CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   OpenFOAM stress analysis results are showing 8-10% error with analytical solution. (http://www.cfd-online.com/Forums/openfoam/106836-openfoam-stress-analysis-results-showing-8-10-error-analytical-solution.html)

Sargam05 September 10, 2012 12:52

OpenFOAM stress analysis results are showing 8-10% error with analytical solution.
 
Hello everyone,

I am running solidDisplacementFoam problem for thermal stress analysis. My solution is converged but not showing good agreement with analytical solution.

Simulation results shows 8-10% error with analytical solution. Does 8-10% error is fine for solid problems?

Has anybody an idea?

Regards,
Sangeeta

bigphil September 10, 2012 14:32

1 Attachment(s)
Hi Sangeeta,

I presume you are referring to the "hollow cylinder with internal temperature" case as described by Timoshenko, Theory of Elasticity.

I tried this case and I found that the errors become very small (much much less than 1%) as you refine the mesh.

I have attached an graph showing a comparison between the results from a coarse and fine mesh.

To answer your question about "is 8-10% error ok?", in general for solid mechanics, it depends on the assumptions made in the analytical solutions (and in the numerical model). But for relatively straight-forward linear elastic and thermal elastic, the results should be very close to the analytical, especially as the mesh is refined.

Hope it helps,
Philip

Sargam05 September 10, 2012 15:06

Quote:

Originally Posted by bigphil (Post 381100)
Hi Sangeeta,

I presume you are referring to the "hollow cylinder with internal temperature" case as described by Timoshenko, Theory of Elasticity.

I tried this case and I found that the errors become very small (much much less than 1%) as you refine the mesh.

I have attached an graph showing a comparison between the results from a coarse and fine mesh.

To answer your question about "is 8-10% error ok?", in general for solid mechanics, it depends on the assumptions made in the analytical solutions (and in the numerical model). But for relatively straight-forward linear elastic and thermal elastic, the results should be very close to the analytical, especially as the mesh is refined.

Hope it helps,
Philip

Hi Philip,

Thanks a lot for the reply. Actually here I am solving another problem of thermal stresses for OpenFOAM validation. This problem is given in Timoshenko (page no. 399). This is for a rectangular plate.

I am going to email this case to you.

Thanks,
Sangeeta


All times are GMT -4. The time now is 22:42.