Computational Modelling '13 - conference diary

Computational Modelling ‘13 was held at the St.Michael’s Hotel, Falmouth, UK from June 18-19, 2013. 

This was the 4th in the series of what are perhaps MEI’s most specialised conferences, attracting a small but elite group of participants involved in the use of fundamental physical equations to predict the behaviour of complex mineral processing systems.

This is my diary of the 2-day event. I am by no means a specialist in this field, so I hope that delegates will add their own comments to provide a definitive report of the meeting.

Tuesday June 18th

Jon Wills opened the conference this morning and welcomed the 43 delegates, representing 13 countries: Australia, Austria, Brazil, Canada, Chile, China, Japan, South Africa, South Korea, Sweden, UK and USA. Although there were a number of new delegates, there is a nucleus of researchers who are regular attendees, notably Prof. Aibing Yu’s group of seven from the University of New South Wales, Australia and Dr. Bertil Palsson with his team of 4 from Lulea University of Technology, Sweden.

Prof. Aibing Yu (centre) with his University of New South Wales team

The first paper this morning was presented by Prof. Luis Cisternas of the Universidad de Antofagasta, Chile, who discussed how different objective functions have been used in the design of concentration plants. Frequently used functions correspond to maximizing revenue or profit. However, there is no study on the effect of the type of objective function on the design of these plants. He analysed the effect of various objective functions, including maximisation of profits, revenue maximisation, maximisation of net present value, maximising the rate of return, minimisation of the payback period, among other functions and how the results show that the objective function has a significant effect on the solution obtained.

Luis Cisternas (right) with Chilean delegates Christian Ihle,
Raimund Burger and Edelmira Galvez

This was followed by four papers on the flow of materials in bulk materials handling systems, the first, from Chile, on the use of computational modelling for efficient long distance ore transport in pipelines, followed by a paper from Australia investigating the use of DEM to model the interactions between particles and process boundaries.

Comminution has been the main theme of the day, with eight papers, from South Africa, Australia, Brazil, China, Sweden and Chile. The first two, which took us to the lunch break, discussed the scale-up of batch grinding data for simulation of industrial milling.

A leisurely lunch break

After an excellent lunch, Andre Silva of Brazil presented a simple iterative algorithm to calculate the closed circuit circulation load which allows the construction of reliable balances of mass, grade and water.

Yicai Wang, of Australia’s JKMRC, presented work aimed at understanding rock breakage behaviour, to provide a means to improve fragmentation efficiency. In the study, numerical modelling technology and image analysis were used to investigate the behaviour of actual rock in breakage processing.

With China's Yuande Zhou

A new member to the conference series is the Tsinghua University, China and Yuande Zhou showed how a coupling approach of the finite element and smoothed particle hydrodynamics methods can be adopted to model the complex deformation and failure process of coal particles in a roller pulveriser.

Par Jonsen (left) at the poster displays

Smoothed particle hydrodynamics was also discussed by Sweden’s Par Jonsen as a novel method for modelling of interactions between pulp, charge and mill structure in tumbling mills. A SPH description of the pulp fluid was introduced. The lifters and the lining were modelled with the finite element method, and the grinding balls with DEM. This combined computational model makes it possible to predict pressure and shear stresses within the pulp fluid.

A further paper from the JKMRC discussed the performance characterisation of AG/SAG mill pulp lifters using CFD techniques applied to a laboratory scale mill. The CFD technique was validated against Laser Doppler Anemometry measurements. This understanding was then used to build a cohesive computational framework for modelling industrial pulp lifters and to investigate unique problems associated with their design and performance.

An intensive first day finished with a paper from Chile on the use of computational fluid dynamics for risk analysis of cross-country ore hydraulic transport operations. In particular, CFD has the potential to predict, with reasonable accuracy, the progression and final condition of spills driven by pipeline leaks at selected locations, at a relatively modest computational cost.

After a final coffee, it was time to unwind with the guided coastal path walk. Despite the misty conditions, around 30 delegates took part on the 8km walk, giving them a chance to chat informally before ending up with beers at the old Chain Locker pub by the harbour.

Overlooking a misty Falmouth

At the 16th century Pendennis Castle

Welcome beers at the 17th century Chain Locker pub

Wednesday 19th June

Yesterday's sea mist has burnt away, and today dawned sunny and clear.

Of all the models of hydrocyclones that have been developed over the past 40 years, the most well-known and useful has perhaps been that of Plitt (CIM Bulletin, Volume 69, December 1976). Combining the first industrial database on cyclones generated at JKMRC with his own laboratory data, Plitt developed a model which over the years has been subjected to many revisions and corrections. In the first of the morning’s presentations, Andre Silva discussed the influence of two operational parameters (percent solids in feed by volume and the pulp feed flow) on seven models based on Plitt’s models, with reference to the corrected classification size (d50c). From the results obtained a different value for Plitt’s model constant was proposed, resulting in another revision for the model.

This was Andre's second paper at the conference. He and his wife Elenice attended Computational Modelling '11, and loved Falmouth so much that this time they have brought their family out with them- a party of seven!

