Application of Computational Intelligence to the Analysis of Friction Measurements

摘要

Friction and the associated wear are complex phenomena, strongly affected by the operating and environmental conditions; the mechanical, physical, and chemical properties of the contacting bodies and of the lubricant used; as well as by the treatment of the mating surfaces. In the present article, computational intelligence (CI) methods were applied in order to analyze tribological data collected during a wear experiment, aiming to demonstrate experimental results description and knowledge extraction. Following this successful initial step, CI methods were employed and proved capable of successfully modeling the behavior of parameters of interest, such as the friction coefficient. It is therefore suggested that the computational procedure proposed could be applied in experimental design and evaluation of experimental strategies, as well as in the identification of malfunctioning experimental devices.