摘要
本文讨论了水轮发电机组故障征兆的特征选择与提取。基于信息融合技术的思想,从设备故障诊断的实际出发,建立了基于信息融合技术的神经网络证据融合故障诊断系统,即通过故障特征信息的有效组合,用各种子神经网络从不同侧面对设备故障进行初步诊断,然后对诊断结果应用DempsterShafer证据理论进行决策融合。诊断实例表明,经过多故障特征信息融合,诊断结论的可信度明显提高,可以有效提高确诊率。
The acquistion of fault symptoms of hydropower generating unit is presented in this paper.The neural network fault diagnosis system is set up based on both information fusion technology and actual fault diagnosis.The sub neural network is considered as primary diagnosis from different sides,then based on Dempster Shafer evidence theory, the decision level fusion is considered as the diagnosis conclusions.It indicates that by diversified characteristic information the reliability of the diagnostic result and the diagnosis rate are improved evidently.
出处
《水力发电学报》
EI
CSCD
北大核心
2004年第6期111-115,共5页
Journal of Hydroelectric Engineering