Building Energy Prediction for Early-Design-Stage Decision Support: A Review of Data-driven Techniques

Building Performance Simulation (used to assess and validate claims about energy requirement of buildings), is often carried out during later design stages, when it is practically difficult to incorporate design changes. In reality, many design decisions affecting Energy requirement of buildings are taken at an early stage, when adequate information to support such decisions is not available or viable due to lack of tools. This paper presents the suitability of data driven techniques, particularly Artificial Intelligence and Machine learning, for early-design-decision support tools through a systematic review of literature. It highlights their potential due to robust prediction capabilities and forms groundwork for further research to develop an early-design-decision support tool.