On heat transfer modelling in large-scale solid-fluid flow systems

  • Imatani, Toshiki (The University of Tokyo)
  • Yunoki, Keita (Kubota Corporation)
  • Shiraiwa, Yuki (Kubota Corporation)
  • Sakai, Mikio (The University of Tokyo)

Please login to view abstract download link

The discrete element method (DEM) coupled with computational fluid dynamics (CFD) is widely applied in numerical simulations to improve the understanding of solid–fluid flow systems, which frequently appear in various engineering fields. However, several issues arise when the DEM is applied to heat transfer modeling. Although thermal conduction within the solid phase depends on particle stiffness, DEM simulations often employ spring constant values much smaller than the real material stiffness. Consequently, complex models that account for actual contact durations and areas are required. Moreover, existing models typically consider only short-range radiative heat transfer within an empirically determined cutoff distance to maintain computational feasibility, while long-range radiation is neglected. Another limitation is that existing heat transfer models show low compatibility with scaling law models such as the coarse-grained DEM. When large coarse grain ratios are used, the accuracy of heat transfer prediction decreases significantly. To overcome these limitations, we develop a new heat transfer model for the DEM that is formulated within an Eulerian framework. In the proposed model, particle motion is resolved by the DEM, whereas heat transfer is represented in the Eulerian framework. A distinguishing feature of the proposed model is that it depends only on the macroscopic solid volume fraction and does not rely on particle contact states in the DEM. In addition, the Eulerian framework allows consistent and simple treatment of both short-range and long-range radiative heat transfer. The validity of the proposed heat transfer model is demonstrated through comparisons between the simulation results and the experiments. Furthermore, we show that the proposed model is independent of the spring constant used in the DEM and is highly compatible with the coarse-grained DEM.