and test velocity  versus equivalent test velocity

-------
 regarded  as  significant.   Figure 48 shows clearly acceptable data.  The
 weights were placed  and  removed after each reading, so a standard test
 for  hysteresis was not performed.  Hysteresis type effects can be observed
 for  the lower loadings,  however, where the force factor appeared higher
 for  decreasing loadings  than for increasing loadings.  This observation
 reinforces the previous  indications that a small amount of hysteresis
 is present.
     The  drag coefficient of the disc and exposed portion of the lever
 arm  were  calculated  to verify the factory supplied value of the assembly
 based only on disc area  of CD = 1.275.  We have for a moment balance
 (see Figure 49):
                       AD *D = CD  AD *D + CD  AA *A                (62)
or
                                         AA £A
                                         AD £D                     (63)
     For the applicable Reynolds number range, "Fluid Dynamic Drag,"
by S. F. Hoerner(Reference 7),  published by the author,  gives
                              V1-17
                              C   -,.0
Physical  measurements of the instrument  give
                             AA =  6.6  cm2
                            AD  =  33.13  cm2
                                 -13]-

-------
  STRAIN
    GAGE
  BRIDGE
^—'
                                            dp = DISC DIAMETER

                                            dA = ARM DIAMETER

                                            hA = EXPOSED  ARM LENGTH
                                            Up = DISC MOMENT ARM

                                            t. = ARM MOMENT ARM
FT
*
I.
I

\
«
*[
A

                                                     MOMENT ABOUT
                                                     STRAIN
                                              " FD *D + Fa  "'A


                                              = I P. U! 
                                            WHERE:  AQ = } (dp)2
                                                    AA = hA dA
Figure  49.  Moment applied at strain gauge  bridge of  Ramapo Mark V
                                   -132-

-------
So we have

                     Cn  =1.17+1.0
                      UF
                         = 1.17 + .10

                         = 1.27

which agrees with the factory experimental  data within the accuracy of
the calculation.

7-3-3  Hastings Raydist Flare Gas Flow Probe Calibration
     Test calibration and factory calibration data are shown for the
Hastings probe in Figure 50.   Accuracy is shown in Figure 51.   The solid
line in Figure 51 represents  the factory stated accuracy, which is based
on a tolerance of +0.1  volts  on the  output  voltage.   The accuracies of all
data points shown are referenced to the experimental  data curve fit in
Figure 50. when an additional 1% tolerance  is added for the accuracy of
the reference measurement system, all but four of the 104 test data
points are within tolerance,  which is very  consistent with a 20 accuracy
for the instrument (i.e., 95% or more of the data points fall  within the
specified tolerance).  It is  clear that the factory data does  not match
the tolerance below  15 m/sec.   The  factory was consulted several  times
and extra test data were obtained to try to resolve the discrepancies,
which are best illustrated in Figure 52.  Here are shown typical  factory
calibration curves for these  probes:  models IK,  6K,  and 10K,  for maximum
speeds of 304.8 m/sec,  1828.8 m/sec, and 3048 m/sec,  respectively.  The test
probe was a 10K.  Consultation with  the factory revealed that  a 10K
probe is identical from a hardware standpoint to  a 6K probe, the difference
being an electrical span change.  The figure shows that the 6K output is
always less than the IK output for a given  speed.  Considering the physical
similarity of a 6K probe and a  10K probe, it would certainly be feasible
that the 10K output would be correspondingly less than the 6K output
for a given speed.  Experimental data confirm this hypothesis, but the
                                -133-

-------
              OUTPUT VOLTAGE, VOLTS
               5
U)
^
I
               1.5
                                                              FACTORY DATA
    •  DATA POINTS

 	  CURVE FIT
                                            10
15
               20
                                                                                                   5-7
                                                                                                   5-8
                                                                                                   5-9
                                                                                                   5-15
                                                                                                   6-7
                                                                                                   6-7
                                                                                                   6-8

                                                                                                CURVE FIT
                                                                                        FLAG ON LEFT--V INCREASING
                                                                                        FLAG ON RIGHT--V DECREASING
25
                                                                                                      30
                                                        Standard Velocity,  v$, METERS/SEC
                   Figure  50.  Factory  and test calibration  curves  for  Hastings-Raydist AFI-10K
                                probe as  probe  output  voltage versus standard  velocity

-------
00
en
               36
               32
               28
               24
               20
               16
               12
                       V - V
                           ref
                               x 100%
                         "ref
TEST DATA POINTS

FACTORY DATA POINTS

FACTORY STATED ACCURACY
                                                        15           20

                                                  STANDARD VELOCITY, v_. METERS/SEC
                                                                                                            35
                  Figure  51.  Absolute difference in  percent between velocity data  points and  reference
                               curve  fit velocity versus  standard  velocity  for Hastings probe

-------
                VOLTAGE OUTPUT, VOLTS
CO
en
i
            2   -
             1
                                                                                                     AFI-6K
                                                    STANDARD VELOCITY, vg, METERS/SEC
                          Figure 52.   Comparison of  calibration curves for three  Hastings probes and
                                       test data for  AFI-10K probe
                                                                                                                   15

-------
factory 10K curve is significantly above the 6K curve at low velocities,
which leads to the possible conclusion that the factory 10K curve is in
error.  The factory did agree that the 10K curve looked "unusual" at the
low end.  The factory also suggested that the problem may have been an
electrical shift during shipment.  Such shifts have been noted by TRW
personnel in items such as pressure transducers.

    The Hastings probe tested is not suitable for accurate measurements
regardless of the test discrepancies.  This unsuitability is shown by the
factory-claimed accuracy as shown in Figure 42, and confirmed by the
scatter of the experimental data.  The loss of accuracy above 15 m/sec
is due to the flattening of the output voltage curve.  The factory has
indicated development of a system with a more accurate (velocity)
linearized output, but literature on such a system has not been received
as of this writing.
7.3.4  Thermo Systems Hot Film Sensor Calibration

     Test calibrations were performed using the test sensors purchased
for the program and TSI anemometers already owned by TRW.  Calibration
data were put in the form which would have been obtained if a VT161 type
anemometer had been used.
     Typical calibration curves are shown in Figures 53  to 55.   Since
only part of a system was purchased, there are no factory calibration
curves.   Figures on accuracy were obtained by curve fitting the test
data points and then determining the accuracy of each point relative to
the curve fit.   Results are shown in Table 21.  The metal  clad  sensor
clearly showed the best repeatability,  and results for all  three sensors
are better than the factory claimed accuracy of 5% for the  linearized
system.

7.3.5  Ellison Annubar Calibration
     Since the Annubar is inherently an averaging instrument, calibration
in a uniform flow is not a true measure of accuracy.  Consequently,
calibration testing could not include determination of accuracy, and had
to be limited to repeatability, shift, hysteresis, etc.   The calibration
                                -137-

-------
         .3 r- SENSOR POWER SQUARED,

              P2, WATTS2
LO
00
I
                                      JO
15            20

 VELOCITY,  METERS/SEC
25
                                                                                             30
                            35
                   Figure 53.   Calibration of TSI metal  clad sensor  as  power squared  versus velocity

-------
       .04
SENSOR POWER SQUARED,
P2,  WATTS2
       .03
       .02
00
UD
I
                                                SYMBOL
                                                  D
                                                  O
                                                  A
                                            DATE
                                            7-25
                                            7-26
                                            7-27
       .01
                                      _L
                                      _L
                                                                  _L
                                                             _L
                                                                                25
                                                                                30
     5             10            15            20
                                 VELOCITY, METERS/SEC
Figure  54.   Calibration  of TSI metal  backed sensor as power  squared versus velocity
                                                                                                              35

-------
      .04  r SENSOR POWER SQUARED,
            P2, WATTS2
      .03
      .02
o
I
      .01
                SYMBOL
                  D
                  O
                  A
DATE
7-25
7-26
7-27
1
1
	 1 	 1 	 1 	 1 	 1
                                     1Q
              15
                 25
                               30
                                                                                                           35
                        Figure 55,
               VELOCITY, METERS/SEC
Calibration of TSI wedge  sensor as  power squared  versus velocity

-------
            Table 21.   TSI  HOT  FILM SENSOR  CALIBRATION SUMMARY
                                    V  -  V
                        Error =  E  =
                                         ref
                                      V
x 100%
                                       ref
V = test velocity
V  f = reference velocity from curve fit
                                a =
n = total number of data points
SENSOR
METAL CLAD

METAL BACKED

WEDGE
max
%
2.7
@ V = 3.66 m/sec
5.2
@ V = 3.66 m/sec
9.4
a V = 3.66 m/sec
at
%
0.6

1.6

2.1
2ot
%
1.2

3.2

4.2
tAdjusted to compensate for ±1% uncertainty in reference value.
                                  -141-

-------
consisted of obtaining an effective calibration factor for the instrument
in a uniform flow, as was done with the "S" pi tot probe:

                                   kAl'coV                        (3D

                           Ap.  =  Annubar reading

                         k^ =  Annubar calibration  factor

                        2" p<» ^«2  =  dynamic pressure

  Test  results are shown  in  Figure 56.   Repeatability  is shown to be good,
  and the pitot factor was  reasonably constant for  data 6  m/sec or  above.
  7.4   STABILITY
       This  test was  performed with each  test  sensor and a reference pitot
  probe mounted in the wind  tunnel.  The  pitot probe output from the
  Baratron pressure transducer was fed  as  a voltage to a Hewlett Packard
  digital voltmeter and printer.  Outputs  from each of the test sensors
  were  handled in  the same manner.  The wind tunnel was turned on,  and
  then  the printers were  turned on at a  known  sampling rate.  These runs were
  made  for each test  sensor  at nominal  velocities of 6, 17, and 30  m/sec.
  Duration was approximately fifteen minutes for each  run.  Fifteen groups
  of fifteen data  points  were then selected at random  for  each sensor for
  evaluation for each run.   Each  group of  fifteen data points was averaged
  and the velocity computed  to give fifteen pairs of velocity per run.
  The ratios of the test  sensor velocity  calculation to pitot probe
  velocith calculation for  the fifteen  points were  normalized so that the
  instrument stability could be evaluated.  A  typical  data reduction sheet
  is shown in Figure  57.  Instrument stability is shown by the variation in
  normalized velocity N.  Results are shown in Table 22  for the sensors
  tested.  Results are considered adequate for all  probes.  Total system
  accuracy is not  high enough  to  allow  quantitative comment on individual
  readings.  All  instruments should be  considered  to have  1% or better
  short term stability over the range tested.
                                  -142-

-------
1.60 ^
      CALIBRATION FACTOR,

1.58
1.54
1.50
1.46
i
CO
' 1.42
1.38
1.34
1 30
u 0
k - ....
A 1 .. 2
7 P v
f. 00 00 -£
A- -**- ^o  - 5.8 m/sec
, . - ' CALIBRATION RUN
(kAj - 1 .531
AVE O 1
MAXIMUM DEVIATIONS IN A 2
TERMS OF VELOCITY:
V 3
-A +1.46%
-1-94% TAG ON LEFT— V INCREASING
4_ TAG ON RIGHT — V DECREASING
-11,1,,
5 10 15 20 25 30
VELOCITY, METERS/SEC
Figure 56. Annubar calibration factor

-------
          Figure 57.  Ramapo probe stability data reduction sheet
6/4/73  Stability Data Reduction
        Ramapo
Speed Control 20
Sample
1
2
3
4

5
6
7
8
9
10
11
12
13
14
15
Ra
.2238
.2247
.2233
.2253

.2247
.2239
.2251
.2237
.2238
.2227
.2233
.2240
.2237
.2253
.2236
Pa
4.2779
4.2695
4.2877
4.2653

4.2778
4.2824
4.2686
4.2727
4.2754
4.2631
4.2774
4.2879
4.2627
4.2736
4.2748
Pa
.2287
.2294
.2282
.2298

.2292
.2287
.2296
.2268
.2268
.2286
.2285
.2286
.2291
.2296
.2287
N
1.0000
1.0030
.9977 <- Min >
1.0048 «- Max /

1.0020
.9997
1.0040
1.0004
1.0003
.9993
.9989
.9993
1.0016
1.0038
.9999



Total
Variation
.71%











            Total run time:  13.8 min
                                 -144-

-------
                                          Table 22.  STABILITY  TEST  RESULTS
trt
PROBE
"S" PITOT
PROBE
ANNUBAR
HASTINGS
RAMAPO
TSI SENSORS
METAL CLAD
METAL
BACKED
WEDGE
TOTAL VARIATION DURING RUN, %
V = 6.10 m/sec
CO
.76
.85
1.54
1.08

0.4
0.4
0.8
V =16.76m/sec
oo
.31
.57
.48
.71

0.5
1-9*
*
4.2
Vm =29.87 m/sec
.22
.47
.62
.77

*
4.0
*
1.3
RUN TIME,
MINUTES
19, 17, 8
13, 12, 13
15, 14, 14
15, 14, 14

22, 19, 19
15, 17, 12
16, 23, -
                    Jf
                     Shifts due to uncompensated temperature variation in freestream.
                     Variation noted for 6.10 m/sec wpuld  be representative  of variation  at
                     higher speed for standard complete factory unit

-------
7.5  TIME RESPONSE
     The setup for each probe was the same as for the stability test.  A
typical run proceeded as follows:  the wind tunnel was started with the
test probe aligned with the flow.  The printer was turned on.  The probe
was turned an angle of 90° to the flow.  It was then quickly turned back
to its normal position (nominal time required = .5 seconds) and the
printer ran until the reading stabilized at its normal value.  This
procedure applied to all probes except the hot film sensors, for which
the response time was determined by covering the sensors and then suddenly
exposing them to the flow.  The response time was then calculated from
the known printer speed for all sensors.  Results are shown in Table  23.
All probes except the Hastings-Raydtst may be considered to have a one
second or better response time.  The Hastings-Raydist time response
acceptability must be determined in individual applications.  It may not
be a drawback for long term continuous monitoring, particularly if flow
conditions are steady.
7.6  SEN S FT IV ITY TO OR IOTAT K)N
     It was  shown in Sectton 4 that the component of velocity parallel to
the duct axis is the most mathematically desirable output for a point
velocity sensor  to have.  This component can be obtained by aligning
the probe directly with the local velocity vector and computing the
axial  component  from the measured total velocity and the angle between
the probe and the duct axis.  This approach is highly impractical under
most circumstances.  A much easier approach is to always align the probe
with respect  to  the duct axis, and have a probe which responds to the
axial  velocity component.  The purpose of this test was to determine
how well the  probes being evaluated responded to this component of
the flow.
     For each case, the probe was rotated with respect to the flow,
rotation being performed about the probe's principal axis since it
was not feasible to rotate the flow itself.   Probes such as the S probe
have two independent axes, as shown in Figure  58, but rotation about the
tilt axis would  have required wind tunnel modifications beyond the scope
of the program.  Runs were made for each sensor at nominal speeds of

                                 -146-

-------
      Table 23.  RESPONSE TIME TEST RESULTS
PROBE
"S" PITOT PROBE
100% RESPONSE
ANNUBAR
100% RESPONSE
HASTINGS
90% RESPONSE
100% RESPONSE
RAMAPO
100% RESPONSE
TSI PROBES
METAL CLAD
METAL BACKED
WEDGE
RESPONSE TIME, SECONDS
V =36.58 m/sec
CD
1.0
0.5
20.8
56.9
1.3

0.2
0.8
0.4
V =16.76 m/sec
oo
0.7
0.2
12.3
24.1
0.6

^0
0.6
0
V =29.87 m/sec
oo
0.8
0.2
8.5
18.1
0.4

^0
0.2
0.1
Time response for change from V  ^  0
                      -147-

-------
                                    6, AXIAL ROTATION

                                       ANGLE
          \.
           \
                                     4, JILT ANGLE
Figure 58.   S  probe orientation angles
                  -148-

-------
6, 17, and 30 m/sec.   The resulting  velocity measurements were
normalized with respect to the measurement at 9  =  0°  and plotted.
     Results are shown in Figures 59-65.   As a reference, cos 0  is  also
plotted.   A summary of results is given in Table 24.   Results indicate
that all  probes should be acceptable for use in  nonuniform  flows with
flow angularities not over 10° in the plane tested.
7.7  ENVIRONMENTAL TESTING
7.7.1  Test Sequence
     Environmental testing consisted of three phases:
     1.  Four hour exposure in duct  at 180°C and 10 m/sec,  ambient
         humidity.
     2.  Four hour exposure in duct  at no°C and 16 m/sec,  humidity
         100 percent + droplet laden.
     3.  Exposure to 30kg of quartz  sand at 40 m/sec,  approximately
         equivalent to four days of  exposure in  a  stream at 16 m/sec
         with a grain loading of Q.Q5 gm/s,cm.
None of the hot film probes was exposed to the sand due to  the  large
particle size (average size«300 microns), and one of the hot film  probes
was not exposed to the high temperature stream due to temperature limitations
of the sensor supports.  Particulate loading was accomplished through  use
of a portable sandblaster in a setup outside the process simulator,
since local air pollution control officials have ruled against  the  use of
particulate material in the process  simulator.
7.7.2  Test Results
     Each of the probes showed physical corrosion  and abrasion,  but not  at
a high enough level to impair performance except for the hot film sensors.
Corrosion and abrasion over a period of months or years could possibly
affect the performance of all instruments tested.   The Hastings  and
Ellison instruments showed no calibration change after environmental
testing.  The Ramapo probe showed approximately  an 80 percent drop  in
output.  The Ramapo factory was consulted on the problem, and said  the
probe was probably clogged internally.  The probe was dismantled in

                                 -749-

-------
            = velocity @ 0=0C
l.OC
1.04
1.02
1.00
 .98
 .96
 .94
 .92
 .90
 .88
 .86
   •	   6 m/sec
          17 m/sec
— n	  30 m/sec
 .84
      -50    -40     -30    -20     -10      0     10     20      30     40
                                                           9. DEGREES
             Figure 59.   S pitot  probe  orientation sensitivity  data
                                      -150-
                   50

-------
1.02
1.00
            VQ = VELOCITY AT 6 = 0°
.78  -
   -40
                                                       V = 6 m/sec
                                           V  =17 m/sec
                                            oo

                                           V  =30 m/sec	D	
-30      -20
Figure 60.
                                  10
20
                                                  30      40
                           -10       0
                            6, DEGREES
                         Ramapo probe orientation  sensitivity data
                         50
                                     -151-

-------
                            VQ  = VELOCITY AT  e = 0
                             8  = FLOW ANGULARITY
1.00  -
                                8, DEGREES

     Figure 61.  Hastings-Raydist probe  orientation sensitivity data
                                    -152-

-------
1.01

1.00



 .98



 .96


 .94


 .92



 .90
V/Vr
 .86
.84
.82
.80
        V  =  6 m/sec
         00
        V  =17 m/sec
         00
        V = 30 m/sec
,00-CK
                                      I
           -30
-20
                      -10
           20
                         0        10
                          9 , DEGREES
Figure 62.   Annubar orientation sensitivity data
30
40
                                 -153-

-------
1.00
 .96
 .92
      V  = VELOCITY AT
           6,  $ = 0°
 .84
.80
.76
.72
.68
.64
.60
  -60
                       CASE A
                              SENSOR
                                                                 NORMALIZED SENSOR
                                                                 OUTPUT. CASE B
                                                         NORMALIZED  SENSOR
                                                         OUTPUT, CASE A
                                                      cose,
                                                      COScp
        CASE A DATA  FROM VIKING HOT FILM SENSOR--GEOMETRICALLY
        SIMILAR TO METAL CLAD  SENSOR
                          _L
•O
                                                        CASE B
                                                                       J
-50    -40    -30
                            -20    -10     0      10

                                 6, <(,, DEGREES
                                              20
         30
                                                          40
                      50    60
        Figure 63.   TSI  metal clad sensor orientation  sensitivity  as
                      normalized  velocity versus  orientation  angle
                                      -154-

-------
1.00
 .98
 .96
 .94
 .92
 .90
 .88
.86
.84
.82
.80
       Vn = VELOCITY AT

        0  6=0°
                                            OO
O  V^ = 6 m/sec


A  V  = 17 m/sec
    oo

17  V  =30 m/sec
       -40     -30
                                                                30       40
     Figure  64.   TSI metal  backed sensor  orientation  sensitivity as

                  normalized velocity versus  orientation  angle
                                  -155-

-------
.96  -
.94  -
.92   -
.90   -
.88
                                                         O  O
                                                       O  V  =  6 m/sec
                                                           CO

                                                   O   A  v  = 17 m/sec
                                                           IB

                                                       V  V  = 30 m/sec
                                                O
               -30       -20
-10       0


  6, DEGREES
10      20      30
                                                                        40
   Figure 65.  TSI wedge sensor orientation sensitivity  as

               normalized velocity  versus orientation  angle
                                -156-

-------
   Table  24.  ORIENTATION  SENSITIVITY  TEST  RESULTS
PROBE
"S" PITOT
PROBE
ANNUBAR
HASTINGS
RAMAPO
TSI SENSORS
METAL CLAD
AXIAL (e)
TILT M
METAL
BACKED
WEDGE
ANGULAR RANGES FOR STATED
MEASUREMENT ACCURACY
ACCURACY:









IX
X
±8°
±8°
±6°

+17°
+8°
+2°
+2°
4%
±14°
±40°
±17°
±12°

+43°
+16°
+ 24°
+8°
10%
±28°
±40°
±26°
±18°

>+60°
+25°
+ 32°
+16°
  ANGULAR VALUES ARE FOR AXIAL ROTATION ANGLE 6
  UNLESS OTHERWISE NOTED
PROBE AXIS
INTO PAGE
FLOW INTO
PAGE
                      -157-

-------
accordance with factory instructions and approximately one gram of sand
was removed.  The probe was then put back together and functioned
normally.  The conclusion is that the Ramapo probe definitely requires
an adequate purge system when used in particle-laden flows.  Both the
Annubar and "S" pitot probe had a net accumulation of particulate.  In
the Annubar, which was horizontal, an equilibrium point was reached
where the probe became self purging and performance was not affected.
The "S" pitot probe would generally require an occasional  purge to prevent
clogging if used for long periods of time.
     Results of the post-environmental calibration are shown in Table 25.
The Hastings data, also shown in Figure 66, showed two points slightly
out of spec when a 1% uncertainty in the reference velocity is taken
into consideration.  The shifts were in the direction of the original
factory calibration, and so are considered not critical, especially in
view of the small magnitude.  The Ramapo post-environmental calibration
was the weight calibration described above and shown in Figure 48.  It
is completely acceptable.  Recall from the last page that the Ramapo
probe clogged during sand testing, and had to be cleaned before it
could be used again.  The factory makes purged probes, as illustrated
in Figure  67, and a probe such as that should have been ordered originally.
No difficulties would be expected with a purged system.  The "S" pitot
probe showed no change in calibration factor, and the Annubar pitot
factor change was within normal repeatability limits.
     Results for the TSI sensors are shown in Figures 68 to 70.  The
test sequence was as follows:
     1.  Each probe was calibrated before environmental testing.
     2.  Environmental testing took place.
     3.  The probes were recalibrated.
     4.  The probes were cleaned in a dilute nitric acid solution and
         a  third calibration was performed.
     Examination of the curves shows that for each case the power required
to maintain overheat went down for the post-environmental tests.  It would
appear that in each case the exposure to high temperature, humidity

                                 -158-

-------
Table 25.  POST ENVIRONMENTAL TEST CALIBRATION RESULTS
INSTRUMENT
HASTINGS
AFI-10K
RAMAPO
MARK V
ELLISON
ANNUBAR
"S" PITOT
TSI SENSORS
RESULTS
90% OF DATA POINTS WITHIN SPECIFICATION.
OTHER POINTS NOT MORE THAN 0.6% OUT OF
SPEC (FIGURE 22).
ALL POINTS ABOVE 3.56 m/sec IN SPEC AFTER
PROBE CLEANED. AMER
% CHANGE IN VELOCITY FROM CHANGE IN
PITOT FACTOR
0.7
0.0
UNACCEPTABLE SHIFTS (FIGURES 24 TO 26).
                        -159-

-------
en
o
i
        32
         28
         24
         20
         16
         12
V -  V
                   ref
                       x 100%
                'ref
                                            O   POST ENVIRONMENTAL TEST DATA POINTS
                                                   FACTORY STATED ACCURACY
                                                                                O
                                                                        O  O
                                                               o
                                                                                        o  o
                                                                                                 o
o
o
	 *-o
0 0
o 8
	 1 	 1 	 . 	 1 	 1 	
                                        10
                                        15
                                                                   20
                                                                   25
                                                                                               30
                                                  STANDARD VELOCITY, Vs,  METERS/SEC
          Figure 66.
                                     Accuracy  of Hastings AFI-10K as percent  difference  between

                                     test data points and calibration curve fit versus velocity
35

-------
                          RAiMAPO  INSTRLJMtlNT COMPANY, 11VC.
                              BLOOMINGDALE         NEW JERSEY
                                                            JUNCTION BOX %" SIZE
                                                            P.N. VTC-2203 OR EQUAL
                                                          	TERMINAL STRIP ENCLOSED
                              PURGE
                              CONNECTION"
                              '/." TUBE, .	.
                              (FLARELESSll
*SHOWN 90° TO ACTUAL LOCATION
                     PIPE
                                                                     8" • 150 ASA
                                                                    .BLIND FLANGE

                                                                     BOLT HOLES STRADDLE
                                                                        C_PIPE
                                                          MODEL:
                                                          LINE SIZE:
                                                          DIM. "A":
                                                    FLOW,
                           Figure 67.   Ramapo purged probe

                                          -161-

-------
I

PO
            POWER2, WATTS2
                      O  PRE-ENVIORNMENTAL

                          POST-ENVIRONMENTAL
                          POST-ENVIRONMENTAL
                          AFTER CLEANING
                                                                                          28% SHIFT
                                                                                               62% SHIFT
                                10
         15
20
25
30
35
                     Figure 68.
                  VELOCITY,  METERS/SEC
Pre- and post-environmental  calibrations of TSI  metal  clas sensor
as sensor power squares versus sensitivity

-------
en
CO
i
              POWER2, WATTS2
            ,04
            .03
            .01
O  PRE-ENVIRONMENTAL


A  POST-ENVIRONMENTAL


D  POST-ENVIRONMENTAL
    AFTER  CLEANING
                                                                       80% SHIFT
                                  10
                       15
20
25
30
35
                                                  VELOCITY,  METERS/SEC
                      Figure 69.
             Pre- and post-environmental  calibrations of TSI metal  backed sensor
             as sensor power squared versus velocity

-------
.04
POWER2, WATTS2

       O   PRE-ENVIRONMENTAL

       A   POST-ENVIRONMENTAL

       D   POST-ENVIRONMENTAL
            AFTER CLEANING
,03
                                                                             21% SHIFT
                                                                                 31% SHIFT
                                                               30
                                                                     35
                                           VELOCITY,  METERS/SEC
               Figure 70.   Pre-  and post-environmental  calibration  of  TSI wedge  sensor
                           as  sensor power squared  versus  velocity

-------
and the amounts of dust present in the process  simulator resulted  in  the
buildup of a film on the sensors,  the film acting  as  insulation  so that
less heat transfer to the air took place.   Cleaning slightly  alleviated
the problem for the wedge sensor,  but worsened  it  for the other  two
sensors.  Shifts were very small  at the lower velocities, which  is clearly
not acceptable.  Also, the cleaning method used was not adequate,  but
even proper methods would have to be used  several  times a day to keep
the sensors clean.  Unless an automatic cleaning system could be devised,
the frequency of maintenance would be unacceptable.  It definitely appears
that additional work is needed to make hot film sensors acceptable for
continuous use in stack environments.
7.8  LABORATORY TEST SUMMARY AND FINAL EVALUATION
7.8.1  S Pitot Probe
     Calibration test results are summarized in Table  26.  Recall  that
the laboratory calibration factor disagreed with the manufacturer's
calibration factory by 3.5%.  The manufacturer, Ellison Instruments,
stated verbally that the same calibration factor is published for all
S type probes  -- they are not calibrated individually.  The discrepancy
noted is therefore probably due to errors in the original calibration
and/or variations among the manufactured probes.
     The following conclusions about probe performance were drawn  from
the accumulated test data, and necessarily apply in the strict sense
only to the probe tested, since not all S probes are alike.  It  would,
however, be non-conservative to assume that other S probes have  better
accuracy characteristics than the probe tested, without having data to
support the assumption.  The primary conclusion about the S probe is  that
it  should  be capable of providing point measurement accuracies on the
order of 10% or better for the flow conditions specified in Section 4.
Nominal expected  accuracy should  be about 5% for a properly operated  probe.
Some specific  drawbacks related to probe design are as follows:
     A.  The probe design is not  standard.
     B.  The probe is  highly sensitive to flow angularities.
     C.  Due to the  design,  the probe  is probably  subject to significant
         Reynolds number and turbulence effects.
                                  -165-

-------
     Table 26.  ELLISON INSTRUMENTS COMBINED REVERSED ("S")  PITOT
                       PROBE CALIBRATION RESULTS
                  Manufacturer's
                 Stated Accuracy*

                  Manufacturer's
                   Stated Range
Calibration Test Results
                 Overall Accuracyt

                 Test Range

                 Probe Range Based
                 on Test Results
                                     ±5% of reading


                                     Not specified
                                      ±5% of reading

                                     2..74 to 3Q.98 m/sec


                                      2.74 to 45.72  + m/sec
     5% accuracy figure includes accuracy, repeatability, short
     term zero and full scale shift, and hysteresis.   Repeatability,
     shift, and hysteresis errors negligible.
          Time Response:

          Short Term Stability:

          Accuracy of Axial
          Velocity Component
          Measurement:
                                   1  sec or better

                                   Better than 1%

                                   Accuracy, %:
±1   ± 4  ±10
                                   Maximum Angularity,   X    14    28
                                     Degrees:
t
Verbal communication—not stated in literature.

Also depends upon accuracy of readout equipment.
                                 -166-

-------
Non-standardization of design leads to variation in many parameters
such as tube curvature, spacing, whether the tubes are connected near
the end, etc.  The result is that uniformity of the calibration factor
cannot be assured unless great care is taken in the manufacturing
process to insure repeatability of all critical design parameters, and
it is highly questionable whether all of these parameters are known.
The downstream tube senses a wake pressure, since a detached wake is
formed behind the two tubes.  Wake pressure measurement has traditionally
been a difficult fluid mechanical problem in uniform streams, and so
must be considered even more difficult in general  industrial applications.
Wake pressures behind many bluff bodies are highly subject to free stream
turbulence, which is responsible for adding energy to the wake and
changing the pressure distribution around the body, and to the Reynolds
number based on the size of the body, since boundary layer characteristics
and consequently the pressure distribution around  the body, are Reynolds
number dependent.  This dependence is noted in Reference  3,   "The
Measurement of for Flow", pn p. 39. inhere it is, stated,
       "... such instruments are unfortunately not to be
       recommended, because they possess certain undesirable
       features.   In the first place, experience shows that
       for most of those so far proposed the value of k varies
       appreciably with Reynolds number, particularly in the
       lower ranges for which the instrument is primarily
       intended.   There is no need to stress the inconven-
       ience of a calibration that changes  with speed, one
       result of which is that at least two calculations
       are needed  to convert a pressure observation into
       velocity:   the first has to be made  with a  guessed
       value of k,  since the speed is initially unknown;
       and the  approximate value of the speed thus  obtained
       has to be  used to obtain, from the calibration curve
       of the instrument,  a  nearer approach to  the  true value
       of k to  be used in  the second,  a more exact  calculation.
                                -167-

-------
            "Another serious disadvantage of such instruments
       is that if k varies appreciably with Reynolds number,
       its value at a given Reynolds number is likely to
       depend much more than that of a standard pitot-static
       tube on the amount of turbulence present in the
       airstream, so that the calibration may be appreciably
       different, even at the same speed, in different pipe
       systems."
The calibration shift which occurred near 6 m/sec should be attributed
to Reynolds number effects.  Turbulence effects could not be investi-
gated with the laminar flow wind tunnel facility used during the test.
     Since probe to probe design variations are difficult to control
due to ambiguity about the relative importance of the numerous parameters
involved, it is evident that for accurate work, probes must be calibrated
individually over the expected flow range as a function of Reynolds
number based on a standard probe dimension.  If the probe must be used
in angular flows, the calibration becomes much more difficult, however.
The problem is that the probe's orientation sensitivity makes such a
calibration highly impractical, because three independent variables
would be involved:  Reynolds number and two orientation angles as shown
in Figure  58.  Investigation of probe response as a function of the tilt
angle has not been performed, but until it is, there is no reason to  assume
that the response would be better than for the investigated axial rotation.
Any calibration of this nature would be prohibitively expensive to
perform.
     On the basis of existing data, it must be concluded that use
of the S probe in angular flows will  result in systematically high
readings due to the probe's angular response characteristics, which were
worse than those of the other probes  tested except for the wedge hot  film
sensor.  This error could be reduced  through use  of an artifically low
pitot factor.   By coincidence, the factory supplied calibration factor
does in fact produce a low pitot factor which was used during the first
field demonstration.
                                -168-

-------
    As a result of investigation of the S probe, it is being recommended
that a standard pitot static probe be used wherever possible to perform
standard velocity traverses.  To justify this, pitot-static probe charac-
teristics, which were not investigated during the program since adequate
characterization has been performed and documented elsewhere, are
presented at the end of this task, in Section 7.8.6.
7.8.2  Ramapo Mark V Drag Meter
    This was without question the most accurate point sensor tested.  The
one significant problem discovered was susceptibility to clogging by
particulate, as explained in the Environmental Test section.  This prob-
lem has apparently since been solved through the introduction of the
Mark VI flowmeter, shown in Figures 71 and  72,  with no loss of accuracy.
The instrument is bidirectional, and calibration tests showed it to be
free of Reynolds number effects except in highly angular flows.  The
main reason that this instrument is free of such affects while the S
probe is not is that the target disc is flat with relatively sharp edges,
which means that boundary layer separation always occur precisely at the
edge of the disc, and so boundary layer characteristics do not influence
performance in most cases.   Page 3-15 in Reference 7  shows that the
drag coefficient is constant above Re £ 3 x 10.   For the unit tested,
a velocity of about 3 m/s corresponded to Re £ 1.2 x 104.
    The instrument also has attractive calibration features for indus-
trial use.  The factory manual  gives procedures for resistance calibration,
requiring only an additional decade box, and it is also possible to
calibrate the unit with actual  weights, as described in the calibration
section.   These techniques  eliminate the need for calibration in an
actual  air stream.It should be noted that the factory recommends that
the probe be turned 180° from the way ft was tested in this program
for normal use.  The orientation used for testing was selected so that
wind tunnel data could be directly correlated with weight testing, and
weight testing could be performed most easily in  the direction chosen.
                                 -169-

-------
           Xj>^wK_>O<>DOO<->O<>	
           ^^ysXaK^Xa&f^OQwgr^                               	-. — ——	..
           ^^^^gggggg^^&^g:	         	._-_                  	.	    ..-
MACOPIN  ROAD. BLOOMINGDALE, NEW JERSEY 07403/201  - 838-2300 /TELEX: 13-8892
                               MARK VI FLOW METER
     Developed specifically for the precise measurement of flow in dirty
               liquids, slurries and high-melt or crystalizing fluids.
The  Ramapo Mark  VI  Flow Meter incorporates  the
basic features of the popular Mark V target type meter
in a  flow meter which  has no moving parts and no
cavities to trap particles and cause  erroneous read-
ings. The  Mark VI uses a bonded strain gage bridge to
translate the fluid forces on the target into an electrical
signal proportional to flow rate squared. It has a useful
and practical 10 to 1 range ratio and can be calibrated
for accuracies to 0.5%. Uni-directional or bi-directional
flows of almost any liquid or gas can be measured from
1 GPM to unlimited ranges.


