In order to further study the dump truck disc brake thermal-structural coupling characteristics

by analyzing the major impact parameter of brake pressure

brake initial speed

plate/sheet friction coefficient and the equivalent moment of inertia on the disc brake performance and using orthogonal experiment principle

the GA-BP network training sample is formed

a GA-BP network model with traditional genetic algorithm BP neural network combined with the introduction of disc brakes thermal-structural coupling finite element predictive model is proposed.The study results show that for the temperature and stress time history results

the overall trend of the training results and finite element simulation results are basically the same

but in the individual condition curve

there are some differences in local curves.Aiming at the specific conditions

the main characteristics of the temperature-time history is basiclly predicted by the GA-BP Network

the data extreme values of stress time history prediction effect and other major information are consistent with the finite element results.The training network has higher approximation performance and better prediction performance

the maximum error of prediction and simulation is only 8%

the computational accuracy is higher.