【目的】实际工程中，部分异步电机未配置转速传感器和转矩测量装置，导致电机能效难以评估，因此研究转速和转矩的辨识方法具有重要意义。本文提出一种基于可测电压和电流的异步电机运行参数在线辨识方法。该方法对电压和电流波形没有限制，且不需要电机的定子电阻、电抗等参数，适用范围广、实用性强。【方法】首先，对定子电流进行快速傅里叶变换分析找到对应的转子槽频率，进而求出转子转速。通过转差率将转子槽频率限制在一定区间内，从而解决齿谐波混叠的问题，提高转速辨识的精度。其次，通过电机铭牌参数求出定子电阻，结合采集到的电压和电流数据计算电磁转矩和输出转矩。最后，为进一步提高所提方法的适用范围，引入基波补偿法，对滤波后的信号进行全局缩放，以精确恢复基波幅值，进一步提高辨识精度。【结果】基于Matlab搭建异步电机直接起动和变频驱动仿真模型，采集输出数据。并且对型号为Y100L1-4的异步电机进行试验，采集电压、电流试验波形。仿真和试验结果表明，不同工况下的转子转速和输出转矩的辨识误差都在2%以内，验证了所提方法的可行性与精确性。【结论】本文所提方法实现了基于可测电气量的异步电机运行参数的在线辨识，为实际工程中实时辨识异步电机运行参数提供了重要参考。

[Objective] In practical engineering, some asynchronous motors are not equipped with speed sensor and torque measuring instrument, making it difficult to evaluate the energy efficiency of motors, so it is of great significance to study the identification method of speed and torque. This paper proposes an on-line identification method of asynchronous motor operation parameters based on measurable voltage and current. The method has no limitation on voltage and current waveforms, and does not need motor parameters such as stator resistance and reactance, so it is widely applicable and highly practical. [Methods] Firstly, the corresponding rotor slot harmonic frequency was found by fast Fourier transform analysis of stator current, and then the rotor speed was calculated. By constraining the rotor slot harmonic frequency within a certain range through the slip rate, the problem of tooth harmonic aliasing was resolved, enhancing the accuracy of speed identification. Secondly, the stator resistance was calculated by the motor nameplate parameters, and the electromagnetic torque and output torque were calculated by combining the collected voltage and current data. Finally, to further enhance the applicability of the proposed method, the fundamental wave compensation method was introduced to globally scale the filtered signal, accurately restoring the fundamental wave amplitude and further improving identification accuracy. [Results] The direct start and variable frequency drive models of asynchronous motor were built by Matlab to collect output data. Furthermore, experiments were conducted on the Y100L1-4 asynchronous motor to collect voltage and current experimental waveforms. Simulation and experimental results demonstrated that the identification errors for rotor speed and output torque remained within 2% under various operating conditions, thereby validating the feasibility and accuracy of the proposed method. [Conclusion] The proposed method achieves on-line identification of asynchronous motor operation parameters based on measurable electrical quantities, providing a significant reference for real-time identification of asynchronous motor operation parameters in practical engineering applications.

国家自然科学基金(52177042)