Tyre/road friction modeling

Literature survey

R. van der Steen DCT 2007.072

Coaches:

I. Lopez B. de Bruijn A.J.C. Schmeitz H. Nijmeijer

Supervisor:

Eindhoven University of Technology Department of Mechanical Engineering Dynamics and Control group Eindhoven, May, 2007

Contents
List of symbols 1 Introduction 2 Materials 2.1 2.2 Elastic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Visco-elastic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 7 11 11 13 15 15 16 18 21 22 26 31 31 31 32 32 35 35 36 39 39 42 43 44 44 47 50

3 Friction models and surface roughness 3.1 3.2 3.3 Amontons-Coulomb friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Friction of viscoelastic materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface roughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 Finite Element Method 4.1 4.2 Contact and friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Literature related to FEM and tyres . . . . . . . . . . . . . . . . . . . . . . . . . .

5 Validation and modeling methods 5.1 5.2 5.3 5.4 LAT1OO and Tekscan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eplexor and DIK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flat plank and tyre test trailer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tyre modeling in ABAQUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 Conclusions 6.1 6.2 Friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . FEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A 32th Tire Mechanics Short Course A.1 Strength, wear and friction of rubber. . . . . . . . . . . . . . . . . . . . . . . . . . A.2 The tire as a vehicle component. . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.3 Tire materials and manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . . . . A.4 Rules and regulations governing tires. . . . . . . . . . . . . . . . . . . . . . . . . . A.5 Advanced tire modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.6 Tire wear, traction and force generation. . . . . . . . . . . . . . . . . . . . . . . . . A.7 Tire stresses and deformation analysis. . . . . . . . . . . . . . . . . . . . . . . . . .

Bibliography

53

List of symbols
Symbol cr cr k f r s T aT ci d f gN gT gT h k l pN tT z Unit [rad] [m] [m/s] [rad] [] [-] [m] [-] [-] [-] [-] [-] [-] [Hz] [rad/s] [-] [] [m] [Hz] [m] [m] [m/s] [-] [-] [m] [Pa] [Pa] [m] Description Slip angle Elastic slip Elastic slip rate Loss angle Field variable Longitudinal slip Length scale Poissons ratio Coecient of friction Kinetic friction coecient Maximum friction coecient Rolling friction coecient Static friction coecient Segmental jump rate at temperature T Spinning velocity WLF shift factor Constant Mean plane distance Frequency Normal gap distance Relative tangential displacement Relative tangential velocity Fit parameter Material constant Characteristic surface length Normal pressure Shear stress Distance to wheel center

Symbol A E E Fn Fx Ft Fr G G G G K M Mz R T Tg Vs W B C D E Ff G

Unit [m2 ] [Pa] [Pa] [N] [N] [N] [N] [Pa] [Pa] [Pa] [Pa] [Pa] [N/m] [N/m] [m] [K] [K] [m/s] [N] [-] [-] [-] [-] [-] [Pa/m]

Description Contact area Youngs modulus Complex Youngs modulus Normal force Longitudinal force Tangential force Rolling friction force Shear modulus Dynamic shear modulus Dynamic loss modulus Complex shear modulus Bulk modulus Torque Self aligning moment Radius Temperature Glass temperature Sliding velocity Normal load Magic formula coecient Magic formula coecient Magic formula coecient Magic formula coecient Slip tolerance Penalty coecient

Chapter 1

Introduction
This literature report has been written in the scope of the CCAR project FEM Tyre Modelling. This project aims to realize a thorough base in modeling of a tyre using a nite element method. An import aspect is the development of a robust friction model to include tyre/road interaction. The report has been written to obtain a basic understanding in the eld of (rubber) friction and the nite element method. An overview is presented with respect to friction models and surface roughness, contact and friction in the nite element method, and tyre analyses using the nite element method. It can therefore be used as a reference guide for future steps within the CCAR project. Since the patent on a pneumatic tyre, by John Boyd Dunlop in 1888, the development of pneumatic tyres started and is still going on today. The worldwide tyre production is now greater than 1 billion units per year with a market value exceeding US $100 billion. Around 73% are passenger cartyres and 70% of these tyres are replacement tyres. There are basically two types of tyres, the bias ply and, since 1948, radial ply tyres. Nowadays most of the passenger car tyres are radial tyres. The main dierence is the orientation of the body plies, see gure 1.1.

(a) Bias tyre.

(b) Radial tyre.

Figure 1.1: Dierence between bias and radial tyres (picture taken from Gent (2007)).

The tyre construction, such as aspect ratio, belt construction and tread compound, depends on the size and target market (speed rating). This information is printed on every tyre, e.g., 215/55 R 16 97 V. The rst number (215) is the nominal section width in mm, the second number (55) is the percentage of the height/width ratio of the cross-section. The R (radial) stands for the tyre construction code. Then the rim diameter (16) is given in inches and 97 is the load capacity index, which indicates the maximum load capacity. The last symbol (V) is the speed

symbol, which stands for the maximum speed. On winter tyres an additional logo is printed together with the mud and snow (M&S) symbol. See table 1.1 for an overview of load indices and the speed symbols for cars.

Table 1.1: Load indices and speed symbols for car tyres. Load index 65 66 67 68 69 70 71 72 73 74 75 76 77 78 Load (kg) 290 300 307 315 325 335 345 355 365 375 387 400 412 425 index 79 80 81 82 83 84 85 86 87 88 89 90 91 92 kg 437 450 462 475 487 500 515 530 545 560 580 600 615 630 index 93 94 95 96 97 98 99 100 101 102 103 104 105 106 kg 650 670 690 710 730 750 775 800 825 850 875 900 925 950 index 107 108 109 110 111 112 113 114 115 116 117 118 119 120 kg 975 1000 1030 1060 1090 1120 1150 1180 1215 1250 1285 1320 1360 1400 Symbol K L M N P Q R S T U H V W Y Speed (km/h) 110 120 130 140 150 160 170 180 190 200 210 240 270 300

Basically a tyre should carry load and transmit forces. However the design of a tyre is a complex trade-o problem, because a tyre needs to perform under varying conditions. In gure 1.2 the main components of a radial tyre can be seen. All these dierent components provide specic properties. For example the tread should have good traction for all weather conditions, low rolling resistance and low wear. But low rolling resistance and a long lifetime require a hard compound, which has less traction. For good wet traction groves in the tread are required to drain o water, but this means less dry traction since the contact area is decreased. And softer compounds, which provide more traction than harder compounds, also wear faster. Besides the engineering aspects, economical factors, (limitations of) the manufacturing process (Walter, 2007a) and government regulations (Walter, 2007b) have to be taken into account. To model a tyre and the interaction with the environment dierent modeling approaches can be used: analytical models, empirical models or physical models. In the following the focus is on physical models to be used in a Finite Element Analysis (FEA). Information of the used materials and geometry of the tyre and an interaction with the road is required for the Finite Element Method (FEM). The basic materials used in tyres and material models are described in chapter 2. In chapter 3 an overview of dierent friction models is given and the relation with surface roughness is shown. Chapter 4 covers the basics of FEM and numerical methods related to contact. Some aspects of contact in ABAQUS are mentioned and an overview of dierent types of tyre problems is given. Validation possibilities are given in chapter 5 and nally conclusions are presented in chapter 6. An additional appendix, which is a short summary of the 32th tire mechanics short course is also included.

1. INTRODUCTION

Figure 1.2: Components of a radial tyre (picture taken from Gent (2007)).

10

Chapter 2

Materials
Dierent materials are used in tyres. The rubber compounds are the most complex parts of the tyre. The principal tyre elastomers are natural rubber (NR), polyisoprene (IR, this is synthetic NR), styrene butadiene rubber (SBR), butyl rubber (IIR) and butadiene (BR). A compound can be dened as the formulation of a mixture of rubber and additives which meets the needs of the tyre component application. Usually a compound consists of one or more polymers, vulcanizing agents, accelerators, reinforcing llers, antidegradents, plasticizers, softeners and tackiers (used to bond pieces together). This explains why the properties of a compound can deviate from a pure elastomer. Bead wires are usually made of steel. Belt ply cords consist of steel or rayon, while body plies cords can be made of nylon or polyester. For a FE model the geometry of the parts, the applied loading and boundary conditions and the material behavior of each of the dierent materials is required. In this chapter the basics of elastic and visco-elastic material behavior are mentioned.

2.1
2.1.1

Elastic
Small strains

For small strains the material properties can be dened by the shear modulus G and the modulus of bulk compression K, which are related to the tensile or Youngs modulus E and Poissons ratio as follows: E = = 2(1 + )G, 1 3K 2G . 2 3K + G (2.1) (2.2)

When a material behaves elastically, there is no permanent deformation and dissipation. This means that if the stress is released the strain will become zero. Furthermore there is no time or path history, thus the stress can be calculated if the strain is known. The strains in the belt and body plies cords, under normal conditions, are in the linear elastic range, which means that they can be described by either G and K or E and .

2.1.2

Large strains

Rubber is an elastic, soft and virtually incompressible material, but it can usually be stretched more than 500%. Its molecular structure consists of long, linear exible molecules forming random coils. The molecular segments are mobile, so-called Brownian motion (rate of segmental jumps,

12

2.1. ELASTIC

Figure 2.1: Polymer properties as function of temperature (picture taken from Gent (2007)).

which depends strongly on temperature), and the molecules are interlinked into a 3D network. Chemical crosslinks are usually made by sulfur linkages also known as vulcanization. The rubbery state of a polymer is determined by the so-called glass transition temperature Tg . If the temperature is above Tg the polymer is rubbery, below Tg the polymer is glassy, see gure 2.1. For rubbers the Tg is below room temperature. For rubber typically K G and it follows that E = 3G and = 0.5. If a rubber block with = 0.5 is compressed, it is sheared outwards and generates an internal pressure. This eect causes a compression modulus, which is generally greater than E. To model this behavior other constitutive models have to be used. Hyperelastic material models are based on energy functions, which dene the strain energy stored in the material per unit of reference volume, instead of the Youngs modulus (E) and poisson ratio () for isotropic linear elastic materials. In ABAQUS several hyperelastic material models are available, an overview can be seen in table 2.1. A specic material model is usually chosen based on available test data, however it is also possible to enter specic material constants directly. It is possible to t material models on experimental data, such as uniaxial, biaxial, planar and volumetric test data. For more details see the ABAQUS Analysis users manual, chapter 17.

2.1.3

Filled rubbers

In tyres rubbers are usually lled with particles like carbon black or silica. This creates two additional eects. The so-called Payne eect and Mullins eect, see e.g. Diani et al. (2006), which are both softening eects. The Payne eect is a softening eect for small strains (0.1%), attributed to breaking apart aggregates of ller particles. The Mullins eect is a substantial softening eect at higher strains, attributed to progressive detachment of rubber molecules from ller particles. The Mullins eect can be modeled in ABAQUS, but it cannot be used together with visco-elasticity or hysteresis options. In Bergstrom (2005) a new constitutive material model is presented, which is compared with the standard ABAQUS material models. In Becker et al. (1998) a material model

2. MATERIALS

13

Table 2.1: Available material models in ABAQUS. Material model: Neo-Hookean Mooney-Rivlin Ogden Marlow Arruda-Boyce Yeoh Van der Waals Polynomial Reduced polynomial User dened

for lled rubbers is presented, which is implemented into an in-house FE code and compared to experiments.

