摘要
近年来研究和开发核电站故障诊断系统成为国际核能界的重要课题。本文针对核电站的实际情况 ,以人工神经网络为基础诊断方法 ,提出超级神经元的概念和分层组网的策略 ,改善了知识采集的条理性和可操作性 。
Recently it has become an important task to study and develop the fault diagnosis systems for the nuclear power plants in the world's nuclear energy area. Based on artificial neural networks as essential diagnosis method, this paper brings forward the concept of supper neuron and the policy of using hierarchical form to build networks according to practice. It is propitious to resolve some difficulties in building diagnostic networks for nuclear power unit using this method. It improves consecution and maneuverability in knowledge gathering. The heavy works to train the supper ANN are divided into smaller tasks and the changes of the whole networks caused by partial knowledge altering are reduced consumedly.
出处
《上海海运学院学报》
北大核心
2001年第3期268-271,274,共5页
Journal of Shanghai Maritime University
关键词
核电站
故障诊断
人工神经网络
超组神经元
实现方法
nuclear power plant, fault diagnosis, artificial neural networks, supper neuron
