EPA-650/2-75-062
August 1975 Environmental Protection Technology Series
REMOTE MEASUREMENT
OF POWER PLANT SMOKE STACK
EFFLUENT VELOCITY
UJ
o
-------
EPA-650/2-75-062
REMOTE MEASUREMENT
OF POWER PLANT SMOKE STACK
EFFLUENT VELOCITY
by
C. R. Miller and C. M. Sonnenschein
Raytheon Company
528 Boston Post Road
Sudbury, Massachusetts 01776
Contract No. 68-02-1752
ROAP No. 26AAP-85
Program Element No. 1AA010
EPA Project Officer: William F. Herget
Chemistry and Physics Laboraotory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U. S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D. C. 20460
August 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. SOCIOEGONOM1C 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.
Publication No. EPA-650/2-75-062
11
-------
ABSTRACT
This report describes the successful demonstration of the abil-
ity a C02 Laser Doppler Velocimeter to measure remotely the velocity
of the effluent from a power plant smoke stack. The basis of the tech-
nique is that laser radiation backscattered from particulates in the
effluent is Doppler shifted in frequency in proportion to the veloc-
ity of the effluent.
Measurements were made against a coal burning power plant
equipped with electrostatic precipitators to remove particulates
from the boiler flue gases. The measurement site was approximately
400 m slant range from the stack. Backscattered signals from the
stack effluents were detected, processed and recorded on magnetic
tape. The taped data was analyzed to determine: (1) agreement
between LDV and in-stack velocity measurements, (2) correlation of
backscatter signal strength and the cross-stack optical transmission,
(3) estimates of the effluent backscatter coefficient at 10.6-jam, (4)
profiles of the stack exit velocity distribution at various heights
above the stack lip, and (5) the effect of turbulence on the back-
scattered Doppler spectra.
As a result of exit velocity measurements in the 25 to 45 m/sec
range, it was concluded that an LDV can remotely measure stack exit
velocities to an accuracy of at least 1.5 m/sec. The potential for
making particulate concentration measurements with the same instru-
ment was also demonstrated by establishing a relationship between
the intensity of the scattered radiation and the optical opacity of
the exit gases.
Based on the results of the measurements a study on the design
of an LDV optimized for the measurement of power plant effluent
velocities was performed.
This report was submitted in fulfillment of EPA contract number
68-02-1752 by Raytheon Company, Equipment Division, Electro-Optics
Department carried out under the joint sponsorship of the Environ-
mental Protection Agency, and the National Aeronautics and Space
Administration. Work was completed as of June 1975.
iii
-------
TABLE OF CONTENTS
Section
1 INTRODUCTION AND SUMMARY 1-1
2 SYSTEM DESCRIPTION 2-1
2.0 Introduction 2-1
2.1 CO2 Laser 2-1
2.2 The Interferometer 2-6
2.3 Telescope 2-7
2.4 Detector and Receiver Electronics .... 2-7
2.5 Optical Scanner 2-8
2.6 Signal Processing 2-11
2.7 The Laser Test Van 2-11
3 RESULTS OF THE FIRST FIELD TESTS 3-1
3.0 Introduction 3-1
3.1 Processing Methods 3-3
3.2 Experimental Results 3-6
3.2.1 Tape #1, Run #1 3-8
3.2.2 Tape #2, Run #1 3-8
3.2.3 Tape #2, Run #2 3-8
3.2.4 Tape #3, Run #1 3-10
3.2.5 Tape #4, Runs #1 & 2 3-15
3.2.6 Tape #5, Run #1 3-18
3.2.7 Tape #6, Run #1 3-18
3.2.8 Tape #6, Run #2 3-21
3.2.9 Tape #7, Run #1 3-21
3.2.10 Tape #7, Run #2 3-25
3.3 Conclusions on the First Field Tests. . . 3-29
4 RESULTS OF THE SECOND FIELD TESTS 4-1
4.0 Introduction 4-1
4.1 System Modifications 4-1
IV
-------
CONTENTS (Continued)
SECTION PAGE
4.1.1 The Modified Mach-Zehnder Interferometer 4-2
4.1.2 The Lead Tin Telluride Detector .... 4-2
4.1.3 The Frequency/Intensity Tracker .... 4-3
4.1.4 The Time Code Generator 4-7
4.2 The Second Field Tests 4-7
4.2.1 Tape #1: 8° CW Run 4-12
4.2.2 Tape #2: 20° CW Run 4-12
4.2.3 Tape #3: 20° Precipitator Run 4-14
4.2.4 Tape #4: 8° Velocimeter Profile . . . 4-17
4.2.5 Tape #5: Velocity Variation by Power
Load Change 4-17
4.2.6 Tape #6: 20° Precipitator Run 4-19
4.2.7 Tape #7: 28° CW Run 4-23
4.2.8 Tape #8: 37° CW Run 4-23
4.2.9 Tapes #9, 10, 11, and 12: 8° Velocity
Profiles 4-26
4.3 Data Analysis 4-29
4.3.1 LDV and Pitot Tube Velocity Measurements 4-29
4.3.2 Signal Strength as a Function of Cross-
Stack Transmission .......... 4-32
4.3.3 The Effluent Backscatter Coefficient. . 4-38
4.3.4 Turbulence Effects 4-43
5 SYSTEM ANALYSIS 5-1
5.1 Introduction 5-1
5.2 Heterodyne Detection 5-1
5.3 Signal to Noise Ratio 5-2
5.4 Beam Size 5-8
5.5 Bandwidth Considerations 5-9
6 CONCLUSIONS 6-1
7 REFERENCES 7-1
-------
ILLUSTRATIONS
FIGURE PAGE
2-1 Block Diagram of the EPA Effluent Dection System 2-2
2-2 Optical Layout of the EPA Effluent Detection System 2-3
2-3 Photograph of Laser Doppler Velocimeter 2-4
2-4 Raytheon Model LS10A C02 Laser 2-5
2-5 Mach Zehnder Interferometer Configuration 2-6
2-6 A Liquid Helium Cooled Coppler Doped Germanium
Detector 2-9
2-7 Scanning Mirror 2-10
2-8 Raytheon Trailer 2-13
2-9 Raytheon Trailer 2-13
3-1 The Raytheon Data Van at the Duke Power River Bend
Steam Station 3-2
3-2 Smoke Stack Geometry for LDV Effluent Velocity
Measurements 3-4
3-3a Typical Composite Photograph of 2 Second Average
Spectrum Analyzer Signals 3-5
3-3b Typical Photograph of 7 Second Average Spectrum
Analyzer Signals 3-5
3-4 Tape #2 Run #1 Duke Power, 28 August 1974, 0.3 m Above
Stack - 28 Second Average 3-9
3-5 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of the
Stack 3-11
3-6 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of the
Stack 3-12
3-7 A Comparison of the 2 Second and 7 Second Average
Effluent Velocity Profile Data 3-13
3-8 Comparison of the 2 Second and 7 Second Averaged
Velocity Data for the Power Load Reduction 3-14
vi
-------
ILLUSTRATIONS (Continued)
FIGURE PAGE
3-9 Tape #3 Run #1, Duke Power, 28 August 1974,
Power Reduction, 28 Second Average 3-16
3-10 Tape #4 Run #1 & 2, Duke Power, 29 August 1974,
Directly Above Stack, 28 Second Average 3-17
3-11 Tape #5 Run #1, Duke Power, 29 August 1974, Directly
Above Stack, 28 Second Average 3-19
3-12 Tape #6 Run #1, Duke Power, 29 August 1974, Directly
Above Stack, 28 Second Average 3-20
3-13 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of the
Stack 3-22
3-14 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of
the Stack 3-23
3-15 A Comparison of the 2 Second and 7 Second Average
Effluent Velocity Profile Data 3-24
3-16 The Average Velocity of Effluents from a Smoke
Stack as a Function of the Distance from the Center
of the Stack 3-26
3-17 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of
the Stack 3-27
3-18 A Comparison of the 2 Second and 7 Second Average
Effluent Velocity Profile Data 3-28
3-19 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of
the Stack 3-30
3-20 The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of
the Stack 3-31
3-21 A Comparison of the 2 Second and 7 Second Average
Effluent Velocity Profile Data 3-32
4-1 Tracker Outputs of Integrated Signal Intensity and
Effluent Exit Velocity as a Function of Time 4-5
vii
-------
ILLUSTRATIONS (Continued)
FIGURE PAGE
4-2 Tracker Output Showing Effluent Exit Velocity
Profile Across the Smoke Stack Lip 4-6
4-3 Block Diagram of LDV System for Second Field Tests 4-8
4-4 Smoke Stack LDV Geometry 4-10
4-5 The Strength and Frequency of Effluent Signals as
a Function of Time for Tape #1: 8°C Run 4-13
4-6 The Strength and Frequency of Effluent Signals as
a Function of Time for Tape #2: 20° CW Run 4-15
4-7 The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #3: 20° CW Run 4-16
4-8 The Strength and Frequency of Effluent Signals as a
Function of the Position above the Smoke Stack Lip
from Tape #4: 8° Velocity Profile 4-18
4-9 The Frequency of Effluent Signals as a Function of
Time for Tape #5: 8° Power Load Change 4-20
4-10 The Frequency of Effluent Signals as a Function of
Time for Tape #5: 20° Power Load Change 4-21
4-11 The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #6: 20° CW Run 4-24
4-12 The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #7: 28° CW Run 4-25
4-13 The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #8: 37° CW Run 4-27
4-14 The Frequency of Effluent Signals as a Function of
The Position and Height Above the Smoke Stack Lip
From Tapes #9, 10, 11, and 12: 8° Velocity
Profiles 4-28
4-15 Effluent Exit Velocity as a Function of Power Load 4-31
4-16 Effluent Exit Velocity Measured by a Laser Doppler
Velocimeter as a Function of the Exit Velocity
Measured by an In-Stack Pitot Tube 4-33
Vlll
-------
ILLUSTRATIONS (Continued)
FIGURE PAGE
4-17 Relative Integrated Doppler Signal Strength as a
Function of the Effluent Attenuation Coefficient 4-37
4-18 Effluent Backscatter Coefficient as a Function
of the Cross-Stack Transmission 4-41
4-19 The Effluent Backscatter Coefficient Plotted
as a Function of the Optical Transmission at the
Smoke Stack Exit 4-42
4-20 Theoretical In-Stack and Exit Velocity Profiles 4-44
4-21 Typical Doppler Spectra from Smoke Stack
Effluents at Various Laser Elevation Angles 4-48
4-22 Relative Turbulence Intensities in Pipe Flow.
(Laufer, J.; Reprinted from NACA Tech. Repts.
1174, pp. 6 and 7, 1954) 4-49
5-1 System Block Diagram 5-1
5-2 System Signal-to-Noise Ratio for Range = 250 m 5-5
5-3 System Signal-to-Noise Ratio for Range = 500 m 5-6
5-4 System Signal-to-Noise Ratio for Range = 1000 m 5-7
ix
-------
LIST OF TABLES
TABLE
3-1 Summary of Data Tapes 3-7
3-2 A Comparison of Pitot Tube Velocity Data With The
Doppler Velocity Data 3-33
4-1 Summary of Data Tapes from the Second Field Tests 4-11
4-2 Comparison of In-Stack and LDV Velocity Measure-
ments For Tape #5 on 17 January 1975 4-22
4-3 Velocity vs. Load Data 4-30
4-4 Smoke Stack Effluent Particle Concentration and
Emission Data. Data Taken from Environmental
Science and Engineering, Inc. Measurements Under
EPA Contract No. 68-02-0232, Task No. 45, Sub-
Task No. 3, 26 - 30 August 1974 4-35
4-5 Relative Integrated Doppler Signal Strength as a
Function of Cross-Stack Transmission and the Optical
Attenuation Coefficient. Data from Tape #3, 16
January 1975 and Tape #6, 17 January 1975 4-36
4-6 Data on Doppler Spectra of Smoke Stack Effluents 4-49
5-1 Trades Between Power and Optics Size No Integration 5-4
5-2 Trades Between Power and Optics Size, Integration 5-8
-------
ACKNOWLEDGEMENTS
The work effort in this project has involved a number of
companies and governmental agencies. The authors would like to
acknowledge the extensive technical assistance of R. E. Schaaf and
the managerial direction of A. V. Jelalian of the Electro-Optics
Department of Raytheon Company's Equipment Division, Sudbury
Engineering Facility. The analysis of turbulence effects was per-
formed under sub-contract by J. A. L. Thomson of Physical Dynamics,
Inc. The cooperation of the personnel of Duke Power's River Bend
Steam Station in support and performance of the experimental measure-
ments is gratefully acknowledged. The authors would like to thank
W. p. Herget and R. Rollins of the EPA's NERC, for their technical
assistance. Acknowledgement is made to Environmental Science
and Engineering, Inc., for the in-stack measurements. The contri-
bution of R. M. Huffaker of NASA, MSFC, as a technical advisor to the
program is greatly appreciated.
XI
-------
SECTION 1
INTRODUCTION AND SUMMARY
Most conventional methods for monitoring particulate flow rates
from stationary sources such as power plants are expensive and quite
time-consuming. These methods require the use of highly specialized
teams to install monitoring equipment on a stack, make the required
measurements, and analyze the data. New methods for stationary
source monitoring are now being developed which are based on remote
measurement of the electromagnetic properties of the polluting species
in the stack effluent. Since the new instrumentation can be operated
without even requiring access to the power plant property, a complete
test can be done with a minimum of preparation and time.
Two types of measurements are needed to calculate a mass flow
rate by remote techniques: species concentration and velocity.
The U. S. Environmental Protection Agency's Chemistry and Physics
Laboratory and Stationary Source Enforcement Division are currently
evaluating various remote measurement techniques. Research over the
past ten years by NASA and other government agencies has shown that
remote measurement of wind velocities at ranges of a kilometer are
possible using a Laser Doppler Velocimeter (LDV). The basis of the
technique is that laser radiation scattered from particulates in the
air is Doppler shifted in frequency, and measurement of this Doppler
frequency shift yields the velocity of the particulates and hence of
the wind. The method has excellent potential for use in the deter-
mination of smoke stack gas exit velocities and particle concentra-
tions. The instrumentation for the remote wind velocity measurement
is in existence. Particulate concentration measurements can be made
with the same instrument by relating the intensity of the scattered
radiation to the emission concentration.
The objective of this program was to prove the feasibility of
remote measurement of smoke stack.velocity using an LDV. To
accomplish this a C0« Laser Doppler Radar system was assembled into
1-1
-------
a mobile van, and measurements were made on an EPA instrumented
smoke stack at the River Bend Steam Station of the Duke Power
Company in Mt. Holly, North Carolina. This facility is a
coal burning power plant with four boilers. Each boiler drives a
turbine generator capable of producing an output up to 150 MW. Each
boiler unit is equipped with electrostatic precipitators which re-
move over 99% of the particulate matter from the smoke. The measure-
ment site was approximately 400 meters slant range from the stack
exit. The elevation angle to the stack lip was about 8 . In-stack
measurements of the flue gas velocities and the cross-stack optical
transmission were supplied by the EPA for comparison to the remote
data.
After alignment and calibration of the LDV system at the Duke
Power site, backscatter signals, representing the particulate char-
acteristics of the effluents from a smoke stack, were detected, pro-
cessed, and recorded on magnetic tape for further analysis. Approx-
imately 5,500 meters of taped data were recorded during the first
field test period of 24 to 31 August 1974. Another 11,000 meters
of taped data were recorded in the second field test period of 12 to
19 January 1975. The data vere analyzed to determine (1) profiles of
the stack exit velocity at various heights above the stack lip, (2)
stack exit velocities as a function of the in-stack velocities, (3)
effluent backscatter coefficients at 10.6 urn, (4) correlation of
backscatter signal strength and cross-stack optical transmission, and
(5) the effect of turbulence on the backscattered Doppler spectra.
