WO2014096751A1 - Zero gas emission continuous steelmaking process - Google Patents

Abstract

Crushed iron ore and ore fines without distinction are incorporated by chemical reaction into a closed- loop melt circulation reactor to form a thin liquid floating oxidic layer comprised initially principally of molten wustite. Hydrogen reduction is gas phase diffusion controlled so is unaffected by total pressure. Energy requirements are supplied by a carrier medium of molten iron circulating within an unmelted steel shell. Post-combustion of reduction off-gas is inherently dangerous so the reduction off-gas has no direct access to post-combustion. A multiple array of axi-symmetric turbulent jets inject clean hydrogen into the gas freeboard in the post-combustion zone to entrain the surrounding gas phase, predominantly turbulent flowing water vapour containing minor oxygen. Major recirculation of post- combustion off-gas is achieved using a gas ejector empowered by compressed oxygen as dictated for safe combustion. Atmospheric pressure operation is recommended with low-pressure steam enshrouding all plant containing hydrogen for enhanced safety.

Description

ZERO GAS EMISSION CONTINUOUS STEELMAKING PROCESS

This invention relates to continuous steelmaking directly from iron ore fines or crushed ore without the need for pelletising, sintering or compaction with a carbonaceous reductant as required by most alternative primary ironmaking processes available as substitutes for the traditional blast furnace and batch steelmaking. At the same time, the need for basic oxygen steelmaking is entirely eliminated, because carbon is not permitted free access to molten iron at any stage, so sub-surface nucleation and growth of carbon monoxide bubbles cannot lead to disruption of continuous processing. This latter issue, in the past has contributed to the lack of successful implementation of a whole range of continuous primary steelmaking attempts starting from principally virgin iron ore or molten pig iron.

The challenges of global climate change and energy conservation demand a paradigm shift in iron and steel technology. In particular, the disruption caused by sub-surface nucleation and growth of carbon monoxide bubbles in previous attempts to implement continuous steelmaking must be addressed. It would appear that no attempt has ever been made to effect continuous steelmaking directly from iron ore employing an essentially carbon-free iron melt.

To make real progress, the iron melt being continuously processed must be similar to any other normal liquid phase. Once this essential requirement is incorporated, rigorous process engineering can then be applied in conceptual designs to meet the challenge. There are a number of ways to proceed. For solid reductants, carbon transfer to the iron melt can be reduced to minimal proportions, if the solid carbon reductant, such as coal or biomass without prior treatment, is admitted carefully to a layer of oxidic melt floating on top of a carrier medium of molten iron in a closed loop melt circulation process. For gaseous reduction, the risk of carbon pick-up into the iron melt does not arise.

BACKGROUND OF THE INVENTION

When referring to the new steelmaking process in his autobiography, Sir Henry Bessemer stated: "The oxygen, next uniting with the carbon, sent up an ever-increasing stream of sparks and a voluminous white flame. Then followed a succession of mild explosions, throwing molten slags and splashes of metal high up into the air, the apparatus becoming a veritable volcano in a state of active eruption.

The scenario depicted above is totally incompatible with continuous processing. Accordingly, to eliminate the risk of such violent behavior, continuous steelmaking should preferably be undertaken only if there is no carbon initially dissolved in the bulk of the molten iron. The same conclusion was reached by the inventor in GB 2 438 570 B, which introduced continuous steelmaking based on natural gas. In this former case, a thin film of liquid ferrous oxide floating on the surface of molten iron carrier medium in a closed loop melt circulation process prevented dissolution of carbon from soot particles dispersed in the gaseous reductant, flowing at relatively high velocity within a small gap between the melt surface and the refractory ceiling immediately above. In such a scheme, solid hematite forms liquid iron oxide by the reaction given in Equation 1.

[FezOdrtu + {Fehiquw = ^FeO^^ (1 )

Reaction 1 is extremely endothermic and in the present context the heat required must be provided by the circulating molten iron carrier medium. Once an oxidic liquid layer is established, carbon no longer has access to the underlying bulk molten iron. This is the foundation for the elimination of oxygen steelmaking from the overall steel process.

No other technology is currently available nor on the horizon, which is capable of responding controllably to the significant thermal demand involved in transforming a hematite or magnetite ore charge, without partial pre-reduction, to a separate liquid phase of molten ferrous oxide saturated with iron, referred to as wustite. In the present invention this occurs in a single multiphase reaction zone at high intensity followed immediately by gravity phase disengagement to yield individual wustite and molten iron separate phases. This invention accomplishes this objective by direct contacting the as- received crushed ore, incorporating both ore fines and lump ore without distinction, with the circulating molten iron. The fundamental phenomenon involved, which ensures a successful outcome, is referred to as interracial turbulence. Solute-induced interfacial turbulence occurs because of the dramatic effect of dissolved oxygen on the surface tension of liquid iron. Fluctuations in the rate of mass transfer of a surface active agent such as oxygen across the interface involving liquid ferrous oxide and molten iron results in unstable interfacial tension gradients leading to convection currents, which cause more solute to be present at one point than in a neighbouring region and hence the onset of spontaneous interfacial turbulence.

Initially in the ore charging zone, under the influence of a steep temperature gradient, a raft of sintered feed material forms, constrained in moving forward by freeze lining protected boundary walls, whilst floating on the molten iron. This raft is attacked on its undemeath side by the circulating molten iron designed to flow forcibly at a velocity somewhat above that required to establish the critical Weber Number for interface instability. As a consequence, any liquid ferrous oxide once formed does not stay attached to the undemeath of the floating raft, but rather is carried forward by being dispersed in the continuous phase of molten iron. To make sure that the downward motion of the raft continues unimpeded in response to the undemeath side being melted and carried away by the circulating molten iron carrier medium, a degree of mechanical assistance is provided to overcome any tendency for solids to adhere to the containment walls involved during subsequent sintering. In its simplest form, vertical oscillation or jigging of the bed of charge solids at a frequency commensurate with the sticking physical properties of the particular ore charge material is conducted to facilitate attainment of this objective.

FURTHER KEY CONSIDERATIONS

Special attention must be focused on those regions in which freeze lining is mandatory. The whole reduction melt circulation loop falls into this category, not only because of the presence of high FeO content oxidic melt but also the molten iron carrier melt containing appreciable quantities of dissolved FeO is potentially aggressive towards conventional refractories. In addition, the launder carrying the overflow molten iron product stream away from the melt reduction loop needs freeze lining until the molten iron is de-oxidized.

Melt circulation technology is the foundation of the inventor's earlier patents relating to smelting reduction of iron ores, but in all cases the focus for energy provision to carry out the reduction of iron oxide to metal has been on post-combustion of reduction off-gases without heat transfer being inhibited by a slag layer on the so-called post-combustion arm of a melt circulation loop employing molten iron as the carrier medium for sensible heat transfer. In all previous cases, the energy requirements to satisfy the endothermicity of iron oxide reductions to metallic iron has been transferred in-situ to the carrier melt of molten iron by a combination of radiative and convective heat transfer and then transported to the reduction arm of the same melt circulation loop.

In the present invention what may appear as straightforward post-combustion of reduction off-gas is purposely avoided on safety grounds. A new approach is called for because post-combustion of hydrogen containing off-gases with high hydrogen contents introduces the risk of accidental serious explosion. This does not necessarily imply that post-combustion itself has to be discarded in favour of alternative modes of energy provision. What needs careful assessment at the process design stage is whether or not in a particular case, adoption of new specialized techniques can be invoked to make post-combustion entirely safe.

To highlight the issues involved, it is first necessary to appraise current state-of-the-art technology associated with hydrogen ironmaking in processes collectively producing what is known as direct reduced iron (DRI). Compacted DRI is referred to as hot briquetted iron (HBI) and is frequently exported to steelmakers worldwide. Alternatively, DRI can be fed directly into a downstream electric arc furnace or other steelmaking facility, encompassing subsequent decarburization and refining to yield a refined molten steel product ready for feeding to the tundish of a continuous casting machine.

According to the internet, a hydrogen-based steelmaking process is currently under serious consideration for commercialization. The reported plan is to produce hot-compacted direct reduced iron (DRI) using established technology, employing a sequence of high-pressure fluidized beds and feed DRI into an electric arc furnace. By these means the carbon content can be adjusted to between 1.0 and 2.5%. However, even with only minor contamination with carbon, violent sub-surface gas bubble evolution is not precluded. Accordingly, disruption of truly continuous steelmaking, if attempted in-line, is probably almost inevitable with the strategy involving DRI. The upstream fluidized bed technology envisaged typically employs input pressures of about 12-13 bar for hydrogen reduction of beneficiated iron ore.

From a health and safety viewpoint, the proposed hydrogen-based reduction process in the current invention has been designed so that all associated gas/solid contactors and gas/liquid contactors operate at essentially atmospheric pressure. Accordingly, there no attempt to pre-reduce the iron ore feed prior to its admission to the melt circulation reactor. Also, within the melt circulation loop itself, gaseous reduction only takes place at the base of a hydrogen containing gas layer a mere 10-20 cm or so in thickness above a quiescent layer of oxidic melt being transported along floating on top of a molten iron carrier medium. Whilst in operation, hydrogen is constrained from entering the flat refractory roof defining the upper surface of the reducing gas layer by maintaining an appropriate pressure gradient between a water vapour protective atmosphere, which acts as an explosion prevention shield, in the structural assembly. This comprises a flat suspended refractory ceiling and its associated steel joist girders, ail rigorously maintained above the steam dewpoint temperature corresponding to the operating pressure. At shutdown, an inert gas is substituted to take on the role of protective atmosphere.

Comprehensive analysis of the process kinetics indicates that at steelmaking temperatures the chemical reaction kinetics have negligible effect on the reduction of a sheet or layer of liquid wustite by hydrogen. This follows from earlier work published by the author (N.A. Warner 'Reduction Kinetics of Hematite and the Influence of Gaseous Diffusion', Transactions Met Soc. AIME, 230 (Feb. 1964), 163- 176) challenging the erroneous conclusion reached in research conducted by US Steel Corporation that dense pellets of hematite were reduced by hydrogen under chemical reaction rate control at both atmospheric pressure and at elevated pressures (W. M. McKewan: "Reduction of Hematite in Hydrogen at High Pressure", 5^th International Symposium on Reactivity of Solids, Munich, August 5-8, 1964). The relative effects of gaseous diffusion increase dramatically when the operating temperature is increased above the melting point of wustite because of a massive increase in the chemical reaction rate at temperatures above the melting point of wustite. The establishment of almost exclusive gaseous diffusion control for reduction by hydrogen at atmospheric pressure immediately removes the necessity for conducting hydrogen steelmaking at elevated pressures. This clearly has immense safety implications.

