Concrete floor deflection - eurocode 1

[hardbutmild]

Hello everyone,

I have a question regarding deflections of floors (or beams) in eurocode, but a general discussion on the philosophy of deflection control is something that would be great so please comment even if you're not familiar with eurocode.

As far as I understand, the code wants me to check the long-term deflections for quasipermanent (QP) load (equivalent to "sustained" load in ACI if I understand ACI correctly) and limit it to L/250.
It also requires that you limit additional deflection to L/300 (EDIT: this seems very large, especially compared to the first limit), but I don't know if I should always check this, for what combination and how to take long-term effects for that?
It makes sense to me to check it for variable loads only (that is additional load), but some of it is supposed to be sustained for a long period of time (30% for residential and office, 80% for a library...). Do I add long term effects for that part of the load or not? I would check 30% of variable load for long-term and the remaining 70% of the load for instantaneous to limit the additional deflection and avoid potential damage to the nonstructural elements.
There is also a problem with the first check (take QP load and limit the long-term deflection) because the additional "not-sustained" part of the load could crack the element more (load history is important) leading to a larger deflection under QP. Some engineers also check for characteristic load (full permanent and variable without load factors), but everything is short-term. I do not understand the point of this check.

I am aware that anything we calculate is very far from realtiy and generally simpler = better, but I feel like a discussion would be valuable, just to hear what others are doing and what your thoughts are.
There is also a question of: "Should we completely avoid nonstructural damage?". This may lead to thick floors, but lowers the cost of maintenance. It is of course easy to say "discuss it with the person paying for the project". Unfortunately they are usually people who do not intend to use the structure and only want to build cheap and sell expensive (at least where I'm from).
This is more intended as a discussion to see how others go about the problems that are not as exact as for example capacity check.

 

[HTURKAK]

I would like to remind some definitions

Quasi-permanent value of a variable action ( ψ2Qk) (EN 1990, Clause 1.5.3.18 )
value determined so that the total period of time for which it will be exceeded is a large
fraction of the reference period. It may be expressed as a determined part of the characteristic
value by using a factor ψ2 ≤ 1
The definition of vertical deflection;

The design loads for the serviceability limit state ( for deflection, cracking and vibration ) are sustained loads.
That is , loads which act for a long time. The loads which should be used in calculating the deflection are:
(a) Characteristic value of dead loads
(b) Quasi-permanent imposed loads (For apartments and office buildings about 25% of the characteristic imposed load
is taken as permanently applied...)



Just saying based on my personnel past experience..



Use it up, wear it out;
Make it do, or do without.

NEW ENGLAND MAXIM

 

[hardbutmild]

Thank you HTURKAK, that is interesting.

Any idea on why does the code (Eurocode 2) give value of beta = 1,0 for the determination of the cracking factor in equation 7.19?
What would I use that for? The code (Eurocode 2) says "beta = 1 for a single short-term loading". When should I check that?

Based on what you said I simply check long-term deflections for the combination shown in eurocode 0 - equation 6.16b called the quasi-permanent combination.

 

[HTURKAK]

Pls look to the following figure to get the concept. The transition from the uncracked state (I) to the cracked state (II) does occur gradually. From the appearance of the first crack, realistically, a parabolic curve can be followed which approaches the line for the cracked state (II).

Use it up, wear it out;
Make it do, or do without.

NEW ENGLAND MAXIM

 

[hardbutmild]

I do not undestand what this means in practical terms.

What is the real problem in which I would have to say zeta = 1-(sigma_sr/sigma_s) instead of zeta = 1-0,5(sigma_sr/sigma_s) in equation (7.19) from eurocode 2? (or vice versa... what is the point of having two different values of beta)

ACI definitions make sense to me. You need to take sustained load and calculate long-term deflections + add short-term deflections for any additional loads to check if partitions will be damaged.

Eurocode is confusing. I calculate the same way as you said - only check QP, but what is the point of having two different values of beta if that is the case? If all loads are long-term, why even perscribe value of beta for short-term loading?

That is what is confusing me, the fact that I have a feeling that code wants all checks made for quasi-permanent combination (so loads sustained for a long period of time), but for some reason it gives a value for short term loads as well. Why?

 

[rapt]

RE you last sentence, the 2 cases are

- a single short term loading - so the the load is only applied once for a brief period

- sustained loads or cyclic loading - the load is applied over a long period or many times over a short period.

This is not referring to the Frequent Load Combination (sometimes called short term) and the Quasi-Permanent load combination (normally called permanent combination). Which of those is to be used is a separate clause.

