A Dynamic Modeling and Angle Identification Method Based on the Crossover of the Roll System of a Hot Rolling Mill

Abstract

The present invention provides a dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill, which relates to the technical field of identifying the crossing angle of the roll system of a hot rolling mill, and includes the following steps: establishing an axial dynamic model according to the correlation relationship between the axial forces of the work rolls and backup rolls, the rolling force and the crossing angle between the roll systems; building an identification system for the dynamic parameter identification model considering the roll system crossing angle, and obtaining the crossing angle between the work rolls and backup rolls of the rolling mill, the crossing angle between the work rolls and the vertical direction of the strip movement direction, and the crossing angle between the upper and lower work rolls from the matrix of parameters to be estimated. The present invention fully exploits the relevant data of the actual production of the hot rolling mill to identify the crossing angle between the roll systems of the rolling mill, and on this basis, it can further analyze the influence degree of the strip on the crossing angle between the roll systems under different rolling widths, thicknesses, materials and different rolling speeds, etc., and dynamically track the change of the crossing angle between the roll systems of the hot rolling mill.

Description

Technical Field

[0001] The present invention relates to the technical field of identifying the crossing angle of the roll system of a hot rolling mill, and in particular, to a dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill. Background Art

[0002] For a rolling mill, especially a hot rolling mill, on the one hand, the crossing between roll systems has an important impact on the stiffness and stiffness difference between the operating side and the driving side of the rolling mill. Especially when rolling strip steel at the tail stand or when rolling thinner strip steel, the control requirements for the roll gap are more stringent. However, due to the crossing between roll systems, the positions of the work rolls and backup rolls of the rolling mill on the operating side and the driving side will change during the rolling process, and ultimately, it will indirectly affect the quality of the strip steel by affecting the stiffness on both sides. Therefore, it is necessary to measure and evaluate the degree of crossing between roll systems. On the other hand, the crossing between roll systems is also one of the main reasons for the deviation and waviness of the strip steel. For this reason, the rolling mill needs to be shut down every once in a while to measure the crossing angle between the rolls, and adjust the spatial attitude of the roll system of the rolling mill according to the measurement results of the crossing angle to ensure the good quality of the final strip steel.

[0003] Currently, there are two methods that enterprises can use to measure the degree of crossing between the roll systems of a rolling mill. One is to use a laser tracker to measure the positions of the left and right lining plates on the operating side and the left and right lining plates on the driving side of the rolling mill, and characterize the degree of crossing between the roll systems by further calculating their relative position relationships. The other is to install piezoelectric sensors on the left and right lining plates on the operating side of the rolling mill and install piezoelectric sensors on the left and right lining plates on the driving side of the rolling mill. By measuring the pressure values of the left and right lining plates on the operating side and the left and right lining plates on the driving side during the rolling process of the rolling mill, further calculate the deformation amounts of the left and right lining plates on the operating side and the left and right lining plates on the driving side, and characterize the degree of crossing between the roll systems by the relative position relationships of the deformation amounts.

[0004] However, the existing methods for measuring the degree of crossing between the roll systems of a rolling mill have certain disadvantages. For the method using a laser tracker, the measurement can only be carried out after the production line is shut down, and the measurement result is the static value of the rolling mill, which has the technical problems of affecting normal production and inaccurate measurement results. For the method of installing piezoelectric sensors, on the one hand, the influence of the initial relative position between the roll systems is ignored, and on the other hand, the serious wear of the lining plates during long-term rolling of the rolling mill is not taken into account, which will cause the service life of the piezoelectric sensors to be difficult to guarantee. Summary of the Invention

[0005] In view of this, the object of the present invention is to propose a dynamic modeling and angle identification method based on the crossing of the roll system of a hot rolling mill, which can reflect the crossing angle between the rolls during the dynamic rolling process of the rolling mill under different rolling conditions, improve the accuracy of obtaining the crossing angle between the rolls of the rolling mill, and at the same time, neither shutdown treatment nor special test experiments are required during the identification process, reducing the working time of personnel. In addition, this method can obtain the crossing angle between the rolls under the dynamic behavior of the rolling mill without calibrating the initial relative position between the rolls of the rolling mill before rolling, avoiding the problem of low service life of the piezoelectric sensors installed on the left and right lining plates on the operating side and the left and right lining plates on the drive side, and at the same time, more crossing angles between the work rolls and the backup rolls can be obtained. To solve the technical problem that the existing methods cannot measure the dynamic numerical values of the production line.

