基于模糊加权有色网和BP神经网络的飞机发动机故障诊断

针对传统知识推理故障诊断方法中参数往往依赖于专家经验,存在着不准确和无法学习的问题,提出了一种基于模糊加权有色网和BP神经网络的故障诊断方法。首先,定义了模糊加权有色网并给出了两种产生式规则对应的模糊加权有色网模型。然后,设计了采用BP神经网络对模糊加权有色网各参数进行学习的算法。最后,给出了使用训练后的各参数进行初始化的模糊加权有色网进行故障推理的具体步骤。通过飞机发动机故障诊断实例仿真实验证明了方法能正确地建立模糊加权有色网推理模型,在采用BP神经网络进行参数训练后,能有效地实现飞机发动机的故障诊断。

The Fault Diagnosis for Aircraft Generator Based on Fuzzy Colored Petri Net and BP Neural

Aiming at the parameters of the traditional knowledge fault diagnosis reasoning method usually relying on the experience of specialist, having the problem of inaccuracy and unable to study, the new fault diagnosis method based on fuzzy colored petri net and BP neural network is proposed. Firstly, the fuzzy colored petri net and the net according to the two kinds of generating rules are defined. Then the algorism for training the parameters of fuzzy colored petri net is designed. Finally, the specified strategy for fault diagnosis reasoning using the fuzzy colored petri net is given. The aircraft generator fault diagnosis experiment shows the method in this paper can build the fuzzy colored petri net model correctly after the training by BP neural network, and it is a effective method for diagnosing the fault of aircraft generator.