JAUE2020-072: A Study on Demand Controlled Ventilation with Vision-based Occupancy Inference for the Korean University Classroom

Abstract

In a university classroom, inadequate ventilation can cause adverse effects such as low learning efficiency, deterioration of health and drowsiness. Ventilation is a major factor affecting the heating and cooling energy consumption of a building. Especially in Korea, ventilation has a great effect on energy consumption due to high heating and cooling load in summer and winter. Demand controlled ventilation (DCV) automatically adjusts the outdoor airflow rate according to the density or preference of occupants. DCV is more energy efficient than constant air volume ventilation because it can prevent over-ventilation. Especially, university classrooms are expected to have a very high potential for energy savings when DCV is applied. However, the university classroom has varying occupancy density and schedule. So, it is necessary to accurately estimate occupancy and to calculate the required outdoor airflow rate for DCV. Vision data can provide highly accurate data about occupants in the building. In this paper, the occupancy inference and ventilation performance of vision-based DCV were evaluated for Korean university classrooms based on actual occupancy schedule. For this end, occupancy density and schedule for two weeks were derived from the videos taken by the CCTV. The target classroom was modeled through the EnergyPlus using the design and measured values. Vision-based DCV strategy that meets the required outdoor airflow rate introduced by ASHRAE Standard 62.1 was configured on the EnergyPlus. Simulation results show that the vision-based approach can provide an acceptable level of indoor concentration of CO2.