Circular Tank: seismic load generation

[TSOFA]

Hi,

I modeled a circular tank in order to design the steel roof. My structure were modeled at "shell design" (structure type) and for the shell I loaded the respective thickness and material ( 8mm steel plate) as a panel. The problem is that when I ask the program to generate automatically the loads to masses for the seismic design, the self weight of the shell is not generated. I open the "mass table" and the self weight of the panes is nowhere! When I check the deformations, only the roof is deforming, when the rest of the structure (shell) is not. How is Robot generates the loads of plates (panels with thickness and material)  to masses?

Thank you in advance

Sofia

[Anonymous]

Spoiler

Hi

You're right there is no SEISME in my area. But I think the warnings are caused made of the number of modes (420) is too much. you Must reduce it and see. I do it and I have no more warnings.

see attached files

[TSOFA]

If I reduse the modes, the percentage of mass participation will not be satisfied (<< 90%) (dynamic analysis results table). That is the reason that I have so many modes. 

[TSOFA]

at this file is impossible to get the mess right. Why?

[Anonymous]

the file is corrupted you sent and you have duplicated some cases charges. Send us the results of the percentage of the mass participation that you said is not enough we will compare with our results (for the original file).

[Rafal.Gaweda]

Such mesh , such results:

[TSOFA]

I also were able to have such a nice mesh when I created the shell of the roof by revolving a polyline which was following the main beams. You see how when I have only 50 or 100 modes the percentage of mass participation will  be very small? I am sending you my deformation results, do you have something similar? Thanks a lot for the support 

[TSOFA]

I run the model with DSC algorithm inactive and the results were better, but still I think that is something wrong. Can we link this topic with " Meshing problem of circular steel tank", or to close this one, because the last posts are concernig the same thing and I don't want to bother more than one collegue. 

[Rafal.Gaweda]

Your model "AUTODESK TANK ROOF_REV 1 2014 ROOF PLATE TEST.rtd" is inaccurate.

Roof is not connected with wall in many nodes

[TSOFA]

I can't understand why is this happening. To create the shell of the walls, I created a closed contour with all the lower nodes of the main beams (UPN) of roof structure. Then I extruded this countour. Respectively, when I wanted to create the shell of the roof, I created a polyline following the exact geometry of a main beam (UPN) and I revolved the polyline 360degrees with 45 divisions, in order to create the exact same geometry of the support roof structure. The reason that I followed this methodology is because the mesh was much better than the previous one (see at the beginning of the topic). What I should do to correct the mistake?   

[Rafal.Gaweda]

Start almost from scratch. I will show you the way. Wait.

[Rafal.Gaweda]

Adjusting bars: http://screencast.com/t/htwLOMzzYqR

Adjusting panels: http://screencast.com/t/xNlm4SiatUHI

Creating top : http://screencast.com/t/jih9V2vzJX

file with my geometry attached

[TSOFA]

So, I have to use the "snaps" more and correct my model often! Thanks a lot for the support, I run the analysis for the new model now, I will let you know! Thanks again 

[TSOFA]

Hi again 

I decided to try the analysis without roof panel and I deleted it from your model and run it again (not great differences from the one that you sent me). Then, I deleted the wall shell from an older model that I did (2010 version), and I created it again with the same way that you showed me at your videos ( snap settings, correct structure). The results were completely different between the two models ( I could not understand the reason, please help). Then (by accident), I asked from the program to display me the automaticaly generated loads for each load case (i.e. dead load, moving load etc). For my big surprice there were such loads on my structure ( yours too but less, you can check it). I can't understand why these loads were created. I don't think that are loads due to seismic loading ( values and way of loading don't match). Maybe that's the reason that our results are not the same.

[Rafal.Gaweda]

 I can't understand why these loads were created.

There are nodes located on these bars so calcaulatuion elements are generated and loaded from main bar.

[TSOFA]

I created these nodes in order to have a better mesh when I tried to have roof panel. I asked the program not to divide the bar. Could you explain more what kind of loads are they? Moreover, why there are such huge differences of the results between mine and your model? Thanks

[TSOFA]

I would expect to replace the one trapezoid load with two equivalent. Here it seems like he added the equivalent load and now to have two trapezoid loads. Is it a graphical "error", or I should try to erase the one load and how I do this?

[Rafal.Gaweda]

I created these nodes in order to have a better mesh when I tried to have roof panel. I asked the program not to divide the bar.

??????????????

Could you explain more what kind of loads are they?

Trapezoidal load generated on each calculation element on bar

Moreover, why there are such huge differences of the results between mine and your model? Thanks

---

I need your file and "my" file.

Keep in mind that the file I sent you has only correct geometry, not correct loads, support, thicknesses, loads to mass conversion etc. I hope you have adjusted "my" modelbefore runing calculations.

[Rafal.Gaweda]

I would expect to replace the one trapezoid load with two equivalent. Here it seems like he added the equivalent load and now to have two trapezoid loads. Is it a graphical "error", or I should try to erase the one load and how I do this?

Everything is correct. See below.

[TSOFA]

I did all the proper changes to your model (support, panel, etc) but the results between two models are very different, even for the static loads (for simplicity). I am sending you the corrected model and some print screens for the results under static loading.

I also did a test. I erased everything from mine and your model except from the primary beams (UPN). I fixed the first and the last node of every UPN and I run the models. The results were exactly the same! So, when I create the shell something is changing...

[Rafal.Gaweda]

To get the the same results:

- delete FEs not connected to any panels from so called "my" model

- delete bar releases from BOTH models -> it means you have made mistake assigning them in both models