Single Angle Orientation in RISA 3D

[chris428]

Hello,
I've run into a situation in RISA that has me confused. This could be a dumb question as I'm relatively new to the software. I'm designing an x-braced frame for a piece of equipment. I would like set the members to “Both Ways.” I'm using single equal leg angles as diagonal x-bracing. I have out of plane loads along with axial loads. I understand that RISA uses the principal axes for the bending unity check. I also know that I can force it to use the geometric axes by setting LcompTop/LcompBot to 0.

My angles will be oriented with the vertical leg pointing down. Graphically this can be drawn 2 ways. I can set the x-axis rotation to either 90 or 180. This results in either the local y or local z axis, respectively, sticking out of plane. I get different unity check results depending on which orientation I choose. I believe, this is due to the orientation of the y' and z' principal axes.
I could figure out which orientation (90 or 180) is more conservative and go with that but is that overly conservative considering my angle is actually 45 degrees off from the worst case principal axis orientation? It also made me question my unbraced length assumptions.

If the angle is oriented with the local z sticking out of plane, I am considering the member is braced by the other diagonal from buckling about local z axis but not braced out of plane (about the y axis). Thus Lbyy would be full length and Lbzz would be 0.5 length. With this line of thought, the opposite should be true if I set it up with the local y axis sticking out.

Local z-axis out of plane:
Lbyy = Full Length
Lbzz = 0.5 Length

Local y-axis out of plane:
Lbyy = 0.5 Length
Lbzz = Full Length

However, if is Lbyy is actually Lbyy’ and Lbzz is actually Lbzz’, the other diagonal does not actually align with a principal axis so I’m not sure how to specify unbraced lengths??? The other diagonal lines up with the geometric axis and not the principal axis.

 

[JoshPlumSE]

Chris -

I have a few comments / questions on your issue:
1) If this is an X braced frame then the primary loading should be axial force. Axial strength of the member should not depend on the orientation at all. Just on the unbraced lengths that are specified.
2) The orientation of the member will affect the bending checks of the member, of course. But, this should solely be due to the member's own self weight.
3) Single angle X bracing connected at the center of the X is a classic engineering dilemma regarding what unbraced lengths to use. Not sure I can provide total clarity here.... ASCE has some documents on transmission line towers that talks about this. TIA has some documents on communication towers that talks about this as well.

Some things I would keep in mind are the following:
a) The members are oriented related to their geometric axes (yy and zz). But, the default bending behavior is about their principal axes.
b) RISA uses the prime (') as a flag to say that we're talking about principal rather than geometric axes. The yy' axis is the axis with minimum moment of inertia. So, this is the one that it will normally buckle about. The zz' axis is the one of maximum I.

Maybe bolting them together can prevent yy' principal axis buckling somewhat, so that it could be based on 50% of the total brace length. That's not an unreasonable assumption.

My tendency (as an engineer) would be to also calculate the geometric axis buckling out of plane based on the full member length and use the larger of this or the 50% yy' buckling value.

 

[chris428]

Thanks Josh! I really appreciate your help on this as it’s had me all turned around. I think that I follow what you’re saying. I’m misapplying unbraced lengths.
Regarding your three comments/question:

1) I agree. I think I was misunderstanding Lbyy vs Lbzz. Based on the diagrams in the user manual, I thought that if I had a support (other member) perpendicular to the local-y axis that braces the member from moving laterally in the local z but allows it to rotate about the local-y axis at that point that I should adjust my Lbyy. Likewise, if my support is perpendicular to the local-z, I thought that I was supposed to adjust my Lbzz. Using that basis, I set it up as shown below. My thought process might actually make sense for everything except angles due to their geometric and principal axes not aligning?

2) Besides self-weight, I’m also applying wind load to the members. This particular brace is on the leeward side of an open frame. As you can see below, the physical orientation is exactly the same but my unity checks are completely different. Unbraced lengths and orientation are the only changes. I should mention again that I'm using equal leg angles.

3) That’s for sure. I don’t want to bring up that whole dilemma. I just want to make sure that I’m accurately modeling my assumptions. Whether my assumptions are correct or not is probably another thread. That said, I felt good about using 50% length in-plane and 100% length out of plane until I realized that the other member would not be perpendicular to either z’ or y’ but 45 degrees from both. That leads me to consider maybe 70% Lbyy and 70% Lbzz. I think I will try the geometric axis buckling as you mentioned and use what ends up conservative.

 

[JoshPlumSE]

Yes, I think your thought process makes sense for other shapes. Even for single angles if they were bent about their geometric axes.

But, the unbraced lengths entries for this model are based on the principal axes and that's what's causing the difference. Lbyy' is always going to be the weak axis with the lowest I value. and Lbzz' is going to be the strongest I value. So, that's very different that what you'd get with geometric axis behavior which would likely yield identical results.