Effect of software implementation on the result of computational fluid dynamics simulation of circulating fluidized bed risers

While in scientific literature much focus is directed toward model validation,
comparison, and even parametric investigations on individual model parameters, the possible effects of the used computational fluid dynamics (CFD)
software on the results are largely neglected. In this article, CFD modeling of
circulating fluidized beds (CFB) are performed with Ansys Fluent and OpenFOAM to investigate the effect of software implementation on the simulation
results. Transient Eulerian–Eulerian simulations are performed of two different laboratory-scale CFB cold models in turbulent and circulating fluidized bed
conditions. The same mesh and as identical models and settings as possible are
utilized on both software. The obtained time-average profiles of pressure, velocity, and solids volume fraction are compared between the software and with
available measurements. A difference in the granular energy equation was identified between the software, and a modification was made to achieve the same
formulation. The effect of boundary conditions was also investigated. It was
found that depending on the case, the software could have a notable effect on
the results. These differences are found especially in particle distribution, visible
in the vertical pressure and solids volume fraction profiles as well as in external
circulation mass flow rates.