基于VMD能量熵和HMM的行星齿轮箱故障识别方法

针对行星齿轮箱振动信号成分复杂、非平稳、非线性的特点，提出一种基于变分模态分解(VMD)能量熵和隐马尔科夫模型(HMM)的故障识别方法。利用VMD算法对不同故障类型的齿轮振动信号进行分解，提取经信号分解得到的各阶本征模态函数(IMF)的能量熵。基于不同故障类型的各IMF分量能量熵在分布上的各异性，将其集合作为故障识别的特征向量。利用不同故障类型的特征向量组成的训练集训练HMM，计算最大对数似然概率值，用于判断测试样本的故障类型。利用该方法对一定转速下行星轮的3种故障进行识别，结果表明：当载荷不同时，它对行星轮齿根裂纹、断齿和齿面磨损3种故障的平均识别率可达到95.83%

Fault Identification Method for Planetary Gearbox Based on VMD Energy-entropy and HMM

Aimed at the complex, non-stationary and nonlinear problem of planetary gearbox vibration signals, a fault identification method based on variational mode decomposition (VMD) energy-entropy and hidden Markov model (HMM) was proposed. VMD algorithm was used to decompose gear vibration signals with different fault types, and the energy-entropy of each order intrinsic mode function (IMF) obtained by signal decomposition was extracted.The sets were used as the feature vector for fault identification based on the dissimilarities of the energy entropy of each IMF components in the distribution of different fault types. The training set composed of feature vectors with different fault types was used to train the HMM, and the maximum log likelihood probability value was calculated to determine the fault type of the test sample.The method was used to identify three faults of the planetary gear at a certain speed.The results show that the average recognition rate for root crack, broken tooth and tooth surface wear of planetary gear under different load conditions can reach 95.83%