摘要
论证了单症兆和多症兆诊断的模糊模型分别与一定条件的单层神经网络等价，从而建立了单症兆和多症兆诊断的模糊神经网络模型。基于模糊诊断原理，阐述了模糊神经网络模型是由若干独立单元组成的可扩充的组合式结构，进而提出一种修改或扩充子网络或子结点及其权重连线的可塑性学习方法。基于此模型建立了旋转机械故障诊断专家系统，以现场实际例子对模型的应用进行了说明。
Its proved that on some spacific conditions the model of single sympton diagnosis is equal to the model of neural network. Thus, a fuzzy neural network model is constructed,so is the model of multi symptom.According to fuzzy diagnosis principles,the paper demonstrates that the neural network model is builtup by a group of units ,and therefore a learning method of plasticity by adding or modifying related unit is proposed.Based on this neural network model,an expert system of fault diagnosis for large rotating machinery is programmed,and the use of the model is demonstrated by an example from production plant.
出处
《振动工程学报》
EI
CSCD
1997年第4期496-500,共5页
Journal of Vibration Engineering
基金
河南省自然科学基金
关键词
故障诊断
神经网络
专家系统
模糊诊断
fault diagnosis
neural network
expert system
fuzzy fault diagnosis
