Finishing roll gap inclination presetting system based on historical rolling data and calibration data

Abstract

The invention discloses a finish rolling roll gap inclination presetting system based on historical rolling data and calibration data, and belongs to the technical field of hot continuous rolling finish rolling control. The system comprises a data module, a manual marking module, a roll gap inclination calculation module, a historical database and a model database; when calibration of the current rack is finished, the data module reads rolling data and calibration data corresponding to the current rack and reads rolling data and calibration data in a historical database; the roll gap inclination calculation module calculates a roll gap inclination value according to the data read by the data module and issues the roll gap inclination value to the current rack; when the data storage signal is received, the data module updates the historical database; and when the manual marking signal is received, the data module updates the model database and retrains the roll gap inclination calculation model by using the updated data in the model database. The stability of the rolling process can be improved, and interference of operators is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of hot rolling finishing control, in particular to a finishing roll gap inclination presetting system based on historical rolling data and calibration data. Background Art

[0002] The steel industry is in a stage of high-quality development. Especially in the context of increasing demand in the fields of aerospace, artificial intelligence, and military industry, the demand for high-precision and high-quality strip steel is increasing. Hot-rolled strip steel accounts for more than 75% of steel production. Among them, the finishing mill is the core of hot rolling production of strip steel, and its technical level determines the quality of the product. The precise measurement and control of strip deviation between finishing mill stands has become a hot issue in the industry. Due to multi-dimensional factors such as raw materials, rolling process, and control system for producing strip steel, the phenomenon of strip deviation is difficult to avoid.

[0003] The purpose of the roll gap tilt presetting is to pre-tilt the roll gap of the rolling mill according to the calibration of the rolling mill, so as to avoid asymmetric bounce of the roll gap during rolling due to the difference in stiffness on both sides of the rolling mill, which in turn causes the strip to deviate and produce asymmetric wave defects. The accuracy of the roll gap tilt value sets the tone for the rolling stability of the strip. If the roll gap tilt value is inaccurate, the strip will produce asymmetric wave defects after biting into the frame. When the wave shape is large, the strip is prone to rolling and breaking. In more serious cases, it will hit the side guide plate and cause rollover, resulting in steel pile accidents.

[0004] The Chinese patent application with application number CN201310100969.1 discloses a method for leveling the gap between the zero-adjusted working rolls. In a stationary state, the oil cylinders on both sides are automatically and synchronously pressed down to make the total rolling force reach the first target value of 500 ± 10T, and judge whether the following conditions are met: the difference in rolling force on both sides is ≤ 10T. If not, the rolls are pre-swung so that the side with large rolling force is lifted and the side with small rolling force is pressed down by the same amount to meet the requirements; then the car is turned and continued to be pressed down so that the total rolling force reaches the second target value, and the second target value is: F1~F4 is 1500 ± 10T, and F5~F7 is 1000 ± 10T. Then, the oil column deviation on both sides of the AGC oil cylinder, that is, the difference between the working side oil column value and the transmission side oil column value, is corrected again to reach the target value of the oil column deviation. The target value of oil column deviation is: the oil column deviation on both sides recorded during the rolling process of the third steel piece in the last rolling plan + the oil column deviation after the completion of 500T static pressure in this rolling - the oil column deviation after the completion of 500±10T static pressure in the last rolling. After the operator confirms that there is no abnormality, the zero adjustment pre-embedding process, that is, the zero adjustment working roll opening roll gap leveling method process is completed, and the rolling is ready.

[0005] A Chinese patent application with application number CN201711476218.4 discloses a method for zero adjustment of a finishing mill, comprising the following steps: recording the average value of the hydraulic cylinder deviation after the fourth to ninth pieces of steel in the previous production unit are thrown; after the previous production unit is rolled, changing the working roll; the hydraulic cylinder is initially swung to the initial position of zero adjustment; the hydraulic cylinder is pressed down to perform a zero adjustment pressure closed loop; the average value of the hydraulic cylinder deviation after the fourth to ninth pieces of steel are thrown is compensated; the hydraulic cylinder roll gap is cleared; the hydraulic cylinder is lifted to the set roll gap position to complete the non-leveling zero adjustment.

