Worldwide, the adoption of energy efficiency strategies in cities is currently an important concern to reduce the buildings’ energy consumption and corresponding CO2 emissions. To achieve this goal, detailed energy models are required such as urban building energy models (UBEMs). However, the data availability and granularity, the complexity of the urban simulation context and the uncertainty related to the modeling assumptions and outcomes contribute to hinder the development of these models. This brings limitations in the role that UBEMs can play in helping decision makers to perform medium-term analysis towards a sustainable urban planning.
This thesis presents the process of collecting, mapping, cleansing and integrating urban data to support an information system for Smart Cities. The goal is to develop a method that considers the available urban data by integrating and transforming it in the specific information required to run a complete urban building energy simulation. Therefore, different archetype-based approaches are implemented and tested to understand the impact and limitations that several data sources, data coverage and building stock simplification levels have in the accuracy of the UBEM outcomes.
The results show that while data oversimplification can contribute to overestimate the energy consumption, there is also no need to consider a very detailed characterization of the building stock to obtain consistent results. The increased data coverage for the case study area and the probabilistic characterization of building archetypes improved the model predictions, namely at higher disaggregated scales, being able to capture the diversity of the building stock electricity demand. Also, for the energy model validation it was demonstrated the importance of considering different spatial and temporal scales for different error metrics.
Keywords: Urban building database (UBD); Urban building energy modeling (UBEM); Energy efficiency; Building archetypes; Smart cities