Danis Ward
Danis Ward
MASTER OF SCIENCE
(Mechanical Engineering)
At the
University of Wisconsin - Madison
2004
Xiaochun Li
(Principal Advisor)
ABSTRACT
Dennis M. Ward
Under the supervision of Professor Xiochun Li
at the University of Wisconsin - Madison
Optical fiber sensors have undergone refinements recently that make them attractive
alternatives for traditional measurement devices. Optical fiber sensors are compact in size,
easy to install and relatively inexpensive. These features have made them an ideal
measurement tool for taking piston temperature measurements.
This project investigated embedding an optical fiber sensor into a metal piston,
calibration of the sensor and the feasibility of measuring piston temperature with three
different light sources. The fiber Bragg grating (FBG) sensor was used in this research was
commercially purchased. A detailed procedure was developed that used a low temperature
electroplating process to fully embed the FBG sensor into a small utility engine piston. The
embedded sensor was calibrated by measuring the wavelength shift of the FBG and
comparing it the temperature as measured by a thermocouple.
ii
Of the light sources investigated, two were wavelength scanning lasers while the third
was a broadband source. Attempts to measure piston temperature with the broadband source
were unsuccessful due to the large light loss through the system. The wavelength scanning
lasers used were both successful in measuring piston temperature in a non-contact manner.
Due to cycle-to-cycle variations in the light transmission efficiency the slow scanning laser
could only measured the piston temperature after the engine had been stopped. After running
the engine at 1700 RPM with a load of 710 kPa IMEP the maximum piston temperature
measured was 149 0C. The fast scanning laser technique measured the piston temperature
with the running engine at the same conditions to be 136 0C. Comparing the results from
both temperature measurements yielded only a discrepancy of 13 0C, which is believed to be
good correlation.
iii
ACKNOWLEDGEMENTS
iv
TABLE OF CONTENTS
Abstract..................................................................................................................................... i
Acknowledgements ................................................................................................................ iii
Table of Contents ................................................................................................................... iv
List of Figures......................................................................................................................... vi
List of Tables ........................................................................................................................ viii
Nomenclature ......................................................................................................................... ix
1
INTRODUCTION........................................................................................................... 1
1.1
Motivation ................................................................................................................. 1
1.2
Research Objectives.................................................................................................. 3
1.3
2.2
2.3
2.4
Test Engine.............................................................................................................. 15
3.2
3.3
3.4
Optical System ........................................................................................................ 18
3.4.1
Fiber Optics..................................................................................................... 19
3.4.2
Light Source.................................................................................................... 21
3.4.2.1 Superluminescent Light Emitting Diode .................................................... 21
3.4.2.2 Slow Scanning Laser................................................................................... 22
3.4.2.3 Fast Scanning Laser .................................................................................... 26
3.4.3
Detector........................................................................................................... 27
v
4
4.2
4.3
4.4
Sensor Embedding .................................................................................................. 37
4.4.1
Piston Preparation ........................................................................................... 37
4.4.2
Chemical Surface Preparation......................................................................... 38
4.4.3
Sensor Preparation .......................................................................................... 39
4.4.4
Electroplating.................................................................................................. 41
4.4.5
Post-Electroplating Preparation ...................................................................... 44
4.4.6
Sensor Placement ............................................................................................ 46
4.4.7
Embedding Issues ........................................................................................... 46
4.5
4.6
4.7
5.2
5.3
Conclusions............................................................................................................. 64
6.2
Recommendations ................................................................................................... 64
REFERENCES.............................................................................................................. 67
vi
LIST OF FIGURES
Number
Page
Figure 1-1 Specific strength of material versus temperature [1] .............................................. 2
Figure 1-2 Piston crown temperature at various engine speeds [2] .......................................... 2
Figure 2-1 Fiber optic cable.................................................................................................... 10
Figure 2-2 Light transmission in fiber optic cable.................................................................. 10
Figure 2-3 Fiber Bragg Grating schematic [22]...................................................................... 12
Figure 2-4 Temperature and strain effect on a Bragg grating [22] ......................................... 13
Figure 3-1 Pressure transducer location in the cylinder head ................................................. 17
Figure 3-2 Temperature measurement technique ................................................................... 19
Figure 3-3 SLED maximum power output ............................................................................. 22
Figure 3-4 New Focus laser at maximum output.................................................................... 23
Figure 3-5 Slow scanning laser trace ...................................................................................... 24
Figure 3-6 Etalon formation mechanism and result................................................................ 25
Figure 3-7 Slow scanning laser trace zoomed ........................................................................ 25
Figure 3-8 Wavelengths versus voltage output from tunable laser......................................... 26
Figure 3-9 Fast scanning laser measured wavelength range................................................... 27
Figure 4-1 Optical engine section view with piston at BDC .................................................. 29
Figure 4-2 Separation of temperature and strain with a stainless steel tube........................... 32
Figure 4-3 Light collection port in the machined engine block.............................................. 33
Figure 4-4 Invar window holders............................................................................................ 34
Figure 4-5 Launching side machined engine block with window and holder ........................ 35
Figure 4-6 Piston channel dimensions .................................................................................... 36
Figure 4-7 Specification of piston ring stopper location ........................................................ 37
Figure 4-8 Electroplating setup............................................................................................... 42
Figure 4-9 Current density variation....................................................................................... 45
Figure 4-10 Excess deposited nickel....................................................................................... 45
Figure 4-11 Piston section to examine sensor placement ....................................................... 46
Figure 4-12 Electroplating defects during manufacturing ...................................................... 48
Figure 4-13 Embedded FBG with good bonding, end view showing fiber ............................ 48
