电工技术学报2026，Vol.41Issue(2)：475-486,12.DOI:10.19595/j.cnki.1000-6753.tces.250135
基于陷波抗停滞的永磁同步电机无模型预测电流滑模控制
Notch Anti-Stagnation Based Model-Free Predictive Current Sliding Mode Control of Permanent Magnet Synchronous Motor Drives
摘要
Abstract
Model-free predictive sliding mode control(SMC)achieves complete independence from physical models and parameters by constructing a data-driven model.However,its modeling and updating processes demand high-quality sampled data.Stagnation and its adverse effects have emerged as a significant challenge.This paper proposes a notch anti-stagnation-based model-free predictive current SMC method for permanent magnet synchronous motor(PMSM)drives.A notch term is designed to extract specific frequency band harmonics generated by the control strategy.These harmonics are inversely injected into the sampled data to force the generation of data gradients.This paper aims to mitigate the risk of stagnation and provide an effective method for high-performance control of PMSMs in complex environments. The proposed method consists of the following key steps.Firstly,a notch term is designed by adjusting the notch width and depth.The harmonics produced by the control strategy do not contribute to model updates.This process generates data gradients,reducing the likelihood of stagnation.Secondly,a universal model is established based on the data gradients,and all model coefficients are estimated using the recursive least squares(RLS)algorithm.The prediction process incorporates time-shift considerations,and the model is ultra-localized to implement an intelligent proportional function as the predictive component in the control function.Thirdly,a non-singular fast sliding mode surface is designed.The sliding mode component of the control function integrates both equivalent and switching control laws.Then,the state variables converge onto the sliding mode surface,forming the complete control function of the system.Finally,the system's stability,robustness,and reachability of the sliding mode component,as well as the convergence and stability of the estimation,are discussed under theoretical conditions. Experimental results on the PMSM platform demonstrate the effectiveness of the proposed method.During steady-states with different load torques and speed references,the proposed method reduces the number of stagnation occurrences from the compared method's 382 and 339 to 8.In terms of phase current quality,5th and 7th harmonic contents are reduced,as evidenced by the total harmonic distortion(THD)of single sample points.Under a load torque of 11.5 N·m,continuous Fourier analysis of 1000 sample points shows that the average THD value and THD covering range are reduced from 7.588％ and 1.980％to 6.69％ and 1.203％,respectively.Similar improvements are observed under a load torque of 4.6 N·m.The proposed method decreases the speed's integrated time and absolute error(ITAE)from 0.02276 to 0.02129.Finally,guidelines for selecting the main parameters of the proposed method are provided. The following conclusions can be drawn.(1)Compared with the model-free SMC method without anti-stagnation,the proposed method significantly reduces the possibility of stagnation by generating data gradients,effectively addressing stagnation and its adverse effects.(2)The adaptability of the universal model in the predictive component is improved,and the current quality and prediction accuracy are both enhanced by combining with the non-singular fast sliding mode surface in the sliding mode component.(3)The proposed method is independent of physical models and parameters,exhibiting enhanced robustness.关键词
无模型预测滑模控制/陷波抗停滞部分/数据驱动模型/永磁同步电机Key words
Model-free predictive current sliding mode control/notch anti-stagnation section/data-driven model/permanent magnet synchronous motor分类
信息技术与安全科学引用本文复制引用
魏尧,付俊荣,王高林,汪凤翔..基于陷波抗停滞的永磁同步电机无模型预测电流滑模控制[J].电工技术学报,2026,41(2):475-486,12.基金项目
国家自然科学基金(52277070)和福建省自然科学基金(2024J08125)资助项目. （52277070）
