期刊
ENERGY STRATEGY REVIEWS
卷 49, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.esr.2023.101160
关键词
Energy consumption; Carbon emission; Carbon peak; Decomposition-ensemble
Accurately predicting energy consumption and carbon emission is crucial for China's carbon peak and carbon neutrality goals. This study proposes a novel three-layer decomposition-ensemble forecasting method that combines the advantages of trend decomposition, empirical mode decomposition, and wavelet decomposition. By breaking down the complex energy consumption prediction task into simpler trend and non-trend subseries prediction tasks, the forecasting performance is improved.
Accurately predicting energy consumption (EC) and carbon emission is essential for China's achieving carbon peak by 2030 and carbon neutrality by 2060. In view of decomposition-ensemble (DE) strategy can effectively enhancing the prediction performance by decreasing the complexity and nonlinearity of the original EC forecasting problems. A novel three-layer DE forecasting approach is proposed, which combines the advantages of trend decomposition (TD), empirical mode decomposition (EMD) and wavelet decomposition (WD). Compared with traditional individual decomposition strategy or secondary decomposition strategy, the proposed three layer DE forecasting method breaks EC dataset into more simpler trend subseries and non-trend subseries. Then the forecasting performance is improved by turning the original complex EC prediction task to some relatively simpler trend and non-trend subseries prediction tasks. The future carbon emission is solved out using IPCC carbon emission coefficient method and EC prediction results. Empirical analysis proves the validity of the proposed forecasting method. Finally, China's EC and carbon emission in 2021-2035 under different EC structure, economic growth, industrial structure, population, energy efficiency, household EC per capita Scenarios are predicted by the proposed three-layer DE forecasting model. And some relevant energy and carbon emission policies and suggestions are put forward.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据