A method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanistic models

Abstract

This invention provides a method for predicting dynamic rolling force in cold continuous rolling mills based on industrial data and a mechanistic model, relating to the field of metallurgical rolling technology. The method includes the following steps: S1, constructing a dynamic rolling force mechanism model; S2, collecting actual production data of the process parameters in step S1 within a cold continuous rolling mill production line; S3, correcting the dynamic rolling force mechanism model in step S1 using industrial data; and S4, calculating the corrected values ​​of the dynamic rolling force for each stand and optimizing the rolling schedule. Based on the dynamic rolling force mechanism model, this method continuously corrects the mechanism model by collecting measured data of process parameters from actual cold continuous rolling processes, using a regression algorithm to improve the prediction accuracy of dynamic rolling force. Furthermore, it optimizes the rolling schedule to reduce fluctuations in dynamic rolling force during the rolling process, thereby improving rolling precision.

Description

Technical Field

[0001] This invention relates to the field of metallurgical rolling technology, and in particular to a method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models. Background Technology

[0002] Currently, with the increasing demand for high-precision strip products, ensuring product quality has become a top priority in the production process. Rolling force, as one of the most crucial process parameters, is fundamental for determining rolling power, verifying the strength of rolling mill equipment, setting rolling processes, and calculating critical rolling speeds. Therefore, accurate prediction of rolling force can effectively improve the dimensional accuracy of strip products and the stability of the rolling process.

[0003] Depending on the application, rolling force can be divided into steady-state rolling force and dynamic rolling force. Steady-state rolling force is often used to calculate the preset value of roll gap and rolling load distribution, and is also applied to automatic thickness control and optimal shape control. Dynamic rolling force is the change in rolling force caused by fluctuations in process parameters. The interaction between the dynamic changes in the rolling mill structure and the dynamic rolling force can cause self-excited vibration, leading to rolling mill vibration. Therefore, dynamic rolling force is often used to study rolling mill vibration problems, and vibration suppression measures can be found by analyzing the dynamic behavior of the rolling system. Currently, research on rolling force prediction and optimization methods mainly includes mechanism prediction methods, data prediction methods, and mechanism and data combined prediction methods. Mechanism prediction methods inevitably involve certain assumptions and simplifications, and may ignore the influence of some minor factors, resulting in insufficient rolling force prediction accuracy. Although data prediction methods can improve prediction accuracy, the constraints between process parameters are not clear, the interpretability of the prediction methods is poor, and they lack clear physical meaning and mathematical relationships, which limits the application of data prediction methods in actual production. The mechanism-data collaborative prediction method, while retaining the rigorous structure of the mechanism prediction method, compensates and corrects the mechanism prediction method through industrial data and advanced algorithms.

[0004] Domestic researchers have conducted extensive work on the prediction and application of rolling force during the rolling process. Chinese invention patent (publication number CN116197254A) discloses a method for predicting the rolling force mechanism in cold continuous rolling. This method considers the influence of process parameters such as roll radius and front and rear tensions, calculates the rolling force in the elastic deformation zone and plastic deformation zone, and then obtains the overall steady-state rolling force. Chinese invention patent (publication number CN111889524A) discloses a steady-state rolling force data prediction method based on machine learning algorithms. Using preprocessed production data as input, it predicts the steady-state rolling force through different machine learning algorithms and optimizes the algorithms, effectively improving the control level of the cold rolling model. Chinese invention patent (publication number CN112711867A) discloses a steady-state rolling force prediction method combining a mechanism model and a data model. This method is based on a steady-state rolling force mechanism model and uses a BP neural network data prediction model to correct and supplement the mechanism model, improving the prediction accuracy of the steady-state rolling force. Chinese invention patent (publication number CN106734194A) discloses a method for predicting and suppressing rolling mill vibration. This method combines a dynamic rolling force mechanism model with a rolling mill dynamic structure model, uses process parameters as independent variables, and predicts the critical rolling speed at which the system becomes unstable based on the Routh stability criterion, thereby improving the stability of the rolling process.

