Back Analysis Of Rock-fill Material Parameters Using Particle Swarm Optimization

Earth-rock dam has some obvious advantages, such as its lower cost, shorter period of construction, local material used adequately and its adaptability for various foundations etc, so it would be adopted widely in future hydroelectric power development of our country. As the increasing of dam's height, the safety of earth-rock dam would be more and more important. For evaluate the operating state, back analysis of monitoring data is one of the main methods, so the study of the back-analysis method is very significant.In this paper, the development of back-analysis research in geomechanical engineering has been reviewed. And then a intelligence optimization algorithm — Particle Swarm Optimization (PSO) has been introduced. For rock-fill material parameters' estimation, the detail of the PSO algorithmic, selecting parameters and application of back analysis in earth-rock dams have been discussed.In order to improve the convergence of the back analysis of the rock-fill material parameters, four improved PSO methods have been proposed in this paper. Two are convergence improved PSO, and the others are the PSO based on multi-objective problem.Finally, taking the Xibei Kou concrete face rock-fill dam as an example, the mechanical parameters of its main rock-fill material have been estimated using four improved PSO methods. Before the back analysis, the creep component of the vertical displacements has been deducted from the total vertical displacement. The results indicate the validity of improved PSO methods in the back analysis of the mechanical parameters of concrete face rock-fill dam. The parameters results can reflect the material qualities full well, by adopting monitoring data at different spaces and time. And using the PSO based on multi-objective problems can bring better effect, if back-analysis is taken as multi-objective problem.