2.2

Visco-elastic

The constitutive laws for large strains cannot be used to fully describe the stress-strain relation, since rubbers do not follow reversible stress-strain relations. When rubber is dynamically stretched and released the returned energy is less than the energy that is put into the rubber. This viscoelastic eect cannot be described by the perfect elastic shear modulus G, but a dynamic shear modulus G and a dynamic shear loss modulus G is introduced to describe this hysteresis. Another term that is frequently used is the loss angle, which is dened as tan = G . For an elastic solid G the strain is in phase with the stress, while for a visco-elastic solid the strain lags behind the stress with a delay of , where is the frequency (rad/sec). The universal WLF relation, named after Williams, Landel and Ferry, (Williams et al., 1955) describes the dependence of the segmental jump rate on temperature, log 17.5(T Tg ) T = , Tg 52 + T Tg (2.3)

where Tg is dened as 0.1/sec. Some polymers, such as butyl rubber, behave according to a dierent relation T 17.5(T Tg ) = . (2.4) log Tg 100 + T Tg This is possibly due to unusually long moving segments. The WLF relation can by used to predict results at other temperatures, e.g. the frequency f at temperature T is equivalent to a frequency f at T , using the following relation T f = = aT , f T (2.5)

where aT is known as the shift factor, this equivalence also holds for strain rates or velocities. To create master curves of material properties of most (pure) polymers this transformation can be used. Normally experiments can only be done for a limited range of frequencies. If the same experiments are conducted at dierent temperatures, a master curve can be made for a chosen reference temperature by shifting the measurements using (2.5). An example is shown in gure 2.2. Details of a FE implementation of 3D visco-elastic behavior for small and nite strains can be found in an article of Kaliske and Rothert (1997).

14

2.2. VISCO-ELASTIC

Figure 2.2: Creation of master curve for E using WLF shifts (picture taken from Gent (2007)).

Chapter 3

Friction models and surface roughness

In this chapter several friction models are discussed. Friction is present in all mechanical systems and originates at an interface between bodies, which are in contact with each other. This contact enables solids to transmit forces through the surfaces. The resultant force can be resolved into a normal and a tangential force. This is referred to as the normal force and the friction force respectively. Friction can now formulated as the ratio between the tangential and normal contact force. Forces can only be transmitted if there is contact, therefore the contact area plays an import role in describing friction aspects. Several theories have been developed to describe surface roughness and the relation to the real contact area. It is well-known that contaminations have a strong eect on friction, e.g. lubricants. One of the classical works on friction and lubrication is the book of Bowden and Tabor (2001).

3.1

Amontons-Coulomb friction

This is the most used friction model, usually referred to as the Coulomb friction model. It is based on the two basic laws of friction (Bowden and Tabor, 2001). When a block is pressed upon a plane with a normal force and is pulled in the horizontal direction, there is a threshold value such that the block starts to move if the tangential force is bigger than this value. This threshold value is proportional to the normal force Ft = Fn , with the so-called coecient of friction. The friction force always works in opposite direction of the moving velocity. Coulomb also showed that kinetic friction could be signicantly lower than static friction, which is sometimes referred to as the third friction law.

3.1.1

Deviations from Coulomb friction

Experiments often show deviations from the basic Coulomb friction model. Variants of the basis model are the dierence between a static s and a kinetic k friction coecient, see gure 3.1 for some examples. This transition can be discontinuous or continuous, using an exponential decaying function between the static and kinetic coecient. More general, friction varies with sliding velocity. An example is the so-called Stribeck curve, rst the coecient decreases to a minimum value and after that it increases with higher sliding velocities. The review paper of Olsson et al. (1998) describes several static and dynamic friction models. The static models include Coulomb friction, viscous friction and the models of Karnopp, where a zero velocity interval (deadzone) is introduced, and Armstrong, where stiction and sliding are decoupled in two separate equations. The dynamic models include the reset integrator model, the Dahl model and two variants of this

16

3.2. FRICTION OF VISCOELASTIC MATERIALS

s k

s k

s k

Vs

Vs

Vs

Vs

Figure 3.1: Coulomb friction model and some variations.

model, the Lugre model and the model of Bliman and Sorine. More details on the transition between sliding and sticking can be found in an article of Rabinowicz (1951). All of these models are often used in control systems, in such a way that they can be used for friction compensation.

3.1.2

Magic Formula

In the book of Pacejka (2006) the empirical magic formula tyre model is explained. This model is widely used for vehicle handling simulations and based on the magic formula F (x) = Dsin(C arctan((1 E)Bx + E arctan(Bx))), (3.1)

where x is an input, e.g. or and F an output such as Fx or Mz . The coecients B, C, D and E are tted using experimental data. This model can accurately capture the global force and moment characteristics of a tyre. Another frequently used model is the brush model, more details can be found in the sheets of Besselink (2003) and the book of Pacejka (2006). This model is useful for a qualitative analysis of tyre behavior when limited measurement data is available.

3.2

Friction of viscoelastic materials

Friction of polymers is closely related to their viscoelastic behavior. Generally speaking, the coecient of friction increases with sliding velocity until a maximum value is reached at a certain speed, followed by a decrease of the friction coecient. This is caused by the exibility of the polymer chains.

3.2.1

Grosch

The article of Grosch (1963) is one of the early reports, which describes a relation between the dependence of the friction coecient due to sliding velocity and temperature and the viscoelastic deformation on strain rate and temperature. In this work experimental results of dierent vulcanized rubbers are presented. Experiments are conducted at dierent temperatures and shifted to a master curve using the WLF equation. This gives a relation between the sliding velocity and temperature in a similar way as the rate and temperature dependence of viscoelastic polymers above their glass temperature, see gure 3.2. Grosch indicated that friction is due to energy dissipated when rubber is compressed and released by asperities. For dry sliding on a smooth surface the curve for vs. sliding speed resembles again the curve for tan versus frequency f . But now it is displaced to much lower speeds than before. Friction in dry sliding on a smooth surface is due to energy dissipated as rubber sticks and slips on a molecular scale. In the case of sliding on a rough dry surface the two dissipative processes appear at dierent speeds, corresponding to

3. FRICTION MODELS AND SURFACE ROUGHNESS

17

dierent length scales: asperities and molecules, see gure 3.3. Several contributions to can arise at the same sliding speed from stick-slip processes occurring at dierent length scales. An early theory to describe the experiments of Grosch has been developed by Schallamach (1963). An unied approach to the entire subject of friction and wear in rubbers and tyres can be found in an article of Moore (1980).

Figure 3.2: Grosch interpretation (picture taken from Gent (2007)).

Figure 3.3: Combined eect of sliding on a rough and dry surface (picture taken from Gent (2007)).

3.2.2

Savkoor

Savkoor proposed a model, based on results of Grosch, where he described the shape of an isothermal master curve with an empirical relation (Lemaitre, 2001, chapter 8), k (Vs ) = s + (m s )exp h2 2 log2 |Vs | Vmax (3.2)

where s is the static coecient of friction and m the maximal coecient for |Vs | = Vmax . h is a dimensionless parameter reecting the width of the speed range in which friction varies signicantly. More details on dry adhesive friction can be found in (Savkoor, 1965, 1966, 1986) and in the thesis of Savkoor (1987). The advantage of (3.2) is that it can be evaluated easily for a given sliding velocity.

18

3.3. SURFACE ROUGHNESS

3.2.3

Rolling friction

Rolling friction Fr can be seen as M , where M is a torque and R the radius of the rolling body. R For an elastic wheel the reaction force acts through the wheel center and therefore the moment M = 0. For a visco-elastic wheel the reaction force moves towards the loading side, because stresses are lower in unloading, M = W z where W is the normal load and z the distance from the wheel center, see gure 3.4. Displacement of the center of pressure, and hence the coecient of friction r , depend on speed and temperature in the same way as the loss factor tan . A formulation for r as proposed by Greenwood et al. (1961), 1 3 W tan , (3.3) r = 0.33 R2 G where G is the complex shear modulus. Rolling friction is mainly due to energy dissipated as rubber is compressed and released in the contact patch.

Figure 3.4: Rolling friction (picture taken from Gent (2007)).

3.2.4

Persson

Persson has published many papers (Persson, 1993, 1995, 1997, 1998, 1999, 2001a,b, 2002a,b; Persson and Tosatti, 2000; Persson and Volokitin, 2000, 2002, 2006; Persson et al., 2002, 2003, 2005) on the subject of rubber friction and the role of the surface, which is in contact with the rubber. These theories are a continuation of the early studies of Grosch, taking into account that sliding friction of rubber has the same temperature dependence as that of the complex elastic modulus. He states that the friction force (under normal circumstances) is related to the internal friction of the rubber, which is a bulk property. The hysteretic friction component is determined by sliding of the rubber over asperities of a rough surface. These oscillating forces lead to energy dissipation. These contributions can be described with a fractal description of the surface. Every length scale , up to the largest particles of asphalt, can be related to a frequency: f Vs . As a result of energy dissipation heat is generated, in a recent article (Persson, 2006a) he also takes the local heating of rubber into account, since the viscoelastic properties are strongly temperature dependent. At very low sliding velocities the temperature increases are negligible because of heat diusion, but already for velocities in the order of 0.01 m/s local heating plays a role. He shows that in a typical case the temperature increase results in a decrease in rubber friction with increasing sliding velocity, if the sliding velocity is above 0.01 m/s.

3.3

Surface roughness

There is a long history on this subject, since it is well-known that almost all surfaces are rough at a certain length scale. This has great inuence on the real contact area and therefore it is closely related to friction. The real contact area is in most cases dierent from the apparent contact area

3. FRICTION MODELS AND SURFACE ROUGHNESS

19

and high local contact pressures can give rise to plastic deformation. The two often used methods to describe surface roughness are statistical and fractal methods. In Archard (1957) it is shown that in order to obey the Coulomb friction law for elastic deformation a rough surface is actually needed. In elastic contact the area A is related to the load W by A kW 3 ,
2

(3.4)

where k is a constant. It is supposed that the friction force is proportional to the contact area, therefore it can not hold that the friction force is proportional to the load. For increasing load the friction coecient has to decrease. To show this, Archard used a at surface in contact with a spherical surface with large radius. However if instead of one spherical contact, several spherical protuberances on this initial sphere are modeled and thus multiple contact areas are created, the following relation is derived 44 (3.5) A W 45 . He showed that is now constant for an increasing load until all the local protuberances are at and than drops again if the load is still increased further. This explains that for soft rubber sliding on a smooth surface (perfect contact) the frictional force is more or less constant, independent of the load W . Thus the coecient of friction decreases continuously as the pressure increases. However there is also a coecient of friction, independent of pressure, for harder rubber compounds, sliding on a rough surface. Apparently in this case contact is incomplete. An increase in pressure creates a larger true area of contact and hence a larger frictional force. The Greenwood and Williamson (GW) contact model (Greenwood and Williamson, 1966) was published in 1966 and since then many revisions to this model have been made by several researchers. It is now known that some of the assumptions of the original model were incorrect (Greenwood and Wu, 2001). The GW model is based on contact between a rough surface and a at rigid plane. It is a statistical description of a rough surface, where the surface is covered with asperities. All these asperities have a spherical shape with the same radius of curvature. Further the summit surface density, the surface roughness and the surface height need to be specied. The surface height is supposed to follow a Gaussian or exponential distribution. Relations for the number of contacts, the real contact area and the applied load are obtained. The model of Bush, Gibson and Thomas (BGT) (Bush et al., 1975) uses elliptical paraboloids asperities with random aspect ratio and orientation, a comparison between these models can be found in an article of McCool (1986). Fractal descriptions to describe a rough surface rst appeared in the papers of Majumdar and Bhushan (1990, 1991) and Majumdar and Tien (1990). If a surface exhibits geometrical selfsimilarity, also referred to as self-anity, the surface appears the same under dierent degrees of magnication. This can be described by the Weierstrass-Mandelbrot function, which satises continuity, nondierentiability and self-anity. The advantage of this approach is that the parameters, which determine the isotropic rough surface, are independent of the roughness length scale and the resolution of the measuring instruments. This is in contrast to statistical parameters where the height, slope and curvature of asperities depend on the the resolution of the measuring instrument. Fractal models are now widely used and compared by a great number of authors, e.g. Klppel u and Heinrich (2000); Gal et al. (2005); Morrow (2003); Jackson and Streator (2006); Ciavarella et al. (2006a,b); Zavarise et al. (2007). Besides his work on rubber friction, Persson has also made a large contribution to the eld of contact mechanics (Persson, 2002a; Persson et al., 2002; Peressadko et al., 2005; Persson, 2006b; Yanga et al., 2006; Persson, 2006c). An extensive overview of contact mechanics and rubber friction can be found in a review paper of Persson (2005). A comparison of experiments and the predictive capabilities of current physical theories is presented in Westermann et al. (2004). This demonstrates the close interaction between rubber friction and surface roughness.