                  FEATURES:

• No cavities or pressure ports in the flow stream

• No rotating parts, diaphragms, or bearing surfaces
  to wear or require servicing and maintenance

• The Mark VI probe mounts directly into a tee,
  or thru a branch fabricated on an existing line

• The Mark VI can be mounted in any position
  without changing calibration

• Installation i* simple and inexpensive

• Mark VI probes are lightweight and can be easily
  installed by one man. (12 kg. for typical 24" unit)

• Range change can be done by simply
  replacing the target

* Usable with liquids, gases, slurries, high melting
  point liquids, food and dairy products at temperatures
  ranging from -200°C to +400"C
• Field calibration is simple;
  no flow checks are required
          ACCESSORIES AVAILABLE

A complete line of transmitters, analog or digital indi-
cators, recorders, controllers and totalizers can be pro-
vided for most system requirements.
          PRINCIPLE OF OPERATION

The Mark V Flow meter measures flow in terms of dy-
namic forces acting on a fixed body in the flow stream:

                Force = C«Ai Vf
                           29
Cj = Drag Coefficient, A = Sensor Area, f = Fluid Den-
sity, V = Velocity. Bonded strain gages in a bridge cir-
cuit outside the fluid stream and shielded by stainless
steel, translate this force into an electrical output  pro-
portional to the flow rate squared.
                               Figure  71.    Ramapo  Mark  VI

                                              -170-

-------
         ELECTRICAL SPECIFICATIONS

Input: 5 to 10 volts A.C. or D.C.
Output: 2.0  mv/v or 20 mv maximum at full scale pro-
portional to velocity squared;  higher or lower outputs
may  be used to meet the requirements  of  specific
applications
0-10 VDC; 1-5, 4-20, 10-50 made outputs available. See
Bulletin SG-1A

Bridge Resistance: 350 ohm standard (120 ohm optional)
Electrical Connections: Barrier type terminal strip in
junction box with %" conduit connection
          GENERAL SPECIFICATIONS

Repeatability: ±0.1% of reading
Accuracy: ±0.5% of full scale How rate with flow cali-
bration

Pressure: to 5000 psi (35.000 KM/m2)—limited by flange
rating

Fluid Temperature: -65 to +350°F (-54° to +177-C) to
-320«F (-195°C) and 4-750-F (400°C) available

Sires: %" to 60"—larger si2es available
Ranges: Any 10101 flow range (maximum recommended
lull scale velocity—15 feet/second water flow equiva-
lenl) [minimum recommended full scale velocity—3
feet/second  water How equivalent]

Materials: 17-4 PH and Type 304 stainless element with
carbon steel  (large and  line  housing standard; Type
316 stainless, Hastelloy C, and Inconel available.
                                                                                 V* CONDUIT SIZE
                                                                                   -FLOW-
Norn.
Pipe
Size
In.
tt
1
IVi
m
2
3
4
6
8
to
18
20
and
larger
Flow Ranges >
Minimum
CPU
.8 - 6
1.5 -15
2 - 20
3 - 30
5 - 50
10 - 100
20 - 200
30 - 300
dm'/tec.
.05- .5
.10- 1.0
.13-1.3
.19-1.9
.30 - 3.0
.63 - 6.3
1.3-13
I 1.9-19
Minimum Recommended
Velocity Range
.3-3fl./sec.
(.1-1 m./sec.)
Maximum
GPM
4- 40
8- 80
10-100
12-120
20 - 200
30-300
60-600
120-1200
dm' /tec.
.25 - 2.5
.50 • 5.0
.63-8.3
.76-7.6
1.3-13
1.9-19
3.8-38
7.6-76
Maximum Recommended
Velocity Range
1.5-15 ft, /sec.
(.45-4.5 m./sec.)
Full Seal*
Prest. Lost '
ptl
15
10
8
6
3
2
1
0.5
.4
and
lower
kN/m>
105
70
56
42
21
14
7
4
3
and
lower
f »
If.
o ™.H
Su.cn
V,
1
1V4
1%
2
3
4
4
4
3
Dim. <
A
3Vi
3V2
3V<
4
4Vi
5V4
6
6

 1  Customer may select any 10 to 1 range desired between the minimum and maximum ranges shown.
   Range in SCFM of air at standard conditions may be estimated at approximately 4 x water flow rate In GPM.
 1  Approximate values—lower pressure loss available.
 3  Mounting flange size may be varied to suit the  requirements of specific applications. ANSI ratings as required.
 '  Plus gasket thickness.
                  Figure  72,   Ramapo  Mark  VI  specification  sheet
                                                -171-

-------
     Final  test results are summarized in Table 27.  The factory modifi-
cation since completion of the test work eliminates the only serious
fault of the Mark V unit.
7.8.3 'Hastings/Raydist Flare Gas Flow Probe
     Unfortunately for this probe, the factory accuracy specifications
speak very clearly for themselves when stated in the right way.   The
factory tolerance of +. .1  volt on the instrument output, which was for
the most part confirmed by the laboratory tests, translates to a very
unacceptable velocity accuracy, as was shown in Figure 46.  There is  no
doubt that the instrument has excellent survival capabilities, but the
same can now be said of the Ramapo Mark VI, which is much more accurate
than the Hastings unit.
     Conversations with the factory have revealed the intention to
introduce a model with improved electronics, but literature on such a
model has not been received as of this writing.  One very noteworthy
Hastings accomplishment has been made since the end of the laboratory
testing, and that is the integrated "suitcase" system shown in Figures  74
and 75.  As will be discussed later in the report, no true measurement
system was tested -- the velocity sensors all required significant
amounts of auxiliary support equipment.  This integrated approach used
by Hastings is to be commended and should be more widely adopted by
the instrumentation industry.
     Results are summarized in Table  28.  The model tested was clearly
acceptable from a survivability standpoint, but just as clearly unaccep-
table from an accuracy standpoint.
7.8.4  Thermo-Systems Hot Film Sensors
     TRW Systems Group is responsible for the Viking Meteorological
Instrument System which will be taking wind speed and direction measure-
ments on Mars starting July 4, 1976.  The system uses hot film sensors
for the measurements.  On the basis of that work, it was hoped that
hot film sensor technology would be applicable to this program, since
specifications required instrument survivability in high dust loadings
up to 150 m/sec.  Unfortunately, the sensors tested, which reasonably
represent state of the art in terms of ruggedness, failed the
                                 -172-

-------
       Table 27.  RAMAPO MARK V FLOW METER CALIBRATION RESULTS


                Manufacturer's       ,~  £     ..
                Stated Accuracy     ±1% of readin9

                Manufacturer's      -j cc <-  ^c co   /
                 Stated Range       3'66 to 36'58 m/sec


Calibration Test Results


                 Overall Accuracy*     ±1% of reading

                 Test Range         2.74 to 30.48 m/sec

                 Probe Range Based  3.66 to 36.58 m/sec
                   on Test Results


     1% accuracy figure includes accuracy,  repeatability,  short
     term zero and full scale shift, and hysteresis.  Hysteresis
     approximately ±0.5% of reading.


      Time Response:             ^1 sec or better  •

      Short Term Stability:     Better than 1%

      Accuracy of Axial         Accuracy, %:          ±1  ±4   ±10
      Velocity Component
      Measurement:               Maximum Angularity,   +6 ±12   ±18
                                  Degrees:
Also  depends  upon  accuracy of  readout  equipment.
                                -173-

-------
-WTELEDYNE
       HASTINGS-RAYDIST
Specification Sheet No. 516
HAMPTON. VIRGINIA 23661
                                 (804) 723-6b31
         HASTINGS  STACK   GAS  VELOCITY   METER
        MODEL GSM-1D5K FOR MEASURING VELOCITY OF WET AND DIRTY GASES
    FEATURES:

    • VELOCITY RANGE: 90-1500 FPM
    • PORTABLE ELECTRO-PNEUMATIC VELOCITY
      INDICATOR
    • CONTINUOUS PURGE PREVENTS ENTRY OF
      STACK GAS  INTO THE PITOT TUBE AND
      INSTRUMENT
    • OPERATES WITH A "V" OR "S" TYPE PITOT
      TUBE
    • PORTABLE UNIT SIMPLIFIES SET-UP FOR
      SPOT CHECKS  OR LONG-TERM SAMPLING
    • SELF-CONTAINED PURGE MANIFOLD AND
      PURGE SOURCE
    • NO AFTER TEST CLEANING REQUIRED
    GENERAL
      The Hastings Stack Gas Velocity Meter is the result of
    nearly two decades of experience in dealing with difficult-
    to-measure. corrosive, wet or dirty gases. A patented
    thermal principle with continuous purging is utilized so
    that line gas is prevented from entering the pitot tube.

      The Hastings Stack Gas Velocity Meter is a portable
    test set which can easily be carried to test  sites for
    periodic checks, or can be left in operation for long-
    term monitoring of stack velocities. The  instrument
    includes solid-state circuitry and a sensitive transducer
    for measuring the differential pressure across the pitot
    tube, which is related to velocity.  An internal pump
    supplies a  continuous purge of air through the pitot
    tube and into the line to prevent line gas from entering
    the instrument, and a connection is provided for using
    some other purge gas if air is not suitable. A manifold
    is also included for ease in balancing the purge. A rugged
    meter provides on-site read-out of velocity  indications,
    and a 0-1 volt d-c output is available to connect to remote
    data logging devices, meters, recorders, etc.

      A pitot tube with a 36" insertion depth is included with
    the instrument and can be interchanged with pitot tubes
    already in use in stacks or gas sampling trains.
 AIR-/   ^-EXTERNAL
INLET      PURGE
         CONNECTION
                   FLEXIBLE TUBING
       S'TYPE PITOT TUBE
    STACK GAS VELOCITY  METER
          BLOCK DIAGRAM
                       Figure 73.   Hastings stack meter
                                       -174-

-------
PRINCIPLE OF OPERATION-
CONTINUOUS PURGE MODE
   Purge gas is injected into a pneumatic bridge arrange-
ment formed by the velocity transducer, manifold, and
pitot tube. At zero line velocity the bridge is balanced so
that no flow occurs through the velocity transducer and
purge gas exhausts equally through both openings of the
pitot tube.

   As flow  across the tip occurs, a differential pressure
is developed, unbalancing the bridge and causing a small
amount of purge gas to flow through the transducer. The
transducer measures the flow which is related to the main
gas flow at the tip of the pitot tube. Purge gas still ex-
hausts through both openings, but at slightly unequal
rates.
   The purge gas continually exhausts into the main line,
thereby preventing contamination in the main line pitot
tube.
CALIBRATION
   Calibration is related to the gas density and the velocity
 profile in the stack.
   Velocity  = kj.k2.Vind where
   kj = Velocity Profile Factor; (typically .8 when probe
        is inserted to  center of stack)
   k2 = Density  Factor:  V075/8as density (Ib/tt1)'
   Vind »  Velocity from probe calibration curve
 Example: What is the full scale (1 volt) range of a GSM-
 1D5K when  measuring a stack gas having a density of
 .092 Ibs/cu.  ft.?
   Velocity  <•= kj..k2.Vind
           * (.8) A/.075/.092 '   (1500)
           = 1083 FPM
       100
D

5

z
HI ^.
u«0
       20

O          SCO         1000        1500

        AM VELOCITY (IfM)
                                                PURGEGASES
                                                  The Hastings Stack Gas Velocity Me'ter has a ouilt-in
                                                pump which pulls ambient air in through a filter to use
                                                .as a purge gas. A gauge indicates the pressure at the inlet
                                                to the purge section, and an adjustable "bleed" valve is
                                                used to adjust this pressure. A three-way valve is supplied
                                                to change the purge source from the internal pump to an
                                                external source so purge gases other than ambient air
                                                can be  used. A 1/4" NPT fitting is mounted on the front
                                                panel for connecting to an external source. The flow  rate
                                                of purge gas is normally 5-10 cfh at 15 psi.

                                                SPECIFICATIONS
                                                  INSTRUMENT
                                                     Modal: GSM - 1D5K
                                                     Rang*: 50-1500 fpm for air
                                                     Power: 115 volts. 50-60 Hz. 1.5 amp
                                                     Output: 0-1 volt d-C
                                                     Cate Dlmentlont: 9" high x 16" wide x  10" deep
                                                     Weight: 27 pounds
                                                  PITOT TUBE
                                                     Type:  S
                                                     Length: 36* Insertion
                                                  Connecting Tubing:  Twin Type. 10 ft. Length
                                                         FOR CONTINUOUS MONITORING OR
                                                           PERMANENT INSTALLATIONS:

                                                        HASTINGS GAS FLOW PROBE

                                                                MODEL AFI-SERIES
                                                         RANGE: 0-1000 fpm or 0-6000 fpm
                                                A dependable, non-clogging flowmeter for contaminated
                                                gas lines.
                                                                               REQUEST CATALOG 0913A
                                                      FEATURES
                                                       • CONTINUOUS PURGE  PRINCIPLE
                                                       • NO EXPOSED SENSORS  or WIRES
                                                       •EASILY INSTALLED  or MOVED
                                                       •EXPLOSION-PROOF TYPE HOUSING
                                                       • 05 VOLT DC OUTPUT SIGNAL
                                                       • PROVIDES LONG LINE TRANSMISSION
                                                        CAPABILITY
                                                       • REMOTE: RECORDING, CONTROL, ALARM,
                                                        INDICATION
                                                       •PURGE WITH AIR,  N, OR  PROCESS GAS
                                                       • CHOICE OF TWO RANGES:  0 1000 or 0-6000 fpm
                                                                    Litinlun tnllibli upon rtqimt:
                                                          Hutinp Vicuum Giuf n
                                                          Hittinp Mcleod Giu(i
                                                          Histinp Giu|t totx tccetsorin
                                                          Hntinp Vicuum Giuf e ftetertnct lubn
                                                          Hiitinfi Air-Mtten
                                                          Hnlinp Mm f Imrimteri for Giiei
                                                          Hlllinp Cllibrited Gil Lukl
                                                                                Cltllot No. 300
                                                                                Sp
-------
 Table 28.    HASTINGS-RAYDIST AFI-10K GAS FLOW PROBE CALIBRATION RESULTS
        Manufacturer's
        Stated Accuracy

        Manufacturer's
         Stated Range

Calibration Test Results

         Overall Accuracy
         Test Range

         Probe Range Based
         on Test Results
See Figure 42--accuracy not better
than 4.8% of reading

0 to 50.6 m/sec
   Test data set met factory
   tolerance but did not agree
   fully with factory calibration

   2.74 to 30.48 m/sec

   0  to 50.6 m/sec
     Accuracy determination includes accuracy,  repeatability,  short
     term zero and full  scale shift, and hysteresis.   Error due
     mainly to repeatability error.
       Time Response:
Short Term Stability:

Accuracy of Axial
Velocity Component
Measurement:
  57 sec @ V =6 m/sec
  24 sec 0 V =17  m/sec
  18 sec @ V =30  m/sec
                              Accuracy,  %:
                       ±1  ± 4  ±10
                              Maximum Angularity,   ±8   ±17   ±26
                              Degrees:
                                -176-

-------
environmental  tests on this program, as shown by the severe calibration
shifts in Figures  68-70.  Tfi.e culprit, which did not exist in the Martian
specifications, was high moisture content at high temperature.   This
resulted in deposition of a film on the heat transfer surfaces.   The
film acted as a thermal  insulator, reducing the power required  to
maintain overheat at a given wind speed.  Efforts used to clean  the
probes succeeded only in worsening the condition.  Even had the  cleaning
been successful, the sensors would not be acceptable for continuous
use due to the need for very frequent cleaning^

    Despite the sensors' unacceptability for continuous monitoring, it
is felt that system development as a diagnostic  tool should be  encouraged.
Thermo-Systems presently offers a total vector anemometer which  is
capable of measuring the true velocity vector -- both speed and  direction
regardless of orientation.  This instrument is the Model  1080 as shown
in Figures 75  and 76 .   Frequent cleaning requirements would not be
detrimental for diagnostic applications.  Possible applications  would
include investigation of blower or other process unit characteristics,
stack swirl properties, etc. to determine possible malfunctions  or
general process effectiveness.
    A general problem which applies to all hot film sensors should
also be considered, and that is the previously mentioned dependence on
thermal conductivity and viscosity.  In general  for hot film sensors,
the calculated velocity is proportional to viscosity and inversely
proportional  to the square of thermal conductivity.  The National
Bureau of Standards is rather reluctant to commit to uncertainties of
less than 3% and 5% for viscosity and thermal conductivity, respectively,
of gases such as air.  This means that significant uncertainties must
be expected in the calculated velocity if the measurements are  taken in
a gas significantly different from the gas in which the sensors  were
calibrated.  Thus to insure accuracy, heat transfer devices such as hot
film sensors should be calibrated in the same gas composition in which
they are to be used.
                                 -177-

-------
  tsT)    SYSTEM  1080  TOTAL  VECTOR ANEMOMETER
        PROBE TIP DETAIL
Streamlined Supports
             \
3 Mutually PerpendicularSplit-Film Sinton
                            Built-in remotely op«r«ted
                            shield and calibrator
                            in each probe
                                                                           Spatial Resolution: lest
                                                                           than 0.3" dia. sphere


      GENERAL DESCRIPTION

         The system 1080 measures the velocity magnitude and direction over the full 360° solid angle in three dimensional
         flow fields. It features solid-state electronic circuitry and a small, rugged transducer which allows fast response, wide
         dynamic range, excellent spatial resolution and high accuracy in both velocity magnitude and direction. The velocity
         probe  feature; Thermo-Systems' split-film* sensors which allows the unambiguous determination of magnitude and
         direction of the instantaneous velocity vector. Six simultaneous velocity dependent analog voltages and a 0-5 volt
         analog temperature signal comprise the system output. The probe includes a pneumatically operated shield with a
         built-in calibration feature.
                                                                                                 'Patented
     PERFORMANCE SPECIFICATIONS

     VELOCITY MEASUREMENT:

     USABLE RANGE: 0 - 500 fpc (Air, 1.0 ATM)

     STANDARD RANGE: 0 - 300 fpt (Air, 1.0 ATM!
                       0- 100 mps
     MAGNITUDE ACCURACY: + 3% of Reading and
        + 0.1% of full scale.O - 150 fps.

