@phdthesis{2138924,
  abstract     = {{Night ventilation to cool buildings attracts growing interest. For, it can improve the summer comfort and can lower the cooling need. However, the extent to which building designers succeed in finding an optimal night cooling design depends strongly on the simulation tool they use. Today, stand-alone building energy simulation (BES) programs are quite popular, but the way they model the convective heat transfer raises questions. They model the complex heat transfer in the boundary layer and the surrounding field by a convective heat transfer coefficient which relies primarily on case-specific experimental data. Therefore, this thesis evaluated whether this modelling approach suffices to accurately predict the night cooling performance and further investigated the impact of the room/system design on the convective heat transfer during night cooling. The work began with a literature review which highlighted the limited applicability of the various convection correlations. It appeared that the researchers involved successively developed correlations for distinct cases which had not been studied yet and that a number of people already suggested to categorize all situations into a discrete number of regimes to which specific correlations apply. The subsequent BES-based sensitivity analysis in this thesis indicated that such a pragmatic approach is indeed no luxury. However, as shown in the experimental study in the PASLINK cell, part of this work, it does not enable to investigate the influence of a parameter (value) other than the ones considered in the experimental setup. So, it is necessary to further investigate how room/system design parameters affect the convective heat transfer and eventually refine the BES approach. The second part of the thesis dealt with the way to do this and presented a pilot study on a night cooled landscape office. Surrogate modelling in conjunction with computational fluid dynamics (CFD) would be a valuable supplement to experiments; on condition that CFD users know how to address the inherent error sources. A fully-automated framework of data sampling, geometry/grid generation, CFD solving and surrogate modelling was set up and then deployed to investigate how the convective heat flux in a night cooled landscape office relates to the room/system design. The resulting surrogate models provided rough-hewn insights and, more importantly, the framework can be reused to derive more globally accurate surrogate models which can be coupled with BES.}},
  author       = {{Goethals, Kim}},
  isbn         = {{9789085785156}},
  keywords     = {{heat transfer,design,full-scale experiments,night cooling,BES,surrogate modelling,CFD}},
  language     = {{eng}},
  pages        = {{205}},
  publisher    = {{Ghent University, Faculty of Engineering and Architecture}},
  school       = {{Ghent University}},
  title        = {{Convective heat transfer modelling in offices with night cooling}},
  year         = {{2012}},
}