基于LMD的风力发电机组振动信号分析
发布时间：2018-12-07 20:14
【摘要】：随着风电的广泛应用，其安全性越来越受到重视。为了保证风力发电机组安全、可靠的运行，对其故障信息的研究具有重要的现实意义。风力发电机组在运行过程中，其振动信号往往包含大量的故障信息，通常表现为多分量的非平稳信号的形式。目前常见的分析方法，例如傅里叶变换、短时傅里叶变换、Winger分布、小波变换等具有一定的局限性。针对以上问题，本文采用了局域均值分解(Local mean decomposition, LMD)和阶比分析(Order Analysis)相结合的方法对风力发电机组的振动信号进行分析。主要研究内容安排如下： 研究LMD的算法，针对滑动步长的选取对LMD的分解的影响，提出了一种自适应选取滑动步长的分解方法；并结合风力发电机在运行过程中会掺入噪声这一问题，对振动信号进行降噪处理。利用改进后的LMD方法对风机的振动信号进行分解。所得到的单分量信号就包含风力发电机组振动信号的特征量。 根据直接法、能量算子解调方法和Hilbert变换法方法的特点，确定提取瞬时频率这一特征量采用直接法，并验证了直接法是提取风力发电机组的旋转机械瞬时频率的有效方法。 通过仿真验证基于瞬时频率估计的阶比分析方法是振动分析的有效方法。并利用LMD算法和直接法获取的瞬时频率，根据瞬时频率确定恒增量角度采样时刻和恒角度增量采样值，，最后进行阶比分析。 利用轴承的振动数据验证基于LMD瞬时频率估计的阶比分析方法是振动信号分析的有效方法。
[Abstract]:With the wide application of wind power, more and more attention has been paid to its safety. In order to ensure the safe and reliable operation of wind turbine, it is of great practical significance to study the fault information of wind turbine. During the operation of wind turbine, the vibration signal of wind turbine often contains a lot of fault information, usually in the form of multi-component non-stationary signal. Some common analytical methods, such as Fourier transform, short time Fourier transform, Winger distribution, wavelet transform and so on, have some limitations. To solve the above problems, the method of local mean decomposition (Local mean decomposition, LMD) and order ratio analysis (Order Analysis) is used to analyze the vibration signal of wind turbine. The main research contents are as follows: the algorithm of LMD is studied, and an adaptive decomposition method of sliding step size is proposed for the influence of the selection of sliding step size on the decomposition of LMD. Combined with the problem that wind turbine will be mixed with noise during operation, the noise reduction of vibration signal is carried out. The improved LMD method is used to decompose the vibration signal of the fan. The obtained single component signal contains the characteristic quantity of the wind turbine vibration signal. According to the characteristics of direct method, energy operator demodulation method and Hilbert transform method, the direct method is adopted to extract the instantaneous frequency, and the direct method is proved to be an effective method for extracting the instantaneous frequency of rotating machinery of wind turbine. Simulation results show that the order analysis method based on instantaneous frequency estimation is an effective method for vibration analysis. The instantaneous frequency obtained by LMD algorithm and direct method is used to determine the sampling time of constant increment angle and the sampling value of constant angle increment according to the instantaneous frequency. Finally, the order analysis is carried out. The order analysis method based on LMD instantaneous frequency estimation is proved to be an effective method for vibration signal analysis by using bearing vibration data.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM315
本文编号：2367792
[Abstract]:With the wide application of wind power, more and more attention has been paid to its safety. In order to ensure the safe and reliable operation of wind turbine, it is of great practical significance to study the fault information of wind turbine. During the operation of wind turbine, the vibration signal of wind turbine often contains a lot of fault information, usually in the form of multi-component non-stationary signal. Some common analytical methods, such as Fourier transform, short time Fourier transform, Winger distribution, wavelet transform and so on, have some limitations. To solve the above problems, the method of local mean decomposition (Local mean decomposition, LMD) and order ratio analysis (Order Analysis) is used to analyze the vibration signal of wind turbine. The main research contents are as follows: the algorithm of LMD is studied, and an adaptive decomposition method of sliding step size is proposed for the influence of the selection of sliding step size on the decomposition of LMD. Combined with the problem that wind turbine will be mixed with noise during operation, the noise reduction of vibration signal is carried out. The improved LMD method is used to decompose the vibration signal of the fan. The obtained single component signal contains the characteristic quantity of the wind turbine vibration signal. According to the characteristics of direct method, energy operator demodulation method and Hilbert transform method, the direct method is adopted to extract the instantaneous frequency, and the direct method is proved to be an effective method for extracting the instantaneous frequency of rotating machinery of wind turbine. Simulation results show that the order analysis method based on instantaneous frequency estimation is an effective method for vibration analysis. The instantaneous frequency obtained by LMD algorithm and direct method is used to determine the sampling time of constant increment angle and the sampling value of constant angle increment according to the instantaneous frequency. Finally, the order analysis is carried out. The order analysis method based on LMD instantaneous frequency estimation is proved to be an effective method for vibration signal analysis by using bearing vibration data.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM315
【引证文献】
相关硕士学位论文 前1条
1 王亚超;基于局部均值分解的旋转机械故障诊断技术研究[D];燕山大学;2015年
本文编号：2367792
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2367792.html
教材专著
