摘要
将人均GDP、人口数量、能源消耗量、房屋竣工面积和建筑业能源消费强度作为新疆建筑业碳排放的主要影响因素,并建立改进队列要素法模型,对新疆人口数量进行预测,引入加权MLP⁃ARIMA时间序列预测模型对新疆GDP进行预测。同时,建立改进柔性灰色多变量预测模型对新疆建筑业未来能源消费进行预测。运用情景分析法结合基于Attention机制的长短期记忆神经网络碳排放预测模型,对未来新疆建筑业碳排放趋势进行动态模拟分析。情景预测结果表明,保持经济水平与城市化水平稳步发展,严格控制建筑业能源消费量与加速优化建筑业能源消费强度是新疆建筑业最优减排路径,其将在2032年实现碳达峰。
The per capita GDP,population,energy consumption,completed area of houses and energy consumption intensity of construction industry are taken as the main influencing factors of the carbon emissions of Xinjiang construction industry,and the improved cohort⁃component method model is established to predict the population of Xinjiang.The weighted MLP⁃ARIMA time series prediction model is introduced to predict the GDP of Xinjiang.At the same time,an improved flexible grey multivariable prediction model is established to predict the future energy consumption of Xinjiang construction industry.The scenario analysis method and the long short⁃term memory(LSTM)carbon emission prediction model based on the attention mechanism(AM)are used to dynamically simulate and analyze the carbon emission trend of Xinjiang construction industry in the future.The scenario prediction results show that maintaining the steady development of the economic level and urbanization level,strictly controlling the energy consumption of the construction industry and accelerating the optimization of the energy consumption intensity of the construction industry are the optimal emission reduction paths for Xinjiang construction industry,which will reach the peak carbon dioxide emissions in 2032.
作者
吕言
宋华
南新元
LÜYan;SONG Hua;NAN Xinyuan(School of Electrical Engineering,Xinjiang University,Urumqi 830017,China;Mechanical and Electrical Branch Institute No.1,Xinjiang Architectural Design Institute,Urumqi 830002,China)
出处
《现代电子技术》
2023年第15期121-127,共7页
Modern Electronics Technique
基金
国家自然科学基金项目(61863033)。
