摘要
论文主要介绍采用BP神经网络对变压器油中溶解气体进行分析来进行变压器故障诊断。论文所用BP算法是在传统的最速下降BP算法中加入可变学爿率以及采用动量因子的改进BP网络算法，即将自适应学习率法和动量BP法相结合的学习率可变的动量BP算法，建立了一种具有较强学习能力、泛化能力和适应能力的神经网络模刚，并通过实验证明了其对变压器故障诊断的准确性。
This paper mainly describes that use BP neural network to analyse transformer oil dissolved gas and diagnosis transformer fault. In this paper, the BP algorithm is a improved BP network algorithm which is the traditional steepest descent BP algorithm with variable learning rate and the momentum, estabilish a neural network model with strong learning ability and adaptability, the generalization ability, and experiment shows the accuracy of transformer fault diagnosis.
出处
《电气试验》
2012年第6期11-14,共4页
Electrical Test
关键词
可靠性
县级电网
提高
措施
neural network
transformer
dissolved gas in oil
algorithm
