Working roll surface roughness forecasting method suitable for hot galvanizing temper mill

Abstract

The invention discloses a working roll surface roughness forecasting method suitable for a hot galvanizing temper mill, and the method comprises the steps: building a working roll surface roughness forecasting model in combination with a field production working condition, and then building an optimization objective function; and solving the influence coefficient in the working roll surface roughness forecasting model through field actual production data by using a big data regression method, and finally obtaining the working roll surface roughness forecasting model conforming to the hot galvanizing temper mill. According to the method, the surface roughness change of the working roller of the hot-galvanized strip steel temper mill is analyzed, the method suitable for forecasting the surface roughness of the working roller of the hot-galvanized temper mill is established, the surface roughness of the working roller in the hot-galvanized strip steel temper mill is calculated, and the surface quality of finished strip steel is improved.

Description

Technical Field

[0001] The present invention relates to the technology of skin pass rolling, and more particularly to a method for predicting the surface roughness of work rolls applicable to a hot-dip galvanized skin pass mill. Background Art

[0002] In recent years, with the rapid development of industries such as automobiles and household appliances, the increasing demand for strip steel by users and the continuous improvement of performance requirements, the requirements for the stable rolling of skin pass mills and the control of the shape and surface quality of finished strip steel are also getting higher and higher. The surface roughness of strip steel has an important influence on the subsequent processing procedures of finished strip steel. However, the roughness of strip steel during the skin pass process is mainly controlled by the surface roughness of the work rolls of the skin pass mill. During the rolling process, the surface characteristics of the strip steel will be imprinted by the work rolls. Different users have different requirements for the surface roughness of strip materials, so effective control is required on site. Before being put into operation, the work rolls have a set initial roughness. As the rolling mileage increases, the roughness of the roll surface will gradually transfer to the strip steel surface, resulting in a gradual decrease in the friction coefficient of the work roll surface. At the same time, during the skin pass process of strip steel after the hot-dip galvanizing process, there is a phenomenon of powdering (the coating peels off). The zinc-iron alloy is obtained by annealing treatment after hot-dip galvanizing, and the coating forms various alloy phases through alloying. In addition, the zinc-iron alloy layer has better weldability and paintability, so it has been gradually used more and more widely. However, during the skin pass process of hot-dip galvanized strip steel, problems such as powdering and peeling often occur. The powdering and peeling of the coating reduce the corrosion resistance of the coating, affect the appearance after painting, and because the peeled zinc powder adheres to the strip steel surface and the work rolls of the skin pass mill and accumulates in the work rolls, different alloying temperatures will change the phase structure of the alloy layer on the strip steel surface, affecting the powdering phenomenon of the strip steel surface. At the same time, the Al content in the plating solution also has a certain influence on alloying and affects the powdering phenomenon of the strip steel surface. This phenomenon causes an increase in the friction coefficient during the skin pass process. To maintain a certain reduction rate, the rolling pressure becomes larger, increasing the attenuation of the roughness of the work rolls. As a result, the friction coefficient between the rolls and the strip steel gradually decreases, affecting the shape and surface roughness of the finished strip steel. The decrease in the friction coefficient also affects the rolling force, resulting in a decrease in the stability of the rolling process of the mill, leading to a decrease in the stability during the rolling process of the mill.

[0003] At present, the research on the roughness of work rolls in the temper mill mainly takes into account the equipment and process characteristics of the temper mill, and deeply analyzes the causes of problems from two aspects: the rolling process and the work roll usage process. Based on the mathematical model of the one-to-one correspondence between the roughness of the strip steel and the work roll and the friction coefficient, corresponding model calculation strategies are proposed and applied to production practice to quantitatively analyze the influence of the surface roughness of the work roll and the strip steel on the friction coefficient; for the problem of powdering of hot-dip galvanized steel, mainly through annealing treatment after hot-dip galvanizing of zinc-iron alloy, the coating often shows problems such as powdering and peeling. And the degree of powdering of the coating phase structure was studied at different alloying temperatures.

