摘要
针对当前电力系统中长期负荷预测过于依赖模型、数据利用效率低的不足,提出了一种中长期负荷的大数据预测方法.该方法采用分析-综合的预测模式,在数据分析的基础上实现负荷预测的建模.首先对负荷进行分层分区与分类以解构负荷预测大数据,多角度分析类型负荷的增长特性;然后基于增长特性建立分析模型,分别预测分区类型负荷的增长;最后基于数据解构模式建立综合模型,统筹分析模型、协调全局负荷、整合大区域负荷总体增量.研究算例表明,该方法能对负荷增长的细节进行分析,能对区域各负荷增量进行整合,具有较高的精确度.
According to such shortage in mid-long term load forecasting as low efficiency in using of the da- ta and excessive relying on the models, a big data prediction method is proposed. Using the analysis and synthesis prediction mode, the load forecasting models are established based on the analysis of data. First- ly the data of load prediction are deconstructed by partitioning to different levels and classifying to different types so as to analyze their increasing characteristics. Based on the increasing characteristics then the anal- ysis prediction models are built to analyze the increase of different types of load. Finally the synthesis pre- diction models are built based on the data deconstructed mode and the analysis prediction models are coor- dinated. The whole load is harmonized and the increase of overall load in large area is integrated. Case study results show that the method can analyze the details of load growth and coordinate the increase of different types of load in large area to bring high precision.
出处
《武汉大学学报（工学版）》
CAS
CSCD
北大核心
2017年第2期239-244,共6页
Engineering Journal of Wuhan University
基金
国家自然科学基金项目(编号:51107090)
关键词
中长期负荷预测
大数据技术
分析-综合预测
分区分类
mid-long term load forecasting
big data technology
analysis and synthesis prediction mode
partition and classification
