基于谱相关密度-连续HMM的滚动轴承故障诊断

摘要

滚动轴承故障振动信号是典型的调幅信号,而谱相关密度分析对调幅信号具有解调功能,它可以有效地提取出滚动轴承信号的故障特征,结合连续隐马尔可夫模型（Hidden Markov Model,HMM）所具有的强大时序模式分类能力,提出了基于谱相关密度-连续HMM的滚动轴承故障诊断方法。该方法首先利用谱相关密度函数在循环频率处进行切片分析,提取滚动轴承故障振动信号的特征,构成特征向量序列;然后将此序列输入到连续HMM中进行训练,得到各类对应故障的模型,最后利用训练好的模型进行滚动轴承的故障诊断。试验结果验证了该方法的可行性和有效性。%The vibration signal of the rolling bearing is a typical amplitude modulated signal,and the spectral correlation density（SCD） analysis can be used as a tool of demodulation to facilitate effective extraction of fault features of the vibration signal.In combination with the timing pattern classification ability of the continuous hidden markov model（HMM）,a method of diagnosing faults of rolling bearings based on the SCD and continuous HMM is proposed.Firstly,SCD function is used to carry out slicing analysis at the cyclic frequency,and then features of vibration signals of rolling bearing faults are extracted to constitute characteristic vector sequences.Then the sequences are input into the continuous HMM to be trained,so as to acquire various models corresponding to different faults.Finally the trained models are used to diagnose the faults of the rolling bearing.The testing results verify the feasibility and effectivity of the method.