摘要

农村住宅节能对于节约能源、保护环境具有重要意义，但在片面追求能耗降低的同时会导致室内热舒适性差和初投资成本高等问题，需要考虑多维目标相互作用对节能设计参数确定的影响。本文整合性能模拟技术、支持向量机建模、多目标遗传算法优化和多属性决策方法，提出以机器学习和优化算法为核心，协同优化建筑能耗、室内热舒适和初投资成本的建筑多目标优化设计流程及方法。以中国寒冷地区乡村住宅为例对设计流程进行应用分析，将建筑朝向、南向和北向外窗宽度及传热系数、各朝向外墙及屋面保温层厚度、冬季采暖设计温度作为优化变量，通过运算得出包括设计变量及优化目标值的Pareto非支配解集，并提出考虑节能性、综合性、经济性三种不同需求的设计模式，该方法可为寒冷地区农村住宅节能优化设计提供参考。

关键词

寒冷地区；农村住宅；节能设计参数；多目标优化

Abstract

Energy saving in rural houses is of great significance for energy conservation and environmental protection, but the one-sided pursuit of energy consumption reduction will lead to poor indoor thermal comfort and high initial investment cost. It is necessary to consider the influence of multi-objective interaction on the determination of energy conservation design parameters. By integrating the performance simulation technology, support vector machine modeling, multi-objective genetic algorithm optimization and multi-attribute decision method, this paper proposes a multi-objective optimization design process and method for collaborative optimization building energy consumption, indoor thermal comfort and initial investment cost with machine learning and optimization algorithm as the core. Taking a rural house in cold areas of China as an example, the design process is applied and analyzed. The building orientation, south and north exterior windows width and heat transfer coefficient, insulation layer thickness of exterior wall and roof, and heating design temperature in winter are taken as optimization variables. The Pareto non-dominated solution including design variables and optimization objective values is obtained through calculation, thus the design patterns considering three different requirements of energy saving, comprehensiveness and economy are proposed，which can provide reference for the energy-saving optimization design of rural houses in cold regions.

Key words

Cold regions; Rural houses; Energy-saving design parameters; Multi-objective optimization