20

3.3. SURFACE ROUGHNESS

Chapter 4

Finite Element Method

The development of non-linear Finite Element Method (FEM) started about forty years ago together with the development of the digital computer, with one of the rst text books of Zienkiewicz and Cheung (1967). Although there are many books related to FEM nowadays, this book or later editions are still cited in almost every article related to FEM. There are two distinctive methods to solve FE problems, implicit and explicit. The implicit method solves for a statical or dynamical equilibrium, while the explicit method solves transient dynamic response problems using an explicit direct-integration procedure. Besides the geometrical, material and interaction issues additional computational problems arise. Since the start of non-linear FEM many dierent types of elements and solution algorithms have been developed and implemented in specialized software such as ABAQUS, MARC, ANSYS, LSDYNA or NASTRAN. This raises the question why one should study non-linear FEM if it is already available in many commercial packages? The answer to this is that a commercial package is basically a black box program providing an interface for simulations. The user has to make several choices in the model formulation, i.e., the discretization of the mesh, the element types, the solution procedure, the solver tolerances etc. These selections (may) have a tremendous eect on the nal result and can even lead to non-physical, thus meaningless, results (see Kouznetsova, 2006). There are four dierent sources of non-linearity in solid mechanics, geometrical, material, force boundary conditions and displacement boundary conditions, see also gure 4.1. Unfortunately all four play a role in tyre modeling. One example is incompressible material behavior of rubber. When the material response is incompressible, the solution to such a problem cannot be obtained in terms of the displacement history only, since a purely hydrostatic pressure can be added without changing the displacements. In ABAQUS this is solved by removing this singular behavior in the system by treating the pressure stress as an independently interpolated basic solution variable, coupled to the displacement solution through the constitutive theory and the compatibility condition, with this coupling implemented by a Lagrange multiplier. This independent interpolation of pressure stress is the basis of the so-called hybrid elements. More precisely, they are mixed formulation elements, using a mixture of displacement and stress variables with an augmented variational principle to approximate the equilibrium equations and compatibility conditions. The hybrid elements also remedy the problem of volume strain locking. Volume strain locking occurs if the nite element mesh cannot properly represent incompressible deformations. Volume strain locking can be avoided in regular displacement elements by fully or selectively reduced integration (ABAQUS Analysis

22

4.1. CONTACT AND FRICTION

users manual).

Figure 4.1: Schematic view of non-linearities in solid mechanics (picture taken from Kouznetsova (2006)).

Two aspects are considered here, the mathematical formulations of contact and friction and implementation issues in ABAQUS, and computational models for tyres found in literature to give an overview of current possibilities.

4.1

Contact and friction

Starting points for this subject are the article of Doghri et al. (1998), the book of Laursen (2002) and especially the book of Wriggers (2006) on computational contact mechanics. Contact is split into normal contact and tangential contact.

4.1.1

Normal contact
pN

gN

Figure 4.2: Schematic view of Signorini conditions.

Normal contact can be formulated as a geometrical constraint, which describes the non-penetration condition of two bodies in contact, also known as low contact precision. This is the so-called Signorini contact description, which models the non-penetrability of the bodies in contact with the following relations. For contact it holds that gN = 0 and pN < 0, where gN is the gap distance and pN is the normal pressure. If there is a gap between the bodies, than the pressure is equal to zero, see gure 4.2. This can be formulated as gN 0, pN 0, pN gN = 0. (4.1)

4. FINITE ELEMENT METHOD

23

This formulation states that only compressive forces can be transmitted if there is contact (no adhesion) and forms the basis of frictionless contact problems. If for a contact problem knowledge of the micromechanical surface is essential to describe the physical phenomena (high contact precision) an interface law is needed to capture this phenomena. Instead of a geometrical constraint a constitutive equation is used, where the contact pressure is given by a general non-linear function of the mean plane distance pN = f (d). An example, taken from Wriggers (2006), of this general non-linear function: pN = c1 (1617646.152 c3 )c2 c4 5.5891+0.0.0711c2 exp 1 + 0.0.0711c2 2 d , (1.363c3 )2 (4.3) (4.2)

where c1 and c2 are mechanical constants which express the nonlinear distribution of the surface hardness and c3 and c4 are statistical parameters of the surface prole.

4.1.2

Tangential contact

Usually friction is present between two interacting bodies and this leads to tangential forces. Tangential contact is described by constitutive laws. The response in tangential direction can be divided into two cases. The rst case is no tangential displacement in the contact zone (stick) and in the second case there is relative displacement (slip). Stick corresponds to a situation where the relative velocity (gT ) is zero and thus a zero relative displacement gT = 0 gT = 0, (4.4) which poses a constraint on the motion in the contact interface. If the tangential forces reach a certain limit the contacting bodies will no longer stick, but sliding will occur. In the case of Coulomb friction this can be formulated as tT = |pN | where tT is the shear stress. Since the friction coecient usually depends on more parameters, a constitutive law can be written as a non-linear function of several parameters, = (gT , pN , , . . .), where is dened as a eld variable, e.g. temperature. An example from Wriggers (2006) of a velocity and pressure dependent constitutive relation for rubber friction is the following (vT , pN ) = c5 pN c6
c7

gT , ||gT ||

if ||tT || > |pN |,

(4.5)

(4.6)

+ c8 ln

Vs Vs c10 ln . c9 c11

(4.7)

Besides the Coulomb law (4.5), it is possible to formulate tangential constitutive equations by regularization of the stick-slip behavior. This is to avoid the non-dierentiability of Coulombs law at zero velocity, which can prevent numerical problems, some examples of this are square root, hyperbolic tangent of piecewise polynomial regularizations. Other formulations are in the framework of elasto-plasticity, where the key idea is to split the tangential slip in an elastic (stick)

24

4.1. CONTACT AND FRICTION

part and a plastic (slip) part. Articles of Huemer et al. (2001a); Liu et al. (2000) and Huemer et al. (2001b) are related to frictional behavior of rubber, where a phenomenological friction law for rubber tread blocks on snow and concrete is presented for use in a macroscopic model. The coecient of friction depends on normal pressure, sliding velocity and temperature. The friction coecient itself is only a function of normal pressure and sliding velocity. Temperature eects are incorporated using the WLF transformation, i.e., if the current temperature is dierent than the reference temperature, an equivalent new sliding velocity for the reference temperature is calculated. This research has been continued by Hofstetter et al. (2003), where a thermo-mechanical coupling has been introduced. Energy dissipation during sliding is converted into heat and this heat ux causes a temperature rise of the rubber and road. In (Hofstetter et al., 2006) also simulations of abrasion of the tread block are added. Dorsch et al. (2002) derived a phenomenological friction law using data obtained from experiments on the LAT100. The modeling and calculation of the transient rolling contact of tyres on road tracks is the subject of a large multidisciplinary research project (FOR492) at the University of Hannover. One of the projects focusses on computational homogenization procedures, to develop a friction law at a macroscopic level, based on eects on a microscopic level. More details on this subject can also be found in the book of Wriggers (2006).

4.1.3

Constraint handling in ABAQUS

The aforementioned constraints, imposed by contact and friction, can be handled by Lagrange multipliers, which satises the constraint exactly, or with penalty methods, which penalize violation of constraints. All the constraint and interaction methods of ABAQUS can be found in the ABAQUS Analysis users manual, chapters 28-32. One can nd here that normal contact is handled by Lagrange multipliers, penalty methods or an augmented Lagrange method. Friction, in ABAQUS/standard, is normally handled by the penalty method. While the surfaces are sticking, a certain elastic slip cr is allowed and the magnitude of sliding is limited to this elastic slip. The allowable elastic slip is given as cr = Ff l, (4.8)

where Ff is an user-dened slip tolerance and l is the characteristic surface length. The slope G, which is a penalty stiness, is adapted depending on the normal pressure, see gure 4.3. One can see that the smaller cr , the steeper the slope of G. Exact sticking, G = , can be obtained by enforcing a Lagrange multiplier. However there is one important exception: If a so-called Steady-State Transport Analysis (which is commonly used in tyre simulations) is performed, the penalty constraint is based on a maximum allowable slip rate cr = Ff 2R, (4.9) where is the spinning velocity and R the radius of the object. This introduces discontinuities if a Steady-State Transport Analysis is combined with other ABAQUS standard analyses and should therefore be avoided.

4.1.4

Contact possibilities in ABAQUS

Several contact models are available in ABAQUS. In this section the import features are given. There are two contact discretizations, node-to-surface and surface-to-surface contact and two tracking approaches, small sliding and nite sliding.

4. FINITE ELEMENT METHOD

25

2 cr = p2

cr = p cr = Ff l G2
1 cr

G=

cr cr

= p1 G1 cr

Figure 4.3: Adapting penalty stiness, based on a maximal allowable slip.

Some of the mechanical contact property models available in ABAQUS/Standard include: Softened contact, friction and user-dened constitutive models for surface interaction. Surface interaction in thermal or coupled thermal-mechanical contact simulations can include heat exchange by conduction and radiation as well as the generation of frictional heat in coupled simulations. During an increment in which the contact status has changed, ABAQUS/Standard will use the default hard contact criterion to determine whether the change should be reversed. In other words, if the contact status changes from open to closed during an increment, the contact pressure must remain positive for the changed status to persist. In subsequent increments the contact point can again sustain tensile pressures up to a specied value before the surfaces separate. This contact pressure-overclosure relationship is useful for cases where negative pressure values (such as surface cohesion) may be allowed physically. It can also be useful in overcoming numerical problems in dicult contact simulations and in obtaining solutions without excessive iteration. Another possibility to oppose the relative motion between the interacting surfaces is to specify damping. The friction models available in ABAQUS: Include the classical isotropic Coulomb friction model, which in ABAQUS: in its general form allows the friction coecient to be dened in terms of slip rate, contact pressure, average surface temperature at the contact point, and eld variables; and provides the option to dene a static and a kinetic friction coecient with a smooth transition zone dened by an exponential curve. Allow the introduction of a shear stress limit, which is the maximum value of shear stress that can be carried by the interface before the surfaces begin to slide. Include an anisotropic extension of the basic Coulomb friction model in ABAQUS/Standard; Include a model that eliminates frictional slip when surfaces are in contact. Include a softened interface model for sticking friction in ABAQUS/Explicit in which the shear stress is a function of elastic slip. Can be implemented with a stiness (penalty) method, a kinematic method (in ABAQUS/Explicit), or a Lagrange multiplier method (in ABAQUS/Standard), depending on the contact algorithm used.

26

4.2. LITERATURE RELATED TO FEM AND TYRES

Can be dened in user subroutine FRIC (in ABAQUS/Standard) or VFRIC (in ABAQUS/Explicit for the contact pair algorithm only), which allows modeling of very general frictional surface conditions. Besides the friction models it is also possible to make use of user-dened interfacial constitutive behavior. This is provided so that any constitutive behavior across an interface can be added to the library of existing models. To use this the constitutive model (or a library of models) for the interface has to be programmed in user subroutine UINTER(ABAQUS/Standard) or VUINTER. It is available only for a surface-based contact denition involved in stress/displacement, coupled temperature-displacement, or heat transfer analysis. According to the manual the feature is very general and powerful, but it requires considerable eort and expertise and is intended for advanced users.

