Experimental Study of Automotive Brake System Temperatures
Experimental Study of Automotive Brake System Temperatures
OF AUTOMOTIVE BRAKE
SYSTEM TEMPERATURES
WA-RD 434.1
Final Report
July 1997
~
:'I: Washington State
'f/l Department of Transportation
WA-RD434.1
4. TITLE AND SUBTITLE 5. REPORT DATE
This study was conducted in cooperation with the U.S. Department of Transportation, Federal Highway
Administration.
16. ABSTRACT
Modern automotive disk brake systems can generate extremely high temperatures under high
but short duration braking loads or under relatively light but continuous braking. One consequence of
the high disk and pad temperatures is a gradual heating of the brake hydraulic fluid, which can lead to
boiling of absorbed water and loss of braking.
This paper describes an experiment in which fluid and braking system temperatures, pressures,
and operating conditions (vehicle speed, braking energy dissipation) were measured in three different
classes of operating vehicles during brake application and subsequent brake release. Brake failure was
observed and correlated to moisture content of the fluid, severity of brake application, and application
time. In addition, a finite element and lumped capacity analysis of the system was conducted. The
analytically predicted temperature histories agreed well with the measured temperatures and can be
used to estimate the time at which the fluid will boil.
Brake failure, brake fluid boiling, automotive No restrictions. This document is available to the
brakes, disc brake failure public through the National Technical Information
Service, Springfield, VA 22616
19. SECURITY CLASSIF. (of this repon) 20. SECURITY CLASSIF. (of this page) 21. NO. OF PAGES 22. PRICE
None None 34
Research Report
Research Project T9903, Task 52
Brake Test II
by
Ashley F. Emery
Professor
P. Kumar Joseph C. Firey
Graduate Research Assistant Professor Emeritus
July 1997
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible
for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the official views or policies of the Washington State Transportation
Commission, Department of Transportation, or the Federal Highway Administration.
This report does not constitute a standard, specification, or regulation.
111
lV
TABLE OF CONTENTS
Section Pa1:e
Summary...................................................................................................................... 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Experimental Tests....................................................................................................... 5
Test Results.................................................................................................................. 8
Modeling...................................................................................................................... 18
The Modeling Process.................................................................................................. 22
Conclusions.................................................................................................................. 32
Acknowledgments........................................................................................................ 33
References . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . 34
VJ
SUMMARY
Modern automotive disk brake systems can generate extremely high temperatures
under high but short duration braking loads or under relatively light but continuous
braking. One consequence of the high disk and pad temperatures is a gradual heating of the
brake hydraulic fluid, which can lead to boiling of absorbed water and loss of braking.
This paper describes an experiment in which fluid and braking system temperatures,
pressures, and operating conditions (vehicle speed, braking energy dissipation, etc) were
measured in three different classes of operating vehicles during brake application and
subsequent brake release. Brake failure was observed and correlated to moisture content of
the fluid, severity of brake application, and application time. In addition, a finite element
and lumped capacity analysis of the system was conducted. The analytically predicted
temperature histories agreed well with the measured temperatures and can be used to
1
boiling point. New fluid is generally moisture free, but moisture is thought to diffuse over
time through the flexible hoses or to enter when the system is opened, and the moisture
content increases approximately 1to2 percent (Halderman 1996) per year. Given the data
in Table 1, a minimum acceptable boiling point is 275°F. A study that surveyed a number
of vehicles showed that approximately 25 percent of the vehicles examined had fluid that
boiled at less than 275°F and thus were at risk for brake failure (NHTSA 1973). Figure I
Table 1
Boiling Points
Although many drivers may never experience brake failure due to boiling, the
for vehicle operators and fleet schedulers to be able to estimate the maximum component
temperatures to avoid operating failure or damage to their brake systems. One approach is
to develop a simulation method that can predict brake fluid temperatures as a function of
braking energy dissipation rates and vehicle operation. Unfortunately, brake systems are
too complex to make such predictions on the basis of theory only, and some empiricism
must be employed. This report describes a test program that measured the time histories of
the temperatures of brake components in three typical vehicles. These experimental data
were then used to test a thermal model of the braking system to validate its use in predicting
3
500
450 -. --.~ -.. ...
400
II. 350
•
I
300
250 - ·- ..
