摘要
应用频率响应分析法(FRA)测试变压器绕组变形,判别时没有量化的指标体系,而仅仅依靠测试人员的经验。本文引入BP神经网络算法,通过对FRA的各种数据实例进行学习训练,实现变压器绕组是否发生变形的自动诊断。
There is no quantitative indicator to use Frequency Response Analysis (FRA) to determine whether the winding of power transformers have deformation. The only indicator is the experience of testers. This paper introduces the BP neural network algorithm, in terms of the learning and training of dozens of data instances of FRA, an automatic diagnosis on the power transformers winding deformation is presented.
出处
《云南电力技术》
2016年第1期102-106,共5页
Yunnan Electric Power