The calculation method you are talking about is the Eurocode version of the ACI code long term multiplier approach and Bransons formula. My understanding is that the Eurocode version gives better results than the ACI version, but both are very approximate.

Clause 7.4.3(7) allows for you do the calculations properly if you want to, allowing for cracking, tension stiffening, shrinkage and creep! That is what is done in RAPT!

 

[HTURKAK]

I copy and pasted the relevant clause .Since EC-2 is a Law resource, it will not violate the copy rights ;

That is ,
ζ= coefficient for tension stiffening (transition coefficient)

σsr =steel stress at first cracking

σs= steel stress at quasi permanent service load

β=1,0 for single short-term loading and 0,5 for sustained loads or repeated loading.



Use it up, wear it out;
Make it do, or do without.

NEW ENGLAND MAXIM

 

[hardbutmild]

Thank you, but I think we do not understand each other... I guess I cannot explain the problem well enough, but I will try one last time.

Clause 7.4.3(7) allows for you do the calculations properly if you want to, allowing for cracking, tension stiffening, shrinkage and creep!

Sure, but there is still a question of what load case do you check for? That has nothing to do with the calculation method - simple or complex.

For example, imagine a simply supported one-way slab - standard residential use with standard loading. At midspan under that slab there is a large glass panel tightly placed under the floor. If your deflections are too large this glass will be damaged. How do you design this according to eurocode? What is the limit on the deflection and under what load combination?

Imagine that the panel is placed right after all the dead load is placed (all finishes are built). The floor has already deformed under this permanent load at the time when glass is placed, but will deform more due to creep, right? So only this creep induced deflection is causing problems. In addition live load will be placed on that floor. 30% of that live load will be placed long enough to cause creep - this is also a problem (so we have 1*permanent long-term deflection + 0,3*live long-term deflection - 1*permanent short term deflection).
The question now is the following - the remaining 70 % of the live load may be placed on the floor at one time. If this happens the glass will be damaged! Do I need to check for that or does limiting the long-term deflection under 1,0 permanent + 0,3 live to L/500 take this into account implicitly? Or does the code say "I don't care, if damage to that glass occurs just replace the glass"?

I can make simple calculations with numbers by hand if you think that it will be easier to understand my point.

 

[rapt]

Codes define minimum requirements for typical conditions.

You are the engineer. If there are specific conditions you think are a problem, check them!

 

[hardbutmild]

I agree, but there is a catch!
Now that I want to check some additional things how do I calculate zeta (see equation 7.19. in HTURKAK's post)? Is it different for QP check and for additional checks or is it defined for the worst case scenario? Because loading history is important and changes the behaviour.
For example, I have seen people use sigma_s in equation 7.19. as either stress for QP combination or for characteristic combination and both makes sense, but the difference in the end result may be quite large.
Zeta takes into account how much of the element is cracked. If at one point characteristic load acts on a structure it will crack it more than QP. Since we do not know when this characteristic load will occur it makes sense to use characteristic load for zeta even if checking only for QP.
On the other hand if all the checks are made for QP combination maybe you should use sigma_s for QP?
Is beta = 1.0 or 0.5 (in equation 7.19) if I choose to check the remaining 70% of live load? Because it depends on the number of times that this load is repeated and I don't know that.

Of course, I could always go with the more strict solution, but someone else will not... and I do not know if I am overdesigning stuff or are they underdesigning them. Shouldn't the code aim to avoid just that? ACI is quite clear in this respect, EC is not. This ambiguity also means that two softwares will get different solutions for the same "code based" check. That is what first made me think about it.

I don't know... I might be overthinking it. I know there is no clearly better or worse solution, but it frustrates me that the code is not clear on how to use the equations.

 

[rapt]

The code equations are simplified solutions for basic cases.

if you want to consider load history etc, they are too simplified for it. Use 7.4.3(7).

 

[hardbutmild]

Thank you.

If I choose to use simplified solutions for basic cases do I use in equation (7.19) sigma_s derived for QP or characteristic combination of loads?

Sorry for a lot of questions, but this was basically my main question...

 

[rapt]

7.4.1 and 7.4.3(1) appear to answer you questions.

Basically what the designer considers to be appropriate for the situation and what will be affected by the deflection!

7.4.3(1) says to use the load conditions considered appropriate for the check. It is calculating tension stiffening effects (Eurocodes alternative to Branson's formula in ACI). The only time QP load is mention in regard to it is for the very simplified L/D method in 7.4.2 where real calculation can be omitted.

 

[hardbutmild]

Great, thanks.

7.4.1 (4) and (5) also mention QP combination which is what confused me.