adjusting stiffness modifiers for a failed shear wall

[structural87]

Hello,


I would like to take your opinion regarding adjusting stiffness modifiers for failed walls in shear or bending.
I have a transfer system & the shear in some of the walls are exceeding the walls capacity.
The ACI recommends 0.35 for a cracked wall. However, if when this modifier is assigned and the pier is still failing in shear or bending, is it permitted to lower the shear modifier so the wall can take up to its capacity and the residual shear would be transferred to other walls which have additional capacity ?
Increasing the wall stiffness by increasing its length or thickness wouldn't be very helpful since the wall would be stiffer and would absorb more forces -> he would still fail.
Adding more walls isn’t an option due to architectural reasons and a small footprint area.

For me, if the structure is in equilibrium by adjusting stiffness modifiers and check would have been given to final stiffness output by rechecking the torsional irregularity, drift, base shear scaling, I would say this is acceptable.
Thanks for your opinions.

 

[DaveAtkins]

I agree with this approach, so long as you can convince yourself the failure of the overloaded shear wall is a ductile failure. It will start deflecting, and the load will go elsewhere.

 

[KootK]

I vote no, assuming that we are talking about seismic design. Or, at the least, the language that you are using isn't inspiring confidence in me.

The shear failure of concrete members is generally thought to be non-ductile and, as such, is avoided at all costs by way of capacity design principles.

If you intend to create a situation where walls develop flexural hinges well before they develop shear failures then, yeah, there is some scope for redistribution there. Even then, however, you would need to employ capacity / overstrength principles to ensure that the redistribution preventing shear failure in your walls really does come to pass before that shear failure occurs.

 

[EZBuilding]

Based on a linear elastic analysis I would generally agree with the "no-vote" - primarily from a heavy wind background. If your wall is failing based on the ACI recommended stiffness modifiers; the wall has failed. Shear is also particularly challenging as you have an upper limit cap that becomes hard to design against.

ACI has provisions in Table 6.6.3.1.1(b) for which you may be able to increase the stiffness of other shear wall elements - effectively reducing the shear load in your cracked shear wall. Note that within this provision the minimum is still 0.35Ig. You also have the new push towards a Performance Based Design, for which you might be able to justify your approach.

Based on your description of this being a transfer condition - seems like the lateral system should be revised to address the overstress.

 

[KootK]

By transfer condition, are we talking backstay effect? Tocci & Levi

Where that is the case, it's pretty common to address it by monkeying with the diaphragm stiffness.

 

[structural87]

I vote no, assuming that we are talking about seismic design

Yes we are.

The shear failure of concrete members is generally thought to be non-ductile and, as such, is avoided at all costs by way of capacity design principles

what is your common practice for such cases ? I don't see any other way than to add walls which is not acceptable by the building owner.
For me, applying stiffness modifiers for overstressed walls is a way to have another load path but I get your point.

you would need to employ capacity / overstrength principles to ensure that the redistribution preventing shear failure in your walls really does come to pass before that shear failure occurs.

Can you please explain practically how you achieve this condition ?

 

[structural87]

By transfer condition, are we talking backstay effect?

no backstay.
I mean by the transfer system a thick slab used to support the upper walls to make some space for car parkings.

 

[KootK]

what is your common practice for such cases ? I don't see any other way than to add walls which is not acceptable by the building owner.

Thicken the walls? I understand the imperative to keep our clients happy but, sometimes, owners just have to suck it up and deal.

Can you please explain practically how you achieve this condition ?

That's a big 'ol question. Why don't we start by you telling me/us what you know about capacity design principles in seismic design? Once we see where you're at, we'll have a better sense of what you need to hear.

 

[KootK]

I mean by the transfer system a thick slab used to support the upper walls to make some space for car parkings.

Where will this building be constructed? France?

How many stories is this building?

Do you at least have columns under the boundary zones of your shear walls at the transfer level? Or stiff beams under the walls?

It might be time to post some sketches of your situation.

 

[milkshakelake]

You don't arbitrarily assign cracked, uncracked, or arbitrarily low stiffness factors to suit the preferred amount of load going into a shear wall or redistribute the forces. It is based on modulus of rupture of the concrete. If that is exceeded, the wall is cracked. If the wall fails, how do you know that it can support the gravity load of the wall above? Or that the stiffness factor will be what you say it is (say 0.25Ig), but not zero?

Adding more walls isn’t an option due to architectural reasons and a small footprint area.

Doing good engineering is more important than making the owner/architect happy. What happens if this building catastrophically fails during an earthquake? That seems to happen sometimes, and it goes all over the news, all because someone wanted to make someone else happy. Anyway, you can consider increasing the concrete strength, adding link beams between walls to increase the moment arm, adding moment frames on columns (if there are any), and thickening the transfer slab to increase slab-column stiffness (but beware of punching shear due to lateral).

 

[structural87]

Thicken the walls?

They will absorb more loads and they will be still failing. I have already tried this with no success. same for increasing the wall length.

Why don't we start by you telling me/us what you know about capacity design principles in seismic design

I know some stuff in seismic engineering. as you said, this isn't the place to detail this.

Where will this building be constructed? France?

working in France but this will be constructed in KSA.
The transfer consists of a thick 90 cm slab. I have attached 2 images showing the wall and the floor where it has been transferred.

 

[Celt83]

So you are transferring out the L shaped shear wall? If so this leaves you with a very poor lateral resisting layout on level 2 where it likely isn’t going to matter what stiffness modifiers you use as static equilibrium from the torsion is going to demand a couple between your edge shear line and that short interior wall.

I have concerns with the diaphragm even being able to develop the force into the walls with all the openings shown make sure you pay careful attention to the force transfer checks and detailing.

How tall is this structure?

 

[TRAK.Structural]

You don't arbitrarily assign cracked, uncracked, or arbitrarily low stiffness factors to suit the preferred amount of load going into a shear wall or redistribute the forces. It is based on modulus of rupture of the concrete. If that is exceeded, the wall is cracked. If the wall fails, how do you know that it can support the gravity load of the wall above? Or that the stiffness factor will be what you say it is (say 0.25Ig), but not zero?

This.

Your analysis model has to reasonably match (by the code prescribed methods or by other rational analysis) the actual strength and stiffness of the structure. Then you let the forces distribute as they will. It sounds like you need more resisting elements