Identifying dominant spatial and time characteristics of flow dynamics within free-surface baffled stirred-tanks from CFD simulations

In many chemical and biochemical processes, it is fundamental to accurately predict flow dynamics within reactors of different sizes and its influence on reactions and their kinetics. Computational Fluid Dynamics can provide detailed modeling about hydrodynamics. The objective of the present work is to assess the abilities of CFD to simulate free-surface turbulent flow within baffled stirred-tanks reactors. Transient simulations are carried out using a homogeneous Euler-Euler multiphase approach, the Volume-of-Fluid (VOF) method, with a Realizable k-epsilon turbulence model. Two methods are considered to account for the impeller motion, namely the Multiple Reference Frame (MRF) and Sliding Mesh (SM) approaches. Global and local results obtained by CFD are presented by means of statistical analysis, including the estimation of characteristic turbulent length scales. Instantaneous numerical data fields obtained with the SM model are then interpreted using modal decompositions methods, i.e. the Proper Orthogonal Decomposition (POD) and the Dynamic Mode Decomposition (DMD) in order to extract their dominant spatial structures with their time behavior. All simulations are discussed based on comparison with experimental data. (C) 2018 Elsevier Ltd. All rights reserved.