[0006] Although the above-mentioned prior arts all involve the presetting of the tilt of the finishing roll gap, they do not take into account the situation that the stiffness of both sides of the mill changes greatly due to the replacement of the support rolls or step pads before the calibration of the finishing mill, so they are not universal. In addition, the control accuracy of the above-mentioned prior arts is also low. Summary of the invention

[0007] The present invention provides a finishing roll gap inclination pre-setting system based on historical rolling data and calibration data to solve the technical problem that the prior art does not take into account the situation where the stiffness on both sides of the rolling mill changes greatly due to the replacement of support rolls or step pads before the finishing mill is calibrated, and therefore is not universal and has low control accuracy.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] A finishing roll gap tilt presetting system based on historical rolling data and calibration data comprises: a data module, a manual marking module, a roll gap tilt calculation module, a historical database and a model database; wherein:

[0010] The historical database and the model database are used to store rolling data and calibration data;

[0011] The manual marking module is used to trigger the manual marking signal;

[0012] The roll gap tilt calculation module includes a plurality of preset roll gap tilt calculation models;

[0013] During each calibration process of each finishing mill stand, the system performs the following operations on each finishing mill stand:

[0014] When the current rack calibration is finished, the data module reads the rolling data and calibration data corresponding to the current rack, and reads the rolling data and calibration data in the historical database;

[0015] The roll gap inclination calculation module calculates the roll gap inclination value through the data read by the data module as the roll gap inclination value corresponding to the current roll, and sends it to the current frame;

[0016] When receiving a stored data signal, the data module updates the historical database; when receiving an artificial marking signal, the data module updates the model database and uses the data in the updated model database to retrain the roll gap inclination calculation model in the roll gap inclination calculation module.

[0017] Furthermore, each piece of data stored in the historical database includes: time, working roll number, support roll number, upper step position, lower step position, cylinder difference at the end of calibration, and roll gap inclination value.

[0018] Further, when the current rack calibration is completed, the data module reads the rolling data and calibration data corresponding to the current rack, and reads the rolling data and calibration data in the historical database, specifically:

[0019] At the end of the current frame calibration, the data module reads the roll number of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current rolling roll, and the cylinder difference at the end of the calibration corresponding to the current rolling roll; and reads the cylinder difference and the roll gap inclination value at the end of the latest calibration from the historical database, which are used as the cylinder difference and the roll gap inclination value at the end of the previous roll calibration, respectively; then reads the upper step position, the lower step position, the cylinder difference and the roll gap inclination value at the end of the calibration that are the most recent and have the same roll number as the current working roll and the current support roll from the historical database, which are used as the upper step position with the same roll number, the lower step position with the same roll number, the cylinder difference with the same roll number at the end of the calibration, and the roll gap inclination value with the same roll number, respectively; wherein, in the process of the data module reading data, if any data reading is abnormal, the corresponding data will be assigned a null value.

[0020] Furthermore, the roll gap inclination calculation model includes a linear model and a particle swarm-random forest model.

[0021] Furthermore, the finishing roll gap tilt presetting system also includes a field control display module;

[0022] The roll gap inclination calculation module calculates the roll gap inclination value through the data read by the data module as the roll gap inclination value corresponding to the current roll, and sends it to the current rack, specifically:

[0023] If any of the roll numbers of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current roll, and the cylinder difference at the end of the calibration corresponding to the current roll is null, then "data reading abnormality, roll gap tilt value ΔS0" is displayed in the field control display module; wherein,

[0024]

[0025] in, is the roll gap inclination value of the previous roll;

[0026] If the upper step position, lower step position, cylinder difference at the end of calibration and roll gap tilt value of the same roll number are empty, the on-site control display module will display "New roll on the machine, roll gap tilt value ΔS0";

[0027]

[0028] If the upper step position corresponding to the current roll is the same as the upper step position of the same roll number and the lower step position corresponding to the current roll is the same as the lower step position of the same roll number and the remaining data are not null values, then "linear model, roll gap inclination value is ΔS0" is displayed in the field control display module; wherein ΔS0 is calculated by the linear model;

[0029] If the upper step position corresponding to the current roll is different from the upper step position of the same roll number or the lower step position corresponding to the current roll is different from the lower step position of the same roll number and the other data are not null values, then "RFR model, roll gap tilt value is ΔS0" is displayed in the field control display module; wherein ΔS0 is calculated by the particle swarm-random forest model;

[0030] The calculated roll gap inclination value is used as the roll gap inclination value corresponding to the current roll, and is sent to the current stand.

[0031] Furthermore, the formula for calculating ΔS0 by the linear model is:

[0032]

[0033] in, Indicates the roll gap inclination value of the same roll number; Indicates the cylinder difference at the end of calibration corresponding to the current roller; Indicates the cylinder difference at the end of calibration of the same roller number; Indicates the preset roll gap inclination coefficient of the same roll number; Indicates the preset cylinder difference coefficient; Indicates the preset rack factor.

