Ref ACI 318-05: Walls can crack under wind loads or any other loads for that matter.
The moments of inertia of structural members should be representative of degree of cracking at various load levels. Criterion for cracking in flexure is based on modulus of rupture. I = 0.35Ig where cracking is predicted at factored loads. At service loads it is satisfactory to use I = 0.50Ig for cracked walls and 1.0Ig for uncracked walls.
Based on the stresses under various loading conditions, the stiffness of the shear walls needs to be modified on the stories where cracking is predicted.

I used to do this a lot at a previous company - but I don't really have a good answer for you, I don't think there is agreement in the high rise industry. I've done a lot of peer reviews on these as well and have seen people do lots of different things.

The most common thing that I've seen is as you stated to check for tension stresses. That usually means looking at the 'uplift' load combos and finding all of the net tension areas. Again I've seen this done different ways - I've seen people crack anything with any net tension, and others only crack where net tension exceeds the modulus of rupture. Depending on what software you are using you can usually set your option to turn everything above a certain stress level a certain color, say red, to make this faster to identify.

Some people crack all walls below any stories where tension is identified and others actually go panel by panel to identify and modify stiffness. I have also seen a cracking modifier applied to everything, regardless of stresses from analysis. I believe it also varies by region. I did a few buildings in Asia where we had really crazy drift requirements, but at the same time they allowed 1.0Ig for all walls.

People will often cite the ACI recommended values but in practice I didn't find that people were using those, it was more the exception.

In high rise you are probably controlled by wind drifts, and changes in stiffness at the base make a big difference so your assumptions for cracking have a large impact on your overall design. As I said - no good answer, just giving you what I've seen.

Thank you so much.
DST148: does that mean you should have two model: one with 0.35I for seismic and 0.5I for wind drift check?

Bookowski: very good information, what number did you see they put when it is cracked by checking the tension stress? I assume 0.35I for strength design for both wind and seismic; and different values to use for drift check?

Yes, two models are common - service and strength. Most likely you are running p-delta in the analysis model and you want to make sure you are running an appropriately 'soft' model to capture that, especially with something very tall.

As far as the value of cracking I've seen - again it varies. I was just part of a peer review of a relatively tall (50+ story) building where the EOR had no cracking in their service model. The building is a heavy system and it did not show any net tension under wind, this was their reasoning for 1.0Ig. For the strength model they have 0.35 for both bending and shear.

I've also seen 0.5 used, and also the 0.7 number that you mentioned. I'm not sure where the 0.7 comes from but I've seen this several times. I've also seen people first transform the section. If you're doing something tall then you probably have walls packed with reinf. at the base. I've seen people base the cracked section on the transformed section.

One more thing in terms of cracking. Often there is some form of an outrigger system. If you are grabbing columns with those outriggers you need to crack/modify the axial stiffness of those columns. Technically this depends on the loading direction so you could have different models for different loading directions.

I've seen at least one company going to a non-linear analysis in Perform 3D, which is probably the closest that it gets at this point to capturing the different directions/cracking etc. in one model.

Basically I think it's all over the place with what people are doing.

0.7I comes from 10.11.1 of ACI318-05 for uncracked wall. compared to you saw people using 1.0I for uncracked condition.
Thank you so much for your information. Very useful. I like your job: reviewing other people's work and can learn alot. BTW did you saw a shear kick at the interface level between tower and the parking podium, say 30 nstories high-rise with 10 levels parking, the shear on the shearwall has a big change on level 10-normally much bigger than normal. My understanding is tower above deflects like a cantilever from the podium below and a big reaction occurs there. Didi you see same thing on the project you reviewed? Anyway better way to tack it?

Yes, a large shear discontinuity usually occurs at podium levels, or at ground floor slabs etc. You can even get a shear reversal (changes direction). This is very dependent on how and what you model. It's basically like a cantilever with a backspan - where in your case the podium is the backspan support. This is sometimes called the backstay effect - there was an article on this in structure magazine recently http://www.structuremag.org/article.aspx?articleID...

and you can also look in PEER/ATC 72-1 which you can find for free on the web. Somewhere in there they have a pretty good discussion of this issue - forget where but I think it might be in the back in an appendix or commentary.

In terms of dealing with it - you can just design for those forces, that is the most typical approach. By design for it I mean design the walls for that elevated shear and also design/detail the diaphragms to transfer it. You can also disconnect the main tower lateral system from the podium system. Typically this is done by providing corbels from the shear walls (assuming you are talking about a shearwall bldg here) to carry the slab. This isn't very typical but I have seen it done - it might require additional podium lateral elements though, depends on your layout. Generally I'd say just make sure that the modeling makes sense and then design for the forces.

That really helps me. Thnaks alot.