Failure mode analysis support system

Abstract

To efficiently perform failure mode analysis.SOLUTION: A proposed failure mode analysis support method is executed by a computer. The failure mode analysis support method includes: input processing s1 of inputting at least one piece of information to be taken as an object of failure mode analysis; and processing sa of, on the basis of information inputted in the input processing, questioning a conversational AI so as to obtain at least one failure mode which may occur in a product or a production process, as an answer to obtain the answer of the failure mode(s) which may occur in the product or the production process.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a failure mode analysis support system. [Background technology]

[0002] Japanese Patent Laid-Open Publication No. 2005-293403 discloses a design work support device that displays a business process template indicating design work procedures on an operation screen and supports the provision of design information and the accumulation of design results in a database in accordance with the design work procedures indicated in the business process template. The design work support device includes an automatic utilization history recording means that records, as a utilization history for each piece of design information, association information between the design work procedures and the design information provided for the design work procedures or generated as a design result in the design work procedures, and a design information utilization history database that stores and manages the utilization history. This design work support device is said to be able to accumulate the utilization history of design information without requiring the user's effort and to facilitate the reuse of design results.

[0003] International Publication No. 2023 / 218659 discloses a software development support device. The software development support device disclosed in this publication includes a similar task extraction unit configured to calculate a similarity between a workflow of a plurality of first tasks, for which software corresponding to each task has already been developed, and a workflow of a second task, for which software corresponding to each task has not yet been developed, and to extract some of the first tasks from the plurality of first tasks based on the similarity, and a software output unit configured to output a group of software corresponding to the extracted first task workflows. This reduces software development costs. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2005-293403 [Patent Document 2] International Publication No. 2023 / 218659 Summary of the Invention [Problem to be solved by the invention]

[0005] The present inventors would like to efficiently carry out failure mode analysis. [Means for solving the problem]

[0006] The failure mode analysis support method proposed herein is a failure mode analysis support method executed by a computer, which includes the steps of: an input process in which at least one piece of information to be subjected to failure mode analysis is input; A process sa for asking the interactive AI a question based on the information input in the input process so that at least one failure mode that may occur in the product or manufacturing process is answered, and obtaining an answer for the failure mode that may occur in the product or manufacturing process. Includes. According to this failure mode analysis support method, failure mode analysis can be carried out efficiently. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a block diagram of a failure mode analysis support system 10 that embodies the failure mode analysis support method proposed here. [Figure 2] FIG. 2 shows an example of a failure mode analysis format used in the failure mode analysis support method. [Figure 3] FIG. 3 is a flow diagram showing an example of the process sa. [Figure 4] FIG. 4 is a flow diagram showing an example of the process sb. DETAILED DESCRIPTION OF THE INVENTION

[0008] An embodiment of the technology disclosed herein will be described below with reference to the drawings. The embodiment described here is, of course, not intended to limit the present invention. The drawings are schematic and do not necessarily reflect the actual product. Furthermore, the same reference numerals are appropriately used for components and parts that perform the same function, and redundant explanations will be omitted where appropriate.

[0009] <Failure Mode Analysis Support System 10> Figure 1 is a block diagram of a failure mode analysis support system 10 that embodies the failure mode analysis support method proposed here. Failure mode analysis is a system that analyzes the failure modes of products and manufacturing processes. Failure mode analysis can include an analysis method (FMEA) that identifies potential failure modes in products and manufacturing processes in advance, analyzes their impact, extracts possible causes, and then takes measures. Here, FMEA stands for Failure Mode and Effects Analysis.

[0010] FMEA, a type of failure mode analysis, involves, for example, a process of identifying potential failure modes in products and manufacturing processes in advance, a process of considering and evaluating the impact and degree of impact when a failure mode occurs in a product or manufacturing process, a process of considering the cause of failure (failure mechanism), an evaluation of the frequency of occurrence, and measures for current design and management.

[0011] For example, if a product's manufacturing process includes a step of "fastening parts with screws," the event of "poor screw tightening" would be listed as a potential failure mode in the product or manufacturing process during the process of identifying potential failure modes in advance. Then, during the process of analyzing and evaluating the impact, the impact and degree of impact on the product or manufacturing process caused by the event of "poor screw tightening" are considered and evaluated. For example, the impact of the event of "poor screw tightening" could include events such as "detachment of parts." These impacts and the degree of impact are then evaluated. Furthermore, for the event of "poor screw tightening," the causes, mechanisms, and frequency of occurrence are evaluated, and countermeasures for the current design and management are considered. Examples of causes and mechanisms of failure include "operational errors, such as not applying the appropriate torque or fastening a screw in the wrong position," as well as component or design issues, such as damaged screws or missing washers.

