Roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of strip

Abstract

The present invention belongs to the technical field of metal pressure processing and relates to a roll gap self-learning compensation method for reducing the influence of the biting impact on the thickness of the head of a strip. The method is as follows: Before each biting, the set value of the roll gap is subtracted by the biting impact compensation value to obtain the roll gap target value; after the rolling mill bites the steel, when the biting impact compensation time is reached, the biting impact compensation value is cancelled to complete this compensation; after the rolling mill bites the steel, the roll gap feedback data is analyzed to obtain the maximum biting impact value and the maximum biting impact moment of the currently rolled steel plate under the corresponding biting impact compensation value; after the biting impact is completed, a self-learning algorithm is used to perform self-learning on the biting impact compensation value and the biting impact compensation time to obtain the corrected biting impact compensation value and the corrected biting impact compensation time, which are used for the biting impact compensation of the next steel plate. The compensation method provided by the present invention can effectively reduce the influence of the biting impact on the thickness of the head of the steel plate and improve the dimensional accuracy of the steel plate.

Description

Technical Field

[0001] The invention belongs to the technical field of metal pressure processing and relates to a roller gap self-learning compensation method for reducing the influence of steel biting impact on the thickness of a plate strip head. Background Art

[0002] During the rolling process, due to the influence of the length of the rolled piece and the rolling specifications, the rolling mill will enter the no-load operation state during the rolling interval. When the rolled piece is just bitten by the roll, it will produce a large impact, causing certain fluctuations in the rolling force and the roll gap. Without applying roll gap compensation, the rolling mill will go from no-load to loaded operation at the moment of biting the steel, and the roll gap will be "pushed open" by the rolled piece, causing the actual roll gap value to be greater than the roll gap value calculated by the process control model, thereby increasing the thickness of the rolled piece at the exit. In actual engineering, a certain amount of compensation will be applied to the roll gap before the rolling mill bites the steel, so as to reduce the impact of the roll gap change on the rolling during the steel biting impact and reduce the fluctuation of the head thickness.

[0003] With the continuous improvement of the rolling rhythm, the rolling speed is constantly increasing, the rolling interval time is shorter, and there are higher requirements for the thickness control of the rolled piece. Adding the steel bite impact roll gap compensation item can reduce the thickness deviation of the rolled piece at the head position, reduce the length of the head shear, and thus achieve a higher yield rate. The existing steel bite impact roll gap compensation value calculation method does not use a self-learning algorithm, and the compensation values are mostly fixed. In the actual rolling process, the data of the rolled piece are all different. If the roll gap compensation value calculated by the data that changes with the actual rolling situation is not used, although there is a certain compensation effect, the compensation effect is not ideal. Summary of the invention

[0004] In view of the above technical problems, the present invention provides a roll gap self-learning compensation method for reducing the influence of steel bite impact on the thickness of the plate strip head. By calculating the roll gap compensation value of the rolling mill steel bite impact, the roll gap compensation value to be set can be calculated in real time and quickly. And due to the addition of the self-learning algorithm, as the rolling continues, the steel bite impact compensation can continuously obtain a better compensation effect.

[0005] The present invention is achieved through the following technical solutions:

[0006] A roll gap self-learning compensation method for reducing the influence of steel biting impact on the thickness of the head of the plate strip, analyzing the data in the rolling process, and compensating for the steel biting impact by continuously learning the roll gap compensation value and compensation time, thereby reducing the change in the thickness of the steel plate head when biting steel.

[0007] Further, the method comprises:

[0008] (1) Before each bite, set the roll gap to S ref Subtract the steel bite impact compensation value S comp As the roll gap target value Sset ;

[0009] (2) After the rolling mill bites the steel, determine whether the rolling time reaches the bite impact compensation time t comp . When the bite impact compensation time t comp is reached, cancel the bite impact compensation value S comp ;

[0010] (3) Starting from when the rolling mill bites the steel, collect roll gap data in real time. After the compensation is completed, analyze the roll gap feedback data to obtain the maximum bite impact value S imp of the current rolled steel plate under the corresponding bite impact compensation value and the roll gap stabilization time t can ;

[0011] (4) After the bite impact is completed, use the self-learning algorithm and adopt the maximum bite impact value S imp and the roll gap stabilization time t can to perform self-learning on the bite impact compensation value S comp and the bite impact compensation time t comp to obtain the corrected bite impact compensation value and the corrected bite impact compensation time for the bite impact compensation of the next steel plate.

