基于广域信息分析的智能配电网故障自愈技术研究  

Research on fault self⁃healing technology of intelligent distribution network based on wide area information analysis

作　　者：郭杉[1] 贾俊青[1] 思勤[1] GUO Shan;JIA Junqing;SI Qin(Inner Mongolia Electric Power Research Institute,Hohhot 010020,China)

机构地区：[1]内蒙古电力科学研究院,内蒙古呼和浩特010020

出　　处：《电子设计工程》2024年第9期119-123,共5页Electronic Design Engineering

基　　金：内蒙古电力(集团)有限责任公司2021年科技项目(2021-52)。

摘　　要：针对传统电网接地故障检测和自愈方法存在识别速度慢且精确度较低的问题,文中结合信号广域信息提出了一种配电网故障自愈方法。该方法构建了单相接地故障模型,并将零序电流作为分析的对象。将零序电流信号分解为时域、频域以及小波域分量来作为广域特征向量,并采用随机森林算法对其权重进行分类,利用LightGBM算法对分类后的广域特征向量加以训练,进而得到故障预测结果。在仿真测试中,所提算法能够抵抗过渡电阻与初始相位角变化对预测结果的影响,且其故障分析准确率的平均值为98.9%,在对比算法中为最优,表明该算法可为配电网故障自愈提供有效的技术支撑。In order to solve the problems of slow identification speed and low accuracy of traditional methods for detecting and self-healing of power network grounding fault,this paper proposes a method for fault self-healing of distribution network based on signal wide area information.The single-phase ground fault model is established,and the zero sequence current is determined as the analysis object.The zero sequence current signal is decomposed into time-domain,frequency-domain and wavelet domain components as the wide area feature vector.The random forest algorithm is used to classify the weights.The LightGBM algorithm is used to train the classified wide area feature vector to obtain the fault prediction results.In the simulation test,the proposed algorithm can resist the influence of transition resistance and initial phase angle change on the results,and the fault analysis accuracy is 98.9%,which is the best in the comparison algorithm,which fully proves that the algorithm can provide effective technical support for fault self-healing of distribution network.

关 键 词：广域信息 电网故障自愈 小波分析 随机森林 轻量级梯度提升机 

分 类 号：TN-9[电子电信]