**7. Laboratory and field tests for analyzing erodible and special soils such as dispersive**

### **7.1 Analysis of the soil erodibility**

Several devices have been developed to evaluate how resistant earth materials are to water flow. Some of them are the *rotating cylinder* to measure the erosion properties of stiff soils (e.g. Chapuis & Gatien, 1986); the *jet test* to evaluate the erodibility of surface soils (e.g. Hanson 1991), and the *hole erosion test* to measure the erosion properties of stiff soils (e.g. Wan and Fell 2004). Another popular device developed in the early 1990s to measure the erosion function is the called *Erosion Function Apparatus* ‒EFA‒ (Briaud et al., 2001). The EFA test (Fig. 12) consists of eroding a soil sample by pushing it out of a thin wall steel tube and recording the erosion rate for a given velocity of the water flowing over it. Several velocities are used and the erosion function is defined through the results of this test (Briaud et al., 2001).

Fig. 12. Erosion Function Apparatus ‒EFA‒ (Briaud et al., 2001)

#### **7.2 Identification of dispersive soils**

296 Soil Erosion Studies

Fig. 11. Recommendations to protect levees in urban areas exposed to flooding (Auvinet et

In addition, Figures 10 and 11 illustrate respectively some practical recommendations that should be taken into account for the protection of levees on the river banks, and levees that

**7. Laboratory and field tests for analyzing erodible and special soils such as** 

Several devices have been developed to evaluate how resistant earth materials are to water flow. Some of them are the *rotating cylinder* to measure the erosion properties of stiff soils (e.g. Chapuis & Gatien, 1986); the *jet test* to evaluate the erodibility of surface soils (e.g. Hanson 1991), and the *hole erosion test* to measure the erosion properties of stiff soils (e.g. Wan and Fell 2004). Another popular device developed in the early 1990s to measure the erosion function is the called *Erosion Function Apparatus* ‒EFA‒ (Briaud et al., 2001). The EFA test (Fig. 12) consists of eroding a soil sample by pushing it out of a thin wall steel tube and recording the erosion rate for a given velocity of the water flowing over it. Several velocities are used and the erosion

(c)

are built in order to protect urban areas exposed to flooding (Auvinet et al., 2008).

function is defined through the results of this test (Briaud et al., 2001).

Fig. 12. Erosion Function Apparatus ‒EFA‒ (Briaud et al., 2001)

al., 2008)

**dispersive** 

(a)

**7.1 Analysis of the soil erodibility** 

(b)

Additionally, as it was mentioned before, soil erosion is likely to occur in certain types of soils. Among those are certain types of clay which erode by a process called dispersion or deflocculation, that occurs when the clay mass is in contact with water. If water is flowing, individual clay particles are detached and carried away through erosion channels or *pipes* that can form rapidly. As it is established by Cedergren (1989), one of the problems related with dispersive clay action, is that the deflocculation process starts as soon as there is a significant flow of water, as it can occur through poorly compacted or cracked layer in an impervious core, or along inadequate bonded contacts with rock foundations, abutments, or outlet conduits extending across the impervious core.

The practical relevance of dispersive clays in dam engineering, started about 60 years ago after realizing that it was the main cause of piping failure of several small earth dams and levees. Most of the earth embankment failures caused by dispersive soils occur during the first filling. If there are no well designed and constructed filters upstream and downstream of the core embankment that has these clays, the probability of an internal erosion failure will be very high. This probability might increase when preexisting surface erosion caused by rainfall contributes to the formation of superficial channels that become connected to tunnels originated by internal erosion. As this type of soil is not possible to identify through the conventional index tests, it was necessary to develop certain laboratory and field procedures for its identification.

Whereas the susceptibility to erosion in cohesionless soils, such as fine sands and silts, is due to high values of water flow velocity, hydraulic gradients and seepage forces, normal clays are usually erosion resistant, except for water velocity higher than 1 m/sec. Nevertheless, for dispersive clays the erosion phenomena occur due to causes that are different to those associated with granular soils. Such causes are due to the following characteristics:


Internal Erosion Due to Water Flow Through Earth Dams and Earth Structures 299

The wet slope of an earth dam, lake or river banks and channel slopes are frequently subjected to sudden changes of water level (increments or decrements), which modify flow conditions inside the soil mass. Flow velocities, hydraulic gradients and seepage forces might, in extreme conditions, cause soil erosion problems ranging from slight to severe, such as piping or even the total failure of the structure. These phenomena, known as *rapid filling* and *rapid drawdown*, are complex problems in which the magnitude and rate of filling or drawdown, hydraulic conductivity and porosity of materials constituting the earth structure, geometry of slope and initial boundary conditions of flow are involved (Auvinet & Lopez-Acosta, 2010). By using the *Plaxflow* algorithm (Delft University of Technology, 2007), analyses to assess how these two phenomena affect soil erosion in earth structures are carried out through a numerical modeling based on finite element method (FEM) (Auvinet

In this practical example, the effect of rapid drawdown phenomenon on erosion problems in a typical embankment is analyzed (Auvinet & Lopez-Acosta, 2010). Simplified geometry of the studied domain and boundary conditions considered in this analysis are shown in Figure 13. Rate of drawdown was established at 1.1m/day. Thus, a total dewatering of 5.5m in 5 days was assumed in this analysis. In the same way, it was accepted that embankment is constituted by a homogeneous and isotropic material with hydraulic conductivity *k*=1×10-

30 m

Fig. 13. Simplified geometry and boundary conditions of the studied embankment (Auvinet

From results of analyses (Auvinet & Lopez-Acosta, 2010), Figure 14 shows, for a typical time interval during drawdown (*t*=4 d), the free surface line which separates unsaturated material (upper part) from saturated material. Variation of this free surface, called *desaturation line* (for drawdown), obtained at different time intervals during rapid drawdown is illustrated in Figure15. Other authors prefer to call it *phreatic line* (Huang & Jia, 2009; Lam & Fredlund 1984; Lam et al., 1987). It must be underlined that this free surface line is not rigorously a flow line since velocity vectors cross it (see Figs. 16a and b). In the same way, results from analysis demonstrate that during drawdown, when water surface descends, large velocities are generated at the contact between the level of water and the slope (which are proportional to hydraulic gradients, and consequently to seepage forces in that zone); these velocities can

Homogeneous and isotropic soil

Impervious boundary

6 m

**8. Practical examples to illustrate the analysis of earth erosion problem caused by rapid filling and drawdown conditions in embankments** 

**8.1 General settings** 

Final level 0.0 m

& Lopez-Acosta, 2010)

& Lopez-Acosta, 2010; Lopez-Acosta et al., 2010).

5 m/s and porosity *n* = 0.3 (void ratio *e* = 0.43).

12 m

Initial level 5.5 m

=26.57°

**8.2 Analysis considering only rapid drawdown phenomenon** 

embankment, where there exists a high hydraulic gradient; such phenomenon progresses upstream, forming a kind of pipe until it reaches the water source. With dispersive clays, however, the internal erosion is due to a deflocculation process and it might start at the upstream side where there is the water source; the tunnel-shape passage or pipe, that is formed, is propagated toward the downstream side. If such dispersive soils exist in areas where there are already some cracks, or not well compacted zones as those presented along conduits, such cracks might increase and propagate very rapidly causing a dam failure.

#### **7.2.1 Field and laboratory test for dispersive soils**

During the field investigation to construct an earth dam, particularly when looking for the materials that might be used to construct the embankments, it is very important to identify the existence of dispersive soils. This identification should be done first through one of the special field tests that exist for this purpose. Although the results of such tests must be verified through laboratory tests, field tests might give a good preliminary evaluation of the dispersivity of the soils under investigation. Knodel (1988) presents a good description of the most common laboratory and field tests that are used in the engineering practice to identify the dispersivity of soils. Among them are the following: a) for field: *crumb test*, *water drop test*, *dissolved sodium test*, and *turbidity test*; b) for laboratory: *crumb test*, *the double hydrometer test*, *pinhole test* and *chemical test*.

#### **7.2.2 Design considerations when constructing with dispersive clays**

For any earth construction it is necessary to investigate, by using one or more of those methods mentioned above, the existence of dispersive soils; this investigation can be carried out through soil samples obtained in open wells during the soil exploration phase. Once the construction materials have been identified, a decision to use or refuse them has to be taken. Sometimes dispersive soils might be used in earth structures if they are mixed with lime or if well designed filters and drains are installed. If for economical reason it is decided to use dispersive clays, the following conditions have to be taken into account:

