Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 23

1

GHOST CELL IMMERSED BOUNDARY METHODS

Guide : Dr.Simon Peter


Presented by :

Gugulothu Ram Naik


B190288PE
 Contents : 2

- Introduction
- Background Information
- Literature Review
- Objective
- Methodology
- Results/Discussion
- Conclusions
Introduction 3

Aquatic locomotion Windmill farm configuaration


Introduction : 4

 The computation is even more expensive and time consuming for moving boundary problem
because the mesh must be regenerated each time step to adapt the moving boundary.

 In addition, large deformation of the boundary is hardly achievable since high quality of the
boundary-fitted mesh is not guaranteed.

 To overcome these difficulties, some efficient non-boundary conforming methods, e.g.


immersed boundary method (IBM), have been proposed.

 The IBM has the advantage of easily treating complex fluid–structure interaction problems
since the governing equations are solved on a fixed Cartesian grid.
Introduction : 5

 Immersed boundary method (IBM) was first introduced by Peskin, for boundaries
that interact with the surrounding fluid. He performed two-dimensional simulations
of flow through the natural mitral heart valve on uniform cartesian grid.

 Numerically this treatment is suitable for flexible elastic boundaries, but its
implementation on rigid bodies leads to stiffness issues.

 Solid cells having at least one fluid cell in their neighbourhood are then identified
as the ghost-cells.

  Ghost-cell immersed boundary method (GCIBM), which imposes the boundary


condition by an interpolation of the variables at the ghost-cell, was proposed by
Majumdar et al.
Literature review : 6

Recently, some IBMs have been developed and applied to a variety of problems.

Goldstein etal proposed a feedback forcing approach that required two ad-hoc user specified
parameters, for imposing boundary conditions on rigid bodies.

Boththese approaches required spreading of the forcing around the immersed boundary,
hence are hardly justified for high Reynolds number flows.

The incorporation of various boundary conditions in GCIBM was demonstrated by Tseng and
Ferziger.

Mittal et
al and De extended this approach to problems involving highly complex three-
dimensional and moving boundaries.
Literature review : 7

 GCIBM involves matrix inversion, thereby strongly depends upon the distribution of the
computing cells.

 To circumvent this issue Iaccarino and Verzicco suggested the use of inverse distance
weighting (IDW).

 While Gao et al  demonstrated the use of Taylor series expansion coupled with IDW for
reconstruction of variables at the ghost-cell.

 An extensive review on the developments, usage and implications of IBM can be found in the
article of Mittal and Iaccarino

 Kim and Choi developed an IBM in the non-inertial frame of reference to solve the flow
around an arbitrarily moving boundary with complex geometry.
Objective : 9

 There are numerous advantages of IBM, such as ease of grid generation,


incorporating body motions, CPU memory and time savings, ease of implementing
fast and stable iterative linear solvers and amenability to geometric multi-grid
methods.

 To enhance the capabilities of the GCIBM which was formulated by Gao et al.
Methodology : 8

IBM :
 Equations of motion for an elastic ,incompressible structure immersed within a
viscous,incompressible fluid.

• X=(x , y) : Cartesian Coordinates . U(x , t) : Eulerian velocity field


• F(q , t) : Lag. Elastic force density due to deformations of structure.
• Q : Lagrangian (material) point position
• X(q , t) : physical position of a material pt. q at time t
Methodology 10

Fig 1. Fluid grid Fig2. Immersed boundary Fig 3. Eulerian fluid grid
(Euler grid) (lagragian Frame) Lagragian Immersed
boundary
Methodology 11

At each time step:

 Forces exerted on the fluid by the boundary are


determined using the elastic deformation of the
boundary
 Spread the force from the boundary to the Euclerian
grid to obtain the elastic body force density used in the
next step.
 Solve the equations of fluid motion defined on the Eulerian
grid with elastic body force density f(x,t).
 Move the boundary at the local fluid velocity . Determine the
velocity at each lagragian point through interpolation
Parallelized GCIMB for moving boundaries : 12

 The normalized (using D, U and D/U as length, velocity and time scales,


respectively) governing equations for incompressible flow with constant fluid
properties in the non-inertial frame of reference.

 Re =UD/ (mu) is Reynold”s number


 aij is I’th component of the acceleration of a
 body w.r.t.inertial frame of reference.
Methodology : 13

 Taylor series and IDW reconstruction for GCIBM :

. Fig shows a typical boundary


comprises a ghost cell(G),
immediate neighbouring fluid
cells
cells(F) and boundary point(B).

 Fig 4
Cont… 14

 The boundary point is taken as GB points in the surface normal direction at B.

 The fluid neigbours F1,F2,F3,F4…..Fn+1 are at h1<h2<h3<h4…hn<R from B.

 TAYLOR SERIES Expansion abt. B at ghost point..

+ terms…
where i,j expand for spatial directions.
<Re<21020<Re<210), steady planar symmetric (210<Re<270210<Re<270), unsteady planar symmetric (280<Re<420280<Re<420) and unsteady asymmetric (420<
20<Re<21020<Re<210), steady planar symmetric (210<Re<270210<Re<270), unsteady planar symmetric (280<Re<420280<Re<420) and unsteady asymmetric (420

Cont… 15

 At the boundary points, IDW interpolation can be used for the flow variables and
their partial derivatives.
m,(
 Uniform flow past a stationary sphere :
• Wake behind a sphere is marked by several regimes at increasing Re.
• Laminar boundary at Re= 20
• Four regimes of laminar wake :
1.steady axisymmetric(20<Re<210)
2.Steady planar symmetric (210<Re<270)
3.Unsteady planar symmetric(280<Re<420)
4.Unsteady asymmetric(420<Re<800)
Flow over circular cylinder : 16

 Strain function :

 Potential function :

 Boundary conditions and velocity :


Stagnation pts. Vo=0,
Flow under circulating cylinder 17

 The rotation of a circular cylinder in a uniform flow


is expected to reduce the flow-induced
oscillations 
Equations:
The vortex shedding is completely suppressed beyond a
critical non-dimensional rotational speed
Results : 18

 Strength of suction is constant. * Location of suction varies with Reynold’s Number .


 Non uniform profile improves upon
uniform suction.
… 19


Cd decreases while Cdrms increases upto in fig.

Fig.Comparison of the average and rms values of the drag coefficient for
rotating circular cylinder with reported results .
Contd : 21

 Cl increases while CLrms is const. for in fig .


 CL continuously increases with for due to
magnus effect.

Comparison of the average and rms values of the lift coefficient for rotating circular cylinder
with reported results.
Conclusion 22

 The treatment of non-inertial frame is also added in the modified GCIBM-IDW method to
deal with complex moving boundary problems in three dimension.
 The application of IBM to moving boundary problems in the inertial frame of reference
encounters spurious force oscillation.
 The use of non-inertial frame avoids these difficulties as the body remains stationary, and
only the boundary moves which is accommodated in the boundary conditions
 The present work has extended the previously proposed IDW technique for
reconstruction on the immersed boundaries to accommodate zero-gradient pressure
situation
 The methodology is applied to three-dimensional flows at moderate Reynolds number 
 The developed method is tested on flows past transversely oscillating and rotating
cylinders
References : 23

 Peskin C S 1972 Flow patterns around heart valves:a numerical method. J. Comput. Phys. 10: 252–
271
 S Peter and A K DE*,A parallel implementation of the ghost cell immersed boundary method with
application to stationary and moving boundary problems; https://DOI 10.1007/s12046-
016-0484-9
 Tong Gai, Yu-Heng Tseng and Xi-Yun Lu, An improved hybrid Cartesian/immersed boundary method
for fluid-solid flows; https://DOI: 10.1002/fld.1522
 Yu-Heng Tseng, Joel H. Ferziger, A ghost cell immersed boundary method for flow in complex
geometry; https://www.sciencedirect.com/science/article/pii/S0021999103004108
 A parallel implementation of the ghost-cell immersed boundary method with application to stationary
and moving boundary problems | SpringerLink
 Goldstein D, Handler R and Sirovich L 1993 Modelling a no-slip flow boundary with an external force
field. J. Comput. Phys. 105: 354–366

You might also like