Author: Caesar Wiratama ([email protected])
Abstract. Sieve trays act as vital gas–liquid contact devices in distillation columns. In this study, a three-dimensional, two-phase Computational Fluid Dynamics (CFD) model employing the Volume of Fluid (VOF) method was developed tosimulate a distillation column containing two sieve trays, explicitly accounting for vapor–liquid phase change. Keyhydrodynamic parameters—such as velocity profiles, liquid holdup, and froth height—were analyzed and compared withexperimental observations. The simulation results demonstrated strong agreement with experimental data, validating theaccuracy of the developed model.
INTRODUCTION
Distillation columns fitted with sieve trays represent one of the most commonly utilized separation techniques in thechemical and petroleum industries, thanks to their high efficiency and versatility across a wide range of mixture types[1–3]. The trays facilitate efficient gas–liquid interaction, promoting the mass transfer required to separate components with similar boiling points. Given their extensive industrial application, gaining a deeper understanding of sieve tray hydrodynamics is essential for enhancing column design and ensuring reliable operation. The hydrodynamic behaviour in sieve tray columns is highly complex and can manifest in different flow regimes, such as froth or spray [4]. The froth regime, characterized by intense bubble–liquid interactions, generally enhances mass transfer but also increases pressure drop, while the spray regime is associated with dispersed droplets and weaker liquid mixing. These regimes strongly influence efficiency, energy consumption, and overall column performance, making their accurate characterization an essential part of process design.Computational Fluid Dynamics (CFD) has been extensively employed to investigate sieve tray hydrodynamics [4,5].Unlike empirical correlations or simplified models, CFD allows for a detailed description of multiphase flow phenomena and provides insight into variables that are otherwise difficult to measure experimentally. By capturing fluid velocities, liquid levels, and vapor–liquid distribution, CFD models offer a more complete understanding of tray performance, particularly under varying operational conditions.One of the major advantages of CFD is its ability to incorporate the effects of complex geometrical features, such as downcomers, perforations, and tray spacing, which significantly impact flow behaviour [6]. However, simplified modelling approaches, including symmetry assumptions, often fail to predict hydrodynamic parameters with sufficient accuracy [4]. This limitation stems from the inherent asymmetry of liquid circulation and vapor distribution within the tray, which cannot be fully captured when boundary simplifications are imposed.
METHODOLOGY
The numerical simulation in this work was validated against the experimental measurements of Solari and Bell [7],which provide widely recognized benchmark data for sieve tray hydrodynamics. Validation against such experimental studies ensures that the numerical predictions capture realistic vapor–liquid interaction phenomena within the distillation column. The simulations in this study were carried out using Cradle CFD – scFlow, a commercial CFD software part of MSCSoftware, Hexagon design and engineering product. The solver was implemented in a transient framework to capture the time-dependent behaviour of vapor–liquid interactions. It incorporated multiphase modelling with explicit phase-change capabilities, enabling the representation of condensation and evaporation processes. This dedicated setup allowed accurate reproduction of the complex hydrodynamics within sieve trays, making it well-suited for the objectives of the present work.In most previous CFD studies of sieve trays, hydrodynamics has been modelled using Eulerian multiphase approaches in combination with assumed drag coefficients. While these models can approximate overall tray behaviour, they often lack the resolution needed to represent local bubble dynamics accurately.In the present study, the Volume of Fluid (VOF) method was employed to explicitly capture the gas–liquid interface and to account for detailed bubble. This approach allows for improved prediction of froth development, liquid circulation, and liquid height compared to drag-based models.Equation (1) shows the Volume of Fluid model [8].
Where 𝑚̇pq is the mass transfer from the q phase to the p phase, and 𝑚̇qp is the mass transfer from the p phase to the q phase and 𝛼% is the volume fraction of the q phase.The operating conditions were selected to reproduce the experimental setup of Solari and Bell [7]. A superficial liquid velocity factor of FS = 1.015 (m/s)(kg/m3)0.5, and a liquid volumetric flow rate of QL = 0.0178 m3/s were imposed.These parameters correspond to an inlet liquid velocity of UL,in = 0.1914 m/s, and an inlet vapor velocity of UG,in =17.87 m/s..Please note that this detailed modelling consumes more computational effort compared to full Eulerian or mixture model.The column geometry and relevant tray parameters are summarized in TABLE 1, while FIGURE 1 presents the computational domain along with the applied boundary conditions, including vapor inlet, liquid inlet, vapor outlet, liquid outlet, and wall constraints. The detailed boundary conditions are provided in TABLE 2,
For spatial discretization, the computational mesh was generated using the octree globally, and polyhedral locally. Abody-fitted mesh was constructed to resolve the tray perforations, local refinement around holes used to enhance flowdetail around them. The mesh is illustrated in FIGURE 2.
