基于用户画像及变分模态分解的综合能源系统多元负荷预测

Multivariate Load Forecasting of Integrated Energy System Based on User Image and Variational Mode Decomposition

摘要: 针对多因素影响综合能源系统多元负荷预测导致预测不精准的问题，本文提出了基于用户画像及变分模态分解的综合能源系统多元负荷预测。首先，引入用户画像的概念，通过CRITIC赋权法对影响用户用能的各种因素进行加权计算，赋予各类因素的权重信息，生成典型的用户画像。之后，使用变分模态分解(VMD)将用户电热冷的用能数据以及相关因素数据分解为本征模态函数(IMF)，再建立GRU模型对IMF进行预测。最后，基于用户画像所提供的权重信息叠加各因素IMF预测结果得到最终的综合能源系统多元负荷预测。仿真分析结果表明：本文预测模型能够有效地预测综合能源系统多元负荷变化趋势，其平均绝对百分比误差、方均根误差和最大相对误差评价指标均优于通过GRU模型和VMD-GRU模型进行预测的结果。

Abstract: Aiming at the problem that multiple factors affect the multiple load forecasting of integrated energy system and lead to inaccurate forecasting, this paper proposes a multiple load forecasting of inte-grated energy system based on user image and variational mode decomposition. First of all, the concept of user portrait is introduced, and various factors affecting the user’s energy consumption are weighted by CRITIC weighting method, and the weight information of various factors is given to generate a typical user portrait. After that, the energy consumption data and relevant factor data of users’ electric cooling are decomposed into intrinsic mode function (IMF) using variational mode decomposition (VMD), and then the GRU model is established to predict the IMF. Finally, based on the weight information provided by the user portrait and the IMF prediction results of various fac-tors, the final comprehensive energy system multiple load forecast is obtained. The simulation re-sults show that the prediction model in this paper can effectively predict the trend of multiple load changes in the integrated energy system, and its average absolute percentage error, root mean square error and maximum relative error evaluation indicators are better than those predicted by GRU model and VMD-GRU model.