[0005] Among the aforementioned patents, the methods for predicting steady-state rolling force are relatively well-developed, having gone through three stages: mechanism-based, data-based, and mechanism-data-based collaborative prediction. However, the methods for predicting dynamic rolling force mainly focus on mechanism-based prediction, with no research on data-based and mechanism-data-based collaborative prediction. With the development of higher-strength and thinner strip products, the stress-strain state at the rolling interface is becoming increasingly complex, leading to a decrease in the prediction accuracy of dynamic rolling force mechanism models. This, in turn, affects the stability of the rolling process and the accuracy of feedback control, severely restricting the improvement of production efficiency and product quality. Therefore, there is an urgent need to research a dynamic rolling force prediction method that combines mechanism prediction with actual data. Summary of the Invention

[0006] The purpose of this invention is to provide a method for predicting dynamic rolling force in cold continuous rolling based on industrial data and a mechanism model. Based on the dynamic rolling force mechanism model, this method collects measured production data of process parameters in actual cold continuous rolling production, uses regression algorithms to continuously correct the mechanism model, improves the prediction accuracy of dynamic rolling force, and optimizes the rolling process to reduce the fluctuation of dynamic rolling force during the rolling process and thus improve the stability of the rolling process.

[0007] Specifically, the present invention provides a method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanistic models, which includes the following steps:

[0008] S1. Construct a dynamic rolling force mechanism model;

[0009] S2. Collect actual production data of the process parameters of the dynamic rolling force mechanism model in the cold continuous rolling production line;

[0010] S3. Correct the dynamic rolling force mechanism model in step S1 using industrial data, and calculate the corrected value of the dynamic rolling force. This includes the following sub-steps:

[0011] S31. Calculate the initial value of the dynamic rolling force when rolling the j-th steel coil of a certain strength grade using the dynamic rolling force mechanism model in step S1:

[0012]

[0013] in, For the process parameters of the dynamic rolling force mechanism model, F i The influence coefficient of the above process parameters on the dynamic rolling force, i = 1-5;

[0014] S32. Calculate the correction coefficient based on the (j-1)th and (j-2)th steel coils: For the (j-1)th and (j-2)th steel coils that have been rolled, based on the actual production data in step S2, the correction coefficient of the influence coefficient is obtained through measured dynamic rolling force and linear regression:

[0015]

[0016] Where, λ i D is the correction factor, and D is the constant term.

[0017] S33. Based on the correction coefficient calculated in step S32, the dynamic rolling force correction value for the j-th steel coil is obtained as follows:

[0018]

[0019] Where α is the weighting coefficient;

[0020] S34. Solve for the dynamic rolling force correction value for all steel coils under this strength grade;

[0021] S4. Based on step S3, solve for the dynamic rolling force under multiple strength grades.

[0022] Preferably, the dynamic rolling force mechanism model constructed in step S1 is shown in the following equation:

[0023]

[0024] in, h is the fluctuation amount of the roll gap velocity. c,var h represents the fluctuation of the roll gap. e,varσ represents the fluctuation in the thickness of the strip at the inlet. xe,var σ is the fluctuation of the tension at the inlet of the strip. xd,var This represents the fluctuation in the tension at the strip exit.

[0025] Preferably, step S1 specifically includes the following sub-steps:

[0026] S11. Based on the elastic-plastic deformation conditions of the strip, establish the relationship between the front and back tension, deformation resistance and rolling compressive stress of the strip.

[0027] S12. Considering the fluctuation of the roll gap and its fluctuation speed, calculate the expression for the inlet and outlet speed of the strip using the metal flow rate equation;

[0028] S13. Substitute the elastic-plastic deformation condition from step S1 into the force equilibrium equation of the strip micro-element, calculate the distribution expression of the rolling compressive stress along the strip inlet to outlet position, and integrate the rolling compressive stress along the contact arc length to obtain the expression of the steady-state rolling force:

[0029]

[0030] Where, p f (x) represents the rolling compressive stress distribution in the forward slip zone, p b (x) represents the rolling compressive stress distribution in the back slip zone, and B is the width of the strip. e x represents the inlet position of the strip. d x represents the exit position of the strip. n This is the neutral point position of the strip;

[0031] S14. Perform a first-order Taylor expansion on the steady-state rolling force in step S13 to obtain the expression for the dynamic rolling mechanism model.

