Lattice Boltzmann model for dense suspended particles based on improved bounce-back method

Abstract

We develop an efficient and accurate lattice Boltzmann model for dense suspended particles based on improved bounce-back. It is difficult to model flow with dense suspended particles because there are not enough fluid nodes surrounding each particle to apply many existing algorithms, and mass leakage due to interpolated bounce-back is more significant when dealing with many particles. We propose a new curved-boundary treatment and fresh-node initialization scheme, both of which are still effective when two particles get close to contact. We further propose a method to calculate the hydrodynamic interaction accurately when two solid particles get very close to each other without leaving any fluid node between their surfaces. An easy redistributing method is developed to fix mass leakage resulting from curved boundaries. To make the algorithm more stable, we combine implicit particle-velocity updating and Galilean-invariant momentum exchange. We validate the second-order accuracy of this curved-boundary method by channel flow, and other simulations are carried out to show that this model is more accurate and stable in dealing with particle particle interaction. Finally, simulations of magnetic particle flows are conducted to show that this model is suitable for dense suspended particles.