[0034] Furthermore, the input of the particle swarm-random forest model is the upper step position and the lower step position corresponding to the current roller, the cylinder difference at the end of calibration corresponding to the current roller, the upper step position of the same roller number, the lower step position of the same roller number, the cylinder difference at the end of calibration of the same roller number, the roll gap inclination value of the same roller number, the cylinder difference at the end of calibration of the previous roller, and the roll gap inclination value of the previous roller; the output is the roll gap inclination value corresponding to the current roller.

[0035] Further, when receiving the storage data signal, the data module updates the historical database, specifically:

[0036] When the storage data signal is received for the first time, the data module stores the current time, the roll numbers of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current roll, the cylinder difference at the end of the calibration corresponding to the current roll, and the roll gap inclination value corresponding to the current roll as a new data in the historical database;

[0037] When the stored data signal is received for the Nth time, the data module updates the roll gap inclination value in the historical database to the roll gap inclination value on the frame at the current moment; wherein N is a preset integer greater than 1, and when the calibration start signal is received, the number of times the stored data signal is received is cleared; the stored data signal is triggered before the strip enters the F1 frame.

[0038] Further, when receiving the manual marking signal, the data module updates the model database and uses the data in the updated model database to retrain the roll gap inclination calculation model in the roll gap inclination calculation module, specifically:

[0039] When the manual marking signal is received, the data module updates the roll gap inclination value in the historical database to the roll gap inclination value on the frame at the current moment calculated by the roll gap inclination calculation module, and after the update, reads the roll number of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current roll, and the cylinder difference at the end of calibration corresponding to the current roll and the roll gap inclination value corresponding to the current roll; and reads the cylinder difference and roll gap inclination value at the end of the latest calibration from the historical database, as the cylinder difference and roll gap inclination value of the previous roll at the end of calibration, respectively; and then reads the cylinder difference and roll gap inclination value at the end of calibration from the historical database, as the cylinder difference and roll gap inclination value of the previous roll at the end of calibration, respectively; The upper step position, lower step position, cylinder difference at the end of calibration and roll gap tilt value with the latest reading time in the database and the same roll number as the current working roll and the current support roll are used as the upper step position with the same roll number, the lower step position with the same roll number, the cylinder difference at the end of calibration and the roll gap tilt value with the same roll number respectively; wherein, in the process of reading data by the data module, if any data reading is abnormal, the corresponding data will be assigned a null value; the read data is stored in the model database, the model database is updated, and the particle swarm-random forest model is retrained using the data in the updated model database.

[0040] Furthermore, the training process of the particle swarm-random forest model includes:

[0041] First, data is selected from the model database to construct the original data set;

[0042] The bootstrap method is used to randomly extract samples from the original data set to construct a training set for building a single decision tree. When building each decision tree, for each feature in the feature subset, multiple candidate split points are tried, and the feature and split point with the largest mean square error reduction are selected as the splitting criterion, and the splitting is continuously recursively performed until the maximum depth of the decision tree or the minimum number of samples for node splitting or the mean square error in the node reaches a preset threshold. At the same time, the mean square error is used as the optimization target, and the PSO algorithm is used to iteratively update the hyperparameters of the random forest by simulating the movement of particles in the search space. The hyperparameters include: the number of decision trees, the maximum depth of the decision tree, and the minimum number of samples for node splitting.

[0043] The average of the predicted values ​​of all decision trees is the roll gap inclination value calculated by the particle swarm-random forest model.

[0044] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0045] The technical solution of the present invention uses a data-driven approach to establish a scientific finishing roll gap tilt presetting system, which accurately calculates the roll gap tilt value for different working conditions, can effectively compensate for the asymmetric bounce phenomenon of the roll gaps on both sides during the loaded rolling process of the rolling mill, and can be updated according to the actual rolling conditions, thereby improving the stability of the rolling process and reducing operator intervention. The technical solution of the present invention also takes into account the situation where the stiffness of both sides of the rolling mill changes significantly due to the replacement of support rolls or step positions before the calibration of the finishing mill, and has universal applicability. At the same time, the technical solution of the present invention incorporates machine learning methods, and therefore also has higher control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 It is a structural block diagram of a finishing roll gap inclination presetting system provided by an embodiment of the present invention;

