Slope Design Near Railroad

[RobPE]

I have a project at a south american coal mine where the goal is to mine as closely as possible to a nearby rail line to maximize coal reserves. We're dealing with about 80 ft of overburden in the area. My initial thoughts are that conventional slope stability analyses (method of slices or FEM) would be very limited in their effectiveness at analyzing such a case because of the sensitivity of the rail line (ie the factor of safety or probability of failure approach to slope design does not help to anticipate 2cm displacements which could disrupt a rail line). Does anyone have any thoughts or related experience on how to analyze and design slopes in this situation? Does anyone know of any guidelines that may be helpful?

 

[fndn]

Could you drive sheetpiles a few meters away, then do stepped or benched slopes into the coal mine? What type of foundation material are beneath and also make the shoulder of the railway line?

 

[RobPE]

That's an interesting idea. Are you suggesting perhaps say 10-20 ft. of sheetpile supported slope at the crest, then a benched sloped from their down another 60-70 ft. to the coal seam? Any ideas for design criteria for sheet pile slopes in this case (Fs=1.5?)? The overburden sediments are stiff lean-fat clays/silts and minor sands (uniax ~500 psi, residual friction angles from 10-25 deg.). Immediately below the rail itself is an embankment that varies in height from 0 to 50 ft. I'm not sure what the embankment is made of.

The sheet pile idea would of course be contingent on the economics. The cost of the sheetpile would have to be outweighed by the benefit of the coal. It may be cheaper to stay further away from the rail and miss out on some coal, but it's certainly worth consideration. Any other thoughts?

 

[Mccoy]

I would take no chances.
I've heard of some cases where numerical methods did not predict soil displacement and settlements of areas not even too close to the excavation top.
I followed myself a case of a house far enough away from an excavated face which settled, whereas geotechnical theory alone would have it stayed safe. But soil relaxation due to the excavation was the only reasonable cause, at least the most significant one.
Also, excavations sometimes modify groundwater regime with consequent sometimes unpredictable settlements.

 

[RobPE]

Mccoy fndn and others, based on your experience, what do you think would be adequate stability criteria for slope design near critical facilities (rail, houses, misc. buildings, etc)? My initial thoughts are maybe something like the greater distance between say Fs=1.5 at facilities or crest to facilities distance ~= 3x slope height?? What do you think?

 

[fndn]

RobPE;

I have always sampled soil near the crest and near the toe and used the lower C and phi values at residual strength. Then I would take the maximum dry density(proctor) and assume that the slope will be no worse than 90% compaction. Based on those,c, phi and gamma, you may run a FS against rotational failure, against infinite slope and also look for deep seated failure(by taking a 50 foot boring from the toe). in addition, account for soil creep, especially if clay soils or tension cracks exist at the present. Based on those results, you may choose say a minimum of 7-10 m from the slope crest as a setback.

 

[dgillette]

By all means check slope stability with suitably conservative parameters, but regardless of the FS, it would be prudent to also analyze deformation. It may lead you to some particular slope (maybe the suggested 3:1, maybe steeper, maybe flatter) to compensate for the reduction of Sigma-H and resulting small deformations that would occur w/o actual rupture and instability (i.e., governed by stiffness rather than strength).

Imagine a rectangular FEM mesh with a point load to represent a coal train perpendicular to the 2D model. If you were to start at the top lefthand corner, removing elements (by setting their weight to zero), how many could you remove before there were significant shear strains and settlements of the tracks?

I agree with McCoy on the groundwater issue potentially being very important, and possibly contributing much more settlement than the actual excavation. The effect of lowering the water table would, however, probably be generalized settlement rather than differential settlement, and if they're coal and sedimentary rock, the slope and foundation may not be very compressible.

DRG

 

[Mccoy]

The effect of lowering the water table would, however, probably be generalized settlement rather than differential settlement, and if they're coal and sedimentary rock, the slope and foundation may not be very compressible.

I agree about relative uncompressibility of lithified rock, but if we are dealing with loose overburden soil, as RobPE reports, effects may also be localized, if sand stringers or lenses for example are involved.

Also, as far as deformations are concerned, dgillette's suggestions appear very sound: conventional slope analysis is based upon strenght parameters and limit state, stability or instability, whereas a FEM deformation analysis may be more appropriate in your settlement problem.