4.2

Literature related to FEM and tyres

Together with the development of non-linear FEM the use of FEM in tyre development started. There is a lot of literature, somehow related to tyres, on computational developments in the framework of FE and quite some work is also incorporated into ABAQUS. Examples are the works of Oden on friction and rolling contact (Oden and Pires, 1983, 1984; Oden and Martins, 1985; Oden and Lin, 1986; Oden et al., 1988; Faria et al., 1989), Laursen and Simo in the eld of contact problems with friction (Simo and Laursen, 1992; Laursen and Simo, 1993a,b) and Padovan on rolling visco-elastic cylinders (Zeid and Padovan, 1981; Padovan, 1987; Kennedy and Padovan, 1987; Nakajima and Padovan, 1987; Padovan et al., 1992). A special issue (Vol 177, nr 3/4) of the journal Computer Methods in Applied Mechanics and Engineering is devoted to computational modeling of contact and friction. More recent are the works of Wriggers (Wriggers et al., 1990; Zavarise et al., 1992; Haraldsson and Wriggers, 2000; Bandeira et al., 2004; Wriggers, 2006) on constitutive interface laws with friction. Literature specically related to tyres usually does not provide detailed information. The main reason is that most tyre manufacturers use own (in-house) nite element codes and specic methods are kept condential. Even in the publications where ABAQUS is used not much background information can be found. Besides this, most research at universities is done in cooperation with a tyre manufacturer and therefore still not much model details are given, more information can be obtained from articles where specic subjects related to tyres are treated. However the found literature provides insight in the trends and developments of tyre modeling throughout the past decades. According to Padovan (2007), the following list of tyre problems are successfully handled by virtual CAE modeling methodologies Static rim mounting, ination and axle load/deection analysis States of stress/deection/strain in all tyre components Force and moment, on center feel Denition of footprint shape as a function of axle load state and pressure Rolling/handling models both quasi-static and dynamic including braking and cornering eects Thermal analysis/rolling resistance under rolling conditions Model cure press, and other manufacturing steps

4. FINITE ELEMENT METHOD

27

Fatigue life prediction models under full duty cycle Natural frequencies, handling and critical speed analysis Hydroplaning models, snow traction, in mud performance Geometry, and cord layout optimization Tread lug design Acoustical analysis Suspension models incorporating tyres In one of his presentations it is stated that the overall tyre performance is usually measured by global actions, such as force & moment, ride comfort or dynamic response and therefore gross FEA models can be employed to establish such characteristics. In contrast to this some aspects, such as tyre failure, require very localized models. In this sense a form of multi length scale methodology can be used, i.e., a global model is used to drive a local model. A summary according to Padovan of advantages and disadvantages for global and local models: An one scale model has one contiguous mesh and a rened mesh is placed in locations of interest, transition meshes are needed to link rened and coarse zones. It is in general dicult to build a mesh with reliable local renements. The models tend to be very large and time consuming and are not very adoptable making mesh modications. The transition areas tend to yield unreliable results. Multiple scales suer the so-called localization syndrome: localization always requires further localization, so a localization cut-o criteria is needed. Besides that stress and strain take on a random character as localization brings on the probabilistic nature of such scale levels. While probabilistic variances can be very signicant, generally the mean is close to the nominal design value and since stiness of cord or metallic/plastic subcomponents are orders of magnitude higher than rubber, local actions of such subcomponents dominate; rubber is hence strictly a geometric place holder. Also every new level should be correlated with experiments to assess accuracy and sensitivity to meshing. Most of the mentioned tyre problems require a specic type of modeling approach, in this sense there is no universal tyre model. This totally depends on the kind of problem one wants to solve. An overview is given according to some points of the above list, although most articles t in more than one category. Besides the journal of Tire Science and Technology, a good starting point are the two overview papers of Mackerle (1998, 2004) about rubber and rubber-like materials, nite element analyses and simulations, where an extensive reference list can be found.

4.2.1

Static analysis

One of the rst overview papers, where FE models and the contact problem for tyres are described, is the paper of Noor and Tanner (1985). In this article, for a NASA research program for the space shuttle tyres, the current status and developments of computational models for tyres are summarized. This review has been made again by Danielson et al. (1996). Basically all early tyre models are axi-symmetric, with no tread pattern or only circumferential groves due to the limits of computational resources. Another popular way to reduce the number of degrees of freedom is the application of the globallocal analysis. In this approach, an analysis of the complete structure is rst performed with a coarse mesh. After that a part of the structure is meshed ner and interpolated displacements are applied at the boundaries of this region. Such an approach can be used to compute the forces in the contact area of a deected tyre (see Gall et al., 1995 and Meschke et al., 1997). Although a

28

4.2. LITERATURE RELATED TO FEM AND TYRES

local model is able to give detailed numerical results corresponding to a tread block, the accuracy is strongly inuenced by the simple global model. With the ever increasingly computational power it is now possible to mesh a part of even the whole tread with a detailed pattern (e.g. Cho et al., 2004) and perform all kinds of static analysis such as footprint shapes as function of axle load state and pressure.

4.2.2

Force & Moment

In an article of Kabe and Koishi (2000) a comparison between ABAQUS standard, explicit and experiments for steady state cornering tyres is made. Although the results of both implicit and explicit are closer to each other than to the experiments, the implicit method is signicantly faster, 6 hours and 8 days respectively. The prediction of tyre cornering forces is given in a paper of Tnk o u u and Unlsoy (2001), where MARC is used. An comparison with experiments is also presented. The application of a new type of non-linear 3D nite element tyre model for simulating tyre spindle force and moment response during side slip is described by Darnell et al. (2002). The simulation model is composed of shell elements, which model the tread deformation, coupled to special purpose nite elements that model the deformation of the sidewall and contact between the tread and the ground. Despite the (seemingly) simple model the results corresponds quite well with experiments. A FE simulation with ABAQUS, in which the eect of tyre design parameters on lateral forces and moments is studied is presented in a paper of Olatunbosun and Bolarinwa (2004). Parametric studies are performed on a simplied tyre (no tread, negligible rim compliance and viscoelastic properties, shear forces modeled with Coulomb friction with constant ). Explicit simulations to predict tyre cornering forces, using PAM-SHOCK, are presented by Koishi et al. (1998). Besides a comparison with experiments, parametric studies on the eect of ination pressure, belt angle and rubber modulus are performed. Rao et al. (2003) simulated the dynamic behavior of a pneumatic tyre using ABAQUS/Explicit. A study on a passenger car radial tyre to simulate cornering behavior, braking behavior, and combined cornering & braking behavior is presented. Further the eect of camber angle and grooved tread on tyre cornering behavior is studied. To reduce computational load two tyre models are used, one treadless tyre and a tyre with ve circumferential groves. A cleat test with a half rotational response treadless tyre model can be found in a paper of Olatunbosun and Burke (2002), where NASTRAN has been used. An comparison with experiments shows deviations as the traverse speed increases above 20 kph. The prediction of the radial force seems better than the longitudinal force. In an article of Cho et al. (2005) the dynamic response of a fully patterned tyre rolling over a cleat is presented, using ABAQUS/Explicit. To decrease the computational load the ber-reinforced rubber is modeled with composite shell elements, instead of rebar elements. The eect of plysteer for steady-state rolling tyres is investigated by Mundl et al. (2005), where a global-local analysis is used, and by Ohishi et al. (2002), where a detailed tread pattern is used in the contact area.

4.2.3

Thermo-mechanical approaches

More recently thermo-mechanical models for prediction of the temperature inside tyres have been presented (Allen et al., 1996a,b; Park et al., 1997; Futamura and Goldstein, 2004; Lin and Hwang, 2004). Besides the prediction of steady state temperatures for rolling tyres a small literature overview is given by Rao et al. (2006). Experimental results of the temperature rise in an actual contact area between the pavement and the tread of a cornering tyre can be found in an article of Fujikawa et al. (1994). It is shown that a signicant temperature rise occurs for severe slip angles.

4. FINITE ELEMENT METHOD

29

4.2.4

Hydroplaning

Another phenomenon is hydroplaning, where the interaction with uid (water) is taken into account by coupling FEM and Computational Fluid Dynamics (CFD). For examples see Seta et al. (2000), Nakijima et al. (2000) and Cho et al. (2006).

4.2.5

Component modeling

Besides complete tyre models a lot of work focusses on the specic components of a tyre, e.g. the tyre cord reinforcements, which can be modeled using so-called rebar elements (Huh and Kwak, 1990; Helnwein et al., 1993; Mesehke and Helwein, 1994). This is also incorporated in ABAQUS. Asperity modeling to simulate rubber sliding on rough surfaces is considered by Bui and Ponthot (2002) for the 2D case and by Faulkner and Arnell (2000) for two 3D hemispherical asperities. Recently a comparison between experimental and explicit FEM has been published (Kuwajima et al., 2006) to examine the interfacial phenomena between rubber and abrasive papers, such as the real contact length, partial slip, and apparent friction coecient under vertical load and tangential force. Other examples are optimization of the carcass (Cho et al., 2002) and the optimum shape of the tread to avoid lateral slippage (Ahmed et al., 2005).

30

4.2. LITERATURE RELATED TO FEM AND TYRES

Chapter 5

Validation and modeling methods

5.1 LAT1OO and Tekscan

The LAT100 (VMI) is a test machine originally developed by Dr. Grosch. It is designed to relatively rank tread compounds with respect to dry traction, wet traction, ice traction, abrasion and rolling resistance. Under certain test conditions it should be possible to relate these indoor experiments with outdoor tyre tests. On this machine a small rubber wheel, see gure 5.1(b), is placed on a spinning disk which represents a road surface. The temperature, speed, surface of the disk, the slip angle and normal load on the sample can be varied. More details of the LAT100 can be found in the documentation system of Vredestein (Doclib).

(a) LAT100 machine.

(b) Sample wheel.

Figure 5.1: The Laboratory Abrasion and skid Tester.

The Tekscan machine can be used to obtain the normal contact pressure distribution for a statically deformed or slow rolling tyre under zero slip angle. More details on the principle of Tekscan and possibilities of using small LAT100 samples on the Tekscan can be found in the report of Broeze (2007).

5.2

Eplexor and DIK

Material test data of rubber compounds can be acquired from the Deutsches Institut fr Kautschuku technologie (DIK). Visco-elastic material properties can be measured using the Eplexor available at Vredestein or at the university in cooperation with the section of Materials Technology.

32

5.3. FLAT PLANK AND TYRE TEST TRAILER

5.3

Flat plank and tyre test trailer

The global tyre deformation characteristics and cleat experiments can be measured on the at plank machine in the automotive lab at the TU/e. The global force and moments can be measured with the tyre test trailer of TNO. More details are presented in the report of Cremers (2005).

5.4

Tyre modeling in ABAQUS

Guidelines for tyre analyses in ABAQUS can be found in the ABAQUS Example problems manual, chapter 3.1 and in the course notes on Tire modeling using ABAQUS. The way to obtain the global force and moment characteristics of a tyre under dierent driving conditions, such as free rolling or slip and chamber angles, in ABAQUS/Standard is the following: Start with a 2D cross-section of the tyre. Perform a static analysis in which the ination pressure (and rim mounting) is simulated. Revolve the axi-symmetric cross-section about the axis of revolution, with a nonuniform discretization along the circumference, to obtain a 3D model. Apply a vertical load, which represents the weight of a vehicle. Perform a static full 3D footprint analysis of the tyre in contact with a at road. Transfer these results to perform steady-state rolling analysis under dierent driving conditions. The great advantage of the steady-state transport analysis is that it uses a moving reference frame in which rigid body rotation is described in an Eulerian manner and the deformation is described in a Lagrangian manner (the so-called ALE-formulation), see also articles of Nackenhorst (2004); Ziee and Nackenhorst (2005); Laursen and Stanciulescu (2006). This kinematic description converts the steady moving contact problem into a pure spatially dependent simulation. Thus the mesh need to be rened only in the contact region, the steady motion transports the material through the mesh. Frictional eects, inertia eects, and history eects in the material can all be accounted for. The limitation of a steady-state transport simulation is that the circumference of the tyre must be continuous (in the current version). As a result only circumferential groves can be used, instead of a detailed tread pattern. If an uniform discretization along the circumference is used, it is possible to transfer steady-state rolling results to ABAQUS/Explicit and perform dynamic/impact simulations. This is computationally expensive and usually a coarse mesh and a simplied tyre model is used to handle these kind of simulations. Contact must be divided into normal and tangential contact as mentioned in chapter 4. Both normal and tangential contact or friction can accounted for by using the available algorithms or by own user subroutines. Temperature eects can be accounted for in dierent ways. A similar procedure for steadystate solutions can be followed as in Huemer et al. (2001b), where a reference temperature is used and WLF transformations are used to calculate equivalent sliding velocities at the reference temperature. In an implicit thermo-mechanical approach it is possible to convert energy dissipation due to friction into heat. If this option is used, one should be aware that material properties can be aected by temperature changes and ideally this should be taken into account as well. This also holds for the friction law, if temperature eects are taken into account the friction law should be an explicit function of temperature.