[0004] Therefore, in on-site production research, how to reasonably set the initial surface roughness of the work roll, calculate the surface roughness of the work roll during the tempering process of hot-dip galvanized strip steel, and improve the surface quality of the finished strip steel has become an important research topic. Summary of the Invention

[0005] Aiming at the defects existing in the prior art, the purpose of the present invention is to provide a method for predicting the surface roughness of work rolls applicable to hot-dip galvanized temper mills. By analyzing the change of the surface roughness of work rolls in hot-dip galvanized strip steel temper mills, a method for predicting the surface roughness of work rolls suitable for hot-dip galvanized temper mills is established, and the surface roughness of work rolls during the tempering process of hot-dip galvanized strip steel is calculated to improve the surface quality of the finished strip steel.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A method for predicting the surface roughness of work rolls applicable to hot-dip galvanized temper mills:

[0008] Combined with the on-site production conditions, a prediction model for the surface roughness of the work roll is established, and then an optimization objective function is established. The influence coefficients in the prediction model of the surface roughness of the work roll are solved by using the big data regression method through on-site actual production data, and finally the prediction model of the surface roughness of the work roll that meets the hot-dip galvanized temper mill is obtained.

[0009] Preferably, the method for predicting the surface roughness of the work roll specifically includes the following steps:

[0010] S1. Collect data;

[0011] S2. Group the collected data;

[0012] S3. Establish the prediction model for the surface roughness of the work roll;

[0013] S4. Construct the corresponding control function formula of the prediction model of the work roll roughness, and introduce the objective function and the influence coefficient of the working conditions on the basis of the control function formula;

[0014] S5. Define the initial target value and intermediate variables;

[0015] S6. Specify the search step size for the working condition influence coefficient;

[0016] S7. Use the working roll roughness prediction model to calculate the working roll roughness value under the current working condition influence coefficient;

[0017] S8. Calculate the specific value of the objective function under the current condition;

[0018] S9. Define the optimal working condition influence coefficient;

[0019] S10. Determine whether the inequality holds. If it does, go to step S6; if not, go to step S11;

[0020] S11. Output the optimal working condition influence coefficient;

[0021] S12. Substitute the optimal working condition influence coefficient into the working roll roughness prediction model to obtain the surface roughness of the working roll of the hot-dip galvanizing temper mill.

[0022] Preferably, in step S1, the data collection specifically includes:

[0023] Collect the equipment characteristic parameters of the temper mill and the key rolling process parameters of the strip, including: the original roughness Ra of the working roll 0 , the roughness Ra of the working roll during rolling r , the rolling kilometers L of the working roll, the galvanizing temperature T, and the aluminum content C in the plating solution Al .

[0024] Preferably, in step S2, the grouping of the collected data specifically includes:

[0025] Define the n groups of input parameters after rolling as The corresponding n groups of actual surface roughnesses of the temper mill working roll {Ra ri} where i = 1, 2,..., n.

[0026] Preferably, in step S3, establish the working roll surface roughness prediction model Ra r , where

[0027] In the formula, λ 1 , λ 2 are the influence coefficients of the hot-dip galvanizing temperature; α 1 , α 2 are the influence coefficients of the Al content in the plating solution; B L is the working roll roughness attenuation coefficient, which is taken as 0.042 according to the unit experiment.

[0028] Preferably, in the step S4, a control function formula corresponding to the work roll roughness prediction model is constructed. Based on the control function formula, an objective function is introduced. The working condition influence coefficient E = {λ 1 , λ 2 , α 1 , α 2};

[0029] When G(E) is smaller, it means that the calculated value of the work roll roughness prediction model is closer to the actual value on site. Then, a set of optimal working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} is sought, such that is minimized. Then, this set of working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} is the optimal working condition influence coefficient of the work roll roughness prediction model.

[0030] Preferably, in the step S5, an initial target value G 0 is defined. At the same time, four intermediate variables m 0 , m 1 , m 2 , m 3 are defined, and m 0 is set to 0.

[0031] Preferably, in the step S6, specifying the search step size of the working condition influence coefficient specifically includes the following steps:

[0032] S61. Specify the search step size step0 of λ 1 , and set λ 1 = 0.500 + m 0 * step0, and set m 1 to 0;

[0033] S62. Specify the search step size step1 of λ 2 , and set λ 2 = 0.001 + m 1 * step1, and set m 2 to 0;

[0034] S63. Specify the search step size step2 of α 1 , and set α 2 = 0.001 + m 2 * step2, and set m 3= 0;

[0035] S64. Given α 2 with a search step size step3, and let α 2 = 0.001 + m 3 * step3.

