EP4400607A1 - Intra-furnace state inference device, intra-furnace state inference method, and molten steel manufacturing method - Google Patents

Abstract

A furnace state estimation device (1) including: an input unit (11) configured to receive track record information and conditions of refining treatment; a model determination unit (14) configured to determine model parameters using past model parameters; a furnace state calculation unit (15) configured to calculate state quantities, including temperature and component concentration of molten metal and component concentration of slag, in a furnace, using the determined model parameters; and a model parameter calculation unit (13) configured to calculate model parameters in the refining treatment of the target charge, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace from beginning to end of a particular period of time in the refining treatment, using track record information including results of the refining treatment of the target charge.

Description

TECHNICAL FIELD
• The present disclosure relates to a furnace state estimation device, a furnace state estimation method, and a molten steel production method. The present disclosure relates, in particular, to a furnace state estimation device, a furnace state estimation method, and a molten steel production method by which component concentrations in molten metal and in slag in refining equipment in the iron and steel industry are estimated.
BACKGROUND
• At steelworks, the components and the temperature of hot metal tapped out of a blast furnace are adjusted in refining equipment, such as pretreatment equipment, converters, or secondary refining equipment. The converters are a process of blowing oxygen into the furnaces, so as to remove impurities from molten metal and raise the temperature, and they play a very important role, for example, in controlling steel quality and rationalizing refining costs. Here, in controlling the components and the temperature of molten metal in a converter, for example, the flow rate and the speed of top-blown oxygen, the height of a top-blowing lance, the flow rate of bottom-blown gas, or the like are used as manipulated variables. The charge amount and charge timing of auxiliary raw materials, such as lime or iron ore, the timing for sampling molten metal, the timing for terminating blowing, or the like are also used as manipulated variables. These manipulated variables are to be optimized according to the furnace state, such as the temperature and the components of molten metal or the components of slag. In order to estimate the furnace state with high accuracy and optimize the manipulated variables, a method has been proposed in which mass balance and heat balance in the furnace are calculated using measurement information regarding the refining equipment, including measured values of exhaust gas that can be continuously measured during refining treatment. This method is generally considered to be capable of estimating the temperature and the components of molten metal and the components of slag with high accuracy and in real time. However, the accuracy of the state estimation model often deteriorates due to changes in the environment of the refining equipment, such as wear of refractory materials in the furnace of the refining equipment, fluctuations in the composition of charged auxiliary raw materials, or degradation of measurement accuracy.
• As a method to optimize model parameters to maintain the accuracy of such a model, in an example, Patent Literature (PTL) 1 proposes a method of calculating values of parameters for an independent model formula for a converter, by extracting track records with similar treatment conditions from past track record information.
• In another example, Patent Literature (PTL) 2 proposes a method of determining multiple parameters for a molten metal temperature estimation model for secondary refining equipment, by finding approximate solutions to simultaneous equations set up so that heat balance is balanced.
CITATION LIST Patent Literatures
• PTL 1: JP 2005-036289 A
• PTL 2: JP 2004-360044 A
SUMMARY (Technical Problem)
• Here, the model formula in PTL 1 is intended to be applied to an independent physical reaction model or a linear combination model. It is therefore difficult to apply the technology of PTL 1 to models, such as state estimation models based on calculation of mass balance and heat balance in a furnace, in which the reaction amount and the heating level in the furnace interact with each other in a complex manner.
• PLT 2 proposes the method for finding approximate solutions to simultaneous equations set up so that heat balance is balanced for temperature estimation model formulas in which multiple model formulas and parameters interact with each other. PTL 2, however, does not describe calculation of mass balance in the furnace using measurement information, including measured values of exhaust gas, in calculating the reaction amount in the furnace. It is therefore difficult to apply the technology of PTL 2 to models in which heat balance and mass balance in a furnace interact with each other.
• Therefore, there is a need for a method of determining model parameters that is effective even for models in which mass balance and heat balance in a furnace interact with each other in a complex manner.
• It would be helpful to provide a furnace state estimation device, a furnace state estimation method, and a molten steel production method by which the component concentrations in molten metal and in slag can be estimated with high accuracy and continuously.
(Solution to Problem)
1. (1) A furnace state estimation device according to an embodiment of the present disclosure includes:
   • an input unit configured to receive track record information and conditions of refining treatment, the track record information including measurement results of temperature and component concentration of molten metal and component concentration of slag before start of or during the refining treatment in refining equipment and measurement results regarding the refining equipment that include flow rate and component concentration of exhaust gas discharged from the refining equipment;
   • a model determination unit configured to determine model parameters in the refining treatment of a target charge using past model parameters acquired from a database that stores model parameters for models related to blowing reaction in the refining equipment, the track record information, and the conditions of refining treatment;
   • a furnace state calculation unit configured to calculate state quantities, including temperature and component concentration of the molten metal and component concentration of the slag, in a furnace, using the determined model parameters; and
   • a model parameter calculation unit configured to calculate model parameters in the refining treatment of the target charge, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace from beginning to end of a particular period of time in the refining treatment, using the track record information including results of the refining treatment of the target charge.
2. (2) As an embodiment of the present disclosure, in (1),
   the model determination unit is configured to determine the model parameters in the refining treatment of the target charge, by averaging model parameters for past refining treatment wherein conditions are similar to the conditions of the refining treatment of the target charge among the past model parameters stored in the database.
3. (3) As an embodiment of the present disclosure, in (1),
   the model determination unit is configured to model relationship between the past model parameters stored in the database and the conditions of refining treatment, including counts of times the treatment is performed, dates and times of the treatment, and/or counts of times the refining equipment is used in the refining treatment, and determine the parameters for the refining treatment of the target charge based on a model.
4. (4) As an embodiment of the present disclosure, in any one of (1) to (3),
   the model parameters include a coefficient or a constant term that corrects an integrated amount of carbon charged into the furnace over a particular period of time, an integrated amount of carbon discharged out of the furnace over a particular period of time, an amount of oxygen charged into the furnace, an amount of oxygen discharged out of the furnace, an amount of oxygen used for oxidation of various metallic impurities in the molten metal, and/or an integrated amount of change in the temperature of the molten metal over a particular period of time due to changes in an amount of heat in the furnace.
5. (5) As an embodiment of the present disclosure, in any one of (1) to (4),
   the evaluation function is formed by a weighted sum of a term representing carbon balance, a term representing oxygen balance, and a term representing heat balance.
6. (6) A furnace state estimation method according to an embodiment of the present disclosure is
   a furnace state estimation method to be executed by a furnace state estimation device, the furnace state estimation method including:
   • an input step of receiving track record information and conditions of refining treatment, the track record information including measurement results of temperature and component concentration of molten metal and component concentration of slag before start of or during the refining treatment in refining equipment and measurement results regarding the refining equipment that include flow rate and component concentration of exhaust gas discharged from the refining equipment;
