betaFeedback

WorksPhysics and Astronomy

Dynamics of fluid-fluid displacements in a model rough fracture beyond the quasistatic limit: A spectral approach

Abstract

In fluid-fluid displacements in porous and fractured materials, viscous friction in microscale interfacial (Haines) jumps is intimately linked to macroscale energy dissipation and the associated pressure-saturation (retention) hysteresis. The mode of control (flow rate vs. pressure) and the driving rate (from quasistatic to finite) can substantially affect hysteresis. Despite the significance of hysteresis and dissipation in various technological and natural processes, a quantitative understanding of the link between the micro- and macro-scales, and of the impact of the inherent heterogeneity of porous media, remains elusive. An imperfect'' Hele-Shaw cell of variable aperture is a simple model system which allows to study all these in details. However, simulating fluid-fluid interface evolution in heterogeneous media, even in such a simple system, is computationally prohibitive, as multiple length and time scales need to be resolved simultaneously. We develop here a spectral computational approach for interface evolution and energy dissipation and validate it via comparison to computational fluid dynamics simulations and experiments. Computational efficiency is demonstrated by following interface evolution in a cell with a single defect'', as well as with random roughness; in both, disparate length scales lead to nontrivial dynamics over many orders of magnitude in time. We also show theoretically that while viscous forces during Haines jumps fully account for the energy dissipated between consecutive metastable equilibria, viscosity does not change the total dissipated amount, merely the relaxation time. Our approach provides a stepping stone towards upscaling of fluid-fluid flows in porous media.

Banner appropriate to article type will appear here in typeset article

Corresponding author: Mykyta Chubynsky, [email protected]; Ran Holtzman, [email protected]

Introduction

The flow of multiple fluid phases in porous materials is an important topic in a multitude of natural and engineered applications. One inherent intriguing aspect is its history-dependent nature, typically described in terms of pressure–saturation hysteresis and the corresponding energy dissipation [1, 10, 26, 27]. Hysteresis and dissipation substantially affect the rate and pathways by which one fluid displaces another, which is crucial for planning and controlling processes such as soil irrigation and evaporation, carbon dioxide sequestration, and underground hydrogen storage [52].

In idealised systems like a Hele-Shaw (HS) cell of uniform aperture or a cylindrical capillary of uniform cross section, the interfacial dynamics can be obtained analytically for pressure-driven flows [9, 13]; in such homogeneous media, hysteresis and dissipation emerge only at finite rates of pressure change. In contrast, displacements in porous systems display hysteresis and dissipation even in the quasistatic driving limit (when the fluids are driven at zero flow rate, an idealised limit that cannot be fully exercised in practice), because the heterogeneity (in conduit geometry, e.g. pore sizes) gives rise to multiple equilibrium interfacial configurations, and energy is dissipated in out-of-equilibrium interfacial jumps between consecutive metastable equilibria. These so-called Haines jumps are triggered whenever the driving pressure pushes the two-fluid configuration beyond its metastability limit [22].

Phenomenological models of hysteresis, based on the collective behaviour of independent or interacting hysteretic units (hysterons), rely on the link between interfacial jumps and energy losses [4, 10, 24]. However, it is not clear what hysterons represent at microscopic scales, and thus how to establish a connection between model parameters and actual physical and geometrical quantities and obtain quantitative predictions without fitting of parameter values. Progress has been made using mildly heterogeneous confined systems that provide a physical understanding of the interfacial dynamics [16, 50, 20, 18, 21, 35, 47, 53, 9, 8] (Soriano et al. 2002). The convoluted geometry of the fluid-fluid interfaces in three-dimensions (3D) makes quasi-two-dimensional (2D) systems an attractive model system. A prominent example of such media is the 2D array of “pillars” (cylindrical objects of various radii), which can be easily manufactured and studied experimentally by standard microfluidics techniques [12, 6].

Recently, in particular, the “imperfect Hele-Shaw cell” (IHSC; Figure 1) has allowed a thorough investigation of the pore-scale processes underlying Haines jumps and hysteresis [46, 26, 27, 33, 25]. The IHSC is an idealised model for an open fracture of random aperture, where fluids move slowly enough (in the Stokes regime) for flows to be quantified via Darcy’s law, and fluid-fluid interfaces are distorted by capillary forces arising from aperture variations. IHSCs prove to be highly useful in understanding instabilities in quasi 2D systems, and a promising stepping stone towards understanding 3D flows in disordered media [e.g. 40]. Using IHSCs composed of single, pair, and multiple randomly-distributed mesa-shaped heterogeneities (“defects”), we have studied both experimentally and numerically individual and collective capillary jumps, and related them to their macroscopic manifestation as pressure-saturation hysteresis and energy dissipation [46, 26, 27, 33, 25]; these works, however, are limited to quasistatic driving [except Lavi et al. [33] which treats a single constriction].

