Modelling heating and cooling energy demand for building stock using a hybrid approach

The building sector accounts for 30% of final energy consumption and 28% global energy-related carbon dioxide emissions, with space heating cooling consuming a large share total buildings’ consumption. Building stock modelling prediction provides critical insights on the aid retrofit policy-making process evaluation energy-saving potential. By combining physical approach data-driven approach, hybrid is applicable stock, including both residential buildings non-residential buildings. Within this framework, Urban Modelling Interface (UMI) tool has been used to generate use intensity. Then, ten different machine learning models, Gaussian radial basis function kernel support vector regression, linear polynomial random forests, extreme gradient boosting, ordinary least-squares ridge least absolute shrinkage selection operator, elastic net artificial neural network, have applied predict intensity (EUI). demonstrated using case study in Chongqing, China. results show that models can achieve accurate EUI prediction, regression showing best accuracy at level single building, performing level. Machine generated by proposed not only provide quickly level, but also decision makings evaluate saving potential various options.