Finite element based modeling of multi-asperity contact and validation of Greenwood-Williamson model

Tribology;Asperity interaction;Finite element;Greenwood-Williamson;contact;asperity

Chandrasekar, Sujana ; Polycarpou ; Andreas A.

关键词: Tribology;    Asperity interaction;    Finite element;    Greenwood-Williamson;    contact;    asperity;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/24467/Chandrasekar_Sujana.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

The Greenwood-Williamson (GW) model of contact between rough surfaces is a widely acclaimed model that has been used to predict contact stiffness and other contact parameters. A main limitation of this model is often cited as it is ignoring of asperity interaction effects. This thesis aims to validate the Greenwood-Williamson model by using the finite element method to study multi-asperity contact and asperity interaction effects. The first step in the study of multi-asperity contact was the development of the geometry of rough surfaces as a collection of hemispherical caps with a statistical distribution of heights, as visualized in the GW model. In this study, different geometries corresponding to different asperity densities but same roughness were created. All geometries consisted of 25 asperities positioned on a 2-dimensional grid. Finite Element (FE) models corresponding to different geometries were created and executed. Each model consisted of ~1-3 million degrees of freedom. To our knowledge, this kind of intense FEA simulation has not been executed for a contact mechanics problem till date. Results of the FE models were compared to results with an extension to an existing analytical model of asperity interaction originally proposed by Ciavarella et al. The extended model includes elastic, elastic-plastic and plastic contacts and it assumes that asperity interaction effects are primarily due to substrate deformation. A comparison of FE model results and the extended Ciavarella et al. model indicated that the analytical model matched the FE model results well at small asperity spacing (i.e., spacing less than half of the asperity radius) but overestimated interaction effects for asperity spacing greater than half the asperity radius. An explanation of this effect is provided using the difference between substrate and asperity pressure profiles at different interferences. It is also shown that the GW model matches the FE model results with an error of less than 10% for very high values of asperity spacing due to negligible interaction effects at high asperity spacing. The conclusion of this work is that substrate deformation is the dominant factor in the interaction of asperities but knowledge of the pressure profile at the substrate is essential to develop an analytical asperity interaction model that matches physical reality more closely than the extended elastic-plastic analytical model.