Pasar al contenido principal

Forecasting and managing induced seismicity for gigatonne-scale carbon storage

RYC2024-048711-I

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de Recursos Humanos
Subprograma Subprograma Estatal de Formación, Atracción y Retención del Talento Investigador e Innovador
Convocatoria Ramón y Cajal
Año convocatoria 2024
Plan Estatal 2024-2027 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDAD DE LAS ISLAS BALEARES

Resultados relacionados

Resultados totales (Incluyendo duplicados): 3
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Datos de investigación Set de datos (Dataset)

Dataset for Thermoporoelastic flow control strategies for enhancing the longevity of hydraulic fracture-based enhanced geothermal systems

  • Tangirala, Sri Kalyan
  • Kivi, Iman Rahimzadeh
  • Vilarrasa, Víctor
ResFrac is a commercial software and requires a license to use it., [Relationship between files, if important] Folders named no_FCDs and with_FCDs contain the models where no FCDs were used and the models where FCDs were used. The no_FCDs folder contains 6 subfolders, each of which have their respective model files. The with_FCDs folder has 3 subfolders where 135kgps and 150kgps, each have 3 of their own subfolders that contain the model files corresponding to when the FCDs were activated during circulation. The pre existing fracture folder in the with_FCDs folder has its model files where FCDs were activated from the beginning of circulation., This dataset includes the input files of the numerical models used in the manuscript for simulating different cases of the hydraulic fracture-based Enhanced Geothermal Systems (EGS). There are two main folders named ‘no_FCDs’ and ‘with_FCDs’ that contain subfolders corresponding to their different circulation flow rates and pre-existing fracture cases. In the no_FCDs folder, there are 6 subfolders where 5 of them are named with their circulation flow rates (90 kgps, 105 kgps, 120 kgps, 135 kgps and 150 kgps) and another subfolder named ‘pre existing fracture’. The folder named with_FCDs has three folders named ‘135 kgps’, ‘150 kgps’ and ‘pre existing fracture’. The 135 kgps folder has subfolders named ‘from start’, ‘from 21 yr’ and ‘from 24 yr’. The 150 kgps folder has subfolders named ‘from start’, ‘from 13 yr’ and ‘from 16 yr’. Within each subfolder, the model is defined by two text files in a ResFrac-specific keyword format. The 'settings' file and the 'input' file together contain the simulation-specific parameters, including material and formation properties, initial and boundary conditions, the mesh definition (matrix region extent and element sizing), the fluid and relative permeability models, and the well controls and treatment schedule. Because the model was built on a rectilinear grid within the user interface, no separate mesh file is generated and the gridding is contained directly within these input files., S.K.T. and V.V. acknowledge funding from the European Research Council (ERC) under the European Union’s Horizon Europe Research and Innovation Program through the Proof of Concept NOSHAKE under Grant agreement No. 101248689. I.R.K. acknowledges funding from MICIU/AEI /10.13039/501100011033 and from the FSE+ through the ‘Ramón y Cajal’ fellowship (reference RYC2024-048711-I). V.V. acknowledges support from the Spanish Ministry of Science, Innovation and Universities through the project HydroPoreII (PID2022-137652NB-C44). IMEDEA is an accredited "Maria de Maeztu Excellence Unit" (Grant CEX2021-001198, funded by MICIU/AEI/10.13039/501100011033)., With funding from the Spanish government through the "Maria de Maeztu Centre of Excellence" accreditation (CEX2021-001198)., File List: README.txt (this file) no_FCDs input_TE_on_90_fln2_noFCDn_45.txt settings_TE_on_90_fln2_noFCDn_45.txt input_TE_on_105_fln2_noFCDn_45.txt settings_TE_on_105_fln2_noFCDn_45.txt input_TE_on_120_fln2_noFCDn_45y.txt settings_TE_on_120_fln2_noFCDn_45y.txt input_TE_on_135_fln2_noFCDn_45y.txt settings_TE_on_135_fln2_noFCDn_45y.txt input_TE_on_150_fln2_noFCDn_45y.txt settings_TE_on_150_fln2_noFCDn_45y.txt input_TE_on_120_fln2_noFCDn_crack1.txt settings_TE_on_120_fln2_noFCDn_crack1.txt with_FCDs input_TE_on_135_fln2_FCDn0p125_45y.txt settings_TE_on_135_fln2_FCDn0p125_45y.txt input_TE_on_135_fln2_21yFCDn45y.txt settings_TE_on_135_fln2_21yFCDn45y.txt input_TE_on_135_fln2_FCDn_24_45.txt settings_TE_on_135_fln2_FCDn_24_45.txt input_TE_on_150_fln2_FCD0p125_45y.txt settings_TE_on_150_fln2_FCD0p125_45y.txt input_TE_on_150_fln2_13yFCDn0p125_45.txt settings_TE_on_150_fln2_13yFCDn0p125_45.txt input_TE_on_150_fln2_FCDn_16_45.txt settings_TE_on_150_fln2_FCDn_16_45.txt input_TE_on_120_fln2_FCDn_crack.txt settings_TE_on_120_fln2_FCDn_crack.txt These models were simulated using a commercial software ResFrac. In order to run the models, a valid license needs to be taken from ResFrac to use their simulator., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Datos de investigación Set de datos (Dataset)

