A REPORT TO
DEPARTMENT OF HEALTH, EDUCATION AND WELFARE
OF FIELD EVALUATION OF
BARRINGER CORRELATION SPECTROMETERS
PREPARED BY
BARRINGER RESEARCH LIMITED
304 CARLINGVIEW DRIVE METROPOLITAN TORONTO REXDALE, ONTARIO, CANADA
-------
A REPORT TO
DEPARTMENT OF HEALTH, EDUCATION AND WELFARE
OF OPTICAL MEASUREMENTS
OF S02 AND N02 AIR POLLUTION USING
BARRINGER CORRELATION SPECTROMETERS
PREPARED FOR
UNITED STATES DEPARTMENT OF HEALTH,
EDUCATION AND WELFARE
Under Contract No. PH-22-68-44
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
. . Cincinnati, Ohio
PROJECT OFFICER: John S. Nader, Division of Chemistry and Physics
PREPARED BY
BARRINGER RESEARCH LIMITED
304 Carlingview Drive
Rexdale, Ontario, Canada
TR69-113 DECEMBER 1969
-------
TABLE OF CONTENTS
SECTION
DESCRIPTION
PAGE NO.
1
2
2.1
2.2
2.3
SUMMARY
INTRODUCTION
GENERAL
INSTRUMENT TECHNIQUES
CONTRACTUAL REQUIREMENTS
4
4
5
6
3.1
3.2
3.2.1
3.2.2
3.2.3
3.3
3.4
3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.5
3.5.1
3.5.2
3.5.3
3.5.4
3.5.5
3.5 6
3.5.7
3.5.8
3.5.9
3.5.10
3.5.11
3.5.12
3.5.13
3.5.14
3.6
3.7
PHASE I
OBJECTIVES
INSTRUMENTATION
IN-STACK MONITOR
THE REMOTE SENSOR
WET CHEMICAL REFERENCE
EXPERIMENTAL APPROACH
PHASE I (PART 1)
GULF OIL REFINERY
LAKEVIEW GENERATING STATION
INVESTIGATIONS
GASEOUS REACTIONS WITH A STACK
CONCLUSIONS
PHASE I (PART 2)
INSTALLATION
FLOW MEASUREMENTS
PROCEDURES
RESULTS - OCT. 1/69
RESULTS - OCT. 2/69
RESULTS - OCT. 3/69
RESULTS - OCT. 6/69
RESULTS - OCT. 7/69
RESULTS <- OCT. 9/69
LAKEVIEW GENERATING STATION
RESULTS - Oct. 22/69
RESULTS - Nov. 24/69
RESULTS - Nov. 27/69
RESULTS. - Nov. 28/69
CONCLUSIONS
ACKNOWLEDGEMENTS
FIGURES AND TABLES
8
8
8
8
10
11
12
14
14
15
15
19
20
21
21
22
24
25
30
30
31
31
31
32
33
33
34
35
36
36
37
4.1
PHASE II
OBJECTIVES
55
55
- 1 -
-------
TABLE OF CONTENTS
SECTION DESCRIPTION PAGE NO.
4.2 THEORETICAL BACKGROUND 55
4.3 INSTRUMENTATION 56
4.4 EXPERIMENTAL SETUP 57
4.5 RESULTS 59
4.5.1 FLIGHT I - LAKEVIEW GENERATING
STATION 59
4.5.2 FLIGHT II- LAKEVIEW GENERATING
STATION 63
4.5.3 FLIGHT III - GULF REFINERY 65
4.6 COMPARISON OF AIRBORNE AND IN-STACK
NO MEASUREMENTS 68
4.7 THEORETICAL ASPECTS OF PLUME DISPERSION 70
FIGURES 73
PHASE III 81
5.1 OBJECTIVES 81
5.2 TEST SITE 81
5.3 INSTRUMENTATION 82
5.4 EXPERIMENTAL PROCEDURE 85
5.5 RESULTS 85
5.5.1 POINT SAMPLE AMBIENT MONITOR VS. 85
LONG PATH AMBIENT MONITOR
5.5.2 POINT SAMPLE AMBIENT MONITOR VS.
WET CHEMICAL 90
5.5.3 BECKMAN COULOMETRIC VS. 90
CHEMICAL METHOD 90
5.5.4 LONG PATH AMBIENT MONITOR VS. WET 91
CHEMICAL METHODS
5.5.5 POINT SAMPLE VS. COULOMETRIC SAMPLER 92
5.5.6 SLIDING CORRELATIONS 92
5.6 DISCUSSION 93
5.7 CONCLUSION 93
FIGURES 94
PHASE IV 104
6.1 INTRODUCTION 104
6.2 PHASE IV (A) OBJECTIVE 104
6.3 LOCATION 105
6.4 INSTRUMENTATION 106
6.5 EXPERIMENTAL PROCEDURE 106
- 11 -
-------
TABLE OF CONTENTS
SECTION DESCRIPTION PAGE NO.
6.6 RESULTS 108
6.6.1 LAPCD MONITOR 108
6.6.2 POINT SAMPLE AMBIENT MONITOR 109
6.6.4 AIRBORNE INSTRUMENT 110
6.6.4.1 FLIGHT 10 110
6.6.4.2 FLIGHT 11 111
6.6.4.3 FLIGHT 12 112
6.6.5 POINT SAMPLER VS. LONG PATH SAMPLER 112
FIGURES 114
PHASE IV B 120
7.1 OBJECTIVES 120
7.2 LOCATION 120
7.3 INSTRUMENTS 121
7.4 EXPERIMENTAL PROCEDURE 121
7.5 RESULTS 122
7.5.1 AIRBORNE RESULTS 123
7.6 CONCLUSION 124
FIGURES - 127
PHASE IV C 131
8.1 OBJECTIVES 131
8.2 INSTRUMENTATION 131
8.3 EXPERIMENTAL PROCEDURES 131
8.4 RESULTS 132
8.5 COMPARISON WITH GROUND STATIONS 134
8.6 CONCLUSIONS 135
FIGURES 136
9 CONCLUSIONS 140
9.1 SUMMARY 140
10 ACKNOWLEDGEMENTS 143
APPENDIX A
APPENDIX B
REFERENCES
- ill -
-------
Section 1
SUMMARY
This final report for the US Department of Health, Education and Welfare outlines
work carried out under Contract No. PH-22-68 on the application and evaluation
of a new measurement technique based on correlation spectrometry and developed
by Barringer Research Limited. The Barringer technique of correlation spectro-
metry is a molecular absorption technique in which a portion of the desired
spectrum containing rotation-vibration band structures is compared against a
stored replica of the sought spectral signature contained within the spectrometer,
thereby generating a real time readout of the quantity of target gas within the
field of view of the instrument. The tests reported herein were designed to obtain
preliminary data on certain air pollution problems which had been hitherto un-
obtainable by existing conventional measuring techniques.
The spectrometer can be operated in several modes, both active and passive, and
hence provide for the first time means to examine basic air pollution problems
for which data could not heretofore be obtained by existing conventional methods.
The field work provided a more definative grasp on the basic air pollution problems
of making representative SO, and NO measurements. Tests were made to compare
< 4£
data derived from the operation of the spectrometers in their various modes with
data derived from existing commercial instruments and wet chemical referee analyses.
Tests on boiler stacks have shown that the in-stack monitor can be used success-
fully for the continuous monitoring of the sulphur dioxide content of flue gases
without any form of preconditioning. The remote sensor tests showed it was possible
to measure the gas content of flue gas at the top of the stack during daylight
when under conditions of good visibility. In the presence of aerosol and restricted
visibility, the scattering effect of aerosol leads to low values. However, the
correct values can be obtained by taking readings at various ranges and extrapolating
the data back to zero range.
- 1 -
-------
The airborne correlation spectrometer was used to study the oxidation of nitric
oxide emissions to nitrogen dioxide downwind of the stack. Dispersion of the
plume under the action of the gustiness of the wind leads to dilution of the
flue gas and oxidation is substantially complete within a period of about 1 to
2 hours. The rate constant for the oxidation of nitric oxide to nitrogen dioxide
by molecular oxygen is too small to account for the rapid buildup of NO recorded
by the airborne monitor. This suggests that some stronger oxidizing agent is
present in the Ontario atmosphere to effect the rapid oxidation indicated by our
data.
Comparisons of the use of the point sample ambient monitor and the long path
ambient monitor indicate that while the two instruments remain substantially in
phase to better than 90% (and especially when short optical path lengths were
used on the long path ambient monitor), with pathlengths of the order of 1 mile,
considerable concentration gradients can exist within the compass of the long
path instrument. Then the average concentration can vary substantially from the
point sample ambient monitor readings and occurs when isolated gas pockets or
plumes are within its pathlength. The results indicate that the long path monitor
will give much more meaningful averages and avoid extreme fluctuations associated
with point samplers which are more strongly influenced by nearby sources and
changeable micrometeorology.
Field tests of nitrogen dioxide concentration carried out in southern California
indicate that the concentrations of the nitrogen dioxide vary considerably from
point to point. On occasions, this variation can approach a factor of 5 in the
vicinity of freeways. Thus, the advantage of the long path instrument for measuring
average values was demonstrated.
Field tests in the vicinity of a southern California freeway showed that nitrogen
dioxide concentration changes of the order of several hundred ppb can take place
within a few minutes on the freeway. Tests to the side of the freeway indicate
that concentrations a few hundred yards from the freeway are lower by a factor of
three to five.
- 2 -
-------
Field tests in southern California conducted with the airborne monitor indicated
that while useful qualitative values may be obtained, the operation of the air-
craft was severely limited by low visibility which was often below 2 miles and
prevented the aircraft from taking off. Aerosol scatter and optical dilution
by the California smog enahanced the amount of backscattered radiation returned.
to the spectrometer from the upper layers within the inversion. Hence, this
radiation contained little or no signal modulation resulting from NO close to
ground level. Consequently the spectrometer was saturated with light which
represented backscatter from approximately a few thousand feet of atmosphere
immediately below the aircraft. This atmospheric backscatter is a function of
many complex inter-relating variables, such as solar zenith angle, spectral
wavelengths used, particulate (mie) and molecular scattering, field-of-view of
the spectrometer etc. Notwithstanding these limitations peculiar to this work
in California, we could delineate gas clouds and follow their movement along
wind trajectories and also obtain integrated gas profiles which do have meaning-
ful significance.
- 3 -
-------
Section 2
INTRODUCTION
2.1 GENERAL
This final report has been prepared by Barringer Research Limited for the United
States Department of Health Education and Welfare under contract No. PH-22-68.
In essence the contract called for a series of field studies involving various
configurations of a new measurement technique, a commercial analyzer and a wet
chemical referee method. The ultimate aim of the contract was to use the new
measurement technique to obtain air pollution measurements for S02 and N0_ under
dynamic conditions heretofore unobtainable with conventional instrumentation.
The Barringer technique of correlation spectrometry is a new powerful method for
qualitative and quantitative measurements upon a gas, in which an absorption
spectrum is cross-correlated in real time against the spectral replica of the
species sought. Barringer1s development of several instruments utilizing this
principle of analysis in various modes of sampling provided the opportunity to
study specific questions on the comparisons of in-stack to remote-stack sampling
techniques and for long-path to point sampling techniques in air pollution measure-
ments. This report deals only with measurement of SO? and NO , two gases of im-
mediate interest in air pollution studies. The field studies involved the
measurement of S09 and NO (as total oxides of nitrogen) both in boiler flue
£ £
gases and in urban atmospheres. In the stack emission studies comparisons were
made between simultaneous in-stack measurements in the breeching of a boiler,
remote measurements at the chimney exit and wet chemical analyses of the flue gas.
Airborne measurements downwind of the stack, were made to follow the conversion
of NO to N00. In the urban atmospheric NO, and SO studies, comparisons were made
2, £2.
between a point sampler, a long path sampler using path lengths of up to 1400
meters, commercially available conventional analyzers and wet chemical referee analyses.
- 4 -
-------
2.2 INSTRUMENT TECHNIQUES
The basic concept of correlation spectrometer has been incorporated in the follow-
ing types of instruments developed by Barringer Research Limited, namely:
(a) In-stack monitor.
(b) Remote Stack monitor.
(c) Point sample ambient monitor (PSAM).
(d) Long Path ambient monitor (LPAM).
(e) Remote sensor for airborne operation.
The instrumental technique of correlating the incoming spectra against a stored
replica is similar in each instrument listed above. The essential differences
are in the method of achieving this spectral correlation and the mode of
obtaining the operation measurements.
In the in-stack monitor the optical path of the spectrometer is contained
in a short probe which is introduced directly into the breeching or stack. This
provides a continuous real time measure of the flue gas constituents passing
by the probe and eliminates the need for cooling, drying or preconditioning the
flue gas. A light beam is projected along the probe and reflected off a mirror
within the probe to return to the spectrometer housing located outside
the stack. The slot in the probe through which the flue gas passes is 1/4 meter
in length and the effective optical path length (2 pass) is 1/2 meter.
The remote stack monitor is a passive telescopic spectrometer using sky light as
an energy source. It is used for measuring the gas content of plumes and clouds
at the exit of a chimney stack by scanning the field of view and measuring the
signal increase recorded through the plume. The plume measurement is compared
with the no-gas reading obtained from the sky background at either side of the
plume.
- 5 -
-------
The point sample ambient monitor physically draws into a two pass 1.25 meters
sample cell the ambient air under measurement. It has an active light source
which projects a light beam through the sample cell. The threshold sensitivity
of the PSAM is close to 10 ppb.
The long path ambient monitor uses a correlation spectrometer boresighted onto
an active light source positioned at any distance up to a maximum of 1400 meters
from the spectrometer. This forms the long optical light path and thereby allows
the spectrometer system to measure very low average concentrations over this path.
For a pathlength of 1400 meters the instrument threhold sensitivity is 1 ppb.
The airborne monitor is a high sensitivity passive spectrometer mounted in an
aircraft with facilities for upward and downward looking measurements. If the
aircraft is flown above the inversion layer, one can reasonably assume that no gas
is present above the aircraft. Consequently an upward looking reading was used
for a zero reference. However, more exact zeroing can be achieved using other
methods that were not available at the time of this contract. A measure of the
total gas burden below the aircraft can thus be made by comparison of the upward
(zero gas) and downward looking readings. The instrument can also be used for
the remote measurement of plumes and gas clouds from changes in the vertical burden
recorded as the aircraft flies over the plume or gas cloud. It should be noted
here that since these airborne measurements were made, significant advances have been
perfected in instrumentation and technique which negates the need for a sky . zero
and which is much less sisceptible to spectral distribution errors).
2.3 CONTRACTUAL REQUIREMENTS
The contract called for the execution of six experiments as follows:
(1) PHASE I - Stack SO Emissions
A comparison of instack sampling and remote (stack exit) sampling using a
standard wet chemical method as reference in measuring S02 in the flue
gases from coal and oil fired furnaces. This was to be performed under
a variety of furnace operating conditions and for different environmental
background illuminations.
- 6 -
-------
(2) PHASE II - NO Plume Chase
An experiment involving the airborne surveillance of the exhaust plume
from a large power station. The measurements were to be made downwind
to determine the rate of formation of nitrogen dioxide from the nitric
oxide emitted by the stack.
(3) PHASE III - Ambient SO
An intercomparison between point sampling and long path sampling of S02
in an urban atmosphere. The comparatively new technique of correlation
spectroscopy enabled long path integrated measurements to be made for the
first time. The object was to obtain a basic understanding of the
relationship between a long integrated path of Ambient S0_ and measurements
of SO^ at a point in the ambient air.
(4) PHASE IV - Ambient NO
A group of three experiments carried-out in southern California to obtain
data on the concentration of N0».
Part A - similar to Phase III but for N0_ and with vertical distributions
determined by airborne measurements at various selected altitudes.
Part B - intercomparisons of N02 levels over a freeway and off freeway
using integrated long path
- vertical distribution from airborne instrument
- point sampling at various points along freeway
- wet chemical measurements along freeway
Part C - repeated airborne flights over a wind trajectory at a constant
altitude using the airborne method synoptic profiling to show
the build-up and degradation of NO concentrations with time
and location.
- 7 -
-------
Section 3
PHASE I
3.1 OBJECTIVES
The objective in Phase I is to compare the results from the in-stack monitor
with data from the remote sensor for typical industrial stacks. Among the
variables to be included in this series of tests are the use of coal and oil-
fired furnaces burning high and low sulphur fuels and producing flue gases of
variable aerosol content. Measurements are to be carried out also under a
variety of sky conditions to include varying degrees of cloud cover from clear
to total overcast. The Shell wet chemical method is to be used as a reference.
A series of simultaneous measurements conducted by DHEW personnel using the
Dalmo Victor IR remote sensor were also planned.
Serious difficulties were encountered with the in-stack instrument and in
interpreting the readings of the remote sensor early in the program and work
on Phase I was discontinued pending results of investigations. Phase II, III
and IV were then accomplished and Phase I was then repeated. This section includes
some data from the original Phase I and all data from the repeat of Phase I.
3.2 INSTRUMENTATION
The following is a brief description of the in-stack monitor and remote sensor.
For a more complete discussion of design and operational features refer to
references 1 and 2.
3.2.1 In-Stack Monitor
The correlation stack monitor is a relatively low sensitivity instrument built to
withstand dust, heat, corrosion and vibration. The maintenance of optical align-
ment in the presence of intense thermal gradients proved to be a major problem
in its development. Designs have now been evolved which obviate the necessity
8 *
-------
for a temperature controlled compartment for the optical components as was
originally found necessary. The instrument is now capable of being bolted
directly to a high temperature duct with negligible affect on instrument per-
formance. (Figure 3.1).
The probe forms a part of an integral package with the spectrometer, light
source, and control electronics. (Figure 3.2). The unit is designed for
single-hole mounting on the wall of the breeching with the probe inserted
into the gas flow. The probe which is about 15 inches long carries a mirror
at the far (in-stack)end. This mirror is protected from particulate fouling
by an air curtain provided by an external blower. The one-half meter (2 pass)
optical absorption path length is defined by the air curtain which protect the
optical train from the hostile environment.
The use of a one-quarter meter of sample length (i.e. 1/4 meter slot) is based
on the assumption that the gas mixture across the stack is homogenous with respect
to the gaseous constituents. This had induced some scepticism, presumably due
to the necessity for accurate isokinetic sampling across the whole area of the duct
when measuring particulate matter. A series of wet chemical traverses carried
out by York Research Corporation at Stamford, Connecticut hower, has confirmed
our belief that there is no significant stratification or uneven distribution of
gaseous components across the flue diameter.
The in-stack instrument was tested for stability and temperature sensitivity in
an environmental test chamber within the range from -10 to +100 C. The
environmental limitations of the current probe design are temperature, pressure
and sensitivity. The probe is presently rated for use at up to 800 F, and will
probably be suitable for operation up to 1000 F. The pressure limitation arises
from the fact that the blower for the air curtain works against the pressure inside
the flue. The upper pressure limits of 10 inches W.G. is set by the size of the
blower but where paint air is available this limit could be relaxed.
- 9 -
-------
The solid state electronic unit follows standard coherent detection practice
and the signal from the spectrometer is fed to an A.G.C. circuit to compensate
for changes due to lamp intensity and dust loading. After band-limiting
amplifiers, the signal is fed to a synchronous detector where it is gated with
a signal from the reference photomultiplier. A low pass filter drives a standard
100 mv potential recorder which can be calibrated in parts per million.
The in-stack monitor used for the initial Phase I work used a quartz prism for
dispersion of the light beam and vibration was effected by means of a torque motor
driving a diffractor plate situated just behind the entrance slit. Calibration of
the instrument was carried out by means of an optical cell which could be inserted
through the slots in the probe (referred to herein as a "slot cell") into which
known amounts of SO could be injected. For routine checking of the instrument
sensitivity in the field, a one centimeter reference cell of known gas content
can be inserted into the optical path. The Phase I repeat program used an improved
version of the stack monitor in which the quartz prism was replaced with a diff-
raction grating, correlation was then achieved by vibrating the grating upon its
torsion bar mount by means of a torque motor.
3.2.2 The Remote Sensor
The remote sensor (Figure 3.3) operates in a passive mode, utilizing the back-
ground sky as the light source. Spectral correlation is achieved by vibrating
refractor plates mounted on the tine of a tuning fork as shown in Figure 3.4.
This instrument when used as a remote sensor enables the quantity of gas within
the field of view to be determined. Zeroing is performed by looking to either
side of the target plume and adjusting the instrument output to a convenient re-
ference value. For routine sensitivity checks and calibration in the field, the
instrument is equipped with two 1 cm fused-silica reference cells, (as shown in
Figure 3.4) which can be easily inserted into the optical path either singly or
together while sighting on the sky background to one side of the plume. The
instrument is then sighted on the plume and the signal generated as a result of
- 10 -
-------
SO in the plume is recorded.
Because of the generally rapid mixing of flue gas with ambient air following
emission at the stack exit, the field of view of the instrument is usually
closed down so as to observe only a small section of the central portion of
the plume as near the stack exit as possible. The horizontal path length
through the plume (assuming a vertical plume) is then assumed equal to the
internal diameter of the stack.
3.2.3 Wet Chemical Reference
The primary wet chemical method used in Phase I..was the Shell Thornton procedure
intended specifically for the determination of SO and SO in stack gases. A
detailed description of the method is given in Reference 3. The essential feature
of this technique is the analysis of flue gases for sulphur oxides by oxidizing
the dioxide to the sulphuric acid which is determined by titration. The deter-
mination of the proportion of the two oxides poses some problems because of the
ease with which the dioxide is converted to sulphuric acid, especially in the
presence of the calatylic metal oxides which are present in fly ash. The Shell-
Thornton method uses a short heated silica probe to bring out the flue gases
while still above the dew point, into an isopropyl alcohol bubblers which removes
the sulphuric acid aerosol and effectively prevents oxidation of the sulphur
dioxide. A second bubbler containing hydrogen peroxide traps the sulphur dioxide
by oxidizing it to sulphuric acid. After a suitable sample has been collected -
a procedure requiring about 20 minutes - a flow of purified air is drawn through
both bubblers to transfer any sulphur dioxide trapped in the first bubbler into
the second bubbler. The acid contents of both bubblers are then determined by
titration.
During the earlier part of the Phase I programme the She11-Thornton method
(abbreviated "Shell" hereafter) was supplemented with the simpler Reich method.
This latter technique, based on the decoloration of iodine, can be accomplished
- 11 -
-------
in about three minutes as compared with 20 minutes for the Shell. This is an ad-
vantage in investigating short term fluctuations in gas concentrations. It is im-
portant to note that while the Reich test was used in conjunction with the Shell
test it was not used to replace the Shell procedure. Also the experience gained
on this programme over a large number of simultaneous measurements showed that the
average of a series of Reich tests agreed closely with the Shell results with dis-
crepancies usually less than 5%.
3.3 EXPERIMENTAL APPROACH
A number of different boiler installations were chosen to satisfy the contract
requirements for both oil and coal fired sources. These installations were:
(1) Gulf Oil Refinery - Clarkson Ontario, Boiler No. 5 and CO Boiler
(2) Ontario Hydro Lakeview Generating Station, Toronto, 300 Megawatt units
No's 3, 4 and 7, 8.
(3) Lever Brothers - Toronto, process steam at 100,000 Ib/hr, oil fired.
In all cases the experimental techniques developed early in the Phase I programme
were used throughout (with the exception of the Gulf CO Boiler) and involved the
installation of the in-stack correlation spectrometer in the breeching near the
base of the stack. The wet chemical sampling probe was installed in close prox-
imity to the spectrometer probe. The remote sensor was tripod-mounted and trans-
ported by a mobile vehicle.
In general the experimental procedure followed was to continuously monitor the
S0_ concentrations in the breeching with the spectrometer probe while taking re-
peated samples of flue gas with the wet chemical system. Concurrently, measure-
ments of the flue gas were made at the stack exit using the mobile remote sensor.
To obtain a measure of the effect of atmospheric scattering on the remote sensor
measurement, a series of remote sensor readings were taken at different distances
from the stack. When feasible each series of readings was taken with the angle
- 12 -
-------
subtended at the stack by the sensor and the sun held roughly constant. It was
not feasible to vary the fuel sulphur content or alter combustion conditions at
these production facilities in a programmed fashion. Consequently the duration
of the experimental period was extended to obtain as wide a range of S02 concen-
trations and dust loadings as was practicable as a result of normal fluctuations
in the fuel sulphur content, fuel burning rate/ excess air and from periodic
purging of soot accumulations. On occasions large changes in SO level were ob-
tained. At the Lever Brothers oil-fired test site this was achieved by releasing
SO directly into the boiler inspection door from a portable high pressure con-
tainer. At the coal fired test site (Lakeview) this was achieved by monitoring
flue gas while the turbine generator system was undergoing a series of special
tests. The test period for both oil and coal buring sources were of sufficient
length to obtain remote sensor plume results over a wide range of sky, wind and
visibility conditions.
Whenever possible, the timing of the wet chemical samples, spectrometer probe
calibrations and remote sensor set-ups were coordinated so as to permit a mean-
ingful comparisons of the measurements. The following discussion of the work
performed is presented in chronological sequence and describes the identification
and investigation of problem areas which revealed themselves during the early
stages of the Phase I program and treats in considerably greater detail the work
performed and results achieved during the Phase I continuation period. Since the
earlier stage of Phase I was basically a trouble shooting and investigation period
and the later part of Phase I was entirely a data gathering and data analysis
period, these two periods are presented in two parts as Phase I (Part 1) and Phase
I (Part 2).
- 13 -
-------
3.4 PHASE I (PARTH)
3.4.1 Gulf Oil Refinery, July 1968
Our first field work was undertaken at No. 5 Boiler of the Gulf Oil Refinery,
Clarkson, Ontari, which exhausts through a single stack. The in-stack monitor
was initially aligned and calibrated in the BRL laboratories and then mounted on
the breeching of this boiler. The remote sensor was set up to view the stack's
exit. This boiler is equipped for both oil and gas firing, the preferred fuel
being natural gas which is used exclusively during the summer. The oil fuel is
residual oils with 3% sulphur or slop oils of unknown and highly variable comp-
osition. During this initial testing program some of the burners were using oil
and the remainder natural gas.
To create variations in the SO content of the flue gas, the number of oil and gas
burners in operation at any given time were varied. To meet the contract require-
ments of varying plume aerosol content, it was planned to generate #2 Ringleman
smoke by adjusting the fuel/air ratio and by manipulation of the dampers. More-
over attempts to regulate the plume by adjusting the fuel/air ratio were unsuc*
cessful because of automatic controller action and only short periods of very
dense smoke resulted. Since the exit velocity of the stack was less than 10 ft/sec,
even very light winds resulted in a stretched and turbulent horizontal plume.
Consequently our field work was considerably complicated by the above effects.
After review of the initial data obtained at this site marked disagreements between
the Shell results and the in-stack monitor were apparent. The degree of disagree-
ment was shown to be a direct function of the flue gas temperature.
Errors in the remote sensor measurements were also apparent. These were found to
stem from two causes; the extremely low stack exit velocity and atmospheric scat-
tering. In general, the remote sensor results were lower than the Shell values
and showed particularly low values when conditions of low visibility or dense smoke
- 14 -
-------
occured.
