基于故障层次模型的集成神经网络诊断

Diagnosis Application of Integrated ANN Based on Hierarchical Fault Model

复杂系统具有层次结构特点 ,基于故障层次模型的诊断策略是复杂系统故障诊断可行和有效方法。集成神经网络模型以故障层次模型为参考 ,可以大大缩小诊断推理的求解空间 ,最终快速定位发生故障的根本部位。集成神经网络的模型结构为多模型或方法集成应用提供了一个良好的融合机制。基于模糊ART神经网络及BP神经网络构建的集成神经网络更进一步具有发现新故障的能力 ,且能够对故障引发部位实现精确诊断和定位。

The configuration of complex manufacturing system has hierarchical characteristic. The diagnostic method based on hierarchical fault model is effective to fault diagnosis of complex manufacturing system. Refer to hierarchical fault model, the integrated ANN diagnostic model can contract the scope of diagnostic reasoning, and find quickly the fault components. The structure of integrated ANN model is very convenient to integrate various diagnostic models and methods in the diagnosis system. The integrated ANN consists of fuzzy ART2 and BP network has ability of finding new fault type, accurate diagnostic reasoning and locating fault.