[0032] Preferably, in step S34, steps S31-S33 are repeated to solve for the dynamic rolling force correction value of the (j+1)th steel coil:

[0033]

[0034] This continues until the dynamic rolling force correction value for all steel coils under that strength grade is obtained.

[0035] Preferably, in step S4, the correction factor for the first rolled steel coil under the new strength grade is obtained from a steel coil with a similar strength grade.

[0036] Preferably, the process further includes step S5, calculating the correction value of the dynamic rolling force for each stand and optimizing the rolling schedule.

[0037] Preferably, the specific steps for optimizing the rolling process in step S5 are as follows:

[0038] S51. With the goal of reducing the fluctuation range of dynamic rolling force in each stand during continuous rolling and improving the stability of the rolling process, based on the dynamic rolling force correction model, the correction value of the dynamic rolling force for each stand is determined by the dynamic fluctuation of the process parameters of each stand for the (j-1)th and (j-2)th coils:

[0039]

[0040] S52. Calculate the maximum dynamic rolling force difference for each stand:

[0041]

[0042] in, This represents the maximum value of the dynamic rolling force correction for each stand. This represents the minimum value of the dynamic rolling force correction for each stand.

[0043] S53. To optimize the rolling process, the optimization range is set so that the maximum dynamic rolling force difference of each stand is reduced and the sum of the reductions of the maximum dynamic rolling force difference of each stand is maximized. Within the optimization range, steps S51-S53 are repeated until the optimization objective is met, and the optimized rolling process is output.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0045] (1) This invention provides a method for predicting dynamic rolling force in cold continuous rolling based on the collaborative driving of industrial data and mechanism model. Compared with the existing dynamic rolling force mechanism prediction method, this method continuously corrects the mechanism model by introducing industrial data from actual production, making the prediction more accurate and the application range more extensive. It can be applied to the prediction of dynamic rolling force in different stands and different strength grades.

[0046] (2) The method of the present invention is based on the accurate prediction of the modified model, which can optimize the strip rolling process before production, effectively reduce the fluctuation of dynamic rolling force, that is, keep the rolling force stable, reduce the thickness difference of strip products along the rolling direction, improve the dimensional accuracy of strip products, and also improve the stability of the rolling process. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the overall process of the dynamic rolling force prediction method for cold continuous rolling based on industrial data and mechanism models of the present invention.

[0048] Figure 2 This is a schematic flowchart of the dynamic rolling force prediction method of the present invention;

[0049] Figure 3 This is a schematic flowchart of the dynamic rolling force optimization method of the present invention;

[0050] Figure 4 This is a schematic diagram of a cold continuous rolling mill in an embodiment of the present invention;

[0051] Figures 5a-5f These are schematic diagrams illustrating the dynamic rolling force prediction results for six continuously rolled steel coils in embodiments of the present invention. Figure 5a This is a schematic diagram showing the predicted dynamic rolling force of steel coil 1. Figure 5b This is a schematic diagram showing the predicted dynamic rolling force of steel coil 2. Figure 5c This is a schematic diagram showing the predicted dynamic rolling force of steel coil 3. Figure 5d This is a schematic diagram showing the predicted dynamic rolling force of steel coil 4. Figure 5e This is a schematic diagram showing the predicted dynamic rolling force of steel coil 5. Figure 5f This is a schematic diagram showing the predicted dynamic rolling force of steel coil 6.

[0052] Figures 6a-6f These are schematic diagrams illustrating the dynamic rolling force prediction error of six continuously rolled steel coils in embodiments of the present invention. Figure 6a This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 1. Figure 6b This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 2. Figure 6c This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 3. Figure 6d This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 4. Figure 6e This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 5. Figure 6f This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 6.

[0053] Figures 7a-7c This is a schematic diagram illustrating the dynamic rolling force prediction results of the first rolled steel coil under the new strength grade in an embodiment of the present invention. Figure 7a This is a schematic diagram illustrating the prediction results for intensity level 1CD61. Figure 7b This is a schematic diagram of the predicted intensity level M3A33. Figure 7c This is a diagram illustrating the comparison of prediction errors.