     DIRECTION ACCURACY: Two independent
        directional angles are within 3° over
        the complete solid angle I 4 7T
        Stand land

     RESPONSE: DC - 1 KHz

     SPATIAL RESOLUTION: Three orthogonal
        sensors fit within 0.3 inch (7.6mm)
        diameter sphere.

     TEMPERATURE MEASUREMENT:

     RANGE: 0°F to 200°F (20°C to 100°C)

     ACCURACY: + 2°F (+ 1°C)

     TIME CONSTANT: 20 millisecond! at 60 fpt
        (Air, 1.0 ATM)

• 	 64(162) -»

f t t
* t
fl «•
«»•
0 ••
*:
r
1134)
1
» 16 3BI3B1I 	 «

CONTROL CIRCUIT

                                      DATA OUTPUT
                               Fooooooooou-
                               1 ooooooooogfc
                                000060005Jo
        SERIES 1080 PROBE


(Dimensions in Brackets an in Millimeters!
                       Figure  75.   TSI  total  vector anemometer

                                                  -178-

-------
               SYSTEM  1080 TOTAL  VECTOR ANEMOMETER
                                              SPECIFICATIONS
OUTPUT:
   Velocity: There are six analog voltage signals available at the
   rear output connector as well as the front  panel pin jacks.
   These voltages constitute the data necessary  to calculate the
   instantaneous 3D velocity vector. Output range  is 0-20 volts
   DC, output impedance is  50 ohms. Output signals can be
   monitored on an oscilloscope or voltmeter. Data storage can
   be via tape recorder or oscillograph. (A 0-5VDC output range
   is optional.)

   Temperature:  A copper-constantan thermocouple - 0.002"
   (.05mm) dia.-is mounted on the probe tip.  Automatic  cold
   junction compensation is provided. The output voltage of
   0-5V DC corresponds to  a  temperature  range  of 0°F to
   200°F.  Output impedance is 25 OHMS. The  temperature
   output signal is available at the rear output connector as well
   as at a front pin jack.

PROBE  SHIELD: A pneumatically actuated  shield is standard
for all systems. The shield switch on the front panel can be used
to actuate  an external air source (not supplied) to power the
pneumatic  cylinder  on the probe. A gas supply of 15 psig or
greater is required.

GENERAL:

   Weight:

   Control  Circuit	10 Ibs.
   Probe	3 Ibs.
   Cable (15 foot)	2 Ibs.

   Power Required:  115V/220V AC ± 15%. 50-400 Hz, 1.0
     amperes.

   Circuitry: All solid state

   Size: See drawing, opposite side.

   Cabinetry:  Bench  type, etched aluminum  panels, vinyl-clad
     aluminum chassis.

   Probe Construction: All  stainless steel  anodized aluminum
   materials.

   Environmental: Control circuit will operate as specified over
   40°F to 120°F, (5°C to 50°C). Probe is  built to operate over
   temperature extremes of -50°F to 200°F f-45°C to 100°C).

   Cable length: Standard length is 15 ft. (4.6m). Cable lengths
   up to 50 ft.  can  be provided. For longer cables, please
   consult  TSI (Model 10117-15).
CALIBRATION  FEATURES: Probe  shields  have  a  built-in
orifice to provide velocity calibration. The "calibrate" switch on
the front panel can be used to actuate the calibration pressure
source.  The  reference  pressure  can  be  changed to  provide
multiple velocity calibrations.

CONTROLS:  Three balance controls are provided on the front
panel of the control circuit. These are used to insure that the
split-film sensors operate at the same temperature.
                                                     i
SENSOR ARRAY: Three mutually perpendicular single-ended
sensor rods constitute the velocity vector sensing element.
Significant parameters of the sensors are:
   Diameter = 0.006 in. (.15mm)
   Sensitive length = 0.08 in. (2mm)
   Total length » 0.2 in. (5.1mm)
   Substrate - quartz
   Conducting film - platinum
   Protective coating - quartz*

STANDARD  EQUIPMENT  SUPPLIED: Each  1080 system  is
furnished with a 15' probe cable, one probe with pneumatic
shield actuator and built-in  orifice,  power supply,  control
circuit,  bench  type cabinet,  0-5V DC  analog temperature
output,  probe calibration constants, power cord  and  output
connector, instruction  manual and data  reduction procedure
manual.

PNEUMATIC  SYSTEM: Thermo-Systems'  Series 1126 Calibra-
tion  systems  are used to  operate  the pneumatic shield and
calibrator.  The Model  1126 consists of a filter/trap, pressure
regulator, pressure gauge and solenoid valves which are actuated
by the"CALIBRATE" and  "SHIELD" switches located on the
control circuit front panel. Consult TSI for the Muriel 1126 best
suited for your needs.
"Patented
                 Figure  76.    TSI  total   vector  anemometer  specifications

                                                    -179-

-------
    Final  results are summarized in Table 29.   The sensors were found to
be unacceptable for use in this program due  to fouling in a stack environment,

 7.8.5  Ellison Annubar
    As was  previously  mentioned, the Annubar was the only  averaging
 instrument  evaluated during the program.  As a result of this,  it was
 the only  test sensor directly  evaluated during the mapping  tests.  All
 of the point sensors could be, and were, evaluated for mapping  purposes
 through substitution of a pi tot probe.  Throughout the mapping  test
 phase of  the program,  the Annubar was found to be comparable in accuracy
 to an array of up  to eight point sensors, and the hardware  cost impli-
 cations of  this  fact are obvious.  The mapping tests were  in fact
 calibrations for both  the Annubar and the point sensor arrays.  For
 the circular duct  case, it was a matter of comparing actual results with
 predicted results  for  both systems.  For the rectangular case,  it was a
 matter of producing new calibration data for each type of system, since
 none existed previously.  This was done for the point sensor arrays by
 choosing a desired output and obtaining "calibration" data on the
 required array location to achieve that output.  For the Annubar, cali-
 bration was performed  at a fixed location by determining correction
 factors based on the instrument output.  Success for the point sensor
 array was determined by locational repeatability, and for the Annubar
 by correction factor repeatability.  Each system was found to perform
 best in a rectangular  duct when the flow was reasonably two dimensional,
 such is immediately downstream of an elbow.   Test results justified
 the attention given to both the Annubar and  point sensor arrays.
    Other laboratory testing revealed no significant drawbacks for the
Annubar, although field testing did show a significant problem related
 to blockage of the rear static orifice, which is discussed in Section 8,
which did not show up in the laboratory tests.   Minor errors due to
the static orifice also showed up during the mapping test as discussed
in Section 6, where it was noted that the static orifice "overreacts"
to high or low pressure regions along the duct axis.   This  feature is
inherent in the probe design.
                                 -180-

-------
        Table 29.  THERMO SYSTEMS HOT FILM SENSOR CALIBRATION RESULTS
Manufacturer's Stated Accuracy
Manufacturer's Stated Range
Calibration Test Results
Overall Accuracy
     Metal Clad Sensor
     Metal Backed Sensor
     Wedge Sensor
Test Range
Systems Accuracy Based on
  Test Results
       ±3% of reading and ±0.1% of full
       scale for nonlinear system;
       ±5% of reading for linear system
        0 to 60.96 m/sec-htgher and  lower
       ranges available
System Range Based on
  Test Results
Manufacturer's Specifications
Time Response:
Short Term Stability:
Accuracy of Axial Velocity
  Component Measurement
       ±1.2%
       ±3.2%
       ±4.2%
       3.05  to  36.58 m/sec
       Meets manufacturer's specs for
       metal  clad sensor
       Meets manufacturer's specs
Given for System Using Metal  Clad Sensor
       Less than 1  sec
       Better than  1%
       Accuracy, %      ±1   ± 4    ±10
       Maximum Angularity
         Metal  Clad
           x-y plane  ±17°   ±43°
           x-z plane   ±8°   ±16°
         Metal  Backed  ±2°   ±24°
         Wedge         ±2°    ±8°
                                                                >±60°
                                                                 ±25°
                                                                 ±32°
                                                                 ±16°
                                  -181-

-------
     The position of the static orifice makes the Annubar calibration
factor more dependent on the local  Reynolds number based on  the Annubar
diameter and the free stream velocity ahead of the statis orifice than
on any other parameter.  Ellison came to this realization during a
research program performed during the past year, and has revised the
instrument calibration factor on the basis of tests performed in various
facilities in Colorado and elsewhere.  The new calibration data applies
only to circular pipes, and was used as reference data for the circular
duct mapping tests and the reference section measurements.  Other results
of the research program were a small design change in the static port
to improve instrument-to-instrument repeatability, and development of  an
equation to predict Annubar output in any known flow field (i.e. pre-
dictions from mapping data).
     Test results are summarized in Table  30.  The Annubar has demonstrated
adequate accuracy in both circular and rectangular ducts, although greater
calibration factor variations must be expected for the latter case.  Since
the instrument is relatively inexpensive, it would appear that for most
applications a measurement system incorporating the Annubar as the flow
sensor would be less expensive than a point sensor array designed to achieve
a comparable accuracy.
7.8.6  Pitot-Static Probe
     This probe was not evaluated during the program since its properties
are rather well known.  It is being discussed here primarily to allow
comparison to the S probe, which has been  shown to have several undesirable
characteristics.  All  data in this section was taken from Chapter III  of
Reference  3, "The Measurement of Air Flow," unless otherwise stated.
     Common pi tot static tubes are generally of two types — either
hemispherical nosed or ellipsoidal nosed, which is descriptive of the
shape near the tip of  the probe.  There are standard designs for both
shapes, and the designs are  simple enough  to insure repeatable instrument
to  instrument performance.   Careful design has led to instruments which
are free  of Reynolds number  effects for speeds above about 2 m/sec  for a
.61 cm  diameter probe.
                                 -182-

-------
                Table  30.   ANNUBAR  CALIBRATION RESULTS
Manufacturer's Stated Accuracy
Manufacturer's Stated Range

Calibration Test Results
Overall Accuracy
Duct Mapping Accuracy
Repeatability, shift, hysteresis
  in uniform stream
Test Range
Probe Range Based on Test Results
Time Response
Short Term Stability
Accuracy of Axial Velocity
  Component Measurement
Circular duct—varies with  size:
2.4% for 180" duct
Rectangular duct--no claims for
accuracy
MIN--not stated
MAX--varies—at least  61 m/sec
Determination out of program scope
±7.2% for runs performed

-0.5%
2.7 to 30.5 m/sec
5.8 61 + m/sec depending on size
Better than 1 sec
Better than 1%
Accuracy, %:    .
                     ±1  ± 4  ±10
                                   Maximum Angularity,   +8   +4Q   +4Q
                                     Degrees:            ~
                                -183-

-------
     The  pitot-static  probe  is  effectively  axisymmetric,  so  that  it
 responds in  the  same  manner to all  flow  angularities  at  a constant angle
 to  the probe axis,  regardless  of  the  actual  orientation.  As mentioned
 previously,  the  angular  response  of an S probe  is  dependent upon two
 independent  angles, which makes its response more  complex.  As the yaw
 angle of a pitot-static  probe  increases  from zero,  both  the sensed
 impact and static pressures decrease.  However,  the static  pressure
 initially decreases faster, so that the  differential  pressure increases.
 At  higher yaw angles,  the impact  pressure  begins to drop faster, and the
 differential  pressure  decreases.  The same phenomenon occurred during
 laboratory testing  of  the S probe,  but in  a  much more radical manner.
 Pitot-static  probe yaw characteristics are shown in Table  31.   Since
 for the  current  type of  application the  desired output is V cos e (=u),
 the ellipsoidal  probe  is to be preferred.  It would be expected that
 the hemispherical probe  in  particular would  tend to give systematically
 high output  in angular flows.   Recall that this was observed during the
 1974 mapping  tests, although the  errors  were only around 2%.  The
 discovery of  a preference for  ellipsoidal  probes was made late in the
 program  or they  would  have  been used rather  than hemispherical  probes.
 Table 31  shows  that both probes more closely follow V than V cos 0.
 This was  rather  intentional  from  a  design  standpoint since pitot-static
 probes have  traditionally been  used mainly in wind  tunnel and aircraft
 applications  where the flow  is well directed and the desired output is
 the  total flow speed,  rather than a component of it.  The flat response
 with respect  to  V is then desirable since  it minimizes errors due to
 probe misalignment.
     Since highly angular flow was  occasionally encountered during the
 program,   it was  necessary to extend the calibration data for yaw  to
much larger angles.   This was performed with one of the test probes,  and
 the  results revealed that response continued to drop when the angle was
 increased above 30°.  The output went negative around  60°,  reached  a
negative  peak at 90°,  and became positive again around 120°.  The
symmetry  of the curve  around 90° led to a decision  to  assume u  =  0
 (i.e. V  cos  e = 0)  any time the probe  output was negative.   Since cos  e
is at its minimum absolute value,  zero,  at  90°,  it  was felt  that  any
errors  incurred would  be  minimal.   Data are shown in Figure  77.
                                  -184-

-------
         Table  31.   YAW  CHARACTERISTICS OF PITOT-STATIC PROBES
AXIS
                V = magnitude of local velocity
                e = yaw angle
          V  cose = axial velocity component
e
degrees


0
5
10
15
20
25
30
Error with Respect
to V cos e, %
El lipsoidal
Nose
0
+ .6
+2.0
+3.3
+2.7
+1.7
-
Hemispherical
Nose
0
+ .7
+2.7
+5.7
+8.4
+10.5
+13.2
Error with Respect
to V, %
Ellipsoidal
Nose
0
+ .2
+ .4
-.2
-3.5
-7.8
-
Hemispherical
Nose
0
+ .4
+1.2
+2.1
+1.8
+ .2
-2.0
                                 -185-

-------
  .a
  .6-
  .4-
       Ap = differential
           pressure @ Q

      APO = differential
           pressure @ 9 = 0


        C = +1 if Ap >0

        C = -1 if Ap > 0
 -.4
                                     HEMISPHERICAL
                                     NOSED PITOT-
                                     STATIC PROBE
                                                      \
                                                       \ COS 9

                                                         \
                                                           \
                                                            \
                                                              \
                                                                \
                                                                  \
-1.0
           2C
40
60
CO
100
120
                                                      140     160

                                                       0, DEGREES

Figure  77.  Approximate  pitot-static  probe response
             at  large yaw angles
                                                                              180
                                    -186-

-------
 It should be noted  that  the  data were  obtained  in  the mapping  test
 facility, due to unavailability of  the laminar  flow wind  tunnel.
 Consequently the accuracy  is not as  good  as would  have been attained
 in a  wind tunnel.
      There should be  no  question that  a pitot static probe is  preferable
 to an S  probe from  an accuracy standpoint.  The S  probe is more widely
 used  for stack sampling  since it is  easier to handle (i.e. no  "hook"
 on the end to make  insertion and removal  difficult), is less subject
 to clogging,  and gives a higher output than a pitot-static probe.
 Electronic pressure transducers are  now sufficiently accurate  and reliable
 to adequately measure the  differential  pressure from either type of
 probe, so that the  higher  S  probe output  does not  compensate for the
 instrument's  lack of  accuracy.  Traverse  data with a pitot-static probe
 was obtained  during the  field test with no clogging problems,  as is
 discussed in  Section  8,  and  the extra  care needed  to avoid damage during
 insertion and removal did  not result in significant delays.  On the
 basis  of the  accumulated data, it is being concluded that data taken with
 an  ellipsoidal  nosed  pitot-static probe will  probably be 3 to 4 per cent
 more  accurate than  data  taken with an S pitot probe.   Consequently the
 pitot-static  probe  should  be used for normal  manual  traversing except
 when  fouling  or handling problems require use of the S type probe.  It
 must  be  noted  that  the expected difference of 3 - 4% does not explain
 the 10%  difference  observed  during the field  demonstration and discussed
 in Section 8.  This higher error may be due to flow angularities in the
 "tilt" direction.
     As  a  further result of  flow angularity investigations, it is being
 recommended that a  pitot-static probe be developed whose  angular
 response  directly follows the axial  velocity  component.   This  could help
 to significantly improve system accuracy in angular flows.
 7.8.7  General Laboratory Test Summary
     Final laboratory results are given in Table 32. The  Ramapo Fluid
Drag Meter was definitely the most accurate point sensor.   The Annubar
was found acceptable as an averaging instrument  for use  in both circular
and rectangular ducts, the latter solely as a  result  of calibration

                                 -187-

-------
                                        Table 32.  LABORATORY TEST SUMMARY
              ACCEPTABLE AVERAGING SENSOR
SENSOR
Ellison Annubar
ATTRIBUTES
Reasonable accuracy and wide
applicability; good
survivability
LIMITATIONS
In place calibration desirable;
static port may need purge
00
oo
              ACCEPTABLE POINT SENSOR
SENSOR
Ramapo Mark VI
ATTRIBUTES
Very good accuracy and
survivability
LIMITATIONS
High cost per channel

-------
                                   Table 32 (continued).   LABORATORY TEST SUMMARY
I

CXI
              CONDITIONALLY ACCEPTABLE  POINT SENSORS
                  SENSOR
                S Probe

                 Hastings
                 Raydist
                AFI-10K
      ATTRIBUTES
Inexpensive, widely used

Very good survivabllity
    LIMITATIONS
Relatively poor accuracy

Poor accuracy
              UNACCEPTABLE POINT SENSORS
SENSOR
Thermo Systems
VT 161
ATTRIBUTES
Linear output,
simple operation
LIMITATIONS
Poor survivability

