Preferences in multi-objective evolutionary optimisation of electric motor speed control with hardware in the loop

Abstract

This paper presents the design of a robust Pareto-optimal controller with the designer preference articulation for a Permanent Magnet Synchronous Motor (PMSM). An evolutionary multi-objective optimisation (EMOO) algorithm is used to tune the proportional integral (PI) speed regulator in the Direct Torque Control drive system. Approximation of the Pareto front with hardware in the loop (HiL) is chosen as an alternative to the time-consuming software simulation studies. Thanks to this approach problems of un-modelled plant dynamics are alleviated and additional manual tuning on-line is not required. The weak point of the HiL approach is caused by disruptive presence of noise which affects the performance of EMOO. This influence is strongly problem-dependent; therefore no generalized results have yet been presented in the literature. In this paper the robustness features of the proposed design approach are verified using the state-of-the-art multi-objective evolutionary Non-dominated Sorting Genetic Algorithm (NSGA-II; Deb, 2001 [1]). The on-line optimised motor drive speed controller is shown to be effective, possessing good dynamic characteristics, demonstrating applicability of the a priori preference articulation technique to the controller design. The final Pareto-optimal solution is selected according to the designer's preference articulation before the Pareto front is approximated by EMOO.