Cohesive forces are implemented in a discrete-particle fluidized bed simulation using both a square-well potential and a Hamaker model for van der Waals forces. This simulation is used as a tool to gain greater understanding of the hysteresis behavior that has been observed experimentally during fluidization–defluidization cycles. For the parameters investigated, the results show that cohesive forces are significant in the pressure overshoot observed during the fluidization process. Furthermore, the mechanisms by which the specific cohesive mechanisms causing the pressure overshoot are different depending on the cohesion model being used. Although both models indicate that particle–particle cohesion dominates the overshoot, the square-well model predicts that cohesive interactions between particles and distributor plate also play a role, while the Hamaker model for van der Waals forces predicts that particle–sidewall friction may be enhanced by the presence of cohesion. The results are in accordance with a one-dimensional force balance on the system, and provide alternative explanations for trends in existing datasets.

Cohesive forces are implemented into a discrete-particle, fluidized-bed simulation using a square-well potential. The square-well description treats cohesive interactions as instantaneous, binary events, thereby making it a viable option for the incorporation of cohesion into a kinetic-theory-based continuum model. Cohesive forces are also incorporated into the simulation using the more elaborate Hamaker description of van der Waals forces in order to provide a basis for assessing the square-well model. Both cohesion models are implemented in the discrete-particle framework of the MFIX software package. A mapping method is also developed to convert material-specific Hamaker constants into equivalent square-well parameters. The corresponding results from the two models are compared both qualitatively and quantitatively. The predictions of the square-well model are on par with the Hamaker model with respect to mixing level, particle mobility and minimum fluidization velocity. Subtle differences are observed between the two models in cases that involved such high levels of cohesion that the particle bed could not fully fluidize.

Bubble formation in a microfluidic flow-focusing device is simulated using the volume-of-fluid approach to achieve a complete solution of the Navier–Stokes equations for both the gas and liquid phases. The results of the simulation show good agreement with previous experimental results. A detailed examination of the predicted pressure and velocity profiles from the simulation also provide further validation for the conclusions drawn previously with experimental results. The simulation results show the existence of two distinct modes of bubble formation. Simulations of systems an order of magnitude smaller than those investigated experimentally indicate that such reduced systems sizes are a viable approach that would result in much smaller bubble sizes.

In the present study, rapid granular flows with attractive inter-particle forces are investigated. In particular, cohesive forces are incorporated into hard-sphere (molecular dynamics) simulations via a square-well potential. The square-well potential treats cohesive forces as both binary and instantaneous. For simple shear flows, an investigation of the input parameter space indicates that two distinct flow regimes are present. For relatively large cohesive forces, the formation of a large, single agglomerate is observed. For moderate cohesive forces, the sheared system is composed of mostly 2-particle, dynamic agglomerates that are fairly evenly distributed throughout the domain. Furthermore, the results for this latter regime indicate that cohesion attenuates the magnitude of the stress components at higher solids fractions (in the collisional regime) as compared to the non-cohesive case. At lower solids fractions (kinetic regime), however the presence of cohesive forces has little impact on the observed stress.