JP7585086B2 - Apparatus and method for producing molten cast iron - Google Patents

Description

The present invention relates to a composition adjustment device and method, and in particular to a composition adjustment device and method for adjusting the composition of molten spheroidal graphite cast iron for each cup.

Spheroidal graphite cast iron, or ductile cast iron, has excellent strength and ductility and is used in a variety of products, including iron pipes for water supply and sewerage systems. Ductile cast iron is manufactured through a graphite spheroidization process in which magnesium (Mg) is added to the molten metal to turn the graphite that crystallizes into spheroids.

As shown in JP 2006-316331 A (Patent Document 1), one method of graphite spheroidization is to add iron-coated Mg wire to molten cast iron using a wire feeder method and spheroidize the graphite. This iron-coated Mg wire is a wire material that uses metallic Mg or an Mg alloy such as an Fe-Si-Mg alloy as the core material, and is coated and formed using a thin steel plate or thin steel pipe as the covering material.

As shown in Japanese Patent Laid-Open Publication No. 3-130344 (Patent Document 2), it has also been practiced to produce molten spheroidal graphite cast iron by adding a siliconizing agent (Fe-Si) as an inoculant before adding a spheroidizing agent containing Mg to the molten iron.

JP 2006-316331 A Japanese Patent Application Publication No. 3-130344

When producing molten spheroidal graphite cast iron, additives such as siliconizing agents are used in addition to spheroidizing agents containing magnesium. In other words, the composition of the molten spheroidal graphite cast iron is adjusted by supplying additives containing siliconizing agents and spheroidizing agents to the molten iron.

In graphite spheroidizing processing, because the yield of magnesium varies for each processing cup depending on various conditions, the amount of spheroidizing agent supplied is generally not constant but is changed for each cup. Therefore, when using a spheroidizing agent containing magnesium and silicon in a predetermined content ratio, such as the iron-coated Mg wire in Patent Document 1, the amount of silicon added to the molten metal when the spheroidizing agent is supplied also varies. For this reason, if the amount of siliconizing agent supplied to the processing cup is set according to a predetermined standard without taking into account the amount of spheroidizing agent supplied, there is a possibility that the composition of the spheroidal graphite cast iron molten metal will not be as desired.

The present invention has been made to solve the above problems, and its purpose is to provide a composition adjustment device and method that can appropriately adjust the composition of molten spheroidal graphite cast iron in each cup.

Another objective is to provide a molten cast iron manufacturing device equipped with such a composition adjustment device.

A composition adjustment device according to one aspect of the present invention adjusts the composition of the spheroidal graphite cast iron molten metal for each cup in a cast iron molten metal manufacturing apparatus that supplies a spheroidizing agent containing magnesium and silicon and a siliconizing agent to the molten metal in a ladle to produce spheroidal graphite cast iron molten metal. This composition adjustment device includes a spheroidizing agent amount determination means, a memory means, and a siliconizing agent amount determination means. The spheroidizing agent amount determination means determines the amount of spheroidizing agent to be supplied for each cup. The memory means stores the silicon content of the spheroidizing agent and the target silicon content for the processing cup. The siliconizing agent amount determination means determines the amount of siliconizing agent to be supplied to the processing cup based on the amount of molten metal in the processing cup, the silicon content of the molten metal in the processing cup before the composition adjustment, the amount of spheroidizing agent to be supplied to the processing cup determined by the spheroidizing agent amount determination means, and the silicon content of the spheroidizing agent and the target silicon content for the processing cup stored in the memory means.

Preferably, the memory means further stores the magnesium content of the spheroidizing agent and the target magnesium content for the processing cup, and the spheroidizing agent amount determination means determines the amount of spheroidizing agent to be supplied to the processing cup based on the magnesium content of the spheroidizing agent stored in the memory means and the target magnesium content for the processing cup.

Preferably, the component adjustment device further includes a data storage means for storing performance data including a pair of the amount of magnesium input in past cups and the magnesium content of the molten metal after graphite spheroidization treatment, or result data indicating the yield, and the spheroidizing agent amount determination means predicts the magnesium content or yield of the molten metal after treatment for each treatment cup based on the past performance data stored in the data storage means, and determines the amount of spheroidizing agent to be supplied using the prediction result.

More preferably, the component adjustment device further includes a model storage means for storing the learning model generated by machine learning, and the spheroidizing agent amount determination means inputs at least past performance data into the learning model to predict the magnesium content or yield of the molten metal after processing.

Preferably, the storage means stores the planned amount of molten metal and the target silicon content for each cup as processing cup schedule information.

In this case, it is more preferable that the storage means prestores target composition information including the required magnesium content and silicon content for each component type of the molten metal, and that the composition adjustment device further includes a schedule setting means for setting processing cup schedule information based on the component type and amount of molten metal for each cup input by the user and the target composition information stored in the storage means.

According to another aspect of the present invention, a cast iron molten metal manufacturing apparatus includes the above-mentioned component adjustment device, and includes a spheroidizing agent supplying means for injecting a spheroidizing agent into the molten metal in accordance with a supply amount determined by the spheroidizing agent amount determining means, and a siliconizing agent supplying means for injecting a siliconizing agent into the molten metal in accordance with a supply amount determined by the siliconizing agent amount determining means before the spheroidizing agent is supplied by the spheroidizing agent supplying means.

A composition adjustment method according to yet another aspect of the present invention is a composition adjustment method for adjusting the composition of molten spheroidal graphite cast iron for each cup in a molten cast iron manufacturing apparatus that produces molten spheroidal graphite cast iron by adding a spheroidizing agent containing magnesium and silicon and a siliconizing agent to the molten metal, and includes the steps of determining the amount of spheroidizing agent to be supplied for each cup, and determining the amount of siliconizing agent to be supplied to the processing cup based on the amount of molten metal in the processing cup, the silicon content of the molten metal in the processing cup before the composition adjustment, the determined amount of spheroidizing agent to be supplied to the processing cup, the silicon content of the spheroidizing agent stored in the memory means, and the target silicon content for the processing cup.

