一种W型结构元素的自适应数学形态学及其在风电机组轴承故障诊断中应用

为了提高风电机组滚动轴承故障诊断的有效性和可靠性,提出一种W型自适应数学形态学特征提取方法,并与谱相关分析相结合形成风电机组滚动轴承故障诊断策略。该方法首先针对传统三角型结构元素在故障特征提取中易出现对脉冲信号的漏查,提出一种W型结构元素,旨在捕捉更多特征信息;之后依据各故障信号的实际波形得到结构元素的高和最优开闭运算加权因子,构建自适应形态学模型;最后对测试信号与训练信号进行频域内谱相关性分析,依据相关系数识别故障。将该方法通过数值例、西储大学实验台轴承数据和实际风场采集数据进行算法验证,并与传统的三角型结构元素进行比较,实验结果表明W型结构元素能更有效地提取信号中的脉冲成分、降低噪声干扰,故障诊断算法可准确识别出故障类别,提高结果的可靠性。

MATHEMATICAL MORPHOLOGY ANALYSIS BASED ON IMPROVED STRUCTURING ELEMENT AND ITS APPLICATION IN WIND TURBINE BEARING FAULT DIAGNOSIS

A new method for fault diagnosis is proposed based on an adaptive mathematical morphology with a structure element of W shape(W-SE)and spectral correlation analysis. Firstly,the W-SE is constructed to capture more feature information while the pulse signal is apt to be missed by means of traditional triangular structure element(triangular-SE).Then,according to the actual waveform of each fault signal,the high and optimal open and close operation weighting factors of W-SE are obtained,and the adaptive morphological model is established. Finally,the fault type is confirmed based on the size of correlation coefficient through spectral correlation analysis between the test signal and the training one in frequency domain. The proposed method is validated by numerical examples,bearing data from Case Western Reserve University and data collected from actual wind field,and compared with traditional triangular-SE. The results show that the W-SE is an more effective and reliable algorithm for fault feature extraction in respect of extracting impulse information from vibration signals,and the proposed method performs better on diagnosis accuracy than the traditional method.