基于MFC的墙体导热系数辨识程序设计
来源期刊：中南大学学报(自然科学版)2012年第9期
论文作者：程建杰 张小松 许大伟
文章页码：3687 - 3692
关键词：微软基础类库；辨识；导热系数
Key words：microsoft foundation class library(MFC); identification; heat conduction coefficient
摘 要：编制适用于夏热冬冷地区现场快速检测墙体导热系数的辨识软件。基于ARMAX的墙体热特性时间序列模型，利用系统辨识原理，通过现场检测实验结果，辨识室内壁面墙体热流，得出墙体的导热系数。针对传统的遗传算法收敛过早、个体不丰富的缺点，提出改进原则，制定改进的遗传算法辨识过程，分析面向对象的编程的特点，用Visual C++语言编制基于MFC的辨识程序。研究结果表明：壁面热流密度辨识的相对误差为4%左右，辨识速度快，程序界面友好。
Abstract: An identification software of the wall thermal conductivity was put forward for in-site measurement in hot summer and cold winter zone. Based on ARMAX time sequence model of wall, the volumes of the indoor wall heat flow were identified and the thermal conductivity of wall was worked out. Based on the convergence of traditional genetic algorithm defects in premature and few individual results, the improved principles for genetic algorithm identification were suggested. The characteristics of object-oriented programming were analysed. The identification software based on MFC with Visual C + + was programmed. The results show that the relative error of heat-flow density of wall is about 4% and the identification speed of the software is fast. The interface of software is friendly.
程建杰1,2，张小松1，许大伟2
(1. 东南大学 能源与环境学院，江苏 南京，210096；
2. 南京工业大学 城市建设与安全工程学院，江苏 南京，210009)
摘 要：编制适用于夏热冬冷地区现场快速检测墙体导热系数的辨识软件。基于ARMAX的墙体热特性时间序列模型，利用系统辨识原理，通过现场检测实验结果，辨识室内壁面墙体热流，得出墙体的导热系数。针对传统的遗传算法收敛过早、个体不丰富的缺点，提出改进原则，制定改进的遗传算法辨识过程，分析面向对象的编程的特点，用Visual C++语言编制基于MFC的辨识程序。研究结果表明：壁面热流密度辨识的相对误差为4%左右，辨识速度快，程序界面友好。
关键词：微软基础类库；辨识；导热系数
CHENG Jian-jie1, 2, ZHANG Xiao-song1, XU Da-wei 2
(1. School of Energy and Environment, Southeast University, Nanjing 210096, China;
2. College of Urban Construction & Safety Engineering, Nanjing University of Technology, Nanjing 210009, China)
Abstract:An identification software of the wall thermal conductivity was put forward for in-site measurement in hot summer and cold winter zone. Based on ARMAX time sequence model of wall, the volumes of the indoor wall heat flow were identified and the thermal conductivity of wall was worked out. Based on the convergence of traditional genetic algorithm defects in premature and few individual results, the improved principles for genetic algorithm identification were suggested. The characteristics of object-oriented programming were analysed. The identification software based on MFC with Visual C + + was programmed. The results show that the relative error of heat-flow density of wall is about 4% and the identification speed of the software is fast. The interface of software is friendly.
Key words:microsoft foundation class library(MFC); identification; heat conduction coefficient