According to the present invention, the composition of the molten spheroidal graphite cast iron can be appropriately adjusted in each cup.

FIG. 2 is a diagram showing a schematic diagram of a manufacturing process of ductile cast iron in each embodiment of the present invention. 1 is a functional block diagram showing a functional configuration of a component adjustment device according to a first embodiment of the present invention. 5A and 5B are diagrams showing examples of data structures of a product table and processing schedule information in each embodiment of the present invention. 4 is a flowchart showing a component adjustment process according to the first embodiment of the present invention. 1A is a block diagram showing the functional configuration of a spheroidizing agent supply control device in a component adjustment device according to embodiment 2 of the present invention, and FIG. 1B is a block diagram showing the functional configuration of a learning device according to embodiment 2 of the present invention. FIG. 13 is a diagram conceptually illustrating a method for predicting Mg yield using a prediction model in the second embodiment of the present invention, in comparison with a comparative example. FIG. 11 is a diagram showing a simulation result when predicting Mg yield using a prediction model in the second embodiment of the present invention, in comparison with a comparative example. FIG. 11 is a diagram showing types of error functions (loss functions) used for evaluating a prediction model in the second embodiment of the present invention. 10 is a flowchart showing a spheroidizing agent supply process in the second embodiment of the present invention.

The embodiment of the present invention will be described in detail with reference to the drawings. Note that the same or corresponding parts in the drawings are given the same reference numerals and their description will not be repeated.

<Outline of the ductile cast iron manufacturing process>
First, an overview of the manufacturing process of ductile cast iron will be described with reference to Fig. 1. Fig. 1 is a diagram that shows a schematic diagram of the manufacturing process of ductile cast iron in the present embodiment.

In this embodiment, ductile cast iron is manufactured through the steps of melting process P1, desulfurization process P2, molten metal holding process P3, molten metal component analysis process P4, component adjustment process P5, spheroidizing process P6, slag removal process P7, and casting process P8 in that order. Processes P4 to P7 are carried out by the molten cast iron manufacturing apparatus 10.

The molten cast iron manufacturing apparatus 10 supplies (adds) a spheroidizing agent containing magnesium and silicon and a siliconizing agent to the molten metal in a ladle 15 to produce molten spheroidal graphite cast iron, and is equipped with a composition adjustment device 20, a QV (countback) analyzer 14, a siliconizing agent supply device 17, and a cored wire device 21. The molten cast iron manufacturing apparatus 10 also supplies additives other than the siliconizing agent, such as carbon (C), to the molten metal to produce molten spheroidal graphite cast iron. The siliconizing agent supply device 17 may also function as a supply device for additives other than the siliconizing agent. The composition adjustment device 20 is realized by a computer including a processor such as a CPU (Central Processing Unit) and a memory.

In the melting process P1, molten metal such as pig iron, scrap iron, steel, etc. is melted in a melting furnace 11 such as a cupola. As the melting furnace 11, for example, an electric furnace may be used.

In the desulfurization process P2, the molten metal produced in the melting furnace 11 is poured into the desulfurization ladle 12, and the sulfur content in the molten metal in the ladle 12 is removed. Note that the desulfurization process P2 may be omitted.

In the molten metal holding process P3, the desulfurized molten metal is poured into a low-frequency furnace 13 for holding the molten metal and is held there.

In the molten metal composition analysis process P4, the composition of the molten metal (original molten metal composition) in the low-frequency furnace 13 is analyzed by the QV analyzer 14. After the composition analysis, a portion of the molten metal in the low-frequency furnace 13 is poured into the ladle 15. The amount of molten metal required to produce ductile cast iron (product) in the final casting process P8 is poured into the ladle 15, and the subsequent processes P5 to P7 are carried out for each ladle 15 (for each cup).

In the composition adjustment process P5, the composition of the molten metal is adjusted based on the analysis results by the QV analyzer 14. Specifically, the amounts of elemental components other than magnesium (Mg), such as carbon (C) and silicon (Si), to be added to the molten metal in the ladle 15 are adjusted. The amount of each additive added to the ladle 15 is calculated by the composition adjustment device 20. In this embodiment, in the composition adjustment process P5, a siliconizing material is supplied to the molten metal as a silicon adjuster via the siliconizing material supply device 17. The siliconizing material is typically an Fe-Si alloy.

In the spheroidizing process P6, a process is carried out to spheroidize the graphite in the molten metal using the wire feeder method. In this embodiment, magnesium is added to the ladle 15 using a cored wire device 21. The cored wire device 21 is controlled by the composition adjustment device 20.

The cored wire device 21 includes a feeder 23 that pulls out the wire 22 from the wire coil 26 and supplies it to the ladle 15, and a guide member 24 that guides the movement of the wire 22 from the wire coil 26 to the feeder 23. The wire 22 is a spheroidizing agent composed of powder containing magnesium and silicon contained within an iron skin. When the wire 22 is supplied to the ladle 15, the iron skin gradually melts, and magnesium is added (put in) together with silicon to the molten metal in the ladle 15. The wire 22 as a spheroidizing agent contains magnesium and silicon in a predetermined content ratio (weight %). The feeder 23 functions as a spheroidizing agent supplying means and includes a drive unit for pulling out the wire 22 (not shown).

The composition adjustment device 20 determines the amount of wire 22 to be supplied from the supply device 23 each time the spheroidizing process is performed (i.e., for each processing cup) and instructs the supply device 23. The supply device 23 supplies the wire 22 to the ladle 15 for a length that corresponds to the supply amount instructed by the composition adjustment device 20. This adds magnesium to the ladle 15 and produces molten spheroidal graphite cast iron.

