Hematite compacts were isothermally reduced with hydrogen at 973-1273 K. The reduction was carried out up to 90-95% reduction degree using thermogravimetric technique. The reduced products were subjected to re-oxidation in dry air. The rates of re-oxidation at 473-1073 K as a function of time was studied. Microscopic examination, and porosimetric measurements were used to elucidate the kinetics and mechanisms of reduction and re-oxidation. During reduction of iron oxide, the rate was increased with temperature and the structure of sponge iron produced was temperature dependent. Three transition temperatures were identified during re-oxidation. Reoxidation at the initial stages was controlled by interfacial chemical reaction whereas at both the intermediate and later stages, solid-state diffusion was the rate controlling step.

View full abstract

Several properties of liquid Fe-S-O at 1473 K provided the motivation for using iron-oxysulfide melts as smelting medium for iron oxides. Experiments were carried out to determine rates of smelting of Fe-S-O melts by solid carbon at temperatures in the range 1361-1525 K. Effect of oxygen concentration, bath agitation rates, and carbon reactivity on the smelting rates were examined. The smelting rates increased with increasing oxygen concentration, bath temperature, and carbon reactivity, but with decreasing bath agitation rates. Microscopic observations and chemical analysis of the melt samples collected after the experiments revealed that sulfur content of the product metal was around the thermodynamic equilibrium value and the metal was not saturated with carbon. It was proved based on a micro-mixing model that the transport of oxygen within the liquid boundary layer adjacent to carbon sample was not a rate limiting process. The reactivity of carbon gasification reaction at the liquid-solid interface was found to be the most likely rate limiting step.

View full abstract

A laboratory scale fine particles-gas conveyed system was utilized to measure the reduction rates of liquid wustite with hydrogen at high temperatures. N₂-H₂ mixtures having various flow rates and compositions were flowed downward through a cylindrical reactor maintained at a constant temperature of 1723 to 1823 K. A batch of pure spherical wustite particles (mean dia.; 58μm) was concurrently fed into the reactor at a small constant rate and reduced in a hot zone.
The reduction process was found to proceed in such a manner that metallic iron particles were enclosed inside a wustite droplet. Rate analysis was made of one dimensional mass balance equations for particle taking the shrinkage into consideration. Under relatively small reducing potentials, it was concluded that the major fraction of overall reaction resistance is attributable to chemical reaction. However, under higher reducing potentials, the reduction process was estimated to include an appreciable diffusion resistance within the liquid phase. >From the temperature dependence of forward chemical reaction rate constants, the activation energy was evaluated to be 110 kJ/mol.

View full abstract

The pore structure of sinter cake was observed in three-dimensional manner by high energy X-ray computerized tomography. Most pores of 5 mm or more in two-dimensional size were stem pores linked with each other, forming complicated three-dimensional networks.
A new analysis method employing X-ray computerized tomography has been developed for evaluating the sinter cake pore structure, based on a network model. This method has successfully made it possible to quantify the complexities of the pore structure.
Applying this method to the analysis of sinter cakes different permeabilities, the importance of not only porosity but also pore branch structure for the permeability was elucidated. Promoting the coalescence of iron ore particles to secure open pores (stem pores) improves the permeability of sinter cake because of producing an effectively permeable pore network.

View full abstract

Presented new ironmaking process on the basis of 100% usage of oxygen and hot reducing gas injection into the No. 2 blast furnace (1033 m³) at RPA Toulachermet. The process involves blast furnace top gas cleaning, compressing, removal carbon dioxide through chemical absorption, preheating the gas in the hot blast stove regenerators, and blowing into the furnace through tuyeres in place of the hot blast. This use of hot reducing gas (HRG) is accompanied by oxygen injection through each tuyere. There have been 13 trials of this technique in the period 1985-90 producing 250000 t of hot metal. In one particular trial the coke rate of 367 kg/thm was achieved with 251 m³/thm of oxygen while producing 1700 thm/day of 2.2% silicon iron.

