a‘A..

f i an” .Jv.» .

. .(9‘
.

‘ V is.“
3a...»

:NQ‘ ,
Avwnafif .

‘

In ‘
.. ywwvdk»

.

7..

. ls”) 2a. .
. +3

.

 

’l lv -' ‘
2.5L (fir...

4:.
|

 

 

 

 

 

 

 

 

 

 

4. $th A
at»; .
‘ 4!
u$ ’l

.5
.1

,. 5.9 1

 

931D.

. u. < 2.
.§ . ,. 1. ,
v . .nflwtfimmmwcn
. . on“; nixolxih .o.
$~Iuka . ul :1
3! £45.. . . xf
.
. 17 V «"5. 1.3.?

 

 

THESlS J memo»: STA‘i' u RABIES

2) lllllllllllllllllllllllll‘lllllllllllll

3 1293 01771 0892

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Experimental Measurement of the
Slit Response Function and
Corrected Infrared
Thermographic Measurements

presented by

Paul Bryan Hoke

has been accepted towards fulfillment
of the requirements for

MS Mechanical Engineering

 

 

degree in

j)? QC M4 6M

Date August 21, 1998

 

0.7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN REI'URN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE bus

DATE DUE

 

MAY 0 5m?

 

 

00kg; 2003

 

NPR—144M-

 

 

033039129"

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1]” COMM

 

EXPERIMENTAL MEASUREMENT OF THE SLIT RESPONSE FUNCTION AND
CORRECTED INFRARED THERMOGRAPHIC MEASUREMENTS

By

Paul Bryan Hoke

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

1998

ABSTRACT

EXPERIMENTAL MEASUREMENT OF THE SLIT RESPONSE FUNCTION AND
CORRECTED INFRARED THERMOGRAPHIC MEASUREMENTS

By

Paul Bryan Hoke

Infrared thermography utilizes emitted radiation to measure temperatures
optically. An experimental facility and procedures were defined to measure emissivity
and temperature of unresolved (target angle < 0.05 rad) targets. The facility consists of a
flow system with instrumentation for measuring flow velocity, temperature, and
humidity. The emissivity is determined with the infrared camera and the target
temperature set by the known flow conditions. Once the specimen emissivity and size
are known, its temperature can be measured in any environment.

An Inframetn'cs 6OOL infrared camera and the associated data acquisition system
were characterized to maximize thermal measurement sensitivity. The minimum
specimen size that can be resolved thermally with the infrared camera was determined.
The effect of specimen size on thermal signature was quantified for the infrared detector
utilizing the slit response function (SRF). A standard procedure to measure emissivity
was validated and a SRF measurement correction was verified. These two pieces of
information increased the measurement capabilities of the infrared system and minimized
experimental error when working with thermal targets smaller than the thermal

measurement resolution of the system.

ii

To Carol,
for always believing

and of course, Madeleine

iii

ACKNOWLEDGMENTS

I would like to thank my advisor Dr. John McGrath for having the patience to let
me try things on my own and guiding me when I needed it. His advice, insight and
assistance to this work have been invaluable.

I would also like to thank Dr. Craig Somerton and Dr. Andre Benard for joining
my committee and editing this work. Dr. Somerton has been an aid through out my stay
at Michigan State University as an advisor, reference and an open door. Dr. Benard’s aid
and insight are greatly appreciated.

Dr. Foss for his advice, guidance, and equipment. Dr. Foss was a wealth of
support and knowledge, especially on the finer points of fluid mechanics.

Mike McClean for his mechanical assistance and precision work.

The aide I have received from my colleagues has been immeasurable. Heidi Relyea,
Jon Darrow, Al Aksan, Ramez AbdulNour, Mark Minor, Gloria Elliot, Bin Lian, Ingo
John, Doug Neal and Scott Morris have all helped me out along the way. It ranged fi'om
moving equipment, helping solve nasty mathematical problems, studying for class, or just

listening, but it all helped me through and for that I say thanks!

TABLE OF CONTENTS

LIST OF TABLES. .................................................................. vii
LIST OF FIGURES .................................................................. viii
NOMENCLATURE .................................................................. xi
CHAPTER 1. INTRODUCTION .................................................... 1
CHAPTER 2. WIND TUNNEL TEST FACILITY .................................. 4
2.1 DESIGN OBJECTIVES ................................................. 4
2.2 WIND TUNNEL BASE ................................................ 6
2.3 TEST MODULE ........................................................ 7
2.4 THROTTLE ASSEMBLY ............................................... 9
2.5 FLOW CONDITIONING ................................................ 10
2.6 DATA ACQUISITION .................................................. 12
2.7 PRESSURE MEASUREMENT .......................................... 13
2.8 HOT WIRE SENSORS .................................................. 13
2.9 CALIBRATION NOZZLE AIRFDOW HEATER ........................ 14
2.10 CHARACTERIZATION OF THE WIND TUNNEL FACILITY ......... 15
2.1 1 SUMMERY ........................................................... 19

CHAPTER 3. CALIBRATION 0F INFRAMETRICS IR CAMERA AND

ASSOCIATED DATA ACQUISITION SYSTEM ...................................... 20
3.1 INTRODUCTION ......................................................... 20
3.2 STATEMENT OF PROBLEM ............................................. 20
3.3 EXPERIMENTAL PROCEDURE ......................................... 22
3.4 RESULTS ................................................................ 27
3.5 OPTIMAL CONFIGURATION TESTING ................................ 31
3.6 RESULTS OF OPTIMIZED EQUIPMENT CONFIGURATION ........... 31
CHAPTER 4. SLIT RESPONSE FUNCTION ....................................... 34
4.1 INTRODUCTION ........................................................ 34
4.2 IMAGE PROCESSING EFFECTS ON DATA ........................... 38
4.3 EXPERIMENTAL TEST PROCEDURE ................................. 39
4.4 RESULTS ............................................................... 41
4.5 PREDICTTVE CAPABILITIES ........................................... 44
4.6 CONCLUSIONS ......................................................... 47

CHAPTER 5. EXPERIMENTAL VERIFICATION 0F THERMOGRAPHIC
TEMPERATURE MEASUREMENT UTILIZING

SLIT RESPONSE CORRECTION ................................................... 49
5.1 INTRODUCTION ....................................................... 49
5.2 EXPERIMENTAL PROCEDURE ....................................... 49
5.3 EMISSIVITY MEASUREMENT ........................................ 51
5.4 EXPERIMENTAL VERIFICATION ...................................... 53
5.5 DISCUSSION ............................................................ 63
CHAPTER 6. CONCLUSIONS ...................................................... 69
CHAPTER 7. RECOMMENDATIONS FOR FUTURE WORK .................... 73
7.1 SLIT RESPONSE FUNCTION .......................................... 73
7.2 INFRARED CAMERA RESPONSE TO TEMPERATURE GRADIENTS. 73
7.3 HOT-WIRE SENSORS .................................................. 74
7.4 DATA ACQUISITION SYSTEM ........................................ 75
APPENDICES ........................................................................ 79
APPENDIX 1. Two DIMENSIONAL FLow CONTRAcnON .................. 80

APPENDIX 2. QUICKBASIC‘D PROGRAM To PERFORM DATA

ACQUISITION AND STEPPER MOTOR CONTROL ............................. 84
APPENDIX 3. FLAT PLATE ISOTHERMAL VELOCITY PROFILES ........... 108
APPENDIX 4. HOT-WIRE CONSTRUCTION AND CALIBRATION ............. 111

APPENDIX 5. MAXIMIZING THE DYNAMIC RESPONSE OF THE
INFRAMETRIX IR CAMERA AND DATA RECORDING SYSTEM ............. 121

APPENDIX 6. VELOCITY AND TEMPERATURE CALCULATION
PROGRAM FOR Two HOT-WIRE METHOD .................................... 124

BIBLIOGRAPHY ................................................................... 129

LIST OF TABLES

Table l. SRF summary for the Infiametrics 6OOL infrared camera
in the HTRL ............................................................ 48

Table 2. Emissivity measurement results (e=0.88) ........................... 54

Table 3. Test results from test 1a (MAF sensor)
(11 amps to flow heater, 42.1-62.10C IR temperature range) 56

Table 4. Test results fi'om test lb (MAF sensor)
(4 amps to flow heater, 20.0-40.00C [R temperature range) 57

Table 5. Emissivity measurement results (e=0.95) ........................... 59

Table 6. Test results from test 2a
(11 amps to flow heater, 32.9~52.9°C IR temperature range) 61

Table 7. Test results from test 2b
(6 amps to flow heater, 23.3-43.30C IR temperature range) 62

Table 8. Null hypothesis statistical test results .............................. 68
Table 9. Two dimensional contraction machining coordinates ............ 81

Table 10. Post calibration data ................................................ 119

vii

LIST OF FIGURES

Figure 1. Wind tunnel test facility concept drawing ............................. 4
Figure 2. Wind tunnel test facility ................................................. 5
Figure 3. Modular concept utilized to guide flow facility design .............. 6
Figure 4. Structural drawing of sub-atmospheric test section ................. 8
Figure 5. Sluice throttle mechanism used to control

airflow rates (top View) ................................................. 10
Figure 6. Airflow conditioner ..................................................... 11
Figure 7. Calibration nozzle airflow joule heater ............................... 15
Figure 8. Axi-symmetric calibration nozzle and associated

coordinate system ...................................................... 17
Figure 9. Calibration nozzle exit velocity profile .............................. 17
Figure 10. Calibration nozzle exit temperature profile ....................... 18
Figure 1]. Ideal gray scale-temperature response for the Infiametrics

infiared camera and associated data acquisition system ........... 21
Figure 12. IR camera tool for area temperature measurement ............... 22
Figure 13. Schematic of different equipment configurations tested .......... 23

Figure 14. Playback data image sequence from the equipment test
experiment conducted on 2/11/97 ..................................... 25

Figure 15. Equipment configuration effects on the IR camera
gray scale-temperature response ...................................... 28

Figure 16. Comparison of equipment tests with and without the
For-A VTG-33 video timer ........................................... 29

Figure 17. IR camera gray scale output as a function of contrast setting.
(Low end of Infiametrics camera temperature scale) .............. 3O

viii

Figure 18. Gray scale-temperature relationship with
optimal contrast settings ............................................... 32

Figure 19. Optimal equipment configuration gray scale-temperature

relationship ............................................................... 33
Figure 20. Schematic of IR camera optical configuration and

definition of terms ....................................................... 35
Figure 21. Comparison of scan direction effect on IR camera response 36

Figure 22. Temperature profile of narrow dimension of the MAF sensor.
Sensor parallel to scan direction ...................................... 37

Figure 23. Temperature profile of narrow dimension of the MAF sensor.
Sensor perpendicular to scan direction .............................. 37

Figure 24. Schematic of the experimental configuration
for SRF measurement .................................................. 39

Figure 25. Slit response function with 95% confidence region (n=5)
for the Infi'ametrics 6OOL IR camera ................................ 41

Figure 26. SRF confidence region/mean value for Inframetrics 600L
IR camera with external optics ...................................... 42

Figure 27. SRF of the Inframetrics 6OOL infrared camera with
3x Telescopic and 6” close lenses ................................... 43

Figure 28. Thermographic image of MAF sensor ............................ 44

Figure 29. Comparison of SRF between system configurations
(no optics compared to external optics) .............................. 47

Figure 30. Emissivity variation observed in semi-conducting materials
by Inagaki et.al., 1994 ................................................... 65

Figure 31. Correction factor (l/SRF) and % uncertainty in CF for
Infi'ametrics 600L IR camera with close up optics in place as

a function of target Size ................................................... 71
Figure 32. Cumulative average of the hot-wire data ........................... 77
Figure 33. Two-dimensional flow contraction .................................. 8O

ix

Figure 34. Compiled jet velocity profiles (w=2cm) ........................... 109

Figure 35. X/W=l .59 velocity profiles (w=2cm) .............................. 109
Figure 36. X/W=4.52 velocity profiles (w=2cm) .............................. 110
Figure 37. X/W==10.95 velocity profiles (w=2cm) ............................ 110
Figure 38. Schematic of hot-wire probe ....................................... 111
Figure 39. Tungsten wire copper plating station ............................. 112
Figure 40. Precision solder station ............................................. 113
Figure 41. TSFL hot-wire calibration flow contraction ..................... 114

Figure 42. “Coke Bottle” contraction utilized for hot-wire calibration 115

Figure 43. TSFL contraction hot-wire calibration curve .................... 1 17
Figure 44. HTRL facility hot-wire calibration curve ......................... 118
Figure 45. Optimal equipment configurations for

data acquisition and processing ...................................... 122
Figure 46. Results of optimized equipment configuration ................... 123

ENGLISH:

A
A,B,n
ATF

Cd

R0.)
Re

SRF

NOMENCLATURE

Area

Hot-wire cah'bration coefficients fi'om Collis and Williams [1959]
Aperiodic transfer fimction
Discharge coefficient

Diameter

Voltage

Blackbody emissive power (W/mz)
Convective heat transfer coeflicient
Infiared

Radiosity

Confidence region

Mass air flow

Pressure

Prandtl number

Thermal energy

Radiated flux

Wavelength dependent radiosity
Reynolds Number (=Ux/v)

Slit response function

Temperature

T Non-dimensionalized temperature [(T - Trange Milo/Tunge]

Tm...“ True temperature value

Tm Infi'ared measured temperature

Tm; MAF sensor temperature

T"Inge Temperature span on infrared camera

Tum“... Lower limit of temperature span on infrared camera
Tm: Infi'ared measured, SRF corrected temperature

U Streamwise velocity

V Velocity

Vm Maximum velocity

W Wall jet Slot width

f Focal distance

g Acceleration due to gravity

k Thermal conductivity

Id Working distance

1. Slit width

1.- Internal focal length

n Number of data points

t Statistical parameter from student T distribution
x Streamwise direction

y Direction orthogonal to streamlines and test plate
2 Direction aligned with gravity or direction orthogonal to xy plane - context

willclarify

xii

GREEK:

(I) Measurement error

O Target angle

.0 Electrical resistance

a Absorbtivity

e Emissivity

A Wavelength

p Density or Reflectance - Context will clarify

0' Standard deviation or Stephan-Boltzmann constant - context will clarify
r Transmissivity

v Kinematic viscosity

xiii

CHAPTER 1. INTRODUCTION

1.1 MOTIVATION

The original motivation for this study was provided by a project for General
Motors Corporation. The manufacturing process of mass air flow (MAF) sensors for an
engine intake manifold requires that each sensor be individually calibrated. The cah’bration
method utilized was expensive and time consuming. The method consisted of mounting
the MAP sensor in a mold and filling the mold with an isothermal, electrically neutral and
environmentally friendly liquid. This created an environment of known temperature in
which the sensor resistance could be measured. The method provided a two point
temperature-resistance calibration curve. The material cost for the fluid utilized was
approximately 1.5 million dollars per year not incorporating the energy costs of heating it.
Infrared thermography and an alternate heating method were proposed to obtain the
necessary temperature/resistance information in order to replace the current cahbmtion
method.

The project was complicated by the size and material characteristics of the MAF
sensor. The sensor is a 1500 um long and 500 pm diameter cylinder constructed from
platinum wire surrounded by a glass body. The glass surface could not be coated with any
materials and was expected to experience a large amount of manufacturing variability.
Thus, the problem of measuring the surface temperature the small sensor Size was
complicated by unknown and variable emissivity properties of the specimen.

A wind tunnel facility was constructed to support and augment the Heat Transfer
Research Laboratory (HTRL) and Specifically to investigate the capabilities of the infi'ared

measurement equipment. This facility was designed to provide a flow of known velocity

and temperature and the sensors necessary to monitor this flow. Chapter 2 details the wind
tunnel facility and its construction, the data acquisition system, and the measurement
transducers utilized. This facility development was essential as a test platform for the
current project.

The laboratory utilizes an Inframetrics 600L infiared camera and the associated
data acquisition system to obtain Spatial temperature information. This system was
characterized and optimized to obtain the full dynamic temperature measurement range.
The effects of specimen Size on the measurement capabilities were determined
experimentally utilizing the Slit response function (SRF) of the infrared sensor. The SRF is
the attenuated response of the sensor to targets smaller than its measurement resolution.

The combination of the flow facility and the characterization of the infrared system

allow the goals of the original proposal to General Motors to be realized.

1.2 PREVIOUS WORK

The capabilities and applications of infi'ared thermographic equipment have been
explored by a previous investigators. The sensors response to an unresolved target
(subtends an angle less than the measurement capabilities of the sensor) has been explored
extensively by Holst [1993]. The target angle to produce a 50% response (the 50% Slit
response fimction (SRF)) is generally provided by the manufacturer of the equipment as a
measure of the performance (Inframetrics [1988]). These quantities are provided to define
the limitations of the equipment so as to avoid inaccurate measurements. It is the goal of

this study to quantify the Slit response function to allow meaningful measurements to be

taken in these circumstances and thus to exploit the maximum capabilities of the infrared
camera systems.

Infiared thermography has also been applied in wind tunnel studies of fluid
mechanics and heat transfer. One application has been the study of supersonic heating of
structures (Lafferty and Collier [1991]). Infrared thermography has also been applied to
study the surface temperature of structures in flows to measure the convective heat
transfer coefficient (Willenborg, et al. [1996], Okamoto and Inagaki [1994]).

