skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Machine learning-based prediction of CO2 fugacity coefficients: Application to estimation of CO2 solubility in aqueous brines as a function of pressure, temperature, and salinity
Title: Machine learning-based prediction of CO2 fugacity coefficients: Application to estimation of CO2 solubility in aqueous brines as a function of pressure, temperature, and salinity
Fugacity is a fundamental thermodynamical property of gas and gas mixtures to determine their behavior and dynamics in complex systems. Fugacity can be deduced experimentally from the measurements of volume as a function of pressure at constant temperature or calculated iteratively using analytical equations of states (EOS). Experimental measurement is time-consuming, and analytical models based on EOS are computationally demanding, especially when an approximate but quick estimation is desired. In this work, machine learning (ML) is employed as a viable alternative to analytical EOSs for quick and accurate approximation of CO2 fugacity coefficients. Five different ML algorithms are used to estimate the fugacity coefficients of pure CO2 as a function of pressure (≤ 2000 bar) and temperature (≤ 1000 °C). A combination of experimental and pseudo-experimental (obtained from an analytical EOS) data of CO2 fugacity coefficients is used to train, validate, and test the models. The best results were found using the Extreme Gradient Boosting algorithm, which showed a mean square error of only 0.0002 in the validation data and an average deviation of only 1.3% in the test data (pure prediction). To quantify the effectiveness of the machine learning techniques, results from the best-performing model are compared with two state-of-the-art analytical models. The ML model with significantly less computational complexity showed similar accuracy to the analytical models. The estimated fugacity data are then used to compute the CO2 solubility in aqueous NaCl solution of different concentrations, and a maximum deviation of only 3.2% from the experimental data is observed.  more » « less
Award ID(s):
1946093
PAR ID:
10495368
Author(s) / Creator(s):
; ; ; ;
Publisher / Repository:
ResearchGate
Date Published:
Journal Name:
International Journal of Greenhouse Gas Control
Volume:
128
Issue:
C
ISSN:
1750-5836
Page Range / eLocation ID:
103971
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; (, Volcanica)
    We present DiadFit—an open-source Python3 tool for efficient processing of Raman spectroscopy data collected from fluid inclusions, melt inclusions and silicate melts. DiadFit is optimized to fit the characteristic peaks from CO2 fluids (Fermi diads, hot bands, 13C), gas species such as SO2, N2, solid precipitates (e.g. carbonates), and Ne emission lines with easily tweakable background positions and peak shapes. DiadFit's peak fitting functions are used as part of a number of workflows optimized for quantification of CO2 in melt inclusion vapour bubbles and fluid inclusions. DiadFit can also convert between temperature, pressure and density using various CO2 and CO2-H2O equations of state (EOS), allowing calculation of fluid inclusion pressures (and depths in the crust), conversion of homogenization temperatures from microthermometry to CO2 density, and propagation of uncertainties associated with EOS calculations using Monte Carlo methods. There are also functions to quantify the area ratio of the silicate vs. H2O region of spectra collected on silicate glasses to determine H2O contents in glasses and melt inclusions. Documentation and worked examples are available (https://bit.ly/DiadFitRTD, https://bit.ly/DiadFitYouTube). 
