Information exchange scenarios between machine learning energy prediction model and BIM at early stage of design

ABSTRACT The building design process incorporates various analysis activities for design space exploration. The need of sustainable built-facility has made energy efficiency an important factor through building life-cycle. Building information modelling (BIM) facilitates energy analysis by reducing re-modelling efforts to create energy model. However, the lack of information makes energy prediction a challenging task in the early design phase with a deterministic approach. The research work analyses various information exchange scenarios at different levels of detail (LOD) that link to an approach of machine learning energy prediction model with BIM data. At any level of detail, information is distinguished by the labels “available”, “developing” and “unknown”. Monte Carlo method will be used to generate samples of energy analysis for unknown information. The uncertainty of energy prediction is represented by mean, maximum and minimum values of heating load. The research will be useful for design space exploration at the early stage of design.