基于小波域平稳子空间分析的风力发电机齿轮箱故障诊断

发布时间：2018-03-01 07:50

本文关键词： 风力发电机齿轮箱故障诊断连续小波变换平稳子空间分析　出处：《机械工程学报》2014年11期 　论文类型：期刊论文

【摘要】：风力发电机齿轮箱故障信号为非平稳瞬态微弱信号,容易被齿轮啮合信号及其他噪声淹没。提出一种融合连续小波变换(Continuous wavdot transform,CWT)和平稳子空间分析(Startionary subspace analysis,SSA)的信号分解方法并应用于风力发电机齿轮箱故障诊断中。平稳子空间分析作为一种盲源分离技术可将高维数据分解成平稳源部分和非平稳源部分,对待分析信号各分量间的独立性没有要求且不需要任何先验信息。连续小波变换则可利用其所具有的多尺度分析特性把一维时间序列转换为不同尺度下的多维时间序列。对观测得到的一维时间序列数据进行连续小波变换得到多维时间序列作为平稳子空间分析的输入,利用平稳子空间分析方法将该多维时间序列分解为平稳源信号分量和非平稳源信号分量,对非平稳源信号进行包络谱分析得到齿轮箱故障的特征频率。该小波域平稳子空间分析方法被应用于一个实际风力发电机齿轮箱振动信号的分析,试验结果表明该方法可有效地诊断出齿轮箱中的轴承故障。
[Abstract]:The fault signal of gearbox of wind turbine is non-stationary transient weak signal. It is easy to be submerged by gear meshing signal and other noise. A signal decomposition method based on continuous wavdot transform (CWT) and stationary subspace analysis (SSA) is proposed and applied to wind turbine gearbox fault diagnosis. As a blind source separation technique, subspace analysis can decompose high-dimensional data into stationary source and non-stationary source. The independence of each component of the analytical signal is not required and no prior information is required. Continuous wavelet transform can be used to convert one dimensional time series to multidimensional time at different scales by using its multiscale analysis characteristics. Using continuous wavelet transform to obtain multi-dimensional time series as the input of stationary subspace analysis. The multidimensional time series is decomposed into stationary source signal component and non-stationary source signal component by using stationary subspace analysis method. The characteristic frequency of gearbox fault is obtained by envelope spectrum analysis of non-stationary source signals. The wavelet domain stationary subspace analysis method is applied to the analysis of vibration signals of a practical wind turbine gearbox. The test results show that this method can effectively diagnose the bearing faults in the gearbox.
【作者单位】：东南大学仪器科学与工程学院;康涅狄格大学机械工程系;
【基金】：国家自然科学基金资助项目(51175080)
【分类号】：TM315

【参考文献】