Total Building Collapse from a Single Column Failure

[JSanderO]

NIST makes the case that the failure of column 79 on floor 13 apparently caused by a girder walking off its beam seat at column 79 led pretty quickly to the collapse of the entire building leaving nothing standing at all.

I wonder... how universal this actually is?

Would column 79 failing at floor 29 have caused the global collapse?

Would any other single column failing on any floor lead to global collapse?

Could any single column failing on any other floor NOT lead to global collapse? (I don't suspect the failure of a column at the roof level would.) If so why or why not?

Is this single column failure applicable to any multi story high rise? Would it have to be steel framed? Would it have to be a minimum building height? Would there have to be a minimum number of floors above the failed column?

If the single column failure global collapse outcome is not more or less universally applicable what was it about 7 WTC's design and column failure at floor 13 that allowed for a single column failure to lead to global collapse?

Should NIST have discussed this or not?

 

[erwinl]

Isn't this a rewritten question of NIST said that this was the straw that broke the camels back and how universal it would be that a single straw actually could break a camel's back?

 

[slickbcfc]

This is just rubbish, all of the structural steel in that building is designed for a failure here and there.

 

[chrismohr]

JSanderO,
Let's see if we can all stay on track with your questions and try to avoid derails.
I haven't looked at Ryan Mackey's lengthy white paper for awhile, but there may be some answers here: https://sites.google.com/site/911guide/ryanmackey
In his private emails to me, he was the first one to explain to me the general principles underlying Building 7's collapse. Even he had some questions about details of the NIST Report in this matter.
Others can explain this better, but very generally, my understanding is that if a single column fails, it shifts its weight-bearing load to other columns at nearly the speed of sound. If those columns are strong enough to bear the weight, then the building stands. In the case of Building 7, Column 79's failure may have led to a very fast cascade of failures... as we say, the straw that breaks the camel's back.
I doubt that the principles of global failure apply only to steel frame buildings, and certainly almost all steel frame buildings on fire do not suffer global collapse. The fact that Building 7's long-span trusses in an unfought fire were left to burn for seven hours makes the conditions that day unique.
I am obviously not talking from a level of great knowledge here. I do wonder how anyone could correctly analyze exactly what happened in Building 7 behind that intact curtain on the outside perimeter of the building, but I also think that very general structural principles can be used to help understand this catastrophe a little better.

 

[DGM]

If the single column failure global collapse outcome is not more or less universally applicable what was it about 7 WTC's design and column failure at floor 13 that allowed for a single column failure to lead to global collapse?

Should NIST have discussed this or not?

I don't think it was limited to floor 13, that just happened to be where evidence suggests it occurred.

NIST and others have expressed concern about long span truss designs being prone to failure in fires (causing global collapse).

That building just happened to be the one that the protection system failures allowed it to happen.

Single point failure is not something that's always designed out (structurally). It's something that measures are taken to prevent the conditions that can cause it.

If you remember, this is why NIST quoted security and safety reasons as to why they wouldn't release their models. Publishing these vulnerabilities would not be a good idea.

 

[JSanderO]

JSanderO,
..
I haven't looked at Ryan Mackey's lengthy white paper for awhile, but there may be some answers here: https://sites.google.com/site/911guide/ryanmackey
In his private emails to me, he was the first one to explain to me the general principles underlying Building 7's collapse. Even he had some questions about details of the NIST Report in this matter.
Others can explain this better, but very generally, my understanding is that if a single column fails, it shifts its weight-bearing load to other columns at nearly the speed of sound. If those columns are strong enough to bear the weight, then the building stands. In the case of Building 7, Column 79's failure may have led to a very fast cascade of failures... as we say, the straw that breaks the camel's back.
I doubt that the principles of global failure apply only to steel frame buildings, and certainly almost all steel frame buildings on fire do not suffer global collapse. The fact that Building 7's long-span trusses in an unfought fire were left to burn for seven hours makes the conditions that day unique.
I am obviously not talking from a level of great knowledge here. I do wonder how anyone could correctly analyze exactly what happened in Building 7 behind that intact curtain on the outside perimeter of the building, but I also think that very general structural principles can be used to help understand this catastrophe a little better.

The so called load redistribution is the key concept here. Let's look at that from a conceptual basis.

A column low down in a structure with many floors above us supporting the local floor loads around it.. sharing the support of those floor loads with adjacent columns. Those loads are not terribly great when compared to the AXIAL load on the column from the columns above all contributing THEIR local floor loads. To redistribute a load laterally the loads would have to move through the beams which support the floor loads and braces which connect the column to the adjacent columns. So when a column fails all the braces connected to all the columns above the failed column now have to redistribute their local floor loads laterally to adjacent columns AND support the column they are connected to above the failed column. In order for the local floors to NOT drop the slabs and braces need to remain in place and hold the column at their center over the missing-failed column. I have some doubt that this would be the case, but it could be. It's hard to conceive that the column above the failed one would not drop at least some distance. If it did this would promote cracking of slabs, breaking of beam stub connections and perhaps the dropping of the columns directly above the failed one and all the local floors. This seems to be what happened under the EPH. A single column failed 34 stories below could not lead to the entire EPH foot print plunging through the building. Clearly more of the building under the EPH had to have collapsed in order for the entire footprint of the EPH to drop through the tower. The area drops... what loads are there to redistribute? Would this arrest at the line of the surrounding columns? That is... is it possible to remove a column line and the surrounding floor area to the next columns (typically 8) and have the rest of the building remain with a square hole right up to the roof (stack of rectangular doughnuts)? Why not?

1x2x3
xxxxx
4x5x6
xxxxx
7x8x9

to

1x2x3
x----x
4-o-6
x----x
7x8x9

numbers are columns x's are slabs -'s are dropped slabs, o is failed column

What makes this failure move laterally? Is it the pulling of the 8 columns into the hole and the floors attached to them? A widening sink hole?

Only braces strong enough to carry the stresses and connections strong enough to function as cantilevers could redistribute loads laterally... Yes or no?

 

[DGM]

The so called load redistribution is the key concept here. Let's look at that from a conceptual basis.

Why would you do such a thing? Doing so would assume that this building at the time was "as built". Was WTC7 "as built" at 5 PM?

 

[chrismohr]

JSanderO, I am way out of my element here, but your idea that there has to be some kind of "movement" through the beams after one column fails does not make sense to me. Let's say that you have 20 columns and one column buckles or snaps. The weight load it is carrying has to be taken up by something. What possible condition would prevent that load from being handed off to another column? If column 79 goes, then colums 78 and 80 just have more weight on them, period. Just like if four guys are carrying a piano and one guy just walks away, the full weight of the piano is now on thr other three guys in some fashion, and each one says holy alf, this is really really heavy now! I freely admit I may be way way off on this hypothesis, but bottom line, it seems that a shifting of the load to remaining columns somehow is inevitable, regardless of lateral movement???? Of course, loads do shift, through the hat truss or the beams or whatever, but regardless of the condition of the rest of the building, the load WILL shift SOMEWHERE almost instantly when one column loses its ability to support any load.

Now could someone with more technical background please answer this better than I can ever hope to?

 

[JSanderO]

JSanderO, I am way out of my element here, but your idea that there has to be some kind of "movement" through the beams after one column fails does not make sense to me. Let's say that you have 20 columns and one column buckles or snaps. The weight load it is carrying has to be taken up by something. What possible condition would prevent that load from being handed off to another column? If column 79 goes, then colums 78 and 80 just have more weight on them, period. Just like if four guys are carrying a piano and one guy just walks away, the full weight of the piano is now on thr other three guys in some fashion, and each one says holy alf, this is really really heavy now! I freely admit I may be way way off on this hypothesis, but bottom line, it seems that a shifting of the load to remaining columns somehow is inevitable, regardless of lateral movement???? Of course, loads do shift, through the hat truss or the beams or whatever, but regardless of the condition of the rest of the building, the load WILL shift SOMEWHERE almost instantly when one column loses its ability to support any load.

Now could someone with more technical background please answer this better than I can ever hope to?

The load if is redistributed through the structure to the closest columns which can carry the load. The key here is that the column and the connection must be strong enough to function such that the load can be carried down to the foundation. If say the floor slab was not able to be self supporting with a column missing... it would collapse. If it was strong enough, rigid enough NOT to collapse its mass would then be carried by the column it remained connected to.

Do you think, for example that the twin tower floors would NOT collapse if all the perimeter columns of the core were "magically" removed? Would the floor slabs be rigid enough... and the truss seats and the bar trusses strong enough to be supported ONLY at the facade? I think the answer is pretty obviously NO... the floors would collapse for any number of reasons... probably the truss seats were not able to carry the increased loads. This is fairly obvious because why WOULD there be such massive perimeter core columns if they were NOT carrying substantial load?


So you can not simply assert removing a column means the load HAS to be carried by adjacent columns. The alternative would be the load "DISAPPEARS:.. ie collapses to the ground. No load redistribution...a column failure leads to load REDUCTION or load DESTRUCTION. This outcome is not certain but it IS a possibility. If there is load destruction or load reduction the failure would NOT propagate or even add stress to the structure.

Therefore for failure to propagate laterally in a structure it likely involves beams pulling and or pushing adjacent columns causing them to move out of axial alignment or bend etc. If there is a collapse it seems more likely to be a pulling action...

This means that the failure might not be from load redistribution causing overload and buckling to other columns but a DIFFERENT mode of failure. The steel in the debris would show different sort of "damage". Stress overload from redistribution would exhibit the signs of classic buckling... and pulling or pushing would not show the web crippling of buckling.

Maybe

 

[DGM]

The load if is redistributed through the structure to the closest columns which can carry the load. The key here is that the column and the connection must be strong enough to function such that the load can be carried down to the foundation. If say the floor slab was not able to be self supporting with a column missing... it would collapse. If it was strong enough, rigid enough NOT to collapse its mass would then be carried by the column it remained connected to.

Do you think, for example that the twin tower floors would NOT collapse if all the perimeter columns of the core were "magically" removed? Would the floor slabs be rigid enough... and the truss seats and the bar trusses strong enough to be supported ONLY at the facade? I think the answer is pretty obviously NO... the floors would collapse for any number of reasons... probably the truss seats were not able to carry the increased loads. This is fairly obvious because why WOULD there be such massive perimeter core columns if they were NOT carrying substantial load?


So you can not simply assert removing a column means the load HAS to be carried by adjacent columns. The alternative would be the load "DISAPPEARS:.. ie collapses to the ground. No load redistribution...a column failure leads to load REDUCTION or load DESTRUCTION. This outcome is not certain but it IS a possibility. If there is load destruction or load reduction the failure would NOT propagate or even add stress to the structure.

Therefore for failure to propagate laterally in a structure it likely involves beams pulling and or pushing adjacent columns causing them to move out of axial alignment or bend etc. If there is a collapse it seems more likely to be a pulling action...

This means that the failure might not be from load redistribution causing overload and buckling to other columns but a DIFFERENT mode of failure. The steel in the debris would show different sort of "damage". Stress overload from redistribution would exhibit the signs of classic buckling... and pulling or pushing would not show the web crippling of buckling.

Maybe

This post makes no sense to me. I'm not sure where to start as to where it fails me. The first paragraph, maybe.

Where are you going with this?

 

[JSanderO]

This post makes no sense to me. I'm not sure where to start as to where it fails me. The first paragraph, maybe.

Where are you going with this?

I am suggesting that a single column failure may NOT lead to load redistribution if the loads it was carrying... local floor areas around it... collapse down with the column... there is no load to redistribute.

Make sense or not?

 

[Quad4_72]

I am suggesting that a single column failure may NOT lead to load redistribution if the loads it was carrying... local floor areas around it... collapse down with the column... there is no load to redistribute.

Make sense or not?

I think the question is what is your underlying point? Are you suggesting that something else led to the collapse? Are you suggesting a different mechanism? What are you getting at?

 

[DGM]

I am suggesting that a single column failure may NOT lead to load redistribution if the loads it was carrying... local floor areas around it... collapse down with the column... there is no load to redistribute.

Make sense or not?

Where did the loads go or are you saying it was not sharing loads to begin with. This would only be a case if that was the only member carrying the load.

 

[Richard the G]

I am suggesting that a single column failure may NOT lead to load redistribution if the loads it was carrying... local floor areas around it... collapse down with the column... there is no load to redistribute.

Make sense or not?

I don't really understand what you are saying but I am a structural engineer that designs tall buildings for a living. Let me tell you how I look at it.

Some codes like the British codes tell you that you have to design for column removal, and the way to design for that is to assume that every floor can act as a catenary and span around the missing column... If you back calculate the assumed floor deflections under this scenario, they are very large ... about 1m or so. In the US we pretty much ignored this in design prior to 9/11 and some still do. No provisions in WTC7

So under certain circumstances removal of a primary girder could be expected to lead to building collapse... but quite a lot of things need to happen and some other things could happen.
- beam girder makes effective length twice as long.
- column load exceeds buckling capacity..due to heat or just load
- failing girder damages floor below (this is probably a secondary effect)
- column fails and catenary tries to develop, but in the process of developing will shift large loads to adjacent columns. There is plenty of capacity in the perimeter so the adjacent internal columns are most at risk.
- several seconds after failure of beam, many of internal columns will have collapsed ( evidenced by failure of the penthouse)
- after the internal columns fail, debris and catenary forces are pulling down perimeter frame, ( or perhaps debris at the base reaches the perimeter)
-Then all the perimeter frame fails as a single element, because it has no diaphragms

Yes its something like that ... Or it could have been the special forces using a newly invented top-secret material and conspiring with Bush and conspiring with the lift repair man and hundreds of others..

