Modification of sweat evaporative heat loss in the PMV/PPD model to improve thermal comfort prediction in warm climates

Deficient performance of Predicted Mean Vote (PMV) in estimating thermal sensation in warm environments has always been considered as one of the major drawbacks of this model. This well-known shortcoming is often attributed to PMV overlooking thermal adaptation, but simplistic treatment of thermoregulatory sweating may also partly contribute to this error. In this paper a new approach to estimating sweating heat loss in the PMV model is proposed based on a piecewise fuzzy regression model. The effect of sweating at low physical activity levels and the impact of environmental parameters on whole-body evaporative heat loss are also accounted for. The new modified PMV model is validated with previously published experimental data. The result of our analysis demonstrated that the proposed model performs better than the original PMV model. The modified PMV model predicted thermal sensations more accurately than the original version of the PMV model, even without the expectancy factor being applied. This implies that a considerable part of the PMV error is associated with the simplification of sweating calculations. Ignoring the structural drawback of the original PMV calculations can lead to over-estimation of adaptation (or expectancy) coefficients when adjusting the PMV model.