BP神经网络在大型发电机主绝缘击穿电压预测中的初步应用  

Preliminary application of BP neural network in predicting breakdown voltage of large generator ground wall insulation

作　　者：刘丽兵[1] 高乃奎[1] 马小芹[1] 谢恒堃[1] 

机构地区：[1]西安交通大学电力设备电气绝缘国家重点实验室,陕西西安710049

出　　处：《电工电能新技术》2004年第2期73-76,共4页Advanced Technology of Electrical Engineering and Energy

基　　金：国家自然科学基金重点资助项目 (5 98372 6 0 )

摘　　要：击穿电压在大电机主绝缘寿命评估中有着重要的作用。本文讨论了BP神经网络在谏壁 7号机更换线棒主绝缘击穿电压预测中的应用。首先作者用所测得的试样介质损耗参量、局部放电参量和击穿电压值作为样本提供给BP神经网络学习。然后 ,用训练好的神经网络对击穿电压进行预测。结果表明训练好的BP神经网络对该批更换线棒击穿电压的预测是可行的 ,并有较高的准确度。The breakdown voltage plays an important role in predicting remaining life of the large generator ground wall insulation. In this paper, we discussed a BP neural network that was used to predict the breakdown voltage of the large generator ground wall insulation of rewind bars of the Jianbi Power Plant No.7 generator. At first the neural network has been trained by the samples that include the parameters of dielectric loss factor tan δ , the parameters of partial discharge (PD) and breakdown voltage. We used the parameters of dielectric loss factor and partial discharge (PD) as the inputs of the BP neural network and breakdown voltage as the output. Then we tried to predict the breakdown voltage by the neural network that has been trained already. We found that it's feasible and much accurate of neural network to predict breakdown voltage. This method can be applied to predict breakdown voltage of other generators ground wall insulation.

关 键 词：神经网络 介质损耗 局部放电 击穿电压 

分 类 号：TM307[电气工程—电机]