(585x) Accelerated Process Innovation through Hybrid Computational Modeling

Process innovation in the chemical, food, pharma and consumer good industries has become increasingly dependent on modeling. High fidelity computational models are being regularly used to evaluate novel process concepts and designs. Such models though suffer from two key drawbacks: first, they tend to be computationally very intensive and second, the numerical techniques involved in their solution tend to be complex and require significant domain expertise. Taken together, these two challenges have hampered the widespread adoption of advanced computational models by the larger engineering community.

With desktop computers becoming increasingly powerful of late, the first challenge has been met with some degree of success. The second, however, is a more fundamental challenge. Process development engineers are not always neccessarily well versed in advanced numerical computing. Nevertheless, they are very well versed in building process/engineering models that are based on engineering fundamentals as applied to their specific processes. Thus the second challenge noted above can be overcome by bridging the gap between the two paradigms - computational modeling and process/engineering modeling.

Hybrid computational models which combine the best aspects of computational modeling paradigms (CFD, DEM, LBM etc.) with traditional process/engineering modeling ideas present a viable means to bridge the gap mentioned above. CFD- and/or DEM-based "compartmental" or zonal models are a good example of this class of models. These models use flow information from CFD, DEM models of process equipment to partition the equipment into "zones" (each zone is uniform with respect to some physical quantity characterizing turbulence, flow, mixing etc.). The equipment is represented as a network of zones with each zone exchanging flow information with its neighbours. Each zone can now be modeled as an idealized flow vessel to perform calculations pertinent to the multiphysics of interest (reactions, heat and mass transfer, phase transfer etc.).

In this talk, we show how such zonal models can be used to describe multiscale, multiphysics engineering problems of different levels of complexity. We also describe ZoneIT, a software package that streamlines the hybrid zonal modeling workflow described above. Using ZoneIT, engineers can easily and readily implement hybrid zonal modeling to analyze their processes. We present case studies where ZoneIT has been successfully employed to analyze processes in the food, pharma and chemicals industries.