基于AGA-RBF算法的柴油机故障诊断研究

Diesel Engine Fault Diagnosis Research Based on AGA-RBF Algorithm

针对在柴油机故障诊断中径向基函数(Radial Basis Function,RBF)神经网络泛化能力不足的问题,提出一种基于AGA-RBF算法的柴油机故障诊断方法。在该方法中将自适应遗传算法(Adaptive Genetic Algorithm,AGA)和RBF神经网络有机地结合起来,利用自适应遗传算法对RBF神经网络的基函数宽度和中心进行优化,将优化后的RBF神经网络应用于柴油机故障诊断。通过实验仿真表明,该算法收敛速度快,改善了RBF神经网络的泛化能力,提高了故障诊断准确率,实用性强,易于工程实现。

In viewof the problem which the generalization ability of radial basis function（RBF） neural network is insufficient in the fault diagnosis of diesel engine,a AGA-RBF algorithm based on fault diagnosis method of diesel engine is proposed. In the method,the RBF neural network and genetic algorithm are organically combined. Adaptive genetic algorithm is applied to optimize the width and center of the basis function;and the optimized RBF neural network is used to diagnose the diesel engine fault. Through the simulation analysis,we can draw,that the algorithm converges fast,improves the generalization ability of RBF neural network,improves the accuracy of fault diagnosis,has strong practicality,easy to realize on the project.