In the slag removal process P7, the ladle 15A after the spheroidizing process is removed from the cored wire device 21, and the slag is removed from the molten metal in the ladle 15A. The ladle 15B after the slag removal process is transported to the casting process P8 by a transport means such as a crane. The above-mentioned QV analyzer 14 may analyze the components of the molten metal in the ladle 15A after the spheroidizing process and feed back the analysis results to the composition adjustment device 20. In this case, a QV analyzer that analyzes the components of the molten metal before the spheroidizing process (original molten metal components) and a QV analyzer that analyzes the components of the molten metal after the spheroidizing process (components of the spheroidal graphite cast iron molten metal) may be provided separately.

In the casting process P8, the molten metal in the ladle 15B, i.e., the molten spheroidal graphite cast iron, is poured into a mold to cast ductile cast iron.

If the magnesium content (hereinafter also referred to as "residual Mg amount") of the molten metal in the ladle 15A after the spheroidizing process is less than the specified amount, the quality of the ductile cast iron will deteriorate. Therefore, in the spheroidizing process step P6, it is desirable to add as little magnesium as possible to the ladle 15 so that the residual Mg amount is the target residual Mg amount (above the specified amount). In the spheroidizing process step P6, the residual Mg amount and yield of magnesium vary depending on various conditions, so the appropriate amount of spheroidizing agent differs for each cup. In this embodiment, since the spheroidizing agent contains silicon, if the amount of siliconizing agent supplied in the composition adjustment step P5 is set according to a specified standard without considering the amount of spheroidizing agent supplied, the composition of the molten spheroidal graphite cast iron may not be as desired.

The component adjustment device 20 according to this embodiment has a function of determining the amount of spheroidizing agent to be supplied to the processing cup prior to the component adjustment process P5, and determining the amount of siliconizing agent to be supplied in the component adjustment process P5 based on the result of that determination. The functional configuration and operation of such a component adjustment device 20 will be described in detail below. Note that the component adjustment device 20 may also have a function of determining the amount of additives to be supplied other than the siliconizing agent.

<First embodiment>
(Functional configuration of the composition adjusting device)
Fig. 2 is a functional block diagram showing the functional configuration of the component adjusting device 20 according to this embodiment. Referring to Fig. 2, the component adjusting device 20 includes a storage unit 41, an input unit 48, a schedule setting unit 42, a modifier component acquisition unit 43, a spheroidizing agent amount determination unit 44, and a siliconizing agent amount determination unit 45. The storage unit 41 is realized by a non-volatile memory or a recording medium that is removable from the computer. The input unit 48 is a user interface that accepts input of instructions from a user, and is realized by, for example, a keyboard or a touch panel.

The memory unit 41 prestores spheroidizing agent information 51, which includes the magnesium content and silicon content of the wire 22, and siliconizing agent information 52, which includes the silicon content of the siliconizing agent used in the component adjustment process P5. Note that it is sufficient that the spheroidizing agent information 51 is provided one-to-one with respect to the spheroidizing agent (lot) used, and multiple pieces of spheroidizing agent information 51 may be stored, and may be added and updated as needed. The same applies to the siliconizing agent information 52.

In addition, the memory unit 41 stores in advance target component information for each component type of the molten metal prepared in this equipment as a component type table 53. In the following explanation, the "magnesium content ratio" is represented as the "Mg ratio" and the "silicon content ratio" is represented as the "Si ratio."

An example of the data structure of the component type table 53 is shown in FIG. 3(A). Referring to FIG. 3(A), the component type table 53 includes an item for recording the component types of the molten metal, an item for recording the Si percentage, and an item for recording the Mg percentage. In other words, in the component type table 53, target component information including the required Si percentage and the required Mg percentage is pre-stored in association with component type identification data for identifying the component types of the molten metal (components A, B, ...). Note that the information included in the component type table 53 may also be added and updated at any time.

The memory unit 41 further stores processing cup schedule information 54 for cups (first cup, second cup, ..., nth cup) scheduled for the day, for example. The processing cup schedule information 54 is information set by the schedule setting unit 42. The schedule setting unit 42 sets the processing cup schedule information 54 as shown in FIG. 3(B) based on the component type and molten metal amount for each cup input by the user via the input unit 48 and the target component information in the component type table 53.

Specifically, the schedule setting unit 42 accepts input of the amount of molten metal for each cup via the input unit 48, and also accepts selection of the component types of the molten metal. Then, by referring to the component type table 53, it searches for target component information corresponding to the selected component types for each cup, and sets the Si percentage and Mg percentage contained in the searched target component information as the target Si percentage and target Mg percentage for that cup. As a result, processing cup schedule information 54 including the amount of molten metal (amount of tapped molten metal), component types, target Si percentage, and target Mg percentage for each planned cup is stored in the memory unit 41.

In the processing cup schedule information 54 of FIG. 3(B), an example is shown in which "3200 kg" is input as the amount of molten metal for the first cup, and "component B" is input as the component type. In this case, the target Si percentage and target Mg percentage for the first cup are set as the Si percentage (Sb) and Mg percentage (Mb) associated with component B in the component type table 53 of FIG. 3(A). Note that the processing cup schedule information 54 only needs to include the amount of molten metal and the component type for each cup, and does not need to include the Si percentage and Mg percentage (target component information) previously stored in the component type table 53.

The base metal component acquisition unit 43 acquires the analysis results of the QV analyzer 14 in the molten metal component analysis process P4, i.e., the base metal components.

The spheroidizing agent amount determination unit 44 determines the amount of spheroidizing agent to be supplied for each cup so as to achieve the target Mg ratio in the processing cup schedule information 54. The spheroidizing agent amount determination unit 44 determines the amount to be supplied to the processing cup prior to the component adjustment process P5, and outputs the determined amount of spheroidizing agent to the instruction output unit 47 and to the siliconizing agent amount determination unit 45.

