摘要
人工神经网络以其独特的优点在故障诊断中的应用越来越受到关注。提出了以改进BP神经网络为基础的多征兆域输入、分层诊断的故障诊断模型,该模型具有网络结构简单,收敛速度快,便于管理等优点。经实验和现场数据验证,该模型不仅对单故障具有诊断能力,而且对于并发故障也有较强的分类功能。
Artificial neural network （ANN） receives many researchers＇ interests because of its unique merits. This paper takes BP improved network as base of diagnosis network and divides diagnosis progress into two steps. This will simplify network structure and accelerate calculation. The reliability and precision of diagnosis with this model is proved by experimentation.
出处
《汽轮机技术》
北大核心
2008年第6期451-454,共4页
Turbine Technology
关键词
神经网络
故障诊断
振动特征
汽轮发电机组
neural network
fault diagnosis
vibration character
turbine generator unit
