Quality factor determination device and quality factor determination method

Abstract

The quality factor determination device 1 is used in batch production equipment having setup calculation and dynamic control functions. The quality factor determination device 1 includes a calculation unit 30. The calculation unit 30 acquires first classification information based on an operational fluctuation range within a batch calculated based on operational performance data and a first threshold value, acquires second classification information based on an operational representative value within a batch calculated based on the operational performance data and a second threshold value, and, upon obtaining a first determination result, presents the first threshold value as an allowable value for the operational fluctuation range to be suppressed by dynamic control, and, upon obtaining a second determination result, presents the second threshold value as an optimal setting value in the setup calculation.

Description

[Technical Field]

[0001] The present disclosure relates to a quality factor determining device and a quality factor determining method used in batch production equipment. [Background technology]

[0002] Conventionally, techniques for improving quality defects in products produced in production facilities have been known. For example, Patent Document 1 discloses a support system that supports an operator in estimating the cause of a quality abnormality. The support system described in Patent Document 1 focuses on the correlation between operational variables contained in past operational performance data and product quality, extracts operational variables that have a strong correlation with product quality, and displays a time series chart of the operational variables to the operator. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6116445 Summary of the Invention [Problem to be solved by the invention]

[0004] The support system described in Patent Document 1 merely extracts operational variables that are highly correlated with quality from product quality data and operational condition data for each product ID, and displays a time series chart of the operational variables to the operator, in order to identify the cause of quality abnormalities and encourage the operator to make improvements. However, in actual quality improvement work, two methods can be considered, including a method of improving quality by improving setup calculations that pre-set operational conditions for each steel strip immediately before manufacturing the steel strip, and a method of improving quality by improving dynamic control that suppresses operational fluctuations within the steel strip.

[0005] The support system described in Patent Document 1 has a problem in that it is not easy for the operator to determine which of the above quality improvement methods should be attempted to improve the relevant operational variable. In addition, in the support system described in Patent Document 1, the operational condition data used to calculate the correlation with quality is data of values representative of each steel strip. Therefore, there is also a problem in that it is not possible to properly extract operational variables for operational conditions that cause product quality defects due to fluctuations within a steel strip, as opposed to fluctuations between steel strips. For these reasons, the convenience of the system for users, such as operators, attempting to improve product quality has been reduced.

[0006] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a quality factor determination device and a quality factor determination method that improve the convenience of the device for users attempting to improve the quality of products. [Means for solving the problem]

[0007] (1) A quality factor determination device according to an embodiment of the present disclosure, A quality factor determination device for use in a batch production facility having functions of setup calculation and dynamic control, comprising: The calculation unit includes: acquire first classification information that classifies the quality of the product for each batch based on an operational fluctuation range within the batch and a first threshold value calculated based on operational performance data including actual operational condition data when the product is manufactured through one or more manufacturing processes and product quality data corresponding to the batch of the manufacturing process; acquiring second classification information that classifies the quality of the product for each batch based on an operation representative value within the batch calculated based on the operation performance data and a second threshold value; When a first judgment result is obtained indicating that the variation in the actual operational conditions within a batch is likely to be a cause of the quality defect of the product, the first threshold value is determined as an allowable value of the operational variation range to be suppressed by the dynamic control, When a second judgment result is obtained that indicates that there is a high possibility that the batch-to-batch variation in the actual operating conditions is the cause of the quality defects of the product, the second threshold value is determined as the optimum setting value in the setup calculation.

[0008] (2) As one embodiment of the present disclosure, in (1), The calculation unit acquires the first classification information from a cross-tabulation table based on the first threshold value.

[0009] (3) As one embodiment of the present disclosure, in (2), The calculation unit obtains the second classification information from a cross-tabulation table based on the second threshold value.

[0010] (4) As an embodiment of the present disclosure, in (3), The calculation unit obtains the first judgment result based on the fact that the value of the first classification accuracy calculated by an accuracy test based on the first classification information is smaller than the value of the second classification accuracy calculated by an accuracy test based on the second classification information.

[0011] (5) As an embodiment of the present disclosure, in (4), The calculation unit obtains the second determination result based on the value of the second classification accuracy being equal to or less than the value of the first classification accuracy.

[0012] (6) As an embodiment of the present disclosure, in any one of (1) to (5), The calculation unit When the first determination result is obtained, the first threshold value is presented as a tolerance of the operational fluctuation range to be suppressed by the dynamic control, When the second determination result is obtained, the second threshold value is presented as an optimum setting value in the setup calculation.

[0013] (7) As an embodiment of the present disclosure, in any one of (1) to (6), the product comprises a steel product; The quality includes at least one of surface defects, internal quality, and material quality.

[0014] (8) As an embodiment of the present disclosure, in any one of (1) to (7), The batch is based on at least one of a hot metal ladle unit and a torpedo unit in the ironmaking process in steel production, a charge unit in the steelmaking process, a continuous unit and a slab unit in continuous casting, a hot rolled steel strip unit in the hot rolling process, a cold rolled steel strip unit in the cold rolling process, and an annealed steel strip unit in the annealing process and surface treatment process.

[0015] (9) As an embodiment of the present disclosure, in any one of (1) to (8), The quality factor determination device further includes a storage unit in which the operational performance data is stored.

