基于DVMD和SSAE的柴油机混合故障诊断

Hybrid fault diagnosis of Diesel engine based on DVMD and SSAE

针对柴油机缸盖振动信号非平稳非线性的特点,提出一种基于散布熵改进的变分模态分解(DVMD)和堆叠稀疏自编码器(SSAE)相结合的柴油机混合故障诊断方法。利用散布熵确定变分模态分解的层数K,并根据散布熵转折点选取有效模态分量。分别对选取的各模态分量提取常用14个时域特征和小波包分解后的能量特征,构建混合多特征向量,输入基于堆叠稀疏自编码器和Softmax层构建的深度神经网络(DNN)中,实现了柴油机7种混合故障模式识别。与其他常见方法进行对比,结果表明该方法能够有效提取故障特征,具有较高的诊断准确率。

Here,aiming at non-stationary and nonlinear characteristics of Diesel engine cylinder head vibration signals,a hybrid fault diagnosis method of Diesel engine based on combination of the dispersion entropy-improved variational modal decomposition(DVMD)and the stacked sparse autoencoder(SSAE)was proposed.The number of VMD layers K was determined using dispersion entropy,and effective modal components were selected according to turning points of dispersion entropy.14 commonly used time-domain features and energy features after wavelet packet decomposition were extracted from the selected modal components,respectively to construct a hybrid multi-feature vector,input it into a deep neural network(DNN)built based on SSAE and Softmax layer,and realize recognition of 7 hybrid fault modes of Diesel engine.Compared with other common methods,the results showed that the proposed method can effectively extract fault features and have higher diagnosis accuracy.