摘要
提出了一种基于概率神经网络的变压器故障诊断方法,并用遗传算法优化概率神经网络的平滑因子,从而提高概率神经网络诊断的正确率。结果表明,实际的油色谱数据验证了此变压器故障诊断方法的可行性和正确性。
A new method based on Probabilistic Neural Networks （PNN） to transformer fault diagnosis is presented in this paper. And the parameters of PNN is determined by genetic optimization to increase the diagnosis accuracy of PNN. The diagnosis accuracy and feasibility of PNN are certified by the practicable chromatography date.
出处
《继电器》
CSCD
北大核心
2006年第3期9-11,共3页
Relay