[0012] Regarding measures for current design and management, preventive management and detection management are considered based on the results of an investigation into the causes and mechanisms of failure. Preventive management is a measure to prevent "poor screw tightening" itself. Measures to prevent "poor screw tightening" itself, for example, include "checking by the operator" if the cause is "operational error." Similarly, if the cause is "loosening due to vibration," examples include "design changes, such as changing to screws that do not loosen." Detection management is a measure to detect "poor screw tightening." Measures to detect "poor screw tightening" during manufacturing include "establishing a process to check for loose screws in a later process." Furthermore, for post-production poor screw tightening, examples include "conducting regular inspections." In this way, it is desirable for preventive management and detection management to take effective measures depending on the cause and mechanism of failure.

[0013] In this way, FMEA, which is a type of failure mode analysis, considers, for example, the process of "fastening parts with screws," then considers the failure mode of "poor screw fastening," the effect of which is "detachment of parts," its importance, the cause of "poor screw fastening," the mechanism, frequency of occurrence, preventive measures, and inspection methods.

[0014] Figure 2 is an example of a failure mode analysis format used in failure mode analysis support methods. In FMEA, which is one type of failure mode analysis, this step-by-step examination is summarized in a table, such as the one shown in Figure 2. Here, for the item "screw fastening," the table is used to fill in the items such as failure mode, failure impact, impact level, failure cause (failure mechanism), occurrence frequency, preventive control, and detection control in order. While this is an example of failure mode analysis using FMEA, FMEA can be applied not only to the item "screw fastening," but also to various items in products and product manufacturing processes. FMEA then organizes and presents potential problems that may occur in products and product manufacturing processes and the solutions to address them.

[0015] Failure mode analysis, such as FMEA, is used in various areas, including design, manufacturing processes, and product evaluation. In design, design FMEA is applied to product design. The purpose of design FMEA is to identify and correct potential bugs and quality risks during the design phase. In manufacturing processes, process FMEA, for example, applies the FMEA concept to process management. Process FMEA primarily aims to prevent potential problems during the manufacturing process. Process FMEA can identify failure modes for each element of the process—equipment, workers, materials, methods, and measurements—to evaluate problems and risk. In product evaluation, functional FMEA is used to evaluate risk by focusing on the components that make up a specific product or system. Functional FMEA can apply to the components of a product, or to the functions and programs that make up a system. It is recommended that an appropriate format be prepared for the FMEA, depending on the context in which failure mode analysis is used. The table format used for failure mode analysis can, for example, follow previous FMEA examples.

[0016] For example, FMEA, which is performed in the analysis of product development processes, requires experience and skill to identify all possible failure modes related to the product and develop countermeasures, resulting in a large amount of work. In particular, as products and manufacturing processes become more complex, the number of items to be considered increases, and the task of developing countermeasures also requires a wealth of experience and skill. In failure mode analyses such as FMEA, results vary depending on the experience and knowledge of the operator, with more experienced operators tending to perform better analyses. Given these circumstances, the inventors hope to make failure mode analyses such as FMEA more efficient and easier, and to quickly obtain consistent, useful analysis results regardless of the operator's experience or knowledge.

[0017] The failure mode analysis support method proposed herein is executed by a computer. One means for realizing the failure mode analysis support method may be a computer program executed by a computer. The failure mode analysis support system 10 may be realized by a computer system. The computer system may include a memory for storing the computer program and one or more processors capable of executing the computer program stored in the memory.

[0018] More specifically, a computer embodying the failure mode analysis support system 10 may include, for example, an interface (I / F) for receiving data from an external device, a central processing unit (CPU) for executing program instructions, a ROM for storing the program executed by the CPU, a RAM used as a working area for expanding the program, and a storage device (recording medium) such as a memory for storing the program and various data. Each function of the failure mode analysis support system 10 can be realized through cooperation with a computer (hardware) that executes a predetermined program (software). Although not shown, the failure mode analysis support system 10 may also be implemented by multiple control devices working together. The failure mode analysis support system 10 may also be implemented by, for example, the cooperative processing of multiple computers connected via a network. Computers capable of performing intelligent processing through mechanical learning include those known as artificial intelligence. Artificial intelligence is sometimes referred to as AI (artificial intelligence).

[0019] Here, the computer program may be stored in, for example, a non-transitory computer readable medium. The program may also be supplied to a computer through such a non-transitory computer readable medium. Examples of non-transitory computer readable media include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), CD-ROMs (Read Only Memory), etc. The program may also be installed in a computer via a network such as the Internet. The program may also be executed by multiple processors.