-------
work  performed  during  this  program.   Both  the S probe and the Hastings-
Raydist  probe were  found  environmentally acceptable but exhibit un-
desirable  accuracy  potential,  especially the Hastings probe.  The Thermo
Systems  sensors were found  unacceptable due to fouling problems.  As
a  sidelight of  the  test,  a  clear preference for the pitot-static probe
over  the S probe was indicated for reference traverse work.  As a result
of the laboratory tests,  the Ramapo unit and the Annubar were recommended
for field  test  evaluation,  with reference  data to be supplied by a pitot-
static probe.   The  test probes were to be  used in accordance with
techniques developed in Task III.
7.4.9 Pre-Field Test  Demonstration
      The purpose of this  demonstration was to verify in the laboratory the
acceptability of the sensors and measurement techniques to be demonstrated
in  the field.  Environmental acceptability of the Annubar and Ramapo probe
had been demonstrated, so further evaluation of this type was not performed.
The field site selected was the Moapa station of the Nevada Power Company,
Moapa, Nevada, described in Section 8.  The mapping facility inlet was
configured to represent expected field conditions as closely as  possible,
and the normal test data were taken.   The demonstration runs were runs
45-48, using the straight rectangular inlet.   Average results for the
four  runs showed an Annubar calibration factor of .5765,  and a location
for the row averaging technique of 20.8% of the  duct width from  one side
wall.   It was immediately recognized  that the Annubar calibration factor
was significantly lower than the normal .65 value,  but this  was  expected
since the geometry of the duct work indicated clearly that a high velocity
region was expected in the center of  the duct, which would result in a
high Annubar output and consequently  a lower than normal  calibration
factor.   It was  decided that the Ramapo probe would be attached  to  a
reciprocating traversal unit to obtain the  row data for that measurement
technique.
                                  -190-

-------
                               SECTION VIII
                      TASK V - FIELD DEMONSTRATION
8.1  GENERAL
     Two field demonstrations were performed at the Reid-Gardner Station
of the Nevada Power Company, located at Moapa,  Nevada.   The plant is  about
72 km. northeast of Las Vegas.   Each test lasted approximately 18 days.
The first was begun in September, 1974, and the second  was begun in
February, 1975.
     The primary purpose of the first test was  to demonstrate hardware and
techniques for volumetric flow measurement which had been successfully
tested in the laboratory.  The primary purpose  of the second test was to
demonstrate techniques for gas composition measurement.   All  composition
data other than that required for flow calculations are to be discussed in
a separate report.
8.2  FACILITY DESCRIPTION
     The plant presently consists of two Foster-Wheeler 120 megawatt  boilers,
with a third under construction.  All work was  performed on the #2 unit,
shown schematically in Figure 78.  Flow from the boiler is separated  into
two streams which pass through Lunjstrom rotary air preheaters.   The  ducts
at the two preheater outlets go through a shape transition and then  rejoin
upon entering a mechanical dust collector.  A row of test ports is located
on each duct just ahead of the dust collector.   This is shown in Figure 79.
All rectangular duct mapping was done at these  locations.
     When the plant was first constructed, the  dust collector exhausted
directly into the stack.  Subsequently, a venturi scrubber and separator
were added by Combustion Equipment Associates.   When the scrubber is  on,
the flue gas is diverted at the dust collector  exit and routed to the
scrubber, where it is processed and fed into the stacks.  If the scrubber
is not on, the flue gas goes directly into the  stack after leaving the
dust collector.  The inlet to the stack is about 13 m above ground level.
Sample ports are located in the stack at the 31.4 m level, or about four
stack diameters above the inlet.  All circular  duct testing was performed at
this location in the stack.
                                   -191-

-------
PO
I
        Plane
        A-A
                       Scrubber
                             r
           Flow routed to and from
           scrubber at plane A-A
                                          Stack
                                                    Location  2
                             Preheaters
Dust
Collector
 V
Location
   I
                                                   Boiler
         A =  Annubar
                                                                           A
                                                                    Duct 1                      Duct 2
                                                                                Location  1
                                        Location  2

                                        Figure 78.   Schematic of MOAPA  power plant

-------
  FLOU

  DIRECTION
                                           SAMPLE PORTS (ROW NUMBER)
ANNUBAR LOCATION
                                  R = PORT FOR RAMAPO
                                  1   PROBE - PROBE KEPT AT R-4
                                                                  SAMPLE POINTS

                                                                  1
                                                  TRAVERSE MAP -
                                                  49 POINTS
                          SAMPLING
                          PLANE
Figure 79.   Field  demonstration  rectangular duct geometry

                                -193-

-------
 8.3   TEST  CONDUCT
 8.3.1   Test  Hardware
      Flow  instrumentation  for  the  1974  test  included two Annubars, one for
 stack measurements  and  one for measurements  in the  left hand duct after the
 preheaters  (Duct 1).  The  Ramapo Fluid  Drag  meter which was evaluated in
 the  laboratory  was  modified to increase its  length  and installed in an
 automatic  traversing mechanism (Figure  80).  Thermocouple probes for use
 with each  of the three  test instruments were fabricated to provide temper-
 ature data.
      Reference  flow data were  provided  by a  hemispherical nosed pitot static
 probe with  an integral  thermocouple.  This probe was used for reference
 traverses  in Duct  1.  The  S type probe  evaluated in the laboratory was
 modified  to  increase  its length and was used for both velocity traverses
 and  gas sample  traverses.   Bulk composition  data were provided by a Carle
 thermal conductivity detector  gas  chromatograph which measured concentrations
 of N2,  02,  C02  and  CO.
      All  pressure  sensors  were monitored by  the same Baratron pressure
 transducers  used in the laboratory.   Instrument outputs were routed through
 a  scanner  to a  digital  voltmeter and  then to a paper tape printer.  The gas
 chromatograph had  its own  separate output system.   Flow data output inter-
 vals varied  from ten seconds to one hour.  Data reduction was accomplished
 by use  of  a  digital computer after the  field work was completed.  The two
 Annubars were left  in place upon completion  of the  test.
      Hardware for  the 1975 test included the two Annubars left in place
•previously  plus a  new third Annubar installed in Duct 2.  The Ramapo probe
 was  not used due to malfunction of the  traversing mechanism during the
 first test.   The type S probe  was  also  not used.  Reference pitot traverse
 data for  the 1975  test  were obtained  in Duct 1, Duct 2, and in the stack.
 The  Carle  gas chromatograph was replaced by  a newer model which allowed for
 automatic  sampling, whereas the older model  required manual operation.
 1975 hardware also  included a  purge system to keep  the rear orifice on
 each Annubar clean. Purging was done for five minutes every half hour.
                                    -194-

-------
                                 PULLEYS
                      MOTOR
                  THREADED ROD	
Operation:  Drag meter is attached
to nut inside square tube.  Nut rides
on threaded rod which is fixed to
square tube.  When motor is turned
on, nut travels down until it trips
bottom switch, which reverses motor.
Nut travels up until it trips top
switch, again reversing motor.
Cycle continues
                                                     ELECTRICAL  LEADS
                                            _1  TOP SWITCH
n
         "NUT
                                                      TRAVERSE  TUBE
                                                     BOTTOM SWITCH
                                                     DUCT WALL
                                            RAMAPO DRAG METER
                                                                       FLOW
        Figure 80.  Schematic of traversing mechanism for Ramapo drag meter

                                      -195-

-------
8.3.2  Extent of Testing
     The Annubars were monitored continuously unless the pressure transducers
were required to support manual traverses.   The duct Annubars were removed
for pitot or gas sample traverses in the ports^where the Annubars were
installed.  The Ramapo probe was monitored continuously during the first
test.  Complete gas sample traverses were performed at both full  and half
load in Duct 1 during the first test.  Complete gas sample traverses were
performed at full load in Duct 1 and in Duct 2 during the second test.
Reference pi tot-static traverses were performed as often as the work load
allowed.  Twenty-two were performed in the Duct area during both tests,
and fifteen were performed in the stack during the second test.
     A total of four basic plant conditions occurred during the first test,
each about equally often.  The unit load was held at maximum (~110-120 MW)
during the PM hours and at about half that during the AM hours.  The
scrubber was on about 40% of the time.  The unit was down about five days
during the test.  Full load was maintained during the second test, the
variation being from about 100 to 123 MW.  The scrubber was on full time
except for one five hour period.  Stack traverses were obtained during
this period  (February 28).  There was no down time during the second test.
8.3.3  Problems
     The  automatic traversing mechanism used with the Ramapo probe failed
early during  the first test due  to  the combined effects of particulate
clogging  around the mating seal  and a too-powerful motor used to drive the
mechanism.   The probe was removed from the unit and  installed at a fixed
location, except during manual  row  traverses toward  the end of the test.
The  Ramapo  traverse data had to  be  discarded due  to  internally shorted leads.
The  leads were modified when the probe was lengthened for the test, and
the  insulation around the junctions  failed due to heat.  These difficulties
were clearly the responsibility  of  test personnel,  and do not reflect
adversely on the instrument itself  nor on the manufacturer.  No clogging
 problems  were experienced with the  Ramapo probe.
     Laboratory  testing  showed the  ability of  the Annubar to  "self-purge"
the  impact  holes.  This was confirmed  during the  field test.  A new problem
did  show  up, however, and that was  the effect  of  particulate on the Annubar

                                    -196-

-------
rear orifice.  Data during the first test showed that this  orifice clogged
repeatedly when the scrubber was operating.  Since the particulate loading
in this operation mode was relatively low, it was apparent  that the high
moisture content caused the remaining particulate to become much more
adhesive.  Purging through the pressure line alleviated the blockage of the
stack Annubar.  Accumulation of considerable particulate on the rear side
of the Duct 1 Annubar was observed when the probe was removed during pi tot
traverses.  Although complete clogging was never observed,  the accumulated
particulate definitely altered the shape of the Annubar's rear orifice.
This showed up as a significant variation in calibration factor during  the
first test for both Annubars.  The problem was effectively  solved during
the second test by addition of the automatic purge system.
     The only new problem encountered during the second test was with
installation of the Duct 2 Annubar.  The duct itself was deformed in the
immediately area of the test port in which the instrument was placed.  This
resulted in the top of the instrument having about an 8° offset in the
vertical plane (the duct axis was horizontal), facing upstream.  As the test
got underway, it was immediately apparent that this Annubar had readings
consistently lower than the Duct 1 Annubar.  These two Annubars were
switched after the first week of testing, and no change in  reading from
either location was noticed, which meant that the Annubars  themselves were
identical.  Testing showed a large difference in calibration factors between
these two Annubars.  It would be impossible to say how much of the difference
was due to the deformation of the duct without getting rid  of the deformation
and recalibrating the probe.
8.4  FLOW DATA CORRELATION
8.4.1  1974 Test
     The Duct 1 Annubar was calibrated in place by means of the pi tot tra-
verses performed in the duct.  Calibration factors were obtained for the
stack Annubar during Duct 1 traverses when the scrubber was off by assuming
that half the stack input came from Duct 1.  This was considered valid  since
no additional flow sources or sinks occurred ahead of the stack.  Correla-
tion could not be obtained when the scrubber was on due to  additional water
vapor and reheat air added during the scrubbing process. Failures with
                                   -197-

-------
the Ramapo system precluded its calibration.   The final  two traverses  were
performed simultaneously with the pitot-static and S probes to obtain  a
direct comparison between the two.
     Coal usage per day and coal analysis data were coupled with test  com-
position measurements of CCL in Duct 1  to make independent flow computations.
The coal data were provided by plant personnel.
8.4.2  1975 Test
     The Annubars in Ducts 1 and 2 were calibrated in place by means of
pitot traverses in each duct.  The stack Annubar was calibrated by pi tot
traverses in the stack.  Plant data were used as during  the first test.
8.5  TEST RESULTS
8.5.1  Composition Measurements and PI ant Reference Data

8.5.1.1  1974 Test-
     Dry gas composition traverses in Duct 1  at full and half loads showed
the average CCL concentration by volume to be 15.83% and 9.078$, respectively.
Plant coal analysis showed that the average usable carbon content of the
coal during the test was 84.0% by weight.  Water content was 5.93%.  Total
water content of the wet gas Duct 1 was determined to be 4% at full load
and 2% at half load, respectively, reducing the C02 content of the wet
gas in Duct 1 to 15.20% and 8.81% for those conditions.   The total flow
per day  could then  be  calculated as follows:

                     v = Coal used (moles) X  .0224SH3/mo1e             ,g4j
                             Average CO- Concentration

     where           V = Total volume flow, SCM

8.5.1.2  1975 Test-
     Average C02 content of the dry gas was 13.85% at full load (the only
run condition) which showed that more excess air was being fed into the
boiler during the second test.  Coal composition data during the test period
were not available  at  the  time of this writing, so January, 1975 data
(month prior to test)  were used.  The January data showed the coal to be

                                   -198-

-------
82.81% C and 7.47% H20.  The average wet gas CCL concentration in Ducts 1
and 2 was 12.99%.
8.5.2  Velocity Traverses
8.5.2.1  1974 Test-
     Results are shown in Table 33.  The estimated flow through both ducts
from plant data is based on average coal usage at full  and half load.   The
49 point traverses constituted in place calibration of the Annubars.  Duct  1
Annubar calibration factor is seen to have varied from .531 to .620.  This
wide range is believed to be due to the particulate problem discussed  above.
The calibration factor was much more constant during the 1975 test.
     The calibration factor for the stack Annubar was initially higher
than the factory suggested value, then seemed to stabilize somewhat  at a
relatively low value.  An average value of S = .62 was  used for continuous
data reduction.  Data from the stack Annubar were generally suspect  due to
the clogging problems which occurred during the test.  Testing during  1975
showed that an average of 49% of the flue gas went through Duct 1  and  51%
went through Duct 2.   The stack calibration factors in  Table 33 as shown
were modified on the basis of this new information.
8.5.2.2  1975 Test-
     In place calibrations of all three Annubars were conducted, and results
are shown in Tables 34 and 35.  Forty-nine point duct traverses were per-
formed on four separate occasions.  On each occasion, they were performed
in immediate succession, which resulted in a total variation from the  mean
of less than 3% for each group.  The calibration factors shown were  based
on the average Annubar  reading during each group of traverses.  Duct  1
results were very consistent for the two groups, and were in good agree-
ment with 1974 results.  Duct 2 results were also self-consistent, but
the average calibration factor was 14.8% higher than that for the Duct 1
Annubar.  Since the ducts are identical in design, equality between  the
calibration factors of the two instruments was somewhat expected.  The
difference cannot be explainted, but may have to do with the unavoidable
misalignment of the Duct 2 Annubar, discussed in Section 8.3.
                                  -199-

-------
                  Table  33.   DUCT  TRAVERSE  SUMMARY,  1974
Date
9-25
9-27
9-28
9-28
10-2
10-2
10-2
10-4
10-4




Load
Full
Full
Half
Half
Half
Half
Full
Full
Changing
Pi tot
Traverse
Duct 1
Flow Rate
SCM/Sec
62.8
62.6
42.0
39.8
43.6
41.1
57.4
56.6
52.8
Estimated
Duct Flow From
Plant Data
SCM/Sec
118.5
118.5
85.0
85.0
85.0
85.0
118.5
118.5
*
Average Duct Calibration Factor
Full Load 0.588
Half Load 0.555
Weighted Average 0.575
Annubar
Calibration
Factor
Ductl Stack
0.620 0.717
0.605 +
0.564 +
0.531 +
0.558 0.632
0.566 0.606
0.571 0.619
0.558 0.621
0.584 +




*No data due to rapidly changing load




 Scrubber on-no correlation possible





 Factory calibration  factor  for  stack Annubar:   .661
                                  -200-

-------
           Table 34.  DUCT TRAVERSE SUMMARY, 1975
Date
2-24
2-28
2-25
2-28
Duct
1
1
2
2
Traverses
2
4
3
4
Average Flow
Rate, SCM/SEC
61.72
61.63
63.99
58.81
Duct Annubar
Calibration Factor
0.586
0.583
0.674
0.668
       Duct 1

       Duct 2
Average Annubar
Calibration Factor

      0.584

      0.6705
           Table 35.   Stack Traverse Summary,  1975
Date
2-17
2-28
Scrubber
ON
OFF
Traverses
10
5
Average Flow Rate
SCM/SEC
154.4 ± 2.2
130.1 ± 0.8
  Scrubber ON


  Scrubber OFF
   Stack Annubar
   Calibration Factor

         0.723
         0.729
Factory calibration factor for stack Annubar  =  .661.
                             -201-

-------
     Four test ports were available in the stacks at 90°  intervals.   The
Annubar was installed through two of them, and twenty point traverses were
performed along the diameter through the other two ports.   It  would  have
been preferable to remove the Annubar and perform traverses along both
diameters.  Removal of the stack Annubar was considered to be  too hazardous
to perform more often than necessary, since the probe is  5 meters long,
weights about 60 kg, and must be handled on a one meter catwalk.
     One set of pressure lines of length 60 meters was available from the
stack, so only one instrument could be monitored at a time. As a result,
Annubar readings immediately before and after the pitot traverse were used
to obtain the calibration factor for the instrument.  Results  are shown
in Table 35, from which it is apparent that flow conditions were quite
steady during the traversal periods.  The Annubar calibration  factors were
consistent for both operation modes.  It should be noted that  the first
calibration in 1974 produced a factor close to the 1975 results, further
suggesting that later degradation during the 1974 test was due to par-
ti cul ate  accummulation.
     Although the  1975 calibration  results showed much better consistency,
it must still be noted that poor agreement was obtained with the suggested
factory value.  This can be explained on the basis of Figures 81 and 56.
Figure 81  shows very flat velocity  profiles except in the  immediate vicinity
of  the wall.   Factory calibration  factors are based  on fully developed pipe
flow characteristics, for which  a  parabolic profile  would  be expected.
The  profiles  shown  in Figure 81  are almost  uniform.  Since the calibration
data in  Figure  56  were taken in  a  uniform flow  in  a  wind  tunnel, the
stack  calibration  value would be expected to  approach the  wind tunnel  value.
At  comparable speed, Figure  56  shows  kA = 1.53.   The standard calibration
factor S  and  the  factor  kA  in Figure  56 are related  as follows:
                                s -  _J _                            (65)
 so that for the case of a uniform stream,  we  would  expect  S  =  .81.  Con-
 sequently, the test calibration factor appears  reasonable.   The  problem
 is that the Annubar should not require an  in-place  calibration for  accurate
 measurements.  Laboratory 'test results in  1974  over a  wide variety  of
                                    -202-

