基于BP神经网络的柴油机振动故障诊断

Fault Diagnosis of Diesel Engine Based on Optimization BP Neural Network

针对柴油机振动故障诊断问题,通过采集缸盖上的振动信号,运用小波分解的方法采集特征向量并对数据进行消除噪声。通过小波分解后,减少了振动信号的维数,体现了故障信号的特征,并且不会造成振动信号中故障特征向量的缺失,而且过滤掉了振动信号中的噪声,提高了故障诊断的准确性和故障诊断的速度,把从振动信号中提取出的特征向量带入优化后的BP神经网络中进行训练,在对比中原始的BP神经网络和优化后的BP神经网络,发现优化后的识别率更高。

Aiming at the problem of diesel engine vibration fault diagnosis, the vibration signal of cylinder head is collected, and the eigenvector is collected by wavelet decomposition, and the noise is eliminated from the data. After wavelet decomposition, the dimension of vibration signal is reduced, the feature of fault signal is embodied, and the defect of fault feature vectors in vibration signal is not caused, and the noise in vibration signal is filtered out, the accuracy of fault diagnosis and the speed of fault diagnosis are improved. The feature extracted from vibration signal is extracted. Vectors are brought into the optimized BP neural network for rate training. Compared with the original BP neural network and the optimized BP neural network, the optimized BP neural network has a higher recognition.