Al-Rashed, MHJ;
(1998)
A study of reactive precipitation processes using computational fluid dynamics.
Doctoral thesis , University of London.
Preview |
Text
Al-Rashed_286284.pdf Download (33MB) | Preview |
Abstract
The objective of this research is to study and model a reactive precipitation process and determine the feasibility of applying the Computational Fluid Dynamics (CFD) technique. The CFX package of AEA Technology is adopted in this procedure. CFD is employed to model the batch gas-liquid reactive precipitation process of calcium carbonate in a 1.357 litre stirred vessel. Two major steps are involved in the modelling scheme, viz. the hydrodynamic model and the chemical kinetics and precipitation process model. Then, a link between these two models is created. The hydrodynamic model considers the impeller as a momentum source. To predict the turbulence behaviour, the Reynolds differential stress model is implemented because it is more appropriate than the standard k-E model for swirling flows in stirred tanks. The flow pattern is predicted for two-blade paddle impeller and details of the fluid speed, pressure and turbulent energy are produced. The same procedure is followed in modelling a 45° pitched-blade impeller with six blades, to show the effect of impeller type on the mixing process. Both 2-D and 3-D models are evaluated. Experimental validation is carried out for the CFD hydrodynamic model using three methods: video recording, still photographs and image analysis of 1 mm diameter beads suspended in the vessel using a white light source. Qualitative measurements of the flow pattern give similar results as the predictions of the 3-D CFD model. In addition, flow number calculations from both the 2-D and 3-D CFD simulations show good agreement with literature data. In order to minimise the computation time, the crystallization kinetics and precipitation process simulation is developed for a 2-D problem. The model calculates the extent of reaction, supersaturation, primary nucleation, size-independent crystal growth and the first four moments of the crystal size distribution throughout the vessel. Subsequently, the crystal average-mean size is evaluated. The model predicts an initial thin layer of particles forming in the interfacial region. The crystal slurry then starts to disperse into the bulk of the vessel following the flow pattern of the liquid. This behaviour is qualitatively comparable to the experimental results reported by Wachi and Jones (1991b). The CFD model predictions are compared with those of the film and penetration theories for the same time range. The film theory shows a minor change in the mean size which indicates that it is a nucleation dominated process, i.e. high local supersaturation. Penetration theory predicts higher mean sizes, i.e. a growth dominated process with lower local supersaturation. The CFD simulation on the other hand, adopts an intermediate behaviour highlighting the role of the hydrodynamics in the process. Thus, the model illustrates the feasibility of CFD to tackle the transient reactive precipitation process sufficiently. The main problem currently is excessive computational time.
| Type: | Thesis (Doctoral) |
|---|---|
| Title: | A study of reactive precipitation processes using computational fluid dynamics |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Thesis digitised by EThOS. |
| UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Chemical Engineering |
| URI: | https://discovery.ucl.ac.uk/id/eprint/1546619 |
Archive Staff Only
 |
View Item |
