A hierarchical approach to chemical reactor engineering: an application to micro packed bed reactors

Abstract

Hierarchical modeling is applied for the investigation of micro packed bed reactors. This method allows for the use of Computational Fluid Dynamics (CFD) simulations in the analysis of representative complex geometries, where a full-scale CFD simulation of the entire reactor is not possible. Detailed and computationally demanding analyses are used to study a selected number of conditions and phenomena. Then, lumped parameters are derived from CFD results by means of engineering correlations. These parameters are incorporated in simplified reactor models based on macroscopic conservation equations. We provide evidence for the potential of the approach by using as a show-case a micro packed bed reactor in the context of highly exothermic selective oxidation processes. This reactor configuration consists of catalytic particles packed in the channels of a honeycomb matrix, which is expected to strongly enhance the radial heat transfer. In particular, we first focus on the analysis of energy transfer mechanisms by CFD and their interpretation via a 1D model and we provide an assessment of existing correlations with respect to the unconventional configuration (2 mm channel equivalent diameter and 0.8 mm sphere diameter). These correlations are then implemented in a pseudo-continuous (i.e. macroscopic) 2D model to allow for a systematic investigation of the capabilities of the micro packed bed reactor in dealing with the selective oxidation of o-xylene to phthalic anhydride. We found that due to the enhanced radial heat transfer micro packed bed reactors allow for quasi-isothermal operations, thus extending the range of operating conditions possible without occurring in adverse thermal behavior of the reactor. On a more general basis, we prove that the hierarchical approach to chemical reactor engineering is an effective tool to bring the application of fundamental modeling at a level of complexity relevant to full-scale applications, otherwise not possible because of the impractical computational costs.