 

[JSanderO]

Where did the loads go or are you saying it was not sharing loads to begin with. This would only be a case if that was the only member carrying the load.

DGM... I don't want to call you dense... but.. please consider that each column in a multi story structure carries the loads connected to its "sides"... ie the local floor loads. These are usually shared with the adjacent columns.. simply supported. And the load of the column above it (whatever they may be... the lower down in the structure the greater the imposed AXIAL loads).

If a column "fails" ie buckles or is severed and no long is able to carry the superimposed loads those loads have to be carried by adjacent columns..

or

if they COLLAPSE the adjacent columns do NOT acquire any additional loads. Where did those loads go? Into a pile on the ground around the failed column.

You don't see this possibility? You think that if the columns no longer carries the loads... they simply hang off the adjacent columns... is that your conception? The ONLY possibility....????????

 

[JSanderO]

I don't really understand what you are saying but I am a structural engineer that designs tall buildings for a living. Let me tell you how I look at it.

Some codes like the British codes tell you that you have to design for column removal, and the way to design for that is to assume that every floor can act as a catenary and span around the missing column... If you back calculate the assumed floor deflections under this scenario, they are very large ... about 1m or so. In the US we pretty much ignored this in design prior to 9/11 and some still do. No provisions in WTC7

So under certain circumstances removal of a primary girder could be expected to lead to building collapse... but quite a lot of things need to happen and some other things could happen.
- beam girder makes effective length twice as long.
- column load exceeds buckling capacity..due to heat or just load
- failing girder damages floor below (this is probably a secondary effect)
- column fails and catenary tries to develop, but in the process of developing will shift large loads to adjacent columns. There is plenty of capacity in the perimeter so the adjacent internal columns are most at risk.
- several seconds after failure of beam, many of internal columns will have collapsed ( evidenced by failure of the penthouse)
- after the internal columns fail, debris and catenary forces are pulling down perimeter frame, ( or perhaps debris at the base reaches the perimeter)
-Then all the perimeter frame fails as a single element, because it has no diaphragms

Yes its something like that ... Or it could have been the special forces using a newly invented top-secret material and conspiring with Bush and conspiring with the lift repair man and hundreds of others..

What you are describing is the collapse of the interior mass but the perimeter columns seem to be immune. I am am not referring to a hull and core concept... I am referring to a grid of columns with an internal column failure. Either the floor loads it supported do not collapse because the plate is stiff enough to span the distance without the missing column or they collapse.

But what about the axial load on the column ABOVE the failed one? Why isn't this pulling all the attached floor beams with it on the floors above the failed column down? Are you saying that each floor plate surrounding the failed column redistributes some load to the adjacent columns? And they all remain "static"? I understand that there is reserve capacity. And you are telling us there would be no local (around the failed column) collapse.

If this is the case then how does a single column collapse an entire building?

 

[DGM]

DGM... I don't want to call you dense... but.. please consider that each column in a multi story structure carries the loads connected to its "sides"... ie the local floor loads. These are usually shared with the adjacent columns.. simply supported. And the load of the column above it (whatever they may be... the lower down in the structure the greater the imposed AXIAL loads).

Yes.

If a column "fails" ie buckles or is severed and no long is able to carry the superimposed loads those loads have to be carried by adjacent columns..

Yes

or

if they COLLAPSE the adjacent columns do NOT acquire any additional loads. Where did those loads go? Into a pile on the ground around the failed column.

What happened before the column collapsed? Did it not fail?

You don't see this possibility? You think that if the columns no longer carries the loads... they simply hang off the adjacent columns... is that your conception? The ONLY possibility....????????

Somewhere the load was redistributed, it didn't just disappear. This would be the "hanging off" you mention. You said it yourself.

 

[MileHighMadness]

Keep in mind, when the north tower collapsed, it completely cut seven of the existing exterior columns of Building 7.

 

[DGM]

I have no idea where he's going with this. If these "hanging" floors are no longer there the column is no longer braced. It's going to fail under it's own weight.

 

[JSanderO]

Yes.


Somewhere the load was redistributed, it didn't just disappear. This would be the "hanging off" you mention. You said it yourself.

A load... ie building material sitting on the ground in a pile of rubble is no longer a "service load"...it's not disappeared but is "supported" by the ground. Building loads are supported by beams, slabs, and columns...

You are being stubborn.

What happens to the loads of 7 WTC after it collapsed? Answer: they are supported by the ground.