Building Simulation: An International Journal

Article Title

Metamodeling of mean radiant temperature to optimize glass facade design in PMV-based comfort controlled space


mean radiant temperature, PMV, thermal comfort control, design of experiments, meta-models, sensitivity analysis


In recent years, glass facades and extensive glassing areas have gained popularity in the built environment. However, thermal comfort and energy-savings in such buildings are still questionable. Advanced comfort-based control strategies have been proposed in order to fulfill a tradeoff between energy-saving and occupants’ thermal comfort. Yet, they could consume energy as the conventional control approach if the building is poorly designed. Thus, an adequate design of building envelope, namely glass facades, is essential to achieve the desired trade-off. The objective of this study is to understand and formulate the relationship between the mean radiant temperature (MRT) and glass facades design parameters in a comfort-controlled space in order to optimize building design for a trade-off between energy savings and thermal comfort. The combined use of numerical simulations, the design of experiments (DoE) technique and an optimization approach is adopted. For the investigations, a previously developed and validated dynamic simulation model is used. The combined use of numerical simulation and DoE aims to identify the significant parameters affecting the MRT, as well as to develop a metamodeling relationship between the considered design factors and MRT. Lastly, the developed meta-models are validated and used to determine a set of optimal solutions using the desirability function approach. The results indicate that the optimized design case allowed about 26% reduction of heating energy consumption compared to the base case design. Finally, the results show that an adequate design of the glazed envelope leads to improved thermal comfort conditions and reduce heating energy consumption.


Tsinghua University Press