基于单振动传感器BP神经网络的柴油机失火诊断

Diesel Misfire Diagnosis Based on Single Vibration Sensor and BP Neural Network

为解决多缸柴油机失火诊断问题,提出基于单振动传感器与BP神经网络的柴油机失火故障诊断方法。首先,通过柴油机缸盖上的单振动传感器获取振动信号,对振动信号进行滤波提取点火频率成分;然后,对振动信号按工作周期进行分段处理,进行等角度重采样及同步平均;最后,提取每缸工作信号的能量与峭度,建立特征向量,将其输入到BP神经网络分类器进行诊断识别。应用结果表明,该方法能有效诊断柴油机失火故障。

In order to solve the problem of multi cylinder diesel misfire diagnosis,the paper proposes the method based on single vibration sensor and BP neural network.It firstly obtains the vibration signals through single vibration sensor on the diesel cylinder head,and extracts the ignition frequency components by filtering the vibration signals.Then,it segments the vibration signals according to the operating cycle,and resamples them in equal angle and averages synchronously.Finally,it constructs feature vector by extracting energy and kurtosis of each cylinder,and inputs it into BP neural network classifier for diagnosis and identification.The application result shows that this method can diagnose diesel misfire failure effectively.