Development of Machine Learning Based Microclimatic HVAC System Controller for Nano Painted Rooms Using Human Skin Temperature

In this paper, a microclimatic data based occupancy regression controller (ORC) is proposed for heating, ventilation, and air conditioning (HVAC) systems and is termed microclimatic HVAC (M-HVAC). Microclimatic data consist of various measurements such as PIR, CO2, humidity, wall, floor and roof temperatures, as well as human skin temperature for the prediction of the optimal thermal setpoint temperature in an M-HVAC system. Microclimatic conditions have a major role in building energy consumption and indoor thermal comfort. Human skin temperature, wall, floor and roof temperatures in the room are obtained through a thermal camera. ORC controller performance is evaluated on the SiO2 nanocoated room walls for high energy savings. Up until now, researchers have focused on the optimization of thermal setpoint temperature (SPT) using indoor air temperature and room occupant count data, but have never addressed the microclimatic conditions. ORC predicts the optimal SPT after including the microclimate data. M-HVAC systems implement the ORC using a Raspberry Pi board connected with sensors and a thermal camera. ORC leads to thermal comfort in a room and reduces energy consumption. ORC improves prediction accuracy through regression analysis and reduces the energy cost of about 23.9% when compared to the traditional method. ORC provides high thermal comfort of about 97% with higher energy savings than the traditional method of temperature setpoint.