About erosion mechanism

[slam00000]

Is this the right erosion mechanism explination ? Please advise:

The differential settlement along the base of a highly compacted sand dam will result in a vertical crack breaking up the dam. If the internal compressive stress is not high enough to squeeze the crack shut, it will remain open and water will ingress into it from the surface or from the neighboring saturated zones. The walls of the crack will most likely be under tension and if the dam cannot resist this tension (through its unsaturation-induced suction) and/or if the water flowing into the crack exerts pressure that renders the effective stress vanished, erosion will take place.

 

[cvg]

sounds a bit odd but I suppose possible.

first, dams are rarely built of sand. If so, they usually have a clay core zone which limits the amount of water flowing through cracks.

I suppose differential settlement could create a vertical crack, however subgrade collapse or settlement more often creates a horizontal crack at the base of the dam (usually on the wet side). Differential settlement also causes low spots along the crest of the dam which are prone to overtopping and then failure.

I'm not sure there is any "internal compressive stress" which is "squeezing" cracks shut in a sand dam. What mechanism could create such a force in a sand embankment? A clay core would expand to fill cracks.

There are a couple of typical failure mechanisms for flow through a crack in a dam.
1) Internal erosion is when flow through cracks / along outlet conduits or abutments erodes the material internally and creates a "pipe".
2) High phreatic pressure exiting from the toe of the dam carries away fine materials. Given enough time, this can also create a "pipe" which would allow the reservoir to drain and cause a dam failure.

In order to cause failure of the dam, both of these mechanisms require sufficient hydraulic head to cause the erosion. They also require that the embankment soil can form a stable "pipe" to allow the sediment transport to progress. This second requirement may not be present in a sand dam unless there is sufficient clay mixed in with the sand to provide the necessary cohesion to allow the pipes to form. In fact, clean sand filters are often used as crack stoppers in flood control dams. The sand is installed in a central filter through the embankment or in a filter diaphram around a conduit penetrating the dam to prevent erosion from progressing through the dam and causing failure.

 

[slam00000]

-I am talking about a tailings dam rather than embankment dam: such dam, which retains loose tailings behind with upstream portion commonly resting on such tailings and built from sand, experiences differential settlement of 1.5 m along its base ( 60 m ).

-I do not know really the consequence of this huge settlement and how can be fixed (references will be appreciated)..?

-Regarding "internal compressive stress": with the progress of construction the dam will be raised resulting in compressive pressure on its underlying portion (which starts to act as foundation).I meant that such compressive stress may be large enough to push the walls of the cracks toward each other (and thus close the crack)
- When it comes to settlement outcome, I thought that that compromised freeboard that might cause the overtopping is mainly due to the overall total settlement rather than the differential settlement that causes piping/ erosion concerns. Please confirm
I should admit that my experience on this subject is less than humble and any suggestion, references, comment will be helpful

 

[fattdad]

It seems to me that the settlement of the foundation soils will be greatest where the embankment is the tallest. That means the body of the embankment will be in compression from the settlement. This would not suggest a mechanic for foundation settlement to lead to tension cracks.

I'm not sure about this topic, just wondering. . . . .

f-d

¡papá gordo ain’t no madre flaca!

 

[cvg]

I would question whether your tailings are truly "sand" or more likely SM or SC soils. If so, they behave quite a bit differently than sand. I doubt that a crack in granular material can close up due to compressive forces due to additional load on the embankment.

Loss of freeboard due to settlement is common. However, differential settlement causing just a portion of the embankment to settle is the worse case and more likely to cause failure during overtopping. The overtopping flow is concentrated at one location in this case producing higher velocity flow and more likely to cause erosional failure than if the entire embankment overtops.

 

[slam00000]

If such settlements are anticipated in the design stage, one should modify the desigh.
But if they occur during operation, how one can repair or account for them

 

[cvg]

to prevent piping / internal erosion failure, you need to install a crack stopper. There are many ways to do this. One way is to place a buttress fill of granular material to filter the seepage. Another is to place a cutoff wall or slab. Another is to install a central filter.

To prevent overtopping, simply add material to the top of the dam to increase the freeboard. Or, add an auxilliary spillway.

Or build another dam downstream to augment the original dam.

 

[dgillette]

You have identified a 'relatively' common failure mode for regular water-containment dams. It may also have happened in tailings dams of the sort you have been talking about on other threads - I don't know. One can put filters and drains in the dike of an upstream or centerline tailings dam (see Chapter 7 of Vick's book), but it isn't always feasible. This is one reason that it is good to keep the liquid slimes (and free water) as far back from the crest as possible with the materials and geometry you have to work with (location of decant tower, keeping the pond drained as low as practical, et cetera). The thicker the beach and dike, the less chance there is for a through-going crack that could allow erosion. (This also benefits slope stability.) The looser material of the beach is less likely to sustain an open crack than the compacted dike is. If you are cycloning, the coarse material may work as a filter/drain.

For a water-retaining structure, one would require a filter 99% of the time. (Maybe only 98%.)

 

[cvg]

in my experience, filters are required 100% of the time - unless you can show through incremental damage assessment that piping failure is no more hazardous than the PMF...

 

[dgillette]

I've never seen that particular incremental-damage criterion. Piping failure can be a whole lot less damaging than the PMF, and still be destructive and have potential for loss of life. Some states allow an incremental-damage approach for flood-induced failure, but that's more often for spillway capacity than for piping.

The 2% is reserved for low-hazard dams, in some cases homogeneous dams with flat slopes where there are no suitable materials available for zoning within a very long haul distance, temporary cofferdams, etc.

 

[cvg]

For detention structures or non water containing dams, piping failure cannot occur without sufficient pool elevation to produce the hydraulic gradient necessary to initiate piping. So it just is not likely to occur except during extreme rainfall events. for PMP rainfall producing PMF flooding, downstream areas would be flooded anyway. Piping failure may result in a flood wave that is only slightly higher than the existing flood level due to the rainfall. Especially when routed downstream to the population center. Therefore, constructing an expensive central filter in the dam might not be required by the state.

Again, for overtopping failure, the spillway is theoretically passing the majority of the PMF before the failure occurs, so the downstream channel is already seeing large flows. The wave created by the breach may cause only an incremental increase in the flood elevation downstream.