The results of the measurements definitely prove the feasibility
of using a LDV to remotely measure smoke stack effluent velocities.
The velocity data from the LDV, taken at the top of the stack, and
the in-stack velocity data, taken from pitot tube traverses at the
base of the stack, agreed to within 14%. It is thought that the
majority of this error is due to miscalibration (e.g. in the pitot-
tube velocity measurements or in the measurement of the LDV elevation
angles.) and that the actual LDV measurement accuracy is about 1.5 m/
sec for stack exit velocities in the 25 to 40 m/sec range. Estimates
1-2
-------
for the backscatter coefficient at 10.6-um vary from 10 to 10
-1
m
depending on the number of precipitators in use. Definite
correlations exist between the 10.6-um backscatter signal strength
and the attenuation coefficient measured in the visible. Since a
correlation can be made between the attenuation coefficient and mass
concentration, the LDV can be used to measure the mass concentration
exhausting from a given smoke stack. Turbulence effects were shown
to flatten the exit velocity distribution across the top of the
stack. This flattening relaxes the spatial alignment requirements
necessary for an accurate velocity measurement. The turbulent
velocities in the stack broaden the backscattered Doppler spectra.
The broadening is relatively independent of elevation angle, and it
was found that suitable effluent velocity measurements can be made
with elevation angles between 8° and 45°.
The LDV system used in the first field tests is described in
detail in Section 2. Section 3 presents the results of the first
experimental phase, while Section 4 describes various system changes
and gives the results of the second set of field tests. The design
studies for an optimized LDV for smoke stack emission measurements
are presented in Section 5. Conclusions and recommendations for
additional work are proposed in Section 6.
1-3
-------
SECTION 2
SYSTEM DESCRIPTION
2.0 INTRODUCTION
A block diagram of the LDV system used to make effluent velocity
measurements is shown in Figure 2-1. The optical portion of the
system consists of a 20 watt CO- laser, a Mach-Zehnder interferometer,
a 30 cm diameter, f/8 telescope and a copper-doped germanium de-
tector, all of Raytheon manufacture. This equipment is shown
schematically in Figure 2-2 and pictorially in Figure 2-3. These
equipments are mounted on an aluminum base which is in turn shock
mounted on a support table to protect against vibration. The trans-
mitted beam is directed by a Raytheon scanner utilizing an ellipti-
cal flat with a 46 cm major axis. The signals are viewed on a
Hewlett Packard 8552B/8553B Spectrum Analyzer for visual monitoring
and processed by a Raytheon Frequency Tracker. Data from the spectrum
analyzer and frequency tracker as well as a voice channel was re-
corded on a Precision Instruments Magnetic Tape Recorder. These
various components, as well as the Raytheon laser test van, are
described in the following sections.
2.1 CO., LASER
r~~1- £^ ' ^^"»»
The Raytheon Model LS10A C02 laser shown in Figure 2-4 was a
water cooled semi-sealed unit having a nominal output of 20 watts.
Sealing of the system was through a vacuum valve which permitted re-
filling of the laser in the field if necessary. Refilling was
accomplished using the gas bottle and pumping station located in the
power unit which also contained the laser power supply, the closed
cycle laser cooling system and a high voltage supply for tuning the
laser by PZT control of the cavity length.
The laser head contained a split discharge tube closed off with
non-hygroscopic zinc selenide windows mounted at Brewster's angle.
The orientation of the windows determined the polarization of the
2-1
-------
LASER
POWER
SUPPLY
M
I
INTER-
FEROMETER
TELESCOPE
DETECTOR
AMPLIFIER
BIASING
CIRCUITRY
SPECTRUM
ANALYZER
FREQUENCY
TRACKER
MICRO-
PHONE
SCANNER
TARGET
7 CHANNEL
TAPE
RECORDER
Figure 2-1. Block Diagram of the EPA Effluent Detection System
-------
SCANNER MIRROR
M
f/8 TELESCOPE
INTERFEROMETER
C02 LASER
return path
ALIGNMENT
TELESCOPE
I
CD LENS
6
DETECTOR
Figure 2-2.Optical Layout of the EPA Effluent Detection System
-------
RAYTHEON COMPANY
EQUIPMENT DIVISION
Figure 2-3. Photograph of Laser
Doppler Velocimeter
2-4
-------
Ul
H
I
m
O
z
o
O
^
TJ
Figure 2-4. Raytheon Model LS10A CO2 Laser
-------
output beam. The discharge tube was mounted between a plane output
mirror and a four meter radius of curvature rear reflector. The
mirrors and discharge tube were mounted on a frame utilizing four
Invar bars for thermal stability. The cavity configuration and
mechanical structure assured a stable TEM output beam.
oo
2.2 .THE INTERFEROMETER
A Mach Zehnder interferometer was used in the LDV system.
The configuration is shown in Figure 2-5.
Mirror
transmit path
From
BS3
BS1 BS2
/ local oscillator path \
To
Telescope
receive path
To
Laser
Detector
Figure 2-5. Mach Zehnder Interferometer Configuration
The laser beam enters from the left, is reflected by the beam-
splitter (BS1) and mirror and goes through the beamsplitter BS3 to
the .telescope. The signal beam, which consists of that portion of
the scattered light which has retained the polarization of the out-
put beam, comes from the telescope, is reflected by BS3 and BS2 to
the detector where it is mixed with the local oscillator beam which
has come from the laser through BS1 and BS2. The beamsplitters
2-6
-------
BS1 and BS2 are chosen so that the product of their transmission
yields the correct local oscillator level for the detector being
used.
In order to eliminate the need for independently maintaining
alignment of the various mirrors and beamsplitters, Raytheon con-
structed interferometers from precision.machined blocks of solid
Invar. This resulted in an interferometer which is mechanically
rigid and thermally stable and which is aligned by a simple tilt,
rotation or translation of the interferometer as a unit rather than
of each of the components.
2.3 TELESCOPE
The telescope was a reflecting telescope of the Cassegrain type.
It is partially visible in Figure 2-3 and shown schematically in
Figure 2-2. The primary is a 30 cm diameter, f/8 spherical ip-'-ror
and is fed by a 2.54 cm diameter spherical secondary. Focussing of
the telescope is achieved by a calibrated translation of the secondary
on a micrometer driven stage. Both components of the telescope
are mounted on the aluminum shock mounted optical table.
The use of reflecting elements in the telescope allowed it to
be aligned in the visible as well as aimed visually, since reflecting
elements aligned in the visible are also aligned at the laser wave-
length in the infrared. The use of reflecting elements also permitted
aiming of the system without transmitting a laser beam, thereby
avoiding a possible safety hazard.
2.4 DETECTOR AMD RECEIVER ELECTRONICS
A copper-doped germanium detector with a quantum efficiency of
10% was used in the LDV system. The liquid helium cooled Ge:Cu
detector requires more than 20 milliwatts to become shot noise
limited and can be operated with 50-100 milliwatts if desired. At
these higher powers, the power reflected by the secondary of the
Cassegrainian telescope is negligible and no saturation effects
occur. In addition, the Ge:Cu detector is nearly impossible to
damage if excessive optical power is applied.
2-7
-------
The detector was properly matched to the bias circuit and the
preamplifier in order to obtain the frequency response desired and
to assure that the shot noise will exceed the thermal noise of the
preamplifier and load resistor which are the major noise sources.
The Ge:Cu detector used in the measurements is shown in Figure 2-6.
2.5 OPTICAL SCANNER
An optical scanner was necessary to direct the 30 cm diameter
laser beam toward a target of interest, which, in this case, was the
effluent of a smoke stack. The scanner provided sufficient deflec-
tion in both elevation and azimuth to accommodate various target
sighting geometries and to correct for the rough positioning of the
van.
The elliptical scan mirror was of sufficient aperture so that
the entire laser beam was intercepted at all scan angles. The mirror
surface was flat to X/10 at 10.6-mn under ambient temperature chancres
and various mirror positions. The scanner mechanism provided a pre-
cise, smooth adjustment in both elevation and azimuth so that target
positioning could be easily accomplished. In addition, once the
target had been aligned, the scanner was able to be locked in position
so that the laser beam remained on the target.
A scanner was built and tested which met the general require-
ments and had the following specifications:
Mirror size: 48 cm x 33 cm x 5 cm
Mirror Flatness: VlO i> 10.6 n
Mirror coating: Aluminum with SiO overcoating
Azimuth scan angle: + 8°
Elevation scan angle: 0 to 20
The scanner had micrometer drives on both the azimuth and elevation
scan controls. It was possible for the scanner to easily maintain
alignment accuracies of jflOO yradians. A photograph of the scanner is
shown in Figure 2-7.
2-8
-------
Figure 2-6.
A Liquid Helium Cooled Copper Doped Germanium
Detector.
2-9
-------
RAYTHEON COMPANY
EQUIPMENT DIVISION
Figure 2-7. Scanning Mirror
2-10
-------
2.6 SIGNAL PROCESSING
The return signal from the smoke stack effluent was processed
in two manners. The spectral density of the returned signal was
determined by passing the signal through an HP 8552B/8553B spectrum
analyzer. The velocity of the smoke stack effluents could be de-
termined from the doppler shift of returned signals.
A frequency tracker was also used to process return signals.
The frequency tracker was basically an FM demodulator with negative
feedback. It, therefore, became a tracking filter with a variable
center frequency. It produced a DC voltage proportional to the
mean spectral frequency as well as signals indicating signal dropout
and signal fluctuations.
The outputs of both the frequency tracker and the spectrum
analyzer were recorded on magnetic tape for further data processing.
2.7 THE LASER TEST VAN
The LDV system was incorporated into Raytheon's laser test van.
This van is a specially modified 12 meter trailer from the Fruehauf
Corporation. Some of its features are listed below.
a. An air conditioning system capable of handling all
electrical power dissipation with three men in the
van, for outside air temperatures as high as 52°C.
b. A heating system capable of maintaining a 20°C in-
ternal van temperature at an outside air temperature
as low as -18°C.
c. Full thermal wall and floor insulation.
d. Full fluorescent lighting fixtures.
e. Ample 60 and 400 Hz power outlets in van walls.
f. Three large 0.9m square windows (one in rear,
one each side) for transmitting the laser beam.
2-11
-------
g. Four heavy truck jacks (with sand shoes) were pro-
vided for maximum stabilization of the van when on
site.
h. A 1300 kg capacity hydraulic lift platform was
installed at one of the side doors for heavy
equipment loading.
A 25 kW, 60 cycle generator was used to power the van. The
generator was gasoline powered. Two views of the van are shown
in Figure 2-8 and 2-9.
2-12
-------
RAYTHEON COMPANY
EQUIPMFNT DIVISION
Figure 2-8. Raytheon Trailer
Figure 2-9. Raytheon Trailer
2-13
-------
SECTION 3
RESULTS OF THE FIRST FIELD TESTS
3.0 INTRODUCTION
The first phase of field testing a LDV system for the remote
detection of effluents from smoke stacks was completed during the
week of 23 - 31 August 1974. A C02 Laser Doppler Velocimeter was
assembled in a 12 meter semi-trailer and taken to the River Bend Steam
Station of Duke Power Company in Mount Holly, North Carolina. A
photograph of the test site is shown in Figure 3-1.
After assembly and calibration of the LDV system at the Duke
Power Test Site, backscatter signals representing the velocity
distribution of the effluents from the smoke stacks were detected
and recorded. All data was taken at a slant range of 400 m as
measured by a laser range finder. The elevation angle to the stack
lip was measured with a transit and found to be 8° from the hori-
zontal. Under these conditions, the effluent exit velocity is re-
lated to the doppler IF frequency by the equation:
v(m/sec) = 3.81 x 10"3 £VD (Hz) (3-1)
Measurements of effluent exit velocities were made under a var-
iety of operating conditions including: (1) various power plant load
conditions from 80 MW to 140 MW to vary the output velocity and (2)
various precipitator operating conditions to vary the effluent opacity.
Velocity profiles across the stack lip and at various heights above the
stack were made. In all approximately 5,500 meters of taped data
were obtained for later processing.
The primary direction of the program during the first field
tests was to show the feasibility for remote effluent velocity
measurements. Hence, the reduction of the data was directed pri-
marily toward the determination of the mean effluent velocity from
the power plant smoke stacks. Most of the data were collected from
the EPA instrumented smoke stack, which is the third stack from the
3-1
-------
I
to
Figure 3-1.
The Raytheon Data Van at
the Duke Power River Bend
Steam Station.
-------
right as seen in Figure 3-1. At Duke Power, this stack is commonly
referred to as the number six smoke stack. The EPA contracted with
Environmental Science and Engineering, Inc. to perform in-stack
pitot-tube velocity measurements at the same time as the remote LDV
velocity measurements were being made. The in-stack monitoring was
done through sampling ports located about 1.6m above the base of the
smoke stack as shown in Figure 3-2. The in-stack monitoring also
included the measurement of particle size distributions and mass
emission rates.
The constriction of the smoke stack at the top from 2.94 m to
1.96 m caused an increase in velocity. The gas velocity measured at
the base had to be multiplied by a factor of 2.32 in order to obtain
an estimate of the exit velocity. The use of such a factor assumed
that there is no significant cooling or compression of the exhaust
gases between the sampling ports and the stack exit.
3.1 PROCESSING METHODS
The recorded data was reduced in basically two different
manners. In one procedure the recorded spectrum analyzer traces
were played back into a 564 Tektronix storage oscilloscope. Thirty-
eight sequential traces were photographed in one composite photo-
graph (a 2 second exposure). Typically, three or four composite
photographs were taken at each stack profile position. During
CW runs between 9 and 16 composite pictures were used to determine
the average effluent velocities. A typical composite photograph
is shown in Figure 3-3a.
In order to justify the above procedure as an accurate pro-
cessing method, a one minute interval of data was evaluated.
Thirty 2 second composite photographs were taken at 2 second inter-
vals. These photographs were used to determine the mean velocity and
the velocity distribution width of the smoke stack effluents. A
data interval from Tape #6 on 29 August 1974 was used in these
measurements. The thirty, 2 second composite photographs indicated
a mean effluent velocity of 50.9 + 3.0 m/sec, where the ±3.0 m/sec
factor represents the standard deviation in velocity from the thirty
3-3
-------
LASER PATH
1.94 m
STACK HEIGHT = 24.7 m
1.6 m i
PLANT ROOF
SLANT RANGE = 400 m
SAMPLING PORTS
Figure 3-2. Smoke Stack Geometry for LBV Effluent Velocity
Measurements.
3-4
-------
RAYTHEON COMPANY
EQUIPMENT DIVISION
Figure 3-3a. Typical Composite Photograph
of 2 Second Average Spectrum
Analyzer Signals.
Figure 3-3b. Typical Photograph of 7
Second Average Spectrum
Analyzer Signals.
3-5
-------
photographs. Similarly, the average -3 dB velocity width of the
effluents was 53.0 _+ 5.8 m/sec. Since the standard deviations are
less than 11% of the average values of both the velocity and velocity
width, the process of using randomly spaced 2 second composite photo-
graphs to estimate effluent velocity parameters appears justified.