[0016] (10) A quality factor determination method according to an embodiment of the present disclosure includes: 1. A quality factor determination method for use in a batch production facility having setup calculation and dynamic control functions, comprising: acquiring first classification information that classifies the quality of the product for each batch based on an operational fluctuation range within the batch and a first threshold value calculated based on operational performance data including actual operational condition data when the product is manufactured through one or more manufacturing processes and product quality data corresponding to the batch of the manufacturing process; acquiring second classification information that classifies the quality of the product for each batch based on an operation representative value within the batch calculated based on the operation performance data and a second threshold value; When a first determination result is obtained that indicates that there is a high possibility that the variation in the actual operational conditions within a batch is a cause of the quality defect of the product, determining the first threshold value as an allowable value of the operational variation range to be suppressed by the dynamic control; When a second determination result is obtained indicating that the batch-to-batch variation of the actual operating conditions is highly likely to be a cause of the quality defect of the product, determining the second threshold value as an optimal setting value in the setup calculation; Includes. [Effects of the Invention]

[0017] According to the quality factor determination device and the quality factor determination method according to an embodiment of the present disclosure, the convenience of the device is improved for users attempting to improve the quality of products. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a block diagram illustrating a schematic configuration of a quality factor determination device according to an embodiment of the present disclosure. [Figure 2] FIG. 10 is a first diagram for explaining an example of the operation of the quality factor determining device of FIG. [Figure 3] FIG. 2 is a second diagram for explaining an example of the operation of the quality factor determining device of FIG. [Figure 4] 2 is a flowchart showing an example of processing executed by the quality factor determining device of FIG. 1; [Figure 5] FIG. 3 is a third diagram for explaining an example of the operation of the quality factor determining device of FIG. [Figure 6] FIG. 4 is a fourth diagram for explaining an example of the operation of the quality factor determining device of FIG. [Figure 7] FIG. 5 is a fifth diagram for explaining an example of the operation of the quality factor determining device of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0019] Hereinafter, the configuration and operation of a quality factor determining device 1 according to an embodiment of the present disclosure will be mainly described with reference to the accompanying drawings.

[0020] The quality factor determination device 1 is used in batch production equipment that has functions of setup calculation and dynamic control. In this disclosure, "batch production" refers to, for example, a production method in which the setup and equipment settings of the manufacturing equipment are performed for each predetermined production unit of the product. "Product" includes, for example, steel products.

[0021] In this disclosure, "setup calculation" refers to a function of calculating, immediately before the start of production, the setpoints of the manufacturing equipment so that the product quality, including the final quality of the product and product characteristic values recognized to be strongly correlated with the final quality of the product, reaches the control target value as a control target, and setting up the manufacturing equipment. "Product characteristic values" include, for example, characteristic values such as the temperature of the product obtained during the manufacturing process. "Product quality" includes, for example, at least one of surface defects, internal quality, and material quality. "Internal quality" refers to, for example, the distribution state of non-metallic inclusions trapped inside the product during casting, which affects weldability and the occurrence of defects such as cracks. Other internal quality factors may include segregation and hydrogen embrittlement. "Material quality" includes, for example, the tensile strength and elongation of the product, which affect the processability during press molding. Other material quality factors may include yield strength and toughness.

[0022] In this disclosure, "dynamic control" refers to, for example, a function of periodically manipulating manufacturing equipment settings so that product quality can be maintained at a control target value from the start to the completion of product production. In batch production equipment with setup calculation and dynamic control functions, equipment settings are set for each production unit by setup calculation just before production begins, and the equipment settings are changed by dynamic control during production. This maintains product quality.

[0023] Fig. 1 is a block diagram showing a schematic configuration of a quality factor determination device 1 according to an embodiment of the present disclosure. An example of the configuration and operation of the quality factor determination device 1 according to an embodiment will be mainly described with reference to Fig. 1. The quality factor determination device 1 includes an input unit 10, an output unit 20, a calculation unit 30, and a storage unit 40.

[0024] The input unit 10 includes one or more input interfaces that detect user input and acquire input information based on the user's operation. The input interfaces include physical keys such as a keyboard, capacitive keys, a mouse pointer, a touch screen integrated with the display of the output unit 20, an imaging module such as a camera, and a microphone that accepts voice input. The input unit 10 is an input means for the calculation unit 30.

[0025] The output unit 20 includes one or more output interfaces that output information to provide it to the user, such as a display that visually outputs information as an image, a speaker that audibly outputs information as a sound, and a vibrator that tactilely outputs information as a vibration.

[0026] The storage unit 40 includes storage modules such as a hard disk drive (HDD), a solid state drive (SSD), an electrically erasable programmable read-only memory (EEPROM), a read-only memory (ROM), and a random access memory (RAM). The storage unit 40 may function as a main storage module, an auxiliary storage module, or a cache memory. The storage unit 40 is not limited to being built into the quality factor determination device 1, and may include storage media such as removable media. The removable media include a universal serial bus (USB) memory, a compact disc (CD), a digital versatile disc (DVD), and a Blu-ray (registered trademark) disc (BD).

[0027] The storage unit 40 stores information necessary to realize the operation of the quality factor determination device 1. The storage unit 40 stores information obtained by the operation of the quality factor determination device 1. For example, the storage unit 40 can store an operating system (OS), various programs, various tables, various databases, etc. The storage unit 40 stores an operation DB (database) 41.

