Abstract
With deteriorating ore quality in the PGM industry, it is expected that future concentrate blends will be lower in base metals and sulphur, and higher in chromium. There is uncertainty about whether current technologies can smelt future blends; do we need new technologies, or can current technologies, with minor changes still deliver the required throughput.

In order to gain insights into the performance and behaviour of a PGM smelter furnace, a fully coupled multiphysics model was developed. The model seeks to describe each of the main physical phenomena within a typical PGM smelting furnace. Hence, the model includes representations of the slag bath, matte bath, black top, freeboard, refractories, and electrodes. The physical phenomena solved for include fluid flow within the slag and matte baths, radiation within the freeboard, thermal conduction through the refractories, and magnetohydrodynamics to approximate the flow of current, current-generated magnetic fields, electromagnetic Lorentz forces, Joule heating and power density distributions through the various furnace regions.

Given that the black top is the region within the furnace where the majority of phase change and energy absorption is expected to occur, a novel throughput model was developed. The temperature profile through the black top is described by solving the full global energy conservation equation, effectively describing both the transport of energy and the magnitude of energy associated with phase change. The temperature-enthalpy relationship governing the phase change energy is approximated using computational thermochemistry. This throughput model is therefore capable of estimating the location and magnitude of energy absorbed due to phase change and approximating the spacially varying rate of smelting through the black top.

The model was validated against measured input power, heat losses, temperatures, and throughput of a typical PGM smelting furnace. A large design-of-experiments (DOE) was then performed to assess the main drivers and system sensitivities and investigate the influence of changes in the concentrate on throughput. The main system parameters that were considered as part of DOE included slag material properties (viscosity, thermal conductivity, and thermal expansion), matte material properties, bath depths, freeboard dust loading, slag operating temperatures, resistance set-points, and black top material properties.

Click here to watch te presentation.

Reference
Willem A. Roos [1], Myren Rajh [1], Alfred E.J. Bogaers [1,2], and Johannes H. Zietsman [1,3]

1. Ex Mente Technologies
2. Department of Mechanical and Aeronautical Engineering, University of Pretoria
3. Office of the Dean, Faculty of Engineering, Built Environment and Information Technology, University of Pretoria

Biography
Willem is a mechanical engineer with a newfound interest in thermochemistry and its inclusion into multiphysics and process models. He is part of the Multiphysics team at Ex Mente where he contributes to the solvers and models that are developed to investigate large, complex, high-temperature systems. Through his PhD research at the University of Pretoria, he developed, implemented, and tested an acceleration algorithm that would allow a large number of multi-component, multi-phase, high-temperature thermochemical equilibrium calculations to be included into multiphysics and process models more efficiently.