Optimal classification of radiant time series for cooling load calculation of building air-conditioning systems

Abstract

Building cooling load is caused by complex thermal effect. Design cooling load determines the capacity of air-conditioning assembly units and systems to maintain stable thermal environment and high operating efficiency. The radiant time series method (RTSM) is applicable for design cooling load calculation in engineering application. The radiant time series (RTS) should precisely match various room types by improving their accuracy and practicality. The room construction types require more explicit identification. This study proposes RTS classification method based on heat balance method (HBM) and CatBoost to obtain the representative RTSs for various room types. A new quantitative indicator, interior surface heat storage coefficient, is introduced to quantify the construction type. An accuracy evaluation is performed on the representative RTSs and design cooling loads calculated using them. The results show that interior wall type, ceiling type and floor type are important parameters for RTS classification. Compared to the HBM, the representative RTSs cause relative deviations of design cooling loads within 5 % for over 90 % of the room samples, with mean deviations within 2 %. This has demonstrated accurate RTS classification in this study. The appropriate representative RTSs with refined construction type can ensure accurate cooling load calculation and facilitate extensive application.