Method for pre-detecting quality risk of design conditions

Abstract

The present disclosure relates to a quality risk detection technology and provides an apparatus and method for detecting quality risk, wherein the method includes: acquiring, from a database, quality performance information of products manufactured on the basis of design conditions that have a predetermined level or higher level of similarity to the target design conditions; depending on the result of determining whether the number of the acquired quality performance information items is at least N (where N is an integer equal to or greater than 1), calculating either a first quality risk indicator or a second quality risk indicator via an artificial intelligence model; and outputting the result of quality risk detection for the target design conditions on the basis of either the first quality risk indicator or the second quality risk indicator thus calculated.

Description

Method for preemptively detecting design condition quality risks

[0001] The present disclosure relates to a technology for detecting quality risks of design conditions.

[0002] The manufacturing industry follows a customer-order-driven production model. Therefore, design conditions for potential customer requirements are pre-registered. When an order is received, the product is manufactured by matching the design conditions to those requirements.

[0003] To minimize design errors and product defects, it's necessary to verify that pre-registered design conditions are risk-free. However, manually collecting and analyzing data based on numerous design conditions to determine whether they contain any issues is time-consuming. Furthermore, for new design conditions with no production history, the lack of existing data makes it difficult to assess risk.

[0004] Therefore, the need for a method to assess and manage the risks of pre-registered design conditions is being raised.

[0005] The present disclosure aims to provide a technology for detecting quality risks of design conditions.

[0006] In one aspect, the present embodiments provide a quality risk detection method including a step of obtaining quality performance information of a product produced based on design conditions having a similarity level or higher with the design conditions from a first database, a step of calculating a first quality risk index or a second quality risk index through an artificial intelligence model based on a result of determining whether the number of obtained quality performance information is N or more (N is an integer greater than or equal to 1), and a step of outputting a quality risk detection result of the design conditions based on either the calculated first quality risk index or the second quality risk index.

[0007] In another aspect, the present embodiments provide a quality risk detection device for a design condition quality risk detection device, comprising a transmitting and receiving unit for quality performance information of a product manufactured based on design conditions having a similarity level or higher with the design conditions from a first database, and a calculating unit for calculating a first quality risk index or a second quality risk index through an artificial intelligence model based on a judgment result of determining whether the number of pieces of quality performance information obtained is N or more (N is an integer greater than or equal to 1), and outputting a quality risk detection result of the design condition based on either the calculated first quality risk index or the second quality risk index.

[0008] The present disclosure can provide a technology for detecting quality risks of design conditions.

[0009] Figure 1 is a drawing schematically illustrating a process for deriving a design condition quality risk detection result according to one embodiment.

[0010] Figure 2 is a drawing schematically illustrating the entire process of detecting design condition quality risks according to one embodiment.

[0011] FIG. 3 is a drawing specifically explaining a process for deriving a design condition quality risk detection result according to one embodiment.

[0012] FIG. 4 is a diagram schematically illustrating a process for organizing the derived quality risk detection results according to one embodiment.

[0013] FIG. 5 is a diagram illustrating a process of clustering pre-registered design conditions according to one embodiment.

[0014] Figure 6 is a diagram for explaining a process of organizing a quality risk detection target according to one embodiment.

[0015] Figure 7 is a drawing for explaining order condition similarity criteria according to one embodiment.

[0016] Figure 8 is a drawing for explaining similar criteria for required quality assurance conditions according to one embodiment.

[0017] Figure 9 is a drawing for explaining manufacturing condition similarity standards according to one embodiment.

[0018] Figure 10 is a drawing for explaining the organization of quality unit review targets according to one embodiment.

[0019] Figure 11 is a drawing for explaining the review results of a quality unit review target according to one embodiment.

[0020] FIG. 12 is a drawing for explaining the review results of a similar order unit review target according to one embodiment.

[0021] Figure 13 is a drawing for explaining the review results of the design condition unit review target according to one embodiment.

[0022] FIG. 14 is a drawing for explaining a design condition quality risk pre-detection device according to one embodiment.

[0023] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. When adding reference numerals to components in each drawing, identical components may have the same numerals as much as possible even if they are shown in different drawings. In addition, when describing the present embodiments, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the technical idea of ​​the present invention, the detailed description may be omitted. When "includes," "has," "consists of," etc. are used in this specification, other parts may be added unless "only" is used. When a component is expressed in the singular, it may include a case in which the plural is included unless specifically stated otherwise.

[0024] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the present disclosure. These terms are only intended to distinguish the components from other components, and the nature, order, sequence, or number of the components are not limited by the terms.

[0025] In a description of the positional relationship of components, when it is described that two or more components are "connected," "combined," or "connected," it should be understood that the two or more components may be directly "connected," "combined," or "connected," but that the two or more components may also be further "interposed" with another component to be "connected," "combined," or "connected." Here, the other component may be included in one or more of the two or more components that are "connected," "combined," or "connected" to each other.

[0026] In the description of the temporal flow relationship related to components, operation methods, or manufacturing methods, for example, when the temporal or flow relationship is described as “after”, “following”, “next to”, “before”, etc., it may also include cases where it is not continuous, unless “immediately” or “directly” is used.

[0027] Meanwhile, when numerical values ​​or corresponding information (e.g., levels, etc.) for components are mentioned, even without separate explicit description, the numerical values ​​or corresponding information may be interpreted as including an error range that may occur due to various factors (e.g., process factors, internal or external impact, noise, etc.).

[0028] The embodiments are described in detail with reference to the drawings below.

[0029] While the manufacturing industry may prioritize production and sales of products to customers, many companies follow a method of manufacturing products upon confirmation of customer orders. Furthermore, these products are manufactured based on design conditions that vary depending on customer requirements and the circumstances of the production plant.

[0030] Therefore, manufacturers pre-register design conditions based on customer requirements. When an order is placed, they match the design conditions to the customer's requirements and produce the product. However, in such cases, there may be cases where pre-registered design conditions that meet the customer's requirements are not available, requiring the registration of new design conditions.