   • a model determination step of determining model parameters in the refining treatment of a target charge using past model parameters acquired from a database that stores model parameters for models related to blowing reaction in the refining equipment, the track record information, and the conditions of refining treatment;
   • a furnace state calculation step of calculating state quantities, including temperature and component concentration of the molten metal and component concentration of the slag, in a furnace, using the determined model parameters; and
   • a model parameter calculation step of calculating model parameters in the refining treatment of the target charge, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace from beginning to end of a particular period of time in the refining treatment, using the track record information including results of the refining treatment of the target charge.
7. (7) As an embodiment of the present disclosure, in (6),
   the model determination step is configured to determine the model parameters in the refining treatment of the target charge, by averaging model parameters for past refining treatment wherein conditions are similar to the conditions of the refining treatment of the target charge among the past model parameters stored in the database.
8. (8) As an embodiment of the present disclosure, in (6),
   the model determination step is configured to model relationship between the past model parameters stored in the database and the conditions of refining treatment, including counts of times the treatment is performed, dates and times of the treatment, and/or counts of times the refining equipment is used in the refining treatment, and determine the parameters for the refining treatment of the target charge based on a model.
9. (9) A molten steel production method according to an embodiment of the present disclosure includes
   producing molten steel, by determining flow rate and speed of top-blown oxygen, height of a top-blowing lance, flow rate of bottom-blown gas, charge amount and charge timing of auxiliary raw materials, such as lime or iron ore, timing for sampling molten metal, and/or timing for terminating blowing, based on the temperature and the component concentration of the molten metal and the component concentration of the slag estimated by the furnace state estimation method according to any one of (6) to (8), and performing refining operation.
(Advantageous Effect)
• According to the present disclosure, the furnace state estimation device, the furnace state estimation method, and the molten steel production method by which the component concentrations in molten metal and in slag can be estimated with high accuracy and continuously are provided.
BRIEF DESCRIPTION OF THE DRAWINGS
• In the accompanying drawings:
• FIG. 1 is a schematic diagram illustrating a configuration of a furnace state estimation device according to an embodiment of the present disclosure;
• FIG. 2 illustrates an example configuration of a database; and
• FIG. 3 is a flowchart illustrating processing of a furnace state estimation method according to an embodiment of the present disclosure.
DETAILED DESCRIPTION
• Hereinafter, a furnace state estimation device, a furnace state estimation method, and a molten steel production method according to an embodiment of the present disclosure will be described with reference to the drawings.
[Configuration of Furnace State Estimation Device]
• FIG. 1 is a schematic diagram illustrating a configuration of a furnace state estimation device 1 according to an embodiment of the present disclosure. In the present embodiment, the furnace state estimation device 1 is used as part of equipment for producing molten steel in the iron and steel industry. The equipment for producing molten steel includes refining equipment 2, and a blowing control system that includes the furnace state estimation device 1.
• As illustrated in FIG. 1, the refining equipment 2 includes a converter 100, a lance 102, and a duct 104. The lance 102 is arranged above molten metal 101 in the converter 100. High-pressure oxygen is ejected from a tip of the lance 102 toward the molten metal 101 located below. The high-pressure oxygen oxidizes impurities in the molten metal 101 and incorporates them into slag 103 (refining treatment). The duct 104 is a flue system for guiding exhaust gas, and it is provided above the converter 100.
• An exhaust gas detection unit 105 is arranged in the duct 104. The exhaust gas detection unit 105 detects the flow rate and the component concentration of exhaust gas (e.g., concentrations of CO, CO₂, O₂, N2, Ar, or the like) discharged in the refining treatment. As exhaust gas measurement, the exhaust gas detection unit 105 measures the flow rate of exhaust gas in the duct 104, based on, for example, the differential pressure before and after a Venturi tube disposed in the duct 104. The exhaust gas detection unit 105 also measures the concentration [%] of each component in exhaust gas as exhaust gas measurement. The flow rate and the component concentration of exhaust gas are measured, for example, in cycles of several seconds. Signals that indicate results of detection of the exhaust gas detection unit 105 are transmitted to a control terminal 10.
• Stirring gas is blown into the molten metal 101 in the converter 100 through vent holes 106 formed in a bottom of the converter 100. The stirring gas is an inert gas, such as Ar. The blown stirring gas agitates the molten metal 101 and promotes reaction between the high-pressure oxygen and the molten metal 101. A flowmeter 107 measures the flow rate of the stirring gas blown into converter 100. The temperature and the component concentration of the molten metal 101 are analyzed immediately before blowing starts and after it ends. The temperature and the component concentration of the molten metal 101 are measured once or multiple times in the middle of the blowing, and based on the measured temperature and component concentration, the supply amount (oxygen feed amount) and the speed (oxygen feed speed) of the high-pressure oxygen, the flow rate (stirring gas flow rate) of the stirring gas, or the like are determined.
• The blowing control system includes the control terminal 10, a display device 20, and the furnace state estimation device 1 as main components. The control terminal 10 may be configured by an information processing device, such as a personal computer or a workstation. The control terminal 10 controls the oxygen feed amount, the oxygen feed speed, and the stirring gas flow rate so that the temperature and the component concentration of the molten metal 101 fall within a desired range. The control terminal 10 also collects data on track records of the oxygen feed amount, the oxygen feed speed, and the stirring gas flow rate. The display device 20 may be configured by a liquid crystal display (LCD) or cathode ray tube (CRT) display, for example. The display device 20 may display calculation results or the like that are output from the furnace state estimation device 1.
• The furnace state estimation device 1 is a device that estimates the temperature and the component concentration of the molten metal 101 treated in the refining equipment 2 and the component concentration of the slag 103. The furnace state estimation device 1 is configured by an information processing device, such as a personal computer or a workstation. The furnace state estimation device 1 includes an input unit 11, a database 12, a model parameter calculation unit 13, a model determination unit 14, a furnace state calculation unit 15, and an output unit 16.
• The input unit 11 is an input interface that receives track record information (track record data), which indicates results of various measurements related to the refining equipment 2, or the like. The input unit 11 may include, for example, a keyboard, a mouse, a pointing device, a data receiver, and/or a graphical user interface (GUI). In the present embodiment, the input unit 11 receives track record information, parameter setting values, or the like from the outside and writes the information into the database 12 and transmits it to the furnace state calculation unit 15. The track record information is input into the input unit 11 from the control terminal 10. The track record information includes information regarding the flow rate and the component concentration of exhaust gas that have been measured by the exhaust gas detection unit 105, information on the oxygen feed amount and the oxygen feed speed, information on the stirring gas flow rate, information on the charge amounts of raw materials (main raw materials and auxiliary raw materials), the temperature and the component concentration of the molten metal 101, the component concentration of the slag 103, or the like. The above information corresponds to category 1 through category M in the track record information of FIG. 2, which will be described later. Additionally, the input unit 11 can be used for manual data input (manual input) by, for example, an operator of the refining equipment 2. Parameter setting values for a model formula (hereafter referred to simply as a "model") may be input manually. In the present embodiment, the input unit 11 also receives conditions of the refining treatment and information on manipulated variables that will be described later. Additionally, the input unit 11 may acquire the track record information or the like before the start of, during, or after the end of the refining treatment.
