Modification and verification of the PMV model to improve thermal comfort prediction at low pressure

Abstract

The human body's thermal physiology changes due to atmospheric pressure, which significantly impacts the perception of thermal comfort. To quantify this effect, an improved version of the Predicted Mean Vote model (PMVp), was developed in this study to predict human thermal sensation under low atmospheric pressure conditions. The study employed environmental conditions of 0km/26°C, 3km/26°C, 4km/26°C, and 4km/21°C. Thirteen subjects were continuously monitored for exhaled CO2, inhaled O2, ambient temperature (ta), relative humidity (RH), air velocity (V), black globe temperature (tg), and altitude (H). The predictive performance of PMVp was evaluated by comparing the experimental results from this study with previous experiments. The findings demonstrate that PMVp exhibits lower root-mean-square errors (RMSE) than the original PMV model. Under the four experimental conditions, the RMSE values for PMVp were 0.311, 0.408, 0.123, and 0.375, while those for PMV were 1.251, 1.367, 1.106, and 1.716, respectively. Specifically, at a temperature range of 21∼27°C (altitude: 941m), the RMSE of PMVp (0.354) was smaller than PMV's. Furthermore, the study analyzed the sensitivity of PMVp to input parameters at an altitude of 4 km. PMVp exhibited considerable sensitivity to the metabolic rate (M) and thermal insulation of clothing (ICL). Consequently, a simple sensitivity scale was established: M>ICL>Ta≈V>Tr>H≈RH, indicating the relative importance of these parameters in influencing PMVp's response. The research findings provide comprehensive knowledge and a useful reference for developing a standard to design and evaluate indoor thermal environments in the plateau region.