A novel model for uncertainty propagation analysis applied for human thermal comfort evaluation

The comfort sensation is mainly affected by six variables: air temperature, mean radiant temperature, air velocity, relative humidity, personal metabolism and clothing insulation. These are characterized by different mean values and distributions. To analyze the uncertainty propagation three numerical models are used: the Fully Monte Carlo Simulation MCSs, the Monte Carlo Simulation Trials MCSt, and a novel model named "Adaptive Derivative based High Dimensional Model Representation" (AD-HDMR). In the paper these three different methods are applied to the thermal comfort evaluation, through the PMV Index, they are analyzed and their efficiency was verified in terms of computational time. To allow a revision of this index, the effect of the different variables was then analyzed.