Amanda with Andre and Elenice Silva

Maryam Ghodrat of the University of NSW, Australia presented a numerical study of the gas¬liquid¬solid flow in hydrocyclones with diffifferent designs of vortex finder. It was shown that a cone-shape vortex finder design can improve the performance of the traditional hydrocyclone with its cylinder-shape vortex finder.

The aggregate industry in Sweden is investigating production methods to improve the quality of manufactured sand for concrete production. A common way to improve the shape of the particles is using a VSI crusher; however the crushing process creates large amount of fines (particles under 63 µm) which is negative for the finished concrete. The industry is therefore investigating methods to reduce the amount of fines; one investigated method is air classification, and Robert Johansson of Chalmers University presented work on an investigation on the use of CFD on a centrifugal air classifier to increase the understanding of the influence of the geometrical design on the cut size and the resulting particle size distribution.

Jonh-HwanYoon prepares
for his presentation

This was followed by a further investigation on air classification, presented by Jong-Hwan Yoon of Daegu University, South Korea, whose team conducted computer simulations by using a commercial CFD program on a circulating air classifier in order to control the cut-size of the product.



Vasile Murariu of Metso, USA, then took us to the morning coffee break with a paper on the simulation of a low intensity magnetic separator model using DEM, CFD and FEM magnetic design software.

Coffee in the sun

The efficiency of sensor-based sorting depends on both correct identification and separation of particles. It is known that the distribution of particles will affect the correct separation of particles; when particles are in close proximity there is an increased probability of co-deflection of unwanted material. Robert Fitzpatrick, of Camborne School of Mines, UK, presented work aimed at modelling the air ejection manifold of a sensor-based sorter to predict the effect of particle proximity on co-deflection and separation efficiency.

The dense medium cyclone is widely used to upgrade run-of-mine coal in coal preparation. It is known in practice that wear is a serious problem to the operation of DMCs but not well understood. Kaiwei Chu, of University of NSW showed that the wear rate of DMC walls due to the presence of coal particles is predicted by a combined computational fluid dynamic and discrete element method (CFD-DEM), facilitated with the Finnie wear model in the literature. The numerical results show that the groove wear pattern observed in practice can be predicted by the approach and severe wear locations are also predicted.

Two papers from Chile discussed the use of advanced models for simulation and optimisation of dewatering systems. Water conservation is increasingly important in mining, particularly in arid and semi-arid regions, such as northern Chile, where fresh water scarcity has led to its replacement by other sources of water such as sea water, and optimisation of its use. It is known that in copper concentration plants approximately 40-60% of the water lost is due to water retained in tailings slurries. Luis Cisternas presented a method of designing dewatering systems considering hydrocyclones and thickeners as dewatering equipments. Several case studies were considered, included comparisons with an experimental study of dewatering in a copper concentrator.

The last session of the conference highlighted computational modelling developments in hydro and pyro metallurgy.

Workers at Swansea University, in collaboration with FLSmidth of USA have developed a comprehensive heap leach model, developed within a computational fluid dynamic software framework, providing a modeling tool capturing reactive dissolution in low grade ores of oxide and sulfide minerals. These systems involve suites of very complex reactions, which are closely coupled with the thermal conditions and key microbial populations. Diane McBride showed that once the model has been parameterized it can be utilized to investigate a range of process variables, such as irrigation rate, lift height, air injection, and bacteria limitations, amongst others.

Two papers from Imperial College continued the theme of heap leach simulation, the first showing how fluid flow and mass transfer can be predicted in heaps using smoother particle dynamics, and the second using mesh adaptivity techniques to simulate multiphase flow.


Electrorefining is widely used to produce high purity copper. Shunsuke Kawai of the Mitsubishi Materials Corp., Japan described the development of a CFD model for the industrial copper electrorefining process. During copper electrorefining, natural convection is induced due to the electrochemical deposition and dissolution of copper on the electrode surfaces. Furthermore, the electrolyte is circulated in a way unique to the individual tankhouse. Because the electrolyte flow pattern is formed by both the natural convection and the forced circulation, analysing the interaction between these flows is indispensable in order to understand the flow pattern in the cell. In the study, both electrolyte flow and cupric ion concentration fields were numerically analysed by the CFD model.

Computational modelling of the iron blast furnace is being undertaken at several institutions, and as well as three poster presentations from University of NSW, there were two papers presented in the afternoon on CFD-DEM modelling of various blast furnace elements.

Two more pyrometallurgical papers were presented, both from South Africa. Bertus de Villiers described the development of a validated CFD thermal model for the Anglo American Platinum converter plant, which produces a copper-nickel sulphide converter matte which is slow-cooled in ingots over several days. In the last paper of the conference, Quinn Reynolds, of Mintek, discussed computational modelling of shear-layer instabilities and vortex formation in DC plasma arcs, the principal heating and stirring element in plasma arc furnaces.

Closing the conference, Amanda invited all delegates to return to Falmouth in June 2015 for Computational Modelling ‘15. All 37 papers from the conference are available in draft form from MEI and authors have been invited to submit final papers for peer-review for a special Computational Modelling issue of Minerals Engineering.