Santiago Márquez Damián

Residence Time Distribution Determination of a Continuous Stirred Tank Reactor using Computational Fluid Dynamics and its Application on the Mathematical Modeling of Styrene Polymerization

Abstract The continuous operation of a stirred tank reactor for styrene polymerization was modeled. The proposed approach consists of an iterative procedure between two modules that considers the fluid-dynamics and kinetics respectively. The kinetic module considers a complex kinetic mechanism and is used to predict the time evolution of global variables, such as conversion and species concentrations, physicochemical properties and molecular structure characteristics of the final product. In order to obtain a 3D representation of the flow field, the simulation of the hydrodynamics of the reactor was carried out with the aid of a commercial computational fluid dynamics (CFD) software package. Because CFD is capable to predict the complete velocity distribution in a tank, it provided a good alternative to carry out residence time distribution (RTD) studies. It was found that the stimulus-response tracer method is reasonably accurate to obtain a complete RTD compared to the particle tracking method. The obtained RTD results showed a good agreement when validated with experimental data and literature information.From the estimates of the kinetic module and the RTD predictions, a statistical calculus allows the determination of the average properties at the reactor outlet. The convergence of the iterative procedure was tested and reasonable predictions were achieved for an industrial reactor.

gdbOF: A debugging tool for OpenFOAM®

OpenFOAM(R) libraries are a great contribution to CFD community and a powerful way to create solvers and other tools. Nevertheless in this creative process a deep knowledge is needed concerning with classes structure, for value storage in geometric fields and also for matrices resulting from equation systems, becoming a hard task for debugging. To help in this process a new tool, called gdbOF, attachable to gdb (GNU debugger) is presented in this paper. It allows to analyze classes structure at debugging time. This application is implemented by gdb macros, these macros can access to code classes and also to their data in a transparent way, giving the requested information. This tool is tested for different application cases, such as the assemble and storage of matrices in a scalar advective-diffusive problem, non orthogonal correction methods in purely diffusive tests and multiphase solvers based on Volume of Fluid Method. In these tests several types of data are checked, such as: internal and boundary vector and scalar values from solution fields, fluxes in cell faces, boundary patches and boundary conditions. As additional features of this tool data dumping to file and a graphical monitoring of fields are presented. All these capabilities give to gdbOF a wide range of use not only in academic tests but also in real problems.

Flow Study and Wetting Efficiency of a Perforated-Plate Tray Distributor in a Trickle Bed Reactor

In this work, a computational fluid dynamics analysis (CFD) employing the Eulerian two-fluid model was carried out with the aim to understand the distribution process and to determine the wetting efficiency of the primary tray distributor (perforated plate) of a trickle bed reactor (TBR) under several operating conditions. The overall inlet geometry was considered, and the small holes of the perforated plate were modeled by sinks (drains) and sources, employing CFD and experimental models to obtain the hole discharge flow coefficients. The influence of the ceramic-ball bed above the catalyst bed was considered by a suitable correlation to estimate liquid distribution inside it.Results showed that because of the scarce liquid sloshing above the tray, little difference on liquid flow rate through the tray holes was found. Due to the really low inlet mass flow rate of gas, it has negligible influence on liquid behavior, which drops through holes slowly without spraying. Thus, the ceramic-ball bed above the catalyst bed is exclusively wetted in a small area under the tray holes. Although the ceramic-ball bed improves liquid distribution, which guarantees a minimum liquid volume fraction at all places, significant differences on the liquid mass flow rate across the top of the catalyst bed were found. Additional causes of low efficiency in TBR like the well-known fouling vulnerability of perforated-plate trays and unevenness were analyzed. For the first, two simple modifications were proposed to improve tray performance: reducing the amount of gas chimneys to only one and adding additional drip points and replacing the tray holes by short risers in order to avoid plugging.

An oscillation-free flow solver based on flux reconstruction

Abstract In this paper, a segregated algorithm is proposed to suppress high-frequency oscillations in the velocity field for incompressible flows. In this context, a new velocity formula based on a reconstruction of face fluxes is defined eliminating high-frequency errors. In analogy to the Rhie–Chow interpolation, this approach is equivalent to including a flux-based pressure gradient with a velocity diffusion in the momentum equation. In order to guarantee second-order accuracy of the numerical solver, a set of conditions are defined for the reconstruction operator. To arrive at the final formulation, an outlook over the state of the art regarding velocity reconstruction procedures is presented comparing them through an error analysis. A new operator is then obtained by means of a flux difference minimization satisfying the required spatial accuracy. The accuracy of the new algorithm is analyzed by performing mesh convergence studies for unsteady Navier–Stokes problems with analytical solutions. The stabilization properties of the solver are then tested in a problem where spurious numerical oscillations arise for the velocity field. The results show a remarkable performance of the proposed technique eliminating high-frequency errors without losing accuracy.

Assessment of gas-particle flow models for pseudo-2D fluidized bed applications

ABSTRACT The aim of this work is to provide more insight into the general modeling criteria for simulating pseudo-2D bubbling fluidized beds. For this purpose, two experimental-based problems are studied. First, a fluidized bed with a high-speed central jet problem is analyzed. A qualitative study of the first bubble indicates that the bubble shape prediction is highly sensitive to the frictional model adopted. The most accurate results in terms of bubble shape and detachment time are given by a frictional model that relates the strain-rate fluctuations with the granular temperature. Second, a uniformly fluidized bed problem in bubbling regime is considered. For this case, the drag models and boundary conditions for the particulate phase are investigated. Time-averaged solid phase velocity profiles are compared with the results of the literature where it is found that no-slip conditions (or partial slip with a high specularity coefficient) are more appropriate than slip conditions at the walls for these regimes. Regarding the drag force, although none of the models presented could match the experimental velocity predictions for low gas velocities at the lower region of the bed, the Di Felice model produces the most accurate results for the whole range of regimes considered.