Indoor thermal comfort based on time scale and habitual trajectory : model modification and experiments

Abstract

Accurately assessing and predicting indoor thermal comfort is essential. However, existing thermal comfort models often fail to be applied to dwellings. These models are designed for single-room settings, while typical dwellings consist of multiple functional spaces. Furthermore, these models evaluate momentary thermal comfort, which cannot represent long-term one. Lastly, thermal comfort experiments require that the subjects’ position is unchanged, overlooking the dynamic nature of their location changes at home, a phenomenon referred to as habitual trajectory here. These discrepancies between thermal comfort models and realistic residential conditions have been unexplored. To address these gaps, this thesis made novel modifications to the mathematical structure of the predicted mean vote (PMV) model, generating the PMVt model. In this model, the effects of temporal variation (time scale) and spatial variation (habitual trajectory) were considered, and the overall thermal comfort was exported. Specifically, model modifications include extending a single-moment, single-location PMV value into a multi-moment, multi-location PMV dataset and then deducing the dataset into a representative value. Moreover, a Python-based tool was developed to streamline the calculation process of the PMVt model. Subsequently, the empirical validation of the PMVt model was conducted in mixed-mode ventilation dwellings in Shanghai during both summer and winter. The experiment invited elderly and younger participants, yielding 1,276 valid thermal sensation votes (TSVs). Participants spent most of their time in the main bedroom and living room, with different purposes across age groups. The thermally neutral temperatures for elderly and younger participants were 25.5 °C and 24.2 °C in summer and 21.4 °C and 20.9 °C in winter, respectively. The elderly displayed a broader comfort range in both seasons, reflecting their unique thermal preferences. Weighted and specified TSV indices were introduced to capture both immediate and delayed thermal responses. The statistical relationships between these indices and the PMVt model demonstrated strong significance. By incorporating time scale and habitual trajectory factors, the PMVt model achieves high accuracy in evaluating and predicting overall thermal comfort. Seasonal analysis revealed that the model performed better in winter, likely because heating preferences are more consistent across age groups, unlike the varied cooling preferences seen in summer. Age analysis further confirmed that the model was suited well for both groups. Gender analysis showed that the model was better for females, as they exhibited greater sensitivity to temperature fluctuations. More importantly, decoupling the combined effects of time scale and trajectory significantly reduced the model's correlation coefficient, dropping from 0.85 to 0.62. This underscores the critical role of both time scale and trajectory in enhancing model accuracy. Momentary or segmented thermal states cannot adequately replace long-term thermal conditions, and simplifying complete trajectories diminishes the model’s effectiveness. This thesis presents the PMVt model as a robust and novel framework for assessing overall thermal comfort, addressing temporal and spatial variations for the first time. The findings emphasize the importance of incorporating both time scale and habitual trajectory into thermal comfort models. These insights would advance residential comfort evaluation methodologies, with significant implications for improving occupant well-being.