Figure 4-14 Cylinder wear comparison, left: stock engine bore finish and piston right: stock
engine bore finish with embedded sensor piston ............................................................ 50
Figure 4-15 Embedded FBG piston after wear testing ........................................................... 50
Figure 4-16 Bragg wavelength measurement of embedded sensor with OSA ....................... 51
Figure 4-17 Specifications for FBG 14-033-3-8514 as measured by the manufacturer, O/E
Land ................................................................................................................................ 52
Figure 4-18 Photo detector signal from a laser scan at through the FBG piston at 20 0C ...... 53
Figure 4-19 Embedded FBG calibration................................................................................. 54
Figure 5-1 Expected results with slow scan laser ................................................................... 56
Figure 5-2 Actual fluctuations in transmission from consecutive engine cycles.................... 56
Figure 5-3 Bragg wavelength extrapolation back to engine stop time ................................... 57
vii
Figure 5-4 Piston temperature variation with running time as measured in a stopped engine 58
Figure 5-5 Data as collected from the fast scanning laser and mirror output ......................... 60
Figure 5-6 Fast scanning laser data with corrected mirror output .......................................... 61
Figure 5-7 Computed wavelength function ............................................................................ 61
Figure 5-8 Fast laser scan through the FBG sensor while stationary ..................................... 62
Figure 5-9 Comparison of fast scanning laser through engine with FBG sensor ................... 63
viii
LIST OF TABLES
Number
Page
Table 3.1 Test engine specifications....................................................................................... 15
Table 4.1 Stainless steel tubing specifications........................................................................ 31
Table 4.2 Aluminum A390-T7 casting alloy compostition .................................................... 39
Table 4.3 Desmut and zinc striping solution composition..................................................... 39
Table 4.4 PVD parameters for titanium and nickel films ....................................................... 41
Table 4.5 Copper strike electroplating bath composition ....................................................... 43
Table 4.6 Watts nickel electroplating bath composition......................................................... 44
ix
NOMENCLATURE
BDC
BTDC
CO
COV
ECDL
FBG
FSR
IMEP
NA
OSA
PVD
RPM
SLED
TIR
uHC
1 INTRODUCTION
1.1 Motivation
All materials, including the metals most frequently used in internal combustion
engines decrease in tensile strength per density as the material temperature increases as can
be seen in Figure 1-1. With more widespread use of small utility engines in daily life and
increasingly tighter emissions regulations, engine designs have been forced to maximize
power while minimized material and, therefore, cost. This has resulted in more efficient
designs, but also less of a built-in safety factor in components, forcing each components
function to be understood in detail. One of the most changeling components to analyze is the
piston due to its reciprocating motion, however the piston remains one of the most important
components of small engines with respect to durability. The piston contributes on the order
of 50 to 60% of the mechanical losses of the engine [2]. Energy from combustion heats the
crown of the piston and this heat needs to be dissipated by way of the ring pack and skirt.
Aluminum is a much better conductor of heat than cast iron. If silver is considered to
be 100%, aluminum and cast iron have conductivities or 38 and 11.9% relative to that of
silver. Aluminum, due to its high diffusivity greatly reduces the maximum operating
temperature of the piston crown. Piston temperate is influenced by many factors including
engine speed, load and combustion chamber design. The impact of engine speed is
represented in Figure 1-2. One of the major disadvantages of aluminum is that it has a high
coefficient of linear expansion. This means that the use of aluminum requires additional
clearance be designed into the system so that at operating temperature excess friction is not
generated. This additional clearance usually leads to piston slap when the engine is first
started cold. Thus, to produce low noise with minimal impact on engine performance and to
2
ensure long engine life, being able to accurately measure piston temperature in a small
engine is important.
4
their results are in Chapter 0. The discussion of the results is given in Chapter 6 along with
the conclusions obtained from this work and recommendations for future work.
7
part of the engine cycle when the least pressure and temperature are in the combustion
chamber. Due to the pistons reciprocating motion is the contacts were placed higher in the
stroke, they would interfere with the engines operation. This method is also limited to the
number of contacts that can be used due to the confinement of the crankcase.
Work on smaller engines and two-cycle engines with a minimum amount of room in
the crankcase led to the use of the thermocouple wires being routed via a tube out through the
combustion chamber. This method required a sealing mechanism to contain the combustion
gasses and a system to cool the wires inside the tube to prevent weakening of the wire
insulation. However, this method also suffered wire fatigue at higher engine speeds and was
limited to 300 RPM [9].
The latest method utilized to transfer thermocouple signals outside the engine uses
telemetry [10]. On the underside of the piston a transmitter is mounted, which transmits
thermocouple information as frequency modulated infrared pulses. A photo detector
mounted in the crankcase, which is wired to an external receiving unit, picks up this signal.
This signal was successfully transmitted up to 5000 RPM with an accuracy of 2.8C. An
improvement on this technique replaced the infrared transmitter with a microwave
transmitter [11]. This removed the line of sight constraint for this method. This technique
experiences several problems in that the transmitter batteries need replacement and the
transmitter box can add a substantial amount of mass to the piston making its application use
limited.
A recent method developed to measure piton temperature uses thermistors and
electromagnetic induction. Transmitter and receiver coils are installed in the bottom of the
cylinder and a resonator coil is installed onto the bottom of the piston. When alternating
8
current is applied to the transmitter coil a magnetic field and current is generated in the
receiver coil. When temperature increases the embedded thermistor resistance decreases and
the current flowing through the receiver coil decreases. Piston temperature can then be
determined from a predefined map of current change and piston temperature. As reported by
Kato [12] this method has many advantages including high accuracy of 2.5C, multiple
measurement points per piston (six were demonstrated) and durability at speeds up to 6000
RPM.
This method does suffer from a larger dispersion due to the D/A conversion maps
9
pipes and containers in power generation [16] and engine conditions in the automotive
industry [17].
The dominant use of embedded optical fibers in the last twenty years has been in nonmetallic composites [18]. They are a convenient method to monitor important factors and
components while being easy to embed in the manufacturing process. Optical sensors also
provide the added benefit of being lightweight, immune to electromagnetic interface and
resistant to hostile environments. However only a few researchers have used them as sensors
to monitor conditions without continuous fiber contact [19], and still fewer have tried to use
optical fiber sensors in hostile environments.
n2
n1
(2.1)
where min is the angle shown in Figure 2-2 and n1 and n2 are the indices of refraction of the
core and cladding, respectively. In order for light to couple into the fiber it must be incident
within an acceptance cone. The acceptance angle is related to the refractive indices of the
core, cladding, and medium surrounding the fiber. This relationship is called the numerical
10
aperture (NA) of the fiber. Figure 2-2 illustrates the relationship between the acceptance
angle and the refractive indices. By using Snell's law and basic trigonometric relationships,
the NA of the fiber is given by:
NA = (n ) sin min =
(n
2
1
n 22
(2.2)
where n is the NA of the material outside of the fiber. Light entering the fiber from outside
the acceptance cone strikes the core-cladding interface at too steep of an angle and enters the
cladding, where it quickly dissipates.
