Effective FPGA-based electric motor modeling with floating-point cores

The simulation of electromechanical systems like motor drives often requires sub-microsecond calculation timesteps considering the fast dynamic of such systems and the high-switching frequency involved. Migrating computational load to an FPGA processor has proven to effectively meet the real-time simulation needs of such systems. However, many challenges still must be overcome before broad adoption of FPGA technology for real-time simulation applications occurs. In this paper, a general framework is presented for effective use of FPGA machine drive modeling when the state-space approach is used. Computations are performed in floating-point using commercially available arithmetic cores. Using the discussed framework guarantees that time steps well below 1 µs can be achieved. Two real-world applications examples are given in the paper: an FPGA-based implementation of a BLDC motor, and an FPGA-based implementation of an induction motor.