FVM Presentation

 Computational fluid dynamics (CFD) deals with
solution of fluid dynamics and heat transfer

problems using numerical techniques.
 CFD is an alternative to measurements of solving
large scale fluid systems.
 CFD has evolved as a design tool for various
industries namely aerospace , automobile ,
chemical and many more.
 CFD is becoming a key-element for computer-
aided designs in industries across world .
FLUID
PROBLEM analysis
PHYSICS OF SIMULATION
FLUID RESULTS
Compu
maths ter
NAVIER- COMP.
STOKES EQN. PROGRAM
Num. Langu
method age
DISCRETIZED
GRIDS
FORM
Geometry
CFD EXPERIMENT
Cost Cheap Expensive
Time Short Long
Scale Any Small\Middle
Info. All Measured Points
Repeatable All Some
Security Safe Not safe
 Aerospace
 Automobile
 Biomedical
 Chemical
 HVAC
 Marine
and many more.

 The basic maths is required to represent the fluid
flow system.
 These is basically done by applying conservation
laws.
 In CFD, we apply two conservations to the system.
i. Mass conservation.
ii. Momentum conservation.
 After the basic laws of conservation, we have to see
how is give rise to N-S equation.
 On the basis of Momentum conservation, We get
 So now as PDEs are our governing equations, so we
should have basic idea about PDEs.
Criteria Details Examples
Order The order of
PDE is set by
highest order
derivative in it.
Linearity If the coefficient
are const. or
independent
variables only,
then eqn. is
Linear.
 Linear second order PDEs: elliptic, parabolic,
hyperbolic.
 The general form pf this class of equation is :
 Where coefficints are either constants or function

of independent variables.
 Basic conditions are:
 So after basic classification, we have to look hoew
this PDEs are assosiated with our physical problem.
 So following are some conditions regarding varous
PDEs.
 The complete N-S eqn. in three space coordinates
(x,y,z) and time(t) are a system of 3 nonlinear
second order eqn.
 So, the normal classification rules do not apply
directly to them.
 Nevertheless, they do posses properties such as
 unsteady ,inviscid
compressible flow. Are
essentially hyperbolic in
Hyperbolic flows nature.
 for steady inviscid
compressible flow are
hyperbolic and if speed is
supsonic , it is elliptical.
 Basically Discretization is a process of converting
analytical equations(PDEs) into Linear algebric
equations.
 There are several methods to do the task, which
are as follows:
i. Finite difference method (FDM) (easy to apply, simple)
ii. Finite element method (FEM) (apply to any geometry)
iii. Finite volume method (FVM) (apply to any geometry)
iv. Spectral Method (combination of standard FDM )
NAME OF THE PROCESS ADVANTAGE DISADVANTAGE
METHOD
 Domain is divided Physical Not as straight
into non overlapping soundness as FDM.
control volumes.
The differential
FINITE VOLUME equaion is
METHOD integrated over each
CV.
Piecewise profiles
expressingthe
variation of
unknown between
the grid pts are
used.
 General form of navier stokes equation is :
 Integral form of navier stokes equation :

 So by conservation of integral form of finite volume
method, we get:-
 While solving FVM we have to consider certain
approximation for achieving ease in the
calculation.
 Various approximatio we use while doing
calculation such as,
i. mass approximation.
ii. Momentum approximation
iii. Interpolation
 Let us consider element P in the grid.
 It is having 4faces adjacent to it (e,n,w,s).
 For a particular elemnt we the help of
approximation we can write it is given form;
 This is nothing new but a linear equation in the
form of : AX=B
 After this step we have successfully managed to
convert PDEs which was our governing eqn. into
linear algebraic.
 Further this linear equation are used to find the
solution of our physical problem.
 Grids are basically the formation in which our
domain should be divided so that it will give the
solution as accurate as possible.
 The type of grids mainly depends on the type of
problem which we are dealing with.
 Grid type also depends on type of approach we are
using to slve PDEs.
 There are mainly 3 types of mesh used:-
STRUCTURED GRIDS
(all nodes have same no.
of elements)
UNSTRUCTURED GRIDS
(for all real bodies)
BLOCK STRUCTURED
GRIDS
(for all bodies)
We will discuss the following points :-
 Modelling of the geometric domain and the
physical phenomena of interest.
 Discretization of the geometric domain into a grid
or mesh.
 Numerical or equation discretization to linear
system.
 The solution of the resulting set of equation.
Physical Physical
domain phenomena
Domain Physical
Set of governing
modelling modelling
eqn.
Domain Equation
discretiz. discretiz.
System of
algebric eqn
Solution
 Domain Discretization:-
 Solvers:-
1) DIRECT: cramer’s rule, Gauss elimination,
LU decomposition.
2) ITERATIVE: jacobi method, gauss seidal
method ,SOR method.
 Numerical Parameter:-
1)Relaxation factor, convergence limit.
2)Multigrid Parallelization.
3)Monitor residuals (change of result betw.
Iterations)
4)Number of iterations and its time steps .
5)Single\double precisions.
 By using these solvers, we get the solution for our
linear systems.
 The values we get are for particular domain and we
put is values in computer systems
 The computes are feeded with certain sets of
instruction.
 On giving these values as input, system stimulate
the results.
 These results we compare with origial occurance of
activity and find the difference between them.
 FVM is typically applied in CFD problems and is not
mainly used in other thermal domain.
 FVM is mainly applied in medium to large scale
problems and FEM is applied in small scale
problems.
 FVM is straightforward to implement on non
uniform/unstructured grid but on simpler
geometry FDM is preffered.
Practical
problem
Identify the
governing eq.
Apply FVM
Convert into
linear sys.
Linear solver
Get the results
Final
simulation.

FVM Presentation

Uploaded by

Copyright:

Available Formats

FVM Presentation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FVM Presentation

Uploaded by

Copyright:

Available Formats

 Computational fluid dynamics (CFD) deals with

solution of fluid dynamics and heat transfer

and many more.

 Where coefficints are either constants or function

 Integral form of navier stokes equation :

Get the results

You might also like