Concrete walls - Load distribution to foundations

[onsiteeng]

We are designing a 12 story building in etabs using a system of shear walls and one and two way concrete slabs. In most cases the walls are 50ft long until we get to the parking garage where a 30ft hole is punched through. The walls on each side of the opening are 10ft long and 12 inches thick. Ignoring lateral effects we expected the loads in the wall to be evenly distributed to the foundations below these lower walls. However etabs is transferring all the load to the outside edge of the support walls and not distributing it evenly as we would have assumed. why is this.

 

[engrbon]

Can you provide layouts.

 

[onsiteeng]

I have attached the actual wall elevation from etabs. Slightly differebnt from what I described(trying to keep it simple). See attachments. Our modeller has modelled the piles with short columns on grade beams

 

[onsiteeng]

Here are the support reactions. We do not understand the opposing lateral forces under the walls.

 

[engrbon]

Sorry but I don't quite get the shape of the entire structure a 3D view without grid lines will make it more easier for us to visualize. Do you mean you are already considering the piles in your model using short columns? Why not try to model first the structure down to the base level only. That's why your building seems like i'ts floating. Then copy your model with a different file name and draw the grade beams and run the analysis then try to export it to SAFE. One more thing your building looks irregular to me.

 

[engrbon]

For the lateral reaction you used pin restraint right? It will produce lateral reactions.

 

[engrbon]

Gravity moments are causing your lateral load reactions. Just imagine if you're using roller restraints, the supports will move, which logically means, there is/are force/s that causing them to move. So if you put pin restraints this will resist the force hereby causing lateral reactions.

 

[onsiteeng]

Yes but why opposing forces in the supports holding up a wall that is supposedly a monolithic element.

 

[engrbon]

As per my understanding of your model you placed the restraints at the end of the short columns. Short columns will produce moments sir the load being carried by the beams will be transferred to the short columns.

 

[engrbon]

My idea is remove the short columns and delete its storey. Make a separate copy so you can retrieve any previous designs. Maintain the grade beams. Run the analysis and export the base level including the loads from above to the SAFE program.

From CSI on SAFE:

• How do I input piles for my basemat?

You will need to input springs at pile locations using stiffness values equivalent to that of the pile. In order to draw a spring, you will first need to use ‘Draw Point Object’ and click at pile locations. You can draw the points by snapping to grids, or by drawing the points, then right clicking each of them to modify XY coordinates. After drawing your ‘pile’ points, select them by clicking them or windowing around them, then Assign > Column Supports > Add New Property, and checkbox Spring Constants Option to input spring values for the piles.

 

[onsiteeng]

For our piled foundations design we typically use drive steel piles driven to the tonage specified based on working loads. Given the large safety factor incorporated in the actual placement of the pile and applied by the geotechnical engineer it is more likely that the piles will be a rigid support and would probably not simulate a spring action. This was substantiated on a previous project where the foundations on grade settled 3/4" and the piles did not. Therefore we normally just require the working loads on each pile group. For the model we sent we were assuming three individual pile caps under the lower support walls. Wouldn't modelling three pinned supports be simpler than transferring to safe?

 

[engrbon]

I'm wondering why use short columns? Is this really part of the structure or you're just depicting it to be your piles?

If you encountered the same problems before at least you can use it as your reference.

By the way CSI said use stiffness values equivalent to the pile.

 

[onsiteeng]

Yes, our modeller was using them to depict the piles. Not 100% sure why as the walls over the piled supports are at least 20ft tall and are very stiff. I may get them to remove the grade beams and just add the three pinned supports.

 

[engrbon]

I guess you should remove the short columns instead.

 

[engrbon]

Well...sorry if I'm not that much of a help. Hope you can find the right solution to the problem.

You can at least try to get your reactions directly from the grade beams...well just use it as your pile cap and get the reactions from the wall itself.

I'm not really comfortable with the short column approach.

Best of luck to you guys...

 

[onsiteeng]

engrbon,
Thanks for your help. We removed the piles and put in supports. One issue we found was the walls had been drawn across the supports and not from support to support. This made the wall span to the end supports and applied very littleload to the interior supports.

Drawing them as piers to each support applies more load to the internal supports, however, is this a correct way of modelling a concrete wall. If the wall is monolithic it would have to crack and deflect significantly before any load is applied to the center supports.

Any thoughts on this.

 

[engrbon]

Can you please check the video tutorials from this site, if ever you haven't watched them yet. I just need to go outside for a couple of minute. Just keep me posted. If you can attach your model maybe experts here can give you more guidance.

http://www.csiberkeley.com/Support_WL_ETABS.html

 

[ge0san]

onsiteeng,
I have the feeling that the walls are not meshed. Is this the case?