There are three basic construction methods for tailings dams, although a combination of two or all three methods is common. These methods refer to the raising technique and direction used, whether upstream, downstream or along a centerline. These dams can be built with material from borrow areas or with the tailings from processing, provided they have been treated and meet the geotechnical specifications of the project. To this end, this tailings can be subjected to additional processes such as cycloning, for desliming, and is then called a hydraulic embankment [7]. De Araújo [8] comments that in the use of hydrocyclones, in his case study, the underflow (outlet with coarser material) has a percentage of 78% solids by weight and the overflow has 35%. As it is the portion with the lowest humidity and is typically more granular, the product from the underflow is destined for use as construction material for the elevations.

Fig. 1 shows the construction methods typically used in tailings dams.

Figure 1 Source: Author

As Fig. 1 shows, different from the downstream heightening method, in the centerline and upstream heightening methods, the heightening is carried out partially on tailings that have already been disposed of. Although this generates significant savings in earthmoving (there is a smaller volume of material used to build the dam), it brings a certain complexity in terms of constructive control of the dam and the execution and control of drainage.

In the absence of data to adequately size the dam height, which should be sized according to the disposal rate and the elevation according to the topography of the site chosen for disposal, also in the absence of characterization of the materials to be used, for the conceptual design, parameters based on literature were adopted. In a conceptual way, in the absence of a specific dam location, which due to the topography would allow the dam height to be determined according to the occupied area. Table 1 presents an estimate of the occupied area and dam height, considering hypothetically a dam on flat terrain (which increases considerably the area occupied by the tailings).

Table 1

Occupied Area (ha)	Dam Height (m)
60,57	30
45,43	40
36,34	50

Source: Author

As for geometry and properties of the dam, a range of case studies is available in the literature. Regarding to material characterization (beneficiation tailings) and geometries adopted, works of Silva et al. [8], Albuquerque Filho [10], Naeini and Akhtarpour [11], Rafael and Romanel [4], and Rout, Sahoo and Das [6] were consulted. Such works present characterizations of tailings from different ores (iron, copper and aluminum) and dam geometries. It should be noted that none of the studies listed above is explanatory as to the filter and/or drainage system used and its dimensions, however several textbooks [12,13] present options for drainage design configurations for earth dams (applicable to tailings dams) and filter sizing (referring to particle sizes).

It is also observed that the same materials (tailings), from mines of the same ore, have variability of geotechnical parameters, a fact that goes against the different beneficiation processes, which are responsible for giving the materials different geotechnical and hydraulic characteristics, such as demonstrates the work of Silva et al. [8]. This is due to the different processing routes, which comminute an ore in different particle sizes and use different reagents in the physical-chemical concentration processes.

To determine the hydraulic parameters of the materials that make up the structure (heightening and tailings), an option to laboratory tests or field tests, is the use of mathematical models for indirect determination of the parameters (kh, kv and/or kh/ratio) as presented by Shamsai et al. [14].

Although most of the studies listed above deal with metal ore tailings, it was decided to use them as a basis for the development of the project, above all regarding the geometry and scale magnitude for the geotechnical properties of the design materials. It was decided to adopt, in a conservative way, geotechnical parameters with values close to the lowest values verified in the bibliography, since these referred to iron ore and copper tailings.

For the project, it was decided to consider dams made up of five materials (foundation, soft clay, landfill, disposed tailings and dense tailings) with properties according to Fig. 2, so that the parameters considered do not differ from the reality presented by real materials, although, for simplification, it was decided to consider isotropic materials.

Figure 2 Source: Author

As a construction method, due to its greater relevance (greater use for tailings dams) and economy (constructive, requiring less earth movement), as previously mentioned, the Upstream method of raising was chosen. Using elevations of 10 (ten) meters interspersed with benches of 10 (ten) meters, the slopes being considered with an inclination of 25°. As a preliminary project, a dam with 4 elevations was considered, totaling a global dam of 40 meters high. For the configuration of the tailings beach, the design parameters were used according to Silva [8] and Araújo [15], who based on empirical and observational models, considered in their studies the existence of two distinct phases, where the submerged portion presents greater inclination in relation to the emerged portion. This phenomenon is well described in works such as those by Araújo [15] and Machado [16], where they cite Vick apud Araújo [15].

