| 基于等距K-means和apriori算法的配电网故障规律挖掘方法 |
| |
| 引用本文： | 方鑫,,殷俊,蒋苏,陈健,崔晋利,张龙,戴欣,陈锦铭.基于等距K-means和apriori算法的配电网故障规律挖掘方法[J].陕西电力,2020,0(10):99-104,125. |
| |
| 作者姓名： | 方鑫 殷俊 蒋苏 陈健 崔晋利 张龙 戴欣 陈锦铭 |
| |
| 作者单位： | 1. 东南大学，江苏 南京 211189；2.国网江苏电科院，江苏 南京 210036； 3.国网扬州供电分公司，江苏 扬州 225000； 4. 国网江苏省检修分公司，江苏 南京 211102；5.国网淮安供电分公司，江苏 淮安 223001 |
| |
| 摘 要： | 为了提升配电网巡检及运维效率，分析配电网线路故障的特征规律，挖掘其潜在关联因素，提出了一种基于等距K-means和apriori算法的配电网故障规律挖掘方法。该方法基于支持度、置信度和提升度框架，完成强关联规则筛选，适用于构建各类故障场景的故障规律挖掘模型。相比于传统apriori算法，该方法可处理配电网线路故障数据样本中的连续值类型属性，属性分类计算过程收敛性更强，分类结果更趋于客观性。最后以某电力公司故障数据为例，验证了该算法可获得具有较好解释性与实用性的故障巡检规则，有效缩减了配电网线路巡检范围。
|
| 关 键 词： | 配电网故障 等距k-means apriori 频繁项集 关联规则 |
Data Mining Algorithm for Fault Rules of Distribution Network Based on Combination of Isometric k-means and apriori Algorithm |
| |
| FANG Xin,' target="_blank" rel="external">,YIN Jun,JIANG Su,CHEN Jian,CUI Jinli,ZHANG Long,DAI Xin,CHEN Jinming.Data Mining Algorithm for Fault Rules of Distribution Network Based on Combination of Isometric k-means and apriori Algorithm[J].Shanxi Electric Power,2020,0(10):99-104,125. |
| |
| Authors: | FANG Xin " target="_blank">' target="_blank" rel="external"> YIN Jun JIANG Su CHEN Jian CUI Jinli ZHANG Long DAI Xin CHEN Jinming |
| |
| Affiliation: | 1.Southeast University, Nanjing 211189, China； 2. State Grid Jiangsu Electric Power Research Institute，
Nanjing 210036， China； 3. State Grid Yangzhou Power Supply Company，Yangzhou 225000， China；
4. State Grid Jiangsu Electric Power Maintenance Company，Nanjing 211102，China；
5. State Grid Huaian Power Supply Company，Huaian 223001，China |
| |
| Abstract: | In order to improve the efficiency of inspection tour and operation and maintenance in distribution network, and analyzes the fault feature and associated factors， this paper proposes an algorithm based on the combination of isometric K-means and apriori for mining and analyzing the fault rules of the distribution network. The strong-association rules mining is realized using the algorithm with support degree， confidence level and lift， and the fault rule mining model under various fault scenarios is established using the algorithm. Compared with the apriori algorithm, this algorithm can determine the attributes of continuous value type in the fault data samples in the distribution network and the classification results tend to be more objective. Finally, the application of the algorithm to actual fault data from a power company shows that the rules derived from the algorithm have remarkable effects in reducing the range of distribution network inspection. |
| |
| Keywords: | distribution network fault isometric K-means apriori frequent itemset association rules |
|
| 点击此处可从《陕西电力》浏览原始摘要信息 |
|
点击此处可从《陕西电力》下载全文 |
