基于EWT-OPRCMDE-ELM的风电机组齿轮箱故障诊断研究

Research of Fault Diagnosis of Wind Turbine Gearbox Based on EWT-OPRCMDE-ELM

针对复杂运行工况和强背景噪声下风电机组齿轮箱故障特征提取和故障模式识别困难的问题。提出一种经验小波变换(EWT)、最优参数精细复合多尺度散布熵(OPRCMDE)和极限学习机(ELM)相结合的故障诊断方法。首先,利用经验小波变换将原始振动信号分解为若干子模态分量(EWF),通过相关系数选取EWF进行信号重构。其次,提取重构信号的最优参数精细复合多尺度散布熵构成故障特征向量,并通过Relief-F算法对特征向量作进一步筛选,剔除冗余。最后,利用极限学习机进行故障诊断。试验分析结果表明,所提方法能够有效提取区分度明显的风电机组齿轮箱故障特征,实现了齿轮箱故障的准确识别。该研究为风电机组齿轮箱故障诊断研究提供了参考,同时具有一定的实际工程应用价值。

Aiming at the problems of fault extraction and fault mode identification in wind turbine gearbox under complex operating conditions and strong background noise,a fault diagnosis method based on a combination of empirical wavelet transform(EWT),optimal parameter refined compound multi-scale dispersion entropy(OPRCMDE)and extreme learning machines(ELM)is proposed.Firstly,the empirical wavelet transform is used to decompose the original vibration signal into several submodal components(EWF).The EWF is selected by the correlation coefficient to perform signal reconstruction.Secondly,the refined compound multi-scale dispersion entropy of the reconstructed signal is extracted to form the fault feature vector.The feature vector is further filtered by the Relief-F algorithm to eliminate redundancy.Finally,an extreme learning machine is used for fault diagnosis.The experimental results show that the proposed method can effectively extract the fault characteristics with obvious discrimination of the wind turbine gearbox and realize the accurate diagnosis.The research provides a reference for the research on fault diagnosis of wind turbine gearbox,and has a certain practical engineering application value.