Accounting for Differential Settlement in Structures

[ToadJones]

I have been asked to determine "how much settlement" a structure can withstand before members in the structure become overstressed.
I can rather easily model the structure and force the movement of hte supports to represent differential settlements and thereby capture the effects on the structure.

The issue I am having is whether to include the settlements along with the effects of other loads.
Have others run into this?

 

[msquared48]

You can either:

1. Determine what stresses will be put onto the foundation structure by the differential settlement mentioned in the Geotechnical report you are supposed to have for the project, and then see how much other dead, then live load the system can tolerate without being overstressed, or
2. Determine the stresses seen by the love live loads and see how much settlement, with the different levels of induced stresses assiociated with varying degress of settlement, that you can tolerate.

Years ago had to analyze a fire station slab and beam system that had differential settlement problems - where the vehicles parked. The damage had already occured, so the settlements were known. We just had to figure out what to do to fix it.

Superposition works great for this.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[ToadJones]

Mike-
Not sure I am following you 100%.

For a hypothetical, say you designed a frame with no consideration for differential settlements and took the design ratios of the members up to 1.0.

Now you analyze the frame with one of the foundations displaced along with your normal load combinations. You're design would likely not be suitable any longer.

 

[msquared48]

That's correct as the differential settlements will add additional moments and shears that have to be accounted for at the supports. That's where the use of superposition comes in.

Say you had a two span continous beam where the middle support settles one inch. To see the effect of the settlement, analyze the tewo span structure as a one span beam for the settlement, findint=g out what load will induce the settlen=ment, or deflection at the support, then backfigure to see what moments and shears that generates, adding those to the oned generated for the loce and dead load conditions of the two span condition.

I can do it easier than I can explain it. Sorry for any confusion here.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[BAretired]

There would be a lot of different load cases to consider if you wanted to be thorough. A settlement in one support would overstress some members in some locations and ease the stress in other locations.

One way to do it would be to apply a downward load at each support in turn, noting the stresses caused at every point in the structure and the deflection caused at the support with the added load. Then consider live load for each span in turn and finally consider dead load throughout.

Finally, you could combine them in a way which produced the worst effects at every location. A lot of work for a nebulous purpose.

BA

 

[Ron]

As Mike notes, the settlement induced stresses are in addition to the stresses caused by the typical load inducement. Add directly with no factors or reductions. If doing as a post-event analysis, settlement reduces the available live load capacity.

 

[slickdeals]

Agree with others. You would run the model with support displacements (you can make that a load case). Then you will combine the forces/moments from this load case with a load factor of 1.0 and apply it with imposed loads and their combinations.

However, if you are given a spring stiffness at the foundation level, then the program will automatically account for the additional loads coming for differential spring settlements. The only issue with this is that the it will use load factors to calculate the settlements from factored loads, I think this means you will be conservative while designing it this way.

 

[dirtsqueezer]

I'm actually not sure I'm following this either, but it seems like there's not reason to include diff. settlement data unless you encountered some problem with the foundation prep. or an inherent soils condition that would make you worry. If so, for sure I would include something. Two sides of a coin there. Just don't flip it.

 

[dhengr]

Ron.... You know the current codes far better than I do, but why wouldn’t you factor loads, moments or stresses due to settlement, just as you factor other loads. The question is, are they more like a D.L. or L.L. or wind or EQ loads in terms of the factors to be used? Now we have to rewrite ASCE 7 again yet this week, before its normal cycle (of every 2 or 3 weeks) because all 700 possible load combinations are missing a settlement load value. You might argue that the settlement loading is more like a serviceability loading condition as opposed to a normal factored design loading.

 

[Ron]

dirtsqueezer...there is reason to include differential settlement in the analysis. If we receive the info in the geotechnical report, we have to consider the effect it will have on the structure. For flexible frames, it is less of an issue unless the differential is large. For more rigid frames, the effect can be tremendous. Further, for rigid walls such as masonry, crack control is an issue when considering differential settlement. A rule of thumb is that about 1/4" of differential settlement in 20 feet can crack masonry, so structural isolation might be necessary.

