Data-driven model for the breakage of dry monodisperse agglomerates by wall impact applicable for multiphase flow simulations

The study is concerned with the development of a model for predicting the particle size distribution resulting from the impact of nearly spherical cohesive agglomerates at a wall under a variety of impact conditions. For the derivation of the model an extensive number of discrete element simulations is carried out. The matrix of impact conditions consists of eight agglomerate size classes between 10 and 104 primary particles and seven impact angles between a flat (7°) and a normal impact (90°). To include the effect of the van-der-Waals bond strength, three primary particle diameters in the micrometer range are considered. Furthermore, the range of velocities allows to reproduce the full breakage spectrum from a rebound of an intact agglomerate to a full disintegration. The model is intended for the application in Euler-Lagrange simulations based on the hard-sphere approach ignoring the time-consuming resolution and tracking of the detailed agglomerate structure. This allows the prediction of turbulent flows with high mass loadings. For the model derivation the impact outcome is quantified by measures enabling the determination of the particle size distribution, i.e., the overall number of resulting fragments (fragmentation ratio) and the number of primary particles incorporated in the three largest fragments. The results concerning the effect of different agglomerate and impact parameters are consistent with the reported literature. Relying on this broad data base a new physically meaningful dimensionless number πimp is derived based on an empirical approach to characterize the breakage phenomenon. The advantage of the proposed dimensionless number reasonably unifying the results obtained under widely varying impact conditions is demonstrated. In addition, the disagreement detected for some of the cases can be attributed to the re-agglomeration of detached fragments due to the van-der-Waals force. Since agglomeration is separately accounted for in the multiphase simulation approach targeted by the breakage model, the data influenced by high re-agglomeration rates are not considered in the derivation of the data-driven model.