(421c) Application of Multi-Level Coarse-Grained Model for DEM Heat Transfer Simulation

In the evolving landscape of heat transfer simulation, the discrete element method (DEM) has emerged as a pivotal tool for predicting the thermal behaviour of granular materials. Building upon a recent advancement in multi-level coarse-graining (MCG) techniques for DEM simulations, this study introduces novel scaling laws for thermal conductivity and thermal capacity tailored for coarse-grained particles. These scaling laws are integral to the proposed multi-level coarse-grained simulation framework, enabling the efficient simulation of heat transfer processes in systems comprising a large number of particles. By leveraging the particle location-based method for multi-level coarse-graining, we delineate zones of varying resolution within the simulation domain, allowing for dynamic adjustment of particle granularity and preserving overall thermal temperature. This approach not only preserves the computational efficiency afforded by coarse-graining but also enhances the accuracy of thermal simulations by accounting for the heterogeneity in thermal properties across different scales. In our study, we validated our model through a hopper discharge simulation, featuring cold and hot temperature particles on the left and right halves, respectively. We compared the thermal temperature and heat flux distributions across the bed height and among discharged particles between MCG simulations with scaled thermal parameters and fully resolved simulations, finding strong agreement. Additionally, we examined MCG simulations without scaled thermal parameters against resolved simulations, highlighting the necessity of accurate thermal parameter scaling for reliable simulation outcomes. This comparison highlights the significance of proper parameter scaling in achieving accurate simulation outcomes. This study paves the way for more accurate and computationally feasible simulations of industrial-scale heat transfer applications.

*This research is partially financially supported by ICSC—Centro Nazionale di Ricerca in High Performance Computing, Big Data and Quantum Computing, funded by European Union -NextGenerationEU.