CFD Pressure data to FEM (Nastran)

[VN1981]

Hello,
This ain't a new task...probably been done numerous times. I am in need to convert spatial pressure data derived from CFD results in to a format which can be read by Nastran/Patran.

Essentially, the CFD results will be in the format:

X Y Z pressure

My mesh for FEM will not be matching with CFD grid points or it may be difficult to make sure nodal points will coincide with the list of points contained in the above CFD spatial set.

I tried unsuccessfully to create Fields in Patran which then can be assigned to PLOAD card.

It seems like many users have written scripts/programs which will convert CFD data in to Nastran PLOAD format or a format suitable to be used as Spatial Fields function in Patran.

Can anyone experienced with above task either provide me pointers on steps needed to accomplish the above or point to a good resource. I've been not very successful in my efforts to find such materials/resources.

 

[Erik Panos Kostson]

Never done that.

If you are lucky though, and the CFD analysis was done using Fluent, then you can there map the CFD pressures to an imported Nastran mesh (your structural mesh), and export (from Fluent) everything back out to a nastran format. In fluent this tool is called FSI mapping.

 

[karachun]

What is your preprocessor. In Femap you can apply loads from table of coordinates and pressure values via Data Surface -> Arbitrary 3D Data Surface.

 

[JBlack68]

in Patran one used to be able to create an interpolated 3D field - did that a long tine ago!
[ol 1]
[li]one had to import the CFD mesh making sure that the CFD and structural mesh are as close as possible,[/li]
[li]create the pressure distribution and apply it to the CFD mehs (one-to-one relation)[/li]
[li]With Patran create a 3D interpolate field (to the structural node)[/li]
[/ol]
interpolation was a "hit-and-miss" affair sometime and could be extremely slow (hours) to create the interpolated field

 

[VN1981]

Hi folks,
Thanks for the replies.

The CFD data will be generated by using a proprietary code developed by a vendor of ours. What I've learnt interacting with the vendor is that, the CFD code will divide the surface, lets say a wing in to strips (both along chordwise as well as spanwise...so each cell will be more or less a square or triangle). The code integrates the pressure values over this cell and outputs the resultant value at the centroid of the cell.

So the output will be X,Y,Z,Pressure where x,y & z = coordinates of the cell centroid & p = resultant pressure value over one (that) individual cell.

The above integration cell is totally different from CFD Mesh/Grid cells.

The preprocessor has to be MSC Patran. Further, a mesh sensitivity study has to be conducted for the above pressure loads, which means that element size will be varying with different runs.

Karachun, Femap is a great preprocessor but I am afraid we will have to use Patran for this particular application at least.

JXB0809, it seems the feature/process you are describing is a rudimentary version of what Femap has as suggested by Karachun. Where can I find this 3D interpolate field (is it the same which lies inside of Fields->Spatial->FEM->Options->Interpolate->Project 2D space to 3D?

I guess another option is to create points at the supplied coords & associate the created points to surface & hence creating a mesh seed point. But I cannot apply pressure load on individual points/nodes.

 

[Erik Panos Kostson]

The mapping of the loads is non trivial as you can see (I have written some small code for that and it took many days).

There seems to be a MSC toolkit that does this, so you might be luckier.

Link