Dynamic Parameter Optimization of High-Speed Pantograph Based on Swarm Intelligence and Machine Learning

Abstract

Good pantograph–catenary interaction quality is a fundamental premise for ensuring stable and reliable current collection of high-speed trains, and the optimization of dynamic parameters of high-speed pantographs provides an effective approach to improve the current collection quality of the pantograph–catenary system. In this paper, with the objective of minimizing the standard deviation of the pantograph–catenary contact force, the multi-parameter joint optimization for pantograph at different filtering frequencies and running speeds was carried out by using swarm intelligence optimization algorithm and artificial neural network method. First, the selection operator in genetic algorithm (GA) was introduced into crow search algorithm (CSA), and the selective CSA was proposed, which can effectively improve the solution accuracy and convergence rate of multi-parameter optimization. Second, a radial basis function (RBF) neural network was used to construct a surrogate model of the standard deviation of contact force with respect to the running speed and pantograph dynamic parameters, and a method for optimizing the upper limit of mapping interval of the decision variables by the selective crow search algorithm (SCSA) was proposed, which effectively improves the generalization ability of the surrogate model. Finally, by combining the surrogate model and SCSA, optimization iterations for a total of 630 combined conditions such as cut-off frequency, running speed and pantograph dynamic parameters were conducted, and an optimization method for high-speed pantograph dynamic parameters with universal applicability was proposed.