Robust predictive current control for performance improvement of induction motors with parameter variation

In this paper a robust predictive current control approach is developed to tackle the impacts of parameter variation in induction motors on the control performance. The robustness can be achieved in general model predictive control (MPC) by integrating the disturbance into the prediction model, which is estimated by an observer and whose effect is compensated to the system input or output. In such a way the controller is robust against those uncertainties and the system can converge on the required available operating point. However, due to the variation, a non neglectable overshooting appears inevitably, which is critical for systems with hard constraints. In contrast a robust predictive control based on linear matrix inequalities (LMIs) implemented in this paper shows a significant improvement of current control in induction motors. For the real-time application the approximated multiparametric semidefinite programming (mp-SDP) with orthogonal partitioning is approached.