[0031] In this regard, there is a need to verify in real time or in advance whether pre-registered design conditions are sufficient for producing a normal product or whether newly registered design conditions are suitable. However, manually checking for errors across numerous different design conditions is time-consuming.

[0032] This disclosure proposes a method for identifying quality risks in pre-registered design conditions and newly registered design conditions as a means of minimizing the error rate of design conditions and the defect rate of manufactured products.

[0033] Figure 1 is a flowchart schematically illustrating a process for deriving design condition quality risk detection results according to one embodiment.

[0034] Referring to Fig. 1, a method for detecting quality risks of design conditions may include an acquisition step for acquiring quality performance information (S100).

[0035] For example, the acquisition step can acquire quality performance information on products manufactured based on design conditions that have a certain level of similarity to the design conditions from a database. The quality of products manufactured based on previously registered design conditions can be measured, and the suitability of the design conditions can be determined based on the measured numerical information. The present disclosure can determine whether the design conditions are suitable by quantifying the quality of products manufactured under similar design conditions to the design conditions to be detected.

[0036] A method for detecting quality risks of design conditions may include a calculation step for calculating quality risk indicators (S110).

[0037] For example, the production step can produce the first quality risk indicator or the second quality risk indicator through an artificial intelligence model based on the judgment result of whether the number of quality performance information acquired in the acquisition step is N or more (N is an integer greater than or equal to 1). The number of quality performance information that serves as the basis for the method of producing the quality risk indicator can be conveniently set by the user, and since the number of quality performance information acquired from the database is N or more, it means that past quality performance data is trustworthy, and the first quality risk indicator is produced using quality performance data retrieved from the database, and if there is no data retrieved or the number is small, less than N, the second quality risk indicator can be produced based on data output by a learned and trustworthy artificial intelligence model.

[0038] For example, the first quality risk indicator produced in the production stage may include a first process capability index and a first defect rate, and the second quality risk indicator may include a second process capability index and a second defect rate. The process capability index refers to a numerical value for evaluating the degree of process capability. Or, it refers to a numerical value for determining whether the current production process is suitable by comparing the level required to improve the production process with the actual production result. The defect rate refers to a numerical value expressing the ratio of the number of products that do not meet the standards set by the user among the products produced under specific design conditions as a percentage. The present disclosure can determine whether the design conditions to be detected are good or bad by calculating the process capability index and defect rate using information acquired from a database in which a conventional production history is stored or information output through an artificial intelligence model.

[0039] For example, the first process capability index and first defect rate calculated in the production stage can be calculated based on the average and deviation of quality values ​​included in the quality performance information. Furthermore, the second process capability index and second defect rate can also be calculated based on the average and deviation of values ​​output by the artificial intelligence model. Alternatively, the process capability index or defect rate can be calculated from the artificial intelligence model and only the results can be output.

[0040] For example, in the production step, if the number of quality performance information acquired from the database in the acquisition step is N or more, the first quality risk indicator can be calculated based on the quality performance information acquired from the database, and if the number of quality performance information is less than N, the second quality risk indicator can be calculated based on information output through the artificial intelligence model.

[0041] For example, the artificial intelligence model used to produce the second quality risk indicator in the production stage may include a Bayesian neural network model.

[0042] A method for detecting quality risks of design conditions may include an output step for outputting quality risk detection results (S120).

[0043] For example, the output step can output the quality risk detection result of the design condition based on either the first quality risk indicator or the second quality risk indicator produced in the above production step.

[0044] For example, the output stage may output the quality risk detection result as good performance based on the fact that the first process capability index calculated in the output stage is higher than the first reference value and the first defective rate is lower than the second reference value; the quality risk detection result may output the performance risk based on the fact that the first process capability index is lower than the first reference value or the first defective rate is higher than the second reference value; the quality risk detection result may output the good prediction based on the fact that the second process capability index is higher than the third reference value and the second defective rate is lower than the fourth reference value; and the quality risk detection result may output the risk prediction based on the fact that the second process capability index is lower than the third reference value or the second defective rate is higher than the fourth reference value. Outputting as good performance or performance risk is determined based on the first quality risk indicator, meaning that the number of quality performance information retrieved from the database is N or more, which is set by the user, and outputting as expected good or expected risk is determined based on the second quality risk indicator, meaning that the number of quality performance information retrieved is less than N.

[0045] By varying the method of calculating quality risk indicators according to the number of quality performance information retrieved, the reliability issue of calculation results that occurs when there are insufficient samples can be resolved, and the design error and product defect rate can be reduced by accurately determining whether the design conditions to be detected are appropriate.

[0046] Below, we provide a more detailed and diverse explanation with reference to a drawing showing the entire process of detecting design condition quality risks.

[0047] Figure 2 is a drawing schematically illustrating the entire process of detecting design condition quality risks according to one embodiment.

[0048] Referring to Fig. 2, the quality risk of design conditions can be detected in more detail by adding an extraction step and a composition step to the acquisition step, output step, and output step described above through Fig. 1.

[0049] A method for detecting quality risks in design conditions may include an extraction step (S200) of extracting similar design conditions based on criteria set by the user from a database storing past design history. The step of extracting similar design conditions may extract similar design conditions based on criteria set by the user from the database storing past design history. Furthermore, the extraction of the similar design conditions may be categorized into design conditions, such as order conditions, quality assurance conditions, and manufacturing conditions.

[0050] For example, extraction of similar design conditions based on order conditions can be based on criteria such as type, product specification, thickness, width, and order purpose. Extraction of similar design conditions based on quality assurance conditions can be based on criteria such as component conditions such as the customer's requested steel composition and product composition, material quality such as mechanical properties, and appearance quality such as shape, surface, and dimensions. In addition, each quality can include information on specimen type, test conditions, and warranty scope. Extraction of similar design conditions based on manufacturing conditions can be based on information on key design factors being designed in the minor firing, plant, tapping composition, steelmaking, continuous casting, and rolling processes.

