Thermal comfort modeling in transient conditions using real-time local body temperature extraction with a thermographic camera

Abstract

This work evaluated the use of thermographic cameras as a non-invasive method to automatically model human thermal comfort in transient conditions, using data from 30 healthy subjects tested in an office setup with ambient temperatures between 21.11 °C and 27.78 °C. Office temperature, relative humidity, exposed skin temperature and clothing temperature were automatically measured over approximately 27 min per subject, using remote sensors and avoiding any contact with the subjects. Thermal comfort levels were evaluated using subjects feedback, recorded every minute for the entire experiment. Clothing insulation and metabolic rate were kept relatively constant for this experiment (0.54 clo and 1.1 met). Average skin temperature was extracted from five different locations, with average temperatures of 33.5 °C, 34.5 °C, and 35.6 °C corresponding to cold discomfort, comfort and warm discomfort respectively. Average clothing temperature was also extracted from three different location, with 32.3 °C, 33.8 °C and 35.0 °C corresponding to the same three comfort levels. Relative humidity levels were similar for all subjects, with average values between 38% and 33%. Results showed significant correlation between observed skin temperature, clothing temperature and thermal comfort level. Also, collected data showed that the temperature difference between different body locations was highly correlated with thermal comfort, and the variance of skin temperature over a small area was significantly correlated with thermal comfort. The results suggest that non-invasive thermographic cameras that combine visual and thermal modes are sufficiently accurate in real-world settings to drive control of HVAC systems.