[0036] Preferably, in the said step S9, define the best operating condition influence coefficient λ 1y , λ 2y , α 1y , α 2y , and determine whether G < G 0 holds. If so, then let G 0 = G, λ 1y = λ 1 , λ 2y = λ 2 , α 1y = α 1 , α 2y = α 2 and then transfer to step S10; if not, directly transfer to step S10.

[0037] Preferably, in the said step S10, determining whether the inequality holds specifically includes the following steps:

[0038] S101. Determine whether the inequality m 3 ≤ 0.5000 / step3 holds. If so, then let m 3 = m 3 + 1 and transfer to step S64; if not, transfer to step S102;

[0039] S102. Determine whether the inequality m 2 ≤ 0.5000 / step2 holds. If so, then let m 2 = m 2 + 1 and transfer to step S63; if not, transfer to step S103;

[0040] S103. Determine whether the inequality m 1 ≤ 0.5000 / step1 holds. If so, then let m 1 = m 1 + 1 and transfer to step S62; if not, transfer to step S104;

[0041] S104. Determine whether the inequality m 0 ≤ 0.5000 / step0 holds. If so, then let m 0 = m 0 + 1 and transfer to step S61; if not, transfer to step S11.

[0042] Preferably, in the step S12, the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y are substituted into the working roll surface roughness prediction model to calculate the working roll surface roughness Ra r .

[0043] A method for predicting the surface roughness of a work roll applicable to a hot-dip galvanizing temper mill provided by the present invention combines a theoretical model with on-site actual production data, establishes a calculation model for the surface roughness of a work roll in a hot-dip galvanizing temper mill that conforms to on-site actual production, more precisely ensures the stability of the surface quality of strip products during hot-dip galvanizing temper rolling, and at the same time, this method can be applied to other temper production lines to better serve production. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a schematic flow chart of an embodiment of the method for predicting the surface roughness of a work roll of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] In order to better understand the above technical solutions of the present invention, the technical solutions of the present invention will be further described below with reference to the drawings and embodiments.

[0046] A method for predicting the surface roughness of a work roll applicable to a hot-dip galvanizing temper mill provided by the present invention:

[0047] Combined with the on-site production working conditions, the working roll decays continuously with the rolling kilometers during the rolling process, and considering the pulverization phenomenon of the galvanized sheet, a prediction model for the surface roughness of the working roll is established, and then an optimization objective function is established. The influence coefficients in the prediction model of the surface roughness of the working roll are solved by using the big data regression method through on-site actual production data, and finally a prediction model for the surface roughness of the working roll that conforms to the hot-dip galvanizing temper mill is obtained to ensure the roughness requirements of the strip surface during the tempering process.

[0048] Combined with Figure 1 as shown, the method for predicting the surface roughness of a work roll of the present invention specifically includes the following steps:

[0049] S1. Collect the equipment characteristic parameters of the temper mill and the key rolling process parameters of the strip, including: the original roughness Ra 0 of the work roll, the roughness Ra r of the work roll during the rolling process, the rolling kilometers L of the work roll, the galvanizing temperature T, and the aluminum content C Al in the plating solution.

[0050] S2. Group the collected data, and define the n groups of input parameters after rolling as The corresponding n groups of actual surface roughnesses {Ra ri} of the work rolls of the planishing mill, where i = 1, 2, …, n.

[0051] S3. Establish a prediction model Ra for the surface roughness of the work roll r , where

[0052] In the formula, λ 1 , λ 2 are the influence coefficients of the hot-dip galvanizing temperature; α 1 , α 2 are the influence coefficients of the Al content in the plating solution; B L is the attenuation coefficient of the work roll roughness, and its value is 0.042 according to the unit experiment.

[0053] S4. According to the above description, construct the corresponding control function formula of the work roll roughness prediction model Introduce the objective function on the basis of the control function formula The working condition influence coefficient E = {λ 1 , λ 2 , α 1 , α 2};

[0054] When G(E) is smaller, it means that the calculated value of the work roll roughness prediction model is closer to the actual value on site. Then, seek a set of optimal working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} such that is the smallest. Then, this set of working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} is the optimal working condition influence coefficient of the work roll roughness prediction model.