• The database 12 stores information on models related to blowing reaction in the refining equipment 2, track record information on refining treatment, and calculation results by the furnace state estimation device 1. The database 12 is configured, for example, by a storage device, such as memory or a hard disk drive. The storage device may further store computer programs. The database 12 stores model formulas and parameters for the model formulas (hereinafter, "model parameters") as the information on the models related to the blowing reaction. The model parameters are calculated by the model parameter calculation unit 13. The database 12 may also store various types of information input through the input unit 11 and results of calculation and analysis in blowing track records that have been calculated by the furnace state calculation unit 15.
• FIG. 2 illustrates an example configuration of the database 12. In the present embodiment, the database 12 stores conditions, track record information, and model parameters, which are calculation results, for N times (N charges) of refining treatment, in association with identification numbers of the charges. N is, for example, an integer greater than or equal to 2. In the example of FIG. 2, the leftmost column indicates the identification number of the charge. For example, in the N^th refining treatment, the database 12 has stored track record information and model parameters for the past N-1 times of refining treatment. When the model determination unit 14 determines model parameters in the N^th refining treatment as will be described later, the information from the past N-1 times of refining treatment stored in the database 12 is used as a candidate. When the N^th refining treatment ends, track record information and calculation results for the N^th refining treatment are added to the database 12 (refer to the area inside the bold lines in FIG. 2). After that, when the model determination unit 14 determines model parameters in the N+1^th refining treatment, the information from the past N times of refining treatment stored in the database 12 is used as a candidate.
• The model parameter calculation unit 13, the model determination unit 14, and the furnace state calculation unit 15 are configured by an arithmetic processing device, such as a CPU. The model parameter calculation unit 13, the model determination unit 14, and the furnace state calculation unit 15 may be realized, for example, by the arithmetic processing device reading and executing a computer program. The model parameter calculation unit 13, the model determination unit 14, and the furnace state calculation unit 15 may include dedicated arithmetic devices or arithmetic circuits.
• Based on mass balance and heat balance in the furnace, the model parameter calculation unit 13 calculates model parameters for a model related to blowing reaction so that a balance error is minimized, and stores them in the database 12. After one refining treatment ends, the model parameter calculation unit 13 calculates mass balance and heat balance using track record information, that is, results of the refining treatment.
• The calculation of mass balance refers to calculating the charge amount of each component into the converter 100 and the discharge amount of each component from the converter 100. The charge amount of each component is calculated from the amounts of main raw materials and auxiliary raw materials charged into the converter 100, oxygen supplied from the lance 102, and the amount of air entrained from outside the converter 100. The discharge amount of each component is calculated from the exhaust gas flow rate and the component concentration of the exhaust gas.
• The calculation of heat balance refers to calculating heat input and heat output in the furnace of the converter 100. The heat input is calculated from sensible heat of main raw materials inserted into the converter 100, reaction heat produced by reaction occurring in the furnace, heat of dissolution of auxiliary raw materials charged into the converter 100, or the like. The heat output is calculated from heat released from a furnace body surface, radiant heat from a furnace throat, heat released by the stirring gas, the slag 103 discharged out of the furnace, sensible heat of exhaust gas, or the like.
• The model determination unit 14 acquires past model parameters stored in the database 12. The model determination unit 14 determines model parameters to be used in the furnace state calculation unit 15 using the past model parameters and transmits them to the furnace state calculation unit 15.
• Based on the model parameters that have been determined by the model determination unit 14, track record information and parameter setting values that have been collected by the input unit 11, or the like, the furnace state calculation unit 15 calculates (estimates) state quantities, including the temperature and the component concentration of the molten metal 101 and the component concentration of the slag 103, in the converter 100. The estimated state quantities in the converter 100 are transmitted to the output unit 16.
• The output unit 16 transmits, to the control terminal 10, the state quantities in the converter 100 that have been calculated by the furnace state estimation device 1. In refining treatment, various manipulated variables are determined and operating conditions are changed, based on calculation results output from the furnace state estimation device 1. The output unit 16 also has the function of transmitting the information calculated by the furnace state estimation device 1 to the display device 20, so that the calculation results output from the furnace state estimation device 1 can be displayed.
• The furnace state estimation device 1 with the above configuration estimates the state quantities, including the temperature and the component concentration in the molten metal 101 and the component concentration in the slag 103, in the converter 100 with high accuracy, by executing processing of a furnace state estimation method that will be described below. Operations performed by the furnace state estimation device 1 to execute the furnace state estimation method will be described below with reference to the flowchart of FIG. 3.
[Furnace State Estimation Method]
• FIG. 3 is the flowchart illustrating processing of the furnace state estimation method according to an embodiment of the present disclosure. The flowchart of FIG. 3 starts at any time before refining treatment is started. That is, furnace state estimation processing proceeds to Step S1 at any time before the refining treatment is started.
• In processing of Step S1, the input unit 11 acquires conditions of the refining treatment. In the present embodiment, the conditions of the refining treatment include the form of refining, the amount of auxiliary raw materials to be charged, target values of the component concentrations and the temperatures of the molten metal 101 and the slag 103, the count of times the treatment is performed, the date and time of the treatment, and the count of times the equipment, including the furnace, the lance, and the measuring devices, is used. The input unit 11 transmits the acquired conditions of the refining treatment to the database 12 and the furnace state calculation unit 15. This completes the processing of Step S1, and the furnace state estimation processing proceeds to Step S2. Step S1 corresponds to part of an "input step." The data input in Step S1 is to be used in processing performed by the model determination unit 14.
• In processing of Step S2, the model determination unit 14 determines model parameters to be used in the furnace state calculation unit 15 using past model parameters stored in the database 12, based on the conditions of the refining treatment. Step S2 corresponds to a "model determination step." In detail, the model parameters determined in the processing of Step S2 are obtained by calculation or selection, based on the model parameters corresponding to past refining treatment that are already stored in the database 12. As described above, for example, in the N^th refining treatment, the model determination unit 14 determines model parameters using information from the past N-1 times of refining treatment stored in the database 12. In a case in which model parameters or data on conditions of refining treatment are not available for the past N-1 times because, for example, it takes time to obtain results of the refining treatment, the model parameters may be determined by extracting track record information for which necessary model parameters or data on conditions of refining treatment are available. The N^th refining treatment in this case, that is, the currently executed refining treatment, may be referred to as the refining treatment of the target charge.
• For example, the model determination unit 14 extracts, from the model parameters stored in database 12, model parameters wherein the conditions of the refining treatment are similar to the conditions of the target charge, and determines them by averaging the extracted model parameters. The model determination unit 14 may perform averaging, by extracting only model parameters in the most recent predetermined count of charges, that is, by excluding older model parameters from the extraction. The similarity (Ds) between the conditions of the refining treatment of the target charge and the past track records can be evaluated, for example, by calculating the Euclidean distance as illustrated in the following Formula (1).
  [Math. 1] $$\mathrm{Ds}=\sqrt{{\displaystyle {\sum }_k\frac{{\left({\mathrm{CA}}_k-{\mathrm{CP}}_k\right)}^2}{{{\mathrm{<figure-callout\; id="2"\; label="G\; k"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">G</figure-callout>}}_k}^2}}}$$

