###
DOI:
电力大数据:2023,26(1):-
←前一篇   |   后一篇→
本文二维码信息
基于随机森林的短期电量负荷精准预测方法
沈杰1, 李大任1, 周扬2, 甘泽鸿2, 葛宇达2, 黄光群1
(1.国网浙江省电力有限公司温州供电公司 浙江温州;2.国网浙江省电力有限公司温州市洞头区供电公司)
Accurate short-term power load forecasting method based on random forest
Shen Jie1, Li Daren1, Zhou Yang2, Gan Zehong2, Ge Yuda2, Huang Guangqun1
(1.State Grid Zhejiang Electric Power Co., Ltd. Wenzhou Power Supply Company Wenzhou;2.State Grid Zhejiang Electric Power Co., Ltd. Wenzhou Dongtou district power supply company)
摘要
图/表
参考文献
相似文献
本文已被:浏览 231次   下载 823
投稿时间:2022-05-19    修订日期:2023-02-15
中文摘要: 可再生能源并入电网后,电能供给量增加,短期电量负荷情况难以预测,无法制定准确的电能分配策略,由此,提出基于随机森林的短期电量负荷精准预测方法研究。深入分析短期电量负荷预测影响因素(气象、时间、电价与随机干扰因素),选取适当的模型输入变量(历史电量负荷数据、温度数据与日类型),结合随机森林算法构建短期电量负荷预测模型,并重复确定相似日的选取规则,采用粒子群优化算法寻找预测模型参数最佳值,将样本集输入至模型中,获得精准的短期电量负荷预测结果。实验数据显示:当输入变量数量达到一定值后,应用提出方法获得的短期电量负荷预测时延稳定在0.55s左右,短期电量负荷预测误差几乎为0,充分证实了提出方法应用性能较佳。
Abstract:After the renewable energy is incorporated into the power grid, the power supply increases, and the short-term power load is difficult to predict, so it is impossible to formulate an accurate power distribution strategy. Therefore, a research on the accurate prediction method of short-term power load based on random forest is proposed. Deeply analyze the influencing factors of short-term electricity load forecasting (weather, time, electricity price and random interference factors), select appropriate model input variables (historical electricity load data, temperature data and daily type), construct the short-term electricity load forecasting model combined with random forest algorithm, repeatedly determine the selection rules of similar days, use particle swarm optimization algorithm to find the best value of forecasting model parameters, and input the sample set into the model, Obtain accurate short-term power load forecasting results. The experimental data show that when the number of input variables reaches a certain value, the short-term power load forecasting delay obtained by the proposed method is stable at about 0.55s, and the short-term power load forecasting error is almost 0, which fully proves the good application performance of the proposed method.
文章编号:     中图分类号:    文献标志码:
基金项目:
引用文本: