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.

 

[FixedEarth]

"Engineering Design in Geotechnics" by F. Azizi and "Foundation Analysis" by R. F. Scott cover lateral subgrade modulus & analysis laterally loaded piles. It is vast topic, like slope stability, and is difficult to capture it in few sentences.

http://www.soilstructure.com/

 

[Mccoy]

Oneengineer, I would heed Fixed earth's suggestions.
One simple aspect though is that if you have the elastic modulus you can calculate the subgrade horizontal moduli by correlations. Vesic formula, adapted to piles, is a pretty popular choice. Elastic modulus must be operational, that is degraded in dynamic conditions (and static condition if horizontal loads are static). This is for purely elastic, linear models.

http://www.mccoy.it

 

[ONENGINEER]

Mccoy, fortunately I have the E modulus for soil. Would you be kind to explain which relation between Kh and E are you referring to for sand and for clay. My search did not yield a definite answer. Thanks for all the comments.

 

[Okiryu]

Oneengineer,

I am just doing some lateral analysis for piles and have information in hand.

Here in Japan, we use this equation to relate Kh and E. We normally use the secant modulus (E50):

kh = 80・E・(B・100)^(-3/4)

kh : coefficient of horizontal subgrade reaction (kN/m3)
E : modulus of deformation of the ground. The value used should be one of the following;
however, that of cohesive soil should not be estimated from the N value, but determined using either 1. or 2.:
1.modulus of deformation of the ground measured in a boring hole (pressuremeter test) (kN/m2)
2.modulus of deformation determined using the uniaxial or triaxial compression test (kN/m2)
3.modulus of deformation estimated at E0 = 7N based on the mean N value (kN/m2)
B : diameter of a pile (m)

If deflections are larger than 10mm we need to adjust (decrease) kh.

Would like to see other correlations that are used in other areas...

 

[Mccoy]

Oneengineer,
The Vesic relation is used for either sand and clay. It is the same relation as that used for vertical K in shallow foundations, but a factor 2 is applied (this is explained in the Bowles textbook). Pros are that its use is usually well accepted, is the same for all soils, includes elastic modulus of concrete and moment of inertia of the foundation.

To obtain Kh, you must multiply the above formula by 2. Es is secant elastic modulus, B is pile diameter, Ef is elastic modulus of foundation, If is moment of inertia=0.1B⁴ for circular pile sections. mus is Poisson secant modulus

There are many simpler formulas, one example is the formula illustrated by Okyriu, which I didn't know, I wonder about its source.
Other published simple relations: kh= [0.8; 1.3] Es/B (Bowles 1982) or kh= 1.6 Es/B (Chen, 1978)

Formulas utilizing the n_h factor have parameters which have been determined empirically by soiltype classes. I do not use them usually.

http://www.mccoy.it

 

[ONENGINEER]

Great answers from McCoy and Okiryu.
If I am right the multiple 2 is E/Es or is it a factor of safety if a FOS required at all. I will check how to get mus from mu. My other question is how far from the top of pile on the pile depth would this Kh be effective.

The square symbols in Okiryu should all be multiple signs if I am right.

Thanks again.

 

[Mccoy]

Onengineer, actually the factor 2 is an empirical factor derived by Bowles according to whom the lateral subgrade reaction is twice the vertical subgrade reaction.

Es is a secant operational modulus, it may include degradation (dynamic) or reduction (static) schemes but it has nothing to do with the factor of 2 indicated by bowles, although very often the static reduction is taken as half the small strain E value, so I understand your reasonings.

Poisson secant: I usually use 0.15 to 0.2 for sands and unsaturated clays, 0.4 to 0.45 for saturated clayey soil.


How far from the top. Of course, the lesser the representative Kh, the farthest from the top its effect is displayed, moment of inertia of pile section remaining the same.

Now, you are going to take a single representative Kh value if your model requires it, otherwise you are going to take multiple Kh values, each every node or every homogeneous layer.

If the model requires a single value, than it should be averaged along the lenght of pile which is significantly interested by the displacement (it can be an iterative procedure). An harmonic weighted average will let the softer layers govern, which is sensible since the shallower layers are usually the softer and they govern the head displacement.

http://www.mccoy.it

 

[Okiryu]

Onengineer, the above equation is based on data collected from pressuremeter tests conducted in Japan. The square signs are multiply signs.

Based on Broms theory, for cohesive soils, the soil resistance starts to work from 1.5B. Soil resistance distribution is linear. For granular, the soil resistance is a triangular distribution. Max. value is 3 times the soil passive pressure. See attached for your reference. Be aware that Broms theory does not consider layered soils.

 

[Mccoy]

Okiryu, do you have a literature source for that equation or is that just a in-house correlation?

http://www.mccoy.it

 

[Okiryu]

Hi McCoy, the equation is from the Architectural Institute of Japan. I found that equation more in line with field pressuremeter testing. I normally use that equation. The "nh" equations you mentioned above appear to be too conservative.

