Safe distance for loading beyond crest of a slope

[Ned_Geo]

Long time lurker, first time poster, so any help is appreciated... I'm looking at an existing slope/gully which is essentially marginally stable, both on inspection and when analysing (FOS = 1.0 roughly). We've been asked to provide indication of a safe 'set-back distance' from the crest of the slope for which a crane loading can be safely applied. Normally we'd simply model the loading in SlopeW, aiming for a FoS >1.3; however given that the slope is marginally stable anyway, is this a pointless exercise?

Secondly, on the same slope, we've also been asked to assess the distance from the crest of the slope which may be affected following a seismic event. Looking at nothing more than a crude (pseudo-static, applying peak horizontal PGA) approach, is it valid to suggest that the distance beyond the crest which exhibits a FoS<1.1 may be liable to deformation/movement?

 

[Mad Mike]

The minimum set-back distance behind the embankment crest should be related to the geometry of the potential failure surface(s).

The crane should be positioned sufficiently far back that it doesn't add any further load onto the soils within your theoretical failure plane. I would start by plotting the profile of a potential failure plane (I'm not sure I'd use a FOS as low as 1.1 given all the variables), then setting back your crane footing to avoid loading any soils within this plane. Then I'd set it back a bit further just to be sure.

Another consideration is how your crane footing will be affected once the theoretical failure actually occurs, with the ground level being modified for the failed soil volume. In my opinion if there is a risk of instability, it is not sufficient to place your crane foundation marginally behind the potential failure...you would need to assume the failure will actually occur and then step back progressively. You may also need to consider the construction time-frames (I'm assuming it's a temporary crane).

I'm not familiar with seismic events (living on a passive margin of South Africa), so can't help with point 2.

All the best,
Mike

 

[GeoPaveTraffic]

My suggestion is to add the crane weight to your stability model near the crest and run the model. If the FOS is less than the case without the crane, move the load back and rerun the model. Keep repeating until the adding the crane doesn't result in a lower FOS.

Since you are using Slope/w and its search routines are not great, you will need to run a lot of failure planes. My suggestion is run at least 50,000 (300,000 would be better) surfaces that cover the area starting near the crest and extending back several hundred feet.

Mike Lambert

 

[Doctormo]

The crane is usually a couple of isolated point or strip loads of finite length so modeling as continuous line or strip load in a conventional slope stability program would tend to be overly conservative. A 3D slope stability modeling would probably be more appropriate for a crane on surface behind a slope. Not sure if Slope/W can handle this complexity or not.

Since that is usually not a practical solution most of us, you just have to take the slope stability analysis for what it is, a check on an improper model, and consider that also.

 

[Okiryu]

Doctormo is correct. The crane loading is not a strain plain (2D) problem. I think SLIDE software can add line loads which represents concentrated loads.

Also, is the crane loading, a temporary loading? If so, I think that seismic analysis may be neglected: there are few chances that the earthquake and your temporary loading to act at the same time.

 

[Ned_Geo]

Thanks for the replies Guys.

Unfortunately we don't have the resources to perform a 3D analysis - our report and analyses are caveated as such!

Just to clarify the seismic issue - that's an additional analysis, carried out as a separate analysis to the crane loading.

 

[Lomarandil]

I agree that 3D has the potential to help (especially if a known failure surface focus exists and the crane can be placed with that in mind). In my experience, for shallow failures, you might get an extra 10% compared to the 2D case.

But if you had a relatively uniform slope without many variations in the 3rd dimension, and the failure surface isn't evident (or is substantially granular and so not driven to repetition by a remolded clay layer).. then you'd essentially be forced to consider a slope failure starting adjacent to an outrigger -- which becomes a 2D problem again.

I'd follow Mike's method.

For part 2, I think pseudo-static and 1.1 could be reasonable, depending what the client really wants to know by defining an "affected" area. Is it the contractor wanting to figure out where they should place their laydown area, or is a developer wanting to place a permanent structure there? For the latter, you might need to sharpen the pencil a little further than pseudo-static methods.

----
The name is a long story -- just call me Lo.