[0028] The operation DB 41 stores operational performance data of products manufactured in the past in a searchable batch unit. For example, if the product is a steel product manufactured through multiple manufacturing processes, the "manufacturing process" includes at least one of the ironmaking process, steelmaking process, hot rolling process, cold rolling process, annealing process, and surface treatment process. The "batch unit" includes at least one of the molten iron ladle unit and torpedo unit in the ironmaking process in steel manufacturing, the charge unit in the steelmaking process, the continuous unit and slab unit in continuous casting, the hot-rolled steel strip unit in the hot-rolling process, the cold-rolled steel strip unit in the cold-rolling process, and the annealed steel strip unit in the annealing process and surface treatment process. The batch unit is variable for each manufacturing process.

[0029] In the present disclosure, "operational performance data" includes, for example, actual operational condition data when a product is manufactured through one or more manufacturing processes, and product quality data corresponding to a batch of the manufacturing process. "Actual operational condition data" includes, for example, manufacturing condition data and sensor data in multiple manufacturing processes when a product was manufactured in the past. "Quality data" includes, for example, data on the quality of a product inspected after the final process of multiple manufacturing processes.

[0030] If the product is a steel product, the actual operating condition data includes, for example, data such as product thickness, product width, product length, steel composition, and casting speed in the steelmaking process, data such as product thickness, product width, product length, product temperature, and rolling speed in the hot rolling process, data such as product thickness, product width, product length, and rolling speed in the cold rolling process, and data such as product thickness, product width, product length, surface treatment furnace temperature, and surface treatment speed in the surface treatment process. The quality data includes, for example, data such as inspection results for surface defects of the product at the end of the surface treatment process, internal quality test results, and material test results.

[0031] FIG. 2 is a first diagram for explaining an example of the operation of the quality factor determination device 1 of FIG. 1. FIG. 2 is a table diagram showing an example of the above-mentioned actual operating condition data. In FIG. 2, actual operating condition data for one batch c is shown. In the table shown in FIG. 2, the vertical column marked with "P" corresponds to, for example, the position of the final product corresponding to the quality inspection or the elapsed time of production of batch c. The vertical column includes N+1 in-product position numbers from 0 to N or the elapsed time of production of batch c. In the table shown in FIG. 2, "X1" to "X M The horizontal row marked with " corresponds to the operating conditions for batch c, for example. M For example, "X m n (c)" is the operating condition X at the n position of the product m The values of in batch c are shown.

[0032] As described above, the operating conditions include, for example, the product thickness, product width, product length, steel composition, and casting speed in the steelmaking process; the product thickness, product width, product length, product temperature, and rolling speed in the hot rolling process; the product thickness, product width, product length, and rolling speed in the cold rolling process; and the product thickness, product width, product length, temperature in the surface treatment furnace, and surface treatment speed in the surface treatment process.

[0033] 1 , the calculation unit 30 includes one or more processors. In the present disclosure, a "processor" refers to, but is not limited to, a general-purpose processor or a dedicated processor specialized for a specific process. The calculation unit 30 includes, for example, a CPU (Central Processing Unit). The calculation unit 30 is communicably connected to each component of the quality factor determination device 1 and controls the operation of the quality factor determination device 1 as a whole.

[0034] The calculation unit 30 loads a program into the working area of the main memory module of the storage unit 40, executes the program, and controls each component unit through the execution of the program. In this way, the calculation unit 30 realizes functions that meet a predetermined purpose. Through the execution of the program, the calculation unit 30 functions as an intra-batch variation factor determination unit 31, an inter-batch variation factor determination unit 32, and a quality factor determination unit 33. While FIG. 1 shows an example in which the functions of the three functional units are realized by, for example, one computer, the means for realizing the functions of each functional unit is not limited to this. For example, the functions of the three functional units may each be realized by multiple computers.

[0035] The intra-batch variation factor determination unit 31 of the calculation unit 30 acquires first classification information that classifies the quality of the products for each batch based on the operational variation range within the batch calculated based on the operational performance data stored in the storage unit 40 and a first threshold value. The following describes in more detail the function of the intra-batch variation factor determination unit 31 of the calculation unit 30.

[0036] The intra-batch variation factor determination unit 31 reads the actual operational condition data shown in FIG. 2 stored as part of the operational performance data in the operation DB 41 of the storage unit 40 for each batch c. Based on the read actual operational condition data, the intra-batch variation factor determination unit 31 determines the operational condition X m In this disclosure, the "operational fluctuation range" is, for example, m This includes the difference between the maximum and minimum values in the data, the standard deviation, and other statistics related to the range of variation.

[0037] The intra-batch variation factor determination unit 31 reads the operational performance data stored in the operation DB 41 of the memory unit 40, and calculates a first threshold value of the operational variation range by machine learning to classify the product groups into those in which the product quality, such as surface defects, is minor or the products are undamaged, and those in which the product quality, such as surface defects, is severe, and to obtain first classification information.

[0038] FIG. 3 is a second diagram illustrating an example of the operation of the quality factor determination device 1 of FIG. 1. The intra-batch variation factor determination unit 31 acquires first classification information using a cross-tabulation table based on a first threshold. In the present disclosure, the "first classification information" may include, for example, a cross-tabulation table based on the first threshold, or a scatter plot as shown in FIG. 6, which will be described later. For example, the intra-batch variation factor determination unit 31 classifies the operational variation range data for each batch c using two variables: a newly labeled severity level based on the first threshold and the original severity level based on the quality data included in the operational performance data, as a cross-tabulation table, to acquire the first classification information. In FIG. 3, a, b, c, and d each indicate the number of data items that meet the corresponding condition.

