中文摘要：

随着电网数据收集能力的提升,积累了海量的负荷及相关数据,为负荷预测开辟了新的思路。提出了一种应用大数据技术的中长期负荷预测新方法。首先通过历史负荷序列的增长趋势、波动性等变化特性的参数化表达,实现负荷的标准化处理,形成大数据聚类的样本;然后结合大数据分析平台的数据处理能力设计了基于Map Reduce并行编程模型的改进模糊K-means聚类方法,实现对负荷大数据的聚类划分;最后综合相同聚类负荷,并建立预测模型。计算结果表明,大数据聚类算法能有效地进行大量负荷数据的聚类划分,实现不同增长特性负荷的区分预测,具有较高的预测精度。

英文摘要：

With the improvement of data collection ability,massive data of power load and related data are gradually accumulated,which brings a new way for load forecasting.In this paper,a new method for mid-long term load forecasting is proposed based on big data technology.First,through the parametric expressions of the increasing trends and the volatility of historical load series,a standardized treatment of load is realized so as to form the samples of big data clustering.Then,an improved fuzzy K-means algorithm based on Map Reduce parallel programming model is designed with the combination of data processing ability of big data platform,thus the clustering analysis of big data of load is achieved.At last,forecasting models are built through synthesizing the same category of load.The calculating results show that the proposed algorithm can effectively accomplish the clustering analysis of massive load data,achieve a differentiated forecasting of load with diverse growth properties and improve the prediction accuracy.