Torsional Irregularity

[spaseur]

Can anyone tell me the history of when torsional irregularity was 1st introduced into the building code and more specifically why was the 1.2 limit of (delta max/delta avg) chosen? I cannot see how the difference in deflection at the ends of the building only determine if the building is irregular. I would think it would be depend on the angle of rotation between the ends, which is dependent upon the distance between the points of deflection. I was able to determine that the requirement has been in the code since the 94 UBC.

Thanks for you help.

 

[moments]

There's a 2018 FEMA / NEHRP report, P-2012 “Assessing Seismic Performance of Buildings with Configuration Irregularities - Calibrating Current Standards and Practices”

Direct PDF link is here:

https://www.fema.gov/media-library-...576db50c85ea48448ede5bd90/FEMA_P-2012_508.pdf

This report is a good, free, jumping-off point for the history and the state of the art regarding this factor, known as the "torsional irregularity ratio" (TIR).

 

[spaseur]

Thanks for the link. Downloaded the pdf and read the portions discussing Torsional Irregularity. Some interesting findings and proposals for code changes which seem to make sense. The one thing I did not see was any measuring of the angle of twist in the studies of the different building types. If anyone knows the origins of the 1.2 and 1.4 factors and how they were determined (settled on)

 

[Deker]

The torsional irregularity check was introduced in the 1988 UBC based on recommendations from the SEAOC Seismology Committee. I believe the 1.2 factor was based on judgement that the maximum displacement should not be more than 50% percent larger than the minimum displacement. Since the code equation is based on a ratio it provides insight into how much displacement is caused by torsion compared to direct shear. I don't see an absolute value for angle of twist being meaningful for design since it depends on the stiffness of the lateral system and the force acting on the diaphragm. Are you proposing some kind of ratio for angle of twist, as well?

 

[spaseur]

If I have a building and the deflections at the corners are 1” and 1.6” respectively I am torsionally irregular (1.6/1.3=1.23). If the length of my building is 50, my building rotates 0.012 rad. If my building is 150’, my building rotates/twists 0.004 rad. Are we really concerned about torsion is a building that only twists 0.004 rad? I think that the angle of twist should be a part of the limit. Long narrow building will almost always be torsionally irregular. If my building is long, I am not sure that I am worried about one end of the building deflecting a little more than the other. Six tenths of an inch in 150 is pretty small.

 

[Agent666]

Part of the issue is while the twist is small, if it has a tenancy to twist you can have a bad day if one side of your structure starts to yield and the other doesn't. Inelastic torsional response can significantly change the dynamics of your structure (for example the side that's yielded has the stiffness dropping off, while the other side can continue to take load making the whole system depart from your nice linear elastic model you created for the analysis). I believe this sort of thing is partly the reason for the limits.

 

[spaseur]

Agent666. I agree that torsion can cause some unexpected things to happen. But was is the bases for the 1.2 trigger. The ratio of one end of the building deflection 50% more than the other, as stated by Deker above, does not seem to me a rational limit. I think the angle of twist should be part of the trigger for increasing the torsional stiffness/strength of the building. The FEMA report P-2012 listed above was interesting and there recommendations will reduce some of the triggers when dealing with torsional irregularities. If you know of where I can get a copy of the 1988 NEHRP Recommendations (supposedly this was the first time the limit was incorporated) and the commentary, that would be appreciated. Maybe the commentary has some reasoning behind the 1.2 limit. I have searched NEHRP's website and cannot find the publication.

 

[Deker]

Sensitivity to torsion--which as Agent666 pointed out can significantly impact how forces in the diaphragm get distributed to vertical elements--does not depend on the magnitude of the angle of twist. It depends on how much displacement is caused by twisting compared to how much is caused by direct shear. It doesn't matter if the ratio is 100"/50" or 2"/1"...if torsion is causing displacement at the edge to be twice that at the center of rigidity, it's clear the the building has a low level of torsional restraint. The code attempts to address this by amplifying the accidental torsion and requiring a 25% force increase for diaphragm connections to collectors and vertical elements. Note that these penalties are attempting to address uncertainties in the force distribution to vertical elements.

The code does limit the angle of twist, though not in the way I think you are proposing. The angle of twist is limited by requiring that additional displacements due to twisting be accounted for when satisfying drift requirements for torsionally irregular buildings (ASCE 7 Section 12.8.6). For a constant drift limit, the allowable angle of twist decreases as the building footprint increases.