Influences on the predictive performance of thermal sensation indices

Understanding humans' perception of the thermal built environment is one of the prerequisites for designing satisfactory workplaces and related to their energy use. This paper compares the predictive performance of nine thermal comfort indices with respect to the population mean and individual votes together with the effect of false assumptions of metabolic rate, clothing level, and air velocity on this performance. In addition, the potential to customize the indices to the building level, e.g. by means of the coefficient for psychological adaptation, PSYCH, included in the adaptive thermal heat-balance (ATHB) model is analysed. The analysis of field data from two naturally ventilated and four mixed-mode office buildings showed that the ATHB indices outperform the other indices in terms of predicting thermal sensation. The effect of false assumption is larger for predicted mean vote (PMV)-based models compared with standard effective temperature (SET)-based models. In addition, customization revealed no clear distinction between naturally and mixed-mode buildings, which suggests that other factors are more influential. These results are discussed with respect to the potential of the ATHB approach to account for differences in building control characteristics and the way thermal comfort is addressed in related standards.