11
The photosensitivity of fibers was first observed in 1978 by Hill et al [20,21] and
spawned a new technique for the measurement of temperature and strain. In its simplest
form a FBG consists of a periodic modulation of the refractive index in the core of a singlemode optical fiber. In a uniform fiber Bragg grating, the phase fronts are perpendicular to
the fibers longitudinal axis with grating planes having constant period as shown in Figure
2-3. Light, guided along the core of an optical fiber, will be scattered by each grating plane.
If the Bragg condition is not satisfied, the reflected light from each of the subsequent planes
becomes progressively out of phase and will eventually cancel out. Additionally, light that is
not coincident with the Bragg wavelength resonance will experience very weak reflection at
each of the grating planes because the index mismatch. For wavelengths where the Bragg
condition is satisfied, the contributions of reflected light from each grating plane add
constructively in the backward direction to form a back-reflected peak with a center
wavelength defined by the grating parameters.
12
The Bragg grating resonance, which is the center wavelength of back-reflected light
from a Bragg grating, depends on the effective index of refraction of the core and the
periodicity of the grating. The simple relationship is shown as
B = 2n eff
(2.3)
Where B is the Bragg wavelength, neff is the effective index change and is the
distance between index of refraction changes. The effective index of refraction as well as the
periodic spacing between the grating planes is affected by changes in strain and temperature.
If (2.3) is differentiated the shift in the bragg grating center wavelength due to strain (l) and
temperature (T) changes is given by:
n eff
n
l + 2 eff + n eff
T
B = 2
+ n eff
l
l
T
(2.4)
13
The strain effect on an optical fiber is represented by the first term in equation (2.4).
This change is induced by two phenomena the physical change in grating spacing and the
strain induced change in the refractive index. Experimental results show that for an applied
500 the Bragg center wavelength will shift approximately 1 nm as shown in Figure 2-4.
The temperature effect on an optical fiber is represented by the second term in (2.4).
The Bragg wavelength shifts due to changes in the grating spacing by thermal expansion and
the changes in index of refraction. If this effect is studied alone the wavelength shift for a
temperature change T can be written as:
B = B ( + n )T
(2.5)
Where is the thermal expansion coefficient for the fiber and n represents the thermooptic coefficient, which are respectfully equal to approximately 0.55 x 10-6 and 8.6 x 10-6
[22]. Observing the temperature effect on the Bragg wavelength in Figure 2-4, the sensitivity
of a Bragg grating is approximately 1.28 nm per 100 0C. It should be noted that any change
14
in wavelength caused by an external excitation to the grating is the sum of both the
temperature and strain effects, thus, if used in a sensing application where only one
perturbation is of interest, the separation of temperature and strain becomes necessary.
15
3 PROJECT EQUIPMENT
3.1 Test Engine
The engine used in this investigation was a Briggs & Stratton, Intek model 110602. It
is a 4.8 kW (6.5 hp) vertical crankshaft single-cylinder spark-ignited engine. The engine is
air cooled, with the airflow provided by vanes cast into the flywheel. The complete Briggs &
Stratton Intek engine specifications are given in Table 3.1.
Engine
Bore
68.275 mm
Stroke
45.720 mm
Displacement
167.38 cm3
88.6 mm
0.381 mm
Combustion chamber
Pancake Style
Rated speed
3060 rpm
Power rating
4.8 kW
Table 3.1 Test engine specifications.
The stock carburetor was a non-adjustable type, so the main jet was replaced with an
adjustable jet so that allowed the mixture to be changed. This allowed the fuel-air mixture to
16
be changed. The engine was also equipped with a fixed magneto ignition system that would
produce a spark 190 before top dead center (BTDC).
17
Bottom View
Of Combustion
Chamber
Exhaust Valve
Spark Plug
Port
Additional
Spark Plug
(Not Used)
Pressure
Transducer
Port
Intake Valve
For setting and monitoring the engine operating conditions the air fuel ratio was
measured. A Horiba Mexa-110 lambda sensor in the exhaust was used to determine the
charge air fuel ratio and verify correct operation of the carburetor.
Engine temperatures were closely monitored to evaluate engine performance. These
included the oil sump temperature, intake air and the exhaust gas. The first exhaust gas
measurement was at the exit of the exhaust port and near the lambda sensor. This
measurement made sure that the oxygen sensor was at proper temperature and not producing
erroneous results due to water condensation. The second exhaust temperature located in a
surge tank downstream in the exhaust was used together with the oil sump temperature to
18
give a point of reference as to when the engine had reached its steady state operating
temperature during testing.
19
Light Source
Focusing Assembly
Photo Detector
Focusing Assembly
20
strength compared to a polymer-coated fiber. To increase the overall robustness of the fiber
assembly a stainless steel sheath covered the fiber. These were important considerations
when working with the high temperatures and vibrations generated by the engine. To make
coupling light into the fiber easier each end of the fiber had a FC connector that could be
mated to the light source.
The collection fiber consisted of a 600 m core, 636 m cladding and a 750 m
diameter aluminum jacket with SMA-905 connectors in a stainless steel sheath. This fiber
was chosen for its large core size to facilitate collecting as much light as possible.
To position the end of the single mode fiber carrying the signal light to the sensor
fiber a translation stage was used. This stage allowed for adjustment in all three dimensions
via 80 thread per inch screws. The whole assembly could be fixed in place with locking
screws. For rigidity the translation stage was mounted to the engine block with 0.25 inch
aluminum supports. When the end of the launching fiber was adjust to have 0.73 mm of
space between it and the end of the sensor fiber a 200 m spot was projected onto the end of
the sensor fiber due to the natural dispersion of light exiting the fiber. This large spot size
would help to allow for misalignment and movement of the sensor end on a cycle to cycle
basis.
To collect the light emerging from the embedded FBG sensor fiber a collimation lens
assembly was used. This assembly (Thor Labs F230SMA-C) combined an antireflective
coated lens into an aluminum housing. The lens was positioned such that it was one focal
length away from the end of the fiber.
