Pile Lateral loading

[ONENGINEER]

I would appreciate feedback and thanks very much in advance.

The pile is embedded in 2 m peat, 6 m very soft to soft silt and 12 m in medium dense sand. Ground water is almost at ground elevation. Do soft silt and peat exhibit any lateral resistance? What are the LPile parameters and the subgrade horizontal.modulus for these soils.

 

[ONENGINEER]

McCoy: Thanks for the explanation. I would further appreciate if you could enlighten me on the followings.
I went through Bowels and got a bit lost. He introduces K's and then Ks = K's/B, which is the Vesic formula that you recommended for both sand and clay. Vesic formula did not depend on depth (except what is reflected in Es indirectly).

That was understandable until I saw in the following pages of Bowels (5th edition, 16-32, 16-33...) that Ks = (Ks v) Z**n/B. I guess Ks v means Vesic's Ks. Bowel does not say if it is for sand or clay. I am deciphering what formulation to propose for Kh and whether it is dependent on depth and soil type for design purposes.

Look forward to hearing from you.

 

[Okiryu]

McCoy, thanks for your comments. We do not do too much CPT so I am not so familiar with CPTs correlations with soil parameters. I heard that Robertson and Campanella provide good references for CPT. Also Professor Mayne. Need to check more on this. Thanks again McCoy !!

 

[Mccoy]

That was understandable until I saw in the following pages of Bowels (5th edition, 16-32, 16-33...) that Ks = (Ks v) Z**n/B. I guess Ks v means Vesic's Ks. Bowel does not say if it is for sand or clay. I am deciphering what formulation to propose for Kh and whether it is dependent on depth and soil type for design purposes.

I have the 4th edition of Bowles which is a little different but he illustrates equations with ks as a function of depth.

Fact is that: You already have the variation with depth of Es, calculated form the CPT data and from such a curve Bowles suggests that you can calculate the non linear trend (there is an n exponent on the Z, depth value). Is such a trend really useful in your case? It may be useful when dealing with wide, homogeneous areas with relatively few tests.

The laws of variation of Kh with depth (the nh method) have been developed as a generic law extracted from databases of piles load tests. The p-y curves of Matlock-Reese and Others are such an example. The use of databases has pros and cons. I have used p-y curves to check the values I calculated from CPTs or other lab and site tests.

http://www.mccoy.it

 

[ONENGINEER]

I have used p-y curves to check the values I calculated from CPTs or other lab and site tests
Did you happen to compare the lateral displacements and pressures obtained from Kh values with or without depth dependency? It would be helpful to know.

As you mentioned, the pile I am considering is at the CPT location and the trend is already reflected in Es values themselves.

 

[Mccoy]

Onengineer, I just compared my elastic-plastic model along the pile vertical to the non linear model of the P-y curves. The p-y curves have depth dependency since they are a function of overburden stress. The initial tract of the p-y curves is the tangent kh. So I just made sure my secant kh would not be too much optimistic compared to the non linear trend of the p-y curves. Of course the comparison must be carried out at various depths.

It is one pretty quick way to check that your secant kh is not too far away from the published non linear literature values.

Of course, you might ask, why not to use the p-y values directly? Here in Italy structural engineers have their own geotechnical add-ins to the structural programs and such add-ins only very rarely allow for a non linear input of kh.

I read again the original post and I take it you must input the values in Lpile, which I do not know well but I'm going to have a fresh look at it, not many chances to use it here since the new building code was released.

By the way, I received a brief e-mail on the L-pile issue and which might be iffy (leading to a scam), that's why I did not reply. If anyone reading this post has written it, please write again citing this thread with your username, if you're not a registered user or if you cannot give me the details of your activity, sorry I won't reply.

http://www.mccoy.it

 

[ONENGINEER]

McCoy, you raised an interesting issue. Just wondering if the Vesic's formula above is taking care of the tangent/secant Kh values automatically, if the Es values are derived from CPT data.

 

[Mccoy]

Oneng, Since the Es from CPT are almost certainly secant, the kh is going to be secant. I say 'almost' because I never happened to see a correlation from commercial softwares which explicitly carried the 'tangent' attribute or the E₀ symbol, the subscripot 0 indicating zero strain.

Vesic formula is blind to the secant/tangent issue, so if E is secant kh is going to be secant and the other way around.

http://www.mccoy.it