3.4.2 Lakeview Generating Station
A similar series ef exploratory tests were conducted at the No. 3 generating unit
of the Ontario Hydro Lakeview Generating station whose location is shown in Bigure
3.5. This station has four stacks which are five hundred feet high and each stack
serves two 300 Mwatt generating units. The in-stack unit and Shell probes used
existing sampling ports and staging. The Lakeview plant employs a pulverized coal
fueled system and 2.2% sulphur coal which generates about 1400 ppm SO . The grade
of coal is substantially constant and stock-piling is carried out in such a manner
as to ensure mixing of various batches. It was therefore not possible to alter
the fuel sulphur content neither could the boiler operating parameters be changed
to make smoke.
The results of these tests confirmed the temperature dependency of the in-stack
monitor and indicated that the problem was not associated with the use of oil
fuel. Because of the relative constancy of the Lakeview plume characteristics,
firm identification of the atmospheric scattering problem associated with the re-
mote sensor was possible.
3.4.3 Investigations
In order to investigate the cause of the temperature dependency problem of the
in-stack it was transferred to a. new test site at Humberview, Toronto. The Humber-
view test facility is a laboratory established by the Department of Public Works
of Metropolitan Toronto to study dust emissions from various types of incinerators.
In summary the most important results of these investigations was the replacement
of the quartz prism dispersing element with a grating element and the calibration
of the instrument to take into account the change in sensitivity of the in-stack
monitor due to temperature-induced line broadening of the SO absorption spectrum.
The line broadening effect, termed the "modulation factor", was measured over the
- 15 -
-------
operating temperature range of the instrument and is reproduced in Figure 3.6
The use of the modulation factor appropriate to the flue gas temperature of a
given installation is required to obtain accurate quantitative results. One
can see from Figure 3.6, that when the modulation factor ranges from 1 to 3,
the latter occuring when the gas temperature approaches 800 F. Correction also
has to be made for the usual gas low density changes associated with the flue
gas temperature changes. As a percentage error this is simply the ratio of
in-stack temperature to the outside ambient temperature. For example the
Lakeview Station stacks were at 300 F, giving a temperature correction of
about 15%.
Concurrent with the in-stack spectrometer investigations at Humberview, a theore-
tical study of the atmospheric scattering and range dependency problem of the
remote sensor was conducted. The major part of these investigations and the
associated field test results are covered in detail in reference 4, and will be
reviewed only briefly, in summary the study showed that significant light
scattering can occur between the plume and the remote sensor which will cause a
reduction in the sensor's sensitivity was shown to be a function of the atmospheric
turbidity and the distance between the stack and the sensor. Table 3.1 presents
test results obtained from both Gulf and Lakeview stacks during August 1969
while Table 3.2 provides their analyses. The percentage errors quoted in
Table 3.2 are calculated using the Shell measurement as the reference and the
remote sensor measurement is the average integrated value over the same time
period that the Shell measurement was made. These data show the effect of slant
range and visibility on the accuracy of the remote sensor measurement. It is
obvious from these tables that minimum scattering error occurs at minimum range
during times of good visibility. The last two columns of Table 3.1 are shown
plotted in Figure 3.7. This curve shows that in spite of errors associated
with estimating visibility and the fact that the data was obtained at a meteor-
ological station several miles from the test sites, there is a close relation-
ship between the ratio of range/visibility and the error in the remote sensor
measurement.
- 16 -
-------
TABLE 3.1
REMOTE SENSOR RESULTS AT GULF AND
LAKEVIEW STACKS, AUG./SEPT. 1968
Date
1968
Aug. 12
13
14
15
i
15
-vl
1 28
29
Sept. 12
12
18
18
19
19
20
20
Stack
Gulf
Gulf
Gulf
Gulf
Gulf
IM
LV
LV
LV
LV
Gulf
LV
Gulf
LV
LV
Slant
Range
(ft.)
580
580
580
700
700
1250
1250
1250
1250
1600
700
1600
350
1830
1450
Visibility
(miles)
15+
8
15+
15+
15+
15+
12
15+
15+
3
3
8
8
8
6
Weather
Clear
Hazy
O'Cast
Clear
Clear
Clear
Clear
Clear
Clear
V Hazy
V Hazy
o'cast
o'cast
Hazy
Hazy
Remote
Sensor
(ppm)
280
460
220
350
420
400
555
725
650
185
175
330
240
325
245
Shell
(ppm)
250
935
250
450
655
1385
1330
1450
1390
1360
570
1300
500
1105
1115
Ratio
Shell
Remote
0.9
2.03
1.13
1.30
1.55
3.46
2.40
2.00
2.14
7.35
3.25
3.94
2.09
3.40
4.57
Ratio
Visibility
Slant Range
26+
13.7
26+
22+
22+
12+
9.6
12+
12+
1.8
4.3
5.0
14.5
4.3
4.1
-------
(1) Very Short Range
TABLE 3.2
Range
(ft.)
Weather
(Condition, visibility)
in miles
Error
emote-Shell
Shell
350
overcast 8
-52
(2) Short Range
580
580
580
clear . 15+
overcast 15+
Hazy 8
+12
-12
-51
(3) Medium Range
700
700
700
clear 15+
clear 15+
v. hazy 3
-22
-36
-69
(4) Long Range
1250
1250
1250
1250
clear
clear
clear
clear
15+
15+
15+
12
-76
-50
-58
-58
(5) Very Long Range
1450
1830
1600
1600
hazy 6
hazy 8
hazy 8
v. hazy 3
-62
-70
-75
-86
- 18 -
-------
The theoretical treatment of the scattering problem described in Reference 5,
showed that extrapolation of a curve defined by a series of remote sensor measure-
ments taken at different ranges could be used to largely compensate for the
scattering errors associated with measurements obtained at any one range. Figures
3.8 and 3.9 show the results of several experiments at the Gulf Oil CO Boiler
in which extrapolation to zero range was used to arrive at stack exit concentrations.
The Gulf CO boiler is a gas-fired unit burning sulfurous gases and was used for
this particular experiment because of its high exit velocity and well defined
plume in moderate winds. Figure 3.10 shows similar results for the Lakeview
stack.
3.4.4 Gaseous Reactions Within A Stack '
As a result of the continued low readings obtained with the remote sensor, it
was suggested that significant amounts of sulphur dioxide could enter into
chemical reactions, e.g. calatytic oxidation to sulphur trioxide, so that the gases
issuing from the stack contain much less sulphur dioxide than is measured at the
breeching. This could be a mjaor source of error in the 500 ft. stacks at
Lakeview.
By arrangement with the Ontario Hydro, who had a steeple jack working on the
top of one of the chimneys, simultaneous chemical measurements were made at the
top and bottom of the stack. The boilers feeding the stack were kept on a
steady load for the duration of the test. The sulphur dioxide level in the flue
gas was established by means of Shell determinations before and after the tests
at the top of the stack. During the tests at the top of the stack, the constancy,
of the SO input at the bottom was monitored by the In-stack Monitor, which
showed a slow increase in S0? during the progress of the tests. The She11-Thornton
results were:
Sample 1 Bottom at 0906-0922 hr. 1105 ppm
2 Top at 1225-1245 hr. 1085 ppm
3 Top at 1440-1500 hr. 1115 ppm
4 Bottom at 1630-1650 hr. 1355 ppm
- 19 -
-------
These results confirm that there is no significant reaction resulting in the
removal of sulphur dioxide during the passage up the stack.
3.4.5 Conclusions
To sum up the results of Phase I (part 1) it may be stated that:
(1) Difficulties were encountered in the measurement of in-stack flue gases
because of two basic reasons;
(a) broadening of S02 absorption bands and consequent loss of instrument
sensitivity was not compensated for, and
(b) Instability of the in-stack monitor optical system resulting from
thermal distortion of the quartz prism produced substantial errors.
(2) Difficulties were encountered in the measurement of stack gases as the
exit because,
(a) loss in sensitivity due to atmospheric scattering was not compensated
for, and,
(b) the very low exit velocity resulted in large negative errors with
considerable scatter in the data.
(3) The in-stack monitor was modified to improve its temperature stability and
the temperature modulation factor was measured and an appropriate calibration
curve drawn up.
(4) The physics of atmospheric scattering and its effect on the remote sensor
were examined and techniques developed to permit the effects of scattering
to be compensated for.
The accomplishment of the Phase I (Part I) programme established a firm basis for
proceeding with Phase I (Part 2).
- 20 -
-------
3.5 PHASE I (PART 2)
3.5.1 Installation '
Because of the limitations of the Gulf Refinery facilities in terms of the programno
requirements, arrangements were made to conduct tests at the Lever Brothers plant
in south Toronto.
Figure 3.11 shows a plan of the Lever Brothers plant. The boiler house generates
process steam at the rate of 100,000 Ib/hr for the manufacture of soaps and deter-
gents. It is equipped with three independent oil-fired boilers, two of which are
in use at any given time. Figure 3.12 illustrates the arrangement of ducts and the
probe installation. Boilers No. 1 and 3 were working near full load over the dur-
ation of the test period while Boiler No. 2 was shut down. The spectrometer probe
and wet chemical sampling probe were both installed in the flue of boiler unit No. 1.
Since the steam load is evenly balanced between units 1 and 2 and fuel and combus-
tion conditions are virtually identical, the SO concentration in the flue gas of
unit No. 1 was assumed representative of the SO level in the combined output of
units 1 and 3. It was not feasible to measure the combined effluent in the stack
itself because of the lack of sampling ports and the presence of a baffle in the
stack which prevented mixing of the outputs of units 1 and 3 until reaching a point
in the stack well above the roof level. Although Boiler No. 2 was shut down and its
damper closed there was still a substantial free flow of air reaching the stack
through flue No. 2 under conditions of natural draft. For purposes of comparing
the in-stack results with the remote sensor measurements, the mass flow of both
flue gas and clean air was measured by means of velocity and temperature traverses
in all three flues and the effect of dilution of the stack exit readings allowed
for.
Figure 3.13 shows the completed installation including the Shell sampling unit.
The fuel used was Bunker "C". The analysis is as follows:-
specific gravity @ 60°F 0.9767
water (by volume) 0.05%
- 21 -
-------
sediment
ash
sulphur
BTU/Emp. Gal. @ 60°F
0.06%
0.07%
1.82%
179,400
Figure 3.14 shows the truck mounted remote sensor viewing the Lever Brothers stack.
This photo was taken from remote sensor site P5 of Figure 3.11.
3.5.2 Flow Measurements
Velocity head traverses were made across each flue using a pitot tube and inclined
manometer at , each port location shown in Figure 3.12 . The data obtained are as
follows:-
#1 Flue
(ins. W.G.)
A (Fahrenheit units)
Head 1.0
1.2
1.3
1.2
1.3
1.1
1.0
340
350
360
350
350
350
350
800
810
820
810
810
810
810
Area =4' x 4' = 16 sq. ft.
Av. Temp. = 810 A
Av. Lead = 1.2"
Velocity = 80 F.P.S.
- 22 -
-------
Flow = 76,800 C.F.M.
= 49,920 S.C.F.M. (70°F)
#2 Flue
(ins W.G.)
"A (Fahrenheit units)
Head
0.35
0.35
0.4
0.5
0.4
0.4
0.4
80
80
80
80
80
80
80
540
540
540
540
540
540
540
Area = 4' x 4' = 16 sq. ft.
Av. Temp. = 540 A
Av. Head = 0.4"
Velocity = 40 F.P.S.
Flow = 38,400 C.F.M.
= 37,632 S.C.F.M. (70°F)
#3 Flue
The traverse in this case showed the flow in this flue to be the essential same as
that in flue No. 1.
Dilution of the flue gas by clean air is then:-
total flue gas from #1 & #3 = 99,840 SCFM
by pass air from #2 = 37,632 SCFM
Total gas in stack = 137,472 SCFM
- 23 -
-------
99 840
S02 is diluted by clean air in the ratio ' = 0.726
To compare the stack exit measurements with the Shell or in-stack values the
remote sensor results are multiplied by 1.38.
3.5.3 PROCEDURES
As mentioned previously whenever possible all measurements were conducted in a
coordinated fashion to maximize the significance of comparisons. Since the Shell
method was the reference and took the longest to perform, both the in-stack and
remote units were usually calibrated, adjusted, or moved as necessary during the
down-time intervals of the Shell measurement. In general the Shell samples were
taken as frequently as the technique would permit over each day of data gathering.
The in-stack unit was generally left in continuous operation once the day's data
run had begun and was removed from the flue only occassionally to check the zero
level and instrument sensitivity.
To accomodate the variations in wind direction and sometimes to avoid interference
from neighbouring plumes, remote sensor sites were chosen on three sides,of
boiler house as shown in Figure 3.11. In conducting a series of multi-range
measurements the remote sensor used sites which lay roughly in line with the stack
e.g. P7, P2, S5, S2, S3 and S4 and moved from one site to next as quickly as
circumstances would permit.
- 24 -
-------
To provide more continuous data for comparison with the continuous in-stack
monitor data, the remote sensor was stationed at the nearby sites e.q. P. 5
P7 or P.4. In these positions the remote sensor results generally corresponded
closely with the in-stack readings without any need for correction due to
scatter except during periods of rain fall.
After installing the equipment and conducting a series of trial runs on Sept.
29th and 30th, 1969 the first of a series of daily continuous data runs began on
Oct. 1st.
3.5.4 RESULTS - Oct. lst/69
For the Oct. 1st 1969 run, the remote sensor was stationed for the full day at P5.
Figure 3.15 shows a portion of the in-stack chart. The fluctuations of flue
gas (S0_) density is apparently typical of boiler/fan installations of this
type. The reference cell was an external 1/4 meter cell containing the
equivalent of 500 ppm SO at 70°F and one atmosphere. For each calibration
the reference cell was flushed with clean air and injected with 0.2 ccs of
SO at ambient pressure and temperature.
A typical calculation for SO concentration in the flue gas is performed as
follows:
Dl
SO (ppm) = x 500 x Mod. Factor x Temp. Factor
where: D = Target gas deflection = 42 chart div.
D = Ref. Cell deflection =48.5 chart div.
Mod. Factor= 1.45 from Figure 3.6 @ 300°F (149°C)
Temp. Pact.- Fl?? G«a *«* Ambient Temp. = 21°C
c Ambient Temp.
_ 273 + 149 = 294 A
294
= 1.43
- 25 -
-------
42
Therefore SO (ppm) = x 500 x 1.45 x 1.43
2, 4o O
= 898 ppm
The purpose of the modulation factor was discussed previously. The temperature
factor takes into account the Charles Law expansion of the gas. This is
required since the sensor is actually responsive to the density of the SO component
of the flue gas and uses a reference cell which has been calibrated in terms of
concentration (ppm by volume) at 21 C and 760 mm. Hg. The slight negative
pressure in the flue has a negligible effect on the measurement and pressure
correction is not necessary.
Figure 3.16 shows a portion of the remote sensor chart. V.. and V_ are the
responses obtained from the reference cells while observing the sky to one
side of the plume. V is the response from the plume itself. These measurements
were made at station P at 1004 EST.
The SO2 concentration in the plume is found from the following expression:
(_i _ _i, + _1_ i x r 273 +
VT V2 C2L2
where V = NO. 1 ref. cell rdg. (low value) volts or chart div.
V = No. 2 ref. cell rdg. (high value) volts or chart div.
C..L = No. 1 ref. cell calibration constant ppm-meters
C?L = No. 2 ref. cell calibration constant ppm-meters
V = Target gas delfection volts or chart div.
T = stack exit temperature
L = Depth of target along line of sight meters
CT = Concentration of SO in target plume ppm
£*
- 26 -
-------
Note that in the case of the remote sensor it "looks through" the total
diameter of the gas column and therefore must cope with a much greater
ppm-meter product than does the in-stack unit. Since for very large ppm-m
values, the sensor response is inherently non-linear, two reference cells are
required. For full details of the remote sensor method refer to Reference 2
Processing of the remote sensor data begins with transferring the chart data
/ 3
to a convenient data sheet such as shown in Table 3.2. This table contains
all data pertinent to the remote sensor measurement. Meteorological data
were obtained from the records of the Toronto Island Airport and plume
characteristics were noted at the time of measurement. Note that in the case
of a bent-over plume the sensor station was always chosen so that the sensor's
line of sight intercepted the plume axis at approximately 90 . The "look-up"
angle of the sensor i.e. the inclination of the line of sight was measured with
a hand-held Abney inclinometer. The precision of this measurement is about
10 minutes of arc. The aperture refers to the setting of the sensor vertical
field of view as determined by the vertical height of the entrance slit. This
setting is read on a dial which has a nominal scale of 3.7 milliradians/turn.
Therefore a setting of 0.5 turns = a vertical acceptance angle = 1.85 mR.
At a slant range = 126 meters the vertical F.O.V. = 0.23 meters (0.75 ft.).
The width of the entrance slit is fixed and the azimuth acceptance angle is
a constant 3 mR. The F.O.V. at 126 meters is therefore 0.378 meters (1.24 ft.)
wide x 0.23 meters (0.75 ft.) high, i.e. in the case of a cylindrical plume
of diameter 2.6 meters and a range of 126 meters, the sensor, when pointed for
peak response, observes only the central portion of plume with a F.O.V. no
greater than l/7th of the full plume diameter.
The A.G.C. (Automatic Gain Control) voltage is a measure of the ultra-violet
flux reaching the sensor. Referring again to Figure 3.16 note that an AGC
performance check was made using a 10% transmission neutral density filter.
The fact that the presence of the filter in front of the entrance window did
not induce error but only caused an increase in the noise pattern demonstrates
that there was more than an adequate light level available for the measurement.
With a 90% reduction in incident light level the AGC level rose to 6.4 volts.
The limiting value is 7.0 volts.
- 27 -
-------
The optical pathlength through the plume is shown at X in Table 3. 3 and
LT
was obtained from
*»
cos a
where W = stack internal diameter = 2.6 meters
s
a = lookup angle
X is equivalent to L in the remote sensor equation previously discussed.
LiT T
Distance between the sensor and the stack exit is termed slant range R
S
and R
stack height (Ht)
s sin a
For purposes of comparison with the instack measurements, the remote sensor
result is shown as two values, one which is the measured concentration at the
stack exit and the other which is the stack exit value multiplied by the
clean air dilution factor of 1.38. After tabulating the raw chart data in this
form, the tabular data are fed into a computer program which solves the fore-
going equations for F.O.V., XLT, R and C , the target gas concentration.
S .L
The results of the in-stack measurements and stack exit concentrations
compensated for clean air dilution are shown plotted along with the Shell
averages in Figure 3.17. The in-stack and remote sensor points are two-
minute averages. Both instruments have a response time of about 3 seconds.
The Shell results are averaged over a 15-20 minute sampling interval and
are shown as a horizontal bar in Figure 3.17. Note that the remote sensor
data tend to have greater scatter than the in-stack data. The reason for this
is due mainly to plume fluctuations and to a lesser extent possibly by incoherent
fluctuations in burning rates of the two boilers.
- 28 -
-------
The in-stack and remote sensor data of Figure 3.17 were examined for correla-
tion and regression using the IBM Quicktran Common Library Program, with the
following results: (IBM STATPK)
Correlation Coefficient 0.89
Regression 0.97
Intercept 73.66
It is apparent that a high order of agreement exists between the in-stack
unit and the remote sensor particularly in view of the differences in the two
methods of measurement. The first Shell test was some 75-80 ppm lower than the
in-stack result but was in very close agreement on the two subsequent tests.
The regression or slope of 0.97 indicates that the calibration of the two sensors
were almost identical and the intercept of 73.66 indicates a minor discrepancy
in determination of the zero level.
The abrupt drop in signal and rapid recovery at time 1110 on Figure 3.17 was due to
purging of soot accumulations in the boiler tubes, a daily routine of the
boiler operator. This results in a particularly dense cloud of yellow-brown
to black smoke of short duration.
It is significant that the results of Figure 3.17 were achieved with no correction
for atmospheric scattering. This is no doubt due to the short slant range used
and the good visibility. The small loss of signal due to scattering was at
least partially compensated for in assuming the stack exit temperature to be
the same as the flue gas temperature (300 F). No correction was applied for
temperature drop due to dilution air or to stack losses.
An interesting side-light of the day's run was the negligible effect observed
when the fuel-oil additive used in the Lever furnaces was turned off for a
3^ hour period. Fuel additive is composed of finely divided oxides of magnesium
and aluminum suspended in a light carrier oil and is blended with the fuel oil
in the ratio of 1 gal. of slurry to 1500 gal. of fuel oil. It was turned off
- 29 -
-------
during the data run in the hope that it would have a significant effect on
the S02 content of the flue gas. The fact that its effect was negligible is
because the additive is primarily effective in removing SO from the products
of combustion, an essential function in its role of inhibiting the buildup
of corrosive deposits on the boiler surfaces.
3.5.5 Results - October 2/69
The remote sensor data for this day's run is provided in Table A-l in appendix A.
The in-stack, remote and Shell data is presented in Figure A-l. As shown in
Table A-T sky and. visibility conditions for;accurate remote sensor measurements
were very poor and the fact that surprisingly good results were obtained was
due to the'light winds and fairly well behaved plume. Low sky brightness is
evident in the observations made and the relatively high AGC levels.
The effect of scattering on the remote measurements is clearly evident in
Figure A. 2. The extrapolation to zero,, range was performed with the knowledge
that the tangent to the curve becomes a horizontal straight line as an
asymptote as range reduces to zero. The shape of the lower curve so constructed
was transposed to the higher minimum range values shown in Figure A.2 to arrive
at stack exit concentrations for these times. These zero range values were
then multiplied by the dilution factoruand plotted in Figure A.I. The number
of remote data points was less than the previous.day because of time taken to
move from station to station. The in-stack sensor data is also rather sparse
and what is shown is probably not very accurate because of inadvertent miss-
alignment of the in-stack unit. The remote data agrees with Shell results
within about 100 ppm which is considered good in view of the meteorological
conditions. " , .
3.5.6 Results - October 3/69
The remote data is provided in Table A-2the, ranging plots in Figure A. 3 and
the combined data in Figure A.4. The in-stack data is also shown averaged over
- 30 -
-------
a sample interval similar to that of Shell. The in-stack and remote data tend
to agree with each other more closely than with the Shell.
3.5.7 Results - October 6/69
The remote data are shown in Table A.3 and the combined data in Figure A. 5.
This was a day of bright but very hazy sky and gusty winds. The pronounced
scatter in the remote data points is due to the fluctuating plume, low values
by the plume center moving out of the F.O.V. and occasional high values
probably due to sudden shifts in wind direction which produced greater pathlengths
than the stack diameter. S02 was injected into one furnace from a lecture bottle at
time 1140 and it was recorded by all three measurements. The Shell values at this
time are low because the sampling time exceeded the duration of the SO. peak.
A second injection at 1350 was also recorded by the three methods. The lack of
exact time coincidence in the remote and in-stack data points is very likely
due to inexact time observations. Because all data shown in Figure A.5 was
obtained at close range stations P5 and P6, scatter correction has not been
included.
3.5.8 Results - October 7/69
The remote data are shown in Table A.4 and the combined data including remote
from P6 in Figure A.6. These data agree fairly well. The few low in-stack
values are due to tube blowing. Following the heavy rain in the afternoon a
series of ranging measurements with the remote sensor were performed and the data
are shown plotted in Figure A.7. The zero range value of .760 ppm when referred
to the flue is 1050 ppm which compares reasonably well with the nearest Shell
value of 980 ppm.
3.5.9 Results - October 9/69
Remote data is presented in Table A.5 and combined data in Figure A.8 (a). A part-
- 31 -
-------
icularly interesting feature of this day's run was the injection of 15 Ibs. of
SO2 into boiler NO. 1 inspection door. The response of all three measurements
are shown quite dramatically in Figure A.8 (a). Copies of the chart profiles are
shown in Figures A.9. (a) and A.9 (b) and high resolution plots of the injection
interval are shown in Figure A.8.(b). The shell result again is low because of
the length of sampling time relative to the duration of the S0_ peak. The
plots of Figure A.8'(b) have been normalized in time by superimposing the prominent
peak values. The difference in the height of the profiles in Figure A.8 (b)
is not known with certainty but possible reasons could be the effect of the
chimney baffle which prevents mixing of the two boiler outputs until a point in
the stack is reached well above the roof level. If flue gas flow about this
point tends to be laminar then the remote data would tend to be high or low de-
pending on the part of the plume under observation. A second reason could be
the lack of well defined remote values just before and just after the injection.
Correlation calculations for the data of Figure A.8 (b) are as follows:
Correlation Coefficient 0.925
Regression Coefficient 1.070
Intercept 129.4
3.5.10 Lakeview Generating Station
Installation plan of the Lakeview station was shown previously in Figure 3.5.
Two different instrument installations were used during the program as indicated
in the layout sketch of Figure 3.18(a). Figure 3.18(b) shows the in-stack and
Shell installation in generator unit No. 7.
Remote sensor ranging experiments were conducted on both stacks #2 and #4 using
the stations and traverse lines indicated in Figure 3.5.
Initial tests began with the equipment installed in the breeching of unit No. 3.
- 32 -
-------
80O-
700
600-
500-
400
300-
200-
3
0 100-
UJ
1-0
1-5 2-0 2-5
MODULATION FACTOR -
3-0
FIGURE 3-6
80-
70-
60-
50 H
40H
30-
20-
10-
FIGURE 3-7
VISIBILITY
VS
SHELL READING
SLANT RANGE REMOTE READING
RATIO
'SLANT RANGE
10
15
20
25
-------
LAKE ONTARIO
LAKEVIEW GENERATING STATION FIGURE 3-5
-------
MOBILE REMOTE SENSOR
FIGURE 3-3
< I
i i
M2
n REF
F°RK
u
, REFRACTOR
PLATE
L «* -3 - M.RHOW
//CELLS _£ ^ _^^===l^y
CYLINDRICAL LENS
SLIT
JNCIDENT "~_^J7 /L-
UOHT Mljf IT
PHOTO MULTIPLIER
TUBE
M3
ORATING
CORRELATION
MASK
FIGURE 3-4
Dispersive system for vapor detection using
spectrum correlation filter.
- 38 -
-------
INSTALLATION OF SPECTROMETER IN-STACK UNIT
LEVER BROS. OCT 69 FIGURE 3-1
Probe
Return Sample Slot
Mirror \|-y4meteM
Grating
Torque Motor
(above grating)
Stack Wall
I t Condensing Optics
II Quartz Lamp
Ref Cell
Phototube
(Sig)
Mirror
Entrance
Slits
Exit
Correlation
Masks
Phototube
(Ref)
FIGURE 3-2
SO2 STACK MONITOR LAYOUT
- 37 -
-------
3.6 CONCLUSIONS
The results of the Phase I program may be summarized as follows:
(1) The in-stack monitor is capable of accurate measurements of S02
in flue gas from both oil and coal-fired sources.
(2) The remote sensor is capable of making measurements of SO in
stack plumes over a wide range of sky brightness conditions provided
that wind speeds do not exceed about 10 mph and the plume is well
behaved. The error analyses tabulated in Tables 3.2 and A.8 clearly
show that the errors become progressively worse as the slant range
increases and also as the visibility decreases. It should be
remembered that the error analyses shown in these Tables is based
upon the sensor measurements equivalent to gas concentration at the
stack exit and air dilution effects occuring within the flue have
been omitted. This dilution does not effect the reference used for
Table 3.2 and A.8, which was the in-stack shell measurement.