-------
     Velocity at
16
14
12

10
a
L>
P
ro
CD
00
' 6
4
2
0 .
_ 3 lanuar u i-uiiu I L 1 Uilb , U
M/Sec S
"
0 o 0
- nO 0 0 0 0 ^ 0 0 0 -
0° °
- a D a D a a a o o D a
D D
D

O Scrubber On, u =13.62 m/sec



D Scrubber Off , u = 11.42 m/sec


	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 | 	
O
c
D
D D _
C
~



-
-
-

IU°
  40       20        0       20      40      60




            Stack Radius, Per Cent From Center




Figure  81.  Stack velocity profiles
80
                                                                                    IOC

-------
conditions close to disturbances did show S factors this high on occasion
(Table 5) although the average values were much closer to the factory
prediction.  Figure 56 and the field test results conclusively show that
an almost uniform flow such as those in Figure 81 will  result in a  higher
than normal calibration factor.   It is an interesting property of the
instrument that it works well  in a variety of stratified flows, but shows
a large calibration shift in a uniform flow.   The reason is  the design  of
the rear orifice, which sense a  pressure below local  free stream static
pressure.  This low pressure is  the reason why the instrument has a cali-
bration factor less than one.   If the instrument were designed so that  the
low pressure orifice sensed the  static pressure of the stream, the  calibra-
tion factor in a uniform stream  would be exactly one, and the calibration
in a developed pipe flow would be very close  to one.
8.5.3  Comparison of S Type Probe and Pi tot-Static Probe Traverses
     The last two traverses during the 1974 test were performed simul-
taneously with a pitot-static probe and the S type probe evaluated  in
the laboratory.  The pitot factor used for the S probe was .823, based  on
the manufacturer's data, rather  than the laboratory test value of .852.
This choice was deliberately made to minimize known errors in angular flow
as shown in Figure 59.  For these two traverses, the pitot-static data  gave
                              o
flow values of 56.6 and 52.8 m /sec, as shown in Table 33.  Corresponding
results with the S probe were 62.4 and 58.4 m /sec, respectively.  The
S probe results were 10.3% and 10.6% higher than the corresponding  pitot-
static probe data.  Results were obviously consistent for both runs.  It
is not known why the S probe data showed such a high shift.   Laboratory
data showed that high readings would normally be expected in angular flows,
and this is why a low pitot factor was used.   As will be shown below, most
traverse point velocities were around 7-8 m /sec, which is where the S
probe tested shows a calibration shift, as shown in Figure 43, which would
result in abnormally high readings, but only  by 2-3%.  Since angularity
testing was not performed in the "tilt" direction as shown in Figure 58,
the problem may be due to that characteristic.  Duct geometry in Figure 79
suggestes that significant flow  angularity may be present in that plane.
                                  -204-

-------
8.5.4  Continuous Flow Monitoring
8.5.4.1  1974 Test-
    The pi tot-static traverses discussed above established the Annubar
calibration factors to be used during the test.   Results  are summarized
in Table 36.  For proper instrument operation, the Duct 1 flow would be
half the stack flow.  Comparison of Duct 1 Annubar data and plant data
was performed for two days and shown in Table 37.   The  Annubar data
shown assume that 49% of the flow went through Duct 1, as was determined
for the 1975 test.  The C02 concentrations shown were obtained by taking
an average of the full load and half load concentrations  times the number
of hours each day at each condition.  Good agreement between test and plant
data was obtained.
          Table 36   SUMMARY OF ANNUBAR CONTINUOUS MONITORING
                     DATA WITH SCRUBBER OFF, 1974
         Load
          MW
         115-123

          49-57
    Average Flow Rate, SCM/SEC
Duct 1  Annubar
  62.3 ± 4.1

  44.6 ± 3.0
Stack Annubar
 138.3 ± 8.2
  92.1 ± 6.4
         Calibration factors  used - Stack  :  0.729
                                    Duct 1:  0.575
          Load
Ratio of Duct 1 Annubar Average Flow to
      Stack Annubar Average Flow
         115-123
          49-57
              0.450 ± 0.046
              0.484  ±  0.054
                                     -205-

-------
                             Table 37.  COMPUTATION OF TOTAL FLOW FROM COAL ANALYSIS AND

                                        MEASURED C02 CONCENTRATION, 1974
Date
10-2
10-3
Coal Used
«g
3,251,380
3,571,480
Carbon Used
kg gm-moles
2,752,618 4.729x107
3,023,614 5.195x107
Average C09
% *
11.89
12.74
Total Moles
of Dry Gas
3.977xl08
4.078x108
Total Moles
of Wet Gas
4.122xl08
4.226xl08
Date
10-2
10-3
Annubar
Average Flow
Rate, SCMS
104.5
109.4
Annubar
Total Flow
SCM
9.026xl06
9.451xl06
Total Wet Gas
Volume, SCM
9.234xl06
9. 466x1 O6
Difference
%
+ 2.3
+ 1.6
ro
o
01
t

-------
8.5.4.2  1975 Test--
    1975 test results are shown in Table 38 on a daily basis.  It was
these results on which the conclusion was based that 49% of the flow went
through Duct 1.  Plant data are given in terms of average rather than
total flow as in Table 37.  Very good agreement is again indicated.  Also,
note that the average flow rate in the stack from the pi tot traverses with
the scrubber off was 130.1 SCM/SEC (Table 34).  This shows that there was
no apparent loss of accuracy of the stack traverses from taking data along
only one diameter.  Flowrates for the 1975 test were higher than for full
load during the 1974 test due to higher amounts of excess air in the system
for the 1975 test period.  Absence of data in Table 38 for the Duct 2 Annubar
on some days are solely a result of the pressure transducer being used for
other purposes.
    Average daily flow rate through the stack is shown in Table 39, which
shows flow rates higher than those in Table 38.  Plant data in Table 39
are based on the coal usage and nominal plant specifications for added
moisture and reheat air during the scrubbing process:  8% and 21% of the
flow through the ducts, respectively.  During the test, the reheat fan
was operated at about 60% of full power according to plant personnel.
The fan is throttled to provide a temperature rise of 17°C in the flue
gas leaving the scrubber.  The actual flow is not monitored.  The measured
flow rate of 152.6 SCM/SEC is 18% higher than the average duct flow, and
so is reasonably compatible with a 60% below operation, which would result
in an average flow rate of 159 SCM/SEC.
8.5.5  Short Term Monitoring - 1974
    The Ramapo probe was placed as shown in Figure 80 after the traversing
mechanism failed during the 1974 test.  This corresponded to the row
location, for the Row Average Method, selected prior to the test.  The probe
was left at the center of the row.  Short term stability of the resulting
point measurement was compared with that of the Duct 1 Annubar and results
are shown in Table 40.
                                    -207-

-------
        Table  38.  AVERAGE DAILY  FLOW THROUGH DUCTS, 1975
Date
2-15
2-16
2-17
2-18
2-19
2-20
2-21
2-22
2-23
2-24
2-25
2-26
2-27
2-28
Ave
Average Flow And Standard
Deviation, Test
SCM/SEC
Duct 1
63.4 ± 1.7
63.5 ± 2.8
64.7 ± 2.3
67.3 ± 2.2
65.8 ± 2.3
65.3 ± 1.6
65.1 ± 0.9
63.2 ± 2.6
64.3 ± 5.2
60.1 ± 8.5
62.5 ± 1.7
62.7 ± 1.8
60.8 ± 2.1
60.3 ± 1.6
63.5
Duct 2
X
X
64.6 ± 1.9
64.5 ± 2.4
65.2 ± 3.0
66.7 ± 2.9
64.3 ± 3.9
67.3 ± 8.2
X
67.3 ± 3.0
67.2 ± 0.9
66.7 ± 2.1
X
X
66.0
Total
X
X
129.3
131.8
131.0
132.0
129.4
130.5
X
127.4
129.7
129.4
X
X
129.5
Average Flow In Ducts
From Plant Data
SCM/SEC
132.0
132.6
134.3
135.1
132.8
134.7
133.7
132.9
128.1
130.6
130.1
132.7
130.2
126.3
131.9
^Indicates insufficient data
                               -208-

-------
     Table 39.   AVERAGE DAILY FLOW THROUGH STACK,  1975
Date
2-15
2-16
2-17
2-18
2-19
2-20
2-21
2-22
2-23
2-24
2-25
2-26
2-27
*
2-28
Ave
Average Stack Flow And
Standard Deviation,
Test SCM/SEC
145.8 ± 3.8
149.0 ± 4.6
152.6 ± 4.0
159.4 ± 4.4
158.1 ± 9.1
155.7 ± 5.3
155.9 ± 5.0
156.1 ± 4.1
152.1 ± 4.3
154.3 ± 3.1
144.6 ± 3.2
145.9 ± 7.1
150.8 ± 3.8
155.4 ± 3.8
152.6
Average Stack Flow
From Plant Data
SCM/SEC
180.5
181.4
183.7
184.8
181.6
184.2
182.9
181.8
175.2
178.6
177.9
181.5
178.1
172.7
180.4
Assumes scrubber on full time
                            -209-

-------
                Table 40.  SUMMARY OF ANNUBAR AND RAMAPO SHORT
                           TERM MONITORING  DATA,  1974
Data
Time
Data Points
Load, MW
9 - 27
14:10 - 14:59
50
118
10 - 3
19:00 - 19:59
60
114
Duct 1  Annubar Flow
Stack Annubar Flow
Ramapo Flow	
Stack Annubar Flow
0.589 + 0.022 (3.6%)
0.418 + 0.096 (22.9%)
0.500 + 0.012 (2.4%)
0.414 + 0.071 (17.1%)
    Data in Table 40 show two notable things.  The first is a variation in
the Annubar calibration coefficients,  which has been discussed above.
The second is that the point sensor readings, which on the average main-
tained a constant ratio with respect to the stack Annubar output, showed
a very high degree of scatter with respect to the Duct Annubar readings.
Since the Annubar reacts to the local velocity at five different points,
it is logical that it would show less data scatter than the Ramapo single
point sensor, and the data confirm this.  The wide scatter for the Ramapo
probe also suggest that it was located in a region where velocity gradients
were relatively high.  This was confirmed, as is shown in the following
paragraph, with resulting implications for use of the Row Average Method.
8.5.6  Row Average Analysis of Pi tot Traverse Data
    Due to failure of the Ramapo traversing mechanism, it was necessary to
evaluate the Row Average Method by analysis of the pi tot traverse data, as
was done in the 1973 laboratory testing.  Complete velocity maps are shown
for two of the 1974 runs in Table 41.  Point velocities equal to zero
indicate a negative pi tot probe differential pressure, which was taken  to
correspond to a local flow angularity of  90°, as explained in Section 7.8
and shown  in Figure 77.  A flow angularity of 90° corresponds to zero
local flow through the measurement plane.  1974 data were analyzed verti-
cally  (rows) and horizontally  (columns) as shown in Figure 79.  Data  are
presented  in Table 42.  Arguments could be made for taking the  "Rows" in
either direction according to  the definition of a row being  in  the plane
                                   -210-

-------
Table 41.  FIELD DEMONSTRATION PITOT TRAVERSE DATA,  1974
           RUN  1   9-25  14:00  LOAD = 106MW
Row
1
2
3
4
5
6
7
Sample Point
1
0.000
0.000
9.516
4.011
6.115
7.182
3.780
2
9.009
0.000
6.928
9.166
8.687
7.663
7.059
3
10.657
7.362
6.674
8.973
7.497
7.011
8.299
4
11.490
8.373
7.747
7.552
6.395
6.966
7.189
5
8.941
6.575
9.267
8.939
6.578
7.089
7.020
6
6.448
8.511
9.700
9.463
6.389
6.299
5.928
7
4.227
7.094
8.600
9.453
5.482
4.630
3.495
           RUN  3  9-28  11:00  LOAD = 52MW
Row
1
2
3
4
5
6
7
Sample Point
1
3.147
3.840
2.663
0.000
3.404
3.191
0.000
2
4.200
4.729
8.262
7.043
0.000
5.011
4.768
3
6.649
3.928
3.998
6.047
5.061
4.877
5.013
4
6.546
4.014
6.790
6.269
3.920
3.880
3.785
5
8.473
4.302
6.171
5.274
5.798
5.395
4.788
6
6.439
5.060
6.467
7.682
4.631
4.370
2.431
7
6.479
6.078
7.114
6.661
2.480
1.273
0.000
 Tabular values are velocity at standard conditions,  M/S
                           -211-

-------
              Table 42.   ROW AND  COLUMN ANALYSIS OF  FIELD
                          TEST  PITOT-STATIC TRAVERSE  DATA,  1974
RUN

1
2
3
4
5
6
7
8
9
AVE
a,%
Row
1
1.041
.878
1.285
.961
.952
.929
1.044
.962
.886
.993
11.8
2
.777
.930
.979
1.076
.956
.927
.952
.940
.939
.942
7.6
3
1.193
1.138
1.270
1.136
1.241
1.015
.977
1.072
1.094
1.127
8.3
4
1.180
1.244
1.195
1.184
1.127
1.228
1.228
1.201
1.129
1.191
3.3
5
.967
.953
.775
.872
.923
1.061
.912
1.201
1.013
.943
8.5
6
.960
.939
.858
1.107
.853
1.059
1.034
.996
1.056
.985
8.5
7
.877
.918
.637
.664
.948
.780
.854
.822
.882
.820
12.4
RUN

1
2
3
4
5
6
7
8
9
AVE
a,%
Column
1
.627
.594
.498
.522
.744
.477
.500
.461
.646
.563
16.0
2
.995
1.073
1.043
1.178
1.115
1.196
1.004
1.010
1.293
1.101
8.8
3
1.158
1.077
1.090
.977
1.097
.974
1.157
1.108
1.257
1.009
8.3
4
1.142
1.119
1.079
1.153
1.141
1.246
1.146
1.086
1.092
1.134
4.2
5
1.115
1.107
1.231
1.081
1.097
1.187
1.123
1.170
1.212
1.147
4.4
6
1.081
1.013
1.137
.988
.881
.958
1.072
1.135
1.221
1.054
9.4
7
.881
1.015
.922
1.101
.925
.961
.399
1.029
1.018
.983
6.5
Rows were vertical, columns horizontal
Tabular values are row average divided by overall  average

                                -212-

-------
 of the  highest  gradients,  but  they-were assumed to be vertical for obvious
 hardware  installation  purposes.   Results for  the  1975 traverses are shown
 in Table  43,  the  rows  being  vertical.
     1974  results  showed  higher scatter since  data were taken at both full
 and half  load,  and  only  at full load  in 1975. Rows 2 and 6, at 21.4% of
 the duct  width  from the  wall,  are closest to  the nominal recommended
 distance  of  19%.  The  average  absolute error  for these rows was 5.1% which
 is  considered reasonable, especially  in view  of the unusual duct geometry,
 and the average standard deviation was ± 5.2%.  The average error for row
 4  was +15.4% and  the average standard deviation was ± 3.5%.  If a measurement
 system  using the  Row Average Method were to be installed in the future in
 one of  these ducts,  it would be recommended that it be placed in the
 number  4  port rather than the  number 2 or number 6 port.  The systematic
 error could then  be  calibrated out, and the working system would have a
 minimal data scatter.
 8.5.7   Use of Pitot-Static Probe in Field Applications
     Once  test personnel  got accustomed to the equipment, 49 point pitot-
 static probe traverses were conducted in each duct in about 50 minutes.
 The  probe head was occasionally bent during removal  during the first two
 traverses, but was not a problem thereafter.  No clogging problems were
 experienced during duct  traverses.  Clogging did occur in the stack with
 the  scrubber on after about one hour of continuous operation.  This was
 alleviated by wiping off the probe head and purging  through the pressure
 lines.  No other problems were encountered.
 8.6  SUMMARY OF RESULTS
    There were significant problems  with each instrument system during
 the  first test.  The traversing mechanism failed  to operate  properly,  which
 has  resulted in a general recommendation that such a  device be perfected
 and made generally available for large ducts.   Previous  work with smaller
 devices in wind tunnels was the basis  for construction of the  test unit
 used, but field operation made it  evident that proper design of an adequate
 system was outside program scope.   The basic adequacy of the Row  Average
Method itself was demonstrated through analysis  of pitot traverse  data.
                                   -213-

-------
 Table 43.   SUMMARY OF PITOT TRAVERSE ROW AVERAGE  DATA,  1975

Run
1
2
3
4
5
6
Ave
a, %
Duct 1 Row
1
0.920
0.913
0.910
0.905
0.961
0.963
0.929
2.6
2
1.090
0.994
1.090
1.108
1.014
1.046
1.057
4.0
3
1.191
1.197
1.123
1.103
1.120
1.070
1.134
4.0
4
1.163
1.235
1.070
1.157
1.142
1.154
1.154
4.2
5
0.856
0.861
0.913
0.929
0.913
0.900
0.895
3.2
6
0.882
0.895
0.969
0.950
0.945
1.035
0.946
5.3
7
0.898
0.904
0.926
0.848
0.904
0.833
0.886
3.8

Run
1
2
3
4
5
6
7
Ave
a, %
Duct 2 Row
1
0.921
0.934
0.906
0.969
1.036
1.074
1.116
0.994
7.6
2
0.892
0.927
0.922
0.892
0.959
0.904
0.894
0.899
2.3
3
1.268
1.236
1.247
1.185
1.198
1.197
1.166
1.214
3.0
4
1.093
1.124
1.094
1.133
1.187
1.078
1.105
1.116
3.0
5
0.949
0.961
0.960
0.942
0.936
0.936
0.883
0.938
2.6
6
0.987
0.995
0.982
0.976
0.898
0.988
1.028
0.979
3.7
7
0.891
0.850
0.890
0.903
0.885
0.823
0.808
0.864
4.0
Tabular values are row average divided by overall  average for each
run.
                             -214-