    more » « less
  2. ; ; ; (, 2022 SPE Annual Technical Conference and Exhibition)
    Abstract Geological storage of carbon dioxide (CO2) in depleted gas reservoirs represents a cost-effective solution to mitigate global carbon emissions. The surface chemistry of the reservoir rock, pressure, temperature, and moisture content are critical factors that determine the CO2 adsorption capacity and storage mechanisms. Shale-gas reservoirs are good candidates for this application. However, the interactions of CO2 and organic content still need further investigation. The objectives of this paper are to (i) experimentally investigate the effect of pressure and temperature on the CO2 adsorption capacity of activated carbon, (ii) quantify the nanoscale interfacial interactions between CO2 and the activated carbon surface using Monte Carlo molecular modeling, and (iii) quantify the correlation between the adsorption isotherms of activated carbon-CO2 system and the actual carbon dioxide adsorption on shale-gas rock at different temperatures and geochemical conditions. Activated carbon is used as a proxy for kerogen. The objectives aim at obtaining a better understanding of the behavior of CO2 injection and storage into shale-gas formations. We performed experimental measurements and Grand Canonical Monte Carlo (GCMC) simulations of CO2 adsorption onto activated carbon. The experimental work involved measurements of the high-pressure adsorption capacity of activated carbon using pure CO2 gas. Subsequently, we performed a series of GCMC simulations to calculate CO2 adsorption capacity on activated carbon to validate the experimental results. The simulated activated carbon structure consists of graphite sheets with a distance between the sheets equal to the average actual pore size of the activated carbon sample. Adsorption isotherms were calculated and modeled for each temperature value at various pressures. The adsorption of CO2 on activated carbon is favorable from the energy and kinetic point of view. This is due to the presence of a wide micro to meso pore sizes that can accommodate a large amount of CO2 particles. The results of the experimental work show that excess adsorption results for gas mixtures lie in between the results for pure components. The simulation results agree with the experimental measurements. The strength of CO2 adsorption depends on both surface chemistry and pore size of activated carbon. Once strong adsorption sites within nanoscale network are established, gas adsorption even at very low pressure is governed by pore width rather than chemical composition. The outcomes of this paper provides new insights about the parameters affecting CO2 adsorption and storage in shale-gas reservoirs, which is critical for developing standalone representative models for CO2 adsorption on pure organic carbon. 
    more » « less
  3. ; (, SPE)
    Abstract Large-scale geo-sequestration of anthropogenic carbon dioxide (CO2) is one of the most promising methods to mitigate the effects of climate change without significant stress on the current energy infrastructure. However, the successful implementation of CO2 sequestration projects in suitable geological formations, such as deep saline aquifers and depleted hydrocarbon reservoirs, is contingent upon the optimal selection of decision parameters constrained by several key uncertainty parameters. This study performs an in-depth parametric analysis of different CO2 injection scenarios (water-alternating gas, continuous, intermittent) for aquifers with varying petrophysical properties. The petrophysical properties evaluated in this study include aquifer permeability, porosity, relative permeability, critical gas saturation, and others. Based on the extensive data collected from the literature, we generated a large set of simulated data for different operating conditions and geological settings, which is used to formulate a proxy model using different machine learning methods. The injection is run for 25 years with 275 years of post-injection monitoring. The results demonstrated the effectiveness of the machine learning models in predicting the CO2 trapping mechanism with a negligible prediction error while ensuring a low computational time. Each model demonstrated acceptable accuracy (R2 >0.93), with the XGBoost model showing the best accuracy with an R2 value of 0.999, 0.995, and 0.985 for predicting the dissolved, trapped, and mobile phase CO2. Finally, a feature importance analysis is conducted to understand the effect of different petrophysical properties on CO2 trapping mechanisms. The WAG process exhibited a higher CO2 dissolution than the continuous or intermittent CO2 injection process. The porosity and permeability are the most influential features for predicting the fate of the injected CO2. The results from this study show that the data-driven proxy models can be used as a computationally efficient alternative to optimize CO2 sequestration operations in deep saline aquifers effectively. 
