基于自适应EKF的感应电机无传感器控制  

Sensorless control of induction motor based on adaptive EKF

作　　者：胡堃[1] 徐磊[1] 周君威 田里思 HU Kun;XU Lei;ZHOU Junwei;TIAN Lisi(School of Electrical Engineering,China University of mining and technology,Xuzhou 221116,China)

机构地区：[1]中国矿业大学电气工程学院,江苏徐州221116

出　　处：《实验技术与管理》2023年第9期117-124,共8页Experimental Technology and Management

基　　金：江苏省高等教育教改研究项目(2021JSJG218);中国矿业大学教学研究一般项目(2021YB77);中国矿业大学教学研究重点项目(2022ZDKT03-209);国家自然科学基金项目(62003349)。

摘　　要：针对传统EKF(TEKF)需要长时间试凑噪声协方差矩阵且噪声特性变化时估算性能下降甚至发散的问题,提出了一种基于极大似然估计准则和有限记忆指数加权的自适应EKF算法(EW-MLE-AEKF)。在使用新息序列进行自适应调整时,系统容易因新息序列的不准确而发散,为此可以利用后验残差序列对噪声协方差矩阵Q和R进行实时调整来提高系统的稳定性。同时在加窗方法的基础上使用了有限内存指数加权算法来提高近期数据的权重,加快了估计的收敛速度。仿真和实验验证了算法在电机系统中的可行性,结果证明该算法避免了噪声矩阵的试凑过程,能够适应噪声的变化,提高了滤波精度且系统稳定性强。Aiming at the problem that the traditional EKF(TEKF)needs to try to gather the noise covariance matrix for a long time,and the estimation performance degrades or even diverges when the noise characteristics change,an improved adaptive EKF algorithm is proposed based on the maximum likelihood estimation criterion and finite memory index weighting(EW-MLE-AEKF).When the new information sequence is used for adaptive adjustment,the system is easy to diverge due to the inaccuracy of the new information sequence.Therefore,a posterior residual sequence can be used to adjust the noise covariance matrix Q and R in real time to improve the stability of the system.To speed up the convergence speed of the estimation,a limited memory exponential weighting algorithm is added to the windowing method,which can increase the importance of recent data.The feasibility of the algorithm in the motor system is verified by simulation and experiment.The results show that the algorithm avoids the trial-and-error process of noise matrix,adapts to the change of noise,improves the filtering accuracy and has strong system stability.

关 键 词：极大似然估计 后验残差序列 有限记忆指数加权 无传感器控制 自适应扩展卡尔曼滤波 

分 类 号：TM343[电气工程—电机]