Schematic illustration of the model system

Figure 1. Schematic illustration of the model system: An imperfect Hele-Shaw cell with randomly varying aperture b(x,y)=b0±δb(x,y)b(x,y)=b_0 \pm\delta b(x,y). The cell is tilted to provide an effective gravity ge=gsin⁡αg_e=g\sin\alpha.

In practice processes are never truly quasistatic, as both driving (changing pressure or fluid volume) and the consequent relaxation (from these perturbations) occur over a finite time scale, and, in particular, Haines jumps are dynamic by nature [even if the driving were truly quasi-static, the interface jumps would occur at high speeds [36, 3]. Dynamics of individual Haines jumps in quasistatic driving can provide insights into the various time scales, from which one can deduce when the quasistatic approximation is appropriate. Furthermore, finite-rate displacements go through different paths (in terms of interfacial configurations) than quasistatic displacements, as the system progressively departs from mechanical equilibrium as rates increase. At finite driving rates, viscous dissipation occurs both in the intermittent jumps triggered by the medium heterogeneities and in the continuous displacements of the fluid phases [36, 3]. The question arises as to whether these two contributions — stemming from viscous friction in perfect cells and from Haines jumps in disordered cells — can simply be added, e.g. to quantify energy dissipation.

The driving mode of displacement also plays a significant role. In flow-rate-controlled displacements the global mass per unit time is conserved, while it is not in pressure-controlled displacements. The corresponding displacement paths can thus differ significantly. For example, Haines jumps are suppressed and pressure-saturation hysteresis disappears in volume-controlled quasistatic displacements in a vertical capillary tube of varying cross section [38]. This is a consequence of the fact that the interface between the two fluids has only one degree of freedom, the height h(t)h(t). While in IHSC the interface h(x,t)h(x,t) presents additional degrees of freedom (many different configurations for a given invaded area), differences between the two driving modes may still be expected; how controlling the volume instead of the pressure impacts hysteresis is an interesting open question.

Simulating the dynamics of the displacements at finite rates in large disordered media is computationally prohibitive. Computational fluid dynamics models (CFD; also called “direct methods”) discretising the Navier-Stokes equations are capable of resolving two-phase flows at finite rates in complex pore geometries, however, at a computational cost that limits simulations to very small domains, insufficient to describe complex heterogeneous pore geometry or generate the large number of realizations required for upscaling (Ramstad et al. 2019; Zhao et al. 2019; Giudici et al. 2023). The computational cost of CFD remains prohibitive even when considering 2D systems, or averaging over the third dimension [31]. Pore network models require much less computational resources, but they use a simplified pore geometry that cannot resolve the actual interface configuration (Hughes & Blunt 2001; Ferer et al. 2011; Giudici et al. 2023).

The main purpose of the present work is to provide a computational tool that makes it possible to predict and analyse two-phase displacements in IHSCs, at high accuracy and computational efficiency, for both pressure and flow-rate driving. The proposed approach is applicable to all scenarios discussed thus far. Here, we benchmark its performance against interfacial behaviours observed in IHSCs of diverse geometries and under varying conditions.

The paper is organised as follows. In § 2, we derive the fundamental equations for the IHSC, and use the Hele-Shaw formulation to explicitly show that viscous (frictional) forces during Haines jumps fully account for the energy dissipated between consecutive metastable equilibria in quasistatic displacements. Our analysis also provides a non-intuitive result: while the energy dissipation is viscous in origin, the total dissipation in a Haines jump is viscosity-independent. § 3 presents a novel spectral approach, formulated based on the hydrodynamic equations for the bulk flow in the IHSC and kinematic and dynamic boundary conditions for the interface dynamics. The approach tracks the dynamics of the Fourier modes into which the interface shape is decomposed, providing both high accuracy and high computational efficiency. § 4 validates the model by comparing it for the simple case of a single (isolated) “mesa” defect against experiments and high-fidelity CFD simulations. Next, we consider two examples illustrating the promising power of the spectral approach. First, in § 5, we study in more detail the single defect case, where, despite its apparent simplicity, our numerical and analytical studies with the spectral approach reveal complex, multistage behaviour of a Haines jump that would be extremely challenging to study using traditional approaches. Then, in § 6 we simulate a large disordered IHSC with random roughness (proxy of a rough fracture). Finally, the conclusions are drawn in § 7.