    more » « less
  4. ; ; ; ; ; ; ; ; (, NOAA National Centers for Environmental Information)
    This dataset includes surface underway chemical, meteorological and physical data collected from Autonomous Surface Vehicle (ASV) Saildrone 1038 (EXPOCODE 316420220616) in the Indian Ocean, Southern Ocean from 2022-06-16 to 2022-07-26. These data include xCO2 SW (wet) - mole fraction of CO2 in air in equilibrium with the seawater at sea surface temperature and measured humidity; H2O SW - Mole fraction of H2O in air from equilibrator; xCO2 Air (wet) - Mole fraction of CO2 in air from airblock, 0.67m (26") above the sea surface at measured humidity; H2O Air - Mole fraction of H2O in air from airblock, 0.67m (26") above the sea surface; Atmospheric pressure at the airblock, 0.67m (26") above the sea surface; Atmospheric pressure at the airblock, 0.67m (26") above the sea surface; Temperature of the Infrared Licor 820 in degrees Celsius; MAPCO2 %O2 - The percent oxygen of the surface seawater divided by the percent oxygen of the atmosphere at 0.67m (26") above the sea surface; Sea Surface Temperature; Sea Surface Salinity; xCO2 SW (dry) - Mole fraction of CO2 in air in equilibrium with the seawater at sea surface temperature (dry air); xCO2 Air (dry) - Mole fraction of CO2 in air at the airblock, 0.67m (26") above the sea surface (dry air); fCO2 SW (sat) - Fugacity of CO2 in air in equilibrium with the seawater at sea surface temperature (100% humidity); fCO2 Air (sat) - Fugacity of CO2 in air at the airblock, 0.67m (26") above the sea surface (100% humidity); dfCO2 - Difference of the fugacity of the CO2 in seawater and the fugacity of the CO2 in air (fCO2 SW - fCO2 Air); pCO2 SW (wet) - Partial Pressure of CO2 in air in equilibrium with the seawater at sea surface temperature (100% humidity); pCO2 Air (wet) - Partial Pressure of CO2 in air at the airblock, 0.67m (26") above the sea surface (100% humidity); dpCO2 - Difference of the partial pressure of CO2 in seawater and air (pCO2 SW - pCO2 Air; pH of Seawater (total scale). The Autonomous Surface Vehicle CO2 (ASVCO2) instruments used to collect these data include Bubble type equilibrator for autonomous carbon dioxide (CO2) measurement, Carbon dioxide (CO2) gas analyzer, Humidity Sensor, and oxygen meter. 
    more » « less
  5. ; ; ; ; ; ; ; ; (, NOAA National Centers for Environmental Information)
    This dataset includes surface underway chemical, meteorological and physical data collected from Autonomous Surface Vehicle (ASV) Saildrone 1039 (EXPOCODE 316420220901) in the Indian Ocean, Southern Ocean from 2022-09-01 to 2023-04-27. These data include xCO2 SW (wet) - mole fraction of CO2 in air in equilibrium with the seawater at sea surface temperature and measured humidity; H2O SW - Mole fraction of H2O in air from equilibrator; xCO2 Air (wet) - Mole fraction of CO2 in air from airblock, 0.67m (26") above the sea surface at measured humidity; H2O Air - Mole fraction of H2O in air from airblock, 0.67m (26") above the sea surface; Atmospheric pressure at the airblock, 0.67m (26") above the sea surface; Atmospheric pressure at the airblock, 0.67m (26") above the sea surface; Temperature of the Infrared Licor 820 in degrees Celsius; MAPCO2 %O2 - The percent oxygen of the surface seawater divided by the percent oxygen of the atmosphere at 0.67m (26") above the sea surface; Sea Surface Temperature; Sea Surface Salinity; xCO2 SW (dry) - Mole fraction of CO2 in air in equilibrium with the seawater at sea surface temperature (dry air); xCO2 Air (dry) - Mole fraction of CO2 in air at the airblock, 0.67m (26") above the sea surface (dry air); fCO2 SW (sat) - Fugacity of CO2 in air in equilibrium with the seawater at sea surface temperature (100% humidity); fCO2 Air (sat) - Fugacity of CO2 in air at the airblock, 0.67m (26") above the sea surface (100% humidity); dfCO2 - Difference of the fugacity of the CO2 in seawater and the fugacity of the CO2 in air (fCO2 SW - fCO2 Air); pCO2 SW (wet) - Partial Pressure of CO2 in air in equilibrium with the seawater at sea surface temperature (100% humidity); pCO2 Air (wet) - Partial Pressure of CO2 in air at the airblock, 0.67m (26") above the sea surface (100% humidity); dpCO2 - Difference of the partial pressure of CO2 in seawater and air (pCO2 SW - pCO2 Air; pH of Seawater (total scale). The Autonomous Surface Vehicle CO2 (ASVCO2) instruments used to collect these data include Bubble type equilibrator for autonomous carbon dioxide (CO2) measurement, Carbon dioxide (CO2) gas analyzer, Humidity Sensor, and oxygen meter. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email