电工技术学报2026，Vol.41Issue(2)：487-498,12.DOI:10.19595/j.cnki.1000-6753.tces.242388
基于单电阻采样的永磁同步电机有限集模型预测控制
Finite-Control-Set Model Predictive Control of Permanent Magnet Synchronous Motor Based on Single Resistor Sampling
摘要
Abstract
Finite-control-set model predictive control(FCS-MPC)has garnered significant attention in permanent magnet synchronous motor(PMSM)control due to its simple structure,multi-objective coordination capability,and nonlinear optimization.Like other high-performance control strategies for PMSMs,FCS-MPC requires real-time phase current sampling.Since current sensors increase both cost and system size,low-power drive systems commonly utilize single-resistor DC-link current sampling.However,during phase current reconstruction,the presence of a dead zone distorts current acquisition,thereby compromising control accuracy.Existing dead-time compensation methods are generally designed for modulated systems and cannot be directly applied to modulator-free discrete switching strategies such as FCS-MPC.Therefore,this paper proposes a finite-control-set model predictive control algorithm based on single-resistor current sampling. Firstly,under zero-vector states of the three-phase inverter,no phase current information is available through single-resistor sampling.In such cases,the current is estimated using the predictive model.During active vector states,the current of one phase is measured via the DC-link resistor,while another phase's current is predicted using the FCS-MPC model.The three-phase currents are then reconstructed based on their balanced relationship.Secondly,to improve reconstruction accuracy,a lumped-parameter model is developed along with a gradient-descent-based adaptive parameter identification algorithm,enabling online parameter estimation and reducing reliance on accurate model parameters.Thirdly,a continuous-state penalty mechanism is introduced by adding a penalty term to the cost function,thereby avoiding model update stagnation. Finally,compared with a conventional SVPWM-based single-resistor sampling method,the proposed current-sampling algorithm was tested on a permanent-magnet synchronous motor driving an oil pump.Two Hall sensors measured the phase currents as a reference for accuracy evaluation.Experimental results show that the proposed method accurately reconstructs the three-phase currents under identical operating conditions,keeping errors within a small margin.The root-mean-square error between reconstructed and sensed currents is reduced by approximately 50%compared to the conventional approach.The proposed algorithm also achieved better total harmonic distortion(THD)performance across different speeds.Starting,reversal,and sudden-load tests verified its dynamic performance and support online parameter identification.The continuous state penalty mechanism has been validated during normal motor operation. The following conclusions can be drawn.(1)A finite control set model predictive control algorithm based on single-resistor sampling is proposed.The issue of FCS-MPC unsuitability for single-resistor sampling scenarios is addressed by employing current reconstruction with the predictive model.(2)Online identification using a lumped parameter model reduces the dependency on precise motor parameters during current reconstruction,thereby improving the accuracy of the reconstructed currents.(3)A continuous state penalty mechanism is proposed to resolve model update stagnation caused by underexcitation during specific operating conditions.关键词
有限集模型预测控制/单电阻采样/电流采样/连续状态惩罚机制Key words
Finite-control-set model predictive control(FCS-MPC)/single-resistor sampling/current sampling/continuous state penalty mechanism分类
信息技术与安全科学引用本文复制引用
赵悦,陈卓易,沈非凡,刘骞宇..基于单电阻采样的永磁同步电机有限集模型预测控制[J].电工技术学报,2026,41(2):487-498,12.基金项目
浙江省自然科学基金资助项目(LTGG23F030002,QN25F030021). （LTGG23F030002,QN25F030021）
