Statistical model construction apparatus, method and program, and molten steel phosphorus concentration estimation apparatus, method and program

Abstract

To provide a statistical model construction device or the like capable of constructing a statistical model capable of accurately estimating a dephosphorization rate constant even when the blowing treatment is stopped midway.SOLUTION: A statistical model construction device includes a data collection part 331 for collecting operation data on the blowing treatment of a converter 11, and a statistical model construction part 333 for constructing a statistical model using operation data as an explanatory variable and a dephosphorization rate constant in the blowing treatment as an objective variable. The dephosphorization rate constant is defined as a proportionality constant assuming that an amount of change in the phosphorus concentration in molten steel of the converter relative to the cumulative per-unit oxygen supply to the converter is proportional to the phosphorus concentration in the molten steel of the converter. The statistical model construction part constructs the statistical model using operation data on the past blowing treatment collected by the data collection part.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a statistical model construction apparatus, method, and program for the blowing process of a converter, and to a phosphorus concentration estimation apparatus, method, and program for the blowing process of a converter. In particular, the present invention relates to a statistical model construction apparatus, method, and program that can construct a statistical model capable of accurately estimating the dephosphorization rate constant even when the blowing process is stopped midway through, and to a phosphorus concentration estimation apparatus, method, and program for the blowing process that can accurately estimate the phosphorus concentration in the molten steel of a converter during the blowing process using this statistical model. [Background technology]

[0002] In converter smelting, controlling the components of molten steel at the end of the smelting process (stopping blowing), particularly the phosphorus concentration, is crucial for steel quality control. To control the phosphorus concentration in molten steel, various parameters are used, such as the height of the top blowing lance, the top blowing oxygen flow rate, the bottom blowing gas flow rate, and the timing and amount of additives such as quicklime and sintered ore. These parameters are often determined based on information obtained before the start of the smelting process, including the target phosphorus concentration in molten steel, molten iron data, and standards created based on past smelting operation data.

[0003] However, when the amount of manipulation is determined based only on information obtained before the start of the smelting process, there is a problem in that the variability of the phosphorus concentration in the molten steel at the end of the smelting process becomes large. Therefore, if the phosphorus concentration in the molten steel can be accurately and sequentially estimated during the smelting process using operational data that also includes information obtained during the smelting process, it is thought that the variability of the phosphorus concentration in the molten steel at the end of the smelting process can be suppressed by performing additional operations during the smelting process to reach the target phosphorus concentration in the molten steel.

[0004] As a technique for sequentially estimating the phosphorus concentration in molten steel during the smelting process, for example, the technique described in Patent Document 1 has been proposed. Patent Document 1 describes a technique in which a first-order reaction equation is assumed for the change in phosphorus concentration in molten steel during the smelting process, the dephosphorization rate constant, which is the proportionality constant of the first-order reaction equation, is estimated using a statistical model represented by a multiple regression equation, using operational data related to the smelting process, including information obtained during the smelting process, and the phosphorus concentration in molten steel during the smelting process is estimated using this estimated dephosphorization rate constant. Furthermore, Patent Document 1 also describes a technique for controlling the phosphorus concentration in molten steel by comparing the estimated phosphorus concentration in molten steel with a target phosphorus concentration in molten steel and changing the manipulated amount based on the comparison result.

[0005] Here, a first-order chemical reaction is one in which the rate of change of reactant concentration per unit time is determined in proportion to the concentration of the reactants, and the constant of proportionality is called the rate constant. Although the dephosphorization reaction in the blowing process of a converter is not a first-order reaction, its transition shape is similar to that of a first-order reaction. Therefore, in the technology described in Patent Document 1, the rate of change of phosphorus concentration in molten steel per unit time is applied to the first-order reaction equation to estimate the transition of phosphorus concentration in molten steel during the blowing process. Specifically, in the technology described in Patent Document 1, the elapsed time from the start of the blowing process is used as the time (reaction time) in the first-order reaction equation, the dephosphorization rate constant is estimated by a statistical model (multiple regression equation) constructed using this elapsed time, and the phosphorus concentration in molten steel during the blowing process is estimated using the statistical model constructed using this elapsed time (using the dephosphorization rate constant estimated using this statistical model).

[0006] However, there are cases where operations include non-blown time, which is time when blown steel is not being blown, such as when the blown steel process is temporarily stopped to perform slag removal (i.e., the supply of oxygen to the converter is temporarily stopped) or when the blown steel process is temporarily stopped due to some operational trouble. In such cases, the length of the non-blown time varies from one blown steel process to the next, so the technology described in Patent Document 1, which uses the elapsed time from the start of the blown steel process as the reaction time, was at risk of large estimation errors in the dephosphorization rate constant and, consequently, in the estimation errors in the phosphorus concentration in the molten steel. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2013-23696 [Overview of the Initiative] [Problems that the invention aims to solve]

[0008] The present invention was made to solve the problems of the prior art described above, and aims to provide a statistical model construction apparatus, method, and program that can construct a statistical model capable of accurately estimating the dephosphorization rate constant even when the smelting process is stopped in the middle of the process, and a molten steel phosphorus concentration estimation apparatus, method, and program that can accurately estimate the phosphorus concentration in the molten steel of the converter during the smelting process using this statistical model. [Means for solving the problem]

[0009] To solve the aforementioned problem, for example, one could consider using the oxygen supply time to the converter (acid supply time), excluding non-blown time, as the reaction time in the first-order reaction equation, instead of the elapsed time from the start of the blown process. However, since dephosphorization is a reaction in which phosphorus in molten steel is oxidized using an oxygen source in the converter and incorporated into the slag, even if the oxygen supply time is the same, the amount of dephosphorization will differ if the amount of oxygen blown in per unit time from the top blowing lance, etc., or the amount of auxiliary raw materials (oxygen-containing auxiliary raw materials) added per unit time differs. Therefore, it is considered difficult to accurately estimate the dephosphorization rate constant and, consequently, the phosphorus concentration in the molten steel.

[0010] Therefore, after diligent research, the inventors discovered that by using the cumulative oxygen supply rate per unit weight of molten steel supplied to the converter (the cumulative value of the oxygen supply per unit weight of molten steel) as the reaction time in the first-order reaction equation, instead of the elapsed time from the start of the blowing process, it is possible to accurately estimate the dephosphorization rate constant and, consequently, the phosphorus concentration in the molten steel, even when the blowing process is stopped midway.

[0011] This invention was completed based on the above-mentioned findings of the inventors. In other words, to solve the above problem, the present invention provides a data collection unit that collects operational data relating to the blowing process of a converter, and uses the operational data as an explanatory variable and the dephosphorization rate constant in the blowing process as the objective variable. , for estimating the dephosphorization rate constant The system comprises a statistical model construction unit for constructing a statistical model, wherein the dephosphorization rate constant is a proportionality constant assuming that the change in the phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. , in the following equation (5) Defined, the statistical model building unit uses the operational data related to past blowing processes collected by the data collection unit. The learning data uses the following as the explanatory variable and the actual value of the dephosphorization rate constant calculated by the following equation (7) from the past operation data related to the blowing process collected by the data collection unit as the dependent variable. Using , by multiple regression analysis or machine learning methods The present invention provides a statistical model construction device for constructing the aforementioned statistical model.

number

[0012] According to the statistical model construction apparatus of the present invention, the dephosphorization rate constant is defined as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. In other words, the dephosphorization reaction in the blowing process of the converter is assumed to be a first-order reaction, and it is assumed that the change in phosphorus concentration in the molten steel per unit time is not proportional to the phosphorus concentration in the molten steel as in a normal first-order reaction, but rather that the change in phosphorus concentration in the molten steel is proportional to the cumulative oxygen supply rate per unit, and this proportionality constant is defined as the dephosphorization rate constant. Therefore, as the inventors have found, by using a statistical model constructed according to this definition, the dephosphorization rate constant can be estimated with high accuracy even when the blowing process is stopped midway through the blowing process. Furthermore, the phrase "using operational data as explanatory variables" is a concept that includes not only cases where all data items of the collected operational data are used as explanatory variables, but also cases where only some of the data items are used as explanatory variables. Furthermore, "cumulative oxygen supply per unit" refers to the cumulative value of the amount of oxygen supplied per unit weight of molten steel from the start of the blowing process.

