Systematic review of the efficacy of data-driven urban building energy models during extreme heat in cities: Current trends and future outlook

Energy demand fluctuations due to low probability high impact (LPHI) micro-climatic events such as urban heat island effect (UHI) and heatwaves, pose significant challenges for urban infrastructure, particularly within urban built-clusters. Mapping short term load forecasting (STLF) of buildings in urban micro-climatic setting (UMS) is obscured by the complex interplay of surrounding morphology, micro-climate and inter-building energy dynamics. Conventional urban building energy modelling (UBEM) approaches to provide quantitative insights about building energy consumption often neglect the synergistic impacts of micro-climate and urban morphology in short temporal scale. Reduced order modelling, unavailability of rich urban datasets such as building key performance indicators for building archetypes-characterization, limit the inter-building energy dynamics consideration into UBEMs. In addition, mismatch of resolutions of spatio–temporal datasets (meso to micro scale transition), LPHI events extent prediction around UMS as well as its accurate quantitative inclusion in UBEM input organization step pose another degree of limitations. This review aims to direct attention towards an integrated-UBEM (i-UBEM) framework to capture the building load fluctuation over multi-scale spatio–temporal scenario. It highlights usage of emerging data-driven hybrid approaches, after systematically analysing developments and limitations of recent physical, data-driven artificial intelligence and machine learning (AI-ML) based modelling approaches. It also discusses the potential integration of google earth engine (GEE)-cloud computing platform in UBEM input organization step to (ⅰ) map the land surface temperature (LST) data (quantitative attribute implying LPHI event occurrence), (ⅱ) manage and pre-process high-resolution spatio–temporal UBEM input-datasets. Further the potential of digital twin, central structed data models to integrate along UBEM workflow to reduce uncertainties related to building archetype characterizations is explored. It has also found that a trade-off between high-fidelity baseline simulation models and computationally efficient platform support or co-simulation platform integration is essential to capture LPHI induced inter-building energy dynamics.