JPWO2020129173A1 - Learning control device for rolling process - Google Patents

Abstract

Provided is a learning control device for a rolling process that can avoid updating the learning coefficient due to an abnormal actual value. The learning control device of the rolling process calculates the layered learning coefficient this time value based on the current actual value of the rolling process and the current prediction model, and the layered learning coefficient this time value and the layered learning coefficient this time in the learning table. The layered learning coefficient update value is calculated based on the layered learning coefficient previous value stored in the same layer as the value, and the layered learning coefficient this time value and the learning table are divided into the same layer as the layered learning coefficient current value. Based on the memorized comparison result with the previous value of the stratified learning coefficient, the abnormality in the learning of the prediction model is determined, and when it is determined that there is an abnormality in the learning of the prediction model, the previous value of the stratified learning coefficient is updated. do not do.

Description

The present invention relates to a learning control device for a rolling process.

Patent Document 1 discloses a learning control device for a rolling process. According to the learning control device, even when the stratified learning coefficient exceeds the limit value, the stratified learning coefficient is corrected and the stratified learning coefficient is controlled within the limit value. Therefore, the error of the prediction model can be corrected efficiently.

Japanese Patent No. 5759206

However, in the learning control device described in Patent Document 1, even if the actual value is abnormal, the learning coefficient is updated if the learning coefficient does not exceed the limit value. In this case, the stratified learning coefficient is not stable. If the stratified learning coefficient is not stable, the setting calculation will not be stable. If the setting calculation is not stable, the rolling state becomes unstable.

The present invention has been made to solve the above-mentioned problems. An object of the present invention is to provide a learning control device for a rolling process that can avoid updating the learning coefficient due to an abnormal actual value.

The learning control device for the rolling process according to the present invention has a learning table that manages a learning coefficient for correcting an error between the actual value of the rolling process and the model predicted value obtained by the prediction model as a stratified learning coefficient for each layer. Layered learning coefficient based on the current actual value of the rolling process and the current prediction model. Layered learning coefficient for calculating the current value. Layer calculated by the layered learning coefficient current value calculation unit and the layered learning coefficient. Separate learning coefficient The layered learning coefficient update value that calculates the layered learning coefficient update value based on the current value and the layered learning coefficient previous value stored in the same layer as the current value in the learning table. The calculation unit and the stratified learning coefficient this time value The stratified learning coefficient calculated by the calculation unit This time value and the stratified learning coefficient previously value stored in the same layer as the stratified learning coefficient in the learning table An abnormality determination unit that determines an abnormality in the learning of the prediction model based on the comparison result of the above, and a layer that does not update the previous value of the stratified learning coefficient when the abnormality determination unit determines that there is an abnormality in the learning of the prediction model. It is equipped with a separate learning coefficient update unit.

According to the present invention, the stratified learning coefficient is not updated when it is determined that there is an abnormality in the learning of the prediction model. Therefore, it is possible to avoid updating the learning coefficient due to an abnormal actual value.

It is a block diagram of the rolling process to which the learning control device of the rolling process in Embodiment 1 is applied. It is a flowchart for demonstrating the outline of the operation of the learning control apparatus of the rolling process in Embodiment 1. FIG. It is a hardware block diagram of the learning control apparatus of the rolling process in Embodiment 1. FIG. It is a block diagram of the rolling process to which the learning control device of the rolling process in Embodiment 2 is applied. It is a flowchart for demonstrating the outline of the 1st operation of the learning control apparatus of the rolling process in Embodiment 2. It is a flowchart for demonstrating the update of the information of the update times of the stratified learning coefficient by the learning control apparatus of the rolling process in Embodiment 2. It is a flowchart for demonstrating the outline of the 2nd operation of the learning control apparatus of the rolling process in Embodiment 2. It is a block diagram of the rolling process to which the learning control device of the rolling process in Embodiment 3 is applied. It is a flowchart for demonstrating the outline of the operation of the learning control apparatus of the rolling process in Embodiment 3.

A mode for carrying out the present invention will be described with reference to the accompanying drawings. In each figure, the same or corresponding parts are designated by the same reference numerals. The duplicate description of the relevant part will be simplified or omitted as appropriate.

Embodiment 1.
FIG. 1 is a block diagram of a rolling process to which the learning control device for the rolling process according to the first embodiment is applied.

In FIG. 1, the rolling process is a process of rolling a rolled material. Although detailed description is omitted, for example, the rolling process includes a heating furnace, a rough rolling machine, a finishing rolling machine, a winding machine and the like.