21
University of Wisconsin Madison, has a scan rate of 2.4 x 10 nm/sec which will now be
referred to as the fast scanning laser.
22
0
-5
-10
-15
Power [dBm]
-20
-25
-30
-35
-40
-45
-50
1370
1380
1390
1400
1410
1420
1430
1440
1450
1460
1470
Wavelength [nm]
23
Power [dBm]
-10
-20
-30
-40
-50
-60
-70
1360
1380
1400
1420
1440
1460
1480
1500
Wavelength [nm]
Scanning laser operation was also evaluated by recording a laser scan with time on a
photo detector. The laser was set to scan from 1425.5 to 1428.5 nm with a speed of 1
nm/second. Single mode fiber optic patch cords delivered the light a PDA400 photo
detector, of which specifications can be found in section 3.4.3. The photo detector output
was recorded at 1 MHz with a Hi-Techniques HT-600. The resultant trace can be seen in
Figure 3-5. This trace shows a periodic signal with large peaks arriving every 0.05 seconds,
corresponding to a of 0.05 nm. This periodic signal is caused by an etalon that has
formed in the optical system. When light is passed between two flat, parallel, partially
reflective surfaces an etalon is formed, as depicted in Figure 3-6. When monochromatic light
24
is passed through an etalon, multiple interference occurs producing fringes in intensity. The
spacing of these fringes is related to the free spectral range (FSR) defined as:
FSR =
(3.1)
c
2nd
where c is the speed of light, d the spacing between reflecting surfaces, and n is the index of
refraction of the material between the reflecting coatings. Knowing the FSR a distance and
therefore likely source of the etalon can be calculated. The etalon formed in Figure 3-5 is
from the spacing between the window and photo array in the photo detector. Proof of this is
evident by the lack of the etalon in Figure 3-4. Another higher frequency etalon was also
measured as shown in Figure 3-7. This etalon was calculated to be from the laser cavity it
self.
P h o to d e te c to r o u tp u t [V ]
4.5
4.0
3.5
3.0
Shown in
Figure 3-7
2.5
0.0
0.2
0.4
0.6
Sample time [seconds]
0.8
1.0
25
3.96
3.94
3.92
3.90
3.88
3.86
3.84
3.82
0.600
0.602
0.604
0.606
Sample time [seconds]
0.608
26
The laser outputs a reference signal proportional to the wavelength produced that is
used to determine an absolute temperature change. The voltage output is linear with respect
to the wavelength of laser light and can be accessed through a BNC connector on the back of
the control box. Figure 3-8 shows how the voltage corresponds to the wavelength as
measured by an OSA. The voltage-to- wavelength relationship can be described as
= 16.499 *V + 1380.417
(3.2)
1430
1429
1428
1427
1426
1425
2.75
2.80
2.85
2.90
Output from scanning laser [V]
2.95
27
vibrating at 15.6 kHz. The vibrating mirror forces the diode laser to sweep from 1420.20 to
1432.68 nm in 32 s, and then to return from 1420.20 to 1432.68 nm in the next 32 s. This
scanning range can be seen in Figure 3-9 as measured by an OSA over a one second period.
To avoid water absorption within the cavity the laser system is contained in a sealed box and
purged with dry air.
-50
-100
-150
-200
1420
1425
1430
Wavelength [nm]
1435
3.4.3 Detector
The main detector used in this project was a Thorlabs PDA400. The PDA400 is an
amplified, switchable-gain, InGaAs detector designed for detection of light signals up to 10
MHz. It has a peak response of 0.95 A/W at 1550 nm. For the spectral range used in this
project it has an average response of 0.86 A/W. This detector also has a large active area of
1 mm2, which allows for the use of a large multimode fiber for the collection of light from
the FBG fiber.
28
An optical spectrum analyzer (OSA) was also used for a spectral intensity
measurement. The Agilent model 86143B OSA has many features that make it very versatile
for the measurements taken in this project. It is capable of low light level detection with a
sensitivity of 90 dBm. The OSA has an absolute accuracy 0.5 nm, but after internal
calibration it is capable of a 0.06 nm resolution measurement. However, this resolution is
based on a single mode input fiber. Since the OSA uses a diffraction grating to separate light
into discrete wavelengths, the accuracy scales linearly with input fiber size. An FC port
provides an easy method to connected signal patch cords. This particular OSA is limited to a
sweep rate of 40 nm in 56.3 ms as stated by the manufacture.
29
4 EXPERIMENTAL SETUP
4.1 Outline of FBG Temperature Measurement Setup
In order to make a piston temperature measurement using the FBG technique specific
hardware modification are required. Optical access ports are required in the cylinder and the
FBG needs to be embedded into the piston surface. A schematic of the optical system is
shown in Figure 4-1. The following section, 4.2, outlines the machining process of the
cylinder block necessary and sections 4.3 and 4.4 cover the sensor embedding processes of
machining and nickel electroplating.
Photo Detector
Light Source
Focusing Assembly
Collection
Fiber Optics
Pitching
Fiber Optics
Engine Block
Sealing O-Ring
Metal Shim
Piston With
Embedded FBG
Focusing Assembly
30
The goal of this research was to measure a steady state maximum piston temperature.
It would be essential to not have the measured piston temperature affected by the heat flux
imposed by combustion on a cycle to cycle basis. In order to check the validity of this
condition the materials time response to a heat input was evaluated. Using the properties of
the embedding material, nickel, at an assumed maximum piston temperature of 400 K in the
thermal diffusivity () of nickel could be estimated as in Equation 4.1.
400
W
80.2
m2
k
m K
= 18.6 x10 6
=
=
cp
J
kg
s
8900 3 485
m
kg K
(4.1)
where cp is the specific heat, is the material density and k is the thermal diffusivity. Using
the thermal diffusivity the time response, , can be calculated as in equation 4.2 with L being
the distance below the surface that the FBG resides.
L2
m2
18.6 x10
s
6
= 0.042 [s ]
(4.2)
The maximum engine speed of the test engine used is 3060 RPM, which equates to one crank
revolution in every 0.0196 seconds. Since the characteristic response time is more than 2
engine revolutions, it can be assumed that the measured piston temperature will be steady,
unaffected by the periodic nature of the heat flux.
