(125b) Direct Numerical Simulation of Particle Rotation Effects in Gas-Solid Flows Using An Immersed Boundary Lattice Boltzmann Method
A new and efficient direct numerical method with second-order accuracy is presented for fully resolved simulations of incompressible viscous flows laden with rigid particles. The method combines the state-of-the-art immersed boundary method (IBM), the multi-direct forcing method and the lattice Boltzmann method (LBM). First, using the new IB-LB method, steady Stokes flows in ordered arrays of non-rotational and rotational spheres are examined extensively. The results of the non-rotational spheres show excellent agreement with existing theories. The present data are more accurate than previous computational results in the literature because of the higher accuracy of the IB-LB method. The rotational Reynolds number is employed to characterize rotational movements of spheres. The solid volume fractions are varied from 0.01 to the close-packing limit. For each solid volume fraction, rotational movements of spheres are simulated at several different rotational Reynolds numbers such as 0.1, 1, 10, 50 and 100. It is found that the lift force produced by rotation movement is directly proportional to the rotational Reynolds number and the drag force is barely affected by them. The lift force is very insignificant when rotational Reynolds number is lower than 0.1. However, it can be larger than the drag force as rotational Reynolds numbers increase especially in the presence of low solid volume fractions. It is shown that the lift force is around 43% larger than the drag force at the solid volume fraction of 0.01 and the rotational Reynolds number of 50. In the case with larger solid volume fraction, e.g., 0.5, the lift force can be as big as 39% of the drag force at the rotational Reynolds number of 100. Second, steady Stokes flows in random arrays of non-rotational and rotational spheres are also examined extensively. The lift force is also found to be very significant compared to the drag force when rotational Reynolds numbers are up to 100. It is noted that, through high speed imaging, investigators have found that about 20% of particles are rotating at high speeds in riser flow fields. The corresponding rotational Reynolds numbers of their reported mean and maximum rotations rates are 130 and 520, respectively. Therefore, simulations in the present study at the rotational Reynolds number up to 100 are not outside of the practical range. In previous drag laws, effects of the particle rotation on lift forces are not considered and thus the lift force is completely ignored. This study demonstrates that the lift force caused by the particle rotation can be very significant compared to the drag force and must be considered in practical simulations adopting drag laws, e.g., two-fluid simulations.