基于智能算法的配网单相接地故障辨识

Intelligent Algorithm-based Identification of Single-phase Earth Faults in the Distribution Network

单相接地故障是配电网中最常见的故障,由于故障电流可能很小,其精确诊断存在较大难度。分析了可能对单相接地故障辨识产生影响的其他短路故障与扰动情况,提出了一种基于改进支持向量机(SVM)与卷积神经网络(CNN)的配网单相接地故障辨识方法。利用粒子群算法对支持向量机的参数进行改进,识别出低阻接地故障与扰动;针对时域中不易区别的永久性与间歇性单相接地故障,通过希尔伯特-黄变换提取频域信息,利用卷积神经网络根据提取出的高维特征向量辨识其具体类型。在MATLAB/Simulink中搭建了一个辐射状10 kV配电网模型进行仿真,仿真结果验证了算法的有效性和优越性。

Single-phase earth fault is the most common fault in the distribution network.As the fault current might be very small,its accurate diagnosis remains quite difficult.In this paper,other short-circuit faults and disturbances which might influence single-phase earth fault identification were analyzed,a single-phase earth fault identification method was proposed on the basis of the improved support vector machine(SVM)and the convolutional neural network(CNN),and the particle warm optimization algorithm was used to improve SVM parameters to identify low-resistance earth faults and disturbances.With respect to permanent and intermittent single-phase earth faults,not easy to distinguish in the time domain,frequency-domain information was extracted through Hilbert-Huang transform,and specific types of these faults were identified in the CNN approach,based on the extracted high-dimensional characteristic vectors.A 10 kV radial distribution network model was built in MATLAB/SIMULINK.The simulation results verified the effectiveness and superiority of the algorithm presented in this paper.