[0006] The technical means adopted by the present invention are as follows:

[0007] A dynamic modeling and angle identification method based on the crossing of the roll system of a hot rolling mill, comprising the following steps:

[0008] Vibration sensors are installed axially on the work rolls and backup rolls of the rolling mill, and the vibration data of the axial direction of the roll system of the rolling mill during the actual production process of the hot rolling production line are collected through the vibration sensors;

[0009] Collect the pre-treatment rolling force data of the operating side and the drive side of the rolling mill collected during the actual production process of the rolling mill system;

[0010] Screen out the axial vibration data of the roll system of the rolling mill and the rolling force data on both sides of the rolling mill when rolling the same strip, and perform processing to obtain the processed vibration data and the processed rolling force data;

[0011] According to the relationship between the crossing angle between the work roll and the backup roll of the rolling mill, the crossing angle between the work roll and the vertical direction of the forward direction of the strip, the axial forces of the work roll and the backup roll, and the rolling force, establish a correlation relational expression between the axial forces of the work roll and the backup roll, the rolling force, and the crossing angle between the rolls;

[0012] According to the correlation relational expression between the axial forces of the work roll and the backup roll, the rolling force, and the crossing angle between the rolls, establish an axial dynamic model;

[0013] Convert the axial dynamic model into corresponding state equations and control equations according to the state space model conversion principle;

[0014] Solve the state equations and control equations, further calculate the dynamic response prediction model using the obtained parameter matrix, and further obtain the loss function of the state space model of the rolling mill in this rolling state according to the dynamic response prediction model and the processed vibration data;

[0015] Build an identification system for identifying the axial dynamic parameters of the roll system considering the crossing angle, and optimize the loss function through the identification system to obtain the parameter matrix to be estimated under the minimum loss function of the rolling mill state space model;

[0016] Obtain the crossing angle between the work roll and the backup roll of the rolling mill, the crossing angle between the work roll and the vertical direction of the strip movement direction, and the crossing angle between the upper and lower work rolls according to the parameter matrix to be estimated.

[0017] Furthermore, process the axial vibration data of the rolling mill roll system, the rolling force data and vibration data on both sides of the rolling mill, including the following steps:

[0018] Perform Fourier transform and DC component removal processing on the measured rolling force data and vibration data to obtain a new signal matrix;

[0019] Set an adaptive filter, perform dot product processing on the adaptive filter and the new signal matrix to obtain the filtered frequency-domain signal and time-domain signal;

[0020] Perform inverse Fourier transform on the filtered frequency-domain signal to obtain the processed rolling force data and processed vibration data after filtering;

[0021] Perform detrending processing on the processed rolling force data after filtering to obtain the processed rolling force data.

[0022] Furthermore, the processing formula for the processed rolling force is:

[0023] P1 = P′ - (α + βn′T), n′ = 1, 2, …, N………(1)

[0024] Where: P1 is the processed rolling force, P′ is the rolling force before processing, T is the sampling period, and α and β are the fitting coefficients of the rolling force before processing.

[0025] Furthermore, the rolling force data of the rolling mill is equal to the sum of the rolling forces on both sides of the rolling mill, and the rolling force formula of the rolling mill is:

[0026] P = P D + P O ………(2)

[0027] Where: P is the rolling force fluctuation of the rolling mill, P D is the rolling force fluctuation on the drive side of the rolling mill, P O is the rolling force fluctuation on the operating side of the rolling mill.

