CAJ | 학술논문

文章针对电网运行产生的数据呈爆炸式增长，EMS 系统有效信息往往淹没在海量数据中这一问题，提出一种云计算模式下的聚类分析处理方法，基于 Hadoop 平台的 MapReduce 计算模型与分布式文件存储，将系统聚类法进行拆分，在云环境中对多个计算模块进行并行分析。作为试验性验证，提取某大用户近三年的负荷特征曲线，选取不同数据量、不同节点数，进行算法加速比的测试。结果表明，在云计算架构中该算法可以有效提高计算效率，适用于电力系统海量数据的挖掘分析。
문장침대전망운행산생적수거정폭작식증장，EMS 계통유효신식왕왕엄몰재해량수거중저일문제，제출일충운계산모식하적취류분석처리방법，기우 Hadoop 평태적 MapReduce 계산모형여분포식문건존저，장계통취류법진행탁분，재운배경중대다개계산모괴진행병행분석。작위시험성험증，제취모대용호근삼년적부하특정곡선，선취불동수거량、불동절점수，진행산법가속비적측시。결과표명，재운계산가구중해산법가이유효제고계산효솔，괄용우전력계통해량수거적알굴분석。
In the era of big data, the data generated by grid operation has an explosive growth. Hadoop as an cloud computing platform which is an open source software is able to handle large amounts of data ef ectively. On the basis of depth study of traditional systematic clustering and Hadoop's MapReduce computation model and distributed file storage this paper splits the systematic clustering method and analyzes each section, final y proposes a new systematic clustering suitable for cloud computing. It is implemented on the platform of Hadoop and its speedup is tested by dif erent amounts of data and dif erent numbers of nodes. Also a certain user load characteristic curve for nearly three years is extracted in this paper. Algorithm is applied to massive data mining and analysis of power system.