sheetpile design Kp reduction factors and mobilized phi 1

[danyul]

can someone explain this to me?

for sheetpile design, Kp is calculated several ways but im looking at Caquot and Kerisel. in school i was taught to reduce the phi of the passive soil by using a mobilized phi' = arctan(tan(phi)/FS)

use the charts to get the Kp

and then use the reduction factors to get the final Kp.

this seems like double reduction... is there a reason for this or is that too conservative?

thanks

 

[fattdad]

Sometimes it helps on questions like these to know what you are designing and the consequence of failure.

I am familiar with factoring tan phi, but don't often do this myself. For the earth pressures that I usually calculate, I usually use phi and then a safety factor on the overall structure (e.g., retaining wall).

f-d

¡papá gordo ain’t no madre flaca!

 

[DRC1]

I have not seen Kp = arctan(tan phi)). This may be to account for friction against the sheet. The friction provides a vertical reaction, which reduces the horizontal component on the passive presure. This may be the mobilized phi you are talking about.

Anyway, to use the tables:

1. Compute your Beta (retained slope angle to horizontal) and divide by phi.
2. enter the table from the bottom at the apporpriate phi and travel up to the correct beta over phi line. Go left and write down the result. For 34 degrees and no slope(B=0) and phi = 34, The result is about 9.2
3. Compute delta over phi. Delta is the angle of friction of soil against the sheet. I typically use 17 degrees for quartz sand on steel. Thus if phi = 34 degrees, delta over phi = .5 and R=.746 for phi =30 and R=.674 for phi= 35, thus with linear interpretation for phi = 34, R= .69
4. Kp = the result in 2 times R. For phi = 34 degrees,
Kp = 9.2 x .69 = 6.3
5. I then divide that by a factor of safety, which I typically set to 1.5. Some people do not reduce the Kp by a safety factor, but then add 20-30% to the length of the toe. I prefer to out the safety factor on the Kp.

 

[danyul]

ok i got in touch with my prof... he said to think of the coquot and kerisel Kp reduction not as a FS but the way to get the calculated Kp based on the interface angle. so then you need to factor in your own FS another way... either the tan phi method or on the structure as fattdad said.

 

[PEinc]

If you reduce your Kp by a safety factor before you run your wall calculation, you will get a higher bending moment and a heavier sheet pile or soldier beam than you would without the using the safety factor. However, there are commonly used references which show using the full Kp, calculating the bending moment, pile size, and embedment and then just increasing the toe embedment by at least 20% in order to provide the safety factor to the passive resistance. The second method is common and cheaper. Both methods work.

 

[kleo]

The reason I reduce Kp by a FS(~1.5) is since the wall tip will not fully mobilize Kp unless you are in a failure condition. Obviously for a proper design the wall tip does not move to this extent. Thus a reduced mobilized Kp should be used for design as it takes much more movement to mobilize Kp vs Ka. I then add 20% to the length.

Alternatively if you use a nonlinear elasto-plastic FEM model such as provided in Plaxis, the full phi value may be used in the model and any reduction in passive resistance is accounted for by the computed deflections. Note also the wall bending moments will be substantially lower since the wall flexibility is included in the analysis. I then factor the wall moment by 1.6 per IBC for section design.

 

[PEinc]

kleo,

I have designed and built a few thousand sheeting walls. Almost never have I reduced the Kp unless the project specifications insist. It isn't necessary and is not economical. As I wrote in a previous forum discussion, I do not know any specialty contractors in the design/build anchored wall business who regularly use Plaxis or other FEA programs to design their walls. Using these types of programs is overkill for almost all projects. For anchored walls designed using the FHWA/AASHTO LRFD method, factored loads are used and the full Kp is used with a resistance factor = 1.0.

 

[DRC1]

I agree with PEinc except that I always put the factor os safety on KP, but do not increase the length. Either method is acceptable, to do both is putting a factor of safety on twice, which is not neccesary. I reduce the Kp because I am usually working with little or no data, and often what data I have, I do not have complete confidence in. Using FE programs, to be useful (which is different than getting an answer) requires higher quality data than is available for most wall designs.

 

[PEinc]

Well said, DRC1. If you assume a hinge at the base of the excavation and then determine the ground (or hinge) reaction, then you will still get the lower bending stress in the pile. Then, knowing the ground (or hinge) reaction, you can either solve for the toe embedment without the safety factor on Kp and then increase the embedment by at least 20% OR you can solve for the toe embedment using Kp reduced by the safety factor. Both calculated toe embedments should then be similar. The hinge at subgrade method has been used for many years and is in most text books.

With computers today, tedious calculations are easy. Many engineers forget about, or never knew, the old ways of design which still work fine today. If a proven method gives a better, more economical answer, shouldn't engineers have an obligation to use that method?

 

[PEinc]

P.S. My last response is for braced or tiedback walls.