To secure the full benefits of enhanced safety, post combustion in the melt circulation loop must be conducted in a combustion chamber above the circulating molten iron carrier medium in a fully back- mixed condition; otherwise the risk of explosion is not precluded. In principle, fully developed turbulent gas flow throughout the entire length of the combustion chamber would be equally effective. The latter implies recycling some of the combustion off-gas to the combustion zone so that multiple additions of hydrogen through high velocity turbulent jets can be made along the length of the post combustion arm or zone into the turbulent flowing bulk gas phase. This aspect is crucially important and warrants a full hazards and operability (HAZOP) study in association with appropriate computational fluid dynamic (CFD) evaluation incorporating all relevant safety precautions.

Liquid oxide melts containing significant amounts of dissolved iron oxide are notoriously difficult to process, because of extremely aggressive attack on all commercially available refractory materials. The new process adapts the approach outlined in the author's paper on conductive heating and melt circulation in pyrometallurgy (N. A. Warner Trans. Inst Min. Metall. C, 2003, 112, C141-C154). Un- melted shells of solid steel are used for melt containment, rather than conventional or traditional refractories. The associated heat removal necessary to maintain a steady-state thickness of un-melted steel in a stable condition is by steam generation from strategically located boiler tubes receiving thermal radiation from the cooler surface areas of the steel shell in selected regions, where melt containment poses a potential problem. The associated heat flow is not to be seen as a heat loss but rather as a good way of generating steam for advanced steam turbine power generation and for satisfying other thermal demands.

SUMMARY OF THE INVENTION

This patent provides a method for direct smelting of iron ore as mined continuously through to refined thin slab steel product and thus represents a paradigm change in current steelmaking technology. Clearly, zero gas emission implies no greenhouse gases are generated in the process itself and there is no reliance on carbon capture and storage (CCS) to alleviate concerns about climate change and global warming. At the same time, it potentially represents a massive reduction in capital expenditure by eliminating wasteful semi-batch processing, currently endemic in the steel industry and

concentrating on truly continuous operation at a single site. In doing so, it opens the door to major reduction in overall energy consumption and thus very much enhanced sustainability for the global steel industry.

The key to successful implementation of such an ambitious target lies in the introduction of appropriate generic melt circulation technology. In the present invention, two separate melt circulation loops are required. The iron ore charge is continuously added to the first, causing continuous overflow by gravity of molten iron product stemming from hydrogen reduction of a thin layer of oxidic melt floating on a recirculating molten iron carrier medium. Prior to its entry into the second melt circulation loop, the overflowing molten iron is deoxidized not by addition of aluminium, silicon or other recognized de- oxidant but by countercurrent contacting with hydrogen itself, whilst en route to the second melt circulation loop. This is where in-line continuous desulphurization, dehydrogenation and final compositional adjustments are made in advance of continuous casting employing radically innovative new technology.

Both melt circulation loops in the present invention operate in the open-channel regime, and as such due attention in design needs to be given to ensuring that subcritical flow (dimensionless Froude number less than unity) conditions are implicit throughout. The melt circulation rate in the first of the loops needs to be typically about 30 times larger than that in the second. Gas-lift pumping or established RH steel vacuum degassing technology is used for melt circulation in both. Actual requirements depend on the chemical composition of the iron ore charge, its degree of preheat and various other factors, mainly cost items associated with rationalization of mass and heat transfer.

Typically, for 2 million tonnes per annum of refined steel product, in the region of 600 tonnes per minute molten iron circulation within a freeze lining of solid iron is typical of that required for the carrier melt to transport the necessary heat to the reduction zone from the so-called post-combustion zone. As ready pointed out for safety reasons, a totally new approach is mandatory to establish fully back-mixed conditions when recycled cleaned-up hydrogen is combusted with oxygen at temperatures up to 1800°C within the "post-combustion zone'.

There are, however, many challenges demanding practical solutions. Perhaps, the least obvious is continuous casting of the refined steel product. This aspect has not previously been addressed in the present context. Current technology provides opportunity to replace key items such as submerged entry nozzles (SEN), which persistently have to face clogging problems. These would be totally unacceptable, if truly continuous steel production was the desired objective.

In the steeimaking context, inclusions are non-metallic particles that are trapped in the solid steel matrix. Exogenous inclusions are those that come externally from outside of the steel, such as refractory bricks or flux used in molds and casters. Exogenous inclusions are typically large (> 1 mm in size) because they originate as particles of these outside sources that become entrapped in the liquid steel while it is being processed. Indigenous inclusions are those that are formed from chemical reactions within the liquid steel as it is being processed. Indigenous inclusions are typically 0.001-1.000 mm in size. When striving for clean steel, it is the indigenous inclusions that are the major non-metallic inclusions needing to be controlled. Ideally, what is required is effectively zero indigenous and exogenous non-metallic inclusions. This is a target, probably more difficult to attain than the zero gas emission continuous steeimaking focus in the title. However, there are several key attributes in what is now being proposed that could lead to drastic reduction in inclusion population well below current levels.

1. Firstly, most of the molten iron or steel is contained within permanent linings of solidified iron (freeze linings) rather than refractories.

2. Dephosphorisation is designed to take place in the primary ironmaking loop under relatively low intensity conditions.

3. Reliance on aluminium or silicon "killing" to deoxidize the iron melt implies that chemical reactions take place in the bulk of the liquid phase, resulting in major indigenous non-metallic inclusions. For truly continuous steeimaking, this is totally unacceptable, so hydrogen is introduced as the sole de- oxidant. 4. Similarly, desulphurization must not be permitted to take place within the bulk of the ferrous melt. This chemical reaction must be restricted to the gas/ liquid interface so that CaS is immediately captured by a very thin flux layer, probably a CaO-CaF₂ binary melt.

With specific reference to desulphurization just noted above, making sure that calcium sulphide indigenous inclusions do not disrupt truly continuous operation demands a totally new approach.

Fluxed launder open-channel desulphurization is introduced accordingly. For example, it may be necessary to import a relatively small amount of CaO-CaF₂ solid flux to avoid C0₂ emissions at the iron ore minesite from somewhere requiring C0₂ for enhanced oil recovery. Careful addition of molten flux to provide a layer less than say 0.2mm in thickness floating on the top surface of the turbulently flowing molten steel recirculation in the steel refining is proposed.

For open channel chemical processes taking place under turbulent flow conditions, attention is drawn to plant trials and associated research by Herbertson and Warner (J. G. Herbertson and N. A. Warner Trans. Instn. Min. & Metall. Sect. C, vol. 82, 1973, pp. C16-C20) on the prototype commercial vacuum de-zincing unit (VDZ) installed in the UK at the Swansea Vale Imperial Smelting furnace (ISF). It is relevant to know that the fundamental equation derived by Komori et al. (S. Komori, H. Ueda, F. Ogino and T. Mizushina: Intl. J. Heat Mass Transfer, vol. 25, 4, 1982, pp. 513-520) is representative of the liquid phase mass transfer coefficients actually obtained by Herbertson and Warner for the tray spiral launder of an operating VDZ plant in which zinc was distilled from the circulating lead of the ISF condenser system. This elevated temperature process has certain features, which make it closely related to open channel mass transfer involved in the present context.

Also, open-channel vacuum dehydrogenation at 20 to 50 mbar total gas pressure is introduced to facilitate dissolved hydrogen desorption into argon prior to RH vacuum degassing, which not only removes hydrogen to say 1.5 ppm or even lower but also provides the driving force for the steel refining melt circulation loop.

Returning next to the all-important challenge of energy conservation, current continuous casting consumes far too much thermal energy by employing water cooling. The latent heat of solidification and subsequent sensible heat on cooling must be captured to enhance the energy efficiency of steelmaking. In the present proposal heat is transferred radiatively from both the steel and the slag, independently. In the relatively short term, so-called tracking would appear to make natural gas thermal decomposition the preferred route to provide the hydrogen reductant so essential for zero gas emission steelmaking. A significant fraction of the heat required to thermally decompose the natural gas is provided by the counter-current radiant heat transfer just referred to in a tunnel kiln. The iron ore charge (fines or lump ore as mined) is preheated to at least 1100°C by the same mechanism.

Molten steel of the required chemical composition and cleanliness is continuously siphoned off from the recirculating open-channel stream of refined molten steel to maintain automatically steady state melt depths throughout the refining melt circulation loop. The siphoning is designed to take place under laminar flow conditions not conducive to exogenous non-metallic inclusions being incorporated into the final steel product from the fused alumina lining in this critically important region. Under steady-state operation, the melt depth adjusts itself to that commensurate with exit flow through a number of pairs of impinging jets, which serve the purpose of ultimately distributing droplets of molten steel virtually uniformly across the whole width of a moving very thin steel sheet, on which steel solidification takes place to form a homogeneous slab incorporating the initial very thin steel sheet floating on a well insulated molten lead bath. Effectively, the only external heat transferred from this system goes usefully by radiation and natural convection to contribute to preheating the carbon pellets directly involved in thermal decomposition of natural gas, the initial source of the hydrogen so essential to the zero gas emission process. Energy recovery from slag formed in the first melt circulation loop follows an analogous procedure, both sub-processes within a horizontal, straight-line well insulated tunnel kiln employing heat resistant conveyor belts, travelling grates or apron conveyors and the like moving adjacent to each other to effect countercurrent heat transfer, principally by thermal radiation.

There are various options for supplying heat to the endothermic methane decomposition process. In one approach, the process heat is introduced to the reactor by externally heated catalyst particles similar to fluid catalytic cracking or fluid coking processes widely used in oil refineries. The process employs two inter-connected fluid-solid vessels: a reactor and a heater with catalyst particles circulating between the vessels in a fluidized state. In the present invention, the specific objective is to eliminate C0₂ emission totally. Accordingly, attention is directed towards supplying the necessary heat by contacting natural gas with preheated pellets of carbon in a moving packed bed in an analogous fashion to the well-established Mond Nickel Process. In this process nickel carbonyl gas is contacted with preheated nickel metal spherical pellets on which the thermal decomposition takes place and the newly deposited metal progressively increases the diameter of individual pellets within a downward moving contiguous packed bed. Gaseous nickel carbonyl is introduced at the bottom and in a single pass is decomposed heterogeneously to nickel metal and carbon monoxide gas. Nickel pellets are continuously recirculated extemally from the bottom of the bed back to the top using an enclosed bucket elevator system. In the Mond Process, all thermal demands are supplied by heating the nickel pellets en-route to the top of the "Decomposer". Oversize pellets are screened out continuously and constitute the refined nickel metal product.