The spheroidizing agent amount determination unit 44 may determine the amount of spheroidizing agent to be supplied from a value input by the user via the input unit 48, or may calculate the amount of spheroidizing agent to be supplied using a predetermined formula. Alternatively, the amount of spheroidizing agent may be determined by predicting the magnesium content or yield of the molten metal after processing for each processing cup. In this embodiment, an example is described in which the amount of spheroidizing agent to be supplied is determined from a value input by the user. A method for predicting the magnesium content or yield of the molten metal after processing will be described later in embodiment 2.

In this embodiment, the user can set the supply amount of spheroidizing agent based on the "target residual Mg amount" obtained from the amount of molten metal in the processing cup and the target Mg ratio contained in the processing cup schedule information 54, the "Mg ratio of the spheroidizing agent" contained in the spheroidizing agent information 51, and a previously set "Mg yield." When a predetermined calculation formula is used, the spheroidizing agent amount determination unit 44 can calculate the supply amount of spheroidizing agent by applying these numerical values to the calculation formula.

In the spheroidizing process P6, the instruction output unit 47 instructs the supply device 23 to supply the wire 22 to the ladle 15 with a length corresponding to the amount of spheroidizing agent supplied determined by the spheroidizing agent amount determination unit 44, thereby carrying out spheroidizing of the graphite. The spheroidizing agent amount determination unit 44 and the instruction output unit 47 constitute a "spheroidizing agent supply control device 30" that controls the supply of the spheroidizing agent.

The siliconizing agent amount determination unit 45 determines the amount of siliconizing agent to be supplied for each processing cup, taking into account the amount of spheroidizing agent to be supplied determined by the spheroidizing agent amount determination unit 44. Specifically, the amount of siliconizing agent to be supplied to the processing cup is determined based on the "amount of molten metal in the processing cup" contained in the processing cup schedule information 54 in the memory unit 41, the "Si percentage of the molten metal in the processing cup before the composition adjustment" acquired by the original molten metal composition acquisition unit 43, the "amount of spheroidizing agent to be supplied to the processing cup" determined by the spheroidizing agent amount determination unit 44, the "Si percentage of the spheroidizing agent" contained in the spheroidizing agent information 51 in the memory unit 41, the "Si percentage of the siliconizing agent" contained in the siliconizing agent information 52, and the "target Si percentage for the processing cup" contained in the processing cup schedule information 54.

The "amount of molten metal in the processing cup" may be an actual measured value of the amount of molten metal tapped into the ladle 15. The siliconizing material may be pure silicon, in which case the siliconizing material information 52 in the memory unit 41 can be omitted since the "Si percentage of the siliconizing material" is not required to calculate the amount of siliconizing material.

The supply amount of siliconizing material determined by the siliconizing material amount determination unit 45 is output to the instruction output unit 46. In the composition adjustment process P5, the instruction output unit 46 instructs the siliconizing material supply device 17 to supply siliconizing material to the ladle 15 according to the supply amount of siliconizing material determined by the siliconizing material amount determination unit 45, thereby adjusting the silicon content of the processing cup.

The functions of the schedule setting unit 42, the base molten metal component acquisition unit 43, the spheroidizing agent amount determination unit 44, the siliconizing agent amount determination unit 45, and the instruction output units 46 and 47 can be realized by a processor such as a CPU executing software. Of these functional units, the spheroidizing agent amount determination unit 44 and the instruction output unit 47 constituting the spheroidizing agent supply control device 30 may be realized by a processor separate from the other functional units. In other words, the additive supply control device 60, which controls the supply of additives such as siliconizing agents, and the spheroidizing agent supply control device 30 may each be realized by an independent computer.

(Operation of the composition adjusting device)
The operation of the component adjustment device 20 will be described with reference to Figures 1 to 4. Figure 4 is a flow chart showing the component adjustment process in this embodiment.

First, the schedule setting unit 42 of the component adjustment device 20 sets the processing cup schedule information 54 based on the information input by the user via the input unit 48, and records it in the memory unit 41 (step S101).

During operation of the molten cast iron manufacturing apparatus 10, a predetermined amount of molten metal is poured from the low frequency furnace 13 into the ladle 15 in the order set in the processing cup schedule information 54. For example, for the first cup, 3,200 kg of molten metal is poured into the ladle 15.

In the molten metal composition analysis process P4, the QV analyzer 14 analyzes the composition of the molten metal in the low-frequency furnace 13 before it is tapped into the ladle 15, or the composition of the molten metal in the ladle 15, i.e., the original molten metal composition. This allows the original molten metal composition acquisition unit 43 to acquire the pre-adjustment Si ratio of the treated cup (step S103).

The spheroidizing agent amount determination unit 44 determines the amount of spheroidizing agent to be supplied at least prior to the component adjustment process P5 (step S105). In this embodiment, a value directly input via the input unit 48 is determined as the amount of spheroidizing agent to be supplied. The amount of spheroidizing agent is typically input to the input unit 48 each time (for each processing cup), but as a modified example, the amount of spheroidizing agent for multiple cups may be input in advance, for example, when setting the processing cup schedule information 54. In this case, the amount of spheroidizing agent for each cup may be included in the processing cup schedule information 54 stored in the memory unit 41.

In the component adjustment process P5, the siliconizing agent amount determination unit 45 determines the supply amount of siliconizing agent (step S107). Specifically, the supply amount of siliconizing agent is calculated according to the following formula (1).

Amount of siliconizing agent = {(target Si content - Si content before adjustment) x amount of molten metal - amount of spheroidizing agent x Si content of spheroidizing agent} ÷ Si content of siliconizing agent ... formula (1)
In other words, the siliconizing agent amount determination unit 45 calculates the supply amount of silicon agent as the "silicon shortage" obtained by subtracting the "planned amount of silicon to be added" calculated from the supply amount of spheroidizing agent and its Si percentage from the "target amount of silicon" calculated from the target Si percentage, the Si percentage before adjustment, and the amount of molten metal, and dividing the result by the Si percentage of the siliconizing agent. Note that the "supply amount of spheroidizing agent" and the "Si percentage of the spheroidizing agent" used in calculating the planned amount of silicon to be added may be replaced with the "Mg amount in the spheroidizing agent" and the "ratio of the Si amount to the Mg amount", respectively.

