Optimal controller design for A 2-AXLE railway vehicle suspension system

Abstract

The project involves designing the controller for a two-axle railway vehicle suspension system. The input to the system is the curve radius, its corresponding cant angle and random track irregularities. The performances of the controller is evaluated based on the wheelset lateral displacement relative to the pure rolling line and yaw angle of the front and rear wheelset, both on straight and curved track. An optimized linear quadratic regulator (LQR) has been designed. The LQR is one of the standard techniques for control system design. In LQR design problem, the choices of controller’s weighting parameter Q and R is crucial for the performance of the LQR controller. To optimized the Q value, a population-based search algorithm. Particle Swarm Optimization is used. PSO algorithm is initialized with random particles and search for optimum value by updating the velocities and positions of the particles. In this project, R value is fixed, whereas the parameter Q is optimized using the PSO approach. The optimum Q is chosen based on the least value of integral absolute error (1AE) of the lateral displacement and yaw angle values. This technique is compared with the trial and error approach of choosing the controller's weighting matrices.