  
• In the formula, k denotes the count of conditions of the refining treatment. CA_k denotes conditions in the past track records. CP_k denotes conditions of the refining treatment of the target charge. G_k is a parameter used to weight the conditions of each type of refining treatment. The conditions of the refining treatment include, for example, the date and time of the refining treatment, the weight of loaded hot metal, the weight of loaded scrap, the temperature of the hot metal, the component concentrations of C, Si, Mn, P, and other components in the hot metal, the count of times the refining furnace and the top-blowing lance are used, or the like. Other examples include the temperature of molten metal after treatment in the immediately preceding refining treatment and time elapsed since the refining treatment, the weight and the components of slag to be carried over, the weight of each brand of auxiliary raw material and the weight of each brand of scrap to be charged before the refining treatment is started, and the like. These conditions correspond to category 1 through category L in the conditions of refining treatment of FIG. 2. To evaluate the similarity, only track records with a matching form of refining furnace used, a matching form of top-blowing lance, a matching form of bottom-blowing nozzle, or the like may be covered in the evaluation. Here, the similarity may be evaluated by not only the Euclidean distance illustrated in Formula (1) but also methods, including the city block distance, the Minkowski distance, the Mahalanobis distance, and the cosine similarity, that evaluate the distance between k-dimensional vectors. Here, high similarity is synonymous with a short distance between the computed k-dimensional vectors. In extracting past track records of refining treatment, track records wherein the calculated similarity is higher than a set threshold may be extracted, or any count of past track records with the highest similarity may be extracted. As a method of extracting similar track records, for each category of k conditions of the refining treatment, the difference between the conditions of the treatment to be calculated and the conditions of the past track records can be calculated, and track records for which the k differences are smaller than individually set thresholds may be extracted.
• The model determination unit 14 may also model the relationship between the model parameters stored in the database 12 and conditions of refining treatment, including counts of times the treatment is performed, the dates and times of the treatment, and counts of times the refining equipment, including the furnace, the lance, and measuring devices, is used in the refining treatment. The model determination unit 14 may then calculate optimal parameters by conducting model calculation from input values for the conditions of the refining treatment of the target charge. The model determination unit 14 transmits the determined model parameters to the furnace state calculation unit 15. This completes the processing of Step S2, and the furnace state estimation processing proceeds to Step S3.
• The processing of Step S3 and Step S4 is started at a time when one time of refining treatment starts and is repeated at any cycle during the refining treatment. In the processing of Step S3, the input unit 11 acquires information on manipulated variables in the refining treatment and measurement information in the converter 100. The information on manipulated variables is, for example, information on manipulated variables, such as the height of the lance 102, the oxygen feed speed, the stirring gas flow rate, or the charge amount of auxiliary raw materials. The measurement information indicates, for example, measured values of the flow rate, the component concentration, or the like of exhaust gas. Here, the measured values are not limited to measured values themselves, but may also include results obtained after analysis (analytical values). Information on manipulated values and measurement information are collected at any cycle. In a case in which there is a large time lag between the information on manipulated values and the measurement information, data is created with the time lag being considered. In a case in which the measurement information contains a lot of noise, the measured values may be replaced by values obtained by smoothing processing, such as moving average calculation. Step S3 corresponds to part of the "input step." The data input in Step S3 is used in processing performed by the furnace state calculation unit 15.
• In the processing of Step S4, the furnace state calculation unit 15 calculates the state quantities in the converter 100 using the information acquired by the input unit 11 and a model with the model parameters determined by the model determination unit 14. Examples of state quantities include the carbon concentration in the molten metal 101, the FetO concentration in the slag 103, and the like. Step S4 corresponds to a "furnace state calculation step."
• The carbon concentration in the molten metal 101 is determined, for example, by calculating the amount of carbon remaining in the converter 100. The amount of carbon charged into the converter 100 and the amount of carbon discharged out of the converter 100 can respectively be expressed as illustrated in Formula (2) and Formula (3) below. The carbon concentration in the molten metal 101 can be calculated, by assuming that the amount of carbon remaining in the converter 100 is equivalent to the amount of carbon in the molten metal 101, wherein the amount of carbon remaining in the converter 100 is obtained by subtracting the discharge amount of carbon from the charge amount of carbon. Here, the amounts of carbon charged and discharged to and from the molten metal 101 are assumed to be much smaller than the total load amount. Additionally, unless otherwise specified, "%" represents "mass%", and various flow rates are represented in basic units of flow rate.
  [Math. 2] $$C_{\mathit{in}}=\frac{\left({\rho }_{\mathit{pig}}\times W_{\mathit{pig}}+{\displaystyle {\sum }_i\left({\rho }_i^{\mathit{Cscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle {\sum }_j\left({\rho }_j^{\mathit{Caux}}\times W_j^{\mathit{aux}}\right)}}\right)}{W_{\mathit{charge}}}$$