Fiber Bragg Gratings are sensitive to both temperature and strain. Since this project
only wants to measure temperature the strain affect needs to be decoupled. To accomplish
31
this the FBG sensor was enclosed in a stainless steel tube. The tube was a 30-gauge
hypodermic needle with a 152 m nominal inside diameter and a 311 m nominal outside
diameter with the complete specifications given in Table 4.1. To secure the fiber to the
piston the thrust side fiber end was allowed to extend out of the tube 3 mm and, therefore,
was electroplated to the piston, as illustrated in Figure 4-2. The rest of the sensor fiber was
enclosed in the tube and able to float, negating any strain effect cause by the surrounding
material.
Material
I.D.
Tolerance
O.D.
Tolerance
32
Electroplated Nickel
FBG
Stainless Steel Tube
Aluminum Piston
Figure 4-2 Separation of temperature and strain with a stainless steel tube
33
To seal the combustion chamber or crankcase from the outside environment sapphire
windows were used on both the input and collection sides. To minimize the intrusion into
the engine the windows were kept as small as small as possible, while still allowing
adjustment of the delivery fiber to account for any variation in manufacturing and to adjust
for maximum signal transmission. The windows used were 9.525 mm in diameter and 0.508
mm thick.
To contain the windows and hold them in place are two window holders, which are
shown in Figure 4-4. To place the windows as close as possible to the cylinder bore, two
12.700 mm counterbored holes were machined into the block concentric to the optical access
ports. The depth of the counterbored holes left 0.051 mm of the block material between the
34
cylinder bore and the inside window surface. In addition to decrease the distance from the
cylinder bore to the window face the sapphire windows were joined to the window holders
with a ceramic adhesive (Cotronics Resbond 940). In order for this mounting method to be
feasible with the large temperature changes that would occur, the window holders were
constructed of Invar 36. The coefficient of thermal expansion of Invar 36 is one-tenth that of
carbon steel up to 200 0C, thus making its thermal expansion within 50% that of sapphire.
The cylinder block was also machined to accept the fasteners that secure the window holders.
Each window holder had a viton o-ring to seal the combustion chamber from the outside
environment with the collection side window holder having an additional o-ring to prevent
oil leakage from the pushrod cavity. Final placement adjustments were possible through the
use of metal shims between the engine block and the window holders. The launching side
holder and window can be seen in Figure 4-5.
35
Figure 4-5 Launching side machined engine block with window and holder
36
0.34 mm
0.89 mm
3.12 mm
Figure 4-6 Piston channel dimensions
The engine used in this investigation has a standard ring design of two compression
rings and one oil control ring. The compression rings were pinned to prevent them from
turning as shown in Figure 4-7. This was done because piston rings rotate under normal
operation as proven by various researchers [25,26]. The pin eliminated the possibility of the
piston rings rotating and becoming caught in the optical access hole. The compression rings
were assembled so that their gaps would be opposite at an angle of 180. Due to the
expander spring in the oil control ring a stopper pin was not used, but the free ends were
given a 0.508 mm chamfer.
37
38
used to connect the piston to the power supply in the electroplating process. To stop the
electroplating process from occurring over the entire piston a lacquer mask was applied.
Four coats of Tolber Miccrostop lacquer were applied to the entire piston with extra
precautions taken to thoroughly coat the ring grooves and wrist pin bosses. After the final
application of lacquer the piston was baked at 37 0C for 20 minutes to ensure that the lacquer
had cured.
39
Component
Al
Cu
Weightt %
78
4-5
Fe
Max
0.5
Mg
0.45 0.65
Mn
Max
0.1
Si
Ti
Max
16 - 18 0.2
Zn
Max
0.1
Nitric Acid
Hydrofluric Acid
Dionized Water
% By Volume
20%
2%
Balance
40
0
of polymer material that melts or significantly softens above 100 C. Therefore, optical
fibers cannot be directly embedded in metal. However, removing the polymer coating and
adding a thin metal coating make optic fibers very attractive for use as sensors in metallic
structures.
Removing the polymer coating can be accomplished by either a: mechanical stripping
or a chemical stripping process. Mechanical stripping is the standard way to remove the
polymer coating for telecommunications, however this process tends to leave residual
polymer particles stuck to the fiber that have to be removed manually. This can cause the
fiber to no longer be chemically clean, and increases the chance of the fiber breaking. When
the FBG is written onto the fiber a section of the polymer buffer is stripped away, creating a
strength gradient at the end of the stripped section. Using mechanical stripping can cause the
fiber to break at this point. Therefore a chemical stripping process was used to remove the
polymer coating. When a section of optical fiber containing the FBG was immersed in an
acetone bath for ten minutes the polymer buffer layer would detach from the fiber leaving a
smooth clean surface.
A thin metallic layer was applied by physical vapor deposition (PVD) to provide an
electrically conductive base for embedding. PVD is a method of physically depositing
materials including both metals and insulating materials for example Ti or SiO2. The basic
idea of PVD is that an inert gas is ionized and the ions are accelerated into a solid target
material. The ions strike the surface of the target material and atomic clusters of the target
material are ejected and then deposited onto the sample surface.
The PVD machine used for this research was the CVC 601. The unit has three DC
magnetron target stations that can be used, each being able to accommodate various
41
materials. After argon was injected into the chamber a high vacuum condition was achieved
by a cryo-pump. The optical fiber was then sputtered with a 0.05 m film of titanium and a
0.1 m film of nickel. The thin titanium film was applied to improve adhesion between the
optical fiber and the desired nickel layer. In Table 4.4 the PVD process parameters are
described in detail.
Target Material
Chamber Pressure
Target Diameter
Deposition Power
Ramp Time
Film Thickness
Titanium
0.4 Pa
203.2 mm
1 kW
10 min
50 nm
Nickel
0.4 Pa
76.2 mm
0.5 KW
20 min
500 nm
4.4.4 Electroplating
Electroplating is the process of depositing a coating having a desirable form by means
of electrolysis. A typical electroplating cell consists of anode, cathode, aqueous-metal
solution, and a power supply. In the simplified example shown in Figure 4-8, the sacrificial
anode is made of nickel, the cathode is made of the material to be electroplated, and the
aqueous-metal solution consists of nickel (Ni2+), chlorine (Cl1-) and sulfate ions (SO42-).
