데이터 기반 모델을 활용한 호텔 객실의 열적 거동 예측

There have been significant efforts to apply a building simulation model to operation and control of building systems. The first-principle based approach requires demanding information and time for determining simulation inputs. In addition, it requires in-depth knowledge and expertise to make best use of simulation tools. It is also not easy to take into account uncertain phenomena a (e.g. infiltration, occupant’s behavior, unplanned change of HVAC operation, etc.). Therefore, the data-driven approach has been highlighted as an alterative to the first-principle based approach since it requires far less inputs, time and effort, and expertise. In this paper, the authors address a case study a data-driven model which can predict the thermal behavior of a hotel room conditioned by a ceiling-recessed FCU. The inputs to the model include indoor temperature, outdoor temperature, and occupant’s presence and behavior (opening of windows). The authors employed a Nonlinear Auto-Regressive eXogenous(NARX) method. It is reported in the paper that the data-driven model successfully predicts the room behavior .