The consolidation process could be slow in an upstream tailings dam; therefore, the stability can reduce due to an increase in excess pore pressures when the dam is raised. The safety of the dam can be enhanced by constructing rockfill berms on the downstream side. This paper presents a case study on the strengthening of an upstream tailings dam with rockfill berms. The finite element analyses were performed for modelling the staged construction of the dam and for optimizing the volume of the rockfill berms. The dam was raised in 11 stages; each stage consisting of a raising phase and a consolidation phase. The study shows that the slope stability of the dam reduced due to an increase of excess pore pressures during the raising phase. The stability of the dam was successfully improved by utilizing rockfill berms as supports on the downstream side. A technique has been presented to minimize the volume of the rockfill berms so that the required stability can be achieved at minimum cost. This paper shows that the finite element method can be a useful tool for modelling the consolidation behaviour of an upstream tailings dam and minimizing the volume of the rockfill berms that may be needed to maintain the stability of the dam during staged construction.

No abstract

SUMMARYThe paper presents an optimization routine especially developed for the identi"cation of model parameters in soil plasticity on the basis of di!erent soil tests. Main focus is put on the mathematical aspects and the experience from application of this optimization routine. Mathematically, for the optimization, an objective function and a search strategy are needed. Some alternative expressions for the objective function are formulated. They capture the overall soil behaviour and can be used in a simultaneous optimization against several laboratory tests. Two di!erent search strategies, Rosenbrock's method and the Simplex method, both belonging to the category of direct search methods, are utilized in the routine. Direct search methods have generally proved to be reliable and their relative simplicity make them quite easy to program into workable codes. The Rosenbrock and simplex methods are modi"ed to make the search strategies as e$cient and user-friendly as possible for the type of optimization problem addressed here. Since these search strategies are of a heuristic nature, which makes it di$cult (or even impossible) to analyse their performance in a theoretical way, representative optimization examples against both simulated experimental results as well as performed triaxial tests are presented to show the e$ciency of the optimization routine. From these examples, it has been concluded that the optimization routine is able to locate a minimum with a good accuracy, fast enough to be a very useful tool for identi"cation of model parameters in soil plasticity.

An investigation of the potential to numerically model the no erosion filter test is performed here, where the flow through a large ensemble of particles is considered by applying minimisation of dissipation rate of energy on the ensemble that is discretised with modified Voronoi diagrams and Delaunay triangulation. Low-Reynolds number simulations are applied to each part of the Voronoi diagram using computational fluid dynamics. The mechanical friction between particles is modelled by increasing the effective viscosity for closely spaced particles. Microscopic mechanisms for successful and unsuccessful sealing of filters are obtained. The numerical results agree with previously presented experimental observations by Sherard and Dunnigan. A conformity is that the sealing starts from the end of the channel and continues outwards in the radial direction. The sealing implies that the permeability can be reduced several orders of magnitude during a test.

To enable assess slope stability problems efficiently, various machine learning algorithms have been proposed recently. However, these developments are restricted to two‐dimensional slope stability analyses (plane strain assumption), although the two‐dimensional results can be very conservative. In this study, artificial neural networks are adopted and trained to predict three‐dimensional slope stability and a program, SlopeLab has been developed with a graphical user interface. To reduce the number of variables, groups of dimensionless parameters to express stability of slopes in classic stability charts are adopted to construct the neural network architecture. The model has been trained with a dataset from slope stability charts for fully cohesive and cohesive‐frictional soils. Furthermore, the impact of concave plan curvature on slope stability that is usually found by excavation in practice is investigated by introducing a dimensionless parameter, relative curvature radius. Slope stability analyses have been conducted with numerical calculations and the artificial neural networks are trained with dimensionless data. The performance of the trained artificial neural networks has been evaluated with the correlation coefficient (R) and root mean square error (RMSE). High accuracy has been found in all the trained models in which R > 0.999 and RMSE < 0.15. Most importantly, the proposed program can help engineers to estimate 3D effects of a slope quickly from the ratio of the factors of safety, FS3D/FS2D. When FS3D/FS2D is large (such as larger than 1.2), a 3D numerical modelling on slope stability analyses that can consider complex 3D geometry and boundary condition is advised.