Title of Invention

A METHOD OF PRODUCING ULTRA LOW CARBON STEEL

Abstract

A method of producing ultra low carbon steel by refining molten low carbon steel in a vaccuum degassing process comprising the steps of: bringing a cylindrical reaction chamber (5) with gas-suction plant (8) and two snorkels (6, 7) connected to the bottom of the chamber (5) close to a ladle (1) containing liquid bath (2) and slag layer (3) such that the two snorkels (6,7) are immersed well into the bath (2) to suck up some amount of molten metal into the chamber through the snorkels due to pressure difference between the reduced pressure of the chamber due to gas suction by the suction plant (8) in the chamber and atmospheric pressure of the ladle (1) such that height of metal in the ladle balances the said difference between the external atmospheric pressure and the internal chamber pressure, injecting inert argon gas into the up-snorkel (6) by means of a device (4) connected to the snorkel (6) to create a gas-liquid mixture (9) inside the chamber (5) and a gradual circulation of the liquid steel argon mixture laterally between the up and down snorkels in the chamber down the down snorkel and then laterally from down to up snorkel in the ladle due to difference to lower density of gas liquid mixture near the vicinity of the up-snorkel (6) and higher density of the molten steel when part of argon acquired in up-snorkel (6) is released from steel as it passes from up snorkel to down snorkel within the chamber thus allowing the molten metal (9) inside the chamber (5) to come back to the ladle through the second snorkel (7) and thus decarburizing the liquid steel by speeding up the reaction between the dissolved carbon of liquid metal and oxygen on producing carbon monoxide and partially carbon dioxide (11) and further decarburizing with injected oxygen from means (10) to produce ultra low carbon steel; characterized by controlling the carbon and oxygen concentration and temperature of the liquid steel during decarburization on predicted estimation of continuously sampled exhaust gases from the chamber (5) via exit gas measurement device based on mass spectroscopy to provide the flow rates of carbon monoxide, carbon dioxide and oxygen with a time delay corresponding to the instant of release of gas and its analysis in the gas analyzer; converting the obtained exit gas values on neutralizing time delay into instantaneous values of carbon, oxygen concentrations and temperature of the steel bath, in real time, said estimation during decarburization is carried out through free of emperical models State Estimation Using Evolutionary Mapping (SEEMA).