Gartenburg and Roberts [1990] studied the temperature gradients on a heated wire
in a flow field with an infrared imaging system. The study involved imaging an unresolved
target (the wire) and compensating for the error in the temperature by using an ‘apparent’
emissivity. This method compensated for the targets emittance and Size but the study was
fixed in one configuration because a change in the working distance caused a change in the
apparent emittance.

This study involved calibrating an infrared camera and the associated data
acquisition system. An ideal linear system response to temperature was defined to
maximize the dynamic range and the equipment was configured to obtain this result. The
infrared camera response to unresolved targets was determined experimentally. A wind
tunnel facility with well characterized flow conditions was developed to support the
experimental work and a series of experiments to validate and explore the infrared data

acquisition system were conducted.

CHAPTER 2. WIND TUNNEL TEST FACILITY

2.1 DESIGN OBJECTIVES

A wind tunnel testing facility was built to advance the laboratory testing
capabilities and to generate constant temperature and velocity flow conditions. The
facility was designed to allow cah'bration of various sensors, to achieve specific testing
goals as well as to allow for future expansion. The entire system was designed to be

modular so that any component could be replaced with minimal efl‘ort (Figure 1).

System Inlet
Steam Supply

Armstrong Model 90
Steam Humidifier

Condensate Return
Duct

Flow Conditioner
(Honeycomb and Screens)

2-Dimensional Contraction
Splitter Plate
Thermally Active Test Section
. Sub Atmospheric Test Chamber
. Probe Traverse System
. Calibration Nozzle
. Wind Tunnel Base
. Fluid Collection
. Prime Mover
. Exit Throttle Valve

1 .. i /,2

I
.h
99°.“ 9‘1"? P’P’?‘

 
 
 
 
 
 
   

l

\O
au-II—lu—I
NI—‘o

12

—a
w

11

u—
an

10

u—I—D
05M

/16

13 .

Figure 1. Wind tunnel test facility concept drawing.

Several physical quantities are controlled or measured in this facility so that the
target emissivity and temperature can be obtained with infi'ared thermographic
techniques. These include: pressure (and thus mean flow velocities), velocity and
temperature of the air flow. Humidity of the air flow was measured but not controlled
in this study. Specific controls or measurement sensors are described and their

application to the infrared testing procedure are defined in this chapter.

 

Figure 2. Wind tunnel test facility.

2.2 WIND TUNNEL BASE

The base of the testing apparatus was designed to fulfill several functions. The
base design followed the modularity concept for the entire system. This concept
allowed different flow configurations with respect to the test piece and the
gravitational field to maximize testing flexibility. The design allows for the air flow to
be oriented in the same or opposite direction as the gravitational field (Figure 3). Air
flow orthogonal to gravitational field can be obtained by turning the system on its side.
The base structure houses the prime mover for the system. A condenser and drain
section may also be added to remove condensed water vapor for future experiments.
The test section was supported physically by the wind tunnel base section. The entire

structure was sealed to maintain a vacuum during operation.

 

 

 

 

Figure 3. Modular concept utilized to guide flow facility design.

The base was constructed with 2”x4” boards for the support structure and
covered with sheets of 15/32” plywood. These materials were chosen to reduce costs
and allow for easy and inexpensive modifications. The assembly was completed with
3/8” hex head bolts and wood screws to allow for easy disassembly if necessary.

The joints were sealed with GE silicon sealant. One end of the base structure
was designated as a door and made removable to allow access to the interior. The
access regions were taped with closed cell foam tape to seal the unit while it was
operational. The interior wood surfaces were treated with a wood preservative to

protect against water damage resulting from high humidity experiments.

2.3 TEST MODULE

One of the primary design requirements for the test section was to allow
variable orientation of the test section with respect to gravity. The test section was
thus designed as a cube so that the sides could be interchanged to accommodate
difi‘erent testing configurations. The simplicity of this design (Figure 4) was its greatest
strength. Two sides of the module have 0.375” thick plate glass to allow optical
access to the interior environment. One side was designed with a removable door and
the remaining side was solid. The top and bottom of the module have openings
appropriate to the air passage through the respective sides.

The flame of the test section was constructed of 2”x2” angle iron with a
15/32” skin of plywood. The plywood skin was bolted onto the iron fi'ame for easy
modifications and reconstruction. The interior of the testing module was treated with

shellac to prevent degradation of the wood during experiments with high humidity.

Four extra 2”X2” angle iron support members were included in the original design.
These members (Figure 4) can be utilized for additional support if necessary or to

mount equipment in the testing environment.

To p \71 0 w

 

' T i ' ’ j/ ("eiillbr'alion

if M Nozzle

 

i3 IQ)III'Ir..-IISI()ITIII
CR)HIITICII(HI
IITICI'IarC

f

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Front View Side View
L3 (5 X 2'3 10 ‘ ‘ , 'r‘ ““““““““““““““““ j ‘
Class [“h“ i 53 X 23 l i
.. ‘ » - I I
\l“111<low NWIWH ; Plywood ; i
(I of :3) “Hubs” : Door 3 .
I . «l
I I
a ‘ i l
1‘:
a N; 4
.__ 45.. ..‘ ,1. I
I I i
____L ______,_ L ‘ _ I: ___________________ Id , I_g
' l
i“ - i 1 i «’1'

2' Annie Iron
I‘I‘mIIIJ

Figure 4. Structural drawing of sub—atmospheric test section.

2.4 THROTTLE ASSEMBLY

Fine control of air flow rate through the system was required to allow hot wire
calibration and to insure experimental reproducibility. A “sluice” style throttle valve at
the exit of the facility was chosen for its simplicity and its proven effectiveness in the
Turbulent Shear Flows Laboratory (TSFL) at Michigan State University.

To form the sluice, a plywood gate was mounted between two 2”x 4” boards
with bearings to allow smooth movement. Two 1”x 1” angle irons lined with Teflon
strips were mounted above the plywood gate to seal the edges. Teflon was chosen
because of its low fi'iction characteristics which prevented the throttle fiom binding.
The final throttle assembly is shown in Figure 5. The throttle system was sensitive
enough to allow hot wire calibration and to maintain steady flow conditions during
velocity profile measurements. The pressure drop between the sub-atmospheric test
chamber can be controlled fiom 0.010 to 0.50 inches of H20 which results in a
velocity range of 2.03 tol4.40 m/s through the calibration nozzle. (The velocity was
calculated utilizing the Bernoulli equation rearranged as

2 P -P
Vi—W=—(—'7—?—)-—g(zl—z.>. (1)

Assume Vle, and -g(Z1-Zz) zO results in:

V2 = M (2)
i ,0

These equations assume incompressible, inviscid, steady flow along a stream line
through the nozzle). The pressure can be controlled by sluice position to increments of

approximately 0.001 in-H20 which was within the resolution limits (0.0025 inHzO) of

the pressure transducer. This corresponds to a velocity resolution of approximately 1 5

cm/s at typical experimental flow settings.

 

Figure 5. Sluice throttle mechanism used to control airflow rates (top view).

2.5 FLOW CONDITIONING

A structure was built for flow conditioning at the inlet of the system. This
structure was designed to remove variability caused by changing ambient conditions
and was necessary to insure experimental reproducibility and flow homogeneity. This

system consists of a preconditioner and a flow contraction (Figure 6). The pre-

conditioner consists of 3 wire screen stretchers and a honeycomb straw structure. The

contraction is two-dimensional and constructed from 1/4” thick Lexano.

 

Figure 6. Airflow conditioner.

The design of the preconditioner was based on a design from the TSFL and
the assistance of Professor Foss. The honeycomb structure removes large scale
turbulent momentum eddies from the fluid. The wire screen used consisted of 30 x 30
wires per square inch mesh. This mesh removes smaller scale eddies from the fluid.
The wire meshes were placed 3.5” apart. The disturbances created by the previous
mesh are greatly dampened prior to reaching the next mesh (Blevins [1984]) at this

distance.

A two dimensional contraction was utilized to accelerate the flow and to
provide a uniform velocity profile at the contraction exit plane. The wall curvature was
the critical design criteria for the flow contraction. The design was obtained fiom the
principles defined in “Design of Two-Dimensional Wind Tunnel Contractions” by
Morrel [1977]. The mechanical details of the contraction are presented in Appendix 1.
This curvature prevents thick boundary layers fi'om being produced at the exit plane of
the contraction and it prevents flow separation in the contraction resulting from flow
acceleration. The design accomplishes these ends by maintaining favorable pressure

gradients and smooth transitions throughout the length of the contraction.

2.6 Data Acquisition System and Motion Control

A Keithley MetraByte DAS-TC data acquisition board and a DAS-02 analog
output board were utilized with sofiware written in QuickBasic° (Microsoft [1988])
(the QuickBasic program is presented in Appendix 2). These two boards can sample
16 channels of data and are capable of 2 channels of analog output. The DAS-TC
board was utilized to obtain temperature and voltage measurements and the DAS-02
to output voltage control signals to a VCO-4050 stepper motor controller. The
voltage measurements were from the hot wire anemometer, the pressure transducer,
the humidity probe and the stepper motor control signal.

The VCO-4050 stepper motor controller and a CMD-SO chopper driver from
American Precision Inc. and the Eastern Air Devices Model #LA23ECK-81 step
motor with the custom built traverse system allowed the hot wire sensors to be moved

in increments of 6.6 pm.

12

2.7 Pressure Measurement

A MKS Baratron Type 225A pressure transducer with a V2” H20 full scale
reading was utilized to measure pressure differentials in the system. These pressure
measurements were used to calculate velocities for hot wire calibrations and to meter
the flow for the actual experimental work. The MKS transducer can resolve pressure
measurements with an accuracy of i0.0025” H20. The relationship between velocity
and pressure in the calibration nozzle is defined in Equation 1. The transducer was
utilized to read the pressure drop across the contraction as well as the pressure
difference between the atmosphere and the interior of the test section. Measurement of
the pressure drop through the contraction allows for the inviscid core velocity in the
test section to be calculated. This velocity was utilized to non-dimensionalize the
experimental results. The pressure difference between the atmosphere and the sub-
atrnospheric test section allows the inviscid velocity through the cahbration nozzle to

be calculated.

2.8 Hot Wire Sensors

The hot wire sensors utilized in this facility consist of 5 mm diameter tungsten
wire controlled by a TSI constant temperature anemometer (CTA) system (model
1051-2). The wires are powered electrically to produce an elevated temperature to
maintain the prescribed resistance value. The CTA system utilizes a bridge circuit to
control the current through the hot wire sensor and thus its resistance. The CTA
system outputs the voltage signal required to maintain the sensor resistance. This

voltage Signal can be correlated with a known flow rates (through the calibration

l3

nozzle) to calibrate the hot wire sensor so that it can be used to measure unknown
flow velocities. The calibration results fi'om a balance between the electrical heating of
the sensor and the convective heat transfer. The governing equation used during the

calibration is

E 2 = A —— B V ". (3)
Essentially, the square of the voltage output (E) from the anemometer bridge was
related to the velocity (V) by three calibration coefficients (A, B, and n). The hot
wires are capable of measuring velocities within the range of the calibration to a

resolution of 352%.

2.9 CALIBRATION NOZZLE AIRFLOW HEATER

A stainless steel 30 x 30 wires per inch mesh screen was utilized to heat the
flow through the calibration nozzle (Figure 7) to provide an thermal energy source to
heat the thermographic targets. The mesh has a resistance of 2 Q and was heated
electrically with a HP6274B DC power supply. The low resistance of the screen
requires a power supply with a high amperage output to supply adequate energy to the

flow. The power supply was capable of providing 15 amps (30 watts).

l4

30 x 30 Stainless
Steel Mesh

  
 

Electrical
Input

Calibration
Nozzle

Figure 7. Calibration nozzle airflow joule heater.

The screen was able to create a temperature rise of 30°C above ambient when
powered by 8 amps with an output flow velocity of 10 m/s through the calibration
nozzle. The core of the flow has a measured temperature stability of i l .5 °C. The

design utilized presently was limited to 9 amps due to electrical arcing.

2.10 CHARACTERIZATION OF THE WIND TUNNEL FACILITY
Velocity profiles of the flows through the calibration nozzle and the two-

dimensional contraction exit were obtained. Only the calibration nozzle was used in

the current study. However, the test section results were obtained to explore the
capabilities of the facility and there are presented in Appendix 3 for completeness.
Figure 8 Shows the calibration nozzle schematically and the coordinate system used.
Figure 9 shows the velocity profile at X/D = 0.5 in the calibration nozzle.

The majority of the exit area of the nozzle has a laminar uniform velocity
profile. The flow was determined to be laminar from the hot wire voltage signal.
This signal appeared smooth on an oscilloscope and had an RMS value of 2% of the

mean flow value. It was important to note that the Reynold’s number,

U-D
ReD = U a (4)

 

based on the flow velocity and the nozzle exit diameter was equal to approximately
13,000. This number seems high enough to indicate that the flow should be turbulent,
but because the flow was still developing, the central region of the nozzle was still

unsheared and laminar.

l6

Flow

 

 

 

Figure 8. Axi-symmetric calibration nozzle and associated coordinate system.

 

Calibration nozzle velocity profile

 

 

 

VNInvIscld

 
       
 
  

Nozzle Exit

~~ Dia=21mm ,
l

 

 

 

 

5 10 15 20
Position (mm)

 

 

 

Figure 9. Calibration nozzle exit velocity profile.

17

Temperature profiles of the air at X/D = 0.5 below the exit plane of the
calibration nozzle were also obtained. The profile for 10 m/S inviscid flow velocity and
ambient temperature equal to 22°C with 8 amps and 6 amps provided to the flow

heater are shown in Figure 10.

 

 

Temperature profile at calibration nozzle exit
(10 mls flow velocity)

ac
+8Anperes
m —-x— 6 Anperes

 

 

 

 

 

 

___ __ .- __- L. .._ . _.._.._.,—__.———-r___—._._4

Temperature (C)

 

 

 

-15

Position (m rn)

 

 

 

Figure 10. Calibration nozzle exit temperature profile.

This figure shows that there was a core of relatively constant temperature
approximately 1 centimeter in diameter for each current setting. This was much larger
than the specimens that were tested and was considered adequate for this study. The
stability of :1 .5 0C was larger than the uncertainty of the infrared camera (: 03°C)

utilized but appeared to be the stability limit of the heating system used. The room

18

temperature above the nozzle inlet was monitored and it experienced a standard
deviation of approximately 3: 1 0C over time period of 1 minute. This variation in the
ambient conditions probably accounts for the majority of the fluctuation found in the

nozzle temperatures.

2.11 SUMMARY

A wind tunnel facility was developed with the capability to create and monitor
a flow with acceptably uniform velocity and temperature profiles in core regions of the
calibration nozzle. The velocity, temperature and humidity of the flow can be
individually measured and recorded with the data acquisition system. Fine motion
control to position the sensors was obtained with stepper motors and a translation
stage. This facility was characterized specifically to obtain measurements with infi'ared
thermography of temperature and emissivity of optically unresolved specimens and to

support other research projects in the Heat Transfer Research Laboratory (HTRL).

l9

CHAPTER 3. CALIBRATION OF INFRAMETRICS IR CAMERA AND

ASSOCIATED DATA ACQUISITION SYSTEM

3.1 INTRODUCTION

Infrared cameras produce temperature data as a function of spatial location and
time. The appropriate data recording system typically includes a time marking video
overlay, a video recording system (VCR), video monitors and image processing sofiware
to analyze the data. The video overlay allows time data to be saved simultaneously with
the temperature field information. The video recording system allows data storage and
post processing of the information. The experiments can be viewed in real time on the
video monitors. The sofiware allows digitizing of the video signal and provides necessary
quantitative analysis tools.

The following equipment were available to achieve these functions: an Infiametrics
600L infrared camera; a For-A VTG-33 video timer; a Sony PVM-l343MD Trinitron
monitor; a Panasonic AG-2400 monitor and Image Pro Plus v2.1 (Media Cybernetics
[1995]).

The interfacing and optimization of this system to obtain optimal experimental
analysis is explored in this chapter. An ideal system response is outlined and an equipment

arrangement to achieve that response is defined.

3 .2 STATEMENT OF PROBLEM
The Inframetrics infiared camera outputs an 8 bit digital signal in R8170 video

format that includes the temperature field information. An ideal system output is defined

20

for this work as linear response that spans the entire dynamic range (Figure l 1). The

linearity of the output maintains uniform sensitivity over the experimental range and

 

 

 

 

 

 

 

Simplifies processing.
Ideal Output of the Infi'ametrix IR Camera
250
200 I-
Guy 150 4.
Scale100 0
50 AL L—o— Ideal Response ,
I
0 .L r J. t i
0 0.2 0.4 0.6 0.8 1
Non Dimensional Temperature

 

 

 

Figure 11. Ideal gray scale-temperature response for the Inframetrics infrared camera and
associated data acquisition system.

Different equipment or equipment arrangements are often utilized in the data
recording circuit depending on the experiment, experiment location (the camera is
portable), and equipment available . There was evidence that different equipment
combinations could result in variations in the data when data obtained at the top of the
lnfrarnetrics camera range would saturate at different levels for different users. It was

determined that this was directly caused by the data recording and processing equipment.

