Method for calculating rolling conditions of cold rolling mill, device for calculating rolling conditions of cold rolling mill, cold rolling method, cold rolling mill, and method for manufacturing steel sheet

Abstract

The present application relates to a rolling condition calculation method for a cold rolling mill, including: a step of estimating, to a prediction model learned by taking first multidimensional data generated based on non-stable rolling performance data in past rolling performance in which a rolling material is rolled by the cold rolling mill as explanatory variables and taking stable rolling performance data and rolling constraint condition data at the time of stable rolling as target variables, second multidimensional data generated based on non-stable rolling performance data of a rolling target material, a rolling constraint condition for a target stable rolling condition with respect to the rolling target material; and a step of changing the target stable rolling condition in such a manner that the estimated rolling constraint condition satisfies a prescribed condition.

Description

Technical Field

[0001] This invention relates to a method for calculating rolling conditions of a cold rolling mill, a device for calculating rolling conditions of a cold rolling mill, a cold rolling method, and a method for manufacturing a cold rolling mill and steel plates. Background Technology

[0002] Generally, in cold rolling of thin steel sheets, rolling is carried out under conditions that stabilize the sheet's passability by maintaining good thickness accuracy in both the length and width directions and ensuring good shape (or flatness). Many control elements of the cold rolling mill are automatically controlled by actuators mounted on the mill, reducing the opportunities for operators to set these control elements.

[0003] On the other hand, there is a growing demand for high-strength, thin-walled, and rigid materials with the aim of reducing fuel consumption through lightweighting. In the cold rolling of such high-load, difficult-to-roll materials, the rolling load (and consequently, the calculated slip ratio and torque), roll gap, work roll bending, intermediate roll displacement, and roll deflection correction (represented by roll expansion due to thermal crown) of the cold rolling mill often become the operating conditions that limit the specifications of the cold rolling mill. In this case, the operator sets the pass planning and rolling speed to not exceed the equipment constraints of the cold rolling mill.

[0004] In addition, depending on the process conditions of the preceding rolling steps, the equipment status of the cold rolling mill, and changes in the coolant status, mill vibrations in the vertical or horizontal directions, caused by insufficient or excessive lubrication, can occur at frequencies of approximately 30Hz to 200Hz. This can easily lead to periodic variations in the thickness of the steel plate. In such cases, the operator must set a rolling speed that meets the equipment constraints of the cold rolling mill without hindering productivity.

[0005] Therefore, in recent years, the operating speed and, consequently, productivity of cold rolling mills have become susceptible to the influence of operator experience. Against this backdrop, Patent Document 1 proposes a method for using a neural network to learn past operating conditions and using the learning results to set up the rolling mill.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: Japanese Patent No. 6705519 Summary of the Invention

[0009] The problem that the invention aims to solve

[0010] However, even if the mill becomes the most suitable operating condition at the set time, if the rolling speed is increased during operation, the operating conditions will change due to the lubrication state in the roll gap, the thermal expansion of the rolls, the increase in the temperature of the steel plate, etc., and sometimes it is difficult to achieve the rolling speed that can be achieved through coils.

[0011] The present invention was made in view of the above-mentioned problems, and its object is to provide a method and apparatus for calculating rolling conditions for a cold rolling mill that can calculate rolling conditions to ensure rolling stability and not hinder productivity when rolling difficult-to-roll materials under high loads. Furthermore, another object of the present invention is to provide a cold rolling method and cold rolling mill capable of stably and productively rolling difficult-to-roll materials under high loads. Additionally, another object of the present invention is to provide a method for manufacturing steel plates that can produce steel plates with desired product characteristics and mechanical properties with high yield.

[0012] Methods for solving problems

[0013] The method for calculating rolling conditions of a cold rolling mill according to the present invention includes: an estimation step, wherein a prediction model is obtained by learning a prediction model based on first multidimensional data generated from unstable rolling performance data in past rolling performances of materials rolled using a cold rolling mill, using stable rolling performance data and rolling constraint data during stable rolling as target variables, and inputting second multidimensional data generated from unstable rolling performance data of the material to be rolled, to estimate rolling constraint conditions relative to target stable rolling conditions of the material to be rolled; and a modification step, wherein the target stable rolling conditions are modified in such a way that the estimated rolling constraint conditions satisfy specified conditions.

[0014] Preferably, the rolled material has a welded portion formed by welding the preceding material and the following material together, and the unstable rolling performance data is the rolling performance data when the welded portion passes through the exit side of the cold rolling mill.

[0015] Preferably, the stable rolling performance data includes data representing the target rolling speed during stable rolling of previously rolled materials, and the modification step includes changing the target rolling speed during stable rolling of the material to be rolled in a manner that satisfies the estimated rolling constraints.

[0016] The rolling condition calculation device for a cold rolling mill according to the present invention comprises: an estimation unit that estimates rolling constraints relative to a target stable rolling condition for the material to be rolled by inputting second multidimensional data generated based on the unstable rolling performance data of the material to be rolled into a prediction model that has been learned by learning a prediction model based on first multidimensional data generated from unstable rolling performance data of past rolling performances of the material to be rolled using a cold rolling mill as explanatory variables and stable rolling performance data and rolling constraint condition data during stable rolling as target variables; and a modification unit that modifies the target stable rolling condition in a manner that makes the estimated rolling constraint condition satisfy a specified condition.

[0017] The cold rolling method of the present invention includes the step of rolling the material to be rolled using a target stable rolling condition modified by the rolling condition calculation method of the cold rolling mill of the present invention.

[0018] The cold rolling mill of the present invention is equipped with the rolling condition calculation device of the cold rolling mill of the present invention.

[0019] The steel plate manufacturing method of the present invention includes: a cold rolling process, wherein the material is cold rolled using the cold rolling method of the present invention; and an annealing process, wherein the rolled material after cold rolling is subjected to a homogenization temperature of 600–950°C and an in-furnace tension of 0.1–3.0 kgf / mm. 2 Annealing process.

[0020] Invention Effects

[0021] According to the rolling condition calculation method and apparatus for cold rolling mills disclosed in this invention, rolling conditions that ensure rolling stability and do not hinder productivity can be calculated even when rolling difficult-to-roll materials under high loads. Furthermore, according to the cold rolling method and cold rolling mill disclosed in this invention, difficult-to-roll materials under high loads can be rolled stably and with high productivity. Additionally, according to the steel plate manufacturing method disclosed in this invention, steel plates with desired product characteristics and mechanical properties can be manufactured with a high yield. Attached Figure Description

[0022] Figure 1 This is a schematic diagram showing the structure of a cold rolling mill as an embodiment of the present invention.

[0023] Figure 2 It is shown Figure 1 The diagram shows the structure of the arithmetic unit.

[0024] Figure 3 This is a diagram illustrating an example of multidimensional array information.

[0025] Figure 4 This is a diagram illustrating a structural example of a stable rolling speed prediction model.

[0026] Figure 5 This is a flowchart illustrating the process of transforming multidimensional array information into one-dimensional information.

[0027] Figure 6 This is a flowchart illustrating the processing flow of the prediction model execution unit. Detailed Implementation

[0028] Hereinafter, with reference to the accompanying drawings, a method for calculating rolling conditions for a cold rolling mill, a apparatus for calculating rolling conditions for a cold rolling mill, a cold rolling method, and a cold rolling mill, as embodiments of the present invention, will be described. It should be noted that the embodiments shown below exemplify apparatus and methods used to concretize the technical concept of the present invention, and do not specifically define the materials, shapes, structures, arrangements, etc., of the constituent components as shown in the following embodiments. Furthermore, the accompanying drawings are schematic. Therefore, it should be noted that the relationships and ratios between thickness and planar dimensions differ from reality, and the accompanying drawings also include portions where the dimensional relationships and ratios differ from each other.

[0029] [Structure of a cold rolling mill]

[0030] First, refer to Figure 1 The structure of a cold rolling mill, as one embodiment of the present invention, will be described below. It should be noted that in this specification, "cold rolling" is sometimes simply referred to as "rolling," and both "cold rolling" and "rolling" are synonymous. Furthermore, in the following description, steel plate is used as an example of the material to be rolled by the cold rolling mill. However, the rolling material is not limited to steel plate; other metal strips such as aluminum plates can also be used.

[0031] Figure 1 This is a schematic diagram showing the structure of a cold rolling mill as an embodiment of the present invention. Figure 1 As shown, the cold rolling mill 1, as an embodiment of the present invention, is installed from the inlet side (towards) of the steel plate S. Figure 1 The left side of the paper faces outwards (towards the paper). Figure 1 The cold rolling mill 1 (as shown on the right side of the paper) has five rolling stands in sequence, namely the first rolling stand to the fifth rolling stand (#1STD to #5STD). In this cold rolling mill 1, tension rolls and differential rolls (not shown), a plate thickness gauge and a shape gauge are appropriately arranged between adjacent rolling stands. The structure of the rolling stands, the conveying device for the steel plate S, etc., are not particularly limited, and known technologies can be appropriately applied.

[0032] Emulsion rolling oil (in the following description, "emulsion rolling oil" will sometimes be referred to simply as "rolling oil") 13 is supplied to each rolling stand of the cold rolling mill 1. The cold rolling mill 1 is equipped with a dirty tank (recycling tank) 5 and a clean tank 7 as rolling oil storage tanks, and the rolling oil supplied from these rolling oil storage tanks is supplied to each rolling stand through the supply line 9.

[0033] The rolling oil recovered by the oil pan 10 located below the first to fifth rolling mill stands, i.e. the rolling oil used in cold rolling, flows into the dirty tank 5 through the return pipe 11.

[0034] The rolling oil stored in the clean tank 7 is produced by mixing warm water (diluent) and the original rolling oil (with added surfactant). The mixture of warm water and original rolling oil is adjusted by changing the rotation speed of the stirring blades of the mixer 12, that is, by adjusting the degree of stirring, to produce rolling oil with a desired average particle size and concentration range.

[0035] As a base fluid for rolling, it can be used in common cold rolling processes, such as base fluids based on natural oils, fatty acid esters, and hydrocarbon-based synthetic lubricants. Furthermore, for these rolling fluids, additives commonly used in cold rolling fluids, such as oiliness improvers, extreme pressure additives, and antioxidants, can be added.

[0036] As the surfactant added to the rolling oil, either ionic or nonionic surfactants can be used, and any surfactant used in conventional circulating coolant systems (circulating rolling oil supply methods) is acceptable. Furthermore, it is preferable to dilute the stock rolling oil to a concentration of 2-8% by mass, more preferably 3-6.0% by mass, and then use the surfactant to form an O / W emulsion rolling oil in which the oil is dispersed in water. It should be noted that the average particle size of the rolling oil is preferably 15 μm or less, more preferably 3-10 μm.

[0037] After operation begins, the rolling oil recovered to the dirty tank 5 flows into the clean tank 7 via an iron powder removal device 6, which consists of an iron powder quantity control device and the like. The rolling oil recovered to the dirty tank 5 contains wear powder (iron powder) generated by the friction between the rolls and the steel plate S. Therefore, the iron powder removal device 6 removes the wear powder in a manner that reduces the soluble iron content of the recovered rolling oil to a level permissible for the rolling oil to be stored in the clean tank 7.

[0038] The movement of the emulsion rolling oil from the dirty tank 5 side to the clean tank 7 side via the iron powder removal device 6 can be continuous or intermittent. While it is preferable, the iron powder removal device 6 is a device that uses a magnetic filter, such as an electromagnetic filter or a magnetic separator, to adsorb and remove iron powder, but is not limited to this. The iron powder removal device 6 can also be a known device using methods such as centrifugal separation.

[0039] However, some of the rolling oil supplied to the rolling mill stand is carried out of the system by the steel plate S or lost due to evaporation. Therefore, the clean tank 7 is a structure that appropriately replenishes (supplyes) the raw rolling oil from the raw oil tank (not shown) so that the accumulation level and concentration of the rolling oil in the clean tank 7 are within a specified range. In addition, warm water for diluting the rolling oil is also appropriately replenished (supplyed) to the clean tank 7. It should be noted that the accumulation level and concentration of the emulsion rolling oil in the clean tank 7 can be measured using a sensor (not shown).

[0040] Next, details will be provided regarding the rolling oil supply system of the cold rolling mill 1. The rolling oil supply system of the cold rolling mill 1 includes a dirty tank 5, an iron powder removal device 6, a clean tank 7, and a pump 8 that pumps rolling oil from the clean tank 7. It should be noted that a filter for removing foreign matter may also be installed between the clean tank 7 and the pump 8.

[0041] The rolling oil supply system of the cold rolling mill 1 has a supply line 9 with one end connected to the clean tank 7 and five sets of lubricating coolant heads 3 and five sets of cooling coolant heads 4 that branch off at the other end of the supply line 9 (on the mill side) and are respectively arranged at positions corresponding to each stand.

[0042] Each lubrication coolant head 3 is located on the inlet side of the rolling mill stand, supplying lubricating oil to the roll gap and work rolls by spraying rolling oil as lubricant through separately provided nozzles toward the roll gap. A cooling coolant head 4 is located on the outlet side of the rolling mill stand, cooling the rolls by spraying rolling oil through separately provided nozzles toward the rolls.

[0043] With this structure, the emulsion rolling oil in the clean tank 7 is pumped by the pump 8 to the supply line 9, and supplied to the lubrication coolant head 3 and cooling coolant head 4 located on each rolling stand, and supplied to the spraying parts from the separately provided nozzles. In addition, the emulsion rolling oil supplied to the rolls, except for the portion carried out of the system by the steel plate S or lost due to evaporation, is recovered by the oil pan 10 and returned to the dirty tank 5 via the return pipe 11. Afterwards, a portion of the emulsion rolling oil accumulated in the dirty tank 5 is returned to the clean tank 7 after a certain amount of soluble iron components generated by cold rolling are removed by the iron powder removal device 6.

[0044] The rolling oil supply system described above circulates rolling oil to the rolls after the wear parts have been removed. In other words, the supplied emulsion rolling oil is recycled. It should be noted that the clean tank 7 corresponds to the circulating rolling oil tank in conventional circulating oil supply methods. As described above, the clean tank 7 is appropriately replenished (supplied) with the original rolling oil.

[0045] [Stable Rolling Speed ​​Prediction Model]

[0046] Next, a stable rolling speed prediction model, which is one embodiment of the present invention, will be described.

[0047] The functions associated with the stable rolling speed prediction model, which is one embodiment of the present invention, are provided by Figure 1 The rolling control device 14 and the arithmetic unit 15 shown are implemented.

[0048] The rolling control device 14 controls the rolling conditions of the cold rolling mill 1 based on the control signals from the arithmetic unit 15.

[0049] Figure 2 It is shown Figure 1 The diagram shows the structure of the arithmetic unit 15. Figure 2 As shown, the arithmetic unit 15 includes an arithmetic device 71, an input device 88, a storage device 89, and an output device 90.

[0050] The arithmetic unit 71 is wired to the input device 88, the storage device 89, and the output device 90 via the bus 87. However, the arithmetic unit 71, the input device 88, the storage device 89, and the output device 90 are not limited to this connection method; they can also be connected wirelessly, or a combination of wired and wireless methods.

[0051] The input device 88 functions as an input port for inputting control information based on the rolling control device 14 and information from the operation monitoring device 91. The information from the operation monitoring device 91 includes execution command information of the stable rolling speed prediction model, information related to the steel plate S to be rolled (previous process conditions, steel grade, size), and cold rolling condition information (numerical information, text information, and image information) set by the process control computer or operator before cold rolling.

[0052] The storage device 89 is composed of, for example, a hard disk drive, a semiconductor drive, an optical drive, etc., and is a device for storing the information required in this system (the information required for the realization of the functions of the prediction model creation unit 77 and the prediction model execution unit 78, which will be described later).

[0053] Information required for the realization of the functions of the prediction model production unit 77 includes, for example, information related to cold rolling such as the required characteristics of the steel plate S (steel grade, plate thickness, plate width, etc.), rolling mill equipment constraints, rolling information after the weld point of the steel plate S passes through, coolant properties used by the rolling mill stand, target rolling speed, and other explanatory variables and target variables (constraint determination relative to the target rolling speed).

[0054] Information required for the implementation of the functions of the prediction model execution unit 78 includes, for example, the stable rolling speed prediction model for each rolling state of the steel plate S produced by the prediction model production unit 77 and various information input to the stable rolling speed prediction model.

[0055] The output device 90 functions as an output port for the rolling control device 14 to output control signals from the computing device 71.

[0056] The operation monitoring device 91 is equipped with any display such as an LCD or an OLED. The operation monitoring device 91 receives various information indicating the operating status of the cold rolling mill 1 from the rolling control device 14 and displays this information on an operation screen (operation screen) for the operator to monitor the operating status of the cold rolling mill 1.

[0057] The arithmetic unit 71 includes RAM 72, ROM 73 and arithmetic processing unit 76.

[0058] ROM 73 stores a prediction model creation program 74 and a prediction model execution program 75.

[0059] The arithmetic processing unit 76 has arithmetic processing functions and is connected to RAM 72 and ROM 73 via bus 87.

[0060] RAM 72, ROM 73 and arithmetic processing unit 76 are connected to input device 88, storage device 89 and output device 90 via bus 87.

[0061] The computation processing unit 76 is a functional block that includes a prediction model creation unit 77 and a prediction model execution unit 78.

[0062] The prediction model generation unit 77 is a processing unit that generates a stable rolling speed prediction model based on a machine learning method. This machine learning method involves establishing a correlation between past rolling performance in the cold rolling mill 1 and the rolling constraints corresponding to those past performances. In this embodiment, a neural network model is used as the stable rolling speed prediction model based on the machine learning method. The machine learning method is not limited to neural networks; other known machine learning methods may also be employed.

[0063] The prediction model creation unit 77 includes a learning data acquisition unit 77A, a preprocessing unit 77B, a first data transformation unit 77C, a model creation unit 77D, and a result storage unit 77E. When the prediction model creation unit 77 receives an instruction from the operation monitoring device 91 to create a stable rolling speed prediction model, it executes the prediction model creation program 74 stored in the ROM 73, thereby functioning as the learning data acquisition unit 77A, the preprocessing unit 77B, the first data transformation unit 77C, the model creation unit 77D, and the result storage unit 77E. The stable rolling speed prediction model is updated each time the prediction model creation unit 77 executes.

[0064] The learning data acquisition unit 77A, as a preprocessing step for generating a stable rolling speed prediction model, acquires multiple learning data sets. These data sets use unstable rolling performance data from past rolling performance data as input performance data (description variables) and stable rolling performance data and rolling constraint condition data during stable rolling as output performance data (target variables). Specifically, the learning data acquisition unit 77A acquires multiple learning data sets that use operational performance data of rolling conditions after the weld point passes (rolling performance data when the welded section formed by welding the preceding and following materials passes through the exit side of the cold rolling mill) and the target rolling speed as input performance data, and use constraint determination values ​​at the target rolling speed during cold rolling using this input performance data as output performance data. The learning data acquisition unit 77A acquires the aforementioned input performance data and output performance data from the storage device 89 to create learning data. Each learning data set consists of a group of input performance data and output performance data. The learning data is stored in the storage device 89. The learning data acquisition unit 77A may also supply the learning data to the preprocessing unit 77B and the model making unit 77D without storing the learning data to the storage device 89.

[0065] The input performance data includes multidimensional array information obtained by linking explanatory variables along the time axis. In this embodiment, the multidimensional array information is provided by, for example... Figure 3 The information shown in (a) to (c).

[0066] Figure 3 (a) shows an example of input performance data where the columns (vertical direction) are composed of roll material conditions and the rows (horizontal direction) are composed of explanatory variables selected from the operating conditions of the cold rolling mill 1. The number of columns of explanatory variables is not particularly limited. In this embodiment, a multidimensional array of information is created by linking each roll material condition in the time direction (depth direction) and used as input performance data. The time interval when linking in the time direction is not particularly limited. In the case of no time series data or explanatory variables that do not change over time, the time interval is 0 seconds, and the same data is copied.

[0067] Figure 3 (b) shows an example of input data where the vertical column represents the elapsed time after a weld point of a certain coil passes through, and the horizontal column represents the input data of explanatory variables selected from the operating conditions of cold rolling mill 1. Figure 3 Similarly, in example (a), a multidimensional array of information is created by linking each roll condition in the time direction (depth direction) to become the input performance data.

[0068] Figure 3 (c) shows an example where the vertical column represents the elapsed time after the weld point of a certain coil passes through, and the horizontal column represents the input performance data of the explanatory variables selected from the operating conditions of cold rolling mill 1. The input performance data only needs to be linked in the time direction (depth direction) and does not need to be the same coil. In this example, a multidimensional array of time series operation information of multiple coils is created and used as the input performance data.

[0069] It should be noted that when the storage device 89 does not store past rolling performance data (e.g., rolling conditions or steel grade conditions for which there are no past performance records) or the sample size is small, the learning data acquisition unit 77A may request the operator once or multiple times to perform cold rolling without using the stable rolling speed prediction model. Furthermore, the more learning data stored in the storage device 89, the higher the prediction accuracy based on the stable rolling speed prediction model. Therefore, even if the amount of learning data is less than a preset threshold, the learning data acquisition unit 77A may request the operator to perform cold rolling without using the stable rolling speed prediction model until the data amount reaches the threshold.

[0070] The preprocessing unit 77B processes the learning data acquired by the learning data acquisition unit 77A into a stable rolling speed prediction model. Specifically, in order to enable the neural network model to read the rolling performance data constituting the learning data, the preprocessing unit 77B standardizes (normalizes) the value range of the input performance data between 0 and 1 as needed.

[0071] The input performance data is multidimensional information. Therefore, the first data transformation unit 77C uses a convolutional neural network 300 to compress the dimensionality of the input performance data while retaining the feature quantities, making it one-dimensional information (see reference). Figure 4 The input performance data has been transformed into one-dimensional information. Figure 4 The input layer 101 shown is combined.

[0072] Here, refer to Figure 5 The processing example of the first data conversion unit 77C will be explained. Figure 5 This is a flowchart illustrating the process of transforming multidimensional array information into one-dimensional information. For example... Figure 5As shown, the process of transforming multidimensional array information into one-dimensional information, i.e., the method for storing multidimensional array information, has a structure with multiple filters whose inputs and outputs are interconnected in multiple stages. That is, the process of transforming multidimensional array information into one-dimensional information includes, from the input side, a first convolution step S10, a first pooling step S11, a second convolution step S12, a second pooling step S13, and a fully connected step S14.

[0073] In the first convolution step S10, the first data transformation unit 77C takes a 64×64 multidimensional array of information as input and outputs a 64×64 first feature map through convolution operation. The first feature map indicates what kind of local features exist in which part of the input array. In the convolution operation, for example, a 3×3 pixel, 32-channel filter is used, the filter application interval is set to 1, and the length of the surrounding zero-filling is set to 1.

[0074] In the first pooling step S11, the first data transformation unit 77C takes the first feature map output from the first convolution step S10 as input and sets the maximum value within the horizontal × vertical 3 pixels of the first feature map as a new pixel. The first data transformation unit 77C performs this operation pixel-shifting and throughout the entire map. Thus, in the first pooling step S11, the first data transformation unit 77C outputs a second feature map obtained by compressing the first feature map.

[0075] In the second convolution step S12, the first data transformation unit 77C takes the second feature map as input and outputs the third feature map through convolution operation. In the convolution operation, for example, a filter with 3 pixels horizontally and 3 pixels vertically and 16 channels is set, the application interval of the filter is set to 1, and the length of the surrounding zero-filling is set to 1.

[0076] In the second pooling step S13, the first data transformation unit 77C takes the third feature map output from the second convolution step S12 as input and sets the maximum value within the horizontal × vertical 3 pixels of the third feature map as a new pixel. The first data transformation unit 77C performs this operation pixel-shifting and throughout the entire map. Thus, in the second pooling step S13, the first data transformation unit 77C outputs a fourth feature map obtained by compressing the third feature map.

[0077] In the fully connected step S14, the first data transformation unit 77C arranges the information of the fourth feature map output from the second pooling step S13 into a column. Furthermore, the 100 neurons output from the fully connected step S14 become the input layer 101. It should be noted that the convolution method and the number of output neurons are not limited to those described above. Additionally, known models such as GoogleNet, VGG16, MOBILENET, and EFFICIENTNET can also be used as convolutional neural network methods.

[0078] The model making unit 77D uses machine learning (including information transformed by the first data transformation unit 77C) to obtain multiple learning data from the preprocessing unit 77B to generate a stable rolling speed prediction model. The stable rolling speed prediction model includes unstable rolling performance data as input performance data and rolling constraint determination value under the target rolling speed as output performance data.

[0079] In this embodiment, since a neural network is used as a machine learning method, the model making unit 77D creates a neural network model as a stable rolling speed prediction model. That is, the model making unit 77D creates a neural network model as a stable rolling speed prediction model, and this model is correlated with the input performance data (including rolling performance data at the target rolling speed) and the output performance data (rolling constraint determination values ​​at the target rolling speed) from the learning data used for creating the stable rolling speed prediction model. The neural network model is, for example, represented by a function.

[0080] Specifically, the model creation department 77D sets the hyperparameters used in the neural network model and performs learning based on the neural network model using these hyperparameters. As an optimization calculation of hyperparameters, the model creation department 77D first creates a neural network model with some changes to the hyperparameters in stages on the learning data, and selects the hyperparameters with the highest prediction accuracy relative to the validation data.

[0081] As hyperparameters, the number of hidden layers, the number of neurons in each hidden layer, the dropout rate in each hidden layer (cutting off the transmission of neurons with a certain probability), the activation function in each hidden layer, and the number of outputs are typically set, but are not limited to these. Furthermore, the optimization method for hyperparameters is not particularly limited, but grid search with phased parameter changes, random search with randomly selected parameters, or Bayesian optimization-based search can be used.

[0082] It should be noted that the model making unit 77D is installed as part of the computing unit 71, but the structure is not limited to this. For example, stable rolling speed prediction models can also be made and saved in advance and read out appropriately.

[0083] like Figure 4 As shown, the neural network model that serves as the stable rolling speed prediction model in this embodiment has an input layer 101, an intermediate layer 102, and an output layer 103 sequentially from the input side.

[0084] exist Figure 3The multidimensional array information created in the learning data acquisition unit 77A is compressed into a state with retained feature quantities by using a convolutional neural network, and then saved as a state of one-dimensional information to the input layer 101.

[0085] The intermediate layer 102 consists of multiple hidden layers, each containing multiple neurons. The number of hidden layers within the intermediate layer 102 and the number of neurons in each hidden layer are not particularly limited. In the intermediate layer 102, the transmission from a neuron to neurons in the next hidden layer, along with the weighting of variables by weighting coefficients, is performed via an activation function. For the activation function, a sigmoid function, a hyperbolic tangent function, or a ramp function can be used.

[0086] The output layer 103 combines information from neurons transmitted from the intermediate layer 102 and outputs it as a constraint judgment value relative to the final target rolling speed. The number of outputs constructed within the output layer 103 is not particularly limited. Based on the results of this output, the stable rolling performance of past cold rolling of steel plate S, and the current rolling constraint performance (rolling load judgment, rolling power judgment, forward slip rate judgment, chatter judgment, plate shape judgment, edge crack judgment, plate thickness accuracy judgment), the model learns by gradually optimizing the weighting coefficients within the neural network model.

[0087] After the weighting coefficients of the neural network model are learned, the model making department 77D inputs the evaluation data (the actual rolling conditions of the steel plate S to be rolled, which uses the stable rolling speed prediction model) into the neural network model after the weighting coefficients have been learned, and obtains the estimated result relative to the evaluation data.

[0088] return Figure 2 The result storage unit 77E stores the learning data, evaluation data, parameters (weighting coefficients) of the neural network model, the output of the neural network model relative to the learning data, and the output of the neural network model relative to the evaluation data in the storage device 89.

[0089] In the cold rolling of the steel plate S, the prediction model execution unit 78 uses a stable rolling speed prediction model created by the prediction model creation unit 77 to predict the stable rolling speed of the steel plate S in cold rolling, corresponding to the rolling conditions of the steel plate S to be rolled. Furthermore, the prediction model execution unit 78 determines the stable rolling speed of the steel plate S to be rolled.

[0090] To perform the above processing, the prediction model execution unit 78 includes an information reading unit 78A, a second data conversion unit 78B, a rolling speed prediction unit 78C, a rolling condition determination unit 78D, and a result output unit 78E. Here, when the prediction model execution unit 78 receives a signal from the rolling control device 14 via the input device 88 indicating that cold rolling is being performed, it executes the prediction model execution program 75 stored in the ROM 73, thereby functioning as the information reading unit 78A, the rolling speed prediction unit 78C, the rolling condition determination unit 78D, and the result output unit 78E.

[0091] The information reading unit 78A reads from the storage device 89 the rolling conditions of the steel plate S to be rolled, which are set by the process control computer and the operator through the operation monitoring device 91.

[0092] The second data transformation unit 78B performs a process that convolves the multidimensional array information, which will become the input data to the stable rolling speed prediction model, into one-dimensional information. The processing of the second data transformation unit 78B is the same as that of the first data transformation unit 77C, so a detailed description of the processing is omitted. Alternatively, the first data transformation unit 77C and the second data transformation unit 78B can be subroutineized as a single processing unit.

[0093] The rolling speed prediction unit 78C inputs the one-dimensional information convolved by the second data transformation unit 78B into the stable rolling speed prediction model to predict the stable rolling speed of the steel plate S to be rolled. In addition, the rolling speed prediction unit 78C predicts the constraint determination value relative to the target rolling speed of the steel plate S to be rolled.

[0094] The rolling condition determination unit 78D performs the following processing: until the constraint determination value relative to the target rolling speed becomes below a preset threshold, the target rolling speed setting is changed and the processing of the aforementioned information reading unit 78A, second data conversion unit 78B and rolling speed prediction unit 78C is repeatedly returned.

[0095] If the constraint judgment value relative to the target rolling speed is below a preset threshold, the result output unit 78E operates and outputs the rolling conditions (target rolling speed) of the steel plate S to be rolled.

[0096] Next, refer to Figure 6 The processing of the prediction model execution unit 78 will be explained.

[0097] Figure 6 This is a flowchart illustrating the processing flow of the prediction model execution unit 78. (For example...) Figure 6As shown, when executing the stable rolling speed prediction model, firstly, the information reading unit 78A of the prediction model execution unit 78 reads the neural network model, which is a stable rolling speed prediction model corresponding to the required characteristics of the steel plate S to be rolled, from the storage device 89 during the processing of step S41.

[0098] Next, in step S42, the information reading unit 78A reads the required constraint determination threshold stored in the storage device 89 from the host computer via the input device 88. Next, in step S43, the information reading unit 78A reads the rolling conditions of the steel plate S to be rolled from the host computer via the input device 88, stored in the storage device 89.

[0099] Next, in step S44, the rolling speed prediction unit 78C of the prediction model execution unit 78 uses the neural network model, which is the stable rolling speed prediction model read in step S41, as the input data obtained by multi-dimensionalizing the rolling conditions of the steel plate S to be rolled in step S43, to calculate the constraint determination value relative to the target rolling speed of the corresponding cold-rolled steel plate S. It should be noted that the prediction result based on the neural network model is output to the output layer 103.

[0100] Next, in step S45, the rolling condition determination unit 78D of the prediction model execution unit 78 determines whether the constraint determination value of the steel plate S relative to the target rolling speed, obtained in step S44, is within the constraint determination threshold read in step S42. It should be noted that if the calculation convergence is insufficient, an upper limit can be set on the number of convergence iterations within the actual calculation time that can be executed in step S45. It should be noted that a constraint determination value within the constraint determination threshold is equivalent to satisfying the specified conditions in this invention.

[0101] Furthermore, if the constraint determination value is determined to be within the constraint determination threshold (if the determination result in step S45 is yes), the prediction model execution unit 78 terminates the process. On the other hand, if the constraint determination value is determined to be outside the constraint determination threshold (if the determination result in step S45 is no), the prediction model execution unit 78 causes the process to proceed to step S46.

[0102] In step S46, the rolling condition determination unit 78D changes a portion of the rolling conditions (target rolling speed) of the steel plate S to be rolled, which was read in step S43, and moves to step S47. In step S47, the result output unit 78E of the prediction model execution unit 78 transmits information related to a portion of the determined rolling conditions to the rolling control device 14 via the output device 90.

[0103] When a portion of the rolling conditions is changed during the processing in step S46, the rolling condition determination unit 78D, in the processing in step S47, determines the optimized rolling conditions for the steel plate S based on the changed rolling conditions of the steel plate S. Specifically, the target rolling speed, rolling pass planning, and the operational amounts of the unit tension between rolling stands are selected. Furthermore, the rolling condition determination unit 78D determines the operational amount of the rolling speed based on the current rolling conditions. During the cold rolling stage, the rolling control device 14 changes the rolling conditions based on information related to the rolling speed transmitted from the result output unit 78E.

[0104] As a method for calculating the change in rolling conditions after the welding point passes, the rolling condition determination unit 78D calculates suitable rolling conditions for the steel plate S to be rolled based on the difference between the constraint determination value relative to the target rolling speed obtained in step S44 and the constraint determination threshold read in step S42. Furthermore, the rolling condition determination unit 78D compares the calculated rolling conditions with the rolling conditions of the steel plate S to be rolled read in step S43, and changes the rolling conditions in step S47.

[0105] After returning to step S43, the rolling speed prediction unit 78C reads the rolling conditions of the steel plate S, which have been modified in part, as part of the rolling conditions. Furthermore, in step S44, the rolling speed prediction unit 78C uses a neural network model as a stable rolling speed prediction model to calculate a constraint determination value corresponding to the target rolling speed of the cold-rolled steel plate S, based on the modified rolling conditions read in step S43. In step S45, the rolling condition determination unit 78D determines whether the constraint determination value calculated in step S44 is within the constraint determination threshold read in step S42. Steps S43, S44, S45, S46, and S47 are repeatedly executed until this determination result is yes. Thus, the processing of the prediction model execution unit 78 (rolling speed control determination step) ends.

[0106] As can be clearly seen from the above description, in this embodiment, the prediction model creation unit 77 creates a stable rolling speed prediction model based on a machine learning method, which establishes a correlation between past rolling performance of the steel plate S and the corresponding past stable rolling performance. Furthermore, during the cold rolling of the steel plate S, the prediction model execution unit 78 uses the created stable rolling speed prediction model to determine a constraint judgment value relative to the target rolling speed of the steel plate S. The prediction model execution unit 78 then determines the rolling conditions after the weld point of the steel plate S passes, ensuring that the determined constraint judgment value is within a threshold value. Thus, predicting a stable rolling speed that does not rely on operator experience or subjectivity and satisfies various constraints in the rolling operation can prevent plate thickness variations, fractures, and other malfunctions during cold rolling and maintain productivity. Furthermore, according to this embodiment, the explanatory variable used in predicting the stable rolling speed of the steel plate S in cold rolling is a combination of numerical information collected from rolling performance data and multidimensional array information used as input data. Therefore, it is possible to identify the factors that constrain cold rolling and contribute significantly to achieving the maximum stable rolling speed on a neural network model.

[0107] For steel sheets that have reached their final thickness through cold rolling, annealing is performed to adjust their mechanical and product properties. This annealing is preferably carried out in a horizontal furnace, with a preferred soaking temperature of 600–950°C and a furnace tension of 0.1–3.0 kgf / mm². 2 If the heat spreader temperature is below 600℃ or the furnace tension is less than 0.1 kgf / mm², 2 If recrystallization is insufficient, not only will good magnetic properties be obtained, but the shape correction effect during annealing will also be compromised. On the other hand, if the soaking temperature exceeds 950℃ or the furnace tension exceeds 3.0 kgf / mm², further problems will occur. 2 If the crystal grain size becomes coarser, the mechanical strength of the steel plate will decrease, or residual strain will remain in the steel plate due to tension, resulting in a decrease in product characteristics.

[0108] [Variation Example]

[0109] The embodiments of the present invention have been described above, but the present invention is not limited thereto and various modifications and improvements are possible. For example, in this embodiment, the rolling conditions of the steel plate S to be rolled, read in the process of step S43, are not only the target rolling speed, but also a portion of the rolling conditions (roll gap in each rolling stand of the continuous rolling mill, tension between rolling stands, roll displacement, and roll bending) when the weld point passes through. In addition, in this embodiment, the stable rolling speed prediction of the steel plate S based on the stable rolling speed prediction model is repeatedly performed and the rolling conditions are determined through the initial unstable rolling stage to the final unstable rolling stage, but this can also be done in a partial manner. Furthermore, the cold rolling mill 1 is not limited to a 4-stage type, but can also be a 2-stage (2Hi), 6-stage (6Hi), or other multi-stage rolling mills, and the number of rolling stands is not particularly limited. In addition, it can also be a multi-roll mill or a Sendzimir mill.

[0110] Furthermore, if the calculation unit 15 calculates an abnormal control quantity exceeding the upper or lower limit of the rolling speed change, or if the control quantity cannot be calculated, the rolling control device 14 cannot execute control based on the instructions from the calculation unit 15. Therefore, it is best not to perform this embodiment if the rolling control device 14 determines that the control quantity from the calculation unit 15 is abnormal or that no control quantity is supplied from the calculation unit 15.

[0111] In addition, Figure 2 In the structural example shown, the output device 90 and the operation monitoring device 91 are not connected, but they can be connected in a communicative manner. Thus, the processing results of the prediction model execution unit 78 (especially the stable rolling speed prediction information of the steel plate S under rolling obtained by the rolling speed prediction unit 78C and the changed rolling conditions determined by the rolling condition determination unit 78D) can be displayed on the operation screen of the operation monitoring device 91.

[0112] Example

[0113] The present invention will now be described based on embodiments.

[0114] use Figure 1The cold continuous rolling mill of the embodiment shown, consisting of a total of five rolling mills, was used to conduct experiments on cold rolling of a base steel sheet containing 2.5% by mass Si for electromagnetic steel sheets with a base material thickness of 2.0 mm and a plate width of 1000 mm to a final thickness of 0.300 mm. The following base oil was used as the rolling oil concentrate: a base oil obtained by adding vegetable oil to synthetic ester oil, with 1% by mass of an oiliness agent and an antioxidant added respectively; and a nonionic surfactant added at a concentration relative to the oil at 3% by mass. Additionally, a recycled emulsion rolling oil was prepared to a rolling oil concentration of 3.5% by mass, an average particle size of 5 μm, and a temperature of 55°C. As a prior learning process, learning was first performed using training data (rolling performance data of approximately 3000 past steel sheets) to implement a neural network model, establishing a correlation between past unstable rolling performance and past stable rolling performance of steel sheets, and creating a neural network model used for predicting stable rolling speeds.

[0115] In the invention example, the rolling performance data of the conventional steel plate uses information consisting of the base material thickness, base material crown, deformation resistance, plate thickness, rolling pass planning (rolling load / tension / plate shape / plate thickness accuracy) when the weld point passes through, emulsion properties, work roll size / crown / roughness information, bending amount, and work roll displacement amount. Furthermore, a multi-dimensional array information obtained by concatenating the above rolling performance data in the time direction is used as input performance data. As stable rolling performance data of the conventional steel plate, stable rolling speed and the judgment performance of the operating conditions (rolling load / torque / slip rate / chatter / plate thickness accuracy / plate shape) that become constraints at this time are learned. Using a cold rolling mill to adjust the roll gap, after the weld point of the steel plate passes through, during the stage when the rolling control device 14 is turned on, the stable rolling speed and constraint judgment value of the cold-rolled steel plate are predicted based on the fabricated neural network model. Furthermore, the rolling conditions were changed sequentially in a manner that the predicted constraint judgment value was below the specified threshold, and the rolling conditions after the welding point was passed were set.

[0116] In the comparative examples, similar to the inventive examples, an experiment was conducted to cold roll a raw material steel sheet (the object to be rolled) containing 2.8% by mass Si for electromagnetic steel sheets with a base material thickness of 2.0 mm and a plate width of 1000 mm to a plate thickness of 0.3 mm. In the comparative examples under conditions 1, 4, 7, 10, and 13, past flutter performance data was correlated using input data from a one-dimensional array that did not link past steel sheet rolling performance data in the time direction, and a neural network model for predicting flutter was created. Furthermore, in the comparative examples under conditions 17 to 19, rolling was performed in the same manner as in the inventive examples, except that an acceleration was performed based on the operator's experience until a stable rolling speed was achieved.

[0117] Table 1 shows the average stable rolling speed of the steel plates after rolling 100 coils of the invention example and the comparative example. As shown in Table 1, in the comparative example, insufficient learning was carried out, or a stable rolling speed was set based on experience differences for each operator. As a result, the operation constraints could not be fully utilized, resulting in low-speed rolling, and failures such as chatter and breakage occurred when the operation constraints were exceeded.

[0118] It has been confirmed from the above that, preferably, the cold rolling method and cold rolling mill of the present invention are used to appropriately predict the stable rolling speed and the limiting judgment value in the rolling of steel plates, and the rolling conditions are changed sequentially in such a way that the predicted limiting judgment value is below a predetermined threshold to determine the stable rolling speed. Furthermore, it has been confirmed that, by applying the present invention, not only can product defects such as chatter and plate breakage during cold rolling be prevented, but it can also greatly contribute to improving the productivity and quality of the rolling process and subsequent processes.

[0119] Table 1

[0120]

[0121] The foregoing has described embodiments using the invention made by the inventors, but the present invention is not limited to the description and drawings that constitute a part of the disclosure of this invention. That is, all other embodiments, examples, and applications made by those skilled in the art based on this embodiment are included within the scope of this invention.

[0122] Industrial applicability

[0123] According to the present invention, a method and apparatus for calculating rolling conditions for a cold rolling mill can be provided, which can calculate rolling conditions that ensure rolling stability and do not hinder productivity even when rolling difficult-to-roll materials under high loads. Furthermore, according to the present invention, a cold rolling method and a cold rolling mill can be provided that can stably and productively roll difficult-to-roll materials under high loads. Additionally, according to the present invention, a method for manufacturing steel sheets that can produce steel sheets with desired product characteristics and mechanical properties with high yield can be provided.

[0124] Label Explanation

[0125] 1 Cold rolling mill

[0126] 3. Lubrication coolant head

[0127] 4 Cooling head

[0128] 5. Dirty / Soil Tanks (Recycling Tanks)

[0129] 6 Iron powder removal device

[0130] 7. Cleanroom tanks (storage tanks)

[0131] 8 pumps

[0132] 9 supply lines

[0133] 10 oil pan

[0134] 11 Return to Piping

[0135] 13 Emulsion Rolling Oil

[0136] 14 Rolling control device

[0137] 15 arithmetic units

[0138] 71 arithmetic unit

[0139] 74 Predictive Model Building Program

[0140] 75 Prediction Model Execution Program

[0141] 76 Computational Processing Unit

[0142] 77 Predictive Model Production Department

[0143] 77A Learning Data Acquisition Department

[0144] 77B Pre-processing Unit

[0145] 77C First Data Transformation Unit

[0146] 77D Model Making Department

[0147] 77E Results Preservation Department

[0148] 78 Prediction Model Execution Department

[0149] 78A Information Reading Unit

[0150] 78B Second Data Conversion Unit

[0151] 78C Rolling Speed ​​Prediction Department

[0152] 78D Rolling Conditions Determination Department

[0153] 78E Result Output Section

[0154] 88 input device

[0155] 89 storage devices

[0156] 90 output device

[0157] 91 Operation monitoring device

[0158] S-type steel plate (rolled material).

Claims

1. A method for calculating rolling conditions of a cold rolling mill, wherein, include: The estimation steps are as follows: The prediction model is learned by using first multidimensional data generated from unstable rolling performance data in the past rolling performance of the material rolled by cold rolling mill as explanatory variables and stable rolling performance data and rolling constraint data during stable rolling as target variables. The model is then input with second multidimensional data generated from unstable rolling performance data of the material to be rolled to estimate the rolling constraint conditions relative to the target stable rolling conditions of the material to be rolled. and The steps are modified to change the target stable rolling conditions in a manner that makes the estimated rolling constraints meet the specified conditions.

2. The method for calculating rolling conditions of a cold rolling mill according to claim 1, wherein, The rolled material has a welded portion formed by welding together the preceding material and the following material. The unsteady rolling performance data refers to the rolling performance data when the welded part passes through the exit side of the cold rolling mill.

3. The method for calculating rolling conditions of a cold rolling mill according to claim 1 or 2, wherein, The stable rolling performance data includes data representing the target rolling speed during stable rolling of previously rolled materials. The modification step includes changing the target rolling speed during stable rolling of the material to be rolled in a manner that makes the estimated rolling constraints meet the specified conditions.

4. A device for calculating rolling conditions of a cold rolling mill, wherein, have: The estimation unit, in the following prediction model, learns a prediction model by using first multidimensional data generated from unstable rolling performance data in past rolling performances of the material rolled using a cold rolling mill as explanatory variables and stable rolling performance data and rolling constraint data during stable rolling as target variables. It then inputs second multidimensional data generated from unstable rolling performance data of the material to be rolled to estimate the rolling constraint conditions relative to the target stable rolling conditions of the material to be rolled. and The unit is modified to change the target stable rolling conditions in a manner that makes the estimated rolling constraints meet the specified conditions.

5. A cold rolling method, wherein, The step includes rolling the material to be rolled using a modified target stable rolling condition based on the rolling condition calculation method of the cold rolling mill according to any one of claims 1 to 3.

6. A cold rolling mill, wherein, The device for calculating rolling conditions of the cold rolling mill as described in claim 4 is provided.

7. A method for manufacturing a steel plate, wherein, include: The process includes a cold rolling process, in which the material is cold rolled using the cold rolling method described in claim 5; and an annealing process, in which the rolled material after cold rolling is subjected to a homogenization temperature of 600–950°C and a furnace tension of 0.1–3.0 kgf / mm. 2 Annealing process.

Images