Quantifying streamers instabilities in 2D particle sedimentation using DNS

Particle streamers formed dynamically strongly influence hydrodynamics in gas–solid multiphase flows, yet their quantitative characterization in sedimentation remains incomplete. We performed a direct numerical simulation (DNS) of quasi-2D particle sedimentation process using the lattice Boltzmann method (LBM) coupled with hard-sphere collisions to resolve both gas–particle and particle–particle interactions. Simulations span solid volume fractions of 0.15 and 0.40, density ratios of 100 and 1000, and terminal Reynolds numbers Re 0 = 3 and 30, corresponding to Stokes numbers St = 17–1667. Structure factor analysis identifies distinct instability modes: dilute systems develop bypassing and plug-funnel structures, whereas dense systems form void-like regions with retarded particle-gas coupling at high St. Domain averaged sedimentation velocities from DNS are compared with Richardson-Zaki and Yin correlations, showing enhanced settling at low St and hindered settling due to persistent plugs at high St . Filter based analysis of local drag demonstrates significant deviations from classical drag laws, with drag reduction in dense phases. Statistical analysis of velocity fluctuations reveals anisotropy, negative skewness linked to plug formation, and departure from Gaussian distributions. To quantify mesoscale heterogeneity, we introduced a heterogeneous index (H-index), which captures orientation dependent fluctuations in DNS fields. Results reveal that streamers are scale-dependent, which modify drag and sedimentation velocity through mesoscale heterogeneities in the flow field. These findings provide quantitative insights into the mechanisms governing gas-solid interactions and particle segregations in sedimentation systems.