-------
    The Annubars suffered from fouling problems  in  the  static orifice during
the first test.   Automatic purging during the second test relieved the pro-
blem completely.  The ones left in place after the  first test showed no
long term effects from being out of use during the  4^ month interval
between tests other than requiring an initial cleanup,  which took about
one hour for each probe.  Agreement between continuous  Annubar monitoring
data and plant data was excellent.
    Pi tot-static probes were used for traverses  in  both the ducts and the
stack effectively and quickly.
                                    -215-

-------
                                SECTION  IX
                           DISCUSSION OF RESULTS

9.1  FUNDAMENTAL CONCEPTS
     The problem of continuous total  volumetric flow measurement in large,
complex ducts has been considered from the standpoint of both hardware and
technique.  The basic methodology involved measurement of flow at a number
of discrete points and determining the relationship between these measure-
ments and the total flow rate.  The objective of constructing and demonstrat-
ing an accurate, reliable, and economic system with current state of the
art hardware was achieved.
     The current practice in flow measurement is to avoid taking data near
a  large flow disturbance.  The program was deliberately oriented toward
worst case conditions in this regard, since such situations are often
unavoidable in the real world.  Most laboratory data were taken at about
two effective duct diameters downstream from large disturbances, and the
rectangular duct data in the field was taken 0.2 diameters downstream and
0.1 diameters upstream from a large disturbance.  The general class of flow
considered may be defined as nondeveloped with regions of high angularity
and density stratification.
     The general flow environment made it necessary to emphasize basic
physical concepts throughout the effort.  The major one is that mass must
be conserved.  Application of this principle was stressed in two ways.  The
first is that averaging of flow data from a number of discrete points to
obtain  total flow must be done at a single pressure and temperature to be
correct.  This  is why the term "total volumetric flow at standard conditions"
was so  carefully defined  in Section 4, and illustrated further in Appendix A.
It is also the  reason why a recommendation is being made that point velocity
data  in  EPA Method 2  be averaged after conversion to standard temperature
and pressure rather  than  before, and also why it is being recommended that
point velocity  sensors be manufactured with the capability to measure abso-
lute  temperature and  pressure.
      The  other  application of the principle of  conservation  of mass had to
do with flow angularity with  respect to  the duct axis.   In all cases, data

                                   -216-

-------
were taken in a plane normal to the local axis.  As was emphasized in
Section 4, only velocity components normal to this plane, i.e., parallel to
the axis, result in a flow through the plane.  Consequently, test sensors
were analyzed to determine their ability to measure the correct flow com-
ponent.  It was here that the S type probe exhibited its poorest perfor-
mance, and a preference for the ellipsoidal nosed pi tot-static probe became
apparent.
9.2  HARDWARE EVALUATION
     This part of the program was very straightforward.  Most applicable
devices turned out to be point sensors, and of these the Ramapo Fluid Drag
Meter was found to have superior accuracy.  Factory improvements since com-
pletion of laboratory work have also made the probe's survivability charac-
teristics very acceptable.  The Ellison Annubar, being an averaging sensor,
required separate evaluation as a technique.   It has been found to be a
very good flow measurement device, with the limitation that purging of the
rear orifice is required in particulate laden flows.  Results have indicated
that the most cost effective continuous flow measurement system would probably
use an Annubar as the velocity sensor.
     Construction of an acoustic flowmeter for this program proved to be
impractical.   The analysis work and field test results for that subtask have
been presented in the hope that they may lead to a solution of the problem.
9.3  TECHNIQUE EVALUATION
     All methods considered inferred a total  flow rate from measurements at
a small number of discrete points.  The objective of using not more than
eight points  was achieved in both circular and rectangular ducts in situations
where EPA Method 1 would require forty or more points.  Perhaps due to the
extreme cases considered, no true "Universal" method which would require no
in-place calibration was found, although the Log-Linear 4 Method for circular
ducts comes extremely close.  The next best case was achieved, however, for
all acceptable methods - a single in-place calibration.
     Great emphasis was placed on development of the Row Average Method in
rectangular ducts for two reasons.  The first is that it works best immedi-
ately downstream of an elbow, which is highly desirable in the practical
                                   -217-

-------
sense due to the wide use of the rectangular elbow in industry.   The second
reason is hardware considerations.   Use of the Row Average Method requires
no significant modifications of existing ductwork; at most one or perhaps
two ports are required for instrument insertion.   Since data are taken along
a straight line, a single reciprocating point sensor could constitute the
entire system.
     Early during the program, there was concern  that if an acoustic flowmeter
were built, no one knew how to use it in a rectangular duct.  This same con-
cern extended to other advanced concepts such as  laser velocimeters which
are also line averaging devices.  The Row Average Method was clearly the
solution to the problem, and this was another reason for investigating it
so thoroughly.  As these advanced types of flow sensor become more readily
available, it is important to know that there is  now a method for using them
in rectangular ducts.
     The good performance of the Annubar in rectangular ducts was at first
very surprising.  As development of the Row Average method progressed, it
became clear that both techniques tended to work  best in the same type of
situation and for the same reason.   That reason was a consistent relation-
ship between velocity along the line of interest  and the total flow rate.
Excellent performance of the Annubar was best demonstrated during the second
field test.
9.4  FIELD DEMONSTRATIONS
     The second field demonstration involved correction of several problems
which occurred during the first, and showed uniformly good results.  The
major regret is that the Row Average Method could only be evaluated through
analysis of traverse data rather than with an independent hardware system.
Both tests showed clearly that the standard pitot-static probe can be used
very effectively for reference measurements.  It is our belief that in-place
calibration should be performed with maximum accuracy, and that this can be
best achieved through proper use of an ellipsoidal nosed pitot-static probe
for reference traverse measurements.
                                   -218-

-------
9.5  FINAL COMMENTS
     We have confidence in the methodology that has been developed and hope
to see it put to general use.  Since some of the techniques are new, the
data base for them should be expanded.   Techniques were developed to utilize
both existing off-the-shelf hardware and more advanced instruments which
are not yet in common use.  Program results showed that continuous flow
measurement in large ducts is practical  with presently available hardware
and techniques.
                                   -219-

-------
                          SECTION X

                          REFERENCES


1.   "Standards of Performance for New Stationary Sources",
    Environmental Protection Agency; Federal  Register,  Vol.  36,
    No.  159, Part II, 9-17-71.

2.   The Analysis of Physical Measurements, E.  Pugh  and  G.
    Winslow; Addison-Wesley, 1966.

3.   The Measurement of Air Flow, E.  Ower and  R.C.  Pankhurst;
    Pergamon Press, 1966.

4.   Boundary Layer Theory, H. Schlichting; McGraw-Hill, 1968

5.   Fundamentals of Acoustics, I.E.  Kinsler and A.R.  Frey,
    John Wiley & Sons, Inc., 1962.

6.   "A Simplified Integration Technique for Pipe Flow Measure-
    ment", F.A.L. Winternitz and C.F. Fischl; Water Power,  Vol.
    9, page 225, 1957.

7.   Fluid-Dynamic Drag, S.F. Hoerner; published by the author,
    1958.
                              -220-

-------
SYMBOL
  A =
  A =

  a,b,c, =
  CD =

  D =

  I =
  k =
  1 =
  L =
  m =

  Nu =
  P =

  Ap =

  R =

  R =
  r =
  Re =
  S =
  S =
            SECTION  XI
             GLOSSARY
            USAGE
area
absorption coefficient

constants
drag coefficient, ,—~-
drag force

energy decay parameter
general calibration constant
length
acoustic path length
mass flow rate

Nusselt number,  T—
pressure

differential pressure

gas constant

Ramapo Drag Meter output
radius
                  Vd
Reynolds number,
                  H
area
Annubar calibration factor
                                       DIMENSIONS
                                      m
                                      dimensionless
                                      dimensionless
                                      kg-m
                                      sec2
                                      dimensionless
                                      m
                                      m
                                      ks_
                                      sec
                                      dimensionless
                                        kg
                                      m-sec
                                        kg
                                      m-sec
                                        m2
                                      sec2-°K
                                      dimensionless
                                      m
                                      dimensionless
                                      m2
                                      dimensionless
                                -221-

-------
SYMBOL
V =
VVOUT'VIN
V =
at =
         USAGE


absolute temperature

component of velocity vector
parallel to duct axis (axial
velocity component)

magnitude of velocity vector


voltage


volumetric flow rate


energy transfer parameter


acoustic impedance ratio

density
                                       DIMENSIONS


                                      °K

                                      m
                                      sec


                                      m
                                      sec
                                      volts

                                      m3
                                      sec

                                      dimension! ess


                                      dimension! ess
a =
standard deviation

mean value

value at standard atmospheric
conditions
                              -222-

-------
                            SECTION XII
                            APPENDICES

                                                                Page
A.    Averaging and Error Propagation Analysis Techniques        224
B.    Flow Meter Survey                                          229
                               -223-

-------
                            APPENDIX A
        AVERAGING AND ERROR PROPAGATION ANALYSIS TECHNIQUES

Error Propagation Analysis
The following technique is taken from  Reference  2,  "The Analysis  of
Physical Measurements", Chapter 11.  Assume that it is  desired  to
calculate parameter G, which is a known function of variables m,, M
—-Mr, given by
                     G = f(Mr M2, —-, Mr)                        (66)
where f denotes the functional  relationship.   Define the error in
the measurement of variable Mr as x^.   The standard deviation of the
measurement of M  is then given by

                   or2 = J1m   JLlL_                           (67)
where
                   a  = standard deviation of
                    n = number of measurements
By derivation, the standard deviation of G is then given as
The equation in this form is used in Section 2 to perform an error
analysis of a point mass flow measurement.
In the remainder of the report, mean values and standard deviations
are calculated for much of the test data.  For this analysis, if N
measurements are made of parameter G, then the mean value of G is
given by                          ^  G
                          _  =   n=1  n _                         (69)
where                    ~G = mean value
                          G  = value of nth measurement of G
                           n
                          N  = total number of measurements

                               -224-

-------
The standard deviation of G is given by
                        -      -      2
                                                                   (70)
                  b           N
where           o~ = standard deviation of G
                 b
If a known reference value (or assumed value) of G is known, then the
systematic error in the series of measurements is given by

                      E =   G " GREF	  x 100%                   (71)

where                  E = systematic error, %
                       GREF = reference value of G

and the random error, which is a measure of the data spread, is  given
by
-------
    For the type  of work  performed  in  this  program, the system random
    error can be  at least as  important as the  system systematic error.
    While systematic errors are  certainly undesirable, a constant systematic
    error can often be removed  by adequate  system calibration.  On the
    other hand,  large random  errors can be  much more difficult to correct,
    and often imply an unacceptable weakness in the technique and/or
    hardware being used.

2.   Averaging of  Data for Flow  Measurement
    The general mass flow equation  as  given in Section 4 is
                        m  =  //pudA =
                             A
    The purpose of this section  is  to explain  the  type of  average appro-
    priate to the expression   — .   The main  point  to be  made  is that  in
    general
                        —   i- — .  —                             (72)

    For a multi-point traverse,  the  mass flow  rate is approximated  by
                              N
                        m  ~  n=l pn   n   r
    For equal area segments,  this becomes

                        i  .   JL?
    In a flow field where either the density? or velocity u  are  constant,
    the constant term can be removed from inside the summation  sign.   This
    was done for some cases in laboratory testing when the fluid was  air
    and the static temperature and static pressure were quite constant,
    meaning that the density was constant.   These, however,  must be
    considered special cases.  The most desirable practice is to calculate
    either mass flow itself or volumetric flow rate at standard  conditions.
    The latter calculation was performed most often during the program.
    Volumetric flow at actual conditions is not really meaningful unless
    either the fluid density or velocity is constant and known.
                                   -226-

-------
Consider the case of a duct divided into two equal  area segments, with
fluid density and velocity constant in each section.   By definition,
let
                     A = 1
                    PS = 1   (density at standard conditions)
As shown in Section 4,
Thus for this situation, we have
                      Pnun =  :
                     J^-n'^hV^)
Several cases are shown in Table A-l.   When either p,  =  p?  or u-,  =  u?,
all calculated results are the same.   When both  density  and velocity
are stratified, however, the calculation 	 . 	  is in error.  The
                                          P     u
point is that density stratification  must be taken into  account  in
performing traverses in order to minimize error.   This can  be done  by
calculating either total mass flow or average velocity at standard
conditions.  Even the latter calculation is in error when converted to
total mass flow if there is significant molecular  weight stratification
in the flow field.
                               -227-

-------
                             Table A-l.   TOTAL FLOW CALCULATIONS FOR FLOWS WITH VELOCITY AND/OR
                                         DENSITY STRATIFICATION
CASE
1
2
3
4
pl
1
1
1
1
P2
1
2
3
2
Ul U2
1 2
1 1
1 3
2 1
P ^
1 1.5
1.5 1
1.5 1.5
1.5 1.5
p • U
1.5
1.5
2.25
2.25
A m
1.5
1.5
2.5
2.0
*.
1.5
1.5
2.5
2.0
no
ro
CD
I
                                    FLOW CONDITIONS
pl
ul
Al
P2
U2
A2
                                       =A
                                    A = A] + A2 = 1
AVERAGES
                                                                             P2)
                                                                m =  pu  A = ±
                                                                   = u$A
                                                                                        '(u1 +  u2)]A
                     + P2u2)A
                                     P,  =

-------
    APPENDIX  B






Flow Meter Survey
     -229-

-------
    Table  B-l .   KEY  UORD CLASSIFICATION FOR FLOW INSTRUMENTS
       Number
                         Type
          1
          2
          3
          4
          5
          6
          7
          8
          9
        10
        11
        12
        13
        14
        15
        16
        17
        18
                 electromagnetic
                 flow tubes
                 laminar
                 mass
                 non-obstructing
                 nozzles
                 open channel
                 orifice meter
                 pi tot-type
                 positive displacement
                 turbine
                 variable area rotometer
                 Venturis
                 thermal  anemometer
                 mechanical anemometer
                 acoustic
                 vortex shedding
                 other
FLOW METER SURVEY:
     MEDIUM:


     TYPE:
G - gas
L - liquid
S - Solid
Numbers correspond to key words
in Table.
     FILE ACCESSION NUMBER:
                  TRW Systems internal retrieval  system.
                              -230-

-------
                                         FLOW METER  SURVEY
I
PO
00
MANUFACTURER (No.)
MODEL (No.)
ARC Appareils Precision Control
Aeroflex Laboratories, Inc.
Aeroquip, Barco Div. ,
Barrington, Illinois
Air Products & Chemicals, Inc.
Allentown, Pennsylvania
238-2-05121
238-1-04660
Alnor Instrument Co.
Chicago, Illinois
Velometers
Type 606BP
Type 6070P
American Chain & Cable Co.
Waterbury, Connecticut
American Meter Co.
American Standard,
New Brunswick, New Jersey
SG-1
Astro Dynamics, Inc.
Autotronic Controls
Avien, Inc.
Woods ide, New York
Badger Meter Mfg. Co.
Milwaukee, Wisconsin
Type X701
Type T
MEDIUM




G

G
G
G

G
G




L,G
L,G




L,G

L
L
TYPE
10
11, 14, 15
11 14 15
II) ' " > Iw
12

12
12
9, 14, 15

9
14, 15


8, 10, 11

18
18
4, 10
10
I]
l l
2,6,7,8,10,11,13,
18
20
11
FILE ACCESSION NUMBER
614
612, 614


600

600
600
610, 531, 297, 612

610, 531
610, 531


614

600
600
614
614
613
W i w
87, 600, 614

600
600

-------
 I
IN3
CO
(V)
 I
MANUFACTURER (No.)
MODEL (No.)
Bailey Meter Co.
Wickliffe, Ohio
Type BY
Type
Type CV
Beckman Instruments, Inc.
Fullerton, California
Bel fort Instruments Co.
Bendix Environmental Science Div.
B&F Instruments, Inc.
BIF General Signal Corp.
BIF Industries, Inc.
Providence, Rhode Island
Current Meters
Shunt Current Meter
Biotronex Laboratory, Inc.
Blaw-Knox Co.
Copes-Vulcan Div.
Lake City, Pennsylvania
Blue White Industries
Bolt Assoc.
Norwalk, Connecticut
Bowles Fluidic Corp.
Silver Spring, Md.
The Bristol Co.
Haterbury, Conn.
Primary Elements
MEDIUM TYPE
L,G,S 1,3,8,12,18
L,G 18
L 12
S 18

15
G 15
G 11
L,G 2,6,7,8,11,13
L 11
G 11
1
9,10

G 18
L,G 8,13,18
L,G 18
FILE ACCESSION NUMBER
87, 600, 613, 614
600
600
600

612
612
612, 614
87, 613, 614
87
87
614
614
613
702, 703
600, 614
600

-------
ro
OJ
OJ
 i
MANUFACTURER (No.)
MODEL (No.)
Brooks Instrument Div.
Emerson Electric Co.
Hatfield, Pennsylvania
DS-346
DS-HPB
D8800
3600
3300
7100
9400
Brown Instrument Div, Minneapolis-
Honeywell Regulator Co.
Long Island City, New York
Bubble-0-Meter
Temple City, California
Type 1-10-100
CGS Scientific Datametrics
Clean Room Products, Inc.
Climactronics Corp.
Consolidated Controls Corp.
Cox Instruments Div., Lynch Corp.
Detroit, Michigan
Ind Series
Crane Co.
MEDIUM

