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]
In a message dated 5/5/2004 11:48:17 AM Eastern Daylight Time, rakamaka(--nospam--at) writes:

All that boils down to what Christopher said: One's approach is as good (or as bad) as someone else's.

I don't think this is strictly true.  As long as someone's assumptions are not completely wacky,  then yes,  they will get a reasonable answer.  But if your assumptions don't make sense,  your answers will probably not be reasonable.

Although I mentioned raft foundations, I was actually thinking about slabs on ground.  Some of what has been printed in Concrete International has been complete nonsense.  

In response to one article, I wrote a letter to say that changes in the slab thickness didn't have a huge effect on the required slab thickness.  The response (by the ASCC technical committee, whoever they are)  was that if you had a 10.9-in.-thick slab on a soil with a k of 200,  you would need a 12 in. thick slab on a k of 50.  Which kind of made it clear they hadn't gotten an "A" in Soil Mechanics.  

In the first place,  no one is going to build a 10.9 in thick slab.  Most are 6 inches thick,  a few might be 10,  a few might be 11.  None are going to be 10.9.

But even more importantly,  if you had a slab with a loading that required that thickness, you would not build it on a subgrade with a k of 50 (mud).  In fact,  you are not going to build anything on a k of 50,  because you are not going to be able to drive the mixer truck on it.  You'd probably even have trouble with a buggy.

The ASCC response also cited another article where plate testing had been done and found that the tested values were much higher than design values.  This was in back up for their statement:

"When test results show that as-built k-values are much greater than the k-values estimated for use in design, a decrease in slab thickness can indeed be compensated for by the greater k-values."

The cited article did in fact get values that were considerably higher than the design values.  But they were also doing tests with a 12-in. plate.   The article failed to mention that their values would be expected to be approximately twice as high as design values estimated from soils classifications (there is a chart to this effect in one of Bowles' books.)  The article also didn't dwell on the fact that there was a tremendous amount of scatter in the results - the standard deviations they listed were quite large (i.e. average 600, standard deviation 255.)

In addition, the cited article  failed to mention that at least one of the design values was probably way too low - a k of 200 had been used for a subgrade that had a 9-in.-thick cement treated base.  It would appear that the designer was either intentionally being being conservative, or he or she did not really have much understanding of k values.

The articles being printed in Concrete International have probably not helped him or her at all.  There was one article that used a k value of 30 (thirty)  to prove something (something about shrinkage and curling I think.)  Another used a k value of 80.   

Concrete International has unfortunately gotten pretty junky in recent years.

Gail Kelley

cc:  Ward Malisch, Concrete International