Driven Piles in Liquefied Sands 3

[Gimbli]

I know from literature review that liquefied sands offer negligible resistance to piles under lateral loads though some pseudostatic approaches, reduced PY methods (Liu, Dobry and Wilson) and undrained shear strength approach (Wang and Reese), among others have been proposed. Recent research shows that resistance is negligible at initial deformations and there is no definite, developed or proposed PY curves to attend this condition (Weaver, Ashford and Rollins). I would like to know what you colleagues think about.

 

[GeoPaveTraffic]

I would not assign any lateral resistance to piles in liquified materials.

 

[Gimbli]

Therefore, it should not be practical to design driven piles for a three story building over 45 feet of liquified sands. Isn't?

 

[civilperson]

A battered end bearing driven pile will provide some lateral resistance.

 

[GeoPaveTraffic]

Gimbli,

I would recommend looking at the conditions that induce the lateral load that you are designing the piles for, as well as, the conditions underwhich the sands liquify. Additionally, while not impossible, sands 45 feet deep are difficult to get to liquify.

If the lateral loads and the liquifiacation of the sands can reasonably occur at the same time, then either the foundation system needs to be changed or the ground needs to be improved.

 

[GeoPaveTraffic]

civilperson,

A question about battered piles; I recall that several pile caps with battered piles have failed during earthquakes. The piles punched through the top of the cap. Are my recollections correct? If they are, is it still normal to design battered piles in earthquake prone areas?

 

[moe333]

Rollins does have P-Y curves for liquefied soil bu they are only to a depth of about 25 feet. They are in LPile5. The curves show no initial resistance, then increasing resistance with deformation. The other methods of analysis you discuss do not model the true shape of the P-Y curve, but may produce conservative results at larger deformations.

A relatively new program DFSAP is being used by Washington State DOT and you can download it free on their website. It apparently models the actual shape of the full scale tests by Rollins but uses the strain wedge method rather than P-Y curves. The comparisons it shows with LPILE also indicate more accurate results with non-liquefied loading.

 

[Gimbli]

That is what I understand GeoPaveTraffic that failures through pile cap (punching shear) has been observed due to high amplification of punching forces or so. In addition, settling liquefied sands will add bending moments to battered piles.

 

[civilperson]

The phenomena described by Gimbli is researched by many academic papers. CalTrans is testing and proposing batter piles for the support of the new Bay Bridge. Batter piles are still used cautiously in the design of new wharves. Railway Technical Research Institute, Tokyo, Japan has improved the seismic response of existing vertical pile foundations with addition of battered pile. SEAOC Blue Book Seismic Design Recommendations put some restrictions on using battered piles, (due to large axial loads developing in battered piles in seismic events).

 

[BigH]

Be careful about battered (raked) piles in such situations. See Nagata Earthquake studies and they found that the pile caps were severely damaged where battered piles were used - pile not undergoing "same" stresses due to the rake. In Vancouver, at least when I was there, they used vertical piles only.

 

[dgillette]

There has been a lot of research of late, and even a small ASCE conference on the topic, including loads on the pile caused by lateral movement of the ground towards a free face or down a steep slope. (The lateral support may be worse than zero if the ground isn't dead flat.) Ross Boulanger at Cal Davis has been active in research on this topic.

 

[Groutingsfun]

Why not put the building on stone columns? Many buildings that size and larger have been built on ground that has been improved through the installation of stone columns. During installation, the vibrations actually liquefy the soil, which is then packed with rock. The treated soil is then left non-liquefiable as well as with a higher bearing capacity.

 

[ShawnR]

Is soil consolidation an option? It may be a viable approach.

http://www.americanwick.com/applications/detail.cfm?app_id=21&app_cat_id=7

 

[dgillette]

I don't think so. Wick drains are useful in accelerating consolidation, but granular soils that might be liquefiable would not consolidate enough under static loading to become non-liquefiable, even with a surcharge. There needs to be either vibration to achieve the increase in density, or else cementing, like mix-in-place soil-cement or jet grouting.

Wick drains have been used effectively to aid densification by stone columns or dynamic compaction. We got "textbook" results on one of our stone column projects.

Wick drains are NOT effective for dissipating excess pore pressure in "real time" during an earthquake.

 

[moe333]

Earthquake drains are becoming more popular, they are similar to wick drains but have a 3-inch rigid pipe in a fabric sock. They are vibrated down with a mandrel so there is some densification during installation. They apparently provide sufficient drainage so that you do not get significant pore pressure generation during an earthquake...no liquefaction. You still will get some dynamic settlement. They have been evaluated through centrifuge studies and blasting techniques. They seem to work on paper and in the controlled testing but have not been real world tested. Nilex among other companies supplies, designs and installs.

 

[dgillette]

Count me among the skeptics!

To be effective in preventing high excess pore pressure, the drains and surrounding soil would have to permit enough drainage to allow up to 1% volumetric strain in something like 15 seconds without high seepage gradients. Consider it like a radial-flow consolidation problem. No matter how pervious the drain is, the flow is limited by the permeability of the soil (and of any "smear" from installation). The only way there can be a high gradient to cause rapid seepage flow is to have high excess pore pressure, in which case...it's too late.

Who did the centrifuge tests, and where were they published? Centrifuge scale modeling of a combined seepage and dynamic problem is quite difficult and commonly requires a pore fluid with viscosity different from that of water in order to get everything to scale right.

I've seen video of such drains in tailings being densified by blasting. See, for example, and for entertainment value:

http://www.rapidimpact.ca/inco_tailings_pond.htm

Watch the timing of the discharge. While the drains blew out a lot of water, they did not prevent high excess pore water pressure (hence, the large discharge), and they could not relieve much pressure during a 15-second earthquake.

DRG

 

[BigH]

Another method to mitigate the liquefaction by densifying would be to drive timber compaction piles. This seemed to be a preferred choice in the Lower Mainland when I was there back in the early 80s.

 

[moe333]

Some pretty reputable people/institutions have done work on the Earthquake Drains:

UC Davis did some centrifuge studies:

http://www.neesgc.gatech.edu/content/view/35/33/

UC Berkeley developed finite element software to evaluate the drains:

http://nisee.berkeley.edu/elibrary/Software/FEQDRAINZIP

U-Texas did field testing:

http://www.nilex.com/Uploads\Effect of PVD on PP Generation.pdf

 

[Cardiffer]

I think this problem is quite easy.
Just use piles installed with blasting technique.