As for the liability of the geotech for not including such potential....well.....

 

[wannabeSE]

Differential settlement is included in the definition of "T" (self straining forces). See IBC section 1602 definition of notations or the ASCE 7-05 commentary section C2.2. The code includes factors to use with "T" in the load combinations.

 

[hokie66]

I have never seen differential settlement used as another load case, although it could be if the differential settlement predictions were precise enough. "Precise" is not a usually a word associated with geotechnical predictions. When we know there is a real potential for differential settlement, overcoming the problem involves either better footings (usually in the form of piles), or articulation and/or flexibility in the superstructure.

 

[ToadJones]

For the record, my approach was to create a load case with a forced displacement on a given support and include that as a load case in all load combinations. When combined with a second order (P delta) the results with even a small differential settlement of one footing were drastic.

Oh Lord I hope the codie's aren't reading this...there is going to be another chapter in ASCE 7 in '15

 

[slickdeals]

FWIW, the Florida Building code has a requirement that suggests sizing footings bearing pressure determined by the load from column with highest LL to total load. I think the intent is to account for differential settlements between larger loaded columns and adjacent smaller loaded ones. I don't know if a similar provision exists in IBC.

SECTION 1624 HIGH-VELOCITY HURRICANE ZONES-FOUNDATION DESIGN
1624.1 Design procedure.
The minimum area of a footing or number of piles under a foundation shall be determined in the following manner:

1624.1.1
The total load of the column that has the largest percentage of the live load to the total load shall be divided by the allowable soil pressure or pile capacity.

1624.1.2
The balance soil pressure or pile capacity shall be determined by dividing the total dead load by the area of the footing or the number of piles.

1624.1.3
The minimum area of other footings or number of piles shall be designed on the basis of their respective dead loads only.

1624.1.4
In no case shall the total load of the combined dead, live, wind and any other loads exceed the allowable bearing pressure of the soil for capacity of any pile upon which the foundation is supported.

1624.1.5
The live load used in the above calculations may be the total reduced live load in the member immediately above the foundation.
The minimum area of other footings or number of piles shall be designed on the basis of their respective dead loads only.

1624.1.4

In no case shall the total load of the combined dead, live, wind and any other loads exceed the allowable bearing pressure of the soil for capacity of any pile upon which the foundation is supported.

1624.1.5

The live load used in the above calculations may be the total reduced live load in the member immediately above the foundation.

 

[dirtsqueezer]

Ron,

Yeah, I see your point. Especially if asked I would definitely make up a model that offered some idea of diff. settlement on the structure. I am just saying that buildings are built to last (barring some economic design), and planning for settlement isn't a solution, unless you have special reason to. Why not put your mental energy into designing a foundation system that works? Account for that discrepancy and increase compation or FS. Utopian I know. I must be just in that mode at the moment.

 

[Ron]

dirtsqueezer..I agree that designing a foundation that works properly is important; however, on some sites and some building configurations, you can't "design your way out of" a differential settlement issue...you have to live with it. Increasing the footing size to lower the bearing pressure doesn't necessarily solve differential settlement issues. In some cases, the bearing pressure is almost irrelevant to the settlement issue, so "balancing" bearing pressures from footing to footing has little effect (though it is still a good idea to do so).

Slickdeals....you've given another excellent example of how ridiculous some of our code provisions are. Some provision such as this only show up in the "HVHZ" in the Florida Building Code (High Velocity Hurricane Zone), which by definition is two counties...Dade and Broward; however, they should also apply to all other areas of the state, since the criteria go well beyond wind loading alone. Makes my life on the forensic side a bit difficult at times....no code backup for anything other than HVHZ. Same difficulty for requiring PT wood in contact with masonry or concrete....only shows up in the HVHZ, so opposing attorneys argue that it doesn't apply to the rest of the state. Absurd argument, but effective in reducing settlement costs for their clients.