神经网络预测功能在中国能源结构调整中的应用
董禹¹，张召锋²
（1.北京交通大学，北京 100044；2.中国石油集团长城钻探工程有限公司，北京 100101）
摘要： 首先通过神经网络建模拟合得到了中国GDP 与能源消费总量的关系，其次根据不同GDP 增速预测得到了至
2060 年的能源消费总量，最后根据2019 年能源结构设置了3 种调整方式（保守趋势、现行趋势及严格趋势）。碳排放预
测神经网络的碳排放结果显示，保守趋势的调整方案难以缓解和控制目前中国碳排放量逐渐上升的趋势，现行趋势的调
整方案也仅在经济增速较低的情况下能控制碳排放，严格趋势的调整方案在各种经济增长情景中都能得到显著降低碳排
放量的结果。结果显示，能源结构调整对中国未来碳排放量及其变化趋势具有重要影响。
关键词： 神经网络；能源结构；碳排放
中图分类号： TP183；X321      文献标志码： A      文章编号： 2095-0802-(2023)07-0001-08
Application of Neural Network Prediction Function in Energy Structure
Adjustment in China
DONG Yu¹, ZHANG Zhaofeng²
(1. Beijing Jiaotong University, Beijing 100044, China; 2. CNPC Great Wall Drilling Engineering Co., Ltd., Beijing 100101, China)
Abstract: The relationship between China's GDP and total energy consumption was obtained through neural network modeling
firstly, and then the total energy consumption by 2060 was obtained according to different GDP growth forecasts. Finally, three
adjustment methods (i.e. conservative trend, current trend and strict trend) were set according to the energy structure in 2019. The
carbon emission results obtained through the neural network for predicting carbon emissions show that it is difficult to mitigate and
control the current rising trend of China's carbon emissions with the adjustment scheme of conservative trend, the adjustment
scheme of current trend can only control carbon emissions under the condition of low economic growth, and the adjustment
scheme of strict trend can achieve significant results in reducing carbon emissions in various economic growth scenarios. The
calculation results show that the adjustment of energy structure will play an important role in the amount and change trend of
China's future carbon emissions.

Key words: neural network; energy structure; carbon emissions