FE ANALSIS MARINE PROPELLER BLADE 1

[Adrian77]

Dear Friends,

I would like to carry out a FE analysis of a marine propeller blade in Cosmosworks (in Solidworks); for that I already have the blade as a surface. I want to load it with a pressure distribution calculated with a propeller CFD softwar. In order to avoid doing that (blade loading)node by node (too much work, not practical), I wonder if is possible loading the blade in a semautomatic way, it is modifying my load file in such a way that the FE program can read it.
Somebody even suggested me using the same mesh from the CFD with the same nodes and loads so I skip exporting the blade surface into Solidworks since I already have the geometry.
I have a text file with the nodes, numbers and nodal loads as pressure coefficient.

Any idea or advice would be helpful!

If somebody want more info let me know.


Thanks in advance for the comments

 

[cbrn]

Hi,

first of all, you won't be able to do anything in CosmosWorks if the model is not a solid one. So, in SolidWorks you will first have to create a solid from the propeller's surfaces you already have.

Secondarily, I know no way (and probably there isn't any) to create a mesh in CosmosWorks where the nodes are at specified locations: the mesher meshes "where he wants".

Third point, loading at individual nodes is not possible in CosmosWorks. You can import loads from neutral files, though (try and see in the Help if your CFD supports some of the formats compatible with CW). Otherwise, you can try and see if a polynomial expression exists that approximates or matches your pressure distribution. This expression must be in the form P=ax^2+by^2+cxy+dx+ey+f, but x and y need not be in the principal cartesian CSYS of the model: you can create an arbitrary CSYS centered and oriented as you like, and refer the variable pressure to it. The problem is how to find a "good" orientation...

Fourth, I suppose you are not running FloWorks; if instead you have it, FW can read in some other neutral formats from other CFD, and then you can transfer pressure loads from FW to CW very easily.

Regards

 

[Adrian77]

HELLO cbrn

Thanks for your help, I see that what I want to do is more difficult than expected, will try to get the pressure distribution equation but first I have to convince the software developer to include it and how to do that.
Will write the future advances,..if any....

Would you recommend some other FEM software for this purpose?

Perhaps the whole Cosmos package...

Best Regards,

 

[cbrn]

Hi,

well, I use Ansys and with this software I can see a way in which you could accomplish your task by using a text file formatted as:
NODE# NODE-X NODE-Y NODE-Z PRESSURE
and a routine written by yourself which could perform the interpolation between the CFD nodal coordinates/values and the FEM analogues.

It would be nothing simple, but feasible and almost software-independent on the CFD side. Moreover, it would use only standard program capabilities, without customizations. I know for sure that it would be feasible because at my workplace some experts have developed years ago an "interpolator" like that.

It's only my personal experience, though... Others can suggest you several other FEA that are programmable in such a way...

Good luck for your project,
Regards

 

[jdesouza]

If you apply a mesh control you enter a parameter in for the size of the elements at a specified region. This creates nodes at these locations.

e.g. for a simple case of a cantilevered beam, if you take a beam of length=10" and whatever you want for width and height, then apply a mesh control over the upper surface with mesh size=0.005", nodes will be created at L= 0, 0.005", 0.010"...up to L=10"

 

[cbrn]

jdesouza,

your comment is valuable, but it won't solve Adrian77's problem: if he wants to avoid interpolation routines, he has to place the nodes EXACTLY where they were in the CFD (on the surface of the blades' model). Most probably the CFD solver is Finite Volumes, so it handles "cells" and "partial cells": where the fluid volume domain intersect the boundary (the blades' surfaces), it will most probably generate partial cells and thus the nodes placing will be not evenly spaced.
Secondarily, there is a problem in applying the pressure distribution.
I would definitively say that this problem is almost out of reach for CosmosWorks, unless Adrian77 talks to SolidWorks representatives in order to see if they can develop a macro jointly with the CFD's developers (or, Adrian77 might also develop a macro by himself using the SDK, but it is far from simple !!!)

Regards

 

[kellnerp]

Adrian,

You have the wrong tool.

 

[Adrian77]

OK FRIENDS,,many thanks to everybody for the comments, all of them are valuable. Perhaps something like NX or Ansys would be better. Will investigate this matter deeply with the software dealers.

Regards,

 

[cbrn]

Hi Adrian,

if you are investigating for new software, I'd suggest that you accurately check if, on the CFD side, you can use FloWorks or some other CFD with rotating frame capability (if not sliding mesh...) which is directly integrated with SolidWorks and which can thus interoperate directly with CosmosWorks.
This is the "reverse" side of your first statement where you wanted to find a structural FEM able to interoperate with your current CFD... Don't know which of the two ways is more practical and/or less expensive, though...

Regards

 

[Adrian77]

Hi cbrn,

The CFD we are using is taylored for propeller and cavitaiton analysis so is very fast and give the data in very useful formats, a CFD code like the one yu mention might be usefull for long term research instead or for analysing a particular case.

Anyhow many thanks for thew comment,

Regards