Method and device for detecting chatter of cold rolling mill, cold rolling method, cold rolling mill and steel sheet manufacturing method

Abstract

The present application relates to a chatter detection method for a cold rolling mill, including the steps of: inputting second multi-dimensional data generated based on condition data corresponding to array data related to a rolling object material, into a prediction model learned by taking as explanatory variables first multi-dimensional data generated based on one-dimensional array data representing past rolling performance of a rolling material rolled by the cold rolling mill, and taking as target variables past occurrence performance of chatter corresponding to the rolling performance, thereby predicting occurrence of chatter at the time of rolling the rolling object material.

Description

Technical Field

[0001] This invention relates to a method for detecting vibration in a cold rolling mill, a device for detecting vibration in a cold rolling mill, a cold rolling method, and a method for manufacturing a cold rolling mill and steel plate. Background Technology

[0002] In recent years, the demand for thin-walled, high-strength, and thin-walled rigid materials has been increasing, with the aim of reducing fuel consumption through lightweighting. However, during the cold rolling of such high-load, difficult-to-roll materials, mill vibrations in the vertical or horizontal directions, known as tremors, mainly caused by insufficient lubrication, occur at frequencies of approximately 30–200 Hz, easily leading to periodic variations in the thickness of the rolled material. Therefore, tremors have become a major obstacle to the high productivity of high-value-added products. Against this backdrop, Patent Document 1 proposes a method for detecting precursor vibrations of tremors based on the time waveform of the cold rolling mill's vibration intensity, and for preventing malfunctions caused by tremors by reducing the rolling speed when these precursor vibrations are detected.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2020-104133 Summary of the Invention

[0006] The problem that the invention aims to solve

[0007] However, while the method described in Patent Document 1 can prevent some degree of vibration, it can also cause vibrations to occur rapidly without warning, and the prediction of vibration occurrence can sometimes be delayed. It should be noted that, to address this problem, the threshold for the vibration intensity judged as a warning vibration can be set low. However, setting the threshold low leads to frequent detection of warning vibrations, causing a slowdown in rolling speed and thus hindering productivity.

[0008] The present invention was made in view of the above-mentioned problems, and its object is to provide a vibration detection method and a vibration detection device for a cold rolling mill capable of predicting the occurrence of vibration with high accuracy. Furthermore, another object of the present invention is to provide a cold rolling method and a cold rolling mill capable of preventing the occurrence of vibration and performing cold rolling with high productivity. Additionally, another objective of the present invention is to provide a method for manufacturing steel sheets capable of producing steel sheets with desired product characteristics and mechanical properties with high yield.

[0009] Methods for solving problems

[0010] The method for detecting vibration during cold rolling mills according to the present invention includes the following steps: predicting the generation of vibration during rolling of the material by inputting a prediction model learned by taking a first multidimensional data generated based on a one-dimensional array of data representing past rolling performances of rolled material using a cold rolling mill as an explanatory variable and taking past vibration generation performances corresponding to the rolling performances as a target variable into a second multidimensional data generated based on conditional data corresponding to the array of data related to the material to be rolled.

[0011] Preferably, the conditional data is a one-dimensional array of data representing the rolling conditions of the workpiece or the time-varying operating state of the cold rolling mill when the workpiece is rolled, and the first multidimensional data and the second multidimensional data are data obtained by connecting the conditional data or data after shifting the conditional data in the time direction in a first direction and / or a second direction.

[0012] Ideally, the time interval when the conditional data is shifted in the time direction should be less than 1 second (but not including 0 seconds).

[0013] Preferably, the conditional data is a one-dimensional array representing the signal strength of the FFT, and the first multidimensional data and the second multidimensional data are obtained by concatenating the conditional data or the conditional data collected at time intervals in a first direction and / or a second direction.

[0014] Preferably, the following steps are included: in the event that vibration is anticipated, the rolling speed of the cold rolling mill on the workpiece is reduced.

[0015] Preferably, the cold rolling mill is a continuous cold rolling mill with multiple rolling stands, having a first rolling oil supply system that supplies a first emulsion rolling oil to each rolling stand and a second rolling oil supply system that supplies a second emulsion rolling oil with a different concentration than the first emulsion rolling oil to a portion of the rolling stands, and, in the event of predicted vibration, changing the supply amount of the second emulsion rolling oil from the second rolling oil supply system.

[0016] The cold rolling mill vibration detection device of the present invention comprises the following unit: predicting the generation of vibration during the rolling of the material by inputting a prediction model, which is learned by taking a first multidimensional data generated based on a one-dimensional array of data representing past rolling performances of rolled material using a cold rolling mill as an explanatory variable and a past vibration generation performance corresponding to the rolling performance as a target variable, into a second multidimensional data generated based on conditional data corresponding to the array of data related to the material to be rolled.

[0017] The cold rolling method of the present invention includes the step of rolling a rolled material using the vibration detection method of the cold rolling mill of the present invention.

[0018] The cold rolling mill involved in this invention is equipped with the vibration detection device for the cold rolling mill involved in this invention.

[0019] The steel plate manufacturing method of the present invention includes: a cold rolling process, wherein the rolled material is cold rolled using the cold rolling method of the present invention; and an annealing process, wherein the rolled material after the cold rolling process 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] The vibration detection method and device for cold rolling mills according to the present invention can predict the occurrence of vibration with high accuracy. Furthermore, the cold rolling method and cold rolling mill according to the present invention can prevent the occurrence of vibration and perform cold rolling with high productivity. Additionally, the steel plate manufacturing method according to the present invention can manufacture steel plates with desired product characteristics and mechanical properties with a high yield. Attached Figure Description

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

[0023] Figure 2 This is a block diagram showing the structure of a computing unit as an embodiment of the present invention.

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

[0025] Figure 4 This is a diagram illustrating the processing flow of a neural network 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 vibration early warning monitoring unit.

[0028] Figure 7 It is shown Figure 1 A diagram showing the structure of a modified example of a cold rolling mill. Detailed Implementation

[0029] Hereinafter, with reference to the accompanying drawings, a vibration detection method for a cold rolling mill, a vibration detection device for a cold rolling mill, a cold rolling method, a cold rolling mill, and a method for manufacturing a steel sheet, 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 specify 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 relationship between thickness and planar dimensions, ratios, etc., differ from reality, and the accompanying drawings also include portions where the dimensional relationships and ratios differ from each other.

[0030] [Structure of a cold rolling mill]

[0031] First, refer to Figure 1 The structure of a cold rolling mill, which is one embodiment of the present invention, will be described. Figure 1 This diagram illustrates the structure of a cold rolling mill according to one embodiment of the present invention. It should be noted that in this specification, "cold rolling" is sometimes abbreviated as "rolling," and both terms are synonymous. Furthermore, in the following description, steel plate is used as an example of the rolled material (the material to be rolled) produced by the cold rolling mill. However, the rolled material is not limited to steel plate; other metal strips such as aluminum plates can also be used.

[0032] like 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: the first to the fifth (#1STD to #5STD). In this cold rolling mill 1, tension rolls and differential pressure rolls (not shown), a plate thickness gauge, and a shape gauge are appropriately installed between adjacent rolling stands. Furthermore, the structure of the rolling stands and the conveying device for the steel plate S are not particularly limited, and known technologies can be appropriately applied.

[0033] Each rolling stand of the cold rolling mill 1 is designed to receive the supply of emulsion rolling oil (hereinafter referred to as "rolling oil" sometimes simply as "emulsion rolling oil") 13. The cold rolling mill 1 is equipped with a dirty tank (recovery 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 via the supply line 9.

[0034] The rolling oil recovered by the oil pan 10 located below each rolling mill stand, i.e., the rolling oil used in cold rolling, flows into the sludge tank 5 through the return pipe 11. The rolling oil accumulated in the clean tank 7 is formed by mixing warm water (diluent) and the original rolling oil (with added surfactants). This mixture of warm water and original rolling oil is adjusted by changing the rotation speed of the agitator blades of the mixer 12 (that is, by adjusting the degree of agitation) to produce rolling oil with a desired average particle size and concentration range.

[0035] The rolling oil stock can be any stock used in conventional cold rolling, such as a stock oil based on any of the following: natural oils, fatty acid esters, or hydrocarbon-based synthetic lubricants. Furthermore, additives commonly used in conventional cold rolling oils, such as oiliness improvers, extreme pressure additives, and antioxidants, can be added to these rolling oils. Additionally, any type of surfactant, both ionic and non-ionic, can be used, and any surfactant used in conventional circulating coolant systems (circulating rolling oil supply methods) can be employed. The rolling oil stock is diluted to a preferred concentration of 2-8% by mass, more preferably 3-6.0% by mass, and an O / W emulsion rolling oil with oil dispersion in water is formed by using the surfactant. It should be noted that the average particle size of the rolling oil is preferably 15 μm or less, more preferably 3-10 μm.

[0036] 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, etc. The rolling oil recovered to the dirty tank 5 contains wear powder (iron powder) generated by friction between the rolls and the steel plate S. Therefore, the iron powder removal device 6 removes the wear powder by reducing the oil-soluble iron content of the recovered rolling oil to the permissible oil-soluble iron content for accumulation in the clean tank 7. 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 a magnetic filter such as an electromagnetic filter or magnetic separator is preferred for adsorbing and removing iron powder, the device is not limited to this. The iron powder removal device 6 can also be a known device using methods such as centrifugal separation.

[0037] A portion of the rolling oil supplied to the cold rolling mill 1 is carried out of the system by the steel sheet S or lost due to evaporation. Therefore, the clean tank 7 is designed to appropriately replenish (supply) the raw rolling oil from the raw material tank (not shown) so that the accumulation level and concentration of the rolling oil within the clean tank 7 are within a specified range. Additionally, warm water for dilution 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 within the clean tank 7 can be measured using a sensor (not shown).

[0038] Next, a detailed description of the rolling oil supply system for the cold rolling mill 1 will be provided. It should be noted that the rolling oil supply system includes a dirt 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. Additionally, a filter for removing foreign matter may be installed between the clean tank 7 and the pump 8.

[0039] The rolling oil supply system 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 rolling mill stand.

[0040] Each lubrication coolant head 3 is located on the inlet side of the rolling mill stand, supplying lubricating oil to the roll bite and work rolls by spraying rolling oil as lubricant from separately provided spray nozzles toward the roll bite. Each cooling coolant head 4 is located on the outlet side of the rolling mill stand, cooling the rolls by spraying rolling oil from separately provided spray nozzles toward the rolls.

[0041] Through this structure, the emulsion rolling oil in the clean tank 7 is pumped by the pump 8 to the supply line 9, supplied to the lubrication coolant head 3 and cooling coolant head 4 located on each rolling mill stand, and supplied to the spraying parts from the separately provided spray 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 sludge tank 5 through the return pipe 11. Afterwards, a portion of the emulsion rolling oil accumulated in the sludge tank 5 is returned to the clean tank 7 after the iron powder removal device 6 removes a certain amount of oil-soluble iron from the emulsion rolling oil generated by cold rolling to remove the oil-soluble iron.

[0042] Through the above-described rolling oil supply system, the rolling oil, after being treated to remove wear powder, is circulated to the rolls. That is, the supplied emulsion rolling oil is recycled. Here, the cleaning tank 7 corresponds to the rolling oil tank for circulation in the conventional circulating oil supply method, and as described above, the original rolling oil is appropriately replenished (supplied) to the cleaning tank 7.

[0043] [Flicker Detection Method]

[0044] Next, details will be provided regarding the vibration detection method and vibration prediction model for a cold rolling mill, which is one embodiment of the present invention. Figure 2 This is a block diagram showing the structure of a computing unit as an embodiment of the present invention.

[0045] In this embodiment, the functions associated with the tremor prediction model are provided by Figure 2 The shown calculation unit 15, vibration measuring device 16, load sensor 17, and tension meter 18 are implemented. Rolling control device 14 (see reference) Figure 1 The rolling conditions of the cold rolling mill 1 are controlled based on the control signals from the arithmetic unit 15.

[0046] The arithmetic unit 15 includes an arithmetic device 71, an input device 88, a storage device 89, and an output device 90. The arithmetic device 71 is wiredly connected to the input device 88, the storage device 89, and the output device 90 via a bus 87. The arithmetic device 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.

[0047] Input device 88 serves as input to rolling control device 14 (see reference) Figure 1 The system functions as an input port for the control information of the cold rolling mill 1, the time variation of the vibration waveform of the cold rolling mill 1 obtained by the vibration measuring device 16, the time variation of the rolling load information of the cold rolling mill 1 obtained by the load sensor 17, the time variation of the tension between the rolling stands obtained by the tension meter 18, and information from the operation monitoring device 91. The information from the operation monitoring device 91 includes execution command information for the vibration prediction model, information related to the material being rolled (steel grade, dimensions), and cold rolling condition information (numerical information, text information, and image information) set by the operator before cold rolling.

[0048] 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 making unit 77 and the vibration precursor monitoring unit 78, which will be described later).

[0049] The output device 90 functions as an output port for outputting control signals from the arithmetic unit 71 to the rolling control device 14.

[0050] 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.

[0051] The arithmetic unit 71 includes RAM 72, ROM 73, and an arithmetic processing unit 76. ROM 73 stores a predictive model creation program 74 for computer programs and a vibration precursor monitoring program 75. The arithmetic processing unit 76 has arithmetic processing functions and is connected to RAM 72 and ROM 73 via bus 87. RAM 72, ROM 73, and the arithmetic processing unit 76 are connected to input device 88, storage device 89, and output device 90 via bus 87.

[0052] The arithmetic processing unit 76 is a functional block that includes a prediction model creation unit 77 and a vibration precursor monitoring unit 78.

[0053] The prediction model generation unit 77 is a processing unit that generates a jitter prediction model based on machine learning methods, which establishes a correlation between past rolling records of steel plates S rolled using cold rolling mill 1 and jitter generation records corresponding to the past rolling records. In this embodiment, a neural network model is used as the jitter prediction model based on machine learning methods. The machine learning method is not limited to neural networks; other known machine learning methods may also be used.

[0054] 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 jitter prediction model, it executes the functions of 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 by executing the prediction model creation program 74 stored in the ROM 73. The jitter prediction model is updated each time the prediction model creation unit 77 executes the prediction model creation program 74.

[0055] The learning data acquisition unit 77A, as preprocessing for generating a vibration prediction model, acquires multiple learning data sets (e.g., approximately 10,000 pieces), using rolling condition operation data and time waveform data as input data and vibration generation data during cold rolling corresponding to the input data as output data. The learning data acquisition unit 77A obtains the input and output data from the storage device 89 to create the learning data. Each learning data set consists of a combination of input and output data. The learning data acquisition unit 77A stores the learning data in the storage device 89. Alternatively, the learning data acquisition unit 77A may not store the learning data in the storage device 89, but instead supply the learning data to the preprocessing unit 77B and the model creation unit 77D.

[0056] The input performance data includes multidimensional array information obtained by linking explanatory variables along the time axis. In this embodiment, an example is used as multidimensional array information. Figure 3 The information shown in (a) to (d) is as follows. Figure 3(a) shows an example of input performance data where the vertical columns consist of coil conditions and the horizontal columns consist of explanatory variables selected from operational performance data of rolling conditions. Representative operational performance data for rolling conditions can include information such as the base material thickness of steel plate S, plate thickness before and after cold rolling, plate width, deformation resistance of steel plate S, rolling load during cold rolling, front and rear tension, work roll dimensions, work roll bending amount, work roll thickness, intermediate roll bending amount, and intermediate roll displacement amount; however, the number of columns of explanatory variables is not particularly limited. In this example, a multidimensional array of information is created by linking each coil condition in the time direction (depth direction) as learning data. The time interval when linking in the time direction is preferably 0 to 1 second or less. Furthermore, in the case where there is no time-series data or explanatory variables that do not change with time, the time interval is 0 seconds, and the same data is copied.

[0057] Figure 3 (b) shows an example where the vertical column represents the roll material conditions, and the horizontal column represents the time-series data of operational performance data within a certain time interval as input performance data for explanatory purposes. Figure 3 Similarly, in the example shown in (a), a multidimensional array of information is created by linking each roll condition in the time direction (depth direction) to become learning data. In this example, time-series data of vibration velocity is shown, but it is not particularly limited and can also be time-series data of rolling load, inter-stand tension, plate thickness deviation, vibration displacement, and vibration acceleration, etc.

[0058] Figure 3 (c) shows an example of input data where the vertical column represents the time sequence and the horizontal column represents the signal strength (FFT intensity) obtained by performing a Fast Fourier Transform (FFT) on the signal waveform at a certain time. In this example, the multidimensional array information obtained by copying the signal strength at a specific measurement location at a certain time t at time intervals of 0 to 1 second along the time direction (depth direction) is used as the learning data.

[0059] Figure 3 (d) shows an example where the vertical column represents time sequence and the horizontal column represents signal strength obtained by performing an FFT on the signal waveform at a specific location and time, serving as the input data for explanation. In this embodiment, as long as the data is connected along the time direction (depth direction), it is not necessary for the data to be from the same location; a multidimensional array of signal strength data from multiple locations can be created and used as training data. Furthermore, this example shows signal strength within a certain frequency range, but the frequency range is not particularly limited.

[0060] It should be noted that, in Figure 3In the examples shown in (a) to (d), the multidimensional array information is data obtained by connecting in the vertical direction (two-dimensional data). However, it can also be connected only in the depth direction, or connected in both the vertical and depth directions to form three-dimensional data. Furthermore, in Figure 3 In the examples shown in (a) to (d), the vertical direction corresponds to the first direction (or second direction) involved in this invention, and the depth direction corresponds to the second direction (or first direction) involved in this invention.

[0061] Furthermore, in cases where the storage device 89 does not store actual rolling data (learning data) (e.g., rolling conditions or steel grade conditions for which there are no past records) or the sample size is small, the learning data acquisition unit 77A can also request the operator to perform cold rolling without using the vibration prediction model once or multiple times. Additionally, since the more learning data stored in the storage device 89, the higher the prediction accuracy of the vibration prediction model, if the amount of learning data is less than a preset threshold, the learning data acquisition unit 77A can also request the operator to perform cold rolling without using the vibration prediction model until the amount of learning data reaches the threshold.

[0062] return Figure 2 The preprocessing unit 77B processes the learning data acquired by the learning data acquisition unit 77A into data for creating a tremor prediction model. Specifically, in order to enable the neural network model to read the learning data, the preprocessing unit 77B standardizes (normalizes) the value range of the input performance data constituting the learning data to between 0 and 1. It should be noted that the input performance data is multi-dimensional information. Therefore, the first data transformation unit 77C uses a convolutional neural network 300 to compress the dimensionality of the input performance data with residual feature quantities, making it one-dimensional information (see reference). Figure 4 The input performance data is presented as one-dimensional information to input layer 101 (refer to...). Figure 4 (Combined)

[0063] Here, refer to Figure 5 Let's illustrate a processing example of the first data transformation unit 77C. For example... Figure 5 As shown, the transformation process that converts multidimensional array information into one-dimensional information, i.e., the method for storing multidimensional array information, has a structure with multiple filters connected in multiple stages of input and output. Specifically, the transformation process that converts 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.

[0064] 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-filled area is set to 1.

[0065] 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 takes the maximum value within the horizontal × vertical 3 pixels of the first feature map as the new 1 pixel. The first data transformation unit 77C performs this operation by shifting pixels and applying it to 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.

[0066] 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 perimeter filled with 0 is set to 1.

[0067] 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 takes the maximum value within the horizontal × vertical 3 pixels of the third feature map as a new 1 pixel. The first data transformation unit 77C performs this operation across the entire map while shifting pixels. 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.

[0068] 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 (see reference). Figure 4 It should be noted that the convolutional techniques and the number of output neurons are not limited to those mentioned above. Furthermore, known models such as GoogleNet, VGG16, MOBILENET, and EFFICIENTNET can also be used as techniques for convolutional neural networks.

[0069] return Figure 2The model creation unit 77D uses machine learning (including information transformed by the first data transformation unit 77C) on multiple learning data obtained by the preprocessing unit 77B to generate a jitter prediction model that takes the rolling performance data of the cold rolling mill 1 as input performance data and the presence or absence of jitter precursors corresponding to the input performance data as output data. In this embodiment, since a neural network is used as a machine learning method, the model creation unit 77D creates a neural network model as a jitter prediction model. That is, the model creation unit 77D creates a neural network model that establishes a correlation between the input performance data (rolling performance data) and the output performance data (jitter precursor presence or absence data) in the learning data to be processed into the jitter prediction model. The neural network model is, for example, represented by a function.

[0070] Specifically, the model making 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 making department 77D creates a neural network model with some hyperparameters that are modified in stages for the learning data (e.g., rolling performance data of tens of thousands of pieces), and selects the hyperparameters with the highest prediction accuracy relative to the validation data.

[0071] As hyperparameters, the number of hidden layers, the number of neurons in each hidden layer, the dropout rate (cutting off neuron transmission with a certain probability) in each hidden layer, and the activation function in each hidden layer 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 random parameter selection, or Bayesian optimization-based search can be used. In this structure, the model creation unit 77D is incorporated as part of the computing unit 71, but the structure is not limited to this. For example, tremor prediction models can be pre-created and saved, and then appropriately read out.

[0072] Figure 4 This illustrates the processing flow of the neural network model in this system. For example... Figure 4 As shown, the neural network model used as the tremor prediction model in this embodiment includes, from the input side, an input layer 101, an intermediate layer 102, and an output layer 103. The learning data acquisition unit 77A uses a convolutional neural network to compress the dimensionality of the input data with residual features, and saves the input data to the input layer 101 as one-dimensional information.

[0073] The intermediate layer 102 consists of multiple hidden layers, each containing multiple neurons. The number of hidden layers in the intermediate layer 102 is not particularly limited, but empirically, too many hidden layers decrease prediction accuracy; therefore, five layers or less are preferred. Furthermore, the number of neurons in each hidden layer is preferably in the range of 1 to 10 times the number of the input explanatory variables. In the intermediate layer 102, the transmission from a neuron to neurons in the next hidden layer, along with the weighting of the variables by weighting coefficients, is performed via an activation function. The activation function can be a sigmoid function, a hyperbolic tangent function, or a ramp function.

[0074] In the output layer 103, the information from the neurons transmitted from the intermediate layer 102 is combined to output the final value indicating the presence or absence of jitter prediction. The weighting coefficients within the neural network model are gradually optimized based on this output and past jitter performance during cold rolling of the steel plate S, thereby learning the weighting coefficients of the neural network model. After the weighting coefficients of the neural network model are learned, the model manufacturing unit 77D inputs evaluation data (rolling performance data of the steel plate S to be rolled up to the point where jitter prediction using the jitter prediction model is implemented) into the neural network model after the weighting coefficients have been learned, obtaining an estimated result relative to the evaluation data.

[0075] return Figure 2 The results 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.

[0076] Information required for the implementation of the predictive model generation unit 77 includes, for example, the time variation of the vibration waveform of the cold rolling mill 1 obtained by the vibration measuring device 16, the time variation of the rolling load information of the cold rolling mill 1 obtained by the load sensor 17, the time variation of the tension between the rolling mill stands obtained by the tension meter 18, and the aforementioned explanatory and target variables for each required characteristic (steel grade, plate thickness, plate width, etc.) of the steel plate S to be rolled. The explanatory and target variables are obtained at predetermined intervals from the initial unstable rolling stage to the final unstable rolling stage.

[0077] During the cold rolling of steel sheet S, the vibration prediction monitoring unit 78 uses a vibration prediction model created by the prediction model creation unit 77 to predict whether vibration will occur in the cold-rolled steel sheet S corresponding to the rolling conditions of the target material. Furthermore, the vibration prediction monitoring unit 78 determines the control amount of the rolling speed of the steel sheet S in a manner that converges the predicted vibration. To perform this process, the vibration prediction monitoring unit 78 includes an information reading unit 78A, a second data conversion unit 78B, a vibration prediction unit 78C, a rolling condition determination unit 78D, and a result output unit 78E. It should be noted that the vibration prediction monitoring unit 78 receives data from the rolling control device 14 (see reference 14) via the input device 88. Figure 1 When a signal indicating that cold rolling is in progress is received, the information reading unit 78A, the vibration prediction unit 78C, the rolling condition determination unit 78D, and the result output unit 78E are executed by executing the vibration warning monitoring program 75 stored in ROM 73.

[0078] The information reading unit 78A reads the rolling conditions of the steel plate S to be rolled, which are set by the operator through the operation monitoring device 91, from the storage device 89.

[0079] The second data transformation unit 78B transforms the multidimensional array information of the input data to the jitter prediction model into one-dimensional information. The processing in the second data transformation unit 78B is the same as that in the first data transformation unit 77C, so a detailed description of the processing is omitted. It should be noted that the first data transformation unit 77C and the second data transformation unit 78B can also be subroutineted as a single processing unit.

[0080] The vibration prediction unit 78C inputs the one-dimensional information generated by the second data transformation unit 78B into the flutter prediction model to obtain the flutter prediction value.

[0081] If the vibration prediction value obtained by the vibration prediction unit 78C is above a preset threshold, the rolling condition determination unit 78D performs a process that repeatedly returns to the processing of the information reading unit 78A, the second data conversion unit 78B, and the vibration prediction unit 78C until the vibration prediction value is below the threshold.

[0082] If the vibration prediction unit 78C calculates a flutter prediction value below a preset threshold, the result output unit 78E operates and outputs the determined rolling conditions (rolling speed).

[0083] Information required for the realization of the function of the vibration prediction monitoring unit 78 includes, for example, a vibration prediction model created by the prediction model creation unit 77 that has learned the rolling state of the steel plate S, and various information input to the vibration prediction model.

[0084] Next, refer to Figure 6 The processing of the vibration early warning monitoring unit 78 will be explained.

[0085] Figure 6 This is a flowchart illustrating the processing flow of the vibration early warning monitoring unit. (Example) Figure 6 As shown, in step S41, the information reading unit 78A of the vibration precursor monitoring unit 78 reads the neural network model, which is a vibration prediction model corresponding to the required characteristics of the steel plate S to be rolled, from the storage device 89. Next, in step S42, the information reading unit 78A reads the required vibration precursor judgment threshold stored in the storage device 89 from the host computer via the input device 88.

[0086] Next, in step S43, the information reading unit 78A reads the rolling conditions and time waveform information of the steel plate S to be rolled, which is stored in the storage device 89 from the host computer via the input device 88. Then, in step S44, the vibration prediction unit 78C of the vibration prediction monitoring unit 78 uses the neural network model read in step S41 as a vibration prediction model, and takes the rolling conditions and time waveform information of the steel plate S to be rolled, which was read in step S43, as multidimensional array input data to calculate the corresponding vibration prediction value of the cold-rolled steel plate S. It should be noted that the prediction result of the neural network model is output to the output layer 103.

[0087] Next, in step S45, the rolling condition determination unit 78D of the vibration precursor monitoring unit 78 determines whether the vibration precursor determination value obtained through step S44 is within the vibration precursor 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 repetitions within the actual calculation time that can be performed in step S45. Furthermore, if the vibration precursor determination value is determined to be within the vibration precursor determination threshold (if the determination result in step S45 is yes), the vibration precursor monitoring unit 78 terminates the process. On the other hand, if the vibration precursor determination value is determined to be outside the vibration precursor determination threshold (if the determination result in step S45 is no), the vibration precursor monitoring unit 78 proceeds to step S46.

[0088] In step S46, the rolling condition determination unit 78D changes a portion of the rolling conditions of the steel plate S to be rolled, which was read in step S43, and proceeds to step S47. In step S47, the result output unit 78E of the vibration precursor monitoring unit 78 transmits information related to the determined portion of the rolling conditions to the rolling control device 14 via the output device 90. It should be noted that when a portion of the rolling conditions is changed in step S46, in step S47, the rolling condition determination unit 78D determines the rolled conditions of the steel plate S to be rolled after the portion of the rolling conditions (specifically, the operating amount of the rolling speed) has been changed, as optimized rolling conditions for the steel plate S. Furthermore, the rolling condition determination unit 78D determines the operating amount of the rolling speed based on the current rolling conditions. During the cold rolling stage, the rolling control device 14 controls the rotational speed of the motors in each rolling stand based on the information related to the rolling speed transmitted from the result output unit 78E.

[0089] As a method for calculating the change in the rolling speed, the rolling condition determination unit 78D calculates the appropriate rolling conditions for the steel plate S to be rolled based on the difference between the vibration prediction value of the cold-rolled steel plate S predicted in step S44 and the required vibration prediction 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.

[0090] If the process returns to step S43, the vibration prediction unit 78C reads the rolling conditions of the steel plate S, whose rolling conditions have been partially modified, and sets the modified time waveform information. Furthermore, in step S44, the vibration prediction unit 78C uses a neural network model as a vibration prediction model to calculate the vibration prediction threshold for the cold-rolled steel plate S corresponding to the rolling conditions of the steel plate S, whose rolling conditions have been partially modified and read in step S43. Then, in step S45, the rolling condition determination unit 78D determines whether the vibration prediction value for the cold-rolled steel plate S predicted in step S44 is within the required vibration prediction threshold read in step S42. And, until this determination result is yes, a series of processes including steps S43, S44, S45, S46, and S47 are repeatedly executed. Thus, the processing of the vibration prediction monitoring unit 78 (vibration control determination step) ends.

[0091] As can be clearly seen from the above description, in this embodiment, the prediction model generation unit 77 generates a vibration prediction model based on machine learning, which establishes a correlation between past rolling performance of steel plate S and vibration performance during cold rolling of steel plate S corresponding to the past rolling performance. Next, in this embodiment, the vibration precursor monitoring unit 78 uses the generated vibration prediction model to predict the presence or absence of vibration in the cold rolling of steel plate S, corresponding to the rolling conditions of the steel plate S to be rolled. Furthermore, the vibration precursor monitoring unit 78 determines the rolling speed of the steel plate S to be rolled in such a way that the predicted vibration precursor judgment value of the cold-rolled steel plate S is below a required vibration precursor threshold. Therefore, by appropriately predicting the vibration during the rolling of the steel plate S and determining the rolling speed in such a way that the predicted vibration warning value is below the required vibration warning threshold, even if the vibration becomes rapidly diverging due to interference, the rolling speed can be instantly corrected to an appropriate range, thereby suppressing plate thickness variation and fracture during cold rolling.

[0092] Furthermore, in this embodiment, since the vibration prediction model based on machine learning is set as a neural network model, vibration during the cold rolling of steel sheet S can be predicted more accurately. Moreover, according to this embodiment, the explanatory variables used for vibration prediction of steel sheet S during cold rolling include not only numerical information collected from actual rolling condition data, but also time-series data such as time waveform information. Furthermore, by linking these data and using multidimensional array information as input data, factors that contribute significantly to vibration during cold rolling can be identified on the neural network model.

[0093] 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 inadequate. 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 the application of tension, resulting in a decrease in product characteristics.

[0094] 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 the above embodiments, when the occurrence of vibration is predicted, the rolling condition determination unit 78D changes the rolling speed, but it is also possible to determine the rolling conditions of the steel plate S to be rolled after the mixing ratio of the first emulsion rolling oil and the second emulsion rolling oil has been changed as the optimized rolling conditions of the steel plate S. In this case, the cold rolling mill includes a first rolling oil supply system that supplies the first emulsion rolling oil to each rolling stand, a second rolling oil supply system that supplies a second emulsion rolling oil with a concentration different from the first emulsion rolling oil to a portion of the rolling stands, and a flow control valve that controls the mixing ratio of the first emulsion rolling oil supplied from the first rolling oil supply system and the second emulsion rolling oil supplied from the second rolling oil supply system and supplies it to each rolling stand.

[0095] Specifically, in this case, such as Figure 7 As shown, the cold rolling mill 1 has Figure 1 The rolling oil supply system shown is a first rolling oil supply system, and a second rolling oil supply system 20 is provided to supply the second emulsion rolling oil 31 to the fourth rolling stand and the fifth rolling stand (#4STD, #5STD).

[0096] The second rolling oil supply system 20 includes a rolling oil line 21 with one end connected to the supply line 9, a rolling oil line 23 with one end connected to the emulsion tank 22, a flow control valve 24, an emulsion head 25, and a mixing rolling oil line 26 with one end connected to the flow control valve 24 and the other end connected to the emulsion head 25.

[0097] A rolling oil crude oil tank 27 and a warm water tank 28 are connected to the emulsion tank 22. The rolling oil crude oil stored in the rolling oil crude oil tank 27 and the warm water stored in the warm water tank 28 are supplied to the emulsion tank 22 via a pump (not shown) and a flow control valve 29, and mixed within the emulsion tank 22 by a mixer 30. In the following description, the rolling oil in the emulsion tank 22 will be referred to as the second emulsion rolling oil 31.

[0098] The temperature conditions of the second emulsion rolling oil 31 are preferably the same as those of the emulsion rolling oil (hereinafter referred to as the first emulsion rolling oil) 13. However, from the viewpoint of improving the cooling capacity of the steel plates in the fourth and fifth rolling stands, the temperature of the second emulsion rolling oil 31 can be lower than that of the first emulsion rolling oil 13 via a cooling device not shown. In addition, the concentration conditions of the rolling oil in the second emulsion rolling oil 31 are not the same as those of the first emulsion rolling oil 13.

[0099] Driven by pump 8, the first emulsion rolling oil 13 accumulated in the clean tank 7 is supplied to the flow control valve 24 through the rolling oil pipeline 21. Meanwhile, the second emulsion rolling oil 31 is supplied to the flow control valve 24 by supply pump 32 through the rolling oil pipeline 23. The second emulsion rolling oil 31 mixes with the first emulsion rolling oil 13 within the flow control valve 24, forming a mixed rolling oil containing the second emulsion rolling oil 31 at a predetermined emulsion concentration. This mixed rolling oil is conveyed to the emulsion heads 25 of the fourth and fifth rolling stands through the mixed rolling oil pipeline 26. The emulsion heads 25 are branched on both the upper and lower surfaces of the steel plate S, thereby enabling the spraying of the mixed rolling oil of the desired concentration from multiple nozzles towards both the surface and back surfaces of the steel plate S. The rolling oil recovered to the oil pan 10 is then returned to the dirty tank 5 through the return pipeline 11 for recycling.

[0100] It should be noted that the flow control valve 24 can also control the flow rate of the second emulsion rolling oil 31 relative to the flow rate of the first emulsion rolling oil 13. Alternatively, the second emulsion rolling oil 31 can be supplied directly to the steel plate S without passing through the flow control valve 24 that constitutes the mixing section, but it is preferable to supply emulsion rolling oil that is a mixture of the first emulsion rolling oil 13 and the second emulsion rolling oil 31.

[0101] Furthermore, the rolling condition determination unit 78D determines the mixing ratio of the first emulsion rolling oil 13 and the second emulsion rolling oil 31 based on the current rolling conditions. Rolling control device 14 (see reference) Figure 1 During the cold rolling stage, the opening degree is controlled based on information related to the mixing ratio of the first emulsion rolling oil 13 and the second emulsion rolling oil 31 transmitted from the result output unit 78E, and the flow control valve 24 controls the mixing ratio of the first emulsion rolling oil 13 and the second emulsion rolling oil 31 based on this. As a method for calculating the change in the mixing ratio of the first emulsion rolling oil 13 and the second emulsion rolling oil 31, the rolling condition determination unit 78D calculates the appropriate rolling conditions for the steel plate S to be rolled based on the difference between the vibration prediction judgment value obtained in the processing of step S44 and the vibration prediction threshold read in the processing of step S42. Thus, when vibration is predicted to occur, the supply amount of the second emulsion rolling oil 31 from the second rolling oil supply system is changed.

[0102] Furthermore, in this embodiment, the vibration prediction model iterates the vibration prediction of the steel plate S and determines the rolling conditions from the initial unstable rolling stage to the final unstable rolling stage, but this can also be done in only a portion of the process. Additionally, the cold rolling mill 1 is not limited to a 4-segment type; it can also be a 2-segment (2Hi), 6-segment (6Hi), or other multi-segment mills, and the number of rolling stands is not particularly limited. Alternatively, it can be a multi-roll mill or a Sendzimir mill.

[0103] 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, 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, it is preferable not to execute this embodiment.

[0104] exist Figure 2 In the structural example shown, the output device 90 and the operation monitoring device 91 are not connected, but they can also be connected in a communicative manner. As a result, the processing results of the vibration prediction monitoring unit 78 (especially the vibration prediction information of the rolling steel plate S obtained by the vibration 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.

[0105] [Example 1]

[0106] The present invention will now be described based on embodiments.

[0107] In this embodiment, the use of Figure 1 The cold rolling mill 1 shown conducted an experiment to cold roll an electromagnetic steel sheet containing Si as shown in Table 1 (with a base material thickness of 2.0 mm and a width of 1000 mm) to a finished thickness of 0.300 mm. As the stock oil for rolling, a base oil obtained by adding vegetable oil to a synthetic ester oil was used, with 1% by mass each of an oiliness agent and an antioxidant added, and a non-ionic surfactant added at 3% by mass to the oil concentration. Additionally, a recycled emulsion rolling oil was prepared with a concentration of 3.5% by mass, an average particle size of 5 μm, and a temperature of 55°C. Furthermore, as a pre-learning process, a neural network model was first implemented on the learning data (rolling performance data of approximately 3000 past steel sheets). A correlation was established between the past rolling performance data and past vibration performance data, and a neural network model was created for predicting vibration.

[0108] In the invention example, conventional rolling performance data for steel plates utilizes information comprising the base material thickness, plate thickness before and after cold rolling, plate width, deformation resistance, rolling load during cold rolling, front and rear tension, work roll dimensions, work roll bending amount, work roll thickness, intermediate roll bending amount, and intermediate roll displacement amount. Furthermore, multidimensional array information obtained by concatenating the rolling performance data in the time direction is used as input performance data. Additionally, conventional vibration performance data is used, learning from performance data on plate thickness variations associated with vibration during cold rolling. Moreover, after adjusting the roll gap in the cold rolling mill and accelerating, during the phase when the rolling control device 14 is activated, vibration prediction of the cold-rolled steel plate is performed based on a fabricated neural network model. Furthermore, the rolling speed during cold rolling is set by progressively changing the rolling speed so that the predicted vibration prediction value is below the vibration prediction threshold.

[0109] Regarding the comparative example, similarly to the inventive example, an experiment was conducted in which a raw material steel sheet (the object to be rolled), containing Si as shown in Table 1 below, with a base material thickness of 2.0 mm and a plate width of 1000 mm, was cold-rolled to a plate thickness of 0.3 mm. In the comparative example, past tremor performance data was correlated with past tremor performance data by using past rolling performance data of steel sheets as input data in a one-dimensional array, without linking them in the time direction, and a neural network model for tremor prediction was created. Otherwise, cold rolling was performed in the same manner as in the embodiment.

[0110] As shown in Table 1 below, under the rolling conditions of the inventive example, the number of vibration occurrences was suppressed to less than 1, confirming the effectiveness of the application of the present invention. From the above, it is confirmed that it is best to use the cold rolling method and cold rolling mill of the steel plate involved in this invention to appropriately predict the vibration precursor judgment value during the rolling of the steel plate, and to determine the rolling speed by successively changing the rolling conditions in a manner that makes the predicted vibration precursor judgment value below a predetermined threshold. Furthermore, it is confirmed that by applying the present invention, not only can vibration and plate breakage be prevented during cold rolling, but steel plates with good thickness accuracy can also be stably manufactured, thereby greatly contributing to improved productivity and quality in subsequent processes.

[0111] [Table 1]

[0112] [Table 1]

[0113]

[0114] [Example 2]

[0115] In this embodiment, except that the mixing ratio of the first emulsion rolling oil and the second emulsion rolling oil was changed successively to make the vibration warning judgment value below the predetermined vibration warning threshold, and rolling was performed with target rolling speeds of 200 mpm, 600 mpm, 800 mpm, and 1000 mpm, it was the same as in Example 1. As shown in Table 2 below, under the rolling conditions of the inventive example, the number of vibration occurrences was suppressed to less than 1, confirming that the application of the present invention is effective. As described above, it can be seen that it is preferable to use the cold rolling method and cold rolling mill of the present invention to appropriately predict the vibration warning judgment value in the rolling of steel plates, and to determine the amount of operation by changing the rolling conditions successively to make the predicted vibration warning judgment value below the predetermined vibration warning threshold. Thus, it can be seen that by applying the present invention, not only can vibration and plate breakage in cold rolling be prevented, but steel plates S with good plate thickness accuracy can also be obtained stably, which can greatly contribute to the improvement of productivity and quality in subsequent processes. In particular, in Example 2, vibration can be suppressed without changing the rolling speed, resulting in high productivity.

[0116] [Table 2]

[0117] [Table 2]

[0118]

[0119] The foregoing has described embodiments using the invention made by the inventors, but the present invention is not limited to the descriptions 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.

[0120] Industrial availability

[0121] According to the present invention, a vibration detection method and a vibration detection device for a cold rolling mill capable of accurately predicting the generation of vibration can be provided. Furthermore, according to the present invention, a cold rolling method, a cold rolling mill, and a method for manufacturing steel sheets capable of preventing the generation of vibration and performing cold rolling with high productivity can be provided. Additionally, according to the present invention, a method for manufacturing steel sheets capable of producing steel sheets with desired product characteristics and mechanical properties with high yield can be provided.

[0122] Explanation of reference numerals in the attached figures

[0123] 1 Cold rolling mill

[0124] 3. Lubrication coolant head

[0125] 4 Cooling head

[0126] 5. Dirty / Soil Tank (Recycling Tank)

[0127] 6 Iron powder removal device

[0128] 7. Clean tank (storage tank)

[0129] 8 pumps

[0130] 9 supply lines

[0131] 10 oil pan

[0132] 11 Return to Piping

[0133] 13. Emulsion Rolling Oil (First Emulsion Rolling Oil)

[0134] 14 Rolling control device

[0135] 15 arithmetic units

[0136] 16 Vibration measuring device

[0137] 17 Load Sensors

[0138] 18 tension meter

[0139] 20 Second Rolling Oil Supply System

[0140] 21, 23 Rolling Oil Pipelines

[0141] 22 emulsion tank

[0142] Flow control valves 24 and 29

[0143] 25 lotion nipples

[0144] 26 Mixed Rolling Oil Pipeline

[0145] 27 Rolling Oil Crude Oil Tank

[0146] 28 warm water bath

[0147] 30 mixer

[0148] 31 Second Emulsion Rolling Oil

[0149] 32 supply pump

[0150] 71 arithmetic unit

[0151] 74 Predictive Model Building Program

[0152] 75 Vibration Prediction Monitoring Program

[0153] 76 Computational Processing Unit

[0154] 77 Predictive Model Production Department

[0155] 77A Learning Data Acquisition Department

[0156] 77B Pre-processing Unit

[0157] 77C First Data Transformation Unit

[0158] 77D Model Making Department

[0159] 77E Results Preservation Department

[0160] 78 Vibration Prediction Monitoring Unit

[0161] 78A Information Reading Unit

[0162] 78B Second Data Conversion Unit

[0163] 78C Vibration Prediction Department

[0164] 78D Rolling Conditions Determination Department

[0165] 78E Result Output Section

[0166] 88 input device

[0167] 89 storage devices

[0168] 90 output device

[0169] 91 Operation monitoring device

[0170] S-shaped steel plate.

Claims

1. A method for detecting vibration in a cold rolling mill, comprising the following steps: By inputting a prediction model learned from a first multidimensional data source (which uses one-dimensional array data representing past rolling performances of rolled materials using a cold rolling mill as explanatory variables) and a second multidimensional data source (which uses conditional data related to the array data associated with the rolled material as the target variable) into a prediction model, the generation of vibration during rolling of the rolled material can be predicted.

2. The vibration detection method for a cold rolling mill according to claim 1, The condition data is a one-dimensional array representing the time-varying changes in the rolling conditions of the workpiece or the operating state of the cold rolling mill when the workpiece is rolled. The first multidimensional data and the second multidimensional data are data obtained by connecting the conditional data or the data after shifting the conditional data in the time direction in a first direction and / or a second direction.

3. The vibration detection method for a cold rolling mill according to claim 2, The time interval for shifting the conditional data in the time direction is less than 1 second, but excluding 0 seconds.

4. The vibration detection method for a cold rolling mill according to claim 1, The conditional data is a one-dimensional array representing the signal strength of the FFT. The first multidimensional data and the second multidimensional data are data obtained by connecting the conditional data or the conditional data collected at time intervals in a first direction and / or a second direction.

5. The method for detecting vibration in a cold rolling mill according to any one of claims 1 to 4, comprising the following steps: If vibration is anticipated, the rolling speed of the cold rolling mill on the workpiece is reduced.

6. The method for detecting vibration in a cold rolling mill according to any one of claims 1 to 4, The cold rolling mill is a continuous cold rolling mill with multiple rolling stands, and includes a first rolling oil supply system that supplies a first emulsion rolling oil to each rolling stand and a second rolling oil supply system that supplies a second emulsion rolling oil with a different concentration than the first emulsion rolling oil to a portion of the rolling stands. The vibration detection method for the cold rolling mill includes the following steps: when vibration is predicted, the supply amount of the second emulsion rolling oil from the second rolling oil supply system is changed.

7. A vibration detection device for a cold rolling mill, comprising the following units: By inputting a prediction model learned from a first multidimensional data source (which uses one-dimensional array data representing past rolling performances of rolled materials using a cold rolling mill as explanatory variables) and a second multidimensional data source (which uses conditional data related to the array data associated with the rolled material as the target variable) into a prediction model, the generation of vibration during rolling of the rolled material can be predicted.

8. A cold rolling method, comprising the step of rolling a rolled material using the vibration detection method of the cold rolling mill according to any one of claims 1 to 6.

9. A cold rolling mill, comprising the vibration detection device of claim 7.

10. A method for manufacturing a steel plate, comprising: The cold rolling process uses the cold rolling method described in claim 8 to cold roll the rolled material; and the annealing process applies a homogenization temperature of 600–950°C and a furnace tension of 0.1–3.0 kgf / mm² to the rolled material after the cold rolling process. 2 Annealing process.

Images