摘要
为了通过放电特性来判断大电机定子绝缘状态，对人工神经元网络组识别放电类型和发展程度的能力进行了研究。在屏蔽试验室内，用定子线棒工业仿真模型取得了不同放电模式的大批试验数据，并用三维谱图对放电信息进行有效压缩，以谱图表列数据为特征量构成放电样本。以类型识别主网络和程度识别子网络组成人工神经元网络组，经训练的网络对放电类型和发展程度的识别结果是令人满意的。
WT9.,8.75BZ] For the purpose of judging insulation condition of generator stator on the basis of discharge characteristics, the ability of artificial neural network group, which is used to recognize the different types and serious levels of discharges, is studied. Using industrial simulation models of stator winding, a lot of experimental data of different discharge patterns are obtained in shielding room. The discharge information is effectively suppressed through 3 dimensional pattern and the tabulated data of these patterns are used as characteristic vectors. An artificial neural network group consisting of a main network for type recognition and some sub networks for serious level recognition is established. The trained network group is used to discriminate the types and serious levels of discharge, and the recognition results are satisfied.
出处
《电网技术》
EI
CSCD
北大核心
1999年第10期29-32,共4页
Power System Technology
基金
国家自然科学基金
关键词
发电机
定子模型
线棒
局部放电
模式识别
winding of electrical machine
partial discharge
on-line monitoring
pattern recognition
artificial neural network