[0028] Furthermore, the vibration measuring sensor is one of an acceleration sensor, a velocity sensor or a displacement sensor.

[0029] Furthermore, the acquisition frequency of the rolling force is set to be the same as that of the vibration data.

[0030] Furthermore, the relational expression between the axial force and the rolling force of the work roll caused by the crossing of the work roll and the backup roll is:

[0031] F b = 0.06×(v ω ) 0.43 ×P×tanθ1×f1 + D………(3)

[0032] Where: F b is the axial force of the work roll due to the crossing of the work roll and the backup roll, v ω is the linear velocity of the work roll surface in the contact zone; P is the rolling force of the rolling mill, θ1 is the crossing angle between the work roll and the backup roll, f1 is the influence coefficient of the friction between the backup roll and the work roll on the axial force, and D is the axial force jump value at the zero point of the crossing angle, approximately 8 - 10 KN;

[0033] The relational expression between the axial force generated by the backup roll and the rolling force is:

[0034] F c = 0.06×(v ω ) 0.43 ×P×sinθ1×f1 + D………(4)

[0035] Where: F c is the axial force of the backup roll due to the crossing of the work roll and the backup roll;

[0036] The relational expression between the axial force of the work roll and the rolling force is:

[0037]

[0038] Where: F is the axial force of the work roll, P is the rolling force of the rolling mill, θ2 is the crossing angle between the work roll and the vertical direction of the strip movement, r is the roll neck ratio of the two rolls, and f is the friction coefficient of the work roll generating friction;

[0039] The axial force of the work roll caused by the crossing in the vertical direction between the work roll and the strip movement direction is:

[0040] F a = F - F b ………(6)

[0041] Where: F a is the axial force of the work roll due to the crossing in the vertical direction between the work roll and the strip movement direction.

[0042] Furthermore, the vibration of the axial dynamics model is the axial vibration of the roll and the strip, and the exciting force is the axial force fluctuation of the roll system.

[0043] Further, the loss function of the state space model is one of the sum of squares of the differences between the predicted acceleration and the actual acceleration of each data node, the sum of squares of the differences between the predicted velocity and the actual velocity of each data node, and the sum of squares of the differences between the predicted displacement and the actual displacement of each data node, and the actual acceleration or velocity or displacement is the processed vibration data.

[0044] Further, the formula for the loss function of the state space model of the roll system is:

[0045] Q m = Q 上支撑辊 + Q 上工作辊 + Q 下工作辊 + Q 下支撑辊 ………(7)

[0046] Where: Q m is the loss function of the state space model of the roll system, Q 上支撑辊 is the loss function of the state space model of the upper backup roll, Q 上工作辊 is the loss function of the state space model of the upper work roll, Q 下工作辊 is the loss function of the state space model of the lower work roll, Q 下支撑辊 is the loss function of the state space model of the lower backup roll.

[0047] Compared with the prior art, the present invention has the following advantages:

[0048] The original method of using a laser tracker to measure the crossing angle of the roll system measures the state in a static state, while this method is a dynamic measurement. Compared with the method of using a laser tracker to measure the crossing angle of the roll system, the present invention can reflect the crossing angle between the rolls during the dynamic rolling process of the rolling mill under different rolling states (such as different rolling widths, different rolling thicknesses, different rolling materials, and different rolling speeds), which can improve the accuracy of obtaining the crossing angle between the rolls of the rolling mill. At the same time, there is no need to stop the machine for processing during the identification process;

[0049] Compared with the method of using piezoelectric sensors to solve the deformation of strain gauges and then solve the crossing angle between the rolls, the method of the present invention solves the problem of needing to calibrate the initial relative position between the rolls of the rolling mill before rolling, avoids the problem of low service life of the piezoelectric sensors installed on the left and right liner plates on the operating side and the left and right liner plates on the drive side, and can also obtain more crossing angles between the work rolls and the backup rolls. Description of the Drawings

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the attached drawings required for the description of the embodiments or the prior art. Obviously, the attached drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other attached drawings can also be obtained based on these attached drawings.

