中文摘要：

为提取非平稳、强干扰振动信号的有效故障信息,并满足工程实践中变工况故障诊断需要,提出一种基于双谱分析的智能诊断方法。研究发现双谱幅值和分布特性在不同故障类型时具有显著差异性,在故障相同但工况不同时具有较大相似性。使用了主成分分析方法提取特征向量,通过线性变换将双谱映射到低维数据空间,并依据特征值累积贡献率确定主成分个数。故障辨识采用连续型隐马尔可夫模型,在4种工况下实现了3种故障程度的不同轴承状态判别,还实现了基于零载荷数据模型的工况鲁棒故障诊断。研究表明该诊断方法能适应载荷变化和转速波动,具有工况鲁棒的优点。

英文摘要：

This paper presents a method for the fault diagnosis of rolling element bearings using the principal components analysis of the bispectrum and hidden Markov models（HMM）.It has been shown that the bispectrum magnitudes and distribution characteristics are significantly different under different failure status of bearings,but much more similar for the same type of faults under different operation conditions.Thus,the bispectrum can be used as the observation quantity for HMM to detect the fault category.Then,aprincipal component analysis（PCA）based method is used to extract the significant components so as to reduce the dimension of the feature vector,and the number of the principal components is also determined with the eigenvalue cumulative contribution rate.Finally,we conducted a number of experiments under different operation conditions and fault severities to validate the efficiency of the proposed method.The results show that the proposed method is effective in the detection of the various bearings faults,and also robust to the variation of the operation conditions.