Progressive Collapse Resistance

[cmrdata]

In GSA “Alternate Path Analysis and Design Guidelines for Progressive Collapse Resistance”, 2016 edition, the “Redundancy Requirements” is prescribed.

This requirement is different from the “Enhanced Local Resistance” (ELR), as stipulated in UFC 4-023-03 documents, which makes sense intuitively.

Commentary C3.4 in the GSA documents explain that the rationale of the Redundancy Requirements is “…. to provide robust structures that provide some level of redundancy ….”, and it further stated that “Structural designs where progressive collapse resistance is localized to one floor level such as single ring girder or truss system do not meet this objective ….”

What I don’t understand is that, when we perform the alternate path method design with column removal at specific locations and stories, the analysis (e.g. using SAP 2000 staged construction methodology as demonstrated in the example in both GSA and UFC documents) is executed in a 3D fashion and all affected beams by the removal of one particular column, plan-wise and height-wise, are analyzed and designed to pick up the resulting loadings. It seems to me that the GSA Redundancy Requirements simply re-check the strengths and stiffness of these same beams that have been beefed up because of the loss of column. I therefore do not understand where does the “redundancy” come from? I also don’t understand the rationale behind the 30% rule as prescribed in eq. 3-14 and 3-17.

Any thoughts anyone may have on this one will be appreciated. Thanks.

 

[hardbutmild]

I haven't read the document so I might say something wrong, but I'll give my opinion.

Let's say you remove one element and check the rest of the structure. If it can stand (not fail) without that one element, wouldn't that make the element kind of redundant? I mean, structure can do it without the element. To make it redundant you have to "beef up" other elements as you pointed out. Now, without that element deflection might be too high for functional reasons or cracking might cause faster corrosion so that's why you need the element in the structure. The point (as I see it) is to avoid collapse in an extreme situation.

As to 30% rule I can only suppose that it's because if you have more than one redistribution system, you don't want one of them to take on a large force from that collapse and others to take way less. If stiffnesses differ by a lot, I think this will happen. Since it's the case of external ground column removal, redistribution systems are cantilevers so it seems like a way to avoid inducing huge forces and moments to a single point (or edge).

It's just my two cents, hope it helps.

 

[cmrdata]

Thank you hardbutmild for your opinion.

Regarding your first comment, let me see if I understand you correctly: All the beams connected to the removed column(including the ones above the floor where the column is removed) are beefed up after the alternate path analysis (e.g. reversed SAP2000 staged construction analysis). The Redundancy check is simply to make sure the proper strength/stiffness distribution of these beams …. even though this check reveals that those already strengthened beams do not need further strengthening.

Your 2nd comment on the 30% rule makes perfect sense to me.
Best Regards,

 

[hardbutmild]

I would agree, first you do a 3D analysis with column removed and determine the required strength of all the elements from that analysis. After that you just check if the distribution of stiffnesses and strengths seems reasonable and if not, you additionally strengthen some beams.
Of course, you might get from the 3D analysis that some elements don't need to be beefed up at all.

The Redundancy check is simply to make sure the proper strength/stiffness distribution of these beams …. even though this check reveals that those already strengthened beams do not need further strengthening.

It seems like that to me.


I think this whole process is a "redundancy process", it's not just the check that ensures redundancy. 3D analysis makes sure that collapse isn't progressive, but the way you conceptually solve it (redundancy check ensures this) will influence how optimal that solution is.
Because you don't really know almost anything (from actual loads, to material to craftsmanship) 3D analysis on its own won't provide a truly safe solution so you must form a solution that minimizes those uncertainties.
Both parts of the process seem as important to me.

Now this is just how I see it.