Another procedure utilized a signal averager. The recorded
vertical output of the spectrum analyzer was played back into the
signal averager. During CW runs 512 sweeps were averaged over a
28 second period and were photographed at 30 second intervals. A
typical photograph of the averaged signal is shown in Figure 3-3b.
These photographs were then analyzed to determine the mean effluent
velocity as a function of time. To determine the velocity profile
across the top of the smoke stack, 128 sweeps were averaged over a
7 second period and photographed at approximately 10 second intervals.
Typically, five or six photographs were obtained at each profile
position. The results from these photographs were averaged to deter-
mine a mean effluent velocity for each profile position.
Efforts were also made to utilize the frequency tracker as a
data processor. Strip charts of effluent velocity as a function of
time were made for most of the CW data runs. However, the tracker
was meant to be used against signals which were considerably narrower
than those generated by the smoke stack effluents. As a consequence,
the tracker performance was questionable, particularly in the pre-
sense of widely varying signal amplitudes. Hence the frequency
tracker data are suspect and are not presented in this report.
Frequency calibrations were made by introducing a 3.0 MHz
frequency standard into the spectrum analyzer at the beginning of
each data run.
3.2 EXPERIMENTAL RESULTS
Seven tapes of data were recorded on 28 and 29 August 1974.
Six of these tapes contained reducible data and are discussed in
detail below. A summary of the taped data is given in Table 1.
3-6
-------
Table 3-1
Summary of Data Tapes
TAPE
1
2
2
3
4
5
6
6
7
7
RUN
1
1
2
1
1&2
1
1
2
1
2
TIME
10:00
14:30
16:10
22:00
12:17
14:05
14:50
15:05
15:59
16:43
DATE
08/28/74
08/28/74
08/28/74
08/28/74
08/29/74
08/29/74
08/29/74
08/29/74
08/29/74
08/29/74
SUBJECT
CW Run - 2nd stack from
right
CW Run - 0.3 m above
#6 stack *
Velocity profile -
0.3 m above #6 stack
Power reduction
directly above #6
stack - 140 MW-100MW
power level
CW Run - directly
above #6 stack
CW Run - directly
above #6 stack
CW Run - directly
above #6 stack
Velocity profile di-
rectly above #6 stack
Velocity profile - 1.8 m
above #6 stack
Velocity profile di-
rectly above #6 stack
PROCESSING
None - tape not reduced - cali-
bration questionable
2 second composite photographs
28 sec, 512 sweep average
2 second composite photographs
7 sec, 128 sweep average
2 second composite photographs
7 sec, 128 sweep average
2 second composite photographs
28 sec, 512 sweep average
2 second composite photographs
28 sec, 512 sweep average
2 second composite photographs
28 sec, 512 sweep average
2 second composite photographs
7 sec, 128 sweep average
2 second composite photographs
7 sec, 128 sweep average
2 second composite photographs
7 sec, 128 sweep average
* Stacks on the
in Figure 3-1
main building are numbered from 1 to 8 from left to right as seen
-------
3.2.1 TAPE #1, RUN #1
Tape #1, Run #1 was recorded between 10:00 and 10:30 on 28
August 1974. The laser beam was positioned just above the center of
the second stack from the right. Efforts were made to record back-
scatter signals during the period when the precipitators were being
turned off. Due to coordination difficulties the time when the pre-
cipitators were off was not recorded. Nor was a good calibration
obtained. The data on this tape was therefore suspect and was not
reduced.
3.2.2 TAPE #2, RUN #1
Tape #2, Run #1 was recorded between 14:30 and 14:45 on
28 August 1974. The laser beam was positioned approximately 0.3 m
above the center of the number six smoke stack. The laser beam
was not moved during the run. The power load during this run was
constant at 140 MW.
The data from this run were reduced by means of 2 second dura-
tion composite photographs. The mean effluent velocity measured
by averaging 11 composite photographs was 50.0 _+ 4.3 m/sec. The
±4.3 m/sec margin represents the standard deviation in mean velocity
from the photographs.
The data were also reduced by averaging 512 spectrum analyzer
sweeps over a 28 second period and photographing the results at
30 second intervals. The results of this method of analysis is
plotted as a function of time in Figure 3-4. The mean effluent
velocity measured in this manner was 46.9 _+ 2.4 m/sec for the time
period 14:30 to 14:45 on 28 August 1974.
The velocity distribution of the effluents venting from the
smoke stack was very broad. The averaged spectrum analyzer data
indicated a 73.2 _+ 5.8 m/sec velocity spread between the -3 dB
distribution points.
3.2.3 TAPE #2, RUN #2
Tape #2, Run #2 was recorded between 16:10 and 16:30 on
3-8
-------
\
tto
1.00 ft/sec = 0.305 m/sec
IV
1*3
Ul
t
t<
1
1 1 !
Figure 3-4 . Tape #2 Run #1
Duke Power
28 August 1974
0.3 m Above Stack
28 Second Average
-------
28 August 1974. The laser beam was positioned approximately 0.3 m
above the number six smoke stack. The laser beam was scanned at
approximately 23 cm spacings across the top of the stack. Approx-
imately 1 minute of data were collected at each protile position.
The power load during this run was constant at 140 MW.
The data from this run were reduced by means of 2 second com-
posite photographs. Approximately 4 photographs were taken at each
profile position. The velocity profile evaluated from 2 second
composite photographs is shown in Figure 3-5.
The data were also reduced by averaging l^b spectrum analyzer
sweeps over a 7 second period and photographing the results at
approximately 10 second intervals. The profile determined from
this method of analysis is shown in Figure 3-6.
A comparison of the profiles made in the two different pro-
cessing methods is shown in Figure 3-7. In general, the two pro-
files are similar. The profile made with 2 second composite photo-
graphs indicated slightly higher effluent velocities.
3 .2.4 TAPE #3, RUN #1
Tape #3, Run #1 was recorded between 22:00 and 22:41 on
29 August 1974. The laser beam was positioned above the center
of the number six smoke stack. The laser beam was not moved during
the run. The power load during this run was reduced from 140 MW to
80 MW. Unfortunately the data run was terminated by a tape re-
corder failure when the power load was at 100 MW.
The data from this run were reduced by means of 2 second dura-
tion composite photographs. The mean effluent velocity during
various power load conditions was determined by averaging between
5 and 7 composite photographs. The results are plotted as a
function of power load conditions in Figure 3-8.
3-10
-------
1.00 ft/sec =
: I ' :.
Figure 3-5.The Average Velocity of Effluents from a Smoke Stack
as a Function of the Distance from the Center of the
Stack
-------
.00 ft/sec = 0.305 m/sec :
ftfi
::: B
DISTAMCB FROM THE CENTER OF THE SMOKE STACK IN FEET
m
Figure 3-6. The Average Velocity of Effluents from a
Smcke Stack as a Function of the Distance
from the Center of the Stack
-------
j : ! ; i tr
1.00 ft/sec
Li
-
= 0
!
; ! U
;"."["" ;
r
;
"TT Y:"
w ~i~~
^ *
.w
;......
i
LEFT
. ' 5 rf
i ! ' .
*.
__
^-^
T\Tl
^/i*
-j
.305
".' r
i
~1
m/s
*
!
i
r
t
4
ec
. j- '
. J
p,
i
-
JIANCr. jr
i
r
-i,:
i~
T
'
:L
1
1
' i
flan
.
i . :.
L: _
Irifc
!
L '.
1
^ENT
250
225
200
175 >
(
.
150
125
100'
:
75^
50
25
-g| - TAPE
W PROF:
w !
« o
»
cu
U
J. *
i '
ii-i- *
r
>4! -V ' +
gj - -4-
Sj i |±i: i"l
EH'
1
fe!
' W
ERj tXr rajE STROKE
1
i i ...
# 2 RUN # 2
[LEO. 3 m ABOVE STACK
2 SECOND AVERAGE
7 SECOND AVERAGE
i i : : i
-
... - ...
Ty L i _i
3 ; ,
I
' i
.1
>
'.
StACJK IN ?EET
STAC
i" i
i |
:
i -
i
! '
i
;
I
ae-D-jAME
B
: 1
1
'
:
1
,
?ER
' ' 1~
;
.IGHT
4 I $-
i . I
:-':~!---U-
Figure 3-7.
A Comparison of the 2 Second
and 7 Second Average Effluent
Velocity Profile Data
-------
I
U)
I
TAPE # 3 RUN # 1
BEAM DIRECTLY ABOVE STACK
AVEftAjGB EFSLUEN.T VSLQCJ3?*: IN F^T PEB SECOND
1 I
225
200 '
7R
cn ,
LDU
L25 '
.00 -
50 -
25^
.:'.:.
' j
..!. . i
i
; i
;.
. i__
- -
STB
!
J
. [:;
1 '
'
_Li.
i
;f lil
IAISY
,|.:..j:. L...
1 ! .
: .
!
" i-
«;fe:
STA'
;
.00 ft/s
. L
L_
!
1
1
IE
i'
1
t
. .
-
: .:
-
ec
= 0.3
:
1
r
::;
I '-
140-
- :
....
.
: i . .
::.. '...
-120
i :
i
L ;:
[
-. 1 '
1
ST
T-u-nivwi
05
i_
.-:
R
i/sec
, L .
*
1
i ~
r
!AJ>T
PUWjSK
i
^
:
_
L- :
IT-
.:;_
: I i ;;
1
0
..:.
,
- : -
; "
"
.
1204-100
STATE
3AD IN MpOSWRTTS'
i
:
_ i.J
- -':''-
i
- -
.... .
"~n~
100
STJSADV 9TZ
^
7
1
TE
SJliUUlNlJ AV^KAUii
SECOND AVERAGE
L - ,
1 : .
._..:.. ;.
-,-:
1 r
n j 1 -f "
' L:
-
100-
!; ;;
;
80
' T ;
j. i
!
;
i
'
Figure 3-8. Comparison of the 2 Second and 7
Second Averaged Velocity Data for
the Power Load Reduction
-------
The data were also reduced by averaging 128 spectrum analyzer
sweeps over a 7 second period and photographing the results every
30 seconds. The results of this method of analysis are plotted as
a function of time in Figure 3-9. The mean velocity for
each power load condition was evaluated by averaging the results of
the 7 second average photographs. The results are plotted in
Figure 3-8 as a comparison to the 2 second composite photographic
data.
3.2.5 TAPE #4, RUNS #1&2
Tape #4, Runs #1&2 was recorded between 12:37 and 12:40 on
29 August 1974. The laser beam was positioned directly above the
center of the number six smoke stack. The laser beam was not moved
during the run.
The data from this run were reduced by means of 2 second
duration composite photographs. The mean effluent velocity measured
by averaging 16 composite photographs was 46.94; 7.9 m/sec. The
i 7.9 m/sec margin represents the standard deviation in velocity from
the photographs.
The data were also reduced by averaging 512 spectrum analyzer
sweeps over a 28 second period and photographing the results at
30 second intervals. The results of this method of analysis are
plotted as a function of time in Figure 3-10. The mean effluent
velocity measured in this manner was 54.9 _+ 4.3 m/sec for the time
period 12:17 to 12:40 on 29 August 1974.
The velocity distribution of the effluents venting from the
smoke stack was very broad. The averaged spectrum analyzer data
indicated a 76.2 + 6.4 m/sec velocity spread between the -3 dB
distribution points.
3-15
-------
U)
I
nf
ift
US
1.00 ft/sec = 0.305 m/sec
SO
Figure 3- 9.
Tape #3 Run #1
Duke Power
28 August 1974
Power Reduction
28 Second Average
-------
= 0.305 m/sec
JFf
7JM£
Ht
Figure 3-10. Tape #4 Run #1 & 2
Duke Power
29 August 1974
Directly Above Stack
28 Second Average
-------
3.2.6 TAPE #5, RUN #1
Tape #5, Run #1 was recorded between 14:05 and 14:27 on
29 August 1974. The laser beam was positioned directly above the
center of the number six smoke stack. The laser beam was not
moved during the run.
The data were reduced by averaging 512 spectrum analyzer
sweeps over a 28 second period and photographing the results
at 30 second intervals. The results of this method of analysis are
plotted as a function of time in Figure 3-11. The mean effluent
velocity measured in this manner was 48.8 .+ 8.8 m/sec for the time
period 14:05 to 14:27 on 29 August 1974.
The velocity distribution of the effluents venting from the
smoke stack was very broad. The averaged spectrum analyzer data
indicated a 66.4 +7.0 m/sec velocity spread between the -3 dB
distribution points.
3.2.7 TAPE #6, RUN #1
Tape #6, Run #1 was recorded between 14:50 and 14:55 on
29 August 1974. The laser beam was positioned directly above the
center of the number six smoke stack. The laser beam was not
moved during the run.
The data from this run were reduced by means of 2 second dura-
tion composite photographs. The mean effluent velocity measured
by averaging 9 composite photographs was 53.6 + 5.5 m/sec. The
±5.5 m/sec margin represents the standard deviation in velocity from
the photographs.
The data were also reduced by averaging 512 spectrum analyzer
sweeps over a 28 second period and photographing the results at
30 second intervals. The results of this method of analysis are
plotted as a function of time in Figure 3-12. The mean effluent
3-18
-------
CO
H
a«*
I7J
1.00 ft/sec = 0.305 m/sec
\
Mf \
H* Mm
Figure 3-11.
Tape #5 Run #1
Duke Power
29 August 1974
Directly Above Stack
28 Second Average
-------
Figure 3-12. Tape #6 Run #1
Duke Power
29 August 1974
Directly Above Stack
28 Second Average
-------
velocity measured in this manner was 47.5 _+ 3.0 hi/sec for the time
period 14:50 to 14:55 on 29 August 1974.
The velocity distribution of the effluents venting from the
smoke stack was very broad. The averaged spectrum analyzer data
indicated a 85.6 + 4.0 m/sec velocity spread between the -3 dB
distribution points.
3.2.8 TAPE #6, RUN #2
Tape #6, Run #2 was recorded between 15:05 and 15:38 on
29 August 1974. The laser beam was positioned directly above the
number six smoke stack. The laser beam was scanned at approxi-
mately 26 cm spacings across the top of the stack. Approxi-
mately 1 minute of data was collected at each profile position.
The data from this run were reduced by means of 2 second com-
posite photographs. Approximately 4 photographs were taken at
each profile position. The velocity profile evaluated from 2
second composite photographs is shown in Figure 3-13.
The data were also reduced by averaging 1^8 spectrum analyzer
sweeps over a 7 second period and photographing the results at
approximately 10 second intervals. The profile determined from
this method of analysis is shown in Figure 3-14.
A comparison of the profiles made in the two different pro-
cessing methods is shown in Figure 3-15. In general, the two pro-
files are quite similar.
3.2.9 TAPE #7, RUN #1
Tape #7, Run #1 was recorded between 15:59 and 16:27 on
29 August 1974. The laser beam was positioned approximately 1.8 m
above the number six smoke stack. The laser beam was scanned
at approximately 26 cm spacings across the top of the stack.
3-21
-------
1.00 ft/sec = 0.305 m/sec
Figure 3-13. The Average Velocity of Effluents from
a Smoke Stack as a Function of the
Distance from the Center of the Stack
-------
TAPE # 6 RUN # 2
PROFILE DIRECTLY ABOVE STACK
7 SECOND AVERAGE
Figure 3-14. The Average Velocity of Effluents from
a Smoke Stack as a Function of the
Distance from the Center of the Stack
-------
1.00 ft/sec = 0.305 m/sec
d!
f 7-
LEFT
**4-
_^_ « ^ «
250
_.; |_._. .
rJ i- i
_L
-:
r 200
f^«V
-1-.L-. -l-:-r -
. te.
