Developing a framework for improvement of building thermal performance modeling under urban microclimate interactions

Abstract

Coupled models developed from the building energy simulation (BES) and computational fluid dynamics (CFD) methods are occasionally used for analyzing the buildings thermal performance. Nevertheless, the large uncertainty in the input parameters of BES models and values of the closure coefficients of Reynolds-averaged Navier-Stokes (RANS) turbulence models restrict the accuracy of coupled BES-CFD models for thermal performance prediction in highly dense urban areas.Thus, a systematic framework for improving the accuracy of the coupled BES-CFD models is proposed in this study, consisting of an approximation technique and a stochastic optimization approach. In this framework, at first, a CFD model is improved with a closure coefficients optimization procedure using experimental data. In the second step, the improved CFD model is utilized to conduct a series of CFD simulations for real-geometry buildings in order to calibrate the BES model with the alteration of parameters such as the adaptive discharge coefficient, local wind profile, and convective heat transfer coefficients over the building façades.The developed framework is then applied to a small cross-ventilated office building surrounded by neighboring buildings. Deviations up to 60% are found in the prediction of the energy saving potential of cross-ventilation strategy by the default and calibrated BES models.