Low-Complexity Multistep Model Predictive Current Control for Linear Induction Machines

Abstract

In this article, the predictive horizon of model predictive current control (MPCC) is extended for achieving better operational performance. However, the computational burden of MPCC exponentially increases with its predictive steps, and thus, the predictive step is usually chosen as one. In order to explore the potential advantage of multistep MPCC (M-MPCC), one simplified search method is proposed so that the M-MPCC can be executed in practice with greatly reduced complexity. In the proposed search method, a large number of possible voltage vector sequences can be excluded in advance with the guide of the defined reference voltage vector sequence. Then, one simplified M-MPCC is applied to two 3 kW arc induction machines with a large radius, which can make a full demonstration on the drive performance of linear induction machine in practice. The comprehensive simulation and experiments have clearly indicated the relationship between the predictive step and the performance of the proposed M-MPCC. It has persuasively verified the feasibility of low complexity and decreasing time consumption based on the proposed search method in this work.