Improved Dual-Vector Predictive Flux Control with Parameter Identification for Three-Level Inverter-Fed Linear Induction Machines in Linear Metro

This article proposes an improved dual-vector predictive flux control (IDV-PFC) for three-level inverter-fed linear induction machine (LIM) drives adapted to the linear metro. Existing dual-vector-based methods face the challenges of increased control targets and neutral point voltage (NPV) balance if directly extended to three-level inverter-fed LIM drives for steady-state performance improvement. Moreover, as a model-based method, PFC is easily affected by the drastic nonlinear variation of magnetizing inductance in LIM. To solve these issues, based on the efficient region division, designed cost function, and switching action optimization, a low-complexity PFC with dual vector is developed, where a high control quality of the flux and thrust and a strong balancing of the NPV are achieved. Then, one online magnetizing inductance identification is proposed based on a model reference adaptive system (MRAS), where pure integrator in conventional identification is eliminated by employing back electromotive force (EMF) as a new observed quantity. Further, this online identification is integrated into the PFC scheme to improve its robustness against magnetizing inductance mismatch. Finally, extensive simulation and experiments are carried out, showing that the proposed IDV-PFC can effectively improve the steady-state performance and eliminate the impact of magnetizing inductance mismatch on the system performance.