2-D simulation result does not match

[Anonymous]

I've started with a half-cut of a 3-D model for simulation and the results seem to match well with hand calculation. So to decrease computing time and increase mesh density i've made a 2-D model using the face. Now with the same boundary conditions the result doesn't seem to match at all.

Is there a technique that i am missing? Is there additional constraint that i have to apply in order to achieve the same result? Thank you

-nDoan_Vacco

[John_Holtz · Accepted solution]

Hi,

Assuming that the full model can be represented by 2-D, then you are correct that the 2-D elements should give the same results. My guess is that the Element Definition is not set properly. Please check that the Geometry Type drop down is set properly

• plane strain if the real part is restrained "on the ends" so that there is no strain in the longitudinal direction.
• plane stress if the real part has no stress in the longitudinal direction.
• axisymmetric if the real part can be created by revolving the cross-section around an axis. In Algor, axisymmetric meshes are revolved around the Z axis, and the Y direction becomes the radius.

Otherwise, if you can archive the 2-D model ("File > Archive > Create") and post the archive to the forum, someone will take a look at it.

John Holtz

Product Design Engineer, Algor Simulation

Autodesk, Inc.

[Anonymous]

Hi John,

Thanks for the prompt reply and the help. I guess my problem resides with the Element Definition. Initially i had it set as Plane Stress since i wanted to know the stress within the cross-section but this is the wrong selection. I have chosen the Axisymmetric.My question is can this be used with a specified cylindrical coordinate revolving around the z-axis? Thank you for your help.

[Anonymous]

Thank you John. I was able to get my 2-D results to match with the 3-D model. One question I have is what is usually the percentage of difference between the two results? Again thank you for your help.

[S.LI]

Error analysis/estimate in any field is quite important and not easy.

It's hard to give an accurate percentage here. In general, the difference on displacement should be less than the one in stress/strain, and the difference in simple problems is less than the one in complicated simulations.

Between 2D and 3D, the difference could be from the FEA theory, software implementation or model settings.

My personal feeling is if the difference is larger than 5~10%, we might need to find the reason.