Surface roughness causes contact to occur only at discrete spots called microcontacts. In the deterministic models real area of contact and pressure field are widely evaluated using Flamant and Boussinesq equations for two-dimensional (2D) and three-dimensional (3D), respectively. In this paper, a new 3D geometrical contact approach is developed. It models the roughness by cones and uses the concept of representative strain at each asperity. To discuss the validity of this model, a numerical solution is introduced by using the spectral method and another 3D geometrical approach which models the roughness by spheres. The real area of contact and the pressure field given by these approaches show that the conical model is almost insensitive to the degree of isotropy of the rough surfaces, which is not the case for the spherical model that is only valid for quasi-isotropic surfaces. The comparison between elastic and elastoplastic models reveals that for a surface with a low roughness, the elastic approach is sufficient to model the rough contact. However, for surfaces having a great roughness, the elastoplastic approach is more appropriate to determine the real area of contact and pressure distribution. The results of this study show also that the roughness scale modifies the real contact area and pressure distribution. The surfaces characterized by high frequencies are less resistant in contact and present the lowest real area of contact and the most important mean pressure.

In this work, we study the influence of roughness geometries on the contact force, contact area and pressure distribution during a static contact between a rough surface and a smooth rigid plane by using two elastic geometrical models. The first one is spherical where each asperity can be characterised by the height and the radius of curvature. The second approach is a conical where the rough contact is modelled by cones. In order to discuss the validity of these two geometrical models, they are compared with a numerical solution of the Boussinesq equation which does not take into account the local geometry of the asperities. The results show clearly that the conical model allows having more realistic values of the contact parameters and model correctly the elastic rough contact. However, the spherical approach of the asperities overestimates the pressures undergone by the asperities.

This study characterizes seal performance using a multiscale analysis of surface topography. The performance of two surface morphologies is compared: the first one is obtained with machining only and leads to leakage while the second one is obtained with machining and superfinishing and prevents leakage. It is shown that conventional roughness analysis does not enable to identify the differences between both surfaces. Only the use of a new parameter, the order parameter, and the use of a multiscale analysis of surfaces enable to distinguish the studied surfaces and to identify leakage causes. These causes are checked using a numerical contact simulation. It is shown that microroughness plays a major role in leakage.