Numerical Methods for Improving the Optimization Efficiency of Textured Surfaces

Le résumé fourni par la source

Abstract As an emerging tribological technology, surface texturing has been widely studied and applied. More and more scholars have committed to the theoretical optimization of textures. To improve the optimization efficiency of textured surfaces, a multiscale method combining the finite cell method with Fischer-Burmeister-Newton-Schur, named as FCM-FBNS, the Fourier Amplitude Sensitivity Test (FAST) and genetic algorithm (GA) were combined in this paper. The proposed methods can calculate the first-order sensitivity indices of texturing parameters to the tribological characteristics efficiently and fastly. Results show that the texture density is the most crucial geometric parameter; the effects of the cell size are negligible. In addition, the effects of texturing parameters have a close relationship with operating conditions and the same parameters may have different influences on different static characteristics. The parameters with higher sensitivity indices can be defined as the key ones and the GA method was applied to improve the tribological performance of textured surfaces by optimizing these texturing parameters. The results achieved by optimizing only the key parameters are similar to that of optimizing all parameters, but the time can be saving up to 30 percent. The methods not only can reduce the optimization difficulty and time, but also provide a novel perspective to understand the optimization problem by quantifing the influence of input on the output.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.