In the field of mining and dredging engineering, oftentimes slurries are involved in transport. Understanding the behaviour of these slurries in a flow is important as the energy and water consumption need to be determined. The slurries have an interesting rheology due to the presence of clay. The rheological parameters cause the flow to be non-Newtonian. Further, the presence of coarse solids (sand particles) in these slurries also influences the rheological parameters. Through Computational Fluid Dynamics (CFD) simulations, these aspects can be predicted.
Previous work in non-Newtonian CFD with coarse solids made use of a free surface flow through a rigid-lid approach (van Rhee, 2017). A shortcoming of this is that the flow needs to be uniform and the flowdepth needs to be known beforehand. In the mining and dredging fields, this is not the case and a different approach is required.
This report will focus on multiphase CFD simulations. One of the fluids will be a non-Newtonian model including sand particles, and the other fluid will just be air. The sand particles will be subjected to transport, segregation, and settling behaviour.
OpenFOAM is the weapon of choice, but as it stands it does not have a solver that’s capable of including sand particles. The interFoam solver is chosen as a starting point and it’s code is adjusted. A Bingham Plastic transport model is implemented, including sand particles. A sand transport equation is also implemented.
Using the adjusted interFoam solver, two sets of simulations are ran. A first set of simulations is performed using a 2D mesh for an open channel. The first simulation utilizes a Bingham Plastic fluid without any sand particles included. The resulting velocity profile is compared to the analytical solution and is found to agree well. The second open channel simulation includes sand particles. The results show a sand bed forming as well as the sand concentration to be roughly constant throughout the plug flow. This is compared to experimental work and concluded effects is captured well from a qualitative perspective.
The second set of simulations is performed using a 3D mesh for a pipe section. These results were not as satisfying as the results for the 2D open channel. Unfortunately, all 3D pipe simulations either showed the pipe to fully fill up with the Bingham Plastic fluid, or the solver crashed at some point. Many attempts have been made at this, and no conclusive reason has been pointed out as the cause for this.
It’s recommended to continue research in this direction with the main focus being the simulation stability.