Problem with load combinations and nonlinear analysis, high rise steel structure

[cenationgg]

Hi Team and Users. I have here a problem with loading manual combinations. Model contains 3d steel framed multi-storey building with composite floors. There are many instabilities, "no convergence of nonlinear problem (Link didnt help). I cut on time of calculations, 13h with freezing to ~20min. I dont know if the composite floor works fixed-fixed with beams and columns. I want to compare fixed steel frame to another model-stell tube-in-tube structure, fixed-fixed between composite floors - columns in terms of a bim abilities project, costs, time. Grouped live loads in automatic combinations-manual mode cause a big freeze to my pc with no effects. I assume to change profiles for beams and columns after mes calculation floor after floor. A the main character of "hannibal" says: This is my design" DOWNLOAD. If you have ever seen the solution, which I did not find at this forum yet, just bring it please. Best regards, Bartek

[cenationgg]

After a 3hours sleep back here again and the model deformations show my model as total unstable, which i dont know how to interpret with adding fixed-fixed connections everywhere.. look! I hope that together we can solve this thing in a moment!

[mustafahesenow]

Can you attach the model please

[Artur.Kosakowski · Accepted solution]

Hi Bartek,

I'd say that a good strategy is to start with a smaller scale model first. I'd consider deleting all but two bottom stories and then calculate such model focusing on releases and perhaps missing bracing etc.

I hope this helps.

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[cenationgg]

Thank you for answer mustafahesenow: model size is above 400MB, attachment did not work. I also saw information that there is a way to upload here such a big model, but i case of my own safety, i uploaded file to my google account and shared in a DOWNLOAD link.

Thank you for reply Artur.Kosakowski: that's it, the tactics is right. I constructed simply a single floor with composite beams that calculated properly in Robot addin composite beams module, with steel, composite beams and steel columns as "level 1" floor then i copied this setting with Z axis 50 times. a single floor was nicely working with my test liveload 3kN/m2. After all I made simply a wind simulation based on my hand calc. Then I watched some of your solutions to learn to mesh with IQ, also read some solutions of other people' model problems with nonlinear analysis. Then I placed claddings with a steel s460 panel with 0.1mm thickness. All this fails in calculations, convergence problems. Still no solution to load manual combinations.

Give me a few hours and i will try to construct it again, in a "+2 floors" way.

Hope you have a good day.

[Artur.Kosakowski]

Hi @cenationgg

1. You don't need large displacements analysis. P-Delta will be sufficient enough

2. Set Full Newton-Raphson method (and you may also try to increase tolerances to e.g. 1e-3)

3. If you have the negative stiffness matrix warning increase the sizes of the columns first

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[mustafahesenow · Accepted solution]

1

[mustafahesenow]

1

[cenationgg]

I have another question due to my job/project. Here i upload frame-based model. There are no mistakes now, but the calculation time is about 3 hours, most of this is "attribute assigning". I have seen a few topics on this forum and didnt solve the problem yet. Can you help me solve this? Is there anything I can set in my model to calculate it much quicker <I use now i7 7800x @4.5GHz> ? 

MODEL UPLOAD

[Rafacascudo]

Did you try using the multithreaded solver?

[cenationgg]

Thank you for reply Rafacascudo

Yes, i use multisolver, but i see only 9-13% of CPU usage. Correct me if I am wrong, but I think that i found what i was looking for. Mr Artur Kossakowski explained simply how to avoid generating model every time HERE. So now I have been waiting for 2 hours to see model panels meshing properly, then  I followed on his steps. If I think correctly: with mesh freeze and model generation switched off I can simply put another steel sections to existing beams without attribute assigning every time I make such a change. This improves my workflow heavily. Once again, correct me if I am wrong.

 Once you mentioned "multisolver", I see no improvement while model generating - which takes most of time in my calculations - in "6 core i7 7800x vs 4 core laptop i7 3720 @3,20GHz. Probably because in model generation there is no multithreading in Robot software engine. Yet, I am only 24y old, still most of your knowledge is about to come. 😉

[cenationgg · Accepted solution]

I have a question. With a simpified model of my high rise building - 8 floors - i compared slab models. Look at my results:

slab-flexible diafraghm; IPE180/120x160x3=beam, słup=column

Slab-shell; IPE180/160x120x3=beam, słup=column

The difference is quit a lot, im asking you which panel model is proper in my calculations - shell or flexible diafraghm slab?

I am thinking about flexible diafraghm due to comparing steel frame vs tube structure..