基于SVM和NSGA-Ⅱ算法的西安地区乡村住宅高维多性能协同优化设计研究  

High-dimensional Multi-performance Collaborative Optimization Design of Rural Residences in Xi’an Based on the SVM and NSGA-ⅡAlgorithm

作　　者：邵腾 张妮 张锟 郑武幸 王晋[1] SHAO Teng;ZHANG Ni;ZHANG Kun;ZHENG Wuxing;WANG Jin(School of Mechanics,Civil Engineering and Architecture,Northwestern Polytechnical University,Xi'an 710129,China)

机构地区：[1]西北工业大学力学与土木建筑学院,西安710129

出　　处：《建筑科学》2024年第10期232-243,共12页Building Science

基　　金：陕西省自然科学基础研究计划项目“基于机器学习和优化算法的居住建筑方案阶段多目标优化设计研究——以西安市为例”(2022JQ-549);陕西省自然科学基础研究计划项目“西部寒冷地区农村老年人居住热环境动态评价与优化设计策略研究”(2023-JC-YB-355);陕西省社会科学基金““双碳”目标背景下关中地区传统民居绿色营建智慧传承与当代转译研究”(2022J009)。

摘　　要：随着乡村高质量发展及居民生活品质的提升,乡村住宅节能设计也不单局限于降低用能,在我国“双碳”目标背景下,应将能源、碳排、舒适与成本综合考量。本文基于西安地区的气候资源背景,以建筑能耗、碳排放、热舒适、自然采光和成本为优化目标,搭建以SVM和NSGA-II算法为核心的乡村住宅优化设计框架与方法,并通过实证研究构建最佳设计模式,该方案相对基准情景的节能率为76.0%、全生命期碳排放降低1000.0 kgCO_(2)/m^(2)、热舒适性提升44.0%、自然采光性能提升7.4%、全生命期成本节约63.0元/m^(2)。在验证高维多目标优化结果可行性的基础上,提出多情景下的设计变量取值建议区间,为乡村住宅性能化设计提供科学量化的依据和指导方案。With the high-quality development of rural areas and the improvement of residents'life quality,the energy-saving design of rural residences is not limited to reducing energy use.In the context of China's“carbon peaking and carbon neutrality”goal,energy,carbon emissions,comfort and costs should be considered comprehensively.Based on the climate and resource background of Xi'an,this paper built a framework and a method of rural residential optimization design based on the SVM and NSGA-II algorithm,taking building energy consumption,carbon emissions,thermal comfort,natural lighting and costs as optimization objectives.Moreover,an optimal design pattern was developed through empirical research.Compared with the baseline scenario,the plan achieved an energy saving rate of 76.0%,reduced the life-cycle carbon emissions by 1,000.0 kgCO_(2)/m^(2),improved thermal comfort by 44.0%,increased lighting performance by 7.4%,and saved life-cycle costs by 63.0 RMB/m^(2).On the basis of verifying the feasibility of the high-dimensional multi-objective optimization results,the proposed value range of design variables under multiple scenarios was put forward to provide scientific and quantitative guidance for the performance-based design of rural residences.

关 键 词：乡村住宅 多维性能 智能算法 协同优化 设计模式 

分 类 号：TU831[建筑科学—供热、供燃气、通风及空调工程] TU111.19