Full Text

FIELD OF APPLICATION The present invention relates to a method for producing ultra-low carbon steel. In particular tire invention relates to a method for real time prediction of carbon concentration in molten steel during RH degassing using a model-free state estimation technique. BACKGROUND OF THE INVENTION With continuous increase in the demand for ultra-low carbon steel, the vacuum circulation refining process (also called RH degassing) has been developed and applied widely. This process, a component of secondary steel making technology, has now become an integral part of the whole gamut of steel production activities. The degassing process is performed in an RH chamber and ladle for producing ultra-low carbon steel. Oxygen and carbon already present in liquid steel are made to further react with one another and produce carbon monoxide under conditions of very low ambient pressure and / or blowing of oxygen into the RH chamber. The RH degassing unit essentially comprises a tall degassing chamber of roughly cylindrical shape internally coated with refractory bricks. The upper part of the chamber is connected to a gas suction unit. The function of this gas suction unit is to create a near-vacuum pressure, which promotes reaction between dissolved carbon and oxygen to generate carbon monoxide and also carbon dioxide. Two tubular snorkels internally and externally covered with refractory material, of circular or oval cross section are connected to the bottom part of the chamber. One of the snorkels has a provision to inject inert gas. Before actual use the chamber is heated to 900° C - 1500° C. The vacuum chamber is brought close to the ladle containing molten metal and both are arranged in a way that the two snorkels are immersed to some depth in the liquid metal. The gas suction unit connected to the chamber, as mentioned above, is brought into operation thereby creating near-vacuum conditions inside the chamber. Consequently, certain amount of metal is sucked up into the chamber from the ladle by rising through the snorkels. Next the inert gas is injected into one of the snorkels. The main purpose of this injection is to drive into circulation the molten metal inside the chamber by creating a density gradient between the metal that is heavily mixed (heterogeneously) with inert gas and the parts of liquid metal bath with less inert gas component in mixture; hence this snorkel is called as "ascending snorkel" or "up-snorkel". The molten metal thus circulating through the chamber, after certain residence time, comes back to the ladle due to gravity through the second snorkel, which is also known as "descending snorkel" or "down-snorkel". In this fashion the molten metal circulates continuously between ladle and chamber. The reduced pressure inside the chamber promotes the decarburization reaction. The possible reaction zones between carbon and oxygen to produce mainly carbon monoxide and some carbon dioxide are inside the vacuum chamber, inside the snorkels and inside the ladle. Maximum reaction occurs in the liquid layer closest to the vacuum in the chamber. The gaseous mixture thus produced is taken out through exit gas channel. Along with the oxidization of carbon, the other metallurgical operations that are likely to be performed in the chamber are addition of alloying elements and reheating of metal by Thermit process. Once the target concentration of carbon is achieved, aluminum is added to the steel to completely reduce the oxygen content. Since the reaction between aluminum and oxygen is exothermic, because of this addition the overall temperature of the melt increases. To produce various grades of steel with specific compositions by RH degassing process, and also to avoid steel getting contaminated with alloying agents, it is very essential to have a system which predicts carbon, oxygen and temperature, the three main constituents that define the quality of the end product. Considerable work has been done to find methods of accurately predicting values of dissolved carbon and oxygen in the molten steel during the degassing process. Matsuo and Kiyoto at Sumitomo Metal Industries have built up correlation relations between the CO gas density as observed in a gas density detector provided in the exhaust gas path of the RH vacuum degassing tank, and the concentration of Carbon in the steel (Patent No. JP 1222018, 1986). This correlation is built up from data recorded over a large amount of sample testing, which is costly, time-consuming and subject to inaccuracies when the process conditions approach limits at which the samples were taken. Hisashi and Kosaku at Nippon Steel (JP 2298861, 1990) have used a means of converting the entire carbon-carrying exhaust gas (composed of CO and CO2) into CO2 by oxidation of CO, and thence continuously carrying a sample into a mass spectrometer for measuring CO2 content in real time. The mass flow rate of CO2 is then converted into a decarburization rate from which instantaneous C concentration is estimated. Inaccuracies are introduced, first, in converting CO2 content in the sample into content for the total exhaust gas, and second, from leakages from the vacuum vessel which need to be accounted for. Ryusuke, Shoichi and Masato of Kawasaki Steel (JP 6192724,1994; JP 6192724, 1995) have built a relationship between the difference in C concentration in steel at an instant and the concentration prevailing at the end of process, and the difference between exhaust gas flow rate at the corresponding instant and the exhaust gas flow rate at end of process. The C concentration in steel at different instances is obtained by sampling. This method is also subject to inaccuracies as applicable to the approach of Matsuo and Kiyoto mentioned above. Takeshi, Aklra and Takuji at Nippon Kokan (JP 7242928,1995) have constructed an empirical mathematical model relating exhaust gas content and flow rates, Argon injection rate, vacuum pressure, molten steel