JP7159677B2 - Operating condition calculation device and operating condition calculation method - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an operating condition calculation device and an operating condition calculation method for calculating operating conditions in molten pig iron or molten steel refining.

Desulfurization of hot metal is performed as a refining treatment of hot metal or molten steel. Desulfurization in hot metal pretreatment is carried out using, for example, a mechanical stirring desulfurization apparatus (hereinafter also referred to as "KR", 1 as shown in FIG. 13). A desulfurization agent is added to the hot metal 5 in the ladle 10, and the impeller 21 is rotated by the driving device 25 to accelerate the desulfurization reaction. Therefore, it is important to reduce the S concentration in the hot metal 5 after desulfurization treatment by KR1 (hereinafter also referred to as “post-treatment S concentration”) to the target S concentration when delivered to the converter. Therefore, if the desulfurization agent is excessively added in order to ensure that the post-treatment S concentration is below the target S concentration in the KR treatment, the production cost will increase. In setting the conditions, it is desirable to be able to accurately estimate the post-treatment S concentration.

For example, Patent Literature 1 describes the charging conditions of the treating agent in the hot metal pretreatment operation, which are capable of accurately guiding the charging condition of the treatment agent in the hot metal pretreatment operation, and taking into account practical variations. A guidance method is disclosed. In Japanese Patent Laid-Open No. 2004-100000, the operating conditions including the processing agent input conditions are changed and repeatedly input to a neural network to determine the processing agent input conditions, and by adding practical variations to the evaluation, an appropriate processing agent is determined. Find the input conditions. As a result, the estimation accuracy is enhanced regardless of whether the correspondence characteristic between the processing agent input condition and the component concentration after processing exhibits linearity or nonlinearity.

JP-A-8-269518

Kengo Kato, and two others, "Theory of Lasso quantile regression and its application to non-life insurance," Journal of the Japan Statistical Society, March 2009, Vol. 38 No. 2, p121-149

However, in the method of Patent Document 1, there is a possibility that the conditions in a region where past performance data does not exist may be taken as a solution when finding an appropriate processing agent supply condition. That is, there is a possibility that a solution that cannot be executed in operation is selected, and the reliability of the solution is a concern.

Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to provide a new and improved operating condition calculation device capable of calculating reliable operating conditions. An object of the present invention is to provide an operating condition calculation method.

In order to solve the above problems, according to one aspect of the present invention, there is provided an operating condition calculation device for calculating operating conditions in a refining process of molten pig iron or molten steel, wherein a Preconditions and operating conditions related to refining are input, and an estimation model is constructed that outputs post- refining component concentrations such that component concentrations after refining are greater than or equal to actual component concentrations after refining with a given probability. a model construction unit, a constraint condition setting unit that sets a constraint condition for operating conditions as a likelihood constraint condition using a probability density function and a constraint condition for a post-treatment component concentration as a post-treatment component concentration constraint condition; An operating condition calculation device is provided, comprising: a feasible solution searching unit that searches for operating conditions based on the degree constraint and the post-treatment component concentration constraint.

An evaluation function setting unit that sets the operating cost to be minimized as an evaluation function, and the feasible solution search unit satisfies the estimation model, the likelihood constraint, and the post-processing component concentration constraint, and minimizes the evaluation function. You may make it search for operating conditions.

The estimation model calculates the expected value of the component concentration after treatment using the expected value prediction model with the assumptions and operating conditions as input, and outputs the component concentration after treatment obtained by performing quantile regression with the expected value as input. good too.

Here, the expected value prediction model may use a supervised regression model.

Alternatively, the estimation model may output post-treatment component concentrations obtained by performing quantile regression with the preconditions and operating conditions as inputs.

The probability density function may assume a Gaussian mixture distribution.

The feasible solution searcher may search for operating conditions by an evolutionary algorithm.

Further, in order to solve the above problems, according to another aspect of the present invention, there is provided an operating condition calculation method for calculating operating conditions in a refining process of molten pig iron or molten steel, comprising: Preconditions before refining and operating conditions related to refining are input, and an estimation model is developed that outputs post- refining component concentrations such that component concentrations after refining are equal to or higher than actual component concentrations after refining with a predetermined probability. An estimation model building step to be constructed, a likelihood constraint setting step of setting constraints for operating conditions as likelihood constraints by a probability density function, and setting constraints for post-processing component concentrations as post-processing component concentration constraints An operating condition calculation method including a post-treatment component concentration constraint setting step and a feasible solution search step of searching for an operating condition based on the estimation model, the likelihood constraint, and the post-treatment component concentration constraint. provided.

As described above, according to the present invention, reliable operating conditions can be calculated.

1 is a block diagram showing the functional configuration of an operating condition calculation device according to one embodiment of the present invention; FIG. It is a flow chart which shows operating condition calculation processing concerning the embodiment. FIG. 4 is an explanatory diagram showing the configuration of an estimation model; FIG. 4 is an explanatory diagram showing an example of the relationship between the post-treatment S concentration predicted by the expected value prediction model y^ based on the ElasticNet model and the actual post-treatment S concentration. FIG. 10 is a graph showing analysis results of predicted post-treatment S concentration and post-treatment S concentration by a quantile regression model; FIG. FIG. 10 is a graph showing analysis results of a predicted post-treatment S concentration and an actual post-treatment S concentration by a quantile regression model when a quantile is directly estimated from an input X; FIG. The 99% quantile of post-treatment S is determined using the 99% quantile regression model. 4 is a graph showing accuracy evaluation results of post-prediction S concentrations. FIG. 10 is an explanatory diagram illustrating values obtained by an expected value prediction model and a quantile regression model; FIG. 4 is an explanatory diagram showing an example of a confidence region of an estimation model; FIG. 10 is an explanatory diagram showing an example of distribution of optimum solutions obtained when no constraint is added; FIG. 10 is an explanatory diagram showing an example of distribution of optimum solutions obtained when a constraint is added; It is a block diagram for demonstrating the hardware constitutions of the operating-conditions calculating apparatus which concerns on the same embodiment. It is an explanatory view showing a schematic configuration of a mechanical stirring desulfurization apparatus (KR).

Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. In the present specification and drawings, constituent elements having substantially the same functional configuration are denoted by the same reference numerals, thereby omitting redundant description.

<1. Functional Configuration of Operating Condition Calculator>
First, based on FIG. 1, the operating condition calculating device which concerns on one Embodiment of this invention is demonstrated. FIG. 1 is a block diagram showing the functional configuration of an operating condition calculation device 100 according to this embodiment.

The operating condition calculation device 100 according to the present embodiment is an information processing device for performing optimization processing of operating conditions in the refining process of molten pig iron or molten steel. As shown in FIG. 1, the operating condition calculation device 100 includes a performance data acquisition unit 110, an estimated model construction unit 120, a constraint condition setting unit 130, an evaluation function setting unit 140, a feasible solution search unit 150, and an output unit 160 .

The performance data acquisition unit 110 acquires, from the performance data storage unit 200, performance data in past operations used for constructing an estimation model for estimating post-refining component concentrations. Performance data includes measurements taken during the refining process or values calculated based on the measurements. For example, in constructing an estimation model for S concentration after KR treatment as component concentration after refining treatment, hot metal temperature, hot metal weight, C concentration before treatment, Si concentration before treatment, Mn concentration before treatment, P concentration before treatment, S concentration before treatment, Concentration, melt surface height, function of impeller height and stirring speed (R-dip), impeller usage count, powdered CaO charging amount, secondary slag charging amount, Al concentration, etc. are obtained as performance data. The performance data acquisition unit 110 acquires performance data of a predetermined number of past operations, and outputs the data to the estimation model construction unit 120 and the constraint condition setting unit 130 .

Based on the performance data acquired by the performance data acquisition unit 110, the estimation model construction unit 120 inputs the preconditions before the treatment and the operating conditions related to the treatment, and the component concentration after the treatment has a predetermined probability. Build an estimation model that outputs post-treatment component concentrations that are equal to or greater than the actual component concentrations.

The constraint setting unit 130 sets constraints for the operating conditions using a probability density function, and sets constraints for the post-treatment component concentrations. The constraint on the operating conditions is also referred to as "likelihood constraint", and the constraint on post-treatment component concentration is also referred to as "post-treatment component concentration constraint". A feasible operating condition can be obtained by searching for an operating condition to be a solution within a reliable range of the post-treatment component concentration estimation model. Therefore, the constraint condition setting unit 130 sets a range for searching for a solution as a constraint condition.

To set the likelihood constraint, specifically, the constraint setting unit 130 first performs simultaneous Estimate a probability density function (eg , Gaussian mixture). Then, the constraint condition setting unit 130 calculates the logarithmic likelihood of the training data based on the estimated density function, and the logarithm Based on the one-dimensional probability density function of the likelihood, a control limit with which the estimation model can be trusted is obtained. This control limit is set as a likelihood constraint. At least one of the upper limit value and the lower limit value of the post-treatment component concentration is defined in the post-treatment component concentration constraint condition.