When the supply amount of siliconizing material is determined by the siliconizing material amount determination unit 45, the instruction output unit 46 outputs an instruction to the siliconizing material supply device 17 to add the determined supply amount of siliconizing material (step S109). This adjusts the silicon content of the molten metal in the ladle 15.

Then, the instruction output unit 47 outputs an instruction to the supply device 23 to add the amount of spheroidizing agent determined in step S105 (step S111). This causes magnesium (together with silicon) to be added to the ladle 15, resulting in molten spheroidal graphite cast iron. This produces molten spheroidal graphite cast iron with adjusted composition (step S112).

The processes in steps S103 to S111 are repeated until the spheroidization process is completed for all cups set as the processing cup schedule information 54 (NO in step S113).

As described above, even if the spheroidizing process P6 is performed after the component adjustment process P5, the spheroidizing agent amount determination unit 44 determines the amount of spheroidizing agent to be added to the processing cup prior to the component adjustment process P5, and the siliconizing agent amount determination unit 45 can automatically and appropriately determine (calculate) the amount of siliconizing agent to be added in the component adjustment process P5. This allows the composition of the molten metal to be transported to the casting process P8 to be optimized. In addition, in this embodiment, by performing the component adjustment process P5 first, the transition time from the spheroidizing process P6 to the casting process P8 can be shortened, which has the effect of minimizing the drop in the molten metal temperature and fading of the added magnesium.

The manufacturing process of molten spheroidal graphite cast iron is not limited to the order shown in FIG. 1, and the composition adjustment process may be performed after the spheroidizing treatment process. Even in this case, the siliconizing agent amount determination unit 45 can automatically and appropriately determine (calculate) the amount of siliconizing agent by using the amount of spheroidizing agent determined for each cup by the spheroidizing agent amount determination unit 44, so that the composition of the molten iron can be optimized efficiently.

In addition, in this embodiment, the spheroidizing agent is added to the ladle 15 using the cored wire device 21, but this is not limited to this example.

<Embodiment 2>
In the second embodiment of the present invention, an example will be described in which the spheroidizing agent supply control device of the component adjusting device 20 predicts the magnesium content or yield of the molten metal after processing for each processing cup and determines the amount of spheroidizing agent. The prediction method described in this embodiment has been filed by the applicant as Japanese Patent Application No. 2020-54703.

The basic functional configuration of the component adjustment device in this embodiment is the same as that of the component adjustment device 20 shown in FIG. 2, but instead of the spheroidizing agent supply control device 30 shown in FIG. 2, it is equipped with a spheroidizing agent supply control device 30A shown in FIG. 5(A).

(Functional configuration of the spheroidizing agent supply control device)
The functional configuration of the spheroidizing agent supply control device 30A will be described with reference to Fig. 5(A) , which is a block diagram showing the functional configuration of the spheroidizing agent supply control device 30A according to the present embodiment.

The spheroidizing agent supply control device 30A includes a data storage unit 31, a model storage unit 32, a determination unit 33, an instruction output unit 34, and a result acquisition unit 35.

The data storage unit 31 stores past performance data for each spheroidizing process (each cup). The data storage unit 31 may store only the most recent predetermined number (e.g., two) of performance data (past cup information). The performance data includes molten metal condition data indicating various conditions of the spheroidizing process and result data of the spheroidizing process.

The molten metal condition data includes all or at least one of the amount of molten metal to be processed, the temperature before processing, and the original molten metal composition. The original molten metal composition includes both the C content and the S content, or at least one of them.

In this embodiment, the result data represents the Mg yield. Note that instead of the Mg yield, the amount of Mg input and the amount of residual Mg may be used as the result data. In other words, the data storage unit 31 only needs to store performance data including a pair of the amount of magnesium input in a past cup and the magnesium content of the molten metal after the graphite spheroidization process, or result data indicating the yield.

The model storage unit 32 stores the prediction model 32A. The prediction model 32A is a learning model generated in the learning device 100 described below.

The determination unit 33 predicts the Mg yield of the molten metal after processing for each processing cup based on the past performance data stored in the data storage unit 31, and determines the supply amount of spheroidizing agent. In other words, the determination unit 33 includes a prediction unit 36 that predicts the Mg yield by inputting at least the past performance data into a prediction model 32A, and a calculation unit 37 that calculates the amount of Mg input according to the prediction result (predicted yield) by the prediction unit 36, and calculates the supply amount of spheroidizing agent from the amount of Mg input. The method of predicting the Mg yield by the prediction unit 36 will be described in detail later.

The calculation unit 37 calculates the amount of Mg added using the following formula (2):

Amount of Mg added = (target residual Mg x amount of molten metal to be treated) / predicted yield ... formula (2)
The calculation unit 37 also calculates the amount of spheroidizing agent based on the amount of Mg input calculated by formula (2) and the Mg ratio of the spheroidizing agent (wire 22). In this embodiment, the amount of spheroidizing agent to be added to the processing cup determined in this manner is output to the siliconizing agent amount determination unit 45. The siliconizing agent amount determination unit 45 determines the amount of siliconizing agent to be added to the processing cup by the above calculation formula (1), as in the above-described embodiment.

The instruction output unit 34 instructs the supply device 23 to supply the wire 22 to the ladle 15 with a length corresponding to the amount of spheroidizing agent determined by the determination unit 33, thereby carrying out spheroidization of the graphite.