[0048] Figure 2 It is a working flow chart of the finishing roll gap inclination presetting system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0050] First of all, it should be noted that in the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "exemplarily" is intended to present the concept in a concrete way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0051] This embodiment provides a finishing roll gap tilt presetting system based on historical rolling data and calibration data. The finishing roll gap tilt presetting system is as follows: Figure 1 As shown, it includes: data module, manual marking module, roll gap tilt calculation module, historical database and model database; wherein, the roll gap tilt calculation model includes linear model and particle swarm-random forest model. According to the on-site rolling signal, the data module reads the required data from the on-site PLC and database and transmits it to the roll gap tilt calculation module at the end of the calibration of each rolling mill, stores the required data to the historical database when the data signal is stored for the first time, and updates the data when the data signal is stored for the Nth time; the manual marking module is equipped with a manual marking button on site, which updates the data in the historical database after clicking, stores the required data in the model database, and retrains the model in the roll gap tilt calculation module; according to the on-site rolling signal, the roll gap tilt calculation model determines the trigger logic according to the data transmitted by the data module at the end of the calibration of each frame, calculates the roll gap tilt value and sends it to each rolling mill, and displays it in the on-site control display module.

[0052] Specifically, Figure 2 As shown, the data module is specifically used for:

[0053] According to the on-site rolling signal, at the end of the calibration of each rolling mill, the current working roll and support roll numbers, upper and lower step positions, and calibration end cylinder difference are read from the on-site PLC equipped with the rolling mill. If any of the readings is abnormal, a null value is assigned and the value is passed to the roll gap tilt calculation module. The latest calibration end cylinder difference and roll gap tilt value are searched in the historical data stored in the historical database, and the corresponding values ​​of the calibration end cylinder difference and roll gap tilt value of the previous roll are assigned and passed to the roll gap tilt calculation module. According to the current working roll and support roll numbers, the latest calibration end cylinder difference, upper and lower step positions, and roll gap tilt values ​​of the same working roll and support roll are searched in the historical data stored in the historical database, and the calibration end cylinder difference, upper and lower step positions, and roll gap tilt values ​​of the same roll number are assigned. If not found, a null value is assigned and passed to the roll gap tilt calculation module.

[0054] When storing the data signal for the first time, the current time, the work roll and the backup roll number, the upper and lower step positions, the calibration end cylinder difference and the roll gap tilt value are stored in the historical database. The stored data signal is the signal before the industrial field strip enters the F1 rolling mill, including: the last pass of the rough rolling R2, the flying shear head, the temperature meter, and the vertical roll;

[0055] When the data signal is stored for the Nth time, the current roll gap tilt value is updated to the historical database. Wherein, N is a configurable parameter, and it is considered that the roll gap tilt value has been accurately adjusted before the strip is rolled, and after the stored data signal is received for the Nth time, the signal will no longer be received in the future.

[0056] The manual labeling module is specifically used for:

[0057] The site is equipped with a manual marking button. By clicking the manual marking button, the current roll gap tilt value is updated to the historical database, and the current time, working roll and support roll numbers, upper and lower step positions, calibration end cylinder difference and roll gap tilt value are stored in the model database. The calibration end cylinder difference and roll gap tilt value of the previous group of working rolls are searched in the historical database, and the calibration end cylinder difference and roll gap tilt value of the previous roll are assigned accordingly. The calibration end cylinder difference, upper and lower step positions and roll gap tilt values ​​with the latest time and the same working roll and support roll are searched, and the calibration end cylinder difference, upper and lower step positions and roll gap tilt values ​​with the same roll number are assigned accordingly. If not found, a blank value is assigned and stored in the model database. According to the updated model database, the particle swarm-random forest model in the roll gap tilt calculation module is retrained.

[0058] Among them, the particle swarm-random forest model is specifically:

[0059] First, data is selected from the field database to construct a data set. The Bootstrap method is used to randomly extract samples from the original data set to construct a training set for building a single decision tree. When constructing each decision tree, for each feature in the feature subset, multiple candidate split points are tried, and the feature and split point with the largest mean square error reduction are selected as the splitting criterion, and the split is continuously recursively performed until the maximum depth of the decision tree or the minimum number of samples for node splitting or the mean square error in the node reaches the threshold. At the same time, the mean square error is used as the optimization target, and the PSO algorithm is used to iteratively update the hyperparameters of the random forest by simulating the movement of particles in the search space: the number of decision trees, the maximum depth of the decision tree, and the minimum number of samples for node splitting. The average of the predicted values ​​of all decision trees is the roller gap inclination value calculated by the particle swarm-random forest model.

