The Application of Computational Fluid Dynamics to Comfort Modelling

Abstract

This thesis studies thermal comfort in heating, ventilation and air-conditioning (HVAC) scenarios with computational fluid dynamics (CFD) models at domain and occupant levels. Domain level comfort modelling, where the details of the occupant are not modelled, is investigated utilising Fanger’s Predicted Mean Vote (PMV) and Predicted Percentage of Dissatisfied (PPD) comfort models. Occupant level comfort modelling, where the occupant geometry and skin temperature are required, is explored using two different models. The first model termed the thermal manikin model couples the University of California Berkeley (UCB) psychological model to a new physiological model which neglects the thermal regulation of the human body, and consists of a central core at constant temperature surrounded by a layer with thickness and corresponding thermal properties to allow the skin temperature to vary over the modelled human body. The second model based on Gagge’s two-node model, which includes thermal regulation, yet assumes the skin temperature of the occupant to be spatially uniform. The models are validated with the experimental results from the Technical University of Denmark, which provides the data of the air flow, and the Indoor Environmental Quality (IEQ) laboratory at the University of Sydney, which offered the actual votes of human subjects for a range of environmental conditions. To conclude, the prediction of the skin temperature and its spatial variation is the most important parameter to predict occupant comfort correctly. The occupant level comfort modelling approach employing the thermal manikin is found to be the superior method to evaluate thermal comfort as it can still be accurate when the environment is complex. However, the computational cost and model setup time is high. Further work employing multi-node thermal manikin models would be a fruitful area of research if the accuracy of occupant comfort prediction in complex thermal environments is of interest.