[0012] Further, in step (1), before the rolling mill bites the steel, set the roll gap. Subtract the bite impact compensation value S ref from the roll gap setting value S comp to obtain the roll gap target value S set , that is, S set = S ref - S comp ;

[0013] In step (2), after the rolling mill bites the steel, when it is determined according to the bite signal moment that the rolling time reaches the bite impact compensation time t comp , remove the bite impact compensation value S comp , quickly restore the roll gap setting value S ref , and complete this compensation. At this time, the roll gap recovery speed should be set according to the maximum speed allowed by the hydraulic cylinder equipment during rolling.

[0014] Further, for the first rolling, the initial value of the bite impact compensation value S comp is selected as 100 μm, and the initial value of the bite impact compensation time t comp is selected as 30 ms;

[0015] For the steel plates rolled subsequently, adopt the corrected bite impact compensation value and the corrected bite impact compensation time obtained by the self-learning of the impact compensation process of the previous steel plate.

[0016] Further, in step (3), the roll gap feedback data is analyzed to obtain the actual roll gap value S within the time from the start moment of steel biting to the maximum steel biting impact compensation time act and the target roll gap value S set The steel biting impact value at the time when the subtraction value is the largest is obtained, and the maximum steel biting impact value S imp is obtained; the maximum steel biting impact value S imp is the difference value when the subtraction value between the actual roll gap value S act and the target roll gap value S set is the largest;

[0017] After reaching the maximum steel biting impact value S imp , the moment when the actual roll gap value S act first reaches a difference less than or equal to 10 μm from the set roll gap value S ref is the maximum steel biting fallback moment; taking the steel biting signal as the starting moment and the maximum steel biting fallback moment as the ending moment, the roll gap stabilization time t can is calculated.

[0018] Here, it is limited that the actual roll gap value S act first reaches a difference less than or equal to 10 μm from the set roll gap value S ref because in the actual feedback of the controller, the feedback data is discrete and randomly fluctuating, and it is impossible to obtain a definite feedback of "equal to" a certain fixed value.

[0019] Further, the maximum steel biting impact compensation time is the maximum boundary value t max of the steel biting impact compensation time, and the maximum boundary value t max is determined according to the historical steel biting impact situation of this rolling mill, and the value range is 120 - 180 ms.

[0020] Further, in step (4), after the steel biting impact is completed, the self - learning algorithm is used to perform self - learning on the steel biting impact compensation value and the steel biting impact compensation time. Specifically:

[0021] S comp (k + 1)=(1 - α)S comp (k)+α*(S imp -S mar );

[0022] t comp (k + 1)=(1 - β)t comp (k)+β*t can ;

[0023] In the formula, k represents the current rolling times, and k + 1 represents the next rolling times; S comp (k + 1) is the corrected steel biting impact compensation value obtained through self - learning; tcomp (k + 1) is the corrected compensation time for the steel biting impact obtained through self - learning; S comp (k) is the compensation value for the steel biting impact adopted for the currently rolled steel plate; t comp (k) is the compensation time for the steel biting impact adopted for the currently rolled steel plate; S imp is the maximum steel biting impact value, t can is the roll gap stabilization time; S mar is the allowable overshoot after compensation, which is a constant and generally can take values from 10 to 50μm; both α and β are coefficients, and α, β ∈ (0, 0.3).

[0024] Furthermore, the maximum value of the compensation value for the steel biting impact is limited to 350μm. Limiting the maximum value of the compensation value for the steel biting impact to 350μm is to prevent the compensation value from being too large in case of instrument failure. If the compensation value for the steel biting impact is greater than 350μm, 350μm is used for compensation.

