Multi-Scale Surface Roughness Optimization Through Genetic Algorithms

Abstract

Artificial intelligence is changing perspectives of industries about manufacturing of components, introducing emerging techniques such as additive manufacturing technologies. These techniques can be exploited to manufacture not only precision mechanical components, but also interfaces. In this context, we investigate the use of artificial intelligence and in particular genetic algorithms to identify optimal multi-scale roughness features to design prototype surfaces achieving a target contact mechanics response. Exploiting an analogy with biology, the features of roughness at a given length scale are described through model profiles named chromosomes. In the present work, the mathematical description of chromosomes is firstly provided, then three genetic algorithms are proposed to superimpose and combine them in order to identify optimal roughness features. The three methods are compared, discussing the topological and spectral features of roughness obtained in each case.