基于单神经元的无刷直流电机无感矢量控制研究

微特电机 ›› 2026, Vol. 54 ›› Issue (2): 76-82.

基于单神经元的无刷直流电机无感矢量控制研究

刘　峰¹,赵彦普¹,刘宗琛¹,张　响²,翟天嵩¹,刘　源¹

1. 南阳理工学院智能制造学院,南阳 473004;2. 郑州大学力学与安全工程学院,郑州 450001

出版日期:2026-02-28 发布日期:2026-02-28
作者简介:刘峰( 1983—) ,博士,副教授,研究方向为电机高精度控制技术。
基金资助:
河南省科技公关项目( 242102221052 ) ; 南阳市协同创新重大专项( 22XTCX12005 ) ; 河南省自然科学基金 ( 252300421932, 252300420919) ;南阳市科技计划项目 ( 23JCQY2018 ) ; 河南省校企协同创新项目 ( 26AXQXT079、 26AXQXT077 ) ; 南阳市重大科技专项( 25ZDZX010)

Research on Sensorless Brushless DC Motor Vector Control System Based on a Single Neuron

LIU Feng¹,ZHAO Yanpu¹, LIU Zongchen¹,ZHANG Xiang²,ZHAI Tiansong¹,LIU Yuan¹#br#

1. Intelligent Manufacturing College,Nanyang Institute of Technology,Nanyang 473004,China;2. College of Mechanics and Safety Engineering,Zhengzhou University,Zhengzhou 450001,China

Online:2026-02-28 Published:2026-02-28

摘要/Abstract

摘要： 针对无刷直流电机无感矢量控制系统中滑模观测器因为电机参数时变特性导致转子位置估计精度下降的问题,本文提出一种基于单神经元的位置环自适应控制策略。该策略利用单神经元的在线学习能力,实时在线调整权值以补偿参数扰动引起的位置估计偏差,不仅实现了电机转矩与磁通的精确解耦控制,更显著提升了位置跟踪精度,并且有效增强了系统应对负载与参数变化的自适应能力与鲁棒性能。本系统在硬件方面,选择 STM32F103单片机作为主控芯片,搭配电源电路、驱动逆变电路和采样电路等关键组件;在软件与控制算法方面,以矢量控制作为基础框架,在位置环控制中创新性的引入了单神经元自适应控制器。仿真与实验结果表明,在转速为 1 200 r / min的稳态工况下,系统突加负载扰动后,能在时间 0. 01 s 内通过自调整迅速恢复稳定,且整个恢复过程的转速动态降落被限制在 5. 4%以内。表明该控制方法在时变参数工况下,仍能保持良好的转速控制的稳定性与系统鲁棒性,为高性能无刷直流电机驱动系统的设计提供了有效的解决方案。

关键词: 无刷直流电机, 矢量控制系统, 逆变电路, STM32F103

Abstract: Aiming at the problem of reduced rotor position estimation accuracy in sensorless vector control systems for brushless DC motors due to time-varying motor parameters in sliding mode observers,this paper proposes an adaptive control strategy for the position loop based on a single neuron. This strategy leverages the online learning capability of the single neuron to adjust weights in real-time,compensating for position estimation deviations caused by parameter disturbances. It not only achieves precise decoupling control of motor torque and flux but also significantly improves position tracking accuracy,while effectively enhancing the systems ’ adaptive ability and robustness in response to load and parameter variations. In terms of hardware,the system adopts the STM32F103 microcontroller as the main control chip,paired with key components such as the power supply circuit, drive inverter circuit, and sampling circuit. In terms of software and control algorithms,the system is built on a vector control framework, with the innovative introduction of a single-neuron adaptive controller in the position loop control. Simulation and experimental results show that under steady-state conditions at speed 1 200 r / min, the system can rapidly recover stability through self-adjustment within time 0. 01 s after a sudden load disturbance,with the dynamic speed drop during the entire recovery process limited to within 5. 4%. This demonstrates that the proposed control method maintains good speed control stability and system robustness under time-varying parameter conditions,providing an effective solution for the design of high-performance brushless DC motor drive systems.

Key words: brushless DC motor, vector control system, inverter circuit, STM32F103

中图分类号:

TM351

刘峰, 赵彦普, 刘宗琛, 张　响, 翟天嵩, 刘　源. 基于单神经元的无刷直流电机无感矢量控制研究[J]. 微特电机, 2026, 54(2): 76-82.

LIU Feng, ZHAO Yanpu, LIU Zongchen, ZHANG Xiang, ZHAI Tiansong, LIU Yuan. Research on Sensorless Brushless DC Motor Vector Control System Based on a Single Neuron[J]. Small & Special Electrical Machines, 2026, 54(2): 76-82.