Rafter design in portal frame

[Ilight]

Hey everyone:

How do you consider the buckling length of rafter under axial compression? Would it just be the distance between purlins multiplied by a factor of 0.7 ~ 1?

I'm trying to work out the combined action of axial force and bending moment on rafter, however, somehow SpaceGass decides to consider an effective length factor of larger than 1.


Any insight is appreciated!

 

[MIStructE_IRE]

As an aside, since you are talking about axial force, are you considering the “length between purlins” as the actual distance between purlins in general - or - distance between purlins which include a torsional V restraint?

While purlins of course adequately restrain a top compression flange, I’m reluctant to say that purlins sitting on the top flange alone can adequately restrain the member when subject to axial force. I would typically take the length between points of torsional restraint as my length when considering axial. Curious to hear others views.

To answer your question, depending on the code used you may find that your effective length is 1.2 - 1.5L since the structure is in “sway mode”. This was the approach taken in BS 5950 in the past. Eurocode as far as I recall gives alternative methods.

 

[Ilight]

Hmm upon further research, I found out that if the bottom flange doesn't have any restraint, then the buckling length for the analysis would be the rafter length. If I had Fly brace for twist and lateral rotation restraint, would the buckling length become shorter?

I know that the axial force may be insignificant in rafter design, but just curious to know how it is considered in combined capacity.

Also for the effective length factor of greater than 1, I can't seem to picture how a "restrained" rafter can still act like a sway member

 

[MIStructE_IRE]

I would say yes, a fly brace (torsional restaint as I’ve called it above) reduces your effective length for axial.

BS 5950 used the Le > 1.0 approach to account for Second Order effects in sway frames. Even though the member is restrained, it is not part of a braced frame and is defined within that particular code as in “sway mode”. There are no doubt many approaches to this. AFAIK Eurocode permits Le = 1.0 if you’re considering second order effects separately.