Neural network-sliding mode control of permanent magnet synchronous linear motor

Abstract

Direct driving system consisted of permanent magnet synchronous linear motor (PMSLM) eliminate the power transmission devices so that drastically reduce the gearing range. Compared with a traditional indirect transmission system composed of rotary motor and ball screw, it has a significant advantage over the traditional indirect transmission on dynamic precision, rapidity and stability [1]. According to the characteristics of permanent magnet linear synchronous motor with strong coupling and interference, the sliding mode control is adopted in the basic control algorithm in this article. As a result of the obvious chattering phenomenon in traditional sliding mode controller, the neural network control is adopted to control the sliding mode vector. The sliding mode surface is established by thrust and flux integral functions. The sliding mode vector is set as input of RBF neural network and ud, uq are set as the outputs of RBF neural network. In order to make the system reach the sliding surface, adaptive algorithm is adopted according to the accessible condition to adjust the connection weights of RBF neural network online. Compared with the general neural network control, the initial value of connection weights can be set arbitrary in this paper. Simulation results demonstrate that the system designed in this paper can effectively overcome the instability caused by load disturbance and system parameter change and has good adaptability and robustness.