As for the configuration of the tailings disposal, a tailing beach of 150 meters in length was considered between the theoretical pivot point of the tailings, at the dam, and the water line over which the tailings are submerged, typically with greater angulation as discussed by Silva [8]. As for the slopes adopted, a slope (i) of 0.5% was considered for the emerged tailings and 3.0% for the submerged tailings. Subsequently, the hypothesis of a water line under the tailings was observed, that is, a situation of greater risk and greater requirement for a filter/drainage system.

Regarding to the filter/drainage system, a chimney drain with a blanket was used in the project, in order to cover half of the foot and the entire central extension of the elevation, thus enabling its continuity during the elevations. The use of foot filters was also analyzed, as well as the option without a filter (hypothetical, system collapse). Fig. 1 shows the configuration of the proposed dam and a table with the constituent materials of the model.

A dam project has several options and geometry solutions that can change the dam performance in different ways, typically the most evaluated aspect is the Safety Factor (SF), however other aspects such as the percolation of water by the massif (dam), pore pressure stresses and hydraulic gradients, must also be evaluated and may be of extreme interest when evaluating the piping and liquefaction phenomena.

To the project in question, it was decided to evaluate the variation of the water line or “tailings beach” position, and its influence on the stability of the dam; as well as the internal drainage system of the dam, by foot drain system and by its hypothetical failure. The hypothesis of an additional elevation and the existence of a low resistance layer on the foundation was also evaluated.

The suggested tailings dam project met the safety requirements by the analytical method used, Limit Equilibrium. Under ideal operating conditions, water line at 150 m from the dam and internal drainage/filters in the proper configuration, the project presents a Safety Factor (S.F.) of 2.2; while under adverse conditions (collapse of internal drainage system and elevation of the water level), it still presents an S.F. of 1.8.

Although considering that the values obtained in the S.F.'s are above those recommended by standard (S.F. between 1.2 and 1.5), and thus the project is considered acceptable, other factors must be taken into account. First, since it is a hypothetical dam, several variables (topography, geology, materials, among others) are uncertainties. In this study, all parameters were estimated from the literature, while a real tailings dam project will certainly have laboratory test data to support the proposed solutions and access to greater resources for analysis.

Regarding to the analyzed scenarios, as expected, impacts on the S.F.'s, concentration of hydraulic gradients and pore pressures were verified; with variations in water level and position (tailing beach) and according to the internal drainage configuration. A scenario with an additional heightening (for a total of five heightening) was analyzed, raising the dam to a final relative height of 50 m. In this case, there is an S.F. of 2.1 in an ideal situation and an S.F. of 1.6 under adverse conditions, still within the safety limits.

Regarding to hydraulic gradients and pore pressures, it appears that these were sensitive to the changes tested. The lower dams and the internal filters are the most critical positions. As for pore pressure, it can be seen that in the proposed scenarios there was a variation along the critical rupture surface. These scenarios could (in the case of critical hydraulic gradients) cause triggers that lead to a dam rupture.

Based on the above analyses, it appears that the best measure to be taken to ensure the safety of the project is to move the water line away from the dam, which not only increases the dam's safety factors, but also reduces the pore pressures and hydraulic gradients. As for the internal filter/drainage systems, three options were evaluated: chimney drain with blanket, horizontal blanket drain and blanket drain only in the initial dike. According to the analysis carried out, a chimney drain with a blanket not only guarantees the highest safety factor, but also results in the smallest pore pressures within the dam slope and, is therefore, the most recommended option. The last analysis to be made would refer to safety due to liquefaction and piping; however, considering the design using a chimney drain with a blanket system in all heightening, and the respective hydraulic gradients (lower than the other options), which reduces the risk of landfill liquefaction; the dam can be considered stable at the level of the conceptual design.