基于EMD-HMM的转盘轴承故障诊断方法

Application of Empirical Mode Decomposition and Hidden Markov Model in Fault Diagnosis of Slewing Bearing

针对转盘轴承故障振动信号非线性、非平稳性的特点,提出了一种将经验模态分解与隐马尔可夫模型相结合的故障诊断方法。首先对故障信号进行经验模态分解,提取固有模态函数的能量作为故障特征向量;然后将故障特征向量输入HMM分类器进行模式识别,输出各状态的似然概率;最后,以最大似然概率所对应的故障状态作为诊断结果。试验表明:该方法能够有效、准确地识别转盘轴承的故障类型,但训练样本数及故障类型数对HMM的诊断精度都有一定的影响。

A fault diagnosis method based on empirical mode decomposition（ EMD） and hidden markov model（ HMM）is proposed in terms of nonlinear and non- stationary of signals of slewing bearing. First of all,fault signals are decomposed by EMD,the energy IMFs are extracted to form the fault feature vector,and then input the feature vectors into HMM classifier for malfunction recognition,the maximum likelihood probability which is output by HMM classifier is in the fault state. Experimental results show that this method can identify the slewing bearing faults effectively,and the number of training samples and fault type has certain influence on the HMM diagnostic accuracy。