The approach outlined in the previous paragraph, involving gas/solid contacting in a moving bed of pellets is still the preferred approach, but now to reduce gas phase pressure drop and avoid fluidization of the solid carbon particles or pellets the gas must be induced to flow radially across the gravity flow of the moving pellet bed. In some respects, this is similar to the pellet bed heater introduced by

Stevanovic (D. Stevanovic and K. Brotzmann: "Pebble-Heater Technology in Metallurgy", Metalurgija, 10 (1), 2004, pp. 19-36), but in the reference case, pebble pellet beds were fixed in position so that regenerative heat transfer, involving cyclic use of two static beds, could be achieved between outlet gas from a combustion chamber or metallurgical reactor and input gas requiring preheating.

The introduction of radial gas flow rather than axial gas flow to a moving bed Mond-type pellet decomposer for continuous methane decomposition is believed to be novel and provides the ideal method for thermally decomposing natural gas to hydrogen and elemental carbon. Carbon pellet recirculation on a travelling grate or heat resistant conveyor provides the mechanism to permit the required radiative heat input to be added to the carbon pellets before re-entering the radial gas flow pebble reactor or decomposer, operating typically at around 1100°C. The whole concept is aided by the availability of metal and slag at temperatures around 1540C or slightly higher in the new continuous steelmaking process.

Hydraulic fracturing of shale (fracking) means that natural gas is going to become increasingly more available and price reductions are almost inevitable. Accordingly, the focus of attention in the near term must be on a scheme for heating the carbon pellets involved in methane thermal decomposition to ensure truly zero gas emissions by using combustion of natural gas-derived hydrogen to provide the thermal energy required rather than incurring energy losses in power generation to supply electrical conductive heating of a packed bed of carbon pellets, which may become the way forward in the longer term with introduction of a low carbon economy.

Therefore in the near term, additional natural gas is used to provide additional hydrogen to supply the thermal energy demand for methane thermal decomposition. Again a regenerative scheme employing its own pebble pre-heater system may be the preferred option. Combustion of the required amount of hydrogen in association with the gas off-take, principally water vapour from the post-combustion zone, is proposed. The requisite amount of oxygen required for safe combustion drives the gas ejector to enable a large recirculation of water vapour as the diluent in a closed circuit, completely analogous to the scheme outlined for the post-combustion zone of the melt circulation loop.

Preliminary calculations have been undertaken to identify the circulation rate of alumina pellets in a packed bed to cool the combustion gases. The heated alumina pellets enter a channel adjacent to a channel containing the circulating carbon pellets immediately before the carbon pellets are returned to the methane thermal decomposition. The heat transfer involved in this latter step is entirely radiative. The alumina pellets will supply the additional heat, over and above, the radiative heat transfer contributed by the product steel and the fluxed reduction zone slag as these individual streams are conveyed independently in a travelling grate arrangement inside a relatively long tunnel kiln. This will ensure efficient heat recovery.

To avoid potential problems with lime dissolution into the slag in the reduction zone, a calcium aluminate flux (available commercially) is the preferred additive to the iron ore charge to the melt circulation loop. Importing such flux material to the mine site avoids carbon dioxide emissions, which can be better accommodated elsewhere at a carbon capture and storage hub (CCS) rather than at the mine site itself. Attempts to introduce hydrogen-based continuous steelmaking will be thwarted unless special steps are taken to alleviate the potential problem of solids deposition within pebble bed regenerative heat exchanger systems and the like. Whereas solids deposition on the moving packed bed immediately above the tuyere zone and in the upper reaches of a traditional blast furnace undoubtedly occurs to a major extent, the problem is alleviated by continuous movement of the solid charge downwards followed by high intensity melting of all solids in the tuyeres region. Unfortunately such action is not available in association with pebble bed degenerative heaters vital to achieve the required level of heat recovery from reduction off-gas, which is mandatory if new hydrogen-based continuous steelmaking is ever to achieve successful commercialization.

There are several key reasons why what may appear as a mundane issue is believed to warrant the introduction of an innovative response to the potential problem at the process design stage so that disruption is not permitted to occur in full-scale plant operation. Without successful implementation of a further novel energy efficient recovery scheme addressed in this patent, it is highly unlikely that the potential global benefits of truly zero gas emission continuous steelmaking will ever come to fruition.

(i) When hydrogen is used as the principal or sole reductant of molten ferrous oxide, for example, the off-gas is composed of the hydrogen water vapour binary system, which has the thermodynamic characteristic, that as the gas phase is cooled for heat recovery purposes, the ratio of hydrogen to water vapour does not change. This permits the re-use of any unused hydrogen after separating out the water vapour by straightforward condensation. With continuous charging of solid iron oxide to a reactor, hydrogen is consumed at a rate which is proportional to the solid charge rate. The maximum amount of hydrogen consumption in a continuous steelmaking reactor, for example, is determined by the reducible oxide content of the solid charge and any excess hydrogen added merely passes through the reactor and may be recycled back to the reactor as pure hydrogen once the water vapour is condensed out. On the other hand, this is clearly not feasible with carbon monoxide/carbon dioxide gas mixtures, because of the thermal instability of CO as the temperature is reduced.

Recycling of cooled hydrogen, however, attracts a thermal penalty in that the sensible heat in cooling is lost unless provision is made for efficient heat recovery by recuperative or regenerative heat exchange. Obviously, the thermal penalty associated with recycling a large excess of hydrogen places a limitation on such action, unless a very efficient heat exchange system is utilized. Such is the rationale behind the preferred choice of a radial gas phase flow pebble regenerative heat exchange system in new technology being proposed by the inventor in related patent applications on continuous steelmaking.

Stevanovic and Brotzmann in their published paper (D. Stevanovic and K. Brotzmann: etalurgija, 10 (1), 2004, 19-36) claim that it has been demonstrated that it is feasible to approach a temperature differential of 20 C with their pebble heater technology in metallurgy. The Pebble Heater referred to in their paper consists of a cylindrical vessel with two permeable grids and a pebble bed between them. In their arrangement the inner grid, the so-called "hot grid", is composed of alumina bricks with "honeycomb segments" set in the prevent movement to pebbles. They claim that with improved honeycomb segment quality that the expectation is that even higher temperatures than the already proven 1500°C limit referred to is feasible, possibly up to 1700°C. The present invention assumes that progress has been made so that temperatures of 1540°C, that is just above the iron melt point, are now acceptable, so that anticipated solid iron deposition within the pebble bed can be dealt with by the innovative approach now being proposed.

(ii) Generic melt circulation technology described in a review article (N. A. Warner JOM, vol. 60 (10), October 2008, 14-22) permits countercurrent contacting of a layer of liquid iron oxide with high velocity preheated hydrogen without surface instability giving rise to entrainment of molten oxidic material, provided due attention is given to what is known by those skilled in the art as Kelvin-Helmholtz instability criteria. Under these circumstances, the reduction off-gas will not be contaminated with liquid or solid phases. This is provided the liquid oxide layer is first established in advance of any contact with the flow of reducing gas. However, the gas phase will be contaminated to some extent by iron vapour due to metallic iron being formed at the gas/liquid interface.

According to theoretical analysis for the likely conditions involved in melt circulation continuous steelmaking, the process will proceed kinetically under predominately gaseous molecular diffusion control. However, the oxidic melt is itself likely to be virtually iron-saturated at close to the interface, so it would be prudent to assume the worst case scenario, implying that the off-gas is saturated with iron vapour. At 1540°C, for example, the iron partial pressure at equilibrium with an interface at unit activity is a mere 3.56 x 10 ^s, but even this will deposit about 1 tonne of metallic iron per day within the pebble regenerator for a plant producing a nominal 2 million tonnes per annum of steel product.

(iii) According to the paper by Stevanovic and Brotzmann "for high temperature applications, the pebble bed consists of alumina pebbles. They can sustain extremely high temperature cycling (even over 400 K) without any damage." However, from personal experience in laboratory experiments (N. A Warner Stream Break-up in Steel Vacuum Degassing, Jnl Iron and Steel Inst., 208, 44, 1969, 44-49) mullite was vastly superior to recrystallized alumina for containing molten iron at temperatures near the melting point and conducting experiments without thermal shock damage. Even if mullite pebbles were used, it is improbable that a sudden change of temperature from say 200 to 1540°C could be sustained without thermal shock, so radial gas flow reversal is unlikely to be a realistic option for melting solid iron deposits within the pebble beds. Accordingly, it is necessary to specify at least one off-line pebble heater, so that overall temperature raising of the whole bed to above the melting point of iron (1538°C) can be conducted cautiously within an argon protective atmosphere.

Solid iron will melt and form globules or liquid metal with run-off of excess melt from the bed to a containment vessel so that ail that is left within the bed is the so-called "static holdup" of molten iron. The applicant has conducted numerous experiments associated with non-wetting irrigation of packed bed, which demonstrate relatively small amounts of static holdup are attained, provided the system is truly non-wetting. This certainly applies to irrigation of solid spheres in packed beds irrigated with molten lead, mercury and with water under non-wetting conditions, when the liquid flow is halted. By analogy, the holdup of iron droplets within the pebble beds of regenerative contactors will induce only a relatively small increase in gas phase pressure drop in addition to that associated with the same bed that has not been irrigated with a non-wetting liquid phase. This phenomenon is the basis of the novel procedure now proposed to ensure that plant shutdown is avoided by a routine maintenance spread over the whole pebble heater system, which now should be specified as having a minimum of three pebble bed heaters, two in operation at any one time and one in standby mode or undergoing thermal treatment.

It is now proposed that a degree of stop/start pressure surging in the flow of argon to the pebble packed bed, once its temperature is above the iron melting point, will enhance the establishment of low static holdup in this non-wetting system. Accordingly, gas phase pressure surging, probably involving intermittent multiple stop/start conditions, should be implemented in the off-line third pebble heater undergoing thermal treatment, as this should be highly beneficial in the overall continuous steelmaking process, because of the combination of lower gas phase pressure drop and greater length of time between necessary cyclic removals of accumulated iron from within the pebble packed beds.

Plant operation will establish when the above initial part of the thermal treatment has been successfully accomplished. This will be obvious, if the molten iron run-off comes to an end. The next step in the thermal treatment is vitally important in order to preclude thermal shock damage to the refractory linings as well as to the ceramic pebbles themselves.

A relatively short reheating process must next be included into the routine once the first stage steady state process is concluded and the preheater temperature lowered to an acceptable extent, which can be evaluated by detailed unsteady state heat flow numerical calculations along with operating experience described in the paper by Stevanovic and Brotzmann. The prime objective is to re-establish an acceptable lateral temperature gradient within the pebble preheater by radiative heat input from a graphite or silicon carbide electrode system placed temporarily inside the central vertical free volume to establish heat flow, which simulates that which occurs normally when the pebble bed heater being thermally treated returns to on-line operation.

