基于SSA-PSO-GRU的短期电力负荷预测

Short⁃term power load forecasting based on SSA⁃PSO⁃GRU

为了提升短期电力负荷预测的精度,该文提出一种基于SSA-PSO-GRU的短期电力负荷预测方法。针对电负荷的非线性和不确定性问题,该文采用奇异谱分析对实测的电力负荷进行分解,把复杂度高、波动性较强的电力负荷分解成若干平稳性好、可预测性强的周期分量、趋势分量及噪声分量,再采用基于粒子群优化算法寻找最优超参数的GRU模型,对分量进行预测并重构得到最终预测结果。通过对西班牙瓦伦西亚市2018年电力负荷数据仿真分析,与其他预测方法对比,结果表明,该文所提方法有效提高了短期电力负荷预测精度。

In order to improve the accuracy of short⁃term power load forecasting,a short⁃term power load forecasting method is proposed based on SSA-PSO-GRU.Aiming at the nonlinear and uncertain problems of electric load,this paper uses singular spectrum analysis to decompose the measured electric load,and decomposes the electric load with high complexity and strong volatility into several periodic components,trend components and noise components with good stability and strong predictability.Then,the GRU model based on particle swarm optimization algorithm is used to find the optimal hyperparameters,and the components are predicted and reconstructed to obtain the final prediction results.Through the simulation analysis of the power load data of Valencia City,Spain in 2018,compared with other forecasting methods,the results show that the proposed method effectively improves the short⁃term power load forecasting.