基于高频信号注入永磁电机无传感器控制降噪策略

发布时间：2019-03-05 18:24

【摘要】：永磁同步单机具有高功率密度,高效率等优点在工业领域扮演重要角色。无位置传感器驱动技术可以降低成本并且提高系统可靠性,近年来已成为研究热点。高频信号注入法是低速区的较有效的无传感器控制技术。然而,该方法会引起高频噪音和电磁兼容问题,致使该方法在使用场合上受到限制。因此,高频信号注入法需要结合相应的降噪策略以降低其负面影响。本文针对可降低高频噪音的随机频率电压注入法进行了研究。首先分析了传统高频信号注入法的基本原理。建立永磁同步电机高频数学模型,在高频电压的激励下,电机绕组可以简化为纯电感模型。内置式永磁同步电机具有凸极特性,使其高频电流响应中含有转子位置信息,通过合理的信号处理方式可以得到电机转子位置,从而实现电机无传感器控制。第二章分析了高频方波电压注入法和高频正弦波电压注入法的基本原理,并通过Matlab/Simulink软件对两种方法进行了验证。在上述分析结果的基础上,本文对随机频率电压信号注入法进行研究。传统高频信号注入法使用固定频率注入信号,使其电流功率谱密度过于集中,因此产生的噪音尖锐刺耳。使用频率随机变换的电压信号可以有效拓展并降低电流的频谱分布。基于以上原理,本文研究了随机频率电压注入法转子位置提取方式。针对高频注入法受数字系统延时影响机理进行了分析,并提出了相应的解决策略。从电流功率谱密度的角度对传统固定频率注入法和随机频率注入法进行比较分析。尽管随机频率注入法可以有效降低噪音问题,但其电流功率谱密度仍然存在离散谱成分,即仍有集中分布的频率成分。本文从理论上分析了离散谱产生的原因,并提出了相应的解决优化策略。实验结果和理论分析结果吻合,皆证明半周期切换随机频率注入法降低噪音的有效性。最后,通过基于ARM的永磁同步电机实验平台对本文研究内容进行了实验验证。实验表明,随机频率电压注入法可以有效拓展电流功率密度谱,使噪音得到降低。同时,该方法可以较准确的估计电机的位置以及转速信息,保证电机控制性能。
[Abstract]:Permanent magnet synchronous single machine plays an important role in industry because of its high power density and high efficiency. Position sensorless driving technology can reduce the cost and improve the reliability of the system, which has become a hot research topic in recent years. High frequency signal injection is an effective sensorless control technique in low speed region. However, this method will cause high frequency noise and electromagnetic compatibility problems, which limits the use of this method. Therefore, the high frequency signal injection method needs to combine the corresponding de-noising strategy to reduce its negative impact. The random frequency voltage injection method which can reduce high frequency noise is studied in this paper. Firstly, the basic principle of the traditional high frequency signal injection method is analyzed. The high frequency mathematical model of permanent magnet synchronous motor is established. Under the excitation of high frequency voltage, the winding of permanent magnet synchronous motor can be simplified to pure inductance model. The built-in permanent magnet synchronous motor (PMSM) has salient pole characteristic and contains rotor position information in its high frequency current response. The rotor position of the motor can be obtained by means of reasonable signal processing, thus the sensorless control of the motor can be realized. In the second chapter, the basic principles of the high frequency square wave voltage injection method and the high frequency sine wave voltage injection method are analyzed, and the two methods are verified by Matlab/Simulink software. On the basis of the above analysis results, the random frequency and voltage signal injection method is studied in this paper. The traditional high frequency signal injection method uses fixed frequency injection signal, which makes the current power spectrum density too concentrated, so the noise is sharp and harsh. The voltage signal with random frequency transformation can effectively expand and reduce the frequency distribution of the current. Based on the above principles, the method of rotor position extraction by random frequency voltage injection is studied in this paper. In this paper, the influence mechanism of high frequency injection method on digital system delay is analyzed, and the corresponding solving strategy is put forward. The traditional fixed frequency injection method and random frequency injection method are compared and analyzed from the angle of current power spectral density. Although the random frequency injection method can effectively reduce the noise problem, there is still a discrete spectral component in the current power spectral density, that is, there is still a centralized distribution of the frequency component in the current power spectral density. In this paper, the reasons for the generation of discrete spectrum are analyzed theoretically, and the corresponding optimization strategies are put forward. The experimental results are in good agreement with the results of theoretical analysis. It is proved that the half-period switching random frequency injection method is effective in reducing noise. Finally, the research content of this paper is verified by the experiment platform of permanent magnet synchronous motor based on ARM. The experimental results show that the random frequency voltage injection method can effectively expand the current power density spectrum and reduce the noise. At the same time, this method can accurately estimate the motor position and speed information, and ensure the motor control performance.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM351

【参考文献】