基于参数辨识的永磁同步直线电机循环神经网络多维观测器

doi:10.19595/j.cnki.1000-6753.tces.231621

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (5762 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract In recent years, with the development of computer technology, control theory, and material technology, the performance requirements for computer numerical control (CNC) machine tools have become increasingly high. The permanent magnet linear synchronous motor (PMLSM) has been widely used in aerospace, semiconductor, and high-speed automated processing equipment due to its fast response, accurate positioning, and high reliability. During the operation of PMSLM, the parameters change correspondingly with operation conditions and environment, which affects the high-performance control of PMSLM and even causes serious machine failures. Neural network observers are studied for speed and motor parameter identification to improve the robustness and dynamic performance of the system. This paper studies a high-precision recurrent neural network (RNN) for multi-dimensional observation of PMSLM based on intelligent online extended Kalman filtering (IOEKF).
Since PMSLM is a complex nonlinear time-varying system with multiple variables and strong coupling, it is necessary to simplify the actual system for analysis when establishing the mathematical model. Starting from the basic structure and equivalent circuit of the motor, a three-phase PMSLM model with mutual inductance is derived in a synchronous rotating coordinate system. A closed-loop control system for the three-phase PMLSM is established.
Secondly, based on the principle and structure of RNN, a multi-dimensional observer for three-phase PMSLM is established. The weight coefficient update of the neural network online is solved using the EKF algorithm. A semi-physical PMSLM control platform based on MT 1050 is built. Simulation and experimental results show that the proposed multi-dimensional observer has higher accuracy than the full-order state observer and the observer without EKF algorithm update under steady-state and dynamic conditions. The Lyapunov stability criterion provides theoretical validation for the observer's stability. The stability of the observer is theoretically ensured by constructing an appropriate Lyapunov function and demonstrating the negativity of its derivative.
Based on the variations of resistance and inductance, the corresponding identification models are established, and the EKF multi-parameter online identification method is proposed. The least squares method is used as a comparison. The results show that the EKF multi-parameter identification algorithm for motor parameter identification has better dynamic and steady-state performance.
The proposed observation method for PMLSM enhances the overall performance and reliability of CNC machine tools. The RNN-based multi-dimensional observer, supported by IOEKF, reduces the training time of the control algorithm, achieving real-time and high-precision performance in the control of linear motors.

Key words： Permanent magnet linear synchronous motor recurrent neural network multidimensional observer parameter identification convergence analysis

Received: 06 October 2023

PACS:

TM351

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Song Lin
	Nie Ziling
	Sun Jun
	Zhou Yangwei
	Li Huayu

Cite this article:

Song Lin,Nie Ziling,Sun Jun等. Multidimensional Observer of Permanent Magnet Synchronous Linear Motor Recurrent Neural Network Based on Parameter Identification[J]. Transactions of China Electrotechnical Society, 2024, 39(22): 7059-7072.

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.231621 OR https://dgjsxb.ces-transaction.com/EN/Y2024/V39/I22/7059

