18-20 June, 2003, Budapest, Hungary OSSKI Center (Törley Palace)

with Exhibition and Pre-Session on Thermal Energy in Hungarian
"THERMO-BRIDGE"
between East and West for technology transfer and information exchange

The heat generated in frictional organs like brakes and clutches induces thermal distortions which may lead to localized contact areas and hot spots developments. Hot-spots are high thermal gradients on the rubbing surface. They count among the most dangerous phenomena in frictional organs leading to damage and early failure. It has been shown that the thermomechanical solicitation due to these hot spots may induce a cycling of tensile and compressive stresses with plastic strain variations¹. Consequently, thermal low cycle fatigue may occur and first results show a relation with the developments of cracks on the disc surface^2,3. These high local temperatures may also lead to unacceptable braking performances such as brake fade or undesirable low frequency vibrations called hot judder⁴.

Experimental observations of hot spots have been reported in many practical applications, particularly in railway and aircraft brakes. Nevertheless, hot-spots occurrence is not well understood up to now in spite of the various approaches presented in previous works^5,6.

In this paper, a classification of the different types of hot spots observed in railway disc brakes is firstly described. Explanation of theses several hot spots phenomena is discussed upon the existing theories. In order to validate the conformity of theses theories with practical observations, a specific experimental test program has been done on a full-scale test bench.

A series of brake drag tests has been done on TGV brake discs. The temperature field on the rotor surface has been measured with an infrared camera coupled with a high-speed data acquisition system. Various friction materials and different geometric combinations have been examined in order to predict their influence on the hot spots development.
Analysis of the experimental observations allows defining scenario of hot spots occurrence from local to macroscopic large thermal gradients.
The conditions of hot spots development seem to agree with a recent approach proposed by authors⁷.

For several years, the increase of railway commercial speed and capacity requires the improvement of braking performances. Even if dynamic braking systems are often largely used in normal service braking, their performances are not sufficient to ensure an emergency braking at high speed. Consequently, friction braking systems are important security systems, which have to match severe criteria imposed by the security rules, in terms of stopping distance associated to a maximum average deceleration, under all sorts of environmental conditions. As an example, in the case of an emergency braking at 300 km.h^-1 of the Thalys TGV, the maximum stopping distance is 3500 m with an average deceleration of 1 m.s^-1 and a braking time of 80 s, corresponding to a dissipated energy of 14 MJ per braking disc.

The trailer bogies of the Thalys TGV include two axles, equipped with four disc braking systems (Fig. 1). Each system is constituted of one disc and two pairs of pads. The disc, with an outer diameter of 640 mm and a thickness of 45 mm, is made of 28CrMoV5-08 steel, manufactured by a forging process.

A TGV disc brake system is tested on a full-scale test bench of Flertex company. The energy to dissipate is provided to the brake disc by mechanical inertia and electric engine.

Observations have been made using thermal imaging equipment provided by Cedip company. Rotor temperature profiles were observed on a thermal imaging screen of the front side of the disc. The infrared system was set to take temperature readings in snap shot mode at a frequency of 90 Hz. It corresponds to one shot every 120° of rotation at the highest rotational velocity used (1788 rpm). Minimum integration time is 12 ms that gives an angular error of 0.13° at the maximal angular velocity.

The experimental apparatus is illustrated in Fig. 2, with on the right the 120° acquisition window. Infrared data acquisition system is described in Fig. 3. The main drawback of this equipment is the sensibility of temperature measurement with the emissivity of the disc. Emissivity is non uniform due to the formation of thin layers of third body on the disc during braking.

A way to estimate the mean disc surface emissivity is to heat the disc to a global uniform temperature and to measure this temperature with inside thermocouples. Global uniform temperature is obtained with continuous brake applications. According to several measurements at various temperatures, the mean emissivity of the disc has been fixed to 0.75.

In order to consider separately the influence of energy and sliding speed, two types of test have been achieved, firstly with increasing speed and constant energy and secondly with both increasing speed and energy. Taking into account the constraints and capacities of the test bench, two programs have been defined:

Railway initial speed from 60 to 300 km/h with constant dissipated energy of 6.4 or 15 MJ per disc,
Railway initial speed from 60 to 120 km/h with dissipated energy from 1.39 MJ to 16.7 MJ per disc.

