Energy-Efficient Model Predictive Speed Control of Permanent Magnet Synchronous Machine Based Automotive Traction Drives

The speed control problem of permanent magnet synchronous machine drives is a challenging problem which has to deal with physical and computational constraints. Moreover, the speed control has to provide also the references of the stator currents. The goal of this paper is to develop an effective control methodology for speed control in permanent magnet synchronous machine drives using a two step approach. Firstly, a model predictive speed control scheme based on control Lyapunov functions is developed from which results the torque reference which has to be converted in references for the stator currents in order to be used by the inner current control loop. Then, secondly, a piecewise affine approximation of the torque function, which is a non-linear non-convex function, is made by gridding the current and torque spaces. As a result, the current reference generation problem is reduced to a convex optimization problem providing guarantees about the error bound introduced by the reference generation procedure. Simulation results are reported and analyzed in order to illustrate the effectiveness of the proposed solution. The obtained results based on realistic simulation scenarios show that the proposed solution can handle both the performance and the existing constraints showing an improvent when comparing to the classical PI control.