[0013] In the statistical model construction apparatus according to the present invention, the cumulative oxygen supply unit amount supplied to the converter is, for example, the cumulative oxygen blown into the converter unit amount. Furthermore, "cumulative oxygen injection rate per unit weight" refers to the cumulative value of the amount of oxygen injected per unit weight of molten steel from the start of the blowing process. For example, it corresponds to the cumulative value of the amount of oxygen injected from the top blowing lance, the cumulative value of the amount of oxygen injected from the tuyeres, or the cumulative value of the amount of oxygen injected from both the top blowing lance and the tuyeres, divided by the weight of the molten steel in the converter.

[0014] Preferably, in the statistical model construction apparatus according to the present invention, the cumulative oxygen supply unit amount supplied to the converter is the sum of the cumulative oxygen blown into the converter and the generated oxygen unit amount generated from the auxiliary raw materials fed into the converter. According to the preferred configuration described above, by considering not only the cumulative amount of oxygen injected into the converter but also the amount of oxygen generated from the auxiliary materials fed into the converter, the dephosphorization rate constant can be estimated with greater accuracy. Furthermore, "oxygen generation rate per unit weight generated from auxiliary raw materials" refers to the amount of oxygen generated per unit weight of molten steel from auxiliary raw materials (oxygen-containing auxiliary raw materials). For example, it can be calculated by dividing the product of the amount of auxiliary raw materials input (input weight) and the amount of oxygen contained per unit weight of the auxiliary raw materials (volume of oxygen contained) by the weight of molten steel in the converter.

[0015] Furthermore, in order to solve the above problems, the present invention provides a data collection unit that collects operational data related to the blowing process of a converter, and uses the operational data as an explanatory variable and the dephosphorization rate constant in the blowing process as an objective variable. , for estimating the dephosphorization rate constant The system includes a phosphorus concentration estimation unit that estimates the phosphorus concentration in the molten steel of the converter during the blowing process using a statistical model, and the dephosphorization rate constant is a proportionality constant assuming that the change in the phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. , in the following equation (5) Defined, The statistical model is constructed using training data that uses operational data related to past blowing processes collected by the data collection unit as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected by the data collection unit using the following equation (7) as the dependent variable, using a multiple regression analysis method or a machine learning method. The phosphorus concentration estimation unit calculates an estimated value of the dephosphorization rate constant in the new smelting process using the operational data related to the new smelting process collected by the data collection unit and the statistical model, and determines the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter during the new smelting process. According to the following equation (6) It is also provided as an estimation device for phosphorus concentration in molten steel.

number

[0016] According to the molten steel phosphorus concentration estimation device according to the present invention, the dephosphorization rate constant is defined as a proportionality constant when it is assumed that the change amount of the phosphorus concentration in the molten steel of the converter with respect to the cumulative oxygen supply amount per unit supplied to the converter is proportional to the phosphorus concentration in the molten steel of the converter. Therefore, by using the statistical model constructed according to this definition, as the inventors have found, even when the blowing process is stopped during the blowing process, the estimated value of the dephosphorization rate constant can be accurately calculated, and thus the phosphorus concentration in the molten steel can be accurately estimated.

[0017] In the molten steel phosphorus concentration estimation device according to the present invention, the cumulative oxygen supply amount per unit supplied to the converter is, for example, the cumulative blown oxygen amount per unit blown into the converter.

[0018] Preferably, in the molten steel phosphorus concentration estimation device according to the present invention, the cumulative oxygen supply amount per unit supplied to the converter is the sum of the cumulative blown oxygen amount per unit blown into the converter and the generated oxygen amount per unit generated from the auxiliary materials charged into the converter.

[0019] Further, in order to solve the above problems, the present invention includes a data collection step of collecting operation data related to the blowing process of a converter, and a statistical model construction step of constructing a statistical model having the operation data as an explanatory variable and the dephosphorization rate constant in the blowing process as an objective variable. The dephosphorization rate constant is defined as a proportionality constant when it is assumed that the change amount of the phosphorus concentration in the molten steel of the converter with respect to the cumulative oxygen supply amount per unit supplied to the converter is proportional to the phosphorus concentration in the molten steel of the converter. , for estimating the dephosphorization rate constant 統計モデルを構築する統計モデル構築ステップと、を有し、前記脱りん速度定数は、前記転炉に供給された累積酸素供給量原単位に対する前記転炉の溶鋼中りん濃度の変化量が前記転炉の溶鋼中りん濃度に比例すると仮定した場合の比例定数として In equation (5)Defined, the statistical model building step includes the operational data relating to past blowing processes collected in the data collection step. The learning data uses the following as explanatory variables, and the actual value of the dephosphorization rate constant calculated by equation (7) from the past blowing process operational data collected in the data collection step as the dependent variable. Using , by multiple regression analysis or machine learning methods This is also provided as a method for constructing the aforementioned statistical model.

[0020] Furthermore, in order to solve the above problems, the present invention includes a data collection step of collecting operational data related to the blowing process of a converter, and using the operational data as an explanatory variable and the dephosphorization rate constant in the blowing process as an objective variable. , for estimating the dephosphorization rate constant The process includes a phosphorus concentration estimation step of estimating the phosphorus concentration in the molten steel of the converter during the blowing process using a statistical model, wherein the dephosphorization rate constant is a proportionality constant assuming that the change in the phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. In equation (5), Defined, The statistical model is constructed using training data in which the operational data related to past blowing processes collected in the data collection step is used as explanatory variables, and the actual value of the dephosphorization rate constant calculated by equation (7) from the operational data related to past blowing processes collected in the data collection step is used as the dependent variable, by a multiple regression analysis method or a machine learning method. In the phosphorus concentration estimation step, the operation data for the new smelting process collected in the data collection step and the statistical model are used to calculate an estimated value of the dephosphorization rate constant in the new smelting process. Based on the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter during the new smelting process, the phosphorus concentration in the molten steel of the converter during the new smelting process is calculated. According to equation (6) It is also provided as a method for estimating the phosphorus concentration in molten steel.

[0021] Furthermore, in order to solve the above problems, the present invention provides a data collection unit that collects operational data related to the blowing process of a converter, and uses the operational data as an explanatory variable and the dephosphorization rate constant in the blowing process as an objective variable. , for estimating the dephosphorization rate constant A statistical model construction unit for constructing a statistical model, and a statistical model construction program for causing a computer to function as such, wherein the dephosphorization rate constant is a proportionality constant assuming that the change in the phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. In equation (5),Defined, the statistical model building unit uses the operational data related to past blowing processes collected by the data collection unit. The learning data uses the following as explanatory variables, and the actual value of the dephosphorization rate constant calculated by equation (7) from past operation data related to the blowing process collected by the data collection unit as the dependent variable. Using , by multiple regression analysis or machine learning methods It is also provided as a statistical model building program for constructing the aforementioned statistical model.

[0022] Furthermore, in order to solve the above problems, the present invention includes a data collection unit that collects operational data related to the blowing process of a converter, and a method that uses the operational data as an explanatory variable and the dephosphorization rate constant in the blowing process as an objective variable. , for estimating the dephosphorization rate constant A phosphorus concentration estimation unit that estimates the phosphorus concentration in the molten steel of the converter during the blowing process using a statistical model, and a molten steel phosphorus concentration estimation program for causing a computer to function as such, wherein the dephosphorization rate constant is a proportionality constant assuming that the change in the phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter. In equation (5), Defined, The statistical model is constructed using training data that uses operational data related to past blowing processes collected by the data collection unit as explanatory variables and the actual value of the dephosphorization rate constant calculated by equation (7) from the operational data related to past blowing processes collected by the data collection unit as the dependent variable, using a multiple regression analysis method or a machine learning method. The phosphorus concentration estimation unit calculates an estimated value of the dephosphorization rate constant in the new smelting process using the operational data related to the new smelting process collected by the data collection unit and the statistical model, and determines the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter during the new smelting process. According to equation (6) It is also provided as a program for estimating phosphorus concentration in molten steel. [Effects of the Invention]

[0023] According to the present invention, a statistical model capable of accurately estimating the dephosphorization rate constant can be constructed. Furthermore, this statistical model can be used to accurately estimate the phosphorus concentration in the molten steel of the converter during the smelting process. [Brief explanation of the drawing]

[0024] [Figure 1] This figure schematically shows the general configuration of a refining facility equipped with a statistical model construction device and a molten steel phosphorus concentration estimation device according to the first embodiment of the present invention. [Figure 2] Figure 1 shows an example of the operational data 321. [Figure 3] This is a flowchart illustrating the schematic steps of the method for estimating phosphorus concentration in molten steel according to the first embodiment of the present invention. [Figure 4] Figure 3 is a flowchart showing the specific steps of Step ST6. [Figure 5] This figure schematically shows the general configuration of a refining facility equipped with a statistical model construction device and a molten steel phosphorus concentration estimation device according to a second embodiment of the present invention. [Figure 6] Figure 5 shows an example of information on the different brands of auxiliary raw materials (323). [Figure 7] This is a scatter plot showing the relationship between estimated values ​​and actual values ​​obtained by the phosphorus concentration estimation method in molten steel according to the examples and comparative examples of the present invention. [Modes for carrying out the invention]

[0025] Hereinafter, embodiments of the present invention (the first and second embodiments) will be described with reference to the attached drawings as appropriate. In this specification and the drawings, components having substantially the same configuration or function are denoted by the same reference numerals, and their descriptions are omitted as appropriate.

