KPCA法在风电机组齿轮箱故障诊断中的应用

将核主成分分析法应用于风电机组齿轮箱的故障诊断中,通过计算齿轮箱振动信号原始数据空间的内积核函数来实现原始数据到特征空间的非线性映射。利用某风场齿轮箱的正常工作状态、初期磨损状态以及断齿状态下的振动数据进行测试,对主成分分析法和核主成分分析法的分类结果进行了分析比较。实验结果表明,核主成分分析法能够有效地对齿轮故障信号进行特征提取和模式分类,更适合于故障信号非线性特征的提取。

Application of KCPA Method in Fault Diagnosis for Wind Turbine Gearbox

In this paper, the Kernel Principal Component Analysis（KPCA） method is applied to the turbine gearbox fault diagnosis, through calculating the inner product kernel function of the original data space of gearbox vibration signal, the nonlinear mapping from original to high dimension space is realized. The experimental data of a gearbox in normal state, gear wear conditions and broken teeth conditions are used for testing. And the classification results of Principal Component Analysis（PCA） and KPCA are compared. The experimental results show that the KPCA can be more effective on the fault feature extraction and pattern classification; it is more suitable for nonlinear feature extraction of fault signal.