The brake pads were made of organic resin bounded composite material provided by Flertex company. The pad geometry is rectangular with a surface of 400 cm2. The shape of the pad was optimized in order to satisfy a uniform heating of the disc⁸. As shown in Fig. 4 on the left, the pad is made of three parts :

• metallic integral support
• substrate with low elastic modulus
• frictional material

Three kinds of pad, respectively A, B and C, were tested. The pad A is a commercial standard model. Pad B has been specially developed for these tests. It is the same as pad A but with a softer substrate, with a Young’s modulus of 30 MPa against 90 MPa for pad A. Pad C is the same as A with a reduced length (Fig. 4, on the right). Pads B and C allow to investigate respectively the influence of the pad stiffness and the contact length on hot spots development.

The general test procedure was firstly to define the inertia (mechanical and electrical) to simulate the required energy and secondly to bring the brake rotor to the chosen speed and finally to apply the brake pressure corresponding to the required braking maximal power that decreases linearly with time. After each braking, the disc cooled down to a temperature of 80°C before the next run. Data from infrared camera have been monitored and recorded continuously during the tests. The environmental conditions such as room temperature and humidity were kept constant as far as possible.

5.1 Hot spots classification
From experimental investigations on the rubbing surface of railway brake discs in the testing bench, authors proposed a classification of the hot spots observed (Tab.1).

Tab. 1. Railway disc hot-spots classification

The infrared system was set to take temperature readings in snap shot mode precisely synchronized with the rotation of the disc. These experimental results illustrate the proposed classification (Fig. 5).

Five kinds of hot spots are considered:

• Asperity type results from discrete asperity contacts. Temperature rises rapidly but briefly on very small areas of the rubbing surface.

• Gradients on hot bands correspond to small contact sites which appear along a single rubbing path.

• Hot bands appear as reduced contact friction areas of the pad in the radial direction. It is seen on the disc as narrow rings at high temperatures in the direction of sliding. They can move along the radial direction during braking, according to the evolution of the bearing surface.

• Macroscopic hot spots (MHS) are large thermal gradients regularly distributed on the disc surface. They are fixed on the disc and appear as a buckling pattern of the disc. This phenomenon reduces drastically the contact surface area with high local temperatures.

• Regional hot spots are low thermal gradients on the whole surface of the disc, due to inhomogeneous cooling. Such distributions appear at the end of braking due to thermal diffusion.

The most damaging thermal gradients correspond to types 2, 3 and 4.

Type 4 presents the high thermal gradients, commonly considered as major in the mechanisms of disc failure^3,4. Experimental investigations have been done with simultaneous thermographs of the two sides of the disc. It has shown that the MHS are located alternatively on the two sides of the disc in the direction of sliding (Fig. 6). The anti-symmetrical distribution of MHS and levels of temperature indicate a circumferentially 'buckled' deformation pattern with plastic flow and metallurgic transformations.

5.2 Classical scenario of hot spots evolution

The continuous observation of braking indicates that the scenario of hot spot development is not identical for every tested coupled disc-pad. However among all observed scripts, one of them is more frequently observed, it will be qualified of "classic".

The braking often starts by hot bands formation on the surface of the disc, relatively uniform in the angular direction and that can move radially on the disc (Fig. 7a). Then, small thermal gradients on hot bands can appear. Their angular disposition is very regular and fixed on the disc (Fig. 7b). Next, macroscopic hot spots (MHS) appear on the surface, also very regularly distributed (Fig. 7c). They are fixed on the disc from one braking to the next as described on Fig. 7d corresponding to the braking following the one illustrated from figures 7a to 7c. These MHS are then qualified “stationary”.

5.3 Analysis of the critical braking actions

Critical braking actions correspond to the appearance of MHS. Table 2 summarizes the history of brakings until MHS occurrence, depending on the pad type and test procedure. As detailed previously, pad A is a conventional organic pad, pad B differs from pad A by a softer substrate and pad C is the same as pad A with a 1/3 shorter angular length. For pad A and pad B, tests were made at a constant dissipated energy of 15 MJ and 6.4 MJ respectively. For pad C, tests have been achieved with increased speed and energy. A new disc has been used for each pad.

Tab. 2: Historic braking sequence for each pad-disc couple

For each pad type, the last braking corresponds to the appearance of MHS. Thermal level is given at the time of the highest mean surface temperature. For pads A and B, in order to simulate the required energy at low speed without using a too high simulated mass, braking is separated in two phases. A first braking at constant speed is done followed by the up-to-stop braking.
   Globally, it appears that the MHS occur at relatively low velocity and that the main criterion seems to be the level of energy. Actually for a same initial speed, MHS may occur only if the dissipated energy is high enough (see sequence of pad C). Sequence of pad B seems to indicate that at the same energy level and initial speed, MHS appear if braking power is high enough.
Nevertheless, for a same energy and a high braking power, the fifth braking for pad C did not give MHS. This can be explained by the lower level of temperature when the following braking (6^th) leads to sufficient thermal loading.
   Theses observations lead to define several conditions for MHS appearance: a sufficient level of energy dissipated in a sufficient short time (given by the level of maximal braking power) and conditions giving sufficient thermal loading.
Note that these conditions are amply combined in the case of pad A.

