Computational Modelling '15 Conference Diary

Computational modelling is perhaps the most highly specialised of all MEI Conferences, and over the years has attracted a small nucleus of around 45 dedicated scientists involved in the use of fundamental physical equations to predict the behaviour of complex mineral processing systems. This year's event was the 5th in the series and was held for the 3rd successive time in Falmouth, at the St. Michael's Hotel from June 9-10.

Tuesday June 9th

I welcomed the 35 delegates from 11 countries to a warm and sunny Cornwall this morning, then introduced the chairmen for the first session Yuande Zhao and Pablo Brito-Parada.

With Pablo Brito-Parada and Yuande Zhao

Philip Schwarz of CSIRO Mineral Resources Flagship, Australia, set the scene by reviewing multi-scale modeling approaches to process simulation, showing the similarities and differences compared with the more common multi-scale approaches to materials modelling. The methods were illustrated with reference to mineral flotation, where multi-phase CFD models of large-scale cells have been complemented by micro-scale CFD simulations of bubble-particle collision, and experimental and modeling studies of the bubble-particle attachment process itself. Other examples of multi-scale modeling were also summarized, highlighting progress on unit operations including aluminium reduction cells, leaching heaps and copper electrowininng.

Philip Schwarz with McGill's Joshua Sovcehles, Darryel Boucher and Kristian Waters

A characteristic of most minerals processing systems is that they are multi-phase. Conference regular Stephen Neethling, of Imperial College, UK described the use of Smooth Particle Hydrodynamics (SPH) to model multiphase mineral processing systems. The Lagrangian nature of SPH means that it is well suited to modelling these systems as it can naturally track interfaces and free surfaces. Stephen described a massively parallel SPH simulator and highlighted some of the features that have been implemented, such as a novel method for modelling surface tension and contact angles, as well as two way coupled solid fluid-interactions. The capabilities of this simulator were illustrated by means of examples showing its ability to simulate various minerals processing related systems at a wide range of different length scales from those resolving the behaviour of individual particles, bubbles and droplets all the way up to the equipment scale.

A 3D numerical analysis of the microwave induced damage in inhomogeneous hard rocks was presented by Michael Toifl of Montanuniversitaet Leoben, Austria . The aim of the research was to identify cracks on the microstructure level of a two component model rock introduced by the thermo-mechanical stresses during microwave irradiation. After microwave irradiation with a source of 25 kW, cracks along the grain boundaries that propagate across the boundaries of the microwave irradiated area were observed.

Comminution was the theme after the coffee break, with 5 papers from Sweden, China, Canada and Brazil.

The aim of the first paper, presented by Erik Hulthén, of Chalmers University of Technology, Sweden identified the fundamental tools needed to increase the resolution of how crushing plants behave and provide a framework for analyzing and suggesting model improvements that will improve performance when controlling the process.

Erik is one of four Chalmers representatives at the conference. Both he and Magnus Bengtsson are on the organising committee for the European Symposium on Comminution and Classification, which takes place in Sweden in September.

Chalmers University's Simon Grunditz, Erik Hulthén, Gauti Asbjörnsson and Magnus Bengtsson

It was good to see Yuande Zhou of Tsinghua University back in Falmouth, as since the last Modelling conference he has been appointed Chinese representative on the Editorial Board of Minerals Engineering. The numerical study given in his paper was aimed at helping the understanding of the pulverizing mechanism of various polycrystalline particles in roller mills and the optimization of the grinding machines for better efficiency.

In a further paper from Chalmers University Gauti Asbjörnsson analysed the incidents that can cause continuous crushing operations to experience altered performance due to changed conditions over a long operating period. A novel method for combining discrete probability simulations with time-continuous simulations for optimizing the process was presented.

Canada's McGill University is also represented by four delegates, and Arash Rafiei presented the first McGill paper, describing the modeling of the production rate of a continuous grinding system based on breakage rate of a semi-continuous grinding system. He showed that there is a direct connection between the breakage rate of the batch and continuous systems, such that the minimum grinding time can be obtained from the applied methodology, which may be helpful for designing a continuous grinding system.

In the final paper of the morning André Carlos Silva of Goiás Federal University, Brazil discussed optimisation of a ball mill at the Anglo American Phosphate Catalão-GO Company through simulations using the Moly-Cop optimization tools, thereby improving energy consumption and grinding efficiency.

