A Study of Time-Dependent Variations in Personal Thermal Comfort via a Dynamic Bayesian Network

摘要

The current practice of defining operational settings for HVAC systems is to use fixed set points, which assumes static and same comfort requirements for building occupants. The majority of the research efforts in the literature study thermal comfort through time-invariant learning algorithms. However, thermal comfort has been shown to vary from person to person, and change over time due to climatic variations or acclimation. In this paper, we present thermal comfort variation results by studying the data from 33 human subjects and statistically evaluate and study the variations to learn similarities and differences among these individuals. In order to quantify the variations, we briefly describe our adaptive stochastic modeling technique. The technique uses a systematic approach for detecting time dependent thermal comfort variations for an individual. The results confirm that personal comfort vary over time (average: 0.061 °C per day). In addition, we observed a high standard deviation (0.159 °C) across the subjects' preference variations.