Steelmaking processes for developing countries BF-LD versus DR-EL by Henry G. Chlala, B.Sc. THE PATTERN of demand for steel products in developing countries is initially simple and leans heavily towards a relatively high usage of reinforcement bars, wire rod and small sections. While the magnitude of the market for these products justifies the setting-up of adequate rolling mills, usually fed by electric arc scrap melting furnaces, that for flat products is generally too small to warrant the installation of continuous hot and cold rolling facilities except in countries such as India and Brazil where the demand is high in absolute terms. As the economy of a developing country with no iron ore resources progresses, steel consumption increases to a point where even the demand for the common range of steel products cannot be met entirely by melting local scrap. Importing steel scrap cannot be a practical solution except over intermittent periods, and the problem ultimately arises as to when and how any developing country should embark on steelmaking proper. A similar problem exists when a country already producing primary iron and steel has to expand its capacity. The conventional long-established blast furnace/ oxygen converter route (BF/LD) to steelmaking has been developed to a high efficiency level through the design of larger units and the introduction of fuel oil/oxygen injection to reduce coke consumption; on the other hand, direct reduction methods to produce sponge iron have been devised and brought to the commercial stage relatively recently. Direct reduction (DR) processes utilise either gaseous or solid reductants. Among the former type are the semi-continuous HyL and the continuous Midrex (Midland-Ross) processes, based on reformed natural gas, and among the latter, the SLRN method utilising non-metallurgical coal or lignite. Sponge iron is produced and is subsequently melted in an electric arc furnace to produce liquid steel. The main advantages of DR processes over the BF/LD process are that they are less capital intensive and that they can be viable for small capacities (0.2-0.5m. tpy). On the other extreme side of the scale, blast furnaces of 3.5m. tpy of hot metal are being currently erected in Japan and other industrially advanced countries, thus resulting in a considerable reduction in the capital charges per unit of output. Other advantages of DR processes are the possibility of utilising indigenous reductants instead of metallurgical coke (often in short supply and characterised by rising prices), and that they lend themselves to expansion by small increments. Their disadvantage is the dependence on the supply of high grade feed which should be in the pellet form (minimum 65% Fe), as *Director, Techno-Economic Studies Division. L. H. Manderstam and Partners Ltd. the presence of silica in lower grade ores has an adverse effect on processing costs, mainly due to a higher power consumption for electric melting (power requirements per ton of sponge iron increase by about 30 kwh for every 1% Si02). World production of iron oxide pellets is growing steadily as a result of increased processing of lower grade ores into concentrates. Pelletizing followed by hardening is a most convenient way of disposing of these concentrates, and many iron ore deposits in various parts of the world are now being exploited economically in this manner. The high crushing strength inherent to the pellets induced blast furnace operators to incorporate them into the burden and to achieve some savings in fuel costs. It can be expected that more and more oxide pellets will become available on the world market, particularly from Australia, Canada, Liberia and North Africa. The choice between various direct reduction pro-cesses will depend primarily on their respective cost performance ratio per unit of output for a given fuel, and on their commercial availability for a particular plant size. While the HyL process has not been developed so far beyond the 0.2m. tpy limit, the other two well known methods, typically Midrex based on shaft kilns and natural gas, and SLRN based on rotary kilns and coal, have reached capacities of 0.4-0.5m. tpy of sponge iron per unit. The advantage of the latter two processes over HyL is that they are continuous, hence with lower capital costs and fuel consumption. Up-to-date capital costs for a HyL plant of sizeable capacity are not available; those of Midrex and SLRN are comparable and are estimated to be in the range of US $35-$40 per ton-year (excluding pelletization, i.e. assuming an oxide pellet feed) for 0.5-1.0m. tpy of metallized pellets. Approximate fuel consumptions are as follows: HyL 4.5 x 10 Kcal/ton product (NG) Midrex 3.0 x 10 Kcal/ton product (NG) SLRN 3.7 x 10 Kcal/ton product (Coal) Thus, unless the HyL process is further developed to a competitive level, preference should be given to a continuous shaft reduction process if natural gas is available, or to a continuous kiln reduction process if a suitable coal can be secured. When both natural gas and coal are available, the choice is pre-conditioned largely by their relative cost on a calorific basis (the break-even point being reached when the cost of coal is roughly 20% lower than that of natural gas). Other considerations might intervene, such as government policy as regards the utilisation of fuel resources. Processing costs per ton of metallized pellets from 16 MB MONTHLY