21

3.3 EXPERIMENTAL PROCEDURE
3 .3 . 1 EXPERIMENTAL CONFIGURATION

An PIP-6236B Triple output power supply powered a plate heater to a temperature
between 5 and 10 degrees centigrade above ambient temperature. The plate temperature
was monitored with a calibrated T-type thermocouple (i0.1 0C accuracy). The IR camera
was then set with its temperature range lower limit just below the plate temperature. The
system was allowed to stabilize for 15 minutes to insure it was at steady state. The IR
camera was then used to define the plate temperature field with several different
equipment arrangements. The area fiInction of the IR camera was used to define the

region of interest (Figure 12). This insured that the same surface area was analyzed each
time. This avoided variability due to non-uniform surface properties on the plate.

 

Figure 12. IR camera tool for area temperature measurement.

22

Data were obtained with a variety of equipment configurations. These equipment
arrangements were chosen to represent potential data acquisition arrangements or systems
that are currently being utilized in the laboratory. The different equipment configurations

that were tested are shown schematically in Figure 13.

Soliciiinlic of equipment l(’~'.\‘t (-onfiglirritiOIiS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[P
Plus
IR
I: (fi‘mmnrd Computer
IP
Plus
[R Sony
l: (i‘amerzi \lonilor Computer

 

 

 

 

 

 

 

 

 

 

 

1P
Plus

IR Sony
C (Mignon-T \-'(_‘R Monitor Computer

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IP

Plus

IR Video Sony ‘ .,
E ("x-”nerd ”‘l Timer ‘ Monitor Computer

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fr‘-l'ir*I'I'ie-ili<f‘- ol‘ I)lifi§'l)tl("l\' C'Olliltllll‘itlIOII

 

 

 

 

 

 

1P
Plus
Sony
\v’t'R ‘ Monitor COIIIIHIII’I"

 

 

 

 

 

 

Figure 13. Schematic of the different equipment configurations tested.

23

The HP power supply was then adjusted to supply more electrical energy to the
plate heater and the system was allowed to stabilize at a higher temperature. The IR
camera was again used to measure the plate temperature with a variety of equipment
configurations. Approximately 45 seconds were required to obtain data with each
equipment configuration. Thus data with all of the different configurations could be
obtained within 3 minutes. The temperature (as measured with the area function of the IR
camera and a thermocouple) always remained stable during this time interval.

This process was repeated at various points in the temperature range defined by
the IR camera Typically a minimum of five data points were taken within the Inframetrics
camera temperature range. Temperature ranges of 5 and 10°C were utilized for the
experiments presented here for convenience, any of the camera temperature ranges could

have been utilized. Typical images are shown in Figure 14.

24

 

e. 473°C f. 50.4°C

Figure 14. Playback data image sequence fi'om the equipment test experiment conducted
on 2/11/97.

25

3.3.2 BRIGHTNESS AND CONTRAST TESTS

The image processing software has three settings that affect how data are captured
and imported: the brightness, contrast and gamma fimctions. The gamma function affects
the linearity of the gray scale as a function of temperature. The gamma fiinction was set at
1.0 and not adjusted Since a linear curve was desired to maintain uniform sensitivity over
the complete range. The brightness setting affects the y-intercept of the gray scale as a
fimction of temperature. The contrast setting affects the slope.

The HP power supply was again used to raise the temperature of the plate heater
fiom 5 to 10 0C above ambient temperature. The Infiametrics camera was set with the
lower temperature range limit just below the temperature of the plate heater. The system
was allowed approximately 20 minutes to come to thermal steady state. The IR camera
area tool was employed to monitor the stability of the plate temperature.

The contrast was adjusted from settings of 65 to 100 (or until saturation occurred)
with the brightness set at a fixed level. Brightness settings of 48, 50,55, and 65 were
selected for testing. These brightness test levels were chosen because settings above 65
caused saturation almost immediately and brightness settings below 48 clipped 40% off
the dynamic range.

Experiments were conducted primarily with the plate temperature near the lower
temperature limit of the IR camera. This was primarily due the fact that the loss of
sensitivity in the IR camera resulted from the upward offset of the RS-170 signal. Tests

were run with the plate temperature near the upper limit of the Inframetrics camera’s

26

temperature scale. Preliminary results were similar to the lower limit tests. Focus was
then placed on the lower limit tests.

It is important to note that the thermal target was at steady state for both the
equipment configuration and the brightness/contrast tests. It is expected that the results

and conclusions of these experiments would also apply for transient experiments.

3.4 RESULTS

These experiments showed that the data acquisition and processing equipment
available in the HTRL affected the data. This effect can be seen in Figure 15 as each piece
of equipment was added to the data acquisition circuit the gray scale-temperature response
changed. The only piece of equipment that did not effect the signal was the Sony PVM-
1343MD Trinitron monitor. The For-A VTG-33 affected the data in different ways
depending on the setting of the 7 50 switch on the back panel. This switch placed 75 ohms
of impedance between the device and the data signal. The remainder of the equipment and
processing software had definite affects on the data and thus, the results of any experiment

conducted with the IR camera system

27

 

IR camera gray scale response to terrperature input

 

 

 

  

 

 
 

 

 

 

 

 

 

300.00 1
250.00 7.
200.00 A
0
E
5, 150.00 «-
O + P Plus Direct
100-00 *‘ .. . - anitor and PFlus
-——Ci— Through VCR(M) Playback)
500° ‘“ + Saturation Level
—0— From Playback
0.00 + + I; .* ¢ F + , ; i
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Tenperature

 

 

 

Figure 15. Equipment configuration effects on the IR camera gray scale-temperature
response.

The results of each steady state test were compared to determine which equipment
arrangement best matched the ideal system response defined in Figure 11. This ideal was
defined to maximize the dynamic range of the IR camera and to maintain uniform
sensitivity over the entire temperature measurement range.

It was found that the For-A VTG-33 Video Timer with the 75 ohm switch on did
not affect the RSl70 signal (Figure 16). The video timer allows time and date overlays to
be placed on the data. This is highly desirable for transient tests, marking data with

identification for later use, or time marking data for sequential information.

28

 

 

 

 

 

 

 

 

 

Effect of ForqA VTG video timer on RS-170 dgnal
300 1:
250 ~ :
200 ..
.9
8
m 150 ~~
>~
E
O
100 «» + Playback w/o ‘
Titer l
50 «» —I— Playback wl Trier i
(750hm on) I
0 P t i 1r l A. t i
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Nondimensional Tern perature

 

 

 

Figure 16. Comparison of equipment tests with and without
the For—A VTG-33 video timer.

The Panasonic AG-2400 VCR did affect the RSI 70 signal and thus the gray scale
to temperature relationship. Specifically, it created a negative offset on the RS-170 signal.
This counteracted the positive offset fi'om the camera and was necessary to capture the
entire dynamic capabilities of the IR camera. The best results were obtained with the
VCR in the data acquisition circuit but not recording the data stream. This arrangement
would be possible during steady state tests, but is impractical due to the time required to
process the data from a typical experiment with the Inframetrics camera.

Based on these results, the recommended equipment configuration for data
acquisition is to connect the Infiametrics IR camera serially to the For-A VTG-33 Video

Timer, to the Panasonic AG-2400 VCR, to the Sony PVM-l 343MD Trinitron monitor (in

that order).

29

The results of the brightness and contrast tests are shown in Figure 17. The
brightness setting moved the overall curve upward, maintaining a constant slope. The
contrast setting altered the slope of the resulting line. These settings appeared to operate
independently of one another. These tests showed that the best settings to maximize the
sensitivity were Brightness 48, Contrast 66. This setting was selected to lower the y-
intercept of the gray scale as a function of temperature response. Figure 17 may make a
contrast setting of 65 look like the best way to lower the response. However, this setting
limited the upper response to a gray scale value of approximately 150 (thus sacrificing a
large portion of the dynamic range). Based on these results, the optimal data processing
system to achieve the ideal response shown in Figure 11, is the Panasonic AG-2400 VCR

connected to the Sony PVM-1343MD monitor to the Image Pro Plus software (Figure

 

 

 

 

 

 

 

 

 

 

 

 

13).
Gray Scale as a function of Contrast Setting
250.00 1
5 re H—w—fi ". ‘
1
200.00 x + Brightness=48
+ Brightness-=50
a Bnghtness=55
150.00 T + Bnghtness=65
Gray '
Scale
100.00 R l
50.00» g
l
0.00 r i t i J.
60.00 70.00 80.00 90.00 100.00 110.00

Contrast Setting

 

 

 

Figure 17. IR camera gray scale output as a function of contrast setting.

(Low end of Infrarnetrics camera temperature scale).

30

3.5 OPTIMAL CONFIGURATION TESTING

An optimal data recording system and a data playback system were defined from
the results discussed above. The data recording system consisted of the Infiametrics IR
camera connected to the For-A VTG-33 Video Timer and the Panasonic AG-2400 VCR
(in that order). The data playback system consisted of the Panasonic AG-2400 VCR
connected to the Sony PVM-l343MD Trinitron monitor with the signal then fed to the
Image Pro Plus software (Figure 13).

This optimal equipment configuration was then tested. The test procedure was
altered to reduce the time required. The plate was powered with the PIP-6236B Triple
output Power Supply to approximately 10 degrees above ambient. The plate was allowed
20 minutes to stabilize thermally. During these experiments, the temperature range of the
Inframetrics IR camera was adjusted instead of changing the plate temperature. This
procedure had the same effect on the IR sensor as adjusting the plate thermally. Several
verification tests were run to insure that the results were identical for each procedure.
This modified test procedure allowed for many more data points to be taken throughout
the temperature range. This procedure allowed verification of the linearity of the

Infrarnetrics IR camera output.

3.6 RESULTS OF OPITMIZED EQUIPMENT CONFIGURATION

Previous work (White [1996]) had Shown linear gray scale response to
temperature but with a different Slope and concluded that the offset was a characteristic of
the radiometer itself. This previous study did not incorporate the affects of the contrast

and brightness settings of the Image Pro Plus software which allow the offset to be

31

adjusted. Figure 18 and Figure 19 Show the IR system response with optimal brightness
and contrast settings and a comparison to previous calibrations obtained without these
controls.

The equations in the upper corner of Figure 18 Show the system insensitivity to
contrast settings between 66 and 70. The correlation coefficient (R2) values Show the
degree of linearity. The x-axis of these plots is presented in terms of non-dimensional

temperature (T') which is defined as

 

T. T- T(lower IR range limit)

_ . 5
IR camera Range (i. e. 10° C) ( )

 

Optimal contrast tarts

250.0 ,_
B48066 Result /)T

y = 246.36x + 7.6393
200.0 ,_ R2 = 0.9992

 

 

Gm y Scale

“DDT

500"

 

 

 

 

 

 

0 OJ 02 03 0A 05 03 OJ OS 09 1

Non Dimensional Temperature

 

 

Figure 18. Gray scale-temperature relationship with optimal contrast.

32

 

 

 

 

 

 

 

 

 

 

Optima] equrpment tests
20 degree IR range
250.0 ..
Test I Result
y = 240.07x + 5. I882
200° T R2 = 0.9996
9.)
3 150.0 _0_ Ideal R5311]!
y = Z55x
E
E3 100.0 ~-
—x-— B48C6’6‘ Tea 1
500 w ' - B48C6'6‘Tcst2
I '—i—- Heal Resrxmse
0.0 '// r I 4 + l
0 0.2 0.4 0.6 0.8 1
Non Dimensrbnal Tempera tare

 

 

Figure 19. Optimal equipment configuration gray scale-temperature relationship.

The system response,
Gray Scale = 2455- T‘ + 7.7, (6)

is almost identical to the optimal system response defined at the beginning of the
investigation. The ideal response includes 100% of the dynamic range over the entire
temperature scale. The optimized equipment configuration response is linear and provides
95% of the dynamic range for experimental work.

The results presented here should also be considered typical for any range setting
on the IR camera but the specific experimental configuration should be characterized prior
to each experimental sequence with the relevant equipment. The procedure described can

be run in a reasonably Short time flame and will help remove bias error created by the data

acquisition system.

33

 

CHAPTER 4. SLIT RESPONSE FUNCTION - THEORY AND EXPERIMENTAL
DERIVATION

4.1 INTRODUCTION

The Slit Response Function (SRF) is a measure of the ability of an infrared system
to sense radiation from targets that subtend an angle which is smaller than the
measurement resolution. An observable target smaller than the measurement resolution
will produce a data signal that is attenuated as a function of target size. The SRF must be
known so that these Signals can be processed correctly and accurate measurements
obtained. The lower thermal and optical measurement limits of the system can be defined
if the SRF is known and corrections can be applied.

The instantaneous field of view (IFOV) and the field of view (FOV) are important
factors in determining system performance. Generally a large field of view is desired with
a small instantaneous field of view. The field of view is the total area the camera can
sense and the instantaneous field of view is the projected detector area (Figure 20). As an
analogy to a digitized image, consider the field of view the number of pixels and the

instantaneous field of view as the pixel Size.

34

 

lnstmrtaneous
Field of View
(projected

detector size)

 

 

 

 

 

IR Camera Field Target
of View

Figure 20. Schematic of IR camera optical configuration and definition of terms.

Scanning infrared systems tend to have different SRFs in the scan and cross scan
directions due to factors such as detector size and shape, line to line interpolation
schemes, optics and scan rates. The SRF for a given system defines the limits on imaging
and measurement resolution. The optical (imaging) resolution is typically defined as the
target angle that will produce a 50% response (that is, it will measure 50% of the full scale
temperature above or below background). The measurement resolution is defined as a
target width that will produce a 99% response (Holst [1993]). The measurement
resolution defines the smallest target angle that may be thermally measured accurately.

The effects of the scanning direction and target alignment can be seen in Figure 21,
Figure 22 and Figure 23. The scan direction is from left to right. The MAF sensor is at
the same temperature for both images (T ~ 56 0C). The sensor aligned with its narrow
dimension in the scan direction did not produce as large a response as the sensor aligned

with the scan direction.

35

 

Figure 21. Comparison of scan direction effect on IR camera response.

The MAF sensor alignment with respect to the scan direction affects the length of
time that the infrared detector has to respond to the radiation emitted. Figure 22 shows a
much flatter response to the target than Figure 23. This is because the SRF is created by
step changes or steep gradients in the emitted radiation. When the sensor is aligned with
the scan direction, its long dimension allows the infrared detector more time to equilibrate.
Therefore, the scan direction defines the critical dimension related to accurately measuring

temperature and was the focus of the study.

36

 

 

Non-dimensional temperature

0.7

Gray scale profile of narrow dimension of sensor
Sensor parallel to scan direction

 

0.6 <L

0.5 ..

0.4 ..

0.3 +

 

 

 

Figure 22. Temperature profile of narrow dimension of the MAF sensor.

Sensor parallel to scan direction.

 

 

Non-dineneionel temperature

Gray scale profile of narrow dimension of sensor
Sensor perpendcular to scan (Erection

 

.0 .0 9 .0 .0
N on as or m

.0
.5

 

Pixel

 

Figure 23. Temperature profile of sensor narrow dimension.

Sensor perpendicular to scan direction.

 

 

 

The slit response firnction for the 600L Infiametrics IR camera was determined

experimentally. An anticipated SRF response can be calculated fiom

ALe (’19 T)Ao AIFOV
(fcol ) 2

 

l2
AEm = [a jam) TWA/OT,“(xi)d/1]ATFS,S. (7)
1.1

However since this equation utilizes other measured characterisitics of the IR system and
the target (i.e. the Aperiodic Transfer Function (ATF) and the spectral emittance), it is
more useful and much simpler to determine the SRF experimentally. Experimental

determination also accounts for issues such as clouded or misaligned lenses

4.2 IMAGE PROCESSING EFFECTS ON DATA

The Inframetrics 600L camera has a field of view of 200 horizontal x 150 vertical
(0.34907 x 0.26180 radians) (Inframetrics [1994]). There are 175 instantaneous fields of
view (IFOV) per line and 131 lines vertical lines that comprise the entire field of view
(F OV). The electronics sample 256 times per line (1.46 samples per IFOV). These data
are output in an analog RS-170 Signal in US television format. The RS-l70 signal has
525 lines which are interlaced so that there are approximately 240 real lines of data. The
other lines are used by synchronization signals and repositioning of the scanner (Russ
[1995]). The original 175 IFOV x 131 lines of optical samples are now divided into 256 x
232 data bins. The signal was then acquired by the digital frame grabber board and split
again into 640 x 480 discrete pieces of information (pixels). Each IFOV was now spread
across 3.66 digital pixels. Each of the 256 data samples per line was divided into 2.5

pixels horizontally and each vertical line of data was Spread over two lines of digital pixels

38

and interlaced. Although no information was lost, it was converted to analog and

resampled into a different number of information packets as it was processed.

4.3 EXPERIMENTAL TEST PROCEDURE

The procedure to determine the SRF experimentally was taken fiom Testing and
E_v_al_u_aftion of Infrared Systems by Holst [Holst 1993]. The test consists of utilizing a
micrometer traverse system to manipulate a mechanical slit. The slit is placed between the
IR camera and the blackbody source (Figure 24). The IR camera is focused on the edges
of the slit to capture collirnated (parallel) IR radiation emitted from the source. The slit

now appears as a target source to the system. The mechanical slit is closed so that the

zero position is known and then opened in incremental steps and the response of the IR
system is recorded. It is important to insure that one measurement is taken of the full

scale (unobstructed) view of the blackbody to non-dirnensionalize the results.

