开关磁阻电机有源自回归双矢量无模型预测电流控制方法OA

Driven by the global low-carbon transition and carbon-neutrality goals,the new energy vehicle(NEV)industry has rapidly emerged as one of the fastest-growing and most promising technological frontiers.Since rare-earth materials are non-renewable strategic resources,the switched reluctance motor(SRM)requires zero rare-earth elements.SRM offers low cost,high durability,simple rotor structure,wide speed regulation range,high reliability,and simple maintenance and repair.It also demonstrates strong potential for electric vehicle applications.Traditional predictive control for SRM has the following disadvantages.(1)Predictive control demonstrates inherent sensitivity to motor parameter variations,resulting in deteriorated control performance.(2)SRM has severe nonlinear characteristics,making it challenging to implement predictive control strategies.(3)Conventional single voltage vector optimization period causes large phase current ripples.To improve the dynamic response of SRM drive systems,this paper investigates high-performance current control strategies. This paper proposes a dual-vector model-free predictive current control(DV-MFPCC)based on an auto-regressive with exogenous input(ARX)function.The active autoregressive model is established based on phase-voltage and current measurements.Then,the normalized least mean square(NLMS)algorithm is used to estimate the parameter vector in the ARX model,and the predicted phase current is obtained.To reduce phase current ripple,a dual-voltage vector control strategy is designed based on the basic voltage vector characteristics.The dual-voltage vector is optimized.The allocation time for the dual-voltage vector combination is calculated by minimizing the evaluation function.The optimal dual-voltage vector combination is determined. An experimental platform for a three-phase 12/8 structure SRM drive system is established.The control chip is the TMS320F28335,and the sampling chip is AD7606.The simulation model mainly includes the electromechanical equation,asymmetric half-bridge power converter,control signal generation,and phase winding modules.In the experiments,the proposed DV-MFPCC-ARX strategy is compared with MFPCC-ARX and DV-MPCC strategies in terms of current ripple and control robustness under steady-state,acceleration,loading,and parameter-mismatch conditions.The experimental results show that:(1)By combining the ARX function with the switched reluctance motor system,a current prediction model is obtained.The normalized least-mean-squares algorithm is used to estimate the coefficient vector in the autoregressive function.This algorithm has low computational complexity and requires only one control parameter for current prediction,eliminating dependence on motor parameters.(2)The proposed method develops an optimized dual-vector control strategy.Within each control period,it selects the optimal dual-vector combination based on voltage-vector-pairing principles and minimization of the evaluation function,while determining the corresponding time allocation.(3)The proposed method demonstrates enhanced robustness during steady-state,acceleration,loading,and parameter mismatch conditions.(4)The developed dual-vector model-free predictive control strategy can be readily integrated with speed closed-loop control and other strategies,facilitating online implementation.