Width control parameter pre-adjusting method for long-time non-rolling specification

Abstract

The invention relates to the technical field of steel rolling automatic control, in particular to a width control parameter pre-adjusting method for a long-time non-rolling specification, and the method comprises the steps: building a specification layer classification system; rolling records with the high width hit rate are screened from the recent data to serve as reference data, and set width control parameters of recent rolled specifications are determined to serve as pre-adjustment calculation benchmarks; according to the layer difference between the to-be-rolled specification and the reference specification, a parameter mapping rule is designed; and constructing a pre-adjustment amplitude calculation model, introducing a layer difference coefficient to dynamically correct a set parameter, and generating a width control parameter suitable for the current working condition. The method has the advantages that the specification similarity is quantitatively evaluated through the layer difference coefficient, so that the width control parameter for recovering and producing the first coil of strip steel is closer to the current working condition; by taking the SPHC specification as an example, the average deviation of a first roll is reduced to-4-8 mm from-6-8 mm in a traditional method, the average deviation of a second roll is stably controlled within + / -4 mm, the width hit rate is increased to 96.8% from 89.3%, and the problem of width fluctuation at the initial stage of specification switching is solved.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology for steel rolling, and in particular to a method for pre-adjusting the width control parameters of a specification that has not been rolled for a long time. Background Technology

[0002] Hot-rolled strip steel is one of the main products of the steel industry, widely used in automobiles, shipbuilding, pipelines and other fields. In the production process of hot-rolled strip steel, width control is one of the core indicators for measuring the dimensional accuracy of the product, which directly affects the yield and applicability of subsequent processing.

[0003] Currently, hot-rolled strip steel production lines typically employ a combination of process control systems (Level 2) and basic automation systems (Level 1) to achieve width control. The process control system establishes a mathematical model based on incoming material properties, finished product targets, and rolling processes to calculate and set key parameters such as the opening degree of the vertical rolls in the roughing mill. However, in actual continuous production, due to the variability of production plans and the diversification of market demands, rolling lines often need to switch between producing different specifications of products.

[0004] Existing technologies face the following technical challenges when dealing with specification switching, especially when resuming production of specifications that have not been rolled for a long time: 1. The factors affecting width are time-varying and complex; The chemical composition of different batches of slabs, fluctuations in heating furnace conditions (such as differences in tapping temperature), changes in equipment status (such as wear of vertical rolls and thermal expansion), and slight drifts in process parameters can all cumulatively affect the final strip width accuracy. When a certain specification has not been produced for a long time, all of the above factors have changed significantly, resulting in a serious mismatch between the original mathematical model settings and the new operating conditions.

[0005] 2. Traditional control methods lack effective pre-adjustment mechanisms; Existing process control systems typically rely on self-learning functions to correct model parameters. However, for specifications that have not been rolled for a long time, the self-learning function cannot converge quickly due to the lack of recent rolling data. This results in a significant drop in the width hit rate in the early stages of production resumption (such as the first and second steel pieces after changing specifications), and even the width deviation along the entire length, causing a large number of scraps or downgrades.

[0006] Therefore, how to quickly and accurately provide roughing setting values ​​that adapt to new working conditions when resuming production of specifications that have not been rolled for a long time, so as to improve the width control accuracy, is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a method for pre-adjusting the width control parameters of long-unrolled specifications. This method establishes a specification stratification system based on steel grade, thickness, width, and lateral pressure. High-hit-rate records of similar specifications are selected from recent production data as references. A pre-adjustment range calculation model is constructed, and a stratification difference coefficient is introduced to dynamically correct the set parameters. This generates width control parameters suitable for the current operating conditions, which are then written into the rolling mill control system. Feedback and iterative optimization are performed using the measured deviation of the first coil. This significantly improves the width control accuracy of long-unrolled specifications during the initial production recovery phase, reduces scrap generation, and increases the yield of hot-rolled strip steel and the continuous operation capability of the production line.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A method for pre-adjusting width control parameters for specifications that have not been rolled for a long time, comprising: Collect historical production data to identify factors affecting the accuracy of strip width control; Establish a classification system based on specification tiers, using steel type, thickness, width, and lateral pressure as tier indexes; Rolling records with high width hit rates are selected from recent continuous monitoring of historical production data as reference data. The reference data is used to determine the set width control parameters for recently rolled specifications and serves as the input benchmark for pre-adjustment calculations. Based on the layer differences between the current specifications to be rolled and the reference specifications, design parameter mapping and conversion rules are established. Construct a pre-adjustment amplitude calculation model to generate width control parameters applicable to long-term unrolled specifications; The pre-adjusted width control parameters are written into the rolling mill control system, and a feedback mechanism is established for dynamic correction.

[0009] Factors affecting the accuracy of strip width include one or more of the following: raw material characteristics, heating process, equipment status, and process parameters.

[0010] The classification system of specifications was established using cluster analysis, with steel grade, thickness, width, and lateral pressure as the key dimensions for stratification.

[0011] The selection of reference data includes the following priority order: Data within the same stratum > Data in adjacent stratums > Data across stratums; And conduct a comprehensive evaluation in conjunction with operating conditions; Operating conditions include rolling mileage and equipment status.

[0012] The pre-adjustment calculation is based on the stratification difference between the current specification to be rolled and the reference specification. Through the constructed pre-adjustment range calculation model, the set width control parameters are corrected to generate width control parameters applicable to specifications that have not been rolled for a long time.

[0013] The parameter mapping and conversion rules include proportional adjustments to key control parameters to accommodate differences between different specifications.

[0014] Key control parameters include one or more of the following: vertical roll opening, rolling force setting, tension setting, and speed setting.

[0015] The expression for the pre-adjustment amplitude calculation model is: ①; in: Steel grade layer difference coefficient The value range is [0, 0.2]. When the span between the steel grade and the reference grade of the specification to be rolled is less than or equal to two grade levels, ; Thickness layer difference coefficient The value range is [0, 0.2], and the thickness layer difference coefficient is... There is a positive correlation between the difference and the target thickness of the strip; the larger the difference, the better. ; Width layer difference coefficient The value range is [0, 0.2], and the width layer difference coefficient is... There is a positive correlation between the difference and the target width of the strip; the larger the difference, the better. ; Lateral pressure layer difference coefficient The value range is [0, 0.2], and the width layer difference coefficient is... There is a positive correlation between the difference in the amount of side pressure on the strip and the amount of side pressure; the larger the difference, the greater the difference. .

