Using artificial intelligence to optimize geothermal energy production from hot sedimentary aquifers in Western Canada while reducing risk (2024-2025)

Our team will be focusing on (1) developing a coupled THM model with flexible working fluids for geothermal systems using Fourier-GRN and (2) maximizing net present value and managing the geomechanical risks of water-based open-loop geothermal systems by developing a reinforcement-learning based optimization algorithm. Our objective by the end of the project is to implement this data-driven optimization algorithm into a real case study in Estevan area and inform the geothermal system design in WCSB. This data-driven optimization algorithm will significantly improve the economic feasibility of geothermal systems by informing the optimum working fluid, well location, injection and production rate at different geological settings, and ultimately empowering the scaling-up of geothermal systems as the baseload energy supply in Canada and beyond.