摘要
针对用电大数据平台关键技术研究以及大数据技术和电力业务的融合应用中面临的困难,提出了有效的解决方案,并结合案例进行了用电数据的深度挖掘。首先,根据用电大数据平台建设过程中面临的多源异构数据抽取、不同业务需求数据存储、多条件快速组合查询这3方面技术难题,提出利用Golden Gate实现对源数据库低影响的高效数据抽取,采用HBase和Hive协同合作的数据存储方案以满足快速查询和离线计算需求,在数据导入阶段构建高效立体索引以提高快速组合查询效率;其次,结合大数据技术,实现了配变负荷影响模型构建、短期网供负荷预测、居民住房空置率分析及行业产能利用率分析4个应用场景,为提升负荷建模精确度和负荷预测准确率提供了技术支撑,为政府掌握居民住房情况和行业发展状况提供了数据支持。
In view of difficulties faced by key technology research of electricity big data and fusion application of big data technology and power system, an effective solution was proposed to solve the problems, and deep data mining of electricity big data combined with case analysis was applied. Firstly, based on construction and operation of electricity big data platform, technological problems were presented, such as multi-source heterogeneous data extraction, data storage scheme for meeting different business requirements and multi conditional fast combination data query. Golden Gate was applied to realize low impact and high efficiency data extraction, HBase and Hive were collaborated to meet requirements of fast query and off-line calculation, and efficient stereo index system was constructed in data import phase to improve efficiency of fast combination query. Secondly, combined with big data technology, construction of load impact model of distribution transformer, short-term load forecasting, residential housing vacancy rate analysis and industry capacity utilization rate were realized. Above application scenarios provided technical supports for improving accuracy of load modeling and forecasting and data supports for government to control housing situation and industry development status.
出处
《电网技术》
EI
CSCD
北大核心
2015年第11期3147-3152,共6页
Power System Technology
基金
江苏省电力公司科技项目(J2014018)的资助
作者简介
郑海雁（19791，男，硕士，高级工程师，研究方向为电力系统信息化：
金农（1956），男，本科，高级工程师，研究方向为电力系统营销；
季聪（1988），男，硕士，研究方向为电力系统信息化，E-mail：icxx01@163．com。