  

  
  [Math. 3] $$C_{\mathit{out}}=\left(\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO\; OG"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{\mathit{CO}\mathit{OG}}^{\mathit{}}}{2}+\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{{\mathit{CO}}_{}}^{{\mathit{}}_2}}{2}\right)\times \left(12/11.2\right)/10$$

  
• Here, Cin [%], which is the charge amount of carbon, denotes a value obtained by converting the sum of the amount of carbon in the main raw materials and the amount of carbon in the charged auxiliary raw materials into the concentration in the molten metal 101. ρ_pig [%] denotes the carbon concentration in the loaded hot metal. ρ_i ^Cscr [%] denotes the carbon concentration in loaded scrap (brand i). ρ_j ^Caux [%] denotes the carbon concentration in a loaded auxiliary raw material (brand j). W_pig [t] denotes the weight of the loaded hot metal. W_i ^scr [t] denotes the weight of loaded scrap (brand i). W_j ^aux [t] denotes the integrated weight of the loaded auxiliary raw material (brand j). W_charge [t] denotes the weight of molten metal loaded into the converter 100. The carbon concentrations (ρ_i ^Cscr, ρ_j ^Caux) in the loaded scrap of brand i and the charged auxiliary raw material of brand j are stored in the database 12, and the furnace state calculation unit 15 acquires information regarding the brands used in the target charge. C_out [%], which is the discharge amount of carbon, is a value obtained by converting the amount of carbon contained in the exhaust gas into the concentration in the molten metal 101. V_CO ^OG [Nm³/t] and V_CO2 ^OG [Nm³/t] respectively denote the integrated flow rate of CO and the integrated flow rate of CO₂ in the exhaust gas up to the time of calculation.
• The FetO concentration in the slag 103 can be calculated, by assuming that the amount of oxygen remaining in the converter 100 is equivalent to an amount obtained by subtracting the discharge amount of oxygen from the charge amount of oxygen. For example, the amount of oxygen charged into the converter 100 and the amount of oxygen discharged out of the converter 100 can respectively be expressed as illustrated in the following Formula (4) and Formula (5).
  [Math. 4] $${\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{in}}={\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}+{\displaystyle {\sum }_i\left({\rho }_i^{\mathit{Oscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle {\sum }_j\left({\rho }_j^{\mathit{Oaux}}\times W_j^{\mathit{aux}}\right)+\frac{21}{79}\times \left(\begin{array}{l}{V}_{\mathit{rem}}^{\mathit{OG}}-\\ V_{\mathit{bot}}\end{array}\right)}}$$

  

  
  [Math. 5] $${\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{out}}=\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO\; OG"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{\mathit{CO}\mathit{OG}}^{\mathit{}}}{2}+{\mathrm{<figure-callout\; id="2"\; label="V\; CO"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{{\mathit{CO}}_{}}^{{\mathit{}}_2}+{\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}$$

  
• Here, O₂ ⁱⁿ [Nm³/t], which is the charge amount of oxygen, denotes the sum of the integrated amount of top-blown oxygen V_O2 ^blow [Nm³/t] from the lance 102, the integrated amount of oxygen in the charged auxiliary raw materials, and the integrated amount of oxygen in air entrained into the furnace from the outside of the converter 100. ρ_i ^Oscr [%] denotes a converted value of oxygen content in the loaded scrap (brand i). ρ_j ^Oaux [(Nm³/t)/t] denotes a converted value of oxygen content in the charged auxiliary raw material (brand j). The oxygen contents (ρ_i ^Oaux, ρ_j ^Oaux) in the loaded scrap of brand i and in the charged auxiliary raw material of brand j are stored in the database 12, and the furnace state calculation unit 15 acquires information regarding the brands used in the target charge. For ρ_i ^Oscr [%] and W_j ^aux [t], analytical values or calculated values regarding the components and the weight of the slag 103 carried over from the previous charge may be included. In a case in which the N₂ concentration and the Ar concentration cannot be obtained as exhaust gas measurements in calculating the charge amount of oxygen, as in the present embodiment, for example, the amount of oxygen in the entrained air may be calculated as in the fourth term of Formula (4). Here, in the fourth term of Formula (4), it is assumed that the amount of N² and Ar in the entrained air is equivalent to an amount obtained by subtracting V_bot [Nm³/t], which is the bottom-blown gas flow rate, from V_rem ^OG [Nm³/t], which is the amount of unanalyzed exhaust gas other than O₂, CO, and CO₂ in the exhaust gas.
• O₂ ^out [Nm³/t], which is the discharge amount of oxygen, is calculated from the amount of oxygen contained in the exhaust gas. V_O2 ^OG [Nm³/t] is the integrated flow rate of O₂ in the exhaust gas up to the time of calculation. V_CO ^OG [Nm³/t] and V_CO2 ^OG [Nm³/t] are the same as in Formula (3). The amount of oxygen remaining in the converter 100 is an amount obtained by subtracting the discharge amount of oxygen from the charge amount of oxygen. The oxygen remaining in the converter 100 is used for oxidation of metallic impurities, such as Si, Mn, or P, in the molten metal 101 and for oxidation of iron. Among them, the amount of oxidation of the metallic impurities is calculated using an oxidation reaction model for impurity metals among models stored in the database 12. For example, V_O2 ^Si [Nm³/t], which is the amount of oxygen used for Si oxidation in the molten metal 101, is expressed as in the following Formula (6).
  [Math. 6] $$\begin{array}{l}{\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}=\frac{{\rho }_{\mathit{pig}}^{\mathit{Si}}\times W_{\mathit{pig}}+{\displaystyle {\sum }_i\left({\rho }_i^{\mathit{Siscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle {\sum }_j\left({\rho }_j^{\mathit{Siaux}}\times W_j^{\mathit{aux}}\right)}}}{W_{\mathit{charge}}}\times \mathit{exp}\left(-K_{\mathit{Si}}\times {\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}\right)\times \\ \left(22.4/28\right)\times 10\end{array}$$