[0051] Figure 1 This is the flowchart of the method of the present invention.

[0052] Figure 2 This is the flowchart of the data preprocessing of the present invention.

[0053] Figure 3 This is the diagram of the axial dynamics model of the rolling mill of the present invention.

[0054] Figure 4 This is the diagram of the cross type of the roll system of the rolling mill of the present invention. Detailed implementation manners

[0055] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the attached drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0056] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above attached drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily need to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0057] As Figure 1 shown, a dynamic modeling and angle identification method based on the cross of the roll system of a hot rolling mill of the present invention specifically includes the following steps:

[0058] Step 1: Install vibration sensors axially on the working rolls and backup rolls of the rolling mill to collect the vibration data of the axial direction of the roll system of the rolling mill during the actual production process of the hot rolling production line (the sensor type can be an acceleration sensor, a velocity sensor or a displacement sensor);

[0059] Step 2: Collect the rolling force data on the operator side and the drive side of the rolling mill system collected during the actual production process (it should be noted that the acquisition frequency of the rolling force should be set the same as that of the vibration data).

[0060] Step 3: Screen out the axial vibration data of the rolling mill roll system and the rolling force data on both sides of the rolling mill when rolling the same strip, and perform processing to obtain the processed axial vibration data of the rolling mill roll system and the rolling force data of the rolling mill. The specific processing methods are as follows:

[0061] a. Processing flow of the axial vibration data of the rolling mill roll system: (1) Perform Fourier transform and DC component removal processing on the measured axial vibration data of the rolling mill roll system to obtain a new roll system vibration signal matrix; (2) Set its adaptive filter according to the characteristics of the axial vibration data of the rolling mill roll system, and perform dot product processing with the new roll system vibration signal matrix; (3) Perform inverse Fourier transform on the frequency domain signal of the filtered axial vibration of the roll system to obtain the filtered axial vibration data of the rolling mill roll system;

[0062] b. Processing flow of the rolling force data of the rolling mill: (1) Perform Fourier transform and DC component removal processing on the measured rolling force data on the operator side and the drive side of the rolling mill respectively to obtain a new rolling force data signal matrix on both sides of the rolling mill; (2) Set its adaptive filter according to the characteristics of the rolling force data on the operator side and the drive side of the rolling mill, and perform dot product processing with the new rolling force data signal matrix on both sides of the rolling mill respectively; (3) Perform inverse Fourier transform on the frequency domain signals of the rolling forces on both sides of the rolling mill after filtering respectively to obtain the filtered rolling force data on both sides of the rolling mill; (4) Perform detrending processing on the filtered rolling force data on both sides of the rolling mill according to Equation (1); (5) Calculate the rolling force fluctuation of this rolling mill according to Equation (2).

[0063] Step 4: Considering the relationship between the crossing angles of the work roll and the backup roll of the rolling mill, the crossing angle between the work roll and the vertical direction of the strip forward direction, and the axial forces and rolling force between the work roll and the backup roll, obtain the correlation relational expressions between the axial forces of the work roll and the backup roll, the rolling force and the crossing angle between the roll systems. The specific correlation relationships are shown in Equations (3) to (6);

[0064] Step 5: Considering that the crossing between the roll systems of the rolling mill includes the crossing angle between the work roll and the backup roll, the crossing angle between the work roll and the vertical direction of the strip forward direction, and the crossing angle between the work rolls, the present invention establishes an axial dynamics model considering the roll system crossing of the rolling mill, as Figure 3 shown, and write the corresponding dynamic equations, as shown in Equation (8):

[0065]

[0066] Further, let e1 = 0.06×(vω上 ) 0.42 ×tanθ 1上 ×f1′, e2 = 0.06×(v ω下 ) 0.42 ×tanθ 1下 ×f1″, where f1′ is the influence coefficient of the friction between the upper backup roll and the upper work roll on the axial force, f1″ is the influence coefficient of the friction between the lower backup roll and the lower work roll on the axial force, f′ is the friction coefficient that generates frictional force on the upper work roll, f″ is the friction coefficient that generates frictional force on the lower work roll. Dividing both sides of Equation (8) by e1 or e2 or e3 or e4 or e5 correspondingly gives a new dynamic equation as shown in Equation (9):