Flow and energy dissipation in an imperfect Hele-Shaw cell

We consider an IHSC in which one inner surface is flat, and the other is rough such that the aperture (inter-surface distance) b(x,y)b(x,y) varies in space. Here xx is horizontal and yy is directed along the surface upwards (Figure 1). The cell is filled with two fluids, one fully wetting with viscosity μ\mu and density ρ\rho (which we will refer to as ‘liquid’) and the other nonwetting with comparatively negligible density and viscosity (‘air’). The interface between the two fluids is assumed to be given by a single-valued function of xx, h(x,t)h(x,t), such that liquid fills the space y<h(x,t)y<h(x,t), namely disallowing interface disconnections and trapping.

We first recall the basic equations describing the above system and introduce two types of boundary conditions corresponding to pressure and flow-rate control (§ 2.1). As considered previously, when the controlling inlet pressure is constant or the flow rate is zero, it is possible to write down an equation for the shape of the interface in equilibrium; in the quasistatic limit, the interface may follow the equilibrium shape that now changes slowly, except during Haines jumps (§ 2.2). Starting from an equation for the energy dissipation rate, we then show that the total viscous dissipation amount during a Haines jump in the quasistatic limit is viscosity-independent (§ 2.3). A simplifying assumption, important, in particular, for the derivation of the spectral approach, is made in Sec. 2.4. This assumption is used to derive an equation for the “instantaneous equilibrium” interface shape, generalising the concept of equilibrium to nonzero flow rates (§ 2.5). The concepts of true and instantaneous equilibria are important in the derivation of the spectral approach in § 3, as perturbations of the interface shape with respect to an equilibrium are considered. An expression for the dissipation rate involving only the evolution of the interface shape is derived in § 2.6; this is used later with the spectral approach (§ 3.5) and also shows explicitly that the dissipation amount matches the work of external forces.

Finally, an expression for the amount of dissipation in a Haines jump is derived and compared to previous work in § 2.7.

Governing equations

We simplify the problem to two dimensions (2D) — the x−yx-y plane — by neglecting velocity gradients along the cell compared to those in the transverse direction, as well as transverse pressure gradients (Homsy [28]). Assuming Stokes flow, we get the following expression for the 2D flow velocity (averaged in the zz direction) in the liquid phase (Darcy’s law):

u(x,y)=−b2(x,y)12μ(∇p(x,y)+ρgey^).(1)\mathbf{u}(x,y)=-\frac{b^2(x,y)}{12\mu}\left(\nabla p(x,y)+\rho g_e\hat{\mathbf{y}}\right). \tag*{(1)}

Here p(x,y)p(x,y) is the pressure in the liquid phase, ge=gsin⁡αg_e=g\sin\alpha is the effective gravity (α\alpha is the tilt angle), and y^\hat{\mathbf{y}} is the unit vector along the yy direction. The continuity equation is

∇⋅[b(x,y)u(x,y)]=0.(2)\nabla\cdot\left[b(x,y)\mathbf{u}(x,y)\right]=0. \tag*{(2)}

These two equations are valid for y<h(x,t)y<h(x,t). At the interface [y=h(x,t)y=h(x,t)], the dynamic boundary condition is Young-Laplace law, which states that

pa−p=γ(π4κ(x,t)+2b(x,h(x,t))).(3)p_a-p=\gamma\left(\frac{\pi}{4}\kappa(x,t)+\frac{2}{b(x,h(x,t))}\right). \tag*{(3)}

where γ\gamma is the surface tension, κ\kappa is the in-plane curvature of the interface, and the expression in the parentheses is an approximation for the sum of the principal curvatures of the interface assuming complete wetting (Park & Homsy [42]). From here on, we take the non-wetting phase pressure pap_a to be atmospheric, and pa=0p_a=0. Hereafter, for brevity we do not explicitly write the functional dependence of these variables on xx, yy, and tt, except where it is essential.

The interface moves according to the kinematic boundary condition,

∂h∂t=uy−ux∂h∂x.(4)\frac{\partial h}{\partial t}=u_y-u_x\frac{\partial h}{\partial x}. \tag*{(4)}

The boundaries at the sides, x=0x=0 and x=Wx=W, are impermeable (ux=0u_x=0), enforced via ∂p/∂x=0\partial p/\partial x=0.

We will be considering throughout this paper two types of boundary conditions for the inlet (bottom boundary, y=0y=0): (i) pressure-controlled, setting the pressure at the inlet, p(y=0)=Pp(y=0)=P; and (ii) flow-rate-controlled, setting the inflow speed, uy(y=0)=Vu_y(y=0)=V. The latter also includes the case of a closed inlet (V=0V=0).

The above relies on a number of assumptions, conditions and approximations:

  • We assume a single interface with a shape that is a single-valued function of xx. While this limitation is not inherent in the approach, it simplifies the formulation; it is also necessary for the spectral approach we develop in § 3 in any case.

