Computational Fluid Dynamics Simulation of Structured Packing

An electronic representation of Sulzer’s Mellapak N250Y structured packing was generated by subjecting a physical packing element to CT scan. The resulting file was imported into a three-dimensional imaging program and then copied, translated, and rotated to create a stack of three packing elements. This stack was then inserted into an open-ended cylindrical surface to approximate a packed column. The resulting three-dimensional image file was imported into the Star-CCM+ CFD software and meshed using the Surface Remesher and Polygonal Volume Meshing routines to create a high-quality volume mesh. The mesh was then utilized to produce a series of CFD simulations. Flow rates were chosen so that the velocities studied would match those commonly employed in the vapor phase of industrial distillation columns. The pressure drop across the packing was monitored during iterative computations, and its rate of change was used to judge convergence. The simulation predictions were shown to be in good agreement with experiments that measured pressure drop in an analogous geometry and flow range. ■ MOTIVATION Distillation is the most common unit operation employed to achieve the necessary chemical separations in the refining and chemical industry. This fact is borne out by the over 40 000 distillation columns in the United States, which require more than 5.06 × 10 J (4.8 × 10 BTU) annually. That translates to over 40% of the energy used in refining and chemical processing industries. Because of the extensive use of distillation columns and their long lifespan, even modest improvements in design or operation will result in significant cost savings. Most distillation columns are heated by steam produced by fossil fuels. Reducing the consumption of these expensive fuels will yield immediate cost savings and will have the added benefit of limiting CO2 emissions. Optimization of distillation performance has long been constrained by a lack of comprehensive and robust models of fluid flow. The complexity of the two-phase flow field coupled with a dearth of methods to directly observe it have precluded a thorough understanding of the flow phenomena and the development of design and operation principles rooted in fluid flow theory. Current distillation design and operation models employ a handful of correlative relationships. The combination of high-performance computing clusters and computational fluid dynamics (CFD) simulation software is employed in the aerospace, automotive, and marine industries to enhance performance, decrease design costs, and accelerate the adoption of new designs. These tools promise similar improvements in distillation design and performance. This work demonstrates a means by which complicated flow systems, specifically whole elements of structured packings, can be quickly and accurately simulated. Previously, the complicated natures and large simulation volumes required to model structured packing have prevented this method of study. This method does not require simplifying assumptions, and the accuracy of the geometry can easily be numerically quantified and studied as an independent parameter on simulation accuracy. A high-fidelity computational geometry was generated by importing the results of an X-ray computed tomography (CT) scan of a common structured packing. Grid generation was performed automatically using adaptive mesh generation software with userspecified minimum and maximum grid spacings, as well as cell growth rate. Using this method, a high-fidelity, unstructured CFD mesh can be generated and ready for simulation in hours