[0067]

[0068] In the formula:

[0069] Simplifying Equation (9) gives the motion differential equation of the system as shown in Equation (10):

[0070] In the formula:

[0071] C1 = a×M + b×K, where a and b are coefficients related to the mass and stiffness of the system respectively;

[0072] Step 6: Convert the established axial dynamic model considering the cross of the rolling mill roll system into the corresponding state equation and control equation according to the state space model conversion principle to describe the relationship between the system input data and output data, specifically including:

[0073] a. Establish a state space model as shown in Equation (11):

[0074]

[0075] In the formula: x(t) is the state variable of the state space model; u(t) is the input of the axial force fluctuation of the rolling mill roll system at time t, N; y(t) is the actual output of the axial acceleration (or velocity, displacement) of the rolling mill roll system at time t, m / s 2 (or m / s, m); is the perturbation matrix; e1(t) and e2(t) are the error values of the state space model at time t; A, B, C, D are the parameter matrices to be estimated of the state model, and the elements in the parameter matrix are parameters containing the cross angle between the roll systems. Obtaining this parameter matrix can further solve for the cross angle between the roll systems;

[0076] b. Obtain the parameter matrices A, B, C, and D to be estimated of the model through the subspace algorithm, and obtain their expressions as shown in Equation (12):

[0077]

[0078] In the formula: 0 is a 5th-order zero matrix, and I is a 5th-order identity matrix;

[0079] Step 7: According to the state equation and control equation established by the state space model, solve the prediction model of the axial dynamic response of the rolling mill roll system under the action of the axial excitation force of the roll system (this response can be the displacement, velocity, or acceleration response of the dynamic model under the action of the axial excitation force of the roll system), and further solve to obtain the loss function of the state space model of the rolling mill in this rolling state. The specific process is as follows:

[0080] a. Obtain the expression of the transfer function R between the input u(t) and the output y(t) in the measured state space model, as shown in Equation (13):

[0081]

[0082] In the formula: is the spatial frequency, m -1 ; I is the identity matrix;

[0083] b. After obtaining the transfer function of the rolling mill state space model, the axial force fluctuation input u(t) of the rolling mill roll system can be used to predict the output of the axial dynamic response of the rolling mill The specific prediction model is as shown in Equation (14):

[0084]

[0085] In the formula: is the predicted output of the axial dynamic response, and R(t) is the transfer function of the state space model;

[0086] c. Define the loss function of the state space model. When the loss function of the roll system state space model reaches the minimum value, the parameter matrices A, B, C, and D of the rolling mill state space model can be calculated, and then the cross angle between the rolling mill roll systems can be obtained. The expression of the loss function of the state space model is as shown in Equation (15), where the loss function of the rolling mill roll system state space model is calculated by Equation (7):

[0087]

[0088] In the formula, N is the total number of data samples participating in the calculation of the loss function;

[0089] Step 8: Build an identification system for the dynamic parameter identification model considering the roll crossing angle, optimize the loss function of the state space model established for the rolling mill, obtain the parameter matrices A, B, C, and D to be estimated under the minimum loss function of the rolling mill state space model. Further, obtain the crossing angles between the work rolls and backup rolls of the rolling mill, the crossing angles in the vertical direction between the work rolls and the moving direction of the strip, and the crossing angles between the upper and lower work rolls. The construction of the identification system is mainly based on the particle swarm optimization algorithm, and the specific construction process is as follows:

[0090] a. Conduct parameter initialization settings, specifically including: the number of particle swarm parameters (L), the particle population size (M), the maximum number of iterations (T), the weight learning factors c1 and c2, the defined range of the adaptive inertia weight coefficient [ω max， , ω min , the defined range of the parameters to be identified [X min X max , the defined range of the particle optimization speed [v min v max , etc.;