As shown in FIG. 1, the learning control device 1 includes a learning table storage unit 2, a stratified learning coefficient, a current value calculation unit 3, a stratified learning coefficient update value calculation unit 4, a stratified learning coefficient difference determination unit 5, and a layer. A separate learning coefficient updating unit 6 is provided.

The learning table storage unit 2 stores the information of the learning table. The training table is stratified by multiple conditions in the rolling process. The learning table is a learning coefficient that corrects the error between the actual value obtained from a measuring instrument (not shown) provided in the rolling process and the model recalculation predicted value obtained by the same prediction model as the set calculation based on the actual value. Is managed as a stratified learning coefficient for each stratum. For example, the rolling load learning table is composed of two layers, steel type and plate thickness. The number of layers in the rolling load learning table is appropriately set according to the learning coefficient. For example, the number of layers in the rolling learning table may be set to a larger number, such as three or four.

The stratified learning coefficient This time value calculation unit 3 includes the actual value obtained from a measuring instrument (not shown) provided in the rolling process and the model recalculation predicted value obtained by the same prediction model as the set calculation based on the actual value. The stratified learning coefficient this time value is calculated using. Specifically, the stratified learning coefficient current value calculation unit 3 calculates the stratified learning coefficient current value using the following equations (1) and (2).

In the rolling process, the layered learning coefficient update value calculation unit 4 stores in the same layer (m, n) in the learning table storage unit 2 when the current layering of the steel type is m and the layering of the plate thickness is n. The stratified learning coefficient update value is calculated based on the stratified learning coefficient previous value and the stratified learning coefficient current value calculated by Eq. (1). Specifically, the stratified learning coefficient update value calculation unit 4 calculates the stratified learning coefficient update value using the following equation (3).

As an abnormality determination unit, the stratified learning coefficient difference determination unit 5 determines an abnormality in learning of the prediction model based on the comparison result between the stratified learning coefficient previous value and the stratified learning coefficient current value. Specifically, the stratified learning coefficient difference determination unit 5 determines whether or not the learning coefficient difference absolute value is within the threshold value by using the following equation (4).

When the learning coefficient difference absolute value on the left side is within the threshold value in the stratified learning coefficient updating unit 6, the stratified learning coefficient update unit 6 calculates the stratified learning previous value stored in the learning table storage unit 2 in the equation (3). Update to the stratified learning coefficient update value. The stratified learning coefficient updating unit 6 does not update the stratified learning coefficient stored in the learning table storage unit 2 when the learning coefficient difference absolute value on the left side exceeds the threshold value in the equation (4).

Next, the outline of the operation of the learning control device 1 will be described with reference to FIG.
FIG. 2 is a flowchart for explaining an outline of the operation of the learning control device for the rolling process according to the first embodiment.

In step S1, the learning control device 1 calculates the stratified learning coefficient this time value. After that, the learning control device performs the operation of step S2. In step S2, the learning control device 1 calculates the stratified learning coefficient update value.

After that, the learning control device 1 performs the operation of step S3. In step S3, the learning control device 1 determines the stratified learning coefficient. Specifically, the learning control device 1 determines whether or not the absolute value of the difference between the stratified learning coefficient current value and the stratified learning coefficient previous value is within the threshold value. For example, the threshold is set to 0.2.

When the absolute value of the difference between the current value of the stratified learning coefficient and the previous value of the stratified learning coefficient is within the threshold value in step S3, the learning control device 1 performs the operation of step S4. In step S4, the learning control device 1 updates the stratified learning coefficient. After that, the learning control device 1 ends the operation.

If the absolute value of the difference between the current value and the previous value of the stratified learning coefficient is not within the threshold value in step S3, the learning control device 1 ends the operation without updating the stratified learning coefficient.

According to the first embodiment described above, the stratified learning coefficient difference determination unit 5 determines an abnormality in the prediction model using the equation (4). Therefore, if some abnormality occurs in the rolled state and the change in the stratified learning coefficient is large, the stratified learning coefficient is not updated. As a result, it is possible to prevent the stratified learning coefficient from becoming unstable. Therefore, it is less likely to use the stratified learning coefficient updated by the abnormal rolling state. As a result, it is possible to maintain a stable rolling state by stable setting calculation and improve the quality of the product.

It should be noted that it may be determined whether or not an abnormality has occurred in the prediction model based on the ratio of the learning coefficient previous value and the learning coefficient current value. In this case as well, it is possible to maintain a stable rolling state by stable setting calculation and improve the quality of the product.

