引用本文: | 叶万余,苏 超,罗敏辉,等.基于关联规则挖掘的输电线路缺陷状态预测[J].电力系统保护与控制,2021,49(20):104-111.[点击复制] |
YE Wanyu,SU Chao,LUO Minhui,et al.Transmission line defect state prediction based on association rule mining[J].Power System Protection and Control,2021,49(20):104-111[点击复制] |
|
摘要: |
对输电线路缺陷状态进行关联因素的分析和预测工作,可以为输电线路的巡维工作提供重要的技术支持。在现有输电线路状态分析和影响因素研究的基础上,提出了基于关联规则挖掘的输电线路缺陷状态预测方法。首先根据历史缺陷数据评价得到输电线路缺陷状态。结合各种影响因素,构建线路缺陷状态与相关因素的特征库。然后引入FP-Growth算法挖掘各因素与缺陷状态间的关联规则,并将得到的规则用于预测线路的缺陷状态。最后以某地区架空输电线路为例,通过历史缺陷等数据评价得到缺陷状态样本,提取相关条件特征作为输入特征,并用于预测线路的缺陷状态。结果验证了该方法的有效性,对输电线路的巡维检修有一定的参考价值。 |
关键词: 输电线路 关联规则 缺陷状态 支持向量机 分类预测 |
DOI:DOI: 10.19783/j.cnki.pspc.210023 |
投稿时间:2021-01-08修订日期:2021-05-07 |
基金项目:国家自然科学基金项目资助(61773126);广东电网有限责任公司科技项目资助(031800KK52180074) |
|
Transmission line defect state prediction based on association rule mining |
YE Wanyu,SU Chao,LUO Minhui,ZENG Yongbin,YAN Zuming,LIU Yi |
(1. Qingyuan Power Supply Bureau, Guangdong Power Grid Co., Ltd., Qingyuan 511500, China;
2. School of Automation, Guangdong University of Technology, Guangzhou 510006, China) |
Abstract: |
The analysis and prediction of the factors of a transmission line defect state can provide important technical support for transmission line patrolling. Based on the analysis of a transmission line state and the study of influencing factors, a method for the state prediction based on association rule mining is proposed. First, the transmission line defect state is evaluated based on historical defect data. Combined with various influencing factors, the characteristic database of the line defect state and related factors is constructed. Then a FP-Growth algorithm is introduced to mine the association rules between each factor and defect state, and the obtained rules are used to predict the defect state of the line. Finally, taking an overhead transmission line in a region as an example, defect state samples are obtained through historical defect data evaluation, and the relevant condition features are extracted as input features, and used to predict the defect state of the line. The results verify the effectiveness of the method presented, and this has a reference value for transmission line patrolling and maintenance.
This work is supported by the National Natural Science Foundation of China (No. 61773126) and the Science and Technology Project of Guangdong Power Grid Co., Ltd. (No. 031800KK52180074). |
Key words: transmission lines association rules defect state support vector machine classification prediction |