基于谱估计与核模糊聚类的往复压缩机轴承故障评估方法

Bearings Fault Evaluation of Reciprocating Compressor Based on Spectral Estimation and KFCM

针对往复机轴承性能衰退评估中模型适应性和指标量化困难等问题,提出一种基于奇异谱分布与核模糊C均值聚类算法(KFCM)的性能衰退评估方法。利用变分模态分解(VMD)算法提取并优选主模态多重分形奇异谱(MSS)构成谱形态参量,经奇异值分解降噪处理,建立KFCM与二叉树支持向量机相结合的评估模型,并给出完整的轴承性能衰退评估流程;进行压缩机轴承磨损故障模拟及算法对比分析。结果表明:该方法能有效评定滑动轴承磨损故障性能衰退程度。

In view of the model adaptability and the difficulty of index quantifying for reciprocating compressor bearings performance degradation evaluation,a performance degradation evaluation algorithm was proposed based on singular spectrum distribution and kernel-based fuzzy C-means clustering method(KFCM).The multifractal singular spectrums(MSS)was extracted by using variation mode decomposition(VMD),the evaluation model combined with KFCM and binary tree SVM was established after denoised through singular value decomposition,and a complete bearings performance degradation evaluation process was given;compressor bearing wear fault simulation and algorithm comparative analysis were carried out.The results show that by using this method,the sliding bearing wear failure performance degradation degree can be effectively evaluated.