[0016] The feedback mechanism includes monitoring the actual width control of the first coil of strip and adjusting the pre-adjustment parameters of subsequent strips based on the width deviation value obtained from the monitoring.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. Traditional methods, when resuming production of specifications that have been idle for a long time, suffer from a lack of recent data support, which prevents the self-learning function of the process control system from converging quickly, resulting in a significant drop in the width hit rate of the first coil and the first few coils. This invention constructs a pre-adjustment amplitude calculation model based on layer differences and uses successful control experience of similar specifications in the recent period to dynamically correct the set parameters, making the width control parameters of the first coil of strip more closely reflect the current working conditions. Taking SPHC specifications as an example, after adopting this invention, the average deviation between the actual width and the target width of the first coil of strip is reduced from -6~8mm in the traditional method to -4~8mm, and the second coil is stably controlled within ±4mm. The width hit rate is increased from 89.3% to 96.8%, effectively solving the width fluctuation problem in the initial stage of specification switching. 2. A specification stratification system based on steel grade, thickness, width, and lateral pressure was established using cluster analysis, and a steel grade stratification difference coefficient was introduced. Thickness layer difference coefficient Width layer difference coefficient Lateral pressure difference coefficient This method quantitatively assesses the similarity between different specifications. When there are layer differences between the specification to be rolled and the reference specification, the reference parameters are adjusted proportionally using a difference coefficient. This avoids misadjustments caused by directly applying data from different specifications and expands the usable range of reference data. For example, within the ultra-low carbon group, this method considers the similarity between specifications of different thicknesses. The coefficient can accurately reflect subtle differences in breadth characteristics, while the larger span between the low-alloy and ultra-low-carbon groups is achieved through... The coefficients are effectively suppressed to enable the safe and accurate reuse of historical experience; 3. After the pre-adjustment parameters are written into the control system and actual rolling is carried out, a feedback mechanism based on the measured data of the width measuring instrument is further established. By monitoring the deviation between the actual width and the target width of the first coil of strip in real time, the deviation value is used as the basis for correction to iteratively optimize the pre-adjustment parameters of subsequent strips. This closed-loop control process enables the parameters to be adaptively adjusted according to the small changes in the current working conditions, avoiding the cumulative error that may exist in open-loop control. For example, in Example 1, after the deviation of the first coil of -4~8mm is fed back to the second coil, the opening degree of the vertical roll is corrected in time, and the second coil hits the target. The updated parameters in subsequent batch production can also be used as the initial setting value for the next production of the same specification, forming a virtuous cycle. 4. Width deviation is one of the main quality defects of hot-rolled strip steel, which can lead to the scrapping or downgrading of the entire coil in severe cases. This invention improves the width control accuracy of specifications that have not been rolled for a long time, effectively reducing the amount of scrap generated in the early stage of production resumption. Taking a 2150mm hot-rolling production line as an example, production can be quickly stabilized after specification switching, reducing rolling interruptions and adjustment time caused by width fluctuations, and improving the production line's operating rate and capacity utilization. 5. Based on the existing process control system and data acquisition platform, no additional hardware equipment is required. It is mainly achieved by embedding a pre-adjustment algorithm in the L2 level model. The layer division method, difference coefficient setting and pre-adjustment formula are universal and can be adapted to the parameters according to the equipment characteristics and data accumulation of different production lines. It is suitable for various hot-rolled wide strip steel production lines and has good prospects for promotion and application. Detailed Implementation

[0018] The present invention will now be described in detail, but it should be noted that the implementation of the present invention is not limited to the following embodiments.

[0019] The following embodiments are implemented based on the technical solution of the present invention, providing detailed implementation methods and specific operation processes. However, the scope of protection of the present invention is not limited to the following embodiments. Unless otherwise specified, the methods used in the following embodiments are conventional methods.

[0020] Example 1: A method for pre-adjusting width control parameters for long-term unrolled specifications, wherein the activation condition for the pre-adjustment algorithm is determined to be strip steel with no rolling record within 7 days, or strip steel with no rolling record within the current vertical roll cycle, specifically including: S1. Collect historical production data to analyze and identify factors affecting the accuracy of strip width control; Production records for the past 12 months of a hot rolling production line were collected, yielding a total of 184,323 coils of valid rolling data. Data dimensions included: slab number, steel grade code, incoming thickness, incoming width, target thickness, target width, actual width, vertical roll opening setting, measured rolling force, furnace exit temperature, final rolling temperature, coiling temperature, and measured width gauge value.

[0021] Statistical analysis of data on strip steel with width deviation (defined as a deviation between actual width and target width > 6mm) revealed the following: (1) Raw material factors: The width hit rate of slabs produced by different continuous casting machines differed by approximately 14.7%; the width deviation of slabs produced by continuous casting machine #1 was 10.5 mm, that of continuous casting machine #2 was 13.39 mm, and that of continuous casting machine #3 was 9.1 mm. (2) Heating process: When the furnace temperature fluctuates by more than ±25℃, the width deviation increases by 15%~20%; the width expansion of slabs tapped at 1150℃ and 1230℃ differs by about 4~7mm. (3) Equipment status: In the initial stage after vertical roll calibration (first 1 / 3 of the rolling cycle), the width hit rate was 93.6%, and in the later stage of calibration (second 2 / 3 of the rolling cycle), the width hit rate rose to 95.5%, while the equipment condition remained relatively stable within a short period of time. (4) Process parameters: When the rolling force fluctuates by ±500kN, the width fluctuates by approximately ±1.5mm.

