Digital Twin For Direct-Use Geothermal Project On The TU Delft Campus
Please login to view abstract download link
More than 60% of the final energy demand by residential households in Europe is related to heating. The vast majority of this demand is currently supplied by natural gas, with direct-use geothermal energy being able to completely replace it, reducing carbon emissions to low or negligible levels. A geothermal well doublet has been drilled at the TU Delft campus for research purposes and direct-use heat supply. The research will be dedicated to the optimal use of geothermal resources and include investigation of natural and engineered materials, monitoring techniques, and production optimization in the presence of uncertainties. The project includes a comprehensive research program, involving the installation of a wide range of geophysics instruments at the surface and in the wellbore alongside an extensive logging and coring program. To address the main research question, a digital twin framework will be designed and linked with all observations including dynamic measurements in the wells using fibre-optic sensing. All observations will be included in the digital-twin framework which will allow us to assimilate observations, reduce uncertainties, and optimize the production schemes. This paper presents the initial attempt in the development of a digital twin for our geothermal project. The geological concept includes conventional geological models together with sketch-based geological scenarios allowing us to address the high uncertainty of the data. Simulation results include rigorous evaluation based on high-fidelity representation and fast proxy models based on Machine Learning. Different data-assimilation techniques are utilized in our digital-twin framework to calibrate the model with real observations. While our study is still at the initial stage, we will show two examples where the proposed framework is utilized to address several operational challenges.
