Liquefaction question 2

[pelelo]

Hello,

I am working on a project which consists on the construction of 5 story buildings.

The Stratigraphy on this project is basically Silty sand (SM), silt (ML) and sandy silt (ML),with SPT values lower than 10 blows per ft.

This situation occurs during the first 30 feet. Going deeper, SPT values increase significantly to 40 and 50 blows per foot (and sometimes even more).

Also the water table is on average 2 feet off the surface.

No doubt this is potentially liquefiable escenario.

I performed the liquefaction analisys posted by Idriss and Seed in the 1996 and 1998 NCEER Workshops.

I computed the Liquefaction Factor of Safety every 5 feet for each boring. In some borings I got several layers with the FS less than 1.

In other borings, i got the FS < 1 in only 1 layer. (15-20 feet depth typically).

Other than excavating and removing the loose soils and filling back up with an adequate material, I know deep foundations can be another solution. Maybe the last one is better because excavating 20 feet would be kind of expensive.

So my question is, is it necessary to improve the soil conditions if only a 5 feet layer could liquefy?.

I am not sure what would be the minium loose soil thickness I need to keep in mind in order to improve the soil conditions (e.g replacing the material or recomending deep foundations).

Please let me know.

 

[Mccoy]

Great links bigH, the 2nd one is sure pretty exaustive

 

[muuddfun]

That second reference is one main reason why I do not like to use piles when liquefaction is a concern. Just do the ground improvement and be done with it. It is usually the most cost effective if liquefaction closer to the surface is a real concern like the OP.

If its light wood frame residential construction of 1 or 2 story and you have some room below the foundation to the liquefaction layers, then PT or mat slabs works quite well.

I have also used a large mat 2 foot thick for a very heavy transformer foundation (300 tons). We made a bathtub out of the thing, because they were going to have an oil catchment basin anyway. We just thickened up the bottom slab. I am happy to say that these foundations passed the test of the 7.3 in Baja recently. I think the OP groundwater levels may make it to difficult to use the box foundation.

 

[howardoark]

I'm pretty sure EZ Frisk covers the Caribbean. You can rent it for 30 days for a reasonable fee. It's simple to use.

 

[geoman110]

If there is no potential for lateral spread then you can calculate the post-liquefaction settlement associated with the a few liquefiable layer(s). From this point, you can have an estimate for the order of magnitude of differential settlement (worse case scenario) for a mat foundation. If the calculated numbers works (e.g., 2” differential settlement from one end of the building to another end) then, why not use mat foundation.
Furthermore, even if the differential settlement associated with liquefaction becomes high enough that is not acceptable, then you can mitigate the liquefaction by using wick-drains. In this case, you can still use mat foundation in conjunction with wick-drains. I do not think you can use conventional footing and wick-drains as the low SPT-N values does not provide high enough bearing capacity (even static condition) in order to use footing for a 5 storey building.
In respect to using piles, I have a question for my colleagues: if the entire upper 30’ was liquefiable, I would not be using piles as mentioned above. However, if the liquefiable layers are only one or two confined 5’ thick layers (as noted at the beginning), this should not jeopardizes the entire bearing capacity of the pile. You should still be able to use end bearing piles (good soil below 30’ depth). Just make sure that everything above the liquefiable layer would be contributing to negative friction. Furthermore, the structural design of the pile should consider within 5’ length of the pile, there would not be any lateral confinement and from structural point of view, this portion of the pile should be designed like a short column (pillar). What do you think?

 

[dgillette]

The foundation soils are silty sand (SM), silt (ML) and sandy silt (ML), as stated by the OP. I doubt rather strongly that wick drains could drain this material quickly enough to keep excess PWP to some manageable level. I would not rely on them. The research on earthquake drains that I am familiar with (U Texas in-ground model with their vibroseis vehicle, UC Davis centrifuge) incorporated very pervious clean sand.

DRG

 

[geoman110]

It is possible to design the wick drain to mitigate the liquefaction for SM and even ML soils. The design process results to a close spacing for ML soils which may appear too tight at the beginning. However, it may be still a cheaper option than others. One should design this various options (e.g., wick & mat, soil improvement, etc) and compare the cost.

Notwithstanding the above, I am still not convinced that if the post-liquefaction settlements are small (and there is no lateral spread), why can not use mat foundation (without anything else such as wick drain).

Does anybody have a comment in regard to my question for piles. I am in doubt about it.

 

[dgillette]

"It is possible to design the wick drain to mitigate the liquefaction for SM and even ML soils. The design process results to a close spacing for ML soils which may appear too tight at the beginning."

Where has this been applied? Has it been tested by an actual earthquake? What research has been done on using wick drains in silty soils? They have been used successfully at a number of sites to enhance densification by stone columns or dynamic compaction, but for "real-time" drainage in the short duration of an earthquake, I'd be very hesitant to rely on them.