(3) Performance of the in-stack unit and remote sensor was not evaluated
over a controlled range of flue gas aerosol content because it was
not feasible to re-arrange the furnace combustion parameters of a
production oriented facility.
3.7 ACKNOWLEDGEMENTS .
It .is difficult to express adequately, our appreciation to the following
organizations for their very valuable assistance and co-operation provided
to Barringer Research over the duration of the program. Without such generous
assistance this program would not have been possible.
Gulf Oil of Canada Limited
The Hydro-Electric Power Commission of Ontario
Lever Brothers Limited
- 36 -
-------
3.5.14 Results - November 28/69
Remote data is presented in Table A.8, ranging data in Figure A. 16 and combined
data in Figure A. 17. Visibility was good all day but high gusty winds made
remote measurements very difficult. Figure A. 16 shows the very large scatter ,
in the data caused by fluctuating plume. A reasonably good run was made between times,
1000 and 1100. As the day wore on, gustiness increased and the data became so
scattered as to be unusable. The in-stack data in Figure A. 17 shows a substantially
lower reading relative to the Shell. The cause for this is not known with
certainty, but a likely reason was excessive air leakage around the in-stack
mounting flange or through the openings in the instrument itself. Since the
location of the in-stack unit in generator unit No. 7 was on the inlet side of
the fan, the in-stack unit was operating into a negative duct pressure of
some 12 in. of water. Excess air flow into the air curtain passages of the
in-stack probe results in distortion of the air curtains, reduced optical path
length and low readings. Low readings have been traced to this cause on previous
occasions.
The remote data presented in Tables A.6, A.7 and A.8 has been re-grouped in
Table A.9 to show more clearly the relationships between slant range, weather
conditions and percentage error.
- 35 -
-------
data and Figure A. 13 shows the combined data. Meteorological conditions were
near ideal for these measurements. Two ranging runs were made as shown in
Figure A. 12., one in the morning and one in the afternoon. Figure A. 13 shows
the close agreement achieved between the zero range values and the in-stack
results. Figure A. 13 also shows a very interesting comparison of data obtained
while the No. 7 generator was undergoing special tests. Over the duration of this
wide swing in SO levels, the remote sensor was stationed at site S6 thereby
permitting continuous remote data to be generated.
In-stack to remote correlation calculations covering the period 1050-1530
and using only S6 remote data, with no allowance for scatter, resulted in the
following:
Correlation Coefficient 0.987
Regression Coefficient 1.035
Intercept 96.57
The agreement between the two sensors is very close indeed. Shell readings
tended to be about 100 ppm high over the entire period.
3.5.13 Results - November 27/69
The remote data are shown in Table A.7 the ranging data in Figure A. 14
and the flue profiles in Figure A.. 15. As indicated in Table A. 7 there was
continuous moderate snow fall over the duration of the remote sensor measure-
ments. The intent here was to measure the degradation in sensor performance
under extreme scattering conditions. Surprisingly enough there was still an
adequate ultraviolet flux level to operate although AGC levels were approaching
maximum. The effect of extreme scattering is evident in the low zero range
determinations of Figure A.14. The curves of Figure A.14 have been extrapolated
using the usual rule but under conditions of such heavy scattering additional
measurements are required at closer range to better define the curve. Obviously
remote sensing in a snow storm is not too profitable. Figure A.15 shows
reasonably close agreement between the in-stack unit and Shell.
- 34 -
-------
As shown in Figure 3.18(a) the location is on the output side of the fan. During
the test period generator unit No. 4 was shut down. As a result the hot flue
gases from unit No. 3 mixed with clean dilution air which entered the stack via
unit No. 4 under natural draft. The dampers in the No. 4 flue were closed as in
normal practice but since they do not seal, a very significant flow of dilution
air was present during the tests. It was not feasible to measure the flow with
acceptable accuracy because of the duct size (51 x 15' approx.) and attempts
to obtain approximate values using a standard 3' pitot tube probe were unsuccessful.
Remote sensor data obtained under these circumstances i.e. without dilution
correction, produced low zero range values. After several days of tests in this
location Hydro officials reported operational difficulties with the generator unit
and all sensor measurements were discontinued when the unit shut down for repairs.
Because of the extended repair period of No. 3 unit arrangements were made to
install the in-stack equipment in the breeching of generator unit No. 7 and all
subsequent data runs were conducted in this location. On the three days listed,
where measurements were made on generator unit No. 7, unit No. 8 was shut down.
Consequently, for the same reasons quoted above, dilution effects were not known.
3.5.11 Results - October 22/69
Figure A. 10 shows the in-stack results obtained from an initial data run on
generator unit No. 3. A sample of the strip chart is shown in Figure A. 11.
The Shell average is shown as a horizontal bar extending over the sample interval.
The dotted data points were taken off the chart using the appropriate calibra-
tion values while the "x" points were obtained by using the same calibration
value throughout. There is reason to believe that the calibration performed
at 1000 AM was in error causing the data between 1000 and 1200 to be low.
Using the afternoon calibration for the morning readings resulted in the higher
values and closer agreement with the Shell results.
3.5.12 Results - November 24/69
The equipment was installed in the breeching of generator unit No. 7 for these
tests. Table A.6 shows all remote sensor data, Figure A. 12 shows the ranging
- 33 -
-------
- 3000
- 2800
_ 2600
- 2400
- 2200
^ 2000
- 1800
- 1600
- 1400
- 1200
- 1000
- 800
- 600
- 400
- 200
0
REMOTE SENSOR PLUME RESULTS OIL REFINERY OCT 24/68
SENSOR
Demo
STACK
CO Boiler
221°C
190°C (exit)
24.4/sec
MET
Wind
Visibility
Sky
4-5.4 m/s @ 085°
25 Km
overcast
Plume vertical over 3 stack dia's
Target bearing 045° all points
Time 1210 - 1450 EST
Shell Result 1330 - 1350 EST, 1820 ppm
FIGURE 3-8
SLANT RANGE-meters
.00
200
300
I
400
500
- 41 -
-------
REMOTE SENSOR PLUME RESULTS OIL REFINERY OCT 28/68
_ 3000
_ 2800
_ 2600
2400
2200
2000
SENSOR
Demo
STACK
CO Boiler
218°C
188°C (exit)
24.4 m/sec
1445 EST
~ mo
1400
1200
1000
_ 800
_ 600
_ 400
_ 200
0
1100 EST
MET
Wind - variable 7.6 - 10 m/s @ 235°
Visibility - 16.4 Km
Sky - overcast with bright patches
Plume vertical within 20°
Target bearing 045° all points
Shell test 1 1237 - 1255 EST, 1660 ppm
" " 2 1446 - 1512 " , 1800 ppm
FIGURE 3-9
100
SLANT RANGE-meters
200 300
400
500
- 42 -
-------
160CU
REMOTE SENSOR PLUME RESULTS.THERMAL e.s. NOV.I/BB
SENSOR
Demo
140C
STACK
ONo.l
XNo.2
104°C
MET
Wind 4.5 - 6.7 m/s t 080°
Visibility 11.5 Km
Sky cloudless, hazy
H-
UJ
O
U
120C
100C-
800.
600
400
Target Bearing 135°
No. 1 Plume vertical 3-4 dia
No. 2 " " 1 - 2 "
Time 1150 - 1515 EST
Shell 1 1200 1300 EST 1430 ppm
" 2 1300 - 1400 " 1430 "
FIGURE 3-10
200
SLANT RANGE-meters
I
ion
xnn
400
600
43 -
700
800
900
1000
-------
R4 STACK AND
BOILER HOUSE
= Remote Sensor Sites
LEVER BROS. (OIL FIRED SOURCE)
FIGURE 3
- 44 -
-------
SPECTROMETER
PROBE
LOCATION-
STACK
BAFFLE
ROOF
PORT
FAN
BOILER
UNIT
No. I
GROUND
PORT
FAN
BOILER
UNIT
No. 2
(SHUT DOWN
DURING
TEST)
PORT
BOILE
UNIT
No. 3
LEVER BROS (OIL FIRED)
FIGURE 3-12
- 45 -
-------
LEVER BROS
OCT. 69
1 SPECTROMETER WET CHEMICAL
IN-STACK UNIT UNIT
REMOTE SENSOR 8 OIL FIRED SOURCE. OCT. 69
- 46 -
-------
BAUSCH & LOMB (
LEVER BROS. OCT. 1/69
TYPICAL IN-STACK TRACE
FIGURE 3.'15
- 47 -
-------
LEVER BROS. STACK
(FROM P5)
OCT. 1/69
r I i^i i i i i i i i i i i i i i i i i i i i ~i i 11
-------
~
O
<*
13001
1200
1100-
1000
900-
800
8
0
600
500-
400-
300-
200
100-
0
900
SOa CONCENTRATION (IN FLUE) VS TIME.
DATE OCT. l
LOCATION LEVE*
FUEL OIL
IN STACK MONITOR AVGE.
SHELL AVERAGE
* REMOTE SENSOR (*** 12*
*
FIGURE 3-I7
IOOO
noo
1200 I300 I400
TIME (MRS) EST
I500
I600
I700
- 49 -
-------
STACK No. 2
IN STACK
PROBE
LOCATION
PRECIPITATOR
GENERATOR
UNIT No.3
STACK No.4
IN STACK
PROBE
LOCATION H
DAMPER^ FAN
FAN j J FAN
\
^m
i 1
\ t
i /
/vv
PRECIPITATOR
GENERATOR
UNIT No.7
PRECIPITATOR
GENERATOR
UNIT No.4
! FAN j ! FAN
t
, I
k 1
V
V
V
PRECIPITATOR
GENERATOR
UNIT No.8
FIGURE
LAKEN/IEW GENERATING
(COAL FIRED)
3-18 (a)
STATION
- 50 -
-------
WET CHEMICAL
EFERENCE SYSTEM
IN-STACK
MONITOR
FIGURE 'i^H NOV. 69
IN-STACK SPECTROMETER INSTALLATION LAKEVIEW G. S. UNIT No. 7
-------
TABLE 3.3
0^2 *v«-"»1-'ic. o£.iMoun. t-ittitt an£>Cii w ..»..» ~~»,. . .. ~^., . ...........
Sensor Ser. No. "DEKO REF. CELL C-|Li =2f«OC2Lp = /5S<) DATE QffT. I &?
O
CO
METEOR
UJ
H
00
SENSOR
RESULTS
NO.
Ht fn)
ni ^ (m)
Terno. (°C)
Skv
Vis.
Wind- (mph/°)
PLUME
No.
Vaearee
\
Tir-.e
Acerture
Vj_ Div.
V2 Div.
VT Div.
AP-C fvoH-O
FOV. (m)
XLT. (m)
He (m)
Temp. Factor
ST/^"^^ "T
bo.H^OK (ppin ) AT
FLUti
SHELL (ppm)
Note
M >
(2)
(3)
(4)
(5)
(M
(7)
/
4Z.-7
z-a
/S"©
7
/"*"
tof
if:
at
PC
iR-f
940
5"
. 6^
*3G
56
^T-
Z/ 7?
2--7fc
L^-(o
/.2 -
-------
TABLE 3.3
SO2 REMOTE SENSOR DATA SHEET STACK L<
Sensor Ser. No. REF. CELL C-)Li =29OtP2^2 = »SSO DATE./*
^
O
CO
METEOR
Ld
CO
SENSOR 1
| RESULTS
NO.
Ht (n)
n.a (m)
Temo. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
Ho.
& dearee
Tir.°
Aoerture
V]_ Div.
V2 Div.
VT Div.
acr fvn^^c^
FCV. (m)
XLT. (n)
Rc (m)
Tenp. Factor
STSdr^Jv1?^ "T
aj-jjg^--< ippm) AT
SHELL (pprn)
Note
n >
m
(3)
(4)
(5)
ffil
(7)
,
42-7
2-6
-3^o°
ts~+
^~-
fr
i i
p^
/^-A
Il2o
/*r
62.
3,5-
3-4
S~'Q
'23
2-76
726
1-41
£63
Ql^.
f/32.
&/
J53-5"
-^S*-r5
6&7
Q^fi:
//^Lo
&l
SZ'S"
-5S
74-d
/o37»
f/SQ
34
37
762
/OSO
S"
/
IS/OB
//&S
-
&s
7/2.
9*1
&7S
rats'
>l
33 &
Sfi»^"
767
/oso
/22.£T
^5^.
6S7
^V-a
/24O
^
4»S5"
76 /
/OSb
7
/
A4/O7'
/SOQ
JT9
J»2
4'^
73S
DCATION.4»J ,.Y£(\?. . !
acT..t.3.G& '
/3/Q
-3SJ3L
-
s*o
f>9£»
/J5P<=>
2a
.SV3Q
8/2.
Notes: (1) Cloud Cover in lOths. Smoke (k) . Haze (hi. Foa (f) . Etc.- (2) (31 Note where taken: (4) Sketch elevation and
^^^^
plan (5) Hand-held Abney, (6) 2400 hr. clock local time, (7) volts from moter or divisions from chart. Opt.
- 53 -
-------
TABLE 3.3
S02 REMOTE SENSOR DATA SHEET
Sensor Ser. No.
REF. CELL
/350
STACK LOCATION
DATE./rr..£?<
No.
Note
Ht (m)
(m)
2-6
Temo. (°C)
36oV
Skv
n
Vis.
(2)
Wind (mph/°)
(3)
PLUME
(4)
17/07
vv
7A/Z>
SUI/A/Y;
No.
& degree
(5)
Time
ffil
14.10
Aperture
Div.
(7)
V2 Div.
.2S/
Jl/
VT Div.
2.7*3
SL7
FOV. (m)
XLT. (m)
S (m)
Temp. Factor
£62
777
637
6 a
tppm; AT
9/5
Io7o
9/S
SHELL (ppm)
Notes:
(1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog (f ) , Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
- 54 -
-------
Section 4
PHASE II
4.1 OBJECTIVES
The objectives of the Phase II tests were to monitor the .emissions of nitrogen
dioxide from sources using coal and oil fuels with high and low sulphur contents
using the airborne sensor and to track the plume down wind to zero response of
the instrument; analysis of the nitrogen oxides in the stack by wet chemical
means to be included for reference. Thus from these measurements it would
appear possible to check upon the conversion of NO to NO_.
In most boiler plants and in a a number of furnaces, excess oxygen is kept to a
minimum for reasons of economy. This results in sulphur being emitted as SO^
rather than as SO and nitrogen as NO rather than as N0_. The NO is oxidized
in the atmosphere to NO which can then initiate photochemical reactions and
promote the development of smog. There are several ways in which NO can be
oxidized, but the method and the rate of oxidation are unknown and measurement
is difficult during the natural diffusion and dilution of a plume. The Barringer
airborne remote sensing N0_ spectrometer provided an instrument technique
whereby the concentration-pathlength of N0« below the aircraft could be followed
as the aircraft overflew the plumes.
4.2 THEORETICAL BACKGROUND
If a remote correlation spectrometer is set looking vertically upwards, then
ignoring the light scatter into the instrument by particulate matter in the light
path, the light falling on the phototube originates from a small part of the
sky defined by the instruments field of view and the signal generated by the
instrument will be a function of the amount of target gas within the field
of view. In the case of an aircraft flying above the inversion layer or above
- 55 -
-------
most of the atmospheric pollution, this upward looking signal is very small
and the instrument can be zeroed under these conditions. If the airborne
spectrometer is now put in the downward looking mode, the increase in signal
will be a function of the additional target gas in the extended light path.
As a first approximation this additional light path is twice the height of
the aircraft. However, since the terrain beneath the aircraft within the field
of view of the instrument gives diffuse and not specular reflection the light
reflected to the spectrometer from the ground target contains components from
the 2IT dome of the sky. It can be shown that for, a gas cloud in the form of
an infinite layer between the aircraft and the ground, the effective optical
pathilength through the gas layer is greater than twice the height of the aircraft
since the direction of maximum illumination of the ground is at an oblique angle
through the gas layer. The effect of this oblique traverse can be obtained from
a computer program which is used to handle the airborne data.
4.3 INSTRUMENTATION
The airborne sensor used was a Barringer NO« remote sensor optimized for high
sensitivity to nitrogen dioxide and a fast response time of about one second.
The NO. spectrometer was installed in an Aerocommander aircraft as shown in
Figure 4.1; it has been successfully used in the past and is described in detail
in reference 1.
Briefly, the airborne system consists of the remote sensor looking into a 45
mirror which can be rotated to reflect light from either the ground below the
aircraft or the sky above. The mirror is housed in a pod on the side of the
aircraft and is clearly visible in Figure 4.2 which shows a low level flight along the
Toronto waterfront. Data is collected on a multi-channel oscillograph recorder
which records, in addition to the remote sensor output, the AGC voltage which is
a measure of the total light flux. It also records the barometric pressure and
an event marker records the frames of the flight path recovery camera.
- 56 -
-------
In operation, the aircraft flies above the inversion layer which is normally
clearly visible from the aircraft and could also be checked by air temperature
measurements. The assumption is then made that the amount of gas above the
aircraft is negligible, thus a zero can be established by setting the mirror
to reflect skylight from above into the aircraft and adjusting the zero accordingly.
The zero is normally checked at each end of a flight line and the sensitivity
is checked by inserting a calibration cell of known gas concentration and path-
length product into the optical path.
The simultaneous analyses of oxides of nitrogen in the stacks under surveillance
was performed.by wet chemical methods. The total oxides of nitrogen were determined
by exposing a known volume of gas to a sodium hydroxide solution. Absorption
of the nitrogen dioxide by the sodium hydroxide upsets the equilibrium and causes
the oxidation of more nitric oxide. After a period of about an hour, the oxida-
tion is complete and the total nitrogen dioxide is determined by the addition
of phenol disulfonic acid and subsequent colorimetric analysis at 4100 A. The
colorimetric response is calibrated from standard solutions of sodium nitrite.
4.4 EXPERIMENTAL SETUP
Arrangements were made to follow plumes from both the Lakeview Generating Station
and the Gulf Refinery. Both these sources are located on the north shore of Lake
Ontario and to eliminate the noise associated with changes in ground albedo when
flying over land, all flights were made over the Lake.
Neither one of the two pollution sources selected is ideal since both are comprised
of multiple stack sources. The Lakeview Generating Station had five boilers on
line exhausting through three stacks, only one of which has sampling facilities.
The three plumes however, quickly merge and the airborne survey consequently
covers the combined plumes. Since this is a base load station with all units
evenly loaded and burning the same fuel, the error in taking one stack as represen-
tative can be expected to be quite small. At the time of the tests, the total coal feed
- 57 -
-------
was 500T/hr. with 3% excess oxygen. A typical measured nitrogen oxide content
of 190 ppm is equivali
later in Section 4.6.
of 190 ppm is equivalent to a nitrogen dioxide release of 14.9 M /min., as shown
At the Gulf Refinery, there are three boilers and an assortment of process
heaters and tube stills. These plumes quickly coalesce and no meaningful estimate
of the total production of nitrogen oxides can be made. The airborne data
however, would indicate the rate of conversion of NO to N0_.
The aircrew attempted to fly a sinuous path over the plume. However, due to the
micrometerological effects of the large body of cold water of Lake Ontario,
pronounced curvature of the plume was found and it was found advantageous to
locate the downwind end of the plume and develop the flight path from that point
back towards the point where the plume could be located by direct vision. In
this case, the gas at the downwind end of the plume had been emitted from the
stack some 90 minutes before measurement. Arrangements were therefore made to
sample the stack for a period of at least 90 minutes before the flight started and
to continue until the aircraft arrived at the stack.
The success for these series of measurements depended on the selection of
satisfactory meterological conditions. In many cases tests were started under
satisfactory conditions but subsequently had to be aborted due to the buildup of
cloud or a change in wind direction, these conditions being particularly difficult
to predict because of the total elapsed time of the airborne survey.
Some difficulty was encountered in deciding how far the plume could be followed.
Under flight conditions the recognition of small changes in the signal when the
plume is widely dispersed is a matter of some difficulty and in many cases the
processing of the data in the laboratory showed satisfactory indications of the
plume when these indications were not apparent .in the aircraft. This caused
many flights to be discontined based upon the airborne operator's judgement, while
subsequent data analyses at the laboratory indicated that further flight continua-
tions could have obtained useful measurements.
- 58 -
-------
The decision to carry out the airborne plume survey over Lake Ontario to take
advantage of the steady reflective background of the Lake surface,necessitated
a steady north or northwest wind and in general flights were only attempted
during these conditions.
Three flights have been selected for detailed examination and they were flown
on August 27 (flight I), September 30 (Flight II) and October 31, 1968 (flight
III). During Flight I, the wind was northwesterly but during Flights II and III,
wind shear was encountered. During Flight H,the plume started initially from
the northwest, but rose to come under the influence of a southwesterly wind,
while during Flight III the plume started inland under the influence of a
southerly on-shore breeze and reversed direction under the influence of a north
wind at higher altitudes.
4.5 RESULTS
4.5.1 Results - Flight I Lakeview Power Station - August 27, 1968
August 27th was a clear cool day with a 70 temperature, with north and northwest
winds at 12 miles per hour. Eight parallel flight lines were laid down at 1-1/2
mile intervals downwind from the Lakeview Generation Station. The complete survey
consisted of traversing the plume along each of eight flight lines in the outward
direction from the stack with a flight on every other flight line on the return
back to the start a total of eleven passes. The location of the flight lines in
relation to the coast and Lake Ontario are shown in Figure 4.3. From the correlation
spectrometer chart the plume width, height, and area under the curve have been
computed for each traverse of the plume along the various flight lines.
Table 4.1 gives the data derived from the eleven passes. Columns 2, 4 and 6 show
the trace dimensions in centimeters taken from original record while columns 3, 5
and 7 give the quantities derived from the calibration constants. Figure 4.4
shows a plot of the NO burden (col. 7 of Table 4.1) plotted against distance. The
plot indicates that there is a general increase in NO- burden with distance, but
there is no evidence that the reaction is approaching completion. Figure 4.5
shows the relative plume shapes and indicates that the plume width and concen-
tration apparently do not change in a regular manner. Visual observation of
- 59 -
-------
TABLE 4.1
DATA FOR PLUME MEASUREMENT
LAKEVIEW G. S. - AUG. 27, 1968
1
Pass
No.
1
2
3
4
5
6
7
8
9
10
11
2
Trace
Length
Cms.
1.1
3.0
8.1
5.7
8.4
9.5
11.6
9.2
8.7
4.8
4.8
3
Plume
Width
M
315
860
2320
1640
2410
2720
3330
2640
2570
. 1390
1390
4
Peak
Ht.
Cms.
4.0
3.4
3.1
5.1
4.7
4.6
5.1
6.0
3.5
2.5
3.1
5
N02
ppm-m
31.1
26.4
24.1
39.6
36.5
35.7
39.6
46.6
27.2
19.4
24.1
6
Planimeter
Area
Cm
2.4
5.0
10.8
14.0
27.0
30.5
34.4
34.2
21.0
6.2
6.3
7 *
N02
2
ppm-m
xlOOO
2.4
5.0
11
14.3
27.3
30.8
34.7
34.5
22.0
6.3
6.4
8
Distance
Miles
0.5
2
3.5
5
6.5
8
9.5
11
8 .
5
2
Wind Velocity = 15 mph. 12 mph.
= 27 m/sec 5.4 m/sec
* Divided by 2.2 for the 'm1 factor
- 60 -
-------
TABLE 4.2
ADJUSTED DATA FOR AUG. 27 PLUME
1
Pass
No
1
2
3
4
5
6
7
8
9
10
11
2
ADJ
Plume
Width
M
320
850
1300
1750
2200
2650
3100
3550
2650
1750
85
3
ADJ
Peak
N02
ppm-M
31
26
23
25
30
35
39
46
35
25
26
4
=2x3
xlO3
10
22
30
44
66
93
121
163
93
43
22
5 *
Corrected
N02
Burden
ppm-M
xlOOO
2.4
4.9
5.9
9.7
20.5
29.5
31.5
46.1
29.5
10.0
4.4
6
NO
Flow
M /min
0.8
1.6
1.9
1.9
6.6
9.3
10.0
14.5
9.3
5.2
1.4
* Divided by the 'm1 factor
- 61 -
-------
plumes indicates that under the influence of the wind, a plume may undergo
looping or fanning and that these effects may stretch or break the plume so
that on a given traverse, an airborne sensor of the type in use for these
tests might record a double thickness through a loop or a double width due to
fanning or might record almost no plume. In an attempt to overcome these
effects the plume width and NO burdens are plotted against distance in Figure 4.6
and smooth curves drawn through the points. The plume width shows a fairly
regular increase with distance, reaching 3500 meters at a distance of 16,000
meters to give a plunu
theory. (see Ref. 6)
meters to give a plume half angle of approximately 6.5 in good agreement with
The values for the NO burden are variable but show a well defined minimum at a
distance of about 3-1/2 miles from the stack. At a point about 4 miles from the
stack, the burden of NO begins to increase in an apparent linear manner with
distance in spite of the increasing width of the plume. Table 4.2 shows a re-
calculation of the total NO burden based on the smoothed curves of Figure 4.6
and the total NO flow calculated from the corrected NO burden in column 5 and
a wind velocity of 12 m.p.h. The effect of smoothing the width and NO burden are
shown by the reduced scatter of the points in Figure 4.7 compared to Figure 4.6
This graph appears not to go through the origin but to show an initial NO rate
3
of 0.5 m / min. It also indicates that the rate is not linear with time but
increases for a period of at least 30 minutes after leaving the stack.
Calculation for smoothing plume width and vertical burden:
From Table 4.1.
Plume width = 1640 metres
Vertical Burden = 39.6 ppm-m
Product = 65,000 ppm-m
2
Planimeters = 14,300 ppm-m
Factor = 14,300
65,000 ~
Smoothed Plume Width (Table 4.2) = 1750 metres
Smoothed Vertical Burden = 25 ppm-m
2
Product = 44,000 ppm-m
Smoothed Planimeter = 44,000 x .22
= 9,700 ppm-m
= Corrected Burden
- 62 -
-------
4.5.2 Results - Flight II Lakeview Generating Station - September 30, 1968.
This flight indicates the usefulness of the airborne spectrometer system in
chasing plumes after they cease to be visible in the naked eye. By observation
the plume was blown out over the lake under the influence of a northwest wind,
estimated to be about 4 miles per hour. However, about 4 miles over the lake
the plume passed through a wind shear and then proceeded under the influence of
a southwest wind at about 20 miles per hour. The nitrogen dioxide signal from
this plume was traced for a distance of 25 miles where the width of the plume
had increased to approximately 8000 meters. Traverses were then made at 4 mile
intervals back to the stack. The locations of the traverses in relation to the
coast of Lake Ontario are shown in Figure 4.8.
TABLE 4.3
LAKEVIEW G.S. - SEPTEMBER 30, 1968
Trace Path Peak N02 Planimeter NO2 2 Dist.
Length Length Height ppm-M Area ppm-M
Cms M Cms Cm xlOOO Miles
1.
2.
4.
5.
6.
7.
8.
9.
10.
11.
12.