The result acquisition unit 35 acquires the analysis results of the molten metal components in the ladle 15A after the graphite spheroidization process from the QV analyzer 14 and detects the amount of residual Mg. It also calculates the Mg yield based on the detected amount of residual Mg. The calculated Mg yield is stored in the data storage unit 31 as result data.

The functions of the determination unit 33, instruction output unit 34, and result acquisition unit 35 of the spheroidizing agent supply control device 30A can be realized by a processor such as a CPU executing software. The data storage unit 31 and model storage unit 32 are realized by a non-volatile memory or a recording medium that is removable from the computer.

(Method of predicting Mg yield)
A method for predicting Mg yield using prediction model 32A in this embodiment will be described with reference to Fig. 6. Fig. 6(A) conceptually shows a prediction method according to a comparative example, and Fig. 6(B) conceptually shows a prediction method according to this embodiment (example).

Referring to FIG. 6(A), a typical prediction model is machine-learned to predict the objective variable, Mg yield (or residual Mg amount), using the processing cup information to be processed as an explanatory variable. The processing cup information corresponds to the molten metal condition data Da, and the Mg yield (actual value) corresponds to the result data Db. The molten metal condition data Da includes, for example, basic information for each process (date, time, product type), the original molten metal composition indicating the ratio of each element component (C, Si, Mn, P, S, Cu, etc.) of the molten metal before processing, the amount of processed molten metal specified by the remaining molten metal amount and the tapped molten metal amount, the original molten metal temperature indicating the molten metal temperature before processing, and further includes the amount of additives.

In contrast, referring to FIG. 6(B), the prediction model 32A in this embodiment is machine-trained to predict Mg yield using not only processed cup information but also past cup information, i.e., performance data, as explanatory variables. Here, the performance data includes Mg yield, which is the answer (actual measurement result) to the objective variable.

When the performance data for the most recent two cups are included in the explanatory variables, the performance data for the cup before includes molten metal condition data Da1 and result data Db1 related to the spheroidizing process for the cup before, and the performance data for two cups before includes molten metal condition data Da2 and result data Db2 related to the spheroidizing process for the cup before.

(Method of generating a predictive model)
A method for generating the prediction model 32A will be described with reference to Fig. 5B. Fig. 5B is a block diagram showing the functional configuration of the learning device 100.

The learning device 100 includes a history data storage unit 101, a model generation unit 102, a model storage unit 103, and an evaluation unit 104. The learning device 100 is realized by a computer including a processor such as a CPU and a memory.

The model generation unit 102 generates processing history data for each spheroidization process performed by the spheroidizing agent supply control device 30A, and stores the generated processing history data 1, 2, ..., n in the history data storage unit 101.

Each processing history data (k) corresponds to the performance data described above, and includes molten metal condition data and result data stored in association with unique identification data. The identification data includes, for example, the date and product type included in the basic information shown in FIG. 5(A). The molten metal condition data includes all or at least one of the amount of molten metal processed, the temperature before processing, and the original molten metal composition. The molten metal condition data may also include the processing time.

The model generation unit 102 generates a learning model, i.e., a prediction model 103A, for predicting the Mg yield of the processed cup based on multiple processing history data stored in the history data storage unit 101. Specifically, the prediction model 103A is generated by machine learning the correlation between the molten metal condition data and past performance data (including Mg yield) of the processed cup and the Mg yield, which is the result data of the processed cup. In other words, as shown in FIG. 6(B), a prediction model 103A is generated in which the molten metal condition data and past performance data of the processed cup are explanatory variables and the Mg yield of the processed cup is the objective variable.

As an algorithm used for machine learning, for example, (multiple) regression analysis may be adopted, but other methods such as neural networks, support vector machines, decision trees, k-NN, etc. may also be adopted.

The evaluation unit 104 evaluates the accuracy of the prediction model 103A generated by the model generation unit 102, and outputs the evaluation result to the model generation unit 102. The model generation unit 102 appropriately corrects the prediction model 103A to eliminate the error between the predicted value and the actual measured value (result) of the Mg yield.

Here, it is desirable for the evaluation unit 104 to evaluate the error (loss) as greater when the actual measured value (result) is lower than the Mg yield predicted using the prediction model 103A than when the actual measured value (result) is higher by the same amount. The same is true when the objective variable of the prediction model is the amount of residual Mg in the molten metal after processing.

The error function (loss function) used in the evaluation of the prediction model 103A will be described with reference to FIG. 8. Examples 1 and 2 below are typical error functions used in the evaluation of machine learning. The error function of Example 1, as shown in the graph of FIG. 8(B), expresses the difference between the predicted value and the actual value as an absolute value and calculates the average of the absolute values as the evaluation result. The error function of Example 2, as shown in the graph of FIG. 8(C), squares the difference between the predicted value and the actual value and calculates the square root of the average as the evaluation result.

In contrast, the error function of this embodiment, as shown in the graph of FIG. 8(D), multiplies the value by a constant (e.g., doubles it) only when the difference between the predicted value and the actual measured value is positive, and calculates the average of the absolute values as the evaluation result. Therefore, as shown in FIG. 8(A), if there are several cases in which the difference between the predicted value and the actual measured value is positive (on the positive side), the evaluation result (error) by the evaluation unit 104 may be higher than the evaluation results (errors) of Examples 1 and 2.

By adopting this evaluation method, it is possible to prevent a decrease in the quality of the finished product obtained through the subsequent casting process P8.

The prediction model 103A generated by the model generation unit 102 is stored in advance as a prediction model 32A in the model storage unit 32 of the spheroidizing agent supply control device 30A. Note that the spheroidizing agent supply control device 30A may have the functionality of the learning device 100, so that the prediction model 32A can be updated at any time in the spheroidizing agent supply control device 30A.

(Simulation results)
FIG. 7(A) is a table showing the simulation results when predicting Mg yield using a prediction model of a comparative example, and FIG. 7(B) is a table showing the simulation results when predicting Mg yield using prediction model 103A of the present embodiment.

