基于VMD-PSO-SVR模型的短期负荷预测

Short-term Load Forecasting Based on VMD-PSO-SVR Model

准确的短期负荷预测结果可以为电网内机组的调度提供基础,制定出合理的调度方案,从而提高电网运行效率。作者提出了短期电力负荷预测的VMD-PSO-SVR组合模型。首先,对原始负荷数据进行预处理,组合各类特征构建负荷数据集,利用VMD对负荷数据集进行分解,降低数据的非光滑性;其次,利用SVR算法对分解后的每个IMF分量进行单独预测,并使用PSO算法对SVR算法的超参数进行优化,提高SVR算法的预测精度;最后,对所有IMF分量所对应的预测结果进行叠加处理,从而获得最终预测结果。实验结果表明,该模型MAPE为1.55%,RMSE为38.56 MW,优于其他预测模型。

Accurate short-term load forecasting results can provide a basis for the dispatch of units within the power grid,formulate reasonable dispatch plans,and thereby improve the efficiency of power grid operations.This paper proposes a combination model of VMD-PSO-SVR for short-term power load forecasting.First,the original load data is preprocessed,and various features are combined to construct the load dataset.VMD is used to decompose the load dataset to reduce the non-smoothness of the data.Second,the SVR algorithm is used to individually forecast each IMF component after decomposition,and the PSO algorithm is used to optimize the hyperparameters of the SVR algorithm to improve its forecasting accuracy.Finally,the predicted results corresponding to all IMF components are superimposed to obtain the final forecast result.Experimental results show that the MAPE of this model is 1.55%and the RMSE is 38.56 MW,which is better than other forecasting models.