[0055] S5. Define the initial target value G 0 , and at the same time define four intermediate variables m 0 , m 1 , m 2 , m 3 , and let m 0 = 0.

[0056] S6. Given the search step size of the working condition influence coefficient, which specifically includes the following steps:

[0057] S61. Given the search step size step0 of λ 1 , and let λ 1 = 0.500 + m 0 *step0, and let m1 = 0;

[0058] S62. Given the search step size step1 of λ 2 and let λ 2 = 0.001 + m 1 * step1, and let m 2 = 0;

[0059] S63. Given the search step size step2 of α 1 and let α 2 = 0.001 + m 2 * step2, and let m 3 = 0;

[0060] S64. Given the search step size step3 of α 2 and let α 2 = 0.001 + m 3 * step3.

[0061] S7. Use the work roll roughness prediction model to calculate the work roll roughness values {Ra 1 , λ 2 , α 1 , α 2} in the condition of E = {λ r , i = 1, 2,..., n}.

[0062] S8. Calculate the specific value of the objective function in the current condition

[0063] S9. Define the optimal condition influence coefficients λ 1y , λ 2y , α 1y , α 2y , and judge whether G < G 0 holds. If so, then let G 0 = G, λ 1y = λ 1 , λ 2y = λ 2 , α 1y = α 1 , α 2y = α 2 and then go to step S10; if not, then directly go to step S10.

[0064] S10. Judge whether the inequality holds, which specifically includes the following steps:

[0065] S101. Judge whether the inequality m 3 ≤ 0.5000 / step3 holds. If so, then let m 3 = m3 After adding 1, go to step S64; if not, go to step S102;

[0066] S102. Determine whether the inequality m 2 ≤ 0.5000 / step2 holds. If so, let m 2 = m 2 + 1, then go to step S63; if not, go to step S103;

[0067] S103. Determine whether the inequality m 1 ≤ 0.5000 / step1 holds. If so, let m 1 = m 1 + 1, then go to step S62; if not, go to step S104;

[0068] S104. Determine whether the inequality m 0 ≤ 0.5000 / step0 holds. If so, let m 0 = m 0 + 1, then go to step S61; if not, go to step S11.

[0069] S11. Output the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y .

[0070] S12. Substitute the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y into the prediction model of the work roll surface roughness, and calculate the work roll surface roughness Ra r .

[0071] Example 1

[0072] Continue to refer to Figure 1 as shown. A method for predicting the work roll surface roughness applicable to the hot-dip galvanizing temper mill provided in this Example 1 specifically includes the following steps:

[0073] S1. Collect the equipment characteristic parameters of the temper mill and the key rolling process parameters of the strip, including: the original roughness Ra 0 = 1.3 of the work roll, the rolling kilometer number L = 20 km of the work roll, the galvanizing temperature T = 530 °C, and the aluminum content C Al = 0.12% in the plating solution.

[0074] S2. Group the collected data. Define the 200 groups of input parameters after rolling as The corresponding surface roughness of the work rolls of 200 actual planers {Ra r1 = 1.06}, where i = 1.

[0075] S3. Establish a prediction model for the surface roughness of the work roll Ra r , where

[0076] In the formula, λ 1 , λ 2 are the influence coefficients of the hot-dip galvanizing temperature; α 1 , α 2 are the influence coefficients of the Al content in the plating solution; B L is the attenuation coefficient of the work roll roughness, and the value is 0.042 according to the unit experiment.

[0077] S4. According to the above description, construct the corresponding control function formula of the work roll roughness prediction model Introduce the objective function on the basis of the control function formula The working condition influence coefficient E = {λ 1 , λ 2 , α 1 , α 2};

[0078] When G(E) is smaller, it means that the calculated value of the work roll roughness prediction model is closer to the actual value on site. Then, seek a set of optimal working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} such that is the smallest. Then, this set of working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} is the optimal working condition influence coefficient of the work roll roughness prediction model.

[0079] S5. Define the initial target value G 0 = 0.1, and at the same time define four intermediate variables m 0 , m 1 , m 2 , m 3 , and let m 0 = 0.

