基于Gumbel copula联合概率分布的电力系统综合净负荷预测

Aggregated Net Load Forecasting for Modern Power Systems Based on Gumbel Copula Joint Probability Distribution

近年来,太阳能和风能在电力系统中的渗透率不断提高。由于风能和太阳能具有的高度不确定性和间歇性,在其并网时需要采用复杂多变的操作策略才能维持电网稳定运行。对于多个或单个不确定变量进行净负荷的汇总可以有效降低系统运行规划的复杂性。由此,提出了一种基于Gumbel copula联合概率分布的新型净负荷预测(net load forecasting,NLF)模型用以对负荷、风能和太阳能发电的预测误差进行有效应对,此外所提模型还引入了基于预测值的Grey指数修正模型以提高预测的精准度。其中由于Gumbel copula联合概率模型的稳定性好,其基本能够覆盖所有极端预测误差,在保证系统可靠运行前提下,基于有限的输入实现了快速精准的预测。实验结果表明,所提模型在极短期净负荷预测上具有较好的效果。

Wind and solar powers possess a growing share in power system.Due to their high uncertainty and intermittence,it takes complicated and flexible operation strategies to keep grid stable when wind and solar energy are connected to the grid.However,summary of net loads for multiple or single uncertainties effectively simplifies system operation planning.This study proposed a new net load forecasting(NLF)model by Gumbel-Copula-based joint probability distribution to eliminate load,wind and solar generation forecasting error.Besides,the proposed model brought in modified Grey index models to improve forecasting accuracy.As Gumbel copula covers all extreme forecasting errors due to max-stable property,the forecasting is fast and accurate with limited input under the premise of ensuring the reliable operation of the system.Results show that the proposed model is excellent in very short-term NLF.