Method for calculating the plate thickness schedule of a tandem rolling mill

Abstract

To calculate and set a limit value for limit check which can maintain a stable rolling / operational state with reduced rolling troubles.SOLUTION: A plate thickness schedule calculation method includes a calculation step of calculating at least one of an upper limit value and a lower limit value of a draft ratio of each rolling stand, based on past rolling data, for each product specification, and a setting step of setting at least one of the calculated upper limit value and lower limit value as a limit value. The method extracts the rolled past rolling data without causing a rolling trouble. The method sets the preset draft ratio when the number of past rolling data is less than a first threshold, sets at least one of the upper limit value and the lower limit value calculated by statistic analysis when it is equal to or more than the first threshold and less than a second threshold, and sets at least one of the upper limit value and the lower limit value calculated by adding or subtracting a constant numeric value to / from the draft ratio calculated using machine learning when it is equal to or more than the second threshold, as limit values.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] This disclosure relates to a method for calculating the plate thickness schedule of a tandem rolling mill in which multiple rolling stands are arranged side by side. [Background technology]

[0002] In the plate thickness schedule calculation method for a tandem rolling mill disclosed in Patent Document 1 below, the plate thickness schedule is calculated using a set of mathematical formulas for predicting the material temperature, rolling load, rolling torque, etc., of each rolling stand in order to achieve the target plate thickness given as a rolling order.

[0003] The load ratio distribution method is used to calculate the plate thickness schedule, and the plate thickness schedule is calculated based on the distribution ratio γi of the load Pi at each rolling stand. In addition, the exit plate thickness hi and roll peripheral speed Vi of each rolling stand must satisfy the constant volume velocity rule (also called the "constant mass flow rule") in order to maintain uniform speed between rolling stands. Here, i is an identifier used to distinguish between multiple rolling stands, and the number of the rolling stand (i=1~N) is substituted for i.

[0004] In the above calculation method, the relationship between the rolling load and load ratio of each stand, as well as the constant mass flow rule, are used to obtain relational equations, and the same number of unknowns are numerically solved using methods such as the Newton-Raphson method. Furthermore, the above calculation method performs limit checks (also called "parameter limits") on parameters such as the reduction ratio, rolling load, and rolling torque of each rolling stand. If the limit value of the limit check is exceeded, the plate thickness schedule is automatically corrected by lowering the target load ratio value of that rolling stand. This plate thickness schedule calculation method had a problem in that the calculation performance deteriorated depending on the number of rolling stands to be corrected or the amount of correction. Here, deterioration in calculation performance includes, for example, when there are more than half of the rolling stands to be corrected or when the amount of correction is somewhat large, the calculation load becomes high or the iterative calculation becomes difficult to converge.

[0005] Patent Document 2, described below, discloses a technique for appropriately correcting calculations by changing the function used in the calculation when parameters related to rolling exceed limit values. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2000-167612 [Patent Document 2] International Publication No. 2021-084636 [Overview of the project] [Problems that the invention aims to solve]

[0007] However, neither Patent Document 1 nor Patent Document 2 clearly describes what numerical value the limit value used for limit checks should be, including how to set it. If a sufficiently large value is set as the limit value, the limit check itself will cease to function. On the other hand, if a relatively small value is set as the limit value, the frequency of exceeding the limit value will increase, and the calculation of the plate thickness schedule will not converge. In either case, it may lead to rolling troubles, potentially resulting in reduced yield and increased downtime. Furthermore, currently, the same numerical value is used for all materials regardless of steel type or size, and this common value cannot be said to be a limit value suitable for rolled materials.

[0008] This disclosure was made to solve the problems described above, and aims to provide a plate thickness schedule calculation method that can calculate and set limit values ​​for limit checks that are suitable for maintaining a stable rolling and operating state with fewer rolling troubles. [Means for solving the problem]

[0009] The first aspect relates to a method for calculating the plate thickness schedule of a tandem rolling mill in which rolling is performed continuously by multiple rolling stands. The plate thickness schedule calculation method includes a calculation step that uses a rolling model equation that includes an advance rate model and a rolling load model, and calculates the plate thickness schedule based on the rolling load or motor power. The calculation step includes a calculation step that calculates at least one of the upper and lower limits of the reduction ratio at each rolling stand for each product specification, including the steel grade and plate thickness to be rolled, based on past rolling data, and a setting step that sets at least one of the upper and lower limits calculated in the calculation step as the limit value for limit checks on the reduction ratio.

[0010] The second perspective, in addition to the first perspective, has the following further characteristics: The calculation process includes a step of extracting rolling data from past rolling data when rolling was performed without rolling trouble, and is configured to calculate at least one of an upper limit and a lower limit from at least one of the set calculated value and actual value of the reduction ratio of the extracted rolling data by statistical analysis.

[0011] The third perspective, in addition to the first perspective, has the following further characteristics: The calculation process includes a step of extracting rolling data from past rolling data when the deviation between the set calculated value and the actual value of the reduction ratio is smaller than a reference value, and is configured to calculate at least one of the upper limit and lower limit from at least one of the set calculated value and the actual value of the reduction ratio of the extracted rolling data by statistical analysis.

[0012] The fourth perspective, in addition to the first perspective, has the following further characteristics: In the calculation process, at least one of the past set calculation values ​​and actual values ​​from past rolling data, when rolling was possible without rolling trouble, is used as input, and the reduction ratio is calculated sequentially using machine learning, and at least one of the upper and lower limits is calculated by adding or subtracting a certain value to the calculated reduction ratio.

[0013] The fifth perspective, in addition to the first perspective, has the following further characteristics: In the calculation process, at least one of the past set calculation value and actual value from past rolling data, where the deviation between the set calculation value and the actual value is smaller than the standard value, is used as input, and the reduction ratio is calculated sequentially using machine learning, which outputs the reduction ratio. At least one of the upper and lower limits is calculated by adding or subtracting a certain value to the calculated reduction ratio.

[0014] The sixth perspective, in addition to the first perspective, has the following further characteristics: In the calculation process, rolling data from past rolling data for the categories in which rolling troubles occurred is extracted, the average value of the reduction ratio for the set calculation is calculated using the rolling data from when rolling was performed without rolling troubles, and percentile values ​​are calculated from the reduction ratio for the set calculation using the rolling data from when rolling troubles occurred, and at least one of the upper and lower limits is calculated by adding or subtracting a certain value from the calculated average value and percentile value.

[0015] The seventh aspect, in addition to the first aspect, has the following further characteristics: If at least one of the upper and lower limits calculated at the downstream rolling stand is greater than the upper limit at the upstream rolling stand, the upper limit of the downstream rolling stand is replaced with the upper limit of the upstream rolling stand, or the upper limit of the downstream rolling stand is replaced with a value obtained by subtracting a certain number from the upper limit of the upstream stand.

[0016] The eighth perspective, in addition to the first perspective, has the following further characteristics: When the number of past rolling data is less than the first threshold, the system is configured to set a pre-calculated reduction ratio as the limit value. When the accumulation of rolling data progresses and the number of past rolling data is greater than or equal to the first threshold but less than the second threshold, the system is configured to set at least one of the upper and lower limits calculated by statistical analysis from at least one of the set calculated value and actual value of the reduction ratio of the rolling data calculated by statistical analysis as the limit value. When the accumulation of rolling data progresses further and the number of past rolling data is greater than or equal to the second threshold, the system is configured to set at least one of the upper and lower limits calculated by adding or subtracting a certain value from the reduction ratio calculated using machine learning as the limit value.

[0017] The ninth aspect comprises multiple rolling stands, a reduction device installed at each of the multiple rolling stands, an electric motor for rotating the rolls of each rolling stand, a process computer configured to calculate the plate thickness schedule for each rolling stand based on one of the values ​​of the rolling load ratio of the reduction device and the motor power ratio of the electric motor, and a database for accumulating past rolling data. The process computer is configured to perform a limit check on the reduction ratio of each rolling stand, calculate at least one of the upper and lower limits of the reduction ratio at each rolling stand for each product specification including the steel grade and plate thickness to be rolled, based on past rolling data accumulated in the database, and set at least one of the calculated upper and lower limits as the limit value for the limit check on the reduction ratio. [Effects of the Invention]

[0018] According to the first and ninth perspectives, by setting at least one of the upper and lower limits of the reduction ratio calculated using past rolling data for each product specification as a limit value, it is possible to set a limit value suitable for the rolled material, thereby avoiding rolling troubles. This makes it possible to improve yield and reduce downtime.

[0019] From the second and fourth perspectives, using rolling data from processes that did not experience rolling problems can reduce the likelihood of rolling problems occurring, enabling more stable operations.

[0020] From the third and fifth perspectives, by using rolling data in which the deviation between the set calculated value and the actual value is small, at least one of the upper and lower limits of the reduction ratio can be determined with high accuracy.

[0021] From a sixth perspective, even when using rolling data from the section where rolling trouble occurred, it is possible to calculate at least one of the upper and lower limits of the reduction ratio.

[0022] From a seventh perspective, preventing a reversal of the reduction ratio can prevent the rolling balance from being disrupted and causing rolling problems.

[0023] From the eighth perspective, limit values ​​can be set in stages according to the number of rolling data points. [Brief explanation of the drawing]

[0024] [Figure 1] This is a schematic diagram showing the configuration of a rolling mill according to an embodiment. [Figure 2] This figure shows an example of the hardware configuration of a process computer installed in a rolling mill. [Figure 3] This is a flowchart illustrating the process of calculating the plate thickness schedule in Example 1. [Figure 4] This is a flowchart illustrating the process of calculating the plate thickness schedule in Example 2. [Figure 5] This is a flowchart illustrating the process of calculating the plate thickness schedule in Example 3. [Figure 6] This is a flowchart illustrating the process of calculating the plate thickness schedule in Example 4. [Modes for carrying out the invention]

[0025] Embodiments of the present invention will be described in detail below with reference to the drawings. In each drawing, elements common to all figures are denoted by the same reference numerals, and redundant explanations are omitted.

[0026] [System configuration of the embodiment] Figure 1 is a schematic diagram showing the configuration of a rolling plant 1 according to an embodiment. The rolling plant 1 uses steel or other metal material as the material to be rolled 10 and rolls the material to be rolled 10 into a plate shape while hot. The material to be rolled 10 is the material to be rolled in the rolling plant 1. The rolling plant 1 may also be configured to roll the material to be rolled 10 into a plate shape while cold.

[0027] The rolling plant 1 comprises a heating furnace 2, a roughing mill 3, a finishing mill 4, a cooling device 5, a winding machine 6, and a roller table (not shown) for transporting the rolled material 10 between them.

[0028] The heating furnace 2 heats and raises the temperature of the material to be rolled 10. The roughing mill 3 has one or more rolling stands. The rolling stand has multiple rolls 31, a reduction device 32, and an electric motor 33 for rotating the rolls.

[0029] The finishing rolling mill 4 is a tandem rolling mill equipped with a plurality of rolling stands F1 to F5 connected in the direction of transport of the rolled material 10. Each rolling stand F1 to F5 is equipped with a plurality of rolls 41, a reduction device 42, and an electric motor 43 for rotating the rolls. In the following description, the rolling stands may also be referred to as i or i-1.

[0030] Furthermore, the number of rolling stands for the heating furnace 1, winding machine 6, roughing mill 3, and finishing mill 4 is not particularly limited. In this embodiment, a rolling plant 1 is given as an example, which includes one heating furnace 1, one rolling stand for the roughing mill 3, five rolling stands F1 to F5 for the finishing mill 4, and one winding machine 6.

[0031] In the following description, the reduction devices 32, 42 and electric motors 33, 43 of each of the rolling mills 3 and 4 described above may be referred to as "equipment" of the rolling plant 1 for convenience. Depending on the specific structure of each of the rolling mills 3 and 4, the equipment may include various other components besides the reduction devices 32, 42 and electric motors 33, 43, such as actuators which are not shown in the illustration.

[0032] Various sensors are installed as measuring instruments at key points in the rolling plant 1. Key points in the rolling plant 1 include, for example, the outlet of the heating furnace 2, the outlet of the roughing mill 3, the outlet of the finishing mill 4, and the inlet of the winding machine 6. Various sensors may also be installed between the rolling stands F1 to F5 of the finishing mill 4. The various sensors include a pyrometer 71 that measures the surface temperature of the rolled material 10 at the inlet of the finishing mill 4, a thickness and width gauge 72 that measures the thickness and width of the rolled material 10, a pyrometer 73 that measures the surface temperature of the rolled material 10 at the outlet of the finishing mill 4, a rolling load sensor 74 that measures the rolling load at each rolling stand F1 to F5, and a pyrometer 75 that measures the surface temperature of the rolled material 10 at the inlet of the winding machine 6. Various sensors sequentially measure the state of the rolled material 10 and each piece of equipment.

[0033] The rolling mill 1 is operated by a computer-based control system. The computer system includes a host computer 20 and a process computer 21, which are connected to each other via a network. The process computer 21 is connected via the network to an interface screen 21a, which is the operation screen, and a database 23. The database 23 is configured to sequentially store past rolling data. The past rolling data includes the set values ​​and actual values ​​of the reduction ratio for each rolling stand F1 to F5.

[0034] The host computer 20 issues a rolling command to the process computer 21 based on a pre-set production plan. The rolling command includes, for example, target dimensions and target temperatures for each rolled material 10. Target dimensions include, for example, target plate thickness, target plate width, and target crown. Target temperatures include, for example, the exit temperature of the roughing mill 3, the exit temperature of the finishing mill 4, and the entry temperature of the winding machine 6.

[0035] When the rolled material 10 is extracted from the heating furnace 2, the process computer 21 calculates the set values ​​for each piece of equipment in the rolling plant 1 according to the rolling command from the higher-level computer 20. The process computer 21 outputs the calculated set values ​​to the controller 22. The set values ​​include the reduction position of the reduction device 42, the roll rotation speed, the bending force, the work roll shift amount, and the amount of cooling water in the cooling device 5.

[0036] When the rolled material 10 is transported to a predetermined position in front of each piece of equipment, the controller 22 operates the actuators (not shown) of each piece of equipment in the rolling plant 1 based on the set values. When rolling starts, the controller 22 sequentially operates each actuator so that the target dimensions and target temperature of the rolled material 10 conform to the rolling command, based on sensor measurements such as radiation thermometers, X-ray plate thickness gauges, and load cells.

[0037] There are no specific limitations on the structure of the process computer 21, but as an example, it may be as follows. Figure 2 shows an example of the hardware configuration of the process computer 21 provided in the rolling plant 1. The arithmetic processing function of the process computer 21 can be realized by the processing circuit shown in Figure 2. This processing circuit may be dedicated hardware 20a. This processing circuit may include a processor 20b and memory 20c. This processing circuit may be formed in part as dedicated hardware 20a and further include a processor 20b and memory 20c. In the example in Figure 2, part of the processing circuit is formed as dedicated hardware 20a, and the processing circuit also includes a processor 20b and memory 20c.

[0038] At least a portion of the processing circuit may be at least one dedicated hardware 20a. In this case, the processing circuit may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.

[0039] The processing circuit may include at least one processor 20b and at least one memory 20c. In this case, each function of the process computer 21 is realized by software, firmware, or a combination of software and firmware. The software and firmware are written as programs and stored in the memory 20c. The processor 20b realizes the functions of each part by reading and executing the programs stored in the memory 20c.

[0040] The processor 20b is also called a CPU (Central Processing Unit), central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, or DSP. Memory 20c includes non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, and EEPROM. It is also possible to configure memory 20c to also function as the database 23.

[0041] Thus, the processing circuit can realize each function of the process computer 21 through hardware, software, firmware, or a combination thereof. The functions of the process computer 21 also include machine learning functions, which will be described later.

[0042] In the rolling plant 1 described above, the material to be rolled 10 is heated in a heating furnace 2 and then drawn onto a roller table (not shown) on the rolling line. At this stage, the material to be rolled 10 is, for example, a steel billet. When the material to be rolled 10 reaches the roughing mill 3, it is repeatedly rolled while changing the rolling direction. At this stage, the material to be rolled 10 is, for example, a bar with a thickness of several tens of millimeters. Next, the material to be rolled 10 is sequentially fed into the rolling stands F1 to F5 of the finishing mill 4 and rolled to the desired plate thickness. At this stage, the material to be rolled 10 is also called a strip. After that, the material to be rolled 10 is cooled in a cooling device 5. The cooled material to be rolled 10 is wound up in a winding machine 6 to obtain a coiled product.

[0043] [Method for calculating plate thickness schedule according to the embodiment] Before the rolling process in the finishing rolling mill 4, the process computer 21 calculates the plate thickness schedule to be executed in the finishing rolling mill 4. The plate thickness schedule is calculated using a mathematical model. The plate thickness schedule includes the exit plate thickness of each rolling stand F1 to F5. This mathematical model is a set of formulas for predicting the temperature, rolling load, and rolling torque of each rolling stand F1 to F5. In the plate thickness schedule calculation based on the load ratio distribution method, the load ratio γ i The following is used: Load ratio γ i The load P at each rolling stand F1 to F5 is i This is the allocation ratio.

[0044] As described in the Conventional Technology, the rolling load in the "load ratio distribution method" is one of the factors that change the plate crown, and the higher the rolling load of a particular rolling stand, the larger the plate crown on the exit side of that rolling stand. Therefore, in order to minimize changes in the crown ratio and maintain good flatness, it is desirable that the rolling load changes in the same way at each stand. However, the rolling load changes moment by moment for each individual rolled material and each stand due to fluctuations in the temperature of the rolled material, which can lead to a deterioration in flatness. Therefore, a plate thickness schedule calculation method has been devised that automatically adjusts the plate thickness on the exit side of each stand and keeps the ratio of rolling loads (i.e., the rolling load ratio) as constant as possible, even if fluctuations in the temperature of the rolled material occur. According to this calculation method, when the rolling load fluctuates due to some disturbance, the trend of increase or decrease in the rolling load will be almost the same at all rolling stands, so the deterioration of flatness can be suppressed. Such a plate thickness schedule calculation method is called the "load ratio distribution method". The method for calculating the plate thickness schedule, excluding the calculation and setting of the limit values ​​for the limit check described later, is publicly known as described in the prior art, so no further explanation is provided here.

[0045] [Limit Check] In the calculation of the plate thickness schedule described above, known limit checks are performed on various parameters such as the reduction ratio, rolling load, and rolling torque.

[0046] In this embodiment, the upper and lower limits, which are the threshold values ​​for the reduction ratio in the limit check, are calculated as follows: The set values ​​or actual values ​​for each rolling stand F1 to F5 are extracted and used from the database 23, which stores past rolling data. Unless otherwise specified below, either the set values ​​or the actual values ​​may be used. Furthermore, calculations may be performed using both set values ​​and actual values, in which case the amount of data handled will be twice that of when only one of them is used.

[0047] [Calculation and setting of upper and lower limits of reduction ratio according to Example 1] Figure 3 is a flowchart illustrating the flow of plate thickness schedule calculation in Example 1. Figure 3 shows the flow of calculation for the upper and lower limits of the reduction ratio according to Example 1.

[0048] When the routine shown in Figure 3 is activated, past rolling data stored in database 23 is extracted as target data (step S1). There are two types of target data that can be extracted in step S1.

[0049] <Target Data (Part 1)> The first target data is rolling data from rolls that were rolled without any rolling troubles. This reduces the probability of rolling troubles and enables stable operation. The presence or absence of rolling troubles is limited to data that was measured when the rolled material 10 was wound into a coil by the winding machine 6.

[0050] <Target Data (Part 2)> The second set of target data consists of rolling data where the deviation between the set value and the actual value is small. This makes it possible to calculate highly accurate upper and lower limits. "Small deviation" means that the deviation (error ratio) between the actual value and the set calculated value is small, and is expressed by the following formula (1).

number

[0051] Here, ri SET is the reduction ratio set value (%) in the rolling stand i, r i ACT represents the actual reduction ratio value (%) in the rolling stand i. The deviation ε i r can be set arbitrarily. For example, a value of about 10% can be used.

[0052] For the above two types of target data, it is preferable to further narrow down to more detailed target data in consideration of the characteristics of the material. For example, the target data is classified for each product specification including the steel grade classification AAA and the plate thickness classification bb (hereinafter, this hierarchical classification is described as "classification (AAA, bb)"). Also, the classification may be added, such as the hierarchical classification (AAA, bb, cc) with the plate width classification cc added, to further subdivide the target data.

[0053] Next, the upper limit value and the lower limit value of the reduction ratio are calculated (step S2). In step S2, it is calculated for each rolling stand using the following statistical analysis or machine learning.

[0054] <Calculation of upper limit value and lower limit value (Part 1)> A combination of the average value and the standard deviation can be used as the statistical analysis. In this case, the average value and the standard deviation of the reduction ratio of the target data are calculated. For example, the average value r i AVE and the standard deviation r i STD are calculated by the following equations (2) and (3). <00003​​​​​​​​​​​​​​​​​​

[0056] When calculating the lower limit, the standard deviation from the mean is expressed as a constant β, as shown in equation (5) below. i Subtract the doubled value.

number

[0057] At this time, the constant β i This value is used as a common value in the calculation of the upper and lower limits, for example, β i Use =2, however, a different constant β at each rolling stand. i You may set it to that.

[0058] <Calculation of Upper and Lower Limits (Part 2)> Percentile values ​​can be used for statistical analysis. In this case, out of the target data (number of data points n) for the target category (AAA, bb), N pct,i (50 <N pct,i The percentile value at <100%) is used as the upper limit for calculation, and (100-N pct,i The percentile value at )% is used as the lower limit for calculation. For example, N pct,i If =95, the upper limit is the 95th percentile value and the lower limit is the 5th percentile value. pct,i This value may be a common value regardless of the rolling stand.

[0059] The reduction rate r of the 95th percentile value, which is the upper limit. i 95% This is calculated using the following formula (6).

number

[0060] At this time, (n+1) × 0.95 (=N pct,i If the integer part of ( / 100) is q and the decimal part is s, then the q-th data is r i q That is the case.

[0061] Similar to the upper limit, the reduction ratio r of the 5th percentile value, which is the lower limit, is also relevant. i 5% This is calculated using the following formula (7).

number

[0062] At this time, when the number of data points is n, (n+1) × 0.05 (= 1-N) pct,i If the integer part of ( / 100) is k and the decimal part is l, then the k-th data is r i k That is the case.

[0063] <Calculation of Upper and Lower Limits (Part 3)> Quartiles can be used for statistical analysis. In this case, the first quartile, second quartile, and interquartile range are calculated from the target data (number of data points n) for the target category (AAA, bb).

[0064] The first quartile r of the reduction ratio for n data points i Q1 and the third quartile r i Q3 In the aforementioned formula for calculating percentile values, N pct,i It is calculated as either 25 or 75. r i Q1 =r i 25% r i Q3 =r i 75%

[0065] Interquartile range of reduction ratio r i IQR r is the third quartile. i Q3 and the first quartile r i Q1 That is the difference.

number

[0066] Therefore, the third quartile r i Q3 The interquartile range r calculated using the above formula (8) i IQR A constant multiple of N Q,i The value obtained by adding this to the upper limit is calculated (see formula (9) below).

number

[0067] Furthermore, as shown in equation (10) below, the first quartile r i Q1 The interquartile range r i IQR A constant multiple of N Q,i The value obtained by subtracting this amount is used as the lower limit for calculation.

number

[0068] The constant multiple N at this time Q,i This is common to the calculation of the upper and lower limits, for example, N Q,i Use a value like =1.5. However, it may also be set at each rolling stand.

[0069] <Calculation of Upper and Lower Limits (Part 4)> The following data is used as input data to generate a machine learning reduction ratio calculation model, with the reduction ratio of each rolling stand F1 to F5 as output data: rolling orders including the steel type, target dimensions, and chemical composition of the slab, actual surface temperature data of the rolled material 10 before the finishing rolling mill 4, setting calculation data for material speed, rolling load, reduction ratio, and material temperature at each rolling stand F1 to F5, and actual rolling data such as material speed, rolling load, reduction ratio, plate thickness and plate thickness deviation at the exit of the finishing rolling mill, and material temperature at each rolling stand F1 to F5.

[0070] For machine learning methods, well-known machine learning techniques such as random forests can be used. Other examples of techniques include decision tree learning, neural networks, and support vector regression.

[0071] As shown in equation (11) below, the reduction ratio r output by machine learning i ML For this, a certain value c i ML (%) is added to determine the upper limit.

number

[0072] Furthermore, as shown in equation (12) below, the reduction ratio r output by machine learning i ML For this, a certain value c i ML Subtract the percentage (%) to obtain the lower limit.

number

[0073] This constant value c i ML (%) is used in common for calculating the upper and lower limits, for example, c i ML Use a value like (%)=5. However, it may also be set at each rolling stand.

[0074] Furthermore, the latest set values ​​and actual values ​​may be used as input and updated as needed each time the number of rolled materials (number of rolled material 10) increases.

[0075] In step S2, at least one of the upper and lower limits of the reduction ratio is set as the limit value (limit range), which is the threshold for the limit check in the plate thickness schedule calculation (step S3).

[0076] [Calculation and setting of upper and lower limits of reduction ratio according to Example 2] FIG. 4 is a flowchart for explaining the flow of sheet thickness schedule calculation in Example 2. FIG. 4 shows the flow of calculation of the upper limit value and the lower limit value of the reduction ratio according to Example 2. In the above Example 1, data without rolling troubles or data with a deviation between the set value and the actual value smaller than the reference value were extracted as target data, whereas in this Example 2, it is different in that rolling data when rolling troubles occur in passing the tip part is also considered.

[0077] When the routine shown in FIG. 4 is driven, among the past rolling data stored in the database 23, rolling data of the layer classification (TTT, bd) (number of data N) in which rolling troubles occurred (relatively many rolling troubles) is extracted as target data (step S11). The target data extracted in step S11 is divided into data that could be rolled without rolling troubles (number of data N OK ) and data with rolling troubles (number of data N NG ), and it is determined which data it is (step S12).

[0078] Next, for the N OK pieces of data that could be rolled without rolling troubles, the average value of the reduction ratio in the setting calculation of each rolling stand is calculated by the following formula (13) (step S13).

Equation

[0079] On the other hand, for the N NG pieces of data with rolling troubles, the percentile value r pct,i NG at N i Npct,i NG % and the percentile value r pct,i NG at (100 - N i 100-Npct,i NG )% of the reduction ratio in the setting calculation of each rolling stand are calculated (step S14). However, here, N pct,i NG < 50 is assumed. The percentile value is calculated by the above formula (6).

[0080] Next, based on a numerical value close to the average value of the reduction rate, the upper limit value and the lower limit value are calculated (step S15).

[0081] For example, when r i OK,AVE <r i Npct,i NG when the reduction rate becomes larger than the average value, it means that the possibility of rolling trouble is high. Therefore, the upper limit value is calculated by the following formula (14).

Equation

[0082] On the other hand, when r i OK,AVE >r i 100-Npct,i NG when the reduction rate becomes smaller than the average value, it means that the possibility of rolling trouble is high. Therefore, the lower limit value is calculated by the following formula (15).

Equation

[0083] This constant numerical value c i OK (%) is common in the calculation of the upper limit value and the lower limit value, and any numerical value can be used. However, a value smaller than (r i Npct,i NG -r i OK,AVE ) or a value smaller than (r i OK,AVE -r i 100-Npct,i NG ) is desirable. Also, the numerical value c i OK (%) may be set for each rolling stand.

[0084] In step S16, at least one of the upper and lower limits of the reduction ratio calculated in step S15 is set as the limit value (limit range), which is the threshold for the limit check in the plate thickness schedule calculation.

[0085] [Calculation and setting of upper and lower limits of reduction ratio according to Example 3] Figure 5 is a flowchart illustrating the flow of plate thickness schedule calculation in Example 3. Figure 5 shows the calculation flow for the upper and lower limits of the reduction ratio according to Example 3.

[0086] The routine shown in Figure 5 is activated when the upper and lower limits of the reduction ratio are calculated in the above examples 1 and 2. If the calculated upper or lower limit of the reduction ratio for a certain rolling stand is greater than the upper limit of the reduction ratio for an upstream rolling stand, a reversal of the reduction ratio may occur, disrupting the rolling balance in the finishing rolling mill 4 and potentially causing rolling problems.

[0087] Therefore, this embodiment specifies how to handle the case where the calculated upper limit of the reduction ratio is larger than that of the upstream rolling stand. In this case, although both are based on past data and are reliable data, it is desirable for the reduction ratio of the downstream rolling stand to be smaller than that of the upstream rolling stand, so the process shown in Figure 5 is executed.

[0088] When the routine shown in Figure 5 is activated, it is determined whether the upper limit of the reduction ratio calculated at a certain rolling stand i is greater than the upper limit of the reduction ratio calculated at the rolling stand i-1 one step upstream, that is, whether the following equation (16) is true (step S17).

number

[0089] If relation (16) holds, the upper limit of the downstream rolling stand i is replaced as shown in equation (17) below (step S18).

number

[0090] In step S18, a constant value α is used, as shown in equation (18) below, so that the upper limit is lower than that of the rolling stand one step upstream. i You can also subtract (%).

number

[0091] This constant value α i For example, α i Use a value like =3. However, the value α will be used for each rolling stand. i You may set it to that.

[0092] Thus, the new upper limit of the reduction ratio calculated using equation (17) above or equation (18) above is set as the limit value (limit range), which is the threshold for limit checking in the plate thickness schedule calculation.

[0093] [Calculation and setting of upper and lower limits of reduction ratio according to Example 4] Figure 6 is a flowchart illustrating the flow of plate thickness schedule calculation in Example 4. Figure 6 shows the calculation flow of the upper and lower limits of the reduction ratio according to Example 4.

[0094] In this embodiment 4, the calculation method for the upper and lower limits of the reduction ratio is switched in stages according to the number of past rolling data N stored in the database 23.

[0095] When the routine shown in Figure 6 is activated, it is determined whether the number of rolling data N stored in the database 23 is less than the threshold Nb1 (step S21). If the number of rolling data N is less than the threshold Nb1, the upper and lower limits of the reduction ratio, which have been calculated in advance, are set as limit values, as in the conventional method (step S22).

[0096] In step 21, if the number of rolling data points N is greater than or equal to the threshold Nb1, it is determined whether the number of rolling data points N is less than or equal to the threshold Nb2 (>Nb1) (step S23). If the number of rolling data points N is less than the threshold Nb2, i.e., if the accumulated number of rolling data points N is small, the upper and lower limits of the reduction ratio are calculated using the statistical analysis <Calculation of upper and lower limits (any one of methods 1 to 3)> described in Example 1 above (step S24). If the accumulated number of rolling data points N increases and the number of rolling data points N reaches greater than or equal to the threshold Nb2, the upper and lower limits of the reduction ratio are calculated using the machine learning <Calculation of upper and lower limits (method 4)> described in Example 1 above (step S25). Subsequently, at least one of the upper and lower limits of the reduction ratio calculated in step S24 or step S25 is set as the limit value (limit range) (step S26).

[0097] According to this embodiment, at least one of the upper and lower limits of the reduction ratio, calculated by a stepwise calculation corresponding to the number of rolling data stored in the database 23, can be used as a limit value.

[0098] Although embodiments and examples of the present invention have been described above, the present invention is not limited to the above embodiments and examples, and can be implemented in various modified forms without departing from the spirit of the invention. When numbers such as the number of elements, quantities, amounts, or ranges are mentioned in the embodiments described above, the present invention is not limited to the numbers mentioned unless they are specifically stated or clearly defined in principle. Furthermore, structures and the like described in the embodiments described above are not necessarily essential to the present invention unless they are specifically stated or clearly defined in principle. [Explanation of Symbols]

[0099] 1...Rolling plant, 4...Finishing rolling mill (tandem rolling mill), 21...Process computer, 23...Database, 41...Rolls, 42...Reduction device, 43...Electric motor, F1~F5...Rolling stand

Claims

1. In a method for calculating the plate thickness schedule of a tandem rolling mill that continuously rolls plates using multiple rolling stands, The process includes a calculation step to calculate the plate thickness schedule based on the rolling load or motor power, using a rolling model equation that includes an advance rate model and a rolling load model. The calculation process includes a calculation step of calculating at least one of the upper and lower limits of the reduction ratio at each rolling stand based on past rolling data, for each product specification including the steel type and plate thickness to be rolled, and a setting step of setting at least one of the upper and lower limits calculated in the calculation step as the limit value for the limit check on the reduction ratio. The calculation process described above is: The process includes extracting rolling data from past rolling data when rolling was performed without any rolling problems. At least one of the set calculated value and the actual value of the reduction ratio of the extracted rolling data is used to calculate at least one of the upper limit and the lower limit by statistical analysis. Using machine learning, which takes at least one of the calculated and actual reduction ratios of the extracted rolling data as input and outputs the reduction ratio, the machine sequentially calculates the reduction ratio, and then calculates at least one of the upper and lower limits by adding or subtracting a certain value to the calculated reduction ratio. The aforementioned setting process is: When the number of past rolling data is less than the first threshold, the reduction ratio calculated in advance is set as the limit value. When the number of past rolling data is greater than or equal to the first threshold and less than the second threshold, at least one of the upper limit and lower limit values ​​calculated by statistical analysis from at least one of the set calculated value and actual value of the rolling reduction ratio of the rolling data is set as the limit value. A plate thickness schedule calculation method in which, when the number of past rolling data is equal to or greater than the second threshold, at least one of the upper limit and lower limit values ​​calculated by adding or subtracting a certain value to the reduction ratio calculated using machine learning is set as the limit value.

2. In a method for calculating the plate thickness schedule of a tandem rolling mill that continuously rolls plates using multiple rolling stands, The process includes a calculation step to calculate the plate thickness schedule based on the rolling load or motor power, using a rolling model equation that includes an advance rate model and a rolling load model. The calculation process includes a calculation step of calculating at least one of the upper and lower limits of the reduction ratio at each rolling stand based on past rolling data, for each product specification including the steel type and plate thickness to be rolled, and a setting step of setting at least one of the upper and lower limits calculated in the calculation step as the limit value for the limit check on the reduction ratio. The calculation process described above is: The process includes extracting rolling data from the aforementioned past rolling data when the deviation between the calculated value of the reduction ratio and the actual value is smaller than a standard value. At least one of the set calculated value and the actual value of the reduction ratio of the extracted rolling data is used to calculate at least one of the upper limit and the lower limit by statistical analysis. Using machine learning, which takes at least one of the calculated and actual reduction ratios of the extracted rolling data as input and outputs the reduction ratio, the machine sequentially calculates the reduction ratio, and then calculates at least one of the upper and lower limits by adding or subtracting a certain value to the calculated reduction ratio. The aforementioned setting process is: When the number of past rolling data is less than the first threshold, the reduction ratio calculated in advance is set as the limit value. When the number of past rolling data is greater than or equal to the first threshold and less than the second threshold, at least one of the upper limit and lower limit values ​​calculated by statistical analysis from at least one of the set calculated value and actual value of the rolling reduction ratio of the rolling data is set as the limit value. A plate thickness schedule calculation method in which, when the number of past rolling data is equal to or greater than the second threshold, at least one of the upper limit and lower limit values ​​calculated by adding or subtracting a certain value to the reduction ratio calculated using machine learning is set as the limit value.

Images