[0020] As shown in FIG. 1, the failure mode analysis support system 10 includes an input unit 21, an output unit 22, and a processing unit 12. In this embodiment, the processing unit 12 includes processing units A, B, B1, C, and C1, each of which performs a required function. The failure mode analysis support system 10 also includes a communication function with an interactive AI. In the embodiment shown in FIG. 1, the interactive AI is incorporated into an external system 11. When the interactive AI is incorporated into an external system, the failure mode analysis support system 10 may be configured to be connectable to an API (Application Programming Interface) of the interactive AI. The interactive AI may be incorporated into the failure mode analysis support system 10. The failure mode analysis support system 10 may also be provided by a server available via a network such as the Internet. The API of the interactive AI may be used by the failure mode analysis support system 10.

[0021] Here, the input unit 21 executes various input processes s1 of the failure mode analysis support system 10. In the input process s1, for example, at least one piece of information to be the target of failure mode analysis may be input. For example, the information input in the input process s1 may be information input for an item to be the target of risk analysis and evaluation in a failure mode analysis such as FMEA. In a failure mode analysis such as FMEA, it is desirable that the item to be the target of risk analysis and evaluation be information that can be broken down into specific causes. For example, if a product uses a screw fastening structure or a welded structure, terms such as "screw fastening" or "welding" may be items in the FMEA.

[0022] The information input in input process s1 may be information selected and input by the user for risk analysis and evaluation in failure mode analysis such as FMEA. As a process executed by a computer, the user may input the required information via a display or the like.

[0023] The information input in input process s1 is, for example, at least one of information related to the product structure, information related to the product function, and information related to the manufacturing process. Information related to the product structure includes product design drawings and parts diagrams. Information related to the product function includes product specifications and instruction manuals. Manufacturing processes include "processing," "assembly," "inspection," "storage," and "shipping." Information related to the manufacturing process can include a process chart listing work procedures.

[0024] For example, in the case of a lithium-ion secondary battery, information about the product structure includes the external dimensions (size), the structure of the positive electrode, the structure of the negative electrode, the composition of the electrolyte, the structure of the electrode terminals, and the structure and sealing of the battery case. Information about the product's function includes the storage capacity, operating voltage, energy density, and cycle life. Information about the manufacturing process includes steps such as kneading the active material, applying the active material, drying, pressing, winding the sheet, inserting into the battery case, welding the battery case, injecting liquid, aging, sealing the injecting hole, and modularization. Here, a lithium-ion secondary battery is given as an example of a product, but the products that can be subjected to failure mode analysis using the failure mode analysis support method proposed here are not limited to lithium-ion secondary batteries.

[0025] In this case, the input process s1 may input character information extracted by a computer from text information related to the product structure, product function, and manufacturing process. In the input process s1, it is preferable that the computer automatically extracts terms that can be used as items in a failure mode analysis. For example, a computer (AI (artificial intelligence)) may be used in the process s1a (information extraction process) for extracting terms that can be used as items in a failure mode analysis. The AI ​​may be trained, for example, using previously created FMEA data as training data, to extract terms that can be used as items in a failure mode analysis from various text data related to the product and manufacturing process. For example, as shown in FIG. 1, it is preferable that the text information related to the product structure, product function, and manufacturing process is recorded in a predetermined database 420. In the information extraction process s1a, it is preferable that the pre-trained AI extracts terms that can be used as items in a failure mode analysis from various text data related to the product and manufacturing process.

[0026] In this way, the information input in the input process s1 may be manually input of phrases that can be used as items in a failure mode analysis. Alternatively, the input process s1 may be calibrated so that candidate keywords are extracted by AI from various text data related to products and manufacturing processes. Keywords to be used as FMEA items may be manually selected from the extracted keywords. In this case, a list of keywords extracted by the machine-learned computer is displayed on a display. A user may then select phrases that can be used as items in a failure mode analysis from the list of keywords displayed on the display. In this way, the input process s1 of the failure mode analysis support method may be configured so that a trained, machine-learned computer lists phrases that can be used as items in a failure mode analysis as input process s1 based on various text data related to products and manufacturing processes. Alternatively, the input process s1 of the failure mode analysis support method may be configured so that a trained, machine-learned computer automatically inputs phrases that can be used as items in a failure mode analysis based on various text data related to products and manufacturing processes.

[0027] In this way, by using pre-trained AI (artificial intelligence) in the input process s1, it is possible to prevent necessary items from being omitted from the items to be entered into the table in the predetermined format of the FMEA. Also, by using pre-trained AI (artificial intelligence) in the input process s1, it is possible to compensate for lack of experience or skill of the operator, thereby improving the quality of the failure mode analysis. It is also possible to improve the work efficiency of the failure mode analysis.

