ANN with the Error Contracting Gradually Algorithm and Its Application in Generator Fault Diagnosis

摘要：

Based on analysis of conventional back-propagation (BP) network, the causes of error curve oscillation and excessive learning are first proposed. Next, a new BP with the error contracting gradually algorithm is put forward, through setting up neuron error threshold function, only when neurons error is bigger than the error threshold, the neurons parameters can be adjusted, otherwise the neuron which error is smaller than the error threshold cant be adjusted. The proposed algorithm can avoid the excessive learning and learning error oscillation. Finally, two fault diagnosis models based on the new BP algorithm is set up respectively, which are turbine-generator set vibration fault diagnosis model and rotor winding inter-turn short circuit fault diagnosis model. The results of verification show that the model has faster speed and higher diagnosis precision.

关键词： Artificial neural network The error contracting gradually algorithm Turbine-generator set vibration fault Rotor winding inter-turn short circuit fault Fault diagnosis

作者: Shuting Wan Yonggang Li

作者单位: College of Mechanical Engineering, North China Electric Power University, Baoding 071003, P. R. China;Key Laboratory of Condition Monitoring and Control for Power Plant Equipment of Ministry of Education , Baoding 071003, P. R. China

会议类型: 国际会议

会议名称: The 2007 International Conference on Intelligent Systems and Knowledge Engineering(第二届智能系统与知识工程国际会议)

会议地点: 成都

会议语种:英文

页码: 1162-1168

在线出版日期: 2007-10-15（万方平台首次上网日期，不代表论文的发表时间）