quantity, temperature, etc. with Carton and Oxygen concentration in molten steel. This empirical model is constructed from a large number of measurements made on these quantities during the degassing process; needless to say, it is costly and time consuming, and to replicate the accuracies attained, will need to be repeated for every new or modified degassing unit. Ono et al at Kawasaki (US 5522915; 1996) have used an alternate method wherein an inert carrier gas is circulated (bubbled) through the molten steel for a long enough time for the concentration of different gases dissolved in the steel to attain equilibrium with their concentration in the carrier gas bubbles; then the carrier gas is collected and analyzed separately for content of each gas - including nitrogen, oxygen, hydrogen, carbon monoxide and water vapour. From these the Carbon concentration in molten steel is recovered. The inability of this method to provide results in real time strikes as an obvious disadvantage. Existing methods try to model the chemical reactions and juxtapose them against the cumulative mass withdrawal of carbon-carrying gases. These methods tend to be inaccurate due to uncertainties in the estimate of the latter. A need therefore, existed for steering dear of chemical and mass based approaches and instead using a model-free technique predicting the point when the desired level of carbon concentration is attained. SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide as estimate of the carbon and oxygen concentrations, and the temperature, in liquid steel in real time with high accuracy during the RH degassing process. Robustness across different RH chamber designs and process conditions is a characteristic of this invention. Another object of this invention is to provide a method to make the above estimate that is free of empirical models based on either measurements or physics or both, and unlinked with chemical, molecular, mass or energy balance principles. Instead, this method belongs to the class of state estimation techniques frequently used in the aerospace or transportation domains that relate dynamic system states to measurables and inputs within a control theoretic framework, as typical of Kalman Filters for example ("An introduction to the Kalman Filter". Welch, Greg and Bishop, Gary, University of North Carolina, http://www.cs.unc.edu/~welch/kalman/. The above method is called state estimation using evolutionary mapping approach (SEEMA). It develops an (m x n) piecewise linear map from n states to m measurables in the initiation phase and then apportions higher-order components of the evolution of the measurable parameters partly to the evolution of states and partly to the evolution of this piecewise linear map. Here n, the number of states, is three-carbon reduction rate, oxygen change rate and temperature change rate, and m, the number of measurable parameters, is also three - CO, CO2 and O2 volumetric flow rates. An RH vessel at near-vacuum pressure inside has two snorkels dipped in liquid steel contained in a ladle held below the vessel. One of the two snorkels is infused with a pipe through which inert gas Argon is injected that gradually sets the liquid steel into circulation between the ladle, up the Argon-blowing snorkel, the RH vessel and then down the other snorkel. Gases emitted from steel are Sucked out past an exit gas measurement system based on mass spectrometry that provides the flow rates of carbon-monoxide, carbon-dioxide and oxygen at high refreshment rates. The exit gas analysis is obtained with a time delay corresponding to the instant of release of the gas and its analysis in the gas analyzer. According to the present invention after neutralizing time delay, the exit gas values are converted into instantaneous values of carbon and oxygen concentrations, and temperature of the steel bath, in real time. The method for mapping real time exit gas measurements into instantaneous dissolved non-metal concentrations is based on the novel technique SEEMA - state estimation using evolutionary mapping approach. SEEMA does not use any form of chemical, molecular or mass balance models as typical of extant methods for carbon prediction in the RH degassing process. SEEMA learns an initial map between the vector of measurables (exit gases) and the vector of states (dissolved concentrations) from available initiating data. SEEMA guides this initial map to evolve with time based on certain assumptions on the nature of evolution of states and measurables of dynamical physical systems, along with the actual characteristics of evolution of the measurables. At any instant of the evolution, the prevailing map is used to convert the current measurables into current states. The initial relational data structure for use by SEEMA is generated using another novel technique, wherein available data on evolution of measurables with time is averaged across a number of RH process heats to create the first component of the data structure. Specially measured values of states at the initiating phases of various RH process heats, mostly different from the set of heats used to generate the first component, are averaged and their evolution with time is used to create the second component of the data structure. The measurables and the states are then linked through the common element - time - the scale for evolution of either - into a relational data base as needed by SEEMA. The relationship between state parameter data and measurement parameter data, necessary to initiate SEEMA in a degassing process run, is embodied in an initiation table that is constructed from sampled data taken from a few vacuum degassing process runs. The linkage between the two sets of variables is provided by a common factor - time. In the estimation technique the present invention maps the measured instantaneous flux of exit gases into a corresponding removal rate of carbon and other dissolved gases in real time. The integration