Scholarlink

<p>Accurate estimation of rotor position is vital for sensorless control of Permanent Magnet Synchronous Motor (PMSM) drive. In this paper, an Adaptive Neuro-Fuzzy Inference System (ANFIS) enhanced Sliding Mode Observer (SMO) for sensorless control of PMSM is presented. SMO acts as a primary estimator for estimating back electromotive force (back-EMF), speed, and position, whereas ANFIS acts as a secondary estimator, serving as a post-processor to refine the estimated position for smooth sensorless control. Additionally, this proposed ANFIS-SMO-based observer performs well in the low-speed region, which is particularly critical for other observers. The proposed methodology is validated through simulation and experimental analysis for different dynamic conditions. From the analysis, it is observed that there is reduction in torque ripple, improvement in speed tracking, and a reduction of approximately 2% error in position estimation. These improvements made the proposed controller robust for sensorless control of PMSMs in various applications.</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

An Improved Sensorless PMSM Drive Performance Using Data-Driven ANFIS-SMO Framework

  • 期刊论文

  • 发表日期：

    2026年4月2日

查看全文
  • Shilpa Ranjan,
  • Madhusudan Singh,
  • Mini Sreejeth

摘要

Accurate estimation of rotor position is vital for sensorless control of Permanent Magnet Synchronous Motor (PMSM) drive. In this paper, an Adaptive Neuro-Fuzzy Inference System (ANFIS) enhanced Sliding Mode Observer (SMO) for sensorless control of PMSM is presented. SMO acts as a primary estimator for estimating back electromotive force (back-EMF), speed, and position, whereas ANFIS acts as a secondary estimator, serving as a post-processor to refine the estimated position for smooth sensorless control. Additionally, this proposed ANFIS-SMO-based observer performs well in the low-speed region, which is particularly critical for other observers. The proposed methodology is validated through simulation and experimental analysis for different dynamic conditions. From the analysis, it is observed that there is reduction in torque ripple, improvement in speed tracking, and a reduction of approximately 2% error in position estimation. These improvements made the proposed controller robust for sensorless control of PMSMs in various applications.