[0080] S6. Given the search step of the working condition influence coefficient, which specifically includes the following steps:

[0081] S61. Given the search step of λ 1 step0 = 0.005, and let λ 1 = 0.500 + m 0*Step0, let m 1 = 0;

[0082] S62. Given the search step size step1 = 0.005 for λ 2 and let λ 2 = 0.001 + m 1 *step1, let m 2 = 0;

[0083] S63. Given the search step size step2 = 0.005 for α 1 and let α 2 = 0.001 + m 2 *step2, let m 3 = 0;

[0084] S64. Given the search step size step3 = 0.005 for α 2 and let α 2 = 0.001 + m 3 *step3.

[0085] S7. Use the work roll roughness prediction model to calculate the work roll roughness value {Ra 1 , λ 2 , α 1 , α 2} under the condition of {λ r = 1.25, i = 1}.

[0086] S8. Calculate the specific value of the objective function under the current condition

[0087] S9. Define the optimal condition influence coefficients λ 1y , λ 2y , α 1y , α 2y , and judge whether G = 0.082 < G 0 = 0.1 holds. If so, let G 0 = G, λ 1y = 0.05, λ 2y = 0.012, α 1y = 0.002, α 2y = 0.01, and then go to step S10.

[0088] S10. Judge whether the inequality holds, which specifically includes the following steps:

[0089] S101. Judge whether the inequality m 3 = 0 ≤ 0.5000 / 0.005 holds. If so, let m 3 = m 3After adding 1, transfer to step S64; if not, transfer to step S102;

[0090] S102. Determine whether the inequality m 2 = 0 ≤ 0.5000 / 0.005 holds. If so, let m 2 = m 2 + 1, then transfer to step S63; if not, transfer to step S103;

[0091] S103. Determine whether the inequality m 1 = 0 ≤ 0.5000 / 0.005 holds. If so, let m 1 = m 1 + 1, then transfer to step S62; if not, transfer to step S104;

[0092] S104. Determine whether the inequality m 0 = 0 ≤ 0.5000 / 0.005 holds. If so, let m 0 = m 0 + 1, then transfer to step S61; if not, transfer to step S11.

[0093] S11. Output the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y .

[0094] S12. Substitute the optimal working condition influence coefficient λ 1y = 0.0096, λ 2y = 0.00018, α 1y = 0.032, α 2y = 0.17 into the prediction model of the work roll surface roughness, and calculate the work roll surface roughness Ra r = 1.12.

[0095] It can be seen from the simulation calculation that before the optimization condition coefficients λ 1 , λ 2 , α 1 , α 2 , the deviation between the predicted work roll roughness and the actual work roll roughness is large. After the optimization coefficients λ 1 , λ 2 , α 1 , α 2 , the work roll roughness of the hot-dip galvanizing temper mill obtained by the simulation calculation is basically consistent with the actual work roll roughness, and the error is only ξ 1 = (1.12 - 1.06) / 1.06 = 5.6%, which fully meets the production requirements.

[0096] Example 2

[0097] Continue to refer to Figure 1 As shown, a method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill provided in Embodiment 2 specifically includes the following steps:

[0098] S1. Collect the equipment characteristic parameters of the temper mill and the key rolling process parameters of the strip, including: the original roughness Ra of the work roll 0 = 1.3, the rolling mileage L of the work roll = 20 km, the galvanizing temperature T = 530 °C, and the aluminum content C in the plating solution Al = 0.12%.

[0099] S2. Group the collected data. Define the 200 groups of input parameters after rolling as the corresponding 200 groups of actual surface roughnesses {Ra r1 = 1.13} of the work rolls of the temper mill, where i = 1.

[0100] S3. Establish a prediction model Ra for the surface roughness of the work roll r where

[0101] In the formula, λ 1 , λ 2 are the influence coefficients of the hot-dip galvanizing temperature; α 1 , α 2 are the influence coefficients of the Al content in the plating solution; B L is the attenuation coefficient of the work roll roughness, and its value is 0.042 according to the mill experiment.

[0102] S4. According to the above description, construct the corresponding control function formula for the work roll roughness prediction model Introduce the objective function on the basis of the control function formula The working condition influence coefficient E = {λ 1 , λ 2 , α 1 , α 2};

[0103] When G(E) is smaller, it means that the calculated value of the work roll roughness prediction model is closer to the actual value on site. Then, seek a set of optimal working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} such that is the smallest. Then, this set of working condition influence coefficients E = {λ 1 , λ 2 , α 1 , α 2} is the optimal working condition influence coefficient of the work roll roughness prediction model.

[0104] S5. Define the initial target value G 0 = 0.1, and at the same time define four intermediate variables m 0 , m 1 , m 2 , m 3 , and let m 0 = 0.

[0105] S6. Given the search step of the working condition influence coefficient, which specifically includes the following steps:

[0106] S61. Given the search step of λ 1 step0 = 0.005, and let λ 1 = 0.500 + m 0 * step0, and let m 1 = 0;

[0107] S62. Given the search step of λ 2 step1 = 0.005, and let λ 2 = 0.001 + m 1 * step1, and let m 2 = 0;

[0108] S63. Given the search step of α 1 step2 = 0.005, and let α 2 = 0.001 + m 2 * step2, and let m 3 = 0;

[0109] S64. Given the search step of α 2 step3 = 0.005, and let α 2 = 0.001 + m 3 * step3.

[0110] S7. Use the work roll roughness prediction model to calculate the work roll roughness value {Ra 1 , λ 2 , α 1 , α 2} under the condition of {λ r = 1.26, i = 1}.

[0111] S8. Calculate the specific value of the objective function under the current condition

[0112] S9. Define the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y , and judge G = 0.075 < G 0If =0.1 holds, then let G 0 =G, λ 1y =0.05, λ 2y =0.012, α 1y =0.002, α 2y =0.01, then transfer to step S10.

[0113] S10. Determine whether the inequality holds, specifically including the following steps:

[0114] S101. Determine whether the inequality m 3 =0 ≤ 0.5000 / 0.005 holds. If so, then let m 3 =m 3 +1, then transfer to step S64. If not, transfer to step S102;

[0115] S102. Determine whether the inequality m 2 =0 ≤ 0.5000 / 0.005 holds. If so, then let m 2 =m 2 +1, then transfer to step S63. If not, transfer to step S103;

[0116] S103. Determine whether the inequality m 1 =0 ≤ 0.5000 / 0.005 holds. If so, then let m 1 =m 1 +1, then transfer to step S62. If not, transfer to step S104;

[0117] S104. Determine whether the inequality m 0 =0 ≤ 0.5000 / 0.005 holds. If so, then let m 0 =m 0 +1, then transfer to step S61. If not, transfer to step S11.

[0118] S11. Output the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y .

[0119] S12. Substitute the optimal working condition influence coefficient λ 1y =0.0083, λ 2y =0.00021, α 1y =0.030, α 2y =0.17 into the work roll surface roughness prediction model to calculate the work roll surface roughness Ra r =1.10.

[0120] It can be seen from the simulation calculation that under the optimized working condition coefficient λ1 and λ 2 and α 1 and α 2 Before that, there was a large deviation between the predicted work roll roughness and the actual work roll roughness. After optimizing the coefficients λ 1 and λ 2 and α 1 and α 2 the work roll roughness of the hot-dip galvanizing temper mill obtained by simulation calculation is basically consistent with the actual work roll roughness, and the error is only ξ 1 = |(1.10 - 1.13)| / 1.13 = 2.6%, fully meeting the production requirements.

[0121] Those of ordinary skill in the art should recognize that the above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. As long as it is within the scope of the spirit of the present invention, changes and modifications to the above embodiments will fall within the scope of the claims of the present invention.

Claims

1. A method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill, characterized in that: Combined with the on-site production conditions, a prediction model for the surface roughness of work rolls is established, and then an optimization objective function is established. The influence coefficients in the prediction model for the surface roughness of work rolls are solved by using the big data regression method through on-site actual production data, and finally the prediction model for the surface roughness of work rolls that meets the hot-dip galvanizing temper mill is obtained.

2. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 1, characterized in that, the method for predicting the surface roughness of work rolls specifically includes the following steps: S1. Collect data; S2. Group the collected data; S3. Establish the prediction model for the surface roughness of work rolls; S4. Construct the corresponding control function formula for the prediction model of work roll roughness, and introduce the objective function and working condition influence coefficients on the basis of the control function formula; S5. Define the initial target value and intermediate variables; S6. Given the search step of the working condition influence coefficient; S7. Calculate the work roll roughness value under the current working condition influence coefficient by using the work roll roughness prediction model; S8. Calculate the specific value of the objective function under the current condition; S9. Define the best working condition influence coefficient; S10. Judge whether the inequality holds. If so, go to step S6; if not, go to step S11; S11. Output the optimal working condition influence coefficient; S12. Substitute the optimal working condition influence coefficient into the work roll roughness prediction model to obtain the surface roughness of the work rolls of the hot-dip galvanizing temper mill.

3. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 2, characterized in that, in step S1, collecting data specifically includes: Collect the equipment characteristic parameters of the recoiling and leveling unit and the key rolling process parameters of the strip, including: the original roughness Ra of the work roll 0 , the roughness Ra of the work roll during rolling r , the rolling kilometers L of the work roll, the galvanizing temperature T, and the aluminum content C in the plating solution Al .

4. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 3, characterized in that, in step S2, grouping the collected data specifically includes: Define the n sets of input parameters after rolling as The corresponding n sets of actual surface roughness of the temper mill work rolls {Ra ri}, where i = 1, 2,..., n.

5. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 4, characterized in that: In the step S3, a prediction model Ra of the surface roughness of the work roll is established r , where where λ 1 , λ 2 are the influence coefficients of the hot-dip galvanizing temperature; α 1 , α 2 are the influence coefficients of the Al content in the plating solution; B L is the attenuation coefficient of the work roll roughness, and its value is taken as 0.042 according to the unit experiment.

6. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 5, characterized in that, In the step S4, a control function formula corresponding to the work roll roughness prediction model is constructed. On the basis of the control function formula, an objective function is introduced. The working condition influence coefficient E = {λ 1 、λ 2 、α 1 、α 2}; When G(E) is smaller, it means that the calculated value of the work roll roughness prediction model is closer to the actual value on site. Then, a set of optimal working condition influence coefficients E = {λ 1 、λ 2 、α 1 、α 2} is sought to make the smallest. Then, this set of working condition influence coefficients E = {λ 1 、λ 2 、α 1 、α 2} is the optimal working condition influence coefficient of the work roll roughness prediction model.

7. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 6, characterized in that, In the step S5, an initial target value G is defined. 0 , and at the same time, four intermediate variables m 0 , m 1 , m 2 , m 3 are defined, and let m 0 = 0.

8. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 7, characterized in that, in step S6, given the search step of the working condition influence coefficient specifically includes the following steps: S61. Given λ 1 with a search step size step0, and let λ 1 = 0.500 + m 0 * step0, and let m 1 = 0; S62. Given λ 2 with a search step size step1, and let λ 2 = 0.001 + m 1 * step1, and let m 2 = 0; S63. Given α 1 with a search step size step2, and let α 2 = 0.001 + m 2 * step2, and let m 3 = 0; S64. Given α 2 with a search step size step3, and let α 2 = 0.001 + m 3 * step3.

9. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 8, characterized in that, In the step S9, the optimal operating condition influence coefficient λ is defined. 1y , λ 2y , α 1y , α 2y , and it is judged whether G < G 0 holds. If so, then let G 0 = G, λ 1y = λ 1 , λ 2y = λ 2 , α 1y = α 1 , α 2y = α 2 . Then, it transfers to the step S10; if not, it directly transfers to the step S10.

10. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 9, characterized in that, in step S10, judging whether the inequality holds specifically includes the following steps: S101. Determine whether the inequality m 3 ≤ 0.5000 / step3 holds. If so, let m 3 = m 3 + 1, then go to step S64. If not, go to step S102; S102. Determine whether the inequality m 2 ≤ 0.5000 / step2 holds. If so, let m 2 = m 2 + 1, then go to step S63. If not, go to step S103; S103. Determine whether the inequality m 1 ≤0.5000 / step1 holds. If it does, then let m 1 = m 1 + 1, and then go to step S62. If not, go to step S104; S104. Determine whether the inequality m 0 ≤ 0.5000 / step0 holds. If so, let m 0 = m 0 + 1, then go to step S61. If not, go to step S11.

11. The method for predicting the surface roughness of work rolls applicable to a hot-dip galvanizing temper mill according to claim 10, characterized in that, In the step S12, the optimal working condition influence coefficient λ 1y , λ 2y , α 1y , α 2y are substituted into the prediction model of the work roll surface roughness to obtain the work roll surface roughness Ra r .

Images