Why there aren't any safety factors for accidental verification? 1

[Pretty Girl7]

I have two questions regarding this equation for ultimate limit state, accidental verification (EN 1990:2002+A 1 :2005 (E), page 48)

1. Why there aren't any partial safety factors (γ) in this equation?
2. Is the equation common for "EQU, STR, GEO, FAT, UPL and HYD etc" verification?

 

[GeorgeTheCivilEngineer]

Because partial factors on permanent and variables actions are supposed to allow for the variability of a 'known' value (strength, density and size of materials, how many people should be using a floor etc).

Whereas accidental loads are by definition entirely unknown - so adding a factor would be meaningless.

Accidental loads aren't supposed to happen, even within the lifetime of the structure. Because they are one-off events, while the leading variable action is included in the combination, a phi value is applied as it is unlikely to be at the full variable value at the time of an accidental action. This makes sense when you consider the sort of event that causes accidental action - impact, blasts, extreme environmental.

To my knowledge (ie without checking), accidental actions are always treated this way.

 

[ThomasH]

Since the question relates to Eurocode I will start with Eurocode EN 1991-1-7 Accidental actions. There is actually an entire standard that describes how to design for accidental loads. And the actions or loads are divided into identified actions and unidentified actions, so they are not considered as entirely unknown .

The approach is that you either design for a specified load, explosion, impact or whatever. Or you have a strategy for limiting the extent of localised failure. The last part means that if something enitrely unexpected should happen and a column collapses, the entire building should not collapse.

Regarding the load factors. In my experience regarding this type opf loading, the design is often based on a specified event. An explosion is based on a "known" possible accident or a likely amount of something to explode. And you don't have to combine the accicent with characteristic load values for other loads, hence the psi factors.

 

[GeorgeTheCivilEngineer]

Yeah, it was hard to word, I'll try again... We know the forces we allow for types of accidental incidents, but we don't know what an actual accidental force will be. Partial factors are to allow for the variability of a known value so they don't make sense to apply to accidental forces.

 

[Pretty Girl7]

Thank you @George for your thoughts.
Thank you @ThomasH for your references and insights.

I have another question for @ThomasH regarding the same,

So the "Ed" effects (the answer from this equation) will be used to compare it with the resistance of the structure. like "Ed < Rd" etc. That means the answer from this equation will not be tested under "EQU, GEO" etc, but it's a seperate test like "EQU, GEO etc"

So is it ok to update my structural checks (verifications) list as "ACCIDENTAL, SEISMIC, EQU, STR, GEO, FAT, UPL and HYD"?
Previously I thought I need to check "Accidental" under each verification like EQU, GEO etc.

 

[ThomasH]

@George
First, I don't want to steal Pretty_Girl7 thread so I'll try to keep this short .
I don't know how familiar you are with the Eurocode and accidental loads. But if you look in Eurocode EN 1991-1-7 Table 4.1 you see the term "Indicative equivalent static design forces". Those forces are based on assumptions for not only the vehicles velocity and mass but also the stiffness of the vehicle and the structure that it collides with. It is not possible to determine an equivalent static load based on the dynamic load alone. So, Eurocode makes some assumptions. If you now go to Annex C I think it is reasonable, based on material I have seen, the loads in Eurocode are based on "Hard impact. I won't go any further right now because the subject is not impact.

@Pretty Girl7
Due to that you have Civil/Environmental I want to ask, what is the purpose of your questions? You seem to be asking questions that an experienced structural engineer wouldn't ask. So it would probaly be easier to help knowing your background and the purpose for this. Loads and load combinations are the base for a lot of what a structural engineer does.

You mention a list of structural checks, but I would call it different design situations. To give a simple example, FAT is short for FATigue. That is usually a comparably small load that has a high number of repetitions. Can be a wind load causing a chimney to vibrate or the trains rolling over a bridge. It is far away in load amplitude from a seismic load or a accidental explosion which is only expected to happen once if it happens at all. But they are both in the Ultimate Limit State, ULS, because they represent a structural failure.

You mention Accidental load and EQU. EQU is "loss of equilibrium when the structure is considered as a rigid body".

If you look at EQU for a chimney with wind load a simplified load combination could be: 0.9 * Selfweight (stabilizing) + 1.5 * Wind (overturning)
If we assume that the chimney is also subject to an explosion the load combination (ALS) could be: 1.0 Self weight + 1.0 * Explosion + psi2 * Wind (psi2 = 0.0)

The factor "0.9" is to get a margin of safety for the Selwweight and "1.5" for the wind is the normal factor for Wind in ULS. The wind load is actually based on what is typically called a 50 year wind velocity.
For the ALS case there are no factors for permanent actions, the accidental loads are also treated different because they are not based on the same type of assumptions as the variable actions. And the variable actions are reduced, typically the cariable loads are not governing in ALS state. But that does not mean the ALS is always a governing situation.

All this comes down to probability, will there likely be a 50 year storm on the same day as an accidental explosion. It can happen but society does not deem the cost for that level of safety reasonable.

 

[GeorgeTheCivilEngineer]

My comment(s) was more a philosophical and esoteric one. We have accidental forces to design to, but they are not known in the same as a permanent or variable action so the partial factor is not relevant. Yes we can design for an impact from a given vehicle, but that has even less control than knowing how a floor will be used in a certain way.