Saxe-Coburg Publications Logo
Saxe-Coburg Publications
Computational Technology Publications
TRENDS IN ENGINEERING COMPUTATIONAL TECHNOLOGY
Edited by: M. Papadrakakis and B.H.V. Topping
Chapter 8

Computational Modelling of Reactive Porous Media in Hydrometallurgy

C.R. Bennett1, D. McBride1, M. Cross1, T.N. Croft1 and J.E. Gebhardt2
1School of Engineering, Swansea University, United Kingdom
2Process Engineering Resources Inc, Salt Lake City UT, United States of America

Keywords: heap leaching, CFD modelling, variably saturated flow, reactive porous media, process modelling, model validation.

In the race to recover the world's base metal resources as efficiently as possible, an increasingly popular method is based on the hydrometallurgical process of heap or stockpile leaching [1]. This process essentially involves the mining of ore through typical open pit mining methods. Mined ore, in the form of piles of run-of-mine ore or sometimes crushed and agglomerated ore, is placed on a "heap" or leach pad that is sprayed or exposed to a chemical solution. The solution dissolves the valuable metals as it percolates through the heap and is collected at the bottom by a geo-membrane lining placed under the heap prior to the construction of the heap pad. The metal-containing or pregnant solution is then processed to concentrate and remove the metal, and finally, to produce a metal product that might go to a refinery. For more details on the heap process, Bartlett [1] provides a good overview on the various aspects of solution mining. These processes mine millions of tonnes of ore to recover metals which are measured in less than 1% of the mined volumes and in the case of gold, the output is measured in parts per million!

The process essentially involves creating the conditions within the heap, so that the liquid solution (containing reactants) is able to contact the ore and enable the leach reactions to occur. The leaching process is extremely complex, especially for sulphide ores, and is a fine balance of a range of physical and chemical interactions involving microbial populations as catalysts. Hence, the leaching process might involve some or all of the following:

  1. liquid solution flow down through the porous heap - this flow may well vary from unsaturated to fully saturated flow and this physics has to be captured carefully
  2. air may be injected through the base and will essentially flow counter-current to the solution, providing an oxygen source to the enable the reaction process
  3. the ore may well contain pyrite, which is exothermic as it reacts and affects the heat balance of the system, and
  4. microbial populations which reside on the surface of the solid complex and catalyse the leaching reactions.

The overall heap leach process typically works by the building of heaps in layers that are 8-15 m deep and the laying of solution application lines along the surface. The optimisation of these processes depends upon the chemical and physical characteristics of the ore, and although heap leaching is a fairly tolerant process, its optimisation over time is really not straightforward. Ironically, it is the recent requirements of accountants to enable a more discriminating and accurate assessment of the metal values both within the ore matrix and in solution within the heap that have provided a motivation to develop and use advanced 'physics' based models as the basis for tracking assets.

Given both the technical and the financial imperatives to capture the behaviour of the heap leaching process, it is not surprising that in the last few years there has been a substantial effort at the development of computational models that are sufficiently comprehensive to be exploited in optimising the industrial family of such operations.

The objective of this paper is to provide an overview of the work by the authors and their colleagues on the computational modelling of both sulphide and oxide ore complexes arising from a sustained research and technology development programme over the last decade. Aside from the technological challenge of building the computational model, in this case within a finite volume based discretisation procedure, together with embedding the effects of a range of transported multi-scale phenomena, the key issues of model parameterisation and validation are addressed. In such complex processes these are not straightforward issues. However, once such a model has been developed, parameterised and validated, then it becomes a really powerful tool for the optimisation of what is a very complex process family.

References
[1]
R.W. Bartlett, "Solution Mining", 2nd Edition, Gordon & Breach Science Publishers, Amsterdam, The Netherlands, 1998.

Return to the contents page
Return to the book description