5.4 Influence of the contact length

The influence of contact length has been studied with a shorter pad called C. Its angular arc angle has been reduced from 66° to 44°, so that the contact angular length of pad C is two third of pad A. A series of braking actions have been done at increasing speeds and dissipated energy levels (Tab. 2). For speed varying from 60 to 90 km/h with an energy from 1,39 to 6,25 MJ, only hot bands appeared. During the following braking (at 90 km/h and for 9.38 MJ), 9 macroscopic hot spots occurred on the friction surface, and not 6 as for pad A and B (Fig. 8).
  Such investigation allows proposing correspondence between angular friction length of the pad and hot spots number. In the present case, the contact arc angle of pad C is one third less than pad A and the number of hot spots is one third higher. Even if the ratio of the disc perimeter with angular contact length is not exactly 9 or 6, it should be noticed that effective contact length is commonly less than the angular length of the pad due to thermal distortions and that hot spots number is the first higher whole number in order to ensure a way of continuity of the contact of the pad. 6 or 9, respectively for pad A and C, is close to the minimum number that at least one hot spot is located in the contact zone at all times.

Experimental investigations have been conducted on TGV disc brakes using a full scale test bench. Thermal measurements have been carried out with a high speed infrared camera. The aim of this study was to better classify and to explain the thermal gradients appearance on the surface of the disc. Tests have been done with virgin discs coupled with three types of pads.
A classical script of thermal gradients evolution has been defined. It was commonly observed firstly hot bands occurrence and then thermal gradients on hot bands and finally macroscopic hot spots.
Analysis of the thermographs allows giving new highlights on thermal gradient explanations.
The focal type commonly called gradients on hot bands seems to be associated with thermoplastic instabilities according to the geometry of the pins of the pad. A model of the pad behavior under fretting loading with dynamical effects coupled with the thermoelastic response could be representative.
Upon the macroscopic hot spots, experiments allowed to precise the condition of their appearance, strongly depending on the thermal loading due to the level of dissipated energy and the maximal braking power. The observations are in good agreement with the progressive waviness distortion approach proposed by the authors.

The authors are pleased to express their gratitude to the companies Flertex S.A. and Valdunes who provided, respectively, the test bench, and the brake discs.

Fig. 1. TGV braking disc

Fig. 2. Flertex full scale test bench

Fig. 3. Infrared acquisition system

Fig. 4 : Flertex half-pad

Fig. 5. Thermal gradients classification illustrated by thermographs.

Fig.6. Simultaneous thermographs of the 2 sides of the disc.

Fig. 7. Classical scenario of macroscopic hot-spots development

[1] P. Dufrénoy, D. Weichert, “A thermomechanical model for the analysis of disc brakes fracture”, Thermal Stresses 2001, Osaka, 2001.
[2] A.E. Anderson, R.A. Knapp, “Hot spotting in automotive friction systems”, Wear, 135, 319-337, 1990.
[3] P. Dufrénoy, G. Bodovillé, G. Degallaix, “Damage mechanisms and thermomechanical loading in brake discs”, Temperature-Fatigue interaction, L. Rémy & J. Petit , 167-176, 2001.
[4] T. Kao, J.W. Richmond, A. Douarre, “Brake disc hot spotting and thermal judder: an experimental and finite element study”, Int. J. Of Vehicle Design, 23, 276-296, 2001.
[5] X. Fan, H. Lippmann, “Elastic-Plastic Buckling of Plates under Residual Stress”, Advances in Engineering Plasticity and its Applications, Pergamon Press, Amsterdam, 1996.
[6] K. Lee, J.R. Barber, “Frictionally excited thermoelastic instability in automotive disc brakes”, J. of Tribology, 115, 607-614, 1993.
[7] S. Panier, P. Dufrénoy, D. Weichert, “Macroscopic hot-spots occurrence in frictional organs”, Thermal Stresses 2001, Osaka, 2001.
[8] N. Benseddiq, D. Weichert, J. Seidermann, M. Minet, “Optimization of design of railway disc brakes pads”, Proc. Inst. Mech. Eng. Part F, 210, 51-61, 1996.

.

OSSKI Center (Törley Palace)
"Fodor József" National Center of Public Heath
"Frédéric Joliot-Curie" National Research Institute for
Radiobiology and Radiohygiene. (OKK-OSSKI)
www.osski.hu