FPGA implementation of model predictive control with constant switching frequency for PMSM drives

Field programmable gate array (FPGA) implementation of a model predictive control with constant switching frequency (MPC-CSF) for a permanent magnet synchronous machine (PMSM) is proposed. The basic finite states MPC (FS-MPC) can be combined with a pulsewidth modulation (PWM) modulator because of an effective cost function optimization algorithm in which voltage vectors are dynamically selected and calculated through iteration based on the idea similar to dichotomy. Using model-based design (MBD), MPC-CSF is implemented on an FPGA with parallel and pipeline processing techniques in short execution time. Functionality simulation analysis presents that MPC-CSF is much robust to parameter variations. Experimental results illustrate that MPC-CSF has good dynamic performance for PMSM drives.