基于小波分析与神经网络的气阀机构故障诊断研究

运用小波分析对柴油机缸盖振动信号进行分析与讨论，计算二进小汉分解后尺度1信号在各个时间段内的能量百分比；将能量百分比作为神经网络的输入进行训练和故障识别，用BP网络及自组织聚类算法实现了气阀机构的故障诊断，取得了较好的效果。

Fault Diagnosis for Valve Train Based on Neural Networks and Wavelet Analysis

Vibration signal measured on cylinder head was analyzed and processed with wavelet analysis.The energy distribution in time domain of different scales has a certain principle.According to the air-matching of 6135G diesel,signals of different scales were divided into 6 segments.Energy percentage was computed.The BP neural network was trained with the energy percentage as the ideal input and with the codes of fault states as the ideal output.After the neural network was successfully trained,it can be used to diagnose valve train faults.And the ART network was also used to carry out the clustering of different valve train states.The results show that they can carry out the diagnosing of valve train effectively and concisely.