Simultaneous solution of the Navier-Stokes and elastic membrane equations by a finite element method

Fluid flow through a significantly compressed elastic tube occurs in a variety of physiological situations. Laboratory experiments investigating such flows through finite lengths of tube mounted between rigid supports have demonstrated that the system is one of great dynamical complexity, displaying a rich variety of self-excited oscillations. The physical mechanisms responsible for the onset of such oscillations are not yet fully understood, but simplified models indicate that energy loss by flow separation, variation in longitudinal wall tension and propagation of fluid elastic pressure waves may all be important. Direct numerical solution of the highly non-linear equations governing even the most simplified two-dimensional models aimed at capturing these basic features requires that both the flow field and the domain shape be determined as part of the solution, since neither is known a priori. To accomplish this, previous algorithms have decoupled the solid and fluid mechanics, solving for each separately and converging iteratively on a solution which satisfies both. This paper describes a finite element technique which solves the incompressible Navier-Stokes equatikons simultaneously with the elastic membrane equations on the flexible boundary. The elastic boundary position is parametized in terms of distances along spines in a manner similar to that which has been used successfully in studies of viscous free surface flows, but here the membrane curvature equation rather than the kinematic boundary condition of vanishing normal velocity is used to determine these diatances and the membrane tension varies with the shear stresses exerted on it by the fluid motions. Bothy the grid and the spine positions adjust in response to membrane deformation, and the coupled fluid and elastic equations are solved by a Newton-Raphson scheme which displays quadratic convergence down to low membrane tensions and extreme states of collapse. Solutions to the steady problem are discussed, along with an indication of how the time-dependent problem might be approached.     

A finite element solutionof three-dimensional mixed convection gas flows in horizontal chnnels using preconditioned iterative metrix methods

An algorithm is presented for the finite element solution of three-dimensional mixed convection gas flows in channels heated from below. The algorithm uses Newton's method and iterative matrix methods. Two iterative solution algorithms, conjugate gradient squared (CGS) and generalized minimal residual (GMERS), are used in conjunction with a preconditioning technique that is simple to implement. The preconditioner is a subset of the full Jacobian matrix centered around the main diagonal but retaining the most fundamental axial coupling of the residual equations. A domain-renumbering scheme that enhances the overall algorithm performance is proposed on the basis of and analysis of the preconditioner. Comparison with the frontal elimination method demonstrates that the iterative method will be faster when the front width exceeds approximately 500. Techniques for the direct assembly f the problem into a compressed sparse row storage format are demonstrated. Elimination of fixed boundary conditions is shown to decrease the size of the matrix problem by up to 30%. Finally, fluid flow solutions obtained with the numerical technique are presented. These solutions reveal complex three-dimensional mixed convection fluid flow phenomena at low Reynolds numbers, including the reversal of the direction of longitudinal rolls in the presence of a strong recirculation in the entrance region of the channel.     

Spiral thermocapillary flows in two-layer systems

Parallel thermocapillary flows in infinite layers can be calculated easily. However, in reality thermocapillary flows occur in channels or closed cavities, and they are three-dimensional. In the present paper, a two-layer fluid system filling a channel with a rectangular cross section is considered. A numerical investigation of three-dimensional spiral thermocapillary flows generated by a temperature gradient imposed along the channel has been performed. Both the case of a zero longitudinal pressure gradient (a through flow in the channel) and that of a zero longitudinal fluid flux (a flow in the closed cavity) are investigated. Steady and oscillatory thermocapillary motions, and transitions between them are studied.     

Open-channel flows and waterfalls

Steady two-dimensional free-surface flows of an inviscid incompressible fluid are studied here, using the complex potential theory. The first flow is a uniform free stream flow in a channel of finite depth. The second is similar, but terminated abruptly as a downstream flow exiting into a falling jet, with and without the effect of gravity. These problems have already been solved for polygonal walls. This paper presents an iterative process for computing flows over arbitrarily shaped channels, with and without the presence of a waterfall at the exit. This process is based on the solution of a mixed boundary problem in the unit disk. The method emphasizes the correspondence between the walls in the physical plane and in the unit disk. The numerical data agree with previously published results and extend them to arbitrary curved walls. This method yields solutions only for supercritical flows.     

