Sap2000 - residential structure, pinned joints assumed everywhere, now no deflections 1

[Couhiced]

I'm modeling a house in SAP2000 with a steel frame that is bolted together. I've assigned rigid joints to the foundation connections, and pinned joints everywhere else.

Now, there are no deflections at the non-internal joints, and my model acts "frozen". But, joint reactions are present everywhere and their magnitudes are acceptable.

If I allow the joints to move freely, all reactions are of course transferred to the rigid foundation connections and the joint reactions occur only at those connections. The flexibility is returned to the structure, but it must be wrong.

Apologies if this question is ridiculous, but what restraints should I be applying to the bolted joints? Why am I not getting global deflections when using pinned joints?

 

[WillisV]

Are you actually assigning pin restraints to each and every node of the structure?

 

[msquared48]

If you restrain all the nodes from lateral movement, of course there will be no deflection.

The other situation is that the structure is seeing no load.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[FlashSet]

Are you trying to output beam end forces? You do not want to pin every node. Run the model and look at the internal forces within the beam elements.

 

[Couhiced]

If you restrain all the nodes from lateral movement, of course there will be no deflection.

Yeah, that's the problem...

The other situation is that the structure is seeing no load.

Loads are working fine when there's movement allowed in the joints.

-------

The part where I'm confused is, if every bolted joint is essentially identical, what's the basis for assuming pinned or free? It seems I'm just arbitrarily picking joints to move freely.

 

[Couhiced]

Are you trying to output beam end forces?

This question emerged when I was looking for base reactions for slab design. The magnitudes of the results were quite large, because all joints were free to move. Also, joint reactions throughout the model were zero, except for the rigid foundations. Danger signs, I thought.

When I pin the joints, the base reaction results are more reasonable (by a factor of 6), the joint reactions return, but the model is unrealistic - there's obviously no deflections anywhere, so the forces are not propagating through the frames as I'd expect.

If it's relevant, the model is in three dimensions. Every joint is bolted, except for riveted light steel trusses. It's a two story structure with roof trusses and shell areas and load combinations per ASCE.

Obviously, I don't want to pin them all. Should I assume translation along Z is restrained? Only pin trusses? Leave everything free to move in all DOF? Is there a precedent, something painfully obvious, or a fundamental tenet I'm forgetting? I've spent a lot of time trying to find more information but to no avail. Also, not many people I can ask "stupid" questions to. Would appreciate any push in any direction - unrestrained

 

[paddingtongreen]

You are not very specific in explaining your problem, are you speaking of the vertical or the horizontal axes? The only reactions you should get on a normal frame are at the column bases. The internal nodes of beams should be fully fixed, only the ends are up for grabs except you must fix enough of them to stop the frame from falling over.

You don't appear to understand that Σf and Σm must both be zero for equilibrium. Add up the total vertical load and check the total vertical reaction is the same. do the same for the horizontal loads.

Your task is not to make the model run, it is to analyze the structure. If you must distort the model till it does not represent the frame, you are wasting your time. You have to look at the member outputs to find the forces. I don't know if your program uses local or global axes for this.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[Couhiced]

OK, thanks everyone. It turns out I WAS doing things correctly and had descended into a deadly spiral of second guessing myself all over the place.

@paddingtongreen's slap in the face brought me back to reality - the equilibrium forces all check out, there's no need to pin anything (except the rigid foundation), and it turns out that the base reactions were reasonable.

Thanks for the help! A little bit balder, but little bit wiser.