200
150
100 V o I u na ' % W a t• r
0 1 2 3 41 5 8
4
EXPERIMENTAL TESTS
Three representative vehicles were instrumented and tested. Details of the vehicles
are given in Table 2. Full details of the instrumentation and the testing procedure are
described in a report by Kumar ( 1997). The vehicles were towed at a constant speed of 30
mph with their brakes applied sufficiently to produce a force equivalent to braking on a
1. The vehicle was weighed and then driven to observe the behavior of the brakes.
During this drive, the maximum deceleration was measured with a G-analyst. (G-
analyst readings of the maximum vehicle deceleration were taken to confirm that the
braking system functioned consistently before and after the test. Good agreement
was assumed to mean that unacceptable pad or rotor degradation had not occurred.)
The instrumentation was then installed, ambient conditions were recorded, the
brake fluid was replaced with a fluid containing 5 percent moisture, and the brake
boost system was disabled. Brake fluid boiling temperatures were measured at the
master cylinder and at each of the wheel cylinders with a device specifically
2. The vehicle engine was started to provide fairly normal air flow within the engine
3. The vehicle was then towed in a straight line until it reached a speed of 30 mph.
4. The brakes were then applied by actuating a ram (an air pressurized
piston/cylinder), which pressed against the brake pedal with a constant force, and
5. AU-turn was made and then the vehicle was stopped. The brakes were left applied
for 2 minutes and then released for 1 minute. During this time, tire pressures were
5
measured and rotor temperatures were measured with a hand-held infrared sensor to
6. Steps 1 through 5 were repeated until brake failure, as indicated by a loss of brake
7. After brake failure, the vehicle was allowed to rest and the brakes to cool.
Intermittently, the brake pedal was depressed (either by the air cylinder or by foot)
and the hydraulic pressures were recorded. This was continued until the hydraulic
pressure recovered, indicating that the vapor had condensed and the brakes were
again functioning.
8. The power boost was reactivated and the brake performance was tested with the G-
analyst.
9. Steps 1 through 8 were then repeated using a different brake fluid, first with 3
percent moisture and then with 0 percent moisture. Each time the fluid was
changed, the entire system was flushed with about 5 times its fluid capacity, and the
boiling point was measured at each wheel cylinder. Rushing continued until the
Table 2
Vehicle Specifications
6
. The instrumentation consisted of the following:
• strain gauge pressure transducers mounted at the master cylinder, each wheel
• thermocouples inserted through the bleed screws into each wheel cylinder as close
to the back face of the piston as possible. Although the SAE (Standard 1291 1992)
specifies that the measurement be taken at the tip of the bleed screw, these
thermocouples were inserted as far as possible into the fluid chamber to measure the
measure the temperatures of the visible portions of the brakes after the vehicle had
stopped
• anemometers to record the speed of air flowing over the brake components .
The raw data were read at a rate of 20 per second, recorded on hard disk, and
7
TEST RESULTS
Figure 2a shows the results for the truck with 5 percent fluid-; these results are
typical of all the tests. The figure shows the towing force, the speed, the front and rear
wheel cylinder hydraulic pressures (both left and right pressures are shown, but they were
almost always equal), and the pressure applied to the ram to indicate when the brakes were
1. The vehicle quickly reached a towed speed of 30 mph and when the ram was
actuated the towing force jumped up to about 300 lbf and then settled down to
2. At the point labeled H, the truck reached the end of the straight section and was
traversed the return straight section and came to a halt. After the vehicle stopped,
the conduction of heat into the components increased the fluid temperature and thus
the pressure in both front and rear brakes, point A. The brakes were then released
for 1 minute.
3. The towing resumed, the brakes were reapplied, and the truck was towed down the
straight section, made another U-tum, and stopped, points I to J'. The brake
application continued for 2 minutes after stopping and then the brakes were released
for 1 minute. Again, heating increased the brake pressure after the truck had
stopped.
4. Another straight section was traversed, points K to L'. This time the truck was
2The different fluids will be termed 5 percent, 3 percent, and 0 percent fluids, referring to the
moisture content. The fluids were supplied by the Union Carbide Corporation.
8
Pressure in Brake Lines, Towing Force, Speed
Truck with 5% Moisture Content Fluid
900
800
Towing Force
( lbs/2 + 600)
-~
QI
I:!
0
600
u.
