An investigation of indoor thermal environment in semi-cold region in Japan – Validity of thermal predictive indices in Nagano during the summer season

Abstract

Thermal comfort can be evaluated using two models—heat balance models and adaptive thermal comfort models. A heat balance model is based on the results of laboratory research. Furthermore, the most common predictive index is the Predicted Mean Vote (PMV). However, PMV cannot be used to predict occupants' thermal sensation well in naturally ventilated buildings. In a previous study, it was considered that the prediction accuracy is improved by associating PMV with two-node models or the psychological and behavioral adaptations of occupants. Our aim here was to evaluate the validity of the prediction accuracy of indices. Three predictive indices were compared in terms of occupants' thermal sensation (TSV): PMV, Predicted Mean Vote Gagge (PMVG), which was modified by Gagge, and predicted thermal sensation (TSENS), which is related to the two-node model. In summer, questionnaire-based subjective votes and thermal environment surveys were conducted in a naturally ventilated high school. Regression analysis showed that PMV and TSENS tended to underestimate TSV, whereas PMVG predicted TSV more accurately than PMV and TSENS. TSV was predicted more accurately PMVG, which is based on the two-node model. In addition, the adaptive predicted mean vote (aPMV), which considers occupants’ adaptive abilities, was used to predict TSV. At high temperatures, the predictive accuracy was improved. Finally, the thermal environment of a naturally ventilated classroom was evaluated using an adaptive thermal comfort model. By modifying the upper limit of the acceptable range, we showed the possibility of applying it more accurately in this study.