[0026] <Principle for estimating phosphorus concentration in molten steel> First, the principle for estimating the phosphorus concentration in the molten steel of the converter during the blowing process in this embodiment will be explained. In the prior art described in Patent Document 1 and other documents mentioned above, it is assumed that the time change d[P] / dt of the phosphorus concentration in molten steel in a converter during the blowing process (hereinafter also referred to as [P][%]) is expressed by the first-order reaction equation shown in equation (1) below. That is, it is assumed that the amount of change per unit time of the phosphorus concentration [P] in molten steel d[P] / dt is proportional to the phosphorus concentration [P] in molten steel. In equation (1), k P [sec -1 ] is the dephosphorization rate constant, which is the proportionality constant. From equation (1), [P] after t [sec] elapsed from the start of the blowing process is expressed by the following equation (2). In equation (2), [P]ini [%] represents the initial value of [P] (i.e., the phosphorus concentration in the molten iron of the converter before the blowing process).

number

[0027] Therefore, the precise dephosphorization rate constant k P If this is obtained, the phosphorus concentration [P] [%] in the molten steel after t [sec] elapsed from the start of the blowing process can be estimated with high accuracy based on equation (2) above. However, generally, the dephosphorization rate constant k P The value is not the same for all blowing processes, and is thought to fluctuate depending on differences in operational data related to the blowing process. Therefore, for example, as disclosed in Patent Document 1, the dephosphorization rate constant k is calculated using operational data related to the blowing process that is the target of the estimation of phosphorus concentration in molten steel. P This is estimated for each blowing process. Note that the dephosphorization rate constant k P The actual values ​​are calculated after the smelting process is completed (more precisely, after the molten steel composition at the end of the smelting process is determined) by substituting the values ​​into the right-hand side of the following equation (3), which is a modified version of the above equation (2). Here, in equation (3), [P] ini [%] represents the phosphorus concentration in the molten iron before blowing, [P] represents the phosphorus concentration. end [%] represents the phosphorus concentration in the molten steel at the end of the blowing process, t end [sec] should contain the elapsed time from the start to the end of the blowing process. [P] ini Ya [P] end This can be obtained, for example, by analyzing samples taken from molten iron before the smelting process or from molten steel after the smelting process is completed.

number

[0028] Dephosphorization rate constant k P To estimate this for each blowing process, operational data related to the blowing process is used as an explanatory variable, and the dephosphorization rate constant k PA statistical model is constructed with the dependent variable as follows. This statistical model can be constructed using various statistical methods, but for example, it can be expressed by the multiple regression equation shown in equation (4) below. Then, the operational data related to the smelting process, which is the target for estimating the phosphorus concentration in molten steel, is used as the explanatory variable X. i By substituting (i=1,2,···,M) into equation (4), we obtain the different dephosphorization rate constant k for each blowing process. P The estimated dephosphorization rate constant k is then determined. P By applying this to equation (2) above, the phosphorus concentration in the molten steel during the blowing process can be estimated. The dephosphorization rate constant k is given by equation (4) below. P The regression coefficient a is a parameter used to estimate the regression coefficient. i (i=0,1,···,M) can be obtained by known multiple regression analysis methods. For example, the dephosphorization rate constant k in past blowing processes obtained from equation (3) above. P The actual values ​​and explanatory variable X, which is operational data related to the blowing process. i By using training data for constructing a statistical model, which associates the two processes for each blowing process, the regression coefficient a i It is possible to make a decision.

number

[0029] As described above, according to the prior art described in Patent Document 1, etc., the dephosphorization rate constant k can be determined by the statistical model shown in equation (4). P We estimate the dephosphorization rate constant k. P By applying this to equation (2), the phosphorus concentration in the molten steel during the blowing process can be estimated. However, as mentioned above, when the blowing process is temporarily stopped during operation, the length of the non-blowing time varies from blowing process to blowing process. Therefore, in the conventional technology described in Patent Document 1, etc., the dephosphorization rate constant k P There was a risk that the estimation error, and consequently the estimation error of the phosphorus concentration in molten steel, would become large. To solve the above problem, the elapsed time t from the start of the blowing process in equations (1) to (3) is... end Alternatively, it is conceivable to use the oxygen supply time to the converter excluding non-blown time. However, as mentioned above, dephosphorization is a reaction in which phosphorus in molten steel is oxidized using an oxygen source in the converter and incorporated into the slag. Therefore, even if the oxygen supply time is the same, the amount of dephosphorization will differ if the amount of oxygen blown in per unit time from the top blowing lance, etc., or the amount of auxiliary raw materials (oxygen-containing auxiliary raw materials) added per unit time differs, thus affecting the dephosphorization rate constant k. P Consequently, it is considered difficult to accurately estimate the phosphorus concentration in molten steel.

[0030] Therefore, in this embodiment, the elapsed time t from the start of the blowing process, t end Instead, the cumulative oxygen supply rate per unit is used. Specifically, equations (5) to (7) below are used instead of equations (1) to (3). In other words, assuming that the change in phosphorus concentration [P] in molten steel with respect to the cumulative oxygen supply rate per unit is proportional to the phosphorus concentration [P] in molten steel, the dephosphorization rate constant k P We define this as the constant of proportionality.

number

[0031] In equations (5) and (6) above, O2 is the cumulative oxygen supply from the start of the blowing process [Nm³]. 3 / ton] (cumulative value of oxygen supply per ton of molten steel), and in the above equation (7), O 2end This is the cumulative oxygen supply per unit [Nm³] from the start to the end of the blowing process. 3 It is / ton]. Dephosphorization rate constant k P Actual value [(Nm 3 / ton) -1 ] is calculated by substituting the values ​​into the right-hand side of equation (7) above after the blowing process is completed. Then, similar to the prior art described in Patent Document 1, etc., the dephosphorization rate constant k in past blowing processes is calculated. PBy using training data with the actual values ​​of as the target variable, we construct the statistical model shown in equation (4) above. Furthermore, for the new smelting process that is the target of the estimation of phosphorus concentration in molten steel, the explanatory variable X of the statistical model shown in the constructed equation (4) is... i By substituting the operational data obtained from the blowing process, the dephosphorization rate constant k can be calculated. P The estimated value of [P] in the molten iron before the blowing process was calculated. ini The calculated dephosphorization rate constant k P By substituting the estimated value of and the cumulative oxygen supply rate per unit O2 supplied during the blowing process into the right-hand side of equation (6) above, the phosphorus concentration in the molten steel during the blowing process can be estimated.

[0032] As described above, in this embodiment, instead of the elapsed time t from the start of the blowing process, the cumulative oxygen supply rate O2 supplied to the converter is used, so that even if the blowing process is stopped midway, the dephosphorization rate constant k can be determined. P This, in turn, allows for accurate estimation of the phosphorus concentration in molten steel.

