Methods to check punching shear of loads close together

[EngDM]

Hello,

I have a slab on grade supporting some racking where the legs of adjacent racks are relatively close together, such that their punching cones would overlap. I am running the slab in safe, and after reviewing their documentation and contacting support, it turns out that they do not consider or check punching when the loads would overlap.

They did not offer any insight on how you would check this.

My first thought was to have SAFE check punching, and ensure that the summed utilization of any loads within slab thickness distance (clear from edge/edge of plate/punching application area) of eachother does not exceed 1.0. However this is likely overly conservative, as the loads would share 1, maybe 2 sides with eachother of they are diagonally spaced.

Has anyone thought of a rationale for justifying a punching check with loads in close proximity to eachother?

 

[Tomfh]

What sort of dimensions?

 

[TRAK.Structural]

If I understand your idea correctly, I don't think that would be overly conservative. It may in fact be unconservative.

If the critical perimeters overlap than I would think that you can realistically only count on 3 out of 4 sides for the perimeter for each load point. Maybe draw up a plan of the critical perimeters showing how much they overlap to help reason a solution.

 

[Just Some Nerd]

I'd do a hand calc on the punching check. To be extra conservative consider treating a line mid-way between the two loads as an edge, so that the assumption is the perimeters stop where they meet instead of overlapping

 

[JoshPlumSE]

My thought is that I would would combine the loads together into a load at moment at the approximate geometric center of the load and then just do the punching check the way we normally do.... by hand of course.

 

[EngDM]

What sort of dimensions?

Say I've got 2 5"x5" baseplates, that are spaced at 6" center to center. So you'd have 1" clear between edge of plate.

If I understand your idea correctly, I don't think that would be overly conservative. It may in fact be unconservative.

If the critical perimeters overlap than I would think that you can realistically only count on 3 out of 4 sides for the perimeter for each load point. Maybe draw up a plan of the critical perimeters showing how much they overlap to help reason a solution.

Yea I think you're right. I'd have a pair of parallel sides that would overlap, and the top and bottom sides might also overlap a bit too depending how close the loads are together, since you go out at a 45° from corner of plate to get your shear perimeter.

And then I'd also have to use a method for punching on a slab on grade.

 

[EngDM]

My thought is that I would would combine the loads together into a load at moment at the approximate geometric center of the load and then just do the punching check the way we normally do.... by hand of course.

Geometric center of the load group?

 

[KootK]

I would:

1) Check each load as though the neighbor were absent.

2) Check both loads on one, common failure frustum which is likely to have an eccentricy (moment). I suspect that this is what JP was getting at).

A bracketed design.

 

[EngDM]

I would:

1) Check each load as though the neighbor were absent.

2) Check both loads on one, common failure frustum which is likely to have an eccentricy (moment). I suspect that this is what JP was getting at).

A bracketed design.

I'm not sure I understand the terms in option 2.

 

[KootK]

I'm not sure I understand the terms in option 2.

Imagine one failure cone that has both loads on it and a reasonable punching shear perimeter that includes both loads.

 

[EngDM]

Imagine one failure cone that has both loads on it and a reasonable punching shear perimeter that includes both loads.

Ah okay, would you just add some stress due to eccentricity as well via M*y/I?

 

[KootK]

Ah okay, would you just add some stress due to eccentricity as well via M*y/I?

Maybe? The center of action of the combined load is likely to be non-coincident with the centroid of the failure perimeter. It's that eccentricity that I'm thinking of.

 

[TRAK.Structural]

Ah okay, would you just add some stress due to eccentricity as well via M*y/I?

I would think the check for the eccentricity would be handled in the flexural design of the slab. Lots of different methods for slab on grade to determine thickness/reinforcing.

 

[dold]

Say I've got 2 5"x5" baseplates, that are spaced at 6" center to center. So you'd have 1" clear between edge of plate.

Probably suggesting the same as others, just more words.

I'd probably start with these two cases:
1. 5"x5" baseplate with load 1P.
2. 5"x11" baseplate with load 2P.

Not sure where your eccentricity is coming from?

 

[Luceid]

Agree with previous - would check once individually (which SAFE can handle) and do the combined check assuming the shared perimeter

 

[Tomfh]

Not sure where your eccentricity is coming from?

If the loads are unequal.

The 2P load with no eccentricity case should cover it though.

 

[EngDM]

If the loads are unequal.

The 2P load with no eccentricity case should cover it though.

Yea if the loads are unequal then it will shift to the higher load but the center of shear failure perimeter is the geometrical centroid.