Page 1 of 1

cfd

Posted: 20 Dec 2004, 17:48
by sandeep
can someone pls explain the procedures involved in designing right from
CAD , etc... to CFD and after CFD ,if any . i dont mean to ak u anything about foundry etc , but computer methods.

Posted: 21 Dec 2004, 21:10
by Monstrobolaxa
Well to give you a general idea....very general (I don't fell like writting too much today)....

1 - general concept design - this is more or less defined between all the departments involved in the design of the new car (just a theoretical idea of the car...no computer involved yet).

2 - overall design of the car - this will be done using some kind of CAD software and will give you a close to finish idea of the car....at this point usually teams start running the CFD software....

3 - detail design - at this stage the car has already been designed from an aerodinamic point of view....it's shape is already know and small parts will be drawn.....wheel nuts, washers, bolts, fuel pumps, etc.....at this stage teams will already be using CFD and FEA (Finite Element Analysis)

This is more or less the general idea of the design fases, usually teams have diferent approaches to the design of the car....comunly if you have an engine supplier you'll start designing around the engine...cause you can't risk in spending money and time and the engine won't fit....

So it really varies from team to team.....some teams go from the general concept to the detailed design and at the end put everything together.

To me (at least my approach is) general concept...general aerodynamic concept....basic suspension parametres....monocoque/engine assembly........then to the rest of the stuff....

So (like I previously said) it really is down to the technical director and the way he thinks it's best to do the work.

Posted: 24 Dec 2004, 13:34
by sandeep varma
thanks for the replay , but it would be helpful if someone could make it a bit more elaborate . something more on the heirarchy of cfd , what is meshing etc .

i would like to learn CFD and this is my first step.

i got one more doubt . while doing aero analysis isnt it necessary to know weight if different components and its position which , of course can be later adjusted for best performance


what is the exact difference between CFD and FEA?( i guess fea is force analysis and cfd flow analysis)

Posted: 24 Dec 2004, 15:38
by hardhgear
hello every body,
Fist what is cfd?
1-CFD stands for Computational Fluid Dynamics.
2-CFD uses computers to represent (or model) the flow of any fluid around a body or inside it in cas eof internal flow or it can model the heat
distibution.

It start with an equation which define the system Like streamline equation in case of air flow about any solid body for example and we need to solve this equation over defined domain which we call computational doamin .
the next step is to see what solution technic we have to use to solve the equation depend on the classification of the equation.

Then time for meshing has come we creat a grid over the computational domain and good mesh mean more acurate solution so then solution is try to solve the points of the grid. and try to reach Convergence to get an accurate solution.

this steps done by the software all you have to do is gove the input and analyze the output and try to enhance your design.

FAE is a soultion technic used to solve the grid point and it use in case of complex geomerty i think this is what i know about FEA method.

I think that's good for today.
BYeBYe BYe :D :roll: :shock:

Posted: 24 Dec 2004, 16:53
by hardhgear
I'm sorry i forget smoething .It's the boundry condition after the meshing and befor the solver run you have to define the boundry condtion and this is the input data you enter.
About the equation which define the case we need to solve it's Partial Differntial equation and we change to algebric equation to solve it over the grid and this where we use solution Method Like (FTCS),(CTCS) ................And so on .depend on the equation type.
:? :lol: :lol: :o

Posted: 26 Jan 2005, 04:04
by hardhgear
Computational Fluid Dynamics or CFD is the analysis of systems involving fluid flow, heat transfer and associated phenomena such as chemical reactions by means of computer-based simulation. The technique is very powerful and spans a wide range of application areas.

The CFD process begins with geometry, which has to be either created or imported from an external CAD package. The 3D CAD then forms the framework around which a mesh is constructed. The mesh is a set of small blocks, which fill the volume through which the fluid flows. The mesh for a full race car can contain millions of blocks (or cells). Next, a CFD code applies the governing equations of fluid flow, the Navier Stokes Equations, to each cell within the mesh. The computer process communicates information across all the cells and proceeds in an iterative manner towards solving the problem. After much computation a solution is reached where forces and mass flows balance in every cell, and across the whole flow domain. Once the calculation is finished, the CFD solution contains all the pressures and velocities both on and off the surfaces of the object being studied, this is usually millions of numbers. Making sense of such a large data set is not a trivial task and the science of Computer Visualisation has developed to deal with this challenge. Modern visualisation packages allow the engineer to interactively explore the wealth of data created by the CFD codes. Visualisation is the key to fully exploiting the potential of CFD.

Posted: 26 Jan 2005, 07:43
by sandeep varma
thanks mohammed

i wanted to hav a basic idea abt CFD before i started learning and these served my purposes . thanks a lot

Posted: 06 Mar 2005, 03:10
by bernard
what on earth does that poll have to do with cfd? :P

Posted: 07 Mar 2005, 08:40
by West
Isn't FEA finite element analysis?

NM saw Monstro's post.

BTW judging by Coulthard's performance, I hope RBR Cosworth is the "dark horse" this year.

Disappointed by Williams, although much more for Toyota.

Posted: 07 Mar 2005, 09:25
by ReubenG
Finite Element Analysis can be used for much more than force/stress analysis. It has its roots in structural analysis but it can also be used to solve thermal problems or acoustic problems. FEA can't be used for fluids because the large displacements associated with fluid flow cause a breakdown of the mesh. If you want to read more, I suggest the following:
R. Cook - Concepts and Applications of Finite Element Analysis
J. N. Reddy - An Introduction to the Finite Element Method

[/b]

Posted: 09 Mar 2005, 17:18
by kilcoo316
Mohamed has it pretty much right - only its usually (well - nearly always) not the pure navier-stokes equations that are used - they are much too large for all current computers to use except in very small basic problems - like boundary layer growth over a small area

turbulence models must be used to shorten and approximate the N-S equations, traditionally this was RANS based turbulence models, like k-E based on turbulence kinetic energy and turbulence dissipation rate - however, more modern systems can use Large Eddy Simulation, where only the smallest degrees of turbulence are approximated using the turbulence models - it does give a more accurate simulation - but still consumes much more computing power

Posted: 29 Jul 2005, 16:59
by hardhgear

CFD, FEA or whatever......

Posted: 30 Jul 2005, 08:29
by riff_raff
Sandeep,

The basic process behind FEA, CFD or whatever.... is to take a complex problem and break it down into multiple simple problems that will approximate the complex solution. Since there will always be some margin of uncertainty in the results of a simplified analysis, FEA and CFD analysis results are never accepted as absolute. They usually have a margin of safety applied to their values, or their results are validated by testing.

For the work I'm familiar with (aircraft loads analysis), we apply a MS of 2.00 by analysis and a MS of 1.25 by testing.