Scholarlink

<p>To enhance the accuracy of modeling and control for a wheel-side motor drive system, an electromechanical coupling dynamic model is established utilizing the lumped mass method in this paper. The model incorporates factors such as the bearing system, gear system, unbalanced magnetic pull, and motor torque fluctuation, thereby improving modeling accuracy. The system’s dynamic response is investigated under varying speeds and bearing clearance conditions. Furthermore, a three-inertia modeling approach is proposed to capture the powertrain’s configurational characteristics. In terms of control, a combined strategy integrating a nonlinear disturbance observer with neural network sliding mode control (NDOB-NNSMC) is developed. The disturbance observer is employed to estimate system disturbances in real time. The neural network identifies and compensates for uncertain components to enhance control precision. The stability of the control strategy is proved utilizing the Lyapunov theorem. Simulation results indicate that compared with the sliding mode control and sliding mode control with nonlinear disturbance observer, the proposed control strategy has higher control accuracy. Ultimately, a 200 W three-phase PMSM experimental platform based on the STM32F407 is constructed. Experimental results illustrate that compared with the sliding mode control, the root means square error and absolute errors of NDOB-NNSMC are reduced by 15.16% and 10.60% respectively. Compared with the NDOBSMC strategy, the two indexes decreased by 18.20% and 14.50%.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Electromechanical coupling modeling and analysis base on neural network sliding mode control of wheel-side electric drive system

  • 期刊论文

  • 发表日期：

    2026年5月23日

查看全文
  • Lai Wei,
  • Xiaopeng Li,
  • Jiaqi Liu,
  • Haozhe Wang,
  • ing Fan

摘要

To enhance the accuracy of modeling and control for a wheel-side motor drive system, an electromechanical coupling dynamic model is established utilizing the lumped mass method in this paper. The model incorporates factors such as the bearing system, gear system, unbalanced magnetic pull, and motor torque fluctuation, thereby improving modeling accuracy. The system’s dynamic response is investigated under varying speeds and bearing clearance conditions. Furthermore, a three-inertia modeling approach is proposed to capture the powertrain’s configurational characteristics. In terms of control, a combined strategy integrating a nonlinear disturbance observer with neural network sliding mode control (NDOB-NNSMC) is developed. The disturbance observer is employed to estimate system disturbances in real time. The neural network identifies and compensates for uncertain components to enhance control precision. The stability of the control strategy is proved utilizing the Lyapunov theorem. Simulation results indicate that compared with the sliding mode control and sliding mode control with nonlinear disturbance observer, the proposed control strategy has higher control accuracy. Ultimately, a 200 W three-phase PMSM experimental platform based on the STM32F407 is constructed. Experimental results illustrate that compared with the sliding mode control, the root means square error and absolute errors of NDOB-NNSMC are reduced by 15.16% and 10.60% respectively. Compared with the NDOBSMC strategy, the two indexes decreased by 18.20% and 14.50%.