Cracked section properties for wind load analysis

[Gaurav Malya]

Hi everyone,

When evaluating shear walls for wind loads, should cracking checks and adjustments to section properties be based on ultimate load combinations or service-level load combinations? If sections crack under ultimate loads but not under service loads, wouldn’t using uncracked properties result in incorrect force and moment distributions for design?

I’d appreciate your guidance or any references on this.

Thanks in advance.

 

[ANE91]

 

[Gaurav Malya]

Hi @ANE91!
I am enquiring about use of cracked section properties for determining the design forces for Wind load analysis.
And not about the servicability considerations.

 

[Gaurav Malya]

To elaborate more on the above, when checking stresses in the walls and comparing them with the cracking stress of concrete (fcr), which load combinations should be used to ensure accurate assessment and design?

 

[HTURKAK]

AFAIK, In case of strength load level , it is mandatory to use cracked section properties,( effective stiffness ) for reinforced concrete buildings in order to limit the sway also.
Relevant section of ACI 318,

 

[Gaurav Malya]

@HTURKAK the section you shared is for approximating the stiffness for reinforced concrete building system loaded to near or beyond yield level, i.e for seismic loads.
I am working on a project governed by wind loads.

So to evaluate the extent of cracking of the walls using ETABS should I be checking the stresses in walls at service level wind load combinations or strength level wind load combinations?

That is whether I need to use the strength level wind loads (700 year return period wind load) or service level wind loads (50 year return period wind load)?

 

[ANE91]

I must be missing something, because your questions don’t make sense to me. The wind forces that go into your shear wall, applied at the diaphragm, have nothing to do with cracking of the shear wall. Further, nobody can tell you if your shear wall is cracked with the information provided. If it’s cracked, then use cracked section properties. If it’s uncracked, then use uncracked section properties. You don’t mention whether this is a masonry wall or a concrete wall, though the analysis is essentially the same. Degree of cracking makes a big difference in the calculation of drift.

Whatever the material, it’s going to be cracked at ultimate, and it might be uncracked at service. (URM shear walls can’t be cracked, iirc.) If you’re designing the wall to be uncracked at ultimate, assuming it’s RM or RC, then I would question the design, because it would be extremely overkill for anything that isn’t client-driven or performance based.

For RM or RC shear walls subject to wind, strength @ ultimate tends to govern over drift @ service. (You must check both, regardless.) Again, service-level wind loads are used to check drift. You need to perform a basic strain-compatibility analysis at the critical section to determine how cracked the wall is, if at all. Go back to first principles and quit relying on the software!

 

[Agbsh]

You run two independent models, one for ultimate load combinations to design lateral elements, in which the level of cracks are based on one of the 3 methods prescribed in chapter 6 (ACI 318). For service load combinations which will be needed for drift, the factors allowed to be increased by 10/7 = 1.4. Notice that for the ultimate load, the wind MRI is what the code tells you, based on the building risk category (700, 1700 or 3000 years), for service load, wind decided by the owner (based on your recommendations), could be anything even 10 years.

In either cases, if you opt to use Table 6.3.1.1.1, you run your model assuming all walls are uncracked (factor 0.7), envelope your result to check against concrete rupture (f_r), any wall in floor exceed (f_r); you cracked to 0.35, rerun until 2 consecutive runs close enough. If your building not large enough, then cracking everything to 0.5 is much easier and you run it once (for ultimate) and with 0.7 for service to check drift.

 

[milkshakelake]

It's tedious to make two models - one for serviceability and one for strength. If one model changes, the other will need to change as well. For large projects, this requires a lot of coordination and extra modeling. What I do is use the cracked model but with reduced wind for drift. The wind return period for serviceability is in the appendix of ASCE 7 and this can be handled with a combination. I personally use D+0.5L+0.42W. I don't remember how I got to this; it's documented in my extensive standards somewhere, but it's kind of a clue for you. I'm basically just scaling the load combination using V_{serviceability}² / V_strength².

Since I am using walls cracked by strength level forces, this approach is conservative. However, it can be adjusted by using lower level wind, and ASCE 7 gives a few different values you can use.

You mentioned that cracked section refers to seismic loads. This is not true. A wall does not care if it's hit with seismic or wind forces. Wind loads can cause cracking, whether it is strength or service level.

@ANE91 mentioned checking cracking by hand, and not relying on software. Well, we do need the software to run the load distribution in the first place. We can calculate seismic/wind loads at each level and do a distribution based on the center of rigidity, but we wouldn't be using the FEM software in the first place if we were doing that. I do agree that you should know what the software is doing, but there are massive speed benefits to using the software. I agree with people who say that we shouldn't rely on software for a lot of questions that come up, but this is too complicated to do by hand, and projects won't be profitable that way.

The way I check cracking in ETABS is comparing shell stresses to the modulus of rupture, as @Agbsh mentioned. I always forget which thing to use; is it S22, S23, F22, or something else? The easy way to remember which one to use is to make a test model with only a wall and apply a lateral load of 10 kips or something. Do a hand calc of the tension/compression stress at ends, and compare it with the value you got by hand. Then you know which shell stress to use in ETABS. I find this method easier than learning what all the different stresses in ETABS are.

 

[Agbsh]

@milkshakelake it is S22, assuming local axis 2 is vertical (default setting).

 

[Gaurav Malya]

Thanks a lot for all your responses guys