Metallurgical Research and Its Contribution to Industrial Development

Nic Barcza
DOI: https://doi.org/10.2139/ssrn.3930154
2021-01-01
SSRN Electronic Journal
Abstract:Scandinavian countries have contributed very significantly to the metallurgical research, development and industrial-scale implementation of technologies for all the major and several minor ferroalloys over the past seventy five years. This includes ferronickel (FeNi) with the RKEF process (ELKEM Norway) and ferrochromium (FeCr) with the first 48 MVA FeCr furnaces in South Africa. The development of the pre-heated sintered pellet process by Outokumpu and Outotec in Finland has been installed in most of the FeCr producing regions and led to the largest 100 MW SAF.The ELRED process piloted in Sweden based on fluidised bed iron ore fines prereduction followed by DC arc furnace (ASEA/ABB, Sweden and Lurgi, Germany) led to the application of DC furnaces for FeCr, FeNi and Ilmenite smelting on a number of industrial-scale projects between 1983 and 2014.Important contributions were also made to the technologies during the implementation in countries where these advances were applied, in particular South Africa where important research work has continued in collaboration with Scandinavian and other institutions for over forty years. This research work relates to several of the considerations mentioned below: The three main bulk ferroalloys, namely FeNi, FeCr and FeMn/FeSiMn utilise different ore-types and have individual smelting characteristics. However the main ferroalloy applications are reasonably similar since they are mostly used in alloy and stainless steel applications but in some cases at different concentration levels based on the required specifications. The production of these respective ferroalloys requires significant amounts of carbon as the reductant and electrical energy inputs. They also have relatively large carbon foot-prints that contribute to greenhouse gases mostly as CO2. The use of electrical energy for the power generation based on coal also adds to the footprint so the option to use hydro power is usually preferable and the use of hydrogen is also being considered mostly for reduction of the iron oxides but it can be effective if the iron oxide in the ores used for ferroalloys is amenable and some forms of hydrogen can in principle reduce the other ferroalloy oxides if used in an ionised from as H+ produced in a thermal plasma.The thermal pre-treatment can lower the electrical energy requirement significantly but also adds to this carbon footprint if a primary fossil fuel such as coal is used but 2 this can be minimised by using off gases from the smelting stage. The use of available chemical and sensible thermal energy to pre-heat and pre-reduce the ore feed can decrease the carbon foot-print if the heat transfer and pe-reduction are effective. This the case with chromite pre-reduction and metallisation but nickel laterites and manganese ore pre-reduction are more challenging for different reasons.The capital cost for the pre-treatment can be very significant and may not be justified in the context of the ferroalloy plant investment alone depending in electricity prices and availability. The options of combined process stages to utilise the thermal energy in the off gases has been considered over many years. The greatest use of the staged process approach is Ferronickel (FeNi) and Nickel pig iron (NPI) that is also a ferroalloy merely with less Ni (typically 15 vs 25 to 30%) and to a lesser extent for FeCr and FeMn alloys smelting up to recently. The standard nickel ferroalloy production process is the RKEF (Rotary Kiln Electric Furnace). The RKEF stages combine drying, calcining, partial pre-reduction (mostly Fe) and smelting plus refining. There is scope to improve the smelting efficiency using some proposed advanced technologies that will be mentioned as potential ways of reducing the carbon foot-print such as hydrogen based pre-reduction of fine ore and concentrate feed and smelting in a DC arc furnace with fluidised be pre-treatment so called ‘cloud cover’ to minimise heat losses. The scale up of FeNi and FeCr furnaces to around 90 and 100 MW respectively has contributed most significantly to the economy of scale but FeMn alloy furnaces are still not able to reach this MW level but could in principle do so with DC SAF technology.The quantity of off gases produced and their chemical and thermal energy are suitable for drying and partially preheating the ore feed. The smelting of FeCr is quite different from FeNi in that the process temperature is higher but the slag to metal ratio is much lower than FeNi (typically around 1.5 to 2 for FeCr and FeMn but 15 up to almost 30 for FeNi). This results in a much lower energy consumption per ton FeCr and FeMn vs FeNi (typically 2.5 to 4.5 MWh/t alloy vs 25 to 50 MWh/t Ni) depending on the ore grade and thermal pre-treatment process used.The reducing conditions required for Cr compared with Ni are much more stringent and for manganese is much more so as shown in the Gamma graphs for Ni, Co, Fe, Cr and Mn where the respective extents of extraction or recovery are illustrated. The indicative Mass and Energy Balance (MEB) relationships for FeNi, NPI, FeCr and FeMn alloys without and with pre-treatment show the differences and impacts of different processing stages. The effective integration of these stages into one vessel offers significant potential benefits including those referred to in the presentation.However the trade-offs of the benefits vs the challenges need further careful considerations based on the operational and cost aspects and this required further research and development.The analysis of the many bath smelting processes developed for iron units and the challenges experienced when applied to these ferroalloys provides a very useful basis for evaluating this subject so that the appropriate research and development work can continue for some ferroalloys where appropriate. The application to the production of the respective ferroalloys also requires an extensive assessment based on the lessons.
What problem does this paper attempt to address?