Foundation on loose sand 3

[vohai]

Hi all!
We are designing a three-story building in a site
with 17m depth of fine grained loose sand. Ground water table
fluctuate by monsoon at 7<->3m below surface. We are considering
many methods of deep compaction and method of compaction/replacement
surface sand with medium grained sand. Please share your opinion
about which method is best for this case and how to control work in
construction site. Thanks.

 

[cbosy]

What are your N Values? Is liquefaction a concern?

 

[MRM]

We've had success using vibro-compaction to densify loose sands at depth in similar applications to that which you describe.

 

[npthao121]

Are you coming from Vietnam?
What is your main concern, foundation for building only or other items else?
What kind of soil is below your loose sand? If it is a good layer and strong enough, why not consider the pile foundation (small concrete square pile) solution?
However if you still consider some soil improvement techniques as mentioned in your msg, I have some ideas

Deep dynamic compaction is a old, basic & not so expansive method to desify the loose sand both above & below the ground water level. This method can be utilized for the depth up to 20m with various type of soils. Up to know there are not any rigorous theory models proposed to understand the behaviour of deep compaction. This method is based on empirical only. Dobson & Slocombe (1982) proposed an expression to evaluate the range of effective depth related to impact energy as below:
1.26*sqrt(Wh) < D < 3.16*sqrt(Wh) (D is effective depth, W is weight being dropped, h is height of drop) (meters, KN unit)
If your site is surrounded by existing buildings & facilities, not so good idea to utilize this method of compaction because it may affect seriously to the existing building. You may face to a lot of complains & sues later.
As far as I know, in Vietnam there were not any building project that used this method to densify the soil. The reason is that the minimum distance between site & existing buildings is not provided frequently.

Another method may be suitable for the case is vibroflotation. The depth up to 20 meters can be improved by this method. The grain-size distribution is the most important to apply this method. Your loose sand must not be too coarse or too fine. The ideal size is about 80% finer than 5.0 mm & 80% coarser than 0.1mm. however, this method is applied well for the soil below the ground water level. Square or triangular pattern with spacing up to 2.0m can be used in this method.

 

[jdmm]

vohai,

Why do you need to compact! You apparently have at least 3 metres of non-liquefiable soil beneath your footings. Do you have a settlement problem for the building? For low relative density it is generally a low bearing pressure problem so that a large footing or raft will suffice and this will generally be cheaper than densification.

If your problem is dynamic (liquefaction) then you have to assess the problem in terms of potential settlement and collapse of the foundations in the event of liquefaction. Seed et al gives guidelines for&quot;freefield displacements and settlements&quot; in the event of earthquakes of various mag

Let me know what your particular problem is and I will try to direct you.

 

[vohai]

This building is an airport station in central area of
Vietnam and I am checking the design.
This site have no surrounding building and very near sea.
As I know this is not a seismic hazardous zone but I
do not know whether to consider any activity in airport
as dynamic load or resonant problem for soil. The water
content of soil is high. Design is based on very poor
site investigation. They used bore hole drill in sand and
got no undisturbed sample. There were no in-situ test such
as SPT... According to books for fine grained loose sand,
&quot;distortion modulus&quot; (by Soviet system, different from
Young modulus) is high enough and settlment is acceptable.
I suspect the result. Now we emphasise on plate load test
during construction to compensate poor site investigation.

 

[cbosy]

Plate load testing won't get you off the hook since it only tests the soils to a few feet. I personally would walk away from a job if the owner refused to do soil borings.

 

[npthao121]

Chosy is right, in this case the plate load test does not work. I think you should re-investigate the subsoil by bore hole. The cost for this work is not so high in your area. If the owner refuse to do so, you better walk off from the project.

 

[Rawand]

I have worked in the UAE for some time where fine snad is all you get from surface to the very bottom of the desert,,
SPT test (or Cone test) for such conditions is the most reliable one for loose soil, other wise you cannot assess bearing capacity nor settlement,,
In loose soils usually the settlement controls the allowable bearing capacity.
Please note that fine sands (if classified as silt) are high washable.

 

[jdmm]

vohai

Most design in sand is based upon density or relative density. You have characterized the soil as loose. If so then design on the basis of loose. For runways this should not be a particular problem with some improvement of the subgrade. For buildings the same applies except that the depth of improvement may increase depending upon the type and loading conditions of the building. A small terminal might be built on spread footings. A larger terminal may require DC or vibro or stone columns depending upon the loading conditions.

 

[vohai]

The site investigation report (bore hole with disturbed samples) indicates that soil is uniform and homogeneous so I think plate load test is acceptable in this case. (SPT is better but to persuade the owner to re-investigate the site I need strong evidence as written in building codes/standards). The report also tells that ground water relates with surface water from rain fall only.
Relative density of sand is 0.25, vertical load per column is about 100tons. The designer uses spread foundation 3x2m
on fill of medium grained compacted sand with thickness about 2m. Characteristic of natural sand below taken from tables of typical characteristics for loose sand and will be verified by plate load test. I now agree with his design.
Do you think this approach is acceptable?

 

[Focht3]

NO!