5. VALIDATION AND MODELING METHODS

33

To incorporate transient dynamical eects, such as local temperature rise for severe slip angles, ABAQUS explicit can be used, although this probably takes a very long time to simulate and no records of this has been found in literature sofar. Guidelines for heat-transfer problems can be found in the ABAQUS Example problems manual, chapter 5.

34

5.4. TYRE MODELING IN ABAQUS

Chapter 6

Conclusions
To model a tyre and interaction with the environment using the nite element method not only knowledge of materials, geometry and interaction is required, but also a thorough knowledge of the computational mechanics behind the nite element method is required. However the most important conclusion is that the often used Coulomb friction model with a constant coecient of friction is in general not realistic in the case of rubber friction.

6.1

Friction

Friction of polymers is closely related to their viscoelastic behavior. Generally speaking, the coecient of friction increases with sliding velocity until a maximum value is reached at a certain speed, followed by a decrease of the friction coecient. The viscoelastic properties of polymers depend strongly on temperature. Grosch showed the relation between the dependence of the friction coecient due to sliding velocity and temperature and the viscoelastic deformation on strain rate and temperature. The friction force for soft rubber sliding on a smooth surface (perfect contact) is more or less constant, independent of the load. Thus the coecient of friction decreases continuously as the pressure increases. However there is also a coecient of friction, independent of pressure, for harder rubber compounds, sliding on a rough surface. Apparently in this case contact is incomplete. An increase in pressure creates a larger true area of contact and hence a larger frictional force. This shows the inuence of the temperature, sliding velocity and real contact area. The current theory of rubber friction and surface roughness, which capture all these properties can be found in the publications of Persson, (see review paper Persson, 2005). The surface roughness is characterized by a fractal description, which exhibits geometrical self-similarity. These results are based on the early studies of Grosch, taking into account that sliding friction of rubber has the same temperature dependence as that of the complex elastic modulus. He states that the friction force is related to the internal friction of the rubber, which is a bulk property. The hysteretic friction component is determined by sliding of the rubber over asperities of a rough surface. These oscillating forces lead to energy dissipation. Every length scale up to the largest particles of asphalt, can be related to an excitation frequency. As a result of energy dissipation heat is generated, in a recent article this local heating of the rubber is also taken into account. Although this theory provides knowledge of the physical origin of rubber friction it has some drawbacks. Exact knowledge of the contribution of each length scale is needed. Therefore it has, at the moment, limited practical use, since variable amounts of unknown foreign materials in the interfacial region makes it almost impossible to derive quantitative estimates of the absolute magnitude of friction from physical properties of the rubber and surface. So assumptions must still be made.

36

6.2. FEM

6.2

FEM

There are two distinctive methods to solve FE problems, implicit and explicit. The implicit method solves for a statical or dynamical equilibrium, while the explicit method solves transient dynamic response problems using an explicit direct-integration procedure. Although a lot of procedures are implemented in ABAQUS, the user still has to make several choices in the model formulation, i.e., the discretization of the mesh, the element types, the solution procedure, the solver tolerances etc. These selections may have a tremendous eect on the nal result, so care must be taken. Contact problems in FEM are divided into normal and tangential contact and interfacial laws must be chosen to solve the contact problem. If the tangential forces reach a certain limit bodies will no longer stick, but start sliding. Normally a local Coulomb law is used to calculate the shear stress in this case, but it is also possible to formulate tangential constitutive equations by regularization of the stick-slip behavior, e.g. to avoid the non-dierentiability of Coulombs law at zero velocity. Besides a xed friction coecient in ABAQUS some build-in constitutive laws can used to make the coecient of friction dependent on pressure, sliding velocity or other eld variables. ABAQUS also provides the possibility to use own interfacial laws. With the standard options in ABAQUS it is nowadays possible to include a detailed tread pattern on a (part of a) tyre global model, instead of the global-local approach, and perform analyses to capture footprints as a function of axle load and ination pressure. The stress and strain in all tyre components and the force and moment on a center feel can be calculated. The steady state option in ABAQUS makes it possible to eciently calculate steady state solutions for dierent driving conditions, since the mesh need to be rened only in the contact region. With this option frictional eects, inertia eects, and history eects in the material can all be accounted for. This makes it a powerful approach to obtain steady state solutions and determine the global forces and moments for several variables, such as ination pressure, normal load, velocity, slip and chamber angle. A limitation of this so-called steady state transport simulation is that the circumference of the tyre and the road must be continuous. As a result only circumferential groves can be used in these kind of simulations. This also implies that a macroscopic friction law is required, in which surface roughness is embedded in the friction model. A phenomenological approach could be used to derive a macroscopic friction law, which depend on normal pressure, sliding velocity, temperature and surface roughness. The articles of Huemer et al. (2001a,b) are related to this context, a friction law for rubber tread blocks on snow and concrete is presented for use in a macroscopic model. The coecient of friction depends on normal pressure, sliding velocity and temperature. The friction coecient itself is only a function of normal pressure and sliding velocity. Temperature eects are incorporated using the WLF transformation. Hofstetter et al. (2003) introduced a thermo-mechanical coupling where energy dissipation during sliding is converted into heat and this heat ux causes a temperature rise of the rubber and road and this inuences the friction coecient. Dorsch et al. (2002) derived a phenomenological friction law using data obtained from experiments on the LAT100. Despite the still increasingly computational power there are still limitations in the kind of simulations one can perform. Especially explicit simulations, where transient dynamic eects are captured, require a lot of time and as a result coarse meshes are often used to solve these problems. A way to perform explicit simulations is to use an uniform discretization along the circumference. It is than possible to transfer steady state rolling results to ABAQUS/Explicit and perform dy-

6. CONCLUSIONS

37

namic/impact simulations such as rolling over a cleat. To incorporate transient local temperature eects at severe slip angles, ABAQUS explicit can be used as well in theory, although this probably takes a very long time to simulate and no records of this has been found in literature sofar. However experimental results of the temperature in an actual contact area between the pavement and the tread of a cornering tyre show a signicant temperature rise (Fujikawa et al., 1994).

38

6.2. FEM

Appendix A

32th Tire Mechanics Short Course

A.1 Strength, wear and friction of rubber.

The presentation of Gent (2007) deals with the mechanical properties of rubber. Four subjects are addressed: Elasticity and visco-elasticity, strength of rubber components, rolling and sliding friction, and abrasion of rubber and wear of tyres.

A.1.1
Elastic

Elasticity and visco-elasticity

Rubber is an elastic, soft and virtually incompressible material. Its molecular structure consists of long, linear exible molecules forming random coils. The molecular segments are mobile (so-called Brownian motion) and the molecules are interlinked into a 3D network. Chemical crosslinks are usually made by sulfur linkages also known as vulcanization. The rubbery state of a polymer is determined by the so-called glass transition temperature Tg . If the temperature is above Tg the polymer is rubbery, below Tg the polymer is glassy. For rubbers the Tg is below room temperature. The Brownian motion (rate of segmental jumps) depends strongly on temperature (sheet 6). Small strains For small strains the material properties can be dened by the shear modulus G and the modulus of bulk compression K, which are related to the tensile or Youngs modulus E and Poissons ratio as follows: E = = 2(1 + )G, 1 3K 2G . 2 3K + G (A.1) (A.2)

For rubber typically K G and it follows that E = 3G and = 0.5. If a rubber block with = 0.5 is compressed, it is sheared outwards and generates an internal pressure. This eect causes a compression modulus, which is generally greater than E. An example can be found on sheet 10. Large strains For large strains the deformation is usually described by stretch ratios 1 , 2 , 3 in the principle directions. Instead of a modulus a strain energy density function is used. Several forms (constitutive models) are available.

40

A.1. STRENGTH, WEAR AND FRICTION OF RUBBER.

Filled rubbers In tyres rubbers are usually lled with particles like carbon black or silica. This creates two additional eects. The so-called Payne eect and Mullins eect, which are both softening eects. The Payne eect is a softening eect for small strains (0.1%), attributed to breaking apart aggregates of ller particles. The Mullins eect is a substantial softening eect at higher strains, attributed to progressive detachment of rubber molecules from ller particles. The constitutive laws for large strains cannot be used to fully describe the stress-strain relation, because rubbers do not follow reversible stress-strain relations.

A.1.2

Visco-elastic

When rubber is dynamically stretched and released the returned energy is less than the energy that is put into the rubber, see sheet 21. This visco-elastic eect cannot be described by the perfect elastic shear modulus G, but a dynamic shear modulus G and a dynamic shear loss modulus G is introduced to describe this hysteresis. Another term that is frequently used is the loss angle, which is dened as tan = G . For an elastic solid the strain is in phase with the stress, while for G a visco-elastic solid the strain lags behind the stress with a delay of , where is the frequency (rad/sec). The dependence of segmental jump rate on temperature is described by the universal WLF relation (Williams et al., 1955), T 17.5(T Tg ) log = , (A.3) Tg 52 + T Tg where Tg = 0.1/sec. Some polymers such as butyl rubber behave according a dierent relation log 17.5(T Tg ) T = . Tg 100 + T Tg (A.4)

This is possibly due to unusually long moving segments. The WLF relation can by used to predict results at other temperatures, the frequency f at temperature T is equivalent to a frequency f at T , where T f = = aT (A.5) f T aT is known as the shift factor, this also holds for strain rates or velocities. An example is given on sheets 27-30. However the following remark should be taken into account: These shifts are only valid for pure rate (viscous) processes. Strength of rubber components In this section the tear strength and fracture energy are discussed. Also the visco-elastic contribution to strength eects of rate and temperature is treated. However this is not so important at this moment.

A.1.3

Friction
=

F , (A.6) N where N is the normal force and F the tangential force. For soft rubber sliding on a smooth surface (perfect contact) the frictional force F is more or less constant, independent of the load N . Thus the coecient of friction decreases continuously as the pressure increases, see gure A.1. However, there is also a coecient of friction, independent of pressure, for harder rubber compounds, sliding on a rough surface. Apparently in this case contact is incomplete. An increase in pressure creates a larger true area of contact and hence a larger frictional force.

The coecient of friction is dened as:

A. 32TH TIRE MECHANICS SHORT COURSE

41

(a) Friction denition.

(b) Friction law.

Figure A.1: Friction on smooth surface.

A.1.4

Rolling friction

Rolling friction Fr can be seen as M , where M is a torque and R the radius. For an elastic R wheel the reaction force acts through the wheel center and therefore the moment M = 0. For a visco-elastic wheel the reaction force moves towards the loading side, because stresses are lower in unloading, M = N z where N is the normal load and z the distance from the wheel center, see gure 3.4. Displacement of the center of pressure, and hence the coecient of friction r , depend on speed and temperature in the same way as the loss factor tan . A formulation for r as proposed by Greenwood et al. (1961), 1 3 N r = 0.33 tan , (A.7) G R 2 where G is the complex shear modulus. Rolling friction is mainly due to energy dissipated as rubber is compressed and released in the contact patch.

Figure A.2: Rolling friction.

Sliding friction Grosch (1963) noted that the curve for vs. sliding speed on a lubricated rough surface resembles the curve for tan vs. frequency f (and they both obey WLF temperature shifts), see gure A.3. He claims that friction is due to energy dissipated as rubber is compressed and released by asperities. For dry sliding on a smooth surface the curve for vs. sliding speed Vs resembles again the curve for tan vs. frequency f . But now it is displaced to much lower speeds than before. Friction in dry sliding on a smooth surface is due to energy dissipation as rubber sticks and slips on a molecular scale. In the case of sliding on a rough dry surface the two dissipative processes appear

42

A.2. THE TIRE AS A VEHICLE COMPONENT.

Figure A.3: Grosch interpretation.

at dierent speeds, corresponding to dierent length scales: asperities and molecules, see gure A.4. Conclusion: Several contributions to can arise at the same sliding speed from stick-slip processes occurring at dierent length scales.

Figure A.4: Combined eect of sliding on a rough and dry surface.