[0028] <Processing section A> In processing section A, a question is posed to the interactive AI based on the information input in input process s1 so that the AI ​​can answer at least one failure mode that may occur in the product or manufacturing process, and process sa is executed to obtain an answer regarding the failure mode that may occur in the product or manufacturing process. Figure 3 is a flow diagram showing an example of process sa.

[0029] Conversational AI Conversational AI (Conversational artificial intelligence) refers to technologies such as chatbots and virtual agents that can converse with users. Conversational AI is realized by combining natural language processing (NLP) and machine learning to respond to questions from users. Conversational AI can utilize the API (Application Programming Interface) of a large-scale natural language processing model. For example, a conversational AI service is available that can understand the meaning and purpose of a question posed in text and generate an appropriate response. An example of such a conversational AI service is CHAT-GPT, provided by OpenAI.

[0030] The interactive AI used in the failure mode analysis support method proposed here is preferably a trained interactive AI that has been trained using data recorded by linking keywords obtained from product information or manufacturing process information with failure modes as training data. Such an interactive AI outputs failure modes by asking questions that answer possible failure modes that may occur in the product or manufacturing process based on the information entered in the input process.

[0031] For example, if "screw tightening" is included as manufacturing process information, process SA asks the conversational AI to answer the failure mode that occurs in "screw tightening", as shown in Figure 3. The conversational AI answers the failure mode that occurs in "screw tightening". Specifically, for example, the conversational AI is asked a question such as, "Tell me the failure mode that occurs in 'screw tightening'". In response, the conversational AI answers that one of the failure modes that occurs in "screw tightening" is "improper screw tightening". Based on the conversational AI's answer, "improper screw tightening" is output as the failure mode.

[0032] <Processing sa> In process sa, as shown in Figure 3, the question to obtain an answer from the conversational AI may be created by the conversational AI (process sa1). In other words, process sa1 is a process for creating a question to obtain an answer about a failure mode. In this case, in process sa1, the conversational AI generates a question to be asked to the conversational AI based on the information input in input process s1 so that at least one failure mode that may occur in the product or manufacturing process is answered based on the information input in input process s1. Then, the conversational AI may ask the generated question to the conversational AI through the conversational AI's API to obtain an answer about the failure mode (process sa2).

[0033] For example, if "screw tightening" is included as information on the manufacturing process, process SA asks the conversational AI a question to answer so that it will answer the failure mode that occurs in "screw tightening." The conversational AI will answer the failure mode that occurs in "screw tightening." Specifically, for example, the conversational AI is asked a question such as, "Please tell me the failure mode that occurs in "screw tightening." In response, the conversational AI will answer that a failure mode that occurs in "screw tightening" is "improper screw tightening." Note that here, one failure mode is listed as the answer from process SA, but multiple failure modes may be listed. If asked, "Please tell me three failure modes that occur in "screw tightening," the conversational AI will list three failure modes that occur in "screw tightening." The conversational AI may answer, for example, "Failure modes that occur in "screw tightening" include "improper screw tightening," "forgetting to tighten a screw," and "screw falling off."

[0034] <Processing section B> In processing section B, as shown in Fig. 1, a process sb is executed in which a question is posed to an interactive AI configured to answer at least one cause of failure based on the failure mode obtained as an answer in process sa, and an answer to the cause of failure is obtained. Fig. 4 is a flow diagram showing an example of process sb.

[0035] In process sb, the conversational AI should be trained in advance so that it can obtain the cause of failure as an answer to a failure mode. For example, a question to the conversational AI may be, "Please tell me the cause of failure for the failure mode obtained as an answer in process sa." In this case, the conversational AI should create a related question based on the failure mode obtained as an answer in process sa. For example, if the failure mode obtained as an answer in process sa is "poor tightening of a screw," the conversational AI creates a question saying, "Please tell me the cause of poor tightening of a screw." The conversational AI responds to this question by saying, "One of the causes of poor tightening of a screw is 'insufficient torque.'" In response to this response, process sb outputs "insufficient torque" as the cause of failure for the failure mode.

[0036] Note that there may not always be one cause of failure for a given failure mode. Depending on the product, complex factors may overlap, making it difficult to narrow down multiple causes of failure to one. For this reason, in process sb, questions may be posed to the interactive AI so that multiple causes for the failure mode can be obtained as answers.

[0037] For example, if the failure mode is "poorly tightened screws," it would be a good idea to ask the conversational AI a question such as, "List 10 causes of poorly tightened screws." In contrast, conversational AI, for example, "The following are possible causes of "poor screw tightening." 1. Insufficient torque applied 2. The screws are not positioned correctly 3. The screw is damaged 4. Dirty threads 5. There are not enough washers 6. Not using enough thread locker 7. Poor fastening due to material issues 8. Incorrect screw size 9. Bent bolts 10. Proper fastening procedures not followed The example given here is just one example of a conversational AI's response.

[0038] In this case, all answers may be adopted, or the user may select the most likely ones based on the causes listed by the interactive AI and adopt them in the FMEA.In this way, the failure mode analysis support method allows questions to be asked to the interactive AI and failure mode analysis to be performed based on the answers, making it possible to perform failure mode analysis efficiently.

[0039] While the example described here assumes that process sb is performed on the assumption that process sa is performed, the failure mode analysis support method proposed here is not limited to this unless otherwise specified. For example, the failure mode analysis support method may include a process of asking an interactive AI a question to answer at least one failure cause for a failure mode in product information or a manufacturing process, and obtaining the answer to the failure cause. In this case, for example, the failure mode of the product information or the manufacturing process may be input by a user to a computer. The interactive AI may be asked a question to answer the failure cause for the input failure mode, and the failure cause may be obtained as an answer. In this case, a question may be created by the interactive AI based on the input failure mode. Alternatively, a user may input a question to a computer that will obtain a failure cause for the failure mode as an answer.

[0040] In other words, the failure mode analysis support method preferably includes a process sa for asking the interactive AI a question to answer at least one failure cause for a failure mode of product information or a manufacturing process, and obtaining the answer to the failure cause. In this process sa, the user may input a question to obtain the answer to the failure cause for the failure mode, and the question may be input to the interactive AI via the API of the interactive AI, and the answer to the failure cause for the failure mode may be obtained. The question may also be created by the interactive AI from the failure mode.

[0041] For example, if the failure mode is "defective damage to the plastic housing," the user may directly input a question to the conversational AI such as "list 10 causes of the defective damage to the plastic housing." Alternatively, when the user inputs "defective damage to the plastic housing," the program may create a request such as "list 10 causes of the failure for 'defective damage to the plastic housing'" and input it to the conversational AI.

[0042] In this case, the conversational AI would list 10 possible causes of the failure for "faulty damage to the plastic casing," for example, as follows: "The following are possible causes of "faulty damage to the plastic casing." 1. Physical damage caused by external impact or dropping 2. Environmental conditions such as ultraviolet rays and chemicals 3. Inappropriate material selection 4. Defects or imperfections in the manufacturing process 5. Proper reinforcement is not provided 6. Material deterioration due to fluctuations in heat and humidity 7. Deterioration of materials due to aging 8. Design Flaws and Weaknesses 9. Stress from improper handling 10. Maintenance and upkeep is not carried out properly.

[0043] In addition, for a power supply unit "fault that causes no power to be output," the conversational AI is asked (requested) to "list 10 faults that cause no power to be output in a power supply unit." The conversational AI will list 10 causes of failure for the "fault that causes no power to be output in a power supply unit" as follows, for example: The following are possible causes of the power supply not outputting power: 1. Power supply problem (plug not inserted properly, power off, etc.) 2. A faulty power cable or plug 3. Faulty fuse inside the device 4. Broken or damaged power cord 5. Power supply unit failure 6. Battery deterioration or failure 7. Overheating due to insufficient cooling system 8. Failure of circuits or components within the system 9. Protection circuit activated due to overload 10. Equipment damage caused by overvoltage and surges

[0044] Regarding the seal structure, the conversational AI is asked (requested) to "list 10 causes of poor sealing." The conversational AI will then list 10 causes of failure for "poor sealing," for example, as follows: Possible causes of "poor sealing" include the following: 1. Incorrect selection of sealing materials 2. Proper sealing process not followed 3. Deterioration of sealing materials due to environmental conditions 4. Uneven sealing surface 5. Sealing under improper temperature and humidity conditions 6. Problems caused by hardening or shrinkage of sealing materials 7. Defects caused by contamination or oil on the sealing surface 8. Air bubbles are trapped during sealing 9. Improper sealing pressure or speed 10. Deterioration of sealing materials due to end of life

[0045] In this way, conversational AI can quickly pick out multiple failure causes for one failure mode through learning. Based on what it has learned, conversational AI can identify causal events that are highly related to the failure mode as failure causes. The conversational AI is preferably a trained conversational AI that has learned using recorded data linking failure modes with failure causes as teaching data. The conversational AI may have previously learned the structure, function, manufacturing process, etc. of the product that is the subject of failure mode analysis. By having the conversational AI previously learn the structure, function, manufacturing process, etc. of the product, it will be able to pick out more appropriate failure causes for the product that is the subject of failure mode analysis.