Next, an example of the learning control device 1 will be described with reference to FIG.
FIG. 3 is a hardware configuration diagram of the learning control device for the rolling process according to the first embodiment.

Each function of the learning control device can be realized by a processing circuit. For example, the processing circuit includes at least one processor 100a and at least one memory 100b. For example, the processing circuit comprises at least one dedicated hardware 200.

When the processing circuit includes at least one processor 100a and at least one memory 100b, each function of the learning control device is realized by software, firmware, or a combination of software and firmware. At least one of the software and firmware is written as a program. At least one of the software and firmware is stored in at least one memory 100b. At least one processor 100a realizes each function of the learning control device by reading and executing a program stored in at least one memory 100b. At least one processor 100a is also referred to as a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, and a DSP. For example, at least one memory 100b is a non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, EEPROM, magnetic disk, flexible disk, optical disk, compact disk, mini disk, DVD or the like.

If the processing circuit comprises at least one dedicated hardware 200, the processing circuit may be implemented, for example, as a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof. To. For example, each function of the learning control device is realized by a processing circuit. For example, each function of the learning control device is collectively realized by a processing circuit.

For each function of the learning control device, a part may be realized by the dedicated hardware 200, and the other part may be realized by software or firmware. For example, the function of the stratified learning coefficient update unit 6 is realized by a processing circuit as dedicated hardware 200, and at least one processor 100a has at least one memory for functions other than the function of the stratified learning coefficient update unit 6. It may be realized by reading and executing the program stored in 100b.

In this way, the processing circuit realizes each function of the learning control device by hardware 200, software, firmware, or a combination thereof.

Embodiment 2.
FIG. 4 is a block diagram of a rolling process to which the learning control device for the rolling process according to the second embodiment is applied. The same or corresponding parts as those of the first embodiment are designated by the same reference numerals. The explanation of the relevant part is omitted.

When the number of times of updating the stratified learning coefficient is small, the stratified learning coefficient previous value stored in the learning table storage unit 2 is unlikely to change significantly from the initial value depending on the stratified learning coefficient update gain. On the other hand, when the parameters of the prediction model are not adjusted sufficiently and the error of the prediction model is large, the learning coefficient this time value becomes large or small, and the learning coefficient difference absolute value is larger than the threshold in the stratified learning coefficient difference determination unit 5. If it is determined that is also large, the update of the stratified learning coefficient does not proceed.

Further, the rolling state immediately after the roll replacement in the rolling process is different from the rolling state after the number of rolls is increased due to the roll replacement. In this case, the stratified learning update may not proceed at the same threshold value.

Therefore, the learning control device 1 of the second embodiment switches the threshold value based on the number of updates of the stratified learning coefficient or the number of rolling rolls from the roll exchange of the rolling mill. Specifically, the learning control device 1 of the second embodiment includes a layered learning coefficient update number storage unit 7, a rolled number storage unit 8, and a determination threshold value switching unit 9.

The stratified learning coefficient update number storage unit 7 stores information on the number of times the stratified learning coefficient is updated. The rolling number storage unit 8 stores information on the number of rolling rolls after the rolling of the rolling mill is replaced, based on the operation data of the rolling process. The determination threshold value switching unit 9 switches the threshold value based on the number of updates of the stratified learning coefficient or the number of rolls after the rolling of the rolling mill is replaced by using the following equation (5).

The determination threshold value switching unit 9 has a flag for selecting one of the threshold values in the case of unlearned stratification classification and immediately after the roll replacement of the rolling mill. For example, the flag is input based on an external operation.

Next, the outline of the first operation of the learning control device 1 will be described with reference to FIG.
FIG. 5 is a flowchart for explaining the outline of the first operation of the learning control device for the rolling process according to the second embodiment.

In step S11, information on the number of updates of the stratified learning coefficient is stored. In the unlearned stratified division, the number of updates of the stratified learning coefficient is 0. After that, the learning control device 1 performs the operation of step S12. In step S12, the learning control device 1 determines the number of updates of the stratified learning coefficient. Specifically, the learning control device 1 determines whether or not the number of updates of the stratified learning coefficient is larger than the preset number of times. For example, the learning control device 1 determines whether or not the number of updates of the stratified learning coefficient is more than five.

If the number of updates is not greater than the preset number of times in step S12, the learning control device 1 performs the operation of step S13. In step S13, the learning control device 1 sets the threshold value to 0.3.