175
-4
t
w
t«
bi
*
-.-4-
[
-150
1*L_,
1*.... I
t -P5
iM
- 75
50 -J
O
:S
TAPE # 6 RUN # 2
PROFILE DIRECTLY ABOVE STACK
O 2 SECOND AVERAGE
+ 7 SECOND AVERAGE
>*
T
j-
, i !
i_.
i ... ..__.
1 -r
T
h
I STACK DJEAMEJFER
RIGHTLY u
-U4-
i-u
STACK 1$
Figure 3-15. A Comparison of the 2 Second
and 7 Second Average Effluent
Velocity Profile Data
-------
Approximately 1 minute of data was collected at each profile posi-
tion.
The data from this run were reduced by means of 2 second com-
posite photographs. Approximately 4 photographs were taken at
each profile position. The velocity profile evaluated from 2
second composite photographs is shown in Figure 3-16.
The data were also reduced by averaging 128 spectrum analyzer
sweeps over a 7 second period and photographing the results at
approximately 10 second intervals. The profile determined from
this method of analysis is shown in Figure 3-17.
A comparison of the profiles made in the two different pro-
cessing methods is shown in Figure 3-18. In general, the two pro-
files are similar.
It is apparent that the turbulence which occurs in the mixing
region above the smoke stack produces received signals with widely
varying Doppler shifts. The resulting effluent velocity profile at
2.8 m above the stack lip was a distribution which is difficult to
interpret. Clearly further experiments and analysis is required
before these data can be interpreted accurately.
3 .2.10. TAPE #7, RUN #2
Tape #7, Run #2 was recorded between 16:48 and 17:08 on
29 August 1974. The laser beam was positioned directly above the
number six smoke stack. The laser beam was scanned at approximately
26 cm spacings across the top of the stack. Approximately 1
minute of data was collected at each profile position.
3-25
-------
1
oo
1
to
CTi
:
LEFT
c
,
. . . ._ . .
- ---
h
- :*
i
''::
]
....
K
_
'
*
^
DI
. . 1 -
....
STAN
:
--
-
r: '
1
: : i
||
Ji
::' -
: j:'
-i
iiiip
"1,
. !
£8; .FROM
: . 1
.1 ..
IKE
. _.
300 "
275-
*
V
:::
. :
-:
: .
250 -
225 ;
.. l::jj
m
150-
-
: i-
i
:'.'.:
T
-r
GEK1
12 5 r
LOO
1
-50
25
;;!
: <1
p
*
:&
B
i
ill
rf
1
1
1
1
t
^
|
I
r~
n
i
;
i
i-
j
:
I
: ..
:...
*
i- .
; '
f-rr-
^r-
T
. _
. .
ER QF T^E g4QKE
- :
-1
...
-*-
r
jrapj
\l^
12 '$1
.TOO^'fl
i
1
!
s # 7 IOTK 4JL1 ;
*U*BJ6 F^T. A&O1
j*V«"V
:::'
-/£
4-:
i
i
....:._1
STACJ3C IH FE1
'INUi A VKBfl
i - - -
i j
Li
~ ..
...!.] .1"
!
.
-* ! STA<
V&K
KC .
... i
JOS m/sec
- i
.. .1
W
i
i -
- i
I
i
_
.:....
:K okA^
1
I
;
--,1
..{..;.
; -
FSR
Ripm
^ -
ET
I !
. _. j_ ..
El
Figure 3-16
The Average Velocity of Effluents
from a Smoke Stack as a Function
of the Distance from the Center
of the Stack
-------
U)
to
300
____
t/sec = 0.30D m/aec
Fiqure 3-17. The Average Velocity of Effluents from
a Smoke Stack as a Function of the
Distance from the Center of the Stack
-------
- -
. <
~
»
1".
u -
.
-
LE
*
-'
-
U -
FT
- !
' |
-J '-1
L
: : i
'
,\ '"
.'. \ :. .
1
_L_
;
i
;
;!
*
:-
1
4t::
I'rii
K
---
- -: -
;
'
i
&
:-:! ;'*
H
r r
1
T-
-;-;:
1
p
-;-:"
1:
:_
:-i
__
--:
-
ys
.. il :;:
r:
--
::
Z
^
ip
.:
jf
1
~
-:-'-
-
~\ :--.
' K i
-
-
TT-
'
5T,
M?
,L_
' r.
: : ."t
^f
: TH
E
j!r
:
-
: : :
i ;.
i
- -:
i
-:'~
GE
.
-.:
TIT
i
*
jt
- i
'w^
: -u
m
r.':
" '.
-
fprj
tK)M
i|
;-
:--
' -:
---
;::'
THE
--
r :-
IH
1
300-
275-
t :lf
~ 200^
-.
-i
t
j» ir/-»-.
150
T O C
' 1- i
t
-1-
\
\ ]
100
75
i
- - r" '
l
-25
r
"\
CENTER O
i
:--
T'
t
?;;-:8g
';£"
I::
:fi:
f
t»-
1
ji
h-|
i:^
1
: i .
F jTH
.
---
^
1*1
-
: p:r-
-rj
:
E
a-
- ru
-
;
-r:
^
i ' '
r
'r .'.'.
ft
4^
n ..
;;
.' .
OKE
r
r: :
;. ; ;
--
- -'-
9
-
TAPE # 7 RUN # 1
PROFILE 6 FT ABOVE STACK
1.
0
2
7
SECOND AVERAGE
SECOND AVERAGE
00 ft/sec =
-
j
_-..:.
T§
M
SI
P
PAC
- - II
K
i3
:«
ii."
i!"
*;h
»t 1
w~r
i'::
F!
El
. j
.
ii[
~-^-
-.."
-
.
-
&
0.305 m/si
_r
-
-
:':: ir:i
:m ^r:
:::: J
'! 'u
ill
: "
:i:; :,:
i ;
: -
-!~
--
.i.: 1 :
: t!
'::; ':'.::
' --
T '"
,
!
t
LAME
SC
*
r:^,:/;
1
:j;
1
--{-
PEft
! i
RIGHT
1
. ' .
H U
ii
Figure 3-18. A Comparison of the 2 Second
and 7 Second Average Effluent
Velocity Profile Data
-------
The data from this run were reduced by means of 2 second
composite photographs. Approximately 4 photographs were taken
at each profile position. The velocity profile evaluated from
2 second composite photographs is shown in Figure 3-19.
The data were also reduced by averaging 128 spectrum analyzer
sweeps over a 7 second period and photographing the results at
approximately 10 second intervals. The profile determined from
this method of analysis is shown in Figure 3-20.
A comparison of the profiles made in the two different pro-
cessing methods is shown in Figure 3-21. In general, the two pro-
files are similar. The profile made with 7 second averaged
photographs indicates slightly higher effluent velocities.
3.3 CONCLUSIONS ON THE FIRST FIELD TESTS
At the end of the first set of field tests the feasibility of
using a Laser Doppler Velocimeter to remotely monitor smoke stack
effluents was clearly demonstrated. Comparisons of the in-stack
pitot tube velocity data with the Doppler velocity data are shown in
Table 3-2. The Doppler velocity data generally exceed the in-stack
data by an average of 12%. It is not known whether this error is
caused by the miscalibration of either the pitot tube or the LDV, or
is the result of the two measurements being carried out at different
points on the stack. A significant portion of the error appears to be
systematic rather than random, and hence could be removed with more
accurate calibration.
During the course of the measurements, it was obvious that
changing the number of precipitators in operation did effect the LDV
signal intensity. While precise measurements were not made, it
appeared that the system SKR increased about 10 dB each time a pre-
cipitator was turned off. The implication of this result is that
the LDV signal intensity may have the potential of being used as a
measure of effluent concentration.
3-29
-------
Figure 3-Lc
The Average Velocity of Effluents
from a Smoke Stack as a Function
of the Distance from the Center
of the Stack
-------
: '':::-
tffi
rf-:
l.C
10
'sec =
ra/sec
itrt
::::
1^250 4
^te:
225
:Hi
PROFILE DIRECTLY ABCTVE ST.
"
..|. L:
U)
I
OJ
«+
-
175
ffiffi
150 r
lift
25:
1
; ::::
-
I
I
tttt
rirr
7B1
q-;
rpfa
50
'*'::
:£:::
LEFT
25
*-
tt;
RIGHTS
STACK
: I:r
Figure 3-20. The Average Velocity of Effluents from
a Smoke Stack as a Function of the
Distance from the Center of the Stack
-------
UJ
U)
... . j
!
!
I
;i.
i i
LEFT
-i-
4 -i
1.00 ft/sec = 0.305 m/sec
-:, - .v-,-:n;r.:r^,,|...
i i -!
hr
I I
TAPE # 7 RUN # 2
PROFILE DIRECTLY ABOVE STACK
2 SECOND AVERAGE
7 SECOND AVERAGE
rjrrrt
i-: i
*
-' 25
'Tig
AJ~
1
'''i ^
!
"
'
- -
1
DjAME
METER
RIGHT
fct!3TAHtbE FfcOtt
1
tflBRT Of ]TH|E SMOKB
..!! I
Figure 3-21. A Comparison of the 2 Second
and 7 Second Average Effluent
Velocity Profile Data
-------
Date
8/28/74
8/28/74
8/28/74
8/29/74
Time
16:20
22:05
22:35
12:15
Pitot Tube
Velocity
VP
(m/sec)
47.9
38.4
34.4
50.9
Average Ratio =
Doppler
Velocity
VD
(m/sec)
53.3
47.9
36.0
VD/VP
1.11
1.25
1.05
1.14 + 0.10
Averaged
Doppler_
Velocity ,VD
(m/sec)
48.5
47.5
38.1
55.2
VD/VP
1.01
1.24
1.11.
1.08
1.11 + 0.10
1.12 + 0.09
Table 3-2. A Comparison of Pitot Tube Velocity Data
With the Doppler Velocity Data.
V_ data are taken from 2 second composite photographs.
V_ data are taken from 7 sec and 28 sec electronically
averaged data.
3-33
-------
One point which must be emphasized is that the amount of
data analyzed in the first test phase represents only a small frac-
tion of the collected data. While the photographic method used
appears satisfactory, it is far from ideal. The frequency tracker
data was found to be unusable. The tracker was designed to be used
with relatively narrow spectral signals. The broad doppler signals
widths from the smoke stack effluents were wider than the tracker's
discriminator bandwidth. This saturation of the discriminator led
to erroneous output signals. As a result, graphs of continuous
effluent velocity as a function of time could not be made.
It was apparent that more work needed to be done to investigate
turbulence effects and the relationship between LDV signal intensity
and effluent concentration. A second set of field tests was
scheduled to make these tests on a calibrated LDV system with a
suitably modified frequency tracker. The tests and their results
are presented in the next section.
3-34
-------
SECTION 4
RESULTS OF THE SECOND FIELD TESTS
4.0 INTRODUCTION
After the first field tests were completed, the scope of the
program was expanded to include measurements (1) to investigate the
feasibility of obtaining mass flow data from the laser doppler veloc-
imeter, (2) to determine the effluent backscatter coefficient as a
function of effluent opacity, (3) to produce exit velocity profiles
with decreased separation between measurement points, (4) to study
the effect of turbulence on the LDV signal spectrum, and (5) to ob-
tain additional measurements of effluent exit velocity for various
in-stack velocities and laser elevation angles.
In order to carry out these tasks, certain modifications were
made to the LDV system. The modifications included the use of a
modified Mach-Zehnder interferometer; a liquid nitrogen cooled,
lead tin telluride detector; a frequency/intencity tracker suited to
wide bandwidth, rapidly fluctuating, Doppler spectra; and the
addition of a time-code generator.
Field tests were carried out at Duke Power's River Bend Steam
Station during the week of 12 - 19 January 1975. Approximately 4
hours of data were recorded on magnetic tape for later analysis. This
data included runs made at 8°, 20°, 28°, and 37° elevation angles,
high resolution velocity profiles at 0 m, 0.9 m, 1.4 m, and 3.7 m
above the lip of the smoke stack, runs to determine the LDV signal
strength as a function the opacity of the effluent, and runs to
evaluate the effluent velocity as a function of the in-stack velocity.
The system modifications, the second set of field tests, and
their results are presented in this section.
4.1 SYSTEM MODIFICATIONS
Four modifications to the system described in Section 2 were
4-1
-------
incorporated into the LDV for the second set of field tests. A
modified Mach-Zehnder interferometer was installed into the system.
An attempt was made to use a liquid nitrogen cooled lead-tin telluride
detector. Most importantly a frequency/intensity tracker was incor-
porated to provide continuous voltage outputs proportional to the
frequency and intensity of the received doppler IF signals. Finally,
a time code generator was used to obtain better time referencing.
These system modifications are discussed below.
4.1.1 THE MODIFIED MACH-ZEHNDER INTERFEROMETER
The modified Mach-Zehnder interferometer is a polarized config-
uration which reduces the inherent 6 dB loss of the conventional
Mach-Zehnder interferometer to approximately 1 dB. The construction
and operation of the modified Mach-Zehnder interferometer is basic-
ally similar to that of the Mach-Zehnder interferometer used in the
first field tests and described in Section 2. In order to eliminate
the need for independently maintaining alignment of the various mirrors
and beamsplitters, the interferometer was constructed from a precision
machined block of solid invar. The result was an interferometer which
was mechanically ridig and thermally stable.
4.1.2 THE LEAD TIN TELLURIDE DETECTOR
An attempt was made to incorporate a liquid nitrogen cooled
lead tin telluride (PbSnTe) detector into the system to avoid having
to work with liquid helium in the field. A PbSnTe detector was
obtained from Raytheon's Special Microwave Devices Operation and tested
to see if it could replace the copper doped germanium detector used
in the first field tests.
The PbSnTe photodetectors are junction devices whose frequency
response is limited by the capacitance of the junction. In order to
maximize the frequency response, the detector area is kept small and
the devices are generally used with transimpedance amplifiers. The
problem with small area detectors is that they are difficult to get
shot noise limited. They tend to saturate or overheat before
4-2
-------
sufficient local oscillator power can be applied to produce shot
noise limited operation. This problem was observed with this PbSnTe
detector.
The detector size was 0.1 x 0.1 mm. By increasing the local
oscillator power on the chip, signal-to-noise ratios approached within
8 dB of the results from a Ge:Cu detector. The PbSnTe detector could
not be made shot noise limited because of saturation and thermal
runaway problems. Nevertheless it was thought that the detector
would perform well enough to be used for the detection of smoke stack
effluents. Because of the degraded performance, the PbSnTe detector
was used a backup, rather than a replacement, for the liquid helium
cooled Ge:Cu detector. The measurements made during the second field
tests used the Ge:Cu detector.
4.1.3 THE FREQUENCY/INTENSITY TRACKER
In order to obtain continuous monitoring of both the velocity
of the effluents and the intensity of the LDV backscatter signals,
a frequency/intensity tracker was incorporated into the system. This
tracker had the capability of following rapidly fluctuating, wide
bandwidth, doppler IF signals. Furthermore, since the tracker operated
off the vertical output of a spectrum analyzer, rather than the raw,
high frequency, doppler signals, the tracker could be used with tape
recorded spectrum analyzer signals. The ability to use a low fre-
quency (30 kHz) FM tape recorder in analyzing megahertz frequency
signals has several important advantages with regard to signal
processing.