You have indicated the D[sub]R[/sub] of a &quot;fine grained loose sand&quot; is only 25 percent, and the pressure on the footing is about 16,667 kg/m[sup]2[/sup] (~3,400 psf.) You may get the structure completed, but it will likely be heavily damaged in a moderate to large seismic event, or during typhoon winds. The sands need to be improved before you build the structure -

Check out this site:

http://ikfia.ysn.ru/IUGG 2003/abst/ss04/015386-1.html

It is important to note that the Nam Chou fault has a 100 year maximum design earthquake of 6.9 - 7.0 (Richter scale) per the above reference, with peak ground accelerations that could approach 0.4 g. (In Los Angeles, The Big One is expected to have a magnitude of about 7.4) And these values are based on somewhat limited data; the actual values could be considerably higher. I don't know if that fault is close to your site; but it is unreasonable to simply dismiss the risk as &quot;acceptable.&quot;


You cannot get an idea of the uniformity of the site with disturbed samples! And a plate load test is absolutely worthless when evaluating a footing that is 3 meters by 2 meters in size...

Are you building a hotel? If so, what will the hotel's name be? I want to put it on my &quot;Do Not Stay There&quot; list...



Please see FAQ731-376 by [blue]VPL[/blue] for tips on how to make the best use of Eng-Tips Fora.

 

[sc]

OK as a consultant if the data is bad, then include the cost of correct testing in your fee. It simply justified as:

increased $$ now = less $$ due to future disasters.


If the client wont pay for higher quality testing then tell him to take the job elsewhere as it will be you in the court room if it fails due to inadequate testing.

Alternatively suggest that you can design the building based on worst case scenarios which will mean greater costs for construction than the cost of testing.



regards
sc

 

[vohai]

It is easier for me if I was the designer of this project. My company was hired for verifying design.
The site of SonLa hydropower plan as Focht3 mentioned is
in zone of frequent earquake (one for 10 years). But the site for my project have no earquake for hundreds years (or at least no information avalaible). The designer replaced
2m surface sand with medium grained compacted sand so
pressure at surface of natural soil is about 1kg/cm2
(10Tons/1m2). The influent zone of settlment is 2B=4m,
that is only 2m in natural soil. Plate load test will carry out at bottom of sand fill during construction can predict
characteristic of sand upto 1.2m (plate 2500cm2 will be used). It means that only 0.8m of sand we expect have the
same characteristic as 1.2m of sand ontop.

 

[ashjun]

Dear Vohai,
Hold on.

Soil Investigation report:

&quot;The site investigation report (bore hole with disturbed samples) indicates that soil is uniform and homogeneous so I think plate load test is acceptable in this case&quot;
What makes you think like that without the knowledge about the state of compaction?

&quot;Relative density of sand is 0.25...&quot;
THIS IS IMPOSSIBLE....even very very loose uniformly graded sand simply dumped in will exhibit a higher RD (please confirm the value by the use of simple tests at and near the surface by sand replacement method, balloon method, etc.)

&quot;..vertical load per column is about 100tons. The designer uses spread foundation 3x2m....on fill of medium grained compacted sand with thickness about 2m....&quot;
You shall have grave problems to the entire structure (with the utilities severely disrupted) if the geotechnical report is evenly remotely correct.

&quot;..will be verified by plate load test...&quot;
I could almost agree with you because in most of the airport and runway designs, plate load test is given a very high importance; but pay heed to the advice by others, it is not the best option.

&quot;...SPT is better but to persuade the owner to re-investigate the site..&quot;
Perhaps if your client is still interested in the ground improvement, then you could suggest that having SPT/CPT tests both before and after the improvement could verify the degree of improvement and hence settle the matter. (tackle the issue the other way round?)

Regarding seismic load I strongly agree to what Focht3 has said that it may perhaps be unreasonable to simply dismiss the risk.

Regards

 

[Focht3]

Remember that Vietnam, like California and the U.S. west coast, is on what the seismologists and vulcanologists call &quot;the rim of fire.&quot;



Please see FAQ731-376 by [blue]VPL[/blue] for tips on how to make the best use of Eng-Tips Fora.

 

[vohai]

Thank you all for your helpful advices.
I have checked the design and the result is:
settlement is about 8cm (high but ecceptable in our building code). I also tried N_value=8 would give OK result.
About seismic load I cannot insist the designer to
consider because I have no strong argument. In Vietnam
there is little number of design considering for seismic load. Until now I never heard of anyone hurt/injured by construction damage because of earthquake in Vietnam.

To Ashjun:
As I understand Rd=(e_max-e)/(e_max-e_min) and
sand is loose if Rd=0->1/3, so Rd=0.25 is not a wrong number, I think.

 

[BigH]

Vohai - maybe not yet, but what will hit the fan when it does?

 

[MRM]

8cm of settlement? Even if we were to assume that there are no seismic considerations and static loads are all we're worried about; do you think that 8-cm settlement will be uniform between all structural/foundation elements? Loose sand is more unpredictable than dense sands, i.e., it's more difficult to quantify compressibility or settlement with accuracy. It actually reaches a point where a sand becomes loose enough to where conventional settlement analyses based on linear elastic theory cease to give anywhere close to what could be considered a reliable prediction. There are too many things going on within the soil fabric. Couple this with sensitive interior and exterior finishes that are likely to crack if subjected to differential settlements of 2.5cm in 15m or so. Sounds like there could be trouble unless something is done to improve the site.

 

[Focht3]

The predicted 8 cm looks too small given the other parameters...even under static loads.



Please see FAQ731-376 by [blue]VPL[/blue] for tips on how to make the best use of Eng-Tips Fora.