[0033] As the cumulative oxygen supply rate per unit O2, for example, the cumulative amount of oxygen injected into the converter can be used. The "cumulative amount of oxygen injected per unit weight" means the cumulative amount of oxygen per unit weight of molten steel injected from the start of the smelting process, as shown in equation (8) below. For example, it corresponds to the value obtained by dividing the cumulative amount of oxygen injected from the top blowing lance, the cumulative amount of oxygen injected from the tuyeres, or the cumulative amount of oxygen injected from both the top blowing lance and the tuyeres by the weight of molten steel in the converter.

number

[0034] Furthermore, more preferably, as the cumulative oxygen supply unit O2, the sum of the cumulative amount of oxygen blown into the converter and the amount of oxygen generated from the auxiliary raw materials (oxygen-containing auxiliary raw materials) introduced into the converter can be used, as shown in equation (9) below. "Amount of oxygen generated from auxiliary raw materials" means the amount of oxygen per unit weight of molten steel generated from the auxiliary raw materials.

number

[0035] Amount of oxygen generated from auxiliary raw materials V auO2 For example, as shown in equation (10) below, the amount (weight) of auxiliary raw materials added is W. au [ton] and the amount of oxygen contained per unit weight of the auxiliary raw material (oxygen content volume) O2 au [Nm 3 It can be calculated by multiplying it by [ / ton].

number

[0036] <First Embodiment> The first embodiment of the present invention uses the cumulative oxygen supply rate unit O2 as the cumulative oxygen supply rate unit O2 blown into the converter. That is, the cumulative oxygen supply rate unit O2 is calculated according to the formula (8) described above. Figure 1 is a schematic diagram showing the general configuration of a refining facility equipped with a statistical model construction device and a molten steel phosphorus concentration estimation device according to the first embodiment of the present invention. As shown in Figure 1, the refining equipment 100 of the first embodiment comprises a converter equipment 10, a measurement and control device 20, and a calculation device 30. In the first embodiment, the calculation device 30 also functions as a statistical model construction device and a molten steel phosphorus concentration estimation device according to the present invention. The calculation device 30 accumulates operational data related to the blowing process while repeatedly operating the blowing process of the converter, and constructs a statistical model. Furthermore, after the statistical model is constructed and becomes available, the calculation device 30 uses the statistical model to estimate the phosphorus concentration in the molten steel during the blowing process, and further updates the statistical model using operational data obtained after the blowing process is completed, if necessary. The individual components of the refining equipment 100 will be described in detail below.

[0037] The converter equipment 10 comprises a converter 11, an upward blowing lance 12, a flue 13, a tuyeres 14, and an input device having an input chute 15. In the converter equipment 10, the upward blowing lance 12, inserted from the furnace mouth of the converter 11, blows oxygen gas G1 into the molten iron MS inside the converter 11. In the decarburization reaction that occurs during the blowing process, carbon in the molten iron MS reacts with oxygen gas G1 to produce CO gas or CO2 gas, and these gases are discharged via the flue 13. The molten iron MS that has undergone the blowing process is sent to the next process as molten steel MI. In addition, during the blowing process, phosphorus and silicon in the molten iron MS also react with oxygen gas G1 or auxiliary materials contained in the slag SL, and are incorporated into the slag SL and stabilized. Meanwhile, bottom-blowing gas G2 such as nitrogen gas or argon gas is blown in from the tuyeres 14 to stir the molten iron MS and promote the above reaction. From the input chute 15 of the input device, auxiliary raw materials AM, including oxygen-containing auxiliary raw materials such as quicklime and limestone that make up the slag SL, and sintered ore for supplying oxygen to the molten iron MS, are fed into the converter 11. If the auxiliary raw materials AM are in powder form, they can also be blown in together with oxygen gas G1 using the top blowing lance 12.

[0038] The measurement and control device 20 performs various measurements related to the refining process in the converter equipment 10 and controls the refining process. Specifically, the measurement and control device 20 comprises a sublance 21, an exhaust gas analyzer 22, an exhaust gas flow meter 23, and a slag level meter 28 as a measurement system. The sublance 21 is inserted from the furnace opening of the converter 11 together with the upper blowing lance 12, and by immersing the measuring device provided at its tip in the molten steel MI at a predetermined timing during the smelting process, a sample for component analysis of the molten steel MI is taken, and the carbon concentration and temperature of the molten steel MI are measured. In the following description, such measurements using the sublance 21 will also be referred to as sublance measurements. The exhaust gas analyzer 22 analyzes the components of the gas discharged via the flue 13. Specifically, the exhaust gas analyzer 22 measures the concentrations of CO, CO2, and O2 contained in the exhaust gas (the concentration of each component is referred to as the exhaust gas component concentration). On the other hand, the exhaust gas flow meter 23 measures the flow rate of the gas discharged via the flue 13 (referred to as the exhaust gas flow rate). The slag level meter 28 measures the level of slag SL in the converter 11 in a non-contact manner from the furnace opening of the converter 11. The results of the sublance measurement described above, as well as the measurement results of the exhaust gas analyzer 22, exhaust gas flow meter 23, and slag level meter 28, are transmitted to the calculation unit 30 (specifically, to the communication unit 31 described later).

[0039] On the other hand, the measurement and control device 20 comprises a lance drive device 24, an oxygen supply device 25, a bottom-blowing gas supply device 26, and an input control device 27 as a control system. The lance drive device 24 drives the upper-blowing lance 12 in the vertical direction. This allows adjustment of the height of the upper-blowing lance 12 (i.e., the distance from the molten iron MS to the position where oxygen gas G1 is supplied in the converter 11). The oxygen supply device 25 supplies oxygen gas G1 to the upper-blowing lance 12. The blown oxygen flow rate, i.e., the flow rate of the supplied oxygen gas G1 per unit time, is adjustable. The bottom-blowing gas supply device 26 supplies bottom-blowing gas G2 to the tuyeres 14. The flow rate of the supplied bottom-blowing gas G2 is also adjustable. The input control device 27 controls the input of auxiliary material AM from the input chute 15 of the input device. Specifically, the input control device 27 controls the timing and amount of input of auxiliary material AM. The operation of the lance drive unit 24, oxygen supply unit 25, bottom-blowing gas supply unit 26, and input control device 27 described above may be controlled independently of the computing unit 30, or, as shown in Figure 1, may be controlled according to control signals from the computing unit 30 (specifically, transmitted from the communication unit 31 described later).

[0040] The arithmetic unit 30 comprises a communication unit 31, a storage unit 32, an arithmetic unit 33, and an input / output unit 34, and is configured, for example, by a computer. The communication unit 31 is a communication device that communicates with each component of the measurement control device 20 by wire or wireless, receives measurement results obtained by the measurement control device 20, and transmits control signals to the measurement control device 20. The communication unit 31 may also be able to communicate with external devices other than the measurement control device 20.

[0041] The memory unit 32 is a storage device capable of storing various types of data. In the first embodiment, the memory unit 32 stores operational data 321 and parameters 322 related to the blowing process of the converter 11. Figure 2 shows an example of operation data 321. In the example shown in Figure 2, operation data 321 uses a charge number (CHNO) determined for each smelting process as the key, and various operation data related to the smelting process with the same charge number are stored in the same row. Operation data 321 includes analysis results data 3211, which stores sublance measurement results and component analysis results at the end of the smelting process. Note that, as in the record CHNO=00XX in the example shown in Figure 2, for new smelting processes that are the target of estimation of phosphorus concentration in molten steel, only the data that has already been collected is stored in the storage unit 32, and analysis results data 3211 that have not yet been measured or whose analysis results are unknown at that time are not stored.

[0042] As described above, in the first embodiment, the cumulative oxygen supply unit O2 is calculated according to equation (8). Therefore, the operation data 321 also stores data items for calculating the cumulative oxygen supply unit O2 according to equation (8). Specifically, the operational data 321 includes the cumulative amount of oxygen injected from the start of the blowing process (cumulative amount of oxygen injected) V shown in equation (8). O2 [Nm 3 The total cumulative value of the amount of oxygen blown in from the top blowing lance 12 and the tuyeres 14 is stored as [ ]. The total cumulative value can be calculated by the data acquisition unit 331 using the measured value of the amount of oxygen blown in (for example, the measured value of the sensor if the oxygen supply device 25 or bottom blowing gas supply device 26 is equipped with a sensor that measures the amount of oxygen), or if there is no measured value, using the indicated values ​​of the oxygen supply device 25 and the bottom blowing gas supply device 26. In addition, the operating data 321 includes the weight W of molten steel (molten iron) shown in equation (8) st The total weight of the main raw materials (the materials to be refined) charged into the converter 11, calculated by the data collection unit 331, is stored as [ton]. The main raw materials include, for example, molten iron, scrap, and cold iron.

