柴谦益+郑文斌+潘捷凯+陆绍彬+温积群
摘 要: 针对智能配电网传统保护方法整定复杂、容错和适应性差等问题,基于大数据分析,提出并设计一套智能配电网状态监测及故障处理方法。该方法先后经过数据预处理、数据融合、数据分析与可视化以及状态辨识与处理共4个环节,将多电气特征量融合转变成为单个综合特征量,监测配电网运行状态,并能根据各个节点关联情况和局部异常因子大小实现对智能配电网故障区域的判定和定位。经RTDS半实物闭环测试,故障判定和定位准确性及可靠性较高,具有一定的参考价值。
关键词: 智能配电网; 保护方法; 大数据分析; 状态监测; 故障定位; 故障处理
中图分类号: TN915?34; TP393 文献标识码: A 文章编号: 1004?373X(2018)04?0105?04
Abstract: In allusion to the problems of complex setting, poor fault tolerance and poor adaptability of the traditional smart distribution network protection methods, a condition monitoring and fault processing method of smart distribution network is proposed and designed based on big data analysis. In this method, four procedures of data preprocessing, data fusion, data analysis and visualization, and condition identification and processing are performed to make multi?electrical characteristic quantities fused and transformed into single comprehensive characteristic quantity. The operation state of distribution network is monitored, and the judgment and location of fault area in smart distribution network are realized according to the correlation condition of nodes and the values of local outlier factors. The RTDS semi?object closed?loop test was carried out. The results show that the method has high accuracy and reliability of fault judgment and location, which also has a certain reference value.
Keywords: smart distribution network; protection method; big data analysis; condition monitoring; fault location; fault processing
智能配电网中引入新能源类DG容易在局部区域产生双向不定潮流,不仅使网络结构复杂化,改变配电网的众多故障特性,还给保护控制设计提出了一定的挑战[1?2]。而大多数保护方法的故障判据均使用小样本的单一电气特征量,整定所需计算较为复杂,且在运行环境变化时需要重新进行整定,保护可靠性方面也存在着一定的风险。在通信或传感器异常时,也容易产生误动或拒动保护[3?6]。
在智能配电网中加入大量传感设备,采集、上传并分析运行数据,形成智能配电网的大数据。目前已被用于用电行为分析、负荷预测等应用中,潜力较大且也为配电网保护方法提供了一个新的思路[7?10]。因此,本文基于大数据分析设计了一种智能配电网状态监测及故障处理方法。……