The tracker was used in conjunction with a H.P. 141T/8553B/8552B
spectrum analyzer. It could track the frequency and intensity of
signals in the 0 - 100 MHz range. When sufficient signal-to-noise
ratios were available(typically 5 dB in 100 kHz bandwidths), the
tracker could follow signals as small as -100 dBm.
4-3
-------
The tracker worked by detecting the frequency component which
has the peak amplitude. The tracker also integrated the received
signal between preset frequency limits to determine a total received
signal strength. The preset limits were used to remove spurious
signals due to ground winds and spectrum analyzer frequency markers.
The effects of noise were averaged out of the tracker's intensity
channel by setting the integrated noise level to zero.
The outputs from both the frequency and intensity tracker
channels were equipped with low pass filters having variable time
constants. This time constant could be controlled to provide signal
averaging over periods from 10 msec to 2000 sec. Typically, 2 sec
averaging time constants were used in reducing the data.
In order to check out the frequency/intensity tracker, it was
used with data taped on the first field tests. The results of a CW
run with the laser beam just above the smoke stack lip is shown in
Figure 4-1. The top trace is the relative integrated signal strength
as a function of time. The spiking behavior is real and is the result
of a rapping cycle on the electrostatic precipitators used to knock off
ashes and other particulate matter. The bottom trace is the effluent
exit velocity plotted as a function of time. Short term fluctuations
are visible but the exit velocity remains fairly constant on a long-
term basis. This chart was made from data recorded on Tape #4 on
29 August 1974.
The results of a velocity profile run are shown in Figure 4-3.
This chart was made from data recorded on Tape #7 - Run #2 on 29 August
1974. It shows the exit velocity distribution across the lip of the
smoke stack. It should be noted that the position of the laser beam
above the stack was moved in 26 cm jumps. Hence, the spatial
resolution of the profile is not very good. It does prove the useful-
ness of the tracker however.
4-4
-------
Ln
O.B
0.6 .
0.4.
[fl
>
:-.
DUKE POWER 29 AUGUST 1974
RIVER BEND STEAM STATION
NO. 6 SMOKE STACK
8° ELEVATION ANGLE
DATA: TAPE *4
12:23
t
M/SEC
i
100.
BO.
JH
g
I 6°
§
40-
20
FT/SEC
150-
100-
SO--
TIME OF DAY IN HR:MIN
I I
12:39
-------
DUKE POWER 29 AUGUST 1974
RIVER BEND STEAM STATION
No . 6 SMOKE STACK
8° ELEVATION ANGLE
1
VELOCITY PROFILE
i
JUST ABOVE THE STACK LIP
j
i ' " i
DATA: TAPE #7 RUN j42
1
. i
-
..
MX
i i
; ,
. ' *
,
;
i
1
1 !
i :
i ; 1
'SEC FT/SEC
-
... ...
.
70-
1
60 .
50
s
1
250 -
' 225
200 .
175
-
t
_
'
1
-... i
IV : ~ "
----- -r- ---,-- -[ ' -
i li
i IH t \
diiJn
jy j ir wu
VIII 1 i Jflnf nlttli II HUM
40-.
til
20--
10--
150 ,
125 . .
30.. 100.-
75
50
RIGHT -
DISTANCE FROM
, 4.3
1
STACK CENTERLINE IN FEET
-< 1; , i | h
3.4 2.6:1 1.7 , .85 0 .85 ! 1.7
-T-" -
2.6 3.4 4.3
u
POSITION NUMBER
3 :
10
Figure 4-2. Tracker Output Showing Effluent Exit Velocity
Profile Across the Smoke Stack Lip.
4-6
-------
4.1.4 THE TIME CODE GENERATOR
A time code generator was incorporated into the LDV system to
provide an accurate time reference for the measurements. This gen-
erator was synchronized with the power plant clock and was recorded
with all measurements during the second field tests. The time code
playback allowed easy correlation of the LDV data with the in-stack
data.
The incorporation of the frequency/intensity tracker and time
code generator resulted in slight changes from the LDV configuration
(shown in Figure 2-1) of the first field tests. The block diagram
of the LDV system used in the second set of effluent measurements is
shown in Figure 4-3.
4.2 THE SECOND FIELD TESTS
The LDV system was located at Duke Power's River Bend Steam
Station in Mt. Holly, North Carolina during the week of 12 - 19
January 1975 for a second set of smoke stack effluent measurements.
The system was set up as shown in Figure 3-1 at a slant range of
400m from the number six smoke stack. The elevation angle to the
stack lip was 8°.
The primary direction of the first field tests was to prove the
feasibility of remote effluent velocity measurement with an LDV. The
second set of tests were oriented toward collecting data to determine:
(1) effluent exit velocities as a function of in-stack velocity con-
ditions, (2) correlation of effluent backscatter signal strength with
cross-stack optical transmission, (3) the value of the effluent back-
scatter coefficient at 10.6 (jm, (4) high resolution profiles of the
effluent exit velocities at various heights above the stack lip, and
(5) the effect of plume turbulence on the backscattered Doppler
spectra.
The measurement of the effluent backscatter coefficient requires
an intensity calibrated LDV. For this purpose, a known Doppler
reflector was placed on the power plant roof. A known reflectivity
4-7
-------
00
IASER
*-
INTER-
FEROMETER
>
TELESCOPE.
TARGE1]
BIASIISG
CIRCUITRY
AMPLIFIER
SPECTRUM
ANALYZER
FREQUENCY
TRACKER
MICRO-
PHONE
TIME CODE
GENERATOR.
1 CHANNEL
TAPE
RECORDER
Figure 4-3. Block Diagram of LDV System for Second
Field Tests.
-------
allowed the calculation of the LDV system's overall efficiency.
The EPA contracted with Environmental Science and Engineering,
Inc. to perform in-staclc pi tot-tube velocity measurements at the
same time as the remote LDV velocity measurements were being made.
The in-stack monitoring performed according to EPA reference method
No. 2 was done through sampling ports located about 1.6m above the
base of the smoke stack as shown in Figure 4-4. The in-stack mon-
itoring included not only pitot-tube velocity transverses but also
cross-stack optical transmission as measured on a Lear Seigler
Model RM-4 transmissometer.
The constriction of the smoke stack at the top from 2.94m to
1.96m causes an increase in velocity. The gas velocity measured at
the base must be multiplied by a factor of 2,32 in order to obtain
an estimate of the exit velocity. The use of such a factor assumes
that there is no significant cooling or pressure change of the ex-
haust gases between the sampling ports and the stack exit, which was
later experimentally confirmed.
The turbulence measurements required Doppler data at various
elevation angles through the smoke stack plume. A relay mirror was
placed at various points on top of the electrostatic precipitators
as shown in Figure 4-4 to obtain elevation angles between 20° and
40°.
Twelve data runs were made during the second field tests. These
data runs included: (1) five high resolution velocity profiles at
heights above the stack lip from 2 cm to 3.7 m, (2) four continuous
runs through the center of the plume at elevation angles from 8° to
37°, (3) two precipitator variation runs to evaluate the LDV signal
intensity as a function of cross-stack optical transmission, and (4)
one run during changing power plant load condition. Most of these
runs were calibrated to allow evaluation of the backscatter coeffi-
cient. The data runs are summarized in Table 4-1 and described in
detail in the following sections.
4-9
-------
LASER PATHS
SLANT RANGE
= 400 m
1.94 m
2.96 m
HEIGHT = 24.7 m
1.6 m i
SAMPLING PORTS
ROOF
\
RELAY MIRROR
ELECTROSTATIC
PRECIPITATOR
(SIX BANKS)
BLOWER
COAL FIRED
BOILER GASES
Figure 4-4. Smoke Stack LDV Geometry,
4-10
-------
TABLE 4-1
Summary of Data Tapes from the
Second Field Tests
Tape
1
2
3
4
5
6
7
8
9
10
11
12
Time
11:13 to
11:34
12:27 to
12:42
15:55 to
16:48
17:39 to
18:06
9:38 to
13:49
15:20 to
15:59
16:42 to
16:55
17:16 to
17:31
7:40 to
9:00
9:15 to
10:00
10:00 to
10:30
10:30 to
11:30
Date
1/16/75
1/16/75
1/16/75
1/16/75
1/17/75
1/17/75
1/17/75
1/17/75
1/18/75
1/18/75
1/18/75
1/18/75
Subject
CW Run - 8° elevation angle - just
above lip of stack.
CW Run - 20° elevation angle - just
above lip of stack.
Precipitator Run: 7 combinations -
20° elevation angle - just above lip
of stack.
Velocity profile - 8° elevation angle-
just above lip of stack.
Load change: 83 - 137 MW - 8° and
20° elevation angles - just above lip
of stack.
Precipitator run: 4 combinations -
20° elevation angle - just above lip
of stack.
CW run - 28° elevation angle - just
above lip of stack.
CW run - 37° elevation angle - just
above lip of stack.
Velocity profile - 8° elevation angle
0.9 m above lip of stack.
Velocity profile - 8° elevation angle
just above lip of stack.
Velocity profile - 8° elevation angle
1.4 m above lip of stack.
Velocity profile - 8° elevation angle
3.7 m above lip of stack.
4-11
-------
4.2.1 TAPE #1: 8° CW RUN
Tape #1 was recorded between 11:13 and 11:34 on 16 January 1975.
It was a CW run in the sense that all operating conditions remained
unchanged during the course of the run. The laser beam was passed
directly through the smoke stack plume, just above the lip of the
stack. No relay mirrors were used, and the laser beam elevation angle
was 8 above horizontal. The power load was constant at about 92 MW.
The cross-stack optical transmission was steady at 96%.
An in-stack effluent velocity measurement was made between 11:30
and 12:00. An exit velocity of 33.5 m/sec was calculated from this
measurement. The effluent velocity measured remotely with the LDV
for the time interval 11:13 to 11:34 is shown in the lower trace of
Figure 4-5. The mean exit velocity measured by the LDV was approx-
imately 29.0 m/sec.
The upper trace in Figure 4-5 shows the relative integrated signal
strength received by the LDV during the time interval from 11:13 to
11:34.
The data displayed in Figure 4-5 were processed through the
frequency/intensity tracker using a 2 sec averaging time constant.
The tracker follows the Doppler frequency component with the peak
spectral amplitude and integrates all frequency components to give
a total integrated signal strength.
4.2.2 TAPE #2: 20° CW RUN
Tape #2 was recorded between 12:27 and 12:42 on 16 January 1975.
Like Tape #1, it also was a CW run. A relay mirror on top of a plat-
form on the electrostatic precipitators was used to direct the laser
beam through the smoke stack plume. The laser beam passed just
above the lip of the stack at an elevation angle of 20° above hori-
zontal. The power load varied slightly from 91 MW at the start of the
run, to 96 MW at 12:34, to 92 MW at the end of the run. The cross-
stack optical transmission also varied slightly increasing from around
87.5% at the start of the run to 91% at the end of the run.
4-12
-------
10
I
M
OJ
8
b
, 5
DUKE POWER 16 January 1975
RIVBR BEND STEAM STATION
NO. 6 SMOKE STCKK
8° ELEVATION ANGLE
DATA TAPE NO. 1
_ L.
Figure 4-5. The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #1: 8° CW Run.
-------
An in-stack effluent velocity measurement was made between 12:15
and 12:45. A mistake was made with the pitot line connections during
this measurement. However, an exit velocity of 31.7 m/sec was
estimated. The effluent exit velocity measured remotely with the
LDV for the time interval 12:27 to 12:42 is shown in the lower trace
of Figure 4-6. The mean exit velocity measured by the LDV was approx-
imately 26.5 m/sec.
The upper trace of Figure 4-6 shows the relative integrated
signal strength received by the LDV during the time interval from
12:27 to 12:42.
The data displayed in Figure 4-6 were processed through the fre-
quency/intensity tracker using a 2 sec averaging time constant.
4.2.3 TAPE #3: 20° PRECIPITATOR RUN
Tape #3 was recorded between 15:55 and 16:48 on 16 January 1975.
In this run various combinations of precipitator banks were turned
on and off to vary the effluent particle content. The last three
banks of electrostatic precipitators (notated: A, B, and C; A being
the last bank) downstream were switched off in various combinations
to obtain seven levels of cross-stack optical transmission from 60%
to 98.5%. A relay mirror was used to direct the laser beam through
the smoke stack plume. The laser beam passed just above the lip of
the stack at an elevation angle of 20° above horizontal. The power
load was constant at 83 MW.
An in-stack effluent velocity measurement was made between 15:45
and 16:10. A mistake was made with the pitot line connections during
this measurement. However, an exit velocity of 30.5 m/sec was
estimated. The effluent exit velocity, measured remotely with the
LDV at various times during the interval from 15:55 to 16:48, is
shown in the lower trace of Figure 4-7. The mean exit velocity
measured by the LDV was about 27.4 m/sec.
The upper trace of Figure 4-7 indicates the variation of re-
ceived signal intensity from the LDV as a function of the electro-
4-14
-------
DUKE POWER 16 JANUARY 1975
RIVER BIND STEAM STATION
NO. 6 SMOKE STACK
20° ELEVATION ANGLE
DATA TAPE NO. 2
TIME OF DAY IN HRrMIN
VELOCITY GATE
TIME OF DAY IN HR:MIN
Figure 4-6. The Strength and Frequency of Effluent Signals
as a Function of Time for Tape #2: 20° CW Run.
4-15
-------
15:55)30
1
DUKX POMEK 16 JANUARY 1.975
RIVER BEND STEAM STATION
HO. 6 SMOKS STACK
20° EUVATION AJ4GLE
DATA TAPE HO. 3
Figure 4-7. The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #3: 20° CW Run.
-------
static precipitators in use on the smoke stack.
The data displayed in Figure 4-7 were processed through the
frequency/intensity tracker using a 2 sec averaging time constant.
4.2.4 TAPE #4: 8° VELOCITY PROFILE
Tape #4 was recorded between 17:39 and 18:06 on 16 January 1975.
Tape #4 was a high resolution profile of the exit velocity distribution
across the top of the smoke stack. The laser beam was scanned
across the top of the stack just above the stack lip. No relay
mirrors were used, and the laser beam elevation angle was 8 above
horizontal. The power load during the profile was fairly constant
at 83 to 84 MW. The cross-stack optical transmission was steady at
94%.
An in-stack effluent velocity measurement was made between 17:30
and 17:50. A mean exit velocity of 28.4 m/sec was calculated from
this measurement. The velocity of effluents was measured at 10 cm
intervals across the top of the smoke stack using the LDV. The vel-
ocity profile obtained from this remote measurement is shown in the
lower trace of Figure 4-8. The exit velocities can be seen to vary
between 23 to 30 m/sec across the stack.
The top trace of Figure 4-8 shows the total integrated signal
strength received by the LDV as a function of the position of the
laser beam above the smoke stack.
The data displayed in Figure 4-8 were obtained by taking one
minute duration samples at discrete positions separated by 10 cm and
processing them through the frequency/intensity tracker using a
2 sec. averaging time constant.
4.2.5 TAPE #5: VELOCITY VARIATION BY POWER LOAD CHANGE
Tape #5 was recorded between 9:38 and 13:49 on 17 January 1975.
This tape was made during a period of changing electrical power load on
the generating unit exhausting gases through the number six smoke
stack. The power load decreased from 135 MW to 83 MW in four steps
and then increased to 137 MW in four steps. Two sets of Doppler
4-17
-------
W
M/SEC FT/SEC
150,
40-
H
30-
20-
ilO-
125J
100.
DUKE POWER 16 JANUARY 1975.
