基于Hadoop的高压输电线路合闸故障诊断方法  

Research on Diagnosis Technology of Switch-on Fault of High Voltage Transmission Line Based on Hadoop

作　　者：周晓[1] 朱晗雨 ZHOU Xiao;ZHU Hanyu(School of Mechanical and Electronic Engineering,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学机电工程学院,湖北武汉430070

出　　处：《数字制造科学》2022年第2期104-110,共7页

摘　　要：针对高压输电线路合闸故障数据体量大、增长速率快,串行故障诊断算法效率低下的问题,提出了一种基于Hadoop平台下的遗传算法和K-means算法的高压输电线路合闸故障诊断方法。根据录波器采集的数据具有故障时刻前后阶跃性的特点,设计整体及局部相结合的数据特征表达方法构建数据样本集。利用遗传算法的全局寻优特性改进K-means算法初始聚类质心选取的问题,结合Hadoop平台的MapReduce框架特点设计并行化遗传算法和K-means算法。经实验证明所提出的并行化方法能够明显提升算法运行效率,具有较好的工程应用潜力。This paper proposes a high-voltage transmission line closing fault diagnosis method based on genetic algorithm and K-means algorithm under Hadoop platform.According to stepping wave features in prior and after the failure,a characteristic expression method is designed to construct a data sample set.The advantages of the genetic algorithm’s in terms of global optimization characteristics and adaptive search probability have been used to improve the K-means algorithm.Combining the characteristics of the MapReduce framework of the Hadoop platform,parallel genetic algorithms and the improved K-means algorithms have been brought up.Numerical simulation results show that the proposed parallelization method can significantly improve the efficiency of the algorithm and it has good engineering application potential.

关 键 词：HADOOP平台 高压输电线路 故障诊断 K-MEANS算法 遗传算法 

分 类 号：TM75[电气工程—电力系统及自动化] TP277[自动化与计算机技术—检测技术与自动化装置]