When the number of updates is larger than the preset number of times in step S12, the learning control device 1 performs the operation of step S14. In step S14, the learning control device 1 sets the threshold value to 0.2.

After step S13 or step S14, the learning control device 1 performs the operation of step S15. In step S15, the learning control device 1 calculates the stratified learning coefficient this time value. After that, the learning control device 1 performs the operation of step S16. In step S16, the learning control device 1 calculates the stratified learning coefficient update value.

After that, the learning control device performs the operation of step S17. In step S17, the learning control device 1 determines the stratified learning coefficient. Specifically, the learning control device 1 determines whether or not the absolute value of the difference between the stratified learning coefficient current value and the stratified learning coefficient previous value is within the threshold value.

When the absolute value of the difference between the stratified learning coefficient current value and the stratified learning coefficient previous value is within the threshold value in step S17, the learning control device 1 performs the operation of step S18. In step S18, the learning control device 1 updates the stratified learning coefficient. After that, the learning control device 1 ends the operation.

If the absolute value of the difference between the current value and the previous value of the stratified learning coefficient is not within the threshold value in step S17, the learning control device ends the operation without updating the stratified learning coefficient.

Next, with reference to FIG. 6, the update of the information on the number of updates of the stratified learning coefficient by the learning control device 1 will be described.
FIG. 6 is a flowchart for explaining the update of the information of the update number of the stratified learning coefficient by the learning control device of the rolling process in the second embodiment.

In step S19, the learning control device 1 determines whether or not the stratified learning coefficient has been updated in the flow of FIG.

If the stratified learning coefficient is not updated in step S19, the learning control device 1 ends the operation without updating the information on the number of updates of the stratified learning coefficient.

When the stratified learning coefficient is updated in step S19, the learning control device 1 performs the operation of step S20. In step S20, the learning control device 1 updates the information on the number of updates of the stratified learning coefficient. After that, the learning control device 1 ends the operation.

The information on the number of updates of the stratified learning coefficient obtained in the flow of FIG. 6 is used for the next learning.

Next, the outline of the second operation of the learning control device will be described with reference to FIG. 7.
FIG. 7 is a flowchart for explaining the outline of the second operation of the learning control device for the rolling process according to the second embodiment.

In step S21, the learning control device 1 stores information on the number of rolled rolls after the rolls are exchanged. After that, the learning control device 1 performs the operation of step S22. In step S22, the learning control device 1 determines whether or not the number of rolled rolls after the roll exchange is larger than the preset number of rolls. For example, the learning control device 1 determines whether or not the number of rolled rolls after the roll replacement is more than 10.

When the number of rolled rolls after the roll replacement is not larger than the preset number in step S22, the learning control device 1 performs the operation of step S23. In step S23, the learning control device 1 sets the threshold value to 0.25.

When the number of rolled rolls after the roll replacement in step S22 is larger than the preset number of rolls, the learning control device 1 performs the operation of step S24. In step S24, the learning control device 1 sets the threshold value to 0.2.

After step S23 or step S24, the learning control device 1 performs the operation of step S25. In step S25, the learning control device 1 calculates the stratified learning coefficient this time value. After that, the learning control device 1 performs the operation of step S26. In step S26, the learning control device 1 calculates the stratified learning coefficient update value.

After that, the learning control device performs the operation of step S27. In step S27, the learning control device 1 determines the stratified learning coefficient. Specifically, the learning control device 1 determines whether or not the absolute value of the difference between the stratified learning coefficient current value and the stratified learning coefficient previous value is within the threshold value.

When the absolute value of the difference between the current value of the stratified learning coefficient and the previous value of the stratified learning coefficient is within the threshold value in step S27, the learning control device 1 performs the operation of step S28. In step S28, the learning control device 1 updates the stratified learning coefficient. After that, the learning control device 1 ends the operation.

If the absolute value of the difference between the current value and the previous value of the stratified learning coefficient is not within the threshold value in step S27, the learning control device ends the operation without updating the stratified learning coefficient.

According to the second embodiment described above, the determination threshold value switching unit 9 switches the threshold value used for determining the abnormality in the prediction model based on the information stored by the stratified learning coefficient update frequency storage unit 7. Therefore, it is possible to prevent the stratified learning coefficient from not being updated when the number of times the stratified learning coefficient is updated is small.

Further, the determination threshold value switching unit 9 switches the threshold value used for determining the abnormality in the prediction model based on the information recorded by the rolling number storage unit 8. Therefore, it is possible to avoid that the update of the stratified learning coefficient does not proceed due to the change in the rolling state due to the replacement of the roll of the rolling mill.