The short afternoon session was chaired by Magnus Bengtsson of Chalmers University of Technology, who introduced the first speaker Arash Rafiei, of McGill University, who presented his 2nd paper of the day, discussing a study on the experimental modelling of energy consumption of a batch grinding system based on mill parameters and breakage rate. The energy consumption was predicted by applying Bond and a discrete element software as experimental and numerical methods, respectively.

This was followed by two hydrocyclone papers. Zeng Qi of Monash University, Australia, discussed the CFD modelling and analysis of classification of fine coals of different densities in large-diameter hydrocyclones. The effect of density distribution leads to misplaced coal particles at the outlets of a hydrocyclone but is not clearly understood. The potential use of large hydocyclones provides the opportunity for circuit in coal preparation to be implemented in the simplest possible manner to overcome the adverse effect of density.

Jiang Chen and Zheng Qi of Monash University

Bawemi Sichinga-Mtonga of the University of Stellenbosch, South Africa, also presented a CFD approach to study the hydrocyclone underflow discharge, which can be used as the basis for online monitoring and control since the operational state can be inferred from its shape, an umbrella-shaped discharge indicating dilute flow while rope-shaped indicates dense flow. For optimum separation conditions, the transitional state between umbrella-shaped and rope-shaped underflow is desirable.

It was interesting to talk earlier to Bawemi and her colleague Iku, who will be presenting a paper tomorrow. As part of their training to become lecturers at the Copperbelt University in Kitwe, Zambia, they have been seconded to the University of Stellenbosch to undertake PhDs, so I am pleased that they were able to come to Falmouth to present results of their research projects.

With Bawemi and Iku

After an afternoon Cornish cream tea, we set off for the traditional coastal path walk, finishing at the 17th century Chain Locker pub by the inner harbour (see also posting of 9th June).

Overlooking the Fal estuary

 

Wednesday June 10th

The dense medium cyclone (DMC) process is one of the most significant unit operations in the modern coal industry. In recent years, many mathematical methods such as the combined Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM) have been widely used to describe the multiphase flow in a DMC. However, such a CFD-DEM model is very time-consuming in computation for fine particles and not suitable for large scale industrial systems. Alternatively, a CFD-DEM model, facilitated by a “parcel-particle” concept to account for fine particles for simulating the flows in a large diameter DMC, was proposed by Jiang Chen, of Monash University, in the first paper of the morning session, chaired by Johan Zietsman, of University of Pretoria, South Africa.

Darryel Boucher, of McGill University, Canada, described how a multi-fluid Eulerian model combined with a dense discrete phase model (DDPM) was used to simulate water and iron ore particles (hematite and quartz) flowing in a spiral concentrator. Flow behaviour and particle velocities, including particle migration, were compared with experimental data from literature and positron emission particle tracking (PEPT) results. This demonstrated the potential of the Eularian-DDPM modelling approach for accurate simulation of spiral, and other density-based concentrators.

The dynamic behaviour of µm-scale ferromagnetic particles in suspension is relevant for various mineral beneficiation processes. It is, however, difficult to experimentally study such processes at the particle-level. In these instances it can be advantageous to resort to suitable particle simulation methods. Anders Sand, of Luleå University of Technology, Sweden, described a Stokesian dynamics approach for simulation of magnetic particle suspensions. The method inherently considers hydrodynamic interactions, but additional interaction models can be included depending on the system under investigation. The method was shown to be useful for studying ferromagnetic suspensions in mineral processing applications, and for understanding and predicting the efficiency of mineral separation processes.

In the session following the coffee break, chaired by Jiang Chen, of Monash University, Australia, Kirill Ivanov of the Russian Academy of Sciences presented a new mathematical model for dynamics of the vibrofluidized layer of granular material, enabling efficient computational evaluation of separation dynamics of minerals in force fields. The timeliness of such approaches to separation arises from a number of reasons including economical and environmental motives for avoiding the use of water as a disperse phase, importance of proper balance between production and extraction rates etc. The model presented in the work was built on the assumption of low content of magnetic or dielectric particles in the material.

An accurate description of the evolution of bubble size distribution is extremely important for modelling flotation columns. Most of the existing bubble column models make a simplifying assumption of monodispersed bubbles due to the high computational cost associated with simulating such complex systems. While this assumption may be reasonable in some situations, the polydispersed nature of the gas phase can have a significant impact on the flow dynamics in many other cases. Gaurav Bhutani, of Imperial College, UK, described the development of a polydispersed multiphase model of a bubble column which has been solved using a parallelised adaptive mesh finite element method.

