No abstract

Abstract. The viscoelastic properties of an ethylene/propylene/diene rubber (EPDM) containing 30 parts per hundred parts rubber [phr] carbon black (CB) were determined by dynamic mechanical thermal analysis (DMTA) measurements. A 15-term Maxwell-model was created to describe the time-dependent material behavior of this rubber. The frictional behavior under dry rolling conditions was studied on a home-built rolling ball (steel)-on-plate (rubber) (RBOP) test rig. Both normal and tangential forces were detected during the measurements. The rolling test was simulated with the MSC.Marc finite element (FE) software using the evaluated viscoelastic material properties. Results of the experimental tests and of the simulation were compared and a good agreement was found between them.Keywords: modeling and simulation, rubber, viscoelasticity, finite element method (FEM), rolling friction eXPRESS Polymer Letters Vol.2, No.3 (2008) [157][158][159][160][161][162][163][164] Available online at www.expresspolymlett.com DOI: 10.3144/expresspolymlett.2008.21 The linear-viscoelastic properties of the rubber were determined by dynamic mechanical thermal analysis (DMTA). On the basis of DMTA measurements a master curve was created and a 15-term Maxwell-model was fitted to this master curve. To describe the incompressibility and the non-linear behavior of the rubber, i.e. the non-linear stressstrain curve, the Mooney-Rivlin material model was used. The constants of the two term MooneyRivlin material law were calculated by simplified equations [20][21]. For the simulations the FE software MSC.Marc was used. The rolling friction of an ethylene/propylene/diene rubber (EPDM) was measured in an oscillating rolling ball (steel)-on-plate (rubber) configuration (Oscillating-RBOP) and simulated by FEM. Friction force, normal load and coefficient of friction (COF) were determined. The results of the tests and the simulations were compared with each other and discussed. Experimental Rolling friction testRolling friction tests were carried out on a home built test rig with oscillating rolling ball (100Cr6, d = 2r = 14 mm, R a = 1 µm) and stationary rubber plate (cf. Figure 1). The steel ball is driven by the driver part to conduct oscillating rolling motion on the rubber plate. The load is applied by a lever system. The reciprocating linear rolling of the ball occurred with an amplitude of A = 25.06 mm, a cyclic frequency of f = 1/30 Hz under a normal load of 140 N. The normal and friction load are measured by load cells. The load cell which measures the normal load (Load cell 1, cf. Figure 1) is placed under the base plate, while the other load cell which measures the friction force (Load cell 2) is placed at the driver element. The ball is rolling in the guiding edges of the driver part (cf. Figure 1c). Test conditionsThe linear reciprocating movement of the driver part and the reciprocating rolling of the ball have different amplitudes and speeds due to the set-up. As starting point of our calculations we used the known movement of the...

Abstract. This paper has investigated theoretically the influence of sliding speed and temperature on the hysteretic friction in case of a smooth, reciprocating steel ball sliding on smooth rubber plate by finite element method (FEM). Generalized Maxwell-models combined with Mooney-Rivlin model have been used to describe the material behaviour of the ethylenepropylene-diene-monomer (EPDM) rubber studied. Additionally, the effect of the technique applied at the parameter identification of the material model and the number of Maxwell elements on the coefficient of friction (COF) was also investigated. Finally, the open parameter of the Greenwood-Tabor analytical model has been determined from a fit to the FE results. By fitting, as usual, the Maxwell-model to the storage modulus master curve the predicted COF, in a broad frequency range, will be underestimated even in case of 40-term Maxwell-model. To obtain more accurate numerical prediction or to provide an upper limit for the hysteretic friction, in the interesting frequency range, the Maxwell parameters should be determined, as proposed, from a fit to the measured loss factor master curve. This conclusion can be generalized for all the FE simulations where the hysteresis plays an important role.

No abstract