Extending Predicted Mean Vote using adaptive approach

Abstract

Accurate prediction of thermal comfort is the core of thermal comfort and energy efficiency of buildings. The rational approach (e.g. PMV - Predicted Mean Vote) is used prevailingly for thermal comfort prediction, but could result in large errors because of its deficiency in explaining thermal adaptations. To improve the rational approach for thermal comfort prediction, this study proposes a method to extend the PMV to fully account for thermal adaptations. The extended PMV proposed is derived by multiplying an extension factor to the PMV. The extension factor is variable in the function of the ambient temperature according to the adaptive approach. With the adaptive approach, the proposed extended PMV can account for all categories of thermal adaptations (i.e., psychological, physiological and behavioral adaptations) and dynamic characteristics of thermal adaptations. Case studies on naturally-ventilated buildings show that the root mean square errors of the PMV, the original extended PMV, and the proposed extended PMV are 1.05, 0.74, and 0.26 scales respectively; and case studies on air-conditioned buildings show that the root mean square errors of the PMV, the new PMV and the proposed extended PMV are 1.24, 0.95, and 0.13 scales respectively. Thus, compared with the existing methods for thermal comfort prediction, the proposed extended PMV improves the accuracy by at least 65%. Besides the improved accuracy, the proposed extended PMV is friendly to practical applications because the extension factor is explicitly formulated, thereby contributing to the updating of thermal comfort standards and the development of thermally comfortable and energy-efficient buildings.