Nombre: LUCAS HENRIQUE PAGOTO DEOCLECIO
Tipo: MSc dissertation
Fecha de publicación: 23/08/2018
Supervisor:
Nombre |
Rol |
|---|---|
| ANA PAULA MENEGUELO | Advisor * |
Junta de examinadores:
| Nombre |
Rol |
|---|---|
| ANA PAULA MENEGUELO | Advisor * |
| DANIEL DA CUNHA RIBEIRO | Co advisor * |
| KAROLLINE ROPELATO | External Examiner * |
| MARCELO SILVEIRA BACELOS | Internal Examiner * |
| OLDRICH JOEL ROMERO | Internal Alternate * |
Resumen: Gravitational separators of dispersed mixtures of two liquids are present in many industry areas. However, the separation dynamics prediction of these equipment is complex, since it involves the concomitant and of mutual influence (by means of the droplet size distribution (DSD)) mechanisms of creaming (or sedimentation) and coalescence. Therefore, it is important to know the phenomena and the models available for their representation before the model implementation. The objective of this dissertation is to model, through Computational Fluid Dynamics (CFD), the laminar creaming zone of a batch gravitational separation in the Stokes regime. For this, the numerical results (using Ansys Fluent® 15.0 and 18.2) were compared with the experimental results of Jeelani, Hosig and Windhab (2005). To model the influence of the dispersed phase volume fraction on the creaming rate, the Schiller and Naumann drag model (1935) was used with the Richardson and Zaki (1954) drag modifier. The initial DSD was discretized with the Gauss-Legendre Quadrature and the new droplets size born from the coalescence process was estimated using the Wang and Davis (1996) model. Both the DSD representation and the drag models were suitable to model the separation process; however, the determination of the creaming zone limits was shown to be a relevant factor, especially in relation to the determination of the coalescence interface, which seems to be a function of the droplets packaging factor. The coalescence models evaluated were constant coalescing efficiency, critical approach velocity and film drainage time of a deformable drop with partially mobile interface. The three models presented similar results in the modeled flow regime; however, the film drainage model was used with its standard coefficient and did not need to be calibrated. This is because this model is more complete because, besides the relative velocity between the droplets it takes into account the phases viscosities and the surface tension, and the droplets diameter. The coalescence models were calibrated for an experimental condition and then employed with the same fitting parameter for other experiments with different DSD and volume fraction, suggesting the models fitting parameters are function of the liquids properties only. The maximum error between the experimental and numerical creaming and coalescence positions was 5.03%. In addition, the phases volume fraction profile was also estimated with reasonable accuracy by the numerical models.