Abrasion Some general remarks can be found on sheets 64-65. This is not relevant at this moment.

A.2

The tire as a vehicle component.

The presentation of Potts (2007) consists of vehicle dynamics and can be compared with the book of Pacejka (2006) or the lecture notes on vehicle dynamics (Besselink, 2003). The presentation actually starts more or less at sheet 44 with the denition of a tyre: A Compliant roller bearing with tractive properties. Further load support and the so-called carpet plots are explained. The importance of vehicle understeer and the contribution of the tyre is treated in sheets 67-79. Testing equipment and results of forces and moments measurements as well as eigenfrequencies are presented in the remainder of the presentation. The tyre models used for vehicle handling are (semi-)analytical dynamic models.

A. 32TH TIRE MECHANICS SHORT COURSE

43

A.3

Tire materials and manufacturing.

The materials and manufacturing processes used for pneumatic tyres are discussed in this presentation (Walter, 2007a). The worldwide tyre production is greater than 1 billion units with a market value exceeding US $100 billion. Around 73% of this is contributed by passenger tyres and 70% of these tyres are replacement tyres. The presentation is divided in four subjects: the tyre components, tyre elastomers, the cord reinforcements and factory processing.

A.3.1

Tyre components

There are basically two types of tyres: bias and radial tyres, however the focus in this report will be on radial tyres, since these are manufactured at Vredestein. The main dierence is the orientation of the plies, see gure A.5. Every rubber component in a tyre requires specic properties.

(a) Bias tyre.

(b) Radial tyre.

Figure A.5: Dierence between bias and radial tyres.

An overview of these components can be found on sheet 10. Further the tyre construction, such as aspect ratio, belt construction and tread compound, depends on the target market (speed rating). The principal tyre elastomers are natural rubber (NR), polyisoprene (IR, this is synthetic NR), styrene butadiene rubber (SBR), butyl rubber (IIR) and butadiene (BR). The reinforcement materials are: a) The llers, such as carbon black and silica. b) The organic textile plies, such as polyester or rayon bodies, nylon cap or kevlar. And c) Steel, which is used in the belt cord and the bead wires. Some details and denitions are given on sheets 1523. The denition of a compound (sheet 24) is the following: the formulation of a mixture of rubber and additives which meets the needs of the tyre component application. Usually a compound consists of one or more polymers, vulcanizing agents, accelerators, reinforcing llers, antidegradents, plasticizers, softeners and tackiers (used to bond pieces together). This explains why the properties of a compound can deviate from a pure elastomer.

44

A.4. RULES AND REGULATIONS GOVERNING TIRES.

A.3.2

Tyre elastomers

In this part (sheet 2964) the aforementioned rubbers are described in detail and the eect of llers is discussed. Some sheets deal with the dynamic mechanical properties, this overlaps with Gent et al. (2007).

A.3.3

Cord reinforcements and Factory processing

In the last part the dierent cord reinforcements, textile and steel, are discussed. Dierent types of steel cord constructions are explained and nally the manufacturing processes is described. This will not be discussed here, however is should be noted that (limitations of) the manufacturing process greatly aects the nal tyre. So it is worthwhile to scroll through this slides.

A.4

Rules and regulations governing tires.

This is an interesting presentation of Walter (2007b) about all the rules, which again aect tyre design. Although this presentation is focussed on the US federal government regulations, it provides a background on the fair amount of rules, which tyre manufactures have to cope with.

A.5

Advanced tire modeling.

The presentation of Padovan (2007) introduces the course goals of nite element analysis in tyre analysis. There are in total 12 presentations with over 600 sheets, so in the remainder of this chapter references to a presentation are made using the number (#) shown in table A.1. In this overview Ill give some remarks to highlight interesting parts. In the introduction presentation Table A.1: Presentations of Padovan. Number 1 2 3 4 5 6 7 8 9 10 11 12 Title Advanced tire modeling Intro to FEA modeling Multilength scale FEA model of tire Dynamic elastomer properties Critical speed Rolling resistance FEA modeling of tires stochastic eects Optimization of cord spacing Fracture mechanics Modeling stochastic eects on durability Overal fatigue analysis procedure Hydroplaning analysis

there is, besides the outline of the other presentations, one sheet giving an overview of tyre problems, which are successfully handled by virtual CAE modeling methodologies (1): Static rim mounting, ination and axle load/deection analysis States of stress/deection/strain in all tyre components Force and moment, on center feel

A. 32TH TIRE MECHANICS SHORT COURSE

45

Denition of footprint shape as a function of axle load state and pressure Rolling/handling models both quasi-static and dynamic including braking and cornering eects Thermal analysis/rolling resistance under rolling conditions Model cure press, and other manufacturing steps Fatigue life prediction models under full duty cycle Natural frequencies, handling and critical speed analysis Hydroplaning models, snow traction, in mud performance Geometry, and cord layout optimization Tread lug design Acoustical analysis Suspension models incorporating tyres The second presentation is the introduction to nite elements, this presentation is more elaborate than the other presentations. These sketch the general problem and show the FE solution. It is important to determine the levels of deformation/strain to use appropriate stress-strain (linear vs. nonlinear) relations (2). In general tyre problems are large deformation/strain problems: Small deformation Deformation <<<<< Characteristic length Small strain Strain = Deformation/Characteristic length <<<<< 1. Large deformation Deformation = O( Characteristic length) Large strain Strain = Deformation/Characteristic length = O(1.) In the third presentation it is stated that the overall tyre performance is usual measured by global actions, such as force & moment, ride comfort or dynamic response and therefore gross FEA models can be employed to establish such characteristics. In contrast to this some aspects, such as tyre failure, require very localized models. In this sense a multi length scale methodology can be used, i.e., a global model is used to drive a local model. A summary of advantages and disadvantages for global and local model is given in this presentation and also stated here. A one scale model has one contiguous mesh and a rened mesh is placed in locations of interest, transition meshes are needed to link rened and coarse zones. It is in general dicult to build a mesh with reliable local renements. The models tend to be very large and time consuming and are not very adoptable making mesh modications. The transition areas tend to yield unreliable results. Multiple scales suer the so-called localization syndrome: localization always requires further localization, so a localization cut-o criteria is needed. Besides that stress and strain take on a random character as localization brings on the probabilistic nature of such scale levels. While probabilistic variances can be very signicant, generally the mean is close to the nominal design value and since stiness of cord or metallic/plastic subcomponents are orders of magnitude higher than rubber, local actions of such subcomponents dominate; rubber is hence strictly a geometric place holder. Also every new level should be correlated with experiments to assess accuracy and

46

A.5. ADVANCED TIRE MODELING.

sensitivity to meshing. Sheets 30-36 give an example of a global-local approach to investigate local cord-rubber interaction. It is however stated here that with a global model the usual performance characteristics are usable. Presentation 4 deals with material properties of elastomers, such as stress-strain and creep, and friction. The friction depends on interfacial pressure and dynamic eects, see gures A.6 and A.7.

1000 psi
Interfacial pressure
Figure A.6: Quasi static.

O(10 20% slip)

dry icy/wet
Percentage slip
Figure A.7: Dynamic eect for average roughness road.

100%

Presentation 5 is about critical speeds of tyres and describes four methodes to analyze this, but it is not relevant at the moment. Rolling resistance is the subject of presentation 6 and is primary caused (95%) by mechanical work stored in rubber that is partially converted to heat and secondary cause (5%) is the mechanical work stored in cords that is converted to heat and friction work induced in the footprint area during rolling, which generates heat in addition to wear. Two dierent approaches are given to determine the rolling resistance. The other presentations go into specic subjects and are not relevant at the moment.

A. 32TH TIRE MECHANICS SHORT COURSE

47

A.6

Tire wear, traction and force generation.

The footprint is the subject of the presentation by Pottinger (2007). A footprint is a complex object, since the tyre is a doubly curved surface and is pneumatically pre-stressed. And friction aects the deformation of the tyre. Contact leads to a stress distribution across the interface, with normal component Z , and shear components at each point, X and Y . The nature of this distribution has a complex dependence on conditions of use. Two distinctive measurement techniques can be identied: Measure Z at ever ner resolutions or examine the entire stress eld at somewhat coarser resolutions. A method for measuring Z is pressure sensitive lm (like the Fuji-lm) for a static case. Electronic pressure mats (like Tekscan) can be used for static and slow rolling cases. These cannot be used for large shears, which occur in cornering, braking or accelerating. The dramatic dierence between static and slow rolling is shown in gure A.8, this shows why static footprints can be misleading. Sheets 18-30 give some examples of machines and concepts for measuring footprints.

Figure A.8: Static and slow rolling footprint.

Temperature plays a great role in the footprint. The surface temperature, in particular the change through contact, is related to wear. The bulk temperature determines the stiness of the tread compound, see also chapter A.1. Both wear resistance and handling depend on tyre temperature in the footprint because it inuences abradability, stiness and also the coecient of friction. In gure A.9 the eect of temperature rise on dierent surfaces is shown. It can be seen that the conductivity of the surface has great inuence on the tread surface temperature, note that the slip angle is 6 . This slip angle is a very severe condition for wear and will cause gumming in racing tyres. Some years ago Michelin published data showing that over 99% of all driving in the United States occurred at less than 6 slip angle. The low slip angles involved in actual operation are why tyres have a reasonable wear life. Figure A.10 shows basic passenger tyre footprint longitudinal and lateral stress characteristics. These gures are for zero slip and chamber angle. It can be seen that X is positive at the front and negative at the back. The shoulders have a larger eective radius than the crown in this gure. The lateral stresses are pointed outwards. The eect of slip angle can be seen in gure A.11. A

48

A.6. TIRE WEAR, TRACTION AND FORCE GENERATION.

Figure A.9: Temperature eects.

slip angle reshapes the footprint into a rough trapezoid. One of the shoulders becomes longer, for example, the right shoulder, when turning left, and the other grows shorter (the left shoulder in this case). The normal stress increases on the long shoulder and falls on the short shoulder.

A. 32TH TIRE MECHANICS SHORT COURSE

49

(a) Longitudinal stress.

(b) Lateral stress.

Figure A.10: Longitudinal and lateral stress.

Figure A.11: Eect of slip angle.

50

A.7. TIRE STRESSES AND DEFORMATION ANALYSIS.

The tread pattern on tyres exist only as traction aids for contaminated surfaces, such as water or snow. These grooves cause uneven wear and noise. A remarkable citation from sheet 61: Industrial design of the pattern is artistic icing on an engineering necessity. For good dry traction one should put as much rubber on the road as possible. The following traction summary shows the trade-o in tread design: Dry Maximize contact area Keep Z low Use soft compound Wet Medium-hard compound Maximize groove volume fraction Minimize element size Long narrow footprint Ice Low TG rubber Lots of element edges Snow Large void projection times void width product perpendicular to direction of required force Flexible tread elements to insure cleaning, so void doesnt clog Sheets 73-97 shows dierent aspects of wear. A relation between wear and shear energy intensity is presented. Some guidelines for outdoor and indoor testing are given. The last part (98-116) of the presentation focusses on the limits of force generation for several conditions, mostly based on experimental data, such as pavement texture, hydroplaning, compounds and snow traction. This outlines again that a lot of environmental issues inuence the force generation, and as a result friction and wear.

A.7

Tire stresses and deformation analysis.

This last presentation given by Trinko (2007) gives an overview of the stresses and strains in tyres. It consists of three parts: Force topics, composite properties and stresses and strains.

A.7.1

Force topics

This section (sheets 349) is divided into 5 topics: Load transfer, cord tension, bead calculation, Purdy equations and air diusion. Interesting part is air diusion, which is equivalent to a heat transfer problem.

A.7.2

Composite properties

In this part of the presentation (sheets 51-86) the laminate theory is explained. This theory can probably be used to do some basic calculations to rubber/cord laminates and compare that to the so-called rebar elements in ABAQUS.

A. 32TH TIRE MECHANICS SHORT COURSE

51

A.7.3

Stress and strain

The last part gives some FE results about slip velocity and friction. Although this is another program than ABAQUS, it shows that a rolling tyre with friction and dierent slip angles is possible. In gure A.12 the slip and friction for 0 and 1 slip angle are shown. There are two

(a) Slip 0 .

