摘要
现代设备的维护需要对设备所出现的故障能够做出准确而且迅速的判断 ,为满足这些要求 ,单纯地依靠传统的故障诊断专家系统是不够的 ,把人工智能中的神经网络知识应用到专家系统中去 ,建立神经网络专家系统 ;同时 ,把神经网络同规则库联合起来 ,又可达到优势互补的效果。在神经网络诊断中 ,采用牛顿算法和反向传播算法的联合对BP算法进行了一些改造 ,使收敛速度得以提高。在某油田注水泵机组故障诊断专家系统中应用了此种原理。
The maintenance of equipment requires accurate and rapid judgement for the arisen fault.To meet the demands, it's not to have the conventional fault expert system.lt's necessary to apply the neural network to the expert system and build Neural Network Expert System . At the same time,to connect neural network with the rule base can reach the best effect. In the neural network diagnosis, using the method of associating BP with the newton arithmetic, the speed of convergence is improved. In the diagnostic expert system of a set affusion pump employs such a principle.
出处
《北京机械工业学院学报》
2001年第1期6-10,共5页
Journal of Beijing Institute of Machinery
基金
国家自然科学基金项目! [项目编号 :5 9775 0 0 2 ]