The evaluation function setting unit 140 sets the operating cost to be minimized as an evaluation function. The evaluation function setting unit 140 sets an evaluation function related to the operating cost and outputs it to the feasible solution searching unit 150 only when the operating cost is taken into consideration when the operating condition is searched by the feasible solution searching unit 150 described later. .

The feasible solution searching unit 150 searches for operating conditions that satisfy the likelihood constraint and the post-processing component concentration constraint using the estimation model. In addition, the feasible solution search unit 150, when searching for operating conditions in consideration of operating costs, uses an estimation model to satisfy the likelihood constraint and the post-processing component concentration constraint, and to minimize Search for operating conditions that minimize the evaluation function with the set cost. As a result of the search, the feasible solution searching unit 150 sets the optimal and feasible operating conditions as the optimal operating conditions. The optimum operating conditions may be output via the output unit 160. The output unit 160 is, for example, a display capable of displaying information, a speaker capable of outputting audio, or the like.

<2. Operating condition calculation processing>
Based on FIG. 2, the operating condition calculation processing by the operating condition calculating device 100 according to this embodiment will be described. FIG. 2 is a flowchart showing operating condition calculation processing according to the present embodiment. In addition, below, the desulfurization process in KR is demonstrated as an example of the refining process of molten iron or molten steel.

(S100: Estimation model construction processing)
First, when the performance data acquisition unit 110 of the operating condition calculation device 100 acquires past performance data from the performance data storage unit 200, the estimation model construction unit 120 constructs an estimation model based on quantile regression. Figure 3 shows the configuration of the estimation model. As shown in FIG. 3, the estimation model is composed of an expected value prediction model y ^ that calculates the expected value from the explanatory variable X, and a quantile regression model that performs quantile regression on the expected value of the expected value prediction model y ^ Become. In this specification, the expected value prediction model is expressed as "y^", which is the same as the "^" above "y" in FIG.

The expected value prediction model y ^ may be any machine learning model (supervised regression model) that can obtain expected values from actual data, such as gradient boosting regression model, multiple regression model, support vector regression model, ElasticNet A model or the like can be used. In this embodiment, as the expected value prediction model ŷ, an ElasticNet model with a high sparsification effect is used in consideration of a regularization term in the regression of the generalized linear model. The post-treatment S concentration predicted by the expected value prediction model y^ based on the ElasticNet model (hereinafter referred to as the “predicted post-treatment S concentration”) and the actual post-treatment S concentration (hereinafter referred to as the “actual post-treatment S concentration”) ) is shown in FIG. The upper side of FIG. 4 is a performance-prediction scatter diagram for the training data of the expected value prediction model y ^, and the lower side of FIG. . Note that the ElasticNet model was a model represented by a second-order polynomial.

From FIG. 4, the S concentration after the predicted treatment and the S concentration after the actual treatment roughly match. When the value is 5, the post-treatment S concentration has a range of about 1 to 15. Therefore, in order to consider reducing this variation, as shown in FIG. 5, the predicted post-treatment S concentration and the actual post-treatment S concentration are analyzed by a quantile regression model (Quantile Regression; see Non-Patent Document 1.). did. A quantile regression model is a regression model that describes the quantiles of the distribution of the dependent variable. A quantile regression model uses the I am building a linear model with When estimating, the linear model corresponding to the quantile to be estimated is used to estimate the quantile of the distribution of the dependent variable for any explanatory variable.

In step S100, an estimation model that directly predicts the quantiles from the explanatory variable X may be constructed. The calculation accuracy becomes very conservative as shown in FIG. For this reason, it is preferable to once predict the expected value and perform quantile regression on the predicted value.

As a result of the analysis using such a quantile regression model, it was found that as the post-prediction S concentration increased, the variation in error increased. Therefore, in the present embodiment, among the quantiles shown in FIG. 5, the 99% quantile of S after processing is obtained using the 99% quantile regression model. FIG. 7 shows the accuracy evaluation result of the post-prediction S concentration. The upper side of FIG. 7 is a performance-prediction scatter diagram for the training data of the expected value prediction model y ^, and the lower side of FIG. 7 is a performance-prediction scatter diagram for the evaluation data of the expectation prediction model y ^ . Results based on the expected value prediction model y ^ shown in FIG. 4, and the results of obtaining the 99% quantile of S after processing using the 99% quantile regression model for the expected value prediction model y ^ shown in FIG. , by using the 99% quantile regression model, the predicted post-treatment S concentration is distributed close to the actual post-treatment S concentration (slanted 45° straight line).

