Data-driven design of brake pad composites for high-speed trains

Abstract

Brake pads play a vital role in controlling the operation of high-speed trains with over 300 km/h. Currently the copper-based composites produced by powder metallurgy techniques have been proved as one of the ideal materials. However，the braking performance is strongly influenced by the chemical composition, production, microstructure as well as the working conditions. The intrinsic mechanism of the thermal fade and wear loss under thermal and mechanical coupled influence has not yet been fully revealed. As of late, machine learning algorithms have attracted widespread attention and extraordinary results has been accomplished in materials science and engineering. In this study the algorithms for predicting the braking performance of copper-based brake pads are developed. A new PM copper-based alloy, was designed via the models of the coefficient of friction (COF) and wear loss. The braking test results of this alloy show that the average COF was 0.37 at stable stage I and thermal recession decrease only 0.012 at stage II when under 350 km/h/0.48 MPa continuous braking for 10 times. And the total wear loss is only 1.30 g when under braking from 200 km/h to 350 km/h with pressure from 0.31 MPa to 0.48 MPa. Compared with the commercial brake pads, our designed alloy exhibits excellent braking performance with COF for thermal fade reduced to 21% and the total wear loss reduced to 20%.