基于概率神经网络的配电网故障区段定位建模

Modeling of fault section location in distribution network based on probabilistic neural network

配电网故障区段定位方法容易受到冗余数据的影响,导致定位精准度下降,为解决这一问题,提出了基于概率神经网络的配电网故障区段定位建模方法。分析概率神经网络组成结构,设计配电网故障区段定位训练过程,以此优化网络模型,剔除容易受到噪声干扰的冗余数据。通过对故障暂态零序电流序列的分析,以馈线为例采用概率神经网络方法求解配电网故障区间位置的非线性整数规划模型,并结合连续空间上的互补非线性约束条件,构造连续域内配电网故障区段位置互补约束规划模型,完成故障区段定位。由实验结果可知,该模型在故障信号段内,波动范围从0.8下降到0.2,与实际故障信号波动范围一致,定位精准度较高。

The distribution network fault section location method is easy to be affected by redundant data,resulting in the decline of location accuracy.In order to solve this problem,a distribution network fault section location modeling method based on probabilistic neural network is proposed.The structure of probabilistic neural network is analyzed,and the training process of fault section location in distribution network is designed,so as to optimize the network model and eliminate the redundant data vulnerable to noise interference.Through the analysis of fault transient zero sequence current sequence,taking feeder as an example,the probabilistic neural network method is used to solve the nonlinear integer programming model of distribution network fault section location,and combined with the complementary nonlinear constraints in continuous space,the complementary constraint programming model of distribution network fault section location in continuous domain is constructed to complete the fault section location.The experimental results show that the fluctuation range of the model in the fault signal section decreases from0.8 to 0.2,which is consistent with the fluctuation range of the actual fault signal,and the positioning accuracy is high.