[0025] Furthermore, the calculation process of the self - learning algorithm is completed in the PLC controller of the first - level control system.

[0026] Advantageous technical effects of the present invention:

[0027] The calculation method of the compensation value for the roll gap of the steel biting impact proposed by the method of the present invention can calculate the compensation value for the steel biting impact that conforms to the actual rolling situation simply and quickly, occupies less PLC computing power and communication data volume, and can reduce the problem of excessive thickness fluctuation at the head of the steel plate caused by inaccurate setting of the fixed compensation value.

[0028] Since the method provided by the present invention adopts a calculation method matching the rolling data and adds a self - learning algorithm, the roll gap compensation value (compensation value for the steel biting impact) and the compensation time for the steel biting impact can be calculated online in real time according to the rolling situation; it can be realized by adding a corresponding roll gap compensation algorithm program in the PLC controller of the first - level (basic automation level) without additional new equipment, and the function of compensating the roll gap of the steel biting impact can be completed without affecting normal rolling. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is the flowchart of the roll gap self - learning compensation method for reducing the influence of the steel biting impact of the rolling mill in the embodiment of the present invention;

[0030] Figure 2 is the compensation effect diagram of the steel biting impact in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, elaborate in detail on the specific implementation manners, structures, features, and effects of the present invention as follows.

[0032] The method for calculating the compensation value of the biting impact roll gap provided by the present invention can combine relevant data in the process control computer and the PLC controller, and use a self-learning algorithm to calculate the compensation value; the roll gap compensation value can be obtained quickly and relatively accurately during the actual rolling process. The following will illustrate with specific algorithms and diagrams.

[0033] The main calculation process of the present invention is mainly completed in the first-level control system (PLC controller). After calculating the roll gap compensation value, subtract the compensation value from the roll gap set value (the roll gap calculated value of the thickness model) as the final roll gap value, and the controller calculates the parameters to actuate the hydraulic cylinder to the specified position to complete the roll gap setting. After biting, analyze the rolling process data, and perform self-learning on the compensation value and compensation time. When the rolling time reaches the impact time, remove the impact compensation term to complete the biting impact compensation.

[0034] A roll gap self-learning compensation method for reducing the influence of the biting impact of the rolling mill on the thickness of the head of the strip steel analyzes the data during the rolling process, and compensates for the biting impact by continuously learning the roll gap compensation value and compensation time, thereby reducing the change in the thickness of the head of the steel plate during biting. The method includes:

[0035] (1) Before each biting, subtract the biting impact compensation value S ref from the roll gap set value S comp as the roll gap target value S set ;

[0036] (2) After the rolling mill bites, determine whether the rolling time reaches the biting impact compensation time t comp . When the biting impact compensation time t comp is reached, cancel the biting impact compensation value S comp ;

[0037] (3) Starting from the moment the rolling mill bites, collect roll gap data in real time. After the compensation is completed, analyze the roll gap feedback data to obtain the maximum biting impact value S imp of the current rolled steel plate under the corresponding biting impact compensation value and the roll gap stabilization time t can ;

[0038] (4) After the biting impact is completed, use the self-learning algorithm, and adopt the maximum biting impact value S imp and the roll gap stabilization time t canSelf - learn the steel - biting impact compensation value and the steel - biting impact compensation time to obtain the corrected steel - biting impact compensation value and the corrected steel - biting impact compensation time for the steel - biting impact compensation of the next steel plate.

[0039] In step (1) of this embodiment, before the rolling mill bites the steel, the roll gap is set, and the roll gap setting value S is used. ref Subtract the steel - biting impact compensation value S comp to obtain the roll - gap target value S set , that is, S set = S ref - S comp ;

[0040] In step (2), after the rolling mill bites the steel, when it is determined that the rolling time reaches the steel - biting impact compensation time t comp according to the steel - biting signal time, remove the steel - biting impact compensation value S comp , and quickly restore the roll - gap setting value S ref , completing this compensation. At this time, the roll - gap recovery speed should be set according to the maximum speed allowed by the hydraulic cylinder equipment during rolling, and 2mm / s can be selected here.

[0041] In this embodiment, during the first rolling, the initial value of the steel - biting impact compensation value S comp is selected as 100μm, and the initial value of the steel - biting impact compensation time t comp is selected as 30ms;

[0042] For the subsequent rolled steel plates, the corrected steel - biting impact compensation value and the corrected steel - biting impact compensation time obtained by the self - learning of the impact compensation process of the previous steel plate are adopted.

[0043] In step (3) of this embodiment, analyze the roll - gap feedback data to obtain the actual roll - gap value S act and the roll - gap target value S set during the period from the steel - biting start time to the maximum steel - biting impact compensation time, and obtain the steel - biting impact value when the subtraction value is the largest, obtaining the maximum steel - biting impact value S imp ;

[0044] After reaching the maximum steel - biting impact value S imp , the moment when the actual roll - gap value S act first reaches a difference less than or equal to 10μm from the roll - gap setting value S ref is the maximum steel - biting impact moment; taking the steel - biting signal as the starting time and the maximum steel - biting impact moment as the ending time, calculate the roll - gap stabilization time t can .

[0045] Specifically, the maximum steel - biting impact compensation time is the maximum boundary value t max of the steel - biting impact compensation time, tmax The value is 150 ms.

[0046] In step (4) of this embodiment, after the steel biting impact is completed, the self-learning algorithm is used to perform self-learning on the steel biting impact compensation value and the steel biting impact compensation time. Specifically:

[0047] S comp (k + 1) = (1 - α)S comp (k) + α * (S imp - S mar );

[0048] t comp (k + 1) = (1 - β)t comp (k) + β * t can ;

[0049] In the formula, k represents the current rolling times, and k + 1 represents the next rolling times; S comp (k + 1) is the corrected steel biting impact compensation value obtained through self-learning; t comp (k + 1) is the corrected steel biting impact compensation time obtained through self-learning; S comp (k) is the steel biting impact compensation value adopted for the currently rolled steel plate; t comp (k) is the steel biting impact compensation time adopted for the currently rolled steel plate; S imp is the maximum steel biting impact value, t can is the roll gap stabilization time; S mar is the allowable overshoot after compensation, and the value can be 40 μm; both α and β are coefficients, α = 0.2, β = 0.2.

[0050] In this embodiment, the maximum value of the steel biting impact compensation value is limited to 350 μm. If the steel biting impact compensation value is greater than 350 μm, 350 μm is used for compensation.

[0051] From Figure 2 it can be seen that for the rolled piece, the roll gap impact value after applying this method changes from S1 to S2, and S2 is much smaller than S1. Therefore, the phenomenon of excessive head thickness is effectively suppressed.

[0052] The above step explanation can elaborate on the online calculation process and method of the roll gap compensation value. In particular, the data used in the explanation and description of this method is the data of a certain specification of rolled piece, but this method is not limited to a specific rolled piece. For the rolling of other steel plates within a rolling plan, the same method can be used to calculate the required steel biting impact roll gap compensation value through online real-time calculation.

[0053] The compensation method provided by the present invention can effectively reduce the influence of the steel biting impact on the head thickness of the steel plate and improve the dimensional accuracy of the steel plate.

