This thesis presents the development and application of a dynamic model of gold adsorption onto activated carbon in gold processing. The primary aim of the model is to investigate different carbon management strategies of the Carbon in Pulp (CIP) process. This model is based on simple film-diffusion mass transfer and the Freundlich isotherm to describe the equilibrium between the gold in solution and gold adsorbed onto carbon. A major limitation in the development of a dynamic model is the availability of accurate plant data that tracks the dynamic behaviour of the plant. This limitation is overcome by using a pilot scale CIP gold processing plant to obtain such data. All operating parameters of this pilot plant can be manipulated and controlled to a greater degree than that of a full scale plant. This enables a greater amount of operating data to be obtained and utilised.
Two independent experiments were performed to build the model. A series of equilibrium tests were performed to obtain parameter values for the Freundlich isotherm, and results from an experimental run of the CIP pilot plant were used to obtain other model parameter values. The model was then verified via another independent experiment. The results show that for a given set of operating conditions, the simulated predictions were in good agreement with the CIP pilot plant experimental data.
The model was then used to optimise the operations of the pilot plant. The evaluation of the plant optimisation simulations was based on an objective function developed to quantitatively compare different simulated conditions. This objective function was derived from the revenue and costs of the CIP plant. The objective function costings developed for this work were compared with published data and were found to be within the published range. This objective function can be used to evaluate the performance of any CIP plant from a small scale laboratory plant to a full scale gold plant. The model, along with its objective function, was used to investigate different carbon management strategies and to determine the most cost effective approach. A total of 17 different carbon management strategies were investigated. An additional two experimental runs were performed on the CIP pilot plant to verify the simulation model and objective function developed.
Finally an application of the simulation model is discussed. The model was used to generate plant data to develop an operational classification model of the CIP process using machine learning algorithms. This application can then be used as part of an online diagnosis tool.
| Identifer | oai:union.ndltd.org:ADTP/221588 |
| Date | January 2003 |
| Creators | sawan.jonguwa@au.experian.com, Pornsawan Jongpaiboonkit |
| Publisher | Murdoch University |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | http://www.murdoch.edu.au/goto/CopyrightNotice, Copyright Pornsawan Jongpaiboonkit |
Page generated in 0.0024 seconds