RIVER BEND STEAM STATION
No. 6 SMOKE STACK
ELEVATION ANGLE
VELOCITY PROFILE
JUST ABOVE THE STACK LIP
Figure 4-8
The Strength and Frequency of Effluent
Signals as a Function of the Position
above the Smoke Stack Lip from Tape #4:
Velocity Profile.
4-18
8°
-------
measurements were made: one at an 8° elevation angle, the other at
a 20 elevation angle. The laser beam was passed directly through
the smoke stack plume just above the lip of the stack to get an 8°
elevation angle. A relay mirror on top of a platform on the electro-
static precipitators was used to redirect the laser beam through the
plume at a 20° elevation angle. The cross-stack optical transmission
changed during this run from 91% at the start to 98% at 11:00 to
81% at the end of the run. This transmission variation was due to
both the load conditions and the number of precipitators in use.
In-stack effluent velocity measurements were made at each load
level. The effluent velocity was measured remotely with the LDV at
each load level. The velocity data for the 8° elevation angle Doppler
measurements are shown in Figure 4-9. The velocity data for the 20°
elevation angle Doppler measurements are shown in Figure 4-10. The
data in Figures 4-9 and 4-10 were processed through the frequency
tracker with a 2 sec time constant. . The results of the in-stack and
remote Doppler velocity data are summarized in Table 4-2.
4.2.6 TAPE #6: 20° PRECIPITATOR RUN
Tape #6 was recorded between 15:20 and 15:59 on 17 January 1975.
As in Tape #3, various combinations of precipitator banks were turned
on and off to vary the effluent particle content (and size distri-
bution) . The last three banks of electrostatic precipitators (notated:
A, B, and C; A being the last bank) were switched off in various com-
binations to obtain four levels of cross-static optical transmission
from 49% to 94%. A relay mirror was used to direct the laser beam
through the smoke stack plume. The laser beam passed just above the
lip of the stack at an elevation angle of 20 above horizontal. The
power load was constant during the run at 131 MW.
An in-stack effluent velocity measurement was made between 15:15
and 15:30. A mean exit velocity of 41.4 m/sec was calculated from this
measurement. The effluent velocity, measured remotely with the LDV
at various times between 15:20 and 15:59, is shown in the lower trace
4-19
-------
DUKE POWER 17 JANUARY 1975
RIVER BEND STEAM STATION
NO. 6 SMOKE STACK
8° ELEVATION ANGLE
DATA TAPE NO. 5
: : -
i !
M/SEC
70
1
NJ
0
>.
H
i
0
60
rjU
40
30-
20
10-
FT/SEI
250
225 <
200
175
150
125
100
75
50
25
0
! ! =
1 ~
! |
.1
1
, -
|
: ^
' T^ iff ^ WfrllUrXd
-~
i
1
1 MIN L*
Mj^A
:
1
!
!
\ ~
1 ' I
;
1
' !
; 1
!
"
VuuMV ^^
»vm^V^
Vy*(#"\At i»kA_iu T "*"V"Y
i" v ^AV »v^\r *«^^,
i
140MW 125MW .. 110MW : 95MW 80MW 95MW = 110MW 125MW 140MW
POWER LOAD ON GENERATOR
9:42:36 10:19:24 ; 10:49:18 11:15:49 11:47:54 12:15:42 12:46:49 13:18:15 13:48:11
TIME OF DAY IN HR:MIN:SEC
Figure 4-9.
The Frequency of Effluent Signals as a Function of Time
Tape #5: 8° Power Load Change.
for
-------
DUKZ POWER 17 JANUARY 1975
RIV£R SEND STEAM STAT1OH
NO - 6 SMOKE STACK
20° ELEVATION AN3LE
DATA TAPE NO. S
M/SEC,
,,
cjQ
t s40
£ I
S 30
20-
FT/SEC
200
180
160
140
120
:oo
80
60
40
20
1 o
^j 1 MIN V+
rV^J^ftf^ frj^ptyv^
jii/U/Tf^ /r^rWi^V
^^A^^a^
L40MI* 125MW 110MW 9M« 80MW 9SMW UQMW 125MW 14QNH
POWER LOAD ON .5ENER VIVK
9:^8:37 10:12:47 10:43:3^ 11,12:13 11=42 = 07 12:12:15 12:42:34 1 3 - 1 J 5t> li'42-15
TIME OF DAY IN HR:M1N;SEC
Figure 4-10. The Frequency of Effluent Signals as a Function of Time
for Tape #5: 20° Power Load Change.
-------
TABLE 4-2
Comparison of In-Stack and LDV Velocity Measurements
For Tape #5 on 17 January 1975
Time
(Hr:Min)
9:45
10:15
10:45
11:15
11:45
12:15
12:45
13:15
13:45
Load
(MW)
135
124
111
96
83
97
109
122
137
Pitot Exit
Velocity, V
(m/sec) P
44.5
41.4
35.9
32.6
26.6
31.1
35.9
39.3
44.5
8° Elevation Angle
Doppler Exit
Velocity, VD
(m/sec)
36.6
34.1
30.5
25.3
24.4
27.4
30.5
34.4
38.1
V /V
VPX D
1.22
1.17
1.18
1.29
1.09
1.13
1.18
1.14
1.17
20° Elevation Angle
Doppler Exit
Velocity, VD
(m/sec)
39.6
35.9
31.4
27.4
25.3
28.1
32.6
36.2
41.2
Vp^D
1.12
1.08
1.15
1.19
1.05
1.11
1.10
1.08
1.08
4-22
-------
of Figure 4-11. The mean exit velocity measured by the LDV was
about 38.1 m/sec.
The upper trace of Figure 4-11 shows the relative integrated
signal strength received by the LDV as a function of the electro-
static precipitators in use on the stack.
The data displayed in Figure 4-11 were processed through the
frequency/intensity tracker using a 2 sec averaging time constant.
4.2.7 TAPE #7: 28° CW RUN
Tape #7 was recorded between 16:42 and 16:55 on 17 January 1975.
Like Tapes #1 and 2, it was a CW run. A relay mirror on top of the
electrostatic precipitators was used to direct the laser beam through
the smoke stack plume. The beam passed just above the lip of the
stack at an elevation angle of 28° above horizontal. The power load
was constant at 131 MW throughout the run. The cross-stack optical
transmission was steady at 87%.
An in-stack effluent velocity measurement was made between 16:30
and 16:45. An exit velocity of 43.3 m/sed was calculated from this
measurement. The effluent velocity, measured remotely with the LDV
during the time interval from 16:42 - 16:55, is shown in the lower
trace of Figure 4-12. The mean exit velocity measured by the LDV was
about 35.7 m/sec.
The upper trace in Figure 4-12 shows the relative integrated
signal strength received by the LDV during the time interval from
16:42 to 16:55.
The data displayed in Figure 4-12 were processed through the
frequency/intensity tracker using a 2 sec averaging time constant.
4.2.8 TAPE #8: 37° CW RUN
Tape #8 was recorded between 17:16 and 17:31 on 17 January 1975.
It was a CW run similar to Tape #7. A relay mirror on top of the
electrostatic precipitators was used to direct the laser beam through
the smoke stack plume. The laser beam passed just above the lip of
4-23
-------
i
M
IJUKE POWER 17 JANUARY I1
RIVER BEND aTEAM STATION
NO. SMOKE STACK
iO'-' ELEVATION ANGLE
UATA TAPE NO. b
6"C" BANKS UFF
.
"A-li-B- BANKS UFF
- BANK OFF
ALL BAJiKi UN
V '
3J la: 34 15:35
TIME OF DAY IN H« :MIN
TIME OF DAY IN HR:MIN
IS: 57 15:58 IS: 5,9 15:4-' 15:48 15:49 15*5.0 15s 12 15:33 15:34 15:35 15: iO 15:21 15:^^ 15:23
»- . 1 « 1 i 1 ^ M * 1 $$ , 1 . »
Figure 4-11. The Strength and Frequency of Effluent Signals as a
Function of Time for Tape #6: 20° CW Run.
-------
DUKE POWER 17 JANUARY 1975
. RIVER BEND STEAM STATION
NO. 6 SMOKE STACK
28° ELEVATION ANGLE
DATA TAPE NO. 7
Figure 4-12. The Strength and Frequency of Effluent Signals
as a Function of Time for Tape #7: 28° CW Run.
4-25
-------
the stack at an elevation angle of 37° above horizontal. The power
load was constant at 131 MW throughout the run. The cross-stack
optical transmission was steady at 87%.
No in-stack effluent velocity measurement was made during the
run. However, conditions were stable after the previous measurement
which was taken during the time interval from 16:30 to 16:45. An
exit velocity of 43.3 m/sec was calculated from this measurement. The
effluent velocity, measured remotely with the LDV during the time
interval from 17:16 to 17:31, is shown in the lower trace of Figure
4-13. The mean exit velocity measured by the LDV was about 38.1 m/sec.
The upper trace in Figure 4-13 shows the relative integrated signal
strength received by the LDV during the time interval from 17:16 to
17:31.
The data displayed in Figure 4-13 were processed through the
frequency/intensity tracker using a 2 sec averaging time constant.
4.2.9 TAPES #9, 10, 11, and 12: 8° VELOCITY PROFILES
Tapes #9, 10, 11, and 12 were recorded between 7:40 and 11:30 on
18 January 1975. These runs were high resolution profiles of the
exit velocity distribution across the top of the smoke stack. The
laser beam was scanned across the top of the stack at heights from
0 to 3.7 m above the lip. No power load data were collected for these
runs. The cross-stack optical transmission increased gradually from
90 to 94% during the course of the runs.
Three in-stack effluent velocity measurements were made while
the profiles were in progress. The mean exit velocity was calculated
to have the values of 36.9 m/sec during the time interval from 8:15 to
8:30, 36.3 m/sec between 9:00 and 9:15, and 36.9 m/sec from 10:00 to
10:15. The velocity of the effluents was measured at 10 cm intervals
across the top of the smoke stack using the LDV. Profiles were made
at heights of 2 cm, 0.9 m, 1.4 m, and 3.7 m above the lip_of the stack .
The velocity profiles obtained by these remote measurements are shown
in Figure 4-14. The variation of the profiles with increasing height
4-26
-------
B
B
H
in
rtt|
LC;T
DUKE POWER 17 JANUARY 1975
RIVER BEND STEAM STATION
NO. 6 SMOKE STACK
37° ELEVATION ANGLE
DATA TAPE NO. 8
Figure 4-13. The Strength and Frequency of Effluent
Signals as a Function of Time for Tape
t8: 37° CW Run.
4-27
-------
RIVER BEND STEAM STATION
No. 6 SMOKE STACK
8° ELEVATION ANGLE
VELOCITY PROFILE
DATA: TAPES #9, 10, 11, & 12
M/SEC FT/SEC
J_
I
I
40. .
PROFILE 12 FEET 30
ABOVE STACK LIP
20-
PROFILE 4.5 FEET
ABOVE STACK LIP
PROFILE 3 FEET
ABOVE STACK LIP
PROFILE AT LIP
OF STACK
Figure 4_14.
The Frequency of Effluent Signals as a Function of
the Position and Height Above the Smoke Stack Lip
from Tapes #9, 10, 11, and 12: 8° Velocity
Profiles.
4-28
-------
is considerable. Velocities range from 15 m/sec at the edges to
greater than 38 m/sec at points in the middle. The data displayed
in Figure 4-14 were obtained by taking one minute duration samples at
discrete positions 10 cm apart and processing them through the fre-
quency tracker using a 2 sec averaging time constant.
4.3 DATA ANALYSIS
The twelve tapes were analyzed to determine: (1) the degree of
agreement between LDV and pitot tube measurements of smoke stack
effluent exit velocities, (2) the correlation of effluent backscatter
signal strength with the cross-stack optical transmission, (3) the
range of values of the effluent backscatter coefficient at 10.6 pm,
and (4) the effects of plume turbulence on the backscattered Doppler
spectra. The analyses are discussed in detail below.
4.3.1 LDV AND PITOT TUBE VELOCITY MEASUREMENTS
Tapes #1, 3, 5, 6, 7, and 8 were used to determine the degree of
agreement between LDV and pitot-tube measurements of smoke stack exit
velocity. During each of these runs, the power load was steady at a
value between 83 MW and 137 MW and both remote and in-stack effluent
velocity measurements were made.
The results of this analysis are tabulated in Table 4-3 and shown
graphically in Figure 4-15. It appears that the effluent exit veloc-
ity from a power plant smoke stack is a linear function of the power
load on the generating unit exhausting gases through the stack. A
linear least squares fit to the remote LDV velocity data is shown in
Figure 4-15 by the solid line. It indicates that the exit velocity
can be found from the equation:
v (m/sec) = 0.26 load (MW) + 3.05 (4_1)
4-29
-------
TABLE 4-3
Velocity vs. Load Data
16 - 17 January 1975
1
Load
MW
83
83
83
92
96
96
97
97
109
109
111
111
122
122
124
124
130
130
130
135
135
137
137
Pitot
Velocity
(m/sec)
26.6
26.6
30.5
33.5
32.6
32.6
31.1
31.1
36.0
36.0
36.0
36.0
39.3
39.3
41.5
41.5
41.5
43.3
43.3*
44.5
44.5
44.5
44.5
Doppler
Velocity
(m/sec)
24.4
25.3
27.4
29.0
25.3
27.4
27.4
28.0
30.5
32.6
30.5
31.4
34.4
36.3
34.1
36.0
38.1
35.7
38.1
36.6
39.6
38.1
41.1
Tape
5
5
3
1
5
5
5
5
5
5
5
5
5
5
5
5
6
7
8
5
5
5
5
Velocity Ratio = v_/V_
Standard Deviation = a (Vp/Vn)
Percentage (Vp/vD) x 100
8°
109
116
129
113
118
118
114
121
122
117
1.177
0.055
20°
105
111
119
111
110
115
108
115
109
112
108
. 1.112
0.040
28°
121
1.21
-
37°
114
1.14
-
Linear Fit: Pitot v = 0.31L + 2.38
a =
Total Ratio: 1.14+0.056
1.10 m/sec
Doppler v = 0.26L + 3.05 a = 1.40 m/sec
* Same as Tape #7
4-30
-------
= 1.10 n/sec
= 0.26 Megawatts
-T+rillli !l!:ilTl.llll
to = 1.40 m/sec
SAMPLTHB
Figure 4-15
Effluent Exit Velocity as a Function of
Power Load.
-------
The standard deviation of the Doppler data from this equation is
1.40 m/sec, which indicates that the linear equation provides a good
approximation for estimation of the exit velocity.
The in-stack velocity measurements yielded exit velocities which
averaged 14% higher than the LDV measured velocities. The results is
shown in Figure 4-16 where the LDV velocity data is plotted against
the pitot-tube velocity data. A linear least squares fit to the pitot
tube data indicated that exit velocity as a function of load followed
the equation:
v (m/sec) = 0.31 megawatts + 2.38 (4-2)
The standard deviation of the pitot-tube data from this equation is
1.10 m/sec again indicating that a good fit was made.
The cause for the 14% discrepancy between the remote and in-stack
velocity data is not known. A number of possible causes are: (1) error
in measuring the LDV's elevation angles, (2) miscalibration of the
pitot tube used on the in-stack measurements, (3) erroneous correction
factor used in calculating the gas exit velocity from the in-stack
velocity, and/or J4) unaccounted for factors such as compression or
thermal cooling of gases in the stack. In any event the dis-
crepancy appears to be the result of a systematic error rather than
a random error. If such an error can be accounted for, the LDV should
be able to make remote smoke stack effluent velocity measurements
accurate to about 1.5 m/sec or less.