  • Further, we assume negligible interface deformation by gravity, valid when ℓc≫b\ell_c\gg b, where

ℓc=(πγ4ρge)1/2(5)\ell_c=\left(\frac{\pi\gamma}{4\rho g_e}\right)^{1/2} \tag*{(5)}

is the capillary length (de Gennes et al. [17]). This allows us to use the expression for the interface curvature in Eq. (3).

  • The tangential speed on the boundaries cannot be specified in this formulation, meaning that there is slip at the side walls (which, however, only affects flow up to distances of the order of bb).

  • The wetting properties of the side walls are not an input. The wall wettability affects the meniscus only up to a distance of ∼ℓc\sim\ell_c from the walls [17], which is typically small (ℓc≪W\ell_c \ll W). This is further considered in § 3.1.

  • Complete wetting of the top and bottom surfaces of the cell is assumed [while noting that generalisation to partial wetting with a fixed contact angle is straightforward, we do not consider dynamic contact angle variations, including hysteresis of the dynamic contact angle between imbibition and drainage displacement directions [41]].

  • Inertial effects, which may be important during fast Haines jumps, are not included.

  • Equation (1) assumes that in the Laplacian (∂2/∂x2+∂2/∂y2+∂2/∂z2)u(\partial^2/\partial x^2+\partial^2/\partial y^2+\partial^2/\partial z^2)\mathbf{u} the last term dominates.

  • Finally, we do not account for 3D wetting effects, including thin residual coating layers of the wetting liquid along the solid walls left in the drained region during imbibition, which modify the kinematic and dynamic boundary conditions to account for volume conservation and enhanced curvature and viscous stresses at the interface [42, 37] (Reinelt & Saffman 1985), as well as instabilities caused by entrainment of the less-wetting fluid [34, 39] and corner films [7] which may appear along mesa defects and possibly other locations where there are abrupt changes of aperture.

Despite these assumptions, the approach applies to a wide range of conditions of interest, confirmed, for example, by the good agreement with experiments discussed in § 4.2.

Equilibria and quasistatic driving

Before introducing dynamics, we recall the equation for the equilibrium shape of the interface and the concept of quasistatic driving.

For the pressure control case with P=constP=\mathrm{const}, we introduce a quantity measuring the imbalance between the pressure at the interface [Eq. (3)], the pressure at the inlet, PP, and the gravity term, ρgeh\rho g_e h,

pe(x)=ρgeh(x)−γ(π4κ({h(x)})+2b(x,h(x)))−P.(6)p_e(x)=\rho g_e h(x)-\gamma\left(\frac{\pi}{4}\kappa(\{h(x)\})+\frac{2}{b(x,h(x))}\right)-P. \tag*{(6)}

The system is in equilibrium if there is a balance for all xx, i.e.,

pe(x)≡0.(7)p_e(x)\equiv0. \tag*{(7)}

Indeed, the solution of Eqs. (1)–(4) in the liquid phase then gives

p=P−ρgey,(8)p=P-\rho g_e y, \tag*{(8)}
u=0,(9)\mathbf{u}=0, \tag*{(9)}
∂h/∂t=0,(10)\partial h/\partial t=0, \tag*{(10)}

and the interface does not move. Depending on the details of b(x,y)b(x,y), the equilibrium condition (7) may have multiple solutions heq(x)h_{\rm eq}(x). If Eq. (7) is not satisfied, it implies pressure imbalance that produces flow and interface motion.

In the quasistatic limit, changes in the inlet pressure PP in time are applied slowly such that the system remains in equilibrium [Eq. (7) is satisfied] at all times. This limit was considered in previous publications [26, 27]. Note that the factor π/4\pi/4 in front of the curvature term in Eq. (6), correcting for 3D curvature [42], was not included in the 2D model in [26, 27] (except where quantitative agreement with experimental data was sought). The existence of multiple solutions may give rise to hysteresis (history dependence) of the interface evolution with PP, even in this quasistatic limit. At a particular PP a solution branch may terminate [heq(x)h_{\mathrm{eq}}(x) may become non-real], with the interface necessarily undergoing a Haines jump to a different equilibrium solution, on a time scale so short that the change of PP during the jump can be neglected.

For flow-rate control with V=0V = 0 (a closed inlet), the equilibrium condition is still given by Eq. (7), however, PP is no longer a known constant but an unknown determined by the volume of the liquid (conserved in this case).

