AI  for geosciences

Brief project description: Hydrothermal energy is seen as one of the main players for the transition to renewable energies worldwide. A key interest of geothermal plant operators is the long-term sustainability of the produced water volume and the difference between production and injection temperature. This requires the understanding of the behavior of the thermalfluids at any time. To manage that, samples on 3 plant sites, hardware and software and the power of modelling is invented and used. The goal is to develop a digital twin of a hydrothermal plant, which allows an early prediction of potential disturbing processes in the thermal system and a prediction tool for hydrochemical and operational processes during the change of operational parameters. This is necessary to enhance the heat exploitation, the capacity and economic situation of geothermal plants. The digital twin will be based on hydrochemical models, which are coupled with operational parameters and machine learning tools. Based on the artificial intelligence (AI) approaches, the twin will be able to train itself. The Austrian partners are involved in the development, set-up, validation and operation process of the machine learning tools (Geosaic in cooperation with KIT(Ger)), and sampling and assistance of demonstrator experiments on the Austrian site (hydroFilt).