CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   FSI (

raj September 12, 2007 02:27

hi all!

could somebody give me an idea about vat is this FSI? is it something like merging of the FEM and FVM codes?

if it is so, it is of great help to resolve the reentry problems and BL-shock-Surface problems!

Any body in this forum, who worked in this area(FSI), please let us know your experience with it!

thanks in advance, raj.

ganesh September 12, 2007 02:34

Dear Raj,

To the best of my knowledge FSI stands for Fluid Structure interaction. For many typical practical applications, the flexibility of the structure becomes quite important and the assumption of rigidity is not good enough. This leads to a coupling between the fluid and the strucutre, with the fluid exerting forces on the structure, the structure getting deformed and the forces exerted by the fluid again changing and the whole cycle repeats. This problem is a complex one requring both CFD and CSM, with suitable coupling. A good example of FSI in aeronautical applications is the flutter problem. There are applications elsewhere also, and this clearly has nothing to do with hypersonics or shock-BL interactions etc ....

Hope this helps



raj September 12, 2007 04:00

thank you very much ganesh, but i mentioned the reentry problems, as i thought like heat dissipation over the surface of the body obviously needs thermal analysis using FEM. so, a one-go solution can be obtained, if we can merge the two codes. not just the heat transfer cases, even for strucural analysis also. in fluid flow analysis, we just treat the boundary as a surface only.

please give me your opinion on htis view also.

thanks and regards, raj.

abouziar September 12, 2007 05:01

To use different codes (FVM for fluids and FEM for structures) is not good for some problems, there is some generation of parasite ossilations and instability, it was when I tried Godunov for fluids and Wilkins for structures. The FSI solver for such case is not good (diffenerent stencil for space and time). More preferable to use also Godunov type for CSD and exact Riemann's solver for FSI, it is stable for impacts problem (from m/sec to km/sec) and for explosion loading and acoustic pressure level

raj September 12, 2007 05:43

dear abouziar,

it was a nice information, and i thank you for that. but i think there must be a definite way to approach the fluid and solid interaction problems. i hope somebody of this forum will reveal it to us. thanks in advance......raj

abouziar September 12, 2007 06:18

it is a coupling of 2 schemes (for fluids and structures), if they are different - stencil, accuracy, conservative nonconservative, explicit implicit it is rather difficult to make such contact stable, it depends how far youhave to calculate, sometimes it seems like stable. When there are the same equations and scheme for both, it is like equations with dicscontinuities of coefficiants and theoretically such contact is to be stable is the scheme is stable.

Anton Lyaskin September 12, 2007 08:50

The thing you're talking about is called "conjugated heat transfer". It's already supported by all major CFD codes. Usually it uses FVM (not FEM) to solve for heat trasport equation in solids.

barracuda September 13, 2007 16:23

Do you want to use the temperatures to compute thermal stresses? This is usually done as a two-step process:

1) solve a conjugate heat transfer problem to compute the solid temperatures

2) map these temperatures to a FEA model and compute the thermal stresses

SANTOSHKUMAR REDDY September 17, 2007 05:25

basicaly FSI is difficult n complex phenomena, bcoz available codes are not so strong to predict exactly what is what?, but ADINA Software is very good for FSI Problem,

Neazen September 20, 2007 11:06

We speak about fluid-structure interaction when the fluid flow evolution depends on structure's one and vice-versa. In other words, the structure behaviour or form is altered because of the action of the fluid stress and conversely, the fluid follows the structure displacement. (It's a definition that you can find in the majority of papers and books discussing this matter)

In order to simulate FSI problems the common approach is to use two solvers, one for the fluid and one for the structure, plus an algorithm describing how the solvers communicates at the common boundary. You can use one of the approach available (FEM, FDM, FVM...) depending if you want it to be "easy" or not.

There's no point in talking about the fact that approaches are different or not for the fluid and the structure, the choice is yours.

saygin September 21, 2007 02:32

for this kind of coupling approach have a look to the site

hamid1811 October 5, 2012 13:07

i'm challenging with adina you know users of adina are if is possible please guide me. i want to model a tank that only have a inlet so when we flow inside it.fluid is water. we must see deformation and expading in structure.but when i run my fsi model i see that report: pressure is too large ..... . i understand that my model is full of water so how i can set initial condition that my volume be empty at first.
thank you

MakisH December 7, 2017 18:26

In FSI one needs to simulate multiple phenomena at the same time. One approach followed by some tools is to have one solver for the complete problem, which solves all the equations in the regions this is needed.

Assuming that you have two different domains (e.g. solid and fluid), which have a shared boundary (their interface), you can simulate them using a different solver for each domain. You may have many reasons to do that, as you may already have very good tools for the fluid or the solid themselves.

In the latter approach, you need an additional piece of software to connect the two solvers. This can be MpCCI (mentioned above), which is commercial, or preCICE, which is free software and has some additional nice features. There are also other tools, such as EMPIRE or OpenPALM.

There are some step-by-step tutorials on GitHub for using preCICE, which may also help you understand FSI:
1. 1D FSI example: Flow through a deformable artery
2. FSI with SU2 and CalculiX: Flow in a channel with an elastic flap
3. CHT with OpenFOAM: Flow over a heat plate

See the resources page for more information.

nourlil December 8, 2017 06:57

1 Attachment(s)

FSI is not complicated you can do this on one calculation by using overset mesh in transcient analysis so the code update by itself value of fluid pressure applied on the solid surface, there is an example.

All times are GMT -4. The time now is 10:06.