Numerical Simulation Of Dispersive Mixing Using Discrete Lement Method For Particulate Materials

Dispersive mixing is one of the most important steps in material processing, particularlyin polymer mixing and compounding. During dispersive mixing, physical bonding ofdispersed phase, namely particulate agglomerate, is broken under hydrodynamic drag forcesapplied by dispersing fluid, which makes agglomerates erosed, ruptured and reduced in size.Different phases in mixture get refined and homogenized during dispersive mixing, thus,performance of final products is improved. It’s very important to understand the mechanismand dynamic details of dispersive mixing for designing efficient mixing equipments anddetermining appropriate process parameters. In early theoretical research, based onconsidering particulate agglomerates as continuum, many mathematical models wereproposed, most of which taked change rate in size or volume of the hypothetic continuum asdispersity and considered only dispersing rather than mixing. Lots of simulation cases werecarried out based on these models. However, particulate agglomerate is typical discontinuum.Thories based on continuum cann’t explain influence of hydrodynamic drag forces andinteractions between particles on movements of particles and evolution behavior ofagglomerates. The discrete element method (DEM), originated from geotechnical mechanics,has many advantages in dealing with discrete mechanics. Due to algorithm characteristics ofDEM, it can solve large deformation and nonlinear problems of discontinuum, such asmovements, deformation and failure. In this research, DEM is used to simulate the process ofdispersive mixing.In this paper, basic principle, contact models, contact judgment and parameters selection,algorithm implementation and programming thought of DEM are introduced firstly. On thesebasises, main program of discrete element method for particulate packing and dispersing,namely DEMix2D, is programmed by interface function of MATLAB and C/C++, namelyC-MEX. The platform is Microsoft visual C++. It should be noticed that DEMix2D is onlyfor two-dimensional disk or spherical particles. With programmed pre&post processor ofDEMix2D, physical parameters, boundary and initial conditions and control parametersduring simulation can be inputted conveniently. Result datas can also be visualized.After completing above work, vandewaals force is introduced in DEMix2D to simulate andanalysis the random packing process of particulate clusters in a rigid box. The results arecompared with previous similar work to verify the correctness of DEMix2D. Fially,hydrodynamic drag force is considered in DEMix2D to simulate the dispersive mixingprocess of fine particulate agglomerates in shear, enlongational and combinedshear-enlongational fows. Dynamic details of deformation, movement and failure ofparticulate agglomerates in different flows are achieved. It’s demonstrated that theelongational flow has the best dispersion performance in the three flow types and the higherapllied flow stress, the better dispersion performance can de achieved. It’s found that thelooser particulate agglomerate is more likely to disperse. Conclusion can also be drawn thatdispersity, in terms of weighted averaged fragment size w, follows a power-law typerelation asw kt m.