of this removal rate with time provides the net withdrawal of carbon since the start of process till the instant under consideration. The present invention thus provides a method for producing ultra-low carbon steel comprising the steps of: reducing the pressure inside a RH degassing chamber using a suction unit; injecting an inert like argon gas through one of a pair of snorkels dipped in liquid steel contained in a ladle arranged below said RH chamber; creating a gradual circulation in the liquid steel; reduced pressure inside the RH chamber promoting formation of CO and CO2; and extracting exhaust gases continuously for analysis to evaluate proportion of CO and CO2; a a graphic user interface (GUI) displaying the results of the analysis in real time. BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS The invention will now be described with reference to the figures of the drawings where Figure 1 is a schematic diagram of RH degassing unit. Figure 2 is a view illustrating the entire system architecture in a block diagram form. Figure 3 shows the relational structure of process initiating data. DETAILED DESCRIPTION Referring to Figure. 1, description is made below of an outline of a conventional RH degassing process to which the invention is applicable. In this figure, molten steel 2 is contained in a ladle 1 with a floating slag layer 3 exposed to atmospheric conditions. A tall cylindrical chamber - the RH reactor chamber 5, is connected to a gas-suction plant 8 which creates vacuum inside the chamber. Two snorkels 6, 7 are connected to the bottom of the chamber. One snorkel 6 is having a device 4 to inject Argon. Once the operator signals the start of degassing process, as a first step, reaction chamber 5 is brought close to ladle 1, such that the two snorkels 6, 7 are immersed well into the bath 2. Thereafter gas suction unit 8 reduces the pressure inside the chamber. Because of the pressure difference between chamber 5 and ladle 1, some amount of the molten metal 2 is sucked up into the chamber 5 through both snorkels, such that the height of metal balances the difference between the external atmospheric pressure and the internal chamber pressure. At this point of time, inert gas Argon is injected into the up-snorkel 6 by means of the device 4. Thus a gas-liquid mixture 9 is formed inside the chamber. In the vicinity of the up-snorkel, the heterogeneous steel-Argon mixture is of lower density as compared to the rest of the molten steel, hence it tends to rise up creating a gradual circulation - laterally between the up and down-snorkels in the chamber, down the down-snorkel, and then laterally from down - to up - snorkel in the ladle. Part of Argon gas acquired in up-snorkel 6 is released from steel as it passes from up-snorkel 6 to down-snorkel 7 within the chamber. Thus the molten metal 9 inside the chamber 5 comes back to the ladle through the second (down) snorkel 7. Reduced pressure inside the chamber promotes decarburization. In other words the reaction between dissolved carbon and oxygen that produces mostly carbon monoxide and partially carbon dioxide 11 speeds up. A provision 10 to inject oxygen helps in further decarburization of steel to produce ultra low carbon compositions. Exhaust gases are extracted and continuously sampled for gas analysis to evaluate the proportions of carbon monoxide and dioxide. Fig. 2 is a schematic representation of the system architecture. From field 12, i.e. RH degassing unit, signals 13 are tapped continuously in real time, and routed to PLC 14. These signals are tapped from the various instruments that measure vacuum pressure inside the chamber 5, Argon flow rate, carbon monoxide, carbon dioxide and also oxygen proportions from gas analyzer. All these signals are in analog form, hence converted into digital values by A/D converter card 15. A CIM card 16 loaded with modbus driver software is inserted in one of the slots of the above said PLC 14. A two-way data communication is established between PLC 14 and CIM card 16 through back plane communication 17. Another software to fetch the data routed from CIM card is loaded in the PC 19, which acts as computing platform. The communication between CIM card and PC 19 is established by means of serial communication 18. The real time computational algorithm, responsible for estimating Carbon, Oxygen and Temperature is loaded and executes on PC 19. Details of this algorithm are explained in the following paragraphs. Results of real time computation are displayed in a computer 20, through a software program that supports graphical user interface features. The communication between PC 19 acting as computing platform for algorithm software and to the PC 20, acting as Man-Machine-Interface is by means of TCP/IP 21. Details of the computational algorithm SEEMA as applied to the RH degassing process for estimating carbon and oxygen concentrations and temperature in the liquid steel in real time, from measurements of volumetric flow rates of CO, CO2 and O2 are explained below. Let Xk, yK represent the vector of states and the vector of measurables, respectively, at time instant k, where and It may be noted that both states and measurables are in fact rates of change of concentration and rates of flow, as seen in equations (1-2), rather than the absolute values of respective variables. Of course our interest lies in Carbon content and Oxygen content, in parts per million, and temperature in degrees Celsius, these can be simply recovered as and similarly for oxygen and temperature. Now, yK is available at polling frequency, and is likely to show undesirable minor fluctuations due to noise. Assuming that the noise is white, taking moving average will reduce the effect of this noise, and larger the base of the moving average, the more complete will be the cancellation. A moving average across eleven polling instances is taken. also the exact representation of yK-1 is explained in step 5 below. Going over the