If you make a seepage analysis you will be able to find the gradient between soil layers. If the filter criterias are not respected between the soil layers than there might be some piping (interanl erosion). I say might as you must also look at the velocity of particle flow to see if the it is strong enough to push the small particles out. You would also usually see colored water coming out of the dam as the color is for small particles.

Also, for remediation, at the toe of the dam, or any place where the piping comes out, you must install a dissipation "chamber", a pile of boulders with a geotextile to keep the material in the dam. The weight of the boulders should be able to reduce the existing force of the water and the geotextile will keep it in. you could evaluate the lenght need and enbed the boulders in the toe soil 2/3 under surface 1/3 outside.

Hope this helps,

Thanks Olivier for your answer, another question will be, what happened with the dispersive behavior, maybe a change in the gradation band or some recommendation about compaction?.

As you mentioned, compaction is something to focus on (usually 98% PROCTOR). Then further, some people are for, and some against the use of lime to combat the effects of dispersivity and internal erosion with dispersive soils. I haven't formed my opinion on this yet. There are some papers on the topic from the 2012 ICOLD conference (lime treated materials for embankment and hardfill dam, Herrier, Lesueuer et al, and How to prevent embankments from internal erosion failure, Dr. JJ Fry).

What tests did you use to evaluate the dispersivity of the material?

I did the following test for every sample from different boreholes: 

1.Pin-hole test

2.Doble hydrometer test

3.Chemical analysis.

I had the opportunity to read some papers about the use of lime for dispersive soils and I encountered some questions.

Internal erosion is associated with plasticity index, more PI means a low probability of internal erosion, but if I use Lime, this decreases the plasticity index of the sample.

Any suggestion?

Hi Johnatan,

I worked on this subject years ago. I have put all my findings in below paper.

A. RezaGholilou, R.Khorasani (2007), “Seepage design criteria in embankment dams “, Icold symposium, Russia

I remember that some experts from Netherlands suggested some relations for gradient or velocity. You can find them in ICOLD bulletins

Keep me updated if you find something new

Reza, where can I find this paper?

Thanks a lot Reza. I will  check your paper.

Hi Reza, I'm also interested in your paper pls, do me a favour to send a copy to my inbox. Thank you in advance.

Dear Johnatan Rivera

Introduction: Dispersion piping occurs in susceptible clays called dispersive clays. The erosion caused by dispersion can proceed unnoticed within a clayey fill until it is too late to control.

For dispersion piping to occur it is essential that there exists an initial concentration of seepage through a continuous crack in the clay core of an embankment. Dispersion piping will thereafter expand this crack by the continual detachment of clay particles (deflocculation) from the main body of the clay into the minute flow stream. The initial crack might have been caused by one a combination of the following: desiccation, differential settlement, saturation settlement, hydraulic fracturing.

The factors that influence the dispersivity of clay are (i) the clay mineralogy (ii) the clay structure (iii) the clay pore water chemistry, that is the types of cations in the pore water and their concentrations (iv) the degree of cementation of the clay by oxides of iron (v) the plasticity index of the clay. The two predominant factors that influence dispersivity are (i) the percentage of the adsorbed sodium cations, Na+, to the quantities of other cations such as Ca2+, Mg2+, K+. (ii) the concentration of dissolved salts in the reservoir water.

Control of dispersion piping: That erosion due to dispersion piping is occurring can be proved by the excessive turbidity of the seepage exiting from the downstream end of a dam.Dispersion piping can be prevented from developing by the provision of properly designed graded sand filter that performs by a filteration and sealing process. Even though the filter cannot block the passage of the colloidal particles, it is able to restrain the silt-size particles from invading the interior of the filter. These silt-size particles are thus retained in the leakage channel upstream of the filter and gradually seal the leak.A remedial measure for an existing dam that is already leaking is to grout the core using a mixture of equal proportions of cement and bentonite with a water solids ratio of 4 to 5 through holes drilled from the crest. It is essential that the grouting pressure at any dam level does not exceed the overburden pressure. Suitable core material is clayey soil with a plasicity index in the range of 15-25%. Clays with higer plasticity index are unsuitable because they tend to swell as the moisture content increases and such swelling is detrimental to dam stability. During construction, the fill material should be properly compacted at a moisture content about 2 to 3% wet of optimum in order to achieve flexibility. A coefficient of permeability less than 10-4 mm/sec will prevent dispersion piping from developing.

