基于神经网络和D-S证据的电厂凝汽器故障诊断研究
Research of Fault Diagnosis Based on Neural Networks and D-S Evidence for Condenser in Power Plant
李平 1黄国樑 2彭道刚 3夏飞3
作者信息
- 1. 上海电力学院自动化工程学院,上海200090
- 2. 国网上海市电力公司培训中心,上海200438
- 3. 上海电力学院自动化工程学院,上海200090;上海市电站自动化技术重点实验室,上海200090
- 折叠
摘要
火电厂凝汽器是汽轮发电机组的重要辅机之一,其工作状况对整个电厂安全和经济运行都有着决定性的影响.结合信息融合思想,提出一种基于神经网络和D-S证据理论的电厂凝汽器故障综合诊断方法,首先通过BP神经网络和CPN神经网络得到各自的诊断结果作为决策层D-S证据理论的初始证据,再利用证据理论对这些结果进行融合,得到最终的故障诊断结果.通过实例数据诊断结果表明:与单一神经网络诊断结果相比,该方法减少了误差,提高了诊断可信度.
Abstract
Condenser is one of the most important auxiliaries of turbine generator unit in power plant,whose operating condition decisively affects the safe and economically operation of the whole plant.Combination the thought of information fusion,the method of condenser integrated fault diagnosis based on neural networks and D-S evidence theory is proposed.The respective diagnosis results regarded as the D-S evidence theory primary evidences in decision layer according to BP neural network and CPN network are obtained first,and then these results are fused by using of the D-S evidence theory to obtain the final diagnosis result.The diagnosis results show that this method has a smaller error and higher diagnosis reliability comparing with the results from the single neural network.
关键词
信息融合/神经网络/D-S证据理论/故障诊断/凝汽器Key words
Information fusion/Neural networks/D-S evidence theory/Fault diagnosis/Condenser引用本文复制引用
基金项目
上海市教育委员会科研创新重点项目(12ZZ177)
上海市电站自动化技术重点实验室开放课题(13DZ2273800)
出版年
2014
NETL
NSTL
维普
万方数据