基于双树复小波包和PNN的柴油机故障诊断研究

针对传统小波变换在故障特征提取中的不足,提出一种基于双树复小波包和概率神经网络(PNN)的故障诊断方法。首先通过双树复小波包变换将各个工况的柴油机声信号分解得到不同频带的分量,选取各频带分量的能量作为特征向量,再利用PNN对特征向量进行训练,最后通过测试样本得到柴油机典型故障诊断结果。实验表明,该方法可以对柴油机典型故障进行较为准确的诊断,相比传统小波包有着更高的故障诊断率。

Research on Fault Diagnosis of Diesel Engine Based on Dual Tree Complex Wavelet Packet and PNN

Aiming at the shortcomings of traditional wavelet transform in fault feature extranction, a fault diagnosis method based on dual tree complex wavelet packet and probabilistic neural network (PNN) is proposed. Firstly, the diesel engine sound signals under various working conditions are decomposed into components of different frequencybands through dual tree complex packet transformation. Finally through the test samples of typical diesel engine fault diagnosis results are obtained. Experimental results show that this method can more accurate diagnosis a typical fault for diesel engine, which compared with the traditional wavelet packet.