- Title
- A continuous, dynamic and steady state segregation-dispersion model of the reflux classifier
- Creator
- Naveed ul Hasan, Syed
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2017
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- The aim of this study was to develop a generalised 2D segregation-dispersion model of the Reflux Classifier applicable to continuous processing, for describing the internal state of the system, and for describing dynamic, and steady state separations. The Reflux Classifier is a beneficiation device comprising a fluidisation section with a set of parallel inclined channels above, in which particle segregation takes place according to differences in density and/or size. The presence of inclined channels provides a large effective settling area, while also improving the separation performance of the device. Many experimental studies have been conducted using the Reflux Classifier. However, the present study was the first to establish a mathematical model of the Reflux Classifier, for describing separations under continuous conditions. In the present study, a 2D model of the Reflux Classifier was developed, formulated in terms of the segregation and dispersion of the particles. This new model, written in Fortran 77, involved two main steps. The first was concerned with implementing the Kennedy and Bretton (1966) 1D segregation-dispersion model, previously used to describe liquid fluidised beds under batch conditions. The predictions of the 1D model for monocomponent systems in liquid fluidised beds were validated using flux curve analysis. The second step involved the modification of the 1D segregation-dispersion model into a 2D model of the Reflux Classifier. According to the segregation-dispersion model (Kennedy and Bretton, 1966), the net flux of any given particle species relative to the vessel consists of two components, namely the segregation and dispersion fluxes. The segregation flux in a monocomponent system was described using the hindered settling model of Richardson and Zaki (1954), modified to obtain the particle velocity relative to the fluid. For systems comprising particle species of different sizes and densities, the segregation flux was evaluated using a modified form of the Richardson and Zaki model, proposed by Asif (1997). The hindered settling model of Asif (1997) is effectively a constitutive model that incorporates both the effects of particle size and density, including the local suspension density on the slip velocity of a given particle species (Asif, 1997; Galvin et al. 1999a, b). The dispersion flux in the fluidisation section was determined using fixed values of the dispersion coefficient in the x (horizontal) and y (vertical) directions, as fitting parameters. Similarly, within the inclined section, a fixed value of the dispersion coefficient was adopted in the tangential direction (y direction). In contrast, in the normal direction (x direction) within the inclined section, a laminar shear mechanism, i.e. shear-induced lift, was incorporated as a contribution to dispersion in the x direction. The 2D segregation-dispersion model of the Reflux Classifier has been utilised to describe the transport behaviour of single, binary and multicomponent species. In particular, the 2D model was used to investigate the separation of a coal feed. The feed was composed of 35 particle species having 7 different densities and 5 different sizes. The work was initially undertaken in the absence of shear-induced lift, circumstances that arose when the Reflux Classifier was operated at full scale with wide channels (Galvin et al., 2005). The values of the separation relative density, D50, and Ecart Probable Error, Ep, obtained from the model predictions accurately matched the experimental results for particles whose sizes were in the range 0.25-2.0 mm. Additional studies were performed using a feed formed from 42 particles species (coal feed) having 7 different densities and 6 different sizes. These studies included the modelling of the shear-induced lift, previously observed experimentally in the closely spaced inclined channels of the Reflux Classifier (Galvin et al., 2010c). The corresponding partition curves for the particle size range 0.25-2.0 mm and the D50 and Ep values were validated using the reported experimental data sets. In order to explore the generality of the model, a comparative study of a low-density feed (coal) and a high-density feed (iron ore) was also conducted. The study showed that at lower fluidisation velocities the normalised overall Ep* values for both cases were low and consistent with a good separation performance. The segregation-dispersion model of the Reflux Classifier for describing a continuous process provided a realistic model of the system. The model permitted the feed, fluidisation velocity and underflow flux to be specified. The model then described the internal suspension density, total and individual solid volume fractions, concentration changes with time, and provided the partition curves for the individual size fractions, and hence the variation of the D50 and Ep with particle size. From this information, it was also possible to predict the fractional and cumulative product ash values. Overall, the model of the Reflux Classifier provided the essential physics of this complex system, based on a simple hindered settling model. Moreover, the 2D segregation-dispersion model of the Reflux Classifier provided a framework in which alternative hindered settling models, dispersion models and descriptions of the shear-induced lift could be inserted.
- Subject
- segregation-dispersion model; Reflux Classifier; continuous processing; steady state separations
- Identifier
- http://hdl.handle.net/1959.13/1353533
- Identifier
- uon:31113
- Rights
- Copyright 2017 Syed Naveed ul Hasan
- Language
- eng
- Full Text
- Hits: 1250
- Visitors: 1620
- Downloads: 511
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 2 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 352 KB | Adobe Acrobat PDF | View Details Download |