RESULTS AND DISCUSSION
From the simulation results, the hydrodynamic behaviour of the sieve tray column was evaluated in terms of velocity distribution and liquid volume fraction. These parameters provided detailed insight into vapor–liquid interactions and the influence of phase change phenomena on tray performance. The velocity fields shown in FIGURE 3 revealed the development of liquid circulation zones across the tray, while the vapor flow through the perforations generated local turbulence that contributed to bubble formation and froth development. Please note that this simulation is transient with highly random flow on top of the tray, the velocity contour is an instantaneous plot.
The liquid volume fraction contours shown in FIGURE 4 confirmed the formation of froth layers above the tray deck, which varied depending on vapor loading. In particular, regions of higher vapor penetration correlated with increased froth height and enhanced mixing, consistent with expected sieve tray operation. Temperature distributions further demonstrated the presence of condensation and evaporation phenomena within the froth region, illustrating the importance of considering phase change effects in predicting tray hydrodynamics.
Comparison with the experimental data of Solari and Bell [7] showed strong agreement in terms of liquid height, froth height, and general flow patterns. Overall, the simulation results validated the effectiveness of the proposed modelling approach in capturing the hydrodynamic characteristics. TABLE 4 shows the CFD and experimental result, since the free surface is not well defined because of the variation of the liquid volume fraction distribution, transient motion of the froth, and non-linear height along the liquid flow direction, the calculated height is the averaged value.
The CFD simulation shows a good agreement with experimental result
CONCLUSION
The CFD simulations showed strong consistency with the experimental findings of Solari and Bell [7]. By employing the Volume of Fluid (VOF) approach, the model successfully captured essential hydrodynamic behaviors, including liquid circulation patterns, froth height, and velocity distribution. These results demonstrate the capability of advancedCFD modeling to deliver valuable insights into sieve tray hydrodynamics and to assist in the design and optimization of industrial-scale distillation columns.
ACKNOWLEDGEMENT
The authors would like to express their gratitude to PT Tensor Karya Nusantara for providing the computational resources, including both hardware and software licenses, as well as the data supplied for the completion of this project. This project represents one of PT Tensor Karya Nusantara’s initiatives to provide broader access for engineers in both industry and academia to numerical simulation resources and literature.
REFERENCES
1. H. Z. Kister. Distillation Design, McGrad-Hill Inc. 1992
2. M.J. Lockett. Distillation Tray Fundamentals, Cambridge University, Cambridge. 1986
3. J.G. Stichlmair, and J.R. Fair. Distillation Principles and Practice, Wiley-VCH, New York. 1998
4. G.K. Gesite, K. Nandakumar, K.T. Chuang. CFD Modeling of Flow Patterns and Hydraulics of Commercial-scale Sieve Trays. AIChe.J., Vol. 49, pp. 910-924. 2003
5. X.G.Li, D.X. Liu, S.M. Xu, H. Li.CFD Simulation of Hydrodynamics of Valve Tray. Chem. Eng. Proc.: ProcessIntensification Vol. 48, pp. 145–151. 2009
6. R. Krishna, J.M. Van Baten, J. Ellenberger, A.P. Hingler, R. Taylor. CFD Simulations of Sieve TrayHydrodynamics. Chem. Eng. Res. Des., Trans. Inst. Chem. Eng., Vol.77, pp. 639–646. 1999
7. R.B.Solari, R.L. Bell. Fluid Flow Patterns and Velocity Distribution on Commercial Scale Sieve Trays., AIChEJ., Vol. 32, pp. 640. 1986
8. C. Wiratama. Theory and Best Practices of Computational Fluid Dynamics, Sleman. 2025