Pablo Brito-Parada, Bertil Pålsson, Diane McBride, Gourav Bhutani and Francisco Reyes

According to Iku Mwandawande, of the University of Stellenbosch, South Africa, flotation columns are renowned for their improved metallurgical performance compared to conventional flotation cells. However, increased mixing in the column can adversely affect its grade/recovery performance. He described a study where the mixing characteristics of the collection zone of industrial flotation columns were investigated using Computational Fluid Dynamics (CFD). Liquid and particle residence time distribution (RTD) data were generated from CFD simulations and subsequently used to determine the mixing parameters (i.e., the mean residence time and the vessel dispersion number).

In the final paper of the morning André Carlos Silva of Goiás Federal University, Brazil discussed modelling of the hydraulic entrainment phenomenon in micro flotation, the results showing a strong correlation between the airflow, particle size and the hydraulic entrainment of the particles.

The afternoon session, chaired by Stephen Neethling, of Imperial College, commenced with four pyrometallurgical papers.

Slag temperature excursions tend to occur in platinum melting furnaces, which lead to excessive matte superheat, increased refractory wear and run-outs in extreme cases. The furnace walls are protected by a slag freeze lining, maintained by intensive side wall cooling. A computational model was developed at the University of Pretoria, South Africa, and described by Nicole Andrew, used OpenFOAM to simulate furnace operation so that various slag temperature drivers could be investigated to provide insight for an improved slag temperature control strategy. The model has been used to study the factors that drive slag temperatures inside these furnaces.

In a further paper from the University of Pretoria, Wihann Leipolt discussed the multiphysics modelling of a steel belt sintering process, which is used in the ferrochrome industry to sinter pellets of chromite fines, in order to introduce the fines into a submerged arc furnace. The process is complex and interactive in nature, is not well understood and has various operational problems associated with it. To improve understanding and assist in problem solving, a combined CFD and DEM model was developed to investigate the process and interactions between the process and the process equipment.

Johan Zietsman, Nicole Andrew and Wihann Leipolt of University of Pretoria

In an ironmaking blast furnace operation, an uneven charging of material in the circumferential direction has a negative impact on stable operation. Therefore, visualization and prevention of uneven charging are great issues. Yoichi Narita, of the Nippon Steel & Sumitomo Metal Corporation, Japan, showed how particle flow in a bell-less type top charging system with parallel hopper was simulated by using DEM.

Yoichi Narita with Stephen Neethling and Yuande Zhao

The cement industry has come a long way from the conventional wet-kiln process, with the addition of multi-stage cyclone technology, preheating of raw-meal and subsequently feeding ‘preheated and calcined hot-meal’ into calciners, either in an in-line or separate line configurations. In modern precalciner rotary kilns, fuel is fired in both kiln and calciner. In calciners, over 60% of the fuel is fired and most of the endothermic reactions are accomplished during the calcination of the injected pulverised lime stone (CaCO3). Since both mass and heat transfer take place in a calciner, conventional CFD based modelling approaches fall short of adequately capturing the calciner’s flow field, temperature and gas species. A new modelling approach based on mineral interactive CFD was developed by Cinar, UK, and described by Tahir Abbas. 

Tahir Abbas (right) with Johan Zietsman

The final session finished with a paper on heap leaching from Francisco Reyes, of Imperial College, UK who discussed how, although different modelling and simulation tools have been used to predict the leaching process, due to its complexity some assumptions have to be made. A well-established assumption is that of quasi-static behaviour, which implies that accumulation terms are unimportant. This assumption is made in, for instance, the shrinking core model and its variants. Both a dynamic and a quasi-static simulator were implemented within a finite difference framework in which the particle shape, its porosity and the initial grain distribution were obtained from X-ray micro-computed-tomography.

Amanda then closed the conference and invited everyone to attend Computational Modelling '17 in Falmouth in two years time.

As always this specialised conference, although small, attracted a very friendly group of people, and it has been a pleasure to be part of it. A special issue of Minerals Engineering, edited by Dr. Pablo Brito-Parada is planned for early 2016, and unrefereed papers from the conference are available on USB from MEI Online.