Coupled Fluid-Particle Modelling of Debris Flow

The whole research can be divided into two parts – numerical modelling and
laboratory verification (flume test). The numerical model developed in this research is the key product and ultimate goal while laboratory test data serve as tools to improve the model and strong argument to demonstrate the feasibility of the model. A feasible model has to catch key characteristics of flume test result, such as deposition shape, the contour of the final deposition and frontal velocity. The research will start with kaolinite-water suspension, which can be assumed to be homogeneous within short experiment period. On latter stage, particles will be added to the suspension to simulate the real debris flow carrying rocks.
The proposed numerical method is the coupled Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) scheme with consideration of free boundary of fluid. The open source C++ libraries OpenFOAM (for CFD), LIGGGHTS (for DEM), and CFDEMproject (coupling of CFD-DEM) will be used.
Two settings of the model have to be determined along the research. First, the mesh size of CFD-DEM model. A finer mesh size is more accurate but requires longer computation time. 5-mm is a good trade-off which is recommended. However, it is subject to change or need further justification. Second, the roughness of the flume. The research starts with the non-slip condition. The channel roughness is represented by a single variable called valueFraction in the numerical model. This variable governs the relative velocity of the fluid adjacent to the boundary. For an extremely rough surface, the proper boundary condition is the no-slip condition, which assumes zero relative velocity between the surface and the fluid immediately at the surface.
In this research, Bingham model and HB model will be investigated. The two models are commonly used to characterize natural and experimental debris, the differences and modelling results of which will be discussed against experiment result.
In terms of flume test, three variables, i.e. slope angle, the volume of fluid and concentration of fluid, will be considered. Concentration affects the rheology of the kaolinite-water suspension. In the model, the concentration will be represented by viscosity and yield stress of the fluid. Rheometer in cement laboratory will be used to measure the viscosity and yield stress of the kaolinite-water suspension at different concentration. Once the three variables are matched between experiment and modelling, their results can be compared and discussed.
Experiments will be designed in a way that allows parameter study. The range of volume concentration, volume and slope angle experimented would be 16% ~ 21%, 2L ~ 5L and 10 ̊ ~ 15 ̊ respectively. The final deposition will be recorded by taking overview photos; deposition depth will be measured directly. Besides, two cameras, placed on the side and in front of the flume, will be recording during experiments. Image processing techniques, such as optical flow, will be adopted to find out frontal velocity and profile of the flow at different time.