A hall with cantilever columns and the isolated member analysis in fire

[user277418]

Hello

I will talk about Eurocode (EC) provisions in relation to the topic, but most likely other codes have similar provisions too.

Acc. to EC1-1-2:
1.5.1.12
member analysis (for fire)
thermal and mechanical analysis of a structural member exposed to fire in which the member is assumed as isolated, with appropriate support and boundary conditions. Indirect fire actions are not considered, except those resulting from thermal gradients
1.5.1.7
indirect fire actions
internal forces and moments caused by thermal expansion

Basically the member analysis allows to extract any kind member from a system and design it as isolated regardless of any other actions developed in the system. It seems ok if stability of the system assured by some other measures.

Imagine a 2-span single storey frame with columns rigidly fixed at base. No bracings at all i.e stability of the hall rely solely on the cantilever action of the columns. Girders are simply supported on top of the columns. Under fire the girders are going to expand, push top of the edge columns outward and definitely produce additional bending moments in the base of the columns. If the base of the columns become a plastic hinge, there are risk of the columns to fall outward. But according to the "member analysis" we may disregard all of that. This doesn't seem right for me at all. I have tried to find any design examples in literature for the case or something similar without luck.

I have a dispute with a colleague about the matter. He doesn't want to bother of any interpretation of the code provisions. So he isolated the columns applied all external actions (wind, self-weight, dead loads) required for the fire design situation, omitted the indirect fire actions and figured out critical temperature for the columns. Fire painted only the columns for the critical temperature. Asserts this is enough. But the expansion of the girders is not going to be affected by the painting. So the risk is still there. It seems to me wrong to the core.

Are you agree to use the member analysis provisions for the case?
Does anyone have a good reference publication discussing stability of cantilever steel columns in fire? I know P313 "Single storey steel framed buildings in fire boundary conditions", but it can not be directly applied to the case.

Regards

 

[user277418]

No opinions on the matter?

 

[centondollar]

Considering indirect effects (catenary-type action of the girder which pulls the column) would directly lead to a geometrically non-linear analysis, which is both complicated, incompatible with standard structural engineering analysis philosophy (linearity, superposition), and not available in all FE software packages. If you consider the effect of fire on structural stiffness, a material non-linearity must also be added, further complicating the analysis.

 

[user277418]

Thank you for the reply centondollar

I know that. My inquiry is a bit different

 

[user277418]

No opinions? Too controversial topic? Sad

 

[lexpatrie]

A drawing would be helpful.

Your concerns seem worth exploring.

This deals with a lot of similar topics - arches spread and scissor trusses spread, they either have an unrestrained connection to the support in one direction, (that still transfers uplift and diaphragm loads), or they are "glued" to the supporting structure and produce lateral loads related to the relative stiffness.

If you have a beam that expands and the column beneath it is fixed, I agree you create additional moment at the fixed base. In the UL there's some discussion of unrestrained and restrained assemblies that kind of leans into the topic you are concerned with, and any time one engineer out of hand dismisses someone else's concerns, it's probably not good.

I suppose it's conceptually possible if the column remains cool and the beam gets hot quickly enough, it will lose stiffness and be unable to transfer the expansion force, or will eventually be unable to transfer the thermal expansion force, or the detailing (long slotted holes along the beam length and a shear plate connection to the column) would "eliminate" the force incurred by thermal expansion. To me this is similar to the "banging bolt" phenomenon from some years back (oversize and slotted hole "slip" connections have some threshold below which they resist load horizontally, when it is exceeded, they slip and the force is relieved as the beam then expands... See November 1999 Modern Steel Construction, Tide.

https://www.aisc.org/globalassets/modern-steel/archives/1999/11/1999v11_banging_bolts.pdf

Regards,
Brian

 

[BAretired]

To me this is similar to the "banging bolt" phenomenon from some years back (oversize and slotted hole "slip" connections have some threshold below which they resist load horizontally, when it is exceeded, they slip and the force is relieved as the beam then expands... See November 1999 Modern Steel Construction, Tide.

https://www.aisc.org/globalassets/modern-steel/arc...

Similar perhaps, but the banging bolts are not seen as a safety issue, whereas the expanding beam is, at least in a fire situation.

Seems to me that a pin and roller support is needed to resolve the problem described by the OP, but that raises other issues. The column with the roller would be rendered incapable of resisting a horizontal force parallel to the beam. Maybe the roller should consist of a pin top and bottom of the column.

 

[user277418]

Thank you for the comments.

While indeed the girder expansion force may be released by using slotted holes it is important that a designer is aware of the existence of the expansion force in principle and would like to handle it whether by detailing, structural or passive protection considerations. Besides the appearance of the slotted holes is going to have some affects on the connections behavior during normal/service life of the frame which should be handled either as BAretired has stated. So a complex of measures should be designed to handle the problem. But the 1st step is to accept the problem.

My real design case is more complicated and I don't have intention to bother anyone by asking to solve it for me. That is why I have described a simple frame which will behave in a similar manner structurally. In the real case bending resistance of the external columns base is so limited that transfer even 10% of the columns flexural stiffness may be a problem. In this condition my concern is becoming more harsh.

Indeed the girder expansion force is not going to be equal to a pure elastic expansion of the girder due to many factors. But theory from the P313 publication states that on the 1st stage some pushing force develops untill the girder's modulus of elasticity softens enough. Actual quantity of the pushing force depends on the whole frame system stiffness. In my real case the system is such that doesn't allow determination of the force manually. In general to determine the force the frame should be checked by a time-dependant coupled thermal-structural analysis with account of material and geometry non-linearities with imperfections like centondollar has mentioned. Which is not a trivial task on it's own, but considering all the fire scenarios and a lot of other things becomes even more elaborative.

BUT... member analysis allows do not consider all of that. And formally my colleague is right. That is why I wanted to find some kind of publications which may describe exceptions of the member analysis application or behaviour of the girders in such kind frames as I have described (not a real case) under fire or something else indirectly related. To lift up my concern from an engineering judgement/opinion to a real life scenario level in a probable continuation of the dispute.