Energy dissipation: Rate, amount, and the role of viscosity

Out of equilibrium, the moving liquid dissipates energy. The dissipation rate RR is equal to the work per unit time of viscous forces, which in the overdamped limit is equal in magnitude and opposite in sign to the work per unit time of all other forces. Using Eq. (1), we get

∣R∣=−∫b(u⋅[∇p+geρy^]) dA=12μ∫u2b dA(11)|R| = -\int b\left(\mathbf{u}\cdot\left[\nabla p + g e_{\rho}\hat{y}\right]\right)\,\mathrm{d}A = 12\mu\int\frac{u^2}{b}\,\mathrm{d}A \tag*{(11)}

where integration is over the area of the cell filled with liquid. (There are different conventions for the sign of energy dissipation; here we use ∣ ∣|\,| to remove the ambiguity.) To use this expression, the full velocity field u(x,y)\mathbf{u}(x,y) must be known (e.g. from numerical simulations of the Hele-Shaw equations). Equation (11) can also be obtained directly as the work per unit time of viscous forces, using the parabolic (Poiseuille) approximation of the transverse velocity profile, from which Eq. (1) is derived (see Appendix § 8).

In the quasistatic limit, outside of Haines jumps—as the liquid velocity tends to zero—the dissipation rate ∣R∣|R| approaches zero as well. Moreover, considering a process in which the interface moves between two given equilibrium configurations without Haines jumps, in the quasistatic limit the total amount of dissipation tends to zero as well. This is because, even though in this limit the duration of the process (inversely proportional to the interface speed) tends to infinity, it does so more slowly than the dissipation rate (proportional to the square of the interface speed) tends to zero. Thus, in this limit one only needs to consider dissipation during Haines jumps.

Notably, even though energy dissipation in the system we consider is viscous in nature, the total amount of dissipation in a Haines jump under quasistatic driving does not depend on the liquid viscosity μ\mu. This is true in both the pressure-controlled case where PP can be considered constant during the jump, and the flow-rate-controlled case with V→0V \to0. We demonstrate this rather nonintuitive statement by considering systems that only differ in liquid viscosity, all starting with the same initial interface profile h(x)h(x). The interface will move through the same configurations, but on different (viscosity-dependent) time scales, i.e.,

h(x,t;μ)=h(x,t/μ;1).(12)h(x,t;\mu)=h(x,t/\mu;1). \tag*{(12)}

This means that it is possible to consider a process where such systems not only start with the same interface profile, but also end with the same interface profile though at different times (∝μ\propto\mu). Further, the velocity fields at the corresponding times in the evolution process are likewise the same, only being scaled by the viscosity, i.e.,

u(x,y,t;μ)=μ−1u(x,y,t/μ;1).(13)\mathbf{u}(x,y,t;\mu)=\mu^{-1}\mathbf{u}(x,y,t/\mu;1). \tag*{(13)}

Therefore, the total dissipation in this process is the same in all systems. Indeed, Eq. (11) gives for the dissipation rate

∣R(t;μ)∣=12μ∫x=0W∫y=0h(x,t/μ;1)[μ−1u(x,y,t/μ;1)]2b(x,y) dx dy=1μ∣R(t/μ;1)∣,(14)|R(t;\mu)| = 12\mu\int_{x=0}^{W}\int_{y=0}^{h(x,t/\mu;1)} \frac{\left[\mu^{-1}u(x,y,t/\mu;1)\right]^2}{b(x,y)}\,dx\,dy = \frac{1}{\mu}|R(t/\mu;1)|, \tag*{(14)}

and the total dissipation is

∣Ψ(μ)∣=∫0μt0∣R(t;μ)∣ dt=1μ∫0μt0∣R(t/μ;1)∣μ d(t/μ)=∣Ψ(1)∣.(15)|\Psi(\mu)| = \int_{0}^{\mu t_0}|R(t;\mu)|\,dt = \frac{1}{\mu}\int_{0}^{\mu t_0}|R(t/\mu;1)|\mu\,d(t/\mu) = |\Psi(1)|. \tag*{(15)}

Here t0t_0 is defined as the duration of the process for μ=1\mu= 1. Thus, the total dissipation amount is viscosity-independent. A particular case of such a process is a Haines jump during quasistatic driving, where t0→∞t_0 \to\infty because the relaxation to the static equilibrium takes infinite time. Intuitively, a more viscous liquid dissipates less per unit time (while the viscosity is higher, the flow speed has a stronger effect as the dissipation rate is quadratic in the speed), yet the process takes longer by the same factor.

The viscosity independence of the total dissipation is natural, because in the Stokes limit where Eq. (1) was derived the contribution of the kinetic energy to the energy balance is negligible, and the total dissipation simply equals the work of all forces other than viscous friction. This work only depends on the initial and final interface shapes. This is discussed in more detail in § 2.6 and § 2.7 below.