I have recently done some pressuremeter tests so I will check Bowles for the Vesic equation you have provided. The tests were conducted in a reclaimed fill area and due to the inconsistency of the fill material I do not know how to assign the proper modulus though.

 

[Mccoy]

The tests were conducted in a reclaimed fill area and due to the inconsistency of the fill material I do not know how to assign the proper modulus though.

Do you have results of other tests like CPTs, NSPT's... ?

http://www.mccoy.it

 

[Okiryu]

McCoy, I do have old CPTs results for the same area. I can provide those tomorrow once am back in the office.

That is always the main issue: lots of fill (undocumented fill) and do not how to assign the proper soil properties. At the end, where thick fill areas are found; piles or soil treatment plus structural floor slabs are recommended.... however old buildings in the same area were designed on shallow foundations....

Talk again tomorrow... thanks.

 

[ONENGINEER]

The formulation given by Okiryu is really useful due to its simplicity. But I cannot apply to my project, not having insitu measurement of E. I have CPT data and will make it very approximate if I convert it to N and then E =7N.

Mccoy: would the notification given by Okiryu "If deflections are larger than 10mm we need to adjust (decrease) kh" be applied to Kh when we use the Vesic's formula or is there another criteria. I think that is to ensure the linearity of pile behavior(?)

 

[Okiryu]

I think that you can still use E from CPT in that equation. Also the reduction of Kh due to deflections larger than 10mm is based on field test data from lateral loading tests. If you use E=7N for cohesive soils, you need to use a coefficient of 60 instead 80. This equation does not apply if you have liquefaction potential. Try to use that equation and compare the results for previous projects in your area.

McCoy, here is the CPT data we were talking about...

 

[ONENGINEER]

I will definitely try your formula when in my next project. Thank you for letting me know that the 80・E・(B・100)^(-3/4) equation was invalid for liquefiable soil. The soil I am considering is quite liquefiable(fos < 0.5). Not sure if liquefaction has impacts on Vesic's formula. Hope not.

 

[Mccoy]

The reduction in kh if deflections>10 mm based on test data is all but technically sensible, since it reflects a non linearity in the soil behaviour, that is a modulus reduction for larger strains, which is what is well known from abundant literature.
What is specified in that case is a treshold above which a reduction in the modulus must be applied.

Vesic's formula is valid for any soils. Of course, if the layer is liquefiable, in dynamic conditions and with the cautious hypothesis that it will liquefy, then Kh=0, since the soil becomes a fluid providing no lateral resistance to deflection, there is no more any spring opposing deflection.

So the kh is valid before the onset of liquefaction, during liquefaction the pile behaves like a pillar which is not laterally confined in the liquefiable layers. Large shear stresses may occur at the boundary with non-liquefiable soil.

http://www.mccoy.it

 

[ONENGINEER]

McCoy: May you explain why we are using Es instead of E values, despite that the linearity and elasticity are paramount in the Kh solutions. Nevertheless my CPT package produces E values and wondered if one should convert it to Es values and how. Thank you again for answering my question.

 

[Mccoy]

Okiryu, I saw the CPTU results, I believe an approximate conservative value for E in teh fill can be found indirectly calculating the Su values in the fill and from these using a relation such as E_undrained50= 200-600 Su (from Duncan & Buchignani, 1976).

In the soft (apparently non-plastic) silts some direct relationship might be applied such as E=qt*1.5*Cm where Cm should be about 1 at half the silts layer. You should need the raw values of qt, they cannot be guessed from the picture.

http://www.mccoy.it

 

[Mccoy]

McCoy: May you explain why we are using Es instead of E values, despite that the linearity and elasticity are paramount in the Kh solutions. Nevertheless my CPT package produces E values and wondered if one should convert it to Es values and how. Thank you again for answering my question.

Oneneg, usually the CPt packages produce a secant value of E (Es) which might be an E25 or E50, not Always mentioned. If it is a tangent E it should Always bear the subscript =, like E₀. E derived from geophysical tests is Always an E₀. Most probably you do not need to convert yor E value, since it's already a generic operational value.

Kh should reflect the operative values we believe will exist in the design situation. Designing by an E₀ value will result in very high and excessively optimistic values of kh, unless we foresee almost no displacement. Since the actual behaviour of the soil is non-linear and not linear elastic, we are just adopting the most reasonable approximation here, using the secant modulus at the design strain level. The design strain level is usually unknown but can be estimated by so called degradation schemes, like the Fahey & Carter method. This in static conditions. In dynamic conditions the strain level can be estimated by tables illustrated in regulations or more accurately by seismic soil response analyses.

http://www.mccoy.it