Model Predictive Current Control for Linear Induction Machine

Abstract

In this chapter, the model predictive control (MPC) is applied to linear induction machine (LIM) with the equivalent circuit of LIM proposed by Duncan, which can fulfil multiple control goals and consider different constraint conditions at the same time. However, it consumes high computing cost to evaluate and compare each candidate voltage vector produced by inverter so as to achieve the best one with the smallest cost function value. Therefore, it is meaningful to find one simplified method to reduce its heavy computation burden, especially when the current limit is considered as one constraint condition in cost function to ensure the safe operation. Moreover, the MPC applies only one voltage vector within the whole sampling period, resulting in high current ripples. Facing with this problem, the modulation strategies based on two and even three voltage vectors are combined with MPC for improving the control freedom and precision. Besides, the predictive steps of MPC is increased instead of only one step so as to further improve the control performance of MPC. And then, according to the optimal value without constraints of multistep model predictive control (MMPC), one iterative algorithm is proposed to search a suboptimal value, which satisfies both the current and voltage limitations. Finally, one test bench with two 3 kW arc induction motors that can simulate the actual LIM operation traits is built for testing the performance of MPC for LIM.