The newly proposed technology effectively does away with the need to incorporate a traditional steelmaking decarburisation step in the overall process. Instead virgin iron ore materials are reduced to molten iron without carbon contamination. The relatively small amount of dissolved carbon required to bring the molten iron up to the desired carbon level specification for a particular steel product range can be added to the in-line refined molten iron once it has left the principal melt circulation loop. At the same time, other minor alloying elements can also be added, probably in association with continuous deoxidation and desulphurization preceding continuous casting. The new continuous casting process employs innovative technology, including the use of a reduced inert gas pressure or vacuum siphonic tundish containing a number of pairs of self-impinging jets of molten steel to distribute the molten steel product across the full width of the moving horizontal layer of molten steel product undergoing solidification by radiation to the adjacent moving horizontal layer of carbon pellets required for thermal decomposition of natural gas.

The back-end of the current invention involving a radically new approach to continuous casting of molten metals with a reduced pressure or vacuum siphonic tundish is believed to be totally novel but it could perhaps be argued that what then follows involving a pool of molten lead is analogous to

Pilkington's float glass process, in which a layer of molten glass is admitted to a static pool of molten tin under non-wetting conditions. There is very little solubility of lead in solid steel but this is increased somewhat in liquid steel above the melting point. Accordingly, the present invention employs a very thin sheet of solid steel floating on the pool of molten lead on which to distribute the molten steel product initially as droplets from the impinging jet tundish. These steel droplets coalesce and form a continuous layer of molten steel on the moving foundation thin layer of solid steel, so there is no contact in the present invention between liquid steel and molten lead.

It should also be stressed that when the molten steel product first contacts the moving foundation thin sheet of solid steel or substrate, heat is transferred preferentially at the outer edge region so that a rim of solid steel will form initially which stops liquid lead from flowing into the zone where some of product steel is itself still molten. Alternatively, steps may need to be taken to weld a rim of steel onto both edges of the foundation sheet upstream of the arrival of the steel droplets from the impinging jet tundish. The latter is the safer option as any tendency for lead to become physically occluded within the solidifying structure of solid steel is thereby totally eliminated.

In the initial conceptual design the pairs of impinging jet large orifices within the reduced pressure siphonic tundish were arranged so that effectively the whole width of the foundation substrate was irrigated with a sheet of molten steel. The spray pattern of this sheet was a vertical plane normal to the plane of the jets. However, detailed analysis revealed that heat transfer intensity or heat flux was high enough to threaten the solid state stability of the advancing very thin sheet, when it first met the irrigating droplets of molten falling from above. Accordingly in the preferred embodiment, the jet plane is now at right angles to the previous arrangement. The net result is the incoming heat flux is considerably reduced and no longer threatens to melt the substrate solid steel floating on a well insulated bath of molten lead at say nominally 900°C. The flowing lead beneath the steel sheet is force- circulated at a very high rate to effect intense heat removal from the underneath side of the floating steel sheet so that it does not melt as the spray of liquid steel droplets from the impinging jet tundish contacts the top surface.

Under these revised conditions, a pool of molten steel contained within the in-situ welded edge constraints already referred to and floating on molten lead, is maintained at the steel melting temperature as it moves downstream, whist the solid thickness increases at a rate determined by the rate of heat transfer to its surroundings and especially to the coolant molten lead flowing immediately beneath. Provided the whole containment assembly is thermally well insulated, this provides a highly efficient means for reheating by radiation and natural convention the solid iron ore charge and flux materials being transported horizontally and counter currently to the thin steel slab within the well- insulated tunnel kiln.

The steel product passes from entry to exit end of the pool of molten lead drawn by the action of withdrawal pinch rolls within a protective argon atmosphere. The solid wide thin slab then continues on its way running parallel to separate conveyed strands of solidified slag and oversize by-product carbon pellets all transporting heat by radiation and convection counter-currently to the iron ore charge (fines or lump as mined ) and flux materials also being conveyed horizontally in the well insulated tunnel kiln. After heat recovery to a specified extent, a shear or torch-cut assembly is provided in preparation of the steel product towards transportation to world markets

If steady state continuous operation is interrupted for a long period of time, the various assemblies associated with continuous casting must be preheated and a fiat dummy plate guided horizontally into the moid section and aligned with the pinch rollers. A system needs to be incorporated in order to provide a continuous supply of thin steel sheet (1 to 2mm in thickness ) of appropriate width; For example, 2 million tonnes per year of product may need somewhere in the region of Sm width.

With regard to the innovative proposal to employ self-impinging jets in continuous casting of steel, reference is now made to research undertaken by Imperial Smelting Corporation in connection with pyre-metallurgical primary production of zinc metal. For self-impinging jets as described in the paper by Gammon (M. Gammon: Process Engineering of Pyrometallurgy, Institution of Mining and Metall., 1974, pp. 24-32) relating to potential replacement of mechanical rotors in the condenser system of a zinc- lead blast furnace (ISF) with an impinging jet spray of molten lead, 900 mm liquid head was used for jets 25 mm in diameter with a jet angle of 30° to the horizontal. The lead splash test rig was run continuously for five days. There was no need to clean the jets and examination at the end of the run showed little sign of wear of the 50 mm thick steel plate containing the jet apertures, even though the zinc content of the lead remained between 1.0 and 1.5% for the duration of the run, in which the lead temperature was maintained nominally at 500°C. At 600 mm head the measured flow rate of lead from a pair of 25 mm jets was found to be 83 tonnes hr, which was the same volumetric flow rate as with the equivalent water system as predicted theoretically from the Bernoulli equation.

More recently, Broadbent and Warner (CP. Broadbent and N.A. Warner Trens.lnstn.Min.Metall. (Sect. C: Mineral Process. Extr. Metall.), 1984 vol. 93, pp. C130-133) showed that satisfactory droplet formation was obtainable using a self-impinging jet disintegration system from molten slags having viscosities between 0.3 and 2.0 N s m² which is three orders of magnitude greater than for the water or molten lead of earlier tests. When two jets of liquid impinge, a sheet is formed in the plane

perpendicular to their direction. The sheet becomes fan-shaped and is surrounded by a rim of thicker material. Waves develop on the sheet which is thinning as it grows and the lower part of the sheet disintegrates into ligaments which further break up into droplets. Droplet size particularly depends upon impact velocity as well as the physical properties of the liquid. In the initial tests, suitable operation was obtained with 6 mm diameter passages angled at 90° to each other under a head of 2.5 m of slag. In the laboratory, this head was obtained by pressurising the crucible with nitrogen. The majority of the volume of the liquid was contained in droplets possessing diameters close to the recorded maximum and there was no apparent variance in droplet size distribution across the width of the sheet. Model experiments with Glycerol Water mixtures showed that maximum droplet size reduced with reduction in viscosity (4.5 mm at 0.69 N s m² and 3.5 mm at 0.303 N s/m²). As expected reduction in the pressurising head, which reduced the jet velocity and hence the Weber number, led to an increase in droplet diameter.

Self-impinging jets of molten slag disintegrate successfully from the resultant sheet into spherical particles. The experiments showed that droplets of about 3 mm mean particle size can be obtained in this way. As disintegration occurs without appreciable cooling, it was suggested that this would be an appropriate method for the generation of slag droplets prior to heat recovery in a quenching medium, such as a gas fluidized bed. Subsequently a major research project was undertaken. Equipment was constructed to incorporate a chamber that afforded an uninterrupted drop of 4 m beneath the point of impingement to enable fundamental heat-transfer data to be collected and further work on industrial slags was planned to establish the effects of scale-up and to investigate the collection of partially solidified droplets and other operating parameters relevant to an overall scheme for the recovery of thermal energy. A summary from papers presented by Peirce and Warner (T.J. Peirce and N.A.

Warner Applied Energy Research, Proceedings of the Institute of Energy Conference, Swansea, 1989, pp 33-45) and Warner et al.( N.A. Warner, T.J. Peirce, J.W. Armitage, J.S.M. Botterill and Y. Sergeev: Tenth International Heat Transfer Conference, 1994, Paper No. 64) give essential background to the utilization of the analogous system in the current invention for uniformly distributing molten steel across a large width, say in the region of 5 to 10 m, also in continuous casting of molten steel. This highlights the fact that although the present proposal in this invention is conceptual, considerable related testwork has already been undertaken.

The whole process described could readily be fully automated and in the mine of the future be controlled along with the iron ore mining operation itself at a convenient central location, such as Perth, for example, to support autonomous mining in the Pilbara region further north in Western Australia.

An embodiment of the present invention will now be described, by way of example only, with references to the accompanying diagrams, in which:-

Fig. 1 is a schematic sectional plan view of a prior art single melt circulation loop employing forced circulation of molten iron through a straight reduction arm and a parallel straight post combustion arm interconnected at one end by a means for melt circulation such as a gas-lift device and an overflow or siphon at the other. Fig. 2 is a schematic half sectional elevation view across the width of one of the "swimming pool" reactors in the prior art, showing the solid shell of iron for containment of the circulating molten iron, the means for protecting exposed solid iron surfaces from oxidation or sulphidation, the general configuration of the steam boiler tubes for power generation, the arrangements for sealing the liftable top enclosure, the method for accommodating thermal expansion of the solid shell and the general features of the basal assembly.

Fig. 3 returns to the current invention and is a schematic representation of the key features of a newly proposed hydrogen-based single melt circulation loop in an essentially annular configuration employing forced circulation of molten iron through a reduction zone and then a post-combustion zone with continuous overflow of a molten iron product.

Fig. 4 is a schematic representation of the key features of the newly proposed hydrogen-based single melt circulation loop in an essentially annular configuration employing forced circulation of molten iron through a reduction zone and then a post-combustion zone with continuous overflow of molten iron product stream, which then undergoes continuous in-line refining based on established steelmaking technology.

Fig. 5 is a schematic plan view of an annular melt circulation loop, which highlights the essential differences between state-of-the-art and mandatory changes introduced on grounds of safety and energy conservation.

Fig. 6 is a cross-sectional elevation taken across the reduction zone, which very largely parallels state-of-the-art with the very important exception that reduction off-gas is not permitted to freely access the post-combustion zone by a conventional gas cross-over or other means because of the risk of accidental explosion.

Fig. 7 is a cross-sectional elevation of the post-combustion zone, which again may appear to follow prior art, but shows the alumina hydrogen header for forcefully supplying clean pre-heated hydrogen to multiple small diameter alumina pipes/nozzles injecting turbulently flowing axi-symmetric jets of clean pre-heated hydrogen into the surrounding gas phase.

Fig. 8 is a sectional elevation taken across a diameter of the annular configuration employing an inline arrangement of a reduction zone followed by a post-combustion zone.

Fig. 9 is a schematic sectional elevation of a natural gas thermal decomposer, in which methane is heterogeneously decomposed to elemental carbon and hydrogen gas on the surface of carbon pellets in a moving packed bed with radial flow across the annular moving bed typically with a thickness of 1 to1.5 m of carbon pellets moving by gravity downwards.

