Is subgrade modulus still used by structural engineers? 1

[Geoluk]

Hello,

I'm a geotech engineer. From my side of the ground surface, the subgrade modulus, k, also known as modulus of subgrade reaction, is not used anymore for designing foundations. It is not really a property of the soil and depends on so many factors that getting an accurate value requires some amount of work. What's worse, it's not constant throughout the bottom of a footing so that giving a single value isn't exactly correct. Foundations are now just designed using shear failure and settlement analyses. Even mat foundations can be designed without subgrade modulus. Just to be clear, I know subgrade modulus is still used for pavement and sometimes slab-on-grade designs.

However, I've got a structural engineer client who is asking for the subgrade modulus to design a footing. I've already told them the allowable bearing capacity he has to work with but they are insistent on getting a subgrade modulus value. It's making me wonder if structural engineering in general has not caught up with this and is still using it in their designs. Do structural engineering software require a k value to work? Any thoughts on this?

 

[MotorCity]

In traditional spread footing design, it is almost never used.
The only exception I could think of is if you are designing your column base for partial fixity and allowing some rotation of the base (some where between a fixed and pinned end).

Other than that, its used mostly for slab on grade or raft design.

 

[geotechguy1]

I've run into structural engineers getting springs to put in models but back calculation allowable bearing capacity with 25mm settlement. Really depends on where you are and who the engineer is.

 

[dik]

For design of slabs, only... at this end.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[JedClampett]

Sounds like one of my guys. He loves using STAAD and refining undefinable parameters to the gnat's buttocks. To get the same results (I hope) he insists on inputting the soil springs into the model. He doesn't understand (or refuses to understand) that make believe input gives make believe results. He does pretty good work otherwise and I hate to get too deep into his methods.

 

[hawkaz]

I use it for rack and forklift loading on slabs on grade- not for footings

 

[TigerGuy]

I've had a few request it for foundation design.

 

[wth]

Where I find an approximate "spring stiffness" usefull, is if I'm designing a slab that also acts as foundation. For instance a waterproof concrete structure where the baseplate also acts as the foundation for exterior walls, and maybe interior columns.
In such cases the subgrade stiffness will determine the slab internal forces and the reinforcement. If we assume uniform reactions under the entire base, that will probably be an conservative approach for member design, but not for verifying geotechnical bearing pressure.

Interested in alternative ways to doing this other than subgrade modulus, as it's not the first time I've heard about the issues with this approach.

 

[ggcdn]

I use it for large combined footings / raft foundations / core footings which require analysis by FEM.

I'm not sure how you could reasonably design a large raft without it. The results are quite sensitive to it on softer soils. Once you get above 200psi/in or so with typical building footings, it becomes less important.

For isolated pad and strip footings, its not typically used.

 

[strucbells]

To echo a few others, slab on grade with equipment wheel loads or other point loads I will typically take a look at it.

I design a lot of water retaining concrete structures with exterior and interior walls/columns etc, the subgrade modulus is essential to design the mat slab that forms the floor of the basin.

But yeah for a typical small building with spread footings/continuous wall footings I would never look at it.

 

[milkshakelake]

Echoing others here, I use it for mat foundation design, and no, we haven't caught up to somehow using bearing pressure to define soil stiffness (except for the famous formula, I think by Bowles). I'm wondering, if it's so flawed, is there a better alternative? I'd use something better in a heartbeat.

 

[EireChch]

The most accurate way to determine modulus of subgrade reaction is to do a settlement calculation using FEM, then you can have contours of K and how it changes with displacement beneath the raft. We do this and then export it to the structural engineers who import it to their structural software. It is a bit of an iterative process.

 

[milkshakelake]

@EireChch For settlement analysis, wouldn't you need the loading? Or it uses various experimental loads to derive a nonlinear response?

 

[Geoluk]

@EireChch
That's exactly what I did. I'm glad to see I'm not the only one.

@milkshakelake
Yes, you need loading and also footing dimensions. I suppose in a pinch you could use various loads and give a range. I think the idea is that the geotech gives the structural a bearing capacity. Using that and the column loads, the structural sizes the mat foundation and passes it back to the geotech. The geotech uses that loading and size to get the subgrade modulus throughout the matt foundation and gives it to the structural (I provided a contour map). Then the structural can use that in their own FEM to see how the mat foundation will bend and how to reinforce it. If resizing is necessary then the it goes back to the geotech and so on. That's the iterative process that EireChch was talking about.

I think I was a bit surprised that there isn't something other than the subgrade modulus used by structural FEM software to model the bending of a mat foundation.

 

[milkshakelake]

This is something completely new to me. Good to know that it exists. I might bring it up next time I have a mat foundation to design.

I'm not entirely sure about the FEM concept. Hooke's law states that F=Kx. So you get the F from the engineer (or assume something), get the settlement based on geotechnical equations and soil layers, and then you can calculate K. Why would you need FEM for that? It seems like an equation you apply over a 2D space.

 

[Geoluk]

Getting the settlement under a mat is what the FEM is for. It can vary a lot from center to edge. Here's an example:

 

[Celt83]

Pretty much all of the structural software uses a winkler model so any large or mat foundation system is required on the structural side to have some form of spring value to enter into the software.

The process should be iterative but usually the geotechnical engineer is under their own contract with the developer and scope of service doesn't cover much if anything beyond the initial soils report. That leaves us in a position to usually need to do some form of enveloping of the spring stiffness. ACI 336.2R-88 offers us some guidance on variation of the spring constant throughout the plan of the mat to account for local stiffening at the edges.

 

[Geoluk]

@Celt83 "Pretty much all of the structural software uses a winkler model"

I think that is the key. Software companies have not developed a more accurate model from a geotech point of view. It could be there isn't a better theory. I don't know. But this makes sense. Even if we all agree the subgrade modulus isn't ideal, it is what we need for advanced modeling.

 

[Celt83]

My understanding is the subgrade modulus, determined from the 1 ft plate test, isn't really appropriate for use in Mat foundation design. The formulas proposed in most Geotechnical texts to modify the 1 ft plate test for actual foundation dimensions trends to unrealistically low values when modified for full building mat foundations. A common practice that I have seen is to use the 1ft plate test result without modification and then run the model again with double and then half the value and look for any variation.

In the best case scenario we'd have been allowed to open a dialog with the Geotechnical engineer and work to get compatible settlements between their analysis and our structural software results.

 

[Celt83]

these are some good articles from Structure Magazine that go into the subgrade modulus and modeling techniques:
Variable Modulus of Subgrade Reaction - Part 1
Variable Modulus of Subgrade Reaction - Part 2
Subgrade Modulus – Revisited