When the power supply is turned on, the positive ions in the solution are attracted to the
negatively biased cathode. The nickel ions that reach the cathode gain electrons and are
deposited or plated onto the surface of the cathode forming the electrodeposit.
Simultaneously, nickel is electrochemically etched from the nickel anode, to produce ions for
42
the aqueous solution and electrons for the power supply. Hydrogen ions also gain electrons
from the cathode and form bubbles of hydrogen gas.
A copper strike was deposited over the zinc immersion deposit to protect the
underlying aluminum. Since the zinc immersion layer is extremely thin, any plating
treatment that penetrates the zinc layer will attack the aluminum and a poor deposit will
result. Based on literature suggestions [27,34] a copper cyanide solution from Technic Inc.
was used with the composition listed in Table 4.5. The strike solution was used at room
temperature. Electrical contact was made before immersion of the piston into the bath to
limit the deposition that would occur by immersion resulting in the poor adhesion of the
43
2
electrodepost. An initial high current density of 26 mA/cm was used for two minutes
followed by three minutes at 13 mA/cm2 to plate approximately 4.5 m of copper.
Material
Potassium Copper Cyanide (K3Cu(CN)4
Potassium Cyanide (KCN)
Potassium Hydroxide (KOH)
Deionized Water
Amount (Vol %)
6.30%
0.65%
1.00%
Balance
With the preparation steps completed the FBG sensor was embedded with nickel
electroplating. The stainless steel tube was inserted into the locating bosses and the FBG
fiber threaded into the tube, at which time the tube and fiber were positioned such that the
FBG was in the center of the piston. The temperature of the electroplating bath was
controlled to 50 0C. For the first hour of plating the current density was set to 22.5 mA/cm2
then adjusted to 40. mA/cm2 for the duration of the approximate 48 hours of electroplating.
The nickel electroplating setup consisted of a polypropylene tub, Micro pump (model
HG0024), Gelman Sciences 1 m filter (CR Capsule model 12829), a Cole-Parmer DigiSense Temperature Controller (model 2186-00) with a type J thermocouple, and a quartz
laboratory heater (Glo-Quartz model LHP-300).
44
Material
Nickel
Nickel Sulfate
Nickel Chloride
Boric Acid
Deionized Water
Amount (Vol %)
8.90%
35.15%
5.86%
3.90%
Balance
45
transmission a solution of colloidal alumina with particle sizes that ranges from 1 m to 50
nm was used on a 1 m polishing cloth. The final result is an embedded FBG sensor in an
aluminum piston with a mass gain of less than 1 %.
Lines of Current
Electroplated Nickel
Polymer Mask
Aluminum Piston
Electroplated Nickel
Mask
FBG
Stainless Steel Tube
Aluminum Substrate
Aluminum Piston
Mask
46
Center of
Tubing
25 m
575 m
47
end of the optical fibers where they enter and leave the metal part. During manufacturing,
two types of defects, delamination and pits, can damage the optical sensors and lower the
strength of the embedded sensor or destroy it. However, all types of defects are well
documented in literature, along with solutions for each [27,29,31]. Figure 4-12(a) shows a
crack at the interface of the base metal and the electroplated material while Figure 4-12(b)
shows delimitation of the electroplated material. Surface pits were also present in some
samples as seen in Figure 4-12(c). Delamination can be due to inadequate cleaning of the
optical fibers and the substrates or improper electroplating parameters such as solution
temperature or composition. Pitting is cause by either insufficient agitation or improper
cleaning. In addition, unwanted electroplating also occurred on wrist pin and ring groove
surfaces as seen in Figure 4-12(d). The cause of this problem was easily identified by close
examination of the piston during the lacquer application process and covering any small
pores in the lacquer with an additional coating. The embedding quality was improved
significantly, as shown in Figure 4-13, after the cleaning process and the electroplating
parameters were optimized as described in previous sections. The bonding strength of the
electroplated nickel to the substrate was also enhanced by sanding the channel sides before
electroplating the part.
48
Delamination
Crack
Surface
Pits
Unwanted
Ni Deposit
Fiber
Electroplated
Nickel
Figure 4-13 Embedded FBG with good bonding, end view showing fiber
49
50
Figure 4-14 Cylinder wear comparison, left: stock engine bore finish and piston right: stock
engine bore finish with embedded sensor piston
51
transmission loss of 83 dB loss from the source to the OSA. The resulting measurement
result can be seen in Figure 4-16 along with the supplied grating specifications as received
from the manufacturer, O/E Land, in Figure 4-17. In an examination of the two plots it can
be seen that the Bragg wavelength matches, but the measured transmission loss was only 3.9
dB at the Bragg wavelength while the stated manufactures loss is 10.5 dB.
-64
-66
-68
-70
-72
1420
1422
1424
1426
Wavelength [nm]
1428
1430
52
Figure 4-17 Specifications for FBG 14-033-3-8514 as measured by the manufacturer, O/E Land
The sensor calibration test was conducted by placing the entire engine assembly into
an oven to raising the entire piston temperature in specified increments. After the oven
temperature was changed the engine was allowed to soak for 90 minutes to ensure that the
interior piston temperature had reached that of the oven. Temperature readings were then
recorded from two K-type thermocouples cemented onto the piston crown and a K-type
thermocouple placed inside the oven. A total of eight temperatures were measured extending
to 100 0C.
A typical scan of the slow speed laser is shown in Figure 4-18. This scan shows the
Bragg wavelength at 1426.19 nm while at room temperature (20 0C). It should be noted that
at a certain amount of discretization and noise is present in the wavelength output with faster
scan speeds. This is due to circuitry than drives the voltage output, when in fact the laser
scans at a near constant speed. This could is reduced by slower scan speed but to keep the
53
recorded data file sizes manageable on the data acquisition computer a moderate scan speed
was used.
3.00
7.2
2.95
2.90
6.8
2.85
6.6
2.80
7.0
6.4
2.75
Bragg Wavelength
6.2
2.70
0
6
Time [seconds]
10
Figure 4-18 Photo detector signal from a laser scan at through the FBG piston at 20 0C
The composite results of calibration test can be seen in Figure 4-19 with error bars
included for the uncertainty of 2.2 0C associated with using K-type thermocouples. As
expected, the Bragg wavelength shift with temperature has a linear relationship that can be
expressed as:
B = 0.0112 * T + 1426.1
where T is temperature in degrees Celsius and B is in nm.
