ANN-BASED MODEL-FREE THERMAL CONTROLS FOR RESIDENTIAL BUILDINGS.

Residential Building Thermal Control;Artificial Neural Network;Predictive and Adaptive Controls;Architecture;Arts;Architecture

Moon, Jin WooWeymouth, Terry E. ;

University of Michigan

关键词: Residential Building Thermal Control;    Artificial Neural Network;    Predictive and Adaptive Controls;    Architecture;    Arts;    Architecture;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/63664/jwmoon_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

This research aimed to develop an Artificial Neural Network (ANN)-based advanced thermal control method for creating more comfortable thermal environments in residential buildings. The proposed control method, which consisted of a thermal control logic and system hardware framework, was designed to improve residential thermal environments through the reduction of thermal imbalance in various rooms; the achievement of thermal comfort considering humidity or PMV as a control variable; and the reduction of overshoots and undershoots of air temperature, humidity and PMV using ANN-based predictive and adaptive control.In the control logic framework, four logics were employed for the residential thermal controls: (1) temperature and humidity control without ANNs as a conventional method, (2) PMV control without ANN, (3) temperature and humidity control with ANNs, and (4) PMV control with ANN. In addition, the system hardware framework was developed using sensors, data acquisition systems, a control panel, and building climate control systems.The performance of four developed control logics and system hardware was tested through computer simulation incorporating IBPT (International Building Physics Toolbox) and MATLAB, and through experiment. A typical two-story single-family home was modeled for the computer simulation while a thermal chamber was built for the experiment. Variables for the simulation were (1) the change of building conditions such as orientations, R-values for walls, the roof and windows, and window-wall-ratio, and (2) disturbances such as the change of internal load and ventilation rate, the application of setback, the change of setpoint, and the extreme change of exterior thermal conditions. Variables for the experiment were application and non-application of setback. The study reveals that ANN-based predictive and adaptive control strategies created more comfortable thermal conditions than ones without in terms of increased comfort period of air temperature, humidity, and PMV. This improvement was through the reduced ratio and magnitude of overshoots and undershoots out of the specified comfort ranges. In many cases, ANN-based strategies consumed less energy for building climate control systems although not as significantly as expected. Based on this study, it can be concluded that ANN-based predictive and adaptive climate control strategies can improve thermal comfort in residential buildings.