(82c) From Particle Resolved CFD to the Transient Modeling of Dynamic Systems: Dispersion in Fixed-Bed Reactors

The modeling of dynamic processes in fixed-bed reactors is of huge interest. Either the overall process can show a transient behavior (e.g. adsorption, desorption or Chemical-Looping) or local and temporal hot spots can be present leading to process safety issues during the start-up phase of otherwise steady-state processes [1]. Particle resolved CFD simulations have been used extensively in the last years (see [2], [3] and [4]) to accurately predict the steady-state conditions of fixed-bed reactors. However, this approach is characterized by a wide range of time scales and a large numerical demand. Hence, it is uncertain if it can be efficiently used for the simulation of the dynamic process behavior in fixed-bed reactors.

On the other hand, the accuracy of simplified models relies on the knowledge of effective parameters like the effective thermal conductivity, the heat transfer coefficient and the dispersion coefficient. Most often these are determined using correlations. In most cases, their application is limited to simple particle shapes and reactors with a large tube-to-particle diameter ratio. To overcome this limitation, a workflow based on particle resolved CFD and particle tracking is proposed to determine axial dispersion coefficients in fixed-bed reactors that can later be used in simplified models to describe the process dynamics.

The effect of the bed morphology, Reynolds number, axial position and tube-to-particle diameter ratio (N) is thoroughly investigated over a wide range of particle shapes. Furthermore, the limitation of the dispersion model for reactors with a very low N is discussed.

References
[1] Adams et al. (2009) Int. J. Hydrogen Energy, 34, 8877-8891
[2] Wehinger et al. (2015) Chem. Eng. Sci., 122, 197-209
[3] Wehinger et al. (2016) AIChE J., 62(12), 4436-4452
[4] Wehinger et al. (2017) Ind. Eng. Chem. Res., 56(1), 87-99