Abstract/Summary

Human thermal comfort is affected by various interactive variables, revealing the thermal adaptation processes is challenging. Simplifying as a single direct step from triggering factors to assessment, the understanding of interactions and causality among explanatory variables and their link to thermal comfort remains insufficiently explored. An Interaction Matrix-based Path Analysis (IMPA) modeling approach was proposed to examine the direct and indirect effects of variables on thermal comfort and modeling the thermal comfort by combining observed and unobserved factors. To verify the approach, a broader range of variables was investigated in field studies in five climate zones of China. The Back Propagation-Artificial Neural Network (BP-ANN) coding-based interaction matrix described the possible interaction pathways between variables. Based on these interaction pathways and thermal adaptation theory, the results indicated eleven dominant hypotheses with the directed connections. The path analysis modeling method quantified the driving effects and causality between the explanatory variables and thermal sensation under various indoor conditions. It concluded that psychological factors directly affected thermal sensation, while physiological factors displayed an indirect relationship. Environmental and behavioral factors had both direct and indirect effects. Environmental factors contributed the most significant total effects on thermal sensation, followed by psychological and behavioral factors in various environments. The physiological factors had no substantial impact in a neutral environment. The observed variables affecting thermal sensation further underscored the importance of indoor air temperature and thermal expectation. This study could provide new insights into describing the direct and indirect pathways and understanding the thermal adaptation process.