Forecasting Induced Seismicity and Maximizing Production of Electricity in EGS

Abstract

Enhanced Geothermal Systems (EGS) has been for decades considered to be one of the best candidate technologies for exploiting the abundant geothermal energy stored in non-volcanic regions and their further development is now eminently needed for decarbonizing our society. Induced seismicity due to the stimulation of EGS reservoirs has been a major setback for the development of the technology since investors need not only to consider the risk of drilling an underperforming well, but also the seismic hazard and its social aspects. The governmentality of EGS is further restricted by the limited experience from successful EGS stimulations, the large number of possible scenarios that need to be considered, and by the complexity of the analytical models that best describe the physical processes during stimulation. Here, a Discrete Fracture Hybrid Model (DFHM) is employed for numerical studies, which DFHM combines deterministic numerical modeling of flow inside discrete fractures with stochastic modeling of seismicity. Given a stimulation strategy for the wells of an EGS, the DFHM returns in time for real-time applications both forecasts of induced seismicity that are useful for Probabilistic Seismic Hazard Assessments (PSHA) and forecasts of the maximum expected electrical power generation that is useful for Probabilistic Reservoir Performance Assessments (PRPA). Inference, prediction, probabilistic optimization, and uncertainty quantification can be performed with the DFHM for several scenarios related to induced seismicity in unconventional fractured reservoirs, for which scenarios conventional PSHA approaches fail due to their simplifications. The concept of multi-staged soft stimulation is tested with the DFHM, where both PSHA and PRPA are performed and compared with past single-staged strategies.