Simplifying assumption: Computing permeability from mean aperture (b≈b0b \approx b_0)

A simplifying assumption made throughout this paper (unless stated otherwise), which is essential for our derivation of the spectral approach, is that the variations of the cell aperture b(x,y)b(x,y) are small compared to the mean aperture b0b_0, such that we neglect their effects on the permeability [Eq. (1)] and continuity [Eq. (2)]. That is, we approximate Eqs. (1) and (2) by replacing bb with b0b_0:

u≈−b0212μ(∇p+geρy^),(16)\mathbf{u} \approx-\frac{b_0^2}{12\mu}\left(\nabla p + g_e\rho\hat{\mathbf{y}}\right), \tag*{(16)}
∇⋅u≈0.(17)\nabla\cdot\mathbf{u} \approx0. \tag*{(17)}

We stress that the effect of the non-uniformity of b(x,y)b(x,y) on capillary pressure is explicitly included. This may appear inconsistent, since all three contributions due to variations of bb are of the same order (linear) in δb=b−b0\delta b = b - b_0 in the equations they enter.

To justify this apparent inconsistency, we note that while the capillary pressure directly affects the equilibrium interface shape, the permeability and continuity only influence the rate at which this equilibrium is approached. A simple analog with one degree of freedom is ds/dt=−s+ϵ+ϵ ds/dtds/dt = -s + \epsilon+ \epsilon\,ds/dt, where ϵ\epsilon is very small. Formally, both perturbation terms are O(ϵ)O(\epsilon). However, the first shifts the equilibrium position by O(ϵ)O(\epsilon), while the second only affects the rate at which the equilibrium is approached. Starting at the unperturbed equilibrium position s=0s = 0 (or in general for perturbations of the equilibrium of O(ϵ)O(\epsilon)), ds/dtds/dt is O(ϵ)O(\epsilon), and thus the second perturbation term is O(ϵ2)O(\epsilon^2) with a negligible effect in the first-order approximation.

Finally, note we use b=b0b = b_0 in Eq. (11) to calculate the dissipation.

Instantaneous equilibrium for finite flow rates

The concept of equilibrium can be generalised to the flow-rate control case with V≠0V \ne0. We define

References

  1. [1]Albers, B. 2014 Modeling the hysteretic behavior of the capillary pressure in partially saturated porous media: A review. Acta Mech. 225, 2163–2189.DOI
  2. [2]Alvarez-Lacalle, E., Casademunt, J. & Ortín, J. 2001 Systematic weakly nonlinear analysis of interfacial instabilities in hele-shaw flows. Phys. Rev. E 64, 061302.
  3. [3]Berg, S., Ott, H., Klapp, S. A., Schwing, A., Neiteler, R., Brussee, N., Makurat, A., Leu, L., Enzmann, F., Schwarz, J.O., Kersten, M., Irvine, S. & Stampanoni, M. 2013 Real-time 3D imaging of Haines jumps in porous media flow. P. Natl. Acad. Sci. USA 110 (10), 3755–3759.DOI
  4. [4]Bertotti, G. & Mayergoyz, I.D., ed. 2006 The Science of Hysteresis, vol. 1: Mathematical Modeling and Applications. Academic Press.
  5. [5]Bonn, D., Eggers, J., Indekeu, J., Meunier, J. & Rolley, E. 2009 Wetting and spreading. Rev. Mod. Phys. 81 (2), 739–805.DOI
  6. [6]Borgman, O., T. Darwent, Segre, E., Goehring, L. & Holtzman, R. 2019 Immiscible fluid displacement in porous media with spatially correlated particle sizes. Adv. Water Resour. 128, 158.DOI
  7. [7]Chauvet, F., Duru, P., Geoffroy, S. & Prat, M. 2009 Three periods of drying of a single square capillary tube. Phys. Rev. Lett. 103, 124502.pubmed.ncbi.nlm.nih.gov/19792442
  8. [8]Clotet, X., Ortín, J. & Santucci, S. 2014 Disorder-induced capillary bursts control intermittency in slow imbibition. Phys. Rev. Lett. 113, 074501.pubmed.ncbi.nlm.nih.gov/25170710
  9. [9]Clotet, X., Planet, R. & Ortín, J. 2012 Capillary rise in Hele-Shaw models of disordered media. J. Colloid. Interf. Sci. 377 (1), 387–395.DOI
  10. [10]Cueto-Felgueroso, L. & Juanes, R. 2016 A discrete-domain description of multiphase flow in porous media: Rugged energy landscapes and the origin of hysteresis. Geophys. Res. Lett. 43 (4), 1615–1622.
  11. [12]Ferrari, A., Jimenez-Martinez, J., Le Borgne, T., Méheust, Y. & Lunati, I. 2015 Challenges in modeling unstable two-phase flow experiments in porous micromodels. Water Resour. Res. 51 (3), 1381–1400.DOI
  12. [13]Fries, N. & Dreyer, M. 2008 An analytic solution of capillary rise restrained by gravity. J. Colloid. Interf. Sci. 320 (1), 259–263.DOI
  13. [14]Frigo, M. & Johnson, S.G. 2005 The design and implementation of FFTW3. Proceedings of the IEEE 93 (2), 216–231, special issue on “Program Generation, Optimization, and Platform Adaptation”.DOI
  14. [15]Ganesan, V. & Brenner, H. 1998 Dynamics of two-phase fluid interfaces in random porous media. Phys. Rev. Lett. 81 (3), 578–581.DOI
  15. [16]de Gennes, P.-G. 1986 Imperfect Hele-Shaw cells. J. Physique (Paris) 47 (9), 1541–1546.DOI
  16. [17]de Gennes, P.-G., Brochard-Wyart, F. & Quéré, D. 2004 Capillarity and Wetting Phenomena. Drops, Bubbles, Pearls, Waves. New York: Springer.DOI
  17. [18]Geromichalos, D., Muegle, F. & Herminghaus, S. 2002 Nonlocal dynamics of spontaneous imbibition fronts. Phys. Rev. Lett. 89, 104503.pubmed.ncbi.nlm.nih.gov/12225201
  18. [20]Glass, R.J., Nicholl, M.J. & Yarrington, L. 1998 A modified invasion percolation model for low-capillary number immiscible displacements in horizontal rough-walled fractures: Influence of local in-plane curvature. Water Resour. Res. 34 (12), 3215–3234.
  19. [21]Glass, R.J., Rajaram, H. & Detwiler, R.L. 2003 Immiscible displacements in rough-walled fractures: Competition between roughening by random aperture variations and smoothing by in-plane curvature. Phys. Rev. E 68 (6, 1).pubmed.ncbi.nlm.nih.gov/14754183
  20. [22]Haines, W. B. 1930 Studies in the physical properties of soil. V. The hysteresis effect in capillary properties, and the modes of moisture distribution associated therewith. J. Agr. Sci. 20, 97–116.
  21. [23]Hecht, F. 2012 New development in FreeFem++. J. Numer. Math. 20 (3-4), 1–14.DOI
  22. [24]Helland, L., Jettfstuен, E. & Friis, H.A. 2021 A discrete-domain approach to three-phase hysteresis in porous media. Water Resour. Res. 57 (6), e2021WR029560.DOI
  23. [25]Holtzman, R., Dentz, M., Moura, M., Chubynsky, Mykyta V, Planet, R. & Ortín, J. 2024 Emergence of dissipation and hysteresis from interactions among reversible, nondissipative units: The case of fluid-fluid interfaces. Phys. Rev. Fluids 9 (6), 064001.
  24. [26]Holtzman, R., Dentz, M., Planet, R. & Ortín, J. 2020 The origin of hysteresis and memory of two-phase flow in disordered media. Commun. Phys. 3, 222.
  25. [27]Holtzman, R., Dentz, M., Planet, R. & Ortín, J. 2023 The relation between dissipation and memory in two-fluid displacements in disordered media. Geophys. Res. Lett. 50 (16), e2023GL104073.
  26. [28]Homsy, G.M. 1987 Viscous fingering in porous media. Annu. Rev. Fluid Mech. 19, 271–311.
