ESR11 Pablo Agustín Becker
Research project description
An enhanced Particle Finite Element Method for debris flows
Fellow: Pablo Agustin Becker
Department: International Center for Numerical Methods in Engineering (CIMNE)
Polytechnic University of Catalunya (UPC)
Supervisor: Prof. Eugenio Onate and Prof. Sergio Idelsohn
Debris flows are complex multi-phase phenomena: the mobilized mass is a multi-phase mixture, composed by solid particles and a fluid phase. As the flow runs along the slope, the high stress on the soil bed causes erosion that can lead to a displaced mass that is several times the original one. To ensure a correct simulation, numerical models must be able to handle this complex interaction. However, most current approaches are simplified, two-dimensional models that usually rely on empirical “black box” formulas to define the erosion rate. Parameters in these algorithms have to be carefully adjusted in order to obtain results that match real data, given that they limit their analysis to two dimensions and they are not able to simulate the bed. This is an important drawback that highlights the need for full tridimensional models that do not need empirical parameter setting, while at the same time yielding more accurate results. However, this is generally at the expense of a higher computational cost.
The goal of this project is to develop a 3D multi-phase model for debris flows that does not only simulate the fluidized zone, but a layer from the soil as well to compute correct soil stresses. The use of enriched FEM will allow for a correct modeling of both the soil and the debris flow, including the interaction forces between them and therefore predicting an accurate erosion rate. Moreover, this numerical approach naturally allows for the shape change of the eroded slope, which has a critical impact on the behavior of the debris flow. The algorithm will also make use of a large number of tracker “particles” that travel along a fixed mesh, allowing for the computation of interaction forces at a subscale from the FE mesh. By combining these two tools, the proposed model will be able to correctly simulate the debris flow and the soil bed beneath it, therefore allowing for a solution whose limitation is only the accuracy of the input parameters of the soil and the definition of the initial debris flow.. Figure 1 shows preliminary results obtained by using a very high viscosity Newtonian fluid as a first approximation for the soil.
The main collaboration will be with Lorenzo Benedetti (ESR 8), who is focused on the initiation phase. By combining the described fluid solver with his solution strategy to tackle the soil phase, the full multiphase solver will be obtained. Testing different strategies to detect strong points and weaknesses will be done with the other Work Package 3 members: Alessandro Leonardi (ESR 9) and Xiannan Meng (ESR 10). Finally experimental data to validate against real world data will be provided by Miguel Angel Cabrera (ESR14).