'[- 500
H J K L
! M
~ 400 Speed
Ill
e
Q. f".__~~(2-*_m_p_h_+~3o_o_>~~~
- 300
.c
Q.
A
.§.
"C 200 Brake Line
QI
QI
F G
Q. Pressure D
Cl)
100
Figure 2a Experimental Pressures, Forces, Speed and Temperatures for the Truck with 0% Fluid
5. Upon towing through another straight section, application of the brakes produced
6. Towing then ceased, but the brakes were reapplied at points E, F, and G to
determine if the vapor had condensed and the pressure had recovered. As the fluid
cooled, the pressure in the front gradually recovered until, at G, the brakes
functioned normally.
expected, each time the vehicle was towed with the brakes applied, the pad and rotor
temperatures showed a marked increase. When the towing stopped, the pad temperatures
decreased precipitously, points L to L', as the rotor cooled rapidly. At point L', the brakes
were released, and although the pad was no longer in contact with the rotor, the rate of
temperature decrease diminished only slightly because the pad was still very close to the
rotor and the high rate of heat transfer to the rotor continued. On the last towing, the failure
of the front brakes resulted in basically no braking force on the front wheels, and thus the
pad temperatures continued to fall, point D. Even though the pad temperature dropped
when towing ceased, the continued conduction of heat into the wheel cylinder caused the
fluid temperatures to rise. The onset of boiling was easily seen at about 1750 seconds,
when the temperature rise was arrested and the fluid temperatures reached a maximum.
The left front (LF) fluid temperature then slowly dropped until about 2500 seconds, when
the vapor began to condense and the temperatures remained essentially constant.
Figure 2c shows the history of the rear brakes. These brakes were characterized by
gentler rises and falls and a lack of boiling. Although the shoe temperatures were high, the
low conductive path from the shoe to the wheel cylinder effectively insulated the fluid from
10
Temperatures for Front Wheels
Truck with 5% Moisture Content Fluid
1200
/Rotor
Inside Pad
1000 LF
L
Inside Pad
800 RF
-!
LL
,__.
,__
...I!::I 600
D
Cl)
Q.
E
Cl)
I-
400
LF Fluid
Outside Pad
200 RF
Rf flui
0
0 500 1000 1500 2000 2500 3000 3500
Time(Seconds)
350.0
300.0
250.0
-
u:-
- -.,
!::::s 200.0
l!
N Q,
E
~
150.0
~- Fluid
- .
100.0
50.0
right rear are on the same system. In this case, boiling occurred at about 1700 seconds in
the right front system and then in the left front system at about 2200 seconds. The left rear-
right front pressure recovered at about 3400 seconds, and both systems had recovered at
about 3700 seconds. The failure of the right front system is clearly shown in Figure 3b at
1700 seconds when the application of the brakes during the towing period did not generate
a braking force at the right front, and the pad temperatures continued to decrease as the
ambient air cooled the brakes. During the next period of towing, because boiling caused
the failure of all the brakes, all the rotors and pads continued to cool. Even in the absence
of heat production at the brakes, and thus a consequent continual cooling of the rotor and
pad, the conduction of heat from the hot pad and rotor to the cylinder caused a continual
increase in fluid temperature, with the result that both fluid temperatures experienced almost
identical histories.
Table 3 lists the boiling points of the different fluids and the measured temperatures
at which boiling was evident. The boiling points were measured with a commercial tester
before and after each test. In the front brakes, the boiling points of the 5 percent fluid
showed no change, but the 3 percent and particularly the 0 percent fluid showed
substantial decreases in boiling point with use. We suspect that the high temperatures that
these latter two fluids were exposed to caused irreversible chemical changes. The rear brake
fluids, which never experienced high temperatures, showed little change. It is interesting to
note the difference between the peak measured temperatures during the tests and the before
and after boiling points. Good agreement was found for the 5 percent fluid, but the
measured peak temperatures were low by about 60 to 80 degrees for the 3 and 0 percent
fluid. There were substantial temperature gradients across the components and through the
fluid, so it is not surprising that the experimental temperatures, being measured slightly
behind the back face of the piston, were lower than the measured boiling points.
