Hybrid Adaptive Integral Sliding Mode Speed Control of PMSM System Using RBF Neural Network

In this paper, a hybrid adaptive integral sliding mode control (HAISMC) based on radial basis function neural network (RBFNN) is proposed for permanent magnet synchronous motor (PMSM) speed control system. HAISMC is generally divided into the reaching phase and the sliding phase. In the reaching phase, linear integral sliding mode control (LISMC) with switching gain varying linearly is adopted. In the sliding phase, a radial basis function neural network (RBFNN) is applied to predict external disturbances on the basis of LISMC. The parameters of RBFNN are fully tuned online. The linearly varying switching gain of LISMC can cope with external disturbances in the reaching phase and RBFNN approximation error in the sliding phase. The stability of the PMSM system is proved by the Lyapunov stability theorem. At the end of the paper, proportional integral (PI) control, linear integral sliding mode control (LISMC), and HAISMC are compared. Simulation and experimental results show that HAISMC has better robustness and reduces chattering.