基于Hadoop的变电站设备故障状态识别与预测模型  

Substation Equipment Failure State Identify and Prediction Model Based on Hadoop

作　　者：曲朝阳[1] 刘晓庆[1] 辛鹏[2] 

机构地区：[1]东北电力大学信息工程学院 [2]国家电网吉林省电力有限公司吉林供电公司,吉林省吉林市132012

出　　处：《软件导刊》2015年第3期61-63,共3页Software Guide

摘　　要：在智能变电站环境下,各种智能量测装置运行过程中产生了海量的状态监测数据。针对在数据量巨大的情况下,现有故障诊断方法分析效率缓慢且预测精度不高等问题,提出一种大数据环境下设备故障快速识别与预测模型,改进并实现了MapReduce并行模式下设备故障分类算法,通过专家推理机制,依据规则进行准确的故障预测。建立了一个基于Hadoop集群的数据处理实验环境,以SF6断路器的3种故障状态为对象,分析证明了该模型在不同故障模式下识别与预测的正确性和有效性。In the intelligent substation environment, vast amounts of condition monitoring data is produced during intelligent measurement devices operation process. In order to solve the low efficiency and low prediction accuracy of fault diagnosis methods, the paper proposes an equipment failure quickly identify and prediction model under the large data environment. The paper designs a substation equipment state vast information distributed storage structure based on HBase, improves and implements an equipment failure classification algorithm under the MapReduce parallel mode. The accurate failure prediction can be achieved. A data processing experimental environment based Hadoop cluster is also constructed. The model is proved accurate and effective for the different failure mode under three kinds of SF6 circuit breaker fault status.

关 键 词：智能变电站 故障识别 故障预测 HADOOP HBASE 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]