【摘要】 碳排放效率优化是实现低碳城市建设的重要途径。以金堂县居住区住宅建筑形态对单位面积碳排放影响为切入点,基于4个样本社区选取86栋住宅形态指标和3 256户住宅能耗统计数据,运用多元线性回归和机器学习方法在住宅形态层面分析了其对住宅单位面积碳排放的影响,并构建住区低碳规划指标表和碳排放预测模型。研究发现,金堂县居住区住宅体形系数、面宽进深比、梯户数均会对住宅碳排放产生显著影响,且不同层数梯度中影响趋势是不同的;住宅形态与碳排放之间并非简单的线性关系,而是随开发强度梯度变化的,通过机器学习模型可以很好地识别梯度变化下的重要影响因素,并定量化建立预测模型。因此,金堂县可以通过优化住宅形态降低住宅单位面积碳排放,提升碳排放效率,从而实现低碳城市建设。更多还原

【Abstract】 The optimization of carbon emission efficiency is an important way to realize the construction of low-carbon cities. Taking the influence of residential building form on carbon emission per unit area in Jintang county as the breakthrough point, based on four sample communities, 86 residential form indicators and 3 256 residential energy consumption statistical data are selected. The multiple linear regression and machine learning methods are used to analyze the impact on the carbon emission per unit area of residential buildings, and the low-carbon planning index table and carbon emission prediction model are constructed. The results show that the shape coefficient, the aspect ratio and the number of staircases have a significant impact on the carbon emission of residential buildings in Jintang county, and the influence trend is different in different layers gradient; the relationship between residential form and carbon emission is not a simple linear relationship, it changes with the development intensity gradient. The important influence under the gradient change can be well identified by the machine learning model and quantitative establishment of prediction model. Therefore, the residential form can be optimized to reduce the carbon emission per unit area and improve the carbon emission efficiency, so as to realize the construction of low-carbon city.更多还原