CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CD-adapco (
-   -   Fluid forces on exhaust valves (

Lars Ola Liavåg August 30, 2000 07:36

Fluid forces on exhaust valves
Dear colleagues,

I am studying the gas forces acting on the exhaust valves of a 4-stroke diesel engine. The intention is to apply the CFD results as BC's in a FEM analysis aimed at determining the valve deformation. Thus, the resultant force and its point of attack must either be calculated directly from the CFD results, or the wall force data output by STAR-CD (which appears to be the gas force in N acting on each surface shell) must be converted to force-per-area vertice data.

If anyone has experience from this sort of manipulation with STAR-CD results (and how to export them to I-DEAS), your advice would be very much appreciated.

Lars Ola

Robert Bland August 30, 2000 11:38

Re: Fluid forces on exhaust valves
Might I suggest you transfer surface pressure values -force is mesh dependent (use GETB to get the surface pressure and then CAVERAGE these values onto the vertices of the wall shells).

In general you would pull your IDEAS mesh into STAR, use PMAP to map the results across from STAR and then write out a list of force application commands for IDEAS with the GENERIC command

Lars Ola Liavåg August 31, 2000 01:58

Re: Fluid forces on exhaust valves
Hello Robert,

If the contribution of the static pressure to the total gas forces is much larger than that of the shear forces, I guess porting of the results to I-DEAS becomes simpler since I don't have to play around with vectors.

Nevertheless, I would still like to assess the magnitude of the shear force part. As you mention, force is mesh dependent, and interpreting the results as displayed in PROSTAR is therefore not very easy. It would be of value to transform the cell force data into force-per-area data on the vertices regardless of whether or not the results will be ported to I-DEAS. Any suggestions on how to do this?

Lars Ola

Robert Bland August 31, 2000 17:06

Re: Fluid forces on exhaust valves
You could normalise the forces by the cell area (using OPERATE) before mapping instead of using pressure. You could also use PSYS on the forces if you, for example, wanted to map in a cylindrical co-ordinate system.

If the grids had broadly similar densities then you may be OK using this approach - the mapping is not conservative so you need to be careful. An overall force level comparison of the original and mapped object should give a first order estimate as to whether the mapping has been successful. Contour plots would also help.

Lars Ola Liavåg September 1, 2000 06:59

Re: Fluid forces on exhaust valves
Thanks for the advice. I loaded in the cell areas and divided the forces by them obtaining a fairly smooth distribution of surface stress. However, I have a couple of questions regarding PROSTAR's handling of wall forces. When using the getw,sftot command, I get exactly the same forces on the elements as when using getw,fxyz - regardless of the pressure option used. This along with the fact that the lowest value of the surface stresses occurs under the valves, where at the same time the pressure is known to be largest, leads me to think that the actual total force is not at all calculated by PROSTAR in my case. So: ??????????

Another thing is that after performing the post register operations on the force components in registers 1-3, PROSTAR responds to the replot command by saying that no vector data is available. typing wplot works, but the interactivity of the GUI is somewhat hampered by the disliking of the replot command. Is this a known problem?

Yet another question: In order to map the results on to a (coarser) I-DEAS mesh, do I have to use the cave command first?

Lars Ola

All times are GMT -4. The time now is 13:58.