Instabilities

[simjack]

Hello,

I went through a few "Instability Questions / Answers" on the forum already, but haven't found what I was looking for.

I have modeled a steel structure and chose fixed releases for "trusses" and pinned for the rest of the beams. I fixed the rotation above the objects' axis. I'm guessing I will have to fixed a few more releases to get rid of some instabilities and excessive releases and deflections, is that correct?

Could you please advise on how to deal with instabilities such as what Im getting, because I feel like inaccuracies / mistakes could occur while I am "chasing" releases.

Thank you,

[Artur.Kosakowski · Accepted solution]

Hi @simjack

You shouldn't define releases and pinned supports in the same time. Try to replace supports with fixed ones > calculate the model > check bending moment diagrams > release degrees of freedoms for these supports where it is 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.

[Artur.Kosakowski]

Hi @simjack

I just wanted to check if replacing the pinned supports with fully fixed solved your issue.

[simjack]

Hello. I have changed the supports to fixed and left the members releases undefined (by default fixed), run the analysis and no instabilities were detected. 

However, obviously I'm getting moments developed and transferred at supports which is undesirable. You suggested changing the releases for supports which require it, I tried doing this, however that's all of them? I don't know how to overcome this problem with releases when modelling in Robot, I realise mechanisms are created when members are unrestrained. I really need to understand this so i can use the software with confidence. I guess my question is, how to approach structures like this. Pinned / Fixed / other supports, when and how to use releases on members. 

[Artur.Kosakowski · Accepted solution]

Hi @simjack

Try to replace supports with fixed ones > calculate the model > check bending moment diagrams > release degrees of freedoms for these supports where it is needed.

As in the pinned support you released all rotations when you in addition defined a release for a bar attached to it you sort of 'separate' bar end node from a model as there in nothing what could prevent it from spinning around. As the release prevent any moment transfer to the bar you can then replace pinned support with the fixed one to prevent instabilities and still have no moments neither for the support nor for the bar. Of course for the more complex model there still may be some moments at the supports but then in your next step just redefine these supports where you have them from fixed to pinned or with only RX and RY released or use elastic supports with very small rotational stiffness instead.

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.

[simjack]

Thank you, this was very helpful. It certainly helped me decide on the support fixities. Any recommendations how to repair the rest of the model?

Thank you

[Artur.Kosakowski]

Hi @simjack

Any recommendations how to repair the rest of the model?

What do you actually consider as corrupted and what effect of this correction should be? 

[simjack]

I guess I'm just worried that the results I'm getting are incorrect. So I want to make sure I understand every step. I think you gave me a really good explanation of supports / releases. In sample 4 model I am getting a 3 type instability, I'm not entirely sure when this requires attention and when not. Also I don't know what I have done differently but suddenly I am getting no warnings (sample 5). 

As a next step, I want to perform a footfall analysis. So I ticked DSC algorithm (bar releases). And simply defined a footfall analysis. And again I'm getting instability 3 type, and number of missing eigenvalues warning. 

Vibration analysis is very new to me, I have never done it by hand or software. So to answer your question, I guess my main intention is to design a safe structure and in my mind any instability requires attention and careful consideration.

[Artur.Kosakowski]

Hi @simjack

Please watch the recordings of the 16 and 12 webinars available from here first and then let me know if you still nee any further assistance. Thank you.

[simjack]

Hello, I have actually seen those two before.

[simjack]

That's how I actually found out about footfall analysis in Robot. Thanks for doing these webinars, it's a great learning tool 🙂

[Artur.Kosakowski]

Hi @simjack

I corrected the model lack of accuracy

and there is no missing eigenvalues warning in the mod3 model (if it is displayed just press Yes to run the search for it/them - mod3RXRY). For the instability - you may consider having the columns of the 'vertical' frames to be attached to the 'ground' beams / supports that prevent / limit their out of the plane rotations (e.g. add larger spring for RX and Ry in the RX and RY elastic support definitions receptively).

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.

[simjack]

Hello Artur,

Thank you for your reply. I had a look at the model you sent around. I noticed you split the beams at each intersection and this probably helped resolve the instabilities (type 3). In future, is this something I should be doing for each structure? I have kept the beams continuous for now. 

I wished to apply an area load over the floor, I used cladding to do this (as oppose to specifying floor type). I have split the cladding into smaller zones as I had a feeling it dint distribute the load as I would have thought it would. I am not entirely sure it distributes the load as I want it right now either. What do you think? 

I also run the footfall analysis, (still getting "missing eigenvalues warning"), and as you can probably tell Im not an expert on dynamic design, things that I'm unclear about are; 1) Im getting response factor and frequency of 0.0 for several nodes, is this correct? 2) It seems strange to me that so many elements have frequency of 2.20Hz (and nothing above)? Has the cladding (rather than floor panel) have an impact on the results Im getting?

Thank you.

[simjack]

I have just checked my span direction and realised I defined some incorrectly. This would explain part of my problem 😕

[simjack]

Hi Artur,

Seemingly I have done everything you did for your models and Im still getting an instability. Do you know why?

[simjack]

Hi Artur,

I apologise I'm spamming it here with my messages, but as I go I amend the model and things keep changing. 

Please discard the previous messages, but if you could advise me on the following that would be appreciated. 

1) I had a look at the models you sent around (Sample 3). I noticed you split the beams at each intersection. I have done similar thing in model Sample 7, however on contrary to your model Im still getting instabilities. Do you know what to change to get rid of them? In future, is splitting the members something I should be doing for each structure?

2) "I also run the footfall analysis, (still getting "missing eigenvalues warning"), and as you can probably tell Im not an expert on dynamic design, things that I'm unclear about are; 1) Im getting response factor and frequency of 0.0 for several nodes, is this correct? 2) It seems strange to me that so many elements have frequency of 2.20Hz (and nothing above)?"

Thank you. 

[Artur.Kosakowski]

Hi @simjack

Seemingly I have done everything you did for your models and Im still getting an instability. Do you know why?

As there is no instability reported when you delete claddings my assumption is that this is due to such small calculation elements:

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.