You are here
Simple beam simulation in Abaqus
Wed, 2009-07-29 15:50 - Emmanuel E. Ufuah
Hi everyone,
I have modelled a simple beam using continuum elements to assess the mechanical behaviour. The results are quite good since it could predict deflection close to test data. Using ABAQUS general-purpose shell elements, the deflection is quite small compared to test results ie it appears the material has become more stiff. Can any one give me a clue?
Many thanks in advance,
Emmanuel
»
- Emmanuel E. Ufuah's blog
- Log in or register to post comments
- 9940 reads
Comments
Did you use four-noded shell
Did you use four-noded shell elements? If the deformation is bending dominant and you used four-noded shell, you would have a stiffer model. If this is the case, can you use nine-noded shell element to see if there is any luck?
Emmanuel Yes, I actually
Emmanuel
Yes, I actually used four-noded shell elements with reduced integration-S4R. Pls, I may want to inform you of the type of analysis I am running.
The analysis is thermo-mechanical which can be one way coupling(ie the temperature distribution from thermal analysis is independent of the deformation caused by the loading). However, the deflection is mostly governed by catenary action due to degradation of mechanical properties at elevated temperatures.
I have also used the S8R5 elements but Abaqus could only recognize two section points in the mechanical analysis out of the five I specified. The result is still not in good agreement.
Thank you for your suggestion and I shall be glad to receive further clue in this matter.
A couple of questions. Is
A couple of questions.
Is this a nonlinear analysis?
Can you simplify the model a little bit? For instance, make the temperature distribution through the shell thickness consant and compare the results.
Simple beam simulation
Emmanuel
I used the same S4R element type with the proper BC and the deformation is fairly okay, though not giving a perfect result. The analysis is non-linear and thermo-mechanical. Moreover, the temp. result from the thermal analysis was imported from the odb. file into the mechanical analysis while using the same geometry and mesh density. The same integration rule ( Simpson with 5 section pts.) was also used for both analyses, yet the structural analysis could only pick the first two pts across the thickness. Does this in any way have effect on the analysis result, though the temp. difference thru the thickness is insignificant?
Many thanks
Emmanuel, My
Emmanuel,
My understanding about your boundary value problem is the following:
1. Thermal analysis is carried out first to get the temperature distribution
2. Apply temperature field to do the stress analysis
3. The model has bending moment negligible in comparison of the tension/compression due to cantery effect
Is my understanding correct?
To use the continuum element to do the above analysis is straightfoward. What I feel confused about is that you used S4R element to do the therma analysis. Can you observe the same temperture distribution as that from continuum elements?
I think there is no need to use reduced integration since there is no shear locking in your case.
Emmanuel Thanks a lot
Emmanuel
Thanks a lot Tony. I have been a bit indisposed and had to stay off for sometime. Now I am okay.
Yes, but the catenary action is initiated by increased temperature.
I actually did not used S4R for thermal analysis as they are not compatible. I used DS4 heat transfer elements.
I am using shell elements to reduce the computational cost because my real model involves a number of structural components.
I got the same results using solid elements.
The results from thermal analysis are quite good but those from the mechanical analysis are not.
I think the information regarding the test data is not fully known as this test was done by someone else.
I have decided to model a different ,but similar one for which I have reliable test data.