[0060] The roll gap tilt calculation module is specifically used for:

[0061] The data required by the roll gap inclination calculation module include the current working roll and support roll number, the current upper and lower step positions, the current calibration end cylinder difference, the upper and lower step positions of the same roll number, the roll gap inclination value of the same roll number, the calibration end cylinder difference of the same roll number, the roll gap inclination value of the previous roll, and the calibration end cylinder difference of the previous roll, among which:

[0062] If any of the current working roll and support roll numbers, the current upper and lower step positions, and the cylinder difference at the end of the current roll number calibration is null, the on-site control display module will display "Data reading abnormality, roll gap tilt ΔS0", and the previous roll gap tilt value will be sent to each rolling mill:

[0063]

[0064] in, The roll gap inclination of the previous roll during stable rolling, in mm, is obtained through the database;

[0065] If the upper and lower step positions of the same roll number, the roll gap tilt value of the same roll number, and the cylinder difference at the end of calibration of the same roll number are empty values, the current working roll or support roll is a new roll on the machine, and the on-site control display module displays "New roll on the machine, roll gap tilt ΔS0", and the roll gap tilt value of the previous roll is sent to each rolling mill:

[0066]

[0067] in, The roll gap inclination of the previous roll during stable rolling, in mm, is obtained through the database;

[0068] If the current upper and lower step positions are the same as those of the same roll number, and the other data are not empty, the linear model is triggered to calculate the roll gap inclination value, and the on-site control display module displays "Linear model, roll gap inclination ΔS0", and the model calculation value is sent to each rolling mill:

[0069]

[0070] in, The roll gap inclination value of the same roll number, in mm, is obtained through the database;

[0071] The current end of the calibration cylinder difference, in mm, obtained through the communication site PLC;

[0072] The cylinder difference at the end of calibration for the same roller number, in mm, is obtained through the database;

[0073] The roll gap inclination coefficient of the same roll number, its setting value is shown in Table 1:

[0074] Table 1 Roll gap inclination coefficients of the same roll number

[0075]

[0076] is the cylinder difference coefficient, and its setting value is shown in Table 2:

[0077] Table 2 Cylinder difference coefficient

[0078]

[0079] is the rack coefficient, and its setting value is shown in Table 3:

[0080] Table 3 Rack Factor

[0081]

[0082] If the current upper and lower step positions are different from the upper and lower step positions of the same roll number, and the rest of the data are not empty, the particle swarm-random forest model is triggered to calculate the roll gap tilt value, and the on-site control display module displays "RFR model, roll gap tilt ΔS0", and the model calculation value is sent to each rolling mill. The input data required by the particle swarm-random forest model are the current upper and lower step positions, the current calibration end cylinder difference, the upper and lower step positions of the same roll number, the roll gap tilt value of the same roll number, the calibration end cylinder difference of the same roll number, the roll gap tilt value of the previous roll, and the calibration end cylinder difference of the previous roll.

[0083] Next, the implementation process of the solution of the present invention is described by taking the F7 rack of the 1580 production line of a certain factory as an example.

[0084] First, historical rolling data and calibration data are collected to train the particle swarm-random forest model: According to the collected historical rolling data and calibration data, the data related to the preset roll gap tilt are screened out, including the current working roll and support roll number, the current upper and lower step positions, the current calibration end cylinder difference, the upper and lower step positions of the same roll number, the roll gap tilt value of the same roll number, the calibration end cylinder difference of the same roll number, the roll gap tilt value of the previous roll, and the calibration end cylinder difference of the previous roll, and the data set is constructed. Some data are shown in Table 4.

[0085] Table 4 Part of the data in the dataset

[0086]

[0087]

[0088] The Bootstrap method is used to randomly extract samples from the data set to construct a training set for building a single decision tree. When constructing each decision tree, for each feature in the feature subset, multiple candidate split points are tried, and the features and split points with the largest mean square error reduction are selected as the splitting criteria, and the splitting is continuously recursively performed until the maximum depth of the decision tree or the minimum number of samples for node splitting or the mean square error in the node reaches the threshold. At the same time, the mean square error is used as the optimization target, and the PSO algorithm is used to iteratively update the hyperparameters of the random forest by simulating the movement of particles in the search space: the number of decision trees, the maximum depth of the decision tree, and the minimum number of samples for node splitting. The average of the predicted values ​​of all decision trees is the roller gap inclination value calculated by the particle swarm-random forest model. At this point, the training of the particle swarm-random forest model is completed, and the roller gap inclination value can be calculated based on the input data.

