论文部分内容阅读
在融合了模糊逻辑的推理能力和神经网络的自适应、自学习能力。同时采用输出空间模式聚类的快速学习算法,引入补偿模糊神经元,模糊运算采用动态的全局优化运算,使学习后的网络具有更高的容错性,并弥补了神经网络学习耗时的缺点,提高了效率,进行了故障检测的仿真分析,并且将其运用于未建模系统的故障诊断中取得良好的效果。
In the fusion of fuzzy logical reasoning ability and neural network adaptive, self-learning ability. At the same time, the fast learning algorithm based on output space pattern clustering is introduced to compensate the fuzzy neurons. The fuzzy operation uses the dynamic global optimization algorithm to make the network after learning more fault tolerant and to make up for the shortcoming of neural network learning time-consuming. Improve the efficiency, carry on the simulation analysis of fault detection, and apply it to the failure diagnosis of unmodeled system and achieve good results.