이중외피 건물 실내 열 환경 제어를 위한 인공지능이론의 적용 수준별 성능비교분석

This study aimed at proposing an indoor temperature control method for providing a comfortable temperature conditions basedon the integrated controls of the cooling system and the surface openings. Employing different application levels of rules andartificial neural network models, four control logics were developed from the conventional rule-based control method to thethree ANN-based methods. Comparative performance tests were numerically performed in the double-skin-enveloped buildingusing TRNSYS (Transient Systems Simulation) and MATLAB (Matrix Laboratory) softwares in an incorporative manner.Analysis revealed that the ANN-based control methods provide more comfortable indoor temperature conditions with increasedcomfortable temperature periods and decreased standard deviation from the center of the comfortable range. However, theANN-based methods did not present superiority in energy efficiency over the conventional logic. From the analysis, theANN-based temperature control logic can be concluded to be able to keep the indoor temperature more comfortably and stablywithin the comfortable range.