G,L

L
L
G,L
G,L
L
L
L



G
G
G,L
G


L
L

TYPE

1,4,10,11,12,18

1
11
12,13
12
11
1
18



10,18
18
4,14
14
15
3
11
11
10
FILE ACCESSION

270, 276,

270
270
270, 276
600
600
600
600

613

600, 614
600
257, 264,
612
612
614
600, 613
600
614
NUMBER

600, 614













612, 614







-------
I
ro
CO
MANUFACTURER (No.)
MODEL (No.)
Daniel Industries Inc.
Houston, Texas
PT Meter
Decker Corp.
Bala Cynwyd, Pennsylvania
Degamo Inc.
Dieterich Standard Corp.
Ellison Instrument Div.
New Buffalo, Michigan
E-100
E-101
Series 700
Primary Element
Dietz Henry 6. Co., Inc.
Disa - S&B, Inc.
Dresser Measurement Div.
Dwyer, F.W. Mfg. Co., Inc.
Michigan City, Indiana
500 Series
Rate-Master
Eastech, Inc.
MEDIUM
L
L


G,L
G
G
G,L
G,L

G

G.L
G.L
G.L
G,L
TYPE
2,11
11

12
8,9,15,18
8,9,15
8,9,15
18
18
8
4,14
10
9,12
12
12
19
FILE ACCESSION NUMBER
600, 613, 614
600
613
612, 614
74, 88, 269, 600
88
88
600
600
614
300, 612, 614
614
301. 614
600
600
611, 613, 614
       Plainfield, New Jersey
                    VS-21

       Eclipse Fuel Engineering
       Rockford, Illinois
613

-------
 I
ro
OJ
en
 i
Manufacturer (No.)
MODEL (No.)
Electrosyn Technology Lab. Inc.
Canton, Massachusetts
Bourdon Tube
Ellison Instrument Div. .Dieterich
Standard Corp.
New Buffalo, Michigan
El 00
E101
Series 700
Primary Element
Epic Inc.
Erdco Engineering Corp.
Addison, Illinois
Fischer & Porter Co.
Warminster, Pennsylvania
Swirl meter
Model 10C1505
10A1151
10C1
10D1400
1 OS 1000
Flo-Tech, Inc.
Barrington, Illinois

The Flowcon Co.
Anaheim, California
F+ Series
Flow- Dyne Engineering, Inc.
MEDIUM
G.L
G,L

G,L

G
G
G,L
G,L
G


Gi
>L
G
G.L
G.L
L
L
G

G.L
L

L
L

TYPE FILE ACCESSION NUMBER
18
18

8,9,15,18

8,9,15
8,9,15
18
18
9,14,15


1,2,5,7,8,9,11,12,14
17,18
14,17
11
12
11
1
18

15,11
11

11
11
2,6,8,13
600
600

74, 88, 296, 600, 614

88
88
600
600
612, 614


87,274,532, 600, 611,
613, 614
532, 611
274
600
600
600
600

600, 612, 614
600

600, 614
600
614

-------
ro
GO
en
MANUFACTURER (No.)
MODEL (No.)
Flow Measurement Corp.
Kensington, Maryland
Flow Technology, Inc.
Fluid Data Inc.
Hauppague, New York
UF-100
Flui Dynamic Devices, Ltd.
Ontario, Canada
308
308R
Fluidyne Instrumentation
Foster Engineering Co.
Union, New Jersey
The Fox Valve DEvelopment Co.
The Foxboro Co.
Foxboro, Massachusetts
Model FG
Gelman Instrument Co.
GM Mfg. & Instrument Corp.
Geospace Corp.
Houston, Texas
Gems Co. ,Inc.
Farmington, Connecticut
VFI 700
General Air Drying Div.
MEDIUM


G

L
L
Gl
,L
G
G





G.L
G

G



L
L

TYPE


7,11,15

16
16
1O
o
18
13
10


2,6,8,13

8,11
11
3
15



12
12
9
FILE ACCESSION NUMBER
£ 1 *\
613
612. 614

600
600

534
534
534
614

613
614

365, 600, 613
600
614
612

613

600
600
614

-------
W.UFACTURER (NO.)
  TYpE
                                                                  FILE ACCESSION NUMBER
General Electric Co. ,
Industrial Process Control Div.      L,G
West Lynn, Massachusetts
             Series 553              L,G
             Series 554               G
Gilmont Roger Instruments, Inc.       L,G
Greiner Scientific Manostat
Guiton Industries, Inc.
Vibro-Ceramic Div.
Metuchen, New Jersey
W&LE Gurley
A Teledyne Co.                         L
Troy, New York
             622                      L
             655                      L
Hagan Corp.
Pittsburgh, Pennsylvania
Halliburton Co.
Duncan, Oklahoma                      L

Hallikainen Instruments
Richmond, California
Hamilton-Standard,
Windsor Locks,  Connecticut            b
Hays Corp.
Michigan City,  Indiana
   18

   18
   18

   12

   12
15,11

   11
   11
   ''
   lb
                                                                       600,  613

                                                                       600
                                                                       600

                                                                       600,  612

                                                                       614


                                                                       613
                                                                       600,  612

                                                                       600
                                                                       600

                                                                       6I3

                                                                       60°
                                                                       609

-------
ro
CO
oo
MANUFACTURER (No.)
MODEL (No.)
Hastings-Raydist
Hampton, Virginia
Gas Flow Probe
No.AFI Series
LF&HF Series
OF Series
Hersey-Sparling Motor Co.
Dedham, Massachusetts
Hoke Inc.
Honeywell, Inc.
Industrial Div.
Fort Washington, Pennsylvania
Diaphragm
Bellows
Hughes Instrument
Bristol, Pennsylvania
Industrial Measurements Controls
410-3
Industrial Physics & Electronics
Interval Corporation
Agoura, California
Model 44
In-Val-Co. Combustion
Engineering, Inc.
Tulsa, Oklahoma
S-Series
W-Series
MEDIUM
6
G
G
G




L.G

L.G
L,G


L
L

G
G

L

L
L
TYPE
2,3,4,8,9,14
FILE ACCESSION NUMBER
88, 89, 185, 209, 210, 216, 227,
231, 305, 600, 612, 613
88, 89, 185, 209, 210, 216, 227,
231, 612, 614
14
14


2

18

18
18


10
10
5
16
16

11

11
11
600
600
613

614

600

600
600
613

600,
600
614
87,
87,

600,

600
600












614


530
530

614




-------
ro
oo
MANUFACTURER (No.)
MODEL (No.) MEDIUM
ITT Barton Controls & Instr.Div. . r
Monterey Park, California '
F-500 L,G
Ion Exchange Products
Kenics Corp.
King Engineering Corp.
Kingmann-White Inc.
Placentia, California '
515 L,G
509 L,G
Kontes Glass Co. . _
Vine land, New Jersey '
Delta P G
K627900 L,G
K West Corp. p
Westminster, California
Model 325 G
Laboratory Data Control
Larson Aero Development
Leeds & Northrop Co. L,G
North Wales, Pennsylvania
1911 L,G
1913 G
Leopold & Stevens Inc.
Librascope Group, General
Precision Systems, Inc. G
Glendale, California
L60-1 G
TYPE
10,11
10

10
12
8,18
18
18
2,12,18
18
12
16
16
2,10
1,4,9,11,12,13
1

1
1
7

18

18
FILE ACCESSION NUMBER
600,
600

614
614
600,
600
600
600,
600
600
228,
228,
614
614
600,

600
600
614

600

600
614




614


614


229
229


614









-------
o
 p
MANUFACTURER (No.)
MODEL (No.)
Linden Labst
State College, Pennsylvania
Turbine
Rotary
Lindsey Meter Co.
Pasadena, California
Turbine
Rotary
Link-Belt, FMC Corp.
Lion Precision Inc.
LKB Instruments Inc.
Liquid Controls Corp.
North Chicago, Illinois
M-3
M-70
Lube Devices Inc.
Mace Corp.
South El Monte, California
960 Series
Manostat Corp.Div. Greiner
Scientific Corp.
New York, New York
36-541-03
36-541-04
36-541-05
36-541-30
Marine-Electro Mechanical Inc.
Marotta Scientific Controls, Inc.
MEDIUM
L.G
L,G
G
L,G
L,G
G



L.G
L
L

L,G
L.G

L,G

L,G
L,G
L,G
L,G


TYPE
11,12
11
12
11,12
11
12
6
3
2,5
10,18
10
10
2,5,8
19
19

12

12
12
12
12
12
13
FILE ACCESSION NUMBERS
600, 613
600
600
600
600
600
614
614
614
87, 600, 614
600

614
600
600

600

600
600
600
600
614
614

-------
 I
ro
MANUFACTURER (No.)
MODEL (No.) MEDIUM
Martek Instruments, Inc.
Martin-Decker Corp.
Matheson Gas Products Div.
600 Series L
8112 Series G
8116 Series G
Meco Modern Engineering Co. ,Inc.
St. Louis, Missouri
6401 Argon
Mec-0-Matic Inc. ,
Miami, Florida
251 L
Maxson, W.L. Corp.
New York, New York
Meriam Instruments Co.
Cleveland, Ohio
Laminar Flow
Metrology Research Inc.
Altadena, California
Micron Instruments Inc.
McNeil Corp.
Modern Engineering Co., Inc.
Moore Products Co.
Spring House, Pennsylvania
Narco Bio-Systems Inc.
TYPE
15
4
2,4,12,18
12
18
4
12
12
18
18

3,4,8,9
3

1

10
1
1
FILE ACCESSION NUMBERS
612
614
600, 614
600
600
600
600
600


600, 613, 614
600
613
614

614
614
614

-------
 I
PO
PO
MANUFACTURER (No.)
MODEL (No.)
National Instrument Lab, Inc.
General Kinetics Inc.
Rockville, Maryland
10-20-30
20-10-250
10-R-100
New Jersey Meter Co.
Plainfield, NEw Jersey
Neptune Meter Co.
Long Island City, New York
Norrich Plastics Corp.
Nusonic Corporation, Electronics
System Div.
Paramus, New Jersey
6155
Panametrics Inc.
Waltham, Massachusetts
Penn Instrument Div.
Burgess Manning Corp.
Philadelphia, Pennsylvania
Permutit Co.
Potter Aeronautical Corp.
Union, New Jersey
Series 3000
Quantachrome Corp.
MEDIUM
G
G
G
G

L

G
G

L,G

L
L

TYPE
3:18
18
18
18

10,11
6
16
15

4
2,6,7,13
4,11
4
12
FILE ACCESSION NUMBERS
600, 614
600
600
600
613
600, 614
614
612
612

87, 613
614
87, 600
600
614

-------
ro
-P.
CO
MANUFACTURER (No.)
MODEL (No.)
Quantum Dynamics Inc.
Tarzana, California
QL 432 WRG 1C
QL 448 WRC G
Ramapo Instrument Co., Inc.
Bloomingdale, New Jersey
X
Mark V
Republic Flow Meter Co.
Chicago, Illinois
Research Appliance Co.
Allison Park, Pennsylvania
Revere Corp. of America
Wallingford, Connecticut
Robinson Orifice Fitting Co.
Houston, Texas
R
Rockwell Mfg. Co.
Bala Cynwyd, Pennsylvania
Eureka B
Series L
Series M
Rotron Controls Div
Woodstock, New York
P Series
Ross Inc.
Matheson Gas Products Div.
East Rutherford, New Jersey
MEDIUM
L,G
G
L
L,G
L,G
L,G





L,G
L,G
L
L
L
L
L,G
L,G



TYPE
4,11,18
18
18
12,19
12
18


8,9,12


6,13,18
18
11
11
11
11
4,13
13



FILE ACCESSION NUMBERS
600, 614
600
600
600, 613, 614
600
600
C1 *5
613
614
Cl O
0 1 o
600, 614
600
600
600
600
600
600, 613
600




-------
ro
MANUFACTURER (No.) men T MM
MODEL (NO.) MEDIUM
Saratoga Systems Inc. . n
Cupertino, California '
Ultrasonic Flow . r
Ultrameter L'
Scarpa Laboratories Inc. , Q
Metuchen, New Jersey '
Ultrasonic Flowmeter
SFSM-5-RPS L,G
SFM-4-RPS L,G
SFM-3-BP L,G
SFS-2-RPS L,G
Schroeder Brothers Corp. ^
McKeesport, Pennsylvania
Schutte & Koerting Co. . Q
Cornwell Heights, Pennsylvania *
SK-ST Flo-thrutronic G
18000 Series L,G
19000 Series L,G
Selas Corp. of America
Sentry Equipment Corp.
Oconomowoc, Wisconsin
Signet Controls Inc.
Simerl, R.A.
Simmonds Precision Instrument Co.
Tarry town, New York
Simplex Valve & Meter Co.
Lancaster, Pennsylvania
TYPE
16
16
3,4,5,16

16
16
16
16
13
1 ^J
12
12
12
12
11
11

8
15
11



FILE ACCESSION NUMBERS
449,
449,
446,

446,
446,
446,
446,
600

273,
273
600
600
614
600,

614
612
613,
\s t *r f
613

450
450
448, 601-608, 614

448, 601-608
448, 601-608
448, 601-608
448, 601-608


600




614



614




-------
ro
-p»
on
MANUFACTURER (No.)
MODEL (No.)
SK Instruments Div.
SMI Instruments Div.
Smith, A.O. Corp. Meter Systems
Standard Controls
Seattle, Washington
Statham Instruments, Inc.
Oxnard, California
E-3000
SP2200
Taylor Instrument Co. .Process
Control Div.,Sybron Corp.
Series 300
Series 750
Series 737
Technol ogy/Versatroni cs
Yellow Springs, Ohio
MFC
NL
Temp line
Tescom Instruments Div.
Thermal Instruments Co.
MEDIUM TYPE
2,12
2,12
10,11


i
\
L 1
L 1

L,6 1,5,8,9,13,15,18
L,G 18
L.G 18
L 1
4
4
4
9
8
3,4,5,14
FILE ACCESSION NUMBERS
617
614
614
f* T O
613
cnr>
ODD
600
600

226, 612, 613, 614
600
600
600
600, 614
600
600
614
614
612, 614

-------
ro
.£.
cn
MANUFACTURER (No.)
MODEL (No.)
Thermo Systems Inc.
St. Paul , Minnesota
VT-163
VT-161
VT-160
P-2
10
20
1223 & 1229
352 IG
352 IL
Thermoneti cs Corp.
TIP Instruments
Tower Systems Co.
Trans-Sonic Inc.
Lexington, Massacusetts
Trimont Instrument Co.
Tyland Corp.
Torrance, California
FMS 311
RP 781
United Sensor Control Corp.
Wallace & Tiernan Inc.
Belleville, New Jersey
Varea-Meter Bypass
NA-11 & 12
MEDIUM
L,G

G
G
G
G
L,G
L,G
G
G
L







G
L
G
L,G
L,G
UG
TYPE
14

14
14
14
14
14
14
14
14
14
14
15
15


8,9
4,14,18
14
18
2,9,13
12
12
12
FILE ACCESSION NUMBERS
87, 88, 218, 219, 294, 600, 612.
613, 614
219
219
219
218
218
218
294
600
600
612
612
612
/" 1 *1
613
614
600, 614

600
307, 614
249, 600, 614
249
600
        Waugh Engineering Co.
        Van Nuys, California
613

-------
        MANUFACTURER (No.)
                     MODEL (No.)
urnT,,u
MEDIUM
TYPE
FILE ACCESSION NUMBERS
        Waukee Engineering Co.
        Milwaukee, Wisconsin

        Westberg Mfg.  Co.

        Westinghouse Computer Instrument
        Hagan Computer System Div.
        Pittsburgh, Pennsylvania
                     3000  Series
        Wilcoxin Research
        Bethesda, Maryland

        Wilmad Glass Co. ,Inc.
                  15


                6,7,18

                  18
                  600, 614

                  612


                  600, 614

                  600

                  613

                  614
PO

-------
                                 TECHNICAL REPORT DATA
                           (Please read Inslructions on the reverse before completing)
 1. REPORT NO.
  EPA-650/2-75-020
                            2.
                                                        3. RECIPIENT'S ACCESSION-NO.
 4. TITLE ANDSUBTITLE
 Continuous Measurement of Total Gas Flowrate
  from Stationary Sources
                              5. REPORT DATE
                              February 1975
                             6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
                                                        8. PERFORMING ORGANIZATION REPORT NO
 E.F. Brooks, E.C.Beder, C.A.Flegal, D.J.Luciani,
   and R.Williams
 9. PERFORMING OR8ANIZATION NAME AND ADDRESS
 TRW Systems Group
 One Space Park
 Redondo Beach,  CA  90278
                                                        10. PROGRAM ELEMENT NO.
                              1AB013; ROAP 21ACX-AE
                              11. CONTRACT/GRANT NO.
                              68-02-0636
 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 PI
                              Final; 10/72-12/74
                                                                         PERIOD COVERED
                              14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
          The program objective was to evaluate hardware and techniques for the
  continuous measurement of the total gas flowrate from stationary sources, speci-
  fically in large or complex ducts where total flow metering  devices  such as  plate
  orifices  are not practical.  Work consisted of  formulation of operating specifica-
  tions,  evaluation  of commercially available  velocity sensors,  development and
  evaluation of flow mapping techniques,  and field demonstration of both hardware
  and technique.  Results showed that total volumetric flowrate  can be measured
  with accuracies consistently better than 10% in either circular or rectangular
  ducts through proper placement of from one  to eight flow sensors, when standard
  traversal techniques would require twenty to fifty traverse points.   The rectangu-
  lar duct  mapping  techniques developed during the program were found to have
  optimum accuracy immediately downstream  of an elbow.  Several off-the-shelf
  velocity  sensors were found acceptable for use in the specified stack-type
  environment.   The field demonstrations verified the acceptability of both
  hardware and techniques.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.IDENTIFIERS/OPEN ENDED TERMS
                                          c. COSATI Field/Group
 Air Pollution
 Velocity Measurement
 Gases
 Flow Rate
 Flowmeters
 Flow Distribution
Ducts
Air Pollution Control
Stationary Sources
Continuous Measurement
Velocity Sensors
13B,
14 B

20D
13K
 8. DISTRIBUTION STATEMENT

 Unlimited
                 19. SECURITY CLASS (This Report)
                 Unclassified
                         21. NO. OF PAGES
                               262
                                           20. SECURITY CLASS (Thispage)
                                           Unclassified
                                          22. PRICE
EPA Form 2220-1 (9-73)
                                        -248-

-------