同时发生设计日集合覆盖挑选与预测方法  

Selection and prediction methods for coincident design day based on set covering problem

作　　者：陈友明[1,2] 谢广仁 CHEN Youming;XIE Guangren(College ofCivil Engineering,Hunan University,Changsha 410082,China;Key Laboratory of Building Safety and Energy Efficiency of Ministry of Education,Hunan University,Changsha 410082,China)

机构地区：[1]湖南大学土木工程学院,湖南长沙410082 [2]建筑安全与节能教育部重点实验室(湖南大学),湖南长沙410082

出　　处：《华中科技大学学报(自然科学版)》2025年第2期120-126,共7页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(52130802)。

摘　　要：为挑选出能准确计算空调系统设计负荷的同时发生设计日,提出了一种基于集合覆盖问题的同时发生设计日选取与预测方法.该方法用长年逐时气象数据和负荷计算模型生成设计日-房间样本表征矢量,基于集合覆盖模型,将同时发生设计日的挑选问题转化成集合覆盖的求解问题,用蚁群算法得到两个或多个设计组合而成的最优集合.用XGBoost算法构建房间特征参数与最优集合中元素的映射关系,从最优集合中预测匹配待设计房间的设计日.用该方法挑选香港地区的最优集合,并预测匹配待设计房间的设计日.实例研究表明:最优集合中同时发生设计日能准确计算任一特征参数组合房间的设计负荷,用预测设计日计算的设计冷负荷与理论设计冷负荷之间的差异率在1.5%以内,满足工程计算精度要求.In order to get coincident design day to accurately calculate the design cooling load of air-conditioning system,a method based on set covering problem was proposed in this study to select and predict coincident design day.In this method,the representation vectors between design days and room samples were generated through long-term hourly weather data and cooling load calculation models,and the selecting problem of coincident design days was transformed into solving set covering problem by set covering model.The ant colony algorithm was used to obtain an optimal set which contains several coincident design days.XGBoost algorithm was used to construct the mapping relationship between the room characteristic parameters and the elements of the optimal set.The coincident design day of a room in design was predicted from the optimal set.This method was applied to obtain the optimal set of coincident design days for Hong Kong and predict the design days of the rooms in design.The results in case study show that a coincident design day with accurate load calculation can be selected from the optimal set for any characteristicparameters-combination rooms.The difference between the design cooling load calculated by the predicted coincident design day and the theoretical design cooling load is less than 1.5%,which meets the accuracy requirement in engineering application.

关 键 词：同时发生设计日 集合覆盖问题 最优集合 XGBoost算法 设计冷负荷 

分 类 号：TU831.2[建筑科学—供热、供燃气、通风及空调工程]