(b) Friction 0 .

(c) Slip 1 .

(d) Friction 1 .

Figure A.12: Dierence between 0 and 1 slip angle.

approaches to rolling tyres. First a Lagrangian approach, where the mesh rolls. To use this approach equal segments around the whole tyre must be used. This allows a full tread pattern. The second one is a mixed Euler-Lagrange approach, where the material ows through the mesh. The mesh can deform, but does not roll. This can only be used for steady state solutions. The main advantage is that a ne mesh is only needed in the contact area, however this requires an axi-symmetric model with a smooth or circumferential grooved tread pattern. The proposed frictional stress for steady state rolling is F = (P, V )P, (A.8)

where V is Vtyre Vroad and P the normal pressure. Two friction laws are given, the rst one as function of pressure, (P ) = CP , (A.9) where C and are constants and the second one as a function of V , see gure A.13. The power law (A.9), see also gure A.6, seems reasonable, according to Trinko, however the aligning moments are wrong compared to experiments. Trinko thinks this is due to a wrong friction law,

52

A.7. TIRE STRESSES AND DEFORMATION ANALYSIS.

but hes not aware of improved friction laws. Another remark is that there is still no conclusive correlation with wear. After the end of the presentation some additional sheets (143-169) are

V
Figure A.13: Friction law as function of V for rolling tyre. added. These describe some measurement techniques and explanations behind calculations.

Bibliography
ABAQUS Analysis users manual. Version 6.6. 2006. ABAQUS Example problems manual. Version 6.6. 2006. S. Reaz Ahmed, S.K. Deb Nath, and M. Wahhaj Uddin. Optimum shapes of tire-treads for avoiding lateral slippage between tires and roads. International Journal for numerical methods in Engineering, 64:729750, 2005. J. Mc Allen, A.M. Cuitio, and V. Sernas. Numerical investigation of the deformation characn teristics and heat generation in pneumatic aircraft tires part I. mechanical modeling. Finite Elements in Analysis and Design, 23:241263, 1996a. J. Mc Allen, A.M. Cuitio, and V. Sernas. Numerical investigation of the deformation characterisn tics and heat generation in pneumatic aircraft tires part II. thermal modeling. Finite Elements in Analysis and Design, 23:265290, 1996b. J.F. Archard. Elastic deformation and the laws of friction. Proceedings of the Royal Society of London, Series A, 243(1233):190205, 1957. A.A. Bandeira, P. Wriggers, and P. de Mattos Pimenta. Numerical derivation of contact mechanics interface laws using a nite element approach for large 3D deformation. International Journal for numerical methods in Engineering, 59:173195, 2004. A. Becker, V. Dorsch, M. Kaliske, and H. Rothert. A material model for simulating the hysteretic behavior of lled rubber for rolling tires. Tire Science and Technology, 26(3):132148, 1998. J.S. Bergstrom. Modeling of the dynamic mechanical response of elastomers. Tire Science and Technology, 33(2):120134, 2005. I. Besselink. Voertuigdynamica. Collegesheets, Eindhoven, 2003. F.P. Bowden and D. Tabor. The friction and lubrication of solids. Oxford University Press, oxford classics series edition, 2001. J. Broeze. Surface pressure on dr. Grosch tyre. In press, Eindhoven University of Technology, 2007. Q.V. Bui and J.P. Ponthot. Estimation of rubber sliding friction from asperity interaction modeling. Wear, 252:150160, 2002. A.W. Bush, R.D. Gibson, and T.R. Thomas. The elastic contact of a rough surface. Wear, 35: 8711, 1975. J.R. Cho, H.S. Jeong, and W.S. Yoo. Multi-objective optimization of tire carcass contours using a systematic aspiration-level adjustment procedure. Computational Mechanics, 29:498509, 2002. J.R. Cho, K.W. Kim, W.S. Yoo, and S.I. Wong. Mesh generation considering detailed tread blocks for reliable 3D tire analysis. Advances in Engineering Software, 35:105113, 2004.

54

BIBLIOGRAPHY

J.R. Cho, K.W. Kim, D.H. Jeon, and W.S. Yoo. Transient dynamic response analysis of 3-D patterned tire rolling over cleat. European Journal of Mechanics A/Solids, 24:519531, 2005. J.R. Cho, H.W. Lee, J.S. Sohn, G.J. Kim, and J.S. Woo. Numerical investigation of hydroplaning characteristics of three-dimensional patterned tire. European Journal of Mechanics A/Solids, 25:914926, 2006. M. Ciavarella, V. Delne, and V. Demelio. A new 2D asperity model with interaction for studying the contact of multiscale rough random proles. Wear, 261:556567, 2006a. M. Ciavarella, C. Murolo, and G. Demelio. On the elastic contact of rough surfaces: Numerical experiments and comparisons with recent theories. Wear, 261:11021113, 2006b. R. Cremers. Investigating dynamic tyre model behaviour. Masters thesis, Eindhoven University of Technology, 2005. DCT 2005.37. K.T. Danielson, A.K. Noor, and J.S. Green. Computational strategies for tire modeling and analysis. Computers & Structures, 61(4):673693, 1996. I. Darnell, R. Mousseau, and G. Hulbert. Analysis of tire force and moment response during side slip using an ecient nite element model. Tire Science and Technology, 30(2):6682, 2002. J. Diani, M. Brieu, and P. Gilormini. Observation and modeling of the anisotropic viscohyperelastic behavior of a rubberlike material. International Journal of Solids and Structures, 43:30443056, 2006. Doclib. Internal documentation system Vredestein Banden B.V. I. Doghri, A. Muller, and R.L. Taylor. A general three-dimensional contact procedure for implicit nite element codes. Engineering Computations, 15(2):233259, 1998. V. Dorsch, A. Becker, and L. Vossen. Enhanced rubber friction model for nite element simulations of rolling tyres. Plastics, Rubber and Composites, 31(10):458464, 2002. L.O. Faria, J.M. Bass, J.T. Oden, and E.B. Becker. A three-dimensional rolling contact model for a reinforced rubber tire. Tire Science and Technology, 17(3):217233, 1989. A. Faulkner and R.D. Arnell. The development of a nite element model to simulate the sliding interaction between two, three-dimensional, elastoplastic, hemispherical asperities. Wear, 242: 114122, 2000. FOR492, May 2007. URL http://www.ibnm.uni-hannover.de/{FOR}492/index.html.en. T. Fujikawa, A. Funazaki, and S. Yamazaki. Tire tread temperatures in actual contact areas. Tire Science and Technology, 22(1):1941, 1994. S. Futamura and A. Goldstein. A simple method of handling thermomechanical coupling for temperature computation in a rolling tire. Tire Science and Technology, 32(2):5668, 2004. A. Le Gal, X. Yang, and M. Klppel. Evaluation of sliding friction and contact mechanics of u elastomers based on dynamic-mechanical analysis. Journal of Chemical Physics, 123, 2005. doi: 10.1063/1.1943410. R. Gall, F. Tabaddor, D. Robbins, P. Majors, W. Sheperd, and S. Johnson. Some notes on the nite element analysis of tires. Tire Science and Technology, 23(3):175188, 1995. A.N. Gent. Mechanical properties of rubber. Presentation, 32th Tire Mechanics Short Course, Kln, 2007. o A.N. Gent, J.B. Suh, and S.G. Kelly. Mechanics of rubber shear springs. International Journal of Non-Linear Mechanics, in press, 2007.

BIBLIOGRAPHY

55

J.A. Greenwood and J.B.P. Williamson. Contact of nominally at surfaces. Proceedings of the Royal Society of London, Series A, 295(1442):300319, 1966. J.A. Greenwood and J.J. Wu. Surface roughness and contact: An apology. Meccanica, 36:617630, 2001. J.A. Greenwood, H. Minshall, and D. Tabor. Hysteresis losses in rolling and sliding friction. Proceedings of the Royal Society of London, Series A, 259(1299):480507, 1961. K.A. Grosch. The relation between the friction and visco-elastic properties of rubber. Proceedings of the Royal Society of London, Series A, 274(1356):2139, 1963. A. Haraldsson and P. Wriggers. A strategy for numerical testing of frictional laws with application to contact between soil and concrete. Computer Methods in Applied Mechanics and Engineering, 190:963977, 2000. P. Helnwein, C.H. Liu, G. Meschke, and H.A. Mang. A new 3-d nite element model for cordreinforced rubber composites application to analysis of automobile tires. Finite Elements in Analysis and Design, 14:116, 1993. K. Hofstetter, J. Eberhardsteiner, and H.A. Mang. A thermo-mechanical formulation describing the frictional behavior of rubber. Proceedings in Applied Mathematics and Mechanics, 2:238239, 2003. K. Hofstetter, Ch. Grohs, J. Eberhardsteiner, and H.A. Mang. Sliding behaviour of simplied tire tread patterns investigated by means of FEM. Computers & Structures, 84:11511163, 2006. T. Huemer, W.N. Liu, J. Eberhardsteiner, and H.A. Mang. A 3D nite element formulation describing the frictional behavior of rubber on ice and concrete surfaces. Engineering Computations, 18(3/4):417436, 2001a. T. Huemer, W.N. Liu, J. Eberhardsteiner, H.A. Mang, and G. Meschke. Sliding behavior of rubber on snow and concrete surfaces. Kautschuk Gummi Kunststoe, 54:458462, 2001b. H. Huh and K. Kwak. Finite element stress analysis of the reinforced tire contact problem. Computers & Structures, 36(5):871881, 1990. R.L. Jackson and J.L. Streator. A multi-scale model for contact between rough surfaces. Wear, 261:13371347, 2006. K. Kabe and M. Koishi. Tire cornering simulation using nite element analysis. Journal of Applied Polymer Science, 78:15661572, 2000. M. Kaliske and H. Rothert. Formulation and implementation of three-dimensional viscoelasticity at small and nite strains. Computational Mechanics, 19:228239, 1997. R. Kennedy and J. Padovan. Finite element analysis of steady and transiently moving/rolling nonlinear viscoelastic structure-II. shell and three-dimensional simulations. Computers & Structures, 27(2):259273, 1987. M. Klppel and G. Heinrich. Rubber friction on self-ane road tracks. Rubber chemistry and u technology, 73:578606, 2000. M. Koishi, K. Kabe, and M. Shiratori. Tire cornering simulation using an explicit nite element analysis code. Tire Science and Technology, 26(2):109119, 1998. V. Kouznetsova. Non-linear nite element method for solids. Collegesheets, Eindhoven, 2006. M. Kuwajima, M. Koishi, and J. Sugimura. Contact analysis of tire tread rubber on at surface with microscopic roughness. Tire Science and Technology, 34(4):237255, 2006.

