基于ELMD与DHMM的风机齿轮箱故障诊断方法

Fault diagnosis method for the fan gearbox based on ELMD and DHMM

针对风机齿轮箱早期故障信号信噪比低且故障难以准确诊断的问题,提出了基于集合局部均值分解(ELMD)与离散隐马尔科夫模型(DHMM)的风机齿轮箱故障诊断方法。首先对风机齿轮箱故障振动信号进行ELMD分解,得到一系列的乘积分量(PF),再对每个PF与原始信号求取相关系数进而滤除噪声信号以及由分解引起的虚假分量,然后对新信号进行标量量化处理得到特征向量,最后将每种状态下的特征向量输入已训练收敛的离散隐马尔科夫模型库进行状态判别并得出诊断结果。实验结果表明,对于风机齿轮箱早期故障诊断,所提出的方法具有一定的有效性和实用性。

Aiming at the accurately diagnose of the early failure for the fan gearbox,it proposes a fault diagnosis method for the fan gearbox based on Ensemble Local Mean Decomposition(ELMD)and Discrete Hidden Markov Model(DHMM).It uses ELMD decomposition for the fault vibration signal of the wind turbine gearbox,obtains a series of Product Function(PF),takes the correlation coefficients for each PF and the original signal to filter out the noise signal,and realizes the scalar quantization.Inputting the eigenvectors in each state into the discrete hidden Markov model of training convergence for the state discrimination and the diagnosis result,it obtains the eigenvectors.The experimental results show the effectiveness and practicability of the proposed method for early fault diagnosis of wind turbine gearbox.