Figure 7(A) is a table showing the predicted and measured values (objective variables) of Mg yield according to the comparative example shown in Figure 6(A) as the difference from the reference value (target value of Mg yield). Figure 7(B) is a table showing the predicted and measured values (objective variables) of Mg yield according to the prediction method of this embodiment shown in Figure 6(B) as the difference from the reference value (target value of Mg yield).

Referring to FIG. 7(A), in the comparative example, the explanatory variable column CA1 contains only the cup data (processing cup information). The cup data corresponds to the molten metal condition data shown in FIG. 6(A). The objective variable column CA2 shows the actual Mg yield value for each run as the difference from the reference value (target value). The predicted value column CA3 shows the predicted Mg yield for each run when prediction is made using the explanatory variables in column CA1 as input as the difference from the reference value (target value).

Referring to FIG. 7(B), in this embodiment, in addition to the cup data, the explanatory variable column CB1 contains the data from one cup before and the data from two cups before. The data from one cup before and the data from two cups before correspond to performance data, and contain molten metal condition data and yield data (result data). In the objective variable column CB2 and the predicted value column CB3, each value is shown as the difference from a reference value (target value), similar to the columns CA2 and CA3 above.

The evaluation of the simulation results of Figures 7(A) and (B) is shown below.

As shown in Table 1, the magnesium usage rate is the same in both the present embodiment and the comparative example, and the prediction error and the rate of deviation from the target lower limit are higher in the comparative example than in the present embodiment. The prediction error here is the root-mean-square error calculated according to Example 2 in Figure 8, and the error is expressed as the square root of the average of the square of the difference between the predicted value and the actual value. The rate of deviation from the target lower limit represents the rate at which the predicted value falls short of the target lower limit. The magnesium usage rate represents the amount of Mg when the amount of Mg (input amount) for spheroidizing treatment used when operating according to the actual Mg yield value (answer) is taken as 1.

The above evaluation results show that by including past performance data in the explanatory variables of the prediction model 103A, as in this embodiment, it is possible to predict the Mg yield with greater accuracy than in the comparative example. In particular, it was found that the prediction accuracy can be improved by including the objective variable, the Mg yield (actual measured value), in the performance data.

(Operation of Spheroidizing Agent Supply Control Device)
The operation of the spheroidizing agent supply control device 30A will be described with reference to Fig. 5(A) and Fig. 9. Fig. 9 is a flow chart showing the spheroidizing agent supply process in this embodiment.

First, the determination unit 33 of the spheroidizing agent supply control device 30A acquires molten metal condition data for the processing cup from the QV analysis device 14 or the like, and acquires performance data for the past two cups stored in the data storage unit 31 (step S1).

The prediction unit 36 of the determination unit 33 predicts the Mg yield of the processing cup based on the information acquired in step S1 (step S3). That is, the molten metal condition data of the processing cup and the actual data of the past two cups are input to the prediction model 32A, and the output value of the prediction model 32A is obtained as the predicted value.

Then, the calculation unit 37 of the determination unit 33 calculates the amount of Mg to be added based on the Mg yield predicted in step S3, the predetermined target amount of residual Mg, and the amount of molten metal to be processed included in the molten metal condition data for the processing cup (step S5). In addition, the calculation unit 37 calculates (determines) the amount of spheroidizing agent to be supplied based on the amount of Mg added and the Mg ratio of the spheroidizing agent (wire 22) (step S6). The target amount of residual Mg is determined from the amount of molten metal and the target Mg ratio in the processing cup schedule information 54 (Figure 2).

The above-mentioned steps S3 to S6 are executed in place of the step S105 in FIG. 4 (processing for determining the amount of spheroidizing agent supplied). Therefore, in this embodiment, the siliconizing agent amount determination unit 45 determines the amount of siliconizing agent supplied using the amount of spheroidizing agent determined in step S6 (step S107 in FIG. 4).

The instruction output unit 34 instructs the supply device 23 of the cored wire device 21 to supply the wire 22 according to the supply amount determined in step S6 (step S7). As a result, an appropriate amount of magnesium is automatically added to the molten metal in the ladle 15, and the molten metal becomes spheroidal graphite cast iron (step S9).

When the spheroidization of the graphite is completed, the result acquisition unit 35 acquires the results of the component analysis of the molten metal in the ladle 15A and detects the amount of residual Mg (step S11). In addition, the Mg yield (actual value) is calculated based on the detected amount of residual Mg, the amount of molten metal processed, and the amount of Mg added calculated in step S5 (step S13).

The result acquisition unit 35 stores the calculated Mg yield (actual measurement value) in the data storage unit 31 (step S15). In other words, the performance data including the molten metal condition data of the processing cup acquired in step S1 and the result data obtained in step S13 is stored in the data storage unit 31 as the most recent (one cup before) performance data.

In this embodiment, the above process is repeated until the spheroidization process is completed for all cups set as the processing cup schedule information 54 (Figure 2).

(Modification of the second embodiment)
(1) The past performance data used to predict the Mg yield is not limited to the previous two cups. For example, performance data from one cup and three cups before may be used. Note that it is preferable that the performance data used to predict the Mg yield is for the same product type.

(2) The historical data used to predict the Mg yield is preferably multiple cups, but is not limited to two cups and may be three or more cups. Alternatively, the Mg yield may be predicted based on one cup of historical data.

(3) The past performance data used to predict Mg yield includes both molten metal condition data and result data for the processing cup, but may include only result data.

(4) Although both the molten metal condition data of the processing cup and the past performance data are used to predict the Mg yield, it is sufficient that at least the past performance data is used. In other words, the explanatory variables of the prediction model 103A may be only the past performance data.

(5) In the second embodiment, a method was described in which the spheroidizing agent supply control device 30A determines the amount of Mg to be added in real time using the prediction model 103A, but the amount of Mg to be added may also be determined offline.

