Automated conversion of architectural massing models into thermal ‘shoebox’ models

Many attempts have been made to automatically convert architectural 3D models into thermal models for building performance simulation. This paper describes a method that is capable of abstracting an arbitrary building massing into a meaningful group of thermal shoebox models. The algorithm is meant to bridge the existing gap between architectural and thermal representations of the same building and to facilitate the use of energy models during schematic design by providing instant performance feedback from the massing stage onwards. The method uses varying facade insolation levels as the key form-related parameter. Discrete facade segments are then grouped by similarity of their local “solar microclimate”. Each group is represented by a reference shoebox model, which consists of a two-zone thermal model for perimeter and core regions. Computed shoebox results are then extrapolated and mapped back to the architectural model. Thus, the relationship between the simulation output and the provided architectural geometry is strengthened and easier to communicate. Combined with a parametric modeling environment, the method may be used to identify optimized local massing solutions. It can also be applied at the urban level to break down a whole neighborhood into a representative subset of simple thermal models, allowing the estimation of urban energy use intensity in a feasible and timely manner.