[0091] b. Randomly initialize the positions and speeds of each particle swarm, and set this position as the initial local optimal position and global optimal position. Calculate the fitness value (i.e., the loss function value) of the global optimal position of the particle at this time. The formula for randomly initializing the particle position is shown in Equation (16), and the formula for randomly initializing the particle speed is shown in Equation (17):

[0092] X = X min +(X max -X min )×random………(16)

[0093] v = v min +(v max -v min )×random………(17)

[0094] In the formula: random is a random number between 0 and 1, X is the initialized population particle position, and v is the initialized population particle speed;

[0095] c. To enhance the global search ability of the particle swarm optimization algorithm in the initial stage and its ability to quickly converge to the global optimum in the later stage, optimize and update the adaptive inertia weight coefficient. The calculation expression of the adaptive inertia weight coefficient ω(i) is shown in Equation (18), and optimize and update the weight learning factors. The specific update method is shown in Equation (19):

[0096]

[0097] In the formula: ω is the current inertia weight coefficient, ω maxis the maximum inertia weight coefficient, ω min is the minimum inertia weight coefficient, s T is the current number of iteration steps, and T is the maximum number of iterations;

[0098]

[0099] In the formula: is the optimized value of the weight learning factor, c 1，f is the final value of iteration of c1, c 2，f is the final value of iteration of c2, s T is the number of iteration steps;

[0100] d. During each iteration, when considering that the particle velocity is small, the particle will wander too much in the local range. When the particle velocity is large, the particle may converge to the local minimum prematurely. Therefore, to ensure the accuracy and efficiency of the algorithm, the velocity of the particle is updated as follows, specifically as shown in Equation (20):

[0101]

[0102] In the formula: is the new velocity value after the particle velocity is updated, X G is the current optimal individual position, and BestS is the current global optimal individual position;

[0103] e. Using the new velocity value obtained by updating the particle according to Equation (20), the position of the particle is updated, and the fitness value of each updated particle is evaluated. The fitness value of each particle is compared with its passed local optimal position. If it is better, the position of the particle is used as the new local optimal position. Otherwise, it is not updated. After traversing all particles and updating the local optimal position, the new local optimal position is compared with the global optimal position. If it is better, it is used as the new global optimal position. Otherwise, it is not updated. The position update formula is as shown in Equation (21):

[0104]

[0105] In the formula: is the position value after the particle position is updated;

[0106] f. Iterate step by step until the loss function Q of the roll system state space model m converges to the minimum value, or the number of iteration steps reaches the maximum number of iterations set by the system, and the identification process ends;

[0107] g. Furthermore, we can obtain the crossing angle between the work roll and the backup roll of the rolling mill, as well as the crossing angles between the upper and lower work rolls and the vertical direction of the strip movement direction. However, the crossing angle between the upper and lower work rolls needs to be further calculated, and the calculation method is shown in Equation (22):

[0108]

[0109] In the formula: θ is the crossing angle between the upper and lower work rolls, θ 上 is the crossing angle between the upper work roll and the vertical direction of the strip movement direction, θ 下 is the crossing angle between the lower work roll and the vertical direction of the strip movement direction;

[0110] Step Nine: Considering that the crossing angles between the rolls of the hot rolling line mill identified by the identification system may have errors due to various factors, we divide the axial vibration data of rolling the same strip into three groups of data, identify each group of data separately, and average the results of the three groups of data to obtain a more accurate crossing angle between the rolls of the rolling mill. The specific formula is shown in Equation (23):

[0111]

