Neural network prediction of brake friction materials wear

Abstract

Wear of brake friction materials depends on many factors such as temperature, applied load, sliding velocity, properties of mating materials, and durability of the transfer layer. Prediction of friction materials wear versus their formulation and manufacturing conditions in synergy with brakes operating conditions can be considered as a crucial issue for further friction materials development. In this paper, the artificial neural network abilities have been used for predicting wear of the friction materials versus influence of all relevant factors. The neural model of friction materials wear has been developed taking into account: (i) complete formulation of the friction material (18 ingredients), (ii) the most important manufacturing conditions of the friction material (5 parameters), (iii) applied load and sliding velocity of the friction material both represented by work done by brake application, and (iv) brake interface temperature.