基于遗传算法优化BP神经网络的GIS设备放电故障诊断

Discharge Fault Diagnosis of GIS Equipment of Optimum BP Neural Network Based on Genetic Algorithm

为有效GIS设备放电故障诊断的快速性和准确性,采用近几年出现的遗传算法对BP神经网络进行优化,减少了BP神经网络算法陷入局部最优解的风险,显著增强了BP神经网络的泛化能力和全局寻优能力。对比发现,遗传算法优化后的BP神经网络模型具有比较好的快速性和准确的诊断能力。测试结果表明,遗传算法优化BP神经网络对GIS设备放电故障诊断具有可行性和有效性。

For the effective discharge of GIS equipment for fault diagnosis of speed and accuracy,using newly developed genetic algorithm for BP neural network optimized BP neural network algorithm is reduced local optimal solutions risks significantly enhances the generalization ability of BP neural network and a global optimization. The comparison,genetic algorithm optimized BP neural network model compared well with the rapid and accurate diagnostic capability. Test results show that the discharge of GIS equipment based on genetic algorithm optimized BP neural network fault diagnosis is feasible and effective.