讨论了反向传播算法的人工神经网络在局部放电模型类型识别中的应用 ,设计了反映大型发电机定子绝缘局部放电的 5类不同物理仿真绝缘模型。仿真模型包括的局部放电类型有端部放电、槽间放电、在绝缘不同位置的三种内部放电。研究了人工神经网络的识别能力。针对大型发电机定子绝缘局部放电的类型 ,得到了较好的识别率 The application of BP neural network in type identification of partial discharge model is discussed. Five types of different physical simulation insulation models reflecting the partial discharge of stator insulation of large generator are designed. The simulation models include partial discharges with end discharges, inter-tank discharges, and three internal discharges at different locations on the insulation. The recognition ability of artificial neural network is studied. Aiming at the type of partial discharge of large generator stator insulation, a better recognition rate is obtained