While the piles will increase the FOS, they are not likely to increase it much.  This of course depends on the specifics of the situation.  One reason piles do not increase the FOS much is that they are not very stiff perpendicular to the direction of movement of the potential slide planes.  Therefore they cannot take a significant portion of the load.  Another reason is the magnitude of the loads involved with the potential slide planes.

All of that said, it is more likely that the end effects due to having a relatively short section (parallel to the slope crest) for the slide to form in is much more significant than the piles.  However, without doing a 3D slope stability analysis, which I have never seen done well, there is no way to qantify the effect.

Generally bridge approach fills are constructed to a stable configuration. Abutment piles are not looked upon to improve F.O.S.

There are lots of information required to provide a discussion onthe topic - such as height of fill, slopes, type of failure plane determined from stability analysis.

Debate of F.O.S 1.4 versus F.O.S 1.5 should not occur. I see no reason to determine additional piles to achieve the magical 1.5.

Agree with both - no significant difference between 1.4 and 1.5 as FOS.  Errors in modeling, material parameters, etc. would be of greater concern. I remember reading that in Ottawa, they did a probability analysis of a slope and found that a slope with a FOS of 1.5 still had a 20% chance of failure!  Go figure.

check out:  www.ctre.iastate.edu/mtc/papers/2004/thompson.pdf
He references a paper by Poulos in Canadian Geotechnical Journal, 1995.

Thanks all for your advice. It is just to satisfy the Depart of Transport's criteria of having the FOS = 1.5. If not major changes like flatter slope and hence additional bridge spans will be required or building a retaining wall at the bottom of the slope.

If piles do contribute to increasing the FOS by small amount, it will be a economical solution rather than using other means. Increasing the FOS from 1.4 to 1.45 which is closer to 1.5 maybe acceptable to them.

How do you carry out enhancing the FOS using computer analaysis?

thanks

"inclusions" ( piles for example ) is a solution to stabilize slope. We use in france a computer program called TALREN which takes into account any type of inclusion ( piles , nails, anchors, etc ). Although this program is well known in Europe , i don't know whether it will be accepted by your DOT !

ozziz,

Typically, the geotechnical consultant will give recommendations for improvement of overall slope stability if it is less than that required by the DOT.

Some DOTs allow short-term FOS of less than 1.5 if it can be shown that long-term (drained) FOS is greater than 1.5.

Talk with your geotech about how severe the problem is. For a recent project, we were finding that for undrained condition and side slopes of 2H:1V, the FOS was computed to be less than required by the DOT, but was ignored because it was considered a mostly cosmetic issue - shallow sloughs on the slope surface. Deeper failure surfaces under the same conditions had the minimum FOS required.

Jeff

Jeffrey T. Donville, PE
TTL Associates, Inc.
www.ttlassoc.com

good point Jeff on the "infinite slope" vs what I would call the real culprit (deeper).  In granular slopes, we would almost always pick a critical "size" of the slide - a minimum of mass that would be considered significant and relate that to the FOS rather than the lower infinite slope value.

Just a couple of words about the probabilistic analysis mentioned by BigH (BigH, as you know, "probability" always captures my eye!). Saying that a slope with a 1.5 FOS has still a 20% probability of  failure means that, treating the FOS as a random variable, the set of all possible output values for FOS include a 20% subset which are lesser then 1.0, hence indicating a failure state. The 1.5 value in this case would generally represent a statistical parameters such as the mean. The errors or uncertainties cited by BigH are well known to us all and include variability of soil parameters plus measurement errors. Probabilistic analysis rises a problem of treshold value: which would be an accepted max value for probability of failure?

I tuned in late, so I'll sound off on multiple topics in one over-long post.  First off, GeoPaveTraffic is right about the inefficiency of piles for slope reinforcement, particularly if they have to bridge a very thick layer so the moments get very large.  If you had just a single plane of weakness (say, a soft clay layer within the dense sand and gravel), the pile would only need to provide shear reinforcement and might work well.

A few years back, the Corps built a seismic rehab for a dam in MS (not far enough from the New Madrid Fault) and used piles (which were really beams installed like piles) to span a fairly thin weak layer of liquefiable silt.  They did a boatload of 3D numerical analysis to verify the concept (especially transfer of force and moment from beam to soil), and they had to use an awful lot of resteel in the beams, which were not off the shelf.  Anyway, it can be done, but only for fairly thin layers, and expect to fight the analysis all the way.  The analysis is more complicated than rock bolts or piles for vertical loads.

For the probabilistic discussions, remember that not all factors of safety are created equal.  Most of us (especially geotechs who are used to great uncertainty in material properties) tend to add minor conservatism at every step of the way - ignoring the cohesion, using the highest piezometer of the set, etc. - so a calculated 1.2 may be absolutely bulletproof.  Is your 1.4 really 1.4, or is there unnecessary conservatism built in?  If you only have one material parameter to deal with, like phi' for a dry sand, it could be reliably stable with something less than 1.2 -- UNLESS -- there is some overlooked detail, in which case higher FS may not do any good.  On the other hand, if you're dealing with a cut in soft clay or something like that, a calculated FS of 1.5 may not be very reliable at all, as in the example cited by BigH.  With undrained shear strength, there can indeed be uncertainty in strength that exceeds the margin provided by FS=1.5, especially if the clay is at all sensitive.  About 20 years ago, I saw an analysis by one of the early leaders in geotechnical risk analysis that showed an undrained FS of 1.3 for an upstream-type tailings dam gave a more reliably stable slope than 1.5 with a drained analysis, by incorporating the uncertainties associated with each of the input parameters.  This was done for the purpose of showing the state regulators that a good 1.3 was better than a typical 1.5.  This was NOT smoke and mirrors, but a realistic depiction of the uncertainties of soil parameters and different forms of analysis.

An appropriate probability of failure?  Depends on how much a failure would cost and how much it costs to raise the FS.  If failure could kill people, real small.  If it means just a few $K for regrading, not so small.

dgillette,
the one you mentioned is a very true and relevant point: due to customary conservativism in evaluating soil parameters (or neglecting some of them) the usual deterministic analysis would yield an output value nearest to say, the 25th, 30th percentile rather then to the mean value.
By the way, are you in on the "riskanal" mail list (risk analysis forum?) your monicker appears to ring a bell.

Or sometimes even lower than the 25th percentile.

Indeed, Mr./Ms./Dr. McCoy, I am part of riskanal.  So if you read that one, you probably know exactly who I am, and that among my main (professional) interests are embankment dams, earthquake engineering, and risk analysis.  I can no longer hide behind the anonymity of the internet.  (A recent cartoon in the New Yorker or somewhere showed two dogs sitting at a computer.  One says something like "On the internet, nobody has to know you're a dog.")

Nice to see the world of tech fora isn't so wide after all.
I'm a Mr., and my monicker is derived from the legendary jazz pianist McCoy Tyner. I always use the same, except in riskanal where a full name is required (I rarely show up there, though, and I'm glad the geotech field has some representatives).