The energy transition group has a research agenda focused on the development of sustainable and systemic solutions to the energy transition. The foundations for the energy transition are generation of renewable and clean energy and its efficient and well managed used. We target both challenges, with focus on:
- Solar energy, from materials for energy to resource assessment and forecasting, distributed generation, system integration with electric vehicles and long term renewable energy scenarios;
- Energy in buildings, including theoretical and experimental approaches to natural ventilation and building stock energy modelling, applied research and design projects as well as building energy audits;
- System energy and emission analysis including mass and energy flow balances of systems transportation and biorefinery systems.
Group Leader: Miguel Brito
Scientific Highlights 2017
Building stock energy models are predictive tools of the energy consumption of a large group of buildings such as those of a city, a region or a country. Buildings energy performance certification (EPC) collects a massive amount of data on building thermal envelope and energy systems.
A new physics-based building stock model using Monte Carlo was developed and calibrated using EPC data as input. The model calculates the energy indicators. The distribution among classes is successfully reproduced by the model with an error below 3%. The developed model constitutes a reliable tool that helps on further research on energy policies and renewable energy integration, namely, studying the impact evaluation of more restrictive thermal quality requirements; evaluating other methodological approaches to calculate energy indicators; analysing new action plans to promote retrofitting and bottom-up hourly estimations of building stock energy consumption.
Spatially distributed information to improve short-term solar forecasting
Clouds introduce very fast and unpredictable ramps in photovoltaic (PV) generation, which are a challenge for the management of an electrical grid with high penetration of solar energy. PV generation forecasts provides the opportunity to anticipate those ramps and plan mitigation strategies.
This work evaluated a statistical model based on spatially distributed solar information. A network of radiation sensors or PV systems generates information helping models to detect cloud presence and motion, improving forecast accuracy. When data is recorded up to every 1-2 minutes, the layout of the network is very important: placing upwind sensors improves the forecasting skill by a factor of 3; the distance between sensors is related to how farther in the future that information is relevant (since a cloud takes longer to travel from one sensor to another, the more distant they are from each other). Even for longer horizons, spatially distributed forecasting may contribute to the accuracy of the forecasting as it overcomes errors due to local outliers in persistence.
Open Renewables, Portugal
Universidade de Coimbra, Portugal
Universidade do Porto, Portugal
Universidad de Navarra, Spain
Universidad Politécnica de Madrid, Spain
Université de Strasbourg, France
Université Toulouse III Paul Sabatier, France