摘要
建立了一个模糊神经网络,用主成分分析法提取故障发生的特征运行参数,确定所建立的模糊神经网络的输入向量个数,再用动态聚类法对所采集的大样本进行故障分类,确定所建立的模糊神经网络的输出向量个数。根据采集的样本训练出模糊神经网络的连接矩阵,然后对单个的联想记忆网络进行合成,实现故障的诊断。通过具体的实例,给出诊断过程。
A fuzzy-neural network is constructed which extracts characteristics of fault in the work by main component analysis to confirm the amount of input in the fuzzy-neural network. Then dynamic clustering is used to classify faults that we got from practice, so that the amount of output is confirmed. The gathered samples are used to train the connected matrix of the fuzzy-neural network. And every single fuzzy-neural network is assembled to realize fault diagnosis. Finally the diagnosis process is provided with an example.
出处
《重庆科技学院学报(自然科学版)》
CAS
2005年第3期81-85,共5页
Journal of Chongqing University of Science and Technology:Natural Sciences Edition
关键词
模糊神经网络
故障诊断
特征提取
故障分类
fuzzy-neural network
fault diagnosis
characteristics extraction
fault classification