 

 

  

 

 

 

1.
¢
IR i.
Camera ] T
lS '- -
Blackbody
Slit
Traverse

Figure 24. Schematic of the experimental configuration for SRF
measurement.

39

It is crucial to maintain a large enough distance between the blackbody and the slit
to prevent radiative heating of the slit. A 3” gap between the heated plate and the
mechanical slit was utilized for this study and the back of the slit was made of polished
aluminum to further reduce thermal absorption. A change in the slit mechanism
temperature has the effect of changing the temperature step that the sensor must track and
will seriously affect the results. This is also why it is critical to maintain the blackbody at a
constant temperature.

A blackened 3" diameter aluminum plate thermally bonded to an Omegalux
silicone rubber electric heater acted as the blackbody source for the experiments. The
emissivity of the plate has been measured previously as 0.95 utilizing the temperature
comparison method (Inframetrics [1988]). The actual emissivity and temperature of the
thermal source are not critical to this experiment. However, they must be stable.

The internal focal distance of the IR camera must be accounted for to properly

calculate the viewing angle. The target angle ('9 is defined as

l
{—3 l
9 “" (4+4) ' ‘8’

The distance 15 is the target width in the scanning direction. The distance from the target to
the bayonet mount lens on the IR camera is 1d. The internal focal path length of the
camera is l.- . The Inframetrics 6OOL camera has an 1.- of 260.4mm. The non-
dimensionalized system response can then be utilized to calculate the target temperature
for any target at any distance from the IR camera assuming the target emissivity and target

size are known.

40

4.4 RESULTS

The SRF for the Inframetrics 6OOL camera with no external optics is shown in
Figure 25. The error bars show the 95% confidence region determined from the
experimental data. It can be seen that the uncertainty increases significantly at srmller
angles. Figure 26 shows the 95% confidence region divided by the mean SRF value. This

plot defines resolution limits if predictive capabilities of the system are considered.

 

 

 

 

 

 

 

 

 

 

Slit Response Function
lnframetrlcs 600L IR Camera - No External Optics
1
0.9 ..
0.8 V
0.7 4r
E? 0-6 " lnstmtaneous
W 0.5 4» Field ofVIew
o 4 (1.99 mrad)
' i" . . a... +onertsmdy
0-3 .. x mus/96)
0.2 ..
0.1 r l t r ¢ ¢ l
o 2 4 6 8 1o 12 14 16 18
Target Angle (mud)

 

 

 

Figure 25. Slit response function with 95% confidence region (n=5)
for the Infiametrics 600L IR camera.

41

 

95% Confidence Raglan/Mean Value

0.4 T- - —

 

0.35 <r
0.3 T
0.25 4L
0.2 —~
0.15 «-

0.1 ..

95% Confldeneelllean

 

0.05 4~

 

 

0 : ¢ 11 1* ¢ % t
0 2 4 6 8 10 12 14 16

Target Angle (m rad)

 

 

 

Figure 26. SRF confidence region/mean value
for Infi‘ametrics 600L IR camera with no external optics.

The HTRL also has a 3x telescopic lens with a 6" close up attachment for working
with smaller samples. The SRF of this configuration was also determined experimentally
and is shown in Figure 27. The figure also shows the 95% confidence regions. This lens
has only one working distance, so target width and target angle are interchangeable (since
Id and 11 are not adjustable). The internal focal length (1,) is different than in the previous
case without the external lenses. The sum of the internal focal length and the working
distance is equal to 77 mm for this arrangement. The actual working length is 145mm, the
decrease is a result of the magnification factors of the lenses. Due to this complexity, the
results are presented as a function of target width instead of target angle.

Figure 25 and Figure 27 both show data obtained by White (White [1996]) with
the same equipment in the same laboratory. White used only a single data set and the IR

camera system was not optimized in the same manner descrlbed in this document. For

42

example, in Figure 27, White’s data levels out at 0.1. This is indicative of either a poor
calibration or heating of the slit. However, White's data show similar trends and are

presented for completeness of the work on this infiared system.

 

 

 

 

    

 

 

 

 

 

 

SRF response of Inframetrim 600L IR camera
with 3x telescopic and 6" lenses
1 --x-—-———-; ;
0.9
0.8 «r
0.7
0.6
'n‘: o
m .5 l
0.4 'z“ "‘ l
03 Instantaneous + Current Study 1
o 2 fl Field of View x White (6,96) .
f (92.4 microns) ’
01 1M
0 t + 4 s + t t 4
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Target width (m m)

 

 

 

Figure 27. SRF of the Inframetrics 600L infrared camera
with 3x Telescopic and 6" close-up lenses.
Inframetrics specifies that the 50% SRF corresponds to 0.002 radians for this
system when it was manufactured. The system currently has a 50% SRF at 0.00291
radians (Figure 25). There are several possible explanations for this variation The system
is due to be calibrated and it may require that the internal optics be cleaned. Also, the
scanning mirror and the IR detector may be slightly misaligned according to Infi'ametrics

technical support staff.

43

4.5 PREDICTIVE CAPABILITIES

The data obtained allows for temperature measurements to be obtained from
targets smaller than the measurement resolution of the system. The data also allows for an
understanding of the general capabilities of the infrared system and its limitations so that
experiments can be properly designed. Small targets can be measured thermally with an
understanding of the accuracy of the measurement and experiments can be designed to
avoid resolution issues. Most importantly, a general understanding of infrared thermal
measurement techniques can be obtained so that the equipment can be used properly.

The following example will illustrate how to utilize the IR camera for an
unresolved target. Figure 28 shows a pseudo-colored infiared image of an electrically
heated MAF sensor taken with the Infiametrics 600L IR camera operating with the 3x
telescopic and the 6" close-up lenses. The sensor has dimensions of 1.5 mm long by 0.5

mm wide.

J . Umm

r————-~l

 

Figure 28. Thermographic image of MAF sensor.

44

Figure 27 shows that 0.5mm is not large enough for the Inframetrics camera to
have a full scale response. The IR system will produce a 90% percent response with a 3%
uncertainty at a 95% confidence level.

The gray scale response of the camera is related to the temperature measured as
shown in Chapter 3. The system has been calibrated so that a known linear relationship

exists between gray scale and temperature as shown in

 

Gray Scale — 7.6
T = 2 4 6 * Trange + T Lower Limit . (9)

Note that this relationship was obtained with specific equipment and the relationship
should be determined for each particular configuration. The variable nge is the
temperature range setting on the IR camera (i.e. 10°C) and TLC“, mm is the base of that
range (i.e. 40°C for a range of 40-500C).

Consider that the camera measures a gray scale reading of 100 (out of 256) on a
target with an SRF of 66% and an uncertainty of 9% at a 95% confidence level. If the
temperature measurement scale was set on 40-50 0C, this would be equivalent to
measuring 43.9 0C using Equation 9. The correct gray scale value and uncertainty in that

value can be found from

Grey Scale

measured

SRF (10)

 

Gray Scale Adm, =

and

45

Gr Seal i‘VoUhoerta'n
AGray Scalemo, = ay “”st 1y

 

(11)
%U10ertain@=95%oafideme rage

respectively. This would result in a gray scale reading of 152 or 459°C and 95 %
confidence gray scale limits of 137 - 164 ( 454- 464°C) for this example.

The resolution decreases linearly with increasing temperature range settings of the
camera. Consider the previous example but with the temperature range set at 50°C (30-80
0C for example). A gray scale reading of 100 would correspond to a temperature of 48.0
0C (again, using Equation 9). The corrected gray scale value would be 152 as before but
this would result in a temperature of 58.2 0C with a 95% confidence range from 56.0 -
64.3 0c.

It appears logical to use the smallest temperature range possible to reduce the
range and thus reduce the error. However, it is essential that the full scale reading and the
attenuated reading be contained in the IR camera range setting for this procedure to
fiinction optimally.

Experiments were conducted to investigate the predictive capabilities when the
target temperature exceeded the range setting on the infrared camera. The experimental
conditions included a target angle of 0.002 radians and the camera temperature range set
10-20 degrees below the target temperature and not encompassing the target temperature.
These experiments showed that the system was unable to reliably predict the correct
temperature utilizing the SRF data presented here. It is therefore recommended that the
temperature scale on the IR camera be set in such a way that the temperature of the target

is incorporated in the range.

46

4.6 CONCLUSIONS

A universal curve for the Infiametrics 600L IR camera was obtained (Figure 29).
This information specifies the response of the system and the uncertainty associated with
the measurement obtained. Proper non-dimensionalization requires that the working
distance and the internal focal length be utilized. The sum of the working distance and the
internal focal length (11 + 1.1) for the system with the 3x telescopic lens and the 6" close—up

optics is equal to 77mm. The internal focal distance for the system with no external optics

is 260.4mm

 

SRF Comparison Between System Configurations

 

   

 

+wl Extemal Optic
—a—wlo optics

 

 

 

Non-Olmenslonallzed
Gray Scale

 

 

 

 

0 r 1 +
0 0.005 0.01 0.015 0.02
Target Angle (radians)

 

 

 

 

Figure 29. Comparison of SRF between system configurations
(no optics compared to external optics).

Dimensions in the scan direction are specified in Table l for the optical resolution
(50% detector response) and the measurement resolution (95% detector response). Recall

that for the system with external optics, the working distance is fixed at 14.5cm from the

47

external lens. Uncertainty ranges can be obtained from the 95% confidence regions shown

in Figure 25 and Figure 27.

Table 1. SRF summary for the Inframetrics 600L infrared camera in the HTRL.

 

 

 

 

 

Slit Response Target Information (no external Target Information (w/ external
optics) optics)
% Response Angular (rad) Width @ ld Angular (rad) Width @
14.5cm
50% Response 0.00291 1.49mm @ 0.00291 0.224mm
(optical 25cm
resolution) 2.21mm @
50cm
99% Response 0.0105 5.36mm @ 0.0105 0.808mm
(measurement 25cm
resolution) 7.98mm @
50cm

 

 

 

 

 

The information presented here Should allow for the Inframetrics 600L camera to

be utilized effectively to perform thermal measurements of targets as small as 25% of the

measurement resolution. A substantial increase in measurement uncertainty will occur as

the target becomes smaller and less visible to the camera sensor. Chapter 5 will Show an

experimental to confirm that these compensations can be realized for small objects.

48

 

CHAPTER 5. EXPERIMENTAL VERIFICATION OF THERMOGRAPHIC

TEMPERATURE MEASUREMENT UTILIZH‘IG SLIT RESPONSE CORRECTION

5.1 INTRODUCTION

The procedure that will be outlined compensates for two separate variables that
will affect the measurement accuracy of the infrared imaging system: the dimensions and
emissivity of the target. The Size of the target generally can be determined easily and
accurately with either a mechanical measurement device (such as a micrometer) or
optically using a calibrated reticule. Either of these methods are accurate enough for the
study presented here. It is important to note that if the thermographic instrument is
capable of large magnification values, the measurement of the target dimensions becomes

more critical. The target emissivity will be determined as part of the procedure.

5.2 EXPERIMENTAL PROCEDURE
The following procedure was utilized to measure the target dimensions and

emissivity so that general experimental measurements could then be taken:

1. Measure target dimensions.

2. Position target at X/D=0.5 under the calibration nozzle (position a calibrated
thermocouple below the target to monitor flow temperature).

3. Fill Infiametrics 600L infiared camera dewar with liquid nitrogen, turn the camera on
and run for 20 minutes. Refill the dewar.

4. Position Infiarnetrics 600L infrared camera inside the sub atmospheric chamber so that
the target is in focus. Measure the distance between the infrared camera and the target.

Refer to Figures 25 and 27 to determine SRF value and uncertainty.

49

. Seal the wind tunnel. Turn on the prime mover and set the sluice gate so that there is a

pressure differential (measured with the MKS Baratron pressure transducer) of

approximately 0.241 in-H20 across the calibration nozzle (This corresponds to flow

velocity of approximately 10 m/S).

. Turn on the HP6274B DC power supply to provide 8 amps of electrical current to the

flow heater (this corresponds to a flow temperature of approximately 56°C, depending

on the ambient temperature).

. Allow the system to equilibrate for approximately 20 minutes.

. Determine the target emissivity (e).

a. The flow temperature is measured with a calibrated thermocouple.Calculate
the gray-scale value from the infrared camera that will produce the flow

temperature using

T

flow - TRange-Min

 

Gray Scale = L ] -SRF -256 (12)

Range
for the ideal response. If the system was calibrated (Chapter 3), use the

appropriate transfer function such as

- 7Iliange- Min

 

T
Gray Scale =[ ”W J -SRF- 240.07 + 8.19 (13)

Range

for the 20°C calibration.
b. Adjust the camera emissivity setting until the gray-scale output matches the

value calculated in step 8a.

50

9. Adjust the HP6274B DC power supply to provide 6 amps of electrical current to the
flow heater (this corresponds to a flow temperature of approximately 36°C, depending
on the ambient temperature).

10. Repeat step 8 as a verification of the emissivity.

11. The target can now be placed in any environment (for example, its operating
conditions) and its temperature can be determined (within the uncertainty Specified in

the Figure 25 or 27 appropriately).

5.3 EMISSIVI'IY MEASUREMENT

The target emissivity can be determined by two different methods. The first is a
procedure that is outlined in the Inframetrics 600L operator’s manual (Inframetrics
[1988]). This method requires measuring the energy fiom two different thermal sources
(source A and source B) directly and reflected off the test specimen. The reflectance (p)
can be calculated from this information using

Source A Refected Level-Source B Reflected Level
p :

 

(14)

Source A Direct Level - Source B Direct Level

The target emissivity can then be calculated fi'om a simple energy balance such as

QR=(a+r+p)-QR:>a+r+p=1. (15)
The energy incident on the target (QR) must either be absorbed (a), transmitted (r) or
reflected (p). This is further reduced to
8 = 1 — p (16)
if the target is opaque (its transmissivity (r) is equal to zero) and the gray body

assumption ofa=€ (Kirchofl’ S law) (Weast [1986]) is applied.

51

This method can be difficult to implement. Difficulty arises from three
requirements of the method. The reflection of the two reference sources must be viewed
on the test specimen. The reflection dimensions must be known if the slit response
function correction is to be applied. The reflection can be very difficult to measure
accurately. Finally, if the test Specimen has a high emittance (absorptivity) (>0.8) obtaining
a reflection that can be imaged clearly is almost impossible unless very high temperature
reference sources are utilized.

The second method, which was employed in this study and included in step 8 of
the procedure in section 5.2, was to place the target in an environment such that its
temperature was known. The dimensions of the target must be known prior to the
application of this method because the unresolved target requires the SRF correction to
measure temperature. The isothermal environment was obtained by using the calibration
nozzle and the electric flow heater (Chapter 2, section 2.9) to provide a jet flow with a
known, uniform temperature profile (within : 1.5 0C). It is important to raise the target
temperature above ambient to increase the ratio of emitted radiation to reflected radiation
(effectively, the signal to noise ratio). The target temperature can then be measured
directly with the infrared camera (with the slit response function correction applied). The
emissivity setting on the camera can then be manipulated until the gray scale output
matches the calculation from the appropriate transfer firnction. It is important to note that
although the inflated imaging system was calibrated previously (Chapter 3), new software
and a new VCR were utilized during the experimental verification. The addition of the
new equipment allowed for the ideal response (Equation 12) to be realized and thus

Equation 12 was utilized throughout this chapter.

52

It is critical to use the SRF correction to accurately measure the emissivity. An
‘apparent’ emissivity will be measured without the SRF correction. This restricts the
experimental configuration to the original working distance (Gartenberg and Roberts
[1990]). Any deviation from this working distance will cause the ‘apparent’ emissivity to

change and introduce systematic errors to the measurements.

5.4 EXPERIMENTAL VERIFICATION

Two experimental procedures were conducted. The first procedure utilized the
MAF sensor shown in Figure 28. This target required that the external optics were
utilized and produced an SRF value of 0.9. The second target was a metal cylinder
(painted black) with a diameter of 1.75 mm. The IR camera was positioned 34 cm away
fi'om the target and the external optics were not utilized. This produced a target with an
SRF of 0.57. The cylinder could have been viewed from closer or with the external optics
to obtain a well resolved target. However the test was conducted to investigate the SRF
correction.

In the first procedure, the MAP sensor was mounted below the calibration nozzle
in the HTRL facility and the procedure in section 5.2 was followed through step 8. The
IR camera emissivity setting was adjusted, an image was acquired and processed, and the
target temperature was calculated from Equation 12. This procedure was followed until
the target temperature matched (approximately) the flow temperature measured by the
calibrated thermocouple. This required processing each data point prior to proceeding.

Six data points were taken with the final emissivity setting. Each thermographic image was

53

the average of 16 times to reduce the effects of noise. The results of the emittance

determination are Shown in Table 2.

Table 2. Emissivity measurement results (e=0.88).

