Instability Errors

[Anonymous]

Hello,
I cannot deal with instability errors (1,2) of nodes when trying to calculate natural vibrations of the 3D truss structure. I already checked connections and it seems like all bars are connected where they should be (I redrew and opened structure model from Autocad .dxf file in Robot and added supports).

Could you please tell me what am I missing? Should I add supports in every joint? Which ones? Or maybe the parameters of calculations are wrong?

Thank you for all your help!

[Anonymous]

Please help me, this is a serious problem for me and I cannot come up with any solution

[StefanoPasquini6790 · Accepted solution]

Very very easy to solve, follow this steps

1) Edit your stucture type (picture below)

2) choose the highlighted type

3) erase your modal analysis

4) define self weight load case;

5) add a new modal analysis and run calculation.

Pay attention to the material of bars

I can't send you my edited model because your version is student limited, follow the previous steps and you will solve your issue.

Let me know if I help you.

Greetings

[Anonymous]

Thank you so much!

Changing structure type into shell really helped, although I can't understand why I should use shell instead of 3D truss? Can you explain this to me? Is it valid only for vibrations? Or can I use this shell structure to general calculations (i.e. wind, icing)?
I defined self weight, just as it is (-Z direction), and in deformations diagram my structure lowered and bars "melted" - is this the reason I should pay attention to material? 
How did you define self weight, so that the structure deflects on one side? //edit: "How did you define self weight, so that the structure deflects on one side" - I think I haven't selected modal combination...

Should the structure really look like this? (attachment)

Thank you once again! I'm waiting for your reply

Greetings

[Anonymous]

i don't know when i can use "release" ,BFGS,ITERATION,NIUTON. Which type of structure must use release? Sometimes, i have been getting these problems.Please give to me some tips. Thanks so much.

[Artur.Kosakowski]

In most of the cases you use releases when you a connection between elements is int intended to transfer bending and they can be used for frame and shell model types (not truss ones). You may switch from BFGS to FNR when you have difficulty to obtain results for divergent nonlinear analysis as the later method has in some cases better convergence.

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.

[Anonymous]

I'm a little confused now...so my truss tower,now being a shell type, should have released connections? Still don't know why I should've changed it's type for shell...

[Anonymous]

i don't understand.I try to more times but i can not correct model.

http://www.mediafire.com/view/pes4yqsq5jknr3n/test.rtd

[Artur.Kosakowski]

You should avoid the situation where you have all bars having releases in their common node. I would try to delete releases from the columns and replace fixed supports with the pinned ones so that the model can be calculated with no errors. Then you can check the values of bending moments and see if some additional releases are needed.

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.

[Anonymous]

it's meaning.I have to delecte all releases at all nodes, not use release at bars.What situation should i use release at nodes?

[Artur.Kosakowski]

The aim is to release excessive or duplicated definition of releases rather than not define releases where they are needed. For the model where it is difficult to determine which definitions of releases cause instability error it may be easier if you delete the releases first and then apply them in stages on selected part of a model based on obtained bending moment diagrams.

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.

[Anonymous]

thanks for your advise.I try to fix model.