Fig. 10 is a schematic sectional elevation of prior art pebble heater technology in metallurgy, illustrating two regenerative heat exchangers with radial gas flow which has demonstrated as little as 20 K between heating and heated gas with hot gas temperatures up to 1500°C.

Fig. 11 illustrates the proposed amended construction and arrangement of a pair of pebble bed heaters, whilst in continuous cyclic operation, during which time iron deposition within the pebble bed heaters is inevitable.

Fig. 12 is a schematic sectional elevation of the third pebble bed heater, which is off-line in the thermal treatment mode to remove iron deposition in order to re-establish low gas phase pressure drop operation, when it is returned to on-line service.

Fig. 13 is a schematic elevation for safely combusting clean pre-heated hydrogen with oxygen and transferring the heat to a moving bed of recirculated alumina pellets subsequently to provide the major energy input for thermal decomposition of natural gas.

Fig. 14 is a schematic plan view showing the general arrangement of the two separate melt circulation loops in relation to the other major plant items.

Fig. 15 shows both plan and sectional elevation of the reduced pressure or vacuum siphonic tundish for impinging jet continuous casting of steel product with cross-hatching omitted for added clarity. Fig. 16 is a detailed sectional elevation of the siphonic tundish.

Fig. 17 is a schematic elevation showing the both zones of the principal melt circulation loop in section and other major plant items, which together illustrate the new overall continuous steelmaking process.

Referring now to Fig. 1, the plant comprises a single melt circulation loop, a charge reduction arm 1 and a post-combustion arm 2 interconnected together by the gas-lift type pump 3 and a siphon 4 or other appropriate overflow device at the other end. Preheated iron ore fines 5 or other iron ore materials are added to the molten iron carrier medium 6 which flows around the closed loop melt circulation reactor. At the remote end of the charge reduction arm relative to the solid charge location, the molten iron carrier material and its associated now iron oxide-depleted melt layer overflow weir 7 into a phase disengagement region 8 from which molten slag 9 is either continuously or intermittently tapped. The preheated iron ore fines 5 are added continuously to the top surface of the molten iron carrier material via an appropriate distribution means 10 and are almost immediately chemically converted into liquid ferrous oxide containing normally a relatively minor amount of oxide gangue impurities. As the layer of iron oxide melt progresses along the reduction arm 1 floating on the molten iron carrier medium 6, it is progressively reduced and the molten iron thus formed joins the molten iron carrier medium and, at the same time, the equivalent amount of molten iron is normally overflown or otherwise removed continuously from the melt circulation loop at any appropriate location, in this case via a product siphon 11.

Preheated natural gas 12 or other manufactured gas, probably based on coal, petroleum coke or biomass material is added to the reduction arm 6 at the remote end relative to the solid charge location and is forced to flow at high velocity through a confined space having a gap of only a few centimeters between the top surface of the oxide melt and the roof of the reduction arm. This gap would normally measure 5 - 6 cm in height but perhaps somewhat larger depending on gas phase pressure drop considerations or even smaller if appropriate control means were developed and integrated into the gas flow circuit. The reducing gas not consumed along with products of reaction are passed through a crossover gas duct 13 into the post-combustion arm 2 either at a single location or more probably at various points along the post-combustion arm where preheated oxygen 14 is added again via a number of ports, such as 15, distributed along the post-combustion arm. Finally, the fully combusted off-gas is ducted at 16 to the solid charge preheater.

Referring now to Fig. 2, the liquid metal pool 19a is representative of a typical hearth arrangement. The solid iron shell established under steady operating conditions is shown as 20. A purge gas header 21 supplies ultra-low sulphur/non-oxidising to iron gas to a pipe 22, one of a number of such pipes attached to the header at various points along the length of the pool 1 a. This purge gas on being admitted to the gas-tight enclosure, comprised of top-hat enclosure 23 and basal enclosure (hearth) 24, provides an inert gas atmosphere throughout the whole volume of the enclosure except for that containing reactive gas in the gas space 25 above the molten iron. Accordingly, protection is provided to the outer surfaces of the solid shell 20, steam boiler tubes 26 and ancillaries such as the heat resisting alloy components 27, which provide skid mounting at the base of the solid shell 20, and a sideways movement trolley 28 extending the length of the pool 19a, which is actuated so that it moves backwards and forwards as the solid shell 20 expands or contracts in width. The latter features are to combat stresses from thermal expansion and are designed to allow freedom of movement of the solid shell 20, particularly when totally solid at start-up or after prolonged shut-down. Both the skids and the trolley system are used in conjunction with heat resistant alloy plates to control creep and distortion of the solid shell 20 during prolonged operation at elevated temperature. It will be noted that the alloy components comprising the skid arrangement form an edge contact with the underside of the metal shells 20 and provide a small spacing between the boiler tubes 26 and the underside of the shells 20, so that heat transfer to the tubes 26 is radiative.

The top-hat enclosure is provided with a skirt 29, which is immersed in a channel 30 containing fusible alloy such as lead-bismuth eutectic, which forms a continuous seal around the perimeter. The channel or trough 30 containing the fusible alloy is attached to the basal enclosure 24 and is heated at all times by electrical conductive heating so that the top-hat enclosure 23 is free to thermally expand or contract, whilst always maintaining a leak-proof gas seal.

Inside the top-hat enclosure 23 a composite lining 23a of low-thermal mass insulating materials provides lightweight insulation and permits rapid heating after shut-down without fear of refractory damage. On the highest temperature faces, high purity alumina fibrous board currently commercially available or microporous materials currently under development are used.

The purge gas enters the gas space 25 above the molten iron through a small clearance passageway 31, bounded by the top surface of the solid shell 20 on one side and on the other by several layers of ceramic fibre board 32 or comparable material, which is profiled to deliver a shroud of protective gas to the solid iron areas vulnerable to oxidation and possible sulphidation immediately above the molten metal surface. To achieve this, the boards 32 project a short distance into the reactive gas space 25, as shown schematically at 33. The purge gas velocity is controlled by varying the pressure upstream to the header 21 to preclude back diffusion of reactive gases to the exposed high temperature surfaces of solid iron.

For routine inspection or maintenance, means for lifting the top-hat enclosure 23, which is effectively a moveable lid on the "swimming pool" reactor 19a, must be included. The arrows shown as 34 and 35 signify such an arrangement, where the options include hydraulic hoists and gantry craneage. Because of the use of lightweight insulating materials without conventional refractory arches and brickwork, the lifting requirements for handling a one-piece fabricated top-hat enclosure 23 for say a 60 m length "swimming pool" reactor are not excessive and well within the bounds of commercial practice, particularly that associated with dry docking of ships and bulk loading of whole barge cargoes.

Although not absolutely essential, it would be convenient to provide clear areas adjacent to the "swimming pools" so that individual top-hat enclosures can be parked nearby.

The fabricated steel basal enclosure 24 encases a refractory concrete or firebrick base 36 and structural steel members 37, which are the principal load bearing members for the whole "swimming pool" reactor. The structural supports are ventilated, possibly forcibly, to ensure the reinforced concrete floor 38 is not over-heated and not over-loaded. Because the solid shells are typically around 1 m in thickness, it is highly unlikely that a break-out of liquid metal should occur, but in the event, the large thermal mass of the firebrick lining can be regarded as a safety lining.

Referring now to Fig. 5, this is a schematic representation of the key features of a single melt circulation loop employing forced circulation of molten iron 6 through a reduction zone 1 and a post- combustion zone 2 by means of a gas-lift pumping arrangement 3 for melt circulation within an annular hearth 39 comprised of an unmelted steel shell 20 for containment of the molten iron carrier medium 6. At the remote end of the reduction zone 1, relative to the solid charge 5 location 10, the molten iron carrier medium 6 and its associated now somewhat iron oxide-depleted melt layer overflow a weir 17 into the phase disengagement region and associated pumping facility 3.

In Fig. 5, hydrogen 40 is added to the reduction zone 1 at the remote location relative to the solid charge incorporation 10 and is forced to flow at high velocity through a confined space having a small gap between the top surface of the oxidic melt and the roof or ceiling of the reduction zone. To retain operation at essentially atmospheric pressure, this gap would normally have a mean height of about 15 cm over the reduction zone 1 length. However, Kelvin-Helmholtz interfacial stability would still be maintained even if this gap was reduced to below 10 cm at the expense of a somewhat higher gas phase pressure drop. The reducing gas not consumed along with products of reaction (principally water vapour), i.e. reduction off-gas 41 then proceeds through a refractory-lined duct 42 to begin heat recovery in a pebble bed regenerative heat exchanger. The gas is next cooled in other heat recovery units eventually to around 50°C for phase separation after removal of particulates by filtration. The cleaned-up hydrogen is then eventually recycled back to the reduction zone 1 and also the post- combustion zone 2 of the melt circulation loop.

Continuing with reference to Fig. 5, in particular the means are shown for recycling hot post- combustion off-gas 43 using at least the oxygen 44 required for stoichiometric combustion as the motive fluid for a gas ejector 45 to ensure that the gas phase 46 within the full length of the combustion zone is fully back-mixed and turbulently flowing water vapour with minor oxygen content to ensure complete safe combustion of hydrogen, when it is injected into the combustion chamber via a multiplicity of axi-symmetric turbulent flowing hydrogen jets 47, which themselves entrain an appropriate excess of oxygen in association with water vapour.

Still referring to Fig. 5, slag 9 is withdrawn continuously from the phase separator (not shown) used in conjunction with the gas- lift pumping arrangement 3. The carrier iron melt continues via an overflow weir with a slightly increased liquid potential head on its recirculation in the melt reduction loop. There has to be a continuous overflow or alternative removal of product iron melt at 48, if melt depths throughout the reduction zone and elsewhere in the melt circulation loop are to be retained at a steady- state level. Once thus removed, this iron product stream then undergoes deoxidation and

desulphurization at 49 followed by slag/metal separation, to remove the newly formed slag (2) at SO. Thermal energy recovery from slag (2) removed at 50 may then be accomplished in association with slag (1) removed at 9 before ultimate disposal. Final adjustments to carbon content and minor alloy composition to meet product specification are accommodated in the finisher at 51 in advance of continuous flow of molten steel product to the tundish of a continuous caster 52.

In-line continuous desulphurization and deoxidation 49, just referred to, may be conducted using a conventional RH (Ruhrstahl Hereaus) degasser, to which reactants such as FeSi and Al and a flux material such as calcium aluminate are added as required. However, the preferred embodiment using permanent freeze linings employs a novel method involving a turbulent relatively thin layer of molten iron flowing in an open channel to which is added molten calcium aluminate flux or binary CaO/CaF2 melt. Controlled addition of molten flux is carried out so that flux layers not exceeding about 0.1mm in thickness are provided for at least one contact but probably for clean steel multiple contacting with fresh flux addition and removal at a number of locations, as the turbulent layer of molten iron undergoes progressively highly efficient desulphurization. Also an appropriate amount of carbon could be added as well to meet desired product steel specification, if ultra-low carbon steel is not the preferred end product.