13
Pressure in Brake Lines, Towing Force
Car with 3% Moisture Content Fluid
800
700
600
100
1500 4000
-100
Ram ( -2 * psi - 50 )
Time(Seconds)
Figure 3a Experimental Pressures, Forces, Speed and Temperatures for the Car with 3% Fluid
Temperatures for Front Wheels
Car with 3% Moisture Content Fluid
1200
1000
800
Inside Pad
-....e
i'.L
:I
LF
200
0 ._ ...... __
3.50E+02
3.00E+02
--e
LL
2.50E+02
-
0-.,
....:J
f
8.
E
2.00E+02 Fluid-RR
Cl)
I- 1.50E+02
1.00E+02
5.00E+01
recovered. In all cases, this appeared to occur at about 250°F. We do not know what
constituents boil when the fluid is overheated, nor what the chemical compositions and
latent heat of these constituents are, so it is not possible to predict the brake recovery on the
basis of a thermal model. However, it does appear safe to suggest that the brake fluid must
Table 3
Measured Boiling Points (°F)
17
MODELING
One of the major goals of this study was to develop a thermal model that could
predict the temperature history of the brake fluids under a variety of braking scenarios.
Figure 4 shows a typical disk brake rotor and caliper assembly. Clearly, a precise thermal
simulation would require a complex 3D model that could represent both the stationary
caliper assembly and the rotating rotor. Instead, our goal was to determine whether a
simpler model could suffice, given the many uncertainties in braking forces, the variable
In developing the thermal models, besides the usual uncertainties about the thermal
properties and their variation with temperature, there are four major uncertainties:
• What is the state of the fluid in the cylinder, i.e., must we account for convection
• Can the complex shape of the caliper assembly be modeled in a reasonable way,
and what is the effect of the contact resistance between the piston and the steel
• What portion of the energy is absorbed by the rotor and the caliper assembly?
The braking energy that was dissipated at each wheel was determined by assuming
that the braking force was proportional to the pressure in the brake line. For the truck and
the van, the brake failure was recorded in the front brakes, so the constant of
proportionality could be determined for the rear wheels by using the measured braking
energy after failure of the front brakes. With this constant known, it was possible to
determine the fraction of the braking energy that was dissipated by the front brakes when
18
Figure 4 Typical Front Wheel Disk Brake System (from Limpert)
19
The most difficult question to answer concerned the behavior of the fluid. It is
expected that the vibrations induced by a moving vehicle cause a stochastic state of
effects, first because the input data are not available, second because the computation
phase change in a material whose chemical composition and thermal properties in a change
of phase are unknown. From a practical point of view, we were primarily interested in
predicting when the brake fluid would boil and not the thermal behavior after boiling.
In each case, the caliper was modeled by an axisymmetric piston/cylinder with the mass of
the assembly distributed evenly around the cylinder (Figure 5). Although the model was
not geometrically exact, it was judged that the thermal response would be adequately
simulated. This was confirmed by the good agreement described below. The thermal
properties were taken from Limpert (1992) and Incropera and DeWitt (1985). The contact
resistance and its effects were unknown. Calculations performed with a contact resistance
varying from 0 to 220 Btulhr-ft2 -F (Incropera and DeWitt 1985) produced a difference in
peak fluid temperature of only 15°F; therefore, its effect was judged to be small. Previous
models (Burgess 1974~ Wolenweber 1993~ Day, Harding and Newcomb 1999) have
concentrated on modeling the rotor, pad, and average fluid temperatures. In our case, the
goal was for the model to estimate the maximum fluid temperatures.
20
Caliper
Piston
Contact
Resistance
Steel
Backing
(not to scale)
21
THE MODELING PROCESS
First the rotor and the pad/piston/caliper were modeled separately. The finite
element mesh shown in Figure 5, but without the rotor, was used for the pad/piston. A
control volume approach was used to calculate the rotor temperatures. This model
considered the heat lost by convection and radiation from the rotor hub, the ventilated disk,
the rotor rim, and the flat sides of the rotor. All of the convection coefficients were
computed from the correlations given by Limpert ( 1992) that use the measured vehicle
speed to compute the rotor rotational speed. As noted by Limpert the convection
coefficients are difficult to estimate accurately, particularly for such a complex shape. The
adjustments that were found to suffice for the truck brakes, namely increasing the
coefficients by 150 percent when the vehicle was in motion and using a convection
coefficient of 6 Btu/Ft2-Hr-F when it was stationary, were found to apply equally well to
A fraction, F, of the heat generated at the rotor/pad interface was input to the rotor
uniformly over the pad area. The remaining fraction, 1-F, was input to the FEM model.
