Roof bracing/wind truss analysis

[DP202405]

Hi

Do you guys analyse the roof bracing using 2D analysis assuming the pin supports provided by the lateral resisting system (say portal frames) OR
do you run 3D model with lateral resisting systems in?

I find out that if say my portal frames are relatively soft, the 2D analysis will underestimate the axial force by a lot. Not sure why 2D analysis is still commonly accepted.

 

[STpipe]

I would think that pin supports would not be the appropriate boundary condition for a 2D analysis. I think springs are more appropriate, with the spring value representing the lateral stiffness of your vertical LRFS elements.

 

[Eng16080]

I find out that if say my portal frames are relatively soft, the 2D analysis will underestimate the axial force by a lot.

If soft means less stiff relative to other lateral force resisting elements (like shear walls, for example), I would expect the reality to be that the portal frames end up with less load going to them in relation to the stiffer elements.

Not sure why 2D analysis is still commonly accepted.

Because it's simpler and that alone is a huge advantage. While a sophisticated 3D analysis software might give a more realistic result, it will almost certainly be more difficult to use and easier to mess up versus a 2D analysis, and the 2D analysis will likely be more complicated than a simple hand calc. or a spreadsheet or a design table. The simplest approach which gives a reasonable solution is preferable in my opinion.

 

[KootK]

Depends on the particular framing scheme being considered. There are plenty of setups where I do feel that 2D with pinned supports would be appropriate. It comes down to being able to accurately predict the behavior prior to modeling.

 

[STpipe]

KootK,

True - like if you're talking about a concrete shear wall. Depending on the configuration of the building, the deflections in the shear walls might be small, which can be closely approximated with pinned supports. But like Eng16080 said, if there are different elements that have varying stiffnesses, then a pinned support assumption is no longer accurate.

 

[KootK]

But like Eng16080 said, if there are different elements that have varying stiffnesses, then a pinned support assumption is no longer accurate.

Even that statement is not broadly true. A great many of the horizontally trussed diaphragms that I encounter in practice are simple trussed diaphragms. And a simple span truss diaphragm doesn't give a rat's ass what the stiffness of it's supports are because it's statically determinate. Springs and 3D models only become necessary when some form of indeterminacy / diaphragm continuity comes into play.

 

[STpipe]

Very good points KootK, definitely neglected to include the need for indeterminacy.

 

[KootK]

I certainly don't deny that there are some situations that warrant the consideration of VLFRS stiffness and continuity. And 3D FEM has its place for that.

At the same time, I don't think that it's prudent to place a ban on 2D analysis for horizontal trusses or anything.