探索了基于抽取泛化应用相类似的样本和避免对矩阵求逆的改进LMBP算法,取得了良好的学习训练效果,将改进的LMBP算法应用于柴油发动机燃油系统的故障诊断,在LabVIEW平台通过调用Matlab算法节点,高效直观地实现了燃油系统典型故障诊断。
This paper explored the method of extracting samples with high generalization level and the improved algorithm to avoid calculating the inverse of matrix. The new algorithm is applied to fault diagnosis of diesel engine fuel system in the platform of LABVIEW through using the Matlab script node. Experimental results indicate that the diagrtosis network based on new algorithm has preferable diagnosis efficiency and satisfactory accuracy.