Abstract
The scope of this study is to examine the response of a relatively new type of dams, the Face Symmetrical Hardfill Dams (FSHD), and their behaviour during construction, impoundment, normal operation, and seismic loading. The static and dynamic analysis of dams is a complex process, which requires a thorough study of the behaviour of the materials used in the dam body. Parameters such as inhomogeneity, non-linearity under cyclic loading, complexity due to geometry of the terrain, and the interaction of dam – reservoir water - canyon make the mathematical problem particularly complex. Simple methods based on approximate analyses were initially used to study dams. Nowadays, rigorous mathematical methods have been developed in order to simulate the behaviour of such systems more accurately. Finite element analysis is one of them, allowing optimization of the design and construction of these projects by finding the best and most economical solution. In this thesis, Hardfill Dams are investi ...
The scope of this study is to examine the response of a relatively new type of dams, the Face Symmetrical Hardfill Dams (FSHD), and their behaviour during construction, impoundment, normal operation, and seismic loading. The static and dynamic analysis of dams is a complex process, which requires a thorough study of the behaviour of the materials used in the dam body. Parameters such as inhomogeneity, non-linearity under cyclic loading, complexity due to geometry of the terrain, and the interaction of dam – reservoir water - canyon make the mathematical problem particularly complex. Simple methods based on approximate analyses were initially used to study dams. Nowadays, rigorous mathematical methods have been developed in order to simulate the behaviour of such systems more accurately. Finite element analysis is one of them, allowing optimization of the design and construction of these projects by finding the best and most economical solution. In this thesis, Hardfill Dams are investigated to evaluate the effect on the performance of various factors such as: the geometry of the valley, the height of the dam, the inclination of the two slopes, the stiffness and strength of the hardfill material, and the intensity of ground shaking. The Face Symmetrical Hardfill Dams are characterised by their trapezoidal cross-section, with a slope typically of 0.8:1 or 0.7:1 (H:V) on the upstream and downstream faces. The spillway is incorporated in the dam body. A reinforced concrete waterproofing element is used on the upstream face of the dam, while prefabricated reinforced concrete elements are used on the downstream face. The hardfill is a low cement content material, with a content less than 80 kg/m3, which may contain fly ash and achieve typically a compressive strength of 5-6 MPa after 90 days. Aggregates are selected from spoil heaps close to the project area or from necessary borrow areas or quarries, provided that they satisfy strength requirements. An analysis of a full three-dimensional finite element model introducing the interaction of the dam, the valley and the reservoir water is carried out. The general-purpose finite-element software Abaqus was used for the numerical analyses. Initially, the staged construction and reservoir impoundment are investigated, and subsequently the response of the system to intense seismic shaking. The study simulates the behaviour of hardfill under cyclic loading using the Lee & Fenves (1998) damage plasticity constitutive model for concrete. A comparison with the elasto-plastic Mohr Coulomb model, extended with tension cut-off, confirms that both models yield results that are almost identical. Time-dependent effects such as creep and swelling are ignored, but plasticity and hysteretic behaviour are considered. In the dynamic analysis, the simulation uses as seismic excitation the Lefkada 2003 earthquake record, scaled to a PGA = 0.4g. Viscous elements are introduced at the base of the valley, dependent on the wave velocity and material density of the underlying canyon rock, to absorb the radiated energy of the outward propagating waves, according to the formulation by Lysmer and Kuhlemeyer (Lysmer & Kuhlemeyer, 1969). Opposite vertical boundaries are connected in a manner that enforces equal displacements (periodic boundaries). The reservoir water is simulated with acoustic finite elements. These elements can accurately describe the hydrodynamic water pressure, considering water compressibility. The dam body and the concrete panels interact with each other by frictional constrains allowing slippage and separation. By contrast, the dam body is tied to the canyon rock, whereas the water is tied to the upper surface of the concrete panels and the canyon rock. A series of parametric studies using two-dimensional analyses were carried out to investigate the effect on the performance of the hardfill dam of canyon rock stiffness, the compressive strength of hardfill, the intensity of ground shaking, and various dam geometries. The response derived from the 2D analyses demonstrated the good behaviour of this type of structure in cases of seismic excitation with PGA up to 0.6g. Damage and plastic deformations of the material develops only locally at the upstream toe area for low excitation. As the acceleration increases, stresses and damage increase in the dam body. The tensile stresses are higher at the upstream face and the compressive stresses higher at the downstream face, as a result of the hydrostatic and hydrodynamic pressures. Sensitivity analysis in various foundation conditions from Es = 1GPa to 24GPa showed a satisfactory response in all cases analysed. The displacements, the plastic deformations and the tensile damage were found to increase in more flexible foundation material. A sensitivity analysis of an asymmetric cross-section hardfill dam with an upstream slope inclination of 0.4:1 (H:V) was used to examine whether a more economical alternative is feasible. The results showed that the usage of such geometry is possible in low seismic excitations. In very high excitations (e.g., PGA ≥ 0.8g) the accumulated damage within the dam body becomes extensive. The analysis of a zoned hardfill dam with higher material strength in the areas where high stresses are observed, proves that such an increase of strength can limit the damage in the dam body and improve the behaviour of the structure. Two existing typical projects were analysed as representative of two categories of medium height and tall dams, the Filiatrinos Dam in the Peloponnese and the Cindere Dam in Turkey. The Filiatrinos Dam, used as a case study of medium dams, has a height of 55 m and slope inclinations of 0.8:1 (H:V). The Cindere dam, used as a representative of large dams, has a height of 107m and upstream and downstream slopes of 0.7:1 (H:V). In addition, a hypothetical dam of height equal to 150 m and slope inclinations of 0.7:1 (H:V) is used to study the behaviour of taller dams. The results of the analysis of the 3D case study of Filiatrinos dam shows that this dam has a satisfactory response to seismic action. For a seismic excitation with a PGA of 0.4g, the calculated displacements are very small, while the compressive and tensile stresses during seismic excitation are much lower than the material strength. The damage that develops in the dam body is minor and local. The concrete slab does not show any defects during seismic shaking. The results of the case study of a 100 m high dam with a geometry similar to that of the Cindere dam in Turkey showed that the behaviour of this type of dam is satisfactory. For a seismic acceleration with a PGA of 0.4g, the horizontal displacements remain low, and the stresses remain lower than the material strength. Damage is minor, located mainly at the upstream foot of the dam, and can be treated by placing higher strength material at the dam-foundation interface area. An incremental analysis was performed to study the response of the dam under various seismic excitation levels. The results are similar to those of the 2D analysis. Damage is local for low intensity excitations, but for higher intensity it extends towards the upstream face and dam bottom. The influence of transversal, longitudinal, the combination of transversal and longitudinal excitation components, and the full 3D seismic excitation (including the vertical component), is analysed in relevant case studies. The results of the analyses demonstrate the necessity of 3D analysis in this type of projects due to the significant influence of the longitudinal component when applied simultaneously with transversal. It is concluded that neither 2D nor 3D analysis under only transversal excitation can produce fully reliable results. The results of the case study of a 150 m high dam shows that hardfill dams could be constructed in such heights. A variation in zoning is required to provide higher material strength in areas where tensile stresses are excessive. It can be concluded that properly constructed hardfill dams can satisfy safety criteria under strong seismic conditions, even in the case of large heights. At the same time, such dams have several advantages compared to embankment dams and roller compacted concrete gravity dams, making them an attractive alternative. The easiness of construction, adaptability in poor foundation conditions and low cost are important factors that need to be considered in the dam type selection procedure.
show more