Therefore, the estimation model construction unit 120 constructs the estimation model shown in FIG. 3 that outputs the post-processing S concentration y^ _{q (τ=0.99)} at the 99% quantile as the post-estimation processing S concentration. . The estimation model shown in FIG. 3 obtains the quantile of the post-treatment S concentration with respect to the expected value. Originally, the model should calculate the quantile according to the value of the explanatory variable X, but in this embodiment, as shown in FIG. Since it is calculated as a point, an error occurs between it and the true quantile. In A and B of FIG. 8, the value of the expected value prediction model y ^ is highly predicted by 99% quantile regression, and in C of FIG. 8, the expected value prediction model y by 99% quantile regression It is predicted to be low from the value of ^. However, when the quantile probability is set to be high as in this example, even if the post-processing S concentration is predicted to be high as in A and B of FIG. The risk of underestimating the future is considered low.

(S110: Constraint condition setting process)
Next, the constraint condition setting unit 130 sets a probability density function that defines the distribution area of the performance data as a constraint condition (S110). In order to search for a solution within the reliable range of the estimation model, the present embodiment implements reliability determination using a mixed Gaussian distribution that can estimate complex joint probability density functions.

Specifically, first, a joint probability density function (assuming Gaussian mixture distribution) of all or part of the explanatory variables is estimated from the training data used to construct the expected value prediction model. Then, the constraint condition setting unit 130 calculates the logarithmic likelihood of the training data based on the estimated density function, and the logarithm Based on the probability density function of the likelihood, the control limits that the prediction model can rely on are obtained. This control limit is set as a constraint.

FIG. 9 shows an example of the confidence region of the estimation model. FIG. 9 shows the confidence regions for two explanatory variables, the CaO powder charge and the secondary slag charge. As shown in FIG. 9, there is a range (that is, a control limit) in which a predetermined number or more of actual data plots are included. Constraint condition setting unit 130 generates a constraint condition that sets this region as a reliable region and excludes performance data that deviates from this region, and outputs the constraint condition to feasible solution search unit 150 . Constraint condition setting section 130 also sets at least one of the upper limit value and the lower limit value of the post-process component concentration as a post-process component concentration constraint condition, and outputs it to feasible solution search section 150 .

If the operating cost is taken into account when searching for operating conditions, the operating cost to be minimized may be set as an evaluation function by the evaluation function setting unit 140 together with the constraint conditions.

(S120: Feasible solution search process)
After that, the feasible solution search unit 150 searches for optimum operating conditions based on the estimated model and the constraint conditions. That is, the feasible solution searching unit 150 solves the optimization problems represented by the following equations (1) to (4) to find the optimum solution. Note that the function f(u, v) representing the quantile regression model and the probability density function p(u) are non-linear functions and cannot be solved by mathematical methods. Therefore, in this embodiment, a differential evolution method, which is one of evolutionary algorithms, is used as an optimization method.

The above variables are as follows.

The decision variable u is a target variable to be obtained as the optimum operating conditions, for example, impeller initial height, treatment time, stirring speed, powdered CaO charging amount, secondary slag charging amount, and B-Al. Weights are respectively set for each decision variable u in terms of the cost of KR processing. Further, in order to search for a solution within the region where the training data is distributed, the likelihood constraint condition is such that the logarithmic degree function (that is, probability density function) p(u) is equal to or less than the control limit α, A constraint expression of equation (2) is set. Furthermore, the constraint expression (3) is set as the post-treatment component concentration constraint condition so that the post-treatment component concentration is equal to or less than the target upper limit S concentration. Furthermore, the constraint expression of Equation (4) is set as the likelihood constraint condition so that the value of each decision variable falls within the specified range of upper and lower limits [u _i ^lower , u _i ^upper ].

The evaluation function J represented by the formula (1) expresses the secondary material cost as the operating cost in the KR treatment. The feasible solution searching unit 150 solves the optimization problem represented by the above formulas (1) and formulas (2) to (4), and finds the optimum operation that improves the component predictive value of the post-treatment S concentration and achieves cost reduction. Ask for conditions.