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected

Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)

Test 1 225 250 49.6

Test2 238 264.4 50.8

Test 3 240 266.7 50.9

Test 4 225 250 49.6

Test 5 226 251.1 49.7

Test 6 223 247.8 49.5

Mean 229.5 255 50.0

Variability (95% 14.6 16.2 11.58

confidence level)

Thermocouple 49.6 :2.8°C (95%

measurement data confidence)

 

 

 

 

The SRF corrected gray scale values averaged 255 and some values exceeded the

255 limit of the 8-bit scale. Section 4.5 warned about the possibility of introducing error

by extrapolating temperatures beyond the IR temperature range setting (or gray scale

values greater than 25 5). The gray scale - temperature relationship is assumed to be a

linear relationship in this analysis. This relationship potentially becomes non-linear beyond

the IR camera range setting. Preliminary experiments showed that the temperature

calculation became less reliable if temperatures beyond the infrared camera temperature

 

range were extrapolated. That affect may be present in these validation experiments as a
decrease in the predictive accuracy and an increase in the standard deviation.
The variability presented in all of the tables in this chapter are the experimental

variance (L) defined at a 95% confidence level with

4:2 ‘2
L: ”a , (17)

 

where n is the number of data points, a' is the sample standard deviation and t is the
variable from the Student’s t statistic.

The power to the flow heater above the cah'bration nozzle was then set to two
different levels to simulate unknown operating conditions. A calibrated thermocouple was
utilized to monitor the flow conditions. The results of tests 1a and 1b are shown in Table 3

and Table 4 respectively.

55

Table 3. Test results from test 1a (MAF sensor)

(11 amps to flow heater, 42.1 - 62.10C IR temperature range).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected
Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)
Test 1 253 281.1 64.1
Test 2 211 234.4 60.4
Test 3 159 176.7 55.9
Test 4 197 218.9 59.2
Test 5 252 280 64.0
Test 6 215 238.9 60.8
Test 7 190 211.1 58.6
Test 8 209 232.2 60.2
Test 9 232 257.8 62.2
Test 10 219 243.3 61.1
Test 11 222 246.7 61.4
Test 12 250 277.8 63.8
Test 13 206 228.9 60.0
Mean 216.5 240.6 60.9
Variability (95% 52.4 58.2 15.1
confidence level)
Thermocouple 61.0 i 3.4°c (95%
measurement data confidence)

 

 

 

 

56

 

Table 4. Test results fiom test 1b (MAF Sensor)

(4 amps to flow heater, 20.0 - 400°C IR temperature range).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected
Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)
Test 1 210 233.3 38.2
Test 2 191 212.2 36.6
Test 3 210 233.3 38.2
Test 4 229 254.4 39.9
Test 5 228 253.3 49.8
Test 6 181 201.1 35.7
Test 7 201 223.3 37.4
Test 8 225 250 39.5
Test 9 231 256.7 40.0
Test 10 203 225.6 37.6
Test 11 195 216.7 36.9
Test 12 223 247.8 39.4
Test 13 185 205.6 36.1
Test 14 229 254.4 38.1
Mean 208.6 231.8 38.1
Variability (95% 33.0 36.6 :3.15
confidence level)
Thermocouple 36.6: 196°C (95%
measurement data confidence)

 

 

 

 

57

 

The second experimental verification utilized the Inframetrics 600L infrared
camera with no external optics. A metal cylinder with a 1.75 mm diameter was placed
under the calibration nozzle in the HTRL facility and the procedure outlined in section 5.2
was executed. The cylinder was positioned 34 cm in front of the infrared camera. This
created a target angle of 2.92 mrad and the resulting SRF was calculated from a curve fit

fi'om Figure 25 as 0.57.

The IR camera emissivity setting was adjusted, an image was acquired and
processed, and the target temperature was calculated fi'om Equation 12. This procedure
was followed until the target temperature matched (approximately) the flow temperature
measured by the calibrated thermocouple. This required processing each data point prior
to proceeding. Six data points were taken with the final emissivity setting. Each
thermographic image was the average of 16 frames to reduce the effects of noise. The

results of the emittance determination are Shown in Table 5.

58

Table 5. Emissivity measurement results (6:0.95).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected

Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)

Test 1 218 382.5 55.8

Test2 217 380.7 55.6

Test 3 216 379.0 55.5

Test 4 212 371.9 55.0

Test 5 216 379.0 55.5

Test 6 237 415.8 58.4

Mean 219.3 384.8 56.0

Variability (95% 17.43 30.58 12.9

confidence level)

Thermocouple 55.4 :1.7°C (95%

measurement data confidence)

 

The power to the flow heater was then adjusted to two different settings. The
system was allowed to equilibrate and images were taken of the metal cylinder. Sixteen
images were taken and averaged at every data point to reduce electronic noise. The
results of the two experiments are shown in Tables 6 and 7.

Each of the experimental results of the temperature measurement (Tables 3,4,6 and
7) includes a different number of tests. This was due in part to images that were taken but
were saturated and were discarded. The images were post processed and therefore
saturation was not detected until after the experiment was over. Also, each test series

ended when the IR camera displayed a “low liquid nitrogen” signal. This indicates that the

59

 

dewar required refilling and this effectively ended the test series because the IR camera

was closed in the HTRL flow facility.

60

Table 6. Test results fi'om test 2a

(11 amps to flow heater, 32.9 - 52.90C IR temperature range).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected
Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)
Test 1 224 393.0 63.6
Test 2 213 373.7 62.09
Test 3 216 379.0 62.5
Test 4 223 391.2 63.5
Test 5 237 415.8 65.4
Test 6 238 417.5 65.5
Test 7 232 407.0 64.7
Test 8 228 400.0 64.2
Test 9 239 419.3 65.7
Test 10 239 419.3 65.7
Test 11 245 429.8 66.5
Test 12 245 429.8 66.5
Test 13 246 431.6 66.6
Test 14 247 433.3 66.8
Test 15 245 429.8 66.5
Test 16 221 387.7 63.2
Test 17 207 363.2 61.3
Test 18 213 373.7 62.1
Test 19 224 393.0 63.6
Mean 230.6 404.6 64.5
Variability (95% 25.3 44.4 :3.7

confidence level)

 

 

Thermocouple

measurement data

 

 

 

63.5: 232°C (95%
confidence)

 

61

 

Table 7. Test results from test 2b
(6 amps to flow heater, 23.3 - 433°C IR temperature range).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRF Corrected

Measured Gray SRF Corrected Gray Temperature (0 C)
Scale Scale (Equation 12)

Test 1 172 301.8 46.9

Test 2 165 289.5 45.9

Test 3 167 293.0 46.2

Test 4 172 301.8 46.9

Test 5 172 301.8 46.9

Test 6 173 303.5 47.0

Test 7 182 319.3 48.2

Test 8 166 291.2 46.0

Test 9 168 294.7 46.3

Test 10 170 298.2 46.6

Test 11 172 301.8 46.9

Test 12 174 305.3 47.2

Test 13 176 308.8 47.4

Test 14 177 310.5 47.6

Test 15 177 310.5 47.6

Test 16 166 291.2 46.0

Test 17 183 321.0 48.4

Test 18 182 319.3 48.3

Mean 173 303.5 47.0

Variability (95% 11.0 19.4 11.6

confidence level)

Thermocouple 43.7: 16°C (95%

measurement data confidence)

 

 

 

 

62

 

5.5 DISCUSSION

The infrared measured temperatures fiom Table 2 and Table 5 agree with the
temperature measured with the calibrated thermocouple within the combined uncertainty
of the IR camera and the thermocouple. The best match attainable was with the emissivity
set on the IR camera at 0.88 and 0.95 for each specimen respectively. The emissivity
setting is adjustable in increments of 0.01 and the afl‘ect on the gray scale output is highly
non-linear (Herring [1992]). A slight change in the emissivity was noted to have a large
effect on the gray scale (and therefore temperature) output fiom the infi'ared imaging
system during the emissivity determination portion of the experiment. Measuring the
emissivity over a wide variety of temperatures could result in better measurement accuracy
for future work.

The SRF correction was critical to obtain accurate results in tests included in Table
3, Table 4, Table 6 and Table. 7. The effect of the SRF correction is more visible in test 2
(Tables 6 and 7) due to the smaller target angle (and smaller SRF value). If the gray scale

outputs in test 2 would have been utilized directly the results would have been as follows:

Test 2 (part a)
Gray Scale = 230.6 Uncorrected Temperature = 509°C
SRF Corrected Gray Scale = 404.6 Corrected Temperature = 645°C

Flow Temperature = 635°C

63

Test 2 (part b)
Gray Scale = 173 Uncorrected Temperature = 368°C
SRF Corrected Gray Scale = 303.5 Corrected Temperature = 47.00C

Flow Temperature = 437°C

The infrared camera measurements fiom part a in both tests (Table 3 and Table 6)
appear more accurate than the results from part b (Table 4 and Table 7). The infrared
measurement results in Table 7 are the least accurate. The confidence regions of the
temperature measured by the IR camera and the thermocouple do not overlap in Table 7.
This could be the result of increased experimental uncertainty fi'om the lower temperature
in test 2b decreasing the ratio of emitted radiation to reflected radiation (Signal to noise).
The inaccuracy of the SRF prediction in Table 7 may also be compounded by the
extrapolation beyond the temperature range setting on the Inframetrics 600L infrared
camera. It is also interesting to note that the 95% confidence regions on the thermocouple
and the infrared measurements are the smallest in this particular test. It may be that the
narrow confidence regions are somewhat of an anomaly and that re-testing would Show
that the confidence regions would overlap.

A third possibility is a change in the emissivity of the component. Other
researchers have noted that the apparent emissivity of semi-conductors can vary over the
temperature ranges investigated here (Okamoto and Inagaki [1994], Inagaki et a1 [1994],

Daryabeigi et a1. [1992]). Figure 30 presents results fiom Inakaki et.al [1994].

 

 

 

 

 

 

 

1.2
"'- [10030-0345
"" [13553013292
W
“’ 1.0 '
3:
15
n
‘3
1.3 0.8 6
05 ‘ ‘ ‘ ‘ ‘ ‘ 4 ‘ L

 

 

° 280 300 320 340 360 330
Radiation Temperature Ts (K)

Figure 30. Emissivity variation observed in semi-conducting materials
by Inagaki et.al., 1994.

The infrared measured temperature would appear to be insensitive to variance in

the emittance if the radiosity - black body temperature relationship,

T,T, —1— 1;,
5.01:“ ”8): ”E” . (18)

is considered, where E, is the equivalent blackbody emission and J is the radiosity

 

measured by the inflated sensor (Welch and Van Gemert [1995]). The radiosity, J, is
proportional to the fourth power of absolute temperature (the Stephan-Boltzmann law).
The blackbody emission - temperature relationship is also non-linear and for the 8-12pm
wavelength band maybe curve fit with a fifth order polynomial (Welch and Van Gemert

[1995]). The Infrarnetrics infiared camera utilizes an internal transfer fimction in the

65

hardware to manipulate this relationship to obtain temperature and electronically adjusts
for the reflected radiation. At elevated target temperatures, the ratio of the emitted
radiation to the reflected radiation (from the environment) is larger and thereby reduces
error by reducing the signal to noise ratio. A variation in the measured radiosity will be
approximately proportional to the fluctuation in the temperature, a variation in the
emissivity will affect the temperature result differently at different temperatures with the
sensitivity to emissivity variation increasing with the temperature. Valvano and Pearce
[1995] state that “. . .all of the errors in thermal imaging contribute to an under-estimation
of the true surface emitted flux, eEb, and thus an underestimate of surface temperature is
obtained”. A Simple analytical model relating the black body emission and the camera
radiosity measurement was created. A trend similar to the one presented was noted, the
computed IR output was consistently below the “true” temperature and the error was
increased at smaller temperature differences between the target and the ambient
conditions. This could be corrected by decreasing the emissivity at lower temperatures
indicating that the emissivity perceived by the IR camera decreases as the target
temperature approaches the ambient condition. However, these results presented could
not be replicated exactly.

It is also unclear if this is a change in the actual emissivity or the emissivity as
perceived by the infrared camera in specific wavelengths viewed. However, either effect
will alter the measurements.

The SRF correction does improve the accuracy of the infi'ared camera when

imaging unresolved targets. The measurement error ((1)) as defined in

66

(ISRF — new)
(T... — T...) '

was reduced by at least 50% (47.8% in Table 7) in the results presented (Tsar is the SRF

 

(19)

corrected temperature, Tm was the uncorrected IR temperature measurement). Typically
the correction adjusted the infrared measurement so that the actual temperature was
included within the measurement uncertainty range of the infrared camera. In the worse
case, the thermocouple and IR camera confidence ranges do not overlap by a gap of
01°C.

Another method to review the data is to perform a null hypothesis statistical test to
determine if the difference between the corrected value and the SRF corrected value is

statistically significant. A standard 2 test statistic (Miller and Freund [1965]) can be

defined as
z 2 3:11, (20)
a, 0'2
_ + __
"I "2

where z is the test statistic that will be compared to a Student’s t statistic to determine the
confidence level, 5? is the mean value, 0' is the standard deviation and n is the total number
of data points. The results of applying the null hypothesis test are shown in Table 8 which
also includes a summary of the true values (as measured with a calibrated thermocouple)
and the SRF corrected values.

Table 8 indicates that at the higher temperature values the SRF corrected value
and the true value are statistically indistinguishable. Test 1b with the MAF sensor almost

passed the statistical test (it failed by 0.1 and is marked No***). Test 2b with the cylinder

67

failed. It is important to note that the uncorrected value for test 2b was firrther from the

true value than the SRF corrected estimate.

Table 8. Null hypothesis statistical test results.

 

 

 

 

 

 

 

MAF Sensor

0 = 0.007 radians SRF = 0.90

e = 0.88

SRF Corrected True Number of 95% Null

Temperature Temperature Data Points Hypothesis
Test 1a 60.91 5.1 61.01 3.4 n =13 Yes
Test 1b 38.1132 36.6119 n= 14 No“

Cylinder

0 = 0.0029 radians SRF = 0.57

e = 0.95

SRF Corrected True

Temperature Temperature
Test 23 64.5 i 3.7 63.5 1 2.3 n = 19 Yes
Test 2b 47.1116 43.7116 n =18 No

 

 

 

 

 

 

In conclusion, the application of the SRF correction procedure outlined in section
5.2 can reduce thermal measurement errors by at least 50%. This improvement extends
the capabilities of the HTRL and the applications of the Inframetrics 600L infrared
camera. The error created by imaging unresolved thermal images was substantially
reduced. The application of the SRF correction does increase the experimental
uncertainty. The uncertainty in the experiments presented above was approximately :25
0C. This uncertainty can be derived from the variability of the SRF fimction or from the
variability of the particular experiment as shown in Tables 3, 4, 6 and 7 by utilizing
Equation 17. It is recommended that the specific experimental variability be applied if

possible.

68

CHAPTER 6. CONCLUSIONS

A facility was constructed to provide air flows of known velocity profiles and
temperatures with the necessary sensors to monitor these parameters. The facility was
characterized specifically to study small scale devices with the infrared imaging system in
the HTRL. A procedure to calibrate the infrared imaging system and the associated data
acquisition equipment was developed. The slit response function (SRF) for the infrared
system was experimentally measured for the basic system and with the external optics (3x
telescopic lens and the 6” close up lens). A procedure to use the SRF to correct
experimental measurements was defined along with the increase in the uncertainty in the
corrected temperature. Finally, an experimental verification of the procedure was
conducted.

The experiments contained in the investigation presented were all conducted under
the calibration nozzle with the airflow heater in place. This provided an environment with
known and controlled velocity and temperature profiles to compare the infrared
radiometer measurements against. The control of these environmental conditions were
utilized to simulate “unknown” operating conditions.

A wind tunnel facility was designed to expand the experimental capabilities of the
HTRL and to support the current research project. The facility contains a two-
dirnensional flow contraction that is being characterized currently for a separate project.
The facility also has the capability to inject steam into the flow to study the effects of
humidity in experiments. The experimental facility contains a traverse system to take
detailed spatial measurements and custom software to control the traverse system and

acquire data. The details of the system are contained in Chapter 2.

69

The infrared imaging system and the associated data acquisition system were
calibrated to obtain the maximum useful dynamic range. This procedure was outlined in
Chapter 3. The result of this calibration was the inclusion of 95% of the dynamic range
with a linear response over the entire measurement extent. These results agree very well
with the ideal system response that was defined as the goal of the cahbmtion.

The SRF function determined for the Infrarmetrics 600L infrared camera did not
agree with the manufacture’s specifications. This was determined to be the result of
delinquent calibrations and potentially contaminated lenses in the optical path. The SRF
measured was repeatable and appeared to be steady over the time flame of the work
presented here. Understanding the SRF is critical to obtaining valid thermal measurements
on objects near the limits of the IR camera resolution.

Figure 31 presents the SRF correction factor (CF) and associated uncertainty as a
fimction of target size for the Inframetrics camera with the close up optics. The SRF
correction (Figure 31) reduces the amount of error when imaging targets smaller than the
thermal measurement resolution. There is an increase in the uncertainty associated with
these measurements as the target angle is reduced. The mean temperature correction can
be substantial (changing the gray scale value by up to and exceeding a factor of 10) and
the uncertainty of this correction can be increased to as much as 25% of the correction for

target Sizes below 0.25 mm.