[0046] The specific processing process of conversational AI is unknown. When a conversational AI is asked to list around 10 failure causes, some of them may be suitable causes that a user would never have thought of immediately. Therefore, in failure mode analysis, it can identify failure causes that are better than the user's experience and knowledge. On the other hand, conversational AI may include inappropriate answers, such as when it has not yet fully learned. Therefore, it may be possible to allow the user to select an appropriate failure cause from among those listed by the conversational AI. In either case, conversational AI-mediated failure mode analysis enables more efficient and superior failure mode analysis. Furthermore, as the conversational AI's learning progresses, it is expected that selection by an experienced and knowledgeable user will no longer be necessary. Even if a user selects an appropriate failure cause from those listed by the conversational AI, the user does not need to start from scratch. This allows for efficient and superior failure mode analysis. Furthermore, advanced failure mode analysis can be performed even when there is a lack of experienced and knowledgeable users.

[0047] <Processing section C> Processing section C executes processing sc, which queries the interactive AI about the cause of a failure in the product information or manufacturing process and obtains a solution to the failure cause. Here, the interactive AI is preferably a trained interactive AI that has learned using data recording the correlation between the failure cause and the solution to the failure as teaching data. The data recording the correlation between the failure cause and the solution to the failure can be, for example, the results of past FMEAs or literature such as technical books and papers that explain the causes of failure and the solution to the failure.

[0048] <Processing section B1, processing section C1> Processing section B1 executes processing sb1 to create a data table in which the product or manufacturing process information input in input processing s1, the failure mode obtained in processing sa, and the failure cause obtained in processing sb are compiled in tabular form. Furthermore, processing section C1 executes processing sc1 to create a data table in which the product or manufacturing process information input in the input processing, the failure mode obtained in processing sa, the failure cause obtained in processing sb, and the solution to the failure cause obtained in processing sc are compiled in tabular form. Here, data table T1 may be created according to a predetermined format, for example, as shown in FIG. 2.

[0049] In the failure mode analysis support method proposed here, for example, in data table T1 shown in Fig. 2, items to be subject to failure mode analysis are entered in item column e1 based on information entered in input process s1. In input process s1, for example, items to be subject to failure mode analysis may be entered directly by a user. Alternatively, targets of failure mode analysis may be extracted and input by a computer from information on the product structure, information on the product function, information on the manufacturing process, etc. Alternatively, after targets of failure mode analysis are extracted by a computer, the user may select the items to be subject to failure mode analysis.

[0050] In the failure mode analysis support method proposed here, when the item column e1 is input, questions and answers are repeatedly posed to the interactive AI, thereby filling in the data table T1 sequentially. Here, in process sa, when the item column e1 is input, a question is posed to the interactive AI based on the item so that it can answer a failure mode corresponding to the content input in the item column e1, and the failure mode is obtained as an answer. In response to the answer from the interactive AI, the failure mode answered by the interactive AI is entered into the failure mode column e2. Furthermore, in process sb, a question is posed to the interactive AI so that it can answer a failure mechanism based on the failure mode, and an answer about the cause of the failure is obtained. In response to the answer from the interactive AI, the failure mechanism answered by the interactive AI is entered into the failure mode column e6. Furthermore, in process sc, a question is posed to the interactive AI so that it can answer a solution to the failure cause based on the failure cause, and an answer about the solution to the failure cause is obtained. In response to the answer from the interactive AI, the solution to the failure cause answered by the interactive AI is entered into the failure cause solution columns e8 and e9. In the format shown in Figure 2, the column for solutions to failure causes includes management for preventing failures (prevention management e8) and management for detecting failures (detection management e9). Answers to each of these should be obtained and entered using the interactive AI. Although omitted here, the degree of impact of the failure (e4), importance (e5), and frequency of occurrence (e7) should also be asked to the interactive AI in that order, and the answers should be entered.

[0051] <Output section 22> In this manner, in the form shown in FIG. 2, "screw tightening" is input into the item column e1. Thereafter, questions and answers are repeatedly posed to the interactive AI, thereby filling in each column of the predetermined data table T1 (in the example of FIG. 2, each column e2 to e9). In this manner, in the failure mode analysis support method, questions and answers are repeatedly posed to the interactive AI, thereby filling in each column of the predetermined data table T1 in turn. The output unit 22 outputs the filled-in predetermined data table T1 of the failure mode analysis and provides it to the user (se). The output unit 22 may, for example, perform a process se of transmitting the data table T1 to a predetermined user terminal. The output unit 22 may also display the data table T1 on a display of the user terminal.