4.3.2 SIGNAL STRENGTH AS A FUNCTION OF CROSS-STACK TRANSMISSION
Tapes #3 and 6 were used to determine the relationship between
the LDV's received signal strength backscattered from the smoke stack
effluents and the cross-stack optical transmission as measured on a
Lear Seigler RM-4 transmissometer. During these runs the electro-
static precipitator banks on the stack were turned on and off in
various combinations to change the cross-stack optical transmission.
It was assumed that both absorption and single scattering determined
the optical transmission across the stack. Since the attenuation
coefficient is proportional to particle concentration, the optical
transmission should obey the relationship:
4-32
-------
Figure 4-16. Effluent Exit Velocity Measured by a Laser
Doppler Velocimeter as a Function of the Exit
Velocity Measured by an In-Stack Pitot Tube.
4-33
-------
T = exp [-a(N)L ] (4-3)
s
where T is the optical transmission, L is stack diameter, and «(N)
S
is the attenuation coefficient and is linearly proportional to the
particle concentration, N.
Unfortunately varying the number of precipitators in use also
changes the effluent particle size distribution as well as the
particle concentration. This result is shown in Table 4-4. The
precipitators tend to remove the larger particles from the smoke
stack effluents. The scattering at 10.6-|jm from these particles is
described by Mie theory. The Mie scattering functions depend on two
parameters: the index of refraction of the particle and a size
parameter, |_i, given by the equation:
2nr ,.. ..
t-i = (4-4)
where r is the particle radius and X is the wavelength. The Mie
scattering functions for the stack are very difficult to evaluate
since they require exact knowledge of the size distribution and chem-
ical distribution of the effluent. It was hoped, however, that a
qualitative rather than an exact quantitative correlation would exist
between transmission in the visible and scattering at 10.6 urn.
In order to determine the relationship between the backscattered
signal strength and the cross-stack optical transmission, the total
integrated signal strength taken from the intensity tracker was
plotted as a function of the attenuation coefficient measured in the
visible. Assuming that there is no compression of the exhaust gases
through the stack constriction, the attenuation coefficient at the
stack exit has the same value as the in-stack attenuation coefficient
described in Equation (4-3). The data from two different runs is
tabulated in Table 4-5 and is shown graphically in Figure 4-17, where
the relative integrated signal strength is plotted as a function of
the optical attenuation coefficient. The solid line in the figure
represents a least squares linear fit to the data points. The
apparent linearity of the graph in Figure 4-17 indicates a linear
4-34
-------
TABLE 4-4
Smoke Stack Effluent Particle Concentration and Emission Data.
Data Taken from Environmental Science and Engineering. Inc.
Measurements Under EPA Contract No. 68-02-0232, Task No. 45,
Sub-task No. 3. 26-30 August 1974.
Date
8/28/74
8/29/74
Time
(Hr:Min)
16:50 -
19:55
11:15 -
14:30
Load
(MW)
140
140
Exit
Velocity
(m/sec)
48.3
51.5
Opacity
(%)
~ 5
30
Precipitators
Off
Unknown
Unknown
Emission
(kg/hr)
18
99.5
Particle Size
(^im)
> 8.42
8.42 - 5.03
5.03 - 3.47
3.47 - 1.89
1.89 - 1.25
< 1.25
> 5.77
5.77 - 3.43
3.43 - 2.63
2.63 - 1.27
1.27 - 0.83
< 0.83
Percent
Collected
0
3.57
2.23
0.44
1.79
91.97
24.34
14.44
8.87
2.26
3.13
46.96
I
LJ
UI
-------
TABLE 4-5
Relative Integrated Doppler Signal Strength as a Function
of Cross-Stack Transmission and the Optical Attenuation
Coefficient. Data from Tape #3. 16 January 1975 and Tape #6
Precipitator
Conditions
A Bank off
15:56
16 January 1975
A & B Banks off
16:04
16 January 1975
A, B, & C Banks off
16:10
16 January 1975
B & C Banks off
16:17
16 January 1975
C Bank off
16:25
16 January 1975
B Bank off
16:32
16 January 1975
No Banks off
16:38
16 January 1975
A Bank off
16:45
16 January 1975
No Banks off
15:20 - 15:23
17 January 1975
A Bank off
15:32 - 15:35
17 January 1975
A & B Banks off
15:47 - 15:50
17 January 1975
A, B, & C Banks off
15:56 - 15:59
17 January 1975
Cross-Stack
Optical
Transmission
(%)
94
82.5
60
93.5
94
97
98.5
94
93.5
89
74
49
Optical
Attenuation
Coefficient
(m-1)
2.09 x 10~2
6.50 x 10"2
1.73 x 10'1
2.27 x 10~2
2.09 x 10~2
1.03 x 10~2
5.11 x 10~3
2.09 x 10~2
2.27 x 10~2
3.94 x 10"2
1.02 x 10'1
2.41 x 10"1
Relative
Integrated
Signal
Strength
0.180
0.299
0.718
0.191
0.150
0.066
0.060
0.120
0.129
0.243
0.457
1.000
4-36
-------
Data taken from Tape #6 on 17 Jan. 1975
mm mm,
Figure 4-17. Relative Integrated Doppler Signal Strength as a
Function of the Effluent Attenuation Coefficient.
-------
relationship between the received backscattered signal at 10.6 urn
and effluent particle concentration. However, a clarifying point
must be made on this point. The intensity tracker integrates the
signals from the spectrum analyzer between preset frequency limits.
The spectrum analyzer was set to give a linear display. For a heter-
odyne system, a linear spectrum analyzer output provides a signal
output which is proportional to the received signal's electric field
strength rather than optical power. The signal strength plotted in
Figure 4-17 is an integrated electric field strength. Because of the
integration, there is no simple relationship between the integrated
electric field strength and total received optical power.
If a good amplitude calibration can be established for the LDV,
and the smoke stack under test, it appears that the total received
signal strength measured by the LDV could be used to conveniently
determine the optical transmission through the smoke stack plume.
If sufficient information on particulate concentration and optical
transmission is available, " the LDV could be used to measure smoke
stack effluent particle concentrations.
4.3.3 THE EFFLUENT BACKSCATTER COEFFICIENT
Estimates of the backscatter coefficient can be made using
theoretical analysis or a variety of experimental techniques. Eval-
uation of the effluent backscatter coefficient by theoretical analysis
is verv difficult since determination of the Mie scattering
functions requires exact knowledge of the size distribution and chem-
ical composition of the effluent.
With regard to evaluation of the backscatter coefficient by
experimental techniques, the experimentation should be carried out
at the wavelength of interest since the backscatter coefficient can
vary with wavelength. One technique is to calibrate the system
making the measurements, so that the signal amplitude can be related
to the backscatter coefficient. According to Sonnenschein and
Horrigan the signal-to-noise ratio for a focussed, coaxial, laser
heterodyne system which collects scattered radiation from a small
length, AL, around the focus is given by the equation:
4-38
-------
SNK = « ' LJULfifc. (4_5,
1
J
where
TI is the system efficiency,
s
PT is optical power output of the laser,
3 (n) is the backscatter coefficient,
D is the receiver optics diameter,
L is the range to the target (L equals the focal distance
in Equation (4-5),
hv is the photon energy of a quanta of laser radiation, and
B is the noise bandwidth.
The use of Equation (4-5) assumes the use of a shot-noise limited
photoconductive detector which was the case in the smoke stack
effluent measurements. Equation (4-5) can be solved for the back-
scatter coefficient giving the equation:
B(TT) = 8hVB L 2 SNR (4-6)
TT ns PT D^ AL
For the smoke stack effluent measurements the bandwidth, B,
was fixed by the spectrum analyzer IF filter at 100 kHz; the range,
L, to the stack was measured on a laser rangefinder to be 400 m; the
2m stack exit diameter determined the resolution length, ^L; the
receiver optics diameter was 0.3m; and the photon energy, hv for
the 10.6-um CO- laser was 1.875 x 10~ Joule. If these values are
substituted into Equation (4-6) p (TT) follows the relationship:
P(TT) = 4.25 x 10"9 SN* Cm'1] (4-7)
^s T
4-39
-------
The signal-to-noise ratio, system efficiency, and laser power
were measured at various values of cross-stack optical transmission.
The results of these measurements are plotted in Figure 4-18. The
backscatter coefficient at any level of optical transmission has a
wide range of values. This range of values is the result of the
rapping cycles in the electrostatic precipitators. The fluctuations
are filtered out of the transmissometer readings by averaging over
long time intervals. The solid line in Figure 4-18 is a plot of
the mean value of the backscatter coefficient as a function of the
cross-stack optical transmission.
It was originally hoped that the backscatter coefficient would
be a linear function of the effluent particle density. A good indi-
cation that such a relationship would not be found was the linear
relationship established between the integrated electric field
strength output from the intensity tracker and the effluent particle
density. Since the heterodyne signal-to-noise ratio is proportional
to the received optical power backscattered from the target, it was
suspected that the backscatter coefficient would obey a functional
relationship of the form:
3(TO = kj"a(N)Lelm (4-8)
where k is a constant; Ot (N) is the optical attenuation coefficient
proportional to effluent particle concentration, N; L is exit dia-
meter of the smoke stack; and the exponent m is a constant whose
value is approximately 2. In order to confirm this assumption, the
backscatter coefficient, p (n), was plotted as a function of the
natural logarithm of the optical transmission at the stack exit on
log-log graph paper. The results are shown in Figure 4-19. It
appears that the effluent backscatter coefficient at 10.6 urn can be
approximated by the equation:
PM [m"L] = 5.1 x 10~4 (-LnTe)1%6 {4-9}
where T is the optical transmission at the smoke stack exit.
4-40
-------
CROSS STACK OPTICAL TRANSMISSION
Figure 4-18- Effluent Backscatter Coefficient as a
Function of the Cross-Stack Transmission,
4-41
-------
-FF4UENT
Figure 4- 19.
The Effluent Backscatter Coefficient Plotted as
a Function of the Optical Transmission at the
Smoke Stack Exit.
4-42
-------
The reason that the backscatter coefficient follows such a
relationship is not known. It may be the result of variations in
the effluent particles' scattering functions caused by changes in
the effluent particle size distribution.
4.3.4 TURBULENCE EFFECTS
A simple model of the smoke stack exhaust velocity profile and
the LDV's interaction with it can be used to estimate the effects
of velocity spread and turbulence on the Doppler signal. A power
law velocity profile of the form
u(r) = uc(l-r/R)1/n (4-10)
is assumed to describe the velocity in the stack near the sampling
ports. In Equation (4-10), u(r) represents the velocity component
parallel to the axis of the stack at a radius r from the center line
of the stack; u is velocity at the center of the stack; and R is
the radius of the stack. The value of the exponent n is determined
by the Reynold's number for the smoke stack. For large values of
Reynold's number the exponent n is approximately equal to 7. ' The
Reynold's number is the criterion for determining if a flow is
laminar or turbulent. When the Reynold's number is less than 2100
the flow is laminar; when it is greater than 3100 it is turbulent
and n is approximately equal to 7. The Reynold's number is a func-
tion of four parameters: average velocity, density, viscosity, and
a characteristic dimension of passage such as the tube diameter.
The Reynold's number of the smoke stack at Duke Power is in the
(4)
30,000 to 40,000 range indicating a turbulent gas flow. The velocity
profile inside the stack and below the constriction has a flattened
axial velocity profile of the form:
u(r) = uc(l-r/R)1/7 (4-11)
This profile is plotted in the lower distribution of Figure 4-20.
The smoke stack is topped with a converging section 3.05 m in
length which reduces the diameter from 2.96 m to 1.94 m. This con-
striction results in an increase in the average flow velocity by a
factor of tl.52) (assuming no compression occurs). It also further
flattens the velocity profile. Along each streamline Bernoulli's
4-43
-------
Figure 4-20. Theoretical In-Stack and Exit Velocity Profiles.
4-44
-------
equation requires that the exit velocity, v, be related to the
entrance velocity, u, according to the formula:
v2 = u2 + | (Pl-p2) (4-12)
where p is the density of the stack cases and p, and p~ are the
respective entrance and exit pressures on the converging section.
The pressure term in Equation (4-12) can be estimated by assuming a
flat velocity profile in the downstream stack section. Conservation
of mass then requires that:
2(Pl-p2)/p * v2 - u2 = u2 [(v'/u)2 - l] (4-13)
where the u and v are the respective mean entrance and exit ve-
locities. For the Duke Power smoke stack, the mean exit to mean
entrance velocity ratio is 2.32 which determines that:
2(p1-P2)/P * 4.41u2 (4-14)
By substituting Equation (4-14) into Equation (4-12), the exit ve-
locity at a radius r from the stack center line can be found. The
exit velocity is given by the equation:
r _ 11/2
v(r/1.52) = [u(r)2 + 4.41u2J (4-15)
The average flow velocity below the stack construction can be
found by averaging Equation (4-11) over the cross-sectional area
of the stack. Hence:
/
_ -, /7
u = 5s- / r(l-r/R) ' dr = 0.817 u (4-16)
TTI c
By utilizing Equations (4-11), (4-16) and (4-15), the exit velocity
distribution can be found to have the form:
4-45
-------
- r 2/7 11/2
v(r/1.52) = u 1.50(l-r/R) n + 4.4l| (4-17)
The exit velocity distribution is plotted in the upper distribution
of Figure 4-22 for comparison to the in-stack distribution. The
exit profile is flat except near the exit wall. Even near the wall
the exit profile is flatter than entrance profile. These theoretical
results are more or less confirmed by the experimentally obtained
profiles shown in Figures 4-2., 4-8, and 4-14. Neglecting asymmetries
and velocity fluctuations within the profile, the experimental pro-
files are fairly flat and exhibit the high wall velocities predicted
by the simple fluid dynamical model.
The flow out of the stack is turbulent and has fluctuating
velocity components in the direction of the flow and perpendicular to
the flow. The turbulence velocity in the direction of the flow is
assumed to have an RMS velocity, v'. The turbulence velocities
Jt
perpendicular to the flow can be broken up into radial and
tangential components with RMS velocities, v1 and v ', respectively.
The component of velocity, vn / parallel to a laser beam
passing through the centerline of the stack dust above the lip at
an elevation angle, 6, is given by the equation:
v,| = (v + v^) sin 6 + v^ cos 9 (4-18)
Since the stack exit diameter is much smaller than the LDV's range
resolution, the entire distribution of velocities through the stack
will be detected. The backscattered radiation will be Doppler
shifted by amount:
_ £
*VD X VH (4-19)
If the flow and turbulence velocity profiles are assumed to be nearly
flat across the stack, the Doppler shift will have a mean frequency:
MJ = - v sin 9 (4-20)
4-46
-------
where the superimposed bar is used to denote an average. The Doppler
spectra will have a frequency width, Af, given by the equation:
sin 0 + v£ cos 6 (4-21)
By analyzing the Doppler spectra from various elevation angles, es-
timates can be made for the axial and radial turbulent velocity com-
poents.