(6) The method for determining the amount of Mg to be added, which is executed by the spheroidizing agent supply control device 30A, can also be provided as a program. Similarly, the learning method executed by the learning device 100 can also be provided as a program. Such a program can be provided by recording it on an optical medium such as a CD-ROM (Compact Disc-ROM) or a computer-readable non-transitory recording medium such as a memory card. The program can also be provided by downloading it over a network.

The program according to the present invention may execute processing by calling necessary modules in a specified sequence at a specified timing among program modules provided as part of a computer's operating system (OS). In such a case, the program itself does not include the above modules, and executes processing in cooperation with the OS. Programs that do not include such modules may also be included in the program according to the present invention.

The program according to the present invention may also be provided as being incorporated into a part of another program. In this case, the program itself does not include modules included in the other program, and the processing is carried out in cooperation with the other program. Such a program incorporated into another program may also be included in the program according to the present invention.

The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present invention is indicated by the claims, not by the above description, and is intended to include all modifications within the meaning and scope of the claims.

10 Cast iron molten metal manufacturing device, 15, 15A, 15B Ladle, 17 Siliconizing agent supply device, 20 Component adjustment device, 21 Cored wire device, 23 (Spheroidizing agent) supply device, 30, 30A Spheroidizing agent supply control device, 31 Data storage unit, 32, 103 Model storage unit, 32A, 103A Prediction model, 41 Storage unit, 42 Schedule setting unit, 43 Molten metal component acquisition unit, 44 Spheroidizing agent amount determination unit, 45 Siliconizing agent amount determination unit, 48 Input unit, 60 Additive supply control device, 51 Spheroidizing agent information, 52 Siliconizing agent information, 53 Component type table, 54 Processing cup schedule information, 100 Learning device, 101 History data storage unit, 102 Model generation unit, 104 Evaluation unit.

Claims (7)

An apparatus for producing molten cast iron , comprising: a spheroidizing agent containing magnesium and silicon and a siliconizing agent being supplied to molten iron in a ladle to produce molten spheroidal graphite cast iron,
a spheroidizing agent supplying means for supplying the spheroidizing agent to the molten metal;
a siliconizing agent supplying means for supplying the siliconizing agent to the molten metal before the spheroidizing agent is supplied by the spheroidizing agent supplying means;
a composition adjusting device for adjusting the composition of the molten spheroidal graphite cast iron for each cup ,
The composition adjusting device is
a spheroidizing agent amount determination means for predicting the magnesium content or yield of the molten metal after treatment for each treatment cup and determining the supply amount of the spheroidizing agent for each cup using the prediction result ;
a memory means for storing the silicon content of the spheroidizing agent and a target silicon content for the processing cup;
a siliconizing agent amount determination means for determining an amount of siliconizing agent to be supplied to the processing cup based on the amount of molten metal in the processing cup, the silicon content of the molten metal in the processing cup before the components are adjusted, the amount of spheroidizing agent to be supplied to the processing cup determined by the spheroidizing agent amount determination means, the silicon content of the spheroidizing agent stored in the memory means, and a target silicon content for the processing cup ,
The siliconizing agent supplying means supplies the siliconizing agent to the molten metal in a supply amount determined by the siliconizing agent amount determining means;
the spheroidizing agent supplying means supplies the spheroidizing agent to the molten metal after the siliconizing agent has been supplied, in accordance with the supply amount determined by the spheroidizing agent amount determining means . The memory means further stores a magnesium content of the spheroidizing agent and a target magnesium content for the processing cup,
2. The apparatus for producing molten cast iron according to claim 1, wherein the spheroidizing agent amount determination means determines the amount of the spheroidizing agent to be supplied to the treatment cup based on the magnesium content of the spheroidizing agent stored in the memory means and a target magnesium content for the treatment cup. The method further includes a data storage means for storing performance data including a combination of the amount of magnesium added in a past cup and the magnesium content of the molten metal after the graphite spheroidizing treatment, or result data showing a yield,
3. The apparatus for producing molten cast iron according to claim 2, wherein the spheroidizing agent amount determination means predicts the magnesium content or yield of the molten metal after processing for each processing cup based on past performance data stored in the data storage means, and determines the supply amount of the spheroidizing agent using the prediction result. The system further includes a model storage means for storing a learning model generated by machine learning,
4. The apparatus for producing molten cast iron according to claim 3, wherein the spheroidizing agent amount determining means inputs at least the past performance data into the learning model to predict the magnesium content or yield of the molten metal after treatment. The apparatus for producing molten cast iron according to any one of claims 1 to 4, wherein the memory means stores a planned amount of molten metal and a target silicon content rate for each cup as processing cup schedule information. The storage means stores in advance target component information including a required magnesium content and a required silicon content for each component type of the molten metal,
The apparatus for producing molten cast iron according to claim 5, further comprising a schedule setting means for setting the processing cup schedule information based on the component types and molten metal amount for each cup input by a user and the target component information stored in the storage means. A method for producing molten cast iron by adding a spheroidizing agent containing magnesium and silicon and a siliconizing agent to molten iron to produce molten spheroidal graphite cast iron, comprising the steps of:
a first determination step of predicting a magnesium content or a yield of the molten metal after processing for each processing cup, and determining a supply amount of the spheroidizing agent for each cup using the prediction result ;
a second determination step of determining an amount of the siliconizing agent to be supplied to the processing cup based on the amount of the molten metal in the processing cup, the silicon content of the molten metal in the processing cup before the composition adjustment, the determined amount of the spheroidizing agent to be supplied to the processing cup, the silicon content of the spheroidizing agent stored in a memory means , and a target silicon content for the processing cup ;
adding the siliconizing material to the molten metal in accordance with the supply amount determined in the second determination step;
adding the spheroidizing agent to the molten metal after the siliconizing agent has been added in accordance with the supply amount determined in the first determination step .

Images