Embodiment 3.
FIG. 8 is a block diagram of a rolling process to which the learning control device for the rolling process according to the third embodiment of the invention is applied. The same or corresponding parts as those of the first embodiment are designated by the same reference numerals. The explanation of the relevant part is omitted.

The inter-material learning coefficient is different from the stratified learning coefficient, and is a learning coefficient corresponding to changes between materials. The inter-material learning factor is applied to the next rolled material.

Therefore, the learning control device 1 of the third embodiment updates the inter-material learning coefficient even when the stratified learning coefficient is not updated. Specifically, the learning control device 1 of the third embodiment has the inter-material learning coefficient current value calculation unit 10, the inter-material learning coefficient update gain switching unit 11, the inter-material learning coefficient update value calculation unit 12, and the inter-material learning coefficient. The update unit 13 and the inter-material learning coefficient storage unit 14 are provided.

Inter-material learning coefficient This time value calculation unit 10 calculates the inter-material learning coefficient using the following equation (6).

When the learning coefficient difference absolute value exceeds the threshold value in the stratified learning coefficient difference determination unit 5, the inter-material learning coefficient update gain switching unit 11 switches the learning coefficient update gain using the following equation (7).

That is, the inter-material learning coefficient update gain switching unit 11 increases the learning coefficient update gain when the learning coefficient difference absolute value exceeds the threshold value.

Using the learning coefficient update gain determined by the inter-material learning coefficient update gain switching unit 11, the inter-material learning coefficient update value calculation unit 12 calculates the inter-material learning coefficient update value using the following equation (9). To do.

The inter-material learning coefficient update unit 13 updates the inter-material learning coefficient previous value to the inter-material learning coefficient update value calculated by Eq. (9).

The inter-material learning coefficient storage unit 14 erases the information of the previous inter-material learning value and stores the information of the inter-material learning coefficient update value. The updated inter-material learning coefficient is used in the inter-material setting calculation.

Next, the outline of the operation of the learning control device will be described with reference to FIG.
FIG. 9 is a flowchart for explaining an outline of the operation of the learning control device for the rolling process according to the third embodiment.

In step S31, the learning control device 1 calculates the stratified learning coefficient this time value. After that, the learning control device performs the operation of step S32. In step S32, the learning control device 1 calculates the stratified learning coefficient update value.

After that, the learning control device performs the operation of step S33. In step S33, the learning control device 1 calculates the inter-material learning coefficient this time value. After that, the learning control device 1 performs the operation of step S34. In step S34, the learning control device 1 determines the stratified learning coefficient. Specifically, the learning control device 1 determines whether or not the absolute value of the difference between the stratified learning coefficient current value and the stratified learning coefficient previous value is within the threshold value.

When the absolute value of the difference between the current value of the stratified learning coefficient and the previous value of the stratified learning coefficient is within the threshold value in step S34, the learning control device 1 performs the operation of step S35. In step S35, the learning control device 1 updates the stratified learning coefficient. After that, the learning control device 1 performs the operation of step S36. In step S36, the learning control device 1 calculates the inter-material learning coefficient update value with a relatively small learning coefficient update gain. After that, the learning control device 1 ends the operation.

When the absolute value of the difference between the current value of the stratified learning coefficient and the previous value of the stratified learning coefficient is not within the threshold value in step S34, the learning control device 1 performs the operation of step S37. In step S37, the learning control device 1 calculates the inter-material learning coefficient update value with a relatively large learning coefficient update gain. After that, the learning control device 1 ends the operation.

According to the third embodiment described above, the update gain of the inter-material learning coefficient is switched based on the determination result of the abnormality in the prediction model. For example, when the stratified learning coefficient update unit 6 does not update the stratified learning coefficient, the inter-material learning coefficient is updated with a relatively large learning coefficient update gain. Therefore, it is possible to reduce the error of the prediction model with respect to the abnormal actual value that occurs continuously. As a result, the rolled state can be stabilized.

In the rolling process, the learning control device 1 of the first to third embodiments may be applied to a prediction model other than the rolling load.

As described above, the rolling process learning control device according to the present invention can be used for learning the prediction model of the rolling process.

1 Learning control device, 2 Learning table storage unit, 3 Layer-based learning coefficient update unit, 4 Layer-based learning coefficient update value calculation unit, 5 Layer-based learning coefficient difference judgment unit, 6-layer learning coefficient update unit, 7-layer Separate learning coefficient update frequency storage unit, 8 rolling number storage unit, 9 judgment threshold switching unit, 10 inter-material learning coefficient current value calculation unit, 11 inter-material learning coefficient update gain switching unit, 12 inter-material learning coefficient update value calculation unit, 13 Inter-material learning coefficient update unit, 14 Inter-material learning coefficient storage unit, 100a processor, 100b memory, 200 hardware

In the rolling process learning control device according to the present invention, the learning coefficient for correcting the error between the actual value of the rolling process and the model predicted value obtained by the prediction model is stratified by a plurality of conditions of the rolling process. For the learning table managed as the stratified learning coefficient, the stratified learning coefficient for calculating the current value based on the current actual value of the rolling process and the current prediction model, the stratified learning coefficient present value calculation unit, and the layer. Separate learning coefficient Layered learning coefficient calculated by the current value calculation unit Layered learning coefficient based on the current value and the layered learning coefficient previously stored in the same layer as the current value in the learning table. The layered learning coefficient update value calculation unit that calculates the coefficient update value, the layered learning coefficient this time value calculated by the layered learning coefficient current value calculation unit, and the same layer as the layered learning coefficient current value in the learning table. If the absolute value of the difference from the previously stored stratified learning coefficient exceeds the threshold set to a value that does not exclude the necessary update of the stratified learning coefficient, there is an abnormality in the learning of the prediction model. It is provided with an abnormality determination unit for determining, and a stratified learning coefficient updating unit that does not update the previous value of the stratified learning coefficient when the abnormality determining unit determines that there is an abnormality in learning of the prediction model.

Claims (5)

For a learning table that manages the learning coefficient that corrects the error between the actual value of the rolling process and the model predicted value obtained by the prediction model as a stratified learning coefficient for each layer, the current actual value of the rolling process and the current prediction Stratified learning coefficient based on the model Layered learning coefficient that calculates the current value With the current value calculation unit,
Layered learning coefficient Layered based on the layered learning coefficient calculated by the current value calculation unit and the layered learning coefficient previous value stored in the same layer as the layered learning coefficient in the learning table. The layered learning coefficient update value calculation unit that calculates the separate learning coefficient update value,
The layered learning coefficient The layered learning coefficient calculated by the current value calculation unit is compared with the previous value of the layered learning coefficient stored in the same layer as the current value in the learning table. Based on the anomaly determination unit that determines anomalies in the learning of the prediction model,
A stratified learning coefficient update unit that does not update the previous value of the stratified learning coefficient when the abnormality determination unit determines that there is an abnormality in the learning of the prediction model.
A learning control device for the rolling process. The abnormality determination unit is the stratified learning coefficient current value calculated by the stratified learning coefficient current value calculation unit and the stratified learning coefficient previously stored in the same layer as the stratified learning coefficient current value in the learning table. The learning control device for a rolling process according to claim 1, wherein when the absolute value of the difference from the value exceeds the threshold value, it is determined that there is an abnormality in the learning of the prediction model. A stratified learning coefficient update count storage unit that stores information on the update count of the stratified learning coefficient of the learning table by the stratified learning coefficient update unit, and a stratified learning coefficient update count storage unit.
A determination threshold value switching unit that switches the threshold value used for determination of the abnormality determination unit based on the information stored by the layered learning coefficient update frequency storage unit.
2. The learning control device for a rolling process according to claim 2. A rolling number storage unit that stores information on the number of rolled materials after the rolls are exchanged in the rolling process, and a rolling number storage unit.
A determination threshold value switching unit that switches the threshold value used for determining the abnormality determination unit based on the information stored by the rolling number storage unit.
2. The learning control device for a rolling process according to claim 2 or 3. Inter-material learning coefficient for calculating the inter-material learning coefficient based on the error between the actual value of the rolling process and the model predicted value for the inter-material learning coefficient applied to the next rolled material.
An inter-material learning coefficient update gain switching unit that switches the update gain when calculating the inter-material learning coefficient update value based on the determination result of the abnormality determination unit, and
Inter-material learning coefficient Inter-material learning coefficient calculated by the current value calculation unit Material-to-material learning coefficient based on the current value, the inter-material learning coefficient previous value, and the update gain switched by the renewal gain switching unit. Inter-material learning coefficient update value calculation unit that calculates the update value,
Inter-material learning coefficient update unit that updates the previous value to the inter-material learning coefficient update value calculated by the inter-material learning coefficient update value calculation unit, and the inter-material learning coefficient update unit.
The rolling process learning control device according to any one of claims 2 to 4.

Images