70

 

 

 

 

 

 

 

 

 

 

Correction Factor (CF)
and Uncertainty Specifications
12 f 30
1 A
10 ) —-CF 25
. o . .
E 8 T A /o uncertainty In CF 1 20 E
'2 e l ‘ . 15
E '| l
II.
o 4 r + 10
A A a!
2 l e 5
~11
o l____._ _-~ I;— — — * +A‘ we. .__.__‘__. ‘———— LL 0
0 0.25 0.5 0.75 1 1.25
Target width (m m)

 

Figure 3]. Correction factor (l/SRF) and % uncertainty in CF for Infiametrics 600L IR
camera with close up optics in place as a fimction of target size.

Target emissivity can be measured within each thermal measurement resolution
(TMR). The TMR for the Infiametrics camera with the close up optics is 700 am.

The experimental verification of the SRF correction (Chapter 5) shows that the
method provides a valid measurement correction and reduces the error associated with
measuring unresolved thermal targets.

The goal of the original research proposal was to apply infrared thermography to
the calibration of the MAF sensors. The MAF sensor resistance as a function of
temperature can be determined utilizing an ohm-meter and the IR camera. Furthermore,
the sensor can be heated electrically in a flow field and its temperature determined with the
infrared system This flow configuration can be utilized with a standard Nusselt

correlation (Incorpera and Dewitt [1990]) such as

hD m ;
Nuz-k—=CReDPr , (21)

71

with C and m as calibration coefficients, the Reynolds number (Re) defined in Equation 4,

the Prandtl number for air (Pr), and the heat transfer coefficient (h) defined as

I2 R
h = , (22)
A(TM4F _ Tflow)

 

where 12R is the Joule heating occurring and A is the surface area of the sensor. All of
these variables can be measured directly and the constants C and m can be obtained from
an ordinary least squares curve fit procedure. The variable In is approximately 0.5 which
will cause the velocity and the temperature measured by the infiared camera to be related
as

V at 1 2 . (23)
(TA/14F - Tflow)

 

Equation 23 can be utilized to reduce the error in the velocity calculation created by
uncertainty in the MAF sensor temperature by maximizing the (TMAF-Tnow) quantity. This
is advantageous with the results presented here because the SRF correction is more
accurate at higher temperatures as discussed previously. These results indicate that the

original goal of the project can be obtained.

72

CHAPTER 7. RECOMMENDATIONS FOR FUTURE WORK

7.1 SLIT RESPONSE FUNCTlON

The SRF determined for this camera is not within specifications for the system as
manufactured. This could be the result of optical misalignment, optical contamination or
possibly scanner calibration. Inframetrics recommends that the 600L infrared camera be
re-calibrated each year. The camera used for the experiments has not been re-calibrated
for at least 3 years. The multi-point calibration consists of viewing a blackbody at known
temperatures and adjusting the electronic processing to the correct value. The scanning
optics are also realigned and cleaned during this procedure.

Any combination of these issues can cause the SRF function to degrade. Never
the less, the results presented here can still be considered valid for the scanner in question
because it is repeatable. Studies in the future should measure the SRF function to see if it
has drifted further. If the system is re-calibrated or if the optics are cleaned, it is expected

that the SRF will improve.

7 .2 INFRARED CAMERA RESPONSE TO TEMPERATURE GRADIENTS

The SRF determines the infrared system response to a step function in
temperature. Typically a step function will only occur at the edges of the target. An
extension of the SRF work would be to quantify the maximum temperature gradient that
the infiared detector can accurately measure. This would be especially relevant in
experiments involving heat transfer such as steady state conduction with a large heat flux

(Q) through a material with a low thermal conductivity (k) as shown in

73

2.
dx

"1

Q

_. 20

k < >
Gartenberg and Roberts [1990] conducted a study in which electrically heated

wires were suspended in an air flow and imaged with an infrared system. The target was

unresolved Optically but this was corrected for with the use of an ‘apparent’ emissivity.

They found that while the IR imaging system could accurately measure the wire center

(where dT/dx =0), as the gradient increased, the measurement accuracy decreased.

7.3 HOT-WIRE SENSORS

Preliminary work was performed with 2 hot wire sensors in the flow field to
measure both temperature and velocity Simultaneously. Computer code was written to
solve the two equations iteratively for the two variables and is included in Appendix 3. It
appears that the maximum accuracy and sensitivity to both velocity and temperature can
be obtained by setting one over-heat ratio as high as possible (approximately 1.7 for
tungsten hot wires) and the other over-heat ratio reasonably low. Preliminary experiments
have shown that an overheat value of 1.3 provides heighten sensitivity to the flow
temperature.

Five micron diameter platinum wire has been ordered. This wire is more stable in
an oxygen environment and can be run at an over-heat of up to 2.2. This wire will be
evaluated for future work. It is expected that the larger difference in the over heat values
will reduce measurement errors in both temperature and velocity. The platinum wire will

also be evaluated as a cold wire sensor. This is very similar to hot wire measurement

74

methods except that a minimum electrical sensing current is used to measure the resistance
of the wire. The resistance is sensitive to temperature changes and therefore will allow the
wire to be calibrated and used to measure temperature directly. The cold wire is
insensitive to flow velocity and the output can be utilized to temperature compensate the

hot-wire output.

7.4 DATA ACQUISITION
Several improvements are scheduled for the data acquisition system. These include
obtaining two 200MHz Pentium computer configured as follows:
A. Pentium 200MHz computer with 32 Mbytes of RAM w/
a. A Keithley-MetraByte DAS-TC/b board
b. A DAS-l802 333kHz data acquisition board
B. Pentium 200MHz computer with 32 Mbytes of RAM w/
a. Image-Pro Plus v3.0.1 for Windows 95 software

b. Data Translations DT3152 Frame Grabber board

The Pentium computer with the DAS-TC and the DAS-l802 board will greatly
enhance the data acquisition capabilities and the speed at which data can be obtained. A
single DAS-TC board is being used for all thermocouple and hot wire data captures. This
board limits the data acquisition rate to 25Hz on each channel. This is much too slow to
capture any flow features aside from the mean velocity. At this rate, a data file of 1000
data points from a single hot wire and a single thermocouple requires 40 seconds. This is

the equivalent to 20,000 slot widths of flow (10 m/s velocity with an exit width of 2 cm),

75

essentially, 1 data point every 20 Slot widths of flow. This is much too slow to capture
transient events. It is also possible that periodic events in the flow (such as vortex
shedding) could suffer from measurement aliasing and not be captured or understood from
the data obtained.

The DAS-TC/b board will be utilized for thermocouple and humidity sensor
measurements. These sensors have relatively slow response times due to the nature or
mass of the sensor (the bead Size of the thermocouple or the capacitor in the humidity
probe). The DAS-1802 board will be used strictly to control the stepper motor motions
and to read the data from the hot wires and the pressure transducer. The DAS-1802 with
3 channels active is capable of acquiring data at 100 kHz per channel. This will allow 200
data points per slot width of flow.

The increase in data rate as well as the increase in memory will also help to insure
data is convergence. The current system is limited to 2 data channels with 4000 points
each. This results in a cumulative average shown in Figure 31. The cumulative average

appears to be converged within 4000 data points.

76

Cumulative Data Averages

 

41.- _ -..- - _ .~ .__L%L__.._._._____.__-L,__.H-fi_ -— -- .—-_. w .._ ..-1_3
4.08»
(>12
4.064»
>11
4.04 l l
H A
4.02 I
E
v
. >
3.9 . A 0.5

 

 

 

0 500 1000 1 500 2000 2500 3012!) 3500 4000
Data Point:

Figure 32. Cumulative average of the hot-wire data.

A method for quantifying the degree of convergence is presented by Mandel

[1964] as shown in

 

(23)

This equation uses the Student t statistic (t) and the estimated standard deviation (0') with
the number of degrees of freedom (data points) (11) to estimate the uncertainty (L) in the
mean value. The confidence level is obtained from the Student t statistic. The distribution
of the data is assumed to be gaussian and n was always greater than 1000 data points.

The data in Figure 31 has a mean velocity of 1.21 m/s with a standard deviation of
0.718 m/s. When Equation 23 is applied to the data utilized in Figure 31, the estimated

uncertainty in the mean is 0.045 m/s at a 95% confidence level. If a 99% confidence level

77

is desired, the uncertainty becomes 0.058 m/s. Typically in this study, 2000 data points
were utilized for hot-wire and thermocouple measurements. This results in an uncertainty
of 0.063 m/s with a 95% confidence for the hot-wire data shown in Figure 31.

This information also indicates the need for the new data acquisition boards. The
equipment utilized for this study acquired data at 25 Hz, thus, it took 40 seconds to obtain
1000 data points. This is much too long for transient flow tests and made steady state
tests take much longer than necessary. The longer each experiment was run, the greater
the drift in the ambient temperature, the greater the drift in the hot-wire calibration and the

greater the chance that a particle would ruin the hot-wire and the entire data set.

78

APPENDICES

79

APPENDIX 1. TWO-DMENSIONAL FLOW CONTRACI'ION

The design for the two-dimensional contraction was taken from Morrel [1977].
The method designs the contraction by using two cubic functions and matching the slopes
so that the flow experiences a favorable pressure gradient to prevent separation along the
contraction surfaces and a gradual smooth change in the derivative of the profile shape.

The design process outlined by Morrel is iterative. A program was written in
Mathematicao to simplify the design process and to plot the resultant profile. The
Mathematicao program is included at the end of this appendix.

The contraction can be seen in Figure 33 and the manufacturing coordinates are
detailed in Table 9. The contraction was manufactured from '/4” thick Lexan° by
machining the necessary profile in a suction mold. The Lexan was heated until it became
pliable and then was pulled into its final shape by the vacuum applied. The machining and

forming was completed by Prodigy Machine and Gage, Lansing Michigan.

 

Figure 33. Two-dimensional flow contraction.

80

Table 9. Two dimensional contraction machining coordinates.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mmeters) y (meters) x (meters) y (meters)
0 0.1905 0.115 0.12951
0.005 0.190495 0.12 0.121203
0.01 0.19046 0.125 0.112175
0.015 0.190365 0.13 0.102395
0.02 0.190179 ._ 0.1335 0.095086
0.025 0.189873 0.135 0.091924
0.03 0.189417 0.14 0.082231
0.035 0.188781 0.145 0.07378
0.04 0.187933 0.15 0.066484
0.045 0.186846 0.155 0.060259
0.05 0.185487 0.16 0.05502
0.055 0.183828 0.165 0.050682
0.06 0.181838 0.17 0.047159
0.065 0.179487 0.175 0.044368
0.07 0.176745 0.18 0.042222
0.075 0.173582 0.185 0.040637
0.08 0.169968 0.19 0.039529
0.085 0.165872 0.195 0.038811
0.09 0.161265 0.2 0.038399
0.095 0.156117 0.205 0.038208
0.1 0.150398 0.21 0.038153
0.105 0.144077 0.215 0.038149
0.11 0.137124

 

 

 

 

 

 

The next two pages contain the Mathematica program written to Show the profile

and to aid the iterative process.

81

Clear-[81, HZ, 3:, Ba, Ba, EX, L]

 

 

HI: .1905;
2

.2. £92,
2

L= 0.213;

Ex: 0.627;

Ha[x] 82 (El 82) 1 (if;
= + - - — ,‘

" n2

(HI-112) (1- !)3
m3[x]=82+ ,-_ v—J'
- (1-Ex)2

 

I

blag] = Oxfialx] ;

fiblxj = axmlxl ;

Plan: Plot[Ha[x] , {3, O, .1336} , PM» "Cartractim Profile",
Dim-t Identity] ;

Plath: Plot-483m] , (x, .1336, 0.213} , W" Ichttitzy];

Plate: Plouhixl , {3, 0, .1336} , We 1&1:in ;

P191311: Flanagan] , (x, .1336, 0.213}, W» MW] ;

Show[Plota, Plath, OW» Wm, We {0, O} ,

Erase-b Tun, We Autumtic, Flow» {{0, .225}, {0, .225}} ,

W110» 1 i , W» (meters, matersfl

82

Output Graphics

Contraction Profile

 

 

 

 

 

 

 

 

 

 

 

 

l
0'21
l
0.15
m .
Ll
(l)
.1.)
2 0.1 _____
\.
0.05 V
. \ 
l
O . a MA . 1
0.05 0.1 0.15
meters

83

APPENDIX 2. QUICKBASIC© PROGRAM TO PERFORM DATA ACQUISITION
AND STEPPER MOTOR CONTROL

The following is a quick basic program developed and written by Dr. John
McGrath and Paul Hoke to operate the data acquisition and stepper motor boards. This
program allows for the stepper motor to move to position in increments of 6.602 pm (this
is determined by the motor and mechanical translation device). The stepper motor control
consists of a CMD-50 chopper drive and a VCO-4050 controller board fiom American
Precision Industries. The stepper motors are controlled fi‘om the data acquisition board
inputs and outputs. The data acquisition card is a DAS-TC utilized in conjunction with a
DAS-02 card from Keithly-MetraByte. These boards collect 16 channels of input and are

capable of driving 2 channels of analog output.

84

'itttttttit*tt*titttttt********#*##3##titttttt*ttttttt******t***t*t**ttt

*****

REM This is the program recalb.bas

REM WEDNESDAY, Dec 17, 1997: 12:10 PM

' Include the supplied Q41FACE.BI file. This file

' contains all function DECLARAtion supported by the driver for the DASTC board.

' $INCLUDE: 'Q41FACE.BI'

 

' Define ALL local variables required by the program here. NOTE
' that you must avoid declaring and using QuickBASIC variables on the
' fly.

DIM NumOfBoards AS INTEGER

DIM DERR AS INTEGER

DIM DEVHANDLE AS LONG

DIM BoardNum AS INTEGER

DIM CJC AS SINGLE

DIM CHAN AS INTEGER

DIM GGAIN AS INTEGER

DIM ADDATA AS LONG

DIM TEMPGBV AS LONG

DIM TEMPGB AS SINGLE

85

DIM TEMPINV AS LONG
DIM TEMPIN AS SINGLE
DIM TEMPOUTV As LONG
DIM TEMPOUT AS SINGLE
DIM TEMPSURFV AS LONG
DIM TEMPSURF As SINGLE
DIM RELHUMV As LONG
DIM RELHUM AS SINGLE
DIM PDIFFVl AS LONG
DIM PDIFFV AS SINGLE
DIM HWVII AS LONG
DIM va1 As SINGLE
DIM va21 AS LONG
DIM va2 AS SINGLE
DIM de AS LONG

DIM dp AS SINGLE

DIM th AS LONG

DIM J AS LONG

DIM NUMBER AS LONG
DIM SUMI As SINGLE
DIM SUM2 As SINGLE

DIM SUM3 AS SINGLE

86

DIM SUM4 AS SINGLE

DIM HWlAVEL AS SINGLE

DIM HWZAVEL AS SINGLE

DIM dpavg AS SINGLE

DIM HWV1(2000)

DIM HWV2(2000)

DIM dp(2000)

DIM tcfl(2000)

DIM VINVISCID AS SINGLE

DIM DELP AS SINGLE

'OPEN FILES FOR WRITING HOTWIRE DATA
'OPEN "HW1.DAT" FOR OUTPUT AS #1
'OPEN "HW2.DAT" FOR OUTPUT AS #2
'OPEN "dp.dat" for OUTPUT as #3

CLS

 

' This step initializes the internal data tables according to the

' information contained in the configuration file DASTC.CFG.
PRINT "Initializing the DASTC board - - - PLEASE wait"
a8 = "DASTC.CFG" + CHR$(0)
DERR = DASTCDEVOPEN%(SSEGADD(a$), NumOfBoards)

IF DERR <> 0 THEN

87

BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING '..DEVOPEN'"
STOP
END IF
BoardNum = 0
DERR = DASTCGETDEVHANDLE%(BoardNum, DEVHANDLE)
IF (DERR <> 0) THEN BEEP: PRINT "ERROR, DEVICE HANDLE IS null. . .":
STOP
CLS
REM This part of the program is using the DAC-02 board to control sensor motion
REM Set base address at 784 for DAC-02 board to NOT conflict w/ DAS-TC board
BaseAddr% = 784
REM D/A Output Channels are defined
REM Channe1#0 will be used for MOTION: move (>25 v; OutData% = 4095)
REM or stop (<2.5 v; OutData% = 0 )
REM Channe1#l will be used for DIRECTION: up/down or in/out for vertical and lead
screw assemblies
REM Set initial voltages on D/A channels to 0. This Means that ChannelO volts = 0 ->
motor stopped
REM Output 0 volts on Channel#1 (DIRECTION); Ch$l first because of prob detected
Ch1% = 1

OutData% = 0

88

HByte% = INT(OutData%/ 16) 'Convert to high byte
LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte
lowportl = BaseAddr% + 2 * Ch1%

OUT lowport, LByte% 'Write low byte

hiportl = BaseAddr% + 2 * Ch1% + 1

OUT hiportl , HByte% 'Write high byte

REM Output 0 volts on Channel#0 (MOTION)

Ch1% = 0

OutData% = 0

HByte% = INT(OutData%/ 16) 'Convert to high byte
LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte
lowportO = BaseAddr% + 2 * Ch1%

OUT lowportO, LByte% 'Write low byte

hiportO == BaseAddr% + 2 * Ch1% + 1

OUT hiportO, HByte% 'Write high byte

'This segment inputs parameters needed for calculating velocity and local
' temperature in the boundary layer