[0022] S2. Establish a classification system based on specifications and tiers, using steel type, thickness, width, and lateral pressure as tier indexes; The K-means clustering algorithm was used to divide the historical data into specification layers, and the steel grade code, target thickness, target width, and side pressure (the difference between the incoming material width and the target width) were used as clustering feature vectors.

[0023] Taking actual production data as an example, the following stratification table is established: Similarity between strata is calculated using Euclidean distance; the smaller the distance, the closer the strata are. For example, the similarity distance between strata L01 and L02 is 0.02, and the similarity distance between group L01 and L09 is 0.35.

[0024] S3. Select rolling records with high width hit rates from recent continuous monitoring of historical production data as reference data; Suppose the current specification to be rolled is: steel grade M3A33, target thickness 4.0mm, target width 1250mm, side pressure 40mm, and this specification has not been produced for 45 consecutive days.

[0025] Reference data was selected from production data over the past 7 days (the most recent continuous monitoring period), based on the following selection criteria: (1) Prioritize filtering data within the same stratum: The most recent rolling records for layer L03 (ultra-low carbon group, thickness 3.2-3.6mm, width 1200-1300mm) were retrieved, yielding 8 rolls of valid data with a width hit rate (deviation ≤6mm) of 96.4%. (2) If there is no data in the same stratum, then consider adjacent strata: The width hit rate for layer L01 (thickness 4.5-5.1mm, width 1500-1600mm) is 95.2%. (3) Considering operating conditions: Prioritize records with a rolling mileage difference of ≤50km (equipment condition is similar), and remove records with large differences in rolling mileage between different vertical roll calibration cycles and within the same vertical roll calibration cycle.

[0026] Finally, the strip steel with a width hit rate of >95% from the most recently rolled 5 coils in layer L03 was selected as reference data, and its average width control parameters were set accordingly. The value is determined to be -18.

[0027] S4. Based on the layer differences between the current specifications to be rolled and the reference specifications, design parameter mapping and conversion rules; The current rolling specification (target thickness 4.0mm, target width 1250mm) differs from the reference specification (layer L03, average target thickness 3.5mm, average target width 1280mm), requiring parameter mapping and conversion. In existing model algorithms, intervention on the opening degree of the vertical roller is mainly achieved by modifying the width control parameter. To achieve this, the influence of thickness and width group intervals on the mapping coefficients needs to be derived from regression analysis of nearly one year's worth of on-site production data, and gradually adjusted during algorithm debugging to determine the final mapping coefficient results.

[0028] S5. Construct a pre-adjustment amplitude calculation model to generate width control parameters applicable to long-term unrolled specifications; The expression for the calculation model using the pre-adjustment amplitude is: ①; in: This indicates the latest width control parameters after a long period of non-rolling and specification adjustment. This indicates the original width control parameters for specifications that have not been rolled for a long time in the historical database. ; This indicates the optimal width control parameters for recently rolled specifications. ; This represents the coefficient of difference between steel grade layers. ; This represents the thickness difference coefficient, with a thickness difference of 0.3 mm (2.3 mm - 2.0 mm), based on an empirical formula. ,Pick ; This represents the width difference coefficient, with a width difference of 30mm (1280mm-1250mm), based on an empirical formula. ,Pick ; This represents the difference coefficient between lateral pressure layers, with a lateral pressure difference of 5mm (the original reference specification had a lateral pressure of 40mm, while the current specification is 35mm). ; The smoothing coefficient for parameter adjustment is a parameter that controls the pre-adjustment range and is set to 0.8 based on historical experience. Substitute into the formula to calculate: ; ; Determine the latest width control parameters for specifications that have not been rolled for a long time. .

[0029] S6. Write the pre-adjusted width control parameters into the rolling mill control system and establish a feedback mechanism for dynamic correction. The calculated unrolled specification width control parameter -16 is written into the roughing process control system (L2 level) for model calculation of the control parameter of this specification, and the calculation result is sent to the basic automation system (L1 level) for execution.