View full abstract

A model proposed in another publication¹⁾ for steady state uni-directional solidification is evolved to a relationship between solidification variables and dendrite structure/morphology in non-steady state directional solidification. Dendrite structure and inclusion behaviors in columnar zone of continuous cast slabs are discussed with the model.
Behaviors of inclusions in calcium-treated clean steels are discussed based on equilibrium considerations as well as widely accepted results and knowledges hitherto. An idea is introduced to interpret competitive precipitations of oxides and sulfides for calcium in residual melts between dendrite arms where the inclusions precipitate.
Quantitative predictions are enabled with the present model, and several equations are given to volume fraction, size and morphology of the inclusions in calcium-treated clean steels in relation with Ca-S-O balance and solidification variables.

View full abstract

A steady state, two dimensional mathematical model for continuous casting of steel has been developed. Towards this, governing partial differential equations of fluid flow and thermal energy transport together with the appropriate set of boundary conditions were derived and a procedure for their non-dimensional representation outlined. The modelling of (1) turbulence, (2) flows and energy transport within the mushy region, and (3) bulk motion of the descending strand on liquid steel flow and heat transfer phenomena were also discussed. The governing p.d.e's and the associated boundary conditions were solved numerically via a control volume based finite difference procedure. To this end, incorporating the SIMPLE algorithm, a computational procedure was developed in double precision, FORTRAN 77. Finally, three different industrial billet casting operations reported in the literature were mathematically modelled and direct comparison were made between predicted and experimental solidified shell thickness. Such comparisons demonstrated reasonable to excellent agreement between the two. Present estimates were also compared with our earlier predictions derived via an effective thermal conductivity based model. This indicated that for mathematical modelling of transport phenomena in continuous casting of steel, a "conjugate heat and fluid flow model" is the most appropriate.

View full abstract

In order to develop mold powder for high speed slab type continuous casting, the viscosity of molten flux and the melting rate are investigated. The results can be summarized as follows;
(1) The viscosity of molten flux is estimated from the equation as a function of the anion cation interaction parameter and the network parameter.
(2) The melting rate is estimated from the equation as a function of carbon content per unit volume and carbonate cantent.
(3) The improved mold powder was successful at a speed of 5 m/min.

View full abstract

The influence of alloy surface preparation as induced by mechanical polishing and electropolishing on the oxidation behaviour of AISI 316 stainless steel in dry air under non-isothermal heating (6 K·min^-1) followed by isothermal holding at 1423 K is reported. Mechanically polished surfaces exhibit a shorter incubation period for initial oxidation but better oxidation resistance during isothermal holding as compared to electropolished surfaces. Such observation is attributed to enhanced outward diffusion of Cr for easy and early establishment of Cr-rich oxide layer on the mechanically polished surfaces.
The morphologies of the scales and nature of their adherence to the alloy substrates have been characterized by SEM. Distribution of the alloying elements like Ni, Cr, Mn, Mo, Si as well as Fe and oxygen across the oxide layers and the type of compounds formed have been examined by EPMA, EDS and XRD techniques. SEM examinations of the alloy/scale cross section for the mechanically polished and oxidized steel, supplemented by the X-ray images of the respective elements, indicate preferential formation of a continuous Cr-rich layer near the oxide/air interface along with two continuous bands of doped Cr₂O₃ at the scale/alloy region. On the other hand, the scale formed on electropolished surfaces of the steel shows fragmented Ni-rich and Cr-rich oxide areas at the bottom region of the scale with mostly compact Fe₂O₃-rich layer at the oxide/air interface.

View full abstract

Torsion testing was performed to assess the effect of applying a strain rate correction to the pass strains and interpass times when simulating high strain rate mill processing using relatively low strain rate laboratory tests. The experiments were carried out on a low carbon steel grade, and the temperatures, strains, and interpass times selected were representative of those employed in the pre-finishing stands of rod rolling. In the first part of the study, the metadynamic softening kinetics of the steel were established over a range of strain rates, using a new technique whereby comparatively few specimens are needed. Multiple pass deformations were then performed at a series of strain rates using corrected and uncorrrected pass strains and interpass times, and the softening behaviours were compared. It was found that the strain rate corrections appropriate for laboratory tests must take into account the different softening kinetics that apply prior to and after the peak strain of dynamic recrystallization. Otherwise, erroneous conclusions are likely to be drawn about the extent of strain accumulation, which will then lead to errors when laboratory findings are applied to mill practice. A properly corrected simulation of an actual rod mill finishing schedule indicates that static recrystallization is absent in the finishing stands, leading to cycles of strain accumulation and dynamic recrystallization, followed by metadynamic recrystallization and nearly complete softening.

View full abstract