Massive Buoyant Force on Mat Slab below GWT

[Shanman_]

Hello All,

My team is racking our brains trying to rationalize a scenario with the following:

A 5 story-wood building is being designed over 2 stories of concrete. We have a 24" mat slab in the basement and two levels of PT slab above. The depth of the highest elevated level is 12" and the lower level is 8".

The issue to contend with is that the soils report as well as historical groundwater elevation data put the bottom of the mat at ~13 ft below the groundwater level. Thus, the report recommends to design the mat using Hydrostatic uplift of 800 psf! The mass of the wood superstructure above is let's say is 250 psf for argument's sake. Adding up the area weights, we can say 900 psf dead load for the building.

We are bouncing back and forth on the following points:
- As the nature of the buoyant force is less dynamic and more precise than a lateral hydrostatic load or a live load for example, we are contemplating running it as a fluid load and therefore would be factored at 1.2H as opposed to 1.6H. Any opinions on running it this way?
- Does the uplift on the underside of the mat contribute to additional compression in the columns as well as punching shear experienced in the mat slab? This would mean an additional 720 kips axial load (service level) to columns at 30 ft oc each way. We have folks leaning both ways.
- Since the mass of the mat slab is not enough to adequately resist the hydrostatic forces at the basement level, would the net uplift carry into the upper concrete levels as upward thrusts at the columns until there is no net uplift?

Would appreciate input from anyone, thanks.

 

[Celt83]

- Does the uplift on the underside of the mat contribute to additional compression in the columns as well as punching shear experienced in the mat slab? This would mean an additional 720 kips axial load (service level) to columns at 30 ft oc each way. We have folks leaning both ways.

Assuming you don't get enough weight to counteract the buoyancy and the hydrostatic forces actually occur, you get a situation where the structure undergoes a time step where it is not in static equilibrium ie in vertical/horizontal motion. It will be in this dynamic state until it floats up to a level where the hydrostatic force and vertical gravity loading reach equilibrium.

So in the dynamic time step you'd need to account for that loading some how, akin to horizontal ground motion from seismic and how we simplify it down to equivalent static procedures.

Open Source Structural Applications:

https://github.com/buddyd16/Structural-Engineering

 

[hokie66]

The mat slab will be like any other flat plate, but with its loading (from hydrostatic uplift) upward rather than from gravity. The columns and walls are its supports. Adjust your thinking by turning it upside down.

 

[Shanman_]

Thanks for the replies everybody.

Just to follow-up:

As I stated the building is 5 levels wood over 2 podium levels with a 2 ft mat slab foundation at ~13' below groundwater per FEMA flood map data and backed by boring logs. The project is for high-density housing and the lower two levels are for subterranean parking. The highest level of concrete is the entry/countyard level.

The building plan is nearly rectangular. The submerged portion of the building is going to completely flanked by retaining walls. The building/slab dimensions are approximately 250 ft x 500 ft. The height from the bottom of the mat slab to the highest wood level is approximately 80 ft.

The only complexities remaining:
1. The number of levels of wood structure above the upper podium vary by region. In some corner areas, it is only 2 stories. In others, it is 3 or 4. The inner 35% is 5 stories. I am fairly sure that in the areas with less superstructure mass, the buoyant force tributary to a column would be greater than the mass coming down. Does is make sense to have small areas of the mat slab be in uplift with the remainder doing the work in holding it down? My ADAPT model is not giving me any different output between 1.2D + 1.6L vs 1.2D + 1.6L + 1.2HL load cases. The deflected shape is the same, and rightfully so in most of the building. I was convinced that the negative bending caused by the hydrostatic load spanning between columns would be greater than that caused by negative bending between a group of column downward loads. I am seeing no difference in added top bar reinforcement required.

2. Late one night it occurred to me that the buoyant force would significantly decrease our factor of safety for overturning in the D + 0.7E + H ASD load case since our resisting forces are being reduced by the hydrostatic forces. However, after running some numbers, it looks like we are at a F.O.S. of 2.25. The last check was on pure uplift in the 1:1 D vs HL case. We were looking for a factory of safety of 1.25 in this case and are thinking we have just about that. I need to QA/QC these numbers to be certain.

All things considered, I think this was a great exercise for us. Any additional recommendations are welcome. Thanks!

 

[bones206]

My opinion is that the buoyant forces should be resisted entirely by the subterranean concrete "bathtub", with no consideration for the wood superstructure dead weight or any other superimposed gravity loads. In other words, the goal should be that the subterranean structure is stable by itself (with a decent safety factor against flotation).

Three methods I can think of to achieve that goal are:

1) Provide a drainage system to lower the water table like Mad Mike suggested.
2) Increase thickness of mat and/or concrete walls.
3) Some kind of tie-down system. Although I think this would be a last resort option.