[1] 周世炯, 李耀华, 史黎明, 等. 分段式双三相永磁直线同步电机的无模型电流预测控制[J]. 电工技术学报, 2024, 39(4): 996-1009, 1021.
Zhou Shijiong, Li Yaohua, Shi Liming, et al.Model- free predictive current control of segmented dual three-phase permanent magnet linear synchronous motor[J]. Transactions of China Electrotechnical Society, 2024, 39(4): 996-1009, 1021.
[2] 方馨, 王丽梅, 张康. 基于扰动观测器的永磁直线电机高阶非奇异快速终端滑模控制[J]. 电工技术学报, 2023, 38(2): 409-421.
Fang Xin, Wang Limei, Zhang Kang.High order nonsingular fast terminal sliding mode control of permanent magnet linear motor based on disturbance observer[J]. Transactions of China Electrotechnical Society, 2023, 38(2): 409-421.
[3] 武志涛, 李帅, 程万胜. 基于扩展滑模扰动观测器的永磁直线同步电机定结构滑模位置跟踪控制[J]. 电工技术学报, 2022, 37(10): 2503-2512.
Wu Zhitao, Li Shuai, Cheng Wansheng.Fixed structure sliding mode position tracking control for permanent magnet linear synchronous motor based on extended sliding mode disturbance observer[J]. Transactions of China Electrotechnical Society, 2022, 37(10): 2503-2512.
[4] 郭科宇, 李耀华, 史黎明. 分段供电双三相永磁同步直线电机电流解耦与扰动抑制[J]. 电工技术学报, 2022, 37(24): 6332-6344.
Guo Keyu, Li Yaohua, Shi Liming.Current decoupling and force disturbance suppression of dual three-phase permanent magnet linear synchronous motors with section power supply[J]. Transactions of China Electrotechnical Society, 2022, 37(24): 6332-6344.
[5] 鲍明堃, 周扬忠. 基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制[J]. 电工技术学报, 2024, 39(8): 2470-2484.
Bao Mingkun, Zhou Yangzhong.Parallel dis- placement velocity control of permanent magnet synchronous linear motor with variable parameters based on composite neural network reconstruction object[J]. Transactions of China Electrotechnical Society, 2024, 39(8): 2470-2484.
[6] Nguyen N D, Nam N N, Yoon C, et al.Speed sensorless model predictive torque control of induction motors using a modified adaptive full-order observer[J]. IEEE Transactions on Industrial Elec- tronics, 2022, 69(6): 6162-6172.
[7] 杨阳, 赵吉文, 宋俊材, 等. 基于深度神经网络模型的无铁心永磁同步直线电机结构优化研究[J]. 中国电机工程学报, 2019, 39(20): 6085-6094, 6189.
Yang Yang, Zhao Jiwen, Song Juncai, et al.Structural optimization of air-core permanent magnet syn- chronous linear motors based on deep neural network models[J]. Proceedings of the CSEE, 2019, 39(20): 6085-6094, 6189.
[8] Bai Chenyao.AGA-GRU: an optimized GRU neural network model based on adaptive genetic algorithm[J]. Journal of Physics: Conference Series, 2020, 1651(1): 012146.
[9] Volpato Filho C J, Vieira R P. Adaptive full-order observer analysis and design for sensorless interior permanent magnet synchronous motors drives[J]. IEEE Transactions on Industrial Electronics, 2021, 68(8): 6527-6536.
[10] Auger F, Hilairet M, Guerrero J M, et al.Industrial applications of the Kalman filter: a review[J]. IEEE Transactions on Industrial Electronics, 2013, 60(12): 5458-5471.
[11] Geng Yiwen, Lai Zou, Li Yonggang, et al.Sensorless fault-tolerant control strategy of six-phase induction machine based on harmonic suppression and sliding mode observer[J]. IEEE Access, 2019, 7: 110086-110102.
[12] Fu Zhijun, Xie Wenfang, Na Jing.Robust adaptive nonlinear observer design via multi-time scales neural network[J]. Neurocomputing, 2016, 190: 217-225.
[13] Yu Xiangqian, Zhang Chi, Su Tao.Speed sensorless of induction motor with adaptive full order state observer[J]. Journal of Physics: Conference Series, 2020, 1550(4): 042058.
[14] Sun Xiaodong, Wu Jiangling, Lei Gang, et al.Torque ripple reduction of SRM drive using improved direct torque control with sliding mode controller and observer[J]. IEEE Transactions on Industrial Elec- tronics, 2021, 68(10): 9334-9345.
[15] 荆禄宗, 吴钦木. 基于递归神经网络的永磁同步电机参数辨识研究[J]. 电气传动, 2020, 50(3): 87-91, 101.
Jing Luzong, Wu Qinmu.Research on identification of PMSM based on recurrent neural network[J]. Electric Drive, 2020, 50(3): 87-91, 101.
[16] El-Sousy F F M, Abuhasel K A. Intelligent adaptive dynamic surface control system with recurrent wavelet Elman neural networks for DSP-based induction motor servo drives[C]//2017 IEEE Industry Applications Society Annual Meeting, Cincinnati, OH, USA, 2017: 1-16.
[17] 苏有成, 陈志辉. 基于电感扰动的三相横向磁通永磁电机参数辨识与估算位置偏差修正[J]. 电工技术学报, 2023, 38(12): 3165-3175.
Su Youcheng, Chen Zhihui.Parameter identification and estimation of three-phase transverse flux per- manent magnet motor based on inductive disturbance, correction of position deviation[J]. Transactions of China Electrotechnical Society, 2023, 38(12): 3165-3175.
[18] Zheng Tianyu, Yao Yu, He Fenghua, et al.An RNN- based learnable extended Kalman filter design and application[C]//2019 18th European Control Con- ference (ECC), Naples, Italy, 2019: 3304-3309.
[19] 王法胜, 郭权. 基于扩展卡尔曼粒子滤波算法的神经网络训练[J]. 计算机工程与科学, 2010, 32(5): 48-50.
Wang Fasheng, Guo Quan.Neural network training based on the extended Kalman particle filter[J]. Computer Engineering & Science, 2010, 32(5): 48-50.
[20] Archana R, Unnikrishnan A, Gopikakumari R.An improved EKF based neural network training algorithm for the identification of chaotic systems driven by time series[C]//2012 International Con- ference on Power, Signals, Controls and Computation, Thrissur, India, 2012: 1-6.
[21] 张明远, 史黎明, 徐飞, 等. 单变流器驱动多分段并联长初级双边直线感应电动机模型与控制策略[J]. 中国电机工程学报, 2022, 42(22): 8343-8352.
Zhang Mingyuan, Shi Liming, Xu Fei, et al.Modeling and control strategy of long primary double-sided linear induction motor segmented in parallel fed by a single inverter[J]. Proceedings of the CSEE, 2022, 42(22): 8343-8352.