基于变分模态分解的综合能源系统短期电负荷预测
Short-term electrical load forecasting for integrated energy system based on variational mode decomposition
苏子越 1柴琳 1谢亮 1肖凡1
作者信息
- 1. 武汉科技大学信息科学与工程学院,湖北 武汉 430081
- 折叠
摘要
针对综合能源系统负荷复杂多变、耦合性强的特点,提出一种基于变分模态分解(variational mode decomposition,VMD)、Prophet 模型、长短时记忆(long-and short-term memory network,LSTM)神经网络、差分自回归移动平均(autoregressive integrated moving average,ARIMA)模型的Prophet-VAL组合预测模型,用于综合能源系统短期电负荷预测.首先,通过VMD获取不同中心频率和较为稳定的电负荷本征模态函数;接着,根据过零率值的大小将不同模态分量分成高频和低频时序分量,并使用 Prophet 模型将高频分量进行时序特征提取;最后,通过ARIMA预测模型对低频分量进行预测,使用LSTM神经网络模型对高频分量进行预测,将各自的预测结果进行叠加得到最终的电负荷预测结果.将所提方法应用于实际综合能源系统,实际算例分析表明,所提出的组合预测模型预测性能良好.
Abstract
Aiming at the characteristics of complex and variable load and strong coupling of integrated energy system,a combined forecasting model based on variational mode decomposition(VMD),Prophet model,long-and short-term memory network(LSTM)and autoregressive integrated moving average(ARIMA)model is proposed for short-term electrical load prediction.Firstly,the electric load eigen mode functions with different center frequencies and relatively stable ones are obtained by VMD.Then,after calculating the value of zero cross rate,the modal components of each group are superimposed respectively to form the high-frequency and low-frequency timing components,and the Prophet model is used to extract the high-frequency components for timing features.Finally,the ARIMA prediction model is used to predict the low frequency component,and the LSTM neural network model is applied to predict the high frequency component.The final predicted electric load is obtained by superimposing the respective prediction results.The proposed method is applied to the actual integrated energy system,and the example analysis shows that the combined forecasting method presented above has good forecasting performance for the integrated energy system.
关键词
综合能源系统/负荷预测/变分模态分解/LSTM神经网络/Prophet模型Key words
integrated energy system/load forecasting/variational mode decomposition/LSTM neural network/Prophet model引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据