It is considered appropriate to reiterate what was said earlier in this description that there are many ways of conducting the chemistry required for desulphurization and deoxidation. Accordingly, other approaches to that described may well be preferred in some cases. However, if traditional steel refining practice is adopted, downstream of desulphurizatbn and deoxidation, the molten iron depth needs to be increased to reduce the flow velocity of the product molten iron stream in order to improve phase separation of flux and other non-metallic inclusions.

Not withstanding what has just been stated, the now preferred approach for truly continuous steelmaking is to follow the route previously described to virtually eliminate non-metallic inclusions altogether This employs hydrogen as the sole de-oxidant as the molten iron product overflows from the primary melt circulation loop and is contacted counter-currently with hydrogen as it undergoes non- wetting irrigation of a packed bed of alumina spheres or other appropriate packing elements. The deoxidized melt under gravity flow then enters an electrically conductively heated second melt circulation loop for desulphurization, residual dehydrogenation and product compositional adjustment. At no stage is the steel temperature permitted to rise more than 1 or 2 degrees above the melting point throughout the steel refining loop in the interests of energy conservation with its permanent freeze lining.

Referring now to Fig. 6, the means for accommodating differential expansion incorporating as already shown in prior art depicted in Fig. 2 involving skid mounting 27 of the unmelted steel shell 20 stabilized by steam generation in the boiler tubes 26 below which is the refractory/thermal insulation 36 and steel encasement 24. The molten iron carrier medium in this figure is carbon-free, which is analogous to the carbon containing melt 6 shown in Fig. 1 of the prior art. The fibrous ceramic board 33 is analogous to that shown in the prior art in Fig. 3. In the current diagram, roof structure incorporating steel joist girders 51 to which are attached refractory elements 52 with the whole roof structure free to move to accommodate differential expansion by the pontoon support arrangement 53 associated with a low vapour pressure liquid, such as a fusible alloy, in which the support pontoons 53 float.

In Fig. 6, hydrogen which is added to the reduction arm or zone at the remote end relative to the solid charge location is forced to flow at high velocity through a confined space 54 having gap 10-15 cm centimeters between the top surface of the oxide melt 55 and the roof or ceiling of the reduction arm 56. The reducing gas not consumed along with products of reaction (principally water vapour) leaves the reduction zone and then proceeds directly through a refractory-lined duct to a pebble regenerative heat exchanger known as a "pebble heater" to recover thermal energy. After leaving the 'pebble heater" (D. Stevanovic and K. Brotzmann: "Pebble-Heater Technology in Metallurgy", Metalurgija, 10 (1), 2004, pp. 19-36) the gas is eventually cooled to around 50°C for phase separation prior to recycling hydrogen together with make-up hydrogen equivalent to that consumed in the overall process back to the reduction arm of the melt circulation loop as well as a small proportion to the post combustion zone. A recuperative shell and tube heat exchanger fabricated with austenitic stainless steel can be used to raise the recycled hydrogen temperature to around 725°C prior to entering the pebble heater arrangement and being preheated to about 1500°C.

Now referring to Fig. 7, the various items for accommodating differential expansion and for stabilizing the unmelted steel shell by steam generation in boiler tubes 26 are the same as already referred to in Fig. 6. Fig. 7 is a cross-sectional elevation of the post-combustion zone, which again may appear to follow prior art, but not shown in the figure, is the external gas ejector powered by compressed oxygen equivalent to that necessary for eventual complete combustion of hydrogen downstream within the fully back-mixed turbulently flowing gas phase in the post-combustion zone which is principally water vapour diluting a relatively minor oxygen component. What is shown in the figure is the alumina hydrogen header 57 for forcefully supplying clean pre-heated hydrogen to multiple small diameter alumina pipes/nozzles 58 injecting turbulently flowing axi-symmetric jets of clean pre-heated hydrogen into the surrounding gas phase in order to entrain enough oxygen and its associated water vapour to ensure entirely safe combustion.

Fig. 7 also stresses the need for a relatively large gas freeboard 59 above the melt surface to facilitate gas phase radiative heat transfer from the very hot post-combustion gases, principally water vapour containing less than 0.5% hydrogen, to the circulating molten iron. The molten iron is covered by a very thin layer of flux 60 (only a fraction of an mm in thickness and thus not shown). This is to increase the emissivity as well as providing protection from melt oxidation by direct gaseous interaction with the molten iron surface. Complete back-mixing or alternatively off-gas recycling of the gas phase above the melt surface is vital to eliminate any risk of explosion. To reiterate what was stated earlier in the description, important steps must be taken in the design to ensure that hydrogen and oxygen levels do not exceed safe limits in the bulk gas phase at all locations along the length of the combustion chamber. A proportion of the recycled hydrogen is forced to enter the combustion chamber to a maximum of about 1500°C temperature via a large number of relatively small diameter alumina tubes or nozzles 58 so that turbulent jets issue forth into the bulk gas phase 59 and entrain the appropriate amount of oxygen, associated with the major gas phase component of water vapour flowing turbulently lengthwise within the entire gas free board 59 above the circulating molten iron carrier medium to safely combust the hydrogen.

The special safety feature introduced in this invention incorporating encapsulation of whole reactor zones with low pressure steam 61 is also shown in Fig. 7.

Still referring to the post-combustion zone shown in Fig. 7, the very useful feature of in-line dephosphorisation may be accomplished by provision of an extended surface layer of liquid flux 60, which floats on the molten iron canier medium with the spent flux (not shown) and molten iron overflowing together across a weir into a phase disengagement zone.

All the melt circulation loop thermal requirements can be provided by oxygen combustion of only a portion of the hydrogen in the reducing gases originally emanating from the reduction arm of the melt circulation loop after its thermal energy is efficiently recovered and undergone phase separation to recover water and then filtered to entirely remove any particulates. Addition of make-up hydrogen to account for that consumed in the process then takes place. The combustion energy released in the post combustion zone is absorbed by the molten iron carrier medium. Sensible heat is then transported via melt circulation to endothermic reaction sites by highly turbulent convection in the liquid metal with effectively zero thermal resistance. This is then followed by heat conduction across the relatively thin oxidic melt layer 55 in Fig. 6 to the gas/liquid reaction interface, where the reduction kinetics exhibit almost exclusive gas phase mass transfer control. To reiterate the immediate consequence of this kinetic assessment is total gas pressures above atmospheric pressure have little effect on the rate of hydrogen reduction of the liquid wustite layer and therefore on safety grounds should be avoided.

Figure 8 is sectional elevation taken right across the annular hearth diameter from centre line 62 showing both the reduction zone 1 and the PC arm 2. Initially the fused ore layer 55 will be in the region of 10 mm in thickness, progressing ultimately around 1 mm before overflowing with the molten iron carrier medium into the slag separation zone. As discussed previously, in Figs. 6 and 7, features are shown to accommodate differential linear thermal expansion of the unmelted steel shells comprising the hearth and side walls and containing molten iron designed to be maintained close to the melting point at the hot face for melt containment. For a nominal overall length of say 80 m the total lateral movement works out to be approximately 1-5 m in going from cold to 1530°C, which is indicative of the necessity to provide a system that can accommodate differential thermal expansion between the steel shell and the alumina refractory lining the ceiling and roof.

Also shown for both the reduction zone and post-combustion zone in Fig. 8, is an Elastomer ^*0° ring- type seal 63, which establishes a gas-tight seal between the vessel internal wall and the removable lid covering the suspended refractory flat roof system, comprising steel joist girders, thermal insulation and flat refractory roof along the length of the containment vessel. This gas sealing arrangement is facilitated by a small differential in total pressure maintained between the slightly higher gas pressure in the reactor containment vessel relative to the gas pressure inside the melt circulation loop itself, which causes a predetermined force to act downwards on the Elastomer "0° ring-type seal. The removable lid covers the whole extent of the iron oxide gaseous reduction zone and thus affords the means for ready access to the gaseous reduction unit. A similar arrangement applies to the post-combustion zone.

A skid-mounted system permitting thermal expansion or contraction of the unmelted steel shell and a row of steam raising boiler tubes to stabilise the unmelted steel shell at a prescribed steady state thickness are shown schematically in Fig. 8 for both the reduction and post-combustion zones. These boiler tubes are fixed relative to the outer enclosure of the furnaces in accordance with a computer- aided design to ensure dimensional stability of pre-determined profiles, commensurate with the particular functional requirements at various locations throughout the melt circulation loop. Gas space for boiler tubes to facilitate the establishment of the steady state thickness of the unmelted steel shell is adjacent to the cooler sides of the steel shells so that radiative heat transfer is permitted to take place throughout the entire extent of the unmelted steel shell system.

Referring now to Fig. 9, this highlights the new approach to achieve zero gas emission continuous steelmaking employing thermal decomposition of natural gas to elemental carbon and hydrogen. Natural gas, preferably preheated to about 350°C, is admitted to the inlet manifold 64 and then flows radially through the annular moving packed bed of carbon pellets 65, typically in the region of 1 to 10 mm in diameter at temperatures in excess of 1200°C and preferably about 1600 K (1227°C) to ensure virtually complete decomposition heterogeneously on the surfaces of the carbon pellets. The hydrogen produced leaves the annular moving bed of carbon pellets through a silicon carbide internal mesh arrangement 3. The hydrogen produced by thermal decomposition leaves the decomposer through the gas off-take 4. The carbon pellets are recirculated through the natural gas decomposer 68 using a bucket elevator system 69, which discharges pellets to be preheated externally. As pellets continually increase in diameter, the oversize are screened off at 70 and are removed as by-product carbon 71.

The various options relating to what happens next to the carbon pellets cover a wide range of scenarios. For enhanced thermal efficiency, it is highly desirable that the carbon is utilized in associated chemical processing or for electricity generation. However, if this is not feasible, the carbon pellets can be readily stored safely at minimum cost as a means for totally eliminating carbon dioxide emissions to the atmosphere. The various scenarios are as follows:

(i) The sensible heat in the carbon pellets exiting the natural gas thermal decomposition reactor is recovered within the tunnel kiln in conjunction with cast steel and reduction zone slag to contribute to solid charge preheating.

(ii) The oversized carbon pellets exiting the natural gas thermal decomposition reactor at temperatures typically in the range 1100°C to 1200°C to provide input general carbon feed for continuous carbothermic reduction on-site of metal oxides for metal production.

