华电技术 ›› 2015, Vol. 37 ›› Issue (9): 8-10.

• 研究与开发 • 上一篇    下一篇

数据预处理对电力负荷预测精度的影响

  

  1. 1.南京工程学院电力工程学院,南京〓2111672.中兴通讯南京研发中心,南京〓210012
  • 出版日期:2015-09-25 发布日期:2015-11-26

Research on data pre processing to improve the accuracy of load forecasting

  1. 1.Department of E1ectric Power EngineeringNanjing Institute of TechnologyNanjing 211167China2. Nanjing Technical CenterZTE CorporationNanjing 210012China
  • Online:2015-09-25 Published:2015-11-26

摘要:

摘要:电力系统负荷预测精度越高,越有利于提高发电设备的利用率和经济调度的有效性。针对淮阴地区2013年负荷数据进行研究,分析了负荷的特性及不同的数据预处理方法对负荷预测精度的影响。结果表明:异常值平稳化是最直接、有效的方法;原始数据每日样本量对预测结果影响不大;预测重大节假日过后的工作日负荷时可删去双休日及重大节假日数据,预测精度显著提高;春、秋季负荷较平稳时,调整同类型日数据顺序也可适当提高预测精度。

关键词: ">关键词:电力负荷;预测;数据预处理;预测精度;负荷特性;同类型日

Abstract:

Abstract:The higher prediction accuracy is,the more helpful is to improve the availability of facilities and the validity of economic dispatches.
Through the load characteristics analysis of Huaiyin area in 2013,the effects of load characteristics and different data pre
processing methods are analyzed.The results show that stabilizing the abnormal data is the most directive and efficient method.The original data of daily amount of samples had little effect on the predicted results.The prediction accuracy will be improved greatly when the load forecasting of the working days after major holidays is predicted by deleting the original data in the weekends and major holidays Adjustment of data sequence of the same day types can also be appropriate to increase the prediction accuracy when load is stable in spring and autumn.