[0089] According to the on-site rolling signal, when the F7 rolling mill calibration end signal is received, the data module communication on-site PLC reads the current upper working roll number FW61910, lower working roll number FW61907, upper support roll number FB20002, lower support roll number FB20001, upper step position 3, lower step position 3, and calibration end cylinder difference -2.36mm, and transmits them into the roll gap inclination calculation model.

[0090] The latest calibration end cylinder difference -2.63mm and roll gap tilt value -0.6mm are searched in the historical data of the database, and the corresponding values ​​of the calibration end cylinder difference of the previous roll and the roll gap tilt value of the previous roll are assigned and passed into the roll gap tilt calculation model.

[0091] Then, according to the current working roll and backup roll numbers, the latest working roll and backup roll with the same calibration end cylinder difference -2.27mm, upper step position 3, lower step position 3 and roll gap tilt value -0.58mm are searched in the historical data of the database, and the calibration end cylinder difference, upper and lower step positions and roll gap tilt value with the same roll number are assigned accordingly, and the values ​​are input into the roll gap tilt calculation model;

[0092] The current upper and lower step positions are the same as those of the same roll number, and the other data are not empty, triggering the linear model to calculate the roll gap inclination value:

[0093]

[0094] in, The roll gap inclination value of the same roll number is -0.58mm obtained through the database;

[0095] The cylinder difference at the end of the current calibration is -2.36mm obtained through the communication site PLC;

[0096] The cylinder difference for the same roller number calibration end is obtained from the database as -2.27mm;

[0097] The roll gap inclination coefficient of the same roll number is set to 1;

[0098] is the cylinder difference coefficient, and its setting value is 1;

[0099] is the rack coefficient, set the value to 1;

[0100] The calculated roll gap inclination value of F7 rolling mill is .

[0101] “Linear model, roll gap inclination -0.67mm” is displayed on the on-site control interface, and the model calculation value is sent to the F7 rolling mill.

[0102] Before the strip enters the F1 stand, store the current time, upper working roll number FW61910, lower working roll number FW61907, upper support roll number FB20002, lower support roll number FB20001, current upper and lower step positions 3, current lower step position 3, current calibration end cylinder difference -2.63mm and current roll gap inclination value -0.67mm to the database;

[0103] When the data signal is stored for the fifth time or the manual marking signal is received, the current roll gap inclination value -0.69 mm is updated to the database.

[0104] After the completion of this rolling plan, the F7 mill replaced the working rolls and carried out calibration.

[0105] When receiving the calibration end signal of the F7 rolling mill, the data module communication site PLC reads the current upper working roll number FW61940, lower working roll number FW61939, upper support roll number FB20002, lower support roll number FB20001, upper step position 4, lower step position 4, and calibration end cylinder difference -2.06mm, and transmits them into the roll gap inclination calculation model.

[0106] The latest calibration end cylinder difference -2.29mm and roll gap inclination value -0.74mm are searched in the historical data of the database, and the corresponding values ​​of the calibration end cylinder difference of the previous roll and the roll gap inclination value of the same roll number are assigned and transmitted to the roll gap inclination calculation model.

[0107] Then, according to the current working roll and backup roll numbers, the latest working roll and backup roll with the same calibration end cylinder difference -2.31mm, upper step position 3, lower step position 4 and roll gap tilt value -0.69mm are searched in the historical data of the database, and the calibration end cylinder difference, upper and lower step positions and roll gap tilt value with the same roll number are assigned accordingly, and then input into the roll gap tilt calculation model;

[0108] The current upper and lower step positions are different from the upper and lower step positions of the same roll number, and the other data are not empty values, triggering the particle swarm-random forest model to calculate the roll gap tilt value .

[0109] “RFR model, roll gap tilt -0.83mm” is displayed on the on-site control interface, and the model calculation value is sent to the F7 rolling mill.

[0110] Before the strip enters the F1 stand, store the current time, upper working roll number FW61940, lower working roll number FW61939, upper support roll number FB20002, lower support roll number FB20001, current upper and lower step positions 4, current lower step position 4, current calibration end cylinder difference -2.06mm and current roll gap inclination value -0.83mm to the database;

[0111] When the data signal is stored for the fifth time or the manual marking signal is received, the current roll gap inclination value -0.93 mm is updated to the database.