[0043] On the other hand, the parameter 322 is stored in the storage unit 32 separately from the operation data 321, which is organized for each blowing process (for each charge number). As for the parameter 322, as described later, it is the parameter of the statistical model estimated by the statistical model construction unit 333 (the regression coefficient a shown in equation (4) above). i ) is stored.

[0044] The arithmetic unit 33 comprises, for example, one or more hardware processors such as a CPU (Central Processing Unit), one or more memories such as RAM (Random Access Memory) and ROM (Read Only Memory), and performs various calculations by executing one or more programs stored in the memory (for example, the storage unit 32) by one or more hardware processors. By operating according to the stored programs, the arithmetic unit 33 functions as a data acquisition unit 331, a data preprocessing unit 332, a statistical model construction unit 333, and a phosphorus concentration estimation unit 334. The arithmetic unit 33 may be a PLC (Programmable Logic Controller) or may be implemented by dedicated hardware such as an ASIC (Application Specific Integrated Circuit). The functions of each unit implemented by the arithmetic unit 33 will be described below.

[0045] The data acquisition unit 331 collects operational data 321 stored in the memory unit 32. Specifically, for example, data such as the weight of molten iron, which is known before the start of the blowing process, is acquired by the data acquisition unit 331 via the communication unit 31 from a converter blowing database stored in an external device (not shown in Figure 1) before the start of the blowing process, and is stored by the data acquisition unit 331 in the corresponding item in the row for the blowing process (CHNO) in the operational data 321. After the start of the blowing process, data acquired by various measurement system devices of the measurement control device 20 is acquired by the data acquisition unit 331 via the communication unit 31, and is stored by the data acquisition unit 331 in the corresponding item in the row for the blowing process (CHNO) in the operational data 321. The analysis results data 3211 in the operation data 321 is acquired by the data collection unit 331 via the communication unit 31 from the converter blowing database stored in an external device after the sublance measurement results and the component analysis results at the end of the blowing process are known, and is stored by the data collection unit 331 in the corresponding item of the row for the blowing process (CHNO) in the operation data 321. The data preprocessing unit 332 performs preprocessing on the operation data 321 described above. For example, the data preprocessing unit 332 performs various numerical processing operations, such as averaging and cumulative calculations, on the data acquired by the data collection unit 331 at regular intervals from the start to the end of the blowing process, and stores the data that has undergone preprocessing by the data preprocessing unit 332 in the corresponding item of the blowing process (CHNO) row of the operation data 321. The data preprocessing unit 332 may also perform data normalization. In the following description, the data that has undergone preprocessing by the data preprocessing unit 332 will also be treated as operation data 321, but the storage unit 32 may store these unprocessed data and the preprocessed data separately.

[0046] The statistical model building unit 333 uses operational data as an explanatory variable and the dephosphorization rate constant k P A statistical model (a statistical model expressed by the multiple regression equation shown in equation (4) above) is constructed with the dependent variable as k. Specifically, the statistical model construction unit 333 uses data items extracted from the past blowing operation data 321 stored in the memory unit 32 as explanatory variables, and the dephosphorization rate constant kP Using the learning data with the actual value as the target variable, the parameters (regression coefficient a i ) of the statistical model are estimated. Unless otherwise stated, the learning data used for constructing the statistical model is the accumulated past operation data 321 that has undergone preprocessing by the data preprocessing unit 332. The statistical model construction unit 333 uses the phosphorus concentration [P] in hot metal before blowing included in the learning data ini , the phosphorus concentration [P] in molten steel at the end of blowing of the analysis actual data 3211 included in the learning data end , and the cumulative oxygen injection amount V from the start to the end of the blowing process included in the learning data O2 to divide by the total weight W of the main raw materials included in the learning data st to obtain the cumulative oxygen supply amount per unit O 2end and substitute it into the right side of the above formula (7) to calculate the actual value of the dephosphorization rate constant k P in the past blowing process. Furthermore, the statistical model construction unit 333 extracts the explanatory variable X i from the data items included in the learning data related to the same blowing process, and uses the actual value of the dephosphorization rate constant k P obtained above as the target variable, and estimates the parameter a i of the statistical model by a known multiple regression analysis method, and stores the estimated parameter a i in the storage unit 32 as the parameter 322. In the first embodiment, the multiple regression analysis method is adopted as the method for constructing the statistical model. However, if the data items extracted from the operation data 321 are used as the explanatory variable X i [[ID=?]] and the dephosphorization rate constant k P is used as the target variable, it is also possible to construct a statistical model using a known machine learning method.

[0047] The phosphorus concentration estimation unit 334 uses the operation data 321 related to a new blowing process for which the phosphorus concentration in molten steel is to be estimated and the statistical model constructed by the statistical model construction unit 333 to calculate the estimated value of the dephosphorization rate constant k P in the new blowing process. Specifically, the phosphorus concentration estimation uniti and parameter a stored in the storage unit 32 as parameter 322 i By substituting them into a statistical model (i.e., the multiple regression equation shown in Equation (4)), the dephosphorization rate constant k in a new blowing process P is calculated. Furthermore, the phosphorus concentration estimation unit 334 determines the phosphorus concentration [P] in the hot metal before the new blowing process included in the operation data 321 for the new blowing process ini and the calculated estimated value of the dephosphorization rate constant k P and the cumulative oxygen injection amount V from the start of the blowing process to the current time included in the operation data 321 for the new blowing process O2 is divided by the total weight W of the main raw materials included in the operation data 321 for the new blowing process to obtain the cumulative oxygen supply amount per unit O2, and substitutes them into the above-mentioned Equation (6) to estimate the phosphorus concentration [P] in the molten steel at the current time (i.e., any time during the blowing process). For example, when the explanatory variable X st includes data items obtained by sublance measurement, the phosphorus concentration estimation unit 334 starts estimating the phosphorus concentration [P] in the molten steel after the sublance measurement in the new blowing process, and can estimate the phosphorus concentration [P] in the molten steel at a predetermined cycle until the end of the new blowing process. i The input / output unit 34 includes, for example, an output device such as a display or a printer, and an input device such as a keyboard, a mouse, or a touch panel. The output device outputs, for example, the estimated value of the phosphorus concentration [P] in the molten steel at any time during the blowing process calculated by the phosphorus concentration estimation unit 334. The input device can obtain, for example, an operation input for adding or modifying data stored in the storage unit 32.

[0048]

[0049] ​Figure 3 is a flowchart illustrating the schematic steps of the molten steel phosphorus concentration estimation method according to the first embodiment of the present invention, using the refining equipment 100 described above. In the example shown in Figure 3, the process of constructing a statistical model and the process of estimating the molten steel phosphorus concentration using the constructed statistical model are shown as a series of processes, but the method is not limited to this, and each may be performed at a different time. For example, the processes of steps ST1, ST4 to ST7 related to the process of constructing a statistical model and the processes of steps ST1 to ST3 related to the process of estimating the molten steel phosphorus concentration may be performed at different times.

[0050] As shown in Figure 3, in the molten steel phosphorus concentration estimation method according to the first embodiment, first, the data collection unit 331 collects data related to the new smelting process and stores it in the storage unit 32 as operation data 321 (step ST1 in Figure 3). Specifically, the data collection unit 331 acquires data such as the weight of molten iron before the start of the smelting process, and during smelting, it acquires the cumulative amount of oxygen blown in, the amount of auxiliary materials added, the exhaust gas component concentration, the exhaust gas flow rate, etc. When acquiring this data, the data preprocessing unit 332 performs data preprocessing, and the data collection unit 331 stores the preprocessed data as operation data 321. In the first embodiment, the explanatory variable X used in the statistical model is selected from the newly collected operational data 321 related to the blowing process. i Once the data is complete, the dephosphorization rate constant k will be determined using a statistical model. P This enables estimation of the phosphorus concentration [P] in molten steel. If a statistical model has already been constructed and is available (if "YES" is selected in step ST2 of Figure 3), the parameter a of the statistical model stored in the storage unit 32 is used as parameter 322. i and explanatory variable X i Using the data, the phosphorus concentration estimation unit 334 estimates the phosphorus concentration [P] in the molten steel for the new smelting process to be estimated (step ST3 in Figure 3). The estimated result of the phosphorus concentration [P] in the molten steel may be output by an output device such as a display or printer included in the input / output unit 34. The phosphorus concentration estimation unit 334 also uses, for example, the explanatory variable X used in the statistical model.i The phosphorus concentration [P] in molten steel may be estimated only once at a predetermined timing after the data has been collected, or the estimation of the phosphorus concentration [P] in molten steel may be repeated at predetermined intervals. On the other hand, if a statistical model is not available (i.e., "NO" in step ST2 of Figure 3), the estimation of the phosphorus concentration [P] in molten steel as described above is not performed.

