FPGA based Sliding Mode Predictive Control for PMSM Speed Regulation System Using an Adaptive Ultra-local Model

An adaptive ultralocal surface-mounted permanent-magnet synchronous motor (SPMSM)-model-based continuous control set sliding-mode predictive speed control (UL-MPSC) is proposed in this article to promote the performance of conventional continuous control set model-based predictive speed control (CCS-MPSC). First, an ultralocal SPMSM model that represents the lumped disturbance as the sum of rotor speed, and $q$ -axis current is provided. A forgetting factor recursive least squares identification technique is utilized to identify the coefficients of the rotor speed, and $q$ -axis current, simultaneously. Consequently, the ultralocal SPMSM model is updated at each sampling time, according to I/O data. Second, unlike the error-prediction-based cost function, a fast terminal sliding-mode-based cost function is designed by applying a linear sliding-mode surface, and a fast terminal reaching law. Finally, a discrete-time integral sliding-mode predictive observer is developed to realize a one-step prediction of the reference speed, which is fed forward to the predefined cost function. Experiments of the minimum controller synthesis algorithm, discrete-time integral sliding-mode load torque observer-based CCS-MPSC, and UL-MPSC have been carried out on a field-programmable gate array-based hardware prototype, and the experimental results validate the excellent performance of the UL-MPSC strategy.