| Title |
Development of Mean Radiant Temperature Virtual Sensor for Core and Perimeter Zones during the Summer using Random Forest |
| Authors |
성승호(Sung, Seung-Ho) ; 윤우승() ; 유원택() ; 서현철() ; 홍원화() |
| DOI |
https://doi.org/10.5659/JAIK.2023.39.3.171 |
| Keywords |
Mean radiant temperature; Machine learning; Virtual sensor; Thermal comfort |
| Abstract |
Mean radiant temperature (MRT) is one of many significant factors that influence an occupant’s thermal comfort. There is a deviation in the
MRT between the indoor core and perimeter zones depending on a building’s thermal properties; this deviation must be mitigated to ensure
thermal comfort. However, there are various practical limitations involved in directly measuring the MRT of these zones. Therefore, this study
developed a model that virtually sensed the MRT of the core and perimeter zones using the random forest. To verify the model’s
performance, the experiment was conducted during the summer season when the MRT deviation between these zones are often the largest. As
a result, the proposed model showed an MRT inference performance of 0.0568°C in the core zone and 0.123°C in the perimeter zone, based
on the mean absolute error. This study demonstrated the potential of the MRT virtual sensor for evaluating the inference performance of the
core and perimeter zones. The virtual sensor can be used in HVAC control systems to improve an occupant’s thermal comfort. |