Proposta de metodologia de predição de sensação térmica dos usuários em ambientes internos

Abstract

The PMV (Predicted Mean Vote) is an index which aims to predict the thermal sensation of people exposed to the same environment. However, there are discrepancies between the PMV model and thermal sensation responses obtained in field studies for some populations. One of the components for the calculation of PMV is the clothing insulation (Icl), which uses the clothing surface temperature (tcl), which can be a factor which contributes towards these discrepancies. Therefore, the aim of this research was to show the tcl influence on the PMV index. Thus, this research aimed to present a new thermal prediction model minimizing inaccuracies of thermal exchanges through the correct determination of tcl, by using Newton's method. Data collection featured a group of welders, a group of office workers performing sedentary activities and a group of Portuguese Army Military. Having collected environmental and personal variables in Brazil and Portugal, this research developed the Snew1, through a value of tcl without residues generated by Newton’s Method and replaced in convection and radiation heat loss equations and Snew2, through multiple regression between thermal sensation votes collected in field study, the metabolic rate and the mechanisms of heat exchange. After confrontation between the real thermal sensation and the calculated PMV values, for all groups, the results found with the Snew1 and Snew2 were always better than the results found with the Fanger’s Original PMV. The best result obtained in this research was with the military group, where the Snew2 presented an improvement about 46% over the original PMV. This search proved that the clothing surface temperature is a variable that influences the PMV model and minimizing inaccuracies in its obtaining decreases discrepancies between thermal sensation votes and PMV.