(iii) The carbon pellets are milled to produce carbon particle feed specifically for incorporation directly into the flow of molten aluminium associated with a new primary aluminium carbothermic reduction technology (N. A. Warner, Metall. and Materials, Trans. B, 39B, April 2008, pp. 246-267). The new technology is based on melt circulation of a slurry of molten aluminium containing particles of pure carbon to provide by conductive heating to satisfy the thermal demands of a metal producing reactor in which refined aluminium oxide is reduced to metal at temperatures greater than 2150°C so that aluminium carbon AI₄C₃ is thermodynamically unstable. Thus the aluminium oxide feed is reduced directly to aluminium metal by the availability of elemental carbon both in solution in the melt and dispersed in the molten aluminium metal carrier medium. Overall energy consumption figures 25% lower than the Hall-Heroult electrolytic route are predicted with 42% less purchased electricity, supplemented by elemental carbon. Whatever the primary aluminum process, carbon emissions are inevitable, so carbon capture and storage (CCS) must feature in forward planning. The total carbon consumption for Hall-Heroult is in the region of 0.45 kg/kg Al, whereas the new carbon input for the carbothermic route under discussion is about 0.67 kg/kg Al. However, this increased carbon input is more than compensated by the fact that CCS for the new technology is expected to be considerably more straightforward. Hall-Heroult has environmental problems arising from the anode effect in relation to formation of perfluorocarbons. The estimated greenhouse equivalent C0₂ emission due to this phenomenon is 2.20 kg CC kg Al. This brings an equivalent carbon emission of 0.6 kg C, making the total 1.05 kg C kg Al, which is then larger than the estimated value of 0.67 kg C/kg Al for the proposed carbothermic process. Production of 6M tpa of steel with the zero gas emission steel technology under discussion provides enough carbon to produce continuously in excess of 1 tpa of primary aluminium.

(iv) Cold carbon pellets are transported to a remote site in which aluminium is produced using established Hall-Heroult technology in order to provide carbon feed material for anode production.

(v) The cold carbon pellets are transported to a site close to where carbon dioxide injection into operating oil wells is required to enhance oil recovery without involvement of long distance pipeline infrastructure.

(vi) The carbon by-product of the steelmaking process is transported to a Carbon Capture and Storage (CCS) hub where it can been combusted either alone or in combination with other carbonaceous solid fuels to generate electricity with carbon dioxide capture ready for sequestration in an appropriate geological storage site either on-shore or off-shore in a depleted oil or gas basin/well or alternatively directly into a saline aquifer below the off-shore sea storage area.

Again referring to Fig. 9, on reaching product size, pellets are segregated and discharged, while a periodic addition of fines carbon particles replaces them in the bed. Upon leaving the decomposer 5, the stream of pellets at around 1600 K is lifted about 20 m by means of enclosed bucket elevator 6. Over 99% of the pellet flow is returned to a preheater and then back to the decomposer 5. A relatively small stream of the largest pellets (~ 10 mm diameter) is continuously removed from the system at 8 as product, while an equal number of small seed pellets (< 2 mm) is added. The whole system is rigorously sealed because of the hazardous nature of the reacting gases, but particularly the hydrogen product.

Referring to the prior art in Fig. 10, the pebble beds are fixed between a hot grid and a cold grid. The cold grid is usually held at temperatures under 250°C, meaning that the outer shells can be constructed of conventional steel without a refractory lining. Such linings are only required on top of the hot grid (so-called "dome") and the bottom around the hot gas inlet outlet. The pebble bed is a regenerative heater exchanger, which means that at least two units are required for continuous flow. Whilst one is being heated, the other produces hot gas. The operation is continuous and after a certain time period, the units switch and exchange their functions. During the heating phase, hot gas enters the pebble bed through the hot grid and progresses radially across the bed and after passing through the packed bed, the cold gas passes out through the flange outlet at the top. When the Pebble-Heater is reheated, hot gas admission is terminated and after pressurizing the cold gas enters in the reverse direction. In Fig. 10 the preheated gas leaves the Pebble-Heater through the hot blast main at the bottom of the unit.

Referring now to Fig. 11, the prior art has been amended to comply with the requirements of the process now being considered, but as far as possible proven technological aspects are retained so that existing mathematical modelling as well as experimental and industrial experiences reported in the literature become the foundation for its application to zero gas emission employing hydrogen as the sole reductant. This context is now the focus of the description, but clearly other more general high temperature applications in extraction metallurgy could be equal relevance. The principal difference from the prior art is the change from cold grid to a second ceramic grid 1 to accommodate subsequently the means for resolving the potential problem of solid iron deposition within the pebble bed 2 in the steelmaking context. Not shown in Fig 10 is the vitally important third pebble heater 3, which has been taken off-line to effect the thermal treatment mode, which is detailed in Fig. 11.

On the left in Fig. 11, one of the two on-line pebble units 74 is in the heating-up phase with hydrogen reduction off-gas entering 75 at say 1540°C and leaving 76 at say 100°C. On the right hand side, make-up hydrogen together with reclaimed hydrogen following hydrogen/water phase separation and filtration or magnetic separation to remove any iron particles (other solid phase contaminants are in theory are virtually non-existent) but some contamination is inevitable, because of physical interaction with refractory materials in gas duct linings and elsewhere entrained in the gas flow before boosting for recycling to both the reduction zone and post-combustion zone of the melt circulation reduction reactor via the pebble heater 74. The recycled hydrogen enters at 77 at 25°C and leaves 78 at say 1520°C.

Referring next to Fig. 12, the pebble heater unit 74 currently in the thermal treatment mode has a central graphite or silicon carbide electrical heating facility temporarily installed in the previous centrally vacant zone 79. Also provision is made for admitting preheated argon or other inert gas 80 initially starting from a relatively low temperature probably in the region of 200°C or so. This inert gas is gradually raised in temperature up to around 1540°C to avoid thermal shock in order to facilitate a slow melt down of solid iron deposited within the pellet bed. This needs to be accompanied by controlled radiative heat input as required from the centrally placed electrical heating facility. Once the temperature exceeds the iron melting point throughout the entire pebble bed, it is raised slightly even further to ensure that all iron is in the molten state. At this juncture an intermittent start stop gas flow regime with pressure pulsation of the admitted argon is commenced to take advantage of the mobility of so-called "static hold-up" of iron globules or droplets. In the non-wetting mode, the motion of droplets continues under the influence of gravity and proceeds to the bottom of the bed. Coalesced molten iron, which was once the contaminant deposition, leaves the unit via the discharge launder 81, probably tapped intermittently or possibly continuously into an insulated container for subsequent addition to the molten iron product overflow stream of the melt circulation reactor (not shown).

Still referring to Fig. 12, the hot argon or other inert gas exits the pebble heater 74 at 84 to be recycled after progressing through its own independent regenerative/recuperative heat recovery system to be boosted in pressure for recycling after preheating back to 80. The next phase in the thermal treatment mode is commenced once no further molten iron accumulates in the launder 81. The objective then becomes establishment of an appropriate temperature gradient across the pebble bed by a controlled programme involving gradual decrease the temperature of the recirculated argon in conjunction with variations of the radiant heat input from the electrical heater facility within the central region 79.

Referring to Fig. 13, combustion of pre-heated clean hydrogen 98 supplies most of the thermal demand for methane decomposition. Heat balance calculations yield the requisite of hydrogen combustion needed in association with the sensible heat of the off-gas 75, principally water vapour from the post-combustion zone, at around 1800°C. This is added at the base of a moving packed bed 86 of alumina pebbles countercurrent to the upward recirculation of water vapour as the diluent in a closed circuit. The requisite amount of oxygen 88 drives the gas ejector 89 to enable a large recirculation of water vapour as the diluent in a closed circuit, completely analogous to the scheme outlined previously for the post-combustion zone of the melt circulation loop. Preliminary calculations have been undertaken to identify the circulation rate of alumina pellets 83 in the packed bed 85 to cool the combustion gases with the excess gas (not shown) leaving the moving packed bed for transmission to a regenerative pebble heater, probably that used for pre-heating the reductant hydrogen. The reheated alumina pellets 84 enter a channel adjacent to a channel containing the circulating carbon pellets immediately before the carbon pellets are returned to the methane thermal decomposition. The heat transfer involved in this latter step is entirely radiative.

Referring now to Fig. 14, the continuous steel refining melt circulation loop 90, which takes hydrogen deoxidised molten iron (not shown) overflown or siphoned out from the principal ironmaking melt circulation loop and subjected it to the new fluxed launder process described already elsewhere. The whole loop 90 is electrically conductively heated and employs permanent freeze linings of solid steel throughout. The relatively small RH units 93 as used when the RH process was first introduced commercially in the 1960's. This provides the molten steel "hydrostatic" head to overcome friction and permit flux/metal phase separation associated with the flux launder desulphurization (not shown) and then final dehydrogenation following open channel desorption of dissolved hydrogen at reduced pressure into an argon carrier jet (not shown). The reduced pressure or vacuum siphonic tundish 91 is discussed later in detail. The decomposer for natural gas is 68 and three pebble heaters involved in regenerative hydrogen preheating is shown as 74. The reduction zone 1 of the principal melt circulation loop and the post-combustion zone 2 are both shown schematically without further detail. The final refined wide slab steel 92 up to say 20- 50 mm in thickness is shown to be beneath the melt circulation loop immerging from the siphonic tundish 91 continuous casting zone. Referring to Fig. 15, the plan view shows the positioning of the impinging jet siphonic tundish 91. The relatively shallow steel refining loop 90 melt depth is increased in the region 98 to accommodate the siphonic tundish 91 in which the melt in the upleg of the siphon is designed so that laminar flow is established to preclude exogenous non-metallic inclusions being introduced at this late stage in the overall process. The recirculated refined molten steel 94 is pumped around the melt circulation loop with solid steel permanent freeze linings throughout the loop but this is not feasible for the siphonic tundish itself as this has to be available for lifting up and down periodically to replace the impinging jet assemblies fabricated from fused highly wear resistant refractory plugs on a regular basis. This implies the availability of a spare siphonic tundish located adjacent to the tundish in continuous operation in a preheated condition ready for installation so that replacement can be rapidly undertaken with only minor effect on the continuity of the casting operation. The sectional elevation shown in Fig. 15 is purposely not crossed-hatched to enhance clarity, whilst the associated numbers 90 to 97 in this sectional elevation are defined and fully discussed in Fig. 16.