The Doppler spectra from four different elevation angles are
shown in Figure 4-21. These spectra were analyzed to determine the
mean Doppler frequency, the spectral width, and the ratio of the
spectral width to the mean Doppler frequency. These data are tabu-
lated in Table 4-6. In order to determine the turbulence parameters,
Equation (4-21) was divided by Equation (4-20) to produce the nor-
malized equation:
Afn 2v"' 2v'
^P- = -2- + -£- cot 9 (4-22)
AV v v
The data in Table 4-7 were fitted to Equation (4-29) to obtain esti-
mates for the normalized turbulence velocities, v'/v and "v'/v. The
X x
fit shown in Table 4-7 was obtained with the turbulence values:
v^/v = 0.040 (4-23)
vr/v = 0.057 (4-24)
These values of relative turbulence can be compared to previously
measured values of relative turbulence intensities. Figure 4-22
shows relative turbulence intensities measured in a fully developed
pipe flow with a Reynolds number of 5 x 10 . Our fitted values can
be seen to have the same range of values as those measured by Laufer.
Undoubtedly the turbulence effects increase with height above
the stack lip. A mixing region between the stationary air and
effluent gases develops at the lip of the stack and spreads upward
4-47
-------
I
J^
00
O.SMHz/DIV
= 8°
GROUND WIND
0.5MHZ/DIV
6 = 20°
I.OMHz/DIV
0 = 37°
Figure 4-21. Typical Doppler Spectra from Smoke Stack Effluents
at Various Laser Elevation Angles.
-------
TABLE 4-6
Data on Doppler Spectra of
Smoke Stack Effluents
6
(Deg)
8
20
28
37
^D
(MHz)
1.2
2.4
3.2
4.3
*fD
(MHz)
1.06
0.99
0.96
0.95
AV^D
Experimental
0.88
0.41
0.29
0.20
AV**D
Fitted Data*
0.88
0.40
0.29
0.23
* Fit with v'/v = 0.040 and v'/v = 0.057
X JL
0 0.1 0.2 03 04 0.5 06 0.7 08 0.9 1.0
Figure 4-22. Relative Turbulence Intensities in Pipe
Flow. (Laufer, J.; Reprinted from NACA
Tech. Repts. 1174, pp. 6 and 7, 1954.)
4-49
-------
with an angle of about 3°. ' This spreading can be observed in
the velocitiy profiles shown in Figure 4-14. In the mixing layer
the flow velocity is approximately one half the flow velocity at the
exit wall. The turbulence velocities in the mixing region can reach
15% of flow velocity at the exit wall. However, the particle
concentration in the mixing region is reduced, so that the intensity
of returns from this area is diminished. The increased turbulence
in the mixing region could broaden the large elevation angle Doppler
spectra. If the beam passes close to the lip of the stack, the
broadening should not be significant for elevation angles less than
40°.
4-50
-------
SECTION 5
SYSTEM ANALYSIS
5.1 INTRODUCTION
A preliminary system analysis for a OX laser heterodyne
system to be used for the measurement of the velocity, and possibly
particulate mass concentration, of the effluent emerging from smoke
stacks of power plants has been performed. This analysis is based
upon the theory of laser heterodyning and upon the results of the
measurement program described in Section 4. The system to be
analyzed is shown in Figure 5-1.
co2
LASER
[NTERFEROMETER
TELESCOPE
DETECTOR
SIGNAL
PROCESSOR
Figure 5-1. System Block Diagram.
5.2 HETERODYNE DETECTION
Radiation from the laser illuminates a moving target, in this
case the aerosols emitted from the power plant smoke stack. The
laser radiation is scattered by the aerosols, and the scattered
radiation has a frequency shifted from the original laser frequency,
For the system illustrated in Figure 5-1, where a common trans-
mitting and receiving telescope is used, the frequency shift, &vn,
2v_
(5-1)
5-1
-------
where
Vj| = aerosol velocity component parallel to the line
connecting transmitter and target
X = laser wavelength
For a C02 laser, the frequency shift is approximately 33 "kHz per
m/sec of the aerosol velocity component. For a vertical flow,
AVD = Y v sin 9 (5-2)
where
v = magnitude of aerosol velocity
6 = transmitter elevation angle
Radiation collected by the receiving telescope is combined
with radiation at the original laser frequency. When this com-
bination illuminates a square law detector-, energy at the dif-
ference frequency,AVD, is obtained. A measurement of this fre-
quency combined with knowledge of the system geometry enables the
determination of the magnitude of the aerosol velocity.
5.3 SIGNAL TO NOISE RATIO
The ability of the laser system to measure the difference
frequency, and therefore the aerosol velocity, is determined by the
system signal to noise ratio. A theoretical derivation of this
quantity for a laser heterodyne system against a target that is a
collection of aerosols is given in Reference 1. This shows that
N
(5-3)
5-2
-------
where
TIQ = quantum efficiency
Tis = system efficiency
P = transmitter power
p (TT) = effluent backscatter coefficient
h = Planck's constant
v = laser frequency
B = system bandwidth
L- = maximum target range
L. = minimum target range
f = range to focus of laser beam
D = optics diameter
If the target extent, &L ,
AL = L2 - L! <5-4>
is much greater than the mean range, L,
L = f » 6L <5~5>
then as shown in Reference 1
_
8
For a photovoltaic detector, the noise is decreased by a factor
of two; additionally a more exact calculation shows
2AL (5_7)
S_ m -
N " 4hvBL2
It is important to select realistic system parameters that will
be valid for a variety of applications. From systems that have been
actually constructed,
nQ = 0.5
V - o-1
5-3
-------
Reasonable stack parameters and ranges appear to be
250m < L < 1000m
AL = 2m
and from the results of Section 4, the lowest value of backscatter
coefficient and the highest value of bandwidth are
B
= 3 x 10"6 m"1 ster'1
= 5 x 105 Hz
Then
| = 1.01 x 107 PT (S_)
(5-8)
Plots of equation (5-7) appear in Figure 5-2 through Figure 5-4 , for
laser powers of 1 W, 5 w, and 23 W and for ranges of 250 m, 500 m,
and 1000 m.
From these curves, a combination of optics sizes and laser powers
can be obtained for any given signal-to-noise ratio. A signal-to-
noise ratio of approximately ten is desired. For no signal integra-
tion, these values are shown in Table 5-1.
TABLE 5-1
Trades Between Power and Optics Size
tfo Xntearation
Laser Power (W)
1
5
20
Optics Diameter (cm)
L = 250 m
25
11
5.5
L = 500 m
50
22
11
L = 1000 m
100
44
22
The use of a signal processor to integrate the return signal increases
system signal-to-noise ratio by approximately the square cost of the
number of pulses to be integrated. A processor, similar to the one
5-4
-------
3D
TARGET RANGE = 250 m
~:-
~
-' "-; ' . J_-L -.; ' ' '' I " 'r;-:3.4 ': :
:£ll-:E:^_:4:;y :S^
"
-30
Optics Diameter (cm)
Figure 5-2. System Signal-to-Noise Ratio
for Range = 250 m.
5-5
-------
30
i TARGET RANGE
; ;.[... .;;:."';,, : ;
_L_ ' '" '
-30
10
Optics Diameter (cm)
Figure 5-3. System Signal-to-Noise Ratio
for Range = 500 m.
5-6
-------
gii-_ !-;;_j_ ...'.. i-ji.;.!-':. TARGET RANGE = 100P m
100
Optics Diameter (cm)
Figure 5-4.
System Signal-to-Noise Ratio
for Range = 1000 m.
5-7
-------
required for this application, has been fabricated by Raytheon
Company. It integrates a Doppler spectrum for times of the order
of 1 msec and obtains approximately a 10 dB improvement in signal-
to- noise ratio. With such a processor, operation can be accomplished
with a pre-integration signal-to-noise ratio of one. The trade
between laser power and optics diameter for this case is shown in
Table 5-2.
TABLE 5-2
Trades Between Power and Optics Size
Integration
Laser Power (W)
1
5
20
Optics Diameter (cm)
L = 250 m
8
3.6
1.8
L = 500 m
16
7.2
3.6
L = 1000 m
32
14
7.2
Therefore, depending upon choices of laser power, target range, and
integration, optics sizes can run from less than 2 cm to 1 m and
laser power can vary from 1 W to 20 W.
5.4 BEAM SIZE
The analysis in Section 5.3 assumed that the beam size on target
was smaller than the target. In fact, to insure a relatively uni-
form region of flow, it is reasonable to keep the beam size no
greater than one quarter of the stack diameter.
The far field distance of an aperture, dFF<
.2
dFP ~ \
(5-8)
For ranges less than the far field distance, the optical beam can be
focused and will be smaller than the transmitter diameter. For ranges
greater than the far field distance, the beam diameter,
5-8
-------
R X
-§- <5-9'
For operation at a range of 1000 m, the condition that D^R be greater
than 0.5 m (one-quarter of the stack diameter) requires an optics
diameter of 2 cm or greater. Operation at shorter ranges permits the
use of smaller optics. Referring to Tables 5-1 and 5-2 shows that
this requirement presents no problem.
5.5 BANDWIDTH CONSIDERATIONS
The system must have a bandwidth sufficiently wide to handle
all potential Doppler shifts. If a maximum vertical velocity of
45 m/sec at an elevation angle of 45° is assumed, then a maximum sys-
tem bandwidth of 6 MHz is required. However, since the signal fre-
quency varies with elevation angle and the angle is fixed for a set of
measurements, a smaller instantaneous bandwidth, approximately 1 MHz
to 2 MHz, can be used. The center frequency of the processor can be
tuned as a function of elevation angle.
5-9
-------
SECTION 6
CONCLUSIONS
During the course of this program, a C02 Laser Doppler Veloci-
meter was tested against a coal burning power plant equipped with
electrostatic precipitators. The purpose of the tests was to prove
the feasibility of making remote velocity measurements of the ex-
haust gases from power plant smoke stacks. Remote effluent velocity
measurements were made at a slant range of 400 m from the smoke stack
and at elevation angles of 8°, 20°, 28°, and 37° from the horizontal.
These measurements were made under a variety of power plant opera-
ting conditions, including different exit velocities obtained by
varying power plant load conditions from 80 MW to 140 MW, and
different particulate concentrations obtained by varying the pre-
cipitator operating conditions. In-stack measurements of the flue
gas velocity and optical transmission were taken by the EPA for com-
parison purposes. Ground wind velocities were also measured. These
were typically less than 2 m/s and had no effect on the stack exit
velocity. The measurement data from the LDV was processed, recorded
on magnetic tape, and later analyzed. The following conclusions can
be reached as a result of that analysis:
A Laser Doppler Velocimeter has conclusively shown its ability
to remotely measure the velocity of effluents from a power plant
smoke stack. Velocity data was both selfconsistent, as shown from
profile measurements, and in good agreement with pitot tube data.
Deviations between the LDV and the pitot tube were of a systematic,
rather than a random, nature and are probably associated with a
relative miscalibration between the two instruments. A least square
fit analysis of the velocity data indicates an accuracy of approx-
imately 1.5 m/sec, although greater accuracy has been obtained in
other LDV measurement.
6-1
-------
A linear relationship was established between the relative in-
tegrated electric field strength received by the LDV and the attenua-
tion coefficient measured in the visible through the smoke stack plume
Since the attenuation coefficient is proportional to particle con-
centration' it appears that the signal strength received by the LDV
can be used to evaluate smoke stack effluent particle concentrations.
A good amplitude calibration would have to be established for each
smoke stack before such measurements could be made, but an LDV has
the potential for making mass emission rate measurements on power
plant smoke stacks.
The effluent backscatter coefficient at 10.6-|am was evaluated as
a function of the optical transmission through the smoke stack plume.
It appears that the effluent backscatter coefficient at 10.6-|am can
be approximated by the equation:
- 5.1 x ID'4 (-LnT1'6 (6-2}
where Te is the optical transmission at the smoke stack exit. Based
on measurements made during the first field tests an effluent back-
scatter coefficient of 10~4 m~ corresponds to an effluent particle
density of approximately 0.2 g/m .
It was shown that the turbulent flow in the smoke stack flattens
the velocity distribution of the gases in the stack. The converging
section at the top of the stack further flattens the velocity profile
so that the exit velocity at the wall is 88% of the peak exit velocity
at the center. This flattening was observed in the experimentally
measured velocity profiles across the top of the smoke stack. The
flat exit velocity distribution of the exhaust gases from the stack
relaxes the alignment requirements on the LDV.
Estimates were made of the radial and axial turbulent intensity
components. These estimates were based "on -tire broadening of the
Doppler spectra observed at various elevation angles between 8 and
37°. The axial and radial turbulence intensities were found to
have the mean relative values:
6-2
-------
v'/v = 0.040 (6-3)
J\
v^/v = 0.057 (6-4)
These turbulence components produce a broadening of the Doppler
spectra which is relatively independent of elevation angle. The
magnitude of the broadening was sufficiently small that suitable
effluent velocity measurements could be made at any elevation angle
between 8° and 37°.
6-3
-------
SECTION 7
REFERENCES
1. C. M. Sonnenschein and F. A. Horrigan, "Signal-to-Noise Relation-
ships for Coaxial Systems that Heterodyne Backscatter from the
Atmosphere", Applied Optics. Vol. 10, No. 7, July 1971, pp. 1600 -
1604, (see equation 23).
2. R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport
Phenomena, John Wiley and Sons, Inc., New York, New York, 1960.
3. J. P. Lee and P. W. Sears, Thermodynamics, Addison-Wesley
Publishing Co., Inc., Reading, Massachusetts, 1963, pp. 284 - 287,
4. Private communication from Dr. W. F. Herget, EPA, NERC.
5. H. Tennekes and J. L. Lumley, A First Course in Turbulence,
MIT Press, Cambridge, Massachusetts, 1972.
6. The authors would like to acknowledge J. A. L. Thomson's
contribution to this section.
7. W. D. Conner, "Measurement of the Opacity and Mass Concentration
of Particulate Emissions by Transmissometry", EPA Report
No. EPA-650/2-74-128, November 1974.
7-1
-------
TECHNICAL REPORT DATA
(Pltasc read Instructions on the reverie before completing)
I. REPORT NO.
EPA-650/2-75-062
3. RECIPIENT'S ACCESSION NO.
I. TITLE AND SUBTITLE
REMOTE MEASUREMENT OF POWER PLANT
SMOKE STACK EFFLUENT VELOCITY
&. REPORT DATE
August 1975
. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
C.R. MILLER and C.M. SONNENSCHEIN
B. PERFORMING ORGANIZATION REPORT NO.
». PERFORMING ORGANIZATION NAME AND ADDRESS
RAYTHEON COMPANY
Equipment Division
Electro-optics Department
Sudbury, Massachusetts 01776
10. PROGRAM ELEMENT NO.
1A AGIO
11. CONTRACT/GRANT NO.
68-02-1752
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
FINAL REPORT
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
Prepared in cooperation with: National Environmental Research Center
Research Triangle Park, N.C. 27711
This report describes the successful demonstration of the ability of a CO, Laser
Doppler Velocimeter (LDV) to measure remotely the velocity of the effluent from a power
plant smoke stack. The basis of the technique is that laser radiation backscattered from
particulates in the effluent is Doppler shifted in frequency in proportion to the velocity of the
effluent. Measurements were made against a coal burning power plant equipped with electro-
static precipitators to remove particulates from the boiler flue gases. Based on the results
of the measurements a study on the design of an LDV optimized for the measurement of
power plant effluent velocities was performed.
17.
KEY WORDS AND DOCUVtNT ANALYSIS
DESCRIPTORS
IB. DISTRIBUTION STATEMENT This document is avail-
able to the public for sale through the Nationa
Technical Information Service, Springfield,
Virginia 22161
EMBED TERMS
c. COSATI Field/Group
19. SECURITY Ci-ASS (T*iiReport)
UNCLASSIFIED
21. NO. OF PAGES
123
20. SECURITY CLASS ,Tftltpage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (S-73)
------- |