5.2
7.9
24.9
25.4
26.7
22.8
19.7
14.7
12.8
10.2
7.6
1520
2320
7320
7470
7850
6700
5790
4310
3770
2990
2230
0.7
0.8
0.6
0.7
0.9
1.5
0.9
2.2
2.6
2.3
1.0
9.2
10.5
8.5
9.2
11.8
19.7
11.8
15.8
34.2
30.5
13.2
1.7
3.2
9.3
8.2
14.7
17.8
12.2
18.6
21.0
14.6
4.6
2.9
5.5
16.0
13.8
25.1
31.2
21.0
32.0
36.2
25.1
8.0
2
5
25
25
22
18
14
10
6
4
3
* Divided by 2.2 for the 'm1 factor
- 63 -
-------
TABLE 4.4
SEPTEMBER 30, 1969
LAKEVIEW POWER STATION
1
Pass
No
1
2
4
5
6
7
8
9
10
11
12
2
ADJ
Plume
Width
M
1500
2700
7600
7600
7000
6300
5300
4400
3200
2300
1900
3
ADJ
Peak
N02
ppm-M
12
30
9
9
11
16
21
29
34
26
20
4
=2x3
xlO3
13.5
83
73
73
91
-134
132
127
102
50.6
30
5*
Corrected
NO
ppm-M
2.9
19.5
20.15
20.5
19.5
31.8
41.5
37
29
14.5
8.2
6
NO,,
2
Flow
.8
10.5
11.0
11.0
10.5
17.2
22.3
20.0
15.5
3.9
2.0
* Divided by 2.2 for the 'm1 factor
- 64 -
-------
As in the previous flight, the data read from the recorder chart is shown in
Table 4.3 and the integrated N02 times width burden is plotted against distance
in Figure 4.9. The plume width and vertical burden are plotted in Figure 4.10,
which is used as a basis for the adjusted widths and burdens used in Table
4.4. In this graph, the influence of the low wind speed on the vertical NO_
burden is clearly shown. The low wind speed influences the NO concentration by
suppressing turbulent mixing and by its failure to "stretch" the plume. Thus,
the minimum which was a feature in the previous flight in either missing or becomes
of minor importance and the vertical burden builds up to a very high value before
the faster upper winds promote diffusion. The corrected N0_ flow is plotted
against distance and time in Figure 4.11, which indicates a maximum flow rate of
22.3 M3/min.
22.3 M ,/min. = 22.3 x 60 M3/nr.
= 22.3 x 60 x 46 kg/hr.
= 2.73 te/hr. 22'4
= 3.0 tons/hi.
4.5.3 Results: Flight III Gulf Refinery - October 31, 1968.
This flight was carried out during a period of north to NNW winds at 10 m.p.h.
On this occassion there was a slight onshore breeze at stack height which caused
a plume reversal as the smoke passed through a wind shear into the northerly flow.
This results in a plume traverse at a distance of 'minus one mile1 as shown for
pass number 14 in Table 4.5. Figure 4.12 shows the location of the flight lines
and a vertical elevation through the plume. In Figure 4.13 a virtual source has
been postulated at a location represented by 3-1/2 miles, ie 3^1/2 miles upwind
of the stack, and the values for traverse 13 and 14 have been halved since they
appear to represent passes through the double plume.
- 65 -
-------
TABLE 4.5
N02 - GULF OIL
OCTOBER 31, 1968
Pass
No
1
2
5
6
7
10
11
12
13
14
Trace
Length
Cms.
6.1
8.9
22.5
20.5
7.1
11.0
9.1
6.2
3.5
5.2
Path
Length
M
1750
2560
6460
5890
2040
3160
2610
1780
1000
1490
Peak
Ht.
Cms.
1.5
1.4
1.6
1.7
2.0
1.1
1.3
0.9
2.8
3.3
N02
ppm-M
11.6
10.9
12.4
13.2
15.5
8.5
10.1
8.2
21.8**
25.6**
Area
Cm
6.4
6.0
20.4
20.0
6.7
6.3
7.5
2.9
5.3
7.1
N02*
2
ppm/M
22.7
21.2
72.0
70.0
23.8
22.3
26.5
10.2
18.7
25.0
Distance
Miles
1
3
12
15
13
6
3
1
0
-1
* Divided by the "m" factor
** Divided by 2 for reversed plume
- 66 -
-------
TABLE 4.6
SMOOTHED WIDTH AND BURDEN DATA
Pass
No
Distance
from
virtual
Source
Miles
Width
M
Burden
ppm-m
2.
ppm-m *
x 1000
1
2
5
6
7
10
11
12
13
14
4
6
15
18
16
9
6
4
3
2
1600
2400
5700
6800
.6000
3500
2400
1600
1200
800
10.
9.
12
14.
13
8.
9.
10.
11.
13.
5
0
5
5
0
5
5
5
6.8
13
34
49
47
16
9.8
11
7.3
7.5
*Divided by the 'm1 factor
The raw data from Table 4.5 is plotted in Figure 4.13 and smoothed values from
this figure are used as a basis of the adjusted data in Table 4.6 and Figure
4.14. Figure 4.14 shows a continuous increase in NO burden but .even after
smoothing, the scatter is still too .great for the form of the curve to be
determined with confidence.
- 67 -
-------
4.6 COMPARISON OF AIRBORNE AND IN-STACK NO MEASUREMENTS
The steady operation of the generating station permits a direct comparison of the
quantity of oxides of nitrogen measured from the air with a calculated quantity
based on a stack analysis.
Under steady base load conditions, the fuel rate is 100 tons/hr per boiler, and
with about 3% excess oxygen, the total flue gas is 1350 tons per hour per boiler.
The wet chemical measurements, together with relevant facts gathered for the period
under examination, are shown in Table 4.7.
FLIGHT I AUGUST 27, 1968
Flue Gas Oxides of Nitrogen = 190 ppm
Boilers on line 5
Therefore NO2 potential
-6
= 1350 x 5 x 190 x 10 x 46_ = T/hr
29
-6 3
= 1350 x 5 x 190 x 10 x 4£ x 2000 x 1_ MJ/min
60 29 2.2 2.06
= 14.9 M3/min
Aircraft measured maximum N0_ rate = 14.5 M /min
FLIGHT II SEPTEMBER 30, 1968
Flue Gas Oxides of Nitrogen = 390 ppm
Boilers on line 5
- 68 -
-------
NO potential = 1350 x 5 x 390 x 10~6 x 46 x .2000 x ! = 28.3
60
29
2.2
2.06
Aircraft measured maximum NO rate = 22.3 M /min.
FLIGHT III - OCTOBER 31, 1968
Owing to the diverse nature of the sources and the lack of analytical data for
many stacks, no estimate of the NO production is possible.
TABLE 4.7
WET CHEMICAL ANALYSIS OF LAKEVIEW POWER STATION STACKS FOR
OXIDES OF NITROGEN
DATE
1968
August 27
Sept. 4
12
12
17
18
19
25
30
TIME
14:30
14:30
15:00
16:30
16:00
15:00
14:40
10:00
14:00
EXCESS AIR
AS % O
3.5
4
4
3.5
3.5
3.7
3.7
3.7
3.5
FURNACE
TEMP. °F
560
630
630
630
635
640
635
635
625
NO AS
x
NO. ppm
190
400
720
530
490
46.0
410
410
390
- 69 -
-------
4.7 THEORETICAL ASPECTS OF PLUME DISPERSION
The behaviour of a plume under the influence of turbulence of the wind has
been examined by numerous workers who have derived formulas to relate concen-
tration with distance from the stack. A formula which is widely used, was
developed by Pasquill (Ref7) and takes the form
KQ ., , , V .2 . r , .H-Z.2 , ,H+Z.2,
x = FT7 exp ~h (5v } [exp ~h (5~) + exp '* (5~> ] (1)
y Zp * z z
X = concentration
where Q = quantity emitted
y = wind velocity
x = direction along the wind
y = direction across the wind
z = vertical direction
and H = height of emission
The aircraft during its flight effects a vertical integral of the concentration
»
and the vertical burden (B ) is given by
Z=°°
BZ= x s
Z= o
^ <) (3)
y y
(see reference 8.) For temperatures somewhat above atmospheric, such as would
exist in a stack and with low concentrations of NO in atmospheric oxygen,
-4
the rate constant is approximately 2 x 10 per pphm per hour. At a point
about 1 mile from the stack where the flue gas has been diluted 1000 times
by air, the concentration of nitric oxide has been reduced to 200 pphm and
the rate of disappearance of nitric oxide will be proportional to the square
of the nitric oxide concentration and equal to about 8 pphm per hour. This
reaction is obviously far too slow to account for the rate of production of
- 70 -
-------
nitrogen- dioxide as observed by the airborne sensor.
The rate of reaction of nitric oxide with ozone to yield nitrogen dioxide and
molecular oxygen is approximately 10,000 times faster than the reaction of
nitric oxide with oxygen. Provided therefore sufficient ozone is available to
react with the nitric oxide, the reaction could be substantially complete in a
period of an hour.
It has already been established in studies in connection with oxidant damage to
tobacco crops in South Western Ontario, that the air mass associated with an
anticyclonic high pressure system and south westerly winds contains the remnants
of a well aged photochemical smog with ozone levels of up to 10 pphm. It is
probably also significant that for the run of Sept. 30th where the plume passes
through a wind shear into a south westerly air flow, that nitrogen dioxide is
generated at a high rate and that the maximum quantity of NO is obtained within
a period of about 1.5 hours.
The flight for Sept. 30th also shows a progressive decrease in the burden of
NO. and the maximum value of the integral during any traverse occurs where y = o
is given by
V ' f\
(4)
Therefore the decay in NO concentration after its formation is substantially
1
complete will vary as .
Table 4.8 shows the concentrations calculated from equations (4) above compared
to with the actual measurements.
Column 3 shows the value of a at various distances, and column. 4 shows the ratio
of a' (x) to a (6Km). Column 5 shows the measured values of x at various distances
- 71 -
-------
and Column 6 shows the ratio of X (x) to X (6 H) If equation (4) holds, the
product of columns (4) and (6) shown in column 7, should be constant. The
theoretical decay values predicted by equation (4) and based on the value of
X = 34 at 6 Km are given in column 7. These values suggest that although the
maximum burden exists at 6 miles, the decrease is less than theoretical up to
18 miles and faster than the theoretical exponential rate after 18 miles. The
additional NO over that predicted by theory for 6 to 18 mile section is
presumably due to continued production of NO while the reaction rate is decreasing
due to exhaustion of the nitric oxide. The defecit of NO between miles 18 to 25
suggests that NO is entering into further reactions and is being destroyed in
this region.
The rate constant for the reaction NO + 0 = NO_ + 0^ at a value of about .05
-1-1
pphm hr at ambient temperature and while this reaction rate is several orders
lower than the reaction of ozone with nitric oxide, the reaction is nevertheless
still fast enough to be important, for example, with nitrogen dioxide and ozone
at 10 pphm each, the rate of reaction will be about 5 pphm per hour or a decrease
of about 30% per hour. Thus, at least in terms of orders of magnitude the data
obtained by these airborne surveys is consistent with the oxidation of nitric oxide
and nitrogen dioxide by ozone to NO .
TABLE 4.8
THEORETICAL DECAY OF PLUME CONCENTRATION
X
miles
Km
ay
n
max
Vy
max y
Theor.
Xn
6
10
14
18
22
25
30
10
16
23
29
36
40
48
850
1300
1700
2200
2600
2800
3400
1.00
1.53
2.00
2.60
3.10
3.33
4.00
34
29
24
20
13
9
7
1.00
.85
.71
.59
.38
.26
.205
1.00
1.30
1.42
1.53
1.18
0.86
0.82
34
22
17
13
11
10
9
- 72 -
-------
SPECTROMETER
PHOTOLYZER LAMP ^
/// \1
T~^ INTERVALOMETER
GROUND TRACK CAMERA REF CEa CONTROL NQj
FIGURE 4.1 INSTALlATICffl OF AIRSORNE
CORRELATION SPECTROMETER IN
AEROCOMMANDER AIRCRAFT
FIGURE 4.2 AEROCOMMANDER AIRCRAFT
- 73 -
-------
FIG.4.3 LOCATION OF AIRCRAFT PLUME TRAVERSES
LAKEVIEW P.S. AUG. 27.68
i
50,000-
40,000-
30,000-
20,000-
10,000-
\ I
3 4
l I I
567
MILES
l
8
I l l
10 II 12
FIG 4.4 NOa BURDEN IN ppm-m* VARIOUS DISTANCES
LAKE VIEW 6.S. AUG. 27,1968
- 74 -
-------
AUG. 27, 1968 NOz PLUME LAKEVIEW
ui
o
0-
I-
2-
3-
4-
5 -
6 -
7 -
8 -
9 -
10 -
II-
12-
2/11
4/10
CO
6/9
8
ppm-m ppm-m
31.1 18.6
26.4/24.1 383/49.0
24.1 84,0
39.6/19.4 109.0/482
3 6.5 201,0
35.7/27.2 273.0/163.0
39.6 268.0
46.6 266.0
FIG 4.5 RELATIVE PLUME CROSS-SECTIONS
(SIMPLIFIED SHAPE)
- 75 -
SCALE 0.1"= Icm.
-------
UJ
5000-
4500-
4000-
3500 -
3000-
2500 -
2000-
1500 -
1000 -
500 -
50 VERTICAL N02 BURDEN
ppm-m AS MEASURED
45
x--x PLUME WIDTH
00 MAX. VERT. NOt
I 2345 67 89 10 II 12
MILES
FIG 4.6 PLUME WIDTH AND VERTICAL NOt VERSUS DISTANCE
FIG. 4.7
N02 FLOW IN PLUME WITH DISTANCE IN MILES
AND TIME IN MINUTES FROM STACK
AUG. 27/68
CJ
O
15
12
I234567'89IO
4 8 12 16 20 24 30 40
12 I MILES
50 MINUTES
- 76 -
-------
54
6
LOCATION OF AIRCRAFT PLUMES TRAVERSES
LAKEVIEW RS.
34-
32-
30-
28-
26-
24-
22-
20-
E 18-
I 16-
14-
12-
10-
8-
6-
4-
2-
10
20
25
15
MILES
FIG 4.9 NOa BURDEN AT VARIOUS DISTANCES
LAKE VIEW G.S SEPT. 30,1968
30
- 77 -
-------
V)
u
o
i
9000 -
8000 -
7000-
6000-
5000-
4000-
3000 -
2000-
1000 -
0
36!
32
280
20>/
16
12 '
8/
* PLUME WIDTH
-VERT. N02
10
20
25
15
MILES
FIG 4.10 PLUME WIDTH AND PEAK VERTICAL NOz VERSUS DISTANCE
FIG. 4.1
CORRECTED N02 FLOW
10
5 -
_O
0246
0 05 I
8 10 12 14 16 18 20 22 24 26
125 1.5 1.75 2.0
MILES
HOURS
- 78 -
-------
LAKE ONTARIO
FORM OF PLUME
FIG. 4.12
- 79
-------
UJ
liJ
I
g
$
UJ
6000 -
5000 -
z 4000 -
3000 -
2000-
1000-
»» PLUME WIDTH
OO VERT. N02
0
-3 -
FIG 4.
2-101
8 10 12 14 16
13 PLUME WIDTH AND VERTICAL N02 VS DISTANCE
(M
E
a.
a.
LJ
K
CD
50000 -
40000 -
30000 -
20000 -
10000
-3 -2 -I
FIG 4.14
123456789
ADJUSTED N02 BURDEN
10 II 12 13 14 15
OCT 31/68
- 80 -
-------
Section 5
PHASE III
5.1 OBJECTIVES
The objective of Phase III was to compare the sulphur dioxide measurements of the
point sampler with those obtained for the long Path Sampler. Various long paths were
used and the complete work was conducted in a typical urban environment. It was
thought that the long path sampler would provide readings more representative of the
ambient sulphur dioxide concentrations in the urban environment since they were
averaged over a considerable distance.
For a long path sampler, there is a limiting short pathlength at which its data does
not differ significantly from that produced by a point sampler. This will occur
when the limiting short pathlength is representative of the local ambient air imm-
ediately available to the point sampler. Conversely, there is a limiting long path
length determined by the instrument noise resulting from the decrease in signal
energy as the pathlength is increased. Thus, an additional objective of this prog-
ramme was to determine these minimum and maximum pathlengths for significant aver-
aging by the long path sampler. Additional information was also sought on the acc-
uracy of the samplers by the simultaneous operation of a wet chemical method and a
commercially available coulometric sulphur dioxide analyzer.
5.2 TEST SITE
The test site chosen was an open field in Etobicoke, on the western side of Toronto,
Ontario. The site was located between Evans Avenue to the south, the Queen Eliz-
abeth Way to the north and just east of Kipling Avenue. This location is in the
suburban west end of.Toronto about one mile from the shore line of Lake Ontario.
This is an area of intense development of small factories and is subject to only
minor sources of SO pollution except for a major power station about 5 miles to the
- 81 -
-------
.southwest. The site was chosen because it contained an air pollution test facility
of the Ontario Government and also its open location permitted easy line-of-sight for
the long path monitor. Its location promised frequent inversions in the air flowing
off the cold waters of Lake Ontario.
The dominant feature of the site is a meteorological tower carrying three Bendix aero-
vane anemometers and it was the opinion of the Government's meteorologist that the
trailers which housed our monitors should be parked at least 50 yards from the base
of the tower so that the wind trajectory to the aerovanes would not be disturbed.
The trailers were therefore parked near the south boundary and about 50 yards from
the base of the tower. At the tower's base was a small hut which housed the Beckman
SO analyzer, and thereby allowed a comparison with an accepted commercial instrument.
5.3 INSTRUMENTATION
The Point Sample Ambient Monitor (Figure 5.1) has been designed as an alternative to
the commercially available coulometric or conductometric continuous analyzers. De-
finitive air quality criteria now being formulated call for maximum
concentrations of 50, 25 and 5 pphm of SO and NO for one hour, one day and one month
£* £t
exposure periods respectively. The point sample ambient monitor was therefore designed
for a detection limit of 1 pphm with a time constant of 100 seconds. This led to the
selection of a double pass optical cell 1.25 meters long which thus gives an absorption
path length of 2.5 meters; the correlation spectrometer itself encloses the absorption
volume (Figure 5.2). The absorption bands of SO and NO are sufficiently regularly
spaced that multiple entrance slits can be used, leading to a high light throughput
and hence an improved signal-to-noise ratio. Modulation of the spectrum was one of
the major problems of the entire development. In a high sensitivity system, the re-
quirements for a suitable modulator are stringent in that they should not contribute
to random noise, nor must they generate coherent noise which can result in zero off-
sets. The system chosen uses a torque motor to vibrate the grating and a feedback
control maintains the exact amplitude and mean position. The electronic block dia-
gram is shown in Figure 5.3.
- 82 -
-------
The signal channel uses an R.C. amplifier filter synchronous detection system and
a low pass filter that drives a standard 100 millivolt recorder. An automatic ele-
tronic re-zero circuit operates at fixed intervals to correct any zero drifts.
Sensitivity to the ambient temperature has been a problem, but where the equipment
is used in an air conditioned environment with the additional advantage of the auto-
matic zero reset circuit, the effect of this drift on the operation of the instru-
ment has been insignificant. However, we feel that an extended temperature range
should be available and it is hoped to reduce the temperature sensitivity of the
instrument from the present 3 ppb per F to 0.3 ppb per F.
The Long Path Ambient Monitor is a modification of the remote sensor in which an
active light source is positioned some distance away from the sensor and thereby,
the pathlength is physically defined. To permit operation during the day light hours,
a mechanical light chopper shown in Figure 5.4 provides 100% modulation of the source
at 2400 Hz and the spectrometer signal is processed to recognize signals synchronous
with this modulation. Early tests were carried out with a 3400 K quartz-iodine
lamp which had a pathlength limit of about 500 meters. Further tests using a xenon
arc source indicated that pathlengths up to 1500 meters could be successfully used.
One essential difference between the long path ambient monitor and the remote sensing
spectrometer is the inclusion of a scanning device to produce a low speed oscillation
of the spectrometer grating. This device,which is considered in detail in the theore-
tical treatment ^n reference 2, displays the change in correlation as the spectrum is
slowly swept across the fixed mask from the correlation to the anti-correlation pos-
ition. Measurement of the difference in output between correlation and anti-correl-
ation is a direct measure of twice the correlation signal and minimizes zero off-
sets arising from changes in the spectral slope of the light caused by variations in
the light source or diffential absorption by gaseous interferences or particulate
matter in the atmosphere.
For the short pathlengths, the spectrometer was housed in a trailer placed alongside
the Point Source Ambient Monitor trailer at the south side of the site and the modu-
- 83 -
-------
lated light source was set up near the northern boundary. For the intermediate
pathlength, the light source was moved northward to the roof of a factory building
across the Queen Elizabeth Way (a 12 lane freeway). For the long path length the
trailer was moved to a new location about 600 metres south, while the light source
remained on the factory roof.
A Coulometric Analyser is operated at the site by the Ontario Provincial Government
as a part of their air pollution measurement network. This instrument measures
the flow of electricity necessary to maintain a selected iodine level in a reaction
cell by electrolysis of an iodide solution as the iodine level is depleted by rea-
ction with atmospheric sulfur dioxide. The instrument was operated by Government
scientists; calibrated and serviced according to their normal routine. By arrange-
ment with the Air Pollution Control Dept. of the Department of Health, photostatic
copies of their charts were made available to us as they were processed by the
department.
Wet Chemical Analyses were carried out manually at frequent time intervals using
the West Gaeke Procedure. To monitor the performance of the PSAM, the samples were
collected from a point near the inlet tube of the monitor. For evaluation of the
LPAM, the wet chemical equipment was installed in an automobile and a spatially
integrated sample collected by driving as close to the optical path between the
long path sensor and the remote light source as geographical conditions would
permit. The West Gaeke procedure consists of bubbling air at a known fixed rate
for a selected time period through a solution of sodium tetrachloromercurate. For
field work, the chloromercurate solution is measured into a midget impinger and a
flow rate of two litres per minute is set on a lucite flowmeter using a freon pow-
ered ejector or alternatively a 1.5 cu. ft. evacuated bottle as a source of suct-
ion. Formaldehyde solution and an acid bleached pararosaniline hydrochloride dye
are added to the aspirated chloromercurate and the intensity of the colour devel-
oped is compared spectrophotometrically with known standards. This technique has
a sensitivity of 10 ppb of SO using 10 mis of chloromercurate as a sampling solu-
tion in conjunction with a 50 litre air sample.
- 84 -
-------
5.4 EXPERIMENTAL PROCEDURE
The field programme ran for a period of 14 days. The point sample ambient monitor
was operated in the trailer at the south side of the field. Simultaneously, the
Beckman Coulometric analyzer was operated in the hut at the meteorological tower.
During this period the long path ambient monitor was first operated at a path
length of 100 meters for a four day period; at an intermediate path length of 300
meters for a further four days; and at a long path length for three days. The short
o
and intermediate pathlength experiments were carried out using a 500 watt 3400 F
quartz-iodine lamp. Attempts to increase the path length beyond 500 meters using
this light source were unsuccessful since beyond this range the energy recieved
by the spectrometer was insufficient to maintain a satisfactory signal/noise ratio.
The quartz-iodine lamp was replaced with an xenon-arc resulting in a considerable
increase in energy transmitted to the spectrometer and consequently an operating
pathlength of 930 meters was selected.
During the early part of the test, wet chemical analyses were carried out at fre-
quent intervals on a routine basis. Subsequently, wet chemical analyses were
suspended and instituted only when the monitors indicated an increase in the SO
content of the atmosphere.
5.5 RESULTS
5.5.1 Point Sample Ambient Monitor Versus Long Path Ambient Monitor
As with most urban air pollution data the results of the tests carried out at the
Evans Avenue site show long periods of relatively low pollution levels. For the
purposes of this study three periods were selected, one for each of the pathlengths,
namely 100 meters, 300 meters and 930 meters. These periods were selected to con-
tain most of the high peaks as well as periods of low pollution in order to cover
the full dynamic range of the instruments. Having confirmed that the readings of
the two monitors were time synchronized, exact readings were taken from the records
- 85 -
-------
at intervals of about 2 minutes. These values were multiplied by their respective
calibration factors and were entered into an IBM 7044 computer. The library pro-
gramme STATPK was used to draw scatter diagrams and calculate the correlation co-
efficient, regression coefficient and the intercept. In each of the studies de-
tailed in 5.5.1 below, the point sample ambient monitor was arbitrarily chosen
to be the independent variable plotted on the x axis.
RESULTS OF 100 METER PATHLENGTH TESTS
A four hour period containing a moderately sized peak was chosen for this path-
length. Figure 5.5 shows the computer generated scatter diagram. At this short
pathlength the sensitivity of the instrument is low with a calibration factor of
19.4 ppb per scale division, and the recording facility read only to the nearest
division. The statistically derived data are:
Correlation Coefficient 0.909
Regression Coefficient 0.826
Intercept 87.72
Despite the low resolution of the long path instrument at this relatively short
pathlength, the correlation is good. The regression coefficient indicates a dis-
crepancy in the calibration factors of one of the instruments. The intercept in-
dicates that the long path monitor has a zero offset of about 5 divisions (87.7/19.4
= 5 divisions).
LONG PATH - 300 METERS TESTS
An 8 hour period was selected for this correlation containing two peaks of over
200 ppb each. The readings were digitized at a rate of 11 points per hour to
cover the whole period within the 100 data point limitation of the computer.
Figure 5.6 shows the plot of both samples for a two hour period illustrating how
the two curves move in unison.
- 86 -
-------
Figure 5.7 shows the scatter diagram for this period. The 2 black squares define
the origin and the line for a unit regression coefficient. The derived data are:
Correlation Coefficient 0.970
Regression Coefficient 0.898
Intercept 60.19
The correlation at this pathlength is very good. The calibration of the point-
sampler again appears to be lower than that for the long path. The calibration
factor of the long path instrument is 6.5 ppb per division and the zero offset
at this pathlength is about 9 divisions.
LONG PATH - 930 METER TESTS
At this pathlength a xenon arc light source was substituted for the quartz iodine
lamp which had insufficient power to operate at pathlengths in excess of 500 meters.
For this pathlength, two consecutive four hour periods were selected each contain-
ing good peaks. Both were digitized at 22 points per hour. Some points were lost
during periods when the Point Sampler was being calibrated and also the resolution
of the diagram is such that two very close points are plotted as one. Figure 5.8
shows the computer generated scatter diagram for the first period which gave the
following statistical values.
Correlation Coefficient 0.951
Regression Coefficient 0.920
Intercept 18.92
The correlation coefficient is good and is in line with the coefficient for
the shorter pathlengths. The regression coefficient is good and shows some
improvement in calibration in comparison with the shorter pathlengths. The
- 87 -
-------
intercept remains at about 8 divisions.
The scatter diagram for the second 4 hour period is shown in Figure 5.9. It
can be seen that there is an increase in the scatter of the points and the
correlation coefficient is poor. The points circled represent consecutive read-
ings ranging over thirty minute periods in which the long path seems to have
registered a peak which was not detected by the point sampler. This is probably
the explanation of the 2 points (x) which were also consecutive and were the
last 2 points in the period. If the circled and x'ed points are disregarded,
the statistical results become:
Correlation Coefficient 0.095
Regression Coefficient 0.618
Intercept 40.57
The omission of the circled points result in an improvement in the correlation
coefficient to the level previously obtained with other pathlengths. But the re-
gression line is very different from that in the previous periods.
Of the three pathlengths, the best correlation is shown at 300 meters. For the
100 meter pathlength, the results are degraded by the low resolution of the instru-
ment. For the 930 meter pathlength, the correlation is good only for short periods.
