基于蚁群算法的神经网络配电网故障选线方法

为了克服基于神经网络的故障选线方法收敛速度慢、易于陷入局部极小点的缺点,提出了蚁群算法和神经网络相结合的故障选线方法。利用ATP-EMTP做单相接地仿真试验,得到各线路的零序电流信号,通过小波变换和傅里叶变换提取其中的故障特征作为神经网络的输入。利用蚁群算法对神经网络进行训练,完成训练的神经网络模型即可实现故障选线。仿真结果表明,该方法训练速度快、误判率低。

Ant colony algorithm and neural network based fault line detection method for distribution network

To overcome the shortcomings of the slow convergent speed and easy convergence to the local minimum points in the neural network based fault line detection method, fault line detection method by combining ant colony algorithm with neural network is presented. The zero sequence current of every line is obtained in single-phase-to-earth fault experiment by using the ATP-EMTP simulation and the fault features are extracted from zero sequence current through wavelet transform and Fourier transform and are used as inputs of neural network. After the neural network is trained using ant colony algorithm, the neural network model trained can realize fault line detection. The simulation results show that the method reaches higher training speed and lower error rate.