Adaptive sliding mode adhesion control of high-speed train based on disturbance compensation

The safe and stable operation of the train depends on the effective exertion of the traction force, which is limited by the adhesion state between the wheel and rail. To solve the problem of insufficient traction force of high-speed train running on low-adhesion rail surface, considering the difficulty in accurately obtaining parameters such as train weight, resistance, and slope, an adaptive sliding mode adhesion control algorithm based on disturbance compensation was designed. According to the adhesion calculation model and the simplified train model, the wheel-rail dynamic model of high-speed train was established. A robust adhesion observer with higher dynamic observation accuracy was designed to accurately obtain the adhesion force provided by the current rail surface. On this basis, aiming at the characteristic that the train can obtain large traction force when it works near the optimal adhesion point, the gradient observation extremum seeking algorithm was used to obtain the creep speed near the point. At the same time, an adaptive sliding mode controller based on disturbance compensation was designed to control the output torque of the motor. The high-speed train can track the creep speed output by the extremum seeking algorithm under the condition that some parameters of system are unknown, thereby improving the traction force. Adaptive sliding mode control based on disturbance compensation reduces the influence of control law chattering on the adhesion control performance by observing and compensating the uncertain part of the system and adaptively adjusting the switch gain. The results show that the adaptive sliding mode adhesion control algorithm based on disturbance compensation can control the train to reach the optimal adhesion working point of ice and snow rail surface within 10 s and reach the optimal adhesion working point of wet rail surface within 1 s when there are unknown parameters in the system. In addition, compared with the traditional sliding mode adhesion control and adaptive sliding model adhesion control, the proposed algorithm has lower control law chattering and can search and track the optimal adhesion working point more accurately. The proposed algorithm can provide a reference for the design of actual train adhesion control algorithm. © 2024, Central South University Press. All rights reserved.