Now referring to Fig 16, the refined molten steel 94 is recirculated continuously around the steel refining melt circulation loop. For 2 millions tones per annum steel product, the melt circulation in the loop is around 20 tonnes per minute compared with the product removal rate of about 4 tonnes per minute. In the example currently being discussed, the reduced pressure or vacuum tundish 91, which sets automatically by gas pressure selection the molten steel depth above four pairs of impinging jets each initially 25 mm in diameter. In Fig. 16 one of these spray sheets 95 is depicted to be formed in the argon protective atmosphere from one of these pairs of jets, producing a rain of molten steel droplets. These droplets fall onto a solid very thin steel sheet 96 about 2 mm in thickness moving continuously horizontally forward and on which coalescence of the steel droplets takes place to form the wide thin pool of initial molten steel, which on solidification yields the wide steel product slab. The depth of this initial is set by the reduced gas pressure in the gas space immediately above the liquid steel 94 within the tundish. This establishes the flow rate automatically by variation of the liquid steel "hydrostatic" head above the plugs containing the cylindrical conduits at the prescribed angle to the horizontal plane, probably within the range 30-45 degrees, a figure yet to be determined in pilot plant operations.

To effect solidification, whilst maintaining the stability of the underlying initial solid thin sheet 96, recirculated molten lead 93, typically at a temperature of 900-950°C, is pumped around a closed loop. This coolant of molten lead enters the solidification zone just after the thin steel sheets are welded together to form the desired width with in-situ welded steel edge rims for melt containment without lead contamination. All of this preparation must be done in advance of the wide thin steel sheet 96 progressing to the layer of circulating coolant of molten lead 93 immediately before coming into contact with the rain of steel droplets.

The thermal stability of the foundation thin steel sheet 96 is ensured by intense liquid metal heat transfer on its underneath side by the provision of forced recirculation of the molten lead. The resulting steel slab 92 remains in contact by floating on the molten lead until it is cooled to around 1000°C. The initial thin steel sheet once transformed to the resultant product steel slab 92 is pulled through the casting zone by pinch rolls after separation from the molten lead coolant is accomplished and any droplets removed by an "air knife^* employing argon rather than air. The wide steel stab 92 optionally continues its forward horizontal motion in order to preheat incoming ore and flux materials within a very well insulated tunnel kiln. Alternatively, it is subjected to hot rolling to produce sheets of the desired thickness, including perhaps a minor production of the 2 mm sheet 96 employed in the casting process. This would remove the necessity for joining together by welding smaller width material. At a remote location at which the iron ore is mined, it may be preferable to use existing rail infrastructure to deliver rolls of thin steel sheet to the minesite, where the zero gas emission continuous steelmaking is taking place. These rolls of thin steel sheet are welded together in advance of welding on the edge rims in- situ, just in advance of being exposed to the impinging jet rain of molten steel droplets 95 from the siphonic tundish 91.

Referring to Fig. 17, the only item not already discussed in the description is the packed bed contactor 100 in which molten iron overflows or is otherwise removed from the primary ironmaking melt circulation loop. The carrier molten iron at steady state contains a dissolved iron oxide concentration in the region of 0.3 thermodynamic activity, depending entirely on the chemical composition of the ore charge, particularly its phosphorus content, which normally needs to be reduced to an acceptable level in the first melt circulation loop before proceeding to the steel refining loop. Major dephosphorization requires a relatively high oxygen concentration to be established in the molten iron carrier medium of the ironmaking melt circulation loop. Deoxidation is not conducted by conventional addition of aluminium or silicon. This is in the interests of precluding non-metallic inclusions, as discussed at length previously in this description.

Chemical reaction between dissolved oxygen and a continuous phase hydrogen gas effects efficient molten iron deoxidation. Hydrogen is admitted to the packed bed contactor 100 containing dense recrystailized alumina spheres or perhaps Raschig ring packing elements. As the molten iron does not wet the AI2O₃ packing, it is a case of non-wetting irrigation of a packed bed discussed previously by the inventor in a number of technical papers. The hydrogen enters at 101 and leaves containing some water vapour at 102.

After passing through a conventional lute, the deoxidized iron melt flows by gravity into the steelmaking loop at 103 to undergo subsequent desulphurization and removal of dissolved hydrogen. Dehydrogenation involves initially open channel desorption into an argon carrier gas at reduced pressure levels, typically around 30-50 mbar. Finally, dehydrogenation occurs in the RH gas-lift pumping system (not shown in Fig. 17), typically down to less than 1.5 pm residual hydrogen in the steel product.

The other items in Fig. 17 have all been discussed several times already in the description. The reduction zone 1 and the post-combustion zone 2 constitute the essentials of the ironmaking melt circulation loop. The regenerative heat recovery system for preheating incoming hydrogen reductant employs the three pebble heater units 74. The initial thermal decomposition of natural gas to produce the required hydrogen needed for zero gas emission continuous steelmaking is conducted in the decomposer 68 with natural gas entering at 64 and hydrogen leaving at 67. The other items include the steel refining loop comprised of the freeze-lined open channels 90 protected with argon and the all important reduced pressure or vacuum tundish 91, discharging a rain of molten steel droplets employing the impinging jet principle (not shown) to finally yield to the wide steel slab product 92 from the foundation thin steel sheet 96. Molten lead recirculation induces highly intensive liquid metal heat transfer to ensure stability of the substrate thin steel sheet 96, when it first encounters the rain of liquid steel droplets in the continuous casting zone.

Claims

Claims

1. A continuous hydrogen-based iron or steelmaking process employing at least one melt circulation loop with carbon-free molten iron as the carrier medium incorporating iron ore fines or crushed as-mined iron ore without distinction, which captures the dramatic increase in the reduction chemical reaction rate constant at the melting point of wusttte for an oxidic melt layer containing principally molten wustite stemming from the rapid chemical reaction between iron oxide and iron, when a preheated charge is first added into the circulating molten iron, reflecting the overwhelming effect of dissolved oxygen on the surface tension of molten iron with the resulting interfacial turbulence contributing markedly to the formation of a stable thin floating layer of oxidic melt with the result the rate of hydrogen reduction is gas-diffusion controlled and thus independent of total pressure, thereby dismissing the need for elevated pressure in the entire flowsheet and hence enhancing overall safety.

2. A process as claimed in claim 1 , in which hydrogen combustion with highly diluted oxygen in a gas phase composed principally of water vapour provides the thermal energy requirements in both the main melt circulation loop and in the thermal decomposition of methane ensuring the objective of zero gas emission on condensation out the water vapour before recycling filtered clean hydrogen back to the process.

3. A process as claimed in claim 1 , in which natural gas is thermally decomposed to yield carbon and hydrogen using radial gas flow within an annular moving packed bed of carbon pellets continuously recirculated from top to bottom, in which carbon is deposited on the surface of the carbon pellets whilst oversized pellets are segregated screened and removed as carbon by-product, and the remainder are utilized as the carrier means for recovering heat radiatively from associated high temperature process streams to contribute towards satisfying the thermal demands of natural gas decomposition.

4. A process as claimed in claims 1 and 3, in which the hot oversized carbon pellets are milled to produce carbon particle feed specifically for incorporation into the flow of molten aluminium associated with new primary aluminium carbothermic reduction technology based on melt circulation of a slurry of molten aluminium containing dispersed carbon particles to satisfy by electrical conductive heating the thermal demands of an aluminium metal producing reactor, in which refined aluminium oxide is reduced directly to metal at temperatures greater than 2150°C so that aluminium carbide is thermodynamically unstable and thus providing the means for co-production continuously of both steel and aluminium at the same plant location.

5. A process as claimed in claims 1 and 3, in which cold carbon pellets are transported to a remote site, in which aluminium is produced using established Hall-Herault technology in order to provide carbon feed material for anode production.

6. A process as claimed in claims 1 and 3, in which the cold carbon pellets are transported to a site where carbon dioxide injection into operating oil wells is required to enhance oil recovery without involvement of long distance pipeline infrastructure.

7. A process as claimed in claims 1 and 3, in which the carbon by-product s transported to a Carbon Capture and Storage (CCS) hub, where it can been combusted either alone or in combination with other carbonaceous solid fuels to generate electricity.

8. A process as claimed in claim 1 , in which disruption of thermal energy recovery is prevented by employing at least three pebble heaters with radial gas flow, by alteration of the current commercially accepted arrangement, so that the entire packed bed, when off-line, can be cyclically raised to a temperature above the iron melting point in order to dislodge the static hold-up iron accumulation within the pebble bed, thus relieving the bed of its contaminant so that the full voidage for gas flow at minimum gas phase pressure drop is effectively restored.

9. A process as claimed in claims 1 and 8, in which a preheated inert carrier gas is recycled through the off-line pebble heater so that its entire bed of pebbles is gradually brought up to a temperature in excess of the iron melting point and then to enhance removal of so-called "static hold-up" droplets or globules, the gas flow is interrupted by an intermittent stop/start procedure with accompanying gas phase pressure swings that promote movement of the static globules of molten iron under non-wetting behaviour.

10. A process as claimed in claims 1 , 8 and 9, in which the dislodged molten iron discharges into an appropriate container during thermal treatment of the particular third pebble heater, whilst not on-line, which is well insulated to retain molten iron in its liquid state in advance of its return to the molten iron product overflow stream of the continuous steelmaking process.

11. A process as claimed in claims 1 and 3, in which thermal input is obtained from graphite heating elements or electrodes placed above the carbon pellets in the form of a layer of pellets discharged from a bucket elevation onto a slowly rotating refractory hearth with both pellet input and output arranged so that gravity discharge into the natural gas or methane thermal decomposer is facilitated.

12. A process as claimed in claims 1 , in which refined steel is continuously drawn from the steel refining loop into the upleg of a siphonic tundish to establish the liquid head required to sustain the design level of stream breakup via pairs self impingement of jets of liquid steel that produce a rain of droplets falling down onto a wide foundation solid sheet of steel around 1mm or so in thickness to form initially a shallow pool of molten steel with the foundation sheet thermally stabilized in the solid state by forced convection heat transfer employing a recirculated liquid metal coolant such as molten lead or bismuth lead eutectic, thereby ensuring highly efficient thermal energy recovery in contrast to conventional continuous casting dissipating heat to cooling water.

13. A process as claimed in claim 1 , in which thermal energy from slag or residual flux is recovered for preheating solid charge materials following phase separation from molten steel employing self- impinging pairs of jets in an analogous fashion that outlined for the molten steel product stream in claim 12.

14. A process as claimed in claims 1, 13 and 15, in which oxidic melts, become thermodynamically equilibrated with iron, so containment in refractory enclosures lined with steel below its melting point is available to assist with containment in the sub-processes involved.

15. A process as claimed in claim 1 , in which slag formed in the principal melt circulation loop along with residual flux and oxidic meK associated with the steel refining loop are reacted with preheated silica flux under oxidizing conditions to capture the exothermicity of reinstatement of ferric iron rather than ferrous iron domination.

16. A process as claimed in claims 1 and 12, in which efficient heat recovery is achieved in a well- insulated tunnel kiln by promoting radiative heat transfer to horizontally conveyed solid iron ore and flux charge materials from the casting processes by capturing the latent heat of solidification as well as the sensible heat as the solidified molten iron and slags continued on their way to the incoming materials undergoing preheating.