estimation steps, first, a second order filter yk of measurable vector Yk is extracted using the sequence of available measurements and second order Taylor series formula as It may be noted that the first and second derivatives are centered at Yk-1/2 and yK-1 instead of at YK-3/2 and yk-2, as would normally happen in any Taylor series extrapolation. Second, the second order filter of measurables yk is used to obtain a corrected estimate of Xk following the steps outlined below: where fk-1(.) represents the second order Taylor series extrapolation formula from Xk-1 as Next the estimate of xK obtained in equation (5.1) is passed through the quasi- linear map Hk-1 to obtain an estimate of the measurable yk as Where H is a quasi - linear mapping from the vector of states to the vector of measurables of the form The mapping H is obtained from the initial data table for every process run by using the values of CO, CO2 and O2 (their volumetric flow rates from exhaust gas analyzer) as prevailing at 4 selected time instances to extract the corresponding values of Xk, and then using least square minimization coupled with Singular Value Decomposition Method ("Numerical Recipes in C", Press, William, Teukolsky, Saul, Vetterling, William and Flannery, Brian, Cambridge University Press, 1992) of solution of a linear system of equations to get H. The value of H thus obtained at the initial step is then updated at every fourth step of the process as described in step four below. The obtained value of Yk in equations (5.2) is then subtracted from the quadratic filter of measurables yk (from step one) to obtain the error estimate at quadratic level of measurables eyk as The error in y is then transformed into the error in x by the inverse map of H as Finally the corrected estimate of Xk is obtained by adding the obtained error exk to the initial estimate as This completes the second step of the application of SEEMA to the carbon concentration prediction problem in the RH degasser. At the third step, the third order term of the Taylor series expression for the vector of measurables y (refer to equation (4)), denoted as ΔYK, is calculated as 6 This term plays an important role in directing the evolution of Hk with time, as shown below. The fourth step involves updating the elements of Hk. This is performed as follows: Let ΔHk = Hk- Hk-1, implying Hk = Hk-1 + ΔHk (7) Where ΔHk represents the change in Hk-1, and is framed with the objective of neutralizing Δyk for known Xk. Thus, with reference to the basic form of Hk as expressed in equations (5.2.) and (5.2.1), one may write wherein, comprising differences between equation (8) and equations (5.2 & 5.2.1), one may write Equation (9) is used to update the value of H at every fourth time step. In a set of 4 time steps, let us call them 4n+l, 4n+2, 4n+3 and 4n+4, the values of x and Ay at 4n+l, 4n+2 and 4n+3 are used to create 3 sets of linear equations of size 3, to solve for the 9 elements of Ah using Singular Value Decomposition. Then the updated H, as shown in equation (7), is used in step 4n+4. The fifth and final step is a cyclical repetition of steps one to four as the evolution continues. Thus, from previous values of x and y and the measurements at current step, Taylor series extrapolations are used to update the second-order filtered value of y (equation 4) and estimate of x (equation 5.1.1). Next, the existing map H is used to correct the estimate of x from error is prediction of y (equations 5.1 - 5.5). Since y may be changing at orders higher than second, while x is constrained to evolve following a quadratic form, the difference is apportioned to an evolution of the mapping H. This is done by obtaining the third-order filter of y (equation 6), and transforming it into a corresponding delta-matrix of H (equations 7-9). It may be noted here that the value of Yk-1 shown in equation (4) is obtained as Yk-1 = Yk-1 + Δyk-1 (10) where right hand terms are obtained from equations (4) and (6) at the previous time steps. Further, there is a delay of around 36 seconds between the instant of emanation of gas molecules from liquid steel and its analysis at the gas analyzer. Hence, it is understood that the values of Xk obtained from yk at an instant, in reality represent the values of XK pertaining to 36 seconds earlier. This is the deviation from real time conditions, and is accounted for by making a second-order Taylor series extrapolation forward from currently obtained Xk with a time step size - Δt - of 36 seconds, thus using obtained values, apart from the current, at 36 and 72 seconds backward in equation (5.1.1). As mentioned earlier, the SEEMA process is set in motion by creating an initial mapping H between states xk and measurables yk. This mapping is generated from data on xk and yk acquired from the initiation stages of different process runs (heats) and stored in a relational data table. The data on yk obtained from gas analyzer at one second intervals is averaged over a substantial number (about 20) of process runs and its variation with time form the three axes of the table - one axis each for CO, CO2 and O2 - see Figure 3. In effect, on all three axes time increases linearly from starting instant to about 70 seconds, and at each point on an axis (i.e. instant of time) the obtained (average) value of the component of yk corresponding to that axis (say CO) is marked. It is not at all necessary that the measurable values also follow a monotonously linear trend. Next, the values of C, O and T, i.e. components of Xk, of liquid steel are sampled at different points in the first 3-4 minutes of a substantial (around 20) number of heats, and the variation with time at 1 second intervals is obtained by first fitting cubic splines against time, and then averaging across heats. The components of Xk at each instant are then inserted into the diagonal elements of the table (see Figure 3). The off-diagonal elements occupy null value. The basis heats from which averages are extracted and fitted are different for measurables and states. As may be appreciated from the tabular structure, the values of Xk and yk are linked to each other through time. Now, to initiate the process the values of XK and yk have to be obtained at four instances t1, t2, t3 and t4 corresponding to k = 1,.., 4. At an instant k, the values of yk are first read from the analyzer. Next, the position on each of the 3 axes corresponding to the 3 respective values of yk, i.e. the 3 Virtual times' are noted. To each of these 3 Virtual times' correspond one value (actually 3 values - elements of Xk) of vector Xk. The simple average across these 3 Virtual times' are taken as the representative values of Xk corresponding to the measured yk. WE CLAIM: 1. A method of producing ultra low carbon steel by refining molten low carbon steel in a vaccuum degassing process comprising the steps of: - bringing a cylindrical reaction chamber (5) with gas-suction plant (8) and two snorkels (6, 7) connected to the bottom of the chamber (5) close to a ladle (1) containing liquid bath (2) and slag layer (3) such that the two snorkels (6,7) are immersed well into the bath (2) to suck up some amount of molten metal into the chamber through the snorkels due to pressure difference between the reduced pressure of the chamber due to gas suction by the suction plant (8) in the chamber and atmospheric pressure of the ladle (1) such that height of metal in the ladle balances the said difference between the external atmospheric pressure and the internal chamber pressure, injecting inert argon gas into the up-snorkel (6) by means of a device (4) connected to the snorkel (6) to create a gas-liquid mixture (9) inside the chamber (5) and a gradual circulation of the liquid steel argon mixture laterally between the up and down snorkels in the chamber down the down snorkel and then laterally from down to up snorkel in the ladle due to difference to lower density of gas liquid mixture near the vicinity of the up- snorkel (6) and higher density of the molten steel when part of argon acquired in up-snorkel (6) is released from steel as it passes from up snorkel to down snorkel within the chamber thus allowing the molten metal (9) inside the chamber (5) to come back to the ladle through the second snorkel (7) and thus decarburizing the liquid steel by speeding up the reaction between the dissolved carbon of liquid metal and oxygen on producing carbon monoxide and partially carbon dioxide (11) and further decarburizing with injected oxygen from means (10) to produce ultra low carbon steel; characterized by controlling the carbon and oxygen concentration and temperature of the liquid steel during decarburization on predicted estimation of continuously sampled exhaust gases from the chamber (5) via exit gas measurement device based on mass spectroscopy to provide the flow rates of carbon monoxide, carbon dioxide and oxygen with a time delay corresponding to the instant of release of gas and its analysis in the gas analyzer; converting the obtained exit gas values on neutralizing time delay into instantaneous values of carbon, oxygen concentrations and temperature of the steel bath, in real time, said estimation during decarburization is carried out through free of emperical models State Estimation Using Evolutionary Mapping (SEEMA). ABSTRACT TITLE: A METHOD OF PRODUCING ULTRA LOW CARBON STEEL A method of producing ultra low carbon steel by refining molten low carbon steel in a vaccuum degassing process comprising the steps of: bringing a cylindrical reaction chamber (5) with gas-suction plant (8) and two snorkels (6, 7) connected to the bottom of the chamber (5) close to a ladle (1) containing liquid bath (2) and slag layer (3) such that the two snorkels (6,7) are immersed well into the bath (2) to suck up some amount of molten metal into the chamber through the snorkels due to pressure difference between the reduced pressure of the chamber due to gas suction by the suction plant (8) in the chamber and atmospheric pressure of the ladle (1) such that height of metal in the ladle balances the said difference between the external atmospheric pressure and the internal chamber pressure, injecting inert argon gas into the up-snorkel (6) by means of a device (4) connected to the snorkel (6) to create a gas-liquid mixture (9) inside the chamber (5) and a gradual circulation of the liquid steel argon mixture laterally between the up and down snorkels in the chamber down the down snorkel and then laterally from down to up snorkel in the ladle due to difference to lower density of gas liquid mixture near the vicinity of the up-snorkel (6) and higher density of the molten steel when part of argon acquired in up-snorkel (6) is released from steel as it passes from up snorkel to down snorkel within the chamber thus allowing the molten metal (9) inside the chamber (5) to come back to the ladle through the second snorkel (7) and thus decarburizing the liquid steel by speeding up the reaction between the dissolved carbon of liquid metal and oxygen on producing carbon monoxide and partially carbon dioxide (11) and further decarburizing with injected oxygen from means (10) to produce ultra low carbon steel; characterized by controlling the carbon and oxygen concentration and temperature of the liquid steel during decarburization on predicted estimation of continuously sampled exhaust gases from the chamber (5) via exit gas measurement device based on mass spectroscopy to provide the flow rates of carbon monoxide, carbon dioxide and oxygen with a time delay corresponding to the instant of release of gas and its analysis in the gas analyzer; converting the obtained exit gas values on neutralizing time delay into instantaneous values of carbon, oxygen concentrations and temperature of the steel bath, in real time, said estimation during decarburization is carried out through free of emperical models State Estimation Using Evolutionary Mapping (SEEMA).

Documents:

00470-kol-2005-description provision.pdf

00470-kol-2005-drawings.pdf

00470-kol-2005-form 1.pdf

00470-kol-2005-form 2.pdf

00470-kol-2005-form 3.pdf

470-KOL-2005-ABSTRACT.pdf

470-KOL-2005-CANCELLED PAGES.pdf

470-KOL-2005-CLAIMS.pdf

470-KOL-2005-CORRESPONDENCE.pdf

470-KOL-2005-DECISION.pdf

470-KOL-2005-DESCRIPTION (COMPLETE).pdf

470-KOL-2005-DRAWINGS.pdf

470-KOL-2005-EXAMINATION REPORT.pdf

470-KOL-2005-FORM 13.pdf

470-KOL-2005-FORM 18.pdf

470-KOL-2005-FORM 5.pdf

470-KOL-2005-GPA.pdf

470-KOL-2005-GRANTED-ABSTRACT.pdf

470-KOL-2005-GRANTED-CLAIMS.pdf

470-KOL-2005-GRANTED-DESCRIPTION (COMPLETE).pdf

470-KOL-2005-GRANTED-DRAWINGS.pdf

470-KOL-2005-GRANTED-FORM 1.pdf

470-KOL-2005-GRANTED-FORM 2.pdf

470-KOL-2005-GRANTED-FORM 3.pdf

470-KOL-2005-GRANTED-FORM 5.pdf

470-KOL-2005-GRANTED-SPECIFICATION-COMPLETE.pdf

470-KOL-2005-OTHERS.pdf

470-KOL-2005-PETITION UNDER RULE 137.pdf

470-KOL-2005-REPLY TO EXAMINATION REPORT.pdf

470-KOL-2005-SPECIFICATION-COMPLETE.pdf