Predictive control of IPMSM using a precise mathematical model with low implementation demands

The main focus of this paper is to propose high quality, low complexity predictive torque control for IPMSM, which typically exhibits a high level of non-uniformity. Particular solution to this problem may results in high complexity flux mapping, which negatively impacts complexity of the control algorithm. To achieve sufficient accuracy while complexity of the control remain acceptable, we propose to design accurate model which is approximated in a piece-wise manner addressing the most dominant effects on the drive behavior. This is done by state dependent terms using data collected by prior commissioning procedures and verified by FEM simulation. Resulting mathematical model is very simple to implement. We use predictive control with finite control set and cost function following the idea of direct torque control. The proposed controller has been experimentally tested and compared to the controller with conventional mathematical model to demonstrate the effect of the model accuracy on control performance.