基于模糊遗传神经网络的信息融合故障诊断技术及其应用

Fault diagnosis technology and its application based on fuzzy genetic neural network multisensor data fusion

在基于广义模糊加权型推理的模糊神经网络基础上,融合非一致性遗传算法,建立了一种模糊遗传神经网络。利用模糊遗传神经网络技术建立信息融合中心,对多传感器数据进行融合处理,通过多源互补信息减小故障诊断系统的不确定性。讨论了模糊遗传神经网络多传感器信息融合方法中数据处理、特征向量维数压缩与关联、归一化处理方法等。同时,对模糊遗传神经的构造以及学习训练等内容,也作了较为详细的讨论。并对模糊遗传神经网络信息融合技术应用于变压器状态的实时监测的应用前景进行了初步探讨。

Fuzzy genetic neural network is developed based on fuzzy neural network with improving fuzzy weighted reasoning method,syncretizing the inconsistent self-adaptive genetic algorithm. A muhisensor system is dealed for applying to mechanical fault diagnosis.By using the information fusion pattern based on fuzzy genetic neural network,the uncertainty of the diagnosis system can be decreased greatly.The methods for data preproceasing, features extraction and construction of the fuzzy genetic neural network are discussed.And then the application of fuzzy genetic neural network information fusion technology to transformer condition real time monitoring is also given.