In order to confirm the effect of setting the constraint conditions, the optimization target variable u was calculated for the first 50 charges of the evaluation data by optimization calculation based on equations (1) to (4). The estimation model shown in FIG. 3 was used as the machine learning model. As a comparative example, FIG. 10 shows the distribution of the optimum solution obtained when no constraint is added, and as an example, FIG. 11 shows the distribution of the optimum solution obtained when the constraint is added. . In FIGS. 10 and 11, distribution 1 shows the amount of charged CaO powder-secondary slag charge, distribution 2 shows the amount of charged CaO powder-treatment time, and distribution 3 shows the amount of charged CaO powder-B-Al distribution. showing. In FIGS. 10 and 11, ● represents training data and ▴ represents the optimal solution. In FIG. 10, as shown clearly in distributions 1 and 2, optimal solutions were obtained where training data did not exist. By adding constraint conditions, as shown in FIG. 11, an optimal solution was obtained where training data existed.

<3. Hardware configuration>
Next, the hardware configuration of the operating condition calculation device 100 according to the embodiment of the present invention will be described in detail with reference to FIG. 12 . FIG. 12 is a block diagram for explaining the hardware configuration of the operating condition calculation device 100 according to this embodiment.

The operating condition calculation device 100 mainly includes a CPU 901 , a ROM 903 and a RAM 905 . The operating condition calculation device 100 further includes a bus 907 , an input device 909 , an output device 911 , a storage device 913 , a drive 915 , a connection port 917 and a communication device 919 .

The CPU 901 functions as an arithmetic processing device and a control device, and controls all or part of the operation in the operating condition arithmetic device 100 according to various programs recorded in the ROM 903, the RAM 905, the storage device 913, or the removable recording medium 921. . A ROM 903 stores programs and calculation parameters used by the CPU 901 . The RAM 905 temporarily stores programs used by the CPU 901, parameters that change as appropriate during execution of the programs, and the like. These are interconnected by a bus 907 comprising an internal bus such as a CPU bus.

The bus 907 is connected to an external bus such as a PCI (Peripheral Component Interconnect/Interface) bus via a bridge.

The input device 909 is operation means operated by a user, such as a mouse, keyboard, touch panel, button, switch, and lever. Further, the input device 909 may be, for example, remote control means (so-called remote control) using infrared rays or other radio waves, or an external connection device 923 such as a PDA corresponding to the operation of the operating condition calculation device 100. There may be. Further, the input device 909 is composed of, for example, an input control circuit that generates an input signal based on information input by the user using the operation means described above and outputs the signal to the CPU 901 . A user of the operating condition calculation device 100 can input various data to the operating condition calculation device 100 and instruct processing operations by operating the input device 909 .

The output device 911 is configured by a device capable of visually or audibly notifying the user of the acquired information. Such devices include display devices such as CRT display devices, liquid crystal display devices, plasma display devices, EL display devices and lamps, audio output devices such as speakers and headphones, printer devices, mobile phones, and facsimiles. The output device 911 outputs, for example, results obtained by various processes performed by the operating condition calculation device 100 . Specifically, the display device displays the results obtained by various processes performed by the operating condition calculation device 100 as text or images. On the other hand, the audio output device converts an audio signal including reproduced audio data and acoustic data into an analog signal and outputs the analog signal.

The storage device 913 is a data storage device configured as an example of a storage unit of the operating condition calculation device 100 . The storage device 913 is configured by, for example, a magnetic storage device such as a HDD (Hard Disk Drive), a semiconductor storage device, an optical storage device, or a magneto-optical storage device. The storage device 913 stores programs executed by the CPU 901, various data, and various data obtained from the outside.

The drive 915 is a reader/writer for recording media, and is built in or externally attached to the operating condition calculation device 100 . The drive 915 reads information recorded on a removable recording medium 921 such as a mounted magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, and outputs the information to the RAM 905 . The drive 915 can also write records to a removable recording medium 921 such as an attached magnetic disk, optical disk, magneto-optical disk, or semiconductor memory. The removable recording medium 921 is, for example, a CD medium, a DVD medium, a Blu-ray (registered trademark) medium, or the like. Also, the removable recording medium 921 may be a CompactFlash (registered trademark) (CompactFlash: CF), a flash memory, an SD memory card (Secure Digital memory card), or the like. Also, the removable recording medium 921 may be, for example, an IC card (Integrated Circuit card) equipped with a contactless IC chip, an electronic device, or the like.

