摘要
在国家标准GB/T 51350—2019《近零能耗建筑技术标准》中,提出了近零能耗建筑性能化设计理念。本文提出1种基于最短路径算法的近零能耗建筑被动式节能设计参数优化方法。该方法基于能耗模拟软件生成数据集,训练基于人工神经网络的快速建筑能耗预测模型,预测精度可达到误差1%左右,基于快速能耗预测模型,采用最短路径寻找算法,对被动式节能设计参数集合进行优化,大幅缩短了优化策略的寻找时间,便于工程应用。
The National Standard Technical standard for nearly zeroenergy consumption buildings(GB/T 51350-2019)proposed the performance-oriented design concept of nearly zero-energy consumption buildings.This paper presented a parameter optimization method for the passive energy-saving design of nearly zero-energy buildings based on the shortest path algorithm.This method generates data sets based on energy consumption simulation software and trains a fast energy consumption prediction model based on the artificial neural network.The prediction accuracy can reach an error rate of around 1%.This method uses the fast energy consumption prediction model and the shortest path to find algorithm,which greatly shortens the search time of optimization strategy,and provides convenience for engineering practices.
作者
刘伟
于震
龚红卫
吴剑林
LIU Wei;YU Zhen;GONG Hongwei;WU Jianlin(China Academy of Building Research,Beijing 100013,China;College of Urban Construction,Nanjing Tech University,Nanjing 210009,Jiangsu,China)
出处
《建筑科学》
CSCD
北大核心
2023年第2期58-66,73,共10页
Building Science
基金
国家重点研发专项课题“公共建筑技术集成和示范工程研究”(No.2017YFC0702610)。
关键词
近零能耗建筑
人工神经网络
最短路径算法
参数优化
nearly zero-energy building
artificial neural network
shortest path algorithm
parameter optimization