(6.2)
54
1428.00
1427.75
y = 0.0112x + 1426.1
R2 = 0.9762
1427.50
1427.25
1427.00
1426.75
1426.50
1426.25
1426.00
0.0
20.0
40.0
60.0
80.0
100.0
120.0
Temperature [0C]
55
1436
1.5
1434
1432
1.0
1430
Wavelegth [nm]
56
1428
0.5
1426
0.0
0.5
1.0
1.5
Time [seconds]
2.0
2.5
10
8
6
4
2
0
0.0
0.1
0.2
0.3
Time [Seconds]
0.4
0.5
0.6
The slow scanning technique was used to investigate the rise of piston temperature as the
engine was warmed up at a full load condition. For this test the engine was controlled to
1700 RPM and full throttle. A 12.5:1 delivered air/fuel ratio corresponded to an IMEP of
57
719 kPa. To accomplish this measurement a scan was taken at with the engine at rest at
room temperature. The engine was then run for three minutes, stopped, the optical path
aligned and a temperature measurement taken as soon as possible. At least two more repeat
scans were taken to establish a cooling curve for the piston. The piston temperature was
linearly extrapolated back to the time when the engine was stopped. The extrapolation
process was repeated after running three, six, ten, and twenty minutes for a total engine
running time of thirty-nine minutes. The result of the thirty-nine minute running time
condition can be seen in Figure 5-3. At the last measurement point the engine oil and
exhaust temperatures had stabilized for the previous ten minutes. The results of this process
can be seen in Figure 5-4 with the maximum temperature measured to be 149 0C.
1427.7
y = -0.0014x + 1427.6
R2 = 0.985
1427.6
1427.5
1427.4
1427.3
1427.2
1427.1
1427
1426.9
0
50
100
150
200
250
300
350
400
450
58
160
140
120
100
80
60
40
20
0
10
20
30
Engine Running Time [minutes]
40
Figure 5-4 Piston temperature variation with running time as measured in a stopped engine
59
faced in the previous section with the slow scanning laser, thus ruling out just this particular
combination of FBG and OSA as a means to measure piston temperature in a running engine.
60
minimum signal output does not correspond to the trigger point. The minimum signal point
is located by finding the data point that lies between the two larger dips in signal, which is
possible since the wavelength scan is symmetrical. Then the trigger was then shifted using
this symmetric relationship until the trigger lined up with minimum data point. The result
can be seen in Figure 5-6. Once this was accomplished how the wavelength scan changed
with time could be calculated from the known sinusoidal motion of the mirror as shown in
Figure 5-7. Due to the fast scan time two measurements were actually taken; one as the laser
scans from 1420.20 to 1432.68 nm and another as the laser scans back to 1420.20 nm. A
sample trace of this measurement can be seen in Figure 5-8.
2.4
3
2.3
2
2.2
1
2.1
0
-60x10
-6
-40
-20
Time [S]
20
Figure 5-5 Data as collected from the fast scanning laser and mirror output
2.5
61
2.4
3
2.3
2
2.2
1
2.1
0
-60x10
-6
-40
-20
Time [S]
20
Figure 5-6 Fast scanning laser data with corrected mirror output
1432
Wavelength [nm]
1430
1428
1426
1424
1422
-60x10
-6
-40
-20
0
Time [s]
20
2.5
62
95x10
-3
90
85
80
75
70
-60x10
-6
-40
-20
20
Time [S]
Figure 5-8 Fast laser scan through the FBG sensor while stationary
Three measurements were taking using the fast scanning laser: the engine stationary
at room temperature, the engine motoring and the engine firing at the same condition used in
the slow scan laser test. A comparison of the three scans can be seen in Figure 5-9. In this
comparison only the scan from 1420.20 to 1432.68 nm is shown since the reverse scan
generated similar results with a measured wavelength difference of 0.02 nm. The scan taken
with the engine stopped at room temperature resulted in a measured temperature of 21.2 0C
with an overall signal loss of 23 dB. The motoring and firing traces had an overall signal
loss of 30 dB and resulted in a temperature measurement of 30.1 and 136 0C, respectfully.
Comparing the running temperature measurement with that obtained from the slow scanning
laser has the running measurement was 13 0C lower.
63
Stationary Power Output
8.E-06
1.8E-06
7.E-06
1.6E-06
7.E-06
1.4E-06
6.E-06
1.2E-06
6.E-06
1.0E-06
5.E-06
8.0E-07
5.E-06
4.E-06
1420
1422
1424
1426
1428
1430
1432
6.0E-07
1434
Wavelength [nm]
Figure 5-9 Comparison of fast scanning laser through engine with FBG sensor
8.E-06
64
6.2 Recommendations
To eliminate most of the problems associated with cycle variation in transmitted light
intensity a redesign is recommended for the pitching side optics. While relying on the
natural dispersion of light to obtain the desired spot size is the simplest, it by no means
provides the flexibility needed to compensate for the motion in an internal combustion
65
engine. In addition the use of a more precise fiber adjustment stage would be helpful. The
manufacturing tolerances, while small prove to be detrimental in aligning the pitching side,
often requiring a long tedious process of trying to located the sensor by gross adjustments of
the X and Y stages. If some sort of adjustment was also allowed for on the collection side, it
may be possible to use a smaller fiber and thus be able to use the OSA for capturing the
transmitted signal.
The specifications of the FBG to be embedded into a piston should be checked before
the embedding process is undertaken. As seen in this research, the Bragg wavelength
reflection was not as strong as claimed by the manufacture. A prescreening operation would
rule out the use of a damaged sensor before undertaking the precise process of embedding the
FBG.
The project as reported in this thesis involved only taking piston temperature
measurement at only one operating condition. In order to further study engine heat transfer a
further investigation should be undertaken to examine the effects of various operating
parameters. As suggested in other literature the effects of engine speed, load and fuel
mixture are thought to be major influences in piston temperature and should be investigated.
While only limited knowledge can be gained by using the current single point temperature
measurement it may prove insightful to overall engine heat transfer.