[0054] Figures 8a-8d A comparative diagram of the optimization of racks S1 to S4 in an embodiment of the present invention. Figure 8a A comparative diagram showing the optimization of the S1 rack. Figure 8b A comparative diagram showing the optimization of the S2 rack. Figure 8c A comparative diagram illustrating S3 rack optimization. Figure 8d A comparative diagram showing the optimization of the S4 rack. Detailed Implementation

[0055] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0056] This invention provides a method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanistic models, such as... Figure 2 As shown, it includes the following steps:

[0057] S1. The dynamic rolling force mechanism model is constructed as shown in the following equation:

[0058]

[0059] in, h is the fluctuation amount of the roll gap velocity. c,var h represents the fluctuation of the roll gap. e,var σ represents the fluctuation in the thickness of the strip at the inlet. xe,var σ is the fluctuation of the tension at the inlet of the strip. xd,var F represents the fluctuation in the tension at the strip exit. i The coefficient representing the influence of the above process parameters on the dynamic rolling force is i = 1-5.

[0060] Step S1 specifically includes the following sub-steps:

[0061] S11. Based on the elastic-plastic deformation conditions of the strip, establish the relationship between the front and back tension, deformation resistance and rolling compressive stress of the strip.

[0062] S12. Considering the fluctuation of the roll gap and its fluctuation speed, calculate the expression for the inlet and outlet speed of the strip using the metal flow rate equation.

[0063] S13. Substitute the elastic-plastic deformation condition from step S1 into the force equilibrium equation of the strip micro-element, calculate the distribution expression of the rolling compressive stress along the strip inlet to outlet position, and integrate the rolling compressive stress along the contact arc length to obtain the expression of the steady-state rolling force:

[0064]

[0065] Where, p f (x) represents the rolling compressive stress distribution in the forward slip zone, p b (x) represents the rolling compressive stress distribution in the back slip zone, and B is the width of the strip. e x represents the inlet position of the strip. d x represents the exit position of the strip. n This represents the neutral point position of the strip.

[0066] S14. Perform a first-order Taylor expansion on the steady-state rolling force in step S3 to obtain the expression for the dynamic rolling mechanism model, as shown in the following equation:

[0067]

[0068] in, h is the fluctuation amount of the roll gap velocity.c,var h represents the fluctuation of the roll gap. e,var σ represents the fluctuation in the thickness of the strip at the inlet. xe,var σ is the fluctuation of the tension at the inlet of the strip. xd,var F represents the fluctuation in the tension at the strip exit. i The coefficient representing the influence of the above process parameters on the dynamic rolling force is i = 1-5.

[0069] S2. Based on the actual rolling production parameters, collect the actual production data of the process parameters in step S1 in the cold continuous rolling production line.

[0070] S3. Correct the dynamic rolling force mechanism model in step S1 using industrial data, and calculate the corrected value of the dynamic rolling force. This includes the following sub-steps:

[0071] S31. Calculate the initial value of the dynamic rolling force when rolling the j-th steel coil of a certain strength grade using the dynamic rolling force mechanism model in step S1:

[0072]

[0073] in, These are the process parameters in step S1.

[0074] S32. Calculate the correction coefficient based on the (j-1)th and (j-2)th steel coils: For the (j-1)th and (j-2)th steel coils that have been rolled, based on the actual production data in step S2, the correction coefficient of the influence coefficient is obtained through measured dynamic rolling force and linear regression:

[0075]

[0076] Where, λ i The correction factor is D, which is a constant term representing the influence of process parameters not considered in the mechanism model on the mechanism model.

[0077] S33. Based on the correction coefficient calculated in step S32, the dynamic rolling force correction value for the j-th steel coil is obtained as follows:

[0078]

[0079] Where α is the weighting coefficient.

[0080] S34. Repeat steps S31-S33 to solve for the dynamic rolling force correction value of the (j+1)th steel coil:

[0081]

[0082] By continuously collecting production data in this step, the dynamic rolling force of the next steel coil is corrected and predicted. The dynamic rolling force of the next steel coil is then calculated sequentially to solve for the dynamic rolling force of all steel coils under a certain rolling intensity. This can significantly improve the prediction accuracy of the dynamic rolling force, thereby ensuring the accuracy and stability of rolling.

[0083] S4. For rolling at different strength grades, repeat steps S31-S34 based on step S3 to solve for the dynamic rolling force under multiple strength grades. In this step, for the first rolled coil under a new strength grade, its correction coefficient can be obtained from a coil with a similar strength grade. Then, the correction values ​​of the dynamic rolling force for each coil rolled sequentially are solved according to the above method. The above steps constitute the prediction process of the dynamic rolling force. In other embodiments, the entire rolling process can be further optimized based on the prediction of the dynamic rolling force as needed. That is, it also includes step S5, calculating the correction values ​​of the dynamic rolling force for each stand and optimizing the rolling process, thereby improving the stability of the entire rolling process.

[0084] Specifically, the steps for optimizing the rolling process in step S5 are as follows:

[0085] S51. With the goal of reducing the fluctuation range of dynamic rolling force in each stand during continuous rolling and improving the stability of the rolling process, based on the dynamic rolling force correction model, the correction value of the dynamic rolling force for each stand is determined by the dynamic fluctuation of the process parameters of each stand for the (j-1)th steel coil:

[0086]

[0087] In practical applications, the dynamic rolling force after correction for each stand is calculated sequentially using this formula.

[0088] S52. Calculate the maximum dynamic rolling force difference for each stand:

[0089]

[0090] in, This represents the maximum value of the dynamic rolling force correction for each stand. This represents the minimum dynamic rolling force correction value for each stand. The maximum dynamic rolling force difference for each stand is then calculated sequentially using this formula.

[0091] S53. The optimization objective is to minimize the maximum dynamic rolling force difference across all stands and maximize the sum of these reductions. An optimization range for the rolling schedule is set, and steps S51-S53 are repeated within this range until the optimization objective is met. The optimized rolling schedule is then output. The optimization objective in this step refers to reducing the dynamic rolling force at each stand. The sum of these reductions is considered optimal when the total reduction reaches its maximum. During the optimization process, correction values ​​for the dynamic rolling force at each stand are obtained to meet the optimization objective. These corrected dynamic rolling forces are then added to the new rolling schedule to increase rolling stability and improve rolling accuracy. Specific Implementation

[0093] In this embodiment, a cold rolling mill in a certain factory is used as an example. The dynamic rolling force of each stand is predicted sequentially using the above method, and the rolling process is further optimized based on the prediction of the dynamic rolling force.

[0094] The present invention provides a data- and mechanism-driven method for predicting and optimizing dynamic rolling force in cold continuous rolling mills, such as... Figure 2 and Figure 3 As shown, it includes:

[0095] S1. Different steel grades are assigned to strips with different strength levels, and similar steel grades represent approximately the same strength level. This step constructs a dynamic rolling force mechanism model based on relevant process parameters.

[0096] S2, such as Figure 4 As shown, process parameters in actual production are tested and collected to obtain actual production data of process parameters.

[0097] S3. Continuously revise the dynamic rolling force model using actual production data of process parameters, calculate and predict the dynamic rolling force, and continuously roll. Sequentially revise the model using actual production data to obtain the predicted dynamic rolling force for the next steel coil, thereby completing continuous rolling and improving rolling accuracy.

[0098] S4. Solve the dynamic rolling force of the steel coil under multiple strengths in sequence.

[0099] In this embodiment, six continuously rolled steel coils of grade M3A33 are selected. Figure 2 Based on the prediction method shown, the dynamic rolling force of the first stand is predicted and compared using both the modified model and the mechanistic model. The relevant comparison results are as follows: Figures 5a-5f As shown. Figures 5a-5f The dynamic rolling force prediction results for six continuously rolled steel coils in this embodiment are shown sequentially. Figure 5a The predicted dynamic rolling force of steel coil 1 is shown below. Figure 5b The predicted dynamic rolling force for steel coil 2 is shown below. Figure 5c The results show the predicted dynamic rolling force of steel coil 3. Figure 5d The predicted dynamic rolling force for steel coil 4 is shown below. Figure 5e The predicted dynamic rolling force of steel coil 5 is shown below. Figure 5f The results show the predicted dynamic rolling force for steel coil 6. From... Figures 5a-5f As can be seen, compared with a single mechanism model, the dynamic rolling force calculated by the modified model proposed in this invention, which combines actual production data, is significantly closer to the measured value of the dynamic rolling force, demonstrating better prediction performance.

[0100] Then, the prediction errors of the modified model and the mechanistic model at each sampling point are calculated, the probability density function of the prediction error is obtained, and the results are compared. Figures 6a-6f As shown. Figures 6a-6f The following diagrams illustrate the dynamic rolling force prediction errors of six continuously rolled steel coils in this embodiment. Figure 6a This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 1. Figure 6b This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 2. Figure 6c This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 3. Figure 6d This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 4. Figure 6e This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 5. Figure 6f This is a schematic diagram illustrating the error in the dynamic rolling force prediction results for steel coil 6; from... Figures 6a-6f It can be seen that compared with the mechanism model, the modified model using the method of the present invention has a smaller prediction error range, no significant deviation, and a higher probability of smaller prediction error. Thus, it can be seen that the dynamic rolling force predicted after the mechanism model is modified by the present invention in combination with actual data is more accurate.

[0101] For the first rolled coil of a new steel grade, a correction factor obtained from a similar steel grade is used instead. Taking a coil of steel grade M3A30 as an example, the correction values ​​for dynamic rolling force are calculated using correction factors obtained from steel grades M3A33 and 1CD61, respectively. Figures 7a-7c As shown. Among them, Figure 7a This is a schematic diagram illustrating the prediction results for intensity level 1CD61. Figure 7b This is a schematic diagram of the predicted intensity level M3A33. Figure 7c This is a diagram illustrating the comparison of prediction errors. From... Figures 7a-7c As can be seen, the dynamic rolling force prediction effect obtained by using the correction coefficient of steel grade M3A33, which is similar to M3A30, is better, which further broadens the application scope of the correction model proposed in this invention, improves the rolling stability, and ensures the application effect.

[0102] In this embodiment, the rolling process for a steel coil of grade M3A33 is shown in Table 1. The fifth stand is a leveling mill. During the rolling process, most of the strip reduction has already been completed in the first four stands; the reduction in the fifth stand is very small, mainly to improve the strip shape. Therefore, this embodiment further... Figure 3 The optimization process shown optimizes the rolling schedule for the first four stands. The rolling schedule and the maximum dynamic rolling force difference before and after optimization are shown in Table 2. As can be seen from Table 2, the optimization of the rolling schedule effectively reduces the maximum dynamic rolling force difference for the first four stands, making the rolling force more stable and indirectly improving the stability of the entire rolling mill system.

[0103] Figures 8a-8d This diagram illustrates the optimization comparison between racks S1 and S4 in this embodiment. Figure 8a A comparative diagram showing the optimization of the S1 rack. Figure 8b A comparative diagram showing the optimization of the S2 rack. Figure 8c A comparative diagram illustrating S3 rack optimization. Figure 8d A comparative diagram showing the optimization of the S4 rack. Figures 8a-8d The optimization effect is more clearly reflected. Compared to the first two stands, S1 and S2, the rolling forces of the third and fourth stands, S3 and S4, are more stable, with less drastic changes. Therefore, the fluctuation range of dynamic rolling force is inherently small, limiting the optimization potential. After optimization, the fluctuation range of dynamic rolling force in the first and second stands is significantly reduced. While the fluctuation range of the third and fourth stands is also reduced, it is not drastic, as they are already relatively stable. As can be seen from the table and figure, the optimization greatly improves the rolling stability of each stand.

[0104] Table 1 Rolling Procedure

[0105]

[0106] Table 2 Optimization Results

[0107]

[0108] As can be seen from the above embodiments, the method of the present invention, based on the accurate prediction of the modified model, can optimize the strip rolling process before production, effectively reduce the fluctuation of dynamic rolling force, that is, keep the rolling force stable, reduce the thickness difference of the strip product along the rolling direction, improve the dimensional accuracy of the strip product, and also improve the stability of the rolling process.