- 88 -
-------
The apparently spurious points in figure 5.9 could have arisen from the presence
of sulfur dioxide clouds existing within the light path of the long path monitor
but distant from the point sampler. This variation in the quantity of sulphur
dioxide intercepted by the long path instrument but not recorded by the point
sampler results in variation in both the regression coefficient and the intercept.
Over short periods however, when the atmosphere remains homogeneous the correlation
coefficient could remain fairly high. It is expected, that this effect would be
more frequent for longer pathlengths than for the shorter pathlengths.
As a corollary to the averaging effect of the long path, it is to be expected
that readings taken at a point would reach higher and lower extremes than the in-
tegrated readings for the long path instrument. The effect of this would be to
give a lower slope and regression coefficient and higher intercept. This fact
may help to explain why in most cases the slope is less than 1. It may also
contribute to the large zero offsets recorded.
It should be noted that there is a major coal fired power generating station lo-
cated several miles from the test site and that looping of the plume from this
source could cause considerable differences between point and long path readings.
Other factors contributing to the scatter are:
(1) Distance between the point sampler and the light path of the long path
sampler.
(2) Difference in the electronic time constants of the two instruments, namely
100 seconds for the point sampler ambient monitor and 1.5 seconds for the
long path ambient monitor.
(3) Effect of zero or calibration drift in either instrument.
(4) Errors in time synchronization for selecting data points.
Within the limits of the field work all of these factors have been reduced to a
minimum and they are not considered to be significant sources of error.
- 89 -
-------
5.5.2 Point Sample Ambient Monitor Vs. Wet Chemical
In these studies the wet chemical results are arbitrarily taken as the independ-
ent variable and plotted on the x axis. Figure 5.6 is the computer generated
scatter diagram for the point sample ambient monitor and the wet chemical analyses
made in the vicinity of the point samples. The small black squares define the line
for a regression coefficient of 1. All scales are in ppb of SO . The generated
statistical data are:-
Correlation Coefficient 0.823
Regression Coefficient 0.872
Intercept 0.883
The correlation is acceptable but not good. The regression coefficient of 0.87
suggests that the point sample calibration is slightly low but the zero deviation
of less than 1 ppb is extremely good.
5.5.3 Beckman Coulometric Vs. Chemical Method
Figure 5.7 is the scatter diagram and again all values are in ppb SO . The black
squares define the line for the regression coefficient equal to 1. The point
circled was taken to be spurious and excluded from the results, since none of the
three monitors read SO levels over 100 ppb while the wet chemical test was being
performed.
These data show appreciable scatter and the correlation is poor. Under these
conditions, the regression equation and intercept probably do not accurately re-
flect differences in calibration. However, since the wet chemical measurements
were taken at a location about fifty yards from the Beckman instrument, most of
the scatter must be attributed to the physical separation of the sample points
and thus further points to the sensitivity of point samplers to their precise lo-
cation.
- 90 -
-------
The statistical data derived are:
Correlation Coefficient . 800
Regression Coefficient .800
Intercept. 9.00
5.5.4 Long Path Ambient Monitor Vs. Wet Chemical Methods
Figure 5.8 is a scatter diagram for the long path monitor versus the Wet Chemical
analysis and includes 2 types of wet chemical analyses. The circled points are
those for which the wet chemical sampler was moved between the light source and
the sensor of the long path instrument close to, but not coincident with, the light
path in order to simulate the space integrated sample. The remaining point rep-
resents stationary samples taken at the point sample ambient monitor location.
The correlation for the integrated sample is some what better than for the stat-
ionary sample. The derived coefficients are:-
Correlation Coefficient 0.892
Regression Coefficient 1.071
Intercept -0.443
There is a good correlation between these two monitoring methods. The regression
coefficient and the intercept are close to ideal. It should be noted that an al-
lowance of five divisions for the quartz iodine lamp and three divisions for the
xenon arc lamp have been made in the zero offset. These are the values derived
from the earlier study of a correlation between the point sample and long path
ambient monitor. The very small intercept obtained here confirms the correctness
of these allowances.
- 91 -
-------
5.5.5 Point Sampler Vs. Coulometric Sampler
The correlation between these two instruments was carried out for a seventeen hour
period. Since the entire set of data was too large to be stored in the computer,
every third point was read in with allowance for the 18 ppb zero offset previously
noted. The correlation for the whole period was found to be:-
Correlation coefficient 0.987
Repression coefficient 0.925
Intercept 0.20
5.5.6 Sliding Correlations
When statistical procedures are applied over long periods of time, average corre-
lations are obtained which may conceal periods of very high or very low correlations.
To investigate the change in correlation with time,, a computer programme was set up
to calculate the correlation and regression coefficients for batches of 40 points.
The computer progressively discards five early results replacing them by 5 new re-
sults to give a continuous series of coefficients.
Figure 5.10 shows the sliding correlation for the Long path (930 metres) versus the
Point Sampler. In the early part of the period, the slope approaches 1.0 and the
intercept is small. During the middle of the period the correlation coefficient
falls to below 0.3 and the intercept increases precipitously. Towards the end of
the period under review, the correlation coefficient and slope again improve.
Figure 5.11 shows the sliding correlation for the Point Sampler and the Coulometric
analyser. In this case, the slope is low for the early part of the period mainly
due to noise during a period of very low gas concentrations. As the gas concen-
tration increases, the slope goes through a maximum and settles to a value of 0.90.
- 92 -
-------
5.6 DISCUSSION
The Long Path Ambient Monitor is the first practical instrument to measure over a
long path length. At short pathlengths, the results are degraded by low signal/
noise ratios. As the path length is increased from 100m to 300m, the correlation
coefficient against the.Point Sampler Ambient Monitor increases from 0.90 to 0.97.
With furthur increases in path length, the sensitivity is increased but the poss-
ibility of intercepting additional plumes is increased and at 930 metres the cor-
relation coefficient against the PSAM is 0.95 for one four hour period and below
0.6 for a second four hour period. However, by discarding the data contained in
one half-hour period, the correlation coefficient is increased to 0.91. For a path
length of 930m, the Long Path Ambient Monitor is thus more representative of an area
than the Point Sampler.
Correlation between the PSAM and the Coulometric analyser was excellent at 0.987.
Corration between the correlation instruments and the wet chemical tests averaged
0.85 while correlation between the coulometric analyser and the wet chemical method
fell to 0.80 due in past to the distance between the coulometric and the wet chemical
sampling which was carried out in the PSAM trailer.
5.7 CONCLUSION
It is concluded that the Long Path and Point Source Ambient Monitors gives a rel-
iable estimate of the SO concentration but that the Long Path configuration using
a 1000m path length can give poor correlation with the Point Sampler results. -
While the PSAM measures the concentration at a point, the Long Path Monitor gives
a better indication of gas concentrations in a given area.
- 93 -
-------
LJ: LJ
A- III
FIG. 5-1.
Torque Motor
\ Window
Entrance Correlation Mask
Iodine Lamp
Diffraction GratingX Windows"" /Phototube
Collimator Exit Correlation Mask
|. 48 ins. J
Torque
'Motor
^Grating Oscillation Absorption Cell
^Grating / Preamplifier
FIG. 5-2.
SO2 AMBIENT MONITOR SCHEMATIC
- 94 -
-------
^ Avifitv
/J//-1.
, .
F XGUH£) 3 3
- 95 -
ELECTRONIC BLOCK (-PSAM)
-------
- 96 -
-------
SCATTER DIAGRAM
14.9863
197.8800 *
,
V
A
R
176.5400 +
I
A
B
L 155.2000 +
E
VARIABLE 3
38.5042 62.0221 85.5400 109.0579 132.5758 156.0937
X -X X X .
f 197.8800
XXX X
V
A
R
+ 176.5400
X X XX XXX XX X I
A
B
X X XXXX XXXX 155.2000 L
E
133.8600
X XX XXXX X XXX XX X
133.8600
. XX XX XXXXXX XXX X
112.5200
112.5200
. X
91. 1800 + . .
14.9863
38.5042
. . + +M. ... + + + 91.1800
62.0221 85.5
-------
' -5.7143
236.2500 + . .
SCATTER DIAGRAM
VARIABLE 1
29.5238 64.7619 100.0000 135.2381 170.4762 205.7143
. + + + + + + 236.2500
f '
195.0000 +
. V
. A
1 "
, . . 153.7500 +
. I .
: A :
. B .
. L 112.5000 +
. E
.2 .
71.2500 +
. i .
X X
30.0000 +
XX
X X
XX X
XX X
. XX
-11.2500+ + .". . .
-5.7143 29.5238
SCALING FOR VARIABLE
SCALING FOR VARIABLE
X
+ 195.0000
- X
+ 153.7500
X
+ 112. soon
X X
x
X
X .
X
+ 71.2500
+ 30.0000
X
. + + -f + + -11.2500
64.7619 100.0000 135.2381 170.4762 205.7143
VARIABLE 1 .
INTERVAL SPACING MINIMUM MAXIMUM
1 > 5.873016 0. 200.0000
2 5.892857 0. 225.0000
V
A
R
I
A
S
L
E
2
FIGURE 5.6
- 98 -
-------
SCATTER DIAGRAM
-8.7023
;
t
276.4580 +
.
V
.
A
R
217.9765 +
I
A
*
B
i
L 159.4950 + ,
E
VARIABLE 1
44.9618 98.6259 152.2900 205.9541 259.6182 313.2323
+ ....* + + + + 334.9395
0
0 Op
0 0
00
0 00 + 276 .4580
0
0
0 0
0
0 0
00 + 217.9765
0
0 0 00 00
O
t
0 0
00 0 0 + 159.4950
0 .
V
A
R
I
A
P
L
E
101.0135 +
00
00
. 0000
. 0000
0
42.5320 + 00
0
101.0135
+ 42.5320
-15.9495 + + + + + + + -15.9495
-8.7023 44.9613 93.6259 152.2900 205.9541 259.C:132 313.2323
VARIABLE 1
SCALING FOR VARIABLE 1 --
SCALING FOR VARIABLE 2 --
INTERVAL SPACING
8. 944 OK,
8.354500
MINIMUM
o.
-o.
MAXIMUM
304.5800
31S.9900
FIGURE 5.7
- 99 -
-------
SCATTER DIAGRAM
VARIABLE 2
-6.0000 31.0000 68.0000 105,0000 142.0000 170.0000 216.0000
220.5000 + + + + + + + 220.5000
182.0000 +
V
A
P.
143.5000 +
I
A !
B
L 105.0000 +
E
X XX
X X XX X X
XX X
XX X
X
X X
X
X
XX X
X X
X X
+ 143.500"
+ 182.0000
V
A
I
A
B
+ 105.0000 L
E
66.5000 +
28.0000 +
X X
XXX X X
XXXXX X
XX
X XXXX X
X
+ C6.5000
28.0000
10.5000+ + j- + .+ + + -in. sono
-6.0000 31.0000 68.0000 105.0000 142.0000 179.0000 216.1000
VARIABLE 2
SCALING FOR VARIABLE 2 --
SCALING FOR VARIABLE 1
INTERVAL SPACING MINIMUM MAXIMUM
6.166667 0. 210.0000
5.500000 0. 210.0000
FIGURE 5.8
- 100 -
-------
C <\ T T F P D I A C ?. A '1
157.5010+ ..... +
" A P. ! A- T, I. f!
75.nnnn
+'
127.8571
130.nnnn
v
A
. 5nnn
on o
n no
i 3 n _ >. n p <.
75. nnnn
^7.5000
o n
n o o
o fi n
on o
n o o
o n
o on
no o
o on
0 0 00
00 0
7 5 . "> " °"1 I.
2n.nonn
,f- n. L r 2
127. "571
SCAl.TfT FOP. VAP'APl.n ?
scAi.it:' por "An i A.m.:': i
7.nrr.s7i
FIGURE 5.9
1; n . n n n
' n . i n n
- 101 -
-------
FIGURE 5.10
SLIDING CORRELATION
too
UJ
o
K
UJ
50
0 J
LPAM VERSUS PSAM
*
o. .5
o
eft
0^
.5 -
o:
(T UJ
oo -
oo
.1-
300-
CM
9 200 H
JO
o.
Q.
100 -
50 100
N° OF READINGS
- 102 -
150
-------
*s
o
I.I
1.0
ui -9
0.
3 .8
CO
.7
.6
15
; 10
»
5
0
\ 200
M
O
CO
100 -
SLIDING REGRESSION
STUDIES
POINT SAMPLE VERSUS
BECKMAN
BECKMAN 906
POINT SAMPLE MONITOR
1 1
0 0
CVJ ^-
i
O
(0
1
0
CD
i
O
o
1
o
CVJ
^ m~ * *^. & m*
1
O
^
^ ^^ t 4*. 4* *L
1
O
(O
. \
1
o
CO
1
o
0
CM
1
o
CM
FIGURE 5.11 SUDIN6 CORRELATION-
READINGS (22/hr)
PSAM VERSUS BECKMAN ANALYSIS
- 103 -
-------
Section 6
PHASE IV
6.1 INTRODUCTION
The broad object of Phase IV was to study the various configurations of the
correlation spectrometer for the measurement of N0? to give data unobtainable
by classical means. These tests were carried out in the Los Angeles basin ar<;
of southern California where at certain times during the year high levels of
ambient NO can be confidently expected. The tests were divided into three
parts:
(a) measurement in an average urban location
(b) comparison of measurements on and off a freeway
(c) repeated passes by the airborne spectrometer along a
wind trajectory to demonstrate the movement of the
NO peak inland during the morning and outwards during
the afternoon and the influence of land-sea breeze.
The quantity and nature of the equipment needed for these tests indicated that
packing and shipping by normal carriers would involve a considerable amount of
work. Consequently the necessary benches and facilities were constructed in
an air conditioned truck, which carried the electronics, the recorder, and the
water bath containing the permeation tubes for calibration purposes. The point
sample ambient monitor was mounted under a bench suspended on elastic cords.
6.2 PHASE IV (A) OBJECTIVE
The objective of the Phase IV (A) experiments were to repeat the comparison
between the long path and point sample ambient monitors (as in Phase III) for
- 104 -
-------
measurements of nitrogen dioxide in a photochemical pollutant environment.
Simultaneously with the ground work addition airborne measurements were to
be made at various altitudes, directly over the ground test location. In
this manner, studies could be made of both the horizontal and the vertical
variations in the NO concentration.
6.3 LOCATION
During a preliminary visit, Brackett Field was selected for the test site
because:
(1) The airport offered good facilities for flying and for ground support
of the aircraft.
(2) The aircraft and flight team when grounded, were available for other
duties.
(3) The airport approach patterns enabled the aircraft to fly over the
monitors at low levels.
(4) The airport markings and radio control helped the pilot to locate
and overfly the monitors.
(5) There are few open spaces where 4000 ft. line of sight survey can
be set up in the Los Angeles area.
Brackett Field is located north of the San Bernadino Freeway at Pomona and
about 4 miles from an APC District monitoring laboratory in northern Pomona
The runway is in excess of 4000 ft. and is marked by a Lake
at the western end and the open expanse of the Los Angeles Fair Ground at the
eastern end.
- 105 -
-------
6.4 INSTRUMENTATION
The point sample and long path ambient monitors used for the Phase III tests
were converted to N02 operation, for the field work at Brackett Field,
California. Fortuitously, the local District maintains a monitoring laboratory
in North Pomona that has a Beckman NO/NO colorimetric analyzer based on the -
Saltzmann reaction. In this instrument the air stream is reacted with
Saltzmanns reagent to remote the NO , then bubbled through potassium permanganate
solution to convert the NO to NO followed by a further treatment with
Saltzmann's reagent.
6.5 EXPERIMENTAL PROCEDURE
The point sampler, installed in the truck and located at the base of the control
tower was run continuously. (Figure 6.1). The instrument showed
considerable sensitivity to ambient temperature and inspite of the efforts of
the air conditioner, the temperature in the body of the truck sitting unshaded
in the California sun showed diurnal variations of 25 F. The zeroing circuit
functioned satisfactorily and automatically re-established the zero at 2 hourly
intervals. However, under the worst temperature fluctuations, considerable
zero shift was evident during this period, and at times, re-zeroing was
carried out hourly. The instrument was subject to dynamic calibration about once
a week and initially a static calibration using the standard gas cell was carried
out automatically every 2 hours in the re-zeroing cycle. After 10 days, it
was established that the calibration was sufficiently stable to permit manual
calibration only once per day and the loss of data occuring during the 2 hourly
re-zero/calibrate cycle was considerably reduced. Apart from the temperature
sensitivity this instrument functioned perfectly throughout the course of the
experiment.
The long path ambient monitor was set up on a pathlength of 1400 meters,
the point sampler monitor being located half-way along the path and just off to
one side. The system operated in a satisfactory manner; no problems were found
in the two areas of greatest concern, namely maintaining the alignment of
the sensor on the light source and spurious signals from fluctuation in the
- 106 -
-------
power voltage and frequency derived from a 6Kw motor generator set. The
only source of interference was found to be the breakthrough of sunlight
in spite of the instrument's intended rejection of unchopped light. This
resulted in high spurious readings when the sun came within a solid angle of
about 45 to the optical path, subtended at the instrument.
Wet chemical analyses were carried out manually both at the location of the
point sampler and along the optical path of the LPAM. The time interval between
determination was variable but almost continuous determinations were carried
out over long periods during high gas concentrations. With the LPAM and light
source both located at the edge of the taxi strip it was possible to collect
integrated samples from a path only a few feet from the light path, care
being taken not to interrupting the light path.
The airborne work was conducted out of Brackett Field. Initially some ten days
of survey was lost because a mounting in the airborne spectrometer had worked
loose. This loss was of no importance since the N0» levels were low and seldom
exceeded 30 ppb. The operation of the aircraft was severely restricted by the
smog, especially during days of high NO readings. Visibility just after daybreak
in the vicinity of Pomona varied from less than one mile for high smog days to
about 3 miles on low smog days. Take offs and landings are permitted only with
visibilities in excess of 2 miles (Special V.F.R. Conditions) and take offs
were thus delayed sometimes until just before noon during heavy smog. The
intention to follow the buildup of smog from the early morning conditions was
therefore abandoned and most flights were confined to the period of 10:30 a.m.
to 4:00 p.m.
- 107 -
-------
6.6 RESULTS
We were fortunate to have two periods of high pollution levels during our stay in
the Los Angeles area. During the first week only low levels of NO were recorded
(below 30 pphm) and the District station also showed low values. However, the
district station consistently recorded values 50 to 100% higher than our values
and we were promoted to make careful calibration of both our own and the district
instruments to establish the reason for the difference in readings. During the
second week a series of medium to high smog episodes occured giving rise to NO
concentrations over 100 pphm on most days. The third week was characterized by a
period of medium to low smog levels followed by medium to high levels during the
last week when the other phases of the program were being run.
6.6.1 LAPCD MONITOR
During the first week of our operation only low levels of pollution were recorded
but there appeared to be a consistent differences between the Barringer Point
Sample Ambient Monitor readings at Brackett Field and the LAPCD readings using
the Beckman instrument at the district office. It was decided therefore, to
calibrate both instruments against a common wet. chemical method. Sample bags
of 10 cu. ft. capacity were fabricated by sticking together a number of polyethylene
bags using scotch tape. With care, it was possible to make gas tight joints and
produce bags which could be inflated by means of a small plastic pump to hold
a pressure of 1" water gage for a period of several minutes. .The inflated bag
was injected with sufficient NO from a gas bottle to yield a calculated concentra-
tion of 300-400 ppb. This sample was sufficient to obtain simultaneous readings
by both the automatic analyzer and the wet chemical method. Since the point
sample ambient monitor had been routinely calibrated against the wet chemical
method, no discrepancy was expected or uncovered during this calibration.
The LAPCD analyzer is a Beckman N0_ colorimetric system based on absorbing the
NO in Saltzmann reagent, oxidizing the NO to NO by potassium permanganate and
determing the additional NO by further reaction with the reagent. Two calibration
- 108 -
-------
runs carried out at 400 ppb showed that the NO reading derived by the
Beckman instrument agreed with the wet chemical determination within 4%. It
was further noted that the Beckman instrument recorded a nitric oxide concen-
tration of approximately 20 ppb and it was first suspected that this was due
to a low efficiency of the N02 absorber which allowed unreacted NO to pass
over into the NO oxidiser and analyzer. However, the indication disappeared
when the permanganate oxidizer was bypassed, confirming the presence of approxi-
mately 7% of NO in the NO bottle. These calibration tests showed that the
Beckman instrument agrees with the wet chemical method to within 5%.
Consequently, since both instruments agreed with the wet chemical method to
within acceptable limits, it is reasonable to assume that these differences were
real and caused by true differences between the two locations.
6.6.2 Point Sample Ambient Monitor
Apart from the sensitivity to ambient temperature changes, the PSAM functioned
well. The sensitivity (span control) proved to be free from drift so that a
manual check once a day was sufficient. The wide fluctuations in temperature
in the truck caused the base line to drift as much as 20% full scale deflection
in two hours, and since the drift is linear, base line errors of an apprecia-
ble size are possible. A section of the chart with a relatively stable baseline
is shown in Figure 6.2 and illustrates the short-term variation in concentration '
This instrument has a time constant of less than two minutes and gives considerably
more information of short-term concentration fluctuations than the Beckman
instrument with a time constant exceeding 20 minutes.
6.6.3 Long Path Ambient Monitor
Figure 6.3 is a typical chart for a period of high NO, and shows the double de-
flection on either side of the undetermined zero caused by the slow scan of the
grating through correlation to anti-correlation. The actual gas concentration
is proportioned to one half the difference between the maximum and minimum values.
- 109 -
-------
This slow sweep has the effect of creating a series of discrete determinations
at about two minute intervals in spite of the very fast t4.me constant of the
instrument. Fig. 6.4 shows for comparison, a period of very low NO .
Figure 6.5 shows the trace during a period of breakthrough of
sunlight. Under these conditions, output voltages in excess of 10 volts
and resenting concentrations in excess of 100 ppm have been recorded. Since
the direct sunlight does not enter the field of view of the instrument, the
light is presumably directed into the entrance slit by forward scattering. The
effect however, is unlike that of dilution discussed in connection with the use
of the remote sensor for stack measurements. In that case, the total light flux
is enhanced, but the amount of correlation is changed to a lesser extent (if at
all) and a lower percentage of correlation is recorded. In the case of sunlight
breakthrough, enhanced correlation apparently takes place. It is well established
that forward scattered light is strongly plane polarized and consideration is being
given to the use of polarizing filters to further discriminate against natural
light.
6.6.4 Airborne Instrument
The object of the airborne determinations was to determine the variations in
NO burden along a flight line over the airport landing strip and hence over the
ground based instruments. Repeated flights were planned in order to study the
change in vertical burden with time. Both upward and downward looking measurements
were made, as the aircraft altitude was decreased so that vertical sections of
the atmosphere were in effect analyzed. These flights, numbers 10, 11 & 12
of July 3rd, 4th and 7th were selected for detailed examination.
6.6.4.1 Flight 10
Figure 5.6 shows a composite of the runs of Flight 10, carried out on July 3rd
under conditions of moderate smog. Runs were made at 5,5000 , 3,500 and 2,500
ft. altitudes. Figure 6.6 shows that there is a variation in vertical burden
along the flight line; examination of successive runs shows no evidence that the
pattern in repeated either in relation to the ground or to the wind. Table B.I
- 110 -
-------
in Appendix B shows the details of the runs and the most important conclusion
is the scatter in the readings obtained for upward and downward measurements.
In the case of the upward-looking measurements, it can be assumed that for
every photon scattered out of the field of view of the telescope, another photon
is scattered into the field of view and that the average pathlength through the
target gas is substantially unaffected. The upward looking mode is therefore less
sensitive to scatter than the downward looking mode. The data contained in
Table B.I allows the vertical profile to be estimated both from the up-Jooking
modes and it is obvious that the downward looking mode suffers most from scatter.
Thus, at an altitude of 1500 ft. (curve 9), the vertical burden is 140 ppm-m.
Assuming a 'm1 value of 2.4, the indicated levels are 58 ppm-m or an average of
38 pphm. The values for downward observations from high altitudes are less than
this value in spite of the additional gas indicated by the 40 ppm-m recorded in the
upward looking mode at the lower altitude.
6.6.4.2 Flight 11
Figure 6.7 is a composite for the runs of flight 11 carried out on June 4/69. Of
the twelve runs completed in this flight, 10 were considered to be sufficiently
defined for digitization and computer plotting. The details of these 10 flights
at altitudes of 2000 to 6000 ft. above mean sea level are detailed in Tables
B.2 and Table B.3 summarizes the average values for the various traverses. In
Table B.3, columns 2 and 3 shows the downward and upward looking values at various
heights above the ground. The dilution of the signal by scatter is immediately
obvious from column 2, in which the downward integral porgressively decreases as the
height of the aircraft increases. Column 3 shows that the gas recorded by the up-
ward looking sensor at least increases progressively as the amount of gas above
the aircraft increased. Column 4 shows the difference between the upward and
downward mode and.indicates that within the limits of experimental error, the net
downward value is constant irrespective of height. It must be assumed therefore.,
that the spectrometer in the downward looking mode sees virtually no light reflected
from the ground but records only a signal scattered back from below the aircraft.
- Ill -
-------
Since the scattered signal is constant down to 1000 ft. above the ground, the majority
of the back scatter originates from a region not more than 1000 ft. below the
aircraft.
6.6.4.3 Flight 12
Run 12 was carried out on July 7/69 from 4:20 to 5:30 pm, a day of moderate smog
which is confirmed by the profiles which suffer less back scatter than those
in run 11. Six runs were carried out at 6500 ft. and 4 runs at 3000 ft. and a
comparison of these repeat runs suggests that the total burden available was
undergoing considerable fluctuation. (Figure 6.8) Thus, runs 2, 4, 5 and 6
represent repeated traverses at 6500 ft. looking downwards. The average burden
recorded for the 4 runs were 38, 42, 55 and 51 ppm-m respectively. Similarly,
there is an increase from 77 to 89 ppm-m average between runs 8 and 9 looking
downward at 3000 ft. and a change from 18 to 29 ppm-m for runs 7 and 10 at
3000 ft. looking upward. The values of 30 to 43 ppm-m recorded at 5500 ft..
looking downwards maybe contrasted with a value of 25 ppm-m recorded at 5000 ft.
during flight 11. Similarly the 18 to 29 ppm-m values for 20QO ft. upward looking
maybe contrasted with values of 65 to 75 ppm-m for a similar position during flight
11. Thus, the upward looking mode records lower gas levels, but the downward
looking mode records higher gas levels as a result of the lower dilution of the
signal by reduced scattering under moderate smog conditions. Tables B.4 and B.5
summarize the average values for the various traverses.
6.6.5 Point Sampler Vs Long Path Sampler
The correlation between the point sampler and the long path sampler has been exam-
ined in detail for two seven hour periods during the nights of June 28/29 and
June 29/30. These were periods of heavy smog with almost calm conditions under a
radiation inversion and being during the hours of darkness, the effect of break-
through of sunlight or skylight was obviated on the long path monitor. Fig. 5.9
shows direct comparisons of readings from the two instruments and fig.6.10 shows
the result of a sliding correlation carried out for the second period. It will be
seen from figure 6.9 that the long path monitor, while following in general the
fluctuations of the point sample ambient monitor, is subject to minor fluctuations,
indicating that the point sampler reading is not often
- 112 -
-------
representative of the average for the whole long path and consequently is even
less representative of the readings at other points along the long path. The
maximum correlation coefficient is 0.92 but for the majority of the period the
correlation coefficient is less than 0.7. Under these conditions, the intercept
and the slope which are shown for periods when the correlation coefficient is
greater than 0«7 have little meaning. These results are consistent with impressions
gained on the site that higher gas concentrations often existed at the west end
of the runway than at the control tower or the east end of the runway, and that the
point sample values tended to be lower than the long path results. It was also
noticed when taking integrated wet chemical samples that significant changes in
reagent color took place at the west end of the runway while little change in color
was seen during the past of the traverses to the east end of the runway. These
indications are confirmed by a comparison of the values for wet chemical tests
carried out at various points along the runway and for integrated wet chemical
samples.
- 113 -
-------
FIGURE 6-2
PSAM CHART 2000 MRS. JUNE 30/69
- 114 -
-------
" ~
FIGURE 6-3
LONG PATH MONITOR RECORD
HIGH GAS
BAUSCH . LOMB V
10r
FIGURE 6-4
LONG PATH RECORD LOW GAS
- 115 -
-------
A awcn >r HDsnoa
FIGURE 6-5
lE-IO-iC 'ON 1VD
LONG PATH MONITOR
SUN BREAK-THROUGH
RUN 9
-150
HOO
Z
i
2
0.
0.
50
FIG 6.6 AIR BORNE NOz PROFILES FLIGHT 10
- 116 -
-------
-150
RUN 12
FIG 6.7 AIR BORNE N02 PROFILES FLIGHT I
RUN 9
-150
HOO
2
a.
a
^ 50
FIG 6.8 AIR BORNE NOz PROFILES FLIGHT 12
- 117 -
-------
N02
(ppb)
N02
(ppb)
300 -,
200 -
100
2hOO
28/6/69
300 i
200 -
100
LPAM VERSUS PSAM
FIG G.9
50 100
N° OF READINGS (31/hr)
150 04:00
29/6/69
LONG PATH
MONITOR
POINT SAMPLE
MONITOR
I8<00
29/6/69
50 100
N° OF READINGS (31/hr)
150 01=00
30/6/69
- 118 -
-------
Z] 2001
O
I00^
INTERCEPT (ppb)
0
o
m
^
o
z
0 J
1.0-
.8-
.6-
.4-
.2-
SLOPE
m
"D
CO
1.0-
.9-
.8 -
.7 -
.6 -
.5-
I
3001
CORRELATION
200-
N02
PPb
IQO-
50 100
N° OF READINGS
150
-------
Section 7
PHASE IV B
7.1 OBJECTIVES
The primary objective of this phase of the tests was to compare the concentration
of nitrogen dioxide existing on a Freeway with the concentration existing in an
area removed from the Freeway traffic. To accomplish this a site was selected
\
where the long path spectrometer could be placed on the shoulder of a Freeway to
define optical paths of about one kilometer in length, either along the Freeway
to one light source or normal to the Freeway to a second light source. Additional
experiments involved the continued operation of the Point Source Ambient Monitor
on the shoulder of the Freeway, the collection of both stationary and spatially
integrated wet chemical samples along the freeway and traverses across the Free-
way by the airborne monitor.
7.2 LOCATION
The site selected was a stretch of the San Bernadino Freeway between Covington and
Pomona. The Freeway runs eastward down a hill towards the San Dimas Overpass,
curving slightly to the left. The Kellogg Campus of Cal State Polytechnical
Institute lies in a valley to the south. With excellent co-operation from the
'Cal-Poly' staff, the spectrometer was set up on the south side of the highway .
in a lemon grove on the campus. The first light source was set up against the
masonry of the San Dimas bridge so that the optical path was approximately
tangential to the inside of the curve and made a double pass across all six lanes
above the traffic. The second light was set up on a hill on the far side of the
Campus overlooking the music department. The spectrometer was set up near a
sign board, which hid the activity from traffic approaching on the near side.
(The police were perturbed that if the activity was too obtrusive, slowdowns
and traffic jams would result).
Figure 7.1 shows the spectrometer located about seven feet off the ground and
adjusted to swing horizontally from one lamp to the other. Figure 7.2 shows
- 120 -
-------
the approximate location.of the optical paths as seen from the San Dimas Overpass
The path length"on the freeway" was 1200 m. and "off the freeway" 900 m.
7.3 INSTRUMENTS
The Long Path monitor was set up on the Campus as described above. The truck
housing the Point Source monitor was parked alongside the monitor.
The aircraft continued to operate out of Brackett Field which was only a mile or
so north of the Freeway. In fact, the downwind leg of the Brackett landing pattern
normally passes over this section of the Freeway and the over flights by the air-
borne spectrometer had to be adapted and become part of simulated landing circuits.
7.4 EXPERIMENTAL PROCEDURE
In order to compare 'along the Freeway1 readings with the 'off the Freeway1
readings the long path spectrometer was turned manually from the San Dimas
overpass light source to the campus light source at 15 minute intervals. For
the first two days this procedure was carried on during daylight hours only and
the spectrometer was left unattended looking at the campus light source during
the hours of darkness. This was in deference to a promise made to the Freeway
Patrol that the San Dimas light source which was situated about 25 feet to the
right of the Freeway would not be operated during the hours of darkness. This
was made in anticipation that the bright light shining directly into the eyes
of drivers of on-coming traffic would create a hazardous situation after dark.
Under actual operating conditions the collimation of the light source was better
than expected and as these appeared to be no hazard an attempt was made to run
the spectrometer looking along the Freeway during the third night. However,
complaints from motorists during the night prevented further operation of the
San Dimas light during the hours of darkness.
No particular problems were encountered in the wet chemical tests, other than the
danger of traversing along the shoulder of the freeway at very slow speed.
- 121 -
-------
Aircraft flights were made whenever visibility conditions permitted. Some
interference in flight plans were experienced due to the fact that our flights
along the Freeway, especially at 2000 and 3000 feet encroached on the downwind
leg of the Brackett Field landing pattern. This was overcome by reporting into
the Brackett Field landing pattern and flying a little off course to complete the
scan along the Freeway.
In order to extend the analytical data, instantaneous grab samples of 2 cubic
foot size were collected in the large plastic bags which were used for the prep-
aration of low concentration calibration samples. By standing on the shoulder of
the Freeway, passing traffic was found to create sufficient draft to inflate the
plastic bags which were quickly gathered at the neck and sealed. These bags
were transported to the truck for analysis by the wet chemical technique. The
point sample ambient monitor had a much faster response, so that
analyses could be completed quicker with a smaller sample by using the point
sample monitor than by using the wet chemical technique. Subsequent analyses of
grab samples were all made on the P.S.A.M.
7.5 RESULTS
One of the prime aims of this series of tests was to compare the levels of NO
^
existing over the freeway with values taken to one side. Figure 7.3 shows a plot
of the output of the long path ambient monitor for the period from 1100 to 1500
hours on July 9th during which the sensor was alternately sited on and off the
freeway. It will be seen that during the period from 1100 to 12:30 hours the
value off the freeway was approximately 100 ppb and the value over the freeway
500 ppb. Just after 1300 hours there is evidence of an influx gas which brings
the readings to 450 and 750 ppb respectively. This peak coincided with a minor
accident which blocked the west bound lane of the freeway for a period of about
15 minutes and resulted in a major portion of the light path passing over
stationary traffic, bumper-to-bumper. Around 1400 hours there is evidence of a
ventilation with fresh air coming over the freeway and a corresponding drop in
N0_ levels to 400 ppb while the off the freeway values increased to 500 ppb.
Figure 7.4which may be used for an overlay to Figure 7.3 shows the record of the .
point sample ambient monitor which was located just off the shoulder' of the
- 122 -
-------
freeway near the long path sensor. There appears to be a slight error of time
synchronization and figure 7.4 has been moved so that it can be placed over
figure 7.3 to show the close similarity in the various peaks recorded on the
point sampler compared with peaks picked up by the long path sampler.
Also shown on figure 7.4 are three wet chemical determinations, 2 stationary samples
taken at the truck and one integrated sample on the freeway. The first stationary
. '
sample agrees with the long path off the freeway, but it is rather lower than the
point sample reading. The second stationary sample agrees with both the point
sample and the long path off the freeway reading. The third integrated sample
agrees with the point sample but is not in agreement with either of the long path
readings unless the long path reading over the freeway which has declined drastically
from 1320 to 1420 has continued its drop in spite of the increase in values off the
freeway.
Table 7.1 shows the analyses on 11 grab samples taken in 2 cu. ft. plastic bags at
various points along the shoulder of the Freeway and on the Cal Poly campus, the
first seven values being plotted in figure 7.3. The sample locations are indicated
in figure 7.5 together with the optical paths used for the LPAM. Sample II was
high and could have been influenced by the proximity of the generator to the "on-
freeway" light source, while Sample IV shows build up of gas under a high embank-
ment on the north side of the freeway. The campus samples, taken in two pairs at
the library and the Engineering building at 1520 and 1540 hours respectively, show
average values of 47 and 95 ppb for the two sample times. The LPAM readings for
these times are not directly comparable since the instrument was aligned with the
"on the freeway" light source, but were 110 and 250 ppb respectively.
7.5.1 AIRBORNE RESULTS
The airborne results were very disappointing. In the absence of the flight
recovery camera, the precise location of the aircraft is unknown and the limit-
ations on aircraft location and maneouvers in the landing circuit distracting the
pilots attention probably resulted in some error in navigation which can not now
- 123 -
-------
be discovered or allowed for.
Since the Freeway runs along the side of a hill, the actual elevation of the
aircraft above the ground changes rapidly even when in level flight. At the
same time, the prevailing SW wind carries the gas from the Freeway up against
the hill,, so that the long optical path length of the airborne monitor when flying
over a valley coincides with low gas concentrations, while the shorter path length
over the hill corresponds with high gas concentrations. These variables, working
inversely to one another, tend to yield a fairly constant and featureless
"concentration times path length" parameter from which the location of the
aircraft cannot be precisely determined.
Figure 7.6 is an aggregate of a number of runs across the freeway and it is clear
that there is no feature which is common even some of the runs, which might be
related to known topography and gas sources.
7.6 CONCLUSION
The value of the Long Path Monitor to measure gas concentrations, in inaccessible
places, i.e. over busy roads, and to show the differences in gas concentrations
recorded over the Freeway and over adjacent areas is clearly demonstrated. The
value of the data given here is somewhat degraded by the lack of two sensors and
simultaneous measurements. The differences however, are clearly demonstrated.
The variation in the concentrations of the grab samples is large. However, limited
comparisons with the long path measurements suggests that the long path average
does not necessarily differ significantly from the grab samples which represent
a build up of gas against the side of the hill. These high values appeared to
correlate, by visual observation, with the passage up the hill of several large
diesel trucks in quick succession.
The airborne sensor data yields nothing of value. The absence of the flight
recovery data prevents the various flights from being superimposed with
confidence. The flight data is devoid of features from which notations of
location written on the chart can be verified.
- 124 -
-------
The long path monitor has shown itself useful in measuring average concentrations
in inaccessible areas. The data collected shows that during calm conditions the
concentration over a freeway may be higher than that over the surrounding area
but there is also evidence that with a cross wind, the gas will be blown from the
freeway into the neighbouring area. In this particular location, with the
'increasing' elevation of the land coinciding with increasing concentrations of
NO , and in the absence of flight recovery photography it is impossible to locate
the flight path with sufficient confidence to draw any conclusions.
- 125 -
-------
TABLE 7.1
Sample No. Analytical Method ppb NCL
1 Wet Chemical 130
2 PSAM 442
3 " 275
4 " 520
5 " 208
6 " 212
7 " 165
8 " 39
9 " 100
10 " 55
11 " 90
- 126 -
-------
LOCATION OF THE
'OFF THE FREEWAY'
LIGHT SOURCE
LOCATION OF THE
LPAM SENSOR
FIG 7 2 VIEW OF THE SAN BERNARDINO FREEWAY
FROM THE SAN DIMAS BRIDGE
- 127 -
-------
.a
a. -
M
O
700-
600-
500-
400-
300-
200-
100 -
FIG 7.3
LONG PATH SENSOR
ON AND OFF FREEWAY
READINGS FOR A FOUR
HOUR PERIOD
* i > n ON FWY
P.L. I200M
.... OFF FWY
P.L. 900 M
0
11.00
12.00
'2
13.00
TIME
PSAM
WET CHEMISTRY
(S) STATIONERY
(I) INTEGRATED
PLASTIC BAG GRAB SAMPLE
!4.00(Hrs.)
N02
(ppb)
400
300
200
100
(NOON)
IV.O
II. O
VO
O VII.
GRAB I. O
1.00 12.00 13.00 14.00 (Hrs) (OVERLAP)
FIG 7.4 AMBIENT MONITOR ON FREEWAY SHOULDER
-------
MEES AND BUILDING
MINISTRATION BUILDING
CLASSROOM BUILDING
.EsCE BUILDING
ASS3OOM BUILDING
SiNESS BUILDING
CLASSROOM ADDITION*
SNEERING CENTER
TAl PROCESSES LAB
ME, MPE ENGINEERING LAB
NTRAL ENGINEERING LAB
RCSPACE/iND ENG. BIDG.
ND TUNNEL
SA5Y BUILDING
S HORTICULTURE UNIT
jiCENCE HALL NO 1 ENCINITAS
SiOENCE HALL NO 2 MONTECITO
SIDENCE HALL NO 3 ALAMITOS
S'CENCE HALL NO. 4 ALISO
Ji C EUILDING
EECH/DRAMA BUILDING
E STABLES
UiT UNIT
;RONOMY UNIT
'iCULTURE UNIT
JLTRY UNIT/POULTRY HOUSES
F UNIT/FEED SHED
3 MILL
AT LAB
/;NE UNIT/SHELTERS
IEEP/WOOL UNIT ,
YSICAL EDUCATION BLDG
MNASIUM
DIMMING POOLS
ENGINEERING BUILDING
ALTH SERVICES BUILDING
ENGINEERING TRACTOR SHOP
SIDENCE HALL NO. 5 PALMITAS
SIDENCE HALL NO. 6 CEDRITOS
CIENEGA
STORAGE
OLIVOS COMMONS
YSICAL PLANT OFFICE
CEIVING/WAREHOUSE BLDG.
PlEX BUILDING
J'NG HALL
S.NOR HOUSE
LLOGG HALL
IEST HOUSE
OW RING
IDER CONSTRUCTION
RKING AREAS
ADMINISTRATION
FACULTY-STAFF
FACULTY-STAFF
STUDENT
COIN GATE
STUDENT
STUDENT
STUDENT
PARKING METERS
VISITOR
FACULTY-STAFF
STUCENT
t»r.i Tv.sTtsc. <;Tiir,FNT
LPAM 8 PS AM
SENSORS
ON FREEWAY'
LIGHT
'OFF FREEWAY
LIGHT
CAT IFORMA
FIG 7.5
LAYOUT OF
FREEWAY SITE
#
GRAB SAMPLES (T)- (IT
POLYTECHNIC
COLLLG1-:
-------
FIG 76
AIRCRAFT RUNS ACROSS FREEWAY
- 130
JULY 10,1969
-------
Section 8
PHASE IV C
8.1 OBJECTIVES
This ancillary experiment on horizontal profiling of nitrogen dioxide concentration
at a constant altitude involved flying the airborne spectrometer repeatedly along a
selected freeway approximately coincident with a wind trajectory. The objectives of
the flights were to observe variations in the total gas burden with photo chemical
activity and with the diurnal movements of a high NO concentration parcel of air
inland and back under the influence of the land-sea breeze.
8.2 INSTRUMENTATION
As in Phase II, only the airborne correlation spectrometer was involved in these
tests. No surface measurements were made but data derived from LAPCD monitoring
*
stations can be used to compare the ground level concentration variations with the
total burden as recorded by the airborne instrument.
8.3 EXPERIMENTAL PROCEDURES
When this experiment was planned it was hoped that the aircraft could take off be-
fore 9:00 o'clock in the morning and make a first run during the morning rush hour.
' )
As with flights from Brackett Field, visibility below 2 miles was normally exper-
ienced until 10 or 11 a.m. which precluded-the take-off of the aircraft even under
special VFR conditions. A first attempt to carry out the Phase IV C flight was
made on July 5th but was cancelled shortly after take off from Riverside airport
on encountering low cloud over the Chino Hills.
A second attempt on July 8th was delayed by heavy haze in the morning and was
abandoned after two round trips due to low cloud over the hills. The last attempt
was made on July llth, which got off to a good start and yielded three round trips
during the morning. The proposed continuation after lunch had to be cancelled due to
- 131 -
-------
build up of cloud which developed into a severe and unseasonable thunderstorm.
In each case the aircraft flew from the Riverside Airport to one side
of the Riverside Freeway through the Chino Hills over Anaheim and Seal Beach to
the Pacific Ocean. Fortunately major land marks such as Disney Land and Seal
Beach were noted on the record since, as we have previously noted, the flight
recovery camera failed to work.
8.4 RESULTS
The vertical integrals of the NO along the line from Riverside over Anaheim to
Seal Beach (Long Beach) are shown in figure 8.1-8.3. In these figures the
right hand side of the page represents Riverside, the left hand side represents
Seal Beach with Anaheim approximately in the middle of the page. The vertical
direction represents time and heavy oblique lines are the abscissa for zero
N0_ with the passage of time during the flight. It should be noted that in the
absence of flight recovery photography the exact location of the aircraft is never
known with certainty, and the ends of the flights are determined from notations
made on the graph. In most cases the notation for Anaheim is also included on
the record. Thus, the ends and center of the traces are reasonably accurate.
For the remainder of the trace the assumption has to be made that the spee^ of
the aircraft was constant and while this is substantially correct, small changes
in wind velocity will cause minor peaks to be displaced slightly.
Figure 8.1 shows three runs made on the afternoon of July 8th in an attempt to
confirm the westward movement of the polluted air mass under the influence of
the afternoon land breeze. Only three runs were completed before the flight
had to be aborted due to cloud build up and adverse weather forcast. Figure 8.1
shows a very irregular trace indicating abrupt changes in gas concentration
except for short periods of the flight which show very steady NO levels, and
were recorded while the aircraft was flying over the cloud. No attempt was made to
align the traces of these three flights. The NO cloud which the flight was
- 132 -
-------
designed to follow is not as discernable on these . flights as on the morning
flight which followed on July llth.- Oh .run 2, the high gas concentration was
obscured by cloud and between flights 1 and 3 very little movement is discernable.
By using the two constant sources which are defined more precisely in later runs
to locate the Long Beach area, the gas cloud which is clearly defined in Run 3,
is found to be located approximately midway between Long Beach and Anaheim where
the flight path crosses the San Diego Freeway.
The data for the flight of July llth is shown in Figures 8.2 and 8.3. Figure 8.2
shows the raw data as indicated by notations on the graph. Each of these curves
show two constant sources located in the downtown area of Long Beach, and in
Figure 8.3 the curves have been moved to bring these two peaks, defined as B & C
vertically beneath one another. Seal Beach is now located at A.
The broad hump representing a polluted air mass seems to be located about 5 miles
east of Seal Beach in the early flights and. to move in an easterly direction during
the course of the tests. The dotted lines represents 9 ppm meters per division
and the maximum value in the polluted air" mass is about 45 ppm meters. The center
of the hump was located by taking the middle of the intercept of the 36 ppm meter
intercept and dropping the perpendicular to the base line at the point marked^thus
the movement of the center of the polluted air mass demonstrates its movement
inland and the curvature of the locus suggests that the movement is accelerating.
The total movement of the air mass is. indicated to be 6 miles in 2.5 hours or
approximately 2.5 miles per hour, which is in good agreement with normal wind
velocities in the area. There is a further peak located near Riverside (designated
by the letter E) which also appears to drift in an eastward direction. The velocity
however is less than for the coastal air mass being only about 1 mph.
Figure 8.4 shows the profile of the second run of the July 10th flight, plotted
in relation to the freeways and major intersections. There appears to be a good
case for believing that the airborne spectrometer responds not only to the major
polluted air mass but to local pockets of pollution in the vicinity of major
freeways. The maximum vertical integral of 45 ppm meters is lower than would be
- 133 -
-------
expected, but as discussed in the previous, qhapters, considerable dilution of
the signal results from light scattered within the polluted air mass and for
reliable quantitative results, dilutipn factors have to be included.
8.5 COMPARISON WITH GROUND STATIONS
The only ground level data available at this time is from t;he Los Angeles County
Station 72 in North Long Beach. The data for July 8th, 1969 are:-
Time 1200 1300 1400 1500 p.D.S.T.
Wind Direction SE SSE SSE W
Wind Speed 3 .4 4 10
NO PPHM 8 98
The inversion height was 7000 ft. MSL or about 2000 m. The maximum vertical
integral from figure 8.1 is approximately 25 ppm-m, or about 10 ppm-m for a
single pass with allowance for the m1 factqr. Thus, the average concentration
will be 10 x 100 pphm-m = 0.5
2000 m
The data for July llth, 1969 are:-
Time
Wind Direction
Wind Speed
N02 ppm
These results indicate that the peak of the gas cloud has left station 72 by ip afm.
and our results indicate its peak is 5 miles east of. Seal Beach by 10 am. The
maximum vertical integral recorded was 45 ppm-m which with due allowance for the 'm1
factor, gives a burden of 20 ppm-m. The inversion height on this day was about 1500 m,
so that the indicated average concentration is:
20 x J.OO pphm-m = 1.3 pphm
" I \ ' T
1500 m
- 134 -
0800
SW
1
15
0900
W
1
25
1000
W
1
25
1100
SSW
7
12
1200
S
5
13
1300
WSW
8
11
-------
In both these flights, the indicated average concentration is about one tenth of
the recorded ground level readings. Part of this will be due to scatter/dilution
of the signal and part due to non-homogenety of the gas cloud.
8.6 CONCLUSIONS
The airborne system gives precise and useful relative values of gas burden beneath
the plane and the method can be used to trace gas clouds and observe gas sources.
Because the influence and effects of scatter/dilution are not fully understood at
this time, we cannot give quantitative data. What has resulted from our airborne
work over Los Angeles is qualitative data which while being of use and interest,
is somewhat limited in its direct application. Theoretical investigations into
the application of this airborne survey technique to air pollution work is
continuing and the achievement of quantitative measurement is considered to be
with the near future state-of-the-art.
- 135 -
-------
FIG 8.1 WIND TRAJECTORY FLIGHTS RIVERSIDE-LONG BEACH JULY 8,1969
- 136 -
-------
X
u
<
iu
CD
UJ
WJ
FIG. 8.2
RIVERSIDE 8 SEAL BEACH FLIGHT
JULY 11/69
-------
SEAL BEACH
UJ
30
40
RIVERSIDE
AIRPORT
TIME P.D.S.T
FIG 8.3 WIND TRAJECTORY FLIGHTS RIVERSIDE-LONG BEACH JULY 11, 1969
- 138 -
-------
N02 BURDEN PPM-M
31
P
oo
m
3 5
o
m
o
"
dio m
O _ o>
m
§
ro w 4^
o o o o
I I I I
SEAL BEACH
S.DIEGO/S.GABRIEL F.WAYS
BEECH BLVD.
S.ANA FWY
ANAHEIM
RIVERSIDE/NEWPORT
* FWYS
YORBA LINDA
RIVER CANYON
CORONA FNY
RIVERSIDE AIRPORT
-------
Section 9
CONCLUSIONS
9.1 SUMMARY
Tests carried out under this program demonstrate that the Barringer Correlation
technique yields a family of instruments which can carry out air pollution
analyses not possible by other standard techniques. These instruments can be
operated in several operational modes which yield air pollution data of great
interest heretofore unobtainable.
In summary our Phase I work showed conclusively that the sulphur dioxide con-
centrations measured at the tops of various stacks under varying conditions of
daylight agreed closely with concentrations recorded by the in-stack monitor
located directly in the flue gas stream in the breeching. The remote sensor
therefore can provide a valuable method of remote surveillance of stacks for
identification of gross emitters of sulphur dioxide. Both instruments will
have added value when efficient scrubbing of flue gases from high sulphur" '
fuels becomes a commercial proposition 'and inefficient operation, which would
release high concentrations of S02 into the atmosphere, could be quickly detected.
Phase II showed that the airborne system can rapidly monitor the behaviour of a
plume. Changes in concentration, resulting from diffusion and turbulence or
from gas phase reactions can be monitored in space and time without disturbing
the system. The technique has been applied to the atmosphere oxidation of
Nitric Oxide to nitrogen dioxide in a boiler stack plume and it has been shown
that the reaction is too fast to be due to oxidation by atmospheric oxygen.
Phase III demonstrated the superiority of the long path ambient monitor over
the point sampler when measuring sulphur dioxide in urban atmospheres, the
impingement of plumes at ground level causes large variations in concentration
of pollutant gases both in space and time. It has been shown that although
- 140 -
-------
the two units are consistent for pathlengths up to 300 meters for the long
path sampler, considerable variations exist between values from the two
units when the long path sampler uses pathlengths of the order of 1400 meters.
The experiments of Phase IV were carried out under conditions of light to
heavy smog in southern California measuring nitrogen dioxide. Under conditions
of heavy smog, the long path instruments were limited by severe loss of light
due to scattering. The airborne system was also hampered by the inability of
the aircraft to take off in visibility below two miles.
Phase IV A showed that the long path sampler was again more representative of an
area than the point sampler when measuring NO under smogy conditions. The
concentration of NO as measured by wet chemical analysis of grab samples was
found to vary widely in both space and time and led at times to poor correla-
tion between the long path and point sampler ambient monitor. Two problems of
the long path monitor uncovered were the energy limitation due to excessive
light scattering in heavy smog and the response to sunlight when looking close
to the sun.
Phase IV B showed that NO concentrations over the Freeway may be five times
higher than the concentrations to the side of the Freeway. This is probably
particularly true of a Freeway along a wind track. The effect of a cross
wind in blowing NO over the surrounding area has been clearly demonstrated.
The aircraft surveys in Phases IV A and B have shown that in addition to the
interference with the aircraft operation, heavy smog and the intense back
scatter of light from the particulate matter render the airborne data very
unreliable. Qualitatively, areas of heavy NO concentrations were indicated
but subsequent passes failed to confirm these concentrations or to show their
trajectory.
Phase IV C also suffered from particulate interference and the quantitative data
is of doubtful value. However, qualitatively, the buildup of N0_ from the
coast inland, the local concentrations of the gas over major freeways and the
- 141 -
-------
of a large cloud of N0» inland under the influence of the moving landward
breeze were all clearly demonstrated. Attempts to show the return of the gas
cloud under the influence of the afternoon sea breeze were unsuccessful.
Parallel test with both commercially available analyzers and by wet chemical
referee methods show that for all configurations but the airborne, the
correlation spectrometers have an accuracy which compares favourably with the
commercial alternative systems. The airborne system is limited by the dilution
effect generated by scattered light from aerosols and particulates. It is
primarily a comparative system only, since it does not directly give ground
level concentration. The airborne platform necessitates a fast response time
for the spectrometer which presently precludes the use of the slow grating sweep
incorporated into the long path ambient monitor.
- 142 -
-------
Section 10
ACKNOWLEDGEMENT S
The studies discussed in this report have all been carried out away from the
facilities of Barringer Research Limited and we gratefully acknowledge the
co-operation of the various persons and organizations who have given us access
to the facilities which have made these tests possible.
- To the British American Oil Company (now Gulf Oil of Canada Ltd.) who have
given us free acess to their stacks and refinery for the extensive tests carried
out as the original Phase 1 of this project.
- To the Hydro Electric Power Commission of Ontario for permission to carry out
stack tests on their Lakeview generating station
- To Lever Brothers of Canada Limited for permission to carry out stack
tests at their Toronto factory as part of the re-run of Phase I of this
project.
- To the Municipality of Metropolitan Toronto and the Commissioner of Works
for permission to use their research incinerator installation at the
Humberview test site to examine the behaviour of the In-Stack monitor under
precisely control high temperature stack conditions.
- To Mr. Brad Drowley and the Air Pollution control division of the Ontario
Government for permission to set up our equipment on the Evans Avenue site
for Phase 3 and for making available records of the Beckman Coulometric
SO Analyzer for comparison of their results.
- To the Lyons Tea Company for permission to set up our long path sampler light
source on the roof of their factory.
- To Mr. Louis J. Fuller and the staff of the Los Angeles County Air Pollution
Control Dept. for their cooperation and understanding.
- To the air controllers of Brackett Field tower and March Air Force Base tower
for their cooperation and help during the flights of CF OVC.
- 143 -
-------
To the president and staff of the Kellogg Campus of the California Polytechnical
institute for permission to set up our equipment on their grounds; for their
generous provision of scaffolding and electric power.
to the state highway patrol for permission to collect samples along the shoulder
of the San Bernadino Freeway.
To Dr. Haagen-Smidt and his staff at Riverside for their interest and advice.
To John Nader for his continued understanding advice and encouragement.
- 144 -
-------
APPENDIX A
-------
""R.C
1 13001
2: 1200
3 1100-
^ 1000
- 900
O 800
§ 700
o:
LU
600
g 500^
8 400
0300^
200}
100-
0
900
S02 CONCENTRATION (IN FLUE) VS TIME.
OCT 2
DATE
LOCATION
FUEL OIL
HSH IN STACK MONITOR AVGE
SHELL AVERAGE
X (REMOTE SENSOR
WITH coeftecTtoM R>R
;
^
xx
FIGURE A-l
ST*C|t ^LOVJCP OFF
I
1000
1100 1200 1300 1400
LOCAL TIME (MRS) EST
1500
1600
1700
-------
S02 CONC. (AT STACK EXIT) vs RANGE
(REMOTE SENSOR)
DATE.
LOCATION.
FUEL.
1200
. l '
~JIOO
tipoo
" |900
o 800
700
600
500
g 400
^300
200
100
100 200 300 400 500 600 700
SLANT RANGE (METERS)
FIGURE A-2
-------
S02 CONC. (AT STACK EXIT) vs RANGE
(REMOTE SENSOR)
DOTE. OCT 3 '
LOCATION.
FUEL. 0»L
f\
MONITOR
1200
ioo
u i900
o 800
en 700-
600-
o 500-
g 400-
^300-
00 200
100
EST.
0
100 200 300 400 500 600 700
SLANT RANGE (METERS)
FIGURE A-3
-------
~ 1200
3-1100-
^ 1000
- 900
o 800
§'700-
z 600
LJ
^ 500-
8 400-
o 300-
200
100-
0
SOa CONCENTRATION (IN FLUE) VS TIME.
900
X
DATE OCT3
LOCATION
FUEL 0\u
IN STACK MONITOR AVGE
SHELL AVERAGE
x REMOTE SENSOR,
FIGURE A-4
1000
1100
1200 1300 1400
TIME (MRS) EST
1500
1600
1700
-------
SOz CONCENTRATION (IN FLUE) VS TIME.
| 1300-1
- 1200
D 1100-
" 1000
Z o^
900
| 800
| 700^
i 600-
±j
^ 500-
8 400
o 300-
200-
100-
n
DATE
LOCA
FUEL
. ;'; ;: ' :; :' ' "-o~
MH»4M>
X
X. K ^
X JQ
A " . " .. " .
X X-,
XO Xv
1 «rtt *r-x
Rx x^ 5r
*&. ;t-v-»«A.
"ia sf-^9 x
X 18T
x
'x
« ^.
$ ^
It * «
£ ^S ^
? 1^ i
3 o-^ ^
1 «l 1
OCT. b. %9
TION uvefc ^to5.
OIL.
IN STACK MONITOR AVGE.
SHELL AVERAGE
REMOTE SENSOR (Rs*WM».)
STACK ftEAfe*i
-------
1300
1200
^ IIOOJ
^ 1000
- 900
0 800
§700
2 600^
LU
g 500^
8 400
0300^
200
loo^
o
S02 CONCENTRATION (IN FLUE) VS TIME.
900
**X\ V«* ..-
- AA A
DATE OCT -7
LOCATION
FUEL
IN STACK MONITOR AVGE
SHELL AVERAGE
X REMOTE SENSOR,
FIGURE A-6
1000
1100
1200 1300 1400 1500 1600
TIME (MRS) EST
1700
-------
1200-
u ;900^
o 800
co 700
£ 600
o* 500
g 400
^300
10 200
100
CONG. (AT STACK EXIT) vs RANGE
(REMOTE SENSOR)
i
DATE.
OCT. 7
LOCATION.
FUEL.
OIL,
100 200 300 "400 500 600 700
SLANT RANGE (METERS)
FIGURE A-7
COHCB' : BvllO-/5 '!/; tF
-------
METRIC
-1200
=> IKXH
^-1000
- 900
O 800
§700
z 600
UJ
^ 500-
8 400
S 300
200-
100
0
900
SOa CONCENTRATION (IN FLUE) !! VS
OCT
IN STACK MONITOR AVGE
SHELL WERAGE :
REMOTE SENSOR X
C A L. READ. BADLY
&
.
o
FIG;
X REMOTE" (K<,= /26*rt)
STACK MON/TOr2
WE T CHCM/CA/_
STACK
1000
1100
1200 1300 1400
TIME (HRS) EST
1500
1600
1700
-------
£ 2000
(L
u
9
J
a
Z
z
o
P
1800
I600
1400
2
3
1200
H
O
1000
eoo
OCT <\
LCVER
t \"L<* m
X - IHVTACK
FIGURE A-8(b)
1345
13 JO 1355 1400
TIME E.S.T.
1400
1410
(420
-------
-------
S02 CONCENTRATION (IN FLUE) vs TIME
V £ T p IC
r\
Date OCT zi '^ » In stack mon. avge.
Location uv.fs. OMTHO. ?. Shell average
1700-
1600
1500
1400
1300
_ 1200
Jnoo
1000
UJ
3 900
z 800
| TOO
of 600
h-
g 500
o
o 400
o
cvj 300-
o
10 200
100
n
:;. -;:.--:- I--..- ;!i:i:H^MMS^Wll
- ::; . ""?"" i : : :-: -.. i !"'::::
.:..: . . . jl; . -I : ' : :.-:;.; :!:!!-:::
IjjjiliTiniriiS;; |§n^;-^-^'n^iffi
f |pli|i|;^ Ililll
-:*:; T-.r.-.-jiij
;:: *^*tt
.':. : ."::.:. T~
::|; j| : | ;
:;: :: '.- !
.
;p»«i
||||
; : . : . . j
: "r
V.T - v :
" " V ' " . "
I^T: . ;:;..;::::
::i:; :|: rrHsi
| i|l
;.--;; ;
t . ,(....
::.:l::;:
t::::::;:
:.;::::;
;:'.:J:^J
M*
~.rr.. .r
; T i ' 77 : rr
''.:. :rr:.r~'
JjipMjU:
:::!:;:: ::::
1
-:rr
1
[jilj
,. . . t
1
itii
lip
hi
511111
... i.ii!!... t ! . . i j
' - ' [
: : . : i . ' : : :
::.:
'::.!
,'); :,:::: * iHjHll
rn
.;.' 1 :'.:.: :
.:::'.:::!. .:
: : : ;:::.;:.::
.-::!::::!
§||
....
*'
.» .*-
'::
.. |..
:.| . :
.
;
ii.
. . .
: :
TTT
uiu
....... :i
..»-.....
; : r i : : '.
it "
r-:Uii:'
It-fjpi
Fuel >CoAu*
ili LJlrij
H- !;(:::;: i..
:::::::-:|:::-;-:;
.;.i.^Trrr-rr~rr
* - . . 1 1 ; .' ) * ;
liil'LL '
::::::::;.:: : : :
mm |
!!.::::: 1; : ::
ii'-ii :: : : ::
:::;:-::!:: r}::
::::.i:::i
nr:;:::
::::i;::rj
: : : I : ; :'
:::;:::
rrri--:
:::!:.
t. .......
::-,:::
::::.::
:;:is.:
J.-J.UI:-
lljljii
;:::!..;-
]j| i
:::.::;
....::
t=jpf|
Remote sensor
?ll!iH|!j|i:
IpiEii^
:::;.::;::
.:;::.
:::!
.. .1 .:.... |.l ;- .. . . ..
: : : . : . : : : .::!::. ! : : : ' i : i i 1 :;:;::: : : . : :
JJr\Ttf
;i|j}[j::[i
ilPBffi!
[i i^Hjl
-H: irk^f
;::j::::|j!t:
.: -H}f
. ffl
jinfiriijjlitlf;
;:'::;:-r!;!;::^l:::--':-
:_.f.i:!i:u.-. : ; ;.
:::;;: .: :::: .T ' j
:::[;;-
1|:!
:::.! : .
. . .
;>
. ..:-S
.,.:;;
;.. ;.....
:: .:: :r ::j:
|
i
.ii.^i.i; :i::::::;\:.
.:lUt .if!!: U:'.. .
Trrr:::. rrfrrr: : .
i|| ill ,
If :: : . . :: ' ' . . : ! : . : .
'{";
-:::: :j;; :i :.
: ' t. . ',:'". . .:.:.'
14 iliwrr *4b.
._!'.' ... .....i . . " :
.. ::.:: .. . /. .:!:.. :: . -...
5.: . : . .'' ... . .'.. ;
.'.
i::'::
*
: '.: :.,"
FIGURE A-10
900
1000
MOO
1200
1300
1400
1500
1600
1700
-------
IN-
FIGURE A-ll
-------
S02 CONC. (AT STACK EXIT) vs RANGE
(REMOTE SENSOR)
DATE. Nov Z4
LOCATION. L.V. P. S
FUEL. Cof\L.
11200-
Si
1100
11000
u |900
o 800
700
600
o1 500
g 400-
200
100
100 200 300 400 500 600 700
SLANT RANGE (METERS)
FIGURE A-12
-------
S02 CONCENTRATION (IN FLUE) vs TIME
V.ETRIC
.::::::;l::::I::::t:;::[::.:!::::{::;;
1700
1600
1500
1400
1300
_ 1200
| 1100
1000
yj
3 900
z 800
Q TOO
of 600
»-
LJ 500
o
o 400
o
CM 300
o
W 200
100
:^ .-.ut .-
:'d
..!" -,
. -t
..i-:
r:**
* « r
ffifi
it
mMi
*:
if
* "i
:r"-t
1
.4
I'L-li
n-;|:-;n
liii
MS
:;ti;^
fllfi
RttfFtl
JliJ:r!!
HRp-a
iiS
ism
»
tti;
1
fs'
:::T
it::
IF
;i:
;tji|:i:
*J T j »T-
q
1
IS
:ttf !W:,TU{
i?fi ?i?8iS
t:i
rff
1
ii
w
tlfil
lif
.:::;:::::!::.
:';:;_p:'l:i::i
!::-i!";Iiil!ri
.:
iiiiii:;;
ill
P
i
iii
nl!
iimit
iiliiil
nthvi
if:;
n;1,
itft
ilUl-i
mp
»
ii
iiitlijli
--
i:. :;::t:u
i|-a^
Date Mov.
Location I-V.PS.
Fuel
m;
t:t:
; m::t
:!t::!!
inn;:
: :i
f|
!»
ii
i i
[:^
4-i;
»
r
In stack mon. avge.
Shell average
Remote sensor/ ~*sv
-.7::
iiiii
f-
U.L
.-:4
....
'.:'
;
FIGURE A-13
900
1000
1100
1200
1300
f*
1400
1500
1600
1700
-------
S02 CONC.(tfT STACK EXIT) vs RANGE
(REMOTE SENSOR)
u
on
100 200 300 400 500 600
i SLANT RANGE (METERS)
700
-------
M: i .- i c
S02 CONCENTRATION (IN FLJUE) vs TIME Date Kov Zl
Location L.VPS
In stack mon. avge.
Jhell average
E
a.
LJ
1
o
o
o
CM
O
en
1700-
1600
1500-
1400
1300
1200
1100-
1000
900
800
700
600
500
400
300
200
100
n
''~-
:..i :
::; . '.
:i::::
- .':- _
:: ."':
:
--
--
:-. . . i
-~. '.
-~77
||
.;::"-
1
[.-";
'
- - 1
:: t
*" I
:
"..--}: -:: ' -!::
! :! r: ' .. :-': >:-.:-.
: L::;:ji: :' .:':{': ;.':! j :i i :l
'. : : : " : : : ; : 1 : ; : : ; i ; : : : .
.-~:l.. : :..:. :^.:. I ;uu..;
: ; :.::: r .- [ ;...: ! :::; :;
Hllrtjlft
.: . :. :: ::.: 1 :.:;:,:
ii|filiiiHnfiji
m ' :
::... ;: -::.:::f: . :':;:. :.:,
' ; - -/ : : :i -::::j!::
.. .: :;:
.:: . ;;: '.; : ::!, :
-I.::.:-:.
' - . :'. " ;':
' T -,..-
-
: .:. :::. . |:: .; ,: .:
. ' ' . ..::.: ..'..
;'
r-:
p-j
1
14
;
_.
:
: :
r-:
1
:
:.
«i
T- :: '--.
: .. :4 _ j
_:;;;;
. : : ; i
f^
lljjj
: :t;:j :
*
: '.'.:::', : 1
::'!::: i :
fjfftj
... . .1
'. . i
!.:: : . i
ilpjl; ;!
. :!:. : j i
J ; : I
. .; :: ' .. I
:;:;.
yi!:? -
- . ;
j ~i
:
j : i:
! : :i
: ; :i
i t :
:: i : :
rrt*:T
. 1 : :
; i : ;
:: i : :
:: | :1
' ; 'i
| »|
;,'' : :
:; .i ^
t- \
'.'.\
:::: 1
-T:
iL
n
U
:::
*.
J
:*:1
i !:i!
fflj
.lr:r:[
I
r"::j
;:: i
13
rr:rj
1
;'_ t£j|i]I^
H-: - t:'::; '.
t:-.:.::- .,
v:-rfrr:Fv
.
I
-
--
i
*
.
.
Fuel COAu
;.:; 1
I
it-
rr. . - f
:;:.. I
iljij
!.. i
:::. t
i::: (
1
:.\
;__ -j
n:::::
|J|
jfej
jiriBji
i::::- :
i ::!.
-TT-
' -;!-';
t '/I:.
:
:-
::^
....
:::'
|
i:-:
j"'
pi
:. . :
( ''~:
I " :
I :
i
j
'-
r
If
r;
r
» .....
TTr.rrrr
., .. n
.1 !:
. i j
I-
;*
. i
...
|
-.'.:-.
:.:j:; [jj:: :
~~f~. ;~.-. .
''-.':
....... .
::::.:.:: :
.
-rr
-
.
;
r'TT":
.
.:: V!::!!
':-. -
7- -'.r :.-
;: :'. [!:
; ; :ij|:j
rr-. r
r
( .:
... ,.._
r
'.):::
: j::-
:;:!:-::
-?;. -.
|;;
'-.1 .
!
: J:
. i
'-. '.::
ijijili
r --'
........
t-ir-
. . . f .
. . . ,
-
:
.. .
I '
:
j
j;
.
- ;-r*i ::
I! ' '':'-
-.. -
- :.-::
t
. :
;;
i
»-
i
i
!
f
;
i
i-i:
^
Remote sensor
" . . . ..:....
'. .
-
.'
- . . :
: ii: :
..
i .'.;
i
- ; -.
<'
'::'.' :
i
: ; :
'." '. ' ' '
. :
"
:..:. . . -- :--;-:
-.- - :
-
-.- - -
.
' - -~
_
'
-.;
-.
FIGURE A-15
.
:{':Hi ..; . ;
. . .
. . .
-
- --
- : -
--
-
-
900
1000
1100
1200
1300
1400
1500
1600
1700
-------
S02 CONC. (AT STACK EXIT) vs RANGE
(REMOTE SENSOR)
DATE. NOV/ 2«
LOCATION.
FUEL. i: : -:: ' ! ,;.' :
FIGURE A-16
-------
)2 CONCENTRATION (IN FLUE) vs TIME M
1700
1600
1500
1400
1300
_ 1200
| 1100-
1000
Ul
3 900
z 800
| TOO
Si 600
H
5 500
0
o 400-
o
cvj 300
o
" 200
100
o
:;_:
--r~
:
..:;:.:..
, - - 1 I-. ... . [.,,.,_,.
!§§5pi|irj|f§
. : . : . . - ' . . . . i ...'.::;..:::;
-.'. 1 ..
, ..:.:. :.
::::T. :
:;':{: :
:.!:..
.. : '
r ::;::::.
mss
**-*+ ** r
--4.. j. *r.
jjjj ;:::
Liii
frrr
!: :
^i'rJ
d
,:::
|i|
; »i . '. ' ;
......
:
-. :-
'
:: :,;::
.:::{!:::
]'',-**
iJitiiii
--if. f
JrHii;:
Iff r;-t
1 lilili '
ri i::;
liiM
i:: :::;
ifi
n
:t'it! t:
:;;ii:;::
Ttrrrrfr;
;t:;r:;;:
bd
h'}iiii
(;:: tut
1 . . t . , .
ill i I ' t '
k::j rif;1
I
;;:;i:;::
III
::;: it;t
..».,..
::: : ::
iirtrni
rrrt
:.::
*^I!
:-iU,,t;
.!:::(:. ::|
^:^j
|
l
>:
1
.
'
. : : j : " : i . ' . r ; r t +*
::: , ;::;';;:
;|ti
, . ... , _,...< . - .-... » . .
;:,
TT
I
jci
ill:
1
|
:
:rn
FT:]
bate *«* *
Location Lvfs
R
:.-.
...,'
-1
:!^
uit-^ii.
ip
'J7Tti;r:
m
ii.;!i;iH
jtHjin,
ipl
;!::!::::
::::
;:
, " ,' i . '.
i. j:
'. . . 1 : :
jei
iiiiili'::
-.-(.*..
frttHi i
§11
B
1
:";n."
iijiijiitj
|
i;r
; ;
i-1
i
:|
'
i
|
||
H
.:
Jt?:!::.
.I..)::
r::
TTf-
;;:
Im
*Tfff~
*
fni
iHr-
,1.1
;-:{:
uli
rnf
LU
ir
: ii
JliL
' : t
: '*
<
: : : : i : : : :
1 ::. ;.:
I :;: .:::!
I!
;:jj::;.;j
;Lii::|
f:i;ii
if
H
[dJ ;L.,
^i...*_t^, L.,
** * In stack moo. av^
- Shod overoQt
|
;FT:1:"
i
','. rt
f
:-:
1
r rl.T
U
.:i
;ti
;::
.'-!!-
||f
i:!:; :'): ;'
fH
i . " . ; ' '
'.r ...
-::-
:
[Liii
"H
-rn
|j
jfi
.: ;; ::::
: r.
i i-
r.
'
i
r~
1
~
~
:: t
r:
t
j:'
i::
ri
i :
i :
x rtvnior* «n§or , ^t » »w% «
(» I --- ..- » -
--t t -i...i. -,....,
j ! :* ' ----~'t j
m
Bf
rr.i::..:
ill
m
[UHiji
i. ::(::::
' ' '''
' -'Up?
::j;-.:
:iJt
: ;. :::
jUji
.- t :r
.::!::::(..:.-.::
:::;r:;i::.-:::;:
::::;:' I :.:.;::::
iiipitlii
:;:;;:;!::;::.;:;
fifll
:::! :.':::;
:i::;: :.! :.:;n;:;
:t .!.:.
. ! ! » ' :
Li" v : ';:;
' ' ". . ;
FIGURE
. :* ::;;.;:- :::.
:.':::;::
:.::. : :
::::; -
:i:~:-;;;--:-ff^:v;
: :::::::;:- ..::
:;;;i;:r:::i ;:::
: ;:::--:[r-:. :~:
::::::::!::::. ::
: ^
*rr~f*~:'.t*tr'rrr
:;: [.'': j:j: :::
. . , . :' i :;.«;;:
."..:;'"
A-17
"
i
;i:i: :";
.-"-..::
- ..J
::;; ::j
::::-. ^
- H-'-rj
fTTT
-1
. . j
-
--
900
1000
MOO
1200
1300
1400
1500
1600
(TOO
-------
Sensor Ser. No.
PLUME
SO2 REMOTE SENSOR^ DATA SHEET
REF. CELL
TABLE A-l
STACK LOCATION.
DATE.
No
& deqree
(5)
It Oft
II
11
Aperture
Vi Div.
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog (f) , Etc; (2) (3) Note where taken; (4) Sketch elevation ane"
Plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt.
-------
Sensor Ser. No.
SO2 REMOTE SENSOR DATA SHEET
REF. CELL CjLj 3^00C2L;> =
TABLE A-l
STACK LOCATION
DATE
No.
Note
Ht /m)
431-7
(m)
2-e
Temp. (°C)
Skv
m
/o
Vis.
(2)
Wind (mph/°)
(3)
/3/oao
PLUME
(4)
No.
PS
& degree
(5)
Time
ffil
Aperture
Div.
(7)
-47
V Div.
VT Div.
fvolf-O
FOV. (m)
XLT. (m)
2-7A
(m)
Temp. Factor
~r
&SA
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog vf) , Etc; (2) (2) Soce where taker.; (4) Sx^-ch elevation
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt.
-------
TABLE A-2
Sensor Ser. No.
No.
Note
SO2 REMOTE SENSOR DATA SHEET
REF. CELL dH^^Cal* =»*S°
STACK LOCATION
DATE . .3.
S .« .S
Ht /mt
42*
n
-------
TABLE A-2
Sensor Ser. No.
SO2 REMOTE SENSOR DATA SHEET
REF. CELL a*°° J-ao°
STACK
DATE.-5.eP.dCrT
No.
Note
Ht (m)
+TL-
(m)
2 >
Temo. (°C)
Skv
M
Vis.
(2)
^
4/J
Wind (mph/°)
(3)
/45/O
PLUME
(4)
"?t< MI? /HV/AI
No.
& deqree
(5)
Aperture
Div.
V2 Div.
Div.
as
AGC fvolt-cl
FOV. (m)
XLT. (m)
2.
(m)
7/4.
Temp. Factor
I-AI
F>£iu|!
!UK
AT
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog vf) , Etc; (2) (3) ^oce where taken,- (4, Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A-2
Sensor Ser. No.
No.
Note
SO2 REMOTE SENSOR DATA SHEET
REF. CEL
STACK LOCATION
DATE.^Rl>.
Ht fm)
(m)
Temp. (°C)
Skv
n
Vis.
(2)
Wind (mph/°)
(3)
PLUME
(4)
No
a/
s degree
(5)
Aperture
Div.
V2 Div.
(7)
22
2.2.
VT Div.
17
FOV. (m)
^Zi
Q-74
XLT. (m)
- A.
2.
(m)
44
2/7
Temp . Factor
/-a/
/-*/
AT
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h), Fog (f) , Etc; (2) (3) Note where taker.; (4) Skecch elevation arid
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A- 3
Sensor Ser. No.
No.
Note
1
SO2 REMOTE SENSOR DATA SHEET
REF. CELL CiLj = ^°C2L2 ='
STACK LOCATION .
DATE _ gfft QCT / 9 >$
Ht (m)
42-7
(m)
Temp. (°C)
Skv
M
Vis.
(2)
Wind (mph/°)
PLUME
i
fb
No.
& degree
(5)
Time
/Of*.
//o/
II O
//3d
Aperture
Div.
V Div.
AGC (volf-O
FOV. (m)
XLT. (m)
RS (m)
Temp. Factor
AT
SHELL (ppm)
ton
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
Sensor Ser. No.
No.
Note
S02 REMOTE SENSOR DATA SHEET
29OO i*
REF. CELL CjLj = C2L2 =
TABLE A-3
STACK LOCATION^
DATE .<6THl &&J9&. $
Ht fm)
(m)
Temp.
Skv
n
Vis.
(2)
Wind (mph/°)
(3)
Gf/AO
PLUME
No.
& deqree
(4)
O*GffC*S r
(5)
£2*0
USA
-
Aperture
O'JT
Div.
(7)
V2 Div.
V Div.
AW fvr»1t-
-------
Sensor Ser. No.
SO2 REMOTE SENSOR DATA SHEET
REF. CELL CjLl = ° CaLp
TABLE A-3
STACK LOCATION
DATE .£T* .
-------
TABLE A-3
Sensor Ser. No.
No.
Note
S02 REMOTE SENSOR DATA SHEET
REF. CELL CiLi2^°C>C2L2 JAS^°
STACK
DATE.^W.
/m)
(m)
Temp. (°C)
Skv
n
Vis.
(2)
ti
2-3
Wind (mph/°)
(3)
A//OO
PLUME
(4)
HAZY
No.
C3
& degree
(5)
9-0
TIYI
Aerture
Div.
(7)
Div.
VT Div.
14-
/>
2.1
FOV. (m)
O-/4
O-/A
XLT. (m)
*a (m)
Temp. Factor
2.0 /
-4 7 A
ippmj AT
£77
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A-4
v
o
£i
CO
rv*
LL.
O
LU
H
LU
2
UJ
H-
co
GH
Note
m
(2)
(3)
(4)
(5)
fM
(7)
0
1
+2-7
2-6
/ XO
10
or
P6
23-0
09S"4
. ^*
5-9
"5/
50-5T
-1ST
2-^2.
no
/44*
^,75"
331
R
«
"
II
*
M
095*9
50-
1*
"
M
**
6>/a2>
9/4
£F. CELL
It
»*
*,
*
ti
100?
'
t.
»
"51.
*»
»
-
,,
y C' O'
Q ^LQ
ClLl =^
"
"
*
*
IGlO
-
"
2.9.
..
i*
& %tf
W/
^r
"
-
'
H
»
\Ql(o
57.
3/.
29
it
-
-
636
^77
M^^^i^OT^Hi
-/3^
«
»
*
/
1010
tt
»»
«
Z9-5"
(i
11
-
64*?
9^9 ^T
95c7 -
"
M
"
702 3
#
"
"30
"
w
"
Vf
662
9/4
«-^^.^
»
TQIIt
"
/i
10+0
25T
5"7' 5"
"5O«i
5*^9
Vlb
l.LiP.l. t
-
M
"
iOS'O
"
*
«
29
j*
-.
«i
(e^Q
^95"
&9
"
/05"tr*
"
-
2f
'
it
-
(o2*l
*t>/
t
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A- 4 (Co*
S02 REMOTE SENSOR DATA SHEET STACK LOCATION k
Sensor Ser. No. DE"M0 REF. CELL CxLi =2?00C2L2 = 1350 DATE. . . 17 J. I P. J.
O
C/3
METEOR
bJ
1-
co
SENSOR
| RESULTS
No.
Ht (m)
nia (m)
Temo. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& deqree
Time
Aperture
V! Div.
V2 Div.
VT Div.
Ar,r fvoit-o
FOV. (m)
XLT. (m)
Rs (m)
Temp. Factor
J&Sc^^ESff ) :~T
bti\JbUK (ppm) AT
r Li\JLt _
SHELL (ppm)
Note
n i
(2)
(3)
(4)
(5)
,M
(7)
/
4Z-7
2-fc
1 S"O
z-^mi
^
* C3
12-5*
11 tc
. 5*
5"^
51
23
33
2-
3^
"2*fo^
23^
n
5"Z5*
T2.4
^
ii
i
t>
«
112(0
»
M
I 3
M
ll
II
" '
Z^4
39Z
it
M
1
,Tub«
Ik
"
//Z"7
i
»
»
I e>
H
It
II
tl
40Z
5-5%
M
it
it
//Zf
II
II
II
\(o
"
"
H
35^
4S^
«
M
-
H
H
//Z9
H
N
U
/?-5~
(
n
H
H
43?
60£T
»*
**
"
USO
n
>i
I
22.
"
"
it
. .1
TOO
bio
II
1
M
C4
ff-0
//40
'«
5"6
3O
14
w * /
5"/
>i * w O
307
i
3/9
440
M
'
t
u
II
1(41-
Z5-
h
r
zr
-
"
110
/5"Z
«
TXxck
ftlack
ii
V
n^i
n
a
it
O
ft
*
"
1
^
0
"y
6.? l
*
*i
t%
*
M
nil*
»
i
*5"
-------
SO2 REMOTE SENSOR DATA SHEET
TABLE A-4
STACK LOCATION
Sensor Ser. No. P£TMO REF. CELL. CiLi =3
O
METEOR
id
D
7)
2
jJ
S)
J
D
jJ
r
No.
Ht (m)
nia (m)
Terno. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
Vj_ Div.
V2 Div.
VT Div.
AnC fvoli-O
FOV. (m)
XLT. (m)
Rs (m)
Temp. Factor
s^XC^S??' . °T
bt.WtjUK (ppmj AT
SHELL (ppm)
Note
n i
(2)
(3)
(4)
(5)
(^
(7)
<
/
42-7
"2.-(ff
1 5"0
Bl
3, (of
1 1 * (0
. 2.«^
5"S"
50
Z6,
3^
"2-6J
45Z
1-44
A» «%^i
in . j^v
Yfc?
qoo-
M
4
p^
Z"S-0
IZI f
?r
\t>O
3O
Z3
1 «^
2-^A
1 10
M
5*19
-7/7
<>
*
i ^a'
pfc
/2L)T
«*
H
Z'S'
»
H
»
..
5"6»7
7i"3
H
Ttt\*V
^^fNJ
Pfo
n
/2I*7
it
23
*
>i
^"J9
7/7
^C2L? = /35V DATE f J ' ^ / »7
H
l>
*
"
M
'
Cleo/mj
"^rtt^t>5
SI
I4i0
5"
^^r
^2.
ai
7
1-6)/
4Z3
n
45^f
43Z
"
n
"
O
k
sz
^"0
14ZLV
25
5*V
^\
n-s-
41
2-4/
4V9
«
3^ 5"
5*45"
"
u
1*
2. m\
^5
4-7
144^
M
S'S"
"SO
1^
43
rz/
««
2.^3
3^O
i
H
"
S-f
S--7
I ZOO
*
*S"li
z^*s*
n
sv
J2-
-------
TABLE A
_4 £o*£ )
Sensor Ser. No. "DE/"
J
0
J
1J
u
o
1 VIOQM'PC
O
J
0
o
u
No.
Ht (m)
JDis (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
V^ Div.
V2 Div.
VT Div.
ARP ft/nH-cl
FOV. (m)
XLT. (m)
RTAIci8sE8STl> :~T
F£U|'JK ippm; AT
SHELL (ppm)
Note
n >
(2)
(3)
(4)
(5)
ffil
(7)
4
SO2 REMOTE SENSOR DATA SHEET STACK LOCATION . 4£Vfi8. O*.4?5 }
O REF. CELL CiLl =2
-------
TABLE A-5
Sensor Ser. No. REF. CELL CiLj. =2?00c2L2 - ISS"O DATE &CT: ^!* fef. '
O
CO
METEOR
LU
h-
co
SENSOR
RESULTS I
No.
Ht (m)
nia (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
Vi Div.
V2 Div.
VT Div.
AfiC (volt-cl
FOV. (m)
XLT. (m)
R 6
-->6»
-fq
/o
Z-76
/Z6
M
/ozz
J4/0
xo
/^
*/
ccr
Pfef^
11 T
\"330
.^j
66
"3 5"
5-f-S"
-Z/
Z'7^
( Z^
Ml
95Z
"
«
1*53^
r
66
~3^
"^"^
'
-
4%
ft
£5*y
^(07
w
M
1337
.
-------
TABLE A-5
Sensor Ser. No.
No.
Note
SO2 REMOTE SENSOR DATA SHEET
REF. CELL. CiLi = G2l-2_=_
STACK
DATE. J.tt
Ht fn)
(m)
Temp.
Skv
m
/o
Vis.
7:
Wind (mph/°)
(3)
PLUME
(4)
rr=:
No.
& degree
(5)
Time
i-AOS
Aoerture
*
Div.
(7)
V2 Div.
VT Div.
9£*.>
FOV. (m)
21
XLT. (m)
RS (m)
Temp. Factor
SHELL (ppm)
Notes:
(1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog ;f) , Etc; (2) (3) Xoue whera raker.; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local timei (7) volts frcn meter or divisions from chart. Opt. ...
-------
Sensor Ser. No.
SO2 REMOTE SENSOR DATA SHEET
REF. CELL
TABLE A-5
STACK LOCATION.
No.
Ht (m)
nia (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
\/i n-J «»
V2 Div.
VT Div.
AGC fvolt-O
FOV. (m)
XLT. (m)
RS (m)
Temp. Factor
ITSC^^E^?^ AT
F£U|UK (ppm) AT
SHELL (ppm)
Note
n )
(2)
(3)
(4)
(5)
F^ffD
31-C
^no
nn
_
ft* rt
/9. ?
r^V\
<{^
K'i tf
s /
7.^
^iff-
3\-«r
TI ?
(074
/^-
res'"?.
^«*
*2»T
R^To
/"5I9
r4*P
|^fe ^^^*
*3y
%W
/zi-7
ViV4!
3T
^"L'L
127-^
.^c:
1^*1 'T
-ffc
1*vV3
mo
\3^S"
43
UC))T
Ibll.
/ s*
\3Vt
*
^l
IOH^
/T//
^^-
nn
39
Vh"^4
I4I3
£s£* ^^v
1 * ^ V^^^^*
4o
LO^^
1 ^w?tA £~
T ^r
1*3^
-39
-
fo*24
I4l?>
/'O
/r
tf/
^CL:
Hffo
4)
10^4
\^\\
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog {f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt.
-------
Sensor Ser. No.
SO2 REMOTE SENSOR DATA SHEET
C9L2 =
TABLE A-5
STACK LOCATION.
^
u
S
METEOR
UJ
CO
SENSOR
RESULTS
No.
Ht (m)
nist (m)
Temp. (Oc)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
TiFPff
Aperture
Vi Div.
V2 Div.
VT Div.
AfiC ftml*-«il
FOV. (m)
XLT. (m)
R« (m)
Temp. Factor
§T:A1c1£RE8?TlJ aT
&tjii|UK (ppnu AT
SHELL (ppm)
Note
M)
12)
(3)
(4)
(5)
(f>1
(7)
I
/
Kei
»r
^j.r
21
^3<\.«r
1042.
|437
^-
\4OT_
3i-r
<\n^
I*S42.
/
140"!>
40
105*
-------
TABLE A-6
Sensor Ser. No. P6HO*
o
CO
METEOR
LU
CO
SENSOR
RESULTS
NO.
Ht (m)
rH a (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
V^ Div.
V2 Div.
V.T. Div.
Af«- (vnlhcl
FOV. (m)
XLT. (m)
Re; (m)
Temp. Factor
s'fricl^E^T0 :~T
fLite""1* (ppm) AT
SHELL (ppm)
Note
m
(2)
(3)
(4)
(5)
ffil
(7)
S02 REMOTE SENSOR DATA SHEET STACK LOCATION. .^-VP.^ j
REF. CELL CjLl =63/fc2L2 =315^ DATE £4 J 1 1 / fc^ '
4
/ 5"O
S"*4
1 2.*i
O
IS*
0-C A»T
^
Sfc
*%4
/0S"O
.^J~
frl
*
T<
»4*5'
£$""/
26>V~
|«3
»
M
"
-
se
2j4.*i
MOO
i.
Ln
43
4T
fol
fT'^2
"3t*T
*
q *)4
I"SI*T
-
S^
IT-Y
nm
*
fT^
41
40
-11
^T-70
4^1
\'(D7
T'*TV'
6lf>
45"
^T/
ZfefT
...
1043
I3&0-
4,
-
S(0
-*&
\t\^
-
77
T5"
S"7
4S'
Z6^
*
7£*V
-^
-
.
..
-.
..
SG
34
IZ2O
-
77
*rv
4^
4^
6*5T/
76^"
-
702
..
-
^8
24-3
IZ"SO
"
d?7
47
"5S
(of
V'^Z.
3feS^
6 T7
-
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt.
-------
TABLE A-6
o
METEOR
LJ
CO
SENSOR
RESULTS |
Sensor Ser. No. t>CMO
NO.
Ht fm)
nia (m)
Temp. (°c)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& deqree
Time
Aperture
Vi Div.
V2 Div.
VT Div.
AfiC fvnlt-O
FOV. (m)
XLT. (m)
R.S (m)
Temp. Factor
gtfScl^S!??' -~T
SfrWijUK ippm; AT
SHELL (ppm)
Note
m
(2)
(3)
(4)
(5)
ffil
(7)
«*«&<
REF. CELL C-|Li =(
^
I TO
4T-4
124
O
IS'*'
6/160
SET*
S4»
"^4-
I3Z4
. tg
69
47
3
4-S"
4-tf/
2.6 ^r
I-'SV
4"S- Co
?40
»
-
-
»
-
s^
I "52.^
«
64
47
O
«*
-
-
O
-
-
-
1 ll
V4\1
-
47
4T
13
-
*
-.
**
I7*T
'
~
-
I4I2T
14
k
IT
*
'
»
**
: ZO4
- - - -
- - ..
3»S"C2L3 =tt5£ DATE . . . 3A.I J } J.
-------
TABLE A-6
^
o
£
METEOR
UJ
CO
SENSOR
RESULTS
Sensor Ser. No. "DI5HC
No.
Ht fn)
nia (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
& degree
Time
Aperture
Vi Div.
V2 Div.
VT Div.
Afif (voll-O
FOV. (m)
XLT. (m)
R\
(7)
£- -
SO2 REMOTE SENSOR DATA SHEET STACK LI
> REF. CELL CiLj. =63lVC2L2 =^3^ DATE...
4
\TO
«?-4
IZ4
o
\T
c-io/ieo
ff
ST
Z4i
/^"40
T
A>&
4T
^/
6>l
»
M
-
»
.
-
«.
s\\
l"S'f^
/fc<50
«
60
4Z
-rr
^2-6^
/<97
«"-r«r
64S"
« >
/OZS
»
*
-
*
*
*»
sfe
'54-
i6/r
/-d?
YT
foO
76
«*-«*
rQ
A-C/
ifo«r
«.
/263
-
-
»
%
«
-
66
2.4-3
Ib20
M
T6
20
TZ
fi-O-fe-T
1-21
^^2
sur
I3ZI
»
«
»
*
cwoub-
IM(V
OV6R
S9
\T-T
/£>50
i.
TO
*f
76
i^>
f-r
V-7
46<
te
16U*
DCATION . *r .V /r.r ... 1
.w./a./.fef. .. j
-
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog (f) , Etc; (2) (3) Note where taker.; ;4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A- 7
SO2 REMOTE SENSOR DATA SHEET STACK LOCATION .t
Sensor Ser. No. "DOIO REF. CELL CiLi ^iSiCsL? =2T*7 DATE. . .2.7 J. L L 1.
o
METEOR
UJ
H
co
SENSOR
RESULTS
No.
Ht (m)
nia (m)
Temp. (°C)
Skv
Vis.
Wind (mph/°)
PLUME
No.
fi deqree
Time
Aperture
Vi Div.
V2 Div.
VT Div.
A«r fvoH-Ql
FOV. (m)
XLT. (m)
Re, (m)
Temp. Factor
S&kc^^ES?!?' AT
F£I^UK (ppm) AT
SHELL (ppm)
Note
m
(2)
(3)
(4)
(5)
(fil
(7)
4-
/ £"0
'iT'^f
?^f
IQ
4*N|
T£6*T
SHOW
MMMB
NIWM
Z.4.
I"i42.
.<"
^^1 ^2
j*^^%
^2.
^6-^'V"
370
620
«
»
-
»
V4VOZ.
2.0
l^TO
t%
77
/O
^7
V-6
440
i"67
-
*
» *
MM07.
2.0
l^tOO
*«
77
f.fi
60
440
6(94
-
*
*
-
4
M\N)Z.
2.O
KlO
7S"
T
*~2-
^2
440
^7.9
-
«
-
i*
-
H\K)3
IS*
\4-10
-
73
r?
4o
C-*t
no
403
-
-
-
NVO^
n-T
\4^O
i»
7^
^TT
4T/
^4
730
£"3D
70
^TZ.
4f
6-f
*TTO
S"7o
.
NVJ4-
M-T
VTOO
(b7
^T"/
59
^3
-
730
463
VPS.....
-
. ..
-
NVJ1^
?T
\eg\o
*
&r
^T5
*5^r
^5"
w v ^r
TI ^^%
^^^M
^^^^^^^^^^^^^^^^^^^^^MMMM^^Hl^taMBBMMBi^^^HMMnVHBM^M^HIBMHHHHi^HI^BHMBi^HBAHHI^VMBI^HHMMritaMiMMa^
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) ixote where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt.
-------
TABLE A-8
Sensor Ser. No.
feGHO,
SO2 REMOTE SENSOR DATA SHEET
REF. CELL L =311^CL =2T57
STACK LOCATION.
DATE.HfiV. 2T
No.
Note
T
Ht (m)
\-5-0
1TO
IS'O
Ib'O
is-o
(m)
Temp. (°C)
Skv
(1)
10
70
ID
Vis.
(2}
IZ mi
Wind (mph/°)
(3)
PLUME
(4)
No.
S\0
.Sil
se
ss
& deqree
5
ir-i
n-**
2.4-3
Z4-5
Itf't
Time
ffil
IO17
tozv
10 5T
1105"
HOT
l\\T
UZO
Aperture
Div.
(7)
2.- fen
I'd*
7.-T
V2 Div.
%-zz
VT Div.
2-64
(vnlt-cl
c-z
4T-2
FOV. (m)
1-O2.
6,1
fel
74,
XLT. (m)
jo.
(m)
fe40
Temp. Factor
i
ppm) AT
SHELL (ppm)
IOSO
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f>, Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A- 8
Sensor Ser. No.
No.
Note
SO2 REMOTE SENSOR DATA SHEET
REF. CELL dLi
STACK LOCATION.!
DATE.ffgy. 2£..&9.
Ht (m)
ISO
13T
(m)
M
it
it
Temp. (°C)
Skv
n
o
o
Vis.
(2)
Wind (mph/°)
(3)
-nor
PLUME
(4)
No.
KH
HZ
mo
HlO
& degree
(5)
2.1-0
ITO
Time
ffil
1107
IZ1T
Aperture
Div.
(7)
4-31
4-OZ
5-11
li-TI
V2 Div.
2.-T&
VT Div.
4-Ofc
5-54-
ACC.
FOV. (m)
4T
XLT. (m)
(o-^O
T-1T
(m)
ACS'
(70
Temp. Factor
it
~T
612.
151.
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
Sensor Ser. No. T>Ct\
No.
Note
SO2 REMOTE SENSOR DATA SHEET
REF. CELL CiLj «3flSSC2L;> =21^1
TABLE A-8
STACK LOCATION
DATE
\!>T
\S'0
I TO
14\
nia (ml
Temp. (°C)
Skv
m
Vis.
(21
IS
Wind (mph/°)
(3)
Guarr
PLUME
(4)
No.
HlO
HU
HII
HI
NT
H7
NIA
HIO
& degree
15)
ir-n
31-
iz-n
Time
\4lfe
Hit
V410
144^
Aerture
Div.
(7)
Vfci
V3T
V2 Div.
z-r?
2.-TT
2-Ti
1-Co
VT Div.
4-12.
2-T2.
1-73
AftC
s-z
5-2.
FOV. (m)
TT
v-zf
XLT. (m)
6-30
4-30
(m)
4TO
r\o
.115.
rrr
4TO
feTO
TIO
Temp. Factor
J32^
Til
ITT
ppm) AT
SHELL (ppm)
Notes: (1) Cloud Cover in lOths, Smoke (k), Haze (h), Fog (f), Etc; (2) (3) Note where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time; (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A- 8 V?
SO2 REMOTE SENSOR DATA SHEET STACK LOCATION.!
Sensor Ser. No. T>MO REF. CELL CiLi =*MllC2L:> =XBM DATE. . NfiV. A?
M
o
£
METEOR
LJ
CO
SENSOR
| RESULTS 1
No.
Ht (m)
nia (m)
Temp. (Oc)
Skv
Vis.
Wind (mph/°)
PLUME
No.
_& deqree
Time
Aperture
Vi Div.
V2 Div.
VT Div.
AftC fvoH-csl
FOV. (m)
XLT. (m)
RnV\U«
H\l
T--V*
»«>OZ
O«^*
"5-ZT
7«^i
^foS*
6"-TJ
l-CK
S-^Ci
qio
\*1A
fclt4
\?0
m
LI.
\<»\2.
«
i-T
Z-6.C
*-T
«--4
47
(.-^0
a.yy
«
TZ(«
V41
»*
*
M\A
\5-5
15 X^
^35
2.-C.O
S---6
S-<*
rs
?>(»0
4TO
««
Tfe*
-1120
\^»
*»
HVO
\Z-V1
\5%?
3-2.
2-5
2-75
d-2.
I-OQ
?.«?i
fe"70
»
^^
\^7J
H\\
XO-'S'?!
I54S
«*
3- 15
Z--AB-
^.foO
4-5
i-if '
?-49
no
-
67^
151
*#
N\7_
T-M
I5S4
«
Z-IO
^
~J* ) t
nVPS Y..T*
.**. l
Notes: (1) Cloud Cover in lOths, Smoke (k) , Haze (h) , Fog (f) , Etc; (2) (2) Nc^e where taken; (4) Sketch elevation and
plan (5) Hand-held Abney, (6) 2400 hr. clock local time, (7) volts from meter or divisions from chart. Opt. ...
-------
TABLE A-9
V. Short Range
Range
(meters)
265
Weather
(Condition-Visibility
in Miles)
Clear 15
285
265
Short Range
Medium Range
Long Range
360
365
405
365
370
465
11
480
II
550
550
465
465
440
440
650
670
Clear 15
Clear 12
II
O'cast 12
Clear 12
Clear 15
Clear 12
O'cast 12
O'cast 4
Clear 15
Clear 12
O'cast 12
II
O'cast 4
Clear 15
Clear 12
V. Long Range
640
580
810
II
970
O'cast 12
O'cast 4
Clear 15
Error
Remote-Shell
+ 100%
- 18.4
Average
Error
23
37
19
18
10
16
27
33
26
38,
12,
39,
.5
.4
.7
.2
- 34.4
- 20.1
- 29.0
+ 27.2*
- 14.
- 20.
- 22.
- 42.
- 37.2
- 43.6
- 37.3
- 37.4
- 18.4
- 44.8
- 22.5
- 27.0
- 36.
- 34.
-34.8
- 43.6
- 49.1
- 40.7
- 31.2
- 14.1
- 21.3
- 7.3
- 33.5
- 32.8
- 46.2
- 61.2
- 45.2
- 38.2
- 38.5
- 41.8
- 39.0
.5
.2
- 23.0
25.8
36.8
- 21.0
- 36.4
- 37.3
- 31.1
3 - 3,
f - 44.
f- 23.4
-46.2
,2
} - 33.:
- 39.4
-------
TABLE A-9 (Cont'd)
V. Long Range
Ranqe
730
II
880
Weather
O'cast 4
Error
48.6
55.5
63.1
Average
Error
f - 55.7
* Inconsistent
-------
APPENDIX B
TABLE .B.I
RUN IDENTIFICATION, FLIGHT 10
Run
1
2
3
4
5
6
7
8
9
Direction
From/To
E/W
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
Altitude
MSL
5500
5500
5500
5500
3500
3500
3500
2500
2500
Altitude Direction
Ground
4500 Up
Dn
Dn
Up
2500 Up
Dn
Dn
1500 Up
Dn
Ave
ppm-m
Value
10
60
55
20
90
70
40
140
TABLE B.2
RUN IDENTIFICATIONS FLIGHT 11
un
1
2
3
4
5
6
7
8
9
10
11
12
Direction
From/To
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
Altitude
MSL
6000
6000
6000
6000
5000
5100
4000
3800
3000
3000
2000
2000
Altitude
ft . above
ground
5000
5000
5000
5000
4000
4100
3000
2800
2000
2000
1000
1000
Direction
Up
Dn
Dn
Dn
Up
Dn
Up
Dn
Up
Dn
Up
Dn
Ave
ppm-n
Value
-
15
25
25
-
45
45
75
65
90
75
110
-------
TABLE B.3
FLIGHT 11 - AVERAGE VALUES
Col. 1
Height
above ground
Col. 2
D'Ward
Col. 3
Upward
ppm-m
Col. 4
2-3
gas below
A/C
5000
4000
3000
2000
1000
25
45
75
90
110
0
-
45
65
75
25
-
30
25
35
TABLE B.4
FLIGHT 12 RUN IDENTIFICATION
Run
1
2
3
4
5
6
7
8
9
10
Direction
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
W/E
E/W
Altitude
MSL
6500
6500
6500
6500
6500
3000
3000
3000
3000
Altitude
Ground
5500
5500
5500
5500
5500
2000
2000
2000
2000
Direction
UP
Dn
Up
Dn
Dn
Dn
Up
. Dn
Dn
Up
Ave
ppm-
38
42
55
51
18
77
89
29
TABLE B.5
RUN 12 - AVERAGE VALUES
Col. 1
Height
above gd.
5,500
2,000
Col. 2
D ' ward
ppm-m
44
83
Col. 3
Up'wd
ppm-m
0
23
Col. 4
2-3=
gas below
A/C
44
60
Col. 5
4/'m'
18
25
Col. 6
4/'m' *
Ppb
3.3
12.5
1
-------
REFERENCES
1. Moffat, A. J., Robbins, J. R., Watts, R. M., Millan, M. M. - The
Applications of Optical Techniques to the In-Stack Measurements of
SO Using Incandescent Light Sources and the Remote Sensing of S0_
Plumes Using Sky Light. Symposium on Advances in Instrumentation for
Air Pollution Control at Cincinatti, May, 1969.
2. Millan, M. M., T.ownsend, s. T., Davies, J. H. - Study of the Barringer
Remote Sensor. UTIAS Report No. 146. 1969.
3. Shell Oil Company,Thornton Research Center. UK. Internal Report on
Shell-Thornton Method of Determination of SO_ and SO_ in Stack Gases.
4. See Reference 1.
5. See Reference 1.
6. Bosanquet, Gary,Halton. Inst. of Mech. Eng. Paper 1949.
7. Pasguill, F. Atmospheric Diffusion: The Dispersion of Windborne
Material from Industrial and other Sources. D. Van Nostrand, London
1962.
8. Leighton, Photo Chemistry of Air Pollution, P. 184 et seq.
-------
CORRECTION SHEET
Barringer Research Report No. TR69-113 entitled:
"A Report to Department of Health, Education, and Welfare of Optical
Measurements of
Spectrometers".
Measurements of SO and NO Air Pollution Using Barringer Correlation
ERRATA
Page 88: - Figure 5.9 should read Figure 5.9 (a)
" 90: - Figure 5.6 should read Figure 5.9 (b)
" 90: - Figure 5.7 should read Figure 5.9 (c)
" 91: - Figure 5.8 should read Figure 5.9 (d)
" 98: - Figure 5.6 should read Figure 5.9 (b)
" 101: - Figure 5.9 should read Figure 5.9 (a)
The correct Figure 5.6 and Figures 5.9 (c) and 5.9 (d)
are attached.
Page 70: - Equation (2) should read
Z=°°
BZ = JZ=oX ^
85: - Line 8 should read " 3400°K"
" 88: - Correlation Coefficient should be 0.905
" 102: - Lower graph vertical scale should be "ppb SO "
The dotted line represents the LPAM readings.
" 108: - Line 9 should read "....10 pphm...."
" 130: - Vertical axis is NO concentration with a scale
of 1 cm = 10 ppm-m. (Curves are displaced relatively
for clarity). Horizontal axis is distance in miles
(scale unknown).
11 134: - Line 20 should start "NO2 pphm "
-------
30 -
25
20 -
2
f 15
10
5 -
© LPAM 300M
PSAM
PERIOD M00-I300hrs.
26-DT-69
-300
-250
-200
-150
-100
-50
I3OO
1230
1200
TIME
1130
1120 1110
1100
COMPARISON OF LONG PATH AND POINT SOURCE SAMPLERS
FIG. 5-6.
-------
SCATTER DIAGRAM.
VARIABLE l'
-5.7143 29.5238 64.7619 100.0000 135.2381 170.4762 205.7143
210.0000 + + + + + + + 210.0000
173.3333 +
V
A
R
136.6667 +
I
A
B
L 100.0000 +
+ 136.6667
173.3333
V
A
R
I
A
R
f 100.0000 L
E
63.3333
63.3333
26.6667
4-«"»-+ -f + -f »
4-
o
26.6667
10.0000 + + + + -f + + -10.0000
-5.7143 29.5238 64.7619 100.0000 135.2381 170.4762 205.7143
VARIABLE 1
SCALING FOR VARIABLE 1
SCALING FOR VARIABLE 2 --
INTERVAL SPACING MINIMUM MAXIMUM
5.873016 0. . 200.0000
5.238095 0. 200.0000
FIGURE 5.9 (c)
-------
SCATTER DIAGRAM
VARIABLE 1
-6.5714 33. OS?1* 74.4762 115.0000 155.5238 196.0476 236.5714
V
A
R
I
A
B
L
E
2
*
199.3333 +
.
;
157.1667 *
115.0000 4-
72.8333 +
: ©
0 o
30.6667 + 0
0
: ©°^n °°
! '
-6.5714 33.9524
SCALING FOR VARIABLE
SCALING FOR VARIABLE
0
+ 199.3333
© :
4. 157.1667
.
.
4- 115.0000
4- 72.8333
0 !
+ 30.6667
X
74.4762 115.0000 155.5238 196.0476 236.5714
VARIABLE 1
INTERVAL SPACING MINIMUM MAXIMUM
1 6.753968 0. 230.0000
2 6.023809 0. ' 230.0000
V
A
R
I
A
B
L
E
2
FIGURE 5.9 (d)
------- |