Certainly the addition of more FBG would be useful to understand heat flows within
the piston. An embedded FBG array would give a more precise account of how temperature
varies spatially within the piston by adding multiple sensing locations. Having this
additional data would allow simulation boundaries to be examined, leading to more accurate
models. While an array of thermocouples would add to the complexity of the data
66
processing, the fast scanning laser has the additional wavelengths available to accommodate
several additional sensors.
In this project the FBG sensor was embedding using a foreign material not normally
used in a piston thus effecting heat flow through the piston. If the FBG embedding process
could be integrating into the original manufacturing process the FBG technique would
provide a more realistic picture of the temperature distribution in an uninstrumented piston.
This would also allow for the most flexibility in placing the FBG sensors in locations that
would be of most interest.
67
7 REFERENCES
1. Kalpakjian and Schmid, Manufacturing Engineering and Technology, Prentice Hall, 2001.
2. Heisler, Heinz, Vehicle and Engine Technology, Society of Automotive Engineers, 1999.
3. Heywood, J., Internal Combustion Engine Fundamentals, New York, McGraw-Hill Inc.,
1988.
4. Gibson, A., Piston Temperature and Heat Flow in High Speed Petrol Engines.
Proceedings of the Institute of Mechanical Engineers 1926.
5. Baker, H. The Operating Temperatures of Cast Iron and Aluminum Pistons in a 12 Inch
Bore Oil Engine. Proceedings of the Institute of Mechanical Engineers 1937.
6. Hepworth, J., Piston Assemblies for Road Transport Oil Engines. Institute of
Mechanical Engineers Automobile Divisional Proceedings 1950.
7. Pinski, F., Temperature Measurement of Internal Combustion Engine Pistons.
Measurement Techniques. Number 7 1964.
8. Underwood, A., Catlin, A., An Instrument for Continuous Measurement of Piston
Temperatures. SAE Transactions, Volume 48, Number 1, 1941.
9. Walker, J., Piston Temperature Measurements in a Watercooled Two Stroke Cycle Spark
Ignition Engine M.S. Thesis, Mechanical Engineering Department, University of
Wisconsin-Madison, 1966.
10. Barna, G., Brumm, D., Anderson, C., An infrared telemetry technique for making piston
temperature measurements SAE Technical Paper 910051, 1991.
11. Burrahm, R., Davis, J., Perry, W., De Los Santos, A., Development of a piston
temperature telemetry system SAE Technical Paper 920232 1992.
68
12. Kato, N., Matsui, M., Shimura, T., Moritsugu J., Piston Temperature Measuring
Technology Using Electromagnetic Induction, SAE Technical Paper
2001-01-2027,
2001.
13. Templug User Information Guide, Revision 3.0 April 2004.
14. R. Measures, R. Maaskant, T. Alaview, G. Tadros, S. Rizkalla, A. Thakurta, Fiber-optic
Bragg gratings for bridge monitoring, Cement and Concrete Composites, 9, pp21-23, 1997.
15. X. Li, F. Prinz, Embedding of Fiber Optic Sensors in Layered Manufacturing, Proc. Eleventh
Solid Freeform Fabrication Symposium, 2000.
16. T. Bosselmann, Magneto- and electro-optic transformers meet expectations of power
industry, Proc. OFS12, pp111-114, 1997
17. K. Nagashima, K. Tsunemichi, K. Tsuchiya, Indicated Mean Effective Pressure Measuring
Method Using Optical Fiber Pressure Sensor 2003-01-2013, 2003
18. Udd, U., ed., Fibre-Optic Smart Structures, Wiley, New York, 1995.
19. X. Li, F. Prinz, Non-contact Strain and Temperature Measurement for Rotating Parts with
Embedded Fiber Optic Sensors, Measurement Science and Technology, 2002.
20. Hill, K., Fuji, Y., Johnson, D., Kawaski, B., Photosensitivity in Optical Fiber Waveguides:
Application to Reflection Filter Fabrication. Applied Physics Letters Vol. 32, 1978, pp.
647-649.
21. Kawaski, B., Hill, K., Johnson, D., Fuji, Y., Narrow-band Bragg Reflections in Optical
Fibers. Optics Letters, Vol. 3, 1978, pp. 66-68.
22. Othonos, A., Kalli, K., Fiber Bragg Bratings: Fundamentals and Applications in
Telecommunications and Sensing, Boston, Artech House optoelectronics library, 1999
69
23. Albert, B., Residual Gas Effects On Combustion In An Air-Cooled Utility Engine, M.S.
Thesis, Mechanical Engineering Department, University of Wisconsin-Madison, 2004.
24. Kranendonk, L.A., Walewski, J.W., Tongwoo, K., Sanders, S.T., "Wavelength-agile sensor
applied for HCCI engine measurements. " 30th International Symposium on Combustion,
Chicago, IL., 2004.
25. Saito, K., Kohama, T., Basaki, M., Suzuki, T. The visualization analysis of the behavior of
piston ring JSAE Technical Paper No. 9941647, 1999.
26. Schneider, E., Blossfeld, D., Method for Measurement of Piston Ring Rotation in an
Operating Engine, SAE Technical Paper 900224, 1990.
27. Brugger, R., Nickel Plating, Teddington, Robert Draper LTD, 1970
28. Durney, L., Electroplating Engineering Handbook, New York, Van Nostrand Reinhold
Company, 1984.
29. Handbook on Electroplating, Birmingham, W Canning & Co. LTD, 1966
30. Iida, Y., Tanaka, K., Fuse, S., Contact-Point Method for Measuring Sliding Face
Temperature and Its Applications, SAE Technical Paper 830311, 1983.
31. King, R.G., Surface Treatment And Finishing Of Aluminum, Oxford, Pergamon Press
1988.
32. Li, X., Embedded Sensors in Layered Manufacturing, PhD Thesis, Mechanical
Engineering Department, Stanford University, 2001.
33. Sandlin, A., Jeskanen, H., Li, X., Bjrklf, A., Manufacturing and Quality Evaluation of
Component Integrated Fibre Optical Sensors For Condition Monitoring,
34. Lowenheim, F.A., Modern Electroplating, New York, John Wiley & Sons, Inc 1963.
70
35. Mierbach, A., Duck, G., Newman, B., Heat Flow Through Piston Rings and Its Influence
on Shape, SAE Technical Paper 831283, 1983.
71