[0052] The user can check the failure mode analysis data table T1 provided by the failure mode analysis support method and modify it as necessary. In this case, the user can check and modify the failure mode analysis data table T1 after it has been completed, so the user's effort in performing the failure mode analysis is minimal. Furthermore, as the learning of the interactive AI progresses, it is expected that the accuracy of the failure mode analysis data table T1 provided by the failure mode analysis support method will improve. The user's effort in performing the failure mode analysis will be significantly reduced.

[0053] Conversational AI additional learning The failure mode analysis support method may further include a process sd for updating the interactive AI. Here, the process sd for updating the interactive AI includes a process sd1 for acquiring corrected data in which the relationship between input information and output information of the interactive AI has been corrected, and a process sd2 for causing the interactive AI to learn the corrected data as teaching data. In the process sd1 for acquiring the corrected data, the corrected data obtained by the user modifying the failure mode analysis data table T1 may be recorded in a predetermined database 400. Then, in the process sd2 for causing the interactive AI to learn, the corrected data obtained by the user modifying the failure mode analysis data table T1 may be learned by the interactive AI as teaching data. This is expected to improve the accuracy of the data table T1 provided by the failure mode analysis support method.

[0054] The failure mode analysis support method exemplified here may be realized by a computer program executed by a computer. Furthermore, the failure mode analysis support system 10 may be configured as a computer system including a memory for storing such a computer program and one or more processors capable of executing the computer program stored in the memory.

[0055] As explained above, The failure mode analysis support system 10 includes: an input process (s1) of inputting extracted information about at least one of the product structure, the product function, and the manufacturing process steps; A process of asking the conversational AI questions for each piece of input information and obtaining an answer for at least one failure mode (that may occur in the product or manufacturing process), and a process (sa to sb) of asking the conversational AI questions for each failure mode and obtaining an answer for at least one failure cause; A process (sb1) of creating a data table T1 (see FIG. 2) in which the extracted information, the failure modes answered for each of the extracted information, and the failure causes answered for each of the failure modes are compiled in a table format; It may be configured to be executed by at least one computer.

[0056] The failure mode analysis support system 10 includes: The system may be configured such that at least one computer further executes processes (sa to sc) for asking a question to the interactive AI for each cause of the failure and obtaining a solution. In this case, the process (sc1) for creating the data table T1 should be configured to create a data table T1 in which extracted information about the structure and function of the product or the steps of the manufacturing process, the failure modes answered for each extracted information, the causes of failure answered for each failure mode, and the solutions answered for each cause of failure are compiled in tabular form.

[0057] The failure mode analysis support system 10 is further configured to have at least one computer execute an information extraction process s1a in which information is extracted from a database 420 in which data on the product structure, product function, and manufacturing process steps is recorded. The input process s1 is preferably configured to input the information extracted by the information extraction process s1a.

[0058] The invention disclosed herein has been described in various ways. Unless otherwise specified, the embodiments described herein do not limit the present invention. Furthermore, the embodiments of the invention disclosed herein can be modified in various ways, and each component and each process described herein can be omitted or combined as appropriate, unless a particular problem arises.

[0059] As described above, this specification includes the disclosures set forth in the following sections.

[0060] Section 1: A failure mode analysis support method executed by a computer, comprising: an input process in which at least one piece of information to be subjected to failure mode analysis is input; A process sa for asking the interactive AI a question based on the information input in the input process so that at least one failure mode that may occur in the product or manufacturing process is answered, and obtaining an answer for the failure mode that may occur in the product or manufacturing process. Including, Failure mode analysis support method.

[0061] Section 2: Item 1. The failure mode analysis support system according to item 1, wherein the information input in the input process is at least one of information regarding the structure of the product, information regarding the function of the product, and information regarding the manufacturing process.

[0062] Section 3: Item 3. The failure mode analysis support system according to item 2, wherein the input process inputs character information extracted by a computer from text information relating to the product structure, product function, and manufacturing process.

[0063] Section 4: A failure mode analysis support method according to any one of items 1 to 3, wherein the interactive AI is a trained interactive AI that has been trained using data recorded in which keywords obtained from product information or manufacturing process information are linked to failure modes as teaching data.

[0064] Section 5: A failure mode analysis support method executed by a computer, comprising: A process sa is included in which an interactive AI is asked a question so that at least one cause of failure is answered for a failure mode of product information or a manufacturing process, and an answer of the cause of failure is obtained. Failure mode analysis support method.

[0065] Item 6: A process sb is a process of asking a question to an interactive AI configured to answer at least one cause of failure based on the failure mode obtained as an answer in the process sa, and obtaining an answer to the cause of failure. 5. The failure mode analysis support method according to any one of claims 1 to 4, further comprising:

[0066] Section 7: Item 6. A failure mode analysis support method according to Item 6, wherein the interactive AI is a trained interactive AI that has been trained using data recorded in which failure modes and failure causes are linked together as teaching data.