[0112] In summary, this embodiment uses a data-driven approach to establish a scientific finishing roll gap inclination pre-setting system, which accurately calculates the roll gap inclination value for different working conditions, can effectively compensate for the asymmetric bounce phenomenon of the roll gaps on both sides during the loaded rolling process of the rolling mill, and can be updated according to the actual rolling conditions, thereby improving the stability of the rolling process and reducing operator intervention. In addition, the technical solution of this embodiment takes into account the situation where the stiffness of both sides of the rolling mill changes greatly due to the replacement of support rolls or step pads before the calibration of the finishing mill, and is universal. At the same time, the technical solution of this embodiment incorporates machine learning methods, and therefore also has higher control accuracy.

[0113] In addition, it should be noted that the present invention can be provided as a method, an apparatus or a computer program product. Therefore, the embodiment of the present invention can be in the form of a full or partial hardware embodiment, a full or partial software embodiment or an embodiment combining software and hardware. Moreover, when implemented using software, the embodiment of the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center containing one or more available media sets. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium. The semiconductor medium may be a solid state drive.

[0114] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] It should also be noted that, in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "include one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements. In addition, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships, for example, A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship, which can be understood by referring to the context before and after. "At least one" means one or more, and "more" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or plural.

[0117] In addition, it can be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0118] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0119] In several embodiments provided by the present invention, it should be understood that the disclosed equipment, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of functional modules / units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0120] If the method is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0121] Finally, it should be noted that the above is only a preferred embodiment of the present invention. It should be pointed out that although the preferred embodiment of the present invention has been described, for ordinary technicians in this technical field, once the basic creative concept of the present invention is known, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. Therefore, the attached claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the present invention.

Claims

1. A finishing roll gap tilt presetting system based on historical rolling data and calibration data, characterized in that: The finishing roll gap tilt presetting system comprises: a data module, a manual marking module, a roll gap tilt calculation module, a history database and a model database; wherein, The historical database and the model database are used to store rolling data and calibration data; The manual marking module is used to trigger the manual marking signal; The roll gap tilt calculation module includes a plurality of preset roll gap tilt calculation models; During each calibration process of each finishing mill stand, the system performs the following operations on each finishing mill stand: When the current rack calibration is finished, the data module reads the rolling data and calibration data corresponding to the current rack, and reads the rolling data and calibration data in the historical database; The roll gap inclination calculation module calculates the roll gap inclination value through the data read by the data module as the roll gap inclination value corresponding to the current roll, and sends it to the current frame; When receiving a stored data signal, the data module updates the historical database; when receiving an artificial marking signal, the data module updates the model database and uses the data in the updated model database to retrain the roll gap inclination calculation model in the roll gap inclination calculation module.

2. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 1, characterized in that: Each piece of data stored in the historical database includes: time, working roll number, support roll number, upper step position, lower step position, cylinder difference at the end of calibration, and roll gap inclination value.

3. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 1, characterized in that: When the current rack calibration is finished, the data module reads the rolling data and calibration data corresponding to the current rack, and reads the rolling data and calibration data in the historical database, specifically: At the end of the current frame calibration, the data module reads the roll number of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current rolling roll, and the cylinder difference at the end of the calibration corresponding to the current rolling roll; and reads the cylinder difference and the roll gap inclination value at the end of the latest calibration from the historical database, which are used as the cylinder difference and the roll gap inclination value at the end of the previous roll calibration, respectively; then reads the upper step position, the lower step position, the cylinder difference and the roll gap inclination value at the end of the calibration that are the most recent and have the same roll number as the current working roll and the current support roll from the historical database, which are used as the upper step position with the same roll number, the lower step position with the same roll number, the cylinder difference with the same roll number at the end of the calibration, and the roll gap inclination value with the same roll number, respectively; wherein, in the process of the data module reading data, if any data reading is abnormal, the corresponding data will be assigned a null value.

4. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 3, characterized in that: The roll gap tilt calculation model includes a linear model and a particle swarm-random forest model.

5. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 4, characterized in that: The finishing roll gap tilt presetting system also includes a field control display module; The roll gap inclination calculation module calculates the roll gap inclination value through the data read by the data module as the roll gap inclination value corresponding to the current roll, and sends it to the current rack, specifically: If any of the roll numbers of the current working roll and the current backup roll, the upper step position and the lower step position corresponding to the current roll, and the cylinder difference at the end of the calibration corresponding to the current roll is null, "Data reading abnormality, roll gap tilt value ΔS0" is displayed in the field control display module; wherein, ； in, is the roll gap inclination value of the previous roll; If the upper step position, lower step position, cylinder difference at the end of calibration and roll gap tilt value of the same roll number are empty, "New roll on the machine, roll gap tilt value ΔS0" will be displayed in the field control display module; ； If the upper step position corresponding to the current roll is the same as the upper step position of the same roll number and the lower step position corresponding to the current roll is the same as the lower step position of the same roll number and the remaining data are not null values, "Linear model, roll gap inclination value is ΔS0" is displayed in the field control display module; wherein ΔS0 is calculated by the linear model; If the upper step position corresponding to the current roll is different from the upper step position of the same roll number or the lower step position corresponding to the current roll is different from the lower step position of the same roll number and the other data are not null values, "RFR model, roll gap tilt value is ΔS0" will be displayed in the field control display module; wherein ΔS0 is calculated by the particle swarm-random forest model; The calculated roll gap inclination value is used as the roll gap inclination value corresponding to the current roll, and is sent to the current stand.

6. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 5, characterized in that: The formula for calculating ΔS0 by the linear model is: ； in, Indicates the roll gap inclination value of the same roll number; Indicates the cylinder difference at the end of calibration corresponding to the current roller; Indicates the cylinder difference at the end of calibration of the same roller number; Indicates the preset roll gap inclination coefficient of the same roll number; Indicates the preset cylinder difference coefficient; Indicates the preset rack factor.

7. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 5, characterized in that: The input of the particle swarm-random forest model is the upper step position and the lower step position corresponding to the current roller, the cylinder difference at the end of calibration corresponding to the current roller, the upper step position of the same roller number, the lower step position of the same roller number, the cylinder difference at the end of calibration of the same roller number, the roll gap inclination value of the same roller number, the cylinder difference at the end of calibration of the previous roller, and the roll gap inclination value of the previous roller; the output is the roll gap inclination value corresponding to the current roller.

8. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 3, characterized in that: When receiving the storage data signal, the data module updates the historical database, specifically: When the storage data signal is received for the first time, the data module stores the current time, the roll numbers of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current roll, the cylinder difference at the end of the calibration corresponding to the current roll, and the roll gap inclination value corresponding to the current roll as a new data in the historical database; When the stored data signal is received for the Nth time, the data module updates the roll gap inclination value in the historical database to the roll gap inclination value on the frame at the current moment; wherein N is a preset integer greater than 1, and when the calibration start signal is received, the number of times the stored data signal is received is cleared; the stored data signal is triggered before the strip enters the F1 frame.

9. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 4, characterized in that: When receiving the manual marking signal, the data module updates the model database and uses the data in the updated model database to retrain the roll gap inclination calculation model in the roll gap inclination calculation module, specifically: When the manual marking signal is received, the data module updates the roll gap inclination value in the historical database to the roll gap inclination value on the frame at the current moment, and after the update, reads the roll number of the current working roll and the current support roll, the upper step position and the lower step position corresponding to the current roll, and the cylinder difference at the end of the calibration corresponding to the current roll and the roll gap inclination value corresponding to the current roll; and reads the cylinder difference and roll gap inclination value at the end of the latest calibration from the historical database, as the cylinder difference and roll gap inclination value at the end of the calibration of the previous roll, respectively; and then reads the cylinder difference and roll gap inclination value at the end of the calibration of the previous roll from the historical database. The upper step position, lower step position, cylinder difference at the end of calibration and roll gap tilt value that are the latest and have the same roll number as the current working roll and the current support roll are used as the upper step position with the same roll number, the lower step position with the same roll number, the cylinder difference at the end of calibration and the roll gap tilt value with the same roll number respectively; wherein, in the process of data reading by the data module, if any data reading is abnormal, the corresponding data will be assigned a null value; the read data is stored in the model database, the model database is updated, and the particle swarm-random forest model is retrained using the data in the updated model database.

10. The finishing roll gap tilt presetting system based on historical rolling data and calibration data according to claim 9, characterized in that: The training process of the particle swarm-random forest model includes: First, data is selected from the model database to construct the original data set; The bootstrap method is used to randomly extract samples from the original data set to construct a training set for building a single decision tree. When building each decision tree, for each feature in the feature subset, multiple candidate split points are tried, and the feature and split point with the largest mean square error reduction are selected as the splitting criterion, and the splitting is continuously recursively performed until the maximum depth of the decision tree or the minimum number of samples for node splitting or the mean square error in the node reaches a preset threshold. At the same time, the mean square error is used as the optimization target, and the PSO algorithm is used to iteratively update the hyperparameters of the random forest by simulating the movement of particles in the search space. The hyperparameters include: the number of decision trees, the maximum depth of the decision tree, and the minimum number of samples for node splitting. The average of the predicted values ​​of all decision trees is the roll gap inclination value calculated by the particle swarm-random forest model.

Images