High Deflection Problem with material non linear analysis

Question:

Good Morning,

I am writing you because I am experiencing issues with a model I am working on currently. You can please find below a link to download the model in question.

The issue is I am getting very high deflections when I run Non-Linear Plastic analysis. The high Deflection are occurring just in the area where I have applied the plastic properties (Bottom flange of cross girders in this case) and just under static loads which have been derived from load combination.
For simple static load cases like self-weight these high deflections are not occurring.



Would you be able to assist me with this. I am looking forward to hear from you.

Kind Regards,


_________________________________________________________________________________________________________________________________________


Answer:
Dear User,


The abnormal deformation are mainly because, the non linear analysis is not converging.
Following points are to be checked when it involves complex meshing.

1) Every time , when involved in models with complex meshing , one can run a simple modal analysis to check for any abnormal mode shape and correct the model appropriately.

In the model attached, it appears from the mode shapes that , deck plate which is cantilevered deforms beyond tolerance and hence hindering convergence.




So you might provide appropriate support condition at this location or delete this portion which I have in the attached model to ensure convergence.






2) For the given loading, the displacement is very large when compared to the dimensions of the plates.
Hence Geometric non linear analysis as well would be required. Or alternatively you may apply the loads on cross beams rather that directly applying on decks plates to avoid huge deflections.




3) The load has been divided into 2 load steps so as to get an idea at what load the yield begins.
For the given yield stress, as can be observed from the convergence criteria , the yielding for member begins at step 1, implying even for the half of the total load applied, the members begins to yield.




4) The Plastic material property used is "von mises" without hardening.
Now, when a load as huge as 18t /m^2 (approx.) is applied on the deck plate of 9mm, it would definitely yield.
Here, since the hardening property is not considered, once the yield strength is attained, the material strains beyond the yield strength abnormally.
On the other hand, when you take advance of the hardening property, you may get gradual increase in strain with respect to stress and hence convergence would be possible.

5) Apart from above, one may try to increase the tolerance for convergence, however accuracy might be compromised.

Based on the above solution I have modified the model. The hardening property considered is approximate, you may choose an appropriate value.

Do let us know if you need any further clarification.
Creation date: 10/31/2017 3:30 AM ()      Updated: 10/31/2017 1:09 PM ()
Files   
DataImage58.png
DataImage74.png
DataImage78.png
DataImage94.png
image001.png
image001.png
image002.png
image002.png
image008.jpg
image009.jpg
Material-Nonlinear-Analysis.pdf