Summary In this work, physics-informed neural networks (PINNs) are used for history matching data from core-scale countercurrent spontaneous imbibition (COUCSI) tests. To our knowledge, is the first work exploring variation in saturation function solutions COUCSI 1D flow was considered, which two phases opposite directions driven by capillary forces with one boundary open to flow. The partial differential equation (PDE) depends only on a saturation-dependent diffusion coefficient (CDC)....

Abstract In this study, the impacts of solutal Marangoni phenomenon on multiphase flow in static and micromodel geometries have experimentally been studied interactions between oil droplet two different alkaline solutions (i.e. MgSO 4 Na 2 CO 3 ) were investigated. The tests revealed that convection exists presence which should carefully be considered porous media. experiments, observations showed flooding, fluids stayed almost stationary, while a spontaneous movement was detected. changes...

The application of physics-informed neural networks (PINNs) is investigated for the first time in solving one-dimensional countercurrent spontaneous imbibition (COUCSI) problem at both early and late (i.e., before after front meets no-flow boundary). We introduce utilization Change-of-Variables as a technique improving performance PINNs. formulated COUCSI three equivalent forms by changing independent variables. describes saturation function normalized position X T; second Y = T0.5; third...

We propose a workflow based on physics-informed neural networks (PINNs) to model multiphase fluid flow in fractured porous media. After validating the forward and inverse modeling of synthetic problem media, we applied it real experimental dataset which brine is injected at constant pressure drop into CO2 saturated naturally shale core plug. The exact spatial positions natural fractures dynamic in-situ distribution fluids were imaged using CT-scan setup. To targeted system, followed domain...

In recent years, the gap between Deep Learning (DL) methods and analytical or numerical approaches in scientific computing is tried to be filled by evolution of Physics-Informed Neural Networks (PINNs). However, still, there are many complications training PINNs optimal interleaving physical models. Here, we introduced concept Physical Activation Functions (PAFs). This offers that instead using general activation functions (AFs) such as ReLU, tanh, sigmoid for all neurons, one can use...

There are several approaches for the calculation of capillary pressure curves in porous media including centrifuge method. In this work, a new installation test is introduced and compared with traditional setup. first setup, which standard approach labs, core face closest to rotational axis open non-wetting phase, while farthest wetting phase where strictly co-current flow generated rotations; labeled Two-Ends-Open (TEO). second proposed as approach, only outer radius surface exposed light...

Spontaneous imbibition is the main oil production mechanism in water invaded zone of a naturally fractured reservoir (NFR). Different scaling equations have been presented literature for upscaling core scale recovery curves to field matrix blocks. Various dependent parameters such as gravity effects and boundary influences are required be considered process. Fluid flow from blocks fracture system highly on permeability value horizontal vertical directions. The purpose this study include...

Polymer flooding is a popular enhanced oil recovery method because of its impact on improvement sweep efficiency. Polymers are non-Newtonian fluids with different behavior at flow rates. At high shear rate, they reveal shear-thinning behavior, which the apparent viscosity reduction by increasing rate. However, higher rate let them become dilatant. Consequently, an increase in contributes to and thus decreases well injectivity polymer. Hydraulic fracturing reduces mechanical shearing vicinity...

Abstract Partial Differential Equations (PDEs) have a wide list of applications in modeling complex processes including flow porous materials. Solution these equations that are mostly highly non-linear is generally possible using numerical algorithms carried out by iterative approaches like Newton's method, where the calculations to find solution at new time step started an initial guess unknown variables. The computational efficiency dependent on closeness guesses exact values. As routine,...

Mercury injection capillary pressure analysis is a methodology for determining different petrophysical properties, including bulk density, porosity, and pore throat distribution. In this work, distinct parameters derived from mercury tests was considered the prediction of permeability by coupling machine learning theoretical approaches in dataset composed 246 tight sandstone samples. After quality checking dataset, feature selection carried out correlation models statistical with measured...

Abstract We investigated countercurrent spontaneous imbibition (COUCSI) of water displacing oil in a 1D linear system with one side open, and closed. The Physics-Informed Neural Networks (PINNs) technique was used to estimate saturation profiles along the core recovery against time; based on same input information as reservoir simulator. demonstrate usefulness Change-of-Variables an approach improve PINN solutions. problem first normalized, where only saturation-dependent diffusion...

Summary Oil flow from matrix to fractures system in naturally fractured reservoirs is depended on imbibition process. Imbibition occurred due the presence of capillary forces. Recent investigations show that pressure value may be affected by rate change saturation wetting phase addition itself. In this work, impact considering dynamic effects processes will investigated. So, a new method has been used simulate using available commercial simulators. The results recovery highly effects. Also,...

Spontaneous imbibition is the main oil production mechanism in naturally fractured reservoirs especially water invaded zones. In this study and by extending previous works, effect of gravity force on relative contribution co counter current to investigated. It assumed that core placed vertically while lower face contact with upper oil. case, imbibes upward saturated core. After mathematical analysing considered scenario, results show effects can be ignored scale studies. By a way, field...

Summary Capillary pressure is routinely measured using a centrifuge setup where capillary forces retain heavy, wetting phase (e.g. water) and keep light, non-wetting from entering. By increasing the rotational speed of centrifuge, density difference phases heavy fluid out, while light can enter. In this work, we consider compare behaviour two setups: first core face closest to rotation axis open phase, farthest phase; labelled Two-Ends-Open, TEO. At increased rotation, generates strictly...

Summary Global energy demand is increasing annually and enhancing oil recovery (EOR) from mature reservoirs of crucial importance. Polymer flooding a favorable chemical EOR method because its impact on sweep efficiency. Polymeric fluid used to drive water within bypassed regions since they are responsible for viscosity increment relative permeability reduction. A high polymeric improves mobility ratio, reduces viscous fingering, as result enhances areal challenging problem polymer the shear...

The application of Physics-Informed Neural Networks (PINNs) is investigated for the first time in solving one-dimensional Countercurrent spontaneous imbibition (COUCSI) problem at both early and late (i.e., before after front meets no-flow boundary). We introduce utilization Change-of-Variables as a technique improving performance PINNs. formulated COUCSI three equivalent forms by changing independent variables. describes saturation function normalized position X T; second Y=T^0.5; third...