[0051] After the blowing process is completed, the data collection unit 331 collects the analysis results data 3211 (step ST4 in Figure 3). For example, if a statistical model has not yet been constructed, but a sufficient amount of past blowing process operation data 321 has already been stored in the storage unit 32 (if "NO" is selected in step ST2 and "YES" in step ST5 in Figure 3), then the process for constructing the statistical model is executed. In this case, the statistical model construction unit 333 constructs a statistical model based on the stored past blowing process operation data 321, referring to Figure 4 and following the procedure described later (step ST6 in Figure 3), and estimates the parameters a of the statistical model. i This is stored in the memory unit 32 as parameter 322 (step ST7 in Figure 3).

[0052] On the other hand, if a statistical model has not yet been constructed and a sufficient number of past smelting operation data 321 have not been stored in the memory unit 32 (if "NO" is selected in step ST2 and step ST5 in Figure 3), the process for constructing the statistical model will not be executed, and only the storage of operation data 321 will be performed. Furthermore, once the statistical model has been constructed and is available, it will be possible to estimate the phosphorus concentration [P] in the molten steel and perform the smelting process, so the process for constructing the statistical model does not need to be executed (if "YES" is selected in step ST2 and "NO" in step ST5 in Figure 3). Furthermore, if certain conditions are met, for example, such as the smelting process performed by estimating the phosphorus concentration [P] in molten steel being repeated a predetermined number of times (if "YES" is given in step ST2 and step ST5 in Figure 3), a new statistical model may be constructed (step ST6 in Figure 3) by using the new operational data 321 for the smelting process together with the past operational data 321, or in place of the past operational data 321 for the smelting process, and the parameters 322 may be updated (step ST7 in Figure 3).

[0053] Figure 4 is a flowchart showing the specific procedure for step ST6 shown in Figure 3. The process shown in Figure 4 is performed by the statistical model building unit 333. In this process, first, the phosphorus concentration in the molten iron before the blowing process [P] is obtained from the accumulated past blowing operation data 321. ini and phosphorus concentration in molten steel at the end of blowing [P] end Extract the cumulative amount of oxygen blown in from the start to the end of the blowing process. O2 The main raw material is the total weight W st The cumulative oxygen supply unit is the value obtained by dividing by O 2end Calculate the dephosphorization rate constant k in past blowing processes, and substitute these into the right-hand side of equation (7) above. P The actual value is calculated (Step ST61 in Figure 4). Next, the accumulated operational data 321 is divided into training data and evaluation data for constructing the statistical model in an arbitrary ratio (step ST62 in Figure 4), and preprocessing is performed in the preprocessing unit 332 as necessary, and then the explanatory variables X of the statistical model shown in equation (4) above are determined. iDetermine this (step ST63 in Figure 4). From the operational data 321, which data item will be used as the explanatory variable X? i Whether to use it as such can be determined, for example, based on operational knowledge or using machine learning methods such as Lasso. The statistical model building unit 333 then uses the explanatory variable X included in the training data. i The target variable, the dephosphorization rate constant k, is calculated from the training data. P Using the actual values, a statistical model is constructed (Step ST64 in Figure 4).

[0054] Finally, the statistical model building unit 333 generates the explanatory variable X included in the evaluation data. i The objective variable, the dephosphorization rate constant k, is calculated from the evaluation data. P The accuracy of the constructed statistical model is evaluated using the actual values ​​(step ST65 in Figure 4). Specifically, the statistical model construction unit 333 evaluates, for example, the explanatory variable X included in the evaluation data. i Substituting into equation (4) gives the estimated dephosphorization rate constant k P The dephosphorization rate constant k is calculated from the evaluation data. P The actual values ​​are compared with the calculated values, and the error between the two (such as the standard deviation or root mean square error) is calculated. If the calculated error is within a predetermined tolerance range, the statistical model constructed in step ST64 is confirmed as the statistical model used to estimate the phosphorus concentration [P] in molten steel. On the other hand, if the calculated error is not within a predetermined tolerance range, for example, the operational data 321 used to construct the statistical model may be changed, or the same operational data 321 as before may be used, but the splitting of training data and evaluation data is redone (step ST62 in Figure 4), and then the statistical model is reconstructed (step ST64 in Figure 4). The parameter a of the statistical model calculated by processing in step ST64 i As shown as step ST7 in Figure 3, this is stored in the storage unit 32 as parameter 322. Furthermore, in order to reliably and accurately estimate the dephosphorization rate constant and, consequently, the phosphorus concentration in molten steel, it is preferable to divide the accumulated operational data 321 into training data and evaluation data (step ST62 in Figure 4), construct a statistical model using the training data (step ST64 in Figure 4), and evaluate the accuracy of the constructed statistical model using the evaluation data (step ST65 in Figure 4), as shown in the example in Figure 4. However, the present invention is not limited to this, and it is also possible to create only training data from the accumulated operational data 321 and to perform only the construction of a statistical model using this training data. In other words, it is possible to omit steps ST62 and ST65 shown in Figure 4.

[0055] As described above, according to the first embodiment of the present invention, the dephosphorization rate constant k P However, it is defined as the proportionality constant when it is assumed that the change in phosphorus concentration [P] in the molten steel of the converter 11, d[P] / dO2, with respect to the cumulative oxygen supply unit O2 supplied to the converter 11 (cumulative blown-in oxygen unit calculated according to equation (8)), is proportional to the phosphorus concentration [P] in the molten steel of the converter 11. Then, by using a statistical model constructed according to this definition, even if the blowing process is stopped in the middle of the blowing process, the dephosphorization rate constant k P This allows for the accurate calculation of the estimated value of [P], and consequently, the accurate estimation of the phosphorus concentration [P] in molten steel.

[0056] <Second Embodiment> A second embodiment of the present invention uses the sum of the cumulative oxygen supply unit O2, which is the cumulative amount of oxygen blown into the converter, and the amount of oxygen generated from the auxiliary raw materials (oxygen-containing auxiliary raw materials) introduced into the converter. In other words, the cumulative oxygen supply unit O2 is calculated according to the formula (9) described above. In the first embodiment, as described above, the cumulative oxygen supply rate unit O2 is defined as the cumulative amount of oxygen blown into the converter unit V. 02 / W stAlthough this is used, in the blowing process of the converter, oxygen-containing auxiliary materials may be added, and it is thought that dephosphorization will also proceed due to the oxygen generated from the auxiliary materials. For this reason, even if the cumulative amount of oxygen blown in is the same, if the amount of oxygen generated from the auxiliary materials is different, the amount of dephosphorization will be different, and as in the first embodiment, the cumulative oxygen supply unit is O2, and the cumulative amount of oxygen blown in unit is V 02 / W st The dephosphorization rate constant k is used only P Furthermore, estimating the phosphorus concentration [P] in molten steel may reduce the estimation accuracy. Therefore, in the second embodiment, the cumulative amount of oxygen blown in per unit is V 02 / W st Furthermore, by also considering the amount of oxygen generated from auxiliary raw materials, the dephosphorization rate constant k P Furthermore, it is possible to estimate the phosphorus concentration [P] in molten steel with even greater accuracy.

[0057] Figure 5 is a schematic diagram showing the general configuration of a refining facility equipped with a statistical model construction device and a molten steel phosphorus concentration estimation device according to a second embodiment of the present invention. As shown in Figure 5, the refining equipment 100A of the second embodiment includes a converter equipment 10, a measurement and control device 20, and a computing device 30A, similar to the refining equipment 100 of the first embodiment. The computing device 30A of the second embodiment also includes a communication unit 31, a storage unit 32A, a calculation unit 33A, and an input / output unit 34, similar to the computing device 30 of the first embodiment, and is configured, for example, by a computer. The refining equipment 100A of the second embodiment differs from the refining equipment 100 of the first embodiment in that, in addition to operation data 321 and parameters 322, auxiliary raw material type information 323 is stored in the storage unit 32A. Furthermore, the refining equipment 100 of the first embodiment also differs from the refining equipment 100 of the first embodiment in that the calculations performed by the calculation unit 33A have some differences from those of the calculation unit 33 of the first embodiment. The following describes the differences between the refining equipment 100A of the second embodiment and the refining equipment 100 of the first embodiment, while explaining similarities will be omitted as appropriate.