[0054] As described above, it is only the preferred embodiment of the present invention and does not impose any formal restrictions on the present invention. Although the present invention has been disclosed above with the preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments by using the disclosed technical content without departing from the technical solution of the present invention. However, as long as it does not depart from the technical solution content of the present invention, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of strip steel, characterized in that, Analyze the data during the rolling process, and compensate for the biting impact by continuously learning the roll gap compensation value and compensation time, so as to reduce the change in the thickness of the head of the steel plate during biting; The method includes: (1) Before each bite of steel, set the roll gap value S ref minus the bite impact compensation value S comp as the roll gap target value S set ; (2) After the rolling mill bites the steel, determine whether the rolling time has reached the bite impact compensation time t comp , when the bite impact compensation time t comp is reached, cancel the bite impact compensation value S comp ; (3) Starting from when the rolling mill bites the steel, collect roll gap data in real time. After the compensation is completed, analyze the roll gap feedback data to obtain the maximum biting impact value S of the currently rolled steel plate under the corresponding biting impact compensation value imp and the roll gap stabilization time t can ; (4) After the steel biting impact is completed, using the self-learning algorithm, the maximum steel biting impact value S imp and the roll gap stabilization time t can are used to perform self-learning on the steel biting impact compensation value S comp and the steel biting impact compensation time t comp to obtain the corrected steel biting impact compensation value and the corrected steel biting impact compensation time, which are used for the steel biting impact compensation of the next steel plate; In step (3), analyze the roll gap feedback data to obtain the actual roll gap value S during the period from the start time of steel biting to the maximum steel biting impact compensation time act and the target roll gap value S set The steel biting impact value when the subtraction result is the largest is obtained to get the maximum steel biting impact value S imp ; Reach the maximum steel biting impact value S imp After that, the actual roll gap value S act For the first time, it reaches the moment when the difference from the set roll gap value S ref is less than or equal to 10 μm, which is the moment when the maximum steel biting falls back; taking the steel biting signal as the starting moment and the moment when the maximum steel biting falls back as the ending moment, calculate the roll gap stabilization time t can .

2. The roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of strip steel according to claim 1, wherein In step (1), the roll gap is set before the rolling mill bites the steel, and the roll gap setting value S ref is subtracted from the bite impact compensation value S comp to obtain the roll gap target value S set , that is, S set = S ref - S comp ; In step (2), after the rolling mill bites the steel, when it is determined according to the moment of the steel biting signal that the rolling time reaches the steel biting impact compensation time t comp , remove the steel biting impact compensation value S comp , and quickly restore the roll gap setting value S ref , thus completing this compensation.

3. The roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of the strip steel according to claim 2, characterized in that, During the first rolling, the bite impact compensation value S comp is initially set to 100 μm, and the bite impact compensation time t comp is initially set to 30 ms; For the steel plates to be rolled subsequently, use the corrected biting impact compensation value and corrected biting impact compensation time obtained by the self-learning of the impact compensation process of the previous steel plate.

4. The roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of the strip steel according to claim 3, characterized in that, The maximum compensation time for the steel biting impact is the maximum boundary value t of the compensation time for the steel biting impact max , and the maximum boundary value t max is determined according to the historical steel biting impact situation of the rolling mill, and the value range is 120 - 180 ms.

5. The roll gap self-learning compensation method for reducing the influence of bite steel impact on the head thickness of the strip steel according to claim 3, characterized in that, In step (4), after the biting impact is completed, use the self-learning algorithm to perform self-learning on the biting impact compensation value and the biting impact compensation time. Specifically: S comp (k + 1) = (1 - α)S comp (k) + α * (S imp - S mar ); t comp (k + 1) = (1 - β)t comp (k) + β * t can ; Wherein, k represents the current rolling pass number, and k + 1 represents the next rolling pass number; S comp (k + 1) is the corrected bite impact compensation value obtained through self-learning; t comp (k + 1) is the corrected bite impact compensation time obtained through self-learning; S comp (k) is the bite impact compensation value adopted for the currently rolled steel plate; t comp (k) is the bite impact compensation time adopted for the currently rolled steel plate; S imp is the maximum bite impact value of the currently rolled steel plate under the corresponding bite impact compensation value; t can is the roll gap stabilization time of the currently rolled steel plate under the corresponding bite impact compensation value; S mar is the allowable overshoot after compensation, which is a constant with a value range of 10 - 50 μm; both α and β are coefficients, and α, β ∈ (0, 0.3).

6. The roll gap self-learning compensation method for reducing the influence of biting impact on the head thickness of strip steel according to claim 5, characterized in that The maximum value of the biting impact compensation value is limited to 350μm.

7. A roll gap self-learning compensation method for reducing the influence of steel biting impact on the head thickness of strip steel according to claim 4, characterized in that The calculation process of the self-learning algorithm is completed in the PLC controller of the primary control system.

Images