Prediction device, learning device, prediction program, and learning program

Abstract

To accurately predict characteristics for each position in a longitudinal direction of a steel material.SOLUTION: A material prediction device 10 predicts characteristics of a steel material which is manufactured in a manufacturing line and extends in a the manufacturing line direction. The material prediction device 10 includes: a measured data acquisition section 32 which acquires a plurality of pieces of measured data measured by a plurality of sensors in the manufacturing line for a prediction object position in a longitudinal direction of the steel material; and a prediction section 40 which predicts characteristics of the prediction object position of the steel material from the plurality of pieces of measured data acquired for the prediction object position using a neural network preliminarily learnt to predict the characteristics of the prediction object position of the steel material on the basis of the plurality of pieces of the measured data for the prediction object position.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a prediction device, a learning device, a prediction program, and a learning program for predicting properties of a steel material. [Background technology]

[0002] Conventionally, a method has been known for predicting the material properties of a continuous annealing process by constructing a nonlinear relational equation between the material properties of a high-strength cold-rolled steel sheet and multiple material-influencing factors using a hierarchical neural network (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2010-106314 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the method described in the above Patent Document 1 does not take into consideration the fact that the high-strength cold-rolled steel sheet has a material variation in the longitudinal direction.

[0005] The present invention has been made in consideration of the above-mentioned circumstances, and provides a prediction device, a learning device, a prediction program, and a learning program that can accurately predict characteristics at each position in the longitudinal direction of a steel material. [Means for solving the problem]

[0006] In order to achieve the above-mentioned object, a prediction device according to a first aspect of the present invention is a prediction device that predicts the characteristics of steel material manufactured on a manufacturing line and extending in the direction of the manufacturing line, and includes an acquisition unit that acquires multiple measurement data measured by multiple sensors on the manufacturing line for a predicted target position in the longitudinal direction of the steel material, and a prediction unit that uses a pre-trained neural network to predict the characteristics of the predicted target position of the steel material from the multiple measurement data for the predicted target position of the steel material acquired by the acquisition unit, wherein the neural network is a neural network for predicting the characteristics of the predicted target position of the steel material based on the multiple measurement data for the predicted target position in the longitudinal direction of the steel material manufactured on the manufacturing line.

[0007] According to the prediction device of the first aspect of the present invention, multiple measurement data are obtained by multiple sensors on the production line for a prediction target position in the longitudinal direction of the steel material, and a neural network is used to predict the characteristics of the prediction target position of the steel material, thereby making it possible to accurately predict the characteristics for each position in the longitudinal direction of the steel material.

[0008] Here, the measurement data is data measured by a sensor on the steel material being manufactured on the production line, and the measurement data on the predicted target position is data measured by a sensor on the predicted target position of the steel material being manufactured on the production line. The characteristics of the steel material are the properties of the steel material finally manufactured on the production line, for example, the material quality of the steel material. The characteristics of the predicted target position are the properties of the steel material finally manufactured on the production line at the predicted target position.

[0009] In the prediction device according to a second aspect of the present invention, the production line is a hot rolling line, and the steel material is a hot-rolled steel sheet.

[0010] In the prediction device according to a third aspect of the present invention, the plurality of measurement data include at least one of the temperature, size, and conveying speed of the steel material measured by the plurality of sensors in the production line.

[0011] In the prediction device according to a fourth aspect of the present invention, the properties include at least one of yield stress, tensile strength, and elongation.

[0012] A learning device according to a fifth aspect of the present invention is a learning device that learns a neural network for predicting characteristics of steel material manufactured on a manufacturing line and extending in the direction of the manufacturing line, and includes an acquisition unit that acquires multiple measurement data measured by multiple sensors in the manufacturing line for each position including at least one of the front and rear ends of each of a plurality of the steel materials, and a learning unit that learns a neural network based on the multiple measurement data for each position including at least one of the front and rear ends acquired by the acquisition unit and the characteristics of the steel material measured for each position including at least one of the front and rear ends, and the neural network is a neural network for predicting characteristics of a predicted target position of the steel material manufactured on the manufacturing line based on multiple measurement data for the predicted target position in the longitudinal direction of the steel material.

[0013] According to the learning device of the fifth aspect of the present invention, multiple measurement data are obtained by multiple sensors in a production line for each position including at least one of the front and rear ends of each of multiple steel materials, and a neural network is trained based on the multiple measurement data for each position including at least one of the front and rear ends and the characteristics of the steel materials measured for each position including at least one of the front and rear ends, thereby making it possible to accurately predict the characteristics for each position in the longitudinal direction of the steel materials.