 

[Mccoy]

If there is no potential for lateral spread then you can calculate the post-liquefaction settlement associated with the a few liquefiable layer(s). From this point, you can have an estimate for the order of magnitude of differential settlement (worse case scenario) for a mat foundation. If the calculated numbers works (e.g., 2" differential settlement from one end of the building to another end) then, why not use mat foundation.

Actually, Towatha in his 2008 book (Geotechnical earthquake engineering) relates that in Japan some small buidings are just built like that (rigid mat foundation) and if they tilt after liquefaction they'll use jacks to straighten them up.

Sure sounds awkward, but when there is no budget, they'll fall back to remedial works if necessary.
Probably limited to 2-stories buildings max

When I told this to the structural who is building his home in a place with a little potential for liquefaction and asked me to proceed with the geotech investigations, he liked the idea !!!
This is also a place with silty soil, GW to ground level, but deep bearing layer for piles.

 

[geoman110]

dgillette, There are a few research works in regard to successful use of wick drains on preventing “water film” under silt layers (e.g., Seid-Karbasi & Bryne 2006). I believe they used FEM to study it. If these researches have passed an actual earthquake in a real case that is something I do not know at all. Furthermore, Naval Facilities Engineering Centre has issues a technical report “Seismic Design Criteria for Soil Liquefaction-1997” which clearly states (Table-4) wick drains are an option for liquefaction remediation of silt soils. On the other hand, I have also seen several references that completely share your concerns in regard to use of wick drains for silt soils. If I recall correctly, one of them called it unreliable in performance. May be the issue is about what is classified and called as “silt” (55% fines or 90% fines). I have to re-evaluate what I believed about the effectiveness of this method and do a little bit more research.
Mccoy, Thanks for sharing the information. These buildings in Japan might have been subject to noticeable differential settlement and tilting, more than what I generally consider acceptable. I always believed that if a building undergoes a differential settlement of 2” from one end to another end, nobody cares to jack and straighten them up. For an 80’ long building, 2” differential settlement from one end to another, gives L/500 that should not be noticeable. I would like to know what order of magnitude of differential settlement/tilting these buildings in Japan have experienced. However, I generally agree with your comment that the whole concept of mat foundation does only work for small buildings. If the footprint area of the building is large, mat foundation performance is questionable.

 

[moe333]

I would not rely on wick drains to mitigate liquefaction for any soil; clean sand through silt. Earthquake drains are a different animal and they have some potential to mitigate liquefaction of cleaner sand.

 

[dgillette]

Even with the larger-diameter drains, be VERRRRRY careful. Look at the test conditions that seemed to showed they worked.

The Texas model was made of clean medium(?) sand, and I think the sand was placed around the drains, rather than the drains being drilled in. This was to avoid any chance for smear. Also, they were using many cycles of a low cyclic stress, so the excess PWP built up slowly in the untreated area; it wasn't like a large earthquake with 0.4g occurring within a few tenths of a second,the first few cycles.

In the Davis centrifuge tests, they used a fine clean sand, which by the scaling laws of centrifuge testing, gave them permeability equivalent to clean coarse sand. Boulanger told me the numbers once, but I don't recall them. In the first test, with smallish PHA (o.11g), the treated area did well and the untreated area had some spreading. Then they cranked up the PHA to 0.28, and the untreated area went liquid almost immediately. Its EPWP went very high and stayed high. (Can't tell you the ratio, but probably close to 100%.) The treated area showed some significant deformation and water gushed out of the drains, but the area stayed stable. Its EPWP fluctuated between 10% and 80% of the untreated area's with each load cycle, presumably more like cyclic mobility than flow liquefaction.

I draw the following conclusions:

Provided the sand is very clean and pervious and there is no smear or caking at the drain perimeters, the drains may be able to reduce the amount of damage and deformation at a liquefiable site. In the field, you might be able to achieve a clean perimeter by surging, like developing a water well.

For a dam foundation, I would never use them, for two reasons:
1) the consequences of them not working could be failure of the dam, with release of thousands of acre-feet of water onto the people downstream, and
2) stability of a dam requires good resistance to horizontal loading through the entire thickness of the foundation - a single layer of small thickness that is too impervious to drain "in real time" could be enough to allow instability. The gravity load remains on the dam slopes after the earthquake is over.

I downloaded the NAVFAC document from, of all places, Middle Eastern Technical University in Turkey. It does indeed list wick drains, but provides no guidance, references, or analysis procedures for them. I have to wonder if the authors had seen them used as part of a compaction program, e.g., stone columns, and just didn't make the distinction between a remediation method and something used to allow a remediation method to be constructed. Anyone know who wrote it?

End of rant; back to work.