Digital Twin of Urban Energy Systems based on Machine Learning
Master project
Fall 2024
In the context of the energy transition, policymakers generally rely on energy system models to guide decision-making. As these tools become increasingly complex to take account of the multidimensionality of the energy system, it becomes difficult to communicate the results of studies effectively to stakeholders. The development of digital twins aims to bridge this communication gap by building a machine learning interface for the models, where stakeholders can specify their needs and explore the space of available solutions. As part of this master’s project, the student will carry out the following tasks:
- Use the REHO open-source tool to generate energy community solutions contextualized in the Swiss context (connection to grids, biomass potential, geothermal accessibility, solar potential)
- Centralize the solutions obtained in a database of solutions
- Build a machine learning model based on the solutions generated, measure the accuracy of the digital twin
- Initiate and develop a user interface
The main outcome of this master project is the development of a fast-responsive interface with visualization of the results. A good knowledge in python is required.
The project will be supervised by Dr. Bingqian Liu (Postdoc IPESE) and Dr. Eduardo Pina (Postdoc IPESE). If interested, please send your CV, with a short motivation letter, to bingqian.liu@epfl.ch.