Data-driven research into the inaccuracy of traditional models of thermal comfort in offices

• DOI: 10.1016/j.buildenv.2023.111104
• Publication Date: 2023-12-09T23:01:34Z
• Authors: Rosana Caro; María Dolores Red...; Arturo Martínez; Elena Cuerda; María del Mar Bar...; Javier Neila; Jorge Aguillón-Ro...; Carlos Renato Ram...
• Keywords: Variables; Predictive modelling; Thermal sensation

The accurate prediction of thermal sensation among office workers, at design and post-occupancy stages, is crucial for controlling indoor temperature efficiently correcting deficiencies in workspaces, ensuring healthy productive working conditions. Traditional analytical comfort models are still the best tool this purpose given their potential interpretation. However, reliability undermined by poor accuracy. Based on 304 data series point-in-time measurements quantitative qualitative comfort-related parameters collected an experimental campaign three buildings, one air-conditioned two free evolution, San Luis Potosí (Mexico), work aims to identify major error-causing factors steady adaptive models. divergences between predicted reported were set as a dependant variable multiple regressions, each model. Eighteen independent demographic, environmental, contextual subjective variables considered. No multicollinearity problems identified. Our findings show that humidity perception relevant model error. Clothing insulation highly impacted accuracy both while age body mass not statistically significant either them. Metabolic rate was factor with greatest influence error Although covered, other influential played key role models' further research needed integrate these new generation more flexible