[0058] Figure 6 shows an example of auxiliary raw material brand-specific information 323. In the example shown in Figure 6, the auxiliary raw material brand-specific information 323 uses the brand name of the auxiliary raw material AM as the key, and the amount of oxygen contained per unit weight (oxygen volume) O2 of the auxiliary raw material AM of the same brand. au [Nm 3 Information such as ` / ton]` is stored on the same line.

[0059] As described above, in the second embodiment, the cumulative oxygen supply unit O2 is calculated according to formula (9). That is, in the second embodiment, unlike the first embodiment, when calculating the cumulative oxygen supply unit O2, the amount of oxygen generated from the auxiliary raw material AM V is used. auO2 Consider this. Amount of oxygen generated from auxiliary raw material AM V auO2 To calculate this, as shown in formula (10) above, the input amount (input weight) W for each brand of auxiliary raw material AM is used. au And the amount of oxygen contained per unit weight (oxygen volume) O2 for each brand of auxiliary raw material AM. au The following is used. Input amount W for each brand of auxiliary raw material AM. au The data collection unit 331 calculates the total amount of each type of auxiliary raw material AM fed from the input chute 15 into the converter 11, and stores it in the operation data 321. In the example shown in Figure 2 above, the input amount W of auxiliary raw material A of a certain type is calculated by the data collection unit 331. au (W auA ) and the amount of auxiliary ingredient B of a different brand W added. au (W auB This indicates the state in which ) is stored. Input amount W for each brand of auxiliary raw material AM. au The data acquisition unit 331 can calculate the amount of input if a measured value (for example, if the input control device 27 is equipped with a sensor that measures the amount of input, the measured value of that sensor) is used, or if no measured value is available, the value indicated by the input control device 27 is used. The amount of oxygen contained per unit weight for each type of auxiliary raw material AM (O2) au The data is acquired by the data collection unit 331 via the communication unit 31 from a converter blowing database stored in an external device (not shown in Figure 5), and stored by the data collection unit 331 in the corresponding item in the row for each type of auxiliary raw material AM in the auxiliary raw material type information 323.

[0060] The statistical model building unit 333 of the second embodiment, similar to the statistical model building unit 333 of the first embodiment, uses data items extracted from the operation data 321 related to past blowing processes stored in the memory unit 32A as explanatory variables, and the dephosphorization rate constant k P Using training data with the actual values ​​of the target variable, the parameters of the statistical model (regression coefficient a) i We estimate ). The statistical model building unit 333 of the second embodiment, similar to the statistical model building unit 333 of the first embodiment, uses the dephosphorization rate constant k according to equation (7). P Although it calculates the actual value, unlike the statistical model construction unit 333 of the first embodiment, the cumulative oxygen supply unit O shown in equation (7) 2end This is calculated using equation (9). That is, the statistical model building unit 333 of the second embodiment calculates the cumulative oxygen supply unit O shown in equation (7). 2end As such, the cumulative amount of oxygen blown in from the start to the end of the blowing process included in the training data is V. O2 And the amount of oxygen V generated from the auxiliary raw material AM from the start to the end of the blowing process. auo2 (The amount of auxiliary raw material AM added for each brand from the start to the end of the blowing process included in the training data W) au The oxygen content per unit weight for each brand of auxiliary raw material AM included in the auxiliary raw material brand information 323 is O2. au The sum of the product of all brands) and the total weight of the main raw materials included in the training data W st Substitute the value obtained by dividing by k to find the dephosphorization rate constant k P Calculate the actual value.

[0061] The phosphorus concentration estimation unit 334 of the second embodiment, similar to the statistical model construction unit 333 of the first embodiment, uses explanatory variables X extracted from the new smelting process operation data 321 that are the target of the phosphorus concentration estimation in molten steel. i And, parameter a stored in the storage unit 32A as parameter 322 i By substituting these into a statistical model (i.e., the multiple regression equation shown in equation (4)), a new dephosphorization rate constant k in the blowing process can be obtained. PThe phosphorus concentration estimation unit 334 of the second embodiment calculates an estimated value of [P] in the molten iron before the new blowing process, which is included in the new operation data 321 related to the new blowing process. ini The calculated dephosphorization rate constant k P The estimated value and the cumulative oxygen supply rate per unit O2 for the new blowing process are substituted into equation (6) to estimate the phosphorus concentration [P] in the molten steel at the present time (i.e., any point in time during the blowing process). However, unlike the phosphorus concentration estimation unit 334 of the first embodiment, the phosphorus concentration estimation unit 334 of the second embodiment calculates the cumulative oxygen supply rate unit O2 shown in equation (6) using equation (9). That is, the phosphorus concentration estimation unit 334 of the second embodiment uses the cumulative amount of oxygen blown in from the start of the blowing process to the present time, which is included in the new blowing process operation data 321, as the cumulative oxygen supply rate unit O2 shown in equation (6). O2 And the amount of oxygen V generated from the auxiliary raw material AM from the start of the new blowing process to the present. auo2 (The amount of each type of auxiliary raw material AM added from the start of the blowing process to the present, as included in the new blowing process operation data 321) au The oxygen content per unit weight for each brand of auxiliary raw material AM included in the auxiliary raw material brand information 323 is O2. au The sum of the products of all brands is calculated as the total weight of the main raw materials W included in the new blowing process operation data 321. st Substitute the value obtained by dividing by the given factor to estimate the current phosphorus concentration [P] in the molten steel.

[0062] The second embodiment of the phosphorus concentration estimation method in molten steel using the refining equipment 100A described above is the same as the first embodiment of the phosphorus concentration estimation method in molten steel described with reference to Figures 3 and 4, and the cumulative oxygen supply unit cost O2(O 2end The only difference is the method of calculating ), and the other procedures are the same, so a detailed explanation will be omitted here.

[0063] According to the second embodiment of the present invention, the cumulative amount of oxygen blown into the converter 11 is V 02 / W stIn addition, the amount of oxygen generated per unit of production from the auxiliary raw material AM fed into the converter 11 is V. auO2 / W st By also considering the dephosphorization rate constant k, P This allows for more accurate estimation of [the relevant factor], and consequently, more accurate estimation of the phosphorus concentration [P] in molten steel.

[0064] In the first and second embodiments described above, a single calculation unit 33, 33A comprises both a statistical model construction unit 333 and a phosphorus concentration estimation unit 334 (in other words, a single calculation unit 33, 33A functions as a statistical model construction device and a molten steel phosphorus concentration estimation device according to the present invention, performing both the construction of a statistical model and the estimation of the phosphorus concentration in molten steel). However, the present invention is not limited to this. For example, a single calculation unit may comprise only the statistical model construction unit 333 among the statistical model construction unit 333 and the phosphorus concentration estimation unit 334, and the statistical model constructed by this statistical model construction unit 333 may be used by another single calculation unit comprising only the phosphorus concentration estimation unit 334 to estimate the phosphorus concentration in molten steel. In other words, a configuration in which the statistical model construction device and the molten steel phosphorus concentration estimation device according to the present invention are provided separately is also possible. [Examples]

[0065] The following describes an example of the results of evaluating the estimation accuracy of the phosphorus concentration estimation method in molten steel according to the embodiments and comparative examples of the present invention. The comparative example's method for estimating phosphorus concentration in molten steel is a conventional estimation method described in Patent Document 1, etc., and is a method for estimating phosphorus concentration in molten steel using the aforementioned equations (1) to (4). The molten steel phosphorus concentration estimation method according to Example 1 is the estimation method described in the first embodiment, and is a method for estimating phosphorus concentration in molten steel using the aforementioned equations (4) to (8). The molten steel phosphorus concentration estimation method according to Example 2 is the estimation method described in the second embodiment, and is a method for estimating phosphorus concentration in molten steel using the aforementioned equations (4) to (7), equation (9) and equation (10). Using the estimation methods described in the comparative example and Examples 1 and 2 above, the phosphorus concentration in the molten steel at the end of the blowing process was estimated. The estimation accuracy was then evaluated by comparing each estimated value with the actual value obtained by analyzing a sample taken from the molten steel after the blowing process.

