基于压缩感知的小电流接地系统故障选线方法

Fault line selection method for small current grounding system based on compressed sensing

针对电力系统故障选线中存在故障信号幅值小、噪声干扰大、波形不稳定等造成故障选线不准确的问题,在压缩感知理论框架下提出了基于压缩感知的小电流接地故障选线新方法。首先,利用Simulink搭建了10 kV电压等级的小电流接地模型,在其基础上通过设置故障相,提取各条线路中的零序电流;然后利用压缩感知算法对所提取的零序电流进行压缩,从而得到压缩零序电流特征;接着通过调整线路参数得到大量正常线路以及故障线路中经过压缩的零序电流特征样本;最后利用径向基神经网络构造了小电流接地故障选线模型。实验结果表明所提出的小电流接地故障选线方法不仅具有较高的精度,而且有较强的鲁棒性。

In order to solve the problem of inaccurate line selection due to the small amplitude of the fault signal,large noise interference and unstable waveform in the fault line selection of power system,a fault line selection method for small current grounding system based on compressed sensing is proposed under the framework of compressed sensing theory.Firstly,a small current grounding model of 10kV voltage level is built by Simulink.Zero-sequence current in each line is extracted by setting the fault phase on the basis of the built model.Then the zero-sequence current is compressed by using compressed sensing theory with the zero-sequence current characteristics of the signal compression obtained.Then,a large number of compressed zero-sequence current both in normal lines and in fault lines are obtained by adjusting the line parameters.Finally,a radial basis function neural network is used to construct a fault line selection model for small current grounding system.The experimental results show that the proposed method of fault line selection model for small current grounding system are not only higher in accuracy,but also stronger in robustness.