[0030] The feedback mechanism is established as follows: (1) After the first strip is rolled, the actual width is detected in real time by a width measuring instrument (model: wg2010, measurement accuracy ±0.6mm) at the mill exit; (2) The test results show that the width of the first coil of strip is 1256.2 mm at the head, 1255.7 mm in the middle, and 1254.8 mm at the tail. The average width of the whole length is 1255.4 mm, and the average deviation from the target width of 1250 mm is 5.4 mm. (3) Based on the deviation value, the pre-adjustment parameters of the subsequent strip steel are iteratively optimized again: the latest width control parameters of the subsequent specifications are determined in order to calculate the corresponding vertical roll setting data; (4) After the second strip is rolled, the width measuring instrument shows that the actual width is 1247~1254mm, the deviation is within ±4mm, and the target is hit; (5) When producing subsequent batches, update the pre-adjustment parameters of this specification to the model parameter library as the initial setting value for the next production of the same specification.

[0031] Through the above closed-loop control process, the width hit rate of specifications that have not been rolled for a long time is improved, effectively solving the width fluctuation problem in the early stage of specification switching.

[0032] Traditional methods, when resuming production of specifications that have been idle for a long time, suffer from a lack of recent data support, causing the process control system's self-learning function to fail to converge quickly, resulting in a significant drop in the width hit rate of the first coil and the first few coils. This invention constructs a pre-adjustment amplitude calculation model based on stratification differences, dynamically correcting the set parameters using successful control experience of similar specifications in recent times. This makes the width control parameters of the first coil of strip more closely reflect the current operating conditions. Taking SPHC specifications as an example, after adopting this invention, the average deviation between the actual width and the target width of the first coil of strip is reduced from -6~8mm in the traditional method to -4~8mm, and the second coil is stably controlled within ±4mm. The width hit rate increases from 89.3% to 96.8%, effectively solving the width fluctuation problem in the initial stage of specification switching. A specification stratification classification system is established using cluster analysis with steel grade, thickness, width, and lateral pressure as key dimensions, and a steel grade stratification difference coefficient is introduced. Thickness layer difference coefficient Width layer difference coefficient Lateral pressure difference coefficient This method quantitatively assesses the similarity between different specifications. When there are layer differences between the specification to be rolled and the reference specification, the reference parameters are adjusted proportionally using a difference coefficient. This avoids misadjustments caused by directly applying data from different specifications and expands the usable range of reference data. For example, within the ultra-low carbon group, this method considers the similarity between specifications of different thicknesses. The coefficient can accurately reflect subtle differences in breadth characteristics, while the larger span between the low-alloy and ultra-low-carbon groups is achieved through... The coefficients are effectively suppressed to achieve safe and accurate reuse of historical experience. After the pre-adjustment parameters are written into the control system and actual rolling is carried out, a feedback mechanism based on the measured data of the width measuring instrument is further established. By monitoring the deviation between the actual width and the target width of the first coil of strip in real time, the deviation value is used as the basis for correction to iteratively optimize the pre-adjustment parameters of subsequent strips. This closed-loop control process enables the parameters to be adaptively adjusted according to the small changes in the current working conditions, avoiding the cumulative errors that may exist in open-loop control. For example, in Example 1, after the deviation of the first coil of -4~8mm is fed back to the second coil, the opening degree of the vertical roll is corrected in time, and the second coil hits the target. The updated parameters in subsequent batch production can also be used as the initial setting value for the next production of the same specification, forming a virtuous cycle. Width deviation is a problem in hot rolling. One of the main quality defects in strip steel is width, which can lead to the scrapping or downgrading of the entire coil in severe cases. This invention effectively reduces the amount of scrap generated in the initial stage of production resumption by improving the width control accuracy of specifications that have not been rolled for a long time. Taking a 2150mm hot rolling production line as an example, production can be quickly stabilized after specification switching, reducing rolling interruptions and adjustment time caused by width fluctuations, and improving the production line's operating rate and capacity utilization. It is based on the existing process control system and data acquisition platform, without the need for additional hardware equipment. It is mainly achieved by embedding a pre-adjustment algorithm in the L2 level model. The layer division method, difference coefficient setting, and pre-adjustment formula are universal and can be adapted to the parameters according to the equipment characteristics and data accumulation of different production lines. It is applicable to various hot-rolled wide strip steel production lines and has good prospects for promotion and application.