'Clear screen

CLS

' REQUEST AN INITIAL POSTION DEFINITION FROM USER IN CASE
DIFERENT THAN ZERO

LOCATE l, 3: INPUT "Define the initial location in the boundary layer (mm): ",

position

89

THEBEGIN:

'REM Set initial position equal to zero

'position = 0

REM Set initial count equal to zero

count = 0

DIRECTION:

REM Decide which direction to move the motor and use Channel#l to effect decision
REM Set channel number to zero

LOCATE 2, 3: PRINT "Positive direction (+y) is AWAY from the wall"
LOCATE 3, 3: INPUT "(D/A Chl) To move HOTWWIRE: TO WALL- Enter 5; AWAY
FROM WALL- Enter 0 ", decide%

IF decide% < 2.5 GOTO OUTDOWN

' LET SIGN : -1 FOR MOTION TO THE WALL

SIGN = -1

REM Output 5 volts on Channel#1 to move IN or UP

Chl% = 1

OutData% = 4095

HByte% = INT(OutData% / 16) 'Convert to high byte
LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte
OUT BaseAddr% + 2 * Chl%, LByte% 'Write low byte

OUT BaseAddr% + 2 * Chl% + 1, HByte% 'Write high byte

GOTO END.DIRECTION

90

OUTDOWN:

REM Output 0 volts on Channel#1 to move OUT or DOWN

' LET SIGN = +1 FOR MOTION AWAY FROM THE WALL

SIGN = 1

Chl% = 1

OutData% = 0

HByte% = lNT(OutData%/ 16) 'Convert to high byte

LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte

OUT BaseAddr% + 2 * Chl%, LByte% ‘Write low byte

OUT BaseAddr° o + 2 * Chl% + 1, HByte% 'Write high byte
END.DIRECTION:

REM Define distance(s) to be moved (This will require motor-specific calibrations later)
LOCATE 4, 3: INPUT "How far (in mm) should the motor move? ", DISTANCE
count.desired = DISTANCE * 151.352

REM Decide whether or not to move the motor and use Channel#0 to effect decision
REM Set channel number to zero

LOCATE 5, 3: INPUT "Do you wish to move the motor (D/A Ch 0)? (5: move; 0 :
stop) "; decide%

IF decide% < 2.5 GOTO DONTMOVE

MOVE:

REM output 5 volts on Channel#0 to move motor

Chl%=0

91

OutData% = 4095

HByte% = INT(OutData%/ 16) 'Convert to high byte
LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte
OUT BaseAddr% + 2 * Chl%, LByte% 'Write low byte

OUT BaseAddr% + 2 * Chl% + 1, HByte% 'Write high byte
GOTO PULSECOUNT

DONTMOVE:

REM Output 0 volts on C hannel#0 to stop motor

Chl% = 0

OutData% = 0

HByte% = INT(OutData%/ 16) 'Convert to high byte
LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte
OUT BaseAddr% + 2 * Chl%, LByte% 'Write low byte

OUT BaseAddr% + 2 * Chl% + 1, HByte% 'Write high byte
GOTO IN F OOUT

PULSECOUNT:

REM This part of the program counts pulses from an MD channel
REM The DAC-02 board can not be used for this, so the DASTC board will be used.
REM Use Channel#8 for pulse counting
REPEAT:

'get an initial voltage for edge detection in a loop

'The next step gets an initial voltage reading on the pulse counting channel

92

(HLMN=8
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, ADDATA)
IFDERR<>0THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
ENDIF
' Calculate the effective voltage.
CALCULATEVOLTAGEI :
'The next step calculates a voltage from the ADDATA value measured
'Call this first voltage V1
Vl = ADDATA / (1000000)
CHECK:
'this is the beginning of a loop which checks whether successive voltages are different
(>2.5 v) than initial 'voltage
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, ADDATA)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRea '"

STOP

93

END IF
' Calculate the effective voltage.
CALCULATEVOLTAGEZ:
“The next step calculates a voltage from the second ADDATA value measured
'Call this first voltage V2
V2 = ADDATA / (1000000)
'Calculate the difference between the an initial voltage and all successive voltages in loop
DIFF = ABS(V1 - V2)
'If a 0 to 5 volt or a 5 to 0 volt change is not detected on channel 8, then continue to take
measurements,
'Other wise increment a counter because an edge has been detected.
IF DIFF < 2.5 GOTO CHECK
count = count + 1
IF count > count.desired GOTO the.end
GOTO REPEAT
the.end:
This stops the motor
Chl% = 0
REM Output 0 volts on Channel#0 to stop motor
OutData° o = 0
HByte% = INT(OutData%/ 16) 'Convert to high byte

LByte% = 16 * (OutData% - HByte% * 16) 'Convert to low byte

94

OUT BaseAddr% + 2 * Chl%, LByte% 'Write low byte

OUT BaseAddr% + 2 * Chl% + 1, HByte% 'Write high byte

' CLEAR SCREEN BEFORE WRITING MOTION INFORMATION TO THE SCREEN
INFOOUT:

CLS

calibration = count / 151.352

position = position + SIGN * calibration

LOCATE 10, 3: PRINT "Hotwire has moved "; calibration * SIGN; "mm"
LOCATE 11, 3: PRINT "Hotwire is now at "; position; "mm"

LOCATE 12, 3: PRINT "Where +x is away from the wall"

DUMMY = 0

LOCATE 15, 3: PRINT "Move again OR set zero position?"

LOCATE 16, 3: INPUT " -1 to quit; I to set current position; 0 to move again; 2 TO
TAKE DATA: ", DUMMY

IF DUMMY = -1 GOTO FINISH

IF DUMMY = 1 GOTO ZERO

IF DUMMY = 2 GOTO DATALOOP

GOTO THE.BEGIN:

ZERO:

POSITION = 0

count = 0

GOTO THE.BND

95

FINISH:

END

CLS

'WRITE TO SCREEN THE CHOICES MADE FOR DIRECTION AND MOTION
'LOCATE 1, 1: PRINT "YOU CHOSE TO MOVE (UP/DOWN, ETC)"

'LOCATE 2, 1: PRINT "YOUR STARTING STREAMWISE POSITION (X/W) WAS:"
'LOCATE 3, 12 PRINT "YOUR STARTING BOUNDARY LAYER POSITION (Y IN
MM) WAS:"

'LOCATE 4, 1: PRINT "INITIATE LOOPING TO UPDATE EXPERIMENTAL
DATA"

DATALOOP:

CLS

LOCATE 1, 3: PRINT "The Hotwire is now at y= "; position; "mm"

LOCATE 2, 3: PRINT "Where +y is away from the wall"

LOCATE 3, 3: PRINT "The Hotwire is now at X: "; position; "mm"

LOCATE 4, 3: PRINT "Where +x is DOWN- in the F low Direction"

LOCATE 6, 3: PRINT "Current data are being displayed for checkirng"

DO

CHECKS = INKEYS

'This section is for measuring experimental variables

'Measure ambient Cl C temperature

BoardNum = O

96

DERR = DASTCGETCIC%(BoardNum, CJC)
IF (DERR <> 0) THEN BEEP: PRINT "ERROR GETTING CJC TEMPERATURE...":
STOP
LOCATE 8, 5: PRINT USING "AMBIENT CJC TEMP(C)= ###.##"; CJC
'Measure TEMPERATURE IN GIELDA BOX, TEMPGB
CHAN = 7
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, TEMPGBV)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
TEMPGB = ((TEMPGBV/ 1000000) - 1) * 25
LOCATE 9, 5: PRINT USING "TEMPERATURE IN GIELDA BOX (C) = ###.##";
TEMPGB
Measure PLATE INLET TEMPERATURE, TEMPIN
CHAN = 5
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, TEMPINV)
IF DERR <> 0 THEN

BEEP

97

PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
TEMPIN = TEMPINV / 100
LOCATE 10, 5: PRINT USING "PLATE INLET TEMPERATURE (C)= ###.##";
TEIVIPIN
'Measure PLATE OUTLET TEMPERATURE, TEMPOUT
CHAN = 6
GGAIN = l
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, TEMPOUTV)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
TEMPOUT = TEMPOUTV / 100
LOCATE ll, 5: PRINT USING "PLATE OUTLET TEMPERATURE (C)= ###.##";
TEMPOUT
'Measure PLATE SURFACE TEMPERATURE, TEMPSURF
CHAN = 2
GGAIN = 1

DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, TEMPSURFV)

98

IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
TEMPSURF = TEMPSURFV/ 100
LOCATE 12, 5: PRINT USING "PLATE SURFACE TEMPERATURE (C)= ###.##";
TEMPSURF
'Measure ambient relative humidity
CHAN = 9
GGAIN = l
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, RELHUMV)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR"; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
RELHUM = ((RELHUMV / 1000000) - 1) / .04
LOCATE 13, 5: PRINT USING "RELATIVE HUMIDITY (%)= ##1##"; RELHUM
'Measure Pressure DIFFERENCE, PDIF F
CHAN = 12

GGAIN = l

99

DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, PDIFFVl)
IF DERR <> 0 THEN

BEEP

PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"

STOP

END IF

PDIFFV = PDIFFVl / 1000000
LOCATE 14, 5: PRINT USING "PRESSURE TRANSDUCER VOLTAGE (V OLTS)=
##.##"; PDIFFV
DELP = PDIFFV * .1
'CHECK FOR NEGATIVE VOLTAGES
IF DELP > 0 GOTO OKSTATE
DELP = 0
OKSTATE:
VINVISCID = SQR((2 / 1.204) * (DELP) * (24884))
LOCATE 15, 5: PRINT USING "IMPLIES MAX INVISCID JET EXIT PLANE
VELOCITY (In/S): ##1## "; VINVISCID
'IVIEASURE CHANNEL 1 HOTWIRE VOLTAGE, HWVl
CHAN = 11
GGAIN = l
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, HWY] 1)

IF DERR <> 0 THEN

100

BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
HWVl =HWV11 / 1000000
E1 = HWVl
LOCATE 16, 5: PRINT USING "CHANNEL 1 HOTWIRE VOLTAGE (VOLTS)=
###.##";HWV1
'MEASURE CHANNEL 2 HOTWIRE VOLTAGE, HWV2
CHAN = 10
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, HWV21)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
HWV2 = HWV2] / 1000000
E2 = HWV2
LOCATE 17, 5: PRINT USING "CHANNEL 2 HOTWIRE VOLTAGE (VOLTS)=
###.##"; HWV2;
LOCATE 18, 3: PRINT "PRESS ANY KEY TO INITIATE HOTWIRE DATA

ACQUISITION"

101

 

 

GOTO KEYCHECK

KEYCHECK:

LOOP WHILE CHECK$ = 1NKEY$ 'CHECK FOR KEYSTROKE
OOPS:

LOCATE 20, 5: INPUT "ENTER THE NUMBER OF VELOCITY SAMPLES To BE
TAKEN"; NUMBER

IF NUMBER < 2 GOTO OOPS

'READ HOTWIRE VOLTAGES

'INITIALIZE SUMS To ZERO FOR AVERAGES

SUMI = 0

SUM2 = 0

SUM3 = 0

SUM4 = 0

HWV1(0) = 0

HWV2(0) = 0

dp(O) = 0

tcfl(0) = 0

FOR J = 1 To NUMBER

'MEASURE CHANNEL 1 HOTWIRE VOLTAGE ARRAY, HWVA1(J)
CHAN = 11

GGAIN = 1

DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, HWVl 1)

102

IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
HWV1(J) -—- HWVII / 1000000
SUM1= SUM1+ HWV1(J)
'MEASURE CHANNEL 2 HOTWIRE VOLTAGE, HWV2
CHAN = 10
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, HWV21)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR"; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
HWV2(I) = HWV2] / 1000000
SUM2 = SUM2 + HWV2(J)
'MEASURE CHANNEL 2 HOTWIRE VOLTAGE, HWV2
CHAN = 12
GGAIN = 1

DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, dpl)

103

IF DERR <> 0 THEN
BEEP
PRINT "ERROR"; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
dp(J) -—— de / 1000000
SUM3 = SUM3 + dp(J)
CHAN = 6
GGAIN = 1
DERR = KADRead%(DEVHANDLE, CHAN, GGAIN, :61)
IF DERR <> 0 THEN
BEEP
PRINT "ERROR "; HEX$(DERR); " OCCURRED DURING 'KADRead'"
STOP
END IF
160(1) = 1C1 / 100
SUM4 = SUM4 + tCflU)
NEXT
'CALCULATE THE AVERAGES
HWlAVEL = SUMI /NUMBER
HW2AVEL = SUM2 /NUMBER

dpavg = SUM3 / NUMBER

104

h20 = dpavg * .1

tcavg = SUM4 / NUMBER

'CALCULATE THE STANDARD DEVIATIONS
hwlst = 0

hw2st = 0

dpst = 0

tcst = 0

FOR J = 1 TO NUMBER

hwlst = hwlst + ((HWlAVEL - HWV1(J)) A 2) A (.5)
hw2st = hw25t + ((HW2AVEL - HWV2(J)) A 2) A (.5)
dpst = dpst + ((dpavg - dp(J)) A 2) A (.5)

tcst = tcst + ((tcavg - tcfl(J)) A 2) A (.5)

NEXT

hwlstdev = hwlst / (NUMBER - 1)

hw2stdev = hw25t / (NUMBER - 1)

dpstdev = dpst * .1 / (NUMBER - 1)

tcstdev = tcst / (NUMBER - 1)

CLS

LOCATE 11, 5: PRINT USING "CHANNEL 1 HOTWIRE AVERAGE VOLTAGE
(VOLTS)= ##.###"; HWlAVEL

LOCATE 12, 5: PRINT USING "ST.DEV of Hot Wirel reading (VOLTS) = #.###";

hwl stdev

105

LOCATE 13, 5: PRINT USING "CHANNEL 2 HOTWIRE AVERAGE VOLTAGE
(VOLTS)= ##.###"; HW2AVEL

LOCATE l4, 5: PRINT USING "ST.DEV of Hot Wire2 reading (V OLTS) = #.###";
hw2stdev

LOCATE 15, 5: PRINT USING "Pressure difference (in/I120): #.###"; h20
LOCATE l6, 5: PRINT USING "ST.DEV of Pressure reading (in/HZO) = #.###";
dpstdev

LOCATE 17, 5: PRINT USING "Temperature of Flow = ##.#"; tcavg

LOCATE 18, 5: PRINT USING "ST.DEV of Temperature = ##.##"; tcstdev
'WRITE THE DATA ARRAYS TO TWO FILES: HW1.DAT AND HW2.DAT
'FOR J = 1 TO NUMBER

'WRITE #1, J, HWV1(J)

'WRITE #2, J, HWV2(J)

'WRITE #3, J, dp(l)

NEXT

'CLOSE THE DATA FILES

'CLOSE #1

'CLOSE #2

'CLOSE #3

LOCATE 24, 3: PRINT "Take another calibration point?"

LOCATE 25, 3: INPUT " -1 to quit; I to set current position; 0 to move again; 2 TO

TAKE DATA: ", DUMMY

106

CLS

IF DUMMY = -1 GOTO FINISH

IF DUMMY = 1 GOTO ZERO

IF DUMMY = 2 GOTO DATALOOP

GOTO THE.BEGIN:

107

APPENDIX 3. FLAT PLATE ISOTHERMAL VELOCITY PROFILES

The velocity profiles taken of the flow below the two dimensional contraction are
presented here for completeness. The profiles are not non-dimensionalized in the common
format of the velocity divided by the inviscid velocity because at the time of the
experiments, there was no pressure tap in the contraction (this has been added since).

These results show the resolution of the stepper motor system combined with the
hot wire probes. They also Show the flow development below the contraction in a limited
fashion. Figure 34 shows all of the data obtained. Figure 35, Figure 36, and Figure 37
Show the velocity profiles at each specific X/W location.

The affect of the number of data points can be seen in Figure 34, Figure 36 and
Figure 37. The profiles taken with only 200 data points tend to appear jagged, an
indication that the measurements at each point were not converged. The profiles with
2000 data points tend to appear smoother and also appear to define a mean value curve
among the scatter shown in the 200 data point curves.

The profiles at x/w = 1.59 show the inviscid core region of the wall jet as the flat
region where V/V max = 1. The outer sheared layer and the inner boundary layer progress
toward the center of the wall jet as the flow advances in the streamwise (x) direction. This
can be seen as the slope becomes more less steep in the outer region and as the flat

inviscid central core region disappears.

108

 

 

Compiled Profiles (Isothermal)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.2
+xM=4.52
—x—xlw=1.59
+xlw=4.52
+xlw=10.95
+xIw-10.95
x —.—x1w==1.59
g —-— xlw =1o.95 (2000)
2 +xlw=4.52 (2000)
> —I—xM=10.95(2000)
2.5 3
Figure 34. Compiled jet velocity profiles (w=2 cm).
xIw=1.59
1 4 .4...“
0.9 4»
0.8 4»
0.7 4 \
5 0.6 i
g 0.5 \
> 0.4 \
0.3 \,
0.2 \‘
x
0.1 - ‘\
‘.‘V’V‘Vv 7V..- Vita». .-
o : -— -- — — -— —— — 4
o o 5 1 1 5 2 2.5

ylw

 

Figure 35. W = 1.59 velocity profiles (w=2 cm).