[0067] Section 8: Item 6. The failure mode analysis support method according to item 6, further comprising a process sb1 for creating a data table that compiles in tabular form the information on the product or manufacturing process input in the input process, the failure mode obtained in the process sa, and the failure cause obtained in the process sb.

[0068] Section 9: A failure mode analysis support method executed by a computer, comprising: A failure mode analysis support method including a process sc for asking the interactive AI about the cause of a failure in product information or a manufacturing process and obtaining a solution to the cause of the failure.

[0069] Section 10: A failure mode analysis support method according to any one of items 6 to 8, further comprising a process sc for asking the interactive AI about the cause of the failure obtained as an answer in process sb and obtaining an answer for a solution to the cause of the failure.

[0070] Section 11: Item 11. The failure mode analysis support method according to Item 10, wherein the interactive AI is a trained interactive AI that has learned using data that records correlations between failure causes and solutions to the failure causes as teaching data.

[0071] Section 12: A process sc1 is performed to create a data table in which the product or manufacturing process information input in the input process, the failure mode obtained in the process sa, the failure cause obtained in the process sb, and the solution to the failure cause obtained in the process sc are compiled in a table format. The failure mode analysis support method according to item 10 further includes:

[0072] Section 13: The method further includes a process sd for updating the interactive AI; where: The process sd for updating the conversational AI includes: A process sd1 of acquiring corrected data in which the relationship between the input information and the output information of the interactive AI has been corrected; A process sd2 in which the interactive AI learns the correction data as teaching data; 13. A failure mode analysis support method according to any one of items 1 to 12, comprising:

[0073] Item 14 A computer program that causes a computer to carry out the method described in items 1 to 13.

[0074] Section 15: A memory for storing the computer program according to item 14; and one or more processors capable of executing the computer program stored in the memory. [Explanation of symbols]

[0075] 10 Failure mode analysis support system 11 External Systems 12 Processing section 21 Input section 22 Output section 400 databases 420 databases T1 Data Table s1 input processing s1a Information extraction processing sa: A process of asking questions to the conversational AI and getting answers about failure modes sa1: Process to create questions to get answers to the causes of failures sa2: The process of asking the generated questions to the conversational AI and obtaining the answer to the failure mode sb Process of asking questions to the conversational AI and getting answers about the cause of the failure sc: A process of asking the conversational AI about the cause of the failure and obtaining a solution to the cause of the failure. sd Conversational AI update process sd1 Process to obtain correction data sd2 Processing to make conversational AI learn corrected data as teaching data se: Processing for sending data table T1 to a predetermined user terminal

Claims

1. A failure mode analysis support method executed by a computer, comprising: an input process in which at least one piece of information to be subjected to failure mode analysis is input; A process sa of asking an interactive AI question based on the information input in the input process so that at least one failure mode that may occur in the product or manufacturing process is answered, and obtaining an answer about the failure mode that may occur in the product or manufacturing process; Including, Failure mode analysis support method.

2. 2. The failure mode analysis support system according to claim 1, wherein the information input in the input process is at least one of information relating to the structure of the product, information relating to the function of the product, and information relating to the manufacturing process.

3. 3. The failure mode analysis support system according to claim 2, wherein said input process inputs character information extracted by a computer from text information relating to the structure, function, and manufacturing process of a product.

4. 2. The failure mode analysis support method according to claim 1, wherein the interactive AI is a trained interactive AI that has been trained using data recorded in which keywords obtained from product information or manufacturing process information are linked to failure modes as teaching data.

5. A failure mode analysis support method executed by a computer, comprising: A process sa of asking the interactive AI a question so that at least one failure cause is answered for a failure mode of the product information or the manufacturing process, and obtaining an answer to the failure cause; Failure mode analysis support method.

6. A process sb in which an inquiry is made to an interactive AI configured to answer at least one cause of failure based on the failure mode obtained as an answer in the process sa, and an answer to the cause of failure is obtained; The failure mode analysis support method according to claim 1 , further comprising:

7. 7. The failure mode analysis support method according to claim 6, wherein the interactive AI is a trained interactive AI that has been trained using data recorded in which failure modes and failure causes are linked together as teaching data.

8. 7. The failure mode analysis support method according to claim 6, further comprising a process sb1 for creating a data table in which the product or manufacturing process information input in the input process, the failure mode obtained in the process sa, and the failure cause obtained in the process sb are compiled in a table format.

9. A failure mode analysis support method executