-------
Figures (Continued)
No. Page
65 Water Profile After Breaking 112
66 Plastic Ring in Clean Water Before 114
Adding Distillate
67 Navy Distillate Being Added to 114
Water Inside Ring
68 Patches of Distillate Sheen 115
69 Distillate Being Stirred for 115
Uniformity
70 Uniform Distillate Sheet 116
71 Uniform Distillate Sheen in Presence 115
of Waves
72 Plastic Ring in Clean Water Before 117
Adding Diesel Fuel Oil No. 2
73 Diesel Fuel Oil No. 2 Being Added 117
to Water Inside Ring
74 Diesel Fuel Oil No. 2 Spread Evenly 118
Inside Ring
75 Diesel Fuel Oil No. 2 Sheen in 118
Presence of Waves
76 Plastic Ring in Clean Water Before 119
Adding Diesel Fuel Oil No. 4
77 Diesel Fuel Oil No. 4 Being Added to 119
Water Inside Ring
78 Patches of Diesel Fuel Oil No. 4 Sheen 120
79 Diesel Fuel Oil No. 4 Spread Evenly 120
Inside Ring
80 Florida Crude Oil Being Added to 121
Water Inside Ring
81 Florida Crude Oil Blown into Upper Right 122
Corner by Wind
82 Small Portion of Florida Crude Oil 122
Removed
83 Thick and Thin Spots of Florida Crude 123
Oil Undispersed Over Entire Ring
84 Temporary Florida Crude Oil Sheen 123
After Disturbance
IX
-------
Figures (Continued)
No. Page
85 Lumps of Florida Crude Oil 124
Coalescing Into Bright Spots
86 Florida Crude Oil Sheen After 124
Removal of Most of Oil
87 Oil-Water Test Configuration 126
88 Temperatures of Water, Oil, and 127
Air Over a Day's Time and the
Corresponding Variations in Presentation
89 Florida Crude Oil on Water, , 9R
1800 Hours, 1/18/72
90 Florida Crude Oil on Water, 0600 Hours, 129
1/19/72
91 Florida Crude Oil on Water, 0630 Hours, 130
1/19/72
92 Florida Crude Oil on Water, 0650 Hours, 131
1/19/72
93 Florida Crude Oil on Water, 0700 Hours, 131
1/19/72
94 Florida Crude Oil on Water, 0730 Hours, 132
1/19/72
95 Florida Crude Oil on Water, 0900 Hours, 132
1/19/72
96 Florida Crude Oil on Water, 1100 Hours, 133
1/19/72
97 Florida Crude Oil on Water, Noon, 133
1/19/72
98 Some Florida Crude Oil Being Blown 134
from Ring
99 Florida Crude Oil on Water, 1300 Hours, 135
1/19/72
100 Seaweed Sample 135
101 Plywood in Water 136
102 Plywood in Water After One Week 136
103 Oil Volume Test Setup 138
104 No Oil as Presented on Television 140
display
x
-------
Figures (Continued)
No.
105 No Oil as Presented on Oscilloscope 141
106 2 Drops of Florida Crude Oil as 141
Presented on Television Display
107 2 Drops of Florida Crude Oil as 142
Presented on Oscilloscope
108 4 Drops of Florida Crude Oil as 142
Presented on Television Display
109 4 Drops of Florida Crude Oil as 143
Presented on Oscilloscope
110 8 Drops of Florida Crude Oil as 143
Presented on Television Display
111 8 Drops of Florida Crude Oil as 144
Presented on Oscilloscope
112 2 Cubic Centimeters of Florida 144
Crude Oil as Presented on
Television Display
113 2 Cubic Centimeters of Florida 145
Crude Oil as Presented on Oscilloscope
114 5 Cubic Centimeters of Florida Crude 145
Oil as Presented on Television Display
115 5 Cubic Centimeters of Florida Crude 145
Oil as Presented on Oscilloscope
116 Apparent Temperature Versus Oil 147
Thickness
117 Reflectivity and Emissivity of 148
a Smooth Water Surface as a
Function of Slant Angel
118 . 2 Ounces of Florida Crude Oil 148
at 43 Degrees Slant Angle
119 2 Ounces of Florida Crude Oil 149
at 60 Degrees Slant Angle
120 2 Ounces of Florida Crude Oil 149
at 70 Degrees Slant Angle
121 Waves at a Slant Angle of 150
86 Degrees
XI
-------
Figures (Continued)
No.
122 Galveston Bay as Seen From the 150
Flagship Hotel
123 Galveston Bay at Slant Angle of 151
70 Degrees
124 Galveston Bay at Slant Angle of 152
70 Degrees
125 Signal Versus Scan Radius for MOD 2 156
a5 50-Foot Altitude
126 Signal Versus Scan Radius for MOD 2 157
at 100-Foot Altitude
127 Signal Versus Scan Radius for MOD 2 158
at 150 Foot Altitude
128 Signal Versus Scan Radius for MOD 3 159
at 50 Foot Altitude
129 Signal Versus Scan Radius for MOD 3 160
a5 100 Foot Altitude
130 Signal Versus Scan Radius for MOD 3 161
at 150 Foot Altitude
131 Concept for Arc and Area Scans 163
132 Gimbal System for Active Sensor 164
133 Scanner Block Diagram 166
134 Area Scan Concept and Display 167
135 Shipboard Application of Passive 175
Scanner
136 Conceptual Line Scanner Approach 176
to Oil Spill Surveillance System
137 RS-18 Outline Dimensions 178
138 Spectral Radiant Emittance Curves for 138
Blackbodies at Several Absolute Temperatures
139 Liquid Transfer System and Passive 182
Sensor Mounted on Tower
140 Transfer Piping System 183
141 Thermal Conductivity Linde Super 191
Insulation
142 Schematic of Proposed Thermoelectric 195
Cooler
-------
Figures (Continued)
No. Page
143 Input Power Versus Heatsink 196
Temperature for a 0.05-Watt Load
144 Cold Side Temperature Versus 197
Hot Side Temperature for 0.05-Watt
145 Oil Spill Imagery, 3- to 5.5-Microns 198
146 Schematic Representation of an Oil 204
Separator Facility
147 Onshore Oil Separator Facility 205
Xlll
-------
TABLES
No. Page
1 Characteristics of Circuit Elements 31
2 Transmitter 35
3 Characteristics of Tungsten Halogen 36
Lamp
4 Characteristics of Xenon Lamps 41
5 Detectors 47
6 Computer Signatures Versus Wavelength 55
7 Atmospheric Effects 55
8 Signatures for Six Types of Oil and 56
Water
9 Signatures for Various Types of Debris 56
10 Omitted
11 Identification of Oil Types by Observed 125
Characteristics
12 MOD 2 Coverage Tradeoff, PbSe Detectors 169
at 243°K
13 MOD 2 Coverage Tradeoff, PbSe Detectors 169
Cold Filtered at 193°K
14 MOD 3 Coverage Tradeoff, PbSe Detectors 170
at 243°K
15 MOD 3 Coverage Tradeoff, PbSe Detectors 170
Cold Filtered at 193°K
16 RS-18 Specifications 179
17 Dewar Characteristics 184
18 Basic Cooler Specifications 194
19 Oil Spill Sensor Application Matrix 206
xiv
-------
SECTION I
CONCLUSIONS
1. The advantages and capabilities of the active infrared
(IR) oil detection sensor are as follows:
a. Absolute distinction between petroleum oil and
other natually occuring substances such as kelp
beds, salinity and temperature variations, and
debris.
b. Suitable for scanning applications with coverage
to 300 feet from the sensor.
c. Sensor sensitivity is capable of quick detection
of an oil film on the water's surface.
d. Sensor output is suitable for transmitting
an alarm.
e. Successful operation is independent of the
time of day, weather, and water surface conditions,
f. Highly reliable and easy to maintain.
g. Relatively low initial cost and low operating
cost.
h. Can be remotely installed on a- fixed structure
on land, or on an offshore platform.
The advantages and capabilities of the passive IR oil
spill imagery system are as follows:
a. Detection of thermal anomolies associated with
oil spills.
b. Suitable for scanning applications with coverage
to 1000 feet from the system.
c. System output is suitable for transmitting an
alarm, making a film record, displaying on a
monitor, and providing the possibility of oil
spill volume determination.
d. Successful operation is independent of weather
and water surface conditions, and except for a
short period of time at sunrise and sunset is
independent of the time of day.
e. Highly reliable and can be designed for easy
maintainability.
f. Relatively low initial cost and low operating
cost.
-------
g. Can be installed on a fixed structure on land,
on an offshore platform, or in an aircraft.
3. A combination of the active and passive systems can
provide highly reliable oil spill surveillance.
-------
SECTION II
RECOMMENDATIONS
The primary purpose of the evaluation phase of this study program was
to determine the ability of existing active and passive sensors to detect
oil spills. Each sensor demonstrated distinct advantages in oil spill sur-
veillance applications. The active sensor provided absolute oil spill detec-
tion that was not affected by the time of day, weather, or water surface
conditions. The passive sensor demonstrated the possibility of oil volume
determination, ability for a real coverage, and long-range detection.
Since both evaluated sensors demonstrated distinct and equally impor-
tant advantages, a combination active/passive sensor system is recommended
for permanent oil spill surveillance applications. Two systems with these
advantages are described in Sections IX, X, and XI.
Specifically there are several options available in most applications.
The active sensor can be used as a staring system or a scanning system. It
also has two recommended configurations in terms of range; Mode 2 and Mode 3,
where Mode 3 is for extended range. The passive system also has two configu-
rations; one is when the 8 to 14 microns wavelength region is used and the
other is the 3 to 5 microns wavelength region. In either configuration the
system can be used as a line scanner or as an areal framing ..system. The 3
to 5 microns system is the lower performance system but is simpler in terms
of logistics. The following list of recommended sensors for reasonable
applications is taken from Table 19.
o Estuary: Active Mode 3 (Arc Scan W/Alarm) or Passive
3-5 microns (Line Scan W/Alarm)
o Industrial Effluent: Active Mode 2 (Staring W/Alarm)
or Passive 3-5 microns (Line Scan W/Alarm)
o River/Bay Area (Tower): Passive 8-14 microns (Area
Scan CRTI film) or Passive 3-5 microns (Area Scan
CRTI film)
o Drainage Channel <60^) Across: Active Mode 2 (Arc
Scan W/Alarm)
o Drainage Channel <: 150^1) Across: Active Mode 3 (Arc
Scan W/Alarm)
o Drainage Channel ^ 150^) Across: Passive 3-5 microns
(Line Scan W/Alarm)
o Offshore Drilling Platform: Active Mode 2 or Mode 3
(Arc Scan) or Passive 8-14 microns (Area Scan CRT 1
film)
o Offshore Sperator Platform: Active Mode 2 (Staring or
Arc Scan W/Alarm) or Passive 8-14 microns (Line
Scan W/Alarm)
-------
o On Shore Separator Platform: Any of the options with Alarm
o Bridge: 8-14 microns Passive (Line Scan W/Alarm)
o Dock Area: Active Mode 2 or Mode 3 (Arc Scan W/Alarm) or
any of the Passive.
-------
SECTION III
INTRODUCTION
The primary objective of this study program was to demonstrate
a remote sensing oil spill surveillance system for real time
detection, alarm, monitoring, and recording of oil spilled
or discharged in harbors, lakes, rivers, estuaries, and near
shore coastal waters. In addition, a complete thermal imagery
and oil detection system, data analysis and oil spill alarm
concepts were developed.
The demonstration phase of the study program was centered
around two existing systems developed by Texas Instruments
Incorporated; an active infrared (IR) sensor system for detec-
tion of oil on water and a passive IR imaging system for areal
surveillance. The concept development phase was optimized
by existing equipment using new techniques and sensors.
The active IR sensor system was designed explicitly to
distinguish petroleum oil from water. Basically, the system
emits a wide-band IR pulse and receives reflections at two
narrow spectral bands to perform its function. The theory
of operation is based on differences in the reflectance of
water and oil at these narrow spectral bands (wavelengths).
The passive IR imaging system was developed by Texas
Instruments Incorporated for the U.S. Army Night Vision
Laboratory as a night weapon sight. This system receives IR
radiation in the 8-to-14 micron spectral region. Since there
are thermal, as well as emission differences between water and
oil, oil can be detected by this system during the day or
night. Although the passive IR imaging system cannot positively
identify the presence of oil, it can complement the active IR
sensor system by providing real-time areal surveillance with
the added possibility of determining quantity of oil spilled.
Tests of the passive and active IR systems at the Texas
Instruments and Galveston, Texas test sites during the
demonstration phase have verified the capabilities of these
systems to detect oil on the water's surface. Furthermore,
practical limits on the variables of detection and observation
have been determined. The following sections provide the
results of these tests and the resultant system design concepts.
-------
SECTION IV
DEMONSTRATION PROGRAM
The Oil Spill Surveillance System study consisted of two
parts: (1) a demonstration program using existing sensors
and (2) concept development. The specific requirements for
the demonstration program were to test two different sensors
to determine their feasibility and to demonstrate the
application of these sensors to the detection and surveillance
of oil spills on water surfaces. To accomplish this, both
the active oil detection system and the passive surveillance
system were simultaneously evaluated; hereafter, these
systems shall be referred to as "active sensor" and "passive
sensor", respectively.
Test Site
Texas Instruments approach was to construct a tower adjacent
to a suitable body of water. The nature of the study required
the creation of oil spills that could be contained and cleaned
up.
A stock pond located on Texas Instruments property near
Renner, Texas was chosen as the test site. This location
offered one definite advantage over other lakes investigated.
Since the lake was on private property and was self-contained
(ie, fed into no streams or waterways) testing could be
accomplished without ecological damage to property or wild
life.
The pond used was approximately 300 feet by 200 feet. A
tower was erected with platforms at the 50- and 30-foot levels,
Figure 1 shows the location of the tower and a portion of the
pond. A portable building, 20 feet by 10 feet, was
moved to the location (Figure 2) to provide shelter for
monitoring and recording equipment. Figure 3 shows some of
the test equipment used to evaluate the active sensor, and
Figure 4 shows test equipment used to evaluate the passive
sensor. The Gulf of Mexico was used for long range tests.
Active Sensor Demonstration Program
The four objectives of the active sensor demonstration
program were as follows:
1. Prove that the oil-water discrimination technique
works on natural wave surfaces.
-------
Figure 1. Test Site Tower
8
-------
Figure 2. Test Site Building
-------
Figure 3. Active Sensor Test Equipment
Figure 4.Passive Sensor Test Equipment
10
-------
2. Prove that the technique works for a representative
group of oils and that it does not cause a false alarm when
common debris or marine growth is detected.
3. Establish and verify the range and angular scan
capabilities of the technique using state-of-the-art
components at a test site where wave action is unified.
4. Verify the range and angular scan performance of
the instrumentation in a rough wave environment.
The results of this demonstration program has made possible
the extrapolation of system performance for essentially any
range and angle, when the surface condition is known. This
program implemented three different forms of the active
sensor. In each successive modification, the sensor had an
increased signal-to-noise for greater range and/or angle.
Another important feature of the demonstration program was
its conclusiveness. By the time the first three objectives
had been achieved there was no need to perform the fourth
phase of the program; ie, to spill oil. During the first
three phases the independence of the signature from wave
action was established. The only condition being that there
was a sufficient signal-to-noise ratio as determined by the
type of wave action, range, and angle. Hence, at the Gulf
of Mexico test, there was no need to pollute the water, only
to measure the signal-to-noise ratio for the above three
parameters.
Other operational features of the system were not considered
major because of the spectral region in which the system
operates but were confirmed during the demonstration. Tests
confirmed that active sensor operation is independent of
time of day, humidity, rain and cloud cover.
During the course of the demonstration program the primary
objectives was to make the conceptual system inexpensive in
mass production, reliable, and low in maintenance by using
state-of-the-art components, and experience gain from the test,
Passive Sensor Demonstration Program
The demonstration program for the passive sensor was developed
to establish the practical limits of observation of oil spills
using the 8- to 14-micron spectrum.
The choice of detector material used for the test sensor
was based on the atmospheric effects on infrared energy and
Wein's Displacement Law governing the wavelength of maximum
11
-------
emission for targets at ambient temperature.
Figure 5 is a plot of percent transmission versus wavelength
for horizontal paths of IR transmission. This curve shows
that energy from 3 to 5 microns and from 8 to 14 microns is least
attenuated by the atmosphere; therefore, a detector material
that operates at one of these ranges of wavelengths provides
optimum detection of oil.
The temperature of oil spills is approximately the temperature
of water and water temperatures vary from 274°K to approximately
303°K (32°F to approximately 65°F).
Wein's displacement formula gives the relationship between
temperature of a blackbody and the wavelength of its
maximum thermal emission. Converting to degrees Kelvin:
T A max = 2900
(1)
where temperature (T) is in °K and the wavelength (Amax)
is measured in microns. From formula (1) and the temperature
0.8
(2 .OOO YARD PATH WITH I 7-MM PREC(PITABIE WATER)
0.6
in
VI
2
I/I
§0.4
l-
0.
(A
O
(.0381
579
WAVELENGTH i \)' MICRONS)
13
IS
Figure 5 Atmospheric Transmission
12
-------
limits of water, the range of X for maximum emission is
from 9.6 to 10.6 microns.
Thus, the choice of detectors in the 8- to 14-micron range is
more appropriate for oil spill detection than the 3- to
5-micron range.
Tests performed encompassed basically three sets of variables:
(1) environmental conditions, (2) oil film conditions, and
(3) sensor configuration.
Environmental conditions included background (water) temper-
atures ranging from approximately 32°F to 65 F, and ambient
(air) temperatures ranging from approximately 30°F to 85°F.
Both day and night background conditions were examined as well
as various sky backgrounds varying from clear to heavy over-
cast and rain. The second variable considered was oil film
conditions. Various types of oil and thicknesses were observed
during the demonstration program. Temperature signatures of
equal amounts of several oil types, ranging from thin Navy
Distillate to heavy crude oils, were recorded. Radiometric
temperatures of oil with respect to water as a function of
oil film thickness was also investigated. Both field and
laboratory tests, with constant environmental conditions,
were conducted to study the effects of thickness.
Various aspects of the sensor were continuously evaluated to
determine the optimum configuration for maximum sensitivity
and probability of detection. Various modulation levels of
the light emitting diode array were examined. The closed-
circuit TV data line was also investigated for optimization.
The standard vidicon image tube was replaced with a Tlvicon*
tube which has peak sensitivity at the wavelength of the
GaAsP diodes used in the passive sensor. Also, angles of
observation variations were considered in sensor configuration.
Vertical angles from 0 to 70 degrees were recorded.
Imagery from each of the tests was recorded on magnetic
recording tape for later playback and analysis. Photographic
prints were made of the imagery and many of these are
presented in Section VIII of this report.
Long Range Tests
Long range and maximum angle of observation tests were not
possible from the 50-foot tower at the Texas Instruments
test site; therefore, both sensors were taken to Galveston,
Texas. Testing in the Gulf of Mexico allowed both systems
to be evaluated at a height and in an environment similar to
those where an eventual operational system would be installed.
*Texas Instruments Incorporated Trademark
13
-------
Selection of a site to conduct these tests was based on
three desirable characteristics: (1) a minimum height of
100 feet, (2) heavy wave action (2-to 3-foot waves), and (3)
the possibility of an occasional oil slick. Of the three
desired test site characteristics, height and wave action
were considered the most important. Since both systems had
detected oil at 50 feet, the presence of an oil slick was
not necessary. The Houston Ship Channel was considered because
of the presence of oil but was rejected because it is usually
calm. Several locations such as towers, bridges and oil rigs
in various Texas cities were considered, but little encouragement
was obtained in securing permission for use of those sites.
The location selected was the 7th floor of the Flagship Hotel
in Calveston, Texas. Two adjacent rooms were rented and the
equipment was mounted on the terraces overlooking the Gulf
of Mexico from a height of 91 feet. The hotel is built on a
wharf extending into the Gulf and is about 100 yards from the
shore line.
During the test period, the Gulf had a continuous ripple of
at least 3 inches in height. From 7:00 am to 3:00 pm on
most days, the waves were approximately 2 feet high with a
period of 15 feet and had a wave velocity of 5 seconds per
cycle.
Photographs of the Galveston test site and equipment set
up are shown in Figures 6 through 10. Figure 6 shows the
active sensor mounted on the terrace and Figure 7 shows the
associated recording equipment. The passive sensor mounted
on the adjacent terrace is shown in Figure 8. The monitoring
test setup for the passive sensor is shown in Figure 9. A
side view of the hotel with the units mounted is shown in
Figure 10.
The Galveston site test results are presented in Sections VII
and VIII.
Figure 6. Active Sensor Test Setup
14
-------
Figure 7. Active Sensor Recording Equipment
Figure 8. Passive Sensor Test Setup
15
-------
Figure 9. Passive Sensor Monitoring Equipment
Figure lO. Galveston Test Site
16
-------
SECTION V
DEMONSTRATION SYSTEM, ACTIVE SENSOR
The theory of operation of the active sensor is based on the
difference in the reflectance of water and oil at two wave-
lengths in the 3- to 4-micron region. This spectral region
was selected because of the presence of a universal hydrocarbon
absorption band at 3.4 microns. The demonstration models
tested on this program were developed in three phases herein
designated MOD 1, MOD 2 and MOD 3. A photograph of MOD 1 is
shown in Figure 11. This sensor is a laboratory unit that
was implemented prior to this contract and modified for field
operation at a 30-foot height. The components of MOD 2 are
very similar to MOD 1 with the exception that refractive
optics are used instead of reflective optics and a higher
f/number was used to extend the range. A photograph of MOD 2
is shown in Figure 12. MOD 3, shown in Figure 13, uses a
pulsed Xenon arc source rather than a 400- hertz modulated
beam from a tungsten halogen/lamp.
Figure 11.Active Sensor, MOD 1
17
-------
t
Figure 12. Active Sensor, MOD 2
Figure 13. Active Sensor, MOD 3
18
-------
System Theory of Operation
The active sensor is a small, lightweight unit that has been
operated remotely from a tower. This sensor was designed to
discriminate oil from water, debris, etc. The sensor is
inexpensive and easy to maintain, has a remote alarm
capability, and can be operated day or night. ,Sensor
operation is independent of weather, and water surface state.
In addition, the sensor can be operated at a height of up to
300 feet above the surface.
Each of these capabilities is discussed in the following
paragraphs. The ability to discriminate between oil and water
can best be understood as a signature effect which is
equivalent to the remote alarm capability. This is explained
in detail.
The active infrared (IR) oil detection sensor is inexpensive
and will require little maintenance. As originally conceived
it was a staring system, ie, one which has a single resolution
element. However, it has been demonstrated to be capable of
operating in a scan mode for area coverage (see Section VII).
The detectors operate at ambient air temperature; therefore,
requiring no expensive complicated cooling techniques. The
number and expense of individual items have been minimized,
using small optics and simple electronic processing.
The requirements that the system work during day or night and
be independent of metereological conditions have been met by
using an active energy source. The signal magnitude of an
active system which transmits and detects a modulated signal
will not be affected by a variation in background level from
the water surface. Such a system, of course, uses the
reflectance properties of the water and/or oil surface. It
might then be argued that surface condition would affect the
amount of received radiation and, hence, the identification of
the surface substance. However, this is not a problem since
two-color processing is used.
A study1of reflected energy from water surfaces characterized
the nature of the reflecting surface in terms of specular versus
diffuse. The results were thatatl.06 microns the surfaces
were, by a factor of 2 to 3, better reflectors than a
Lambertian surface of the same reflectivity at the specular
surface. This should be coupled with the measured results
from a specular surface available in the literature. The
reflectance of a water surface is shown in Figure 14. The
data are of reflectivity versus wavelength as a function
of aspect angle for a bistatic measurement. As can be seen,
19
-------
to
o
*INFRARED
RADIATION;
MIKAEU A.
BRANSON.
PLENUM PRESS,
1968.
PP 553-594.
S-tff
5-a>'3
123014
Figure 14. Spectral Reflectance of Water in the Infrared Region for Natural
Polarization and Selected Angles of View*
-------
the absolute reflectance at a fixed wavelength varies
strongly with aspect angle, while the ratio of reflectivity
coefficients is essentially independent of aspect. If the
diffuse reflectance is thought of as a series of specular
reflectance events, then the ratio of reflected return at
two different wavelengths should be essentially independent
of surface condition. This would be true even though the
absolute return of each of the single wavelength channels
would vary by several orders of magnitude for a variance in
surface conditions. Hence, if it desired to measure the
specular reflectivity of a liquid surface with wave action
the reflectance at two wavelengths should be measured and
then ratioed. The result will be the ratio of the specular
reflectivity at wavelengths.
To actually implement the effect of ratioing, the real-time
log processors, developed for a multicolor radiometer and
further refined on this program, were used. Prior to
discussing the actual active system and the detailed block
diagram of the analog processing circuitry, the final
question of remote-alarm capability should be resolved. To
have a remote alarm, strong signatures of oil and water
should be found. Oil signature means the ratio of reflected
radiance at two different wavelengths. The voltages used
in forming the signature are obtained by processing the
output of two detectors, one filtered for X, and the other
for X2. This can be expressed as:
V(Xi) = Rir(Xi)S Pt(Xi)
where
V(X.) = the voltage from the detector filtered at X.
R = system response at X.
i 1
r(X.) = surface reflectivity (normal specular) at X.
S = scattering factor due to surface condition and
aspect
ptU•) = total power from the transmitter at the surfaces
which is in the detector's field-of-view
The signature between X, and \~ is, therefore
12 = in VU^/VBI - in v(x2)/vB2
21
-------
where Vg, and V 2 are fixed bias voltages in the log
processors.
cj>12 = In [r(X1)/r(X2)J + In [R^ (X-^/R^ (X2) ]
~ ln[VBl/VB2]
Note that S has dropped out of the signature since it was
the same for both wavelengths. By setting the bias voltages
properly in the log processor
12 = In [r(X1)/r(X2)l .
For reliable remote-alarm capabilities, the signature for
water should be significantly different from that of oil;
therefore, two wavelengths are needed which offer highly
contrasting signatures. All wavelengths shorter than
1.5 microns have not been considered here because of water
subsurface effects.
Oil which is polluting water remains on the surface of the
water. This is unlike other forms of chemical water
pollution where the pollutant is the bulk of the water
itself. The former case is fortunate since this means that
conceivable the areal extent and total amount of the oil can
be remotely determined. However, precautions should be taken
to avoid interference with subsurface conditions. Figure 15
gives the absorption coefficient and normal reflectance
respectively for pure water, from 1 to 5 microns* Water
appears transparent in the visible, and, as can be seen
from the absorption coefficient, this continues until
1.5 microns is reached. From this point on the water
becomes strongly absorbing. For example, the skin depth
for water at 3.4 microns is only 12.0 microns. This strong
absorption continues out into the 8- to 14-micron region.
The conclusion is that to avoid any subsurface effects in
the water, a system should operate at wavelengths beyond
1.5 microns.
Figure 15 shows that a strong log ratio signature for water
can be achieved by using one wavelength, in the region of 3.0
to 3.5 microns and the other in the vicinity of 4.0 microns.
The strong variation in the shape of the reflectance curve
is characteristic of a resonance absorption. In this case
it is the hydroxyl (0-H) vibrational mode. The resonance
absorption affects the dielectric constant of the material,
and hence, the index of refraction. The shape of this 3
curve is well known and is an anomalous dispersion curve.
It has the dependence that the reflectance is minimum for
the wavelengths shorter than the absorption and maximum for
wavelengths that are longer.
22
-------
to
CO
> 0.05
o 0.04
UJ
7? 0.03
0.02
0.01
0.00
LU
<
S
o
1.0
2.0 3.0
WAVELENGTH (MICRONS)
A
4.0
5.0
h-
y 3.0
o
LL.CO
U_ 0
UJ i-H _ _
0 ^ 2.0
o ^
0 '
Od
o
CO
CQ
_
1 -H
< 1.0 2
3.0
WAVELENGTH (MICRONS)
B
4.0
Figure 15. Absorption Coefficient and Normal Reflectivity of Pure Water
-------
The wavelengths for maximum signature from water would be
2.75-and 3.1-microns. The short wavelength sample must,
however, be taken at a wavelength longer than 3.3 microns .
because of the water vapor atmospheric attenuation in the
2.4- to 3.3-micron region. Figure 16 shows the transmission
of a 1000-foot strong water signature,1* practical in terms
of both atmospheric transmission and magnitude, that can be
obtained by ratioing the reflectivity at 3.4-microns to
that at 3.9-microns. This latter point is enhanced when
the characteristics of oil are examined. The reflectance
data were not available in literature; however, Figure 17
gives the transmission data for a typical crude oil.
Figure 18 shows the postulated shapes and locations of the
absorption and reflectance curves in dash lines. The
hydrocarbon (C-H) resonance in the 3.2- to 3.5-micron region
is noted. This will be present in all hydrocarbon compounds.
This anomalous dispersion shape seen in the comparison of
absorption to reflectivity in water will be present in the
oil reflectivity in the 3.2- to 3.5-micron region. That is,
reflectivity for oil at 3.3-to 3.4-microns will be lower than
that at 3.9-microns since it should be on the short wavelength
side of the resonance. This means, that the ratio for water
will be 1.5 while that of oil will be less than 1.0. In
terms of signatures, this is a very strong effect.- These
conculsions have been confirmed with the Texas Instruments
sensor and are discussed in Section VII.
System Characteristics
Short Range System (MOD 1)
The reflectometer sensor incorporates all the features discussed
previously. It is an active system in which simplicity
was emphasized in the optical system. It also uses the
circuitry knowledge gained in analog processing of multicolor
signatures in Texas Instruments multicolor radiometer. A
drawing of the optical system is shown in Figure 19 and a
photograph of the sensor is shown in Figure 11. The radiation
source is a quartz tungsten iodine lamp. An f/1 relay lens
system is used for efficient collection of the radiation
and refocusing for chopping. The chopping or modulation is
accomplished with a tuning fork whose resonance is 400 Hertz.
The primary transmitting and receiving paraboloid mirrors
have a 10-inch focal length at f/4. This allows for a small
optical housing and one that is light in weight.
The two wavelength channels must be simulataneously viewing
the surface area, illuminated by the transmitting portion of
the system. This is achieved by using one of the two
narrowband interference filters as a beamsplitter.
24
-------
NJ
100
90
80
70
60
50
40
30
20
10
0
I I I I
H2O AND CO2
I I
2.0 2.1 2.2 2.3 2.4 2.3 2.6 2.7 2. a 2. 9 3.0
3.0 3.1 3.2
3.5 3.6 3.7 3.8 3.9 4.O 4.1 4.2 4.3 4.4
WAVELENGTH (MICRONS)
Figure 16. Transmission of a 1000-Foot Horizontal Air Path at Sea Level, 5.7 mm
Precipitable Water and 79°F Temperature
-------
100
•I HASH CHUOE CAIBOL1C OIL
SOOO 4000 MOO 1SOO
' 1111
t. I. 1»0-JSO«C
1500 1400
to
CTi
I »
Soure«t B«ri«tt Olvltlon
I* C«ll: 0.04M
Figure 17. Transmission Data of Typical Crude Oil
-------
NORMAL REFLECTIVITY (r)
to
MICRON RANGE
2.0 3.0 4.0
WAVELENGTH IN MICRONS
ABSORBTION
COEFFICIENT
IN (CM-ijX 103
3.0
2.0
1.0
WATER-
I/
¥
v
\
1
\
\
V
^.
x-OIL
-H-- '
1.0 2.0 3.0
WAVELENGTH IN MICRONS
4.0
5.0
Figure 18. Postulated Shape and Location of Normal Reflectance and Absorption
Coefficients of Oil Relative to Those of Water
-------
PB SE DETECTOR
3.4-MICRON FILTER
3.9-MICRON FILTER
PB SE DETECTOR
CHOPPER
Figure 19. Optical Schematic of Active Infrared Oil Detection
System, MOD 1
28
-------
This interference filter passes only radiation in a 0.15
micron bandwidth centered at 3.9 microns. All other radiation
is reflected from that filter. The portion of the reflected
radiation in a similar band at 3.4 microns is then passed
by the second filter. Lead selenium (PbSe) detectors are
then placed behind each filter. These detectors operate well
at earth ambient temperatures.
A block diagram of the electronics is shown in Figure 20.
The preamps use a field effect transistor (FET) input stage
and are mounted directly on the detector housings. All other
electronic stages have integrated circuit (1C) operational
amplifiers. A dual transistor located in each log amplifier
circuit provides an output slope that depends on the
temperature of the junction of these transistors. Hence, to
improve the balance between the two analog channels, the
two dual transistors have a common temperature reservoir.
The characteristics of the processing circuitry are given
in Table 1. Although the system had been used for only
specular surfaces in the laboratory prior to this contract,
minor changes were made to permit operation off of nonspecular
surfaces at a range of 30 feet and in a field environment.
These changes included: the design of power supplies for
the analog electronics, chopper driver and TE coolers in
the detector, improvements on the housing and optical mounts,
the selection of a better light source, the reduction of
microphonic noises, and improved matching of log amplifiers
and RC (resistive-capacitive) time constants.
The original system used a quartz-iodine tungsten filament
lamp as a source. The quartz envelope provides suitable
transmission in the 3- to 4-micron region while the tungsten
has an effective blackbody color temperature of 3000°K.
One of the problems with the lamps was the inefficiency of
the geometry. The filament size was a 3/8- by 1/8-inch
coil. Radiation from only a small portion of the total
filament was being captured by the relay lens system and
collimated by the transmitting mirror. A special purpose
lamp was located with a small active element. By using this
lamp and a larger aperture chopper at the same frequency,
the relay lens system was eliminated with an increase in
the total optical through-put. Note that the constraints
on the system are that the area illuminated and the area
viewed are approximately the same. The focal lengths and
diameters of the transmitting and receiving mirrors were
kept the same for simplicity of alignment. This MOD 1 was used
in the first step in the evaluation of the system and the
technique in a field test at a range of 30 feet.
29
-------
ELECTRONICS
V
BIAS
U)
o
/•v
FILTE
«
il
Ri
T
RD :
PKbAMPLIHtK
;RL
•
- NARROW BAND-
PASS AMP
PEAK
DETECTOR
L
LOG
AMPLIFIER
^\
T
V
V.
3.4/x-
DIFFERENCING
AMPLIFIER
THREE
CHANNEL
STRIP CHART
RECORDER
REMOTE
ALARM
Figure 20. Block Diagram of Processing Circuitry
-------
TABLE 1. CHARACTERISTICS OF CIRCUIT ELEMENTS
3.4 Microns 3.9 Microns
Channel Channel
Narrow Bandpass Amplifier
(a) Gain at 400 Hz 52 dB 70 dB
(b) Q 9.3 16
DC + NBP Amplifier
(a) DC Level for for 1-mV rms input 0.47V 3.7V
(b) Noise for 1-K Load ImV 3mV
Log Amplifier
(a) 0.1-V input -2.30 -2.30
(b) 1.0V input Q.QO 0.00
(c) 2.0V input +0.69 +0.69
Preamplifier
(a) Gain 20. dB 20 dB
(b) Input noise for 400-K source 1/2 1/2
resistance 0.02 V/Hz 0.02 V/Hz
31
-------
Long Range System (MODS 2 and 3)
The main emphasis on extending the range of the system was
placed on improvement of signal-to-noise ratio by
optimizing the optical design in both the transmitter and
receiver. An important part of this design consisted in
determining the best field of view (FOV) for the environment
where the equipment would be installed. This was done by
performing a detailed study of the wave action on seas and
large lakes.
A certain fraction of a natural wave will always present a
facet which will be normal to the view angle. This is
demonstrated in Figure 21. This Figure gives the probability
that the sea surface slope, B, is less than or equal to
3. This is plotted as a function of 3. Figure 22 gives the
same data but plotted as the slope probability density
versus 3. On the basis of this data set it is seen that
the maximum return would be obtained from a beam striking
the surface at 14 from the normal. This angle depends on
the state of the water surface. Examples of this effect are
given in the test sections.
Since a fraction of the wave surface always gives a normal
reflection back, the most direct approach to the problems
of an active system is to have the area illuminated greater
or equal to the largest wave period. Also the area viewed
should be equal to the area illuminated for optimum efficiency,
If the distance from peak to peak of the largest waves to
be encountered is X then the FOV of the transmitter and
receiver should be 8/2. A design goal for the spot size
was on the order of 2.5 to 4 feet for 100 foot altitude
which is 7 to 12 feet at 300 feet. This corresponds to
FOV's of 24 to 40 milliradians. The large FOV values imply
that short focal length systems be used for both the
transmitter and receiver. Also, the aperture must be wide
to give strong capture power.
Design criterions for the optical system were:
1. Wide aperature compatible with filament area of
light source, the desired illuminated area, and the area of
available detectors.
2. FOV greater than or equal to the wave period.
3. FOV of transmitter and receiver equal.
4. Simple, inexpensive, and easy to maintain.
5. Easy to align in lab or in the field.
32
-------
6. Capable of being prealigned in the lab.
7. Easily reset when instrument is relocated at a
different height.
Transmitter Optics
The tungsten-iodine lamp was the source originally used
for this systems application. The reason is that it is
simple to incorporate and at the same time has many good
spectral characteristics. The maximum color temperature
is on the order of 2900°K which gives spectral efficiencies
on the order of 5 x 10~2 for the 4-micron region.
While this is not as good as a globar it is
power that can be used efficiently in a lens system due to
the small filament size. The one problem is that it is
a continuous source that must be mechanically modulated.
This dictates that the f-number for a system using a
tungsten lamp be on the order of 1.5 which thereby reduces
the collection capability of an optimized collecting
optical system. The color temperature is discussed in
more detail later in this section in terms of lamp lifetime,
1 1 1 1 1
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 UO.OO
BETfll DEGREES)
Figure 21. "Probability that.Sea Surface Slope is Less than
or Equal to $ Versus &
33
-------
-, ,
15.00 20.00 25.00 30.00 3r,.00
BETfl(DEGREES)
Figure 22. Slope Probability Density Versus 3
Another f.orm of radiation source is the gas lamp of which
Xenon is the best. A side by side comparison of the
tungsten iodine and Xenon arcs is given in Table 2. The
spectral efficiency is the fraction of total blackbody
emission occurring at 3.4 and 3.9 microns in a 1-micron band-
width. The Xenon system due to its higher temperature is
less efficient by an order of magnitude. This inefficiency
can be compensated for by the fact that the Xenon system
can be electrically modulated while the tungsten system must
be mechanically modulated.
By pulsing a flash lamp a given amount of energy can be
dumped in a short time. This means that by pulsing a
relatively low wattage CW (continuous wave) lamp can produce
high peak wattage signals. It also means, however, that
the detection bandwidth must be increased to allow for the
high frequency content of the signal. Since the noise
voltage of the detector increases as the square root of
the bandwidth it will increase as 1//T where T is the
integration time required to collect the signal pulse. The
peak signal increases as I//T" since the energy is being
dumped in that time period. The signal-to-noise ratio then
goes as I//T".
34
-------
TABLE 2. TRANSMITTER
Co
U1
PARAMETER
Spectral Efficiency
W(X)total
(per micron)
Optical Efficiency
Emissivity
Collection
I/A
Available Energy
in 0.1 x 0.1 inches (joules)
Figure of Merit
(per micron)
Comments
TUNGSTEN
(2900°K)
X = 3.4ym: 6.20 x 10
X = 3.9ym: 4.0 x 10~
= 0.4
=0.003
1.0
(T= lysec)
44
X = 3.4ym: 0.036
X = 3.9ym: 0.023
-2
-3
-3
XENON
(7000°K)
X = 3.4ym: 7.3 x 10
X = 3.9ym: 4.4 x 10
0.1
0.75
200.0
(T = 25ysec)
100
X = 3.4ym: 11.0
X = 3.9ym: 6.6
Possible loss of Optical
Efficiency at high energy
-------
All of the lamp factors are combined to give a final figure
of merit at the two wavelengths. As can be seen, even though
the Xenon system is spectrally inefficient, it can produce
a factor of 10 increase in system range over a system using
tungsten-iodine lamp. There is one problem. Photon flux
data has been taken with a Xenon lamp as a function of input
energy for a low-powered lamp rated at 5 joules at 3.6 microns.
The data was taken and as is shown in Figure 23 the signal
does not increase linearly with input power. This is believed
to be caused by either an increase in opacity due to an
increase in current density or by the increase in black body
temperature with an increase in energy input.5' 6
The key reliability factor in the active system is the lamp
lifetime. In order to obtain good collimation, lamps with
small active areas have been used. These small, high wattage
lamps usually have short lifetimes and hence become the
limiting factor on time between system maintenance. The
lifetime of a tungsten filament unit is inversely proportional
to the thirteenth power of the applied voltage! The optimum
lamp for the active system appears to be special tungsten
halogen lamp made by Sylvania. The number is 60 Q/CL-8.5
volts. The data appears in Table 3.
TABLE 3. CHARACTERISTICS OF TUNGSTEN HALOGEN LAMP
Wattage 60
Design Voltage 8.5
Rated Lumens 1000
Life (hours) 50
Color Temp. (°K) 3100
Min. Bulb Wall Temp. (°C) 350
Max. Seal Temp. (°C) 450
Element (coil) 0.1 inch long
by 0.1 inch diameter
Note that at 8.5 volts the rated life is 50 hours with a
color temperature of 3100°K. The latter value means that the
X , = 0.94 microns which determines the power distribution
ai a function of wavelength. The optimum operating level is
a tradeoff of the power in the region of 3.4 and 3.7 microns
and the life of the lamp. Figure 24 gives the lamp lifetime
as a function of applied voltage, Figure 25 gives the power
output at 3.5 microns as a function of applied voltage and
Figure 26 shows the current versus applied voltage. Note
36
-------
10.0
s.o
.
O
ec
u i .o
u
s ...
K
0.2
CONSTANT EFFICIENCY
• MEASURED DATA AT 3.0
I I I I I I I I
I I I I I I I I
J
0.1 0.2
0.5 1.0 2.0
LAMP INPUT (JOULES)
5..0 10.0 20.0
Figure 23.
Relative Flux Output Versus Input Energy for
Pulsed Xenon Lamp
37
-------
10
10
10
6.3 VOLTS
1000
500
200
100
50
20
10
5
4.0 5.0 6.0 7.0
VOLTAGE (VOLTS)
8.0
9.0
Figure 24. Lamp Life Versus Applied Voltage for a Tungsten
Halogen Lamp
38
-------
24
22
20
co
< 18
OHL
LiJ
0.
16
14
12
10
4.0 5.0
6.3 VOLTS
100
90
80
o
Q£
LU
Q.
70 -
o
60 ^
50
6.0 7.0 8.0 9.0
VOLTAGE (VOLTS)
40
Figure 25.
Emitted Power at 3.5 Microns as a Function of
Lamp Voltage (Tungsten)
39
-------
I
Mr.
!tf
m
m\
:: :tfi-
•t-q
'—3
tt
I
-1!
rt
I!
?HJ
4$
±;ff
#
Ifi:
TJ.
I
:±tl
t:
ili
'•IT
::-ft..
-ttt
i; etr
i; JC
itt
LIT i
r S±
ft n-<-
iiffii
M
-•f -1-H4
Il-
Figure 26. Current Versus Applied Voltage
Halogen Lamp
for a Tungsten
40
-------
that by using 6.3 volts, a hundred days of operation can be
obtained with a loss in signal of only 18%. Note that the
lifetime can be increased to 200 days with a signal loss of
22%. That is, the lifetime has been doubled but the signal
level has dropped only 4%. This, however, is not necessarily
desirable. The 6.3 volts is an off-the-shelf transformer
and, hence, the lamp supply is inexpensive. Also, 100 days
is a reasonable maintenance time period for all electro
optical equipment.
The lifetime of Xenon flash lamps is somewhat more complex
than that of the tungsten lamp. The number of flashes not
only depends on the energy per pulse but also the pulse
duration. These are related by
-8.5
N = y
where N = number of flashes
and Y = E /E which is the ratio of lamp energy to
O X
explosion energy.
Again, small emitting areas are required. Two Xenon lamps,
X-80 and X-160 made by Illumination Industries Inc, were
selected. Their characteristics are given in Table 4. Note
that instead of explosion energy, the single shot explosion
energy constant is given. They are related by
F - r r1/2
Ex - CeT
where C = single shot explosion energy constant and
T = pulse width
TABLE 4. CHARACTERISTICS OF XENON LAMPS
X-80 X-160
Maximum Energy (joules) 25 100
Arc Spacing (mm) 1 1.5
Maximum Average Power, DC
Operation (watts) 60 100
Maximum Average Power,
Pulsed Conditions (watts) 75 150
C = single shot explosion -,/- 4
energy constant (joules/sec ' ) 1.19 x 10
Figure 27 shows N versus C for the X-80 lamp. It should be
noted that 10^ is a practical upper limit for N for any lamp,
not just the X-80 lamp.
41
-------
10'
(A
U
0.
IL
o
K
U
O
3E
3
Z
10
10
0. 1
0.2 0,3
Y (DIMENSION/ESS)
0.4
0.5
Figure 27. Number of Pulses Versus the Single Shot Explosion
42
-------
Assuming a value of 10 for N the value of y is 0.197.
Therefore
T1/2 = E / (0.197 C )
o e
~ (4.22 x 10~4) E
o
Figure 28 shows a plot of T versus E . This isgthe limiting
curve for the X-80 lamp in order to obtain 10 pulses. One
important point is that the signal-to-noise ratio goes as
(E /T-*-/^) . This is constant for a given lamp operating at the
optimum energy point. It is, of course, not a constant if
the lamp is not operated at optimum conditions. A plot of
the relative signal-to-noise ratio as a function of the
number of pulses given in Figure 29 for the X-80 lamp. Note
that in order to gain a factor of two in the signal-to-noise
ratio, the life of the lamp is decreased by almost three
decades.
A study of existing lasers showed that the long-range system
could not be fabricated cheaply with gas lasers and exceeded
the efficiency of different type of source in a long term
time period. Texas Instruments has a laser diode program
for building tri-metal lasers which operate in the 4- to 12-
micron region. The present materials used are PbSnTe and
PbGeTe. In a pulsed mode, the lasers can produce up to 1
watt of power in a line width on the order of 10~5 cm" .
Also considered was the InSbAs diode. If this material
develops, it can be tuned to the 3.4 micron region. Due to
its very narrow line width, it will also be suitable for
heterodyne detection. The application of this type of system
is at least two years away.
In conclusion the following points are obvious. For a long
life time system a tungsten-halogen lamp should be used and
operated at a point below maximum rating. For example the
60 Q/CL-85 lamp operated at 6.3 volts. For applications of
very long range a flash lamp should be used.
Receiver Optics
An improved receiver was designed for the long range system.
An optical diagram of the unit is shown in Figure 30. It uses
a 3.5-inch aperature refractive lens with an f-number of 1.4.
The 3-mm square detectors used in the original unit were used.
However, light pipes were added in front of the detectors to
43
-------
100.0
50
o
LU
CO
20
10.0
Q_
r—
l-
2
1.0
0
10
15
EQ (JOULES)
N = 10
20
25
Figure 28. Pulse Width Versus Pulse Energy for X-80 Lamp
44
-------
-,1L6
105
on
LU
on
1.2
O
o
s
oc
UJ
CO
OO
a:
10
0.1
2 3 4
RELATIVE SIGNAL TO NOISE
Figure 29. Number of Flashes of an X-80 Xenon Lamp Versus the
Signal-to-Noise Ratio for a Pulsed Detection System
45
-------
RECEIVER OPTICS FOR MODS 2 AND 3
SILICON LENS
efl = 5"
f NUMBER = 1.4
3.9/x FILTER
1173 GLASS
DETECTOR
3 mm X 3 mm
3.4//. FILTER
DETECTOR 3 mm X 3 mm
Figure 30. Receiver Optics
-------
give an enlarged field of view (FOV) of 40 milliradians.
This gives a versatility of 7.2- and 12-foot FOV's at 300
feet. In order to simplify the alignment procedure and to
allow for a quick readjustment of the active sensor in the
field when it is relocated at a different height, all of
the receiver optics were mounted on an adjustable metal plate
that was pre-aligned in the laboratory. By using a similar
plate for the transmitter optics and micrometer head for
angular control of each plate, the system can be preset per
Figure 31 for any height between 30 and 300 feet.
In addition to the improvements made on the optical system
discussed above, another area of possible significant
improvement is in a change of detectors. At present the
system uses PbSe with an inexpensive thermoelectric cooler.
An alternative is to use InSb detectors mounted on open cycle
Joule-Thomson coolers for 77°K operation. The advantages of
this mode of operation are shown in Table 5. Note that the
improvement in detectivity (units of cm(Hz)-'-' watts"-1-) of
the InSb detector is approximately a factor of 10 over the
PbSe. However, still further improvement can be made since
there is a cold reservoir available. This is shown in the
cold stop factor and cold spectral filtering. These
improvements result from aperturing the detector to remove
some of the background and then removing all of the background
except that which is in the desired spectral bandpass. The
total increase in the figure of merit is approximately 102
or an increase in range of a factor of 10. This technique,
while attractive, was not evaluated during this program due
to the increase in the cost of mass produced systems.
D*(PEAK)
(180°FOV)
Cold Stop Factor
Cold Spectral
Filter
Figure of Merit
TABLE 5. DETECTORS
PbSe
(2400R)
3.4ym:
3.9ym: 6x10'
cm Hz
1/2
3.4ym: 5x10
3.9ym: 6xl09
InSb
(77°K)
3.4ym: 5.5x10
cm Hz1/2
3.9ym: 7xl010
2.0
3.4ym->-8.9
3. 9ym->-4 .9
3.4ym: 9.8x10:
3.9ym: 6.9x10-
10W
47
-------
10,000)
NOTE:
1) MICROMETERS ARE SET AT
ZERO FOR CASE WHERE
PLATES ARE PARALLEL I.E.
IMAGE AT CO
1,000
UI
It
%•*
ff
P
O
tL
O
III
100
to
o.oot
50 FT
30 FT
O.O1
MICROMETER SETTING IN INCHES
0.1
Figure 31. Optic Plate Adjustment Curve
48
-------
Signature Optimization
As part of this program, a study was made of oil and water
signatures in the 3.3- to 3.9-micron region. As outlined
initially the purpose of the system is to distinguish between
oil and water in a fashion that is independent of the surface
state of the water. This is accomplished by performing the
two wavelength reflectance measurement as described in the
first part of this section. All view factors are removed
by the processing when the two channels are identical in
their fields of view.
The reflectance of water is well known and is shown in
Figure 18, however, previously the only data available on
oil was in the form of transmission (Figure 17). The
location of the C-H bond absorption in the 3.0- to
3.5-micron region could be interpreted to give an absorption
coefficient similar in nature to that shown in Figure 18.
An anomalous dispersion curve for the reflectance should
result as indicated in Figure 18.
In order to optimize the signature difference, the exact
shape of the reflectance curves had to be determined. This
was done by using the MOD 1 system and substituting different
filters. The results are shown in Figure 32. Rather than
take the time to obtain exact reflectance values in which
source color and system spectral bandpass was removed, the
values have been normalized to the known values for water.
The resulting oil curve is plotted in Figure 33. The data
is for aged crude oil; hence, the signature is not as strong
as for a fresh product. It was necessary to use the aged
product since a time period of approximately an hour was
required to obtain the spectral results.
The above quantities have been processed by hand to simulate
the processing signature; ie, the system output. The results
are given in Table 6. The implication is that the short
wavelength should be 3.3 microns; however, the atmosphere
(see Figure 16) is an interfering quantity. Table 7
summarized the atmospheric effects on the signature when the
short wavelength values are either 3.3 or 3.4 microns. As
can be seen, the system would essentially have to be
recalibrated for every range and atmospheric condition (water
vapor) if 3.3 microns were used. Hence, 3.4 microns is a
better choice even though the signature is lower by 0.14
units.
Also, from the data it appears that 3.9 microns is the
optimum long wavelength value. This does not, however,
49
-------
NOTE: The values are not corrected for source color or system
spectral response.
4.0 --
3.0 --
Q
LU
tVI
o;
o
2.0 --
1.0 --
0 OIL SURFACE (AGED)
WATER SURFACE
\
\
N
\
0
4-
t
3.3 3.5 3.7
WAVELENGTH (MICROMETERS)
Figure 32. Measured Reflectances from Smooth Oil and Water Surfaces
3.9
-------
5.0 --
4.0
3.0 --
I 2.0
i.o --
NOTE: The normalization values were obtained from a water surface
and the values in Reference 1.
©
\
'0.
OIL SURFACE (AGED)
Q WATER SURFACE (REFERENCE I)
3.3
3.5
3.7
3.9
WAVELENGTH (MICROMETERS)
Figure 33. Reflectance from a Smooth Oil Surface at Normal Incidence
-------
take into account the signal-to-noise ratio. Figure 34
shows data obtained on both Tungsten and Xenon lamps in
the 3.3- to 3.9-micron region. In both cases the output
peaks in the vicinity of 3.6 microns. This is due to the
transmissions of the lamp envelopes and filters in addition
to the blackbody dependence. This must be considered when
selecting the optimum wavelength. Write the signature to
include the noise such that:
4>(XlfX2) = In
P(X2) + N(X2)_
- In
rv
B(X1)
LVB(X2U
where N values denote an rms value of noise.
The signature will have a worst case error of
e(XlfX2) = In
1 +
S(X2)
- In
1 -
N(X2)
S(X2)
where S(X) represents the signal including the lamp color
and filter transmission. It is a valid assumption to assume
that the noise levels are the same for X, and X2 since
fields of view are the same and unchillea filters are used.
Hence, the values of S(X)/N(X) should be scaled according
to the data of Figure 34. This has been accomplished with
the results shown in Figure 35. Two examples of upper wave-
lengths are used, 3.6 and 3.9 microns. The data is plotted
versus the signal-to-noise ratio of 3.4 microns. The
goodness of the upper wavelength depends on what S/N value
the quantity e = 1/2, that is, one-half the signature.
These levels are shown on the plots with the data coming
from Table 6. The 3.6 micron values are slightly better
for both the tungsten and Xenon lamps. It was concluded
that 3.4 and 3.6 microns are the best wavelength pair for
the identification problem.
Signature Analysis of Various Types of Oil and Debris
Humble Oil supplied Texas Instruments with quart samples of
crude, Navy distillate, diesel fuel oil, numbers 2, 4, and
6. Another crude was obtained from Sun Oil. The signatures
of each of these samples and water were measured with the
MOD 1 system. The 2.5-inch entrance aperture was, however,
stopped down to 0.75 inches to reduce the signal level
since the measurements were made in the laboratory. This
produced an order of magnitude drop in sensitivity. The
filter bandpasses were at 3.3 and 3.6 microns. The results
52
-------
NOTE: Xenon Data is for a short arc, small cathode unit
operating at 25 ysec. pulsewidth.
1.0 --
0.5 --
Xe @ 1300 volts
Xe G> 1000 volts
tn
to
0.2 --
0.1 --
a:
o
Xe @ 500 volts
0.05 __
0.02 __
i
H 1 1-
3.3
Tungsten
Tungsten
3.9
3.5 3.7
WAVELENGTH (MICROMETERS)
Figure 34. Lamp Output Versus Wavelength
6.3 volts
5 volts
-------
TUNGSTEN
Xe
(SI
0.4 4-
UJ
OH
0.3 4-
0.2 4-
o.i 4-
*(3.4, 3.9)
$(3.4. 3.6)
-I 1-
0.5 4-
0.4 4-
0.3
0.2 +
0.1 4-
16 32 4
(S/N) IN 3.4 MICROMETER (CHANNEL)
-t h
16 32
Figure 35. Signature Dependence On Signal-to-Noise Ratio
-------
of the "log difference: signature are given in Table 8. As
can be seen, the variation between samples does not exceed
3 percent. This can be considered negligible for this
type of application.
For the system to discriminate various types of oil from
water is not sufficient for a remote alarm system. The
system must also not confuse materials such as dirt, mud, or
floating debris with oil. Therefore, the same "log
difference" signature was measured for a variety of materials
including the above. The MOD 1 system in the same configu-
ration as was used for the oil signatures was again used.
The results are shown in Table 9.
TABLE 6. COMPUTER SIGNATURES VERSUS WAVELENGTH
X
(ym)
3.3
3.4
3.5
3.6
3.7
3.8
3.9
WATER
-
.08
.18
.27
.44
.48
(3.3ymX)
OIL DIFFERENCE WATER
-
-.06
-.23
-.22
-.04
-.07
-
.14
.41
.49
.48
.55
-.08
-
.10
.19
.36
.40
(3.4ym,X)
OIL DIFFERENCE
+ .06
-
-.17
-.16
.02
-.01
-.14
-
.27
.35
.34
..4.1. .
TABLE 7. ATMOSPHERIC EFFECTS
3.4ym 3.3ym
r(600 FT) 0.72 0.96
. _ . . -.33 -.04
Log Transmission
55
-------
TABLE 8. SIGNATURES FOR SIX TYPES OF OIL
AND WATER
Sample
Water
Crude (Humble)
Navy Distillate
Diesel #2
Fuel #4
Fuel #6
Crude (Sun Oil)
(3.3ym, 3.6ym)
0.41
-0.04
-0.07
-0.07
-0.04
-0.04
-0.04
TABLE 9. SIGNATURES FOR VARIOUS TYPES OF DEBRIS, WATER AND
CRUDE OIL
Sample
Water
Oil
Dirt
Mud
Dirt with Oil
Plywood
Plywood with Oil
Plywood with Water
Paper
V(3.3ym)
(Volts)
0.70
0.72
0.019
0.018
0.045
0.018
0.018
0.013
0.53
0.017
V(3.6ym)
(Volts)
1.57
2.55
0.016*
0.018*
0.080
0.021*
0.018*
0.047
1.10
0.017*
(3.3ym, 3. 6ym)
+0.41
-0.13
+1.30
+1.13
+0.58
-1.12
+1.13
-0.16
+0.41
+1.13
*Voltage approach noise level of system in apertured mode.
The actual voltages were measured at the point prior to the
log amplifiers. The voltages are included in the table to
point out the very low value of reflectance of soil and
plywood primarily at the 3.6-micron wavelength. The result
is that in terms of this system concept, paper, dirt and
plywood give a signature value even more positive than does
56
-------
water. In fact, for those values indicated by the
asterisk, the 3.6-micron voltage approached the noise
level of the MOD 1 system with its 0.75-inch aperture and
resulting decreased sensitivity. There was also a drop in
the 3.3 micron values although not as pronounced. This
latter point is attributed to the diffuse nature of the
reflection.
In addition to the above, natural alga was checked as a
potential false alarm. The material used was that which was
found as naturally occurring at the edges of the pond at the
Texas Instruments test site. The alga had a signature which
was indistinguishable from water. Hence, it will not false
alarm the system.
57
-------
SECTION VI
DEMONSTRATION SYSTEM, PASSIVE SENSOR
General System Description
The system employed for the passive infrared (IR) oil spill
sensor demonstration was a Multipurpose Infrared System
(MIRS) developed by Texas Instruments for the Night Vision
Laboratories. This system, designed for use as a night
sight, has a direct view raster display produced with light
emitting diodes. The primary considerations for design of
the MIRS were weight and power consumption. Effective
performance and low cost were also important design goals.
Staying within design limits required many trade-offs in
terms of cooling, electronics, optics, and water tightness;
therefore, the system may not have been ideal for the purposes
of the oil spill investigative study. This sensor, however,
was sensitive in the 8- to 14- micron region, had good
thermal resolution, and could demonstrate the feasibility
of IR oil spill detection and surveillance.
The MIRS weighs 5.4 pounds and uses Mercury-Cadmium-Telluride
(HgCdTe) detectors in a straight vertical line array. The
system also uses 44 Gallium-Arsenide-Phosphide (GaAsP) light
emitting diodes (LED's) in the same array configuration as
the detectors. This facilitates reproducing the scene as
seen by the detectors. The LED display is direct view and
is scanned by one side of the two-sided scan mirror.
The system produces a framing-type presentation. That is,
it continuously scans relatively small areal section of the
panorama at one time. However, the presentation is a real-
time television-type raster scan. Conventional line scanners
present video for film or magnetic tape recording of continuous
information, one scan line at a time. The MIRS system scan
mirror provides the X axis of areal coverage while the array
of detectors provides the Y axis. The overall MIRS system
is shown in Figures 36 and 37. Figure 38 shows how the
MIRS operates.
The detector array receives a frame of IR energy with a
field of view (FOV) of 5° vertical and 10° horizontal.
The following paragraphs describe how the MIRS operates.
59
-------
Figure 36: Multipurpose Infrared System
Figure 37: MIRS Optics and Electronics Packaging
60
-------
LED
Array
Lens
System
Video
Processing
Detectors
Infrared
Video
Figure 38: MIRS Concept Design
With the system pointed in one direction, infrared energy,
proportional to the scene pattern, passes through a lens
system to a two-sided scan mirror. The scan mirror
oscillates through an angle of 20° about a vertical axis;
thus, the energy reflects from the mirror with an
oscillation in the X-Y plane of 20°. The IR energy pattern
is then focused on a vertical line array of 44 detectors.
As the IR pattern oscillates, the detector line array
sweeps out the entire IR input with each half cycle of
oscillation. The detector line array has a relative
horizontal movement with respect to the IR pattern.
This type of configuration produces a bi-directional scan.
The video signal from each detector is processed and applied
to the associated LED on the vertical array of 44 Gallium
Arsenide Phosphide (GaAsP) LED's. The LED output reflects
from the opposite side of the two-sided scan mirror, through
a lens system, to the eyepiece.
On each half cycle of scan oscillation, the scan mirror is
tilted to shift the IR pattern vertically by 0.01 inches.
At the same time, the light array on the opposite side is
shifted in the opposite vertical direction by the same
amount. This is called interlacing, and produces 88 lines
per frame of video.
61
-------
Since the light output from each LED varies with respect to
another when excited with, identical currents, a resistor is
selected for each, diode to adjust the forward current. This
provides uniform light output throughout the array. The
diodes operate with, this set current at all times. Variations
in brightness due to the video signal are achieved through
a unique pulse width, modulation technique. The diodes are
switched on and off at a constant frequency several times
greater than the system resolution frequency to prevent the
pulses from being visible. The duty cycle of each pulse
is controlled by the video level. Thus, for higher level
video signals, the duty cycle is longer, exciting the LED's
over a longer period. This appears as a brighter light
to the eye. The eye effectively integrates the constant
level pulses over time and reads them as intensity variations.
The purpose of this technique is to eliminate possible non-
uniformities in diode light outputs as a function of bias
current.
The light output from the MIRS was, then, viewed by a closed
circuit TV camera which facilitated presentation on a
remote monitor and allowed video tape recording of the data.
For accurate measurement of voltage level differences between
oil and water, one of the 44 channels of video was diverted
into the video input of an oscilloscope, where polaroid
photographs of one channel of video were taken.
The system of cooling used for the HgCdTe detectors is an
open cycle, Joule-Thompson, cryogenic cooler, using
compressed Nitrogen gas to accomplish cooling at 77°K (-321°F),
Spatial and Thermal Resolution
The modulation transfer function (MTF) of an electro-optical
system simply shows the system relative response as a function
of spatial frequency. The concept is analogous to that
used in the communication field, since many optically
oriented systems essentially perform the function of infor-
mation transmittal. The optical transfer function of the
electro-optical equipment corresponds to the electronic
transfer function of communication equipment. Each relates
the frequency response of a system to an input source of
energy which, in the case of IR imaging systems, is a point
source.
As is the case for electronic transfer functions, the total
system MTF is composed of the MTF's of the various individual
components. Results of the theoretical calculations of the
MTF for the MIRS is shown in Figure 39. Modulation,
62
-------
O IR Lens
Q Video Amp
Display Optics
A Detector Scan D Emitter Scan
(Ti
Fl.O
.8 -
.6 -
.4 _
.2
.2
.4
.6 .8 1.0 1.2
f/fo fo - l/2A9o
Figure 39. System MTF
1.4
-------
normalized to 100%, is plotted as a function of spatial
angular resolution and normalized to the specified value of
system resolution. The fact that the system resolution
is classified is in itself an indication of the capability
of the sensor.
A common parameter of the IR systems used tc measure their
temperature sensitivity is noise equivalent temperature
(NET). NET is defined as the input temperature differential
which produces a unity signal-to-noise ratio out of the system.
NET essentially indicates the minimum detectable temperature
differential capability of the system. The NET of the MIRS
sensor is classified. Refer to Section X for calculation
of NET for an unclassified sensor utilizing similar detector
material.
Detector
The IR detector utilized for the MIRS sensor is fabricated
from mercury-cadmium-telluride (HgCdTe) material. This
detector has a substantial response at wavelengths below
8 microns as shown by the relative spectral response curve in
Figure 40. However, the MIRS sensor is sensitive only in
the 8- to 14-micron region due to the transmission
characteristics of the optical system. Figure 41 shows the
spectral response of the MIRS optics.
HgCdTe detectors are employed in most current system designs
concerned with the thermal infrared portion of the spectrum.
Imagery obtained from a system which utilizes HgCdTe shows
greater contrast of targets above average terrain temperatures,
and greater contrast between foliage and background than
extrinsic photodetectors, also sensitive in the 8- to 14-
micron range.
The chief difference between HgCdTe detectors and other
types of detectors is that HgCdTe is not an extrinsic photo
detector. HgCdTe is a ternary alloy formed by chemically
combining the semimetal mercury telluride (HgTe) with the
semiconductor cadmium telluride (CdTe). HgCdTe can be
fabricated with various wavelengths of peak response in the
4- to 18-micron region, by varying alloy composition.
Extrinsically doped detectors must operate at a temperature
much lower than HgCdTe in order to attain a sufficiently
low noise figure.
64
-------
17.5
15.0
12.5
10.0
HgCdTe Response vs. X
D vs. X
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0
Figure 40. HgCdTe Relative Spectral Response
1.00
.80
.60
.40
.20
7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0
Figure 41. Spectral Response of MIRS Optics
65
-------
The reason is. that the number of free carriers present in
the conduction band of an extrinsically doped semiconductor
directly affect the magnitude of the noise voltage produced
by the detector. The density of these carriers is an exponential
function of the temperature. Thus, the operating temperature
must be kept below 30°K for extrinsically doped detectors.
Due to the intrinsic property of HgCdTe; that is, the photon
absorbing properties being a part of the material and not
introduced as impurities, HgCdTe can be operated at 77°K.
66
-------
SECTION VII
TEST RESULTS, ACTIVE SENSOR
The purpose of this test program was to determine experimen-
tally the full capabilities and potentials of the active
system. The system demonstrated its capability to
absolutely discriminate oil from water, debris, and other
similar material. The system was proved to be inexpensive
and easy to maintain. In addition, its remote alarm
capability was demonstrated. Day or night operation
independent of weather and water surface state were also
demonstrated to a height of up to 91 feet above the surface.
This data was extrapolated to show capability up to 300 feet.
In addition to the above characteristics, the active system,
which was proposed as a staring system, whose line of sight
was normal to the water surface, is also capable of being
operated in a scanning pattern. This pattern can have large
tilt angles, depending on the wave action of the body of
water. Results of these tests are included in this section
in roughly the order in which they were obtained. Hence,
the development of the sensor capabilities can .be followed.
Short Range Tests
The first step in the evaluation of the system and technique
was to install the existing system (MOD 1) on a 30-foot
tower overlooking a body of water at a Texas Instruments test
range near the Dallas Central Expressway site. A photograph
of this installation is shown in Figure 42. The purpose of
these initial steps was to test the reflectance versus
surface state and to evaluate the various system parameters.
A three-channel strip recorder was used. It recorded as a
function of time the signal amplitudes at 3.4 and 3.9
microns and also the output of the log processor. The output
of the log processor indicated a "yes" or "no" state on the
presence of oil. This signal can be used to drive a horn,
siren, bell, light or other desired alarm mechanism to
provide an audible and/or visual signal to a responsible
individual located away from the oil spill site.
While these tests were being performed, the long range
system was being designed and fabricated. The results of
the MOD 1 tests were utilized to improve the design of the
long range system. Of particular significance were the
human factor considerations, the simplification of the
optic alignment and the stabilization of the optical and
electronic subsystems.
67
-------
Figure 42.
MOD 1 Mounted at 30 Feet Overlooking a Body of
Water
Wave Action Data
The primary problem in the field evaluation of the oil spill
surveillance system is the scaling of the performance to wave
amplitude and period. To determine if the scaling is valid,
the signal level prior to the log amplifiers must be
correlated to the wave amplitude and period. A probe was
built which measures wave amplitude and frequency. An
example of the probe output is shown in Figure 43. This
data was recorded simultaneously with the other data.
The MOD 1 system was used to evaluate the effects of wave
action on the two-channel processing system. Examples of
the data are shown in Figures 44 (7 sheets) for MOD 1
system at a height of 30 feet.
The record is for a continuous segment of data. The gains
and log amplifiers were set such that the water signature
occurred at approximately 70 divisions and the oil at 58
divisions. The actual amplitude of the 3.3- and 3.6-micron
channels prior to log processing were also recorded to
indicate amplitude fluctuations.
68
-------
10
i -'• '
i ;
I S ;
_ ._ 1
1
•
•
Figure 43. Example of Wave Probe Output
-------
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 1 of 7)
-------
1C)
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 2 of 7)
-------
to
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 3 of 7)
-------
U)
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 4 of 7)
-------
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 5 of 7)
-------
-o
en
1C)
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 6 of 7)
-------
1C 5
Figure 44. Example of MOD 1 Data at 30 Feet (Sheet 7 of 7)
-------
The systems field of view was on water during the time of
Figure 44, sheets 1 and 2. During the time of Figure 44,
sheet 3, oil was added to the water surface. At first,
light machine oil was used. This condensed into droplets
as indicated. Crude oil was then added as noted in the
upper portion of Figure 44, sheet 1. The fraction of the
water surface within the systems field of view that was
covered by the crude oil varied until the surface was
uniformly covered as indicated in Figure 44, sheet 6.
The data in Figure 44 exemplifies in terms of signatures
the effects of oil covering the water surface in the
systems field of view. This data was obtained such that the
signal levels of both oil and water targets passed through
the same amplitude regions to display signal level
independence. The log amplifiers used for this data were
an improved design for which the environmental temperature
dependence has been reduced.
During the time that the data in Figure 44 was being taken
the wave action was that which was naturally occurring on
the lake. The average signal level was, therefore, high with
the 4 second time constant giving maximum signal variations
on the order of 8 to 1. In order to obtain a larger dynamic
range, a fan mounted at the base of the tower was turned on
to generate increased wave action.
Figure 45 (7 sheets) contains the data for the fan generated
waves. Note that the voltage span for this set of data has
been increased by an order of magnitude for V(3.6) and V(3.3),
Also, the gain of the 3.6 micron channel was increased
which repositioned the water signature at 60 divisions and
dropped the oil signature to a value of less than 50
divisions. These adjustments are indicated in Figure 45
sheets 1 and 2. During Figure 45, sheet 3, the wave probe
was placed on a higher sensitivity. The oil was placed in
the systems field of view as shown in Figure 44, sheet 4.
The data shown in Figure 45, sheets 5, 6 and 7 demonstrates
the discrimination capability of the MOD 1 even under the
strong wave action. By comparing the results of Figure 44
with those of Figure 45 it is seen that the demonstrated
dynamic range of the MOD 1 system was 50 to 1 using the 4
second time constant.
77
-------
CO
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 1 of 7)
-------
ICO
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 2 of 7)
-------
CO
o
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 3 of 71
-------
00
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 4 of 7)
-------
CO
ro
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 5 of 7)
-------
00
CO
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 6 of 7)
-------
00
10)
Figure 45. Example of MOD 1 Data with Fan Generated Waves (Sheet 7 of 7).
-------
Angular Data
A key point to the application of an active system is its
ability to look at angles off of normal. The degree to which
the system can be tilted off of normal will depend on the
surface state. The signature or alarm capability will still
be independent of the surface state so long as the tilt
angle does not exceed some critical value where some signal
is lost. It is anticipated that for some surface conditions
the optimum view angle is not straight down.
Some preliminary data was taken with. MOD 1 system. The wind
velocity was 20 mph. The geometry is shown in Figure 46.
Note that the angle of wave motion with the scan direction
was 45 degrees. Some of the results are shown in Figure 47,
sheets 1, 2, and 3. The elevation angle for the data shown
in Figure 47, sheet is 2.9 degrees and the target is water.
The elevation angle for the data shown in Figure 47, sheets
2 and 3 is 5.7 degrees and the targets are water and oil,
respectively. Note that the signatures are valid. The angle
subtended by the transmit beam in MOD 1 is approximately
0.5 degrees.
Additional data of this type was taken with the MOD 2 system.
Long Range Tests
MOD 2 Performance
The MOD 2 version of the active system has an approximate 24
milliradian field of view which corresponds to a 1-foot spot
at a 50-foot range. The transmitter and receiver each contain
a simple lens of focal length 5.90 inches and f/number of
1.5. The anticipated improvement of MOD 2 over MOD 1 was
a factor of 16 for a specular reflective surface and a
factor of 60 for a Lambertian surface at 30 feet. The specular
case could be checked in the laboratory and this criteria was
met.
In field operation, the system performed well at the 50-foot
test height, implying again that the design goals had been
met. This can only be a subjective evaluation due to the
nature of wave action on the water surface.
One question which had been raised in program reviews concerned
signature dependence on height. The MOD 2 system was set
at the 30-foot range and aligned for optimum signal. The
system was then moved to the 50-foot height and turned on
without any realignment. This corresponded to a 25%
85
-------
variation in height of sensor above the water surface.
The signal level decreased by approximately a factor of 3,
but the signature did not change, meaning that the system
can satisfactorily operate under these extreme changes.
The theoretical change in going from 30 to 50 feet in
height is 2.77 loss in signal. The value cannot be determined
exactly due to wave fluctuations, but appears to be closer
to a loss of only a factor 2 rather than 3. This implies
that the field of view was more optimum for the 1-foot waves
at a height of 50 feet than it was at 30 feet.
Some samples of MOD 2 test data are shown in Figure 48,
sheets 1 and 2. To record this data the large boom described
in a previous section was set under the beam. The log
difference was set at 5-V span (0.5 V per major division)
with zero position at 3 V; ie, the line 60. At the beginning
of the trace, Figure 48, sheet 1 the system was looking at
water and the trace shows a slight negative reading at line
55. A small quantity of oil was placed in the boom and the
trace dropped to 50. It stayed at approximately this value
until a larger quantity of Navy distillate was added,wherein
it dropped to 43 and stayed at this value until the oil
slick moved away from the beam. Note that the reading did
not return to the original reading but remained at 50. This
condition was observed for several minutes and did not
change until the boom was moved and the water agitated to
disperse the scum. Figure 48, sheet 2 shows that the trace
then returned to its original water reading of 55. Amplitude
settings had to be changed when the boom was removed because
of the difference in wave action between inside and outside
of the boom and to prevent the recorder, which has a dynamic
range of only 10 to 1, from overdriving. However, the log
difference output remained the same since the dynamic range
of the system is 100 to 1. Another sample of water and oil
discrimination is shown in Figure 49, where MOD 2 was viewing
the open pond. The oil stayed under the boom approximately
1.5 minutes before natural wave action swept it away.
Angular Data from MOD 2
As first proposed, the active system was to be a staring
unit which looked only at the mean normal to the water
surface. However, satisfactory systems performance and the
knowledge gained concerning wave surfaces prompted the angular
scan data discussed in a previous paragraph of this section.
More extensive data was obtained with the MOD 2 system.
86
-------
The data was obtained at both. 30- and 50-foot levels. The
geometry is shown in Figure 50. Samples of the data are
plotted in Figure 51. For this data, all scan angles were
along the = 0 degree line; ie, in the North direction.
The wind was also coming from the North. The plotted values
are the signal in the 3.4-micron channel. The signal
voltage appears to follow an exponential curve for angles
not close to zero. A possible description of the probability
of backscatter is
PCS) = A
for 0
TEST TOWER
ACTIVE SYSTEM
30'
ELEVATION ANGLE
LINE OF SIGHT FOR
NORMAL INCIDENCE
SHORELINE
Figure 46. Geometry of Scan Data at 30 Feet
87
-------
Figure 47. Water Target at = 2.9°, MOD 1 (Sheet 1 of 3)
-------
00
ID
Figure 47. Water Target at 4> = 2.9°, MOD 1 (Sheet 2 of 3)
-------
Figure 47. Water Target at $ = 2.9°, MOD 1 (Sheet 3 of 3)
-------
Figure 48. Example of MOD 2 Text Data (Sheet 1 of 2)
-------
—-I - I I I I
Figure 48. Example of MOD 2 Text Data (Sheet 2 of 2)
-------
fRUMENTS INCORPORATED. HOUVTOM. TVXA*. U.S.A. CMAKT WUT
Figure 49. Example of Water and Oil Discrimination
-------
SIDE OF TEST TOWER
vo
50 FOOT
LEVEL
SENSOR
SCAN ANGLE
1
N
0
TOP OF TEST TOWER |
I
NORMAL TO
SURFACE
CTIVE BEAM
t
SHORE LINE WATER
Figure 50. Test Scan Geometry
-------
-0.8
NORTH P26 - 27 30 FT. 15 FEB
CO
UJ
o
UJ
LU
°-5 NORTH 18 FEB-V
-.04
-.03
NOISE LEVEL
ut t t t
WNO.. 1.2. .3
t
t t
AJ 1
t
,Z
t
j i
o
10 12 14 16 18 20 22 24 26
TILT ANGLE IN DEGRES
Figure 51.
MOD 2 Signal Amplitude Versus Scan Angle for
30- and 50-Foot Heights
95
-------
2
The range squared spreading loss gives a cos 6 dependence,
hence for this description
V(6) = V
The data appears to be a simple exponential in the region
of 10° <_ 6 <_ 20°.
This angular dependence was also checked for different values
of . Data for MOD 2 at 50 feet is shown in Figure 52. For
this particular data, the scan angles were along a direction
parallel to the wind direction which was from the North
(normal to the plane of the waves) , 45 degrees into the
waves and finally perpendicular to the wind direction which
was East. The results indicate little difference in the
probability function. This result implies that natural
waves are modulated not only in the direction normal to
travel but also modulated perpendicular to that direction.
This means that the system can scan a natural water surface
in any scan pattern independent of wind direction.
The next point is total angular extent of scan. Figure 53
shows two typical curves from Figure 52. They have, however,
been extrapolated to higher angles. This is based on the
probability distribution described earlier. The measured
limit of MOD 2 at 50-foot height is indicated. The
anticipated limits of MOD 3 are also shown. If these
extrapolations are valid, MOD 3 will have an angular scan
capability of + 28° at a height of 150 feet. This
corresponds to + 90 feet at the water surface. Note that
with MOD 3 at 50 feet the angular extent went up to + 35°
which corresponds to only + 35 feet. Hence, by elevating
the system a greater area of water surface can be scanned.
MOD 3 Test Results
A side by side comparison of the tungsten-iodine and the
Xenon arc indicated that the spectral efficiency of the
Xenon system is less by an order of magnitude due to its
higher temperature. This inefficiency can be compensated
for by the fact that the Xenon system can be electronically
modulated while the tungsten system must be mechanically
modulated. By pulsing a flash lamp a given amount of
energy can be dumped in a short time. This means that by
pulsing, a relatively low wattage CW lamp can produce high
peak wattage signals. Even though the Xenon system is
spectrally inefficient, it can produce a factor of 10
96
-------
Figure 52. MOD 2 Signal Amplitude Versus Angle for Three
Scan Directions
97
-------
MODEL
DATE
from N
Figure 53. Angular Scan Limits of MODs 2 and 3
98
-------
increase in system range over a system using tungsten-
iodine lamp. Because of certain advantages unique to each
system, it was decided to build a unit which would accept
either light source with a few minor modifications to effect
the interchange. The system which used a tungsten lamp
is designated as MOD 2 and the Xenon arc system is called
MOD 3. It consists of the Xenon lamp, a parabolic reflector,
the optic plate, a 20 joule capacitor bank and a short
arc lamp pulser. In addition, four printed circuit
cards which are compatible with system wiring are required.
In the MOD 3 setup, the interference at the output caused
by the high voltage firing pulse for the Xenon lamp and,
thus, the noise level at the output, was reduced to an in-
significant value by shielding the high voltage line and
its ground return between the pulser and the lamp, and by
gating the 75 microsecond detected pulse for 25 microseconds
centered around the peak. It was also determined that for
a 75 microsecond pulse 20 joules of energy would give
complete ionization of the gas and thus give a more stable
color temperature than lower energy levels. This power
level is comptible with the 10° pulses per lamp curve
discussed in Section V.
The system was tested in the laboratory and was demonstrated
to distinguish between water and oil. The data shown in
Figure 55 is not quantative since a projection distance
of only 10 feet was available whereas the system was de-
signed for 300 feet. Techniques used to attenuate the beam
by a factor of 1000 produced some distortion and extraneous
reflections. However, the data does show a clear distinction
between oil and water.
Since MOD 1, 2 and 3 had been successfully demonstrated to
distinguish between oil and water in the laboratory, the
MOD 1 and 2 had also been demonstrated at 30 and 50 feet on
a stock pond which periodically had wave action up to 3 or
more inches high, the next step was to field test the sensor
at a height and in an environment similar to final system
application. Since the system concept of oil versus water
discrimination had been established as valid so long as
signal was present, the only remaining factor was signal
level versus environment. The presence of oil was not
necessary for these tests; thus, a site 91 feet above the
Gulf of Mexico was chosen.
99
-------
Figure 54. Modification Kit for MOD 3
This test site was located on the seventh floor terrace of
the Flagship Hotel in Galveston, Texas. The hotel is
built on a wharf extending 100 yards into the Gulf of
Mexico and the seventh floor is approximately 91 feet above
the surface of the water during normal tide. The test set
up for MOD 3 is shown in Figure 56. MOD 3 was operated
during the first two days while moderate waves (6 to 12
inches) were present. Figure 56 also shows typical waves
that were present during those two days.
Test performed show that the system has reserve capabilities
which would permit operation at a 300 foot height at a large
tilt angle. The system was checked at 9 degrees and 14
degrees and strong signal returns were measured during this
period of moderate waves. However, some problems were
experienced with the electronic circuitry, which prevented
the collection of stable data out of the log difference
amplifier. Typical data is shown in Figure 57. The gains
setting of the pulse amplifiers were reduced by a factor of
10 to prevent saturation; thus, this factor of 10 would
allow operation at 300 feet. The chart recorder spans were
100
-------
^^. j_ __• ? _;.~ .i. ..jr...} i.. i . LL..- I
-::-—F—t —r: i .+.--U- —.4
_ 1
fl I'llillfl
jllMBIBltt
—i
IMC !
Hlltll
Ill I El Illl! I !
Ill) 'f! milll I
111111II
iiiiiii
11 »ii: n i s
~—T— -j 1 -
- - . — --- — -~ j - •
Figure 55. MOD 3 Lab Data
101
-------
Figure 56. Test Setup for MOD 3 at Gulf of Mexico Test Site
102
-------
o
u>
r.....i. . _ay. ..,*.. >. ../_._ ....
3.9 OUTPUT; SPAN - tov (IV/INCH)
-A -
3.4 OUTPUT SPAN = 10V (tV/INCH):
CHART SPEED 5 INCHES/M1N
CHART SPEED 1 INCH/MIN
Figure 57. Typical Data from MOD 3 at 91 Feet
-------
set at 10 volts for each channel and at 5 volts for the log
difference. The zero reference point for the log difference
output was 3 volts or at 60 on the chart. Note that the
excursions on the difference output were greater than 1 volt.
This is not satisfactory for proper operation of the oil
spill monitor. Figure 58 shows how the system would have
reacted if sample and hold circuits were used instead of
peak detectors. The main problem was caused by differences
in the RC discharge time constants. Figure 58 shows how
this same data would have been processed if sample and hold
circuits had been used. In Figure 58 there is still an
excursion of 0.35 volts while observing water. This
differential is believed to be caused by a phase difference
between the pulses in the 3.4- and 3.9-micron channels.
Figure 59 shows a line drawing of these pulses and the
relative position of the gate shows that since each peak
detector would be charging from a different slope the
outputs would not trace from pulse to pulse. An attempt to
correct this problem was made by widening the gate.
However, this defeated the purpose of the gate by allowing
the noise spike to pass.
In order to get meaningful data from MOD 3, the peak detector
should be replaced by sample and hold circuits and the gating
circuits should be changed to insure that the same part of
each pulse is being gated.
MOD 2 at 91.5 Feet
The disappointment in not being able to obtain quantitative
field data on MOD 3 was quickly overcome in observing the
results from MOD 2. These results showed that MOD 2 will
fulfill all of the design requirements, as specified in
Section V, including the 300-foot height requirement. It
will also operate at slant angles of 45° or greater with
heavy wave action.
The system was converted to MOD 2 in a manner to illustrate
that the laboratory adjusted optical plates can be easily
interchanged in the field and that the system can be placed
back in operation by merely setting the micrometers to a
predetermined value. The time to perform this interchange
was 1 hour. By using plug-in connectors rather than solder
connections and individual micrometer heads for each spare
plate, this time could be cut down to less than 1/2 hour.
MOD 2 was operated for about an hour pointed straight down
at night and signals out of the peak detectors were
between 3 to 3.5 volts with a signal-to-noise ratio of 13.
An example of this night time data is shown in Figure 60.
104
-------
CHART SPEED 5 INCHES/MIN
Figure 58. Predicted Operation with Sample and Hold Circuits
-------
Span for each scale is 5 volts. Extrapolation of this data
to 300 feet gives a signal reduction of 11 to 1 and
thus a signal-to-noise ratio of 1.2 which is still sufficient
for reliable operation. The noise in the system is an rms
value of a uniform noise source measured in the system
bandwidth. This gives a steady state mean value with
fluctuations about the mean. Both the mean and deviation
are determined by bandwidth. By subtracting out the mean
value, the effective noise used in the processing (log
differing) can be reduced and the effective range increased.
The next morning wave action had increased considerably
and was estimated to between 2- to 2.5-foot peaks with a
period of 15 to 20 feet and a repetition rate of 5 seconds.
This is the equivalent of sea state three. All previous
data had been for sea state zero. A photograph of the wave
action is shown in Figure 61. Slant range tests were very
successful and a continuous plot was obtained from 0 degrees
tilt out to 60 degrees tilt. This data is shown in Figure
62. The experiment was repeated by varying the tilt from
60° to 0°. This curve is shown in Figure 63. From 10° to
40° both curves have a signal-to-noise ratio greater than or
equal to 3 to 1. Azimuth data from -30° to +90° is shown
in Figure 64. The direction where the unit faces the waves
is designated as the zero reference angle.
3,9
3.4
GATE1
J
Figure 59. Pulse Sampling
106
-------
Figure 60. Nighttime Data from MOD 2 at 91 Feet
107
-------
•fc
*
Figure 61. Photograph of Wave Action
108
-------
AMPLITUDE VS TILT ANGLE (MOD 2 AT 91 FEET)
2.2V
2.0V
1.8V
1.6V
1.4V
1.2V
1.0V
0.8V
0.6V
0.4V
0.2V
TILT VARIED FROM (T TO 60°
NOISE LEVEL
II I II I
111 I
0° 5° 10° 15° 20° 25° 30° 35° 40° 45° 50° 55° 60°
Figure 62. Amplitude Versus Tilt Angle for MOD 2 at 91 Feet
Angle 0° to 60°
109
-------
AMPLITUDE VS. TILT ANGLE (MOD 2 AT 91 FEET
2.2V
2.0V
1.8V
1.6V
1.4V
1.2V
1.0V
0.8V
0.6V
0.4V
0.2V
TILT VARIED FROM 60° TO l
NOISE LEVEL
I I I I I I I I I I
0° 5° 10° 15° 20° 25° 30° 35° 40° 45° 50° 55°
60"
Figure 63.
Amplitude Versus Tilt Angle for MOD 2 at 91 Feet
Tilt Angle 60° to 0°
110
-------
0.8V
LU
0.6V --
0.4V —
NOISE LEVEL
0.2V--
-30° 0 30° 60° 90°
AZIMUTH ANGLE
Figure 64. Amplitude Versus Azimuth Angle, Tilt Angle 36(
111
-------
An analysis of the wave action present at the time that the
tilt data was recorded indicates that the largest slope
available is 18° (Arc tan 2.5 ft ) if sinusoidal motion is
15 ft/2
considered. This .would imply that the results obtained were
not possible. However, a close examination of Figure 61
and a recheck in the literature indicates that a "break
wave" condition exists where the wave continues to steepen
in front and eventually breaks. At this point the
mathematical theory for waves ceases to be valid. Figure
65 shows the water surface for a time well beyond the breaking
point.
Summary of Test Results
The active system was demonstrated to have all of the
capabilities listed in the introduction of this section.
All except the test at 300 feet were performed explicity.
The requirement of the test at a height of up to 300 feet
above the surface was shown implicity on the basis of
extrapolations of 30-, 50- and 91-foot data. In addition
to the original set of design requirements the system was
known to have another desirable characteristic. It can be
scanned in angle to a value set by the wave condition. The
concepts developed from the test data are given in Section
IX of this report.
Figure 65. Water Profile After Breaking
112
-------
SECTION VIII
TEST RESULTS, PASSIVE SENSOR
Oil Type Test
The passive, infrared sensor system tested can only detect
thermal anomalies. It cannot determine exact oil types by
spectroscopy or by any other means used by active systems.
With this passive system, various oil types can be deduced
by observing the differences in physical characteristics
of one oil signature from another. The following paragraphs
describe how four oil types were categorized according to
their various differing physical characteristics present in
their signatures.
Navy distillate was the first sample to be observed. Figure
66 shows the plastic sample ring (3 feet in diameter) with
clear water. Figure 67 shows 1/2 ounce of Navy distillate
being administered to the ring. Figure 68 shows patches of
sheen formed shortly after contact with water. The oil
forms a sheen quickly, but does not spread out to cover the
entire ring. Figure 69 shows the experimenter stirring the
sample for uniformity of sheen over the entire area of the
ring. Sheens of oil appear as black, or dark areas, due
to the low emission of thin sheens on water. Thick lumps
appear as bright spots, indicating higher temperature than
water or better emissivity than water.
Figure 70 shows a uniform sheen over the entire area of the
ring after stirring. Figure 71 shows the uniform sheen
remaining in spite of good wave action.
Diesel fuel oil number 2 was the second sample observed.
Figure 72 shows the ring clean again. Figure 73 shows 1/2
ounce of the oil being poured inside of the ring. Figure
74 shows the sample shortly after contact with water.
Notice that it has already spread to the entire area of
the ring. Figure 75 shows the sheen remaining in spite of
the wave action.
Diesel fuel oil number 4 was the third sample observed.
Figure 76 shows the ring cleaned out again. Figure 77
shows 1/2 ounce of oil inside the ring. Figure 78 shows
the oil spread into patches of sheen, not covering the
entire area of the ring. Figure 79 shows the oil spread
evenly over the entire area of the ring.
113
-------
Figure 66. Plastic Ring in Clean Water Before Adding
Distillate
Figure 67. Navy Distillate Being Added to Water Inside Ring
114
-------
Figure 68. Patches of Distillate Sheen
Figure 69. Distillate Being Stirred for Uniformity
115
-------
Figure 70. Uniform Distillate Sheen
Figure 71. Uniform Distillate Sheen in Presence of Waves
116
-------
Figure 72. Plastic Ring in Clean Water Before Adding Diesel
Fuel Oil No. 2
Figure 73.
Diesel Fuel Oil No. 2 Being Added to Water
Inside Ring
117
-------
Figure 74. Diesel Fuel Oil No. 2 Spread Evenly Inside Ring
Figure 75. Diesel Fuel Oil No. 2 Sheen in Presence of Waves
118
-------
Figure 76. Plastic Ring in Clean Water Before Adding Diesel
Fuel Oil No. 4
Figure 77. Diesel Fuel Oil No. 4 Being Added to Water Inside
Ring
119
-------
Figure 78. Patches of Diesel Fuel Oil No. 4 Sheen
Figure 79. Diesel Fuel Oil No. 4 Spread Evenly Inside Ring
120
-------
Florida crude oil was the fourth sample observed. Figure 80
shows 1/2 ounce of the oil poured inside of the ring. Note
the heavy, bright thick spot remaining in tact. After
awhile, the wind blew the oil to the upper right corner, as
shown in Figure 81. Figure 82 shows the removal of a small
portion of the oil to illustrate its effect in an area large
enough to not limit its spread. Figure 83 shows light and
dark spots spread over the entire area of the ring. The
lumps did not disperse. Figure 84 shows a temporary sheen
of oil after disturbance. Figure 85 shows lumps of oil
coalescing into bright spots again after being left alone
for a while. Figure 86 shows the oil sheen after most of
the oil was removed with polyurethane foam chips.
These four samples illustrate how one could identify oil
types by their signatures by merely classifying them by
their physically observed characteristics. Table 11
shows the classification for the four oils, as observed.
Figure 80.
Florida Crude Oil Being added to Water Inside
Ring
121
-------
Figure 81. Florida Crude Oil Blown into Upper Right Corner
by Wind
Figure 82. Small Portion of Florida Crude Oil Removed
122
-------
Figure 83.
Thick and Thin Spots of Florida Crude Oil
Undispersed Over Entire Ring
Figure 84. Temporary Florida Crude Oil Sheen After Disturbance
123
-------
Figure 85. Lumps of Florida Crude Oil Coalescing Into
Bright Spots
Figure 86. Florida Crude Oil Sheen After Removal of Most
Oil
124
-------
TABLE 11. OIL TYPE TABLE
Navy Diesel Diesel Florida
Characteristics Distillate Fuel #2 Fuel #4 Crude
A. Lumpy even after
Agitation 0 0 0 x
B. Light & Dark.
Initially 0 x x x
C. Spreads to Sheen
Fast x x 0 0
D. Covers Entire Area
Without Agitation 0 x 0 0
Thus, one can see how to identify di.es.el fuel oil number 2,
for example. It exhibits characteristics B, C and D in
Table 11, whereas, for example, Florida crude exhibits A
and B. This method is not absolute or fool proof, as there
are probably other oils not listed in the table that exhibit
the same characteristics as one or more of the samples. The
solution to that type of problem is to have a very large
list of characteristics, and identify each known oil
according to its differences from other oils. Even then,
there is uncertainty due to environmental conditions.
This type of test is meant only to be a rough iden-
tifier.
One overriding point to be made in this test is that all
types of oil tested could be seen with the exception of a
light "overhead" oil. this one exception evaporated too
fast for observation.
Time-Of-Day Test
A black body radiator radiates energy according to the
Stephan-Boltzman Fourth-power law: E, = a T4.
where E, = Energy radiated from the black body per unit
time per unit area exposed.
a = constant of proportionality
T = Temperature (absolute) of radiator
A gray body, such as water or oil, radiates according to
the following:
125
-------
G £g b ega
where e is a constant for any particular material which
indicatis the quality of black body it is. Thus, the
radiation of a material like the absorbtion, is a function
of the emissivity (e a) and the absolute temperature (T4).
Let's apply the preceding to the experimental objects
shown in Figure 87.
Water
Figure 87. Oil.- Water Test Configuration
The ambient temperature can heat or cool a small body of oil
in a much shorter time than the water which the oil is in.
As night falls, the ambient temperature drops; the water
and the oil will start to cool with this drop in ambient
temperature, but the oil is able to cool faster due to its
smaller volume. Thus, the temperature of the oil will be
lower than the temperature of the water at night, making
the emission energy of the oil less than that for the water.
Since less energy is being emitted from the oil than from
the water, less energy is evident to the sensor. Thus,
the ring of oil will appear dark with respect to the
water at night. This would not necessarily be true if oil
and water had better thermal conductivity at the boundary.
In the day, of course, the opposite is true. That is,
due to the oil sample's ability to raise its temperature
faster, it is a better radiator in the daytime. Thus, the
oil appears as a bright spot compared to water, as shown
in Figure 88.
126
-------
Dawn '
Dusk
Night
Night
(2)
(4)
Figure 88. Temperatures of Water, Oil, and Air Over a Day's
Time and the Corresponding Variations in Presentation
127.
-------
As can be seen in samples (2) and (4) in Figure 88, the water
and oil cannot be distinguished from each other at two
times in the diurnal cycle. This is at the time when the oil
temperature and water .temperature are equal during the
transition portion of the diurnal cycle. Thus, at two times
(of day) the passive sensor cannot be used effectively.
This duration is very short for a good sensor with low NET.
Figure 89 shows 2 oz of Florida crude oil at 1830 hours on
1-18-72 after approximately five hours of setting. Note
that due to lack of sun which results in drop in temperature,
the oil is darker than the water, which agrees with sample
(5) in Figure 88 for that time of day. Figure 90 shows the
same sample at 0600 hours on 1-19-72. Note that it is hard
to distinguish any intensity difference between the water
and oil. This agrees with sample (2) in Figure 88 for dawn.
Figure 89. Florida Crude Oil on Water, 1800 Hours, 1-18-72
128
-------
Figure 90. Florida Crude Oil on Water, 0600 Hours, 1-19-72
Figure 91 is taken at 0630 hours and still exhibits dawn
characteristics. Figure 92 still exhibits the dawn charac-
teristics of gray versus gray intensity. This was the last
dawn shot taken at 0650 hours. Thus, for this sample and
this system, the (indistinguishable) time lasted approximately
50 minutes. Figure 93 shows the same sample in the small
ring taken at 0700 hours. The sun has heated the sample
slightly. This corresponds to a region between samples (2)
and (3) in Figure 88. Figure 94 shows the same sample, in
the small ring, at 0730 hours. Note the increasing intensity
as the oil heats up. Figure 95 shows the sample at 0900
hours. Note the high intensity of oil. Also note that the
wind is blowing a small portion of the oil out. Figure 96
shows the same sample at 1100 hours. Notice that the
intensity has not increased. This is probably the maximum
temperature point of oil for that particular set of air
and water temperatures and that particular ring area and
wind speed. Figure 97 shows the sample at noon after some
of the sample was blown away. There is also disturbance,
as shown in Figure 98, due to the wind. Note that the
sample is still fairly bright compared to the water.
129
-------
Figure 99 shows the remaining sample at 1300 after wind is
settled. Note that the sample is still fairly bright. Thus,
the illustration in Figure 88 is well supported in Figures
89 through 99.
Many things can affect the results of a test like the one
described in the preceeding paragraph. For example, if
enough oil is blown out of the ring, the sample will exhibit
a sheen effect, which appears dark compared to water,
regardless of the time of day. The sheen effect will be
discussed in detail in the subsection on thickness. The
point is that many variables can and do affect the results
of a particular test, and although much has been done to
isolate variables, it is possible for some un-named
variables to affect the test results somewhat.
Figure 91. Florida Crude Oil on Water, 0630 Hours, 1-19-72
130
-------
Figure 92. Florida Crude Oil on Water, 0650 Hours, 1-19-72
Figure 93. Florida Crude Oil on Water, 0700 Hours, 1-19-72
131
-------
Figure 94. Florida Crude Oil on Water, 0730 Hours, 1-19-72
Figure 95. Florida Crude Oil on Water, 0900 Hours, 1-19-72
132
-------
Figure 96. Florida Crude Oil on Water 1100 Hours, 1-19-72
Figure 97. Florida Crude Oil on Water, Noon, 1-19-72
133
-------
Figure 98. Some Florida Crude Oil Being Blown from Ring
Debris Test
Some selected tests were made to determine what types of
debris, or non-oil types of material would be visible with
the infrared sensor. Targets of temperatures or emissivities
much greater than water should be visible.
The visibility of muddy water was tested by stirring up the
mud in the lake. The mud was not visible. Seaweed, or its
equivalent, was tested. Figure 100 shows a large clump of
seaweed floating inside the test boom in the upper left
quarter. The seaweed was not visible. A piece of plywood
was put in the boom and soaked for a few days. The wood was
visible due to the solar heating of the wood and the solar
reflection from the wood. Reflection is most conspicuous on
the portions of the wood protruding out of the water. This
is shown in Figures 101 and 102.
Ice is visible as very dark, sharply outlined objects in
both the day and night which makes it easy to differentiate
from oil.
134
-------
Figure 99. Florida Crude Oil on Water, 1300 Hours, 1-19-72
Figure 100. Seaweed Sample
135
-------
Figure 101. Plywood in Water
Figure 102. Plywood in Water After One Week
136
-------
In general, there are a few classes of materials whi.cn. are
visible. One class includes those materials capable of
reflecting thermal radiation. Another class includes those
materials with good thermal absorption properties and poor
thermal conductivity with water. This enables the latter
material to store heat; thus, it will have a higher
temperature. All oils fall into this category. Not many other
materials existing in large quantity on the ocean surface
fall into this category.
All single or solid objects have definite, unchanging
silhouettes and are not large enough to be mistaken for an
oil spill.
Thickness Test
The volume of a particular sample of oil on water is the
product of its thickness and its surface area of contact
with the water. Thus, for a given volume of oil, the
greater the thickness, the less surface area in contact with
the water. A water-oil boundary has low thermal conductivity.
Thus, for a sufficient amount of heat to be transmitted
through the boundary to allow the temperature of the oil to
follow the water temperature the contact area must be large.
The larger the volume of oil, the more heat is necessary
to raise or lower the oil temperature. This means that in
order to transport sufficient heat across a water-oil
boundary to allow the oil to follow the water temperature,
the volume must be small in proportion to the contact area
or the contact area; in proportion to the volume, must
be large. Since thickness is volume divided by surface
area, a thin sample of oil will meet the criterion mentioned
above, and the temperature of a sufficiently thin sample
of oil should follow the temperature of the water.
The thermal emissivity of water is greater than that of oil.
Thus, if a sample of oil is on the water, sufficiently thin
to allow its temperature to follow the temperature of the
water, the oil should appear darker than water to an infrared
sensor. A material of a certain temperature and emissivity
will emit less heat than another material of equal temperature,
but higher emissivity.
From the preceding discussion, it is clear that a thin sheen
of oil on water, which has had sufficient time to reach the
temperature of the water, should appear dark with respect
to the water as seen from an infrared sensor.
Conversely, a thick layer of oil should appear brighter than
water in the daylight and darker than water at night. The
reason for this was discussed in the section on time-of-day
137
-------
test. Thus, for a typical daytime experiment, with any
particular type of oil, the brighness of the oil should
vary from the same as "water when thickness is zero, to less
than water for thin sheen, to greater than water for thick
layer.
To isolate sample thickness as the variable by eliminating
wind, waves and solar radiation, a laboratory environment
was used.
The oil thickness variable was controlled by placing water
of equal volume into two identical containers and adding
crude oil to one container using a hypodermic syringe. The
syringe was graduated in cubic centimeters; therefore, by
knowing the surface area of the water, the oil thickness
could be determined. This method of determining thickness,
of course assumes a uniform distribution of the oil. To
insure this, the sample was selected to disperse rapidly
and to cover the entire area of the water surface. The
sample chosen was crude oil from the Jay Field in Florida,
donated by Sun Oil Company, Richardson, Texas.
The 8- to 14-micron sensor was set up to view the water and
oil-on-water samples perpendicular to the surface. Figure
103 illustrates the test configuration.
\
IR Energy-
iMirror
Oscilloscope
Oil
Figure 103. Oil Thickness Test Setup
138
-------
The center channel of the sensor detector array was displayed
on an oscilloscope where voltage measurements were made. The
sensor has been previously lab calibrated using a
controlled-temperature black body source (4°C per volt).
Figures 104 through 115 are the actual data for thicknesses
from 0 to 270 microns. Figure 104 shows the oscilloscope
presentation of a center channel. Voltage is read from the
curve by measuring the distance between a line drawn through
the trough on the left (oil) and a line drawn through the
trough on the right (water). Notice that the "hump" between
them is the empty center trough.
Figure 105 shows the corresponding television video. Due
to misalignment in the oscilloscope, it is necessary to use
the parallel line method instead of reading the face of the
oscilloscope directly. Figure 104 shows no oil. When the
two parallel lines are formed through right and left they
will meet. Figure 106 shows 0.25 micron thickness of oil
on the left side. Notice the parallel line drawn on the left
trough (L) is lower by 0.33°C than the parallel line on the
right trough (R). Figure 108 shows 1.45 microns of thickness
on the left. Notice that (L) is almost 0.5°C lower than (R).
Notice the contrast difference between (R) and (L) in
Figure 108. Figure 110 shows 16.5 microns of thickness on
the left. Here, the oil has passed the optimum thickness
for lowering the heat emission of the oil with respect to
the water. Notice that (L) is only 0.1°C below (R). Figure
112 shows 140 microns of thickness on the left. Notice that
the oil now appears hotter than water. (L) is 0.5° higher
than water. Figure 114 shows 270 microns of thickness on
the left. Here, (L) is 1.25°C higher than (R).
The apparent temperature of the oil surface with respect to
water as a function of the thickness is shown in Figure 116.
This curve was found to be repeatable for the given surroundings,
However, as discussed in previous reports, variations in
environmental conditions are the controlling factors.
Range Test
The effect of increased range is dominantly a slant angle
characteristic rather than a distance characteristic, unless
the infrared sensor is operating in a highly absorbtive
atmosphere. Slant angle affects the emissivity and
reflectivity characteristics of a surface as seen by the
sensor. Figure 1178shows the variation of reflectivity and
emissivity versus angle for a smooth water surface for three
wavelengths (8, 11 and 12.5 microns). From this graph it
is apparent that emissivity is significantly reduced at
slant angles greater than 70 degrees. Correspondingly,
reflectivity is increased significantly at this angle. Thus,
at angles greater than 70 degrees it becomes increasingly
139
-------
difficult to receive energy from the surface of the water.
For this reason, it becomes very difficult to detect thermal
anomalies at angles above 70 degrees.
With the sensor employed on this experiment, imagery at
approximately 85 degrees slant angle was recorded. Figure 118
shows 2 ounces of Florida crude oil in 10- foot by 10-foot
frame at 43 degrees slant angle which corresponds to a
slant range of about 40 feet. Note the excellent clarity
of the three gray shades. Figures 119 and 120 show 2 ounces
of Florida crude oil in a frame at a slant angle of 60 and
70 degrees, which correspond to slant ranges of 60 and 88
feet respectively. Note the clarity of the three gray shades
indicating thick oil, thin oil, and water in the 70 degree
presentation.
Figure 104. No Oil as Presented on Television Display
140
-------
Figure 105. No Oil as Presented on Oscilloscope
Figure 106. 2 Drops Florida Crude Oil as Presented on Television
Display
141
-------
Figure 107. 2 Drops of Florida Crude Oil as Presented on
Oscilloscope
Figure 108. 4 Drops of Florida Crude Oil as Presented on
Television Display
142
-------
Figure 109. 4 Drops of Florida Crude Oil as Presented on
Oscilloscope
Figure 110. 8 Drops of Florida Crude Oil as Presented on
Television Display
143
-------
Figure 111. 8 Drops of Florida Crude Oil as Presented on
Oscilloscope
Figure 112. 2 Cubic Centimeters of Florida Crude Oil as
Presented on Television Display
144
-------
Figure 113. 2 Cubic Centimeters of Florida Crude Oil as
Presented on Oscilloscope
Figure 114. 5 Cubic Centimeters of Florida Crude Oil as
Presented on Television Display
145
-------
Figure 115.
5 Cubic Centimeters of Florida Crude Oil as
Presented on Oscilloscope
146
-------
1000
Figure 116. Apparent Temperature Versus Oil Thickness
-------
60-
100
30 40 50 60 70 80 90
Degree inclination
Courtesy of McSwain and Bernstein, 1960
Figure 117. Reflectivity and Emissivity of a Smooth Water
Surface as a Function of Slant Angle
Figure 118. 2 Ounces of Florida Crude Oil at 43 Degrees
Slant Angle
148
-------
Figure 119. 2 Ounces of Florida Crude Oil at 60 Degrees
Slant Angle
Figure 120. 2 Ounces of Florida Crude Oil at 70 Degrees
Slant Angle
149
-------
•
Figure 121. Waves at a Slant Angle of 86 Degrees
'' i-au&'
Figure 122. Galveston Bay as Seen From the Flagship Hotel
150
-------
Later, samples at greater slant angles were observed with some
success. At the testing site, another infrared sensor was
tested for comparison. Figure 121 shows waves at approxima-
tely 2QO feet horizontal range. The sensor was approximately
14 feet above the water surface. This is a slant angle of
approximately 86 degrees. Note the clarity of wave action
at approximately 86 degrees angle or 200 feet range. Figures
122 through 124 show imagery at the same approximate slant
angle, but a far greater range. Figure 122 shows the view
of Galveston Bay as seen from the Flagship Hotel. Note the
sign on the jetty, the farthest pier, and the three posts
protruding from the water at the end of the farthest pier.
Figure 123. Galveston Bay at Slant Angle of 70 Degrees
151
-------
Figure 124. Galveston Bay at Slant Angle of 86 Degrees
Figure 123 shows imagery taken with the sensor at 70 degrees
slant angle from approximately 90 feet above the water surface.
The vertical lines at the top of the picture are the supporting
members of the pier. The bright square at the bottom of the
picture is the sign on the jetty. The bright spots to the
left are wave action at an angle of 70 degrees and a horizontal
range of over 400 feet. Figure 124 shows the imagery of the
bay at an angle of 86 degrees at the cross hairs. Note the
three posts protruding from the water in the upper right
corner, and the pier in the upper left. Some wave action is
shown in the lower left, which can be approximated at above
80-degrees angle.
Thus, the imagery capabilities of the system at varying
ranges are affected more by the slant angle than the distance.
Figure 121 shows an angle of 86 degrees at 200 feet,
whereas Figure 124 shows imagery at 86 degrees angle at over
1000 feet horizontal range.
152
-------
substantially increase the acoustical noise, and would
introduce forced vibrations to the components. The natural-
convection heatsink would be slightly larger, but would
weigh only 12 ounces and be totally passive in operation.
The heat dissipation capability of the natural-convection
heatsink is limited by fin surface area rather than internal
thermal conduction. Therefore, this type could be fabricated
of aluminum instead of a material with a higher thermal
conductivity without loss in heat dissipating efficiency.
This natural-convection heatsink is proposed for use on the
3- to 5-micron PbSe detectors.
153
-------
SECTION IX
SYSTEM CONCEPTS, ACTIVE SENSOR
Introduction
Unit sensor cost is an important factor in the oil spill
surevillance system concept and application study. When
the unit cost for the active sensor is considered, it has
been determined that this system is optimum for applications
where a staring or limited scanning capability is required.
Examples are estuaries, drainage channels, offshore drilling
platforms and separator platforms. In this section, the
system concepts for these applications will be discussed
in the general terms of performance versus height, scan angle
and meantime between failure. The specifics of which system
for a given application and cost analysis are presented in
Section XI.
Sensor Capabilities
The previous sections discussed the optical efficiencies of
MOD 2 and MOD 3, both from a theoretical and experimental
point of view. These results are summarized in Figures 125
through 130. In these figures, the expected signal-to-noise
ratio for two system MODs are shown for three different
altitudes as a function of scan radius. The graphs show
the point at which the signal reaches such a low level that
processing will no longer distinguish oil from water as
discussed in Section VII. These points are indicated by a
dashed line. The break points for several detector-filter
combinations have been used (See Table 5). The detector
temperatures of 243°K and 193°K can be achieved by using
thermoelectric coolers which have relatively low cost. For
77°K operation in the field, an open cycle cooler is
recommended. Note that there is an uncertainty in the
values for PbSe and InSb when cold filtering is used since
this is dependent upon detector selection.
The curves in Figures 125 through 130 are based on a combi-
nation of measured and extrapolated data for the case of
smooth wave action. This was typical of cases at the Texas
Instruments test facility where the waves did not have
crests. On the basis of the test performed at the Gulf of
Mexico, these curves are conservative. For normal incidence,
that is zero scan radius, the smoother the surface the larger
the signal-to-noise ratio will be. From the Gulf of Mexico
test it was found that for the normal wave action the
15 5
-------
10
-1
10
-2
o
in
> -3
10
-4
10
10
-6
PbSe AT 243 K
\ PbSeAT193 K
lnSbAT77°K
]PbSe AT 193°K
___j(COLD FILTERED)
•"]lnSbAT77°K
JICOLD FILTER ED)
10
20 30 40
SCAN RAD I US (FEET)
50
Figure 125,
Signal Versus Scan Radius for MOD 2 at 50-Foot
Altitude
156
-------
10°
10
-1
10
-2
o
UJ
o:
10
-3
10
-4
10
-5
10
-6
PbSeAT243 K
PbSeAT193 K
lnSbAT77 K
~ PbSeAT193uK
-j(COLD FILTERED)
~HnSbAT77°K
(COLD FILTERED)
20
40 60 80
SCAN RAD I US (FEET)
100
Figure 126,
Signal Versus Scan Radius for MOD 2 at 100-Foot
Altitude
157
-------
10
-1
10
-2
o
to
10
-4
10
-5
10
-6
_ PbSeAT243uK
. PbSeAT193°K
__ lnSbAT77uK
PbSe AT 193K
4 (COLD FILTER ED)
lnSbAT77uK
(COLD FILTER ED)
j I
20 40 60 80 100 120 140 160
SCAN RAD I US (FEET)
Figure 127,
Signal Versus Scan Radius for MOD 2 at 150-Foot
Altitude
158
-------
10
-1
10
-2
GO -3
1Q
10
-4
10
-5
10
•6
PbSeAT193°K (COLD FILTERED)
InSb AT 77 K (COLD FILTERED)
10 20 30 40
SCAN RADIUS (FEET)
50
Figure 128.
Signal Versus Scan Radius for MOD 3 at 50-Foot
Altitude
159
-------
10C
10
-1
10
-2
10
10
-5
10
lnSbAT77K
PbSe AT_193UK (COLDJIJLTERED)
IbSb AT 77°K (COLD FILTERED)
«
20
0
SCAN RAD I US (FEET)
100
Figure 129.
Signal Versus Scan Radius for MOD 3 at 100-Foot
Altitude
160
-------
10°.
10
-1
10
-2
o
CO
10
-4
10
-5
10
-6
_\ PbSeAT243wK
PbSeAT193uK
Ny lnSbAT77°K
] PbSe AT 193°K
—] (COLD FILTERED)
__X lnSbAT77°K(CF)
20 40 60 80 100 120 140 160
SCAN RAD I US (FEET)
Figure 13Q.
Signal Versus Scan Radius fro MOD 3 at 150-Foot
Altitude
161
-------
system performance was a factor of two better than indicated
by the curves at 100 .feet. For the waves that were cresting,
the curves are a factor of two too low at zero scan radius.
However, for this case of cresting waves, the peak signals
which occurred at approximately 35° to 40° were higher by a
factor of 2 to 3. This type of wave action gave a MOD 2
scan radius of 170 feet for PbSe detectors at 243°K. Hence,
it is felt that the curves shown in Figures 125 through
130 give a conservative estimate for system performance on
any natural body of water. These curves will, therefore,
be used as specifications of the active sensor capabilities.
Note that for MOD 2 with PbSe detectors at 243°K an altitude
of 300 feet is possible for zero scan radius and that for
MOD 3 it is significantly greater than the 300 foot value.
Possible Scan and Display Concepts
The three possible configurations for using the angle and
range capabilities of the active sensor are staring, arc
scan, and area scan. The choice depends upon the application,
Similarly, the options available in alarm and/or display
units are alarm, chart, and CRT storage with hard copy.
In some applications it will be essential to scan a region
of water surface, such as an effluent region where the area
exceeds the instruments field of view. To accomplish this,
a simple arc scan is recommended as is shown in Figure 131.
When an indication of direction of flow or size of the slick
is required, an area scan is required. The scan pattern as
shown in Figure 131 is a sequence of arc scans each with a
larger elevation angle. The minimum angle is for essentially
normal incidence and the maximum angle is determined by the
scan radius. The increment is fixed by the spot size of
sensor at the water surface. An example of a gimbal unit
for area coverage by the active sensor is shown in Figure
132.
The primary function of the active sensor is to alert the
staff of a facility that there is an oil spill or leak.
This can easily be done due to the type of sensor output.
The signature output is a well-defined voltage that changes
by a set amount as oil moves into the field of view. Hence,
a simple electronic switch can be closed if oil is detected.
This can be used to sound an alarm such as a horn at the
sensor or can be relayed by a cable or telephone type of line
to a central control station. Specific examples are covered
in Section XI.
162
-------
AREA SCAN
ARC SCAN
PATHS AND
•DIRECTIONS
OF SCANS
OJ
PATH AND
•DIRECTION
OF SCAN
'THERE ARE A
TOTAL OF EITHER
20 OR 40 ELEVATION
'POSITIONS WITH COMPLETE
270 ASMUTH COVERAGE FOR
EACH.
THERE ISA
SINGLE ELEVATION
'POSITION FOR 270
'ASMUTH COVERAGE. THE
ELEVATION ANGLE IS AT MAX-
IMUM.
Figure 131. Concept for Arc and Area Scans
-------
AZIMUTH AXIS
AZIMUTH AXIS MOTOR
AND RESOLVER
Figure 132. Gimbal System for Active Sensor
-------
Some users of the sensor will need a hard copy of the extent
of the spill. For a staring system this will be a 1-inch
strip chart with, a pen deflection when oil is present. For
arc scan coverage the most efficient recording is also a
single pen strip chart recorder, while for area scan a
storage tube display is required. In both instances an "on"
condition indicates the presence of oil. Figure 133 shows
the plot circuitry for a 90° area scanner. The resolvers
give the angular position of the azimuth scan and convert
this to factors related to cos 6 and sin 6. The values of
E are incremented by the tilt position as indicated by EN
where N is the Nth. position. The values of ENcos0 and
ENsine8 drive the X and Y axis of the display. The oil
present condition is indicated by the output of the difference
amplifier which drives the display "on" or "off". The
resulting display is shown in Figure 134. The heavy lines
that appear on the display indicate the presence of oil.
The storage tube display can be adapted to provide hard
copy on 35 mm film. A record will be taken only if oil is
present on the display. A clock annotation will also be
recorded on the film.
For an arc scan, the recording can be simplified in the
interest of cost, although it will have a distorted rep-
resentation. A strip chart recorder is recommended with
the pen deflection being proportional to the arc. The pen
is dropped only when oil is present. If the geometry permits
a line scan could be used to prevent this strip chart
display from being distorted.
Mean Time Between Failures
The two key factors that effect the maintenance of the active
sensor are illuminator usage and the environment. Another
minor factor is the detector response.
The optimum operating points for the tungsten halogen lamp
and the Xenon flash lamp are described in Section V under
transmitter optics. It was determined that the tungsten
halogen lamp would operate continuously for 100 days without
loss of system reliability. The Xenon flash lamp has a
10 pulse rating; hence, the time between failures is
determined by the pulse rate, that is, the application of
the sensor.
165
-------
H.P.
OSCILLATOR
FREQ=10 KC
I
TILT VARIED'
IN 1° STEPS
FROM DIFFERENCE
AMP.
8V-
PULSE SHARER
AND
RELAY DRIVER
RE SOLVER
E.COS&
v
X-Y PLOTTER
_L7
PEN
DOWN
SIGNAL
Figure 133. Scanner Block Diagram
-------
SCAN CONCEPT
TOWER
PLATFORM
CO
o
o
Z
UJ
\\\\yi •,'•. \
\\ \\\\\\\\\
X = E.. SIN0
IM
OIL SLICK
Figure 134. Area Scan Concept and Display
-------
The front surface optics have been set up as refractive elements
so that the inside housing can be sealed off from the
environment. Though not shown in Section V, MOD 3 production
items will have a window in front of the parabolic reflector
in the transmitter optics. It is the coatings on the re-
fractive optics which, therefore, become critical in terms
of the environment. In a harsh salt air environment the
front coatings on the optics should be cleaned and possibly
refurbished every six months. In a les-s severe environment
the time is indefinite. However, since it is anticipated
that the lamp should be checked every 100 days, the external
optical surface should also be cleaned at that time.
The critical item in the electronics is the detector which
degrades slowly due to leaks in the hermetic seal. This
corresponds roughly to a loss of 10 percent in sensitivity
in the first year. The loss in sensitivity can be decreased
by improving the vacuum.
Trade-Off Considerations
As discussed in the preceeding paragraphs of this section,
there are two factors, simplicity and meantime between failures
(MTBF)/which must be evaluated relative to system performance.
Simplicity when considering mass produced quantities is
important in determining the unit cost.
For a staring system the performance is determined by the range
required; that is, the altitude at which the system is
mounted. The MOD 2 sensor composed of PbSe detectors operating
at 243°K can be used at altitudes of up to 3.00 feet in the
staring mode. Since this sensor uses the simple tungsten
lamp for a source, it has the maximum MTBF and the lowest
cost in production. The MOD 2 sensor is, therefore,
optimum for any application that requires a staring sensor.
In addition, the MOD 2 sensor should be used for any scan
application that is required.
Tables 12 through 15 give the tradeoff values for MOD 2 and
MOD 3 sensors at 150-foot altitude. Two values of scan
radius are used, one for systems with PbSe detectors at
243°K and the other for cold filtered PbSe detectors at
193°K. The radius values come from Figures 127 and 130.
The dwell times or pulse rates are based on a 10-minute frame
time whether it is arc or area scan • The optical resolution
is fixed by sensor type as described in Section V. However,
two cases are considered for display resolution. In the
first, the Rayleigh criteria is used and in the second,
twice the Rayleigh criteria is used. It is recommended
that the value of twice the Rayleigh criteria be used such
that the distance between elements equals the size spot.
168
-------
TABLE 12. MOD 2 COVERAGE TRADEOFF
PbSe DETECTORS AT 243°K
Based Upon (1) 10 Minute Frame Time
(2) 30 Foot Scan Radius from 150 Foot Altitude
with. 3 Foot Spot Size
Area Arc
Type of Coverage Rayleigh 2Rayleigh Rayleigh 2Rayleigh
__ Resolution (1.5 Ft) (3 Ft) (1.5 Ft) (3 Ft)
Number of Resolution
Elements 990 260 94 47
Dwell Time
(Seconds/Element) 0.68 2.3 6.4 12.8
TABLE 13. MOD 2 COVERAGE TRADEOFF
PbSe DETECTORS COLD FILTERED AT 193°K
Based Upon (1) 10 Minute Frame Time
(2) 60 Foot Scan Radius From 150 Foot Altitude
with 3 Foot Spot Size
Area Arc
Type of Coverage Rayleigh 2Rayleigh Rayleigh 2Rayleigh
Resolution (1.5 Ft) (3 Ft) (1.5 Ft) (3 Ft)
Number of Resolution
Elements 3860 990 189 94
Dwell Time
(Seconds/Element) 0.16* 0.6 3.2 6.4
*Unattainable
169
-------
TABLE 14. MOD 3 COVERAGE TRADEOFF
PbSe DETECTORS AT 243°K
Based Upon (1) lOgMinute Frame Time
(2) 10 Pulses From the Lamp
(3) 83 Foot Scan Radius From 150 Foot Altitude
with. 6 Foot Spot Size
Area Arc
Type of Coverage Rayleigh 2Rayleigh Rayleigh 2 Rayleigh
Resolution (3 Ft) (6 Ft) (3 Ft) (6 Ft)
Number of Resolution
Elements
Number of Frames
per Lamp
Number of Days
Per Lamp
Pulse Rate
(Pulses per Second)
1910
520
3.6
3.2*
490
2020
14.1
0.8
130
7690
53.4
0.9
65
15380
107
1..0
Marginal Performance
TABLE 15. MOD 3 COVERAGE TRADEOFF
PbSe DETECTORS COLD FILTERED AT 193°K
Based Upon (1) 10 Minute Frame Time
(2) 106 Pulses From the Lamp
(3) 110 Foot Scan Radius From 150 Foot Altitude
with 6 Foot Spot Size
Area Arc
Type of Coverage Rayleigh 2Rayleigh Rayleigh 2Rayleigh
Resolution (3 Ft) (6 Ft) (3 Ft) (6 Ft)
Number of Resolution
Elements
Number of Frames
Per Lamp
Number of Days
Per Lamp
Pulse Rate
(Pulses per Second)
3226
310
2.2
5... 4.
806
1241
8.6
.1.3.
173
5780
40.1
0..3.
86
11628
80.8
0..15
*Marginal Performance
170
-------
It should be noted that for those cases where the dwell
time is longer than one second, the signal-to-noise ratio
could be increased by the square root of the magnitude dwell
time. However, thi.sr is a second order effect in the size of
the scan radius; due to the strong dependence on wave action
at the larger radius values. For example, the most extreme
case is a factor of 12.8-second dwell time. This occurs for
MOD 2 in an arc scan mode. The signal-to-noise ratio could
be increased by a factor of 3.5 which would increase the
the scan radius from 30 feet to 40 feet. For all other
examples, the change is significantly less and, hence, has
been ignored in the tables.
The other way that the dwell time could be interpreted is
that the frame time of 10 minutes could be decreased by the
ratio of 1 second to the dwell time. Again, for the MOD 2
case with the 12.8-second dwell time, the frame time could
be decreased to 0.8 minutes and still have a 30-foot scan
radius. The possible applications are discussed in Section
XI.
171
-------
SECTION X
SYSTEM CONCEPTS, PASSIVE SENSOR
Introduction
Analysis of data collected during the demonstration phase
of the Oil Spill study program has helped establish the
criteria for development of system concepts. Thermal
properties of oil on a water surface have been recorded
for a variety of environmental and oil film conditions.
Several oil types were investigated as well as various
volumes. The primary conclusion drawn from the test program
is that under most conditions, oil causes a thermal anomaly on
water surfaces; thus, the thermal infrared portion of the
electro-magnetic spectrum offers excellent opportunities
for detection and surveillance. Probably the greatest
advantage that thermal infrared offers is the ability to
utilize a passive sensor and work in total darkness under
most weather conditions.
Texas Instruments approach to the problem of defining system
concepts, based on the results of the test phase, has been
twofold. The first approach was to optimize infrared (IR)
detection techniques. The second approach was to define,
based on the applications, the system configuration best
suited to solving the problem for all the various locations
and situations. The Multipurpose Infrared System .(MIRS),
developed by Texas Instruments and used in the demonstration
phase of this program has proven that it can provide useful
real-time imagery of oil-to-water thermal contrast. However,
the Texas Instruments approach was not to necessarily develop
concepts centered around this sensor. The problem has been
evaluated objectively to arrive at the optimum system concept.
The first goal, optimization of IR detection techniques, refers
to basic system concepts. The primary objective here was to
determine the spectral band to be used, which determines the
detector type, which in turn dictates the cryogenic requirements.
Recent research by individuals, in addition to that work done
by Texas Instruments during the test phase of this program,
indicates that thermal IR offers the most potential for
imaging of oil spills on water surfaces. Polcyn and Wezernak
in a report entitled Technological Assessment of Remote
Sensing Systems for Water Pollution Control-Oil Pollution
reported that "The ultraviolet, red, near infrared, and the
8 to 14 micron infrared appear to provide the optimum regions
for oil pollution detection." They further state "Radar, microwave
173
-------
radiometry, and the 8 to 14 micron infrared techniques are
unaffected by hours of darkness; and hence offer a 24 hour
detection capability. Radar and microwave will be affected
by sea-state and thus of doubtful use for monitoring under
calm water conditions." The system described in this section
utilizes the thermal properties of oil as the detection and
surveillance mechanism.
The many and varied possible applications (i.e., rivers,
harbors, estuaries, etc.) located on towers, bridges, buoys,
loading docks, offshore oil platforms, and others have
dictated the requirement for the second goal. That goal was
to define the specific system configuration best suited and
most versatile for the solution of the oil spill surveillance
problem.
The key to attaining this goal was applications of the sensor.
The principles of operation and components of this sensor had
to be such that it could be adapted to the various situations
requiring water surface monitoring. A discussion of how the
sensor can be applied to various situations is in Section XI.
»
The fixed, platform-mounted imaging sensor described in this
section does not propose to positively identify oil spills as
to type or thickness. It is proposed to complement the active
oil detection sensor where applicable or stand alone where
necessary. Texas Instruments feels the approach taken will
provide the most versatile and economical source of imagery
of oil on inland and coastal waterways.
System Description
The system conceived to fulfill the requirements of this
program is a single-channel line-scanning device as opposed
to the multichannel framing sensor used for the demonstration
phase. The line scan approach was chosen to permit versatility
in adapting the system to different applications. A stationary
sensor continuously scanning one line at a time to detect the
presence of oil can be used across rivers and canals, down-
stream from loading docks, or possibly mounted on a ship as
illustrated in Figure 135. To monitor a small bay or inlet,
an area sector scan could be employed. This can be accomp-
lished using the same sensor scanning from vertical radially
outward. The desired azimuth angle can be attained using
a gimbal system. This configuration would cover a full
360 degrees if desired for an application such as an offshore
oil platform. Figure 136 shows the area coverage capability
of this technique.
174
-------
Figure 135. Shipboard Application of Passive Scanner
175
-------
CONCEPTUAL LINE SCANNER
APPROACH TO OIL SPILL
SURVEILLANCE SYSTEM
Figure 136. Conceptual Line Scanner Approach to Oil Spill Surveillance System
-------
The scanner is conceived to be a variation of the RS-18, an
airborne IR line scanner built by Texas Instruments. The
RS-18 is intended for external mounting on aircraft and is
therefore aerodynamically designed. This, of course, would
hot be a requirement for mounting the sensor on a fixed
platform, but a mechanical reconfiguration of the sensor
would produce an applicable system. Outline drawings of
the RS-18 are shown in Figure 137.
The scanner utilizes a single-sided wedge scan mirror with
a cassegramian optical system. Additional parameters of the
existing RS-18 are listed in Table 16. The scan mirror is
rotated by a variable-speed dc motor. Scan rates from
100 scans per second down to a few scans per second are
possible to provide the best presentation for the particular
application.
This system with a 3-milliradian resolution capability will
provide contiguous coverage of a 360-degree area with a
frame time of 21 seconds if operating at 6000 rpm (100 scans
per second). A multichannel framing system with a 10 by 5
degree field of view would require external scanning in both
aximuth and elevation and would require a longer time to
cover the same area. Additionally, the entire area being
surveyed may be continuously monitored in real-time using
a storage tube display system. Other advantages of this
approach over the multichannel imaging system are:
(1) Simplicity - single channel
(2) Bandwidth - 100 kHz maximum
(3) Ease of calibration
(4) Low cost
(5) Reliability
(6) Continuous areal coverage
(7) Single axis external scan, azimuth
(8) Declassified status
(9) Versatility
177
-------
AMBIENT
BLACKBODY
(NO.
•CAN DRIVE
MOTOR
ASSEMBLY.
0>
<^.... i^e; .ry
r^." --.''' :!
SCANNU INTERFACE
CONNECTOR
OEWAR. PREAMP
ASSEMBLY
PRIMARY MIRROR
VSCAN MIHROH ASSEMBLY
Figure 137. RS-18 Outline Dimensions
178
-------
TABLE 16. RS-18 SPECIFICATIONS
Scan speed
V/H
IFOV
Field of view
Roll compensation
Roll rate for 1-mrad perturbation
detectors
Cooling type
Hold time
Aperture
Focal length
Power required
Calibration Sources
5000 rpm
0.015 to 012
2.5 mrad
100 degrees
+ 15 degrees
17 degrees/second minimum
1-HgCdTe (cooled)
1-InSb (cooled)
1-Silicon (uncooled)
Dewar
6 hours, typical
2
7.7 inches
6 inches
Less than 7.4 amperes
at 28 Vdc
1-ambient
1-heated
The field of view of the scanner can also be made variable to
adapt it to various conditions. This can be accomplished by
maintaining the full 100-degree field of view design of the
system and electronically reducing the field of view of the
output signal to any value desired.
In choosing detector material, the two primary tradeoff
considerations were sensitivity to oil and the logistics
problems encountered in cooling the detector with liquid
nitrogen on a remote platform.
Atmospheric transmission of IR energy is greatest in the
3 to 5 micron region and the 8 to 14 micron region of the
electromagnetic spectrum. Thus, .a choice of detector
materials sensitive in one or both of these regions is
recommended. Lead selenide (PbSe) is sensitive in the 3 to
5 micron region and mercury-cadmium-telluride (HgCdTe) is
sensitive in the 8 to 14 micron region.
179
-------
60382
10
6000K
4000K
3000K
WIEN DISPLACEMENT LAW
. GRAY BODY CURVE
€ - 0.8 0
T = 1500 °K
500K
300K
46810 20
WAVELENGTH, MICRONS
Figure 138. Spectral Radiant Emittance Curves for
Blackbodies at Several Absolute Temperatures
180
-------
In choosing between these two materials, one must first choose
which consideration is more important. According to Wein's
Displacement Lav;, the wavelength (A) for maximum emission of
energy from a black body is a function of its ambient tempera-
ture as described by
A max T = 2900 (1)
where A max is in microns and T is in degrees Kelvin. This
relationship is presented graphically in Figure 139. This
graph shows that much less emissivity is available in the
3 to 5 micron region than in the 8 to 14 micron region for
ambient temperatures at about 300°K. Thus, if sensitivity
is the primary consideration, even over logistics problems,
HgCdTe is the logical choice.
For effective sensitivity to targets at ambient temperature,
it is necessary to cool HgCdTe to 77°K, which is accomplished
with a liquid nitrogen cooler. The following paragraphs
describe the cooling requirements.
Cooling Requirements
The following paragraphs describe a conceptual design that
is compatible with the thermal requirements of cooling a single
element HgCdTe detector operating at 77°K that does not require
a closed cycle cooler nor a frequent supply of a liquid coolant
such a liquid nitrogen. (The word "frequent" is also discussed
in more detail.)
In order to avoid the use of a closed-cycle cooler, a liquid
transfer system is proposed. The only way that a liquid
transfer system will be practical will be to have a large
volume of liquid nitrogen placed on the tower itself. An
earlier study indicated that the losses for a 100 foot vacuum
insulated transfer line from a ground storage vessel would be
too large (22 watts) to make this a practical alternative.
After consideration of a number of geometrical arrangements, it
is suggested that the detector/optics/electronics housing be
mounted on top of the storage vessel or on a platform just
above the vessel. Figure 140 shows how this arrangement will
allow an unrestricted 360 degree field of view for the detector
and keep the supply of liquid nitrogen close to the detector
to be cooled.
181
-------
Sensor/Optics Rotates
Full 360° and Reverses.
Rate of Rotation 90°/Min,
5O Liter
Storage Dewar
Flexible,'Vacuum
Insulated Line
1/8" ID.
6' Long
100'
Ground Level
Free Standing Tower
Figure 139. Liquid Transfer System and Passive Sensor Mounted
on Tower
182
-------
1/8" Vacuum
Insulated Line
CO
OJ
P_ = Ambient Pressure
Supply
h = difference in
elevation
Figure 140. Transfer Piping System
-------
LR-10
LR-17
LR-31
LR-40
LR-50
10.4
17.4
30.0
46.6
50.0
Linde Division of Union Carbide offers a whole line of large
volume storage Dewars with relatively long hold times.
Table 17 below summarizes these Dewars with their important
characteristics .
TABLE 17. DEWAR CHARACTERISTICS
Union Carbide Max Hold Time Weight
Catalog No. Capacity (liters) (days) _ (Nom)
30 33
48 50
90 88
21 128
100 130
The LR-50 container is recommended for this application. It
has a capacity of 50 liters and a hold time of 100 days. The
only restrictive factor would be the weight but 130 Ibm
does not seem to be too large for this application. The
hold time represents a static evaporation rate of 0.5 liter/
day.
The detector-Dewar has a measured capacity of 3.5 ounces
of liquid nitrogen by weight. The large supply container
has a capacity of 50 liters (equivalent to a 89 Ibm of
liquid nitrogen) . The problem that presents itself is how
to achieve the transfer of the liquid nitrogen from the con-
tainer to the small Dewar with minimum losses.
Figure 140 shows a transfer piping system. Since the
detector-Dewar must operate at nearly ambient pressure,
valve V~ will be a check valve set at 2 to 3 psig above
atmospheric pressure. This will allow the boil-off to
escape from the Dewar. Valve V, will also be a check
valve but it will be set at a pressure level that will
force fluid through the 1/8 vacuum insulated line against
the pressure in the small Dewar, P2/ and the pressure head
due to the difference in elevations, h. The pressure, P^
must therefore be in excess of the value given by
P >
1-
184
-------
where _
LN2 = 50.61 Ibm/ft
g = 32.2 ft/sec2
gc = 32.2 Ihm ft/sec2 Ib
h = elevation in. ft
2
p = pressure in. lbf/ft
The resultant equation becomes
Pl - P2 + (50-61)h
Once the exact geometry is set, the check valve may be
selected that will control the pressure, P,, to the extent
that it will be great enough to accomplish the fluid transfer.
For purposes of this discussion, the pressure drop due to
the small amount of liquid nitrogen through the 0.125 inch
inside dimension (ID) line has been ignored.
This proposed transfer system is extremely simple in that it
requires no level sensors, no solenoid values, no sequencing
operations and as will be demonstrated in the next section,
a minimum heat input to the system.
The frequency at which the 50 liter container must be refilled
depends on four factors:
(1) Heat load requirements of the detector-Dewar
assembly.
(2) Losses in the vacuum insulated transfer line.
(3) Static evaporation rates from the 50 liter container
itself.
(4) Effect of variable ambient temperature.
These four factors will be discussed independently.
Heat Load Requirements of the Detector Assembly
The total heat load requirements for a single element
HgCdTe detector are determined through calculation of the
following quantities:
(1) Heat generation in the detector due to the
resistance of the detector and the bias power at
185
-------
(2) Conduction of energy through the detector leads.
(3) Radiant heat load on the detector from the window
and the surrounding vacuum enclosure and lead tube.
(4) Joulean heating in the detector leads.
There are some experimental data regarding this particular
system that has been used in the RS-18 system. In that case,
the small Dewar had to be refilled every 6 hours. It is
estimated that approximately 75 percent of the liquid had
been evaporated at that time. The full 3.5 ounces capacity
is equivalent to 0.218 Ibm. The 75 percent usage factor
yields a mass of approximately 0.150 Ibm. Using this mass
and a time of 6 hours the usage rate becomes
usage rate = m = -— lbm = 0'025 lbm/Hr
b rir
This represents a heat load of
q = iti h_
H fg
q = (0.025 ||SL) (85.32 Btu) (1 watt/3.41 Btu/Hr)
q = 0.625 watts
This is a reasonable heat load estimate for the detector.
Heat losses in the Vacuum Insulated Transfer Line
The vacuum insulated line (flexible) suggested for this
application is manufactured by Hofman Division of Minnesota
Valley Engineering Incorporated. For the small flow rates
involved the 1/8 inch nominal hose size line is recommended.
ID = 0.125 inches
OD = 2.25 inches
Assuming that a length of 6 feet is required to allow for the
rotation of the optical scanner housing, the heat loss is
q = (0.3 Btu/hr ft) (6 ft)
=1.8 Btu/hr ^
q - 0.53 watt
This represents a loss in mass of
186
-------
• _ g _ 1.8 Btu/hr _ n n~, Ibm
hf 85.32 Btu/lbm U-UZJ- Hr
Static Evaporation From the Large Container
The product bulletin regarding the 50 liter Linde container
indicates that the static evaporation rate is 0.5 liters/day.
The density of liquid nitrogen is 1.77 Ibm/liter resulting
in a loss given by:
• n HOT Ibm
m = 0. U J / —r-—
The total loss in mass due to all sources of heat input is
therefore given by:
m (vacuum insulated line) = 0.021
m (static tank evaporation) = 0.037
m (detector heat load) = 0.025
= 0.083 Ibm/hr
The total capacity of the 50 liter tank is 89 Ibm of liquid
nitrogen. It is therefore estimated that it will last for
a period of time given by:
Time = 89 lbm =45 davs
lline 0.083 Ibm/hr 4D aays
It is therefore concluded that every month to a month and
a half the large container will have to be refilled.
187
-------
Effect of Variable Ambient Temperature
The static evaporation rate of Q.5 liters/day used previously
is for an ambient temperature of 7Q°F. It is obvious that
the application requiring the placement of the container
on top of a tower will expose the system to greater losses.
Discussions with representatives from Linde Division of
Union Carbide reveal no insight into the problem; it was
indicated that they felt the increase to be rather small.
In order to make an estimate of the change in ambient heat
load, it is necessary to determine the temperature of the
outside surface of the vessel. This will be the radiation
equilibrium temperature. It may be determined by equating
the solar heat load on the vessel to the heat lost by
convection to the ambient air and by radiation exchange with
its surroundings.
q solar = q conv + q rad.
(1) Solar Heat Load: The solar constant is 442 Btu/hr
ft. Assuming that the transmissivity of the atmosphere is
equal to 0.8, the irradiation on the surface of the vessel
is given by
Gi = T Go
» (Q.442 BTU/hr ft2)(0.8)
= 353.6 BTU/hr ft2
Assuming an absorptivity of 1.0 which would be a "worst case"
as far as heat load would be concerned, the solar heat load
may be calculated as
q , = otG.A
nsolar i
= (1.0) (353.6 BTU/hr ft2) (3.37 ft2)
= 1191.6 BTU/hr
(2) Convective Heat Loss: the heat loss due to
connection to the ambient air may be estimated by
q = hA(t - T . )
^conv w amb
Assuming a value of the convection coefficient equal to 2.0,
an area of 10 ft2, and an ambient temperature of 100°F (560 R),
we have
q = (2.0 BTU/hr ft2°F)(10 ft2)(TW-560)°R
conv '
= 20 (T - 560) BTU/hr
W
188
-------
(3) Net Radiation Loss to Ambient: the net radiation
loss to the ambient surroundings may be given by
qrad = CCT AF Tw4 = T^ 4
where
e = surface emissivity = 1.0
a = Stefan-Boltzmann constant = 0.1714 x 10 Btu
hrft2°R4
A = surface area, ft = 10 ft
F = view factor to surroundings = 1.0
Tamb = amt)ient temperature = 100°F = 560°R
The radiation loss is therefore
qrad - (1.0) (0.174 x 10-8 ) dO ft2) T - 5604
= (1.714 x 1Q~8) T 4 - 5604
w
Equating the solar heat load to the sum of the heat losses
through convection and radiation and rearranging we get
T 4 + (1.165 x 109) T - (8.21 x 1011) = 0
w w
Solving by trial and error we get the equilibrium surface
temperature
T = 600°R - 140°F
w
This represents an increase by a factor of two in the surface
temperature; laboratory surface temperature was previously
assumbed to be 70°F.
Lack of knowledge of the surroundings prevents the above
analysis from taking into account reflections from other surfaces,
These reflections will tend to increase the radiant load
and thereby the equilibrium surface temperature. In an effort
to be extremely conservative in the calculation of the heat
load, it is assumed that the equilibrium surface temperature
is equal to 180°F. An estimate of the heat leak into the
container may now be made.
189
-------
The container selected for use in this application will be
filled with liquid nitrogen and will have an interior temp-
erature of -320 F. The container is constructed of aluminum
inner and outer vessels with a plastic neck tube. The
overall dimensions of a typical vessel would be
24 inches high with an 18 inch outer diameter and a 15.25
inch inner diameter.
The vacuum space between the inner and outer walls is 1.375
inches thick and is filled with Linde Super Insulation.
This Super Insulation consists of alternating layers of
radiation shields and spacer materials in a high vacuum of
0.1 micron of mercury or lower. Also contained in the
vacuum space are Linde Molecular Sieves which adsorb trace
amounts of gases that may off-gas from the metal and insulation,
A chemical getter is also used to ensure long vacuum life.
The thermal conductivity of the Super Insulation is shown
in Figure 141. An equation representing these experimental
data in the form of the thermal conductivity as a function
of temperature is
k(t) = (2.2 x 10~5) + (1.06 x 10~7)(t + 320)
This is of the form
kf-M = V M = RTM
JY \ u / i\. ^ j_ — p j. /
Now if
T = t + 320
then kQ = 2.2 x 10
and 3 = 0.00482
o
Assuming a simplified model of one dimensional conduction
through the super insulation which essentially ignores the
losses through the plastic neck tube, the heat transfer
rate is given by
2irko , . .8 ,_ „ . (T - T )
q — r2Yrl u V.L-I T i0; VJ-T -"-oV
where
Ti, = 0 when t-L = - 320°F case (a)
T2 - 390 t2 = 70°F
T-L = 0 tl = ~ 320°F case {*>)
T, = 500 t0 = 180°F
& £*
190
-------
Figure 141. Thermal Conductivity, Linde Super Insulation
191
-------
case (a) represents the temperature difference encountered
when the outside temperature is 7Q°F and case (b) is
representative of an outside temperature of 180°F. The heat
transfer rates for these two cases are given by
(500> (-500)
The percent increase in heat transfer due to an increase in
exterior wall temperature is given by
q... - q
% increase = — - - x 100
This results in the following
% increase = <500 ' 390>
x 100
2
500 + 2 (500)
For 6 = 0.00482 the percent increase in heat transfer is
% increase in heat transfer = 31.5%.
The rate of decrease in liquid mass for the 70°F surface
temperature was
~ * Hr
and the corresponding heat transfer was
q = 3.157
*AJ-
The increase of 31.5% would indicate a heat transfer rate of
q = 4.122
Bty
Hr
192
-------
which would indicate a new rate of decrease of mass of
Ib m
m P' 0.0483
Hr
Comparing this calculation with those previously discussed
indicates that the time elapsed before refill becomes necessary
decreases from 45 days to 40 days. This, of course, is a
decrease of 5 days due to the unfavorable ambient conditions.
As can be seen, the logistics problems can be an important
consideration. If very high sensitivity is not as important
as the logistics problems involved in cooling with liquid
nitrogen on a remote platform, then selecting PbSe as the
detector material might be best. It is only necessary to
cool PbSe to 195°K, which can be accomplished with a
thermoelectric cooler (TEC). A discussion of one particular
TEC for cooling a detector to 178°K is discussed in the
following paragraphs.
The TEC has the following advantages when the heat load is
small and the required detector temperature is between 150°
and 200°K:
No moving parts
Low maintenance
High reliability
No consumable gases
All electric operation
Efficient use of TEC requires consideration of three design
factors. Parasitic heat loads must be minimized, cascading
of cooler stages must be optimized for best thermal and
electrical efficiencies, and a high-conductance heatsink
should be used to dissipate the heat rejected by the base
of the thermoelectric cooler.
These factors have been optimized for the 3- to 5-micron
PbSe detector, resulting in a six-stage cascaded TEC design
using a natural convection heatsink. Nominal cooler operation
will provide a detector temperature of 178°K at an ambient
temperature of 80°F, with an input power of 25 watts.
The performance of a TEC depends on many internal parameters
(including materials, geometry, and number of stages) and
operational parameters (power input, total thermal load,
heatsink temperature, environment, etc). These considerations
have been carefully developed by Click and Marlow.
193
-------
The cooler is derived from the nine-stage cooler previously
developed for the Night Vision Laboratories by Nuclear
systems Incorporated (NSI) of Garland, Texas. The load and
temperature requirements can be obtained by using the
lower six stages of the 145°K cooler previously developed by
NSI.
The cooler and vacuum enclosure are shown in Figure 142. A
low-emissivity shield surrounds the cooler to minimize radiation
loading of the cascaded stages. Constantan conductors are
used to minimize the thermal conduction load from the detector
leads. Performance characterisitcs of this cooler are
shown in Figure 143. The basic cooler specifications are
listed in Table 18.
TABLE 18. BASIC COOLER SPECIFICATIONS
Size
Base 1.0 by 1.0 inch
Height 0.6 inch
Top 0.25 by 0.40 inch
Weight 20 grams
Tambient 300°K
Hot side (Tu) 333°K nominal (on heatsink)
n
Cold side (TC) 178°K nominal
Power 25 watts
Voltage 6 volts
Current 4.2 amperes
Heat load 50 milliwatts
(detector)
The heatsink must be designed to efficiently transfer heat
rejected by the TEC to the air. Tow basic types of heatsinks
have been considered for use with the six-stage TEC, assuming
that the ambient air temperature will vary from 40° to 120°F.
Both natural-convection and forced-convection heatsinks have
been evaluated, and the calculated performance curves are
shown in Figure 144. A weight restriction of 1 pound
maximum was placed on both heatsinks. Each is of the
highest efficiency possible within the weight and volume
constraints of the system.
The forced-convection unit would provide a 10°K lower
detector temperature than the natural-convection unit.
However, it would require additional power for a 10-cubic-
feet-per-minute capacity blower, would weigh 1 pound, would
194
-------
substantially increase the acoustical noise, and would
introduce forced vibrations to the components. The
natural-convection heatsink would be slightly larger, but
would weigh only 12 ounces and be totally passive in
operation.
The heat dissipation capability of the natural-convection
heatsink is limited by fin surface area rather than internal
thermal conduction. Therefore, this type could be fabricated
of aluminum instead of a material with a higher thermal
conductivity without loss in heat dissipating efficiency.
This natural-convection heatsink is proposed for use on the
3- to 5-micron PbSe detectors.
SAPPHIRE WINDOW WITH GOLD FILM
"ON INSIDE TO REDUCE FOV
RADIATION SHIELD
SIX STAGE
THERMOELECTRIC
COOLER
CONSTANTAN
CLAD
POLYFILM
H\\
•1.0 INCH-
TWELVE PIN
TO-5 HEADER
Figure 142. Schematic of Proposed Thermoelectric Cooler
195
-------
100
tn to
1
K
U
0.
0.
z
0. 1
134974
220 240 260 280
TH (K)
300
320
Figure 143. Input Power Versus Heatsink Temperature
for a 0.05-Watt Load
196
-------
200
190
160
170
160
1 3 X 1C X 1 '*CH HEATSINK.1.3' C/WATT.THERMALLOY CO (12 OZ)
2 X 2 X 4 INCHES (HEATSINK AND BLOWER). WAKEFICLD FC 805
2 AND IxlMAX I (1 . 6 LB)
80'F NOMINAL
AMBIENT AIR
4O
60
134975
80
AMBIENT AIR
100
120
Figure 144. Cold Side Temperature Versus Hot Side Temperature
for a 0.05-Watt Load and 25-Watt Input Power
197
-------
If sensitivity is of primary importance, the HgCdTe detector
is recommended. If alleviating logistics problems is of
primary importance, the PbSe detector is then recommended.
Tests conducted using a 3- to 5.5-micron scanner as a part
of this program has verified the principle. Figure 145
shows imagery made at 2200 hours at the oil spill test facility
with a line scanner and recorded on photographic film as is
recommended for the oil spill surveillance system. The
imagery shows the 100-square-foot boom with 2 ounces of
crude oil. The oil appears colder (darker) than the
surrounding water. Note the small amount of oil that has
spilled over the upper right hand corner. The imagery is
very noisy, indicating the high gain required for the low
signal level. The collecting aperture of the scanner used
was nearly an order of magnitude less than that of the RS-18;
thus, increased performance can be attained by the oil spill
system. Detailed specifications for both system
configurations appear in the following subsection.
Figure 145. Oil Spill Imagery, 3- to 5.5-Microns
198
-------
The nature of semiconductor detectors is such that a
tradeoff exists between spatial and thermal resolution
capability. It is suggested that the spatial resolution for
the oil spill surevillance system be 3 to 5 milliradians.
This will yield maximum thermal sensitivity. A 3-milliradian
system will resolve 3-foot targets at a range of 1000 feet
which should be sufficient for most applications.
Technical Specifications
The RS-18 is a reflective optics system utilizing a
single-sided rotating wedge mirror for scanning and
Cassegrainian collecting optics. The collecting optics,
consisting of a 3.5-inch-diameter parabolic primary mirror
and a 1.4-inch diameter flat secondary mirror, are mounted
in an optical tube making the converging optics modular in
nature. The primary is three-point mounted at one end of the
tube, and the secondary is mounted on a three-finger aluminum
spiker attached at the opposite end. Radiation baffling is
incorporated into the tube assembly. One baffle is provided
around the secondary mirror, and the other is located on
the detector assembly surrounding that portion of the
detector assembly passing through the primary. This baffling
is necessary to prevent extraneous radiation from reaching
the detectors without first going through the system scanning
optics. The collecting optics mounting arrangement has been
proven on existing systems to provide a very rugged and
structurally rigid optical assembly. Once elements of the
converging optics assembly are mounted, the assembly is
slipped into the scanner housing. The optical tube provides
support and rigidity to the scanner housing itself. The
system effective aperture is 7.7 square inches. This gives
a system f/number of 1.93.
As discussed in the previous paragraphs, two different
detectors are suggested for operation in different applications
The primary detector choice is HgCdTe cooled to 77°K using
liquid nitrogen. The secondary choice, for applications
where the requirement for liquid nitrogen is prohibitive,
is a PbSe detector thermoelectrically cooled to 195°K.
Each interchangeable detector assembly consists of a Dewar
(thermoelectric cooler for PbSe), detector, and preamplifier.
The preamplifier is a linear, ac-coupled, high-gain, low-
noise amplifier matched to the detector impedance to provide
a detection-noise limited system. The video electronics does
not have any function that can irreversibly alter the
quantitativeness of the data. The raw video signal from
the preamplifier is ac coupled into the video post-amplifier
where the portion of the video corresponding to a heated
calibration source is clamped to a variable dc voltage.
199
-------
This voltage may be set manually or automatically to vary
with changes in the temperature of the surrounding water
surface using an in situ temperature sensing element in
the water. The purpose of the automatic mode is to provide
a constant output level corresponding to water irregardless
of diurnal or seasonal water temperature variations.
The calibration source is provided to permit qualitative
temperature data to be obtained. A calibrated temperature
analog voltage will be provided for monitoring with a
readout range of -25°C to +30°C and an accurace of + 0.4°C.
The readout temperature is obtained from two linear
platinum-wire temperature sensors embedded at the base of
the blackbody to sense the contact temperature of the source.
The sensors are used in a low-power bridge circuit to
accurately measure the source temperature. A low-drift dc
amplifier is then used to provide a dc temperature analog
voltage of approximately 0.1 volts/°C.
The active area of the calibration source measures 4 by 5
inches and is located 2.5 inches from the scan mirror axis.
With the above size and spacing, the IFOV of each source
will dwell 60 resolution elements before vignetting occurs.
This dwell corresponds to 9 degrees of scan per revolution.
The heated blackbody calibration source is mounted above
and to one side of the scan mirror. Temperature sensing and
heating elements are mounted in the base of the blackbody
source. Effective emissivity of the source is better than
0.99. Experiments performed have resulted in measured
emissivities of greater than 0.993.
The effect of this video system is to achieve a de-coupled
system without its disadvantages/ such as dc level shift with
the changing temperature of the electronics and dc level
shift due to changing background radiation seen by the
detector because of Dewar window temperature changes.
A common parameter of IR systems, used to measure their
temperature sensitvity, is noise equivalent temperature (NET),
NET is defined as that temperature difference which produces
a unity signal-to-noise ratio out of the system. The NET
of the oil spill surveillance scanner may be mathematically
determined from the following equation:
Tr(f.l) (Af )1/2
NET =
4aeT3A (A6A)1//2 D*x T r
O a. O
200
-------
where
f.l = optical focal length
Afn = noise bandwidth = Tr/2 B.W.
a = Stefan-Boltzmann constant
e = target emissivity
T = temperature of target in °K
A = area of optical aperture
A0 = across-track instantaneous field of view
A = along-track instantaneous field of view
D* = detectivity of the detector normalized for
unity bandwidth and detector area
(cm-Hz 1//2 -watt'1)
T = transmission of the atmosphere
Cl
T = transmission of the optical system
r = system response to input (y is unity for
large target NET calculations or the MTF
of the system for resolution size targets)
Under worst case conditions, the NET has been calculated
for each detector type. The system configuration was assumed
to be 3-milliradian resolution (A0-A) with a scan speed of
100 scans per second for HgCdTe and 5 scans per second for
PbSe. The 5 scans per second rate is maximum for PbSe in
a 3-milliradian configuration due to the relatively low
frequency response of the detector. The system NET is
specified to be:
NETHgCdTe 1 °-l0c
NETPbSe 1 °-25°C
In practice, the scanner NET is determined by the following
procedure. Two known temperature targets are scanned by^_
the receiver and their temperature difference (AT=T2~T,) Jis
noted. The peak-to-peak signal output (AS) produced by
these two targets is measured. At the same point in the
electronics chain, the system noise (n) is determined by
use of a true rms voltmeter. A signal-to-noise ratio for the
two targets is then calculated (AS/N). This number is then
divided into the temperature differential of the two targets
to calculate NET or
201
-------
SECTION XI
TOTAL SYSTEM APPROACH
The sensors tested in this study program can provide oil spill
surveillance for estuaries, rivers, drainage channels, bridges,
docks, offshore drilling platforms, and onshore or offshore
separator facilities. For each application there is a particu-
lar mix of passive and active sensors that provides the best
cost-effective oil spill surveillance.
For short range staring applications the MOD2 active sensor,
described in Section IX, provides adequate low-cost
surveillance. Conversely, the active sensor does not have
the range to cover a large bridge area; therefore, the passive
sensor in a line scan mode should be used.
One of the more challenging surveillance applications is the
oil separator. Here the need for an unmanned alarm system is
important due to the large number of both offshore and onshore
installations which have continuous water output. Figure 146
shows a schematic representation of a separator. Many offshore
wells feed a single offshore separator platform. The oil and
water from the wells is separated and the oil is then trans-
mitted by pipe to the onshore station. The water emitted
from this offshore separator can contain oil droplets that
immediately rise to the surface and form a detectable sheen.
The existence of this sheen can be monitored by the active
sensor.
The final separation of oil and water is performed at the
onshore facility as indicated in Figure 146. In this case
the water output from the separator station enters the public
supply by means of a channel as shown in Figure 147. An
active or passive sensor, depending on the size of the channel
and the point in the channel where the sensor is located,
should be located downstream from the oil skimmers. This
would provide detection of any oil not detained by the skimmers,
Examples of sensor versus application are given in Table 19.
In addition to the basic sensor, the scan and display types
are treated as options. Consider the case of the offshore
separator. Here an alarm only would be needed since the
application is one of spill prevention as opposed to
surveillance and possible legal documentation. The latter
would be needed in a river area where total areal extent
would be required.
203
-------
ONSHORE TERMINAL
GAS
STOCK
TANK
OIL
>.TO PIPELINE
WATER TO DISPOSAL
GAS
PIPE
LINE
SKIMMER
SHORE LINE
OIL PIPE LINE
PRODUCTION PLATFORM
GAS
OIL
~f SEPARATOR
OIL
WATER TO
DISPOSAL
AND GAS WELLS
SKIMMER
PRODUCTION PLATFORM
Figure 146. Schematic Representation of an Oil Separator
Facility
204
-------
REFINERY
to
o
PRIMARY
SEPARATOR
MASTER
SEPARATOR
Figure 147. Onshore Oil Separator facility
-------
TABLE 19. OIL SPILL SENSOR APPLICATION MATRIX
>s. MODE
APPLICATIONS.
1. ESTUARY
2. INDUSTRIAL
EFFLUENT
3. RIVER/BAY
AREA
(TOWER)
4. DRAINAGE
CHANNEL
< 60 'ACROSS
5. DRAINAGE
CHANNEL
<150' ACROSS
6. DRAINAGE
CHANNEL
> 150 'ACROSS
7. OFFSHORE
DRILLING
PLATFORM
8. OFFSHORE
SEPARATOR
PLATFORM
9. ONSHORE
SEPARATOR
PLATFORM
10". BRIDGE
11. DOCK AREA
ACTIVE SENSOR
MODE 2
STARING
W/ALARM
X
X
X
ARC
SCAN
W/ALARM
X
X
X
X
X
ARC
SCAN
PLOT
X
MODE 3
STARING
W/ALARM
X
ARC
SCAN
W/ALARM
X
X
X
X
ARC
SCAN
PLOT
X
PASSIVE SENSOR
8-14ym
LINE
SCAN
W/ALARM
PROC
X
X
X
X
AREA
SCAN
CRT1
FILM
X
X
X
3-5ym
LINE
SCAN
W/ALARM
PROC
X
X
X
X
X
AREA
SCAN
CRT1
FILM
X
X
to
o
CTi
-------
The sensor configuration, depending on the particular applica-
tion, can be expensive. However, when mass produced, most
sensor configurations would be low cost.
207
-------
SECTION XII
ACKNOWLEDGEMENTS
Grateful acknowledgement is given to Pat Daly and Bill Hawley
of Night Vision Laboratory for granting Texas Instruments
permission to use the MIRS system and release the imagery for
publication.
Hugh Yantis of Texas Water Quality Board, Major General H.R.
Parfitt of the Southwest Division Corps of Engineers, and
Bill Aston of Dallas Power and Light have our thanks for
their assistance in trying to find a suitable test site.
The assistance of Professor M.A.Collin, Professor R.L.
Simpson and Professor W.G.Wyatt, of Southern Methodist
university in the search of a local test site is acknowledged
with thanks.
The advice of Dr. Tim Matzke and Dr. Oppenheimer of the
Environmental Protection Agency on legal limits of oil
spilling for research purposes is greatly appreciated.
Ray Lozano, Dick Peckham and Jerry Thornhill of the EPA
were very helpful in advising us on particular applications
of oil spill sensors.
Jerome Overfield of Pollution Abatement Research is acknowledged
for his consultation.
The contribution of various samples of oil from Humble, Sun,
and Richfield Oil Companies is greatly appreciated.
The support of John Riley of the EPA is acknowledged with
much thanks.
209
-------
SECTION XIII
REFERENCE
1. Albert V. Jelalian, "Sea Echo at Laser Wavelengths",
Pfocee'dings of the IEEE, Vol. 56, No. 5, May 1968.
2. JOSA, "The Refractive Index of Liquid Water in the Near
Infrared Spectrum", Me1chor Centeno, Vol. 31, p 244,
1941.
3. Born and Wolf, Principles of Optics , Pegamon Press,
3rd Edition, pp 90-98.
4. H. W. Yages, The Absorbtion Spectrum From 0.5 to 25
Micron of a 1OOP-Foot Atmospheric Path at Sea Level,
Naval Research Laboratory, September 1957.
5. J. L. Emmett, A. L. Showlow and E. H. Weinburg, "Direct
Measurement of Xenon Flashtube Opacity," Journal of
Applied Physics, Vol. 35, No. 9, September 1964, pp
2601-2604.
6. John H. Gonca and P. Bruce Newell, "Spectra of Pulsed
and Continuous Xenon Discharges", Journal of Optical
Society of America, Vol. 56, No. 1, January 1966, pp
87-92.
7. V. W. Forsythe, "Measurement of Radiant Energy,
McGraw-Hill, N.Y. 1937.
8. Journal of Geophysical Research, March 15, 1965, Vol.
70, No. 6, pp
jphysi
T33TT
211
-------
SECTION XIV
GLOSSARY
Active IR System - the infrared energy used to provide
target illumination is generated internal to the system.
Angular Scan Capability - range of angles from azimuth
for effective signal reception.
Anomalous Dispersion Curve - graph of change in reflectivity
as a function of wavelength in the region of a resonant
absorption.
Areal Coverage - coverage of an area.
Azimuth - vertically downward.
Bistatic Reflection - reflectance measurements where the
optical paths of the transmitter and receiver are not
congruent.
Blackbody - a solid that radiates or absorbs energy with
no internal reflection of the energy at any wavelength.
Chopping - interrupting a current or a light beam at
regular intervals for purposes of modulation.
C-H Resonance - vibrational frequency of carbon-hydrogen
element in a molecule.
Detector - a device that produces an electrical output
that is a measure of the radiation incident on the device.
Diffuse - rough surface.
Estuary - an inlet or arm of the sea, the wide mouth of a
river, where the tide meets the current.
(FOV) Field Of View - the solid angle from which objects
can be acceptably viewed, photographed, or otherwise detected.
f/Number - a ratio of focal length to aperture opening.
Heterodyne Detection - combining the received signal with
a locally generated wave to produce sum and difference frequencies
(demodulation)
imagery - view of objects of interest stored on film, tape,
or other types of storage material.
Infrared (IR) - a form of electromagnetic radiation, at a
frequency lower than visible light, and generated by thermal
agitation.
213
-------
Lambertiari Surface - totally diffuse
Microphoriic Noises - noises due to magnetic shock or
vibration.
MOD 1 - active system used in test section with reflective
optics and a 400 Hz modulated tungsten beam.
MOD II - active system used in test section with
refractive optics and 400 Hz modulated tungsten beam.
MOD III - active system used in test section with a
pulsed Xenon arc source rather than a 400 Hz modulated tungsten
beam.
0-H Resonant Mode - the mode of vibration of the oxygen-
hydrogen element of a molecule.
Passive IR System - the infrared energy used to produce
target illumination is generated external to the system.
Real Time Presentation - presentation generated fast enough
to be viewed at the same time the system is pointing toward
subject.
Reflectance - light or heat bouncing back toward the source
Resonance Absorbtion - an absorbtion which occurs due to
a fundamental vibrational mode in the material.
Sensor - image detecting system
Slick - A layer of oil floating on water.
Signal-To-Noise (S/N) - ratio of magnitude of the signal
to that of the noise
Skin Depth - the depth at which the radiation is 1/e times
the radiation of the surface.
Spectral - pertaining to the frequency or wavelength
aspect of a signal.
Specular - like a mirror surface
Staring System - system which has a single resolution
element which is not scanned in any direction.
Subsurface Effects - effects of physical variations beneath
the surface of the water on signal to be detected on the
surface.
214
-------
Theritioelectri;c Copier (TEC) - a device used to cool
infrared detectors. The; mechanism for cooling is electric
current passing through, a series of diodes in such a way
as to heat one side of the device and cool the other.
Peltier cooler.
Wide Band TR Pulse - an electromagnetic pulse in the
infrared region of the spectrum which consists of a wide
range of frequencies.
215
-------
SELECTED WA TER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
2. 3. Accession No.
w
4. Title 5. Report Date
OIL SPILL SURVEILLANCE SYSTEM STUDY «•
*. Performing Organization
7. Author(s)
Don Mohr, Kent McCormack, Gary Brewster, G. Fournier0-
15080HBP
Contract/Grant No.
9. Organization
Texas Instruments Incorporated
Post Office Box 6015 68-01-0150
Dallas, Texas 75222 13. Type of Report and
> . s^jj j™— -, -rr
-------