  
• Here, ρ_pig ^Si [%] denotes the Si concentration in the loaded hot metal. ρ_i ^Siscr [%] denotes the Si concentration in the loaded scrap (brand i). ρ_j ^Siaux [%] denotes the Si concentration in the charged auxiliary raw material (brand j). K_Si denotes an oxidation reaction rate constant for Si. As is the case with Formula (6), the amount of oxygen used for oxidation of various metallic impurities, such as Mn or P, in the molten metal 101 can be calculated. Here, it is assumed that the total amount of oxygen used for oxidation of various metallic impurities, such as Si, Mn, or P, in the molten metal 101 is V_O2 ^met [Nm³/t], The amount of FetO in the slag 103 can be calculated, by assuming that it is equivalent to an amount obtained by subtracting the discharge amount of oxygen and V_O2 ^met from the charge amount of oxygen.
• At a time when one time of refining treatment (refining treatment of the aforementioned target charge) ends, the processing of Step S3 and Step S4 ends (Yes in Step S5), and the furnace state estimation processing proceeds to Step S6. When one time of refining treatment has not ended (No in Step S5), the furnace state estimation processing returns to Step S3 and Step S4.
• In processing of Step S6, the input unit 11 acquires results of the refining treatment as track record information. In the present embodiment, the results of the refining treatment include the temperature and the component concentration of the molten metal 101, the component concentration of the slag 103, and the flow rate and the component concentration of exhaust gas. The input unit 11 stores the obtained results of the refining treatment in the database 12. This completes the processing of Step S6, and the furnace state estimation processing proceeds to Step S7. Step S6 corresponds to part of the "input step." The data input in Step S6 is to be used in processing performed by the model parameter calculation unit 13.
• In processing of Step S7, based on mass balance and heat balance in the furnace, the model parameter calculation unit 13 calculates model parameters for a model related to blowing reaction so that a balance error is minimized, and stores them in the database 12. Step S7 corresponds to a "model parameter calculation step." As described above, the furnace state calculation unit 15 estimates the state quantities in the converter 100 in the refining treatment of the subject charge using the model parameters determined by the model determination unit 14. The model parameter calculation unit 13 corrects the model parameters that have been used by the furnace state calculation unit 15 using the results (track record information) on the refining treatment of the target charge. The model parameter calculation unit 13 then stores the corrected and more accurate model parameters in the database 12. In other words, the model parameters that are stored in the database 12 in association with the target charge are not the model parameters of the model that have been used for calculation (estimation) by the furnace state calculation unit 15. The model parameters stored in the database 12 in association with the target charge are the model parameters that have been calculated (corrected) by the model parameter calculation unit 13 based on the track record information on the refining treatment of the target charge.
• The model parameter calculation unit 13 may calculate coefficients for correction as the model parameters. The coefficients for correction may include, for example, a correction coefficient A for a measured value of the exhaust gas flow rate and a correction coefficient B for a measured value of the component concentration of the exhaust gas. The coefficients for correction may include, for example, a correction coefficient ΔC for a measured value of the component concentration of the molten metal 101, a correction coefficient D for a measured value of the temperature of the molten metal 101, a constant E related to the yield rate of reaction of the loaded scrap in the furnace, and a constant F related to the yield rate of reaction of the charged auxiliary raw material in the furnace. The coefficients for correction may include coefficients H for an increased amount of heat and an absorbed amount of heat associated with various reactions, such as oxidation reactions of components in the molten metal 101, reduction reactions of components in the slag 103, or dissolution of auxiliary raw materials, in the furnace. Furthermore, the coefficients for correction may include a coefficient I for heat loss, such as sensible heat of gas and the slag 103, or the amount of heat radiation from the furnace throat and the furnace body.
• The model parameter calculation unit 13 may incorporate the above coefficients as variables in an evaluation function as expressed by, for example, Formula (7), and find model parameters that minimize the evaluation function. Here, although in the present embodiment the model parameter calculation unit 13 minimizes the evaluation function, an evaluation function that is maximized by suitable model parameters may also be used. That is, the model parameter calculation unit 13 may find model parameters that minimize or maximize the evaluation function.
  [Math. 7] $$\mathrm{J}=\frac{{\left\{C_{\mathit{in}}-C_{\mathit{out}}-\left[\mathrm{<figure-callout\; id="2"\; label="C"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">C</figure-callout>}\right]\right\}}^2}{{{\mathrm{<figure-callout\; id="2"\; label="\sigma \; C"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_C}^2}+\frac{{\left\{{\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{in}}-{\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{out}}-{\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}-{\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}\right\}}^2}{{{\mathrm{<figure-callout\; id="2"\; label="\sigma \; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_O}^2}+\frac{{\left\{T_{\mathit{fin}}-\left(T_{\mathit{ini}}+D\right)-\mathrm{<figure-callout\; id="2"\; label="\Delta \; T"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\Delta </figure-callout>}T\right\}}^2}{{{\sigma }_T}^2}$$

  

  $$C_{\mathit{in}}=\frac{\left({\rho }_{\mathit{pig}}\times W_{\mathit{pig}}+{\mathit{\Delta C}}_C+{\displaystyle {\sum }_i\left(E_i\times {\rho }_i^{\mathit{Cscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle {\sum }_j\left(F_j\times {\rho }_j^{\mathit{Caux}}\times W_j^{\mathit{aux}}\right)}}\right)}{W_{\mathit{charge}}}$$

  

  $$C_{\mathit{out}}=\mathrm{A}\times \mathrm{B}\times \left(\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO\; OG"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{\mathit{CO}\mathit{OG}}^{\mathit{}}}{2}+\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{{\mathit{CO}}_{}}^{{\mathit{}}_2}}{2}\right)\times \left(12/11.2\right)/10$$

  

  $${\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{in}}={\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}+{\displaystyle \sum _i\left(E_i\times {\rho }_i^{\mathit{Oscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle \sum _j\left(F_j\times {\rho }_j^{\mathit{Oaux}}\times W_j^{\mathit{aux}}\right)}}$$

  

  $${\mathrm{<figure-callout\; id="2"\; label="O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">O</figure-callout>}}_2^{\mathit{out}}=\mathrm{A}\times \mathrm{B}\times \left(\frac{{\mathrm{<figure-callout\; id="2"\; label="V\; CO\; OG"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{\mathit{CO}\mathit{OG}}^{\mathit{}}}{2}+{\mathrm{<figure-callout\; id="2"\; label="V\; CO"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{{\mathit{CO}}_{}}^{{\mathit{}}_2}\right)+A\times {\mathrm{<figure-callout\; id="2"\; label="V\; O"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">V</figure-callout>}}_{O_{}}^{{}_2}$$

  

  $$V_{O_{}}^{}$$ $${{}_2}_{\mathit{Si}}^{}=\frac{{\rho }_{\mathit{pig}}^{\mathit{Si}}\times W_{\mathit{pig}}+{\mathit{\Delta C}}_{\mathit{Si}}+{\displaystyle {\sum }_i\left(E_i\times {\rho }_i^{\mathit{Siscr}}\times W_i^{\mathit{scr}}\right)+{\displaystyle {\sum }_j\left(F_j\times {\rho }_j^{\mathit{Siaux}}\times W_j^{\mathit{aux}}\right)}}}{W_{\mathit{charge}}}$$

  

  $$\mathrm{\Delta }T={\displaystyle \sum _m\left(H_m\times {\mathit{\Delta R}}_m\right)}-{\displaystyle \sum _n\left(I_n\times {\mathit{\Delta L}}_n\right)}$$

  
• Cin denotes the integrated amount of carbon charged into the converter 100 over a particular period of time. C_out denotes the integrated amount of carbon discharged out of the converter 100 by exhaust gas or the like over a particular period of time. O₂ ⁱⁿ denotes the amount of oxygen charged into the converter 100. O₂ ^out denotes the amount of oxygen discharged out of the converter 100 due to discharge of exhaust gas, the slag 103, or the like. V_O2 ^met denotes the amount of oxygen used for oxidation of various metallic impurities, such as Si, Mn, or P, in the molten metal 101. ΔT denotes the integrated amount of change in the temperature of molten metal over a particular period of time due to changes in the amount of heat in the converter 100, including reaction heat generated by reaction in the converter 100, heat released by exhaust gas, the slag 103, or the like discharged out of the converter, heat radiation from the furnace body, and radiant heat from the furnace throat. T_ini denotes a measured value of the temperature of the molten metal 101 at the beginning of a particular period of time in the refining treatment. [C], V_O2 ^FetO, and T_fin respectively denote a measured value of the amount of carbon in the molten metal 101, the amount of oxygen used to produce FetO calculated from a measured value of the amount of FetO in the slag 103, and a measured value of the temperature of the molten metal 101, at the end of a particular period of time in the refining treatment. σ_C ², σ_O ², and σ_T ² are constants that can be freely set. A through I and ΔC correspond to parameter 1 through parameter K of FIG. 2. In the present embodiment, the model parameters include a coefficient or a constant term that corrects the integrated amount and/or the amount of oxygen used in Formula (7).
• Note that ΔR_m denotes the amount of reaction in various reactions m, such as oxidation reactions of components in the molten metal 101, reduction reactions of components in the slag 103, or dissolution of auxiliary raw materials, in the furnace. ΔL_n denotes the amount of heat loss, such as sensible heat of gas and the slag 103 or the amount of heat radiation from the furnace throat and the furnace body, on a heat loss path n in the molten metal 101.
• The evaluation function J of Formula (7) is a weighted sum of the following three terms. In the evaluation function J, the first term and the second term represent a mass balance error, and the third term represents a heat balance error. The first term is a square value of the difference between the amount of carbon remaining in the converter 100, and a measured value of the amount of carbon in the molten metal 101, wherein the amount of carbon remaining in the converter 100 is obtained by subtracting the discharge amount of carbon from the charge amount of carbon. When this term is zero, it indicates that carbon balance is maintained in the converter 100. The second term is a square value of the difference between an amount obtained by subtracting the discharge amount of oxygen and the amount of oxygen used for oxidation of impurity metals from the charge amount of oxygen, and the amount of oxygen used for iron oxidation in the molten metal 101 that is calculated from a measured value of FetO in the slag 103. When this term is zero, it indicates that oxygen balance is maintained in the converter 100. The third term is a square value of the difference between a measured value of the amount of change in the temperature of the molten metal 101 from the beginning to the end of the particular period of time in the refining treatment, and a calculated value of the amount of change in the temperature of the molten metal 101 that is calculated from reaction heat, heat releasing, or the like in the converter 100. When this term approaches zero, it indicates that heat balance is maintained in the converter 100. The "particular period of time" in the above explanation of Formula (7) may be a different period of time for each of the three terms.
• The weighting factors (σ_C ², σ_O ², σ_T ²) in denominators of the respective terms of the evaluation function J are set by, for example, a user. Various algorithms have been proposed for nonlinear programming problems to minimize the evaluation function J under constraints, and calculations to find the model parameters may be executed by known methods.
• The model parameters calculated by the model parameter calculation unit 13 are stored in the database 12, so as to be used for furnace state estimation processing in refining treatment from the next time onwards. This completes the processing of Step S7, and the furnace state estimation processing in the refining treatment is completed.
• Based on the temperature and the component concentration of the molten metal 101 and the component concentration of the slag 103 that are estimated by the above processing of the furnace state estimation method, manipulated variables are determined, and the refining operation is conducted, to thereby produce good molten steel. As the manipulated variables, the optimum flow rate and speed of top-blown oxygen, the height of the top-blowing lance, the flow rate of bottom-blown gas, the charge amount and charge timing of auxiliary raw materials, such as lime or iron ore, the timing for sampling molten metal, and/or the timing for terminating blowing are/is determined. Thus, based on the furnace state calculated by the above furnace state estimation method, a favorable molten steel production method can be realized.
• As described above, with the above configuration and processing, the furnace state estimation device 1, the furnace state estimation method, and the molten steel production method according to the present embodiment optimize model parameters, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace and then, store them in the database 12. In estimating the furnace state in the refining treatment, the past optimized model parameters accumulated in the database 12 can be used. This improves the accuracy of estimating the temperature and the component concentration of the molten metal 101, the component concentration of the slag 103, or the like.
• Although an embodiment of the present disclosure has been described based on the drawings and examples, it is to be noted that various modifications and changes may be easily made by those skilled in the art based on the present disclosure. Accordingly, such modifications and changes are included within the scope of the present disclosure. For example, functions or the like included in each component, each step, or the like can be rearranged without logical inconsistency, and a plurality of components, steps, or the like can be combined into one or divided. An embodiment according to the present disclosure can also be implemented as a program that is executed by a processor included in a device, or as a storage medium in which the program is recorded. It is to be understood that these are included within the scope of present disclosure.
• For example, the type and the count of model parameters to be determined and the form of evaluation function to be minimized are not limited to those listed in the above embodiment, and the same effect can be achieved as long as the mass balance error and the heat balance error in the furnace can be minimized. Furthermore, the models are not limited to those illustrated in Formula (2) through Formula (6) in the above embodiment, but may include molten metal temperature estimation models, scrap dissolution models, auxiliary raw material dissolution and yield rate models, decarburization efficiency models, dephosphorization models, FetO generation and reduction models, and the like. Moreover, although the furnace state estimation device 1 and the furnace state estimation method for the converter 100 are described in the present embodiment, they are also effective for calculating model parameters based on mass balance and heat balance in a furnace in secondary refining equipment or pretreatment equipment.
REFERENCE SIGNS LIST

1

Furnace state estimation device

2

Refining equipment

10

Control terminal

11

Input unit

12

Database

13

Model parameter calculation unit

14

Model determination unit

15

Furnace state calculation unit

16

Output unit

20

Display device

100

Converter

101

Molten metal

102

Lance

103

Slag

104

Duct

105

Exhaust gas detection unit

106

Vent hole

107

Flowmeter

Claims (9)

1. A furnace state estimation device comprising:

   an input unit configured to receive track record information and conditions of refining treatment, the track record information including measurement results of temperature and component concentration of molten metal and component concentration of slag before start of or during the refining treatment in refining equipment and measurement results regarding the refining equipment that include flow rate and component concentration of exhaust gas discharged from the refining equipment;

   a model determination unit configured to determine model parameters in the refining treatment of a target charge using past model parameters acquired from a database that stores model parameters for models related to blowing reaction in the refining equipment, the track record information, and the conditions of refining treatment;

   a furnace state calculation unit configured to calculate state quantities, including temperature and component concentration of the molten metal and component concentration of the slag, in a furnace, using the determined model parameters; and

   a model parameter calculation unit configured to calculate model parameters in the refining treatment of the target charge, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace from beginning to end of a particular period of time in the refining treatment, using the track record information including results of the refining treatment of the target charge.
2. The furnace state estimation device according to claim 1, wherein the model determination unit is configured to determine the model parameters in the refining treatment of the target charge, by averaging model parameters for past refining treatment wherein conditions are similar to the conditions of the refining treatment of the target charge among the past model parameters stored in the database.
3. The furnace state estimation device according to claim 1, wherein the model determination unit is configured to model relationship between the past model parameters stored in the database and the conditions of refining treatment, including counts of times the treatment is performed, dates and times of the treatment, and/or counts of times the refining equipment is used in the refining treatment, and determine the parameters for the refining treatment of the target charge based on a model.
4. The furnace state estimation device according to any one of claims 1 to 3, wherein the model parameters include a coefficient or a constant term that corrects an integrated amount of carbon charged into the furnace over a particular period of time, an integrated amount of carbon discharged out of the furnace over a particular period of time, an amount of oxygen charged into the furnace, an amount of oxygen discharged out of the furnace, an amount of oxygen used for oxidation of various metallic impurities in the molten metal, and/or an integrated amount of change in the temperature of the molten metal over a particular period of time due to changes in an amount of heat in the furnace.
5. The furnace state estimation device according to any one of claims 1 to 4, wherein the evaluation function is formed by a weighted sum of a term representing carbon balance, a term representing oxygen balance, and a term representing heat balance.
6. A furnace state estimation method to be executed by a furnace state estimation device, the furnace state estimation method comprising:

   an input step of receiving track record information and conditions of refining treatment, the track record information including measurement results of temperature and component concentration of molten metal and component concentration of slag before start of or during the refining treatment in refining equipment and measurement results regarding the refining equipment that include flow rate and component concentration of exhaust gas discharged from the refining equipment;

   a model determination step of determining model parameters in the refining treatment of a target charge using past model parameters acquired from a database that stores model parameters for models related to blowing reaction in the refining equipment, the track record information, and the conditions of refining treatment;

   a furnace state calculation step of calculating state quantities, including temperature and component concentration of the molten metal and component concentration of the slag, in a furnace, using the determined model parameters; and

   a model parameter calculation step of calculating model parameters in the refining treatment of the target charge, based on an evaluation function that includes terms representing a mass balance error and a heat balance error in the furnace from beginning to end of a particular period of time in the refining treatment, using the track record information including results of the refining treatment of the target charge.
7. The furnace state estimation method according to claim 6, wherein the model determination step is configured to determine the model parameters in the refining treatment of the target charge, by averaging model parameters for past refining treatment wherein conditions are similar to the conditions of the refining treatment of the target charge among the past model parameters stored in the database.
8. The furnace state estimation method according to claim 6, wherein the model determination step is configured to model relationship between the past model parameters stored in the database and the conditions of refining treatment, including counts of times the treatment is performed, dates and times of the treatment, and/or counts of times the refining equipment is used in the refining treatment, and determine the parameters for the refining treatment of the target charge based on a model.
9. A molten steel production method comprising
   producing molten steel, by determining flow rate and speed of top-blown oxygen, height of a top-blowing lance, flow rate of bottom-blown gas, charge amount and charge timing of auxiliary raw materials, such as lime or iron ore, timing for sampling molten metal, and/or timing for terminating blowing, based on the temperature and the component concentration of the molten metal and the component concentration of the slag estimated by the furnace state estimation method according to any one of claims 6 to 8, and performing refining operation.

Images