Prediction of CO2 diffusivity in brine using white-box machine learning

Amar, Menad Nait; Ghahfarokhi, Ashkan Jahanbani

doi:10.1016/j.petrol.2020.107037

Cited by 39 publications

(10 citation statements)

References 69 publications

(95 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Errors occurred during experimental measurements are the main sources of such data points. Existence of some suspected data is unavoidable when analyzing a large number experimental data from various sources 135,[146][147][148] . In this study, the graphical technique of William's plot was employed to detect the probable suspected data.…”

Section: Resultsmentioning

confidence: 99%

Intelligent modeling of hydrogen sulfide solubility in various types of single and multicomponent solvents

Moradkhani

Hosseini

Ranjbar

et al. 2023

Sci Rep

View full text Add to dashboard Cite

This study aims to study the solubility of acid gas, i.e., hydrogen sulfide (H2S) in different solvents. Three intelligent approaches, including Multilayer Perceptron (MLP), Gaussian Process Regression (GPR) and Radial Basis Function (RBF) were used to construct reliable models based on an extensive databank comprising 5148 measured samples from 54 published sources. The analyzed data cover 95 single and multicomponent solvents such as amines, ionic liquids, electrolytes, organics, etc., in broad pressure and temperature ranges. The proposed models require just three simple input variables, i.e., pressure, temperature and the equivalent molecular weight of solvent to determine the solubility. A competitive examination of the novel models implied that the GPR-based one gives the most appropriate estimations with excellent AARE, R2 and RRMSE values of 4.73%, 99.75% and 4.83%, respectively for the tested data. The mentioned intelligent model also performed well in describing the physical behaviors of H2S solubility at various operating conditions. Furthermore, analyzing the William's plot for the GPR-based model affirmed the high reliability of the analyzed databank, as the outlying data points comprise just 2.04% of entire data. In contrast to the literature models, the newly presented approaches proved to be applicable for different types of single and multicomponent H2S absorbers with AAREs less than 7%. Eventually, a sensitivity analysis based on the GPR model reflected the fact that the solvent equivalent molecular weight is the most influential factor in controlling H2S solubility.

show abstract

Section: Resultsmentioning

confidence: 99%

Intelligent modeling of hydrogen sulfide solubility in various types of single and multicomponent solvents

Moradkhani

Hosseini

Ranjbar

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Numerous studies have been performed to construct the relations between the CO 2 solubility with those parameters that would impact solubility trapping (i.e. diffusivity 196 , oil/gas -brine interfacial tension (IFT) 197 , etc.). The solubility of CO 2 in the oil phase is generally higher than that of brine in mature oil reservoirs 191 .…”

Section: Co 2 Storagementioning

confidence: 99%

“…Research was carried out to study the CO 2 /oil/brine interactions under subsurface conditions. Amar and Ghahfarokhi 196 established the correlation between diffusivity coefficients of the CO 2 in brine water with pressure, temperature and the viscosity of the solvent using the group method of data handling (GMDH) and gene expression programming (GEP). GMDH is one type of ANN that can generate an explicit expression for the correlation between inputs and output.…”

Section: Co 2 Storagementioning

confidence: 99%

Harnessing the power of machine learning for carbon capture, utilisation, and storage (CCUS) – a state-of-the-art review

Yan

Borhani

Subraveti

et al. 2021

Energy Environ. Sci.

124

View full text Add to dashboard Cite

show abstract

“…An overwhelming amount of research is focused on using machine learning (ML) to improve the efficiency and accuracy of subsurface energy-related fluid-flow applications 13 . Machine learning has been used to create reduced order models for geologic CO 2 sequestration [14][15][16][17] , CO 2 enhanced oil recovery, geothermal energy [18][19][20][21] , geothermal energy [22][23][24][25] , and oil and gas extraction 26,27 . For each of these applications, an effective pressure management strategy is required to mitigate the risks associated with injection/extraction operations 1, 10-12, 28, 29 .…”

Section: Introductionmentioning

confidence: 99%

Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management

Pachalieva

O’Malley

Harp

et al. 2022

Preprint

View full text Add to dashboard Cite

Avoiding over-pressurization in subsurface reservoirs is critical for applications like CO2 sequestration and wastewater injection. Managing the pressures by controlling injection/extraction are challenging because of complex heterogeneity in the subsurface. The heterogeneity typically requires high-fidelity physics-based models to make predictions on CO2 fate. Furthermore, characterizing the heterogeneity accurately is fraught with parametric uncertainty. Accounting for both, heterogeneity and uncertainty, makes this a computationally-intensive problem challenging for current reservoir simulators. To tackle this, we use differentiable programming with a full-physics model and machine learning to determine the fluid extraction rates that prevent over-pressurization at critical reservoir locations. We use DPFEHM framework, which has trustworthy physics based on the standard two-point flux finite volume discretization and is also automatically differentiable like machine learning models. Our physics-informed machine learning framework uses convolutional neural networks to learn an appropriate extraction rate based on the permeability field. We also perform a hyperparameter search to improve the model’s accuracy. Training and testing scenarios are executed to evaluate the feasibility of using physics-informed machine learning to manage reservoir pressures. We constructed and tested a sufficiently accurate simulator that is 400 000 times faster than the underlying physics-based simulator, allowing for near real-time analysis and robust uncertainty quantification.

show abstract

Prediction of CO2 diffusivity in brine using white-box machine learning

Cited by 39 publications

References 69 publications

Intelligent modeling of hydrogen sulfide solubility in various types of single and multicomponent solvents

Intelligent modeling of hydrogen sulfide solubility in various types of single and multicomponent solvents

Harnessing the power of machine learning for carbon capture, utilisation, and storage (CCUS) – a state-of-the-art review

Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management

Contact Info

Product

Resources

About