No abstract

No abstract

The percolation threshold strongly affects sealing performance. This paper investigates the relationship between the percolation threshold and the rough surface anisotropy, which is represented by the Peklenik number, γ. A series of anisotropic rough surfaces were generated and the conjugate gradient-fast Fourier transform (CG-FFT) method was used to determine the percolation threshold. The percolation threshold was found to be A/A0≈0.484±0.009 (averaged over 45 surfaces) was established for an isotropic rough surface (γ=1). Furthermore, it was also found that the percolation threshold decreased from A/A0≈0.528±0.011 to A/A0≈0.431±0.008 as 1/γ increased from 0.6 to 2. Our results differ from the theoretical result of Persson et al., where A/A0=γ/(1+γ). Comparing our calculated results with the theoretical results established the presence of an intersection value of 1/γ that was related to the effect of elastic deformation on the percolation threshold. When 1/γ was smaller than the intersection value, our calculated results were lower than the theoretical ones; and when 1/γ was greater than the intersection value, our calculated results were higher than the theoretical ones.

No abstract

Percolation threshold is a very important parameter to estimate the sealing performance. Thus, it is crucial to determine the correct value of the percolation threshold for contact sealing surfaces. In this paper, we applied a numerical generation method, in which the autocorrelation length can be easily controlled, to obtain different Gaussian isotropic rough surfaces. Then, the contact status between a rigid flat half-space and numerically generated rough surfaces were calculated using the conjugate gradient-fast Fourier transform method. Based on the contact status, the percolation threshold was obtained using a search method. The calculated results established that the percolation threshold of [Formula: see text] is determined for Gaussian isotropic contacting rough surfaces. To obtain an exact value of the percolation threshold, the finite size of the generated rough surfaces should be six times greater than the autocorrelation length, and the autocorrelation length should not be smaller than 20 times the sampling interval.