Dataset for coupled hydromechanical modelling of CO2 leakage through fractured caprocks

  • Vafaie, Atefeh
  • Kivi, Iman Rahimzadeh
  • Makhnenko, Roman Y.
  • Vilarrasa, Víctor
This dataset contains the input files for numerical assessment of the potential for CO2 leakage through a fractured caprock. Numerical models solve for coupled hydromechanical processes, including two-phase flow of CO2 and water, mutual dissolution of the two phases, and rock deformation. Models are two-dimensional and use a plane-strain formulation to represent CO2 injection through a horizontal wellbore into a multilayered sedimentary basin. The geological setting comprises of a sequence of aquifers and aquitards overlying the crystalline basement. A fracture dipping at 60º crosses the caprock directly overlying the storage aquifer, and its base is located at a horizontal distance of 100 m from the injection wellbore. We assume two scenarios depending on the sealing properties of the caprock: (1) a high-seal scenario in which aquitard properties represent those of intact Opalinus Clay, and (2) a low-seal scenario, where the aquitard properties represent those of Eau Claire Shale. Hydromechanical properties of the geological formations and the fracture are taken from experimental measurements [ , ]. The dataset has two folders, each containing the input files for one of the two leakage assessment scenarios mentioned above: “High_seal_scenario”, in which the considered caprocks have low intrinsic permeability k=6·10-21 m2 and high capillary entry pressure p0=4.55 MPa, and “Low_seal_scenario” in which the caprocks have higher permeability k=2.8·10-18 m2 and lower capillary entry pressure p0=0.1 MPa. The power-law exponents for the relative permeability relations and the van Genuchten exponent for the capillary retention curves also differ between the two scenarios following direct experimental measurements. In each folder, there is a file with the name of the folder ending in “_gen.dat” which contains the model input data, including material properties, initial and boundary conditions, and the time intervals. There is also a file ending in “_gri.dat” that contains the information on the mesh. The file “root.dat” includes the name of the model. To run the simulation, execute the CODE_BRIGHT executable “CODE_BRIGHT_2024.exe” in the same folder as the three input files., A.V. and V.V. acknowledge funding from MICIU/AEI /10.13039/501100011033 and from the FSE+ through the ‘Juan de la Cierva’ fellowship (reference JDC2024-053739-I). I.R.K. acknowledges funding from MICIU/AEI /10.13039/501100011033 and from the FSE+ through the ‘Ramón y Cajal’ fellowship (reference RYC2024-048711-I). V.V. acknowledges support from the Spanish Ministry of Science, Innovation and Universities through the project HydroPoreII (PID2022-137652NB-C44)., No
DIGITAL.CSIC. Repositorio Institucional del CSIC
Datos de investigación Set de datos (Dataset)

Dataset, code, and figure outputs for “What controls maximum injection-induced earthquake magnitude?”