[0112] In the formula: θ is the crossing angle between the upper and lower work rolls, θ′, θ″, and θ″′ are the crossing angles between the upper and lower work rolls identified from the three groups of data obtained by dividing the same strip, θ 1上 is the crossing angle between the work roll and the backup roll of the upper roll system, θ 1上 ′, θ 1上 ″, θ 1上 ″′ are the crossing angles between the work roll and the backup roll of the upper roll system identified from the three groups of data obtained by dividing the same strip, θ 1下 is the crossing angle between the work roll and the backup roll of the lower roll system, θ 1下 ′, θ 1下 ″, θ 1下 ″′ are the crossing angles between the work roll and the backup roll of the lower roll system identified from the three groups of data obtained by dividing the same strip;

[0113] Furthermore, organize and summarize the rolling performance, and identify and analyze the crossing angles between the rolls of the rolling mill in cases such as rolling strips with the same rolling speed, the same material, the same thickness, and different strip widths, rolling strips with the same rolling speed, different materials, the same thickness, and the same strip width, rolling strips with the same rolling speed, the same material, different thicknesses, and the same strip width, and rolling strips with different rolling speeds, the same material, the same thickness, and the same strip width;

[0114] Furthermore, by analyzing the identification situation when the rolling mill rolls the easily vibrating strip and there are quality defects in the strip product, a threshold value of the cross angle between the roll systems can be summarized for the rolling process of the production line, providing a reasonable support basis for the adjustment of the rolling mill equipment on the production line.

[0115] Furthermore, by completing the identification process of the cross angle between the roll systems of the rolling mill and obtaining the cross angle between the roll systems, it is convenient for the hot rolling production line to adjust the spatial attitude of the roll system of the rolling mill, ensuring the good quality of the final strip. It should be noted finally that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill, characterized in that The method includes the following steps: Axial vibration sensors are installed axially on the work rolls and backup rolls of the rolling mill, and the axial vibration data of the roll system of the rolling mill during the actual production process of the hot rolling production line is collected through the vibration sensors; Collect the pre - processed rolling force data on the operator side and the drive side of the rolling mill collected during the actual production process of the rolling mill system; Filter out the axial vibration data of the roll system of the rolling mill and the rolling force data on both sides of the rolling mill when rolling the same strip, and perform processing to obtain the processed vibration data and the processed rolling force data; According to the crossing angles between the work roll and the backup roll of the rolling mill, the crossing angles between the work roll and the vertical direction of the strip forward direction, and the relationships among the axial forces and the rolling force of the work roll and the backup roll, establish the correlation relational expressions between the axial forces of the work roll and the backup roll, the rolling force, and the crossing angles between the roll systems; Establish an axial dynamic model according to the correlation relational expressions between the axial forces of the work roll and the backup roll, the rolling force, and the crossing angles between the roll systems; Convert the axial dynamic model into corresponding state equations and control equations according to the state - space model conversion principle; Solve the state equations and control equations, further calculate the dynamic response prediction model by using the obtained parameter matrix, and further obtain the loss function of the state - space model of the rolling mill in this rolling state according to the dynamic response prediction model and the processed vibration data; Build an identification system for the axial dynamic parameter identification model considering the roll crossing angle, and optimize the loss function through the identification system to obtain the parameter matrix to be estimated under the minimum loss function of the state - space model of the rolling mill; Obtain the crossing angles between the work roll and the backup roll of the rolling mill, the crossing angles between the work roll and the vertical direction of the strip movement direction, and the crossing angles between the upper and lower work rolls according to the parameter matrix to be estimated; 2. The dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill according to claim 1, characterized in that The processing of the axial vibration data of the roll system of the rolling mill, the rolling force data on both sides of the rolling mill, and the vibration data includes the following steps: Perform Fourier transform and DC - component removal processing on the measured rolling force data and vibration data to obtain a new signal matrix; Set an adaptive filter, and perform dot - product processing on the adaptive filter and the new signal matrix to obtain the filtered frequency - domain signal and time - domain signal; Perform inverse Fourier transform on the filtered frequency - domain signal to obtain the processed rolling force data and the processed vibration data after filtering; Perform detrending processing on the processed rolling force data after filtering to obtain the processed rolling force data; 3. The dynamic modeling and angle identification method based on the roll system crossover of a hot rolling mill according to claim 2, wherein The processing formula for the processed rolling force is: P1 = P'-(α + βn'T), n' = 1, 2, …, N·········(1) Where: P1 is the processed rolling force, P' is the pre - processed rolling force, T is the sampling period, and α and β are the fitting coefficients of the pre - processed rolling force; 4. The dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill according to claim 1, wherein: The rolling force data of the rolling mill is equal to the sum of the rolling forces on both sides of the rolling mill, and the rolling force formula of the rolling mill is: P = P D + P O ·········(2) Where: P is the rolling force fluctuation of the rolling mill, P D is the rolling force fluctuation on the drive side of the rolling mill, P O is the rolling force fluctuation on the operating side of the rolling mill.

