Characterization of tight-gas sand reservoirs from horizontal-well performance data using an inverse neural network

Kulga, Burak; Artun, Emre; Ertekin, Turgay

doi:10.1016/j.jngse.2018.08.017

Cited by 17 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The development potential of tight gas reservoirs is considerable as a result of extremely large proven hydrocarbon reserves . Tight sandstone gas reservoirs typically exhibit poor physical properties, which result in problems such as excessive attenuation of formation energy and gas production in conventional development. , CO 2 injection provides a new approach to enhanced gas recovery in tight gas reservoirs as well as being developed for use in long-term geological storage of CO 2 . However, the pore-throat microstructure of tight gas reservoirs is complex, resulting in anfractuous pathways for gas flow that are also constrained by small pore-throat diameters.…”

Section: Introductionmentioning

confidence: 99%

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks

et al. 2022

View full text Add to dashboard Cite

The simultaneous flow of gas and water is controlled by a threshold pressure gradient (TPG) effect during CO 2 injection of tight gas reservoirs. The TPG effect is dynamic because it varies with both the effective stress and the water saturation. The sensitivity of TPG to effective stress and mobile water is affected by the porethroat microstructure. In this paper, we report the results of dynamic TPG tests on six cores with similar permeability. The influence of the pore-throat microstructure on the sensitivity of the TPG to stress and to mobile water was also quantitatively studied using a fractal method, and the distribution of the threshold pressure and corresponding gas production loss were calculated during CO 2 injection in tight gas reservoirs. The test results show that TPG decreases logarithmically with the increase of the pore fluid pressure during CO 2 injection, a change of 0.1−50 MPa in the pore fluid pressure corresponding to a 1.8−3.5 times increase of the TPG variation. The TPG increases exponentially by 3.5−6.7 times from irreducible water saturation to a mobile water saturation of 30%. The fractal dimension (D) of the heterogeneity of the rock pore-throat microstructure has a linear relationship with both the stress sensitivity coefficient (λ) and mobile water sensitivity coefficient (η), with the larger values of λ and η being associated with more heterogeneous pore-throat microstructures. The reservoir threshold pressure showed a significant nonlinear distribution in near-well reservoirs at low bottom-hole flow pressures of the production well during CO 2 injection. The calculated gas well production loss as a result of a dynamic threshold pressure is 7−16% higher than that of the fixed threshold pressure, and the difference is larger at low pressures.

show abstract

Section: Introductionmentioning

confidence: 99%

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Additionally, some optimization methods, such as those from the gradient descent family, need the gradient vector evaluated using expensive finite differentiation. An alternative approach is to construct an interpolation model, e.g., in the form of neural networks, that intakes a reduced form of the experimental data, i.e., force-displacement or stress-strain data, and outputs their corresponding material parameters [15,29,49,39,61,66]. Once trained, these models are extremely fast in performing inference.…”

Section: Introductionmentioning

confidence: 99%

Constitutive model characterization and discovery using physics-informed deep learning

Haghighat¹,

Abouali²,

Vaziri³

2022

Preprint

View full text Add to dashboard Cite

Classically, the mechanical response of materials is described through constitutive models, often in the form of constrained ordinary differential equations. These models have very limited number of parameters, yet, they are extremely efficient in reproducing complex responses observed in experiments. Additionally, in their discretized form, they are computationally very efficient, often resulting in a simple algebraic relation, and therefore they have been extensively used within large-scale explicit and implicit finite element models. However, it is very challenging to formulate new constitutive models, particularly for materials with complex microstructures such as composites. A recent trend in constitutive modeling leverages complex neural network architectures to construct complex material responses where a constitutive model does not yet exist. Whilst very accurate, they suffer from two deficiencies. First, they are interpolation models and often do poorly in extrapolation. Second, due to their complex architecture and numerous parameters, they are inefficient to be used as a constitutive model within a large-scale finite element model. In this study, we propose a novel approach based on the physics-informed learning machines for the characterization and discovery of constitutive models. Unlike data-driven constitutive models, we leverage foundations of elastoplasticity theory as regularization terms in the total loss function to find parametric constitutive models that are also theoretically sound. We demonstrate that our proposed framework can efficiently identify the underlying constitutive model describing different datasets from the von Mises family.

show abstract

Constitutive model characterization and discovery using physics-informed deep learning

Haghighat

Abouali

Vaziri

2023

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

Characterization of tight-gas sand reservoirs from horizontal-well performance data using an inverse neural network

Cited by 17 publications

References 19 publications

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks

Constitutive model characterization and discovery using physics-informed deep learning

Constitutive model characterization and discovery using physics-informed deep learning

Contact Info

Product

Resources

About

Characterization of tight-gas sand reservoirs from horizontal-well performance data using an inverse neural network

Cited by 17 publications

References 19 publications

Quantifying Controls on Threshold Pressure during CO2 Injection in Tight Gas Reservoir Rocks

Quantifying Controls on Threshold Pressure during CO2 Injection in Tight Gas Reservoir Rocks

Constitutive model characterization and discovery using physics-informed deep learning

Constitutive model characterization and discovery using physics-informed deep learning

Contact Info

Product

Resources

About

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks

Quantifying Controls on Threshold Pressure during CO₂ Injection in Tight Gas Reservoir Rocks