Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: Modeling a subgrade

[Subject Prev][Subject Next][Thread Prev][Thread Next]


A lot depends on the size and importance of your project and how much money you can spend.  Doing the 12 inch plate test on petrochemical and power projects is very common however probably very rare for a small commercial project.  Where plate tests are not performed the values come from the geotech's experience and to be honest it usually does not vary much from what you could get out of a typical handbook such as Bowles.

From a structural engineering point of view, the two key results you are looking for are the slab shears and bending moments.  You will find that a change in the k value (within reason) does not really effect these two values all that much.  Even with plate test results, it is very common to envelope the shears and moments with a k value of plus or minus 20 percent.

Thomas Hunt, S.E.
ABS Consulting


05/05/2004 07:34 AM

Please respond to

Modeling a subgrade

I guess one of my questions about modeling subgrades was what people were using for the k values.  Are they using values from the soils report?  If so,  how is the geotech coming up with the values?  Are they just be taken from the soils classification (if it's a silty sand,  k =  250?).

Nominally, values for k are determined by plate load testing using procedures given in ASTM D 1196, Standard Method for Nonrepetitive Static Plate Load Test of Soils and Flexible Pavement.  Load is applied incrementally and the deformation is recorded; the results are plotted to obtain a load versus deformation curve.  Results of plate-load testing should be reviewed with care, however. The modulus of subgrade reaction is defined as a spring constant and assumes a linear response between load and deformation of the subgrade. In reality, the relationship between load and deformation of a soil is nonlinear and is not a fundamental soil property.   Because the load-deformation results are nonlinear, either an arbitrary deformation or arbitrary load limit must be assumed. ASTM D 1196 does not specify a deformation limit but the final deformation is typically 0.05 in. or less.  Sometimes a load limit of 10 psi is used.

There is no single k value for a subgrade; the load-deformation relationship is a function of the soil's moisture content and prior loading.  The relationship is also strong dependant on the location of the load with respect to a slab edge and the size and shape of the loaded area.  

ASTM 1196 specifies a 30-in-dia. plate and the PCA, WRI, and COE design aids are based on the k value determined with a 30-in. dia. plate.  Testing an in-place subgrade with a 30-inch dia. plate is time- consuming and expensive, however.  Large loads may be needed to obtain even small settlement of the plates and adjustments need to be made for any plate deflections or non-recoverable deformation. Testing with a 12-in.-dia. plate test is much less expensive; however, since the load only stresses the soil to a depth of about twice the plate diameter, the results are not the same as with a 30-in. plate.  The value obtained from a 12-in. test is approximately twice that of a 30-in. test, but this can vary with soil type and compaction.

Even testing with a 12-in. plate is expensive,  which is why a value is typically taken from the soils classification.  I don't know where these values came from,  maybe from testing a long time ago ?  I'm not sure anyone things they are particularly accurate.

But the question still remains - what do the results of a 12- or even a 30-in. plate test have to do with your slab behavior?

How do people interpret their results?  

Gail Kelley