The connection port 917 is a port for directly connecting equipment to the operating condition calculation device 100 . Examples of the connection port 917 include a USB (Universal Serial Bus) port, IEEE1394 port, SCSI (Small Computer System Interface) port, RS-232C port, and the like. By connecting the external connection device 923 to the connection port 917 , the operating condition calculation device 100 directly acquires various data from the external connection device 923 and provides various data to the external connection device 923 .

The communication device 919 is, for example, a communication interface configured with a communication device or the like for connecting to the communication network 925 . The communication device 919 is, for example, a communication card for wired or wireless LAN (Local Area Network), Bluetooth (registered trademark), or WUSB (Wireless USB). Further, the communication device 919 may be a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), a modem for various types of communication, or the like. This communication device 919 can, for example, transmit and receive signals to and from the Internet or other communication devices according to a predetermined protocol such as TCP/IP. A communication network 925 connected to the communication device 919 is configured by a wired or wireless network, and may be, for example, the Internet, home LAN, infrared communication, radio wave communication, satellite communication, or the like. .

An example of the hardware configuration capable of realizing the functions of the operating condition calculation device 100 according to the embodiment of the present invention has been described above. Each component described above may be configured using general-purpose members, or may be configured by hardware specialized for the function of each component. Therefore, it is possible to appropriately change the hardware configuration to be used according to the technical level at which the present embodiment is implemented.

Although the preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, the present invention is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field to which the present invention belongs can conceive of various modifications or modifications within the scope of the technical idea described in the claims. It is understood that these also naturally belong to the technical scope of the present invention.

REFERENCE SIGNS LIST 100 operating condition calculation unit 110 performance data acquisition unit 120 estimation model construction unit 130 constraint condition setting unit 140 evaluation function setting unit 150 feasible solution search unit 160 output unit 200 performance data storage unit

Claims (8)

An operating condition calculation device for calculating operating conditions in a molten pig iron or molten steel refining process,
Based on the performance data in the past operation, the preconditions before the refining process and the operating conditions related to the refining process are input, and the component concentration after the refining process is equal to or higher than the actual component concentration after the refining process with a predetermined probability. An estimation model construction unit that constructs an estimation model that outputs a post-processing component concentration such that
a constraint condition setting unit that sets a constraint condition for the operating condition as a likelihood constraint condition using a probability density function, and sets a constraint condition for the post-treatment component concentration as a post-treatment component concentration constraint condition;
A feasible solution searching unit that searches for the operating condition based on the estimation model, the likelihood constraint, and the post-processing component concentration constraint;
An operating condition computing device. An evaluation function setting unit for setting an operating cost to be minimized as an evaluation function,
The operating condition according to claim 1, wherein the feasible solution searching unit searches for the operating condition that satisfies the estimation model, the likelihood constraint and the post-treatment component concentration constraint, and minimizes the evaluation function. Arithmetic unit. The estimation model is
calculating the expected value of the post-treatment component concentration by an expected value prediction model using the preconditions and the operating conditions as inputs;
3. The operating condition calculation device according to claim 1, which outputs the post-treatment component concentration obtained by performing quantile regression using the expected value as an input. 4. The operating condition calculation device according to claim 3, wherein said expected value prediction model uses a supervised regression model. The operating condition calculation device according to claim 1 or 2, wherein said estimation model outputs a post-treatment component concentration obtained by performing quantile regression with said precondition and said operating condition as inputs. The operating condition calculation device according to any one of claims 1 to 5, wherein the probability density function assumes a mixed Gaussian distribution. The operating condition calculation device according to any one of claims 1 to 6, wherein said feasible solution searching section searches for said operating conditions by an evolutionary algorithm. An operating condition calculation method for calculating operating conditions in a molten pig iron or molten steel refining process,
Based on the performance data in the past operation, the preconditions before the refining process and the operating conditions related to the refining process are input, and the component concentration after the refining process is equal to or higher than the actual component concentration after the refining process with a predetermined probability. An estimation model construction step for constructing an estimation model that outputs a post-processing component concentration such that
a likelihood constraint setting step of setting a constraint on the operating conditions as a likelihood constraint using a probability density function;
a post-treatment component concentration constraint setting step of setting a constraint on the post-treatment component concentration as a post-treatment component concentration constraint;
a feasible solution search step of searching for the operating condition based on the estimation model, the likelihood constraint, and the post-processing component concentration constraint;
Operating condition calculation method , including

Images