HPC large scale simulation of an industrial fluidized bed and applications to chemical engineering processes with NEPTUNE_CFD

Conference Dates

May 22-27, 2016


Fluidized beds and more broadly dilute and dense particle-laden reactive flows are encountered in a wide range of industrial chemical engineering applications such as catalytic polymerization, coal combustion … Nowadays, it is possible to perform realistic 3D simulations of industrial configurations using an unsteady Eulerian multi-fluid approach for polydisperse reactive flows with a good physical modelling. Hence CFD is a powerful tool for studying the optimization of chemical processes, new designs and scaling-up.

To obtain numerical results in an acceptable CPU time, it is important to check the feasibility of CFD simulation of fluidized bed flows in complex geometries at industrial scale. Also we need to estimate HPC capacities of CFD tools. Numerical simulations have been performed with the solver NEPTUNE_CFD: parallelized unstructured code (MPI) using unsteady Eulerian multi-fluid approach. NEPTUNE_CFD is based on the same numerical methods than Code_Saturne. Code_Saturne is an open source CFD software code ready to run on petascale machines. NEPTUNE_CFD’s high parallel computing performances for particle-laden flows have been demonstrated over last years. Recent developments allow overtaking NEPTUNE_CFD’s limitations making it fit for massive parallel computing.

Please click Additional Files below to see the full abstract.

This document is currently not available here.