Study of wall properties on peristaltic transport of a couple stress fluid

In this paper, we investigate the peristaltic transport of a couple stress fluid in a channel with compliant walls. Perturbation method has been used to get the solution. The flow is induced by sinusoidal traveling waves along the channel walls. The effects of wall damping, wall elastance, wall tension and couple stress parameter on the flow are investigated using the equations of fluid as well as deformable boundaries. It is found that the mean velocity at boundaries decreases with increasing couple-stress parameter and wall damping and increases with increasing wall tension and wall elastance, while the mean axial velocity increases with increasing wall tension and wall elastance and decreases with couple-stress parameter and wall damping.     

NUMERICAL SIMULATION OF STEADY FLOW IN A COMPLIANT TUBE OR CHANNEL WITH TAPERED WALL THICKNESS

Flow through compliant tubes with linear taper in wall thickness is numerically simulated by finite element analysis. Two models are examined: a compliant channel and an axisymmetric tube. For verification of the numerical method, flow through a compliant stenotic vessel is simulated and compared to existing experimental data. Steady two-dimensional flow in a collapsible channel with initial tension is also simulated and the results are compared with numerical solutions from the literature. Computational results for an axisymmetric tube show that as cross-sectional area falls with a reduction in downstream pressure, flow rate increases and reaches a maximum when the speed index (mean velocity divided by wave speed) is near unity at the point of minimum cross-sectional area, indicative of wave-speed flow limitation or “choking” (flow speed equals wave speed) in previous one-dimensional studies. For further reductions in downstream pressure, the flow rate decreases. Cross-sectional narrowing is significant but localized. For the particular wall and fluid properties used in these simulations, the area throat is located near the downstream end when the ratio of downstream-to-upstream wall thickness is 2; as wall taper is increased to 3, the constriction moves to the upstream end of the tube. In the planar two-dimensional channel, area reduction and flow limitation are also observed when outlet pressure is decreased. In contrast to the axisymmetric case, however, the elastic wall in the two-dimensional channel forms a smooth concave surface with the area throat located near the mid-point of the elastic wall. Though flow rate reaches a maximum and then falls, the flow does not appear to be choked.     

Effect of spatial discretization schemes on numerical solutions of viscoelastic fluid flows

This study examines the effect of discretization schemes for the convection term in the constitutive equation on numerical solutions of viscoelastic fluid flows. For this purpose, a temporally evolving mixing layer, a two-dimensional vortex pair interacting with a wall, and a fully developed turbulent channel flow are selected as test cases, and eight different discretization schemes are considered. Among them, the first-order upwind difference scheme (UD) and artificial diffusion scheme (AD), which are commonly used in the literature, show most stable and smooth solutions even for highly extensional flows. However, the stress fields are smeared too much by these schemes and the corresponding flow fields are quite different from those obtained by higher-order upwind difference schemes. Among higher-order upwind difference schemes investigated in this study, a third-order compact upwind difference scheme (CUD3) with locally added AD shows stable and most accurate solutions for highly extensional flows even at relatively high Weissenberg numbers.     

Flow-induced deformation of an elastic membrane adhering to a wall

The flow-induced deformation of a two-dimensional membrane with a circular unstressed shape clamped at the two ends on a plane wall at an arbitrary contact angle is considered. Working under the auspices of generalized shell theory, the membrane is allowed to develop in-plane tensions, transverse tensions, and bending moments determined by the curvature of the resting and deformed shapes. A system of ordinary differential equations governing the membrane shape is formulated, and the associated boundary-value problem is solved by numerical methods. Numerical results are presented to illustrate the deformation of a clamped membrane due to gravity or a negative transmural pressure. The shell formulation is coupled with a boundary-integral formulation for Stokes flow, and an efficient iterative scheme is developed to describe deformed equilibrium shapes of a membrane attached to a plane wall in the presence of an overpassing shear flow. Computations for different contact angles and shear rates reveal a wide variety of profiles and illustrate the distribution of the membrane tension developing due to the flow-induced deformation.     

Flow of viscoelastic fluids through a porous channel—I

The flow of viscoelastic fluids through a porous channel with one impermeable wall is computed. The flow is characterized by a boundary value problem in which the order of the differential equation exceeds the number of boundary conditions. Three solutions are developed: (i) an exact numerical solution, (ii) a perturbation solution for small R, the cross-flow Reynold's number and (iii) an asymptotic solution for large R. The results from exact numerical integration reveal that the solutions for a non-Newtonian fluid are possible only up to a critical value of the viscoelastic fluid parameter, which decreases with an increase in R. It is further demonstrated that the perturbation solution gives acceptable results only if the viscoelastic fluid parameter is also small. Two more related problems are considered: fluid dynamics of a long porous slider, and injection of fluid through one side of a long vertical porous channel. For both the problems, exact numerical and other solutions are derived and appropriate conclusions drawn.     

Numerical simulation of non-linear elastic flows with a general collocated finite-volume method

This paper reports the development and application of a finite-volume based methodology for the calculation of the flow of fluids which follow differential viscoelastic constitutive models. The novelty of the method lies on the use of the non-staggered grid arrangement, in which all dependent variables are located at the center of the control volumes, thus greatly simplifying the adoption of general curvilinear coordinates. The pressure–velocity–stress decoupling was removed by the development of a new interpolation technique inspired on that of Rhie and Chow, AIAA 82 (1982) 998. The differencing schemes are second order accurate and the resulting algebraic equations for each variable are solved in a segregated way (decoupled scheme). The numerical formulation especially designed for the interpolation of the stress field was found to work well and is shown to be indispensable for accurate results. Calculations have been carried out for two problems: the entry flow problem of Eggleton et al., J. Non-Newtonian Fluid Mech. 64 (1996) 269, with orthogonal and non-orthogonal meshes; and the bounded and unbounded flows around a circular cylinder. The results of the simulations compare favourably with those in the literature and iterative convergence has been attained for Deborah and Reynolds numbers similar to, or higher than, those reported for identical flow problems using other numerical methods. The application of the method with non-orthogonal coordinates is demonstrated. The entry flow problem is studied in more detail and for this case differences between Newtonian and viscoelastic fluids are identified and discussed. Viscoelasticity is shown to be responsible for the development of very intense normal stresses, which are tensile in the wall region. As a consequence, the viscoelastic fluid is more intensely decelerated in the wall region than the Newtonian fluid, thus reducing locally the shear rates and the role of viscosity in redeveloping the flow. A layer of high stress-gradients is formed at the wall leading edge and is convected below and away from the wall; its effect is to intensify the aforementioned deviation of elastic fluid from the wall.     

On the Perturbation of Hamel Flow Due to Unevenness of the Channel Walls

A method of analyzing the receptivity of longitudinally inhomogeneous flows is proposed. The process of excitation of natural oscillations is studied with reference to the simplest inhomogeneous flow: the two-dimensional flow of a viscous incompressible fluid in a channel with plane nonparallel walls. As physical factors generating perturbations, the cases of a stationary irregularity and localized vibration of the channel walls are considered. By changing the independent variables and unknown functions of the perturbed flow, the problem of the generation of stationary perturbations above an irregularity is reduced to a longitudinally homogeneous boundary-value problem which is solved using a Fourier transform in the longitudinal variable. The same problem is investigated using another method based on representing the required solution in the form of a superposition of solutions of the homogeneous problem and a forced solution calculated in the locally homogeneous approximation. As a result, the problem of calculating the longitudinal distributions of the amplitudes of the normal modes is reduced to the solution of an infinite-dimensional inhomogeneous system of ordinary differential equations. The numerical solution obtained using this method is tested by comparison with an exact calculation based on the Fourier method. Using the method proposed, the problem of flow receptivity to harmonic oscillations of parts of the channel walls is analyzed. The calculations performed show that the method is promising for investigating the receptivity of longitudinally inhomogeneous flow in a laminar boundary layer.     

Horizontal mixing layer in shallow water flows

Horizontal-shear thin-layer homogeneous fluid flow in the open channel is considered. A one-dimensional mathematical model of the development and evolution of the horizontal mixing layer is derived within the framework of the three-layer scheme. The steady-state solutions of the equations of motion are constructed and investigated. In particular, supercritical (subcritical)-in-average flow concepts are introduced and the problem of the mixing layer structure is solved. The proposed model is verified on the basis of comparison with a numerical solution of two-dimensional equations of shallow water theory.     

Density-stratified Sadovskii flow in a channel

Stably density-stratified and nonstratified flows in a channel past a pair of symmetrical closed-streamline vortices on the channel axis are considered. The numerical results obtained cover the whole range of subcritical stratification and eddy lengths. An asymptotic solution for a very long closed-streamline region is found. The results can be used directly in the asymptotic theory of separated flows at high Reynolds number. Sadovskii flows are plane potential inviscid flows past a pair of closed-streamline regions of uniform vorticity. The flow velocity may be discontinuous at the boundary of the closed-streamline region. The analysis below is restricted to the specific case of continuous velocity distribution, so that the Bernoulli constant jump at the eddy boundary is zero. Unbounded nonstratified flows of this kind were studied in [1, 2]. Calculations of the corresponding channel flow were restricted to relatively wide channels. Closely related problems were also considered in [3, 4].Translated from Izvestiya Rossiiskoi Akademii Nauk, Mekhanika Zhidkosti i Gaza, No.4, pp. 118–123, May–June, 1993.     

Secondary flows in a layer with a free surface

The loss of stability of a plane-parallel incompressible viscous heat-conducting fluid flow in a horizontal layer subject to a longitudinal temperature gradient is considered. The lower surface of the layer is assumed to be rigid, while the upper one is free with a surface tension coefficient depending linearly on temperature. Both boundaries are assumed to be thermally-insulated. The critical value of the temperature gradient as a function of other relevant parameters is determined by analyzing the spectrum of the linearized problem. Secondary flows arising after the onset of instability are determined from an analysis of the full nonlinear problem using the expansion of the solution in a power series in terms of a supercritical state parameter in the vicinity of the bifurcation point. Three types of secondary flows are studied: plane two-dimensional waves propagating along the temperature gradient; plane waves travelling at a certain angle to the gradient; and three-dimensional waves propagating along the gradient. A numerical method of problem solution, based on the polynomial approximation, is described.Moscow. Translated from Izvestiya Rossiiskoi Akademii Nauk, Mekhanika Zhidkosti i Gaza, No. 5, pp. 85–98, September–October, 1994.     

Laminar flow through a channel with uniformly porous walls of different permeability

Summary The steady laminar flow of a viscous incompressible fluid through a two-dimensional channel, having fluid sucked or injected with different velocities through its uniformly porous parallel walls is considered. A solution for small suction Reynolds number has been given by the authors in a previous paper. The purpose of this paper is to present a solution valid for large Reynolds numbers for the cases of (i) suction at both walls, and (ii) suction at one wall and injection at the other. A technique of matching outer and inner expansions is used to obtain an asymptotic solution for both of these cases. Further a perturbation solution for the case of suction at one wall and injection at the other is obtained by choosing the difference between two wall velocities as the perturbation parameter. Both asymptotic and perturbation solutions are confirmed by exact numerical solutions. As expected, the resulting solutions show the presence of the usual suction boundary layers in both types of flow considered in this paper.     

Motion of a circular cylinder in a viscous liquid between parallel plates

The two-dimensional motion of a cylinder in a viscous fluid between two parallel walls of a vertical channel is studied. It is found that when the cylinder moves very closely along one of the channel walls, it always rotates in the direction opposite to that of contact rolling along the nearest wall. When the cylinder is away from the walls, its rotation depends on the Reynolds number of the flow. In this study two numerical methods were used. One is for the unsteady motion of a sedimenting cylinder initially released from a position close to one of the channel walls, where the Navier-Stokes equations are solved for the fluid and Newton's equations of motion are solved for the rigid cylinder. The other method is for the steady flow in which a cylinder is fixed in a uniform flow field where the channel walls are sliding past the cylinder at the speed of the approaching flow, or equivalently a cylinder is moving with a constant velocity in a quiescent fluid. The flow field, the drag, the side force (lift), and the torque experienced by the cylinder are studied in detail. The effects of the cylinder location in the channel, the size of the channel relative to the cylinder diameter, and the Reynolds number of the flow are examined. In the limit when the cylinder is translating very closely along one of the walls, the flow in the gap between the cylinder and the wall is solved analytically using lubrication theory, and the numerical solution in the other region is used to piece together the whole flow field.This research was supported by NSF DMR91-20668 through the Laboratory for Research on the Structure of Matter at the University of Pennsylvania and from the Research Foundation of the University of Pennsylvania.     

Unidirectional Heat-Gravitational Motion of Viscous Fluid in A Flat Channel with A Given Flow Rate

This paper touches upon an initial-boundary-value problem that describes the unidirectional heat-gravitational motion of fluid in a plane channel in the case of solid immobile upper and lower walls with temperature distribution thereon and in the case of a heat-insulated upper wall. The motion is caused by a joint effect of the longitudinal temperature gradient and given nonstationary flow rate. The initial-boundary-value problem is inverse relative to the pressure gradient along the channel. An exact stationary solution is obtained. A solution of the nonstationary problems in Laplace images is determined, and the results of numerical calculations are presented.     

Investigation of the influence of the conditions of swirling on the structure of a two-phase flow in an expanding channel

Both experimental and theoretical investigations show that the main feature in the structure of sufficiently strongly swirling gas flows is the presence in then of reverse circulation regions whose configuration depends very strongly both on the law of swirling of the flow and the conditions at the entrance as well as on the channel geometry [1–6]. In expanding channels, the occurrence of such regions is most probable in the axial region [7, 8]. In short annular channels for which the characteristic transverse and longitudinal dimensions are of the same order, reverse flow arises in the exit part of the channel along its inner wall [6, 9]. Hitherto, the investigations have been made for single-phase gas flows. The present paper reports a numerical investigation of the influence of particles of a condensed phase on the intensity of the reverse flow and the structure of the gas flow in an annular expanding channel under conditions of thermal, mass, and mechanical interaction of the phases. The method of stabilization was used to solve the boundary-value problem. The system of equations describing the axisymmetric unsteady flow of the two-phase medium was integrated by means of Godunov's difference scheme [10, 11]. The calculations were made for different conditions of injection of the particles of the condensed phase into the channel.     

A stable unstructured finite volume method for parallel large-scale viscoelastic fluid flow calculations

A new stable unstructured finite volume method is presented for parallel large-scale simulation of viscoelastic fluid flows. The numerical method is based on the side-centered finite volume method where the velocity vector components are defined at the mid-point of each cell face, while the pressure term and the extra stress tensor are defined at element centroids. The present arrangement of the primitive variables leads to a stable numerical scheme and it does not require any ad-hoc modifications in order to enhance the pressure–velocity–stress coupling. The log-conformation representation proposed in [R. Fattal, R. Kupferman, Constitutive laws for the matrix–logarithm of the conformation tensor, J. Non-Newtonian Fluid Mech. 123 (2004) 281–285] has been implemented in order improve the limiting Weissenberg numbers in the proposed finite volume method. The time stepping algorithm used decouples the calculation of the polymeric stress by solution of a hyperbolic constitutive equation from the evolution of the velocity and pressure fields by solution of a generalized Stokes problem. The resulting algebraic linear systems are solved using the FGMRES(m) Krylov iterative method with the restricted additive Schwarz preconditioner for the extra stress tensor and the geometric non-nested multilevel preconditioner for the Stokes system. The implementation of the preconditioned iterative solvers is based on the PETSc library for improving the efficiency of the parallel code. The present numerical algorithm is validated for the Kovasznay flow, the flow of an Oldroyd-B fluid past a confined circular cylinder in a channel and the three-dimensional flow of an Oldroyd-B fluid around a rigid sphere falling in a cylindrical tube. Parallel large-scale calculations are presented up to 523,094 quadrilateral elements in two-dimension and 1,190,376 hexahedral elements in three-dimension.