56

BIBLIOGRAPHY

T.A. Laursen. Computational contact and impact mechanics. Springer, 2002. T.A. Laursen and J.C. Simo. A continuum-based nite element formulation for the implicit solution of multibody, large deformation frictional contact problems. International Journal for numerical methods in Engineering, 36:34513485, 1993a. T.A. Laursen and J.C. Simo. Algorithmic symmetrization of coulomb frictional problems using augmented lagrangians. Computer Methods in Applied Mechanics and Engineering, 108:133146, 1993b. T.A. Laursen and I. Stanciulescu. An algorithm for incorporation of frictional sliding conditions within a steady state rolling framework. Communications in numerical methods in engineering, 22:301318, 2006. J. Lemaitre. Handbook of materials and behavior models, volume 2. Acedemic press, 2001. Y.J Lin and S.J. Hwang. Temperature prediction of rolling tires by computer simulation. Mathematics and Computers in Simulation, 67:235249, 2004. W.N. Liu, G. Meschke, and H.A. Mang. On the approximations of the tangential slip in frictional contact analyses. Computers & Structures, 78:5362, 2000. J. Mackerle. Rubber and rubber-like materials, nite-element analyses and simulations, an addendum: a bibliography (1997-2003). Modeling and Simulation in Materials Science and Engineering, 12:10311053, 2004. J. Mackerle. Rubber and rubber-like materials, nite-element analyses and simulations: a bibliography (19761997). Modeling and Simulation in Materials Science and Engineering, 6:171198, 1998. A. Majumdar and B. Bhushan. Role of fractal geometry in roughness characterization and contact mechanics of surfaces. Journal of Tribology, 112:205216, 1990. A. Majumdar and B. Bhushan. Fractal model of elastic-plastic contact between rough surfaces. Journal of Tribology, 113:111, 1991. A. Majumdar and C.L. Tien. Fractal characterization and simulation of rough surfaces. Wear, 136:313327, 1990. J.I. McCool. Comparison of models for the contact of rough surfaces. Wear, 107:3760, 1986. G. Meschke, H.J. Payer, and H.A. Mang. 3d simulations of automobile tires: Material modeling, mesh generation, and solution strategies. Tire Science and Technology, 25(3):154176, 1997. G. Mesehke and P. Helwein. Large-strain 3D-analysis of bre-reinforced composites using rebar elements: hyperelastic formulations for cords. Computational Mechanics, 13:241254, 1994. D.F. Moore. Friction and wear in rubbers and tyres. Wear, 61:273282, 1980. C.A. Morrow. Adhesive Rough Surface Contact. PhD thesis, University of Pittsburgh, 2003. R. Mundl, M. Fischer, M. Wajroch, and S.W. Lee. Simulation and validation of the ply steer residual aligning torque induced by the tyre tread pattern. Vehicle System Dynamics, 43:434 443, 2005. U. Nackenhorst. The ALE-formulation of bodies in rolling contact theoretical foundations and nite element approach. Computer Methods in Applied Mechanics and Engineering, 193:4299 4322, 2004.

BIBLIOGRAPHY

57

Y. Nakajima and J. Padovan. Finite element analysis of steady and transiently moving/rolling nonlinear viscoelastic structure-III. impact/contact simulations. Computers & Structures, 27 (2):275286, 1987. Y. Nakijima, E. Seta, T. Kamegawa, and H. Ogawa. Hydroplaning analysis by FEM and FVM: Eect of tire rolling and tire pattern on hydroplaning. International Journal of automotive technology, 1(1):2634, 2000. A.K. Noor and J.A. Tanner. Advances and trends in the development of computational models for tires. Computers & Structures, 20(1/3):517533, 1985. J.T. Oden and T.L. Lin. On the general rolling contact problem for nite deformations of a viscoelastic cylinder. Computer Methods in Applied Mechanics and Engineering, 57:297367, 1986. J.T. Oden and J.A.C. Martins. Models and computational methods for dynamic friction phenomena. Computer Methods in Applied Mechanics and Engineering, 52:527634, 1985. J.T. Oden and E.B. Pires. Numerical analysis of certain contact problems in elasticity with nonclassical friction laws. Computers & Structures, 16(1/4):481485, 1983. J.T. Oden and E.B. Pires. Algorithms and numerical element approximations of results for nite contact problems with non-classical friction laws. Computers & Structures, 19(1/2):137147, 1984. J.T. Oden, T.L. Lin, and J.M. Bass. A nite element analysis of the general rolling contact problem for a viscoelastic rubber cylinder. Tire Science and Technology, 16(1):1843, 1988. K. Ohishi, H. Suita, and K. Ishihara. The nite element approach to predict the plysteer residual cornering force of tires. Tire Science and Technology, 30(2):122133, 2002. O.A. Olatunbosun and O. Bolarinwa. Fe simulation of the eect of tire design parameters on lateral forces and moments. Tire Science and Technology, 32(3):146163, 2004. O.A. Olatunbosun and A.M. Burke. Finite element modelling of rotating tires in the time domain. Tire Science and Technology, 30(1):1933, 2002. H. Olsson, K.J. strm, C. Canudas de Wit, M. Gfvert, and P. Lischinsky. Friction models and A o a friction compensation. European Journal of Control, 11:1652, 1998. H.B. Pacejka. Tyre and vehicle dynamics. Butterworth-Heinemann, 2nd edition, 2006. J. Padovan. Advanced tire modeling. Presentation, 32th Tire Mechanics Short Course, Kln, o 2007. J. Padovan. Finite element analysis of steady and transiently moving/rolling nonlinear viscoelastic structure-I. theory. Computers & Structures, 27(2):249257, 1987. J. Padovan, A. Kazempour, F. Tabaddor, and B. Brockman. Alternative formulations of rolling contact problems. Finite Elements in Analysis and Design, 11:275284, 1992. H.C. Park, S.-K. Youn, T.S. Song, and N.-J. Kim. Analysis of temperature distribution in a rolling tire due to strain energy dissipation. Tire Science and Technology, 25(3):214228, 1997. A.G. Peressadko, N. Hosoda, and B.N.J. Persson. Inuence of surface roughness on adhesion between elastic bodies. Physical Review Letters, 95, 2005. doi: 10.1103/PhysRevLett.95.124301. B.N.J. Persson. Elastic instabilities at a sliding interface. Physical Review B, 63, 2001a. doi: 10.1103/PhysRevB.63.104101.

58

BIBLIOGRAPHY

B.N.J. Persson. Elastoplastic contact between randomly rough surfaces. Physical Review Letters, 87(11), 2001b. doi: 10.1103/PhysRevLett.87.116101. B.N.J Persson. Adhesion between an elastic body and a randomly rough hard surface. European physical journal E, 8:385401, 2002a. B.N.J. Persson. Adhesion between elastic bodies with randomly rough surfaces. Physical Review Letters, 89(24), 2002b. doi: 10.1103/PhysRevLett.89.245502. B.N.J. Persson. On the nature of surface roughness with application to contact mechanics, sealing, rubber friction and adhesion. Journal of Physics: Condensed matter, 17:R1R62, 2005. B.N.J. Persson. Rubber friction: role of the ash temperature. Journal of Physics: Condensed matter, 2006a. doi: arXiv:cond-mat/0605273. B.N.J. Persson. Contact mechanics for randomly rough surfaces. Surface Science Reports, 61: 201227, 2006b. B.N.J. Persson. Contact mechanics for randomly rough surfaces. Journal of Physics: Condensed matter, 2006c. doi: arXiv:cond-mat/0603807v1. B.N.J. Persson. Theory and simulation of sliding friction. Physical Review Letters, 71(8):1212 1215, 1993. B.N.J. Persson. Theory of friction: Stress domains, relaxation, and creep. Physical Review B, 51 (19):568585, 1995. B.N.J. Persson. Sliding friction: the contribution from defects. Journal of Physics: Condensed matter, 9:28692889, 1997. B.N.J. Persson. On the theory of rubber friction. Surface Science, 401:445454, 1998. B.N.J. Persson. Sliding friction. Surface Science Reports, 33:83119, 1999. B.N.J. Persson and E. Tosatti. Qualitative theory of rubber friction and wear. Journal of Chemical Physics, 112(4):20212029, 2000. B.N.J. Persson and A.I. Volokitin. Dynamical interactions in sliding friction. Surface Science, 457:345356, 2000. B.N.J. Persson and A.I. Volokitin. Theory of rubber friction: Nonstationary sliding. Physical Review B, 65, 2002. doi: 10.1103/PhysRevB.65.134106. B.N.J. Persson and A.I. Volokitin. Rubber friction on smooth surfaces. European physical journal E, 21:6980, 2006. B.N.J. Persson, F. Bucher, and B. Chiaia. Elastic contact between randomly rough surfaces: Comparison of theory with numerical results. Physical Review B, 65, 2002. doi: 10.1103/ PhysRevB.65.184106. B.N.J. Persson, O. Albohr, F. Mancosuc, V. Peveric, V.N. Samoilov, and I.M. Sivebaek. On the nature of the static friction, kinetic friction and creep. Wear, 254:835851, 2003. B.N.J. Persson, U. Tartaglino, O. Albohr, and E. Tosatti. Rubber friction on wet and dry road surfaces: The sealing eect. Physical Review B, 71, 2005. doi: 10.1103/PhysRevB.71.035428. M.G. Pottinger. Tire wear, traction and force generation. Presentation, 32th Tire Mechanics Short Course, Kln, 2007. o G.R. Potts. The tire as a vehicle component. Presentation, 32th Tire Mechanics Short Course, Kln, 2007. o

BIBLIOGRAPHY

59

E. Rabinowicz. The nature of the static and kinetic coecients of friction. Journal of Applied Physics, 22(11):13731379, 1951. K. Rao, R. Kumar, and P. Bohara. Transient nite element analysis of tire dynamic behavior. Tire Science and Technology, 31(2):104127, 2003. K.V. Narasimha Rao, R. Krishna Kumar, P.C. Bohara, and R. Mukhopadhyay. A nite element algorithm for the prediction of steady-state temperatures of rolling tires. Tire Science and Technology, 34(3):195214, 2006. A.R. Savkoor. Dry adhesive friction of elastomers. PhD thesis, Technische Universiteit Delft, 1987. A.R. Savkoor. On the friction of rubber. Wear, 8:222237, 1965. A.R. Savkoor. Some aspects of friction and wear of tyres arising from deformations, slip and stresses at the ground contact. Wear, 9:6678, 1966. A.R. Savkoor. Mechanics of sliding friction of elastomers. Wear, 113:3760, 1986. A. Schallamach. A theory of dynamic rubber friction. Wear, 6:375382, 1963. E. Seta, Y. Nakajima, T. Kamegawa, and H. Ogawa. Hydroplaning analysis by fem and fvm: Eect of tire rolling and tire pattern on hydroplaning. Tire Science and Technology, 28(3): 140156, 2000. J.C. Simo and T.A. Laursen. An augmented lagrangian treatment of contact problems involving friction. Computers & Structures, 42(1):97116, 1992. Tire modeling using ABAQUS. Lecture notes, 2000. u E. Tnk and Y.S. Unlsoy. Prediction of automobile tire cornering force characteristics by nite o u element modeling and analysis. Computers & Structures, 79:12191232, 2001. M. Trinko. Tire stresses and deformation analysis. Presentation, 32th Tire Mechanics Short Course, Kln, 2007. o VMI, May 2007. URL http://www.vmi.nl/uploads/documents/10.pdf. Vol 177, nr 3/4. Special issue: Computational modeling of contact and friction. Computer Methods in Applied Mechanics and Engineering, 177(3/4):163468, 1999. J.D. Walter. Tire materials and manufacturing. Presentation, 32th Tire Mechanics Short Course, Kln, 2007a. o J.D. Walter. Rules and regulations governing tires. Presentation, 32th Tire Mechanics Short Course, Kln, 2007b. o S. Westermann, F. Petry, R. Boes, and G. Thielen. Experimental investigations into the predictive capabilities of current physical rubber friction theories. Kautschuk Gummi Kunststoe, 57:645 650, 2004. M.L. Williams, R.F. Landel, and J.D. Ferry. The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. Journal of the American Chemical Society, 77:37013707, 1955. P. Wriggers. Computational contact mechanics. Springer, second edition, 2006. P. Wriggers, T. Vu Van, and E. Stein. Finite element formulation of large deformation impactcontact problems with friction. Computers & Structures, 37(3):319331, 1990.

60

BIBLIOGRAPHY

C. Yanga, U. Tartaglinoa, and B.N.J. Persson. A multiscale molecular approch to contact mechanics. European physical journal E, 15(1), 2006. G. Zavarise, P. Wriggers, E. Stein, and B.A. Schreer. Real contact mechanisms and nite element formulation A coupled thermomechanical approach. International Journal for numerical methods in Engineering, 35:767785, 1992. G. Zavarise, M. Borri-Brunetto, and M. Paggi. On the resolution dependence of micromechanical contact models. Wear, 262:4254, 2007. I. Zeid and J. Padovan. Finite element modeling of rolling contact. Computers & Structures, 14 (1/2):163170, 1981. M. Ziee and U. Nackenhorst. A new update procedure for internal variables in an ale-description of rolling contact. Proceedings in Applied Mathematics and Mechanics, 5:7174, 2005. O.C. Zienkiewicz and Y.K. Cheung. The Finite Element Method in Structural and Continuum Mechanics, volume 1-3. McGraw-Hill Publ, London, rst edition, 1967.