In addition, earth dams, constructed of dispersive materials, should be zoned with the core protected by downstream sand filters. According to Sherard et al. 1984 sand filters with D15 less than 0.5 mm provide adequate security against dispersion piping. Homogeneous earth dams are strongly discouraged. Attached is a pdf file that contains (a) chart for selecting filters for clays (b) a chart for identifying dispersive clays (c) and a flow chart for identifying dispersive clays.

Identification of dispersive clays: It is essential that the clay in the existing dams in question bee proved to be dispersive because transverse hydraulic cracking of the core could open up leakage channels for seepage. Hydraulic fracturing occurs due to negative effective stresses in the clay caused by arching. Since clay consolidates it will settle more than the surrounding shell thereby making the latter support the clay and reducing the vertical stress in it. Since substantial pore pressures may exist at the end of construction negative stresses may occur. Also the key factor in new dam design is the ability to identify a dispersive clayey soil prior to its use in an embankment. (1) Dispersive soil can be tentatively identified in the field by the occurence of extensive erosional pattern accompanied by tunnels and deep gullies together with excessive turbidity in any storage water. (2) Laboratory identification of dispersive clay is done using physical tests such as (a) Sherard's pinhole test (b) Emerson's crumb test complying to USBR 5400. (c) Volk's (SCS) double hydrometer test that are carried out using distilled water; (d) chemical tests carried out on the soil's pore water that yield the parameters (i) ESP, exchangeable sodium percentage, defined by the ratio of exchangeable Na+ to the CEC in % (ii) SAR, sodium absorption ratio = Na+/[0.5Ca2+ + 0.5Mg2+]1/2 (iii) percent sodium = 100 times Na+/TDS, where CEC = cation exchange capacity, TDS = total dissolved solids = Na+ + K+ + Ca2+ + Mg2+ , and all quantities are measured in units of meq/litre. It is essential that the tested samples be maintained and tested at their natural moisture content because drying alters dispersive characteristics. In addition, large number of representative samples from the site should be tested.

The interpretation of test results, especially the chemical tests, can be a tricky task . The double hydrometer test has a reliability of about 30%. The pinhole test is considered to be reliable. There is no simple correlation between the chemical parameters and dispersivity. Generally the greater the exchangeable sodium cation content of a clay relative to the other cations the greater to the susceptibility of a clay to dispersion. The lower the dissolved salt concentration of a reservoir water the greater the tendency of sodium-saturated clay to dispersion. The following quantitative classification could be made. (a) A soil is likely to be dispersive if SAR > 10. In Australia it was discovered that a soil is dispersive if the SAR > 2 provided the TDS lies between 0.5 and 3.0 meq/litre. (b) A soil with ESP ≥ 15 and which can have its free salts leached by distilled water is dispersive. (c) A soil with percent sodium > 40 is dispersive provided the TDS > 0.5 meq/litre.

For more details and charts Johnatan Rivera should refer to the (1) Proceedings of 9th European conference on SMFE (2) Advanced dam engineering for design, construction and rehabilitation, 177-182. (3) Sherard, J.L, Dunnigan, I.P. and Talbot, J. R. (1984) filters for silts and clays. ASCE J. geotech. Engrg. Div., 110(6).

Dear Johnatan.

Please, go through my post again. I have added new matter to it and provided a pdf file containing 3 charts for evaluating dispersivity and designing filters for silts and clays.

http://www.nrcs.usda.gov/internet/fse_documents/nrcs142p2_024606.pdf

http://www.usbr.gov/pmts/hydraulics_lab/pubs/R/R-91-09.pdf

http://www.usbr.gov/pmts/materials_lab/pubs/earth.pdf