The thermal histories of the rotor and the pad/piston were computed with different values of
F. The values of F and the convective heat transfer coefficients were varied by trial and
error until the interface temperatures predicted by the two models were equal and matched
the measured temperatures. This procedure was used successfully by Wolak (1985) in
modeling the abrasion of a jet engine caused by the tips of the turbine blades and by Lavine
(1991) in modeling cutting tools. An initial estimate of F can be obtained by considering the
rotor and the pad to be one-dimensional, semi-infinite solids with a heat input at the
interface and by requiring that the interface temperatures be equal. This model gives
F = -V (kpc)rotor Equation 1
'1 (kpc)rotor + '1 (kpc)pad
22
Table 4 lists the fraction of braking energy at the wheel that was applied to the pad
Table4
Fraction of Total Braking Energy Assigned to One Pad/Piston FEM Model
Truck Van
From Equation 1 9.0% 3.9
Separate Models 5.6 4.5
Combined Model 6.4 5.1
material. In the truck the heat quickly conducted through the pad and reached the low
capacitance plastic piston. The effect of the plastic piston was to cause the interface
temperature to rise above that predicted by using the energy computed from Equation 1.
Thus the heat input determined from the FEM was substantially less (Table 4). The van's
piston was steel, and the effect ofthis highly conductive backing was to reduce the
• During the entire test, braking and stationary, there was negligible spatial variation
• The same convection coefficient adjusted to produce good agreement for the truck
• An effective heat transfer coefficient that combined convection and radiation was
nearly constant at one value throughout the braking period. The same was true for
• With the braking energy distributed uniformly over the pad surface area, the
average of the interface temperatures predicted for the rotor and the pad/piston
23
Given these observations, it was possible to
• replace the separate models for the rotor and the pad/piston with the
speed.
In this combined model, the braking energy was applied as a source to the interface nodes,
and the model automatically apportioned the heat to the rotor and the pad. However, a
major question remained. In this model, some of the heat generated at the rotor/pad
interface would be conducted through the pad to the piston, and some would be lost from
the back, unsupported, side of the pad. We assumed, subject to later examination, that the
fraction of the total braking energy that should be used in the combined model was equal to
the fraction of the pad area that was covered by the piston.
Figure 6a compares the predicted and measured temperatures for the truck. Given
the variations in wind speed, vehicle speed, surface temperatures, radiation, and
convection, the agreement was quite satisfactory. The fraction of the braking energy input
to the model was 0.64, and the piston/pad area fraction was 0.69. Figure 6b shows the
predicted and measured fluid temperatures. Because of the low conductivity of the piston,
the peak fluid temperatures were found in the gap between the piston and the cylinder, not
in the chamber behind the piston. Because of the low conductivity of the piston, little heat
reached the rear surface of the piston, and the temperature was relatively isothermal in the
chamber. On the basis of brake failure, boiling was estimated to occur between 1500 and
2000 seconds, which agrees well with the measured boiling point of the 5 percent fluid.
Figure 7a compares the predicted and measured temperatures for the van based
upon an input energy fraction equal to the piston/pad area fraction. In these tests, the pad
thermocouple was recessed slightly below the pad surface, and the peak pad surface
temperatures were considerably higher, as indicated by the curve marked Peak Pad. Figure
7b shows the measured and predicted fluid temperatures. Because of the highly conducting
24
Truck5%
1000
800
.
600 Rotor/Pad (Pred)
--
u.
....
(I)
-....
:J
<U
Cl)
~
400
E
(I)
I-
200
Fluid(Exp)
..
Fluid (Pred)
.
..j
0 Boiling
Energy lnput(Scaled)
500 1000 1500 2000 2500 3000 3500
Time(Seconds)
25
Tru~5% ..(
300
Boiling
B.P. Tester
250
200
..
Fluid (Pred)
50
-50'--~~~--'"~--~--....J.-~~~----.l..-~~~--L~~~-----'-----~--------'
0 500 1000 1500 2000 2500 3000
Time( Seconds)
26
Van0%
1400
1200
1000
-
Q.)
'-
::J
600 Pad (Pred) Right Pad (Exp)
nJ
'-
Q.)
a. 400
E
Q.)
I-
200
0 HBoiling
Energy lnput(Scaled)
-200
-400
0 500 1000 1500 2000 2500 3000 3500
Time(Seconds)
27
Van0%
500
H Boiling
B.P. Tester
400
-100:;---::::::--~L-..-----J-..----...l..---...l....-----L----l
0 500 1000 1500 2000 2500 3000 3500
Time(Seconds)
28
steel piston, the fluid temperatures were greatest in the gap between the piston and the
cylinder and at the rear piston face .. Substantial variations in fluid temperature were
calculated through the thickness of the fluid layer in the chamber. In the experiments it was
not possible to determine the location of the thermocouple, so the curve marked Fluid
(Predicted) represents the nodal temperature that best matched the measured temperature.
This node was used to represent the fluid temperatures in the other van prediction. Brake
failure occurred in the interval marked "Boiling," which agrees well with the time of
Because of the uncertainty about the fluid properties during a phase change, the
model did not account for boiling. Therefore, the predicted temperatures are valid only until
the boiling point is reached. In Figure 7b, when the maximum fluid temperature reaches the
boiling point, a phase change should occur. However, the total energy associated with the
phase change is quite small in this case, and the boiling can be neglected in predicting the
temperature history. This is borne out by the good agreement between the measured and
Figures 7c and d are for tests in which substantial amounts of boiling did occur.
Here ignoring the phase change energy resulted in over-prediction of the system
temperatures. The effect of the latent heat of evaporation was clearly shown by the almost
constant experimental temperature when the brakes were no longer applied. In addition,
the reverse heat liberation when the vapor condensed kept the measured temperature high,
whereas the overestimated model temperatures predicted a high heat loss and thus a too
29
Van3%
400
350
H Boiling
B.P. Tester
300
250
--
e200
u.
-
:J
~
Q)
~150
.,_Q)
100
. .
50
-50
0 500 1000 1500 2000 2500 3000
Time(Seconds)
30
Van5%
350
H Boiling
300 B.P. Tester
250
Fluid (Pred)
50
-SQL-~~~---1.~~~~.......i_~~~~.1.-~~~--1~~~~-'-~~~----'
0 500 1000 1500 2000 2500 3000
Time(Seconds)
31
CONCLUSIONS
model that predicted the average fluid temperature with reasonable accuracy. However,
because of the large temperature variations that exist through the pad/fluid/caliper assembly,
a knowledge of the average fluid temperature is not sufficient to estimate whether and when
boiling will occur. The combined rotor-pad/piston/caliper finite element model can be used
to predict when brake fluid boiling is likely to occur under different braking conditions if
reasonable estimates of convection coefficients and the fraction of braking energy supplied
32
ACKNOWLEDGMENTS
The authors wish to thank the technical monitors, Mr. Jim Shanafelt and Mr. Carl Toney.
In addition, we wish to acknowledge the assistance of the Washington State Patrol Major
Accident Investigation Team members, Sgt. J. Hunter, Det. Scott Cartier, Mr. David
33
REFERENCES
Burgess, I., "Brake Fluid Temperatures Obtained in Alpine Vehicle Trails," SAE Paper
740126, 1974
Day, A. 1., Harding, P. R. J. and Newcomb, T. P., "Combined thermal and mechanical
analysis of drum brakes," Prodc. Inst. Mech. Engr., vol 1980, No. 15, 1984, pp
287-294
Halderman, J. D., Automotive Brake Systems, Prentice-Hall Inc. New Jersey, 1996
Incropera, F. P. and DeWitt, D. P., Introduction to Heat Transfer, John Wiley and Sons,
New York, 1985
Lavine, A. S. and Jen, T. C., "thermal Aspects of Grinding Heat Transfer to Workpiece,
Wheel, and Fluid," ASME J. Heat Transfer, Vol 113, May 1991, pp 296-303
Limpert, Rudolph, Brake Design and Safety, Society of Automotive Engineers, 1992
NHTSA Technical Report "R-11, Field Evaluation of Brake Fluid Moisture Pickup in
Vehicles," 1973
Wolak, J., Zenas, R., Archbold, T. and Etemad, S., "Metallurgical and Mechanical
Phenomena due to Rubbing of Titanium Against Sintered Powder Nichrome",
Wear, Vol 28, pp. 97-103, 1985
34