Simulation of deterministic energy-balance particle agglomeration in turbulent liquid-solid flows

An efficient technique to simulate turbulent particle-laden flow at high mass loadings within the four-way coupled simulation regime is presented. The technique implements large eddy simulation, discrete phase simulation, a deterministic treatment of inter-particle collisions and an energy-balanced particle agglomeration model. The algorithm to detect inter-particle collisions is such that the computational costs scale linearly with the number of particles present in the computational domain. On detection of a collision, particle agglomeration is tested based on the pre-collision kinetic energy, restitution coefficient and the van der Waals' interactions. The performance of the technique developed is tested by performing parametric studies of the influence the restitution coefficient $(e_{n} = 0.2, 0.4, 0.6$ and $0.8)$, particle size ($d_p = 60, 120, 200$ and $316 \mu{m}$), fluid inertia ($Re_{\tau} = 150, 300$ and $590$) and particle concentration ($\alpha_{p} = 5.0 \times 10^{-4}, 1.0 \times 10^{-3}$ and $5.0 \times 10^{-3}$) have on particle-particle interaction events (collision and agglomeration). The results demonstrate that the collision frequency shows a linear dependency on the restitution coefficient, while the agglomeration rate shows an inverse dependence. Collisions among smaller particles are more frequent and efficient in forming agglomerates than those of coarser particles. The particle-particle interaction events show a strong dependency on the shear Reynolds number $Re_{\tau}$, while increasing the particle concentration effectively enhances particle collision and agglomeration. Overall, the sensitivity of the particle-particle interaction events to the selected simulation parameters is found to influence the population and distribution of the primary particles and agglomerates formed.