基于时间序列聚类算法优化下的多变量短期负荷预测模型研究
Research on Multi-variable Short-term Load Forecasting Model Optimized by Time Series Clustering Algorithm
徐洁1
作者信息
- 1. 国能(广东)综合能源有限公司,广东 广州 510000
- 折叠
摘要
为提高短期负荷预测的精度问题,针对短期负荷预测的特点,采用了对海量序列数做数据增强聚类操作,和外部输入变量(天气因素)并行处理,提出了基于时间序列聚类算法优化下的多变量短期负荷预测模型,并对某电力售电公司进行了实际操作.结果表明:该方法大幅提升了模型的预测精度和实用能力.
Abstract
To improve the short-term load forecasting accuracy,considering the characteristics of short-term load forecast,a multi-variable short-term load forecasting model optimized by a time series clustering algo-rithm is proposed.This model performs data enhancement clustering operations on massive sequences and pro-cesses external input variables(weather factors)in parallel.The model has been implemented in practical op-erations for an electricity sales company.The results show that the method significantly improves the forecast-ing accuracy and utility of the model.
关键词
短期负荷预测/时间序列/聚类分析/天气Key words
Short-term Load Forecasting/Time Series/Clustering Analysis/Weather引用本文复制引用
出版年
2024
NETL
NSTL
维普
万方数据