基于EMD、遗传算法及神经网络的柴油机故障诊断

研究了将经验模式分解(Empirical Mode Decom position,EMD)、遗传算法及BP神经网络相结合对柴油机振动信号进行故障诊断的方法。首先运用经验模式分解方法对柴油机缸盖表面振动信号进行分解并提取特征参数;然后利用遗传算法对得到的特征参数进行选择,找到对于故障诊断最为敏感的参数;最后建立了BP神经网络模型对柴油机典型故障进行诊断。通过对某型柴油机的验证,表明该方法能够准确识别柴油机供油系统的典型故障。

Diesel Engine Fault Diagnosis Based on EMD,GA and BP Neural Network

The diesel engine fault diagnosis combined with Empirical Mode Decomposition (EMD), Genetic Algorithms (GA) and BP neural network analysis of vibration signal are explored. First EMD process is used to decompose vibration signal from the face of cylinder head and extract feature parameters from them. Then GA is adopted to select the parameters which are identified easily for diagnosis. At last BP neural network model and diagnosed typical fault of the diesel engine are established. The example proves this m...