Claims

1. A method for pre-adjusting width control parameters for specifications that have not been rolled for a long time, characterized in that, include: Collect historical production data to identify factors affecting the accuracy of strip width control; Establish a classification system based on specification tiers, using steel type, thickness, width, and lateral pressure as tier indexes; Rolling records with high width hit rates are selected from recent continuous monitoring of historical production data as reference data. The reference data is used to determine the set width control parameters for recently rolled specifications and serves as the input benchmark for pre-adjustment calculations. Based on the layer differences between the current specifications to be rolled and the reference specifications, design parameter mapping and conversion rules are established. Construct a pre-adjustment amplitude calculation model to generate width control parameters applicable to long-term unrolled specifications; The pre-adjusted width control parameters are written into the rolling mill control system, and a feedback mechanism is established for dynamic correction.

2. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The factors affecting the accuracy of strip width include one or more of the following: raw material characteristics, heating process, equipment status, and process parameters.

3. The method for pre-adjusting width control parameters for a long-term unrolled specification according to claim 1, characterized in that, The classification system of specifications is established using cluster analysis, with steel grade, thickness, width, and lateral pressure as the key dimensions for stratification.

4. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The selection of the reference data includes the following priority order: Data within the same stratum > Data in adjacent stratums > Data across stratums; And conduct a comprehensive evaluation in conjunction with operating conditions; Operating conditions include rolling mileage and equipment status.

5. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The aforementioned pre-adjustment calculation is based on the layer difference between the current specification to be rolled and the reference specification. Through the constructed pre-adjustment amplitude calculation model, the set width control parameters are corrected to generate width control parameters suitable for specifications that have not been rolled for a long time.

6. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The parameter mapping and conversion rules include proportional adjustments to key control parameters to accommodate differences between different specifications.

7. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 6, characterized in that, The key control parameters include one or more of the following: vertical roll opening degree, rolling force setting value, tension setting value, and speed setting value.

8. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The expression for the pre-adjustment amplitude calculation model is as follows: ①； in: This indicates the latest width control parameters after a long period of non-rolling and specification adjustment. This indicates the original width control parameters for specifications that have not been rolled for a long time in the historical database; This indicates the optimal width control parameters for recently rolled specifications; Indicates the coefficient of difference between steel grade layers; Indicates the thickness difference coefficient between layers; Indicates the difference coefficient between width layers; Indicates the difference coefficient between lateral pressure layers; The smoothing coefficient represents the parameter adjustment and is a parameter that controls the pre-adjustment range.

9. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 8, characterized in that, Steel grade layer difference coefficient The value range is [0, 0.2]. When the span between the steel grade and the reference grade of the specification to be rolled is less than or equal to two grade levels, ; Thickness layer difference coefficient The value range is [0, 0.2], and the thickness layer difference coefficient is... There is a positive correlation between the difference and the target thickness of the strip; the larger the difference, the better. ; Width layer difference coefficient The value range is [0, 0.2], and the width layer difference coefficient is... There is a positive correlation between the difference and the target width of the strip; the larger the difference, the better. ; Lateral pressure layer difference coefficient The value range is [0, 0.2], and the width layer difference coefficient is... There is a positive correlation between the difference in the amount of side pressure on the strip and the amount of side pressure; the larger the difference, the greater the difference. .

10. The method for pre-adjusting width control parameters of a long-term unrolled specification according to claim 1, characterized in that, The feedback mechanism includes monitoring the actual width control of the first coil of strip and adjusting the pre-adjustment parameters of subsequent strips based on the width deviation value obtained from the monitoring.

Images