A building or a group of buildings may be considered as a set of thermal zones which exchange energy with the environment through the envelopes and systems: walls, thermal bridges, glazing, HVAC. A thermal model of a building (or group of buildings) may thus be represented by a graph where the vertices stand the capacitive nodes (thermal zones and wall meshes) and the edges carry the heat flows. Therefore, it is possible to convert a BIM representation of a building such as a gbXML file into a graph holding the physical laws of the heat flows and heat balances involved.
The aim of this paper is to introduce a novel methodology to generate building energy models (BEM) from BIM digital mock-ups. This new approach consists in creating a graph model using the Python NetworkX library from the available geometric and physical data extracted in the BIM representation. The graph model is used to generate a set of linear invariant systems for numerical simulation assuming linear or linearizable heat fluxes (such as radiative exchanges). In this contribution, the approach is applied to a test case building and validated by comparison with a reference model generated with an already tested tool chain.
Keywords Building energy modeling, BIM, gbXML, graph representation