Radial basis function neural network control of chaotic motion of locomotive traction gear system

According to the operation performance monitoring requirements of HXD2 traction gear system, the dynamic model of the single degree traction gear system is established and by combining the bifurcation diagram, phase diagram, and Poincaré cross section diagram, it is able to examine how the damping coefficient and the change in engagement stiffness affect the system's periodic response. Using a radial basis function neural network (RBFNN), we created a parameter feedback controller. The quantum particle swarm algorithm (QPSO) is used to optimize the controller's parameters. By applying a microamplitude perturbation to the damping coefficient, The system chaotic motion is controlled as a stable periodic motion.