Dispersive, erodible, and slaking soils are prevalent over wide areas of the world. Dispersive clay soils collapse or disperse to form dissolved slurry when in contact with water. Using the material increases the risk of erosion in the dams during operation if precaution is not taken properly. The erosion usually occur by cracking resulting from different settlements caused by poor compaction in the vicinity of conduits. The failure is also common when the filling of the reservoir is rapid, however, some embankments have also failed after surviving an initial filling when later subjected to a higher water head than the initial filling.
The recommendation of the codes and references is to avoid using the dispersive soil if possible, however, there are some solution for dealing with dispersive where it is inevitable. Dams and other hydraulic structures can safely be built with dispersive clay materials if certain precautions are taken. The references state that "The measures include, but are not limited to, proper moisture and density control, use of filters and filter drains, select placement of materials, use of sand-gravel blankets or lime-modified soil on slopes, and chemical treatment of dispersive clays" (Icold Bulletin 1990, DPIW 2009).
Many dams have been built using the dispersive material in the world. In Australia, Eildon, Cairn Curran, Ross river, and Mole River dams (SMEC, 2020, Lesleighter, et al. 2004, Swindon, et al. 2003, Fell, et al., 2015) were upgraded using the dispersive soil. The filter solution has been used as an effective method to deal with internal scoring. In Eildon dam, as part of the embankment raising works, the issues with piping were addressed by the installation of properly designed filters to the top of the dam. The use of sand filters was adopted over other options such as lime stabilisation. In Cairn Curran dam, the existing embankment materials were highly dispersive and very fine. To figure out the internal erosion, a two stage fine filter was designed with a very fine filter, produced from rock crusher dust, used in locations where the likelihood of piping was highest. In Ross river dams, the existing embankment materials were highly dispersive with very prevalent “tunnel” erosion of the outer shoulder. Risk of piping through the crest of the dam was considered an unacceptable risk and the existing filters did not meet modern design standards. Therefore, filters were retrofitted in the critical area along the length of the embankment. In Mole River dam, the downstream filter arrangement comprises both a fine filter and coarse filter. The downstream filters were designed to satisfy critical filter criteria as detailed in Fell et al (2015). In view of the potential dispersive nature of the core material, the maximum d15 size of the fine filter has been set at 0.5mm rather than 0.7mm as adopted for non-dispersive core material, based on the recommendations given in Fell et al (2015). However, for design purposes it was considered that no-erosion filter tests would be undertaken to confirm the acceptability of the dispersive core/fine filter arrangement, and hence the required fine filter grading.
Richards, et al. (2007) presented a review of published literature on soil piping phenomena. They indicated that the recent work on piping highlights the limitations of the occurrence of piping and the role that design and construction may play in a large percentage of piping failures. They also showed the standardized laboratory procedures are available to assess piping potential in cohesive materials, but no such methods exist for noncohesive soils, however; methods are available for evaluation of self-filtration potential. Paige-Green (2008) has summarized the fundamental differences in the material, discussed current tests methods to differentiate between them and reviewed the techniques for refining them as construction material. He has concluded that the state-of-the-art has progressed little since the problem of using dispersive was discussed at the 1985 problem soils conference. NAGY, at al. (2015) presented the background of the dispersive soils, and the mechanism of failure, the tunnel erosion. They concluded that the identification of dispersive clays is a multi-faceted assessment process, where none of the methods of the identification ensures 100% certainty. They have indicated that the pinhole test is a well-used method, but for the certainty, it should be confirmed by other techniques, such as the geoelectric measurements, or the standard void ratio. McCook D. k. (2016) has presented the required samples and tests for identifying dispersive materials. Minimum 20-30 samples has been recommended for even small embankment projects and Crump test has been introduced as the best method of establishing whether dispersive clays are likely present. TOSUN (2006) briefly presented tests to determine the dispersibility properties of soils and summarizes the procedures to be adopted for embankment dams with dispersive soils in Turkey. He outlined the geotechnical investigation to determine its failure mechanism and also introduces the construction measures taken into account for Tınaztepe dam in Turkey, which failed due to internal erosion of dispersive soil along the conduit. Guan (2018) described the design and construction challenges of the Marlborough dam in New Zealand by the use of the dispersive loess material, and the innovative approach to processing of complying filter material that were used in the development of a cost constrained dam, in an area of high seismicity. The key risks to dam safety were assessed as the local seismicity and the potential for internal erosion of the dispersive loess material used in the dam shoulders. Singh, et al. (2018) tried to analyze the basic characteristics, problems and stabilization with suitable additives. They concluded that the Double Hydrometer test and Chemical analysis of pore water extract test are more conservative in showing the dispersion of soil. They indicated that the pinhole test is more reliable as it simulates field conditions and the crumb test gives a good indication of the soil for potential tendency to erosion. They also showed that the strength of dispersive soil increases with increase in lime, alum and gypsum content up to certain limit.
This paper presents concerns and challenges of embankment dam design with length of 15 km at the saddles of a big reservoir in east Africa with a total capacity of 30 BCM at normal water level. The main and saddle dams and hydropower plant are under construction at the moment and the saddle dams are deemed as high hazard structure since they hold 40% of the reservoir capacity in the minimum foundation level.