Freight Train Speed Active Disturbance Rejection Tracking and Wheel Anti-slip Based on T-S Fuzzy Neural Network

Abstract

Abstract A speed tracking control scheme is proposed for freight trains in this paper. This speed tracking scheme approach can prevent the wheel from slipping while controlling the traction motor to drive freight trains at an appropriate speed. Direct Torque Control (DTC) used by the HXD1 electric traction locomotive is applied in this strategy to control the asynchronous motor. Velocity tracking control is implemented by a Predictive Auto Disturbance Rejection Control (PADRC). Through an modified Smith estimator, this PADRC can predict the response of time-delay systems. In addition, an Unscented Kalman Filter (UKF) observer is designed. An adaptive parameter adjustment mechanism implemented by Affinity Propagation-Fuzzy Neural Network (AP-FNN) is also integrated into this observer. It is used to solve the problem that is difficult to measure the radial velocity and creep rate accurately. Using the anti-slip parameters obtained by this observer, the control scheme of anti-skid control is determined. Under wet and dry pavement conditions, an actual speed curve of a freight train is used to simulate and verify the effectiveness of this scheme.