[0051] Each design condition extracted according to similar criteria can be stored in a separate database.

[0052] A method for detecting quality risks of design conditions may include a composition step of organizing review items to be detected (S210).

[0053] For example, the review target item may be determined based on at least one of the similar design conditions extracted in the extraction step and the design conditions to be detected. The review target item refers to a criterion for determining the validity of a design condition among multiple items set as design conditions. The review target item may be organized based on the similar design conditions extracted in the similar design condition extraction step or based on the actual design condition to be detected. Once the review target item is determined, related information about the review target item may be stored in a separate database.

[0054] For example, the review target items may include at least one of a quality unit item, an order unit item, and a design unit item. Based on the determined review target items, a quality risk detection result including at least one of a first quality risk detection result for the quality unit item, a second quality risk detection result for the order unit item, and a third quality risk detection result for the design unit item may be output.

[0055] In order to determine suitability by examining design conditions to be detected based on various criteria, the present disclosure determines the detection result as an item including at least one of a quality unit item, an order unit item, and a design unit item, and if at least one of the items subject to review is determined to be dangerous, the design condition to be detected can be output as unsuitable.

[0056] For example, the quality risk detection result may include a fourth quality risk detection result related to the suitability of the design conditions to the design conditions, and the fourth quality risk detection result may include any one of performance good, performance risk, good expected, risk expected, or indeterminate. For example, based on whether at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result is performance risk or risk expected, the fourth quality risk result may be output as either the performance risk or the risk expected.

[0057] The present disclosure has the effect of increasing the accuracy of determining the suitability of design conditions in that, in detecting quality risks of design conditions, the suitability of design conditions can be determined not only by a single criterion but also by two or more criteria.

[0058] In Figure 3 and below, each process for detecting design condition quality risks is explained in more detail and in various ways with reference to the drawings.

[0059] FIG. 3 is a drawing specifically explaining a process for deriving a design condition quality risk detection result according to one embodiment.

[0060] The result of detecting quality risks of design conditions is determined based on the calculated quality risk indicators, and the result of detecting quality risks can be one of the following: good performance, risky performance, expected good performance, expected risk, or indeterminate.

[0061] The process of deriving the results of detecting design condition quality risks is performed as follows.

[0062] 1) Search for quality review targets (S300)

[0063] 2) Search for quality performance data and determine the number of data obtained (S310)(S320)

[0064] 3) Calculate the average and standard deviation of quality performance (S330)

[0065] 4) Calculate quality risk indicators based on the number of quality performance data acquired (S340)

[0066] 5) Query the criteria for judging the quality risk detection results (S350)

[0067] 6) Output quality risk detection results based on the generated quality risk indicators and judgment criteria (S360).

[0068] According to the present disclosure, the steps for outputting the detection results of design condition quality risks can be summarized as a data acquisition step, an indicator calculation step, and a result output step.

[0069] Specifically, design conditions for detecting quality risks can be retrieved (S300). These design conditions may also be referred to as "quality risk review targets." Furthermore, detection can be performed only on design conditions for which quality risk detection has never been performed and therefore review results do not exist in the database.

[0070] After selecting a review target by searching the design conditions for detecting quality risks, the presence of a past product production history based on the design conditions is checked (S310). According to the present disclosure, the presence of a past product production history can be checked from a database storing product production history based on the design conditions. The database storing the product production history can also be referred to as an operational quality performance database.

[0071] Product production history can be checked not only based on identical design conditions, but also on similar design conditions to the target design condition. Similar design conditions can be selected based on user-defined criteria, or through an AI model or algorithm that calculates the similarity between design conditions. Furthermore, users can check for past product production history by specifying a specific time period.

[0072] When quality performance data regarding design conditions to be detected is retrieved from a database storing past product production history, the number of retrieved quality performance data is confirmed (S320).

[0073] The reason for checking the number of quality performance data is that quality risk assessments rely on statistics such as averages and standard deviations, which poses reliability issues. If the number of data is large, quality risk assessments are based on the average and standard deviation of the retrieved quality performance data. If the number of data is small, quality risk assessments are based on data output from the AI ​​model. The user can set the criteria for determining whether the number of quality performance data is large or small, setting the threshold to N (where N is an integer greater than or equal to 1).

[0074] Once the number of quality performance data items is verified, a quality risk indicator is calculated based on the number of quality performance data items (S330). If the number of quality performance data items retrieved is N or more, the average and deviation of the quality performance data stored in the database are calculated. If the number of quality performance data items retrieved is less than N, the average and deviation of quality data are calculated using a quality prediction AI model.

[0075] For example, if the number of quality performance data retrieved is 10 and the criteria set by the user is 30, quality risk assessment cannot be performed with only the 10 retrieved quality performance data. Therefore, the 10 retrieved quality performance data can be input into a quality prediction artificial intelligence model to additionally output 20 predicted data, and a quality risk index can be calculated using the 10 actually retrieved quality performance data and the 20 predicted data output by the artificial intelligence model. Alternatively, the quality prediction artificial intelligence model with the 10 retrieved quality performance data as input data can output 30 predicted data, and a quality risk index can be calculated based on the output predicted data.

[0076] For convenience of explanation, FIG. 3 shows 30 criteria set by the user, but the present disclosure is not limited thereto, and the user may set a variety of criteria for quality performance data using a quality prediction artificial intelligence model as needed.

[0077] Once the quality average and deviation are calculated using quality performance data stored in the database or information output by an artificial intelligence model, the process capability index and defect rate are calculated based on the average and deviation information (S340). As described above, the process capability index is a numerical value used to determine the suitability of the current production process by comparing the level required to improve the production process with the actual production results. The defect rate is a numerical value representing the number of products that fail to meet the user-defined standards among products produced under specific design conditions, expressed as a percentage.

[0078] Once the process capability index and defect rate are calculated, the customer's required quality criteria are checked (S350). The required quality includes any of the following: quality unit, order unit, or design unit, and represents the standard value for determining good or bad. This can be stored in the database in advance by the user and can be modified as needed.

[0079] When the judgment criteria for the required quality are obtained, the quality risk detection result is judged and output. The quality risk detection result is judged based on whether the process capability index and the defect rate are above or below the standard. According to the present disclosure, when the number of quality performance data is N or more and is calculated based on the values ​​stored in the database, it can be called a first process capability index and a first defect rate, a first quality risk index, and a first quality risk detection result, and when it is calculated through an artificial intelligence model and the detection result is output, it can be called a second process capability index and a second defect rate, a second quality risk index, and a second quality risk detection result.

[0080] Specifically, if the first process capability index is higher than the first reference value and the first defective rate is lower than the second reference value, the quality risk detection result may be output as good performance, and if the first process capability index is lower than the first reference value or the first defective rate is higher than the second reference value, the quality risk detection result may be output as performance risk.

[0081] If the second process capability index calculated through the artificial intelligence model is higher than the third reference value and the second defective rate is lower than the fourth reference value, the quality risk detection result may be output as a good prediction, and if the second process capability index is lower than the third reference value or the second defective rate is higher than the fourth reference value, the quality risk detection result may be output as a risk prediction.

[0082] The terms for each of the above criteria and quality risk detection results are not limited to this, and users may set them differently as needed.

[0083] If there is no separately designated quality prediction artificial intelligence model even though the number of quality performance data being searched is less than N, it may be output as impossible to judge.

[0084] FIG. 4 is a diagram schematically illustrating a process for organizing the derived quality risk detection results according to one embodiment.

[0085] When the quality risk judgment results are derived according to Figure 3, the judgment results for each judgment item can be stored in a separate database.

[0086] For example, when quality risk detection results for a quality unit item are derived, user-specified items can be stored in a database. Furthermore, rather than storing the detection results directly in the database, they can be classified into items related to ingredient quality, material quality, and appearance quality and stored in separate databases. According to the present disclosure, these databases can be referred to as the ingredient quality review results database, material quality review results database, and appearance quality review results database, respectively.

[0087] Once quality risk detection results for order unit items are derived, user-specified items can be stored in a database. Furthermore, data from the database, which stores the detection results by categorizing them into the aforementioned ingredient quality, material quality, and appearance quality categories, and the derived quality risk detection results for order unit items can be aggregated and stored in a separate database based on user-defined criteria. According to the present disclosure, this database may be referred to as a "Similar Order Review Results Database."

[0088] Once quality risk detection results for design unit items are derived, user-specified items can be stored in a database. Furthermore, data stored in the similar order review results database and the quality risk detection results for the derived design unit items can be aggregated according to user-defined criteria and stored in a separate database. According to the present disclosure, this database can be referred to as the "Design Condition Review Results Database."

[0089] FIG. 5 is a diagram illustrating a process of clustering pre-registered design conditions according to one embodiment.

[0090] While detection results can be derived by querying a database to determine whether there is a history of products manufactured under the same design conditions as the design conditions to be detected, the number of design conditions can vary widely, sometimes in the tens or hundreds of thousands, and thus the number of data retrieved may be insufficient to derive detection results. Furthermore, design conditions for previously unproduced products are novel, and if the product production history is not retrieved, it may be difficult to determine whether the design conditions to be detected are appropriate. Accordingly, the present disclosure establishes design conditions similar to the design conditions to be detected in advance and retrieves quality history data based on design conditions within a similar range, thereby enabling more accurate derivation of quality risks of design conditions.

[0091] Specifically, design conditions can be categorized into order conditions, quality assurance conditions, and manufacturing conditions, and based on each design condition, a database storing the history of product production can be used to classify order conditions, quality assurance conditions, and manufacturing conditions by similar conditions. According to the present disclosure, the database storing the history of product production can be referred to as an order design result history information database.

[0092] For example, similar order conditions can be extracted from the order design result history information database (S500). The criteria for similar order conditions can be set by the user as needed. Examples of order conditions may include product type, size, order thickness, order width, and order purpose.

[0093] Categorical items, such as variety, size, and order purpose, can be defined as lists of similar user-defined values, while continuous items, such as order thickness and order width, can be defined as ranges of similar user-defined values. Once extraction is complete, this information can be stored in a separate database. According to the present disclosure, this database can be referred to as a "similar order condition database."

[0094] As another example, similar quality assurance conditions can be extracted from a database of order design result history information (S501). The criteria for similar quality assurance conditions can be set by the user as needed. For example, quality assurance conditions can be classified into composition conditions, material quality, and appearance quality. Composition conditions can be extracted based on criteria that have the same warranty scope for the steel composition and product composition for each alloy composition. Material quality conditions can be extracted based on criteria that have the same specimen type, test conditions, and warranty scope for each mechanical property. Appearance quality conditions can be extracted based on criteria that have the same warranty scope for the same shape, surface, and dimension. Once extraction is complete, this information can be stored in a separate database. According to the present disclosure, this database can be referred to as a similar warranty condition database.

[0095] As another example, similar manufacturing conditions can be extracted from a database of order design result history information (S501). To satisfy customer order conditions and quality assurance conditions, key design factors for process conditions, such as tapping components and steelmaking, continuous casting, and rolling processes, designed to be similar, can be extracted and managed based on similar clustering criteria. The criteria for conditions with similar manufacturing conditions can be set by the user as needed. For example, similar tapping components can be extracted based on tapping target numbers with the same deoxidation method, dephosphorization method, and alloy target components. Process conditions can be extracted based on the same criteria for steelmaking (steelmaking plant, secondary refining method, etc.), continuous casting (light pressing conditions, scarfing conditions, etc.), heating furnace (extraction temperature, remelting time), rolling (start temperature, end temperature, reduction amount), and cooling (start temperature, end temperature, speed). Once extraction is complete, this information can be stored in a separate database. According to the present disclosure, this database can be referred to as a similar manufacturing condition database.

[0096] Figure 6 is a diagram for explaining a process of organizing a quality risk review target according to one embodiment.

[0097] According to the present disclosure, when detecting quality risks of design conditions, rather than organizing and reviewing only a single item, the reliability of the quality risk detection results can be increased by organizing two or more items and reviewing whether the design conditions are satisfactory or dangerous.

[0098] For example, review targets for quality risk detection may include quality unit review targets, order unit review targets, and design unit review targets.

[0099] Specifically, quality unit review targets can be organized into a quality review target table in the quality risk analysis database by linking identical ordering conditions, quality assurance conditions, and manufacturing conditions for required quality units such as components, materials, and appearance quality. The criteria for organizing quality unit review targets can be organized based on key design factors that impact quality. Quality unit review targets can be further categorized into component, material, and appearance review targets.

[0100] For example, among material qualities, tensile strength can be organized by similar order conditions such as type, order thickness range, and width range, and quality assurance conditions can be organized by length direction of specimen type, specimen direction, and specimen size, and test conditions can be organized by YP code and warranty range. Manufacturing conditions can be organized by minor firing, factory, tapping composition, steelmaking, continuous casting, and rolling conditions.

[0101] Organizing review targets by order allows for the organization of review targets by similar orders within the order list, each with the same design requirements, similar order and manufacturing conditions, and identical quality requirements. Organizing review targets by order allows for the proactive detection of risky orders with quality risks and avoids the need for duplicate quality risk assessments when the same order occurs.

[0102] The design unit review target can be organized into design condition units managed by the design department, including similar order conditions, required quality assurance conditions, and manufacturing conditions. Organizing design unit review targets can proactively detect hazardous design conditions that pose quality risks.

[0103] When the above quality unit review target, order unit review target, and design unit review target are organized, the organized data can be stored in the required quality review target database, similar order review target database, and design condition review target database, respectively.

[0104] Figure 7 is a drawing for explaining order condition similarity criteria according to one embodiment.

[0105] Referring to Figures 5 and 7, in the process of clustering pre-registered design conditions, the design conditions can be classified into order conditions, quality conditions, and manufacturing conditions, and Figure 7 is a diagram illustrating an example of clustered order conditions. As described above, order conditions can include the type, product specifications, thickness and width of the ordered product, order purpose, order size, etc., and the order conditions are not limited to these, and if there are pre-set conditions, they can be additionally included in the order conditions. In addition, the similarity criteria of the conditions can be set according to the range set by the user.

[0106] For example, categorical items, such as "variety," can be defined as a user-defined list of similar values. For example, one of the order conditions might include items like "hot-rolled," "plate," and "wire rod," and among them, "hot-rolled" might be classified as "FH" and "FD." Continuous items, such as "order thickness," can also be defined as a user-defined range of similar values. Another example: for hot-rolled products, thicknesses between 0 and 1.5 could be classified as a similar group, and so could thicknesses between 1.5 and 6.

[0107] Therefore, two order conditions, one of which is a hot-rolled variety with an FH code and an order thickness of 1.7 and the other of which is a hot-rolled variety with an FH code and an order thickness of 5.5, can be classified as similar order conditions and stored and managed in a separate database.

[0108] Figure 8 is a drawing for explaining similar criteria for required quality assurance conditions according to one embodiment.

[0109] Referring to Figure 8, the quality conditions desired by customers can be classified into character quality, material quality, and appearance quality. Component quality can be classified into steel components and product components. Material quality can be classified into material properties, material interior, material surface, material magnetism, and material coating. Appearance quality can be classified into exterior surface, exterior shape, and exterior dimensions. Steel components can be classified into terminal components, composite components, etc.

[0110] Additionally, other conditions requested by customers regarding quality of character, material quality, and appearance quality can be classified and managed.

[0111] For example, among the ingredient quality conditions, single components of the weak components that have SI as a condition can be classified as similar conditions and stored in a separate database for management.

[0112] In more detail, among the ingredient quality conditions, the single component of the low-strength component is MN, the complex component is PCM, the single component of the product component is SI, and the complex component is CEQ. These can be classified into similar conditions and stored in a separate database for management.

[0113] The classification items and classification conditions for the above quality conditions are not limited to this, and users can set various conditions according to their needs.

[0114] Figure 9 is a drawing for explaining manufacturing condition similarity standards according to one embodiment.

[0115] Referring to Figure 9, manufacturing conditions can be classified into manufacturing plant, steel component, and process conditions. Manufacturing plant can be classified into small firing and factory, steel component can be classified into deoxidation method, dephosphorization method, and target component, and process conditions can be classified into steelmaking, continuous casting, and rolling.

[0116] As with order conditions and quality conditions, non-continuous items can be classified according to categories set by the user, and continuous items can be classified according to ranges set by the user, and based on the classification, the conditions stored in the past production history information table can be classified in various ways and stored in the database.

[0117] For example, design conditions having factory conditions where the minor component is steel mill 1 and the factory is factory 1 can be classified as similar manufacturing conditions, and cases where the target component of the steel composition is C and the extraction temperature of the rolling process condition is between 100 and 200 degrees can be classified as similar manufacturing conditions.

[0118] Figure 10 is a drawing for explaining the organization of quality unit review targets according to one embodiment.

[0119] As described above, the quality unit review target can be organized into the quality review target table of the quality risk analysis database by linking the same order conditions, quality assurance conditions, and manufacturing conditions as the required quality unit.

[0120] Accordingly, each detail can be stored in a database with the quality ID and SEQ ID as the key of the database, and data on each item subject to quality review, and each item regarding order conditions, warranty conditions, and manufacturing conditions can be aggregated and used for retrieval when necessary.

[0121] Figure 11 is a drawing for explaining the review results of a quality unit review target according to one embodiment.

[0122] Referring to Figure 11, this is a diagram showing an example of the results of calculating each quality risk indicator for a quality unit review target and deriving a quality risk detection result accordingly.

[0123] For example, for a design condition where the quality ID is AB1 and the SEQ number is 01, if the number of retrieved quality performance information is determined to be N or more, a performance-based judgment index including a process capability index and a defect rate can be calculated, and a performance good or performance risk can be derived based on the judgment result.

[0124] The process capability index is calculated based on average and deviation information, based on user-defined criteria. The defect rate can be expressed as a percentage of the total number of data points and the number of defective items. If the process capability index is above the first reference value and the defect rate is below the second reference value, the quality risk detection result can be output as "Good," and thus the overall judgment result can be output as "Good Performance."

[0125] As another example, for a design condition with a quality ID of AB2 and a SEQ number of 02, if the retrieved quality performance information is determined to be less than N, additional prediction data may be output based on an artificial intelligence model, a judgment index may be calculated based on the output prediction data, and a judgment result may be derived. Alternatively, a judgment index may be calculated based on the retrieved quality performance information and the output prediction data, and a judgment result may be derived. In this case, the performance-based judgment result may be output as "unable to judge," and the prediction-based judgment result may be output as "risk prediction" or "good prediction," and the comprehensive judgment result may be output as "risk prediction" or "good prediction" depending on the prediction-based judgment result.

[0126] As another example, for a design condition with a quality ID of AB3 and a SEQ number of 01, if the number of retrieved quality performance information is determined to be less than N, the process capability index and defect rate must be calculated based on an artificial intelligence model. However, if there is no artificial intelligence model to be applied, the performance-based judgment result and the prediction-based judgment result may each be output as impossible to judge, and the comprehensive judgment result may also be output as impossible to judge.

[0127] FIG. 12 is a drawing for explaining the review results of a similar order unit review target according to one embodiment.

[0128] Referring to Figure 12, an example of the results of a quality risk review regarding quality assurance conditions and a comprehensive judgment result is shown based on each condition regarding order conditions, quality assurance conditions, and manufacturing conditions for each similar order condition and quality performance information for each condition.

[0129] This disclosure provides a comprehensive design assessment result that determines an unsuitable design condition if even one of the following conditions—component-level review target, material-level review target, and appearance-level review target—is not met: In other words, only when all review targets are deemed satisfactory can the comprehensive assessment result for similar order-level review targets be deemed satisfactory.

[0130] For example, if the similar order identification ID of FIG. 12 is AC1 and the design condition identification ID is BD1, and the results of the quality risk review regarding the quality of components, materials, and appearance are all output as good performance, the quality risk detection result of the comprehensive order unit design condition can be judged as good performance.

[0131] Figure 13 is a drawing for explaining the review results of the design condition unit review target according to one embodiment.

[0132] Referring to Figure 13, an example of the results of a quality risk review regarding quality assurance conditions and a comprehensive judgment result is shown based on each condition regarding order conditions, quality assurance conditions, and manufacturing conditions for each design condition and quality performance information for each condition.

[0133] For example, if the design condition identification ID is BD1, and the quality risk review results for components, materials, and appearance quality are all output as good performance, the overall quality risk detection result of the design condition can be judged as good performance.

[0134] As another example, if the design condition identification ID is BD2, the quality risk review result regarding the quality of components and materials is output as good performance, but the quality risk review result regarding the quality of appearance is output as performance risk, the quality risk detection result of the comprehensive design condition can be judged as performance risk.

[0135] FIG. 14 is a drawing for explaining a design condition quality risk pre-detection device according to one embodiment.

[0136] Referring to FIG. 14, the design condition quality risk pre-detection device (1400) may include an information acquisition unit (1430) that acquires quality performance information.

[0137] For example, the information acquisition unit (1430) can acquire quality performance information on products manufactured based on design conditions that have a certain level of similarity with the design conditions from a database. The quality of products manufactured based on previously registered design conditions can be measured, and whether the design conditions are met can be determined based on the measured numerical information. The present disclosure can determine whether the design conditions are met by quantifying the quality of products manufactured under similar design conditions to the design conditions to be detected.

[0138] The design condition quality risk pre-detection device (1400) of the design condition may include an indicator calculation unit (1440) that obtains quality performance information.

[0139] For example, the indicator calculation unit (1440) can calculate the first quality risk indicator or the second quality risk indicator through an artificial intelligence model based on the judgment result of determining whether the number of quality performance information acquired from the information acquisition unit is N or more (N is an integer greater than or equal to 1). The number of quality performance information that serves as the basis for the calculation method of the quality risk indicator can be conveniently set by the user, and since the number of quality performance information acquired from the database is N or more, it means that past quality performance data is trustworthy, and the first quality risk indicator is calculated using quality performance data retrieved from the database, and if there is no data retrieved or the number is small, less than N, the second quality risk indicator can be calculated based on data output by a learned and trustworthy artificial intelligence model.

[0140] For example, the first quality risk index produced by the index production unit (1440) may include a first process capability index and a first defect rate, and the second quality risk index may include a second process capability index and a second defect rate. The process capability index is a numerical value for comparing the level required to improve the production process with the actual production result to determine whether the current production process is suitable. The defect rate refers to a numerical value that represents the ratio of the number of products that do not meet the standards set by the user among the products produced under specific design conditions as a percentage. The present disclosure can determine whether the design conditions to be detected are good or dangerous by calculating the process capability index and the defect rate using information acquired from a database in which a conventional production history is stored or information output through an artificial intelligence model.

[0141] For example, the first process capability index and the first defect rate calculated by the index calculation unit (1440) may be calculated based on the average and deviation of quality figures included in the quality performance information. In addition, the second process capability index and the second defect rate may also be calculated based on the average and deviation of figures output by the artificial intelligence model, and the process capability index or defect rate may be calculated by the artificial intelligence model and only the results may be output.

[0142] For example, if the number of quality performance information acquired from the database by the information acquisition unit (1430) in the indicator production unit (1440) is N or more, the first quality risk indicator can be produced based on the quality performance information acquired from the database, and if the number of quality performance information is less than N, the second quality risk indicator can be produced based on information output through the artificial intelligence model.

[0143] For example, the artificial intelligence model used to produce the second quality risk indicator in the indicator production unit (1440) may include a Bayesian neural network model.

[0144] A method for detecting quality risks of design conditions may include a result output unit (1450) that outputs quality risk detection results.

[0145] For example, the result output unit (1450) can output a quality risk detection result of the design condition based on either the first quality risk index or the second quality risk index produced by the index production unit (1440).

[0146] For example, the result output unit (1450) can output the quality risk detection result as good performance based on the fact that the first process capability index calculated by the index calculation unit (1440) is higher than the first reference value and the first defective rate is lower than the second reference value; the quality risk detection result can be output as performance risk based on the fact that the first process capability index is lower than the first reference value or the first defective rate is higher than the second reference value; the quality risk detection result can be output as good prediction based on the fact that the second process capability index is higher than the third reference value and the second defective rate is lower than the fourth reference value; and the quality risk detection result can be output as risk prediction based on the fact that the second process capability index is lower than the third reference value or the second defective rate is higher than the fourth reference value. Outputting as good performance or performance risk is determined based on the first quality risk indicator, meaning that the number of quality performance information retrieved from the database is N or more, which is set by the user, and outputting as expected good or expected risk is determined based on the second quality risk indicator, meaning that the number of quality performance information retrieved is less than N.

[0147] The present disclosure may be configured with a design condition quality risk pre-detection device (1400) comprising an information acquisition unit (1430), an indicator calculation unit (1440), and a result output unit (1450), and may further include a design condition extraction unit (1410) and a compilation unit (1420).

[0148] Specifically, the design condition quality risk pre-detection device (1400) may include a design condition extraction unit (1410) that extracts similar design conditions based on criteria set by a user from a database storing past design history. The design condition extraction unit (1410) may extract similar design conditions based on similar criteria set by a user from a database storing past design history. In addition, the extraction of the above-mentioned oil price design conditions may be classified and extracted into types of design conditions, such as order conditions, quality assurance conditions, and manufacturing conditions.

[0149] For example, extraction of similar design conditions based on order conditions can be based on criteria such as type, product specification, thickness, order width, and order purpose. Extraction of similar design conditions based on quality assurance conditions can be based on criteria such as component conditions such as the customer's requested steel composition and product composition, material quality such as mechanical properties, and appearance quality such as shape, surface, and dimension. In addition, each quality can include information on specimen type, test conditions, and warranty scope. Extraction of similar design conditions based on manufacturing conditions can be based on information on key design factors being designed in the minor firing, plant, tapping composition, steelmaking, continuous casting, and rolling processes.

[0150] Each design condition extracted according to similar criteria can be stored in a separate database.

[0151] The design condition quality risk pre-detection device (1400) may include a compilation unit (1420) that compiles items to be reviewed for detection (S210).

[0152] For example, the review target item may be determined based on at least one of the similar design conditions extracted by the design condition extraction unit (1410) and the design conditions to be detected. The review target item refers to a standard for determining which of the multiple items set as design conditions will be reviewed for validity. The review target item may be organized based on the similar design conditions extracted by the design condition extraction unit (1410) or based on the actual design conditions to be detected. Once the review target item is determined, the review target item may store related information in a separate database.

[0153] For example, the review target item may include at least one of a quality unit item, an order unit item, and a design unit item. Alternatively, the review target item may include at least one of a first quality risk detection result, a second quality risk detection result for the order unit item, and a third quality risk detection result for the design unit item. In order to determine whether the design condition to be detected is suitable by reviewing it based on various criteria, the present disclosure determines the detection result as an item including at least one of a quality unit item, an order unit item, and a design unit item, and if at least one of the review target items is determined to be risky, the design condition to be detected may be output as unsuitable.

[0154] For example, the quality risk detection result may include a fourth quality risk detection result related to the suitability of the design conditions to the design conditions, and the fourth quality risk detection result may include any one of performance good, performance risk, good expected, risk expected, or indeterminate. For example, based on whether at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result is performance risk or risk expected, the fourth quality risk result may be output as either the performance risk or the risk expected.

[0155] Through the operation of the aforementioned components, accurate quality risk monitoring of the design conditions to be detected can be performed. This can reduce the defect rate of products produced based on the design conditions to be detected.

[0156] The above description is merely an illustrative example of the technical idea of ​​the present disclosure, and those skilled in the art to which the present disclosure pertains will appreciate that various modifications and variations can be made without departing from the essential characteristics of the technical idea of ​​the present disclosure. In addition, the present embodiments are not intended to limit the technical idea of ​​the present disclosure but rather to explain it, and therefore the scope of the technical idea of ​​the present disclosure is not limited by these embodiments. The scope of protection of the present disclosure should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included within the scope of the rights of the present disclosure.

[0157]

[0158] CROSS-REFERENCE TO RELATED APPLICATION

[0159] This patent application claims priority under 35 USC § 119(a) to Korean Patent Application No. 10-2023-0184030, filed December 18, 2023, the entire contents of which are incorporated herein by reference. Furthermore, this patent application claims priority in countries other than the United States for the same reasons, the entire contents of which are incorporated herein by reference.

Claims

1. In the method for detecting quality risks of design conditions by the device, An acquisition step of acquiring quality performance information of a product manufactured based on design conditions having a certain level of similarity with the design conditions from the first database; A calculation step for calculating a first quality risk indicator or a second quality risk indicator through an artificial intelligence model based on the judgment result of determining whether the number of quality performance information obtained above is N or more (N is an integer greater than or equal to 1); and A quality risk detection method, comprising an output step for outputting a quality risk detection result of the design condition based on either the first quality risk indicator or the second quality risk indicator calculated above.

2. In paragraph 1, An extraction step for extracting similar design conditions based on criteria set by the user from a second database where past design history is stored; and It further includes a formation step for organizing the review target items to be detected, A quality risk detection method, characterized in that the above-mentioned review target item is determined based on at least one of the extracted similar design conditions and the design conditions to be detected.

3. In paragraph 2, The above review target items include at least one of a quality unit item, an order unit item, and a design unit item, A quality risk detection method, wherein the quality risk detection result includes at least one of a first quality risk detection result for the quality unit item, a second quality risk detection result for the order unit item, and a third quality risk detection result for the design unit item.

4. In paragraph 3, The above quality risk detection result includes the fourth quality risk detection result related to the suitability of the design conditions, The above 4th quality risk detection result includes one of the following: good performance, performance risk, good expected, risk expected, or indeterminate. A quality risk detection method, characterized in that the fourth quality risk result is output as either a performance risk or a risk expectation based on at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result being a performance risk or a risk expectation.

5. In paragraph 4, A quality risk detection method, characterized in that the fourth quality risk detection result is output as indeterminable based on at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result being indeterminable.

6. In paragraph 2, A quality risk detection method, characterized in that the above similar design conditions include order conditions, quality assurance conditions and manufacturing conditions.

7. In paragraph 1, The first quality risk indicator includes the first process capability index and the first defect rate. The second quality risk indicator includes the second process capability index and the second defect rate. Based on the above first process capability index being higher than the first reference value and the above first defective rate being lower than the second reference value, the quality risk detection result is output as good performance. Based on the above first process capability index being lower than the first reference value or the above first defective rate being higher than the above second reference value, the quality risk detection result is output as a performance risk, Based on the above second process capability index being higher than the third reference value and the above second defective rate being lower than the fourth reference value, the quality risk detection result is output as a good prediction. A quality risk detection method, characterized in that the quality risk detection result is output as a risk prediction based on the second process capability index being lower than the third reference value or the second defective rate being higher than the fourth reference value.

8. In paragraph 7, A quality risk detection method, characterized in that the first process capability index and the first defective rate are calculated based on the average and deviation of quality figures included in the quality performance information.

9. In paragraph 1, The above production steps are: A quality risk detection method characterized in that if the number of the above judgment results is N or more, the first quality risk indicator is calculated based on quality performance information acquired from the first database, and if the number of the above judgment results is less than N, the second quality risk indicator is calculated based on information output through the artificial intelligence model.

10. In paragraph 1, A quality risk detection method, characterized in that the above artificial intelligence model includes a Bayesian neural network model.

11. In a device that detects quality risks of design conditions, An information acquisition unit that acquires quality performance information of a product manufactured based on design conditions having a certain level of similarity with the design conditions from the first database; and An indicator calculation unit that calculates a first quality risk indicator or a second quality risk indicator through an artificial intelligence model based on the judgment result of determining whether the number of quality performance information obtained above is N or more (N is an integer greater than or equal to 1); and A quality risk detection device including a result output unit that outputs a quality risk detection result of the design condition based on either the first quality risk indicator or the second quality risk indicator calculated above.

12. In paragraph 11, A design condition extraction unit that extracts similar design conditions based on criteria set by the user from a second database where past design history is stored; and Further including a composition section that organizes the review target items to be detected, A quality risk detection device, characterized in that the above-mentioned review target item is determined based on at least one of the extracted similar design conditions and the design conditions to be detected.

13. In paragraph 12, The above review target items include at least one of a quality unit item, an order unit item, and a design unit item, A quality risk detection device, wherein the quality risk detection result includes at least one of a first quality risk detection result for the quality unit item, a second quality risk detection result for the order unit item, and a third quality risk detection result for the design unit item.

14. In paragraph 13, The above quality risk detection result includes the fourth quality risk detection result related to the suitability of the design conditions, The above 4th quality risk detection result includes one of the following: good performance, performance risk, good expected, risk expected, or indeterminate. A quality risk detection device, characterized in that the fourth quality risk result is output as either a performance risk or a risk expectation based on at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result being a performance risk or a risk expectation.

15. In paragraph 14, A quality risk detection device, characterized in that the fourth quality risk detection result is output as indeterminable based on at least one of the first quality risk detection result, the second quality risk detection result, and the third quality risk detection result being indeterminable.

16. In paragraph 12, A quality risk detection device, characterized in that the above similar design conditions include order conditions, quality assurance conditions and manufacturing conditions.

17. In paragraph 11, The first quality risk indicator includes the first process capability index and the first defect rate. The second quality risk indicator includes the second process capability index and the second defect rate. Based on the above first process capability index being higher than the first reference value and the above first defective rate being lower than the second reference value, the quality risk detection result is output as good performance. Based on the above first process capability index being lower than the first reference value or the above first defective rate being higher than the above second reference value, the quality risk detection result is output as a performance risk, Based on the above second process capability index being higher than the third reference value and the above second defective rate being lower than the fourth reference value, the quality risk detection result is output as a good prediction. A quality risk detection device, characterized in that the quality risk detection result is output as a risk prediction based on the second process capability index being lower than the third reference value or the second defective rate being higher than the fourth reference value.

18. In paragraph 17, A quality risk detection device, characterized in that the first process capability index and the first defective rate are calculated based on the average and deviation of quality figures included in the quality performance information.

19. In paragraph 11, The above indicator calculation section, A quality risk detection device characterized in that if the number of the above judgment results is N or more, the first quality risk indicator is calculated based on quality performance information acquired from the first database, and if the number of the above judgment results is less than N, the second quality risk indicator is calculated based on information output through the artificial intelligence model.

20. In paragraph 11, A quality risk detection device, characterized in that the artificial intelligence model includes a Bayesian neural network model.

Images