Quality defect diagnostic device and quality defect diagnostic method

Abstract

To provide a quality defect diagnostic device and a quality defect diagnostic method, in which quality defects of a product can be predicted with high accuracy and factors of the quality defect can be estimated.SOLUTION: A quality defect diagnostic device comprises: a data division unit that two-dimensionally divides product data of a diagnostic target into prescribed blocks; a data aggregation unit that, for each category of predetermined quality defects, extracts operation data and facility maintenance data on the basis of the occurrence tendency of the quality defect in each divided block, and aggregates the extracted operation data and facility maintenance data into areas each including two or more blocks in a specific direction of the product; and a quality defect determination unit that determines occurrence or non-occurrence of the quality defect using multiple prediction models, which are created for each category of the quality defect and in which the relationship between the aggregated operation data and facility maintenance data and quality data is learned.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a quality defect diagnosis apparatus and a quality defect diagnosis method.

Background Art

[0002] Patent Document 1 discloses a method for predicting and diagnosing the quality of a product. In this method, operation condition data during product manufacturing and quality determination data in the final process are associated in consideration of the quality determination position on the product, and a prediction model learned from this data is used to predict the quality.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when predicting the quality of a product, if quality data is uniformly used regardless of its characteristics, information regarding multiple types of quality defects generated by different mechanisms will be missing. Therefore, even if a prediction model is created based on such quality data, a problem may occur in that sufficient prediction performance cannot be achieved.

[0005] The present invention has been made in view of the above, and an object thereof is to provide a quality defect diagnosis apparatus and a quality defect diagnosis method capable of accurately predicting a quality defect of a product and estimating the cause of the quality defect.

Means for Solving the Problems

[0006] In order to solve the above-described problems and achieve the object, a quality defect diagnosis apparatus according to the present invention includes: a data splitting unit that splits product data to be diagnosed, which consists of operation data, equipment maintenance data, and quality data, into a two-dimensional shape for each predetermined block; a data aggregation unit that extracts the operation data and the equipment maintenance data for each predetermined classification of quality defects based on the occurrence tendency of quality defects in the split blocks, and aggregates the extracted operation data and equipment maintenance data for each region including two or more blocks in a specific direction of the product; and a quality defect determination unit that determines the presence or absence of the occurrence of quality defects using a plurality of prediction models created for each classification of the quality defects and in which the relationship between the aggregated operation data and equipment maintenance data and the quality data has been learned.

[0007] Further, the quality defect diagnosis apparatus according to the present invention includes, in the above invention, a cause estimation unit that estimates the cause of the quality defect based on the determination result of the quality defect.

[0008] Further, the quality defect diagnosis apparatus according to the present invention includes, in the above invention, a result display unit that causes an output unit to display at least one of the determination result in the quality defect determination unit and the cause estimation result in the cause estimation unit.

[0009] Further, in the above invention, the prediction model splits product data of past performance, which consists of the operation data, the equipment maintenance data, and the quality data, into a two-dimensional shape for each predetermined block, classifies the quality defects included in the quality data based on the occurrence tendency of the quality defects in the split blocks, aggregates the operation data and the equipment maintenance data extracted for each classification of the quality defects for each region including two or more blocks in a specific direction of the product, and is created by learning the relationship between the aggregated operation data and equipment maintenance data and the quality data for each classification of the quality defects by machine learning.

[0010] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the operation data is data that varies over the length of the product in a specific direction.

[0011] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the equipment maintenance data is data that varies over the period during which the product is processed in the equipment.

[0012] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the quality data is data that varies over the length of the product in a specific direction.

[0013] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the equipment maintenance data is data indicating the equipment state and its maintenance state that affect a specific position of the product by a direct or indirect action on the product.

[0014] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the data aggregation unit extracts the operation data and the equipment maintenance data for each predetermined classification of quality defects based on the quality defect occurrence tendency having a plurality of quality defect occurrence locations over the longitudinal direction of the product or the quality defect occurrence tendency having a plurality of quality defect occurrence locations over the width direction of the product.

[0015] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the quality defect determination unit determines the presence or absence of the occurrence of a quality defect using a plurality of prediction models created for each classification of the quality defect and in which the relationship between the aggregated equipment maintenance data and the quality data has been learned, and the cause estimation unit estimates the cause of the quality defect based on the determination result of the quality defect.

[0016] Further, in the quality defect diagnosis apparatus according to the present invention, in the above invention, the result display unit causes the output unit to display the data group subjected to the alignment process in the data aggregation unit.

[0017] In order to solve the above problems and achieve the object, a quality defect diagnosis method according to the present invention includes: a data division step in which data division means provided in a computer divides product data to be diagnosed, which consists of operation data, equipment maintenance data, and quality data, into a two-dimensional shape for each predetermined block; a data aggregation step in which data aggregation provided in the computer extracts the operation data and the equipment maintenance data for each predetermined classification of quality defects based on the occurrence tendency of quality defects in the divided blocks, and aggregates the extracted operation data and equipment maintenance data for each area including two or more blocks in a specific direction of the product; and a quality defect determination step in which quality defect determination means provided in the computer determines the presence or absence of the occurrence of quality defects using a plurality of prediction models created for each classification of the quality defects and in which the relationship between the aggregated operation data and equipment maintenance data and the quality data has been learned.

[0018] Further, the quality defect diagnosis method according to the present invention includes, in the above invention, a factor estimation step in which factor estimation means provided in the computer estimates the factors of quality defects based on the determination result of quality defects.

Effects of the Invention

[0019] According to the quality defect diagnosis apparatus and the quality defect diagnosis method of the present invention, by creating and using, as prediction models, the relationships between operation data, equipment maintenance data, and quality defects for each classification according to the occurrence tendency of quality defects, it is possible to highly accurately predict quality defects of products and estimate the factors of the quality defects.

Brief Description of the Drawings

[0020]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

DETAILED DESCRIPTION OF THE INVENTION

[0021] A quality defect diagnosis apparatus and a quality defect diagnosis method according to an embodiment of the present invention will be described with reference to the drawings. The quality defect diagnosis apparatus and the quality defect diagnosis method are for diagnosing quality defects occurring in products.

[0022] In this embodiment, the "product" includes, for example, steel products manufactured through a plurality of processes, such as semi-finished products like slabs and products like steel plates manufactured by rolling the slabs. Further, in this embodiment, the "quality defect" includes, for example, surface defects, internal defects, characteristic values of materials (such as determination results of mechanical characteristics), and the like.

[0023] (Information processing apparatus) FIG. 1 shows the configuration of an information processing apparatus 1 that realizes the quality defect diagnosis apparatus according to the embodiment. As shown in FIG. 1, the information processing apparatus 1 includes an input unit 10, a storage unit 20, an arithmetic unit 30, and an output unit 40.

[0024] The quality defect diagnosis apparatus according to the embodiment can be realized by the components of the information processing apparatus 1 excluding the quality defect classification unit 32 and the model creation unit 34 of the arithmetic unit 30. Further, the apparatus for creating the prediction model used in the quality defect diagnosis apparatus according to the embodiment can be realized by the components of the information processing apparatus 1 excluding the quality defect determination unit 35, the factor estimation unit 36, and the result display unit 37 of the arithmetic unit 30. Hereinafter, each component of the information processing apparatus 1 will be described.

[0025] The input unit 10 is an input means for the arithmetic unit 30 and is realized by, for example, an input device such as a keyboard, a mouse pointer, or a numeric keypad. The input unit 10 inputs information necessary for various processes in the arithmetic unit 30.

[0026] The storage unit 20 is composed of recording media such as an EPROM (Erasable Programmable ROM), a hard disk drive (HDD), and a removable medium. Examples of the removable medium include disk recording media such as a USB (Universal Serial Bus) memory, a CD (Compact Disc), a DVD (Digital Versatile Disc), and a BD (Blu-ray (registered trademark) Disc). The storage unit 20 can store an operating system (OS), various programs, various tables, various databases, and the like. A product DB (database) 21 is stored in the storage unit 20.

[0027] The product DB 21 stores product data manufactured in the past. This product data includes, for example, operation data, equipment maintenance data, and quality data.

[0028] The operation data is data obtained by measuring or estimating by some means the manufacturing state during product manufacturing. This operation data preferably varies over the length in a specific direction of the product. The "data that varies over the length in a specific direction of the product" indicates, for example, data that continuously varies with respect to the longitudinal direction or the width direction of the product.

[0029] Examples of the operation data include, for each manufacturing process of steel products, manufacturing condition data in the steelmaking process, manufacturing condition data in the hot rolling process, manufacturing condition data in the cold rolling process, manufacturing condition data in the annealing process, and the like. Specifically, examples of the operation data include data on the components in the steel, the temperature of the steel sheet during hot rolling, the hot rolling speed, the cold rolling speed, the temperature of the steel sheet during annealing, and the annealing time. Also, the operation data is measured at predetermined fixed lengths, for example, according to the moving distance of the steel sheet in the longitudinal direction with respect to the conveyance direction of each process.

[0030] Equipment maintenance data is data that measures or estimates by some means the state of equipment and its maintenance status during product manufacturing. This equipment maintenance data preferably varies over time while products are being processed in the equipment. Also, the equipment maintenance data is preferably data indicating the state of equipment and its maintenance status that affects a specific position of a product due to direct or indirect action on the product, such as data directly or indirectly representing the contact state of equipment that contacts the product in a manufacturing process.

[0031] Examples of equipment maintenance data include data related to the contact state of conveying rolls that convey products by rotating in contact with the products in each manufacturing process of steel products. Also, examples of equipment maintenance data include data related to the contact state of walking beams that convey products by lifting them in contact with the products in a heating furnace in a heating process before rolling of products. Specifically, as equipment maintenance data, data on the shape across the width direction of the conveying roll, data on the contact position of the walking beam with respect to the product, etc. are assumed.

[0032] Among the equipment maintenance data, the data on the shape across the width direction of the conveying roll is, for example, data in which the wear depth of the roll for each predetermined constant length is measured as a roll profile with respect to the width direction of the conveying roll. Also, among the equipment maintenance data, the data on the contact position of the walking beam with respect to the product is, for example, calculated from the measured head and tail end positions in the longitudinal direction of the product in the heating furnace and the length of the product in the heating furnace and the position of the walking beam in the heating furnace. Thus, by using equipment maintenance data when creating a prediction model and when performing quality defect diagnosis, it becomes possible to accurately predict quality defects continuous in a specific direction (longitudinal direction, width direction) of a product as described below.

[0033] Quality data is data representing the quality of a product, based on the result of measuring the quality of the product or estimating it by some means. This quality data is preferably data that varies over the length of the product in a specific direction. Examples of quality data include, for example, surface defect (surface flaw) data of the product. Specifically, as quality data, data such as the position and shape of surface flaws on the product is assumed.

[0034] Note that the above product data includes product data of "past performance" used when creating a prediction model (see Fig. 10) and product data of "object to be diagnosed" used when performing quality defect diagnosis (see Fig. 11).

[0035] In the storage unit 20, in addition to the product DB 21, for example, product data associated with each block by the data division unit 31, information regarding the classification of quality defects by the quality defect classification unit 32, operation data and equipment maintenance data aggregated in block units by the data aggregation unit 33, etc. are also stored. Further, in the storage unit 20, the determination result of quality defects by the quality defect determination unit 35, the cause estimation result of quality defects by the cause estimation unit 36, etc. may be stored as necessary.

[0036] The arithmetic unit 30 is realized by, for example, a processor composed of a CPU (Central Processing Unit) etc. and a memory (main storage unit) composed of a RAM (Random Access Memory), a ROM (Read Only Memory), etc.

[0037] The arithmetic unit 30 loads a program into the working area of the main memory and executes it, and controls each component and the like through the execution of the program, thereby realizing a function that meets a predetermined purpose. The arithmetic unit 30 functions as a data division unit 31, a quality defect classification unit 32, a data aggregation unit 33, a model creation unit 34, a quality defect determination unit 35, a cause estimation unit 36, and a result display unit 37 through the execution of the above-described program. In FIG. 1, for example, an example is shown in which the functions of each unit are realized by one computer (arithmetic unit), but the means for realizing the functions of each unit is not particularly limited, and for example, the functions of each unit may be realized by a plurality of computers respectively.

[0038] The data division unit 31 divides the product data two-dimensionally for each predetermined block. First, the data division unit 31 collects operation data, equipment maintenance data, and quality data from, for example, the product DB 21. Subsequently, the data division unit 31 divides the total length of the product data in the product longitudinal direction into p blocks and divides the total length of the product data in the product width direction into q blocks as shown in FIG. 2, for example. In the figure, the leftmost block in the product longitudinal direction is the first block, and the rightmost block is the q-th block. Also, in the figure, the uppermost block in the product width direction is the first block, and the lowermost block is the p-th block. The data division unit 31 stores the product data composed of operation data and quality data in the storage unit 20 in association with each block obtained by dividing the product as shown in FIG. 2.

[0039] The data division unit 31 divides the product data of past performance when creating a prediction model (see FIG. 10). Also, the data division unit 31 divides the product data of the diagnosis target when performing quality defect diagnosis (see FIG. 11). Also, it is desirable that the data division methods (number of divisions, area of each block, etc.) be the same when creating a prediction model and when performing quality defect diagnosis.

[0040] The data division unit 31 divides the total length of each product data in the longitudinal direction of the product into p blocks and divides the total length in the width direction of the product into q blocks. The number of divisions of the product data (the values of p and q) is not particularly limited, but it is desirable that it be at least divided into three or more parts (the values of q and p > 3). That is, p is an integer of "p ≧ 3", q is an integer of "q ≧ 3", and the values of p and q can be arbitrarily determined in advance. Also, the area of each block after division may be the same for each product data, or may be different. For example, in the classification of quality defects described later (see Figure 3), for the blocks classified as quality defects continuous in the longitudinal direction of the product or continuous in the width direction of the product, the area (region) may be different for each product data.

[0041] Based on the occurrence tendency of quality defects in the divided blocks, the quality defect classification unit 32 classifies the quality defects included in the quality data into two or more categories. Here, the "occurrence tendency of quality defects" indicates, for example, the occurrence position, occurrence distribution, etc. of quality defects in the divided blocks (see Figure 2).

[0042] Specifically, the quality defect classification unit 32 classifies the quality defects in the quality data into those that occur continuously along the longitudinal direction of the product (continuous in the longitudinal direction of the product) and those that occur continuously across the width direction of the product (continuous in the width direction of the product). For example, as shown in Figure 3, the quality defect classification unit 32 classifies the quality defects in the quality data into either continuous in the longitudinal direction of the product or continuous in the width direction of the product. Also, the quality defect classification unit 32 classifies the quality defects continuous in the longitudinal direction of the product and the quality defects continuous in the width direction of the product respectively based on, for example, the following criteria (1) to (4).

[0043] (1) When there are a plurality of quality defect occurrence blocks in the longitudinal direction of the product at each position along the width direction of the product, and the number of quality defect occurrence blocks is s, if s / p ≧ x, it is classified as a quality defect of "continuous in the longitudinal direction of the product" (x is an arbitrary real number where 0 < x ≦ 1 and is set individually according to the type of quality defect). (2) When there are multiple quality defect occurrence blocks in the longitudinal direction of the product at each position along the width direction of the product, if m or more quality defect occurrence blocks are continuously adjacent to each other, it is classified as a quality defect of "continuous in the longitudinal direction of the product" (m is an arbitrary integer of 2 or more, and is set individually according to the type of quality defect). (3) When there are multiple quality defect occurrence blocks in the width direction of the product at each position along the longitudinal direction of the product, if the number of quality defect occurrence blocks is t and t / q ≥ y, it is classified as a quality defect of "continuous in the width direction of the product" (y is an arbitrary real number where 0 < y ≤ 1, and is set individually according to the type of quality defect). (4) When there are multiple quality defect occurrence blocks in the width direction of the product at each position in the longitudinal direction of the product, if n or more quality defect occurrence blocks are continuously adjacent to each other, it is classified as a quality defect of "continuous in the width direction of the product" (n is an arbitrary integer of 2 or more, and is set individually according to the type of quality defect).

[0044] Figure 4 shows the case where the quality defect classification unit 32 classifies it as a quality defect continuous in the longitudinal direction of the product based on the above criterion (1). For example, when "x = 0.7" in the above criterion (1), "s = 5", and "s / p = 0.71 ≥ x", the quality defect in the figure is classified as "continuous in the longitudinal direction of the product". In the figure, an example of the case where the quality defect occurrence blocks are continuous in the longitudinal direction of the product is shown, but even if the quality defect occurrence blocks are not continuous in the longitudinal direction of the product, it may still meet the above criterion (1). For example, even if the quality defect occurrence blocks occur intermittently in the longitudinal direction of the product, if it meets the above criterion (1), it is classified as "continuous in the longitudinal direction of the product".

[0045] Figure 5 shows the case where the quality defect classification unit 32 classifies it as a quality defect continuous in the longitudinal direction of the product based on the above criterion (2). For example, when "m = 3" in the above criterion (2), since the number of continuously adjacent blocks in the longitudinal direction of the product is "3", the quality defect in the figure is classified as "continuous in the longitudinal direction of the product".

[0046] FIG. 6 shows a case where the quality defect classification unit 32 classifies a quality defect that is continuous in the longitudinal direction of the product based on the above criterion (3). For example, when "y = 0.7" in the above criterion (3), "t = 6", and "t / q = 0.86 ≧ y", the quality defect in this figure is classified as "continuous in the product width direction". In this figure, an example where the quality defect occurrence blocks are continuous in the product width direction is shown. However, even when the quality defect occurrence blocks are not continuous in the product width direction, it may still meet the above criterion (1). For example, even when the quality defect occurrence blocks occur intermittently in the product width direction, if they meet the above criterion (3), they are classified as "continuous in the product width direction".

[0047] FIG. 7 shows a case where the quality defect classification unit 32 classifies a quality defect that is continuous in the longitudinal direction of the product based on the above criterion (4). For example, when "n = 4" in the above criterion (4), since the number of adjacent blocks that are continuous in the product width direction is "5", the quality defect in this figure is classified as "continuous in the product width direction".

[0048] The quality defect classification unit 32 stores information regarding the classification of the quality defects processed as described above in the storage unit 20. Also, when creating a prediction model (see FIG. 10), the quality defect classification unit 32 classifies the quality data of past performance. And when performing quality defect diagnosis (see FIG. 11), it uses the classification of the quality defects at the time of creating this prediction model to aggregate the operation data and equipment maintenance data of the diagnosis target.

[0049] The data aggregation unit 33 aggregates the operation data and equipment maintenance data in block units. The data aggregation unit 33 first extracts the operation data and equipment maintenance data for each classification of quality defects (see FIG. 3) based on the occurrence tendency of quality defects in the divided blocks. The classification of quality defects used here is the classification of quality defects (continuous in the longitudinal direction of the product, continuous in the width direction of the product) classified by the quality defect classification unit 32 at the time of creating the prediction model.

[0050] In this case, based on the quality defect occurrence tendency (see FIGS. 4 and 5) having a plurality of quality defect occurrence locations (quality defect occurrence blocks) over the longitudinal direction of the product, the data aggregation unit 33 extracts operation data and equipment maintenance data for each classification of quality defects. Also, based on the quality defect occurrence tendency (see FIGS. 6 and 7) having a plurality of quality defect occurrence locations over the width direction of the product, the data aggregation unit 33 extracts operation data and equipment maintenance data for each classification of quality defects.

[0051] Subsequently, the data aggregation unit 33 aggregates the extracted operation data and equipment maintenance data for each region including two or more blocks in a specific direction (continuous in the longitudinal direction of the product, continuous in the width direction of the product) of the product. The data aggregation unit 33 aggregates the operation data and equipment maintenance data as follows, for example.

[0052] For example, as shown in FIGS. 4 and 5, for quality defects whose classification is "continuous in the longitudinal direction of the product", the data aggregation unit 33 extracts only the operation data and equipment maintenance data of p blocks in the longitudinal direction of the product at each position along the width direction of the product. Then, the data aggregation unit 33 aggregates the extracted operation data and equipment maintenance data in block units.

[0053] Also, for example, as shown in FIGS. 6 and 7, for quality defects whose classification is "continuous in the width direction of the product", the data aggregation unit 33 extracts only the operation data and equipment maintenance data of q blocks in the width direction of the product at each position along the longitudinal direction of the product. Then, the data aggregation unit 33 aggregates the extracted operation data and equipment maintenance data in block units.

[0054] As a method of aggregation in the data aggregation unit 33, for example, basic statistical quantities representing the values at each position within the block can be adopted. Specifically, these basic statistical quantities are values such as the average value, maximum value, minimum value, standard deviation, variance, and range represented by the difference between the maximum value and the minimum value.

[0055] In this way, the data aggregation unit 33 aggregates the operation data and the equipment maintenance data by calculating at least one of the average value, maximum value, minimum value, standard deviation, variance, and difference between the maximum value and the minimum value of the operation data and the equipment maintenance data in block units. Note that when aggregating the operation data and the equipment maintenance data, at least one type of basic statistic is adopted, but multiple types of basic statistics may also be adopted. The data aggregation unit 33 stores the operation data and the equipment maintenance data aggregated up to the block unit in the storage unit 20.

[0056] When creating a prediction model (see FIG. 10), the data aggregation unit 33 aggregates the operation data and the equipment maintenance data of past performance. Also, when performing quality defect diagnosis (see FIG. 11), the data aggregation unit 33 aggregates the operation data and the equipment maintenance data of the diagnosis target.

[0057] The model creation unit 34 creates a plurality of prediction models according to the classification of quality defects. The model creation unit 34 creates a plurality of prediction models according to the classification of quality defects by learning the relationship between the aggregated operation data and equipment maintenance data and the quality data by machine learning for each classification of quality defects.

[0058] That is, the model creation unit 34 constructs a prediction model showing the relationship between the operation data and the equipment maintenance data aggregated in block units and the quality defects occurring in the corresponding block or product for each classification of quality defects. The model creation unit 34 creates, for example, two types of prediction models for predicting quality defects that are continuous in the longitudinal direction of the product and continuous in the width direction of the product, respectively. For creating these prediction models, various machine learning algorithms represented by, for example, deep neural networks, decision trees, support vector machines, etc., and other statistical algorithms can be used.

[0059] The quality defect determination unit 35 determines the presence or absence of quality defects using a plurality of prediction models created by the model creation unit 34. The quality defect determination unit 35 determines the presence or absence of quality defects for each classification by inputting the operation data and equipment maintenance data of the diagnosis target into the prediction models for each classification respectively. Here, the operation data and equipment maintenance data input into the prediction model are data aggregated in block units after data division in the data division unit 31 and data aggregation in the data aggregation unit 33 (see FIG. 11).

[0060] The determination of quality defects by the quality defect determination unit 35 is performed, for example, before the manufacture of the product or during the manufacture of the product. When it is performed before the manufacture of the product, as the operation data and equipment maintenance data of the product to be predicted, for example, the set values of the manufacturing conditions and equipment conditions before processing are input into the prediction model and calculated, so as to predict the quality defects of the product after manufacture. Also, when it is performed during the manufacture of the product, as the operation data and equipment maintenance data of the product to be predicted, the actual values of the processed manufacturing conditions and equipment conditions are input into the prediction model and calculated, so as to predict the quality defects of the product after manufacture.

[0061] The cause estimation unit 36 estimates the cause of the quality defect based on the determination result of the quality defect by the quality defect determination unit 35. The "cause of the quality defect" indicates, for example, the operation data and equipment maintenance data that are important factors for the quality defect of the product.

[0062] The cause estimation unit 36 can estimate, for example, the operation data and equipment maintenance data necessary to suppress the occurrence of the quality defect when it is determined that a quality defect has occurred in a specific product of the diagnosis target. That is, when it is predicted that a quality defect will occur in a product before or during manufacture in the future, the operation data and equipment maintenance data that need to be changed and adjusted to prevent the occurrence of this quality defect are estimated. Thereby, a countermeasure method for suppressing the occurrence of quality defects can be considered.

[0063] In addition, when a quality defect occurs in a specific product, the cause estimation unit 36 can estimate the operation data and equipment maintenance data that caused the quality defect. That is, when there is a quality defect in the product after manufacturing, the operation data and equipment maintenance data that caused the quality defect are estimated. Thereby, the cause of the quality defect can be identified.

[0064] In addition, when quality defects occur in a large number of products, the cause estimation unit 36 can estimate the operation data and equipment maintenance data that caused the quality defects. That is, when there are quality defects in a large number of products after manufacturing, the operation data and equipment maintenance data that caused the quality defects can be estimated. Thereby, the cause of the quality defect can be identified.

[0065] When it is determined that a quality defect has occurred in a specific product to be diagnosed, and when estimating the operation data and equipment maintenance data necessary to suppress the occurrence of the quality defect, for the data of the product, an index corresponding to the contribution degree of each factor in the above determination is calculated. Then, a factor showing a large value of the index is estimated as an important factor. At that time, a general-purpose prediction importance factor calculation method that does not depend on the creation algorithm of the prediction model, such as the method represented by "SHAP (SHapley Additive exPlanations)", is used.

[0066] In addition, when a quality defect occurs in a specific product, and when estimating the operation data and equipment maintenance data that caused the quality defect, for the data of the product, an index corresponding to the contribution degree of each factor in the above determination is calculated. Then, a factor showing a large value of the index is estimated as an important factor. At that time, a general-purpose prediction importance factor calculation method that does not depend on the creation algorithm of the prediction model, such as the above SHAP, is used.

[0067] In addition, when there are quality defects in a large number of products and the operation data and equipment maintenance data that caused the quality defects are to be estimated, the cause estimation unit 36 calculates an index corresponding to the influence degree of each factor in the entire prediction model. Then, a factor that shows a large value of the index is estimated as an important factor.

[0068] In addition, the cause estimation unit 36 may use a prediction importance factor calculation method specific to the algorithm that created the prediction model, such as the importance based on the Gini impurity in a decision tree algorithm. Alternatively, a general-purpose prediction importance factor calculation method that does not depend on the prediction model creation algorithm, such as "Permutation Importance", may be used.

[0069] The result display unit 37 causes the output unit 40 to display (output) at least one of the determination result in the quality defect determination unit 35 and the cause estimation result in the cause estimation unit 36. In addition, the result display unit 37 preferably causes the output unit 40 to display the data group subjected to the alignment process in the data aggregation unit 33 in a unified manner.

[0070] When the determination result of the quality defect by the quality defect determination unit 35 is output to the output unit 40, for example, on a screen where information about the product to be diagnosed can be viewed, with the longitudinal direction of the product as the horizontal direction and the width direction of the product as the vertical direction, a plan view of the product divided into blocks by quality defect classification is displayed. Then, the quality prediction result for each position of the product is displayed on the product plan view.

[0071] In this case, for example, as shown in FIG. 8, on the product plan view, an "×" mark may be added and displayed at the location (block) with a quality defect. Alternatively, on the product plan view, an "○" mark may be added to the location without a quality defect and an "×" mark may be added to the location with a quality defect and displayed. Alternatively, different colors may be used to display according to the prediction result of the presence or absence of a quality defect. Alternatively, a statement such as "It is predicted that the product to be diagnosed (or a specific block of the product to be diagnosed) has a quality defect" may be output from the output unit 40 only in the case of a prediction result of a quality defect.

[0072] Also, for example, as shown in FIG. 9, one or more operation data and equipment maintenance data that vary in the longitudinal direction of the product may be displayed in a chart form at the lower part (or upper part) of the product plan view. Further, one or more operation data and equipment maintenance data that vary in the width direction of the product may be displayed in a chart form at the left part (or right part) of the product plan view.

[0073] Also, when the quality defect cause estimation result by the cause estimation unit 36 by the quality defect determination unit 35 is output to the output unit 40, for example, on a screen where information on the product to be diagnosed can be viewed, with the longitudinal direction of the product as the horizontal direction and the width direction of the product as the vertical direction, one product plan view divided into block units by quality defect classification is displayed. Then, quality data and quality prediction results for each position of the product are displayed on the product plan view. Alternatively, two product plan views may be displayed, and quality data for each position of the product may be displayed on one of the product plan views, and quality prediction results for each position of the product may be displayed on the other plan view.

[0074] Also, for example, as shown in FIG. 9, the quality defect cause estimation result by the cause estimation unit 36 may be displayed side by side next to the product plan view. In this case, the cause estimation result may be displayed in a list together with all items of each input data and the calculated importance index value, or alternatively, among the items of each input data, a predetermined number of items with larger calculated importance index values may be displayed in order. The display method of the cause estimation result is not limited, including other methods.

[0075] Also, when displaying the quality defect cause estimation result by the cause estimation unit 36, one or more operation data and equipment maintenance data that vary in the longitudinal direction of the product may be displayed in a chart form at the upper part or lower part of the product plan view. The data to be displayed at that time may be selected and displayed up to a certain number in order from those with larger calculated importance index values.

[0076] Also, when displaying the result of cause estimation of quality defects by the cause estimation unit 36, one or more operation data and equipment maintenance data that vary in the product width direction may be displayed in a chart form on the left or right part of the product plan view. At that time, the data to be displayed may be selected and displayed in a certain number in order from those with large values of the calculated importance index.

[0077] The output unit 40 is realized by a display device such as an LCD display or a CRT display. The output unit 40 displays, in the form of characters, graphics, etc., for example, the quality defect determination result by the quality defect determination unit 35, the quality defect cause estimation result by the cause estimation unit 36, etc., based on the display signal input from the result display unit 37.

[0078] (Method for creating prediction model) A method for creating a prediction model used in the quality defect diagnosis method according to the embodiment will be described with reference to FIG. 10. In the method for creating a prediction model, a block division step (step S1), a quality defect classification step (step S2), a data aggregation step (step S3), and a model creation step (step S4) are performed.

[0079] In the block division step, the data division unit 31 divides the product data of past performance two-dimensionally for each predetermined block (step S1). Subsequently, in the quality defect classification step, the quality defect classification unit 32 classifies the quality defects included in the quality data according to the occurrence tendency of quality defects in the divided blocks (step S2).

[0080] Subsequently, in the data aggregation step, the data aggregation unit 33 aggregates the operation data and equipment maintenance data extracted for each classification of quality defects for each area including two or more blocks in a specific direction (product longitudinal direction, product width direction) of the product (step S3). Subsequently, in the model creation step, the model creation unit 34 creates a prediction model by learning the relationship between the aggregated operation data and equipment maintenance data and the quality data for each classification of quality defects by machine learning (step S4).

[0081] (Quality Defect Diagnosis Method) The quality defect diagnosis method according to the embodiment will be described with reference to FIG. 11. In the method for creating a prediction model, a block division step (step S11), a data aggregation step (step S12), a quality defect determination step (step S13), a cause estimation step (step S14), and a result display step (step S15) are performed.

[0082] In the following description, it is assumed that a prediction model created in advance at a different timing is used according to steps S1 to S4 in FIG. 10, but a prediction model may be created at the time of diagnosing a quality defect. That is, steps S1 to S4 in FIG. 10 and steps S11 to S15 in FIG. 11 may be performed at different timings, or steps S11 to S15 in FIG. 11 may be performed following steps S1 to S4 in FIG. 10.

[0083] In the block division step, the data division unit 31 divides the product data to be diagnosed two-dimensionally for each predetermined block (step S11). Subsequently, in the data aggregation step, the data aggregation unit 33 aggregates the operation data and equipment maintenance data extracted for each classification of predetermined quality defects for each region including two or more blocks in the specific direction of the product (product longitudinal direction, product width direction) (step S12). In step S12, for example, the operation data and equipment maintenance data are extracted using the information regarding the classification of quality defects created in step S2 in FIG. 10.

[0084] Subsequently, in the quality defect determination step, the quality defect determination unit 35 determines the presence or absence of the occurrence of a quality defect using a plurality of prediction models (step S13). Subsequently, in the cause estimation step, the cause estimation unit 36 estimates the cause of the quality defect based on the determination result of the quality defect (step S14). Subsequently, in the result display step, the result display unit 37 causes the output unit 40 to display the processing results of the quality defect determination step and the cause estimation step (step S15).

[0085] In the quality defect diagnosis apparatus and the quality defect diagnosis method according to the embodiment described above, the location where the quality defect occurs in the product is classified, and the quality defect is diagnosed using a prediction model modeled including the classification result. That is, in the quality defect diagnosis apparatus and the quality defect diagnosis method according to the embodiment, for each classification according to the occurrence tendency of the quality defect, the relationship between the operation data and the equipment maintenance data and the quality defect is created and used as a prediction model. Thereby, the quality defect of the product can be predicted with high accuracy, and the cause of the quality defect can be estimated.

[0086] As described above, the quality defect diagnosis apparatus and the quality defect diagnosis method according to the present invention have been specifically described by the embodiments and examples for carrying out the invention. However, the gist of the present invention is not limited to these descriptions, and should be broadly interpreted based on the description of the claims. Needless to say, various changes and modifications based on these descriptions are also included in the gist of the present invention.

[0087] For example, in the above-described embodiment, the quality defect is diagnosed using a plurality of prediction models that have learned the relationship between the operation data and the equipment maintenance data and the quality data. However, the diagnosis may be performed using a plurality of prediction models that have learned the relationship between the equipment maintenance data and the quality data.

[0088] In this case, in the model creation unit 34, for each classification of the quality defect, a plurality of prediction models corresponding to the classification of the quality defect are created by learning the relationship between the aggregated equipment maintenance data and the quality data by machine learning. Further, in the quality defect determination unit 35, the presence or absence of the occurrence of the quality defect is determined using a plurality of prediction models created for each classification of the quality defect and in which the relationship between the aggregated equipment maintenance data and the quality data has been learned. Further, in the cause estimation unit 36, based on the determination result of the quality defect, the cause of the quality defect is estimated. Thus, even when only the equipment maintenance data and the quality data are used as the input data of the prediction model, the quality defect of the product can be predicted with high accuracy, and the cause of the quality defect can be estimated.

Explanation of Signs

[0089] 1 Information processing device 10 Input unit 20 Memory unit 21 Product DB 30 Arithmetic unit 31 Data division unit 32 Defective quality classification unit 33 Data aggregation unit 34 Model creation unit 35 Defective quality determination unit 36 Cause estimation unit 37 Result display unit 40 Output unit

Claims

1. A data splitting unit that splits product data to be diagnosed, which consists of operation data, equipment maintenance data, and quality data, into a two-dimensional form for each predetermined block, Based on the occurrence tendency of quality defects in the divided blocks, the operation data and the equipment maintenance data are extracted for each predetermined classification of quality defects, and the extracted operation data and equipment maintenance data are aggregated for each area including two or more blocks in a specific direction of the product. A data aggregation unit, A quality defect determination unit that determines the presence or absence of the occurrence of quality defects using a plurality of prediction models created for each classification of the quality defects and in which the relationship between the aggregated operation data and equipment maintenance data and the quality data has been learned. A quality defect diagnosis device comprising the above.

2. The quality defect diagnosis device according to claim 1, further comprising a factor estimation unit that estimates the factors of quality defects based on the determination result of quality defects.

3. The quality defect diagnosis device according to claim 2, further comprising a result display unit that causes at least one of the determination result in the quality defect determination unit and the factor estimation result in the factor estimation unit to be displayed on an output unit.

4. The prediction model is The past performance product data consisting of the operation data, the equipment maintenance data, and the quality data is split into a two-dimensional form for each predetermined block. Based on the occurrence tendency of the quality defects in the divided blocks, the quality defects included in the quality data are classified. The operation data and the equipment maintenance data extracted for each classification of the quality defects are aggregated for each area including two or more blocks in a specific direction of the product. For each classification of the quality defects, it is created by learning the relationship between the aggregated operation data and equipment maintenance data and the quality data by machine learning. The quality defect diagnosis device according to claim 1.

5. The quality defect diagnosis device according to claim 1, wherein the operation data is data that varies over the length in a specific direction of the product.

6. The quality defect diagnosis device according to claim 1, wherein the equipment maintenance data is data that varies during the period when the product is processed in the equipment.

7. The quality defect diagnosis device according to claim 1, wherein the quality data is data that varies over the length in a specific direction of the product.

8. The quality defect diagnosis device according to claim 2, wherein the equipment maintenance data is data indicating the state of the equipment and its maintenance state that affects a specific position of the product by a direct or indirect action on the product.

9. The data aggregation unit extracts the operation data and the equipment maintenance data for each predetermined classification of quality defects based on a quality defect occurrence trend having a plurality of quality defect occurrence locations over the longitudinal direction of the product, or a quality defect occurrence trend having a plurality of quality defect occurrence locations over the width direction of the product. The quality defect diagnosis device according to claim 1.

10. The quality defect determination unit determines the presence or absence of the occurrence of a quality defect using a plurality of prediction models created for each classification of the quality defect and in which the relationship between the aggregated equipment maintenance data and the quality data has been learned. The cause estimation unit estimates the cause of the quality defect based on the determination result of the quality defect. The quality defect diagnosis device according to claim 8.

11. The result display unit causes the output unit to display the data group subjected to the alignment process by the data aggregation unit. The quality defect diagnosis device according to claim 3.

12. A data division step in which data division means provided in a computer divides product data of a product to be diagnosed, which consists of operation data, equipment maintenance data, and quality data, into a two-dimensional shape for each predetermined block. A data aggregation step in which data aggregation provided in the computer extracts the operation data and the equipment maintenance data for each predetermined classification of quality defects based on the quality defect occurrence trend in the divided blocks, and aggregates the extracted operation data and equipment maintenance data for each region including two or more blocks in a specific direction of the product. A quality defect determination step in which quality defect determination means provided in the computer determines the presence or absence of the occurrence of a quality defect using a plurality of prediction models created for each classification of the quality defect and in which the relationship between the aggregated operation data and equipment maintenance data and the quality data has been learned. A quality defect diagnosis method including the above steps.

13. The quality defect diagnosis method according to claim 12, including a cause estimation step in which cause estimation means provided in the computer estimates the cause of the quality defect based on the determination result of the quality defect.

Images