Construction of Thermal Comfort Estimation Model for Individual Air-conditioning System Using Machine Learning

Abstract

Multi-package type air-conditioning systems are installed in various offices. Generally, those systems are operated in such a way that the heating or cooling modes, temperature, and air volume can be set arbitrarily. Even if the room temperatures and air flows are reasonable, the thermal environments are not clear whether it is comfortable for persons in the rooms.

Our goal is to build an automatic control system for individual air conditioning by using IoT devices and machine learning. Therefore, as a first effort, this paper proposes to focus on the prediction performance of machine learning and to build a thermal comfort level estimation model to predict the comfort level of humans in a room.