[0014] In a learning device according to a sixth aspect of the present invention, the acquisition unit acquires the multiple measurement data for each position of at least one of the tip and rear end of each of the multiple steel materials, and the learning unit learns the neural network based on the multiple measurement data for each position of at least one of the tip and rear end acquired by the acquisition unit and the characteristics of the steel material measured for each position of at least one of the tip and rear end.

[0015] In the learning device of the seventh aspect of the present invention, the characteristics of the steel material for each position including at least one of the front and rear ends are measured on a sample taken from that position.

[0016] A prediction program according to an eighth aspect of the present invention is a prediction program for predicting the characteristics of steel material manufactured on a manufacturing line and extending in the direction of the manufacturing line, the prediction program causing a computer to function as an acquisition unit that acquires multiple measurement data measured by multiple sensors in the manufacturing line for a predicted target position in the longitudinal direction of the steel material, and an acquisition unit that acquires multiple measurement data measured by multiple sensors in the manufacturing line for a predicted target position in the longitudinal direction of the steel material, and a prediction unit that predicts the characteristics of the predicted target position of the steel material from the multiple measurement data for the predicted target position of the steel material acquired by the acquisition unit using a pre-trained neural network, and the neural network is a neural network for predicting the characteristics of the predicted target position of the steel material based on the multiple measurement data for the predicted target position in the longitudinal direction of the steel material manufactured on the manufacturing line.

[0017] A learning program according to a ninth aspect of the present invention is a learning program for learning a neural network for predicting the characteristics of steel material manufactured on a manufacturing line and extending in the direction of the manufacturing line, the learning program causing a computer to function as an acquisition unit that acquires multiple measurement data measured by multiple sensors in the manufacturing line for each position of at least one of the front and rear ends of each of a plurality of the steel materials, and a learning unit that learns a neural network based on the multiple measurement data for each position of at least one of the front and rear ends acquired by the acquisition unit and the characteristics of the steel material measured for each position of at least one of the front and rear ends, and the neural network is a neural network for predicting the characteristics of the steel material at a predicted target position in the longitudinal direction of the steel material manufactured on the manufacturing line based on multiple measurement data for the predicted target position. Effect of the Invention

[0018] According to the present invention, it is possible to predict the characteristics of a steel material at each position in the longitudinal direction with high accuracy. [Brief description of the drawings]

[0019] [Figure 1] FIG. 1 is a diagram showing an example of a schematic configuration of a hot rolling line. [Diagram 2] FIG. 1 is a configuration diagram of a material prediction device. [Diagram 3] 13 is a flowchart of a learning process. [Figure 4] 13 is a flowchart of a material prediction process. [Diagram 5] FIG. 1 is a graph showing the correlation between the measured tensile strength and the predicted tensile strength in Comparative Example 1. [Figure 6] FIG. 2 is a diagram showing measured values ​​and predicted values ​​of tensile strength at each position in the longitudinal direction of a hot-rolled steel sheet in Comparative Example 1. [Figure 7] FIG. 2 is a graph showing the correlation between the measured tensile strength values ​​and the predicted tensile strength values ​​in Example 1. [Figure 8]FIG. 2 is a diagram showing the actual measured values ​​and predicted values ​​of tensile strength at each position in the longitudinal direction of the hot-rolled steel sheet in Example 1. [Figure 9] FIG. 13 is a graph showing the correlation between the measured tensile strength values ​​and the predicted tensile strength values ​​in Example 2. [Figure 10] FIG. 13 is a diagram showing the actual measured values ​​and predicted values ​​of tensile strength at each position in the longitudinal direction of the hot-rolled steel sheet in Example 2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that in the embodiment of the present invention, a prediction device and a learning device are described as an example of a material prediction device that predicts the material quality of a hot-rolled member.

[0021] <Outline of the embodiment of the present invention> Until now, the prediction and control of the material properties of hot-rolled steel sheets in factories has been done mainly by observing the behavior of process data and by using the experience and track record of the manufacturing site, although physical models based on fundamental principles have also been used as a reference. However, this method still has poor prediction accuracy, and there have been cases where, for example, the out-of-range parts of the material properties were predicted and discarded based on experience, and then the material properties were evaluated again, but were still out-of-range. Even if the re-evaluation of the material properties shows that the properties are within the product specifications, the possibility of having discarded them too much cannot be denied. Therefore, a highly accurate technology for predicting the material properties of hot-rolled steel sheets in the longitudinal direction is required.

[0022] In order to improve the accuracy of material prediction for hot-rolled steel sheets, the following problems (1) to (5) exist.

[0023] (1) It is necessary to improve the accuracy of physical models based on fundamental principles.

[0024] (2) It is necessary to improve the accuracy of process models that predict the cooling behavior of plates in factories.

[0025] (3) Measurement errors in process conditions, such as process temperature, must be identified.

[0026] (4) It is necessary to clarify hidden variables in each factory.

[0027] (5) It is necessary to identify random factors due to disturbances.

[0028] Up until now, material predictions have been centered on physical models based on fundamental principles, but there are limitations to predicting the material properties of actual products using only these models, because prediction accuracy is greatly affected by factors other than the above-mentioned physics and models.

[0029] In recent years, instead of physical models, material prediction using neural network (NN) models has been attracting attention. In actual manufacturing sites, although there is a certain range of variation, products are produced stably. From this, there should be a correlation between manufacturing conditions and materials. Therefore, if big data related to manufacturing conditions linked to the material of a factory can be used as training data to model the correlation using NN, it is possible to build a material prediction model with higher accuracy than a physical model, and there are several prior documents so far. For example, the above-mentioned Patent Document 1 describes a manufacturing method in which a nonlinear relational expression between the material of a cold-rolled steel sheet and a plurality of material influence factors is obtained using NN, the target material of the cold-rolled steel sheet and the remaining material influence factors excluding the intentional control factor, which is one of the material influence factors, are substituted into this nonlinear relational expression to obtain a target value of the intentional control factor that obtains the target material, and the above-mentioned intentional control factor is controlled to this target value, thereby extremely reducing the material variation between cold-rolled steel sheets.

[0030] However, there is no technology yet available for predicting the material properties of hot-rolled steel sheets.

[0031] Furthermore, in addition to reducing the average variation between individual hot-rolled steel sheets, it is also necessary to reduce the material variation in the longitudinal direction of a single hot-rolled steel sheet, but there is no technology related to this yet.

[0032] Furthermore, there are problems with NN models. Various disturbances exist in manufacturing sites, and the relationship between measurement data on process conditions and materials is not deterministic. Therefore, with a normal NN, there is a possibility that the generalization error cannot be reduced due to overlearning. As a countermeasure, a Bayesian Neural Network (BNN), which combines NN with Bayesian estimation, is effective, and in this embodiment, a BNN is used as the prediction model.

[0033] In addition, the material quality at each position in the longitudinal direction of the hot-rolled steel sheet is predicted from the prediction model using this BNN. To this end, measurement data on the process conditions at each position in the longitudinal direction of the hot-rolled steel sheet (steel sheet temperature, steel sheet thickness, steel sheet transport speed, etc.) are acquired and input into the prediction model using the BNN to predict the material quality at each position in the longitudinal direction of the hot-rolled steel sheet.

[0034] Furthermore, even if this BNN reduces the generalization error of learning, problems with prediction extrapolation may occur. The material at each position in the longitudinal direction of a hot-rolled steel plate may deviate significantly from the product specifications depending on the location. Basically, NNs, including BNNs, have poor prediction extrapolation performance, and in order to predict unexpected material deviations, ingenuity is required in collecting and cleaning data. For this reason, data on hot-rolled steel plates that do not meet product specifications is also taken into account in the learning of the BNN, thereby improving the accuracy of material prediction for out-of-specification parts.

[0035] <Outline of hot rolling line configuration> FIG. 1 is a diagram showing an example of a schematic configuration of a hot rolling line which is an example of an application of the material quality prediction device 10. As shown in FIG.

[0036] In FIG. 1, the hot rolling line includes a heating furnace 11, a width direction rolling mill 13, a roughing mill 14, a finishing mill 15, a cooling device (run-out table) 16, and a winding device (coiler) 17.

[0037] The heating furnace 11 heats a slab (cast piece) S.

[0038] The width direction rolling machine 13 rolls the slab S heated in the heating furnace 11 in the width direction.

[0039] The roughing mill 14 rolls the slab S, which has been rolled in the width direction by the width direction rolling mill 13, from above and below to form a rough bar. In the example shown in Fig. 1, the roughing mill 14 has a rolling stand 14a consisting of only a work roll, and rolling stands 14b to 14e each having a work roll and a backup roll.

[0040] The finishing mill 15 continuously performs hot finish rolling on the rough bar produced by the roughing mill 14 until the rough bar has a predetermined thickness. In the example shown in Fig. 1, the finishing mill 15 has seven rolling stands 15a to 15g.

[0041] The cooling device 16 cools the hot-rolled steel sheet H (hereinafter simply referred to as the steel sheet H) that has been subjected to hot finish rolling by the finishing mill 15 with cooling water.

[0042] The winding device 17 winds the steel sheet H cooled by the cooling device 16 into a coil shape to produce a hot rolled coil.

[0043] The hot rolling line can be realized by known techniques and is not limited to the configuration shown in FIG.

[0044] Further, a sensor 20 is provided at a rear stage of the heating furnace 11 to constantly measure the temperature and size of the slab S extracted from the heating furnace 11. As a result, the sensor 20 measures and outputs the temperature and size at each position in the longitudinal direction of the slab S. Here, size means at least one of the thickness and width.

[0045] A sensor 22 is provided downstream of the roughing mill 14 to constantly measure the temperature, thickness, and transport speed of the rough bar. The sensor 22 measures and outputs the temperature, thickness, and transport speed at each position in the longitudinal direction of the rough bar. The sensor 22 also measures and outputs the passing time of the tip of the rough bar.

[0046] Further, a sensor 23 for measuring the passage time of the leading end of the rough bar is provided upstream of the finishing mill 15. Thus, the sensor 23 measures and outputs the passage time of the leading end of the rough bar.

[0047] Further, a sensor 24 is provided at a downstream location of the finishing rolling mill 15 to constantly measure the temperature, thickness, and conveying speed of the steel sheet H. As a result, the sensor 24 measures and outputs the temperature, thickness, and conveying speed at each position in the longitudinal direction of the steel sheet H.

[0048] Further, a sensor 25 is provided downstream of the cooling device 16 to constantly measure the temperature of the cooled steel sheet H. Thus, the sensor 25 measures and outputs the temperature at each position in the longitudinal direction of the steel sheet H.

[0049] Further, a sensor 26 is provided at a location where the steel sheet H is wound into a coil, and measures the temperature of the steel sheet H at any time when the steel sheet H is wound into a coil. Thus, the sensor 26 measures and outputs the temperature of the steel sheet H at each position in the longitudinal direction of the steel sheet H wound into a coil.

[0050] The sensors 20, 22, 23, 24, 25, and 26 are configured using, for example, infrared sensors for measuring temperature, and sensors for capturing images, analyzing the captured images, and measuring the size, conveying speed, and passing time of the leading edge. Here, the size means the size of the slab S, the plate thickness of the rough bar, and the plate thickness of the steel plate H.

[0051] <Configuration of material prediction device> Next, the configuration of the material prediction device will be described.

[0052] As shown in FIG. 2, the material prediction device 10 includes a measurement data acquisition unit 32, a material data acquisition unit 34, a learning unit 36, a prediction model storage unit 38, a prediction unit 40, and a display unit .

[0053] During learning, the measurement data acquisition unit 32 acquires at least one of the temperature, size, and conveying speed measured by multiple sensors 20, 22, 24, 25, 26 in the hot rolling line for each position of at least one of the front and rear ends of each of the multiple hot-rolled steel plates.

[0054] Specifically, the measurement data acquisition unit 32 acquires the additive concentration input by the input unit (not shown) during learning. The measurement data acquisition unit 32 also acquires the time from the time when the rough rolling of the front end of the rough bar by the roughing mill 14 is completed to the time when the rough bar enters the finishing mill 15 from the output of the sensors 22 and 23. The measurement data acquisition unit 32 also acquires the output (temperature, plate thickness, and conveying speed) of each of the sensors 20, 22, 24, 25, and 26 corresponding to the front end of the steel sheet H. The measurement data acquisition unit 32 also acquires the output (temperature, plate thickness, and conveying speed) of each of the sensors 20, 22, 24, 25, and 26 corresponding to the rear end of the steel sheet H. Note that at least one of the temperature, size, and conveying speed measured by the multiple sensors 20, 22, 24, 25, and 26 in the hot rolling line may be acquired for multiple positions including at least one of the front end and the rear end and other positions.

[0055] During prediction, the measurement data acquisition unit 32 acquires at least one of the temperature, size, and conveying speed measured by the multiple sensors 20, 22, 24, 25, 26 in the hot rolling line for each position in the longitudinal direction of the hot rolled steel plate.

[0056] Specifically, the measurement data acquisition unit 32 acquires the additive concentration input by an input unit (not shown) at the time of prediction. The measurement data acquisition unit 32 also acquires the time from the time when rough rolling of the tip of the rough bar by the roughing mill 14 is completed to the time when the tip enters the finishing mill 15, from the outputs of the sensors 22 and 23. The measurement data acquisition unit 32 also acquires the outputs (temperature, plate thickness, and conveying speed) of the sensors 20, 22, 24, 25, and 26 corresponding to each position in the longitudinal direction of the steel sheet H.

[0057] The material data acquisition unit 34 acquires material data measured on samples taken from at least one of the front end and rear end of each of the multiple hot-rolled steel sheets, inputted by an input unit (not shown) during learning. The material data includes, for example, at least one of yield stress (YS), tensile strength (TS), and elongation (EL). The samples are collected offline while rewinding the hot-rolled coil wound by the winding device 17.

[0058] The learning unit 36 ​​creates teacher data for each of a plurality of hot-rolled steel plates, the teacher data being composed of various measurement data acquired by the measurement data acquisition unit 32 and material data acquired by the material data acquisition unit 34 for each position of at least one of the front and rear ends of the hot-rolled steel plates.

[0059] The learning unit 36 ​​learns a prediction model, which is a neural network for predicting the material of a prediction target position of a hot-rolled steel sheet based on various measurement data about the prediction target position in the longitudinal direction of the hot-rolled steel sheet produced on a hot rolling line, based on the created teacher data, and stores the prediction model in the prediction model storage unit 38. In this embodiment, a Bayesian neural network is used as the neural network. The neural network also receives as inputs the temperature and size of the slab S at the prediction target position at the rear of the heating furnace 11, the temperature, thickness, and transport speed of the rough bar at the prediction target position at the rear of the roughing mill 14, the time from the time when the rough rolling of the tip of the rough bar is completed by the roughing mill 14 to the time when the tip of the rough bar enters the finishing mill 15, the temperature, thickness, and transport speed of the steel sheet H at the prediction target position at the rear of the finishing mill 15, the temperature of the cooled steel sheet H at the prediction target position at the rear of the cooling device 16, and the temperature of the coiled steel sheet H at the winding position by the winding device 17. The neural network also receives as output at least one of the yield stress (YS), tensile strength (TS), and elongation (EL) at the prediction target position.

[0060] In addition, for each training data, the neural network is trained so that a loss function that represents the difference between the output when various measurement data of the training data are input to the neural network and the material data of the training data is minimized.

[0061] The prediction unit 40 uses a prediction model stored in the prediction model memory unit 38 to predict material data of the prediction target position of the hot-rolled steel plate from various measurement data for the prediction target position acquired by the measurement data acquisition unit 32.

[0062] The display unit 42 displays the prediction result by the prediction unit 40 .

[0063] Next, the learning process executed by the material quality prediction device 10 will be described with reference to the flowchart shown in Fig. 3. The learning process shown in Fig. 3 is executed when at least one of the temperature, size, and conveying speed measured by the multiple sensors 20, 22, 24, 25, 26 in the hot rolling line for each position of at least one of the front end and rear end of each of the multiple hot rolled steel sheets is input, and material quality data measured for samples taken from each position of at least one of the front end and rear end of each of the multiple hot rolled steel sheets is input.

[0064] In step S100, the measurement data acquisition unit 32 acquires at least one of the temperature, size, and conveying speed measured by the multiple sensors 20, 22, 24, 25, 26 in the hot rolling line for each position of at least one of the front end and rear end of each of the multiple hot rolled steel plates. In addition, the measurement data acquisition unit 32 acquires the additive concentration inputted by an input unit (not shown) for each of the multiple hot rolled steel plates.

[0065] In step S102, the material data acquiring unit 34 acquires material data measured on samples taken from each position of at least one of the front end and the rear end of each of the input hot-rolled steel plates.

[0066] In step S104, the learning unit 36 ​​creates teacher data for each of the multiple hot-rolled steel plates, the teacher data being composed of various measurement data acquired by the measurement data acquisition unit 32 and material data acquired by the material data acquisition unit 34 for each position of at least one of the front and rear ends of the hot-rolled steel plates.

[0067] In step S106, the learning unit 36 ​​learns a neural network for predicting the characteristics of the predicted target position of the hot-rolled steel plate based on various measurement data about the predicted target position in the longitudinal direction of the hot-rolled steel plate manufactured in the hot rolling process, based on the created teacher data, and stores the neural network in the prediction model memory unit 38.

[0068] Next, the material prediction process executed by the material prediction device 10 will be described with reference to the flowchart shown in Fig. 4. The material prediction process shown in Fig. 4 is executed when at least one of the temperature, size, and conveying speed measured by a plurality of sensors 20, 22, 24, 25, and 26 in the hot rolling line is input for a prediction target position in the longitudinal direction of the hot-rolled steel sheet.

[0069] In step S110, the measurement data acquisition unit 32 acquires at least one of the temperature, size, and conveying speed measured by the multiple sensors 20, 22, 24, 25, and 26 in the hot rolling line for the prediction target position in the longitudinal direction of the hot-rolled steel sheet. The measurement data acquisition unit 32 also acquires the additive concentration inputted by an input unit (not shown).

[0070] In step S112, the prediction unit 40 uses a neural network stored in the prediction model memory unit 38 to predict material data of the prediction target position of the hot-rolled steel plate from various measurement data for the prediction target position of the hot-rolled steel plate acquired by the measurement data acquisition unit 32.

[0071] In step S114, the display unit 42 displays the prediction result by the prediction unit 40, and the material prediction process ends.

[0072] <Example> First, in the learning process for constructing a prediction model, measurement data of the leading ends of multiple hot-rolled steel sheets and material data measured from samples taken from the leading ends are used as teacher data.

[0073] By inputting measurement data at each longitudinal position of a specific hot-rolled steel plate into the predictive model constructed through learning, the material of the hot-rolled steel plate at each longitudinal position is predicted.

[0074] As examples of training data used for learning, the following measurement data and material data are used.

[0075] The measurement data includes the concentrations of each of the additives C, Si, Mn, Al, Mo, Ti, Nb, N, and B, the extraction temperature from the heating furnace and the slab size, the plate temperature, plate thickness, and conveying speed during rough rolling, the time until the leading end enters finish rolling after the end of rough rolling, the plate temperature, plate thickness, and conveying speed during finish rolling, the plate temperature on the cooling table, and the coiling temperature.

[0076] Material data includes yield stress (YS), tensile strength (TS), and elongation (EL).

[0077] The Bayesian neural network structure, which is the prediction model, has three hidden layers between the input layer and the output layer. The number of neurons in each hidden layer, starting from the first layer, is 64, 32, and 16. All activation functions are hyperbolic tangent functions.

[0078] (Comparative Example 1 (average value of the entire steel plate length used as measurement data)) First, Figures 5 and 6 show an example of learning using only data from approved hot-rolled steel plates as training data. For example, Figure 5 shows the correlation between the actual measured tensile strength (TS) values ​​and the values ​​predicted by the prediction model. It can be seen that a good correlation was obtained. This prediction model was used to predict the material properties of each position in the longitudinal direction of the hot-rolled steel plate. In this example, the average value for the entire length of the hot-rolled steel plate was used as the measured data. The results are shown in Figure 6. It can be seen that predictions were made with a fair degree of accuracy, but there is room for improvement in prediction accuracy.

[0079] (Example 1 (values ​​at each position on the steel plate are used as measurement data)) Next, Figs. 7 and 8 show examples in which values ​​at each position in the longitudinal direction of a hot-rolled steel sheet are used as measurement data. Naturally, the correlation between the actual measurements and predictions of TS during learning is the same as in Fig. 5. Fig. 8 shows the material properties at each position in the longitudinal direction of a hot-rolled steel sheet under these conditions. Compared to Comparative Example 1, the prediction accuracy of tensile strength (TS) has improved. However, the prediction of the tip of the hot-rolled steel sheet is still insufficient.

[0080] (Example 2 (Example 1 + Failed Data)) Next, Figures 9 and 10 show the results when data on rejected hot-rolled steel sheets was added as training data during learning of the prediction model, and the data on passed hot-rolled steel sheets was thinned out by 30%. The area surrounded by the solid ellipse in Figure 9 corresponds to the data on rejected hot-rolled steel sheets, and the area surrounded by the dashed ellipse is the range of data on passed hot-rolled steel sheets, with the data thinned out. Figure 10 shows the material properties at each position in the longitudinal direction of the hot-rolled steel sheet obtained using the prediction model constructed through this learning, from the measurement data at each position in the longitudinal direction of the hot-rolled steel sheet. It can be seen that the prediction accuracy has improved significantly.

[0081] In this manner, in this embodiment, multiple measurement data are obtained for the prediction target position in the longitudinal direction of the hot-rolled steel sheet by multiple sensors in the hot rolling line, and a neural network is used to predict the material of the prediction target position of the hot-rolled steel sheet, thereby making it possible to accurately predict the material for each position in the longitudinal direction of the hot-rolled steel sheet.

[0082] Moreover, by acquiring a plurality of measurement data measured by a plurality of sensors in a hot rolling line for each position of at least one of the front end and the rear end of each of a plurality of hot rolled steel sheets, and learning a neural network for predicting the material of a prediction target position of the hot rolled steel sheet based on the plurality of measurement data for each position of at least one of the front end and the rear end and the material of the hot rolled steel sheet measured for each position of at least one of the front end and the rear end, it is possible to predict the material of each position in the longitudinal direction of the hot rolled steel sheet with high accuracy. Moreover, regardless of whether the product is acceptable or unacceptable, it is only necessary to take samples from the ends of a plurality of hot rolled steel sheets and create teacher data using the measured material data, thereby reducing the burden of creating teacher data.

[0083] In addition, since the material quality can be accurately predicted for each position in the longitudinal direction of the hot-rolled steel sheet, the management of the parts of the hot-rolled steel sheet that do not meet the product specifications can be greatly simplified. In other words, the work of taking samples from the hot-rolled coil, discarding the parts that do not meet the product specifications if the taken samples are not within the product specifications, and then taking samples again while rewinding the hot-rolled coil can be omitted.

[0084] In addition, since an appropriate truncation range can be determined from the material prediction results for each position in the longitudinal direction of the hot-rolled steel plate, the amount of waste of parts of the hot-rolled steel plate that do not meet the product specifications can be reduced, thereby improving product yield.

[0085] Therefore, in this embodiment, a Bayesian neural network is used to predict the material quality at each position in the longitudinal direction of a hot-rolled steel sheet.

[0086] In addition, training data consisting of process data and material data for the longitudinal ends of the hot-rolled steel sheet is used in the neural network training. Training data for hot-rolled steel sheets that have been determined to have failed in the material inspection is also used. In this case, data cleaning is performed to reduce the amount of training data for hot-rolled steel sheets that have been determined to have passed the material inspection, thereby increasing the proportion of training data for hot-rolled steel sheets that have been determined to have failed.

[0087] This makes it possible to improve the accuracy of material prediction at each position in the longitudinal direction of the hot-rolled steel sheet. In addition, since the material data measured on the samples taken at the end portions in the longitudinal direction of the hot-rolled steel sheet is used, the teacher data can be easily generated.

[0088] <Modification> In the above embodiment, the material prediction process and the learning process are performed by one device, but the present invention is not limited to this. The device may be separated into a prediction device that performs the material prediction process and a learning device that performs the learning process. In this case, the learning device includes a measurement data acquisition unit 32, a material data acquisition unit 34, a learning unit 36, and a prediction model storage unit 38. The prediction device includes a measurement data acquisition unit 32, a prediction unit 40, and a display unit 42.

[0089] In addition, in this embodiment, the case where the sensors are installed at the rear of the heating furnace 11, the rear of the roughing mill 14, the front of the finishing mill 15, the rear of the finishing mill 15, the rear of the cooling device 16, and the winding location by the winding device 17 has been described, but the installation locations are not limited to these. The sensors may be installed at other locations.

[0090] The measurement data that is the input of the neural network is not limited to the example described above, and the material data that is the output of the neural network is not limited to the example described above.

[0091] In the present embodiment, the case of predicting the material quality of a hot-rolled steel sheet has been described as an example, but the present invention is not limited thereto, and the prediction may be made for a property other than the material quality of the hot-rolled steel sheet, for example, a property such as the shape may be predicted. In this case, the measurement data used as the input for the prediction is the measurement data related to the influencing factor selected for the property to be predicted. In addition, the prediction may be made for a property of a steel material other than a hot-rolled steel sheet, for example, a property of a cold-rolled steel sheet, a plated steel sheet, a wire rod, etc., if the property is a property of a steel material that is manufactured on a manufacturing line and extends in the direction of the manufacturing line.

[0092] The above-described embodiment of the present invention can be realized by a computer executing a program. In addition, a computer-readable recording medium on which the program is recorded and a computer program product such as the program can also be applied as an embodiment of the present invention. As the recording medium, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a magnetic tape, a non-volatile memory card, a ROM, etc. can be used. [Explanation of symbols]

[0093] 10. Material Prediction Device 11 Heating furnace 13. Cross-rolling mill 14 Roughing mill 15 Finishing rolling mill 16 Cooling device 17 Winding device 20, 22, 23, 24, 25, 26 Sensors 32 Measurement data acquisition section 34 Material data acquisition section 36 Learning Department 38 Prediction model memory unit 40 Prediction Department 42 Display section

Claims

1. A prediction device for predicting properties of a steel material produced in a production line and extending in a direction of the production line, An acquisition unit that acquires a plurality of measurement data measured by a plurality of sensors in the production line with respect to a prediction target position in the longitudinal direction of the steel material; a prediction unit that predicts characteristics of the steel material at the prediction target position from the plurality of measurement data for the prediction target position of the steel material acquired by the acquisition unit using a pre-trained neural network, the neural network is a neural network for predicting characteristics of a predicted target position of a steel material produced on the production line based on the plurality of measurement data for the predicted target position in the longitudinal direction of the steel material, In the learning process, the teacher data includes data on rejected steel products, and the prediction device learns a neural network based on multiple measurement data measured by multiple sensors on the production line for each position including at least one of the front and rear ends of each of the multiple steel products, and the characteristics of the steel products measured for each position including at least one of the front and rear ends.

2. A learning device that learns a neural network for predicting properties of a steel material that is manufactured on a manufacturing line and extends in a direction of the manufacturing line, an acquisition unit that acquires, as data on steel materials including rejected products, a plurality of measurement data measured by a plurality of sensors in the manufacturing line for each position including at least one of a front end and a rear end of each of a plurality of the steel materials; a learning unit that learns a neural network based on the plurality of measurement data for each position including at least one of the front end and the rear end acquired by the acquisition unit and the characteristics of the steel material measured for each position including at least one of the front end and the rear end, The neural network is a learning device that is a neural network for predicting characteristics of a predicted target position of a steel material produced on the production line based on multiple measurement data for the predicted target position in the longitudinal direction of the steel material.

3. The acquisition unit acquires the plurality of measurement data for each position of at least one of a front end portion and a rear end portion of each of the plurality of steel materials, The learning device described in claim 2, wherein the learning unit learns the neural network based on the multiple measurement data for each position of at least one of the front and rear ends acquired by the acquisition unit and the characteristics of the steel material measured for each position of at least one of the front and rear ends.

4. 4. A learning device according to claim 2 or 3, wherein the characteristics of the steel material at each position including at least one of the front end and the rear end are measured on a sample taken from the corresponding position.

5. A prediction device for predicting properties of a steel material produced in a production line and extending in a direction of the production line, An acquisition unit that acquires a plurality of measurement data measured by a plurality of sensors in the production line with respect to a prediction target position in the longitudinal direction of the steel material; A prediction unit that predicts characteristics of the steel material at the prediction target position from the plurality of measurement data for the prediction target position of the steel material acquired by the acquisition unit, using the neural network that has been trained in advance by the learning device according to any one of claims 2 to 4; A prediction device comprising:

6. The production line is a hot rolling line, The prediction device according to claim 1 or 5, wherein the steel material is a hot-rolled steel plate.

7. A prediction device according to any one of claims 1, 5 to 6, wherein the plurality of measurement data includes at least one of the temperature, size, and conveying speed of the steel material measured by the plurality of sensors in the production line.

8. The prediction device according to any one of claims 1, 5 to 7, wherein the properties include at least one of yield stress, tensile strength, and elongation.

9. A prediction program for predicting properties of a steel material produced in a production line and extending in a direction of the production line, Computer, An acquisition unit that acquires a plurality of measurement data measured by a plurality of sensors in the production line with respect to a predicted target position in the longitudinal direction of the steel material; and a prediction unit that predicts characteristics of the steel material at the prediction target position from the plurality of measurement data for the prediction target position of the steel material acquired by the acquisition unit using a pre-trained neural network; It is a prediction program to function as a the neural network is a neural network for predicting characteristics of a predicted target position of a steel material produced on the production line based on the plurality of measurement data for the predicted target position in the longitudinal direction of the steel material, In the learning process, a prediction program is used that trains a neural network based on multiple measurement data measured by multiple sensors on the production line for each position including at least one of the front and rear ends of each of the multiple steel pieces, including data on rejected steel pieces as teacher data, and the characteristics of the steel pieces measured for each position including at least one of the front and rear ends.

10. A learning program for learning a neural network for predicting properties of a steel material produced on a production line and extending in a direction of the production line, comprising: Computer, An acquisition unit that acquires a plurality of measurement data measured by a plurality of sensors in the manufacturing line for each position of at least one of a front end and a rear end of each of a plurality of steel materials, as data of the steel materials including the rejected products; a learning unit that learns a neural network based on the plurality of measurement data for each position of at least one of the front end portion and the rear end portion acquired by the acquisition unit and the characteristics of the steel material measured for each position of at least one of the front end portion and the rear end portion. This is a learning program to enable students to function as The neural network is a learning program that is a neural network for predicting the characteristics of a predicted target position of a steel material produced on the production line based on multiple measurement data for the predicted target position in the longitudinal direction of the steel material.

Images