5. The dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill according to claim 1, characterized in that: The vibration sensor is one of an acceleration sensor, a velocity sensor, or a displacement sensor; 6. The dynamic modeling and angle identification method based on the roll system crossing of a hot rolling mill according to claim 1, characterized in that: The acquisition frequency of the rolling force is set to be the same as the acquisition frequency of the vibration data; 7. According to the method for dynamic modeling and angle identification based on the roll crossing of a hot - rolling mill as claimed in claim 1, wherein, The relational expression between the axial force and the rolling force of the work roll caused by the crossing of the work roll and the backup roll is as follows: F b = 0.06 × (v ω ) 0.43 × P × tanθ1 × f1 + D·········(3) Where: F b is the axial force of the work roll due to the cross work roll of the work roll and the backup roll, v ω is the linear velocity of the work roll surface in the contact area; P is the rolling force of the rolling mill, θ1 is the crossing angle between the work roll and the backup roll, f1 is the influence coefficient of the friction between the backup roll and the work roll on the axial force, D is the axial force jump value at the zero point of the crossing angle, about 8 - 10 KN; The relational expression between the axial force and the rolling force generated by the backup roll is as follows: F c = 0.06 × (v ω ) 0.43 × P × sinθ1 × f1 + D ········· (4) Where: F c is the axial force of the cross support roll between the work roll and the backup roll; The relational expression between the axial force of the work roll and the rolling force is as follows: Where: F is the axial force of the work roll, P is the rolling force of the rolling mill, θ2 is the crossing angle between the work roll and the vertical direction of the strip movement, r is the roll neck ratio of the two rolls, and f is the friction coefficient for the work roll to generate frictional force; The axial force of the work roll caused by the crossing between the work roll and the vertical direction of the strip movement is: F a = F - F b ·········(6) Where: F a is the axial force of the cross work roll generated due to the vertical direction of the work roll and the movement direction of the strip.

8. The dynamic modeling and angle identification method based on the roll system crossover of a hot rolling mill according to claim 1, characterized in that: The vibration of the axial dynamics model is the axial vibration of the roll and the strip, and the exciting force is the axial force fluctuation of the roll system.

9. The dynamic modeling and angle identification method based on the roll system crossover of a hot rolling mill according to claim 1, characterized in that: The loss function of the state space model is one of the sum of the squares of the differences between the predicted acceleration and the actual acceleration of each data node, the sum of the squares of the differences between the predicted velocity and the actual velocity of each data node, and the sum of the squares of the differences between the predicted displacement and the actual displacement. The actual acceleration or velocity or displacement is the processed vibration data.

10. The dynamic modeling and angle identification method based on the roll system crossover of a hot rolling mill according to claim 9, characterized in that The formula for the loss function of the state space model of the roll system is: Q m = Q 上支撑辊 + Q 上工作辊 + Q 下工作辊 + Q 下支撑辊 ·········(7) Where: Q m is the loss function of the state space model of the roll system, Q 上支撑辊 is the loss function of the state space model of the upper backup roll, Q 上工作辊 is the loss function of the state space model of the upper work roll, Q 下工作辊 is the loss function of the state space model of the lower work roll, Q 下支撑辊 is the loss function of the state space model of the lower backup roll.

Images