  27. [30]Juanes, R., Meng, Y. & Primkulov, B.K. 2020 Multiphase flow and granular mechanics. Phys. Rev. Fluids 5 (11), 110516.DOI
  28. [31]Krishna, R., Méheust, Y. & Neuweiler, I. 2025 A two-dimensional depth-integrated model for immiscible two-phase flow in open rough fractures. J. Fluid Mech. 1011, A43.
  29. [32]Ladd, A.J.C. & Szymczak, P. 2021 Reactive flows in porous media: Challenges in theoretical and numerical methods. Annu. Rev. Chem. Biomol. Eng. 12 (1), 543.DOI
  30. [33]Lavi, I., Rose, L., Planet, R., Casademunt, J., Santucci, S. & Ortín, J. 2023 Prefiburfocation enhancement of imbibition-drainage hysteresis cycles. Phys. Rev. Fluids 8, 124002.DOI
  31. [34]Levaché, B. & Bartolo, D. 2014 Revisiting the Saffman-Taylor Experiment: Imbibition Patterns and Liquid-Entrainment Transitions. Phys. Rev. Lett. 113 (4), 044501.DOI
  32. [35]Meakin, P. & Tartakovsky, A.M. 2009 Modeling and simulation of pore-scale multiphase fluid flow and reactive transport in fractured and porous media. Rev. Geophys. 47.DOI
  33. [36]Moebius, F. & Or, D. 2012 Interfacial jumps and pressure bursts during fluid displacement in interacting irregular capillaries. J. Colloid Interf. Sci. 377 (1), 406–415.
  34. [37]Morrow, L.C., Cuttle, C. & MacMinn, C.W. 2023 Gas compression systematically delays the onset of viscous fingering. Phys. Rev. Lett. 131, 224002.DOI
  35. [38]Nepal, A., Sheikh, S., Hidalgo, J., Ortín, J. & Dentz, M. 2026 Fluid interfaces under quasi-static driving in capillaries with varying cross-section. J. Colloid. Interf. Sci. pp. submitted (JCIS–26–2819).
  36. [39]Odier, C., Levaché, B., Santanach-Carreras, E. & Bartolo, D. 2017 Forced imbibition in porous media: a fourfold scenario. Phys. Rev. Lett. 119 (November), 208005.DOI
  37. [40]Orozco, D., León-Fon-Fay, D., Zamora-Ledezma, C. & Medina, E. 2026 Two-phase displacement instabilities in variable-gap hele-shaw cells. Trans. Porous Media 153 (3).
  38. [41]Pahlavan, A.A., Cueto-Felgueroso, L., McKinley, G.H. & Juanes, R. 2015 Thin films in partial wetting: internal selection of contact-line dynamics. Phys. Rev. Lett. 115, 034502.
  39. [42]Park, C.W. & Homsy, G.M. 1984 Two-phase displacement in Hele–Shaw cells: Theory. J. Fluid Mech. 139, 291–308.
  40. [43]Pauné, E. 2002 Interface dynamics in two-dimensional viscous flows. PhD thesis, Universitat de Barcelona, available at http://hdl.handle.net/2445/35490.
  41. [44]Pauné, E. & Casademunt, J. 2003 Kinetic roughening in two-phase fluid flows through a random Hele-Shaw cell. Phys. Rev. Lett. 90, 144504.pubmed.ncbi.nlm.nih.gov/12731922
  42. [45]Persson, B. N. J., Albohr, O., Tartaglino, U., Volokitin, A. I. & Tosatti, E. 2005 On the nature of surface roughness with application to contact mechanics, sealing, rubber friction and adhesion. J. Phys.: Condens. Matter 17 (1), R1–R62.
  43. [46]Planet, R., Díaz-Polo, L. & Ortín, J. 2020 Capillary jumps of fluid-fluid fronts across an elementary constriction in a model open fracture. Phys. Rev. Fluids 5, 044002.DOI
  44. [47]Planet, R., Santucci, S. & Ortín, J. 2009 Avalanches and non-gaussian fluctuations of the global velocity of imbibition fronts. Phys. Rev. Lett. 102, 094502.
  45. [48]Ramstad, T., Berg, C.F. & Thompson, K. 2019 Pore-scale simulations of single- and two-phase flow in porous media: Approaches and applications. Trans. Porous Media 130 (1), 77–104.DOI
  46. [50]Rubio, M. A., Edwards, C. A., Dougherty, A. & Gollub, J. P. 1989 Self-affine fractal interfaces from immiscible displacement in porous media. Phys. Rev. Lett. 63 (16), 1685–1688.pubmed.ncbi.nlm.nih.gov/10040644
  47. [51]Saeibehrouzi, A., Abolfathi, S., Denissenko, P. & Holtzman, R. 2024 Pore-scale modeling of solute transport in partially-saturated porous media. Earth Sci. Rev. 256, 104870.DOI
  48. [52]Sahimi, M. 2011 Flow and Transport in Porous Media and Fractured Rock. N.Y.: Wiley VCH Verlag GmbH.
  49. [53]Santucci, S., Planet, R., Måloy, J.K. & Ortín, J. 2011 Avalanches of imbibition fronts: Towards critical pinning. Europhys. Lett. 94 (4), 46005.DOI
  50. [55]Zhao, B., MacMinn, C.W., Primkulov, B.K., Chen, Y., Valocchi, A.J., Zhao, J. Q., Kang, K., Bruning, K., McClure, J.E., Miller, C.T., Fakhari, A., Bolster, D., Hiller, T., Brinkmann, M., Cueto-Felgueroso, L., Cogswell, D.A., Verma, R., Prodanović, M., Maes, J., Geiger, S., Vassvik, M., Hansen, A., Segre, E., Holtzman, R., Yang, Z., Yuan, C., Chareyre, B. & Juanes, R. 2019 Comprehensive comparison of pore-scale models for multiphase flow in porous media. P. Natl. Acad. Sci. USA 116 (28), 13799–13806.

Paper details

vlnc:W_wsmpm78c
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 6

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents