Precise FEM solution of corner singularity using adjusted mesh applied to 2D flow

Similar documents
Application of A Priori Error Estimates for Navier-Stokes Equations to Accurate Finite Element Solution

The effect of irregular interfaces on the BDDC method for the Navier-Stokes equations

On the high order FV schemes for compressible flows

Strömningslära Fluid Dynamics. Computer laboratories using COMSOL v4.4

Backward facing step Homework. Department of Fluid Mechanics. For Personal Use. Budapest University of Technology and Economics. Budapest, 2010 autumn

ANSYS AIM Tutorial Turbulent Flow Over a Backward Facing Step

Computation of Velocity, Pressure and Temperature Distributions near a Stagnation Point in Planar Laminar Viscous Incompressible Flow

Stabilized Finite Element Method for 3D Navier-Stokes Equations with Physical Boundary Conditions

NUMERICAL METHOD FOR FREE SURFACE VISCOUS FLOWS

MESHLESS SOLUTION OF INCOMPRESSIBLE FLOW OVER BACKWARD-FACING STEP

FEMLAB Exercise 1 for ChE366

A NURBS-BASED APPROACH FOR SHAPE AND TOPOLOGY OPTIMIZATION OF FLOW DOMAINS

ALE Seamless Immersed Boundary Method with Overset Grid System for Multiple Moving Objects

Stream Function-Vorticity CFD Solver MAE 6263

Solutions of 3D Navier-Stokes benchmark problems with adaptive finite elements 1

Axisymmetric Viscous Flow Modeling for Meridional Flow Calculation in Aerodynamic Design of Half-Ducted Blade Rows

cuibm A GPU Accelerated Immersed Boundary Method

Simulation of Turbulent Flow in an Asymmetric Diffuser

Flow and Heat Transfer in a Mixing Elbow

Numerical Simulation of Coupled Fluid-Solid Systems by Fictitious Boundary and Grid Deformation Methods

Solving Partial Differential Equations on Overlapping Grids

Terminal Falling Velocity of a Sand Grain

Application of Wray-Agarwal Turbulence Model for Accurate Numerical Simulation of Flow Past a Three-Dimensional Wing-body

OpenFOAM GUIDE FOR BEGINNERS

NUMERICAL 3D TRANSONIC FLOW SIMULATION OVER A WING

THE EFFECT OF THE FREE SURFACE ON THE SINGULAR STRESS FIELD AT THE FATIGUE CRACK FRONT

Driven Cavity Example

Fully discrete Finite Element Approximations of Semilinear Parabolic Equations in a Nonconvex Polygon

Control Volume Finite Difference On Adaptive Meshes

Keywords: flows past a cylinder; detached-eddy-simulations; Spalart-Allmaras model; flow visualizations

Numerical studies for Flow Around a Sphere regarding different flow regimes caused by various Reynolds numbers

CONFORMAL MAPPING POTENTIAL FLOW AROUND A WINGSECTION USED AS A TEST CASE FOR THE INVISCID PART OF RANS SOLVERS

Verification and Validation of Turbulent Flow around a Clark-Y Airfoil

Numerical Investigation of Non-Newtonian Laminar Flow in Curved Tube with Insert

IJISET - International Journal of Innovative Science, Engineering & Technology, Vol. 2 Issue 11, November

PANM 18. Radim Hošek The role of Sommerville tetrahedra in numerical mathematics. Terms of use:

Inviscid Flows. Introduction. T. J. Craft George Begg Building, C41. The Euler Equations. 3rd Year Fluid Mechanics

Numerical and theoretical analysis of shock waves interaction and reflection

Coupling of STAR-CCM+ to Other Theoretical or Numerical Solutions. Milovan Perić

Computational Simulation of the Wind-force on Metal Meshes

Eulerian Techniques for Fluid-Structure Interactions - Part II: Applications

Incompressible Viscous Flow Simulations Using the Petrov-Galerkin Finite Element Method

FLOW SIMULATIONS IN CROSS-FLOW FANS USING THE FINITE ELEMENT AND FINITE VOLUME METHODS

Steady Flow: Lid-Driven Cavity Flow

ANSYS AIM Tutorial Compressible Flow in a Nozzle

Solution of 2D Euler Equations and Application to Airfoil Design

NUMERICAL INVESTIGATION OF THE FLOW BEHAVIOR INTO THE INLET GUIDE VANE SYSTEM (IGV)

Ryian Hunter MAE 598

COMPUTATIONAL FLUID DYNAMICS ANALYSIS OF ORIFICE PLATE METERING SITUATIONS UNDER ABNORMAL CONFIGURATIONS

Computational Fluid Dynamics - Incompressible Flows

Unstructured Mesh Generation for Implicit Moving Geometries and Level Set Applications

Investigation of mixing chamber for experimental FGD reactor

Microwell Mixing with Surface Tension

CHAPTER 3. Elementary Fluid Dynamics

CFD VALIDATION FOR SURFACE COMBATANT 5415 STRAIGHT AHEAD AND STATIC DRIFT 20 DEGREE CONDITIONS USING STAR CCM+

NUMERICAL SIMULATION OF SHALLOW WATERS EFFECTS ON SAILING SHIP "MIRCEA" HULL

Solved with COMSOL Multiphysics 4.2

Ashwin Shridhar et al. Int. Journal of Engineering Research and Applications ISSN : , Vol. 5, Issue 6, ( Part - 5) June 2015, pp.

CFD simulations of blood flow through abdominal part of aorta

Hydro-elastic analysis of a propeller using CFD and FEM co-simulation

SHOCK WAVES IN A CHANNEL WITH A CENTRAL BODY

The viscous forces on the cylinder are proportional to the gradient of the velocity field at the

A parallel iterative approach for the Stokes Darcy coupling

Imagery for 3D geometry design: application to fluid flows.

Optimizing Bio-Inspired Flow Channel Design on Bipolar Plates of PEM Fuel Cells

Case C3.1: Turbulent Flow over a Multi-Element MDA Airfoil

CFD Analysis of a Fully Developed Turbulent Flow in a Pipe with a Constriction and an Obstacle

Using a Single Rotating Reference Frame

Modeling and Simulation of Single Phase Fluid Flow and Heat Transfer in Packed Beds

Finite element method - tutorial no. 1

Modeling a Nozzle in a Borehole

Possibility of Implicit LES for Two-Dimensional Incompressible Lid-Driven Cavity Flow Based on COMSOL Multiphysics

Design optimization method for Francis turbine

Calculate a solution using the pressure-based coupled solver.

Shape optimisation using breakthrough technologies

FLOW PAST A SQUARE CYLINDER CONFINED IN A CHANNEL WITH INCIDENCE ANGLE

Maurice Holt University of California at Berkeley, Berkeley, California

Coupled Analysis of FSI

arxiv: v1 [math-ph] 21 May 2011

FAST ALGORITHMS FOR CALCULATIONS OF VISCOUS INCOMPRESSIBLE FLOWS USING THE ARTIFICIAL COMPRESSIBILITY METHOD

Introduction to Finite Element Modelling in Geosciences

ISSN(PRINT): ,(ONLINE): ,VOLUME-1,ISSUE-1,

NUMERICAL STUDY OF CAVITATING FLOW INSIDE A FLUSH VALVE

Modeling and Simulation of Artificial Hair Cell Sensor for Underwater Applications

Dimensioning and Airflow Simulation of the Wing of an Ultralight Aircraft

Numerical comparison of different solution methods for optimal boundary control problems in thermal fluid dynamics

TUTORIAL#3. Marek Jaszczur. Boundary Layer on a Flat Plate W1-1 AGH 2018/2019

Simulation of Turbulent Flow over the Ahmed Body

PATCH TEST OF HEXAHEDRAL ELEMENT

1 Thomas Chengattu MAE 494 November Project # 2: Transient Simulation using VOF Methods

1.2 Numerical Solutions of Flow Problems

On the order of accuracy and numerical performance of two classes of finite volume WENO schemes

ANSYS FLUENT. Airfoil Analysis and Tutorial

Introduction to CFX. Workshop 2. Transonic Flow Over a NACA 0012 Airfoil. WS2-1. ANSYS, Inc. Proprietary 2009 ANSYS, Inc. All rights reserved.

arxiv: v2 [math.na] 19 Mar 2011

MODELLING OF SUBCRITICAL FREE-SURFACE FLOW OVER AN INCLINED BACKWARD-FACING STEP IN A WATER CHANNEL

CHAPTER 1. Introduction

EXPLICIT AND IMPLICIT TVD AND ENO HIGH RESOLUTION ALGORITHMS APPLIED TO THE EULER AND NAVIER-STOKES EQUATIONS IN THREE-DIMENSIONS RESULTS

arxiv: v1 [math.na] 20 Sep 2016

NUMERICAL SIMULATION OF LOCAL LOSS COEFFICIENTS OF VENTILATION DUCT FITTINGS

Transcription:

Precise FEM solution of corner singularity using adjusted mesh applied to 2D flow Jakub Šístek, Pavel Burda, Jaroslav Novotný Department of echnical Mathematics, Czech echnical University in Prague, Faculty of Mechanical Engineering, Prague, Czech Republic Introduction We present an alternative approach to the adaptive mesh refinement. It is based on the knowledge of singularity near the corner. For steady Navier-Stokes equations we proved in [1] that for nonconvex internal angles the velocities near the corners possess an expansion u(ρ, ϑ) = ρ γ ϕ(ϑ) +... (+ smoother terms), where ρ, ϑ are local spherical coordinates. he local behaviour of the solution near the singular point is used to design a mesh which is adjusted to the shape of the solution. We show an example of 2D mesh with quadratic polynomials for velocity. hen we use this adjusted mesh for the numerical solution of flow in the channel with corners. Model problem We consider two-dimensional flow of viscous, incompressible fluid described by Navier- Stokes equations in a domain with corner singularity, cf. Fig. 1. Fig. 1: Geometry of the channel Due to symmetry, we solve the problem only on the upper half of the channel. Let us denote this domain Ω R 2. he steady Navier-Stokes problem for the incompressible fluid consists in finding the velocity v = (v 1, v 2 ), and pressure p defined in Ω and satisfying (v )v ν v + p = f (1) div v = 0 (2) together with boundary conditions on disjoint parts of the boundary Γ in, Γ wall and Γ out (meaning, in turn, the inlet, the wall, and the outlet part) v = g on Γ in Γ wall (3) ν v n pn = 0 on Γ out ( do nothing boundary condition). (4) We consider kinematic viscosity ν = 0.000025 m 2 /s and v in max = 1 m/s, which give maximum Reynolds number around 760. We don t consider volumetric loads f = (f 1, f 2 ).

Algorithm derivation In [1] we proved for Stokes flow that for internal angle α = 3 π, the leading term of the 2 expansion of the solution for the velocity components is v l (ρ, ϑ) = ρ 0.54448374 ϕ l (ϑ) +..., l = 1, 2, (5) where ρ is the distance from the corner, ϑ the angle. Similar results have been proved for the Navier-Stokes equations. Differentiating by ρ we get v i(ρ,ϑ) ρ for ρ 0. A priori estimate of the finite element error is (cf. [3]) [( (v v h ) 0 + p p h 0 C h 2k v 2 H k+1 () ) 1/2 + ( ) 1/2 ] h 2k p 2 H k (), where k = 2. aking into account the expansion (5), we derived in [1] the estimate r v 2 C H k+1 () ρ 2(γ k 1) 2(γ k) ρ dρ C r, (6) r h where h is the diameter of the triangle of a triangulation h, and r is the distance of the element from the corner. Putting (6) into the a priori error estimate, we derive that we should guarantee [ h 2k r 2(γ k) + (r h ) 2(γ k)] h 2k (7) where h is the diameter of the triangle of a triangulation h, and r is the distance of the element from the corner, in order to get the error estimate of the order O(h k ) and uniformly distributed on elements. After simplifications we used aproximate expression h 2k r2(γ k) h 2k. (8) his lead us in [1] to an algorithm for generating the mesh near the corner in recurrent form h i = h (r i ) 1 γ k, (9) r i+1 = r i h i, (10) i = 1, 2,..., N, where r 1 is the distance of the large element from the corner. Using this algorithm we obtained satisfactory results. Some of them were presented in [4].

At present time, we developed a programme for computing the element sizes directly from expression (7) using Newton s method. his algorithm for mesh refinement is applied to the corner where the channel or tube suddenly decreases the diameter (forward step in Fig. 1). We start with r 1 = 0, 25 mm, h = 0, 1732 mm, k = 2, γ = 0, 5444837. his corresponds to the contribution cca 3 % of individual elements to the global error. his way we get fourteen diameters of elements, cf. ab. 1. i r i (mm) h i (mm) 1 0.25000 0.0600369 2 0.18996 0.0480779 3 0.14189 0.0379492 4 0.10394 0.0294666 5 0.07447 0.0224535 6 0.05202 0.0167410 7 0.03527 0.0121680 8 0.02311 0.0085813 9 0.01453 0.0058366 10 0.00869 0.0037980 11 0.00489 0.0023392 12 0.00255 0.0013434 13 0.00121 0.0007042 14 0.00050 0.0005042 ab. 1. Resulting refinement Design of the mesh In [4] we showed that the best way how to use data given by the algorithm is design of the mesh corresponding to polar coordinate system due to its usage in estimates. We continue this idea and design two dimensional mesh near the corner with singularity as could be seen on Figure 2. Fig. 2: Refined detail of mesh his detail is connected to the whole computational mesh, cf. Figure 3 and 4. Fig. 3: Computational mesh near the corner Fig. 4: Whole computational mesh

Evaluation of the error o evaluate the error on elements we use now the modified absolute error computed using a posteriori error estimates, defined as A 2 m(v1, h v2, h p h, Ω l ) = Ω E2 (v1 h, vh 2, ph, Ω l ), (11) Ω l (v1, h v2, h p h ) 2 V,Ω where E 2 (v1 h, vh 2, ph, Ω l ) is estimate of error on element l, Ω is the area of the whole domain and Ω l is the mean area of elements obtained as Ω l = Ω. Here n means the n number of all elements in the domain. More about the evaluation of error in [2]. Numerical results On Figures 5-8 we present the graphical output of entities that chartacterize the flow in the channel. On Figure 5 there are the streamlines in the channel. Figure 6 with contours of velocity v y shows that the solution is satisfactory smooth on refined area. On Figs. 7-8 we observe how strong the singularity is, both for velocity and pressure (note that here the flow is from the right to the left, to have better view). 0.0105 0.015 0.01 0.01 0.005 0.02 0.03 0.04 0.05 0.009 Fig. 5: Streamlines 0.0295 0.03 0.0305 Fig. 6: Isolines of velocity v y Fig. 7: Velocity component v y Fig. 8: Pressure near the corner

On Figs. 9-11 we show the errors on elements. 0.0097 0.0096 Y 0.0094 0.0093 1.970 2.786 0.496 1.021 1.558 0.311 0.411 0.610 0.172 0.284 0.277 0.209.957 0.285 0.237 0.186 0.265 0.237 0.488 0.309 0.245 0.268 0.243 0.215 0.140 0.345 0.194 0.168 0.3990.499 0.363 0.066 0.341 0.530 0.896 0.082 0.329 0.408 0.793 0.120.733 0.216 0.289 0.122 0.392 0.482 1.222 0.388 0.2420.343 0.596 0.142 0.476 1.070 0.252 0.390 0.782 0.151 0.570 3.776 0.222 0.471 1.380 0.419 1.353 0.499 0.577 1.813 2.495 1.754 1.996 0.0298 0.0299 0.03 0.0301 0.0302 0.0303 Fig. 9: Errors on elements - whole refined area 3 0.174 2 216 0.191 0.192 0.163 0.159 1 0.00949 0.165 0.109 0.169 0.139 0.139 0.125 0.118 0.144 0.112 0.122 0.105 0.085 0.072 0.132 0.149 0.098 0.033 0.127 0.193 0.036 0.042 0.123 0.089 0.110 0.160 0.043 0.054 0.156 0.1030.140 0.194 0.234 0.2 0.00948 0.051 0.068 0.170 0.192 0.238 056 0.085 0.230 0.288 0.00947 0.154 0.205 0.102 0.276 0.02998 0.03 0.03002 0.03004 Fig. 10: Errors on elements - detail

02 0.237 0.134 0.115 0.079 0.032 0.665 0.654 0.521 0.404 2.277 1.912 2.409 2.223 0.274 1.544 1.858 0.585 2.221 0.523 1.574 2.292 2.427 0.652 0.093 0.664 0.104 0.513 0.261 0.407 0.135 0.229 0.034 0.122 0.009498 0.083 0.110 0.10 Conclusions 0.029998 0.029999 0.03 0.030001 0.030002 0.030003 Fig. 11: Errors on elements - with smallest elements Presented results give satisfactory confirmation of developed algorithm. Application of a priori error estimates of finite element method for mesh refinement near singularity is very efficient for our problem what can be seen especially from obtained errors on elements which are very uniformly distributed. Derived algorithm is universal for design of the mesh close to internal angle 3 π. But it 2 affords a way how to generate mesh for different angles as well, in accordance to parameter γ which is necesary to find for each angle. his approach is an alternative to classical one using adaptive mesh refinement, which is still much more robust. Reference [1] Burda, P. (1998): On the F.E.M. for the Navier-Stokes equations in domains with corner singularities. In: Křížek, M., et al. (eds) Finite Element Methods, Supeconvergence, Post-Processing and A Posteriori Estimates, Marcel Dekker, New York, 41-52 [2] Burda, P. (2001): A posteriori error estimates for the Stokes flow in 2D and 3D domains, In: Neittaanmäki, P., Křížek, M. (eds) Finite Element Methods, 3D. (GAKUO Internat. Series, Math. Sci. and Appl., vol. 15) Gakkotosho, okyo, pp. 34-44 [3] Girault, V., Raviart, P. G. (1986): Finite Element Method for Navier- Stokes Equations. Springer, Berlin [4] Šístek, J. (2003): Solution of the fluid flow in a tube with singularity using suitable mesh refinement (in Czech), Student Conference SČ, ČVU Praha Acknowledgements: his research has been supported partly by the GAAV Grant No. IAA2120201/02, and partly by the State Research Project No. J04/98/210000003.