Paper Conference

Proceedings of eSim 2020: 11th Conference of IBPSA-Canada

     

Understanding the Influence of Heating/Cooling Modes on Occupant Thermal Comfort Perception

Wooyoung Jung, Farrokh Jazizadeh
Virginia Tech, United States of America

Abstract: Recent studies have investigated personalized thermal comfort modeling using probabilistic and machinelearning-based techniques to enhance the capability of contextualizing human factors for building systems’ operation and simulation. Different modeling schemes and input parameters have been used to create models with varying degrees of accuracy. Inspired by the thermal acclimation process, we have investigated the influence of short-term thermal history on individuals’ thermal comfort votes as a critical variable in the modeling process. To this end, through an experimental study in a controlled environment, 10 human subjects were exposed to two consecutive modes of thermal conditioning with transient temperature configurations – gradually changing from low to high and then high to low temperatures while collecting environmental and human-related data. Through statistical analyses, it was observed that the temperature ranges, for which participants reported thermal comfort satisfaction were significantly different depending on the conditioning modes. Moreover, participants showed to have different sensitivities (reflected in rates of providing response) in these consecutive conditioning modes. These preliminary findings provide an insight into the necessity of tracking short-term thermal history for personalized thermal comfort modeling to enhance control and simulation processes for same environmental conditions complicate the process of inferring individual comfort characteristic patterns and limit the potentials of comfort-aware HVAC operations. This approach requires precise prediction of occupant thermal comfort to enable energy-efficient adjustment of indoor conditions. As synthesized in our recent review article (Jung and Jazizadeh 2019), the accuracy of predicting individual comfort without using any physiological parameters as input into machine-learning algorithms was reported to be between 60 – 70%. In field studies, it has been shown that temperature is the main driving factor with a dominant impact on thermal comfort vote variations (Jazizadeh, Marin et al. 2013). Therefore, in comfort-aware operations, when occupants provide or change their votes, the corresponding ambient temperature is captured and used in developing personalized comfort profiles. However, as noted, given the fact that thermal acclimation property and the short-term thermal history could affect individual comfort perceptions, in this study, we have investigated the impact of this factor on votes reported by human subjects. The findings will provide an insight into considerations for data collection and modeling of personalized comfort predictors. 30 28 energy efficiency.
Paper:
esim2020_1272