[0066] Figure 7 is a scatter plot showing the relationship between estimated and actual values ​​obtained by the phosphorus concentration estimation method in molten steel according to the examples and comparative examples of the present invention. Figure 7(a) shows the relationship between estimated and actual values ​​for the comparative example, Figure 7(b) shows the relationship between estimated and actual values ​​for Example 1, and Figure 7(c) shows the relationship between estimated and actual values ​​for Example 2. Table 1 shows the error σ (standard deviation) and root mean square error (RMSE) calculated from the results shown in Figure 7. The results shown in Figure 7 and Table 1 were obtained using operational data from 3640 blowing operations as training data and operational data from 1564 blowing operations as evaluation data (therefore, the scatter plot in Figure 7 plots data from 1564 blowing operations). [Table 1]

[0067] As shown in Figure 7 and Table 1, both Examples 1 and 2 showed higher estimation accuracy than the comparative example, with Example 2 exhibiting the highest estimation accuracy. Thus, the estimation method according to the present invention makes it possible to accurately estimate the phosphorus concentration in molten steel in a converter. [Explanation of symbols]

[0068] 10...Converter equipment 11. Converter 20. Measurement and control device 30,30A...Arithmetic unit 33,33A...Arithmetic section 100, 100A... Refining equipment 331...Data Collection Department 333...Statistical Model Construction Department 334...Phosphorus concentration estimation unit MI...molten steel

Claims

1. A data collection unit that collects operational data related to the blowing process of the converter, The system includes a statistical model construction unit that constructs a statistical model for estimating the dephosphorization rate constant, with the aforementioned operational data as explanatory variables and the dephosphorization rate constant in the blowing process as the dependent variable. The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, The statistical model construction unit constructs the statistical model using a multiple regression analysis method or a machine learning method, with the operational data related to past blowing processes collected by the data collection unit as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected by the data collection unit using the following equation (7) as the dependent variable. A statistical model building device. [Number 10] In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process.

2. The cumulative oxygen supply rate per unit supplied to the converter is the cumulative oxygen blown into the converter per unit. The statistical model construction apparatus according to claim 1.

3. The cumulative oxygen supply rate per unit supplied to the converter is the sum of the cumulative oxygen blown into the converter and the generated oxygen rate per unit produced from the auxiliary materials fed into the converter. The statistical model construction apparatus according to claim 1.

4. A data collection unit that collects operational data related to the blowing process of the converter, The system includes a phosphorus concentration estimation unit that estimates the phosphorus concentration in the molten steel of the converter during the blowing process using a statistical model for estimating the dephosphorization rate constant, with the aforementioned operational data as the explanatory variable and the dephosphorization rate constant in the blowing process as the dependent variable. The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, The statistical model is constructed using training data that uses operational data related to past blowing processes collected by the data collection unit as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected by the data collection unit using the following equation (7) as the dependent variable, using a multiple regression analysis method or a machine learning method. The phosphorus concentration estimation unit calculates an estimated value of the dephosphorization rate constant in the new smelting process using the operational data related to the new smelting process collected by the data collection unit and the statistical model, and estimates the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter in the new smelting process using the following formula (6): A device for estimating phosphorus concentration in molten steel. [Math 11] In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process. In equation (6) above, [P] represents the phosphorus concentration in the molten steel, [P] ini represents the phosphorus concentration in the molten iron before blowing, k P represents the dephosphorization rate constant, and O 2 represents the cumulative oxygen supply unit from the start of blowing.

5. The cumulative oxygen supply rate per unit supplied to the converter is the cumulative oxygen blown into the converter per unit. The apparatus for estimating phosphorus concentration in molten steel according to claim 4.

6. The cumulative oxygen supply rate per unit supplied to the converter is the sum of the cumulative oxygen blown into the converter and the generated oxygen rate per unit produced from the auxiliary materials fed into the converter. The apparatus for estimating phosphorus concentration in molten steel according to claim 4.

7. A data collection step to collect operational data related to the blowing process of the converter, The method includes a statistical model construction step, in which the aforementioned operational data is used as an explanatory variable and the dephosphorization rate constant in the blowing process is used as the dependent variable to construct a statistical model for estimating the dephosphorization rate constant, The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, In the statistical model construction step, the statistical model is constructed using training data in which the operational data related to past blowing processes collected in the data collection step is used as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected in the data collection step using the following equation (7) is used as the dependent variable, by multiple regression analysis or machine learning. Methods for constructing statistical models. [Math 12] In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process.

8. A data collection step to collect operational data related to the blowing process of the converter, The process includes a phosphorus concentration estimation step, in which the phosphorus concentration in the molten steel of the converter during the blowing process is estimated using a statistical model for estimating the dephosphorization rate constant, with the aforementioned operational data as the explanatory variable and the dephosphorization rate constant in the blowing process as the dependent variable. The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, The statistical model is constructed using training data in which the operational data related to past blowing processes collected in the data collection step is used as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected in the data collection step using the following equation (7) is used as the dependent variable, by a multiple regression analysis method or a machine learning method. In the phosphorus concentration estimation step, the operation data for the new smelting process collected in the data collection step and the statistical model are used to calculate an estimated value of the dephosphorization rate constant in the new smelting process. Based on the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter during the new smelting process, the phosphorus concentration in the molten steel of the converter during the new smelting process is estimated by the following equation (6). Method for estimating phosphorus concentration in molten steel. 【Number 13】 In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process. In equation (6) above, [P] represents the phosphorus concentration in the molten steel, [P] ini represents the phosphorus concentration in the molten iron before blowing, k P represents the dephosphorization rate constant, and O 2 represents the cumulative oxygen supply unit from the start of blowing.

9. A data collection unit that collects operational data related to the blowing process of the converter, A statistical model construction unit constructs a statistical model for estimating the dephosphorization rate constant, with the aforementioned operational data as explanatory variables and the dephosphorization rate constant in the aforementioned blowing process as the dependent variable, and a statistical model construction program for causing a computer to function as such. The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, The statistical model construction unit constructs the statistical model using a multiple regression analysis method or a machine learning method, with the operational data related to past blowing processes collected by the data collection unit as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected by the data collection unit using the following equation (7) as the dependent variable. A statistical model building program. [Number 14] In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process.

10. A data collection unit that collects operational data related to the blowing process of the converter, A phosphorus concentration estimation unit for estimating the phosphorus concentration in molten steel in a converter during the blowing process, using a statistical model for estimating the dephosphorization rate constant, in which the aforementioned operational data is used as an explanatory variable and the dephosphorization rate constant in the blowing process is used as an objective variable, and a molten steel phosphorus concentration estimation program for causing a computer to function as such, The dephosphorization rate constant is defined by the following equation (5) as a proportionality constant assuming that the change in phosphorus concentration in the molten steel of the converter is proportional to the cumulative oxygen supply rate per unit supplied to the converter, The statistical model is constructed using training data that uses operational data related to past blowing processes collected by the data collection unit as explanatory variables, and the actual value of the dephosphorization rate constant calculated from the operational data related to past blowing processes collected by the data collection unit using the following equation (7) as the dependent variable, using a multiple regression analysis method or a machine learning method. The phosphorus concentration estimation unit calculates an estimated value of the dephosphorization rate constant in the new smelting process using the operational data related to the new smelting process collected by the data collection unit and the statistical model, and estimates the phosphorus concentration in the molten iron of the converter before the new smelting process, the calculated estimated value of the dephosphorization rate constant, and the cumulative oxygen supply rate per unit supplied to the converter in the new smelting process using the following formula (6): A program for estimating phosphorus concentration in molten steel. [Number 15] In equation (5) above, [P] represents the phosphorus concentration in the molten steel, O₂ represents the cumulative oxygen supply rate from the start of the blowing process, and kP represents the dephosphorization rate constant. In equation (7) above, kP is the dephosphorization rate constant, [P]ini is the phosphorus concentration in the molten iron before the blowing process, [P]end is the phosphorus concentration in the molten steel at the end of the blowing process, and O2end is the cumulative oxygen supply unit from the start to the end of the blowing process. In equation (6) above, [P] represents the phosphorus concentration in the molten steel, [P] ini represents the phosphorus concentration in the molten iron before blowing, k P represents the dephosphorization rate constant, and O 2 represents the cumulative oxygen supply unit from the start of blowing.

Images