109

 

 

 

 

 

 

     
 

 

 

 

 

 

 

 

 
 
   
  

 
  
  

 

 

 

xlw=4.52
1
0.9 ,7 l
—1—2oo data points i
0.8 -9— 200 data mine
0 7 +200 data po'nts
‘ +2000 data po'nts
§ 0.6 4
g 0.5 l
> 0.4 4
I
0.3 l
0.2 l
0.1
0 4 4
0 0 5 1 1.5 2 2 5 3
ylw
Figure 36. W = 4.52 velocity profiles (w=2 cm).
xlw=1 0 .9 5
1
l
0.9 ‘
. -o— 2000 data pohts
0.8 +200 data pohts
—-—200 data pohts
0‘7 + 200 data pohts
x 0.6 :-I—2000 data pohts
a A ,,
g 0.5
> 0.4
0.3
0.2
0.1
O .7 a 1
o o 5 1 1.5 2 2 5 3
ylw

 

 

Figure 37. W = 10.95 velocity profiles (w=2 cm).

110

APPENDIX 4. HOT-WIRE CONSTRUCTION AND CALIBRATION

The hot-wires used in this study were constructed of 5 micron diameter tungsten
wire and are schematically shown in Figure 38. The hot-wires were fabricated in the

Turbulent Shear Fluids Laboratory (TSFL) at Michigan State University.

Probe body

Broaches

  
 

_[lmm

. |._ 4T1.--

Figure 38. Schematic of hot-wire probe.

The construction procedure, while simple in theory, was very diflicult to
implement due to the small size of the tImgsten wire. The construction procedure is
outlined below.

1. The broaches were soaked in flux for 30 seconds. The broaches were dipped

slowly into and out of a solder pot to provide a uniform layer of solder.

2. The flux was cleaned off using a flux remover in an ultrasonic bath.

3. A 5 micron tungsten wire was strung across a steel “C” bracket and held in

position by adhesive tape.

4. The span was placed in a device so it could be lowered into a copper oxide

solution holder to be coated. The span was positioned so that the two regions

to be plated were separated by approximately 1 mm (Figure 39).

Ill

 

Figure 39. Tungsten wire copper plating station.

6. Power was applied so that approximately 0.8 mA flowed through the device
for 8 minutes.
7. The tungsten wire was then soldered across the broaches (Figure 38) using the

special solder station to align them properly (Figure 40).

Step seven is by far the most difficult and time consuming. A typical first attempt
at manufacturing a hot-wire might take 3 hours. A high level of manual dexterity and
practice are required to produce hot-wires with any spwd. It is suggested that the entire
process could be greatly simplified if the need for human control of the soldering process

was eliminated.

112

 

Figure 40. Precision solder station.

Hot-wire calibration requires a well-characterized flow with a known velocity
distribution to establish the relationship between the hot—wire output voltage and velocity.
A hot-wire calibration system (Figure 41) fi'om the Turbulent Shear Fluids Laboratory
(TSFL) was utilized to obtain the first cahbration data. This hot-wire cah'brator was
designed to produce a uniform flow of known velocity (measured with a pressure

transducer) with very low turbulence intensity.

113

 

Figure 41. TSFL hot-wire calibration flow contraction.

The hot-wire was taken to the HTRL Flow Facility and mounted beneath the
“coke bottle” contraction (Figure 42). This contraction has been well characterized in the
TSFL to have a discharge coefiicient (Cd) of 0.98 to 1. Assuming that the fluid is
incompressible and inviscid, the velocity along a streamline can be calculated from

(Blevins [1984])

 

 

(22)

114

 

Figure 42. “Coke Bottle” contraction utilized for hot-wire calibration.

Calculating the ( 1-(A2/A1)2)'1 term utilizing the venturi dimensions results in a
value of 1.00196. Therefore, it incorporates little error to simplify Equation 22 to
I7 = {fl} , (23)
p
which is the Bernoulli equation. Furthermore, if area A1 is considered to be a large
hemisphere (control surface) above the nozzle, V12 quickly approaches zero.

A pressure transducer was utilized to measure the difference in static pressure
across the contraction, thus allowing the exit velocity to be calculated by applying
Equation 23. The density of air was calculated from the ideal gas law.

A typical hot-wire calibration performd in the TSFL contraction resulted in the
calibration curve shown in Figure 43. The same hot-wire was then taken to the HTRL

Flow facility and calibrated using the same pressure transducer. This resulted in a

115

calibration curve as shown in Figure 44. The data were curve-fit using TableCurve2D v3

software (AISN [1994]) and the relationship
E=¢A+BV". (24)

The coefficients and correlation factors were then used to compare the HTRL
facility calibration curves and the calibrations obtained in the TSFL contraction. Typical
cahbration curves for the TSFL calibrator and the HTRL facility are shown in Figure 43
and Figure 44 respectively. There was less than a 1% difference between the TSFL
calibration coefficients and the HTRL facility coefficients. The conclusion was drawn that
the HTRL facility could be used for future hot-wire calibration. A pitot tube was later
used to verify that the flow velocity below the calibration nozzle agreed with the Bernoulli

calculation.

116

Hot Wire
Voltage

9l10197 Hot wire calibration in TSFL contraction (test 3)

r2=0.99989089 DF Ad] 12:0.99985452 FIISIGEII=ODO23138116 Fstat=45820.757

 

 

 

 

3:8.3199168 b=2.56?3594
0:0.47403183
4.2
4 .1
4
3.9 f
3.8 /

 

 

3.7 /
3.6

 

/

 

 

 

 

 

 

 

3 .5
3 .4
2.5 5 7.5 10 12.5
Velocity (m/s)

Figure 43. TSFL contraction hot-wire calibration curve.

117

 

15

9118437 Hot wire calibration in Ford Facility (test 1)
rh099950373 DF Adi r2=0.99937966 FltStdErr=0.0075817315 Fstat=13091.133

 

 

 

0:9.2139295 1329585184
c=0.44175237
4 .5 /
4 .25
Hot Wire
voltage
4 1’

/

 

3.75 //
3.5 /
3.25

0 5 10 1 5 20 25
Velocity (m/s)

 

 

 

 

 

 

 

 

Figure 44. HTRL facility hot-wire calibration curve.

Hot-wire calibrations drift with time due to several factors such as oxidation or
stretching or damage of the ttmgsten filament due to impacts with particles. It is
absolutely necessary to calibrate the hot-wire prior to its use. Post calibration is desirable
as well if the hot-wire survived the experiment. The two cah'bration data sets define a drift
that the hot-wire experienced over time. It is necessary to perform this calibration every
time a hot-wire is used and after approximately 3 hours of operation (due to oxidation) in
order to insure that accurate measurements can be taken. The data set fiom the first
calibration is used to define the operating parameters and the post calibration data is

strictly used to define the amount of deviation that has occurred relative to the operating

118

parameters used for the experiment. These two calibration data sets define the maximum
accuracy of the velocity measurement possible with the hot-wire for that experiment.

For example, Figure 44 shows the calibration curve for a hot-wire used in an
experiment. The same hot-wire was re-calibrated after two and a halfhours with the
resultant data shown in Table 10. These hot-wire voltage data and pressure (velocity)
measurements from the post-cah'bration are then used with the curve-fit data from the pre-

cahbration to define an error or drift quantity.

Table 10. Post cahbration data

 

 

 

Velocity
Velocity (m/s, via Hot-wire predicted by % Error
Equation 23) Po stcalibration Precalibration
Output (V) curve (m/s)

1.89 3.5 2.01 6.4
3.09 3.63 3.22 4.1
4.09 3.7 4.03 1.4
4.76 3.76 4.81 1.1
6.04 3.85 6.16 2.0
9.56 4.026 9.49 0.7
11.38 4.11 11.42 0.4
13.85 4.19 13.5 2.6
14.44 4.218 14.29 1.1
16.63 4.28 16.13 3.1
17.37 4.313 17.17 1.1
18.17 4.345 18.22 0.3
19.26 4.371 19.11 0.8
19.81 4.39 19.78 0.2
21.07 4.4 20.13 4.6
21.82 4.44 21.61 0.9

 

 

 

 

119

 

The mean of the error is 2.7%. Therefore, a measured velocity of 10 m/S for this
test can be reported as 10 m/s i 54 cm/s with a 95% confidence level. This error term
does not incorporate variance that occurs during the experiment itself (due to a surging
compressor) so discretion must be applied by the experimenter to insure accurate

reporting of uncertainty.

120

APPENDIX 5. MAXIMIZING THE DYNAMIC RESPONSE OF THE

INFRAMETRICS IR CAMERA AND DATA RECORDING SYSTEM

Previous experiments by members of the Heat Transfer Laboratory have shown
that the equipment used to record data from the Inframetrics IR camera efl‘ects the data.
Following is a brief summary and recommendation from experiments run to optimize the
dynamic range of the system.

It was found that the Panasonic AG-2400 VCR effected the data stream fi‘om the
IR camera. It would seem obvious that the conclusion should be to replace this piece of
equipment. However, the Panasonic AG-2400 VCR increased the dynamic range. The
For-A VTG-33 video timer and the Sony PVM-l343MD Trinitron monitor did not effect
the data stream. The raw Signal fiom the Inframetrics camera has too much offset which
results in an upward shift of the gray scale as a function of temperature output. This
results in a reduction of the dynamic range capabilities of the system. Tests Show that as
much as 40% of the dynamic range was being loss by the previous equipment
configurations being utilized.

The recommended equipment configuration for data recording and playback are as

follows:

121

 

Data Recording Configuration

 

 

 

 

 

 

 

 

 

 

IR ,
C omen ForA Timer ———lPanasonic VCR Monitor

 

 

 

 

 

 

 

 

 

Data Playback Configuration
IP
Sony Plus I
Parasonic VCR MW . z.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 45. Optimal equipment configurations for data acquisition and processing.

On IP Plus Software:
Set Brightness to 48
Set Contrast in the range of 66 to 70

Set Gamma Function to 1.0 (this is the default setting)

0n For-A Video Timer:

Set 7552 switch to ON

This arrangement resulted in obtaining 96.8% of the systems available dynamic

response to temperature input. The results obtained and a comparison to the ideal system

response are shown in Figure 46..

122

 

 

Optima] equipment tests

 

 

 

 

 

 

 

20 degree IR range

250.0 4. *
Test I Result /]
y = 240.07x + 5. 1882 i
200“) " R2 = 0.9996 ,
q, I
3; 1500 4L [dad R8311” ;
y=2flk l
I? l i
23 100.0 1 l
—-*—— mecca Test 1 l

50 0 mecca Test 2
l ——+— Idea! Respmse l
/ l
0.0 I ‘l T I
o 0.2 0.4 0.6 0.8 1

Non Dimensional Temperature

 

Figure 46. Results of optimized equipment configuration on
gray scale - temperature relationship.

123

 

APPENDIX 6. VELOCITY AND TEMPERATURE CALCULATION PROGRAM FOR
TWO HOT-WIRE METHOD

The following is a program written in Mathematica° to calculate the velocity and
temperature from the data obtained fi'om two hot-wires positioned in the flow field. The

program is solves for velocity and temperature values in an iterative fashion.

Clear[al ,a2,b1 ,b2,ts1 ,ts2,tinf,v1 ,v2,v,E1,E2,n1,n2,alpha,Rc1,Rc2,Rh1,Rh2,delta]
(** Iterative Solution for hot wire measurements of velocity and temperature")
(** Input Room Temperature at Calibration “)

tcall=22.9;

tca12=23.0;

(** Input Parameters for specific wires and specific data points **)

E1=3.454;

E2=2.869;

A1=7.1596309;

B1=2.9114248;

n1=0.4197877;

Rc1=2.32;

Rh1=3.94;

A2=5.1325372;

B2=1 .4904883;

n2=0.45644473;

124

Rc2=1.99;

Rh2=2.59;

ts1=(Rh1/Rc1)*(273+tcall);

ts2=(Rh2/R02)*(273+tca12);

delta=.005;

al=A1/(tsl-tcall);

b1=Bl/(tsl-tcall);

a2=A2/(ts2-tcal2);

b2=B2/(ts2-tca12);

vell [tinf__]=(((E1A2)/(ts1-tinf)-a1)/b1)A(1/n1);

ve12[tinf_]=(((E2A2)/(ts2-tinf)-a2)/b2)A(1/n2);

Plot[{ve11[tinf],ve12[tinf]},{tinf,15,35}]

tinf=15;

Do[{v1=(((E1A2)/(tsl-tinf)-al)/b1)A(1/n1),
v2=(((E2A2)/(ts2-tinf)-32)/b2)A(l/n2),
v=Abs[v2-v1],
If[v<delta,i=5000,tinf=tinf+.0075]},{i,4000}]

"Temperature (C)"

N[tinf,5]

"Velocity (m/s)"

"=Hot Wirel velocity" v1

"=Hot Wire2 velocity" v2

125

"=Hot Wire velocity, average" N[(v1+v2)/2,5]

(*0 Olltpllt HI)

2.35719 =Hot Wirel velocity

2.35308 =Hot Wire2 velocity

2.3551 =Hot Wire velocity, average

126

 

BIBLIOGRAPHY

AISN Software Inc., EbleCurve 2D v.3.05 for Win32. San Rafael, CA, 1994

Blevins, Robert D., Applied Fluid Dynamics Handbook, Van Nostrand Reinhold
Company Inc., New York NY, 1984 p. 136-138, 312-316

Daryabeigi, Karnran, Wright, RE, Puram, OK. and D.W. Alderfer, “Directional
emittance corrections for thermal infiared imaging”, Thermosense XVI,
SPIE Vol. 1682, 1992 p.325-331

Foss, J, M. Potter, Fluid Mechanics, Great Lakes Press, Inc. Okemos MI, 1982,
pp.49-50

Gartenberg, Ehud and A. Sidney Roberts Jr., “Influence of Temperature Gradients on the
measurement accuracy of IR imaging systems”, SPIE Vol. 1313, Thermosense
XII, 1990, p.218-224

Herring, Robert J ., “Achieving high temperature measurement accuracy over a wide
ambient temperature range in thermal imaging radiometers”, SPIE proceedings
from Infiared Imaging Systems: Design, Analysis, Modeling, and Testing III,
SPIE Proceedings Vol. 1689, WA 1992

Holst, Gerald G, Testm' g and Evaluation of Infi'ared Systems, JCD Publishing Co.,
Maitland FL, 1993

Inagaki, Terumi, Suzuki, K and Y. Okamoto, “Uncertainty Analysis of
Surface Temperature Measurement Using Infi'ared Radiometer”, Thermosense
XVI, SPIE Vol. 2245, 1994 p.262-273

Incorpera, F. and D. Dewitt, Fundamentals of Heat and Mass Transfer John Wiley and
Sons, NY, NY, 1990, pp. 408-417, 708-749

 

127

Infi'ametrics Inc., Model 600 Opemtor's Manual Document #05250-200 Rev C,
Inframetrics Inc., N. Billenica, MA, 1988

Infiametrics Inc., How Many Smts Make a Picture?, Infi'ametrics reports, Infiametrics
Inc., N. Billenica, MA 1994

Lafi‘erty, John E, Arnold S. Collier, Infrared Surface Temrature Measurement in
NAVSWC’S HyperVelocity Win_d Tunnel #9. ICIASF Record, IEEE, NJ,
1991, pp. 167-177

Mandel, John, The Statistical Analysis of Experimental Data, Dover Publications Inc.,
NY, 1964 p.23

 

Media Cybernetics L.P., Image-Pro Plus ® version 1.3 for windows, Silver Springs, MD,
1995

Microsoft Inc., QuickBasic Version 4.50, Seattle Washington, 1988

Miller, Irwin and Freund, J ., Probabm ° and Statistics for Er_1gineers, Prentice-Hall, NJ,
1965,p.166

Morel, Thomas, “Design of Two-Dimensional Wind Tunnel Contractions”, Journal of
Fluids Engineering, Vol. 99, June 1977

Okamoto, Yosizo and Terumi Inagaki, “Thermal Measurement Techniques of IC package
Boards by means of Infiared Radiometer”, SPIE Vol. 2245 Thermosense XVI,
1994, p.231-240

Russ, John C., The Ime Processmg' HandBook Secong Editio_n, CRC Press,
Ann Arbor MI, 1995

Valvano, J .W. and Pearce, J. “Temperature Measurements”, thic -Ther1n§l Remnse of
Laser Irradiated Tissue edited by Welch, A.J., and van Gemert, M.J.C, Plenum
Press, NY, NY, 1995, pp. 520 - 529.

128

 

White, M.J., The Development of Non-Contactmg' and Non-Destructive Experimental

Techniques Cagble of Mm’ the Thermal Diffusivity of CVD Diamond PM
A thesis submitted to Michigan State University, 1996, pp. Bl-B4

White, Matt, Measurement of Slit Response Function, Internal Laboratory Document -
A30 RCE, Michigan State University, 1996

Willenborg, Klaus, Invegg' ation of the Velocfl' and TMe Fields of a Model
Defioster Flow Studientarbeit, Rhienisch-Westfalische Technische Hochschule,

Aachen, Germany, 1996

 

Weast, Robert c., CRC Handboole Chem and Plumes 66‘" Edition, CRC Press,
FL, 1986, p. F-89

Welch, A.J., and van Gemert, M.J.C, gm -Therma1 Respgnse of Laser Irradiated
Tissue, Plenum Press, NY, NY, 1995, pp. 520 - 529.

129