| Title | Calibration of a high enthalpy geothermal reservoir model utilizing micro-seismicity data |
|---|---|
| Authors | Dimitrios Karvounis, Raymi Castilla, Peter Meier, Sigríður Kristjánsdóttir, Vanille Ritz, Antonio Rinaldi, Vala Hjörleifsdóttir, COSEISMIQ Team |
| Year | 2023 |
| Conference | World Geothermal Congress |
| Keywords | micro-seismicity, reservoir modeling, discrete fractures, high-enthalpy, EGS |
| Abstract | Large fractures play a crucial role in the performance and sustainability of geothermal power plants. They make certain regions more preferential for drilling new wells, they host many of the important flow paths in the reservoir that can disrupt power generation, and when large fractures in the reservoir are undetected then risks are high (e.g., financial, seismic, and contamination risks). Diffusion-induced micro-seismicity sheds light on many of these fractures away from the well. The risk of mismanaging these fractures can then be alleviated with a calibrated Discrete Fracture Model (DFM). Calibrating DFMs requires new workflows due to their novelty. The effort is justified because continuum models have limited validity to scenario testing and under-utilize the expensive to collect micro-seismic data. Here, we present our workflow for calibrating an embedded DFM utilizing micro-seismicity data. The workflow uses the embedded DFM of the HFR-Sim simulator, which is the in-house simulator of ETH Zurich for Enhanced Geothermal Systems. The embedded DFM of HFR-Sim is specially designed for dynamically changing fracture networks and probabilistic predictions for diffusion-induced earthquakes. Necessary inputs to the workflow are the detected micro-seismicity data and the corresponding hydraulic logs. A likely representation of the ruptured fracture network is derived through a combination of physics-based modeling and unsupervised machine learning. The size of these fractures and their orientations are the ones that best explain the shape of the micro-seismic cloud and its migration. The fractures derived from the micro-seismic analysis, the deterministic fractures at the wells, and the continuum part of the embedded DFM are calibrated iteratively with successive HFR-Sim simulations, which simulations can take into account a plethora of typical field measurements. We apply the workflow to the Husmuli (Húsmúli) zone in South West Iceland, where wastewater is injected since September 2011, a lot of micro-seismicity has been recorded, and for which several field measurements and tests exist. The zone is part of the high-enthalpy geothermal field Hellisheidi (Hellisheiði) with a power plant that produces 300 MWe and 130 MWth. Important fractures are extracted with the workflow, and are considered by the high-fidelity model of Husmuli. The model is then calibrated and reproduces the total injectivity of the five wells at the end of the considered period and the overall rate of seismicity. |