92
4 Modelling Dynamically Structured Fluidisation
Besides, model validation progresses in parallel to model development, but
rigorous and systematic validation processes are still subjected to intensive discussion. In general, simple, empirical correlations are considered valid only under a
limited range of conditions, whereas more fundamental models, incorporating the
underlying physics, are expected to be reliable for a broader variety of scenarios. A
considerable number of validation exercises of computational models used relate
to researcher expectations, or sometimes, even subjective interests. Grace and
Taghipour [22] proposed several guidelines for comprehensive CFD code validation to avoid the excessive claims commonly reported in the literature. Some of the
guidelines are more practical in a general situation, while other ones are still inherently subjective. Most importantly, there is no consensus about determining if the
simulated and experimental traits are considered in “reasonable agreement”. Sometimes, agreement could be claimed without reproducing the fundamental physics
correctly. In this context, it is evident that models for multiphase systems require
extra attention when subjected to a validation test. Nevertheless, the community is
still seeking a tool that is sufficiently robust to determine whether a multiphase model
reproduces the physics of gas-solid phases at different scales. This is especially relevant for the validation of the fluid-particle closures that are formulated based on
direct numerical simulations.
In turn, those attempts in reproducing structured flows demonstrate the potential of the phase synchronisation phenomenon, which has been largely ignored by
the community so far. The structured flow is not just a single specific fluidisation
state but arises from the comprehensive coupling of gas-solids dynamics at multiple
spatiotemporal scales [11]. Owing to its unique visual manifestation, the structured
pattern excels as a convenient, yet rigorous tool to facilitate comparisons between
simulated and experimental practices. Such a validation practice can be based merely
on the falsification test, that is whether the model reproduces the experimentally
witnessed phenomenon, and also avoid artefacts induced by indirect measurements
and inappropriate analysis techniques.
The work presents a comprehensive assessment of the performance of EulerianEulerian and Eulerian-Lagrangian approaches in terms of reproducing the structured flow of bubbles in a deep, quasi-2D geometry. The investigation of simulated dynamics reveals the critical role of solid mechanics in pattern formation.
The comparison of the simulated results demonstrates the power of dynamically
structured flows in validating multiphase models.
4.2 Model Implementation
Different modelling packages are used to simulate the quasi-2D pulsed fluidised
beds. TFM simulations are conducted in the Ansys Fluent (version 12) finite-volume
solver, whereas CFD-DEM simulations are run in the CFDEM (version 3.1.0) fourway coupled solver which integrates OpenFOAM finite-volume solver for resolving
the continous phase and LIGGGHTS solver for resolving the discrete phase [21].
4 Modelling Dynamically Structured Fluidisation
Besides, model validation progresses in parallel to model development, but
rigorous and systematic validation processes are still subjected to intensive discussion. In general, simple, empirical correlations are considered valid only under a
limited range of conditions, whereas more fundamental models, incorporating the
underlying physics, are expected to be reliable for a broader variety of scenarios. A
considerable number of validation exercises of computational models used relate
to researcher expectations, or sometimes, even subjective interests. Grace and
Taghipour [22] proposed several guidelines for comprehensive CFD code validation to avoid the excessive claims commonly reported in the literature. Some of the
guidelines are more practical in a general situation, while other ones are still inherently subjective. Most importantly, there is no consensus about determining if the
simulated and experimental traits are considered in “reasonable agreement”. Sometimes, agreement could be claimed without reproducing the fundamental physics
correctly. In this context, it is evident that models for multiphase systems require
extra attention when subjected to a validation test. Nevertheless, the community is
still seeking a tool that is sufficiently robust to determine whether a multiphase model
reproduces the physics of gas-solid phases at different scales. This is especially relevant for the validation of the fluid-particle closures that are formulated based on
direct numerical simulations.
In turn, those attempts in reproducing structured flows demonstrate the potential of the phase synchronisation phenomenon, which has been largely ignored by
the community so far. The structured flow is not just a single specific fluidisation
state but arises from the comprehensive coupling of gas-solids dynamics at multiple
spatiotemporal scales [11]. Owing to its unique visual manifestation, the structured
pattern excels as a convenient, yet rigorous tool to facilitate comparisons between
simulated and experimental practices. Such a validation practice can be based merely
on the falsification test, that is whether the model reproduces the experimentally
witnessed phenomenon, and also avoid artefacts induced by indirect measurements
and inappropriate analysis techniques.
The work presents a comprehensive assessment of the performance of EulerianEulerian and Eulerian-Lagrangian approaches in terms of reproducing the structured flow of bubbles in a deep, quasi-2D geometry. The investigation of simulated dynamics reveals the critical role of solid mechanics in pattern formation.
The comparison of the simulated results demonstrates the power of dynamically
structured flows in validating multiphase models.
4.2 Model Implementation
Different modelling packages are used to simulate the quasi-2D pulsed fluidised
beds. TFM simulations are conducted in the Ansys Fluent (version 12) finite-volume
solver, whereas CFD-DEM simulations are run in the CFDEM (version 3.1.0) fourway coupled solver which integrates OpenFOAM finite-volume solver for resolving
the continous phase and LIGGGHTS solver for resolving the discrete phase [21].