As we know, the wear of high-speed trains cannot be avoided, otherwise it cannot brake and stop. But its operating environments are complex and changeable, leading to insufficient research on friction performance of copper (Cu)-based powder metallurgy (PM) brake pads and difficult prediction of the wear rate. Therefore, we explore the variation trend of the friction performance of the brake pad with different multi-factors coupling braking conditions, and predict the wear rate of Cu-based PM brake pad based on Atom Search Optimization-Back Propagation (ASO-BP) neural network. The Cu-based PM brake pad regarded as the research object whose friction coefficient, friction temperature, braking distance, wear rate, and change mechanism are discussed with multi-factors coupling braking conditions, and wear rate is predicted based on ASO-BP neural network that can obtained by optimizing the weights and thresholds of the Back Propagation (BP) neural network with Atom Search Optimization (ASO) method. The results show that braking speed and braking pressure are main contributors to braking ability, the average coefficient of friction of the Cu-based PM brake pads varied nearly between 0.35 and 0.45 with different braking conditions, the maximum of temperature, braking distance, braking time and the wear rate of the brake pad are 473 °C, 3506 m, 138s and 0.14 cm3/MJ, respectively. And the prediction accuracy of brake pad wear rate based on ASO-BP neural network can reach 97.3%.
Read full abstract