Development of PMV Virtual Sensors for Optimal Seat Reservation in Libraries

Thermal comfort varies according to the seating infrastructure, even within the same space; and individuals may perceive thermal comfort differently, albeit using the same seat.The current seat-reservation systems in library reading rooms do not account for these differences, leading to dissatisfaction among users.A survey of 149 library patrons at K University found that 73% were dissatisfied with the thermal conditions in the library.In this study, we developed a Predicted Mean Vote (PMV) virtual sensor system to provide the important thermal-comfort information required for ensuring optimal seat reservation.Using a Multi-Layer Perceptron (MLP) model, we developed PMV virtual sensors for all seats in the library while measuring the temperature and humidity across the room; a total of 45 models were constructed, with the Coefficient of Variation of the Root-Mean-Square error (cv(RMSE)) being less than 10%.The Computational Fluid Dynamics (CFD) simulations provided the airflow data; the data were incorporated into the PMV calculations.Based on this, the PMV was visualized within an actual seat reservation system to identify and recommend optimal seating, reducing user dissatisfaction by up to 41.7%.This study indicates that providing PMV-based thermal comfort information during seat reservation can effectively reduce occupant dissatisfaction, which in turn can help minimize unnecessary heating and cooling demands and contribute to building energy optimization.