[0109] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanistic models, characterized in that: It includes the following steps: S1. Construct a dynamic rolling force mechanism model; S2. Collect actual production data of the process parameters of the dynamic rolling force mechanism model in the cold continuous rolling production line; S3. Correct the dynamic rolling force mechanism model in step S1 using industrial data, and calculate the corrected value of the dynamic rolling force. This includes the following sub-steps: S31. Calculate the initial value of the dynamic rolling force when rolling the j-th steel coil of a certain strength grade using the dynamic rolling force mechanism model in step S1: ； in, For the process parameters of the dynamic rolling force mechanism model, This is the influence coefficient of process parameters on dynamic rolling force. ; S32. Calculate correction coefficients based on the (j-1)th and (j-2)th steel coils: For the (j-1)th and (j-2)th steel coils that have been rolled, based on the actual production data in step S2, obtain the correction coefficients of the influence coefficients through measured dynamic rolling force and linear regression. ； in, For correction factor, For constant terms; S33. Based on the correction coefficients calculated in step S32, the dynamic rolling force correction model for the j-th steel coil is obtained as follows: ； in, These are the weighting coefficients; S34. Solve for the dynamic rolling force correction value for all steel coils under this strength grade; S4. Based on step S3, solve for the dynamic rolling force under multiple strength grades.

2. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 1, characterized in that: The dynamic rolling force mechanism model in step S1 is shown in the following equation: ； in, This represents the fluctuation amount of the roll gap velocity. This represents the fluctuation in the roll gap. This represents the fluctuation in the thickness of the strip at the inlet. This represents the fluctuation in the tension at the strip inlet. This represents the fluctuation in the tension at the strip exit.

3. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 1, characterized in that: Step S1 specifically includes the following sub-steps: S11. Based on the elastic-plastic deformation conditions of the strip, establish the relationship between the front and back tension, deformation resistance and rolling compressive stress of the strip. S12. Based on the fluctuation of the roll gap and its fluctuation speed, an expression for calculating the inlet and outlet speed of the strip using the metal flow rate equation; S13. Substitute the elastic-plastic deformation condition from step S11 into the force balance equation of the strip micro-element, calculate the distribution expression of the rolling compressive stress along the strip inlet to outlet position, and integrate the rolling compressive stress along the contact arc length to obtain the expression of the steady-state rolling force: ； in, The rolling compressive stress distribution in the front slip zone, The rolling compressive stress distribution in the back slip zone, The width of the strip. This is the inlet position of the strip. x represents the exit position of the strip. n This is the neutral point position of the strip; S14. Perform a first-order Taylor expansion on the steady-state rolling force in step S13 to obtain the dynamic rolling force mechanism model.

4. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 1, characterized in that: In step S34, repeat steps S31-S33 to solve the dynamic rolling force correction model for the (j+1)th steel coil: ； This continues until the dynamic rolling force correction value for all steel coils under that strength grade is obtained.

5. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 4, characterized in that: In step S4, for the first rolled steel coil under the new strength grade, the correction factor is obtained from steel coils with similar strength grades.

6. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 1, characterized in that: It also includes step S5, calculating the correction value of the dynamic rolling force for each stand and optimizing the rolling procedure.

7. The method for predicting dynamic rolling force in cold continuous rolling based on industrial data and mechanism models according to claim 6, characterized in that: The specific steps for optimizing the rolling process in step S5 are as follows: S51. Based on the dynamic rolling force correction model, the correction model for the dynamic rolling force of each stand is determined using the actual process parameters of each stand during the rolling of the (j-1)th and (j-2)th steel coils as follows: ； S52. Calculate the maximum dynamic rolling force difference for each stand based on the modified model of dynamic rolling force for each stand: ； in, This represents the maximum value of the dynamic rolling force correction for each stand. This represents the minimum value of the dynamic rolling force correction for each stand; S53. To optimize the rolling process, the optimization range is set so that the maximum dynamic rolling force difference of each stand is reduced and the sum of the reductions of the maximum dynamic rolling force difference of each stand is maximized. Within the optimization range, steps S51-S53 are repeated until the optimization objective is met, and the optimized rolling process is output.

Images