[0039] The intra-batch variation factor determination unit 31 calculates a first classification accuracy value, which indicates how accurately the first threshold classifies the first classification information into good product quality, i.e., minor / undamaged, and bad product quality, i.e., severe, as a p-value using Fisher's exact test on the cross-tabulation table. p1, the p-value for the first classification accuracy, is calculated using the following formula 1, where i=a1+b1+c1+d1.

[0040]

number

[0041] The batch-to-batch variation factor determination unit 32 of the calculation unit 30 acquires second classification information that classifies the quality of the products for each batch based on a representative operation value within the batch calculated based on the operation performance data stored in the storage unit 40 and a second threshold value. The following describes in more detail the function of the batch-to-batch variation factor determination unit 32 of the calculation unit 30.

[0042] The batch-to-batch variation factor determination unit 32 reads the actual operational condition data shown in FIG. 2 stored as part of the operational performance data in the operation DB 41 of the storage unit 40 for each batch c. Based on the read actual operational condition data, the batch-to-batch variation factor determination unit 32 determines the operational condition X m In this disclosure, the "operational representative value" is, for example, m This includes statistics relating to medians, means, and other representative values in the data.

[0043] The batch-to-batch variation factor determination unit 32 reads the operational performance data stored in the operation DB 41 of the memory unit 40, and classifies the product groups into those in which the product quality, such as surface defects, is minor or the products are undamaged, and those in which the product quality, such as surface defects, is severe, and calculates a second threshold value of the operational representative value by machine learning to obtain second classification information.

[0044] The inter-batch variation factor determination unit 32 acquires second classification information using a cross-tabulation table based on the second threshold, similar to the processing details of the intra-batch variation factor determination unit 31 described using Fig. 3. In the present disclosure, the "second classification information" may include, for example, a cross-tabulation table based on the second threshold, or may include a scatter plot of Fig. 7 described below. For example, the inter-batch variation factor determination unit 32 classifies the data of the operation representative values for each batch c using two variables, the severity newly labeled using the second threshold and the original severity based on the quality data included in the operation performance data, as a cross-tabulation table, and acquires second classification information.

[0045] The inter-batch variation factor determination unit 32 calculates a second classification accuracy value, which indicates how accurately the second threshold classifies the second classification information into good product quality, i.e., minor / undamaged, and bad product quality, i.e., severe, as a p-value using Fisher's exact test on the cross-tabulation table. p2, the p-value for the second classification accuracy, is calculated using the following equation 2, where i=a2+b2+c2+d2.

[0046]

number

[0047] Fig. 4 is a flowchart showing an example of processing executed by the quality factor determining device 1 in Fig. 1. The flowchart shown in Fig. 4 shows the overall flow of processing related to determining the cause of a quality defect in a product by the quality factor determining device 1 and presenting information to a user.

[0048] In step S101, the quality factor determination unit 33 of the quality factor determination device 1 determines whether or not there is a high possibility that intra-batch variation in actual operating conditions is a factor in product quality defects. For example, the quality factor determination unit 33 determines whether or not the value of the first classification accuracy p1 calculated by Fisher's exact test based on the first classification information acquired by the intra-batch variation factor determination unit 31 is smaller than the value of the second classification accuracy p2 calculated by Fisher's exact test based on the second classification information acquired by the inter-batch variation factor determination unit 32. In addition, the quality factor determination unit 33 determines whether or not the value of the first classification accuracy p1 is smaller than 1 / i (the value obtained by dividing 1 by the number of data i). The value used for the determination is not limited to 1 / i and may be any other value. A smaller value used for the determination means that the criterion for determining that intra-batch variation is a factor in product quality defects is stricter.

[0049] When the quality factor determination unit 33 obtains a first determination result indicating that within-batch variation in actual operating conditions is highly likely to be a cause of product quality defects, the quality factor determination unit 33 executes the process of step S102. When the quality factor determination unit 33 obtains the first determination result based on the value of the first classification accuracy p1 being smaller than the value of the second classification accuracy p2, the quality factor determination unit 33 executes the process of step S102. For example, when the quality factor determination unit 33 determines that the value of the first classification accuracy p1 is smaller than the value of the second classification accuracy p2 and that the value of the first classification accuracy p1 is smaller than 1 / i, the quality factor determination unit 33 executes the process of step S102.

[0050] When the quality factor determination unit 33 obtains a determination result other than the first determination result, it executes the process of step S103. For example, when the quality factor determination unit 33 determines that the value of the first classification accuracy p1 is equal to or greater than the value of the second classification accuracy p2 and / or that the value of the first classification accuracy p1 is equal to or greater than 1 / i, it executes the process of step S103.

[0051] In step S102, when the first judgment result is obtained in step S101, the calculation unit 30 of the quality factor judgment device 1 outputs first information using the output unit 20. For example, the calculation unit 30 displays the first information using the display of the output unit 20. In the present disclosure, the "first information" includes, for example, first visual information. The "first visual information" includes, for example, the name of the operating conditions that were the subject of judgment, a message that the intra-batch variation of the operating conditions is highly likely to be a cause of the defective product quality, the value of the first threshold calculated by the intra-batch variation factor judgment unit 31, and a message that the first threshold corresponds to the allowable value of the operational variation range that should be suppressed by dynamic control in order to maintain good product quality.

[0052] As described above, when the calculation unit 30 obtains the first judgment result in step S101 that the intra-batch variation of the actual operating conditions is highly likely to be the cause of the defective product quality, in step S102, the calculation unit 30 presents the first threshold value as the allowable value of the operational variation range that should be suppressed by dynamic control.

[0053] In step S103, if the calculation unit 30 of the quality factor determination device 1 obtains a determination result other than the first determination result in step S101, the calculation unit 30 determines whether or not there is a high possibility that inter-batch variation in actual operating conditions is a factor in the quality of the product. For example, the quality factor determination unit 33 determines whether or not the value of the second classification accuracy p2 calculated by Fisher's exact test based on the second classification information acquired by the inter-batch variation factor determination unit 32 is equal to or less than the value of the first classification accuracy p1 calculated by Fisher's exact test based on the first classification information acquired by the intra-batch variation factor determination unit 31. In addition, the quality factor determination unit 33 determines whether or not the value of the second classification accuracy p2 is smaller than 1 / i. The value used for the determination is not limited to 1 / i and may be any other value. A smaller value used for the determination means that the criterion for determining that there is a high possibility that intra-batch variation is a factor in the quality of the product is stricter.

[0054] When the quality factor determination unit 33 obtains a second determination result indicating that batch-to-batch variation in actual operating conditions is highly likely to be a cause of product quality defects, the quality factor determination unit 33 executes the process of step S104. When the quality factor determination unit 33 obtains the second determination result based on the value of the second classification accuracy p2 being equal to or less than the value of the first classification accuracy p1, the quality factor determination unit 33 executes the process of step S104. For example, when the quality factor determination unit 33 determines that the value of the second classification accuracy p2 is equal to or less than the value of the first classification accuracy p1 and that the value of the second classification accuracy p2 is smaller than 1 / i, the quality factor determination unit 33 executes the process of step S104.

[0055] When the quality factor determination unit 33 obtains a determination result other than the second determination result, it executes the process of step S105. For example, when the quality factor determination unit 33 determines that the value of the second classification accuracy p2 is greater than the value of the first classification accuracy p1 and / or the value of the second classification accuracy p2 is 1 / i or more, it executes the process of step S105. For example, when the value of both the first classification accuracy p1 and the value of the second classification accuracy p2 are 1 / i or more, the quality factor determination unit 33 executes the process of step S105.

[0056] In step S104, when the second determination result is obtained in step S103, the calculation unit 30 of the quality factor determination device 1 outputs second information using the output unit 20. For example, the calculation unit 30 displays the second information using the display of the output unit 20. In the present disclosure, the "second information" includes, for example, second visual information. The "second visual information" includes, for example, the name of the operating conditions that were the subject of the determination, a message that batch-to-batch variation in the operating conditions is highly likely to be a cause of poor product quality, the value of the second threshold calculated by the batch-to-batch variation factor determination unit 32, and a message that the second threshold corresponds to an optimal setting value in the setup calculation for maintaining good product quality.

[0057] As described above, when the calculation unit 30 obtains the second judgment result in step S103 indicating that the batch-to-batch variation in the actual operating conditions is highly likely to be the cause of the defective product quality, in step S104, the calculation unit 30 presents the second threshold value as the optimal setting value in the setup calculation.

[0058] In step S105, when the calculation unit 30 of the quality factor determination device 1 obtains a determination result other than the second determination result in step S103, it outputs third information using the output unit 20. For example, the calculation unit 30 displays the third information using the display of the output unit 20. In the present disclosure, the "third information" includes, for example, third visual information. The "third visual information" includes, for example, a message that the operating conditions that were the subject of the determination are likely to have little impact on product quality. The third information includes information that both intra-batch variation and inter-batch variation of the actual operating conditions are likely to have little impact on product quality.

[0059] To summarize the above, the quality factor determination unit 33 of the quality factor determination device 1 compares the value of the first classification accuracy p1 based on the first threshold calculated by the intra-batch variation factor determination unit 31 with the value of the second classification accuracy p2 based on the second threshold calculated by the inter-batch variation factor determination unit 32. If the value of the first classification accuracy p1 is smaller than the value of the second classification accuracy p2 and is smaller than 1 / i, the calculation unit 30 outputs, as first information via the output unit 20, an indication that intra-batch variation in actual operating conditions has a greater impact than inter-batch variation. The first information also includes the fact that the first threshold calculated by the intra-batch variation factor determination unit 31 is set as the allowable value of the operational variation range to be suppressed by dynamic control.

[0060] If the value of the second classification accuracy p2 is equal to or less than the value of the first classification accuracy p1 and is smaller than 1 / i, the calculation unit 30 outputs, as second information via the output unit 20, information indicating that inter-batch variation in actual operating conditions has a greater impact than intra-batch variation. The second information also includes setting the second threshold calculated by the inter-batch variation factor determination unit 32 as an optimal setting value in the setup calculation for maintaining good product quality. [Example]

[0061] An example of the present disclosure will be described below with reference to Fig. 5 to Fig. 7. In the example, an example will be described in which a quality factor determination device 1 according to an embodiment of the present disclosure is applied to the manufacture of steel sheet products.

[0062] 5 is a third diagram for explaining an example of the operation of the quality factor determination device 1 of FIG. 1. For example, the operation DB 41 of the storage unit 40 stores the actual operating conditions X shown in FIG. m As an example, values of the operating conditions X1 of the casting process are stored as a set. Figure 5 shows a chart of the operating conditions X1 for one batch of slab c among the data of the operating conditions X1 stored in the operation DB 41. The horizontal axis in Figure 5 corresponds to the vertical columns in Figure 2.

[0063] The intra-batch variation factor determination unit 31 reads the data of the operating condition X1 of the casting process stored in the operation DB 41 and calculates the operational variation range for each slab. For example, for the data of slab c corresponding to FIG. 5, the intra-batch variation factor determination unit 31 calculates the difference of 0.15 (=0.66-0.51) between the maximum value of the operating condition X1, 0.66 (position 100° from the front end of the slab), and the minimum value of the operating condition X1, 0.51 (position 10° from the front end of the slab). The operating condition X1 in the embodiment may be, for example, the mold thermocouple temperature measured by a thermocouple embedded in the outer surface of the mold during the casting process. In FIG. 5, the value of the operating condition X1 is normalized to be between 0 and 1.

[0064] Fig. 6 is a fourth diagram for explaining an example of the operation of the quality factor determination device 1 of Fig. 1. The intra-batch variation factor determination unit 31 performs similar calculations on other slab data to obtain data on the operational variation range for each slab, as shown in Fig. 6. Data for 80 slabs is shown in Fig. 6. There are a total of 80 pieces of table data shown in Fig. 2 as actual operational condition data for batch c.

[0065] In Figure 6, the vertical axis is normalized so that the operational fluctuation range ranges from 0 to 1. Each slab is labeled A, B, or C, which indicates the rank of product quality. Product quality is ranked in the order A>B>C. In Figure 6, slabs Nos. 0 to 19 are labeled C, slabs Nos. 20 to 39 are labeled B, and slabs Nos. 40 to 79 are labeled A.

[0066] In the present disclosure, the intra-batch variation factor determination unit 31 is required to calculate a first threshold value for classifying operational variation range data into two categories: "Quality A" and "Quality B or C." "Quality A" corresponds to "minor / undamaged" in the cross-tabulation table of FIG. 3. "Quality B or C" corresponds to "severe" in the cross-tabulation table of FIG. 3. The first threshold value calculated by machine learning is indicated by a solid line in the scatter plot of FIG. 6. The intra-batch variation factor determination unit 31 classifies coils whose operational variation range value is above the first threshold solid line as good, and coils whose operational variation range value is below the first threshold solid line as bad, and acquires first classification information.

[0067] The intra-batch variation factor determination unit 31 calculates 0.63 as the value of the first classification accuracy p1 by Fisher's exact test based on the first classification information as shown in Fig. 6. The smaller the p-value, the higher the accuracy of classifying product quality using the first threshold value.

[0068] Product quality is judged based on surface defects inspected by a surface defect inspection device installed at the exit of the final annealing process. This surface defect inspection device detects only defects from images of the coil surface captured by a line sensor camera and extracts optical features such as the length, width, and brightness of the defects. The surface defect inspection device determines whether the product is good or bad based on the extracted optical features.

[0069] For example, if a steel sheet is used for the body of an automobile, surface defects will cause unevenness in the paint, and the product will be deemed unsuitable for quality, and measures will be taken such as cutting out only the areas with surface defects. Each label indicating the rank of product quality is determined by a comprehensive assessment based on the number of surface defects within the same coil, the degree of accumulation of the defects, and the severity of the flaws, such as their depth.

[0070] The batch-to-batch variation factor determination unit 32 reads the data of the operating conditions X1 of the casting process stored in the operation DB 41 and calculates the operation representative value for each slab. For example, for the data of slab c corresponding to Figure 5, the batch-to-batch variation factor determination unit 32 calculates the median value of the operating conditions X1, 0.56 (position 50 cm away from the front end of the slab).

[0071] Fig. 7 is a fifth diagram for explaining an example of the operation of the quality factor determination device 1 of Fig. 1. The batch-to-batch variation factor determination unit 32 performs similar calculations on the other slab data to obtain data on the representative operation values for each slab, as shown in Fig. 7. Data for 80 slabs is shown in Fig. 7. There are a total of 80 pieces of table data shown in Fig. 2 as the actual operating condition data for batch c.

[0072] In Figure 7, the vertical axis is normalized so that the operational representative value is between 0 and 1. Each slab is labeled A, B, or C, which indicates the rank of the product quality. Product quality is ranked in the order A>B>C. In Figure 7, slabs Nos. 0 to 19 are labeled C, slabs Nos. 20 to 39 are labeled B, and slabs Nos. 40 to 79 are labeled A.

[0073] In the present disclosure, the batch-to-batch variation factor determination unit 32 is required to calculate a second threshold value for classifying the data of the operational representative value into two categories, "Quality A" and "Quality B or C." "Quality A" corresponds to "minor / undamaged" in the cross-tabulation table of FIG. 3. "Quality B or C" corresponds to "severe" in the cross-tabulation table of FIG. 3. The second threshold value calculated by machine learning is indicated by a solid line in the scatter plot of FIG. 7. The batch-to-batch variation factor determination unit 32 classifies coils whose operational representative value is below the second threshold solid line as good and coils whose operational representative value is above the second threshold solid line as bad, and acquires second classification information.

[0074] The inter-batch variation factor determination unit 32 calculates the value of the second classification accuracy p2 by Fisher's exact test to be 0.0052 based on the second classification information as shown in Fig. 7. The smaller the p value, the higher the accuracy of classifying product quality using the second threshold value.

[0075] The quality factor determination unit 33 compares the first classification accuracy p1 (0.63) based on the first threshold calculated by the intra-batch variation factor determination unit 31 with the second classification accuracy p2 (0.0052) based on the second threshold calculated by the inter-batch variation factor determination unit 32. Because the quality factor determination unit 33 determines that the second classification accuracy p2 based on the second threshold is smaller and that the second classification accuracy p2 (0.0052) is smaller than 1 / 80 (=0.0125), the calculation unit 30 executes the process of step S104 in the flowchart of FIG. 4 . Regarding the operating condition X1 of the casting process, the calculation unit 30 presents, as second information via the output unit 20, the fact that the inter-batch variation of the actual operating conditions significantly affects product quality and that the influence of the intra-batch variation is small. The second information also includes the second threshold value of 0.4 calculated by the inter-batch variation factor determination unit 32 as an optimal setting value in the setup calculation for maintaining good product quality.

[0076] According to the quality factor determining device 1 according to the embodiment described above, the convenience of the device is improved for users who are trying to improve the quality of products.

[0077] The calculation unit 30 of the quality factor determination device 1 acquires first classification information that classifies the quality of the product for each batch based on the operational fluctuation range within the batch calculated based on the operational performance data stored in the memory unit 40 and a first threshold value. When the calculation unit 30 obtains a first determination result that indicates that the intra-batch fluctuation of the performance operational conditions is highly likely to be a factor in the poor quality of the product, it presents the first threshold value as the allowable value of the operational fluctuation range that should be suppressed by dynamic control.

[0078] For example, the calculation unit 30 of the quality factor determination device 1 calculates the operational fluctuation range within a batch from time-series data of the operational conditions within the batch, and calculates whether the relationship between the calculated operational fluctuation range and the quality of the product is statistically significant. This enables the calculation unit 30 to present an allowable value for the operational fluctuation range that should be suppressed by dynamic control.

[0079] The calculation unit 30 of the quality factor determination device 1 acquires second classification information that classifies the quality of the product for each batch based on a representative operation value within the batch calculated based on the operation performance data stored in the storage unit 40 and a second threshold value. When the calculation unit 30 obtains a second determination result that indicates that the batch-to-batch variation in the performance operating conditions is highly likely to be a factor in the poor quality of the product, it presents the second threshold value as an optimal setting value in the setup calculation.

[0080] For example, the calculation unit 30 of the quality factor determination device 1 calculates a representative operation value within a batch from time-series data of the operation conditions within the batch, and calculates whether the relationship between the calculated representative operation value and the quality of the product is statistically significant. This enables the calculation unit 30 to present optimal setting values for the setup calculation that minimizes quality defects of the product.

[0081] As a result, the user can easily determine whether to improve the product quality by improving the setup calculation or by improving the dynamic control in order to solve the current product quality problem. As a result, the convenience of the quality factor determination device 1 is improved for users who are trying to improve the quality of their products.

[0082] Although the present disclosure has been described based on the drawings and examples, it should be noted that those skilled in the art can make various modifications and alterations based on the present disclosure. Therefore, it should be noted that these modifications and alterations are included in the scope of the present disclosure. For example, the functions included in each configuration or step can be rearranged so as not to be logically inconsistent, and multiple configurations or steps can be combined or divided into one.

[0083] For example, the shape, size, pattern, arrangement, orientation, type, and number of each of the above-mentioned components are not limited to those shown in the above description and drawings. The shape, size, pattern, arrangement, orientation, type, and number of each component may be configured arbitrarily as long as the function can be realized. Each component of the illustrated quality factor determination device 1 is a functional concept. The specific form of each component is not limited to that shown in the drawings.

[0084] For example, a general-purpose electronic device such as a smartphone or a computer can be configured to function as the quality factor determination device 1 according to the above-described embodiment. Specifically, a program describing the processing content for realizing each function of the quality factor determination device 1 according to the embodiment is stored in the memory of the electronic device, and the program is read and executed by a processor of the electronic device. Therefore, the present disclosure can also be realized as a program executable by a processor.

[0085] Alternatively, the present disclosure may be realized as a non-transitory computer-readable medium storing a program executable by one or more processors to cause the quality factor determination device 1 according to an embodiment to execute each function, etc. It should be understood that these are also included within the scope of the present disclosure.

[0086] In the above embodiment, the calculation unit 30 of the quality factor determination device 1 calculates the p-value based on the Fisher exact test, but this is not limiting. The calculation unit 30 may calculate a value equivalent to the p-value based on another exact test, or may calculate the classification accuracy value based on a method other than the exact test.

[0087] In the above embodiment, the calculation unit 30 of the quality factor determination device 1 presents the first threshold value as the allowable range of operational fluctuations to be suppressed by dynamic control when it obtains a first determination result indicating that intra-batch variation in actual operational conditions is likely to be a factor in product quality defects. However, this is not limited to this. Instead of or in addition to such a presentation process, the calculation unit 30 of the quality factor determination device 1 may actually execute dynamic control based on the first threshold value as the allowable range of operational fluctuations to be suppressed by dynamic control. As a preliminary step to the presentation process and / or execution process described above, the calculation unit 30 determines the first threshold value as the allowable range of operational fluctuations to be suppressed by dynamic control.

[0088] In the above embodiment, the calculation unit 30 of the quality factor determination device 1 presents the second threshold as the optimal setting value in the setup calculation when it obtains the second determination result indicating that batch-to-batch variation in the actual operating conditions is likely to be a factor in the quality defect of the product. However, this is not limited to this. Instead of or in addition to such a presentation process, the calculation unit 30 of the quality factor determination device 1 may actually execute the setup calculation based on the second threshold as the optimal setting value in the setup calculation. The calculation unit 30 determines the second threshold as the optimal setting value in the setup calculation as a preliminary step to the presentation process and / or execution process described above.

[0089] In the above embodiment, the quality factor determination device 1 has been described as having the storage unit 40, in which operation performance data is stored, integrally with the calculation unit 30, but this is not limited to this. The quality factor determination device 1 does not have to have the storage unit 40. The storage unit 40 may be included in a device separate from the quality factor determination device 1. The same applies to the input unit 10 and the output unit 20; the quality factor determination device 1 only needs to have at least the calculation unit 30. [Explanation of symbols]

[0090] 1. Quality factor determination device 10 Input section 20 Output section 30 Arithmetic section 31 Intra-batch variation factor determination section 32 Batch variation factor determination section 33 Quality Factor Judgment Unit 40 Storage section 41 Operation DB

Claims

1. A quality factor determination device for use in a batch production facility having functions of setup calculation and dynamic control, comprising: The calculation unit includes: acquiring first classification information that classifies the quality of the product for each batch based on an operational fluctuation range within a batch calculated based on operational performance data including actual operational condition data when a product is manufactured through one or more manufacturing processes and product quality data corresponding to a batch of the manufacturing process, and a first threshold value of the operational fluctuation range that is calculated by machine learning based on the operational performance data and that classifies the quality as good or bad; and calculating a first classification accuracy of the first threshold value by accuracy testing based on the first classification information; acquiring second classification information that classifies the quality of the product for each batch based on an operation representative value within the batch calculated based on the operation performance data and a second threshold value of the operation representative value that is calculated by machine learning based on the operation performance data and that classifies the quality as good or bad, and calculating a second classification accuracy of the second threshold value by accuracy testing based on the second classification information; comparing the first classification accuracy with the second classification accuracy to obtain a first determination result indicating that it is highly likely that the variation in actual operating conditions within a batch is a cause of the quality defects of the product, or a second determination result indicating that it is highly likely that the variation in actual operating conditions between batches is a cause of the quality defects of the product; Quality factor determination device.

2. A quality factor determination device according to claim 1, The calculation unit When the first determination result is obtained, the first threshold value is determined as an allowable value of the operational fluctuation range to be suppressed by the dynamic control, When the second determination result is obtained, the second threshold value is determined as an optimal setting value in the setup calculation. Quality factor determination device.

3. The quality factor determination device according to claim 1, the calculation unit acquires the first classification information from a cross-tabulation table based on the first threshold value. Quality factor determination device.

4. The quality factor determination device according to claim 3, the calculation unit acquires the second classification information by a cross-tabulation table based on the second threshold value. Quality factor determination device.

5. The quality factor determination device according to claim 4, the calculation unit obtains the first determination result based on the first classification accuracy being smaller than the second classification accuracy. Quality factor determination device.

6. The quality factor determination device according to claim 5, the calculation unit obtains the second determination result based on the value of the second classification accuracy being equal to or less than the value of the first classification accuracy. Quality factor determination device.

7. The quality factor determination device according to any one of claims 1 to 6, The calculation unit When the first determination result is obtained, the first threshold value is presented as an allowable value of the operational fluctuation range to be suppressed by the dynamic control, When the second determination result is obtained, the second threshold is presented as an optimal setting value in the setup calculation. Quality factor determination device.

8. The quality factor determination device according to any one of claims 1 to 6, the product comprises a steel product; The quality includes at least one of surface defects, internal quality, and material quality. Quality factor determination device.

9. The quality factor determination device according to any one of claims 1 to 6, The batch is based on at least one of a hot metal ladle unit and a torpedo unit in an ironmaking process in steel production, a charge unit in a steelmaking process, a continuous unit and a slab unit in continuous casting, a hot rolled steel strip unit in a hot rolling process, a cold rolled steel strip unit in a cold rolling process, and an annealed steel strip unit in an annealing process and a surface treatment process. Quality factor determination device.

10. The quality factor determination device according to any one of claims 1 to 6, Further comprising a storage unit in which the operation performance data is stored. Quality factor determination device.

11. 1. A quality factor determination method for use in a batch production facility having setup calculation and dynamic control functions, comprising: acquiring first classification information that classifies the quality of the product for each batch based on an operational fluctuation range within a batch calculated based on operational performance data including actual operational condition data when a product is manufactured through one or more manufacturing processes and product quality data corresponding to a batch of the manufacturing process, and a first threshold value of the operational fluctuation range that is calculated by machine learning based on the operational performance data and that classifies the quality as good or bad, and calculating a first classification accuracy of the first threshold value by accuracy testing based on the first classification information; acquiring second classification information that classifies the quality of the product for each batch based on an operation representative value within the batch calculated based on the operation performance data and a second threshold value of the operation representative value that is calculated by machine learning based on the operation performance data and that classifies the quality as good or bad, and calculating a second classification accuracy of the second threshold value by accuracy testing based on the second classification information; comparing the first classification accuracy with the second classification accuracy to obtain a first determination result indicating that it is highly likely that the variation in actual operating conditions within a batch is a cause of the quality defects of the product, or a second determination result indicating that it is highly likely that the variation in actual operating conditions between batches is a cause of the quality defects of the product; Including, Quality factor determination method.

12. The quality factor determination method according to claim 11, comprising: When the first determination result is obtained, determining the first threshold value as an allowable value of the operational fluctuation range to be suppressed by the dynamic control; When the second determination result is obtained, determining the second threshold value as an optimal setting value in the setup calculation; further comprising: Quality factor determination method.

Images