Can a layered nonlinear shell model capture shear failure of a reinforced concrete element?

[brTCP]

I am trying to model a reinforced concrete wall in Sap2000 with nonlinear layered shells. I know for a fact that the behavior of this wall is governed by shear, because it failed during a recent earthquake and the failure mode is clearly a shear failure. Is the nonlinear layered shell model adequate for this? I know this model includes nonlinear behavior of the materials and a nonlinear shear response, but I am not sure it captures correctly the shear failure, because CSI states that "Multi-axial failure criteria is not incorporated into the formulation of the nonlinear layered shell objects". What do you think?

 

[WARose]

I've never considered non-linear behavior of reinforced concrete in shear. Only flexural.

 

[271828]

The analysis would output something .

If I was going to try this, I would look for journal papers with measured results. Model the specimen and see if your method is accurate. Also, the papers might have fea so you could see how they did it.

The default assumption needs to be that the predictions are not accurate.

 

[ThomasH]

Since you know the failure mode I guess this is not design. Why do you do this analysis? School / Insurance purpose / Something else ?

I don't use SAP2000 but a guess is that the approach might work for bending. Shear failure is a different beast so I doubt that it will work. Especially since CSI makes that remark.

 

[brTCP]

Thank you for your replies. This is for a repair and strengthening/retrofitting project. I thought it was a good idea to have a numerical model that accurately represents the reality before attempting any solutions. I do agree with you that the safe assumption would be to consider the result inaccurate from the layered shell. I am aware of more advanced techniques, but they are too complex to be applied in this project.

 

[ThomasH]

I know that shear is more complex to model than pure bending. But I don't know SAP so I don't have any opinion regarding the software.

 

[Brad805]

There is a tutorial for this--> SAP2000 - 20 Nonlinear Shear Walls: Watch & Learn I don't think the SAP2000 material models will capture what you are looking for. You could reach out to tech support.

 

[hardbutmild]

I thought it was a good idea to have a numerical model that accurately represents the reality

I would be careful with this... since you can come to the same solution in different ways and you don't know most of the data regarding material properties required for nonlinear analysis.

Personally, I would not care for non-linear behaviour of concrete in shear. I would check how much shear can my wall reinforcement carry and any remaining force would be carried by strengthening. Here, I would also try to ensure that shear failure is delayed until after bending failure occurs.
Concept of strengthening is more important in my opinion than the actual forces from model (this will always be a very rough approximation anyways)

I guess you're talking about diagonal shear failure and not sliding? Just out of curiosity, are the cracks inclined at 45° or some other angle?

 

[brTCP]

Thanks, yes, I am talking about diagonal shear failure (not sliding). The inclination of the cracks is about 45 deg.

 

[TLHS]

I'm not going to give you a hard no, because I haven't looked at this software in particular. However, I can't see how they'd capture that without you having to do some very specific things. Non-linear behaviour is generally analyzed based on yield curves or tested post failure conditions. Shear failure in concrete doesn't have an easily modeled post failure behaviour unless your curve just goes to zero. It's not going to be able to model the shear interface friction and the rebar yield contribution to strength. You could implement something where they look at the reinforced component of the shear resistance, but that would still be very detailing dependent and unless you're putting in a hell of a lot of information specific to a shear wall calculation module for non-linear response I wouldn't be assuming any of that is happening.

When you're doing seismic analysis assuming hinges and things in shear walls, you use simplified elements like rotational springs and hinges to represent the behavior, rather than modelling in a wall and trying to have the software figure the behaviour out.

This is the sort of feature where the vendor would have spent a lot of effort implementing it and you would find it described specifically.

If you're just putting in a wall element or a bunch of plate elements, and you have the option for non-linear materials, it's likely that you just get to put in an elastic-plastic stress-strain curve that's similar to a ductile metal material. That won't capture what you are looking for.