  • Sachdeva, Sanchit
  • Kivi, Iman Rahimzadeh
  • Mukuhira, Yusuke
  • Vilarrasa, Víctor
[Description of methods used for collection or generation of data] Selected variables from the open-access GEoREST database were characterized, and rock names were standardized into 12 categories. Missing values were filled using literature values, deterministic calculations, and random-forest regression. Moment_Max was neither imputed nor used for imputation. Two LightGBM models were compared: an injection-only model and a full 12-feature model. Data were divided into 90% training and 10% testing subsets. Predictors were normalized from 0 to 1 using training-set ranges only., [Description of data processing and preparation] 1. Select and filter the GEoREST variables. 2. Fill supported literature values. 3. Calculate deterministic elastic and stress properties. 4. Impute remaining eligible values using random forests. 5. Merge the completed data and train the LightGBM models. Excel colours show provenance: white = original, yellow = literature, orange = random forest, and green = deterministic. Full processing details are provided in the scripts., [Codebooks or data dictionaries] This README documents the final variables. The original GEoREST workbook and Kivi et al. (2023) provide the source-variable definitions., Fluid injection into the subsurface for energy and resource development has led to increased seismic activity in recent years. Understanding the key factors governing maximum induced earthquake magnitude remains critical for their improved prediction and risk assessment. In this study, we apply a machine-learning model using the LightGBM algorithm to a global induced seismicity database spanning multiple geoenergy applications and settings to identify major factors associated with maximum injection-induced moment magnitude. We find that incorporating operational parameters, site characteristics, and host rock properties as input variables, compared to only operational parameters, significantly increases the predictive model accuracy. Cumulative injected fluid volume and injection rate impose first-order controls on seismic magnitude, while rock porosity, permeability, pore pressure, and in-situ stress field are found to play important secondary roles. These results highlight the importance of physical site characteristics beyond injection parameters in enriching magnitude-scaling relations and developing more reliable, physics-based seismicity forecasting and risk mitigation strategies., - The authors acknowledge support for data collection from the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Program through the starting grant GEoREST under grant agreement no. 801809. - S.S. and V.V. acknowledge funding from the European Research Council through the Horizon Europe Proof-of-Concept project NOSHAKE, Grant Agreement No. 101248689. - I.R.K. acknowledges funding from MICIU/AEI/10.13039/501100011033 and the European Social Fund Plus (FSE+) through the Ramón y Cajal fellowship RYC2024-048711-I. - IMEDEA is a María de Maeztu Excellence Unit supported by Grant CEX2021-001198, funded by MICIU/AEI/10.13039/501100011033., With funding from the Spanish government through the "Maria de Maeztu Centre of Excellence" accreditation (CEX2021-001198)., File list Main files - README.md: workflow guide. - README_DATASETS.txt: dataset metadata. - requirements.txt: Python dependencies. - rf_script.py: random-forest imputation. - principal_stress_rf_script.py: principal-stress imputation. - final_dataset_script.py: final data merge and deterministic recalculation. - lightgbm_base_script.py: baseline LightGBM workflow. - lightgbm_normalization_script.py: normalized LightGBM workflow and figures. 01_Data_Curation - GEoREST_induced_seismicity_database_v20.11.2022.xlsx/.csv: source dataset. - DataCuration.xlsx: selected, standardized, literature-filled, and deterministic data. - Literature/RockProperties_4.png to RockProperties_8.png: literature tables used for filling rock properties. 02_RF_Imputation - RF_Datacurated.xlsx: random-forest-imputed data. - Audit_Outputs/principal_stress_values.csv: imputed principal stresses. - Audit_Outputs/principal_stress_validation.csv: stress-imputation validation. 03_Model_Training - Final_Imputed_Dataset.xlsx/.csv: final model input. - Previous_Baseline/Mean_RockSiteInjection.xlsx: earlier comparison dataset. - Results/model_comparison.csv: model-performance summary. - Results/new_v2_all_row_predictions.csv: model predictions. - Results/normalization_ranges.csv: training-set normalization ranges. - Results/normalized_labelled_model_inputs.csv: normalized model inputs. - Results/comparison_report.json: model settings and validation report. - Results/LightGBM_V2_TestV1_model.txt: trained LightGBM model. - Figures/F10.png, F21.jpg, F31.png, and F32.png: manuscript figures., Peer reviewed