Evaluating the capabilities of a simplification algorithm for Urban Building Energy Modeling in performing building-level Multi-Objective Optimizations at district scale

Since Urban Building Energy Modeling and Multi-Objective Optimization require large computational resources, both could benefit of simplification techniques able also to make them more affordable to professionals. In this work, the capabilities of a simplification algorithm for urban scale application are assessed in the context of the optimization of energy efficiency measures for buildings. A group of buildings from different districts were selected along with a set of energy efficiency measures and four objectives, such as heating and cooling needs, thermal comfort, and costs. The performance of the simplification was assessed by using precision, recall and F1-score as metrics to check whether the simplified models could obtain the same optimal solutions as the detailed ones. Overall, the simplification produced adequate results with an F1-score greater than 0.8 for 85 % of the cases considered in all climates, allowing the simulation time to be reduced up to 15.7 times.