KR-102962696-B1 - ROLLING LOAD PREDICITION SYSTEM FOR CONTROLLING ROLLING MILL AND ROLLING LOAD PREDICTION METHOD USING THE SAME

Abstract

A rolling load prediction system for controlling a rolling mill according to one aspect of the present invention comprises: an operation data collection unit that collects operation data from a rolling mill installed in a rolling process and at least one sensor; a correction coefficient prediction model unit that predicts at least one correction coefficient at the time when a strip is rolled using a correction coefficient prediction model learned based on the operation data; and a rolling load calculation unit that inputs the predicted at least one correction coefficient into a rolling load formula model to calculate a predicted rolling load value at the time when a strip is rolled.

Inventors

• 최영식
• 김광식

Assignees

• 주식회사 포스코디엑스

Dates

Publication Date

20260507

Application Date

20230629

Claims (12)

1. Operation data collection unit for collecting operation data from a rolling mill installed in a rolling process and at least one sensor; A correction factor prediction model unit that predicts at least one correction factor at the time when the strip is rolled using a correction factor prediction model learned based on the above-mentioned operation data; and A rolling load calculation unit that inputs at least one predicted correction factor into a rolling load formula model to calculate a predicted rolling load value at the time when the strip is rolled; A model update determination unit that calculates the deviation between the predicted rolling load value and the actual rolling load value collected from at least one sensor, and if the deviation exceeds a preset maximum deviation, determines that the current correction factor prediction model does not correspond to the current operating state due to a change in the process state, and determines an update to the current correction factor prediction model; and When an update is determined by the above model update determination unit, the correction coefficient prediction model learning unit retrains the above correction coefficient prediction model using recently collected operation data during a set period. A rolling load prediction system for rolling mill control, including
2. In paragraph 1, The above-mentioned operation data collection unit extracts a plurality of first input variables including at least one of physical property information, process control information, process environment information, and process measurement information regarding a strip based on the collected operation data, and provides the extracted plurality of first input variables to the correction coefficient prediction model unit. A rolling load prediction system for controlling a rolling mill, wherein the above correction coefficient prediction model unit inputs the plurality of first input variables into the correction coefficient prediction model to predict at least one correction coefficient.
3. In paragraph 1, The above-mentioned operation data collection unit extracts a plurality of second input variables including the radius of a roll equipped in the rolling mill, the thickness of the strip's entry side, the thickness of the strip's exit side, and a tension correction coefficient based on the collected operation data, and provides the extracted plurality of second input variables to the above-mentioned rolling load calculation unit. A rolling load prediction system for controlling a rolling mill, characterized in that the above rolling load calculation unit inputs the above plurality of second input variables and at least one correction coefficient predicted by the above correction coefficient prediction model unit into a rolling load formula model to calculate a predicted rolling load value at the time when the strip is rolled.
4. In paragraph 1, A rolling load prediction system for rolling mill control, characterized in that at least one correction factor includes a friction coefficient of the strip, a deformation resistance of the strip, and a variation value according to a change in process state.
5. delete
6. delete
7. delete
8. Step to check whether the correction factor prediction model has been updated; If there is an update to the above correction factor prediction model, a step of changing to a new correction factor prediction model and synchronizing the above correction factor prediction model with the state according to the process change; A step of extracting a plurality of first input variables used in the correction coefficient prediction model and a plurality of second input variables used in the rolling load formula model based on operation data collected from a rolling mill installed in the rolling process and at least one sensor; A step of inputting the above plurality of first input variables into a correction factor prediction model to predict at least one correction factor at the time when the strip is rolled; A step of calculating a predicted rolling load value by inputting the plurality of second input variables and the predicted at least one correction coefficient into the rolling load formula model; A step of calculating the deviation between the predicted rolling load value and the actual rolling load value collected from at least one sensor; If the above deviation exceeds a preset maximum deviation, determining that the current correction factor prediction model does not match the current operating state due to a change in process state, and determining an update to the current correction factor prediction model; and When an update is decided, the step of retraining the correction factor prediction model using recently collected operation data from the set period. A rolling load prediction method for rolling mill control, including
9. In paragraph 8, The above plurality of first input variables include at least one of physical property information regarding the strip, process control information, process environment information, and process measurement information, and A method for predicting rolling load for rolling mill control, characterized in that the plurality of second input variables include the radius of a roll provided in the rolling mill, the thickness of the strip at the entry side, the thickness of the strip at the exit side, and a tension correction coefficient.
10. In paragraph 8, A method for predicting rolling load for rolling mill control, characterized in that at least one correction factor includes a friction coefficient of the strip, a deformation resistance of the strip, and a variation value according to a change in process state.
11. delete
12. delete

Description

Rolling Load Prediction System for Controlling Rolling Mill and Rolling Load Prediction Method Using the Same The present invention relates to a rolling load prediction system and a rolling load prediction method, and more specifically, to a rolling load prediction system and a rolling load prediction method for controlling a rolling mill. The steelmaking process consists of multiple processes, such as the ironmaking process, steelmaking process, continuous casting process, and rolling process. The ironmaking process is the process of producing molten iron (pig iron), the steelmaking process is the process of removing impurities from the molten iron, and the continuous casting process is the process of turning liquid iron into a solid. The rolling process is the process of making iron into strips, steel plates, or wire rods, and produces strips by passing slabs, blooms, or billets produced in the continuous casting process through rolls to stretch or thin them. In the rolling process, the rolling load at the point of strip rolling is a critical factor in determining the thickness quality of the strip when controlling the rolling mill. Generally, the rolling load is predicted using values derived from experiments, and the predicted rolling load can be utilized as a key factor in setting the operating conditions of the rolling mill or controlling the strip thickness. However, conventional methods for predicting rolling loads rely on experimentally derived values that fail to account for changes in steel grade, fluctuations in rolling conditions, seasonal factors, and the aging of process equipment; consequently, long-term rolling load predictions may become inaccurate. This can have an adverse effect on product quality. FIG. 1 is a schematic diagram showing a rolling load prediction system according to one embodiment of the present invention. Figure 2 is a block diagram schematically showing the configuration of the rolling load prediction section of Figure 1. Figure 3 is a block diagram schematically showing the configuration of the correction coefficient prediction model learning unit of Figure 1. FIG. 4 is a flowchart illustrating a rolling load prediction method according to one embodiment of the present invention. FIG. 5 is a flowchart illustrating a method for learning a correction factor prediction model according to an embodiment of the present invention. Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. The meaning of the terms described in this specification should be understood as follows. A singular expression should be understood to include a plural expression unless the context clearly defines otherwise, and terms such as "first," "second," etc., are intended to distinguish one component from another, and the scope of rights shall not be limited by these terms. Terms such as "include" or "have" should be understood as not excluding in advance the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. The term “at least one” should be understood to include all combinations that can be presented from one or more related items. For example, the meaning of “at least one of the first item, the second item, and the third item” is not only the first item, the second item, or the third item individually, but also all combinations of items that can be presented from two or more of the first item, the second item, and the third item. Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. Hereinafter, embodiments according to the present invention will be described in detail with reference to the attached drawings. FIG. 1 is a schematic diagram showing a rolling load prediction system according to one embodiment of the present invention. Referring to FIG. 1, a rolling load prediction system (100) according to one embodiment of the present invention includes a rolling mill (110), an operation performance measurement unit (120), a rolling load prediction unit (130), a control unit (145), a correction coefficient prediction model learning unit (140), and an operation information database (150). A rolling mill (110) is installed in a rolling process and includes a plurality of rolls (112). The rolling mill (110) can pass a strip made of iron between the plurality of rolls (112) to make it thin to a desired thickness. At least one sensor (115) may be provided in the rolling mill (110). At least one sensor (115) may be installed in the rolling mill (110) or around the rolling mill (110) to sense physical quantities or changes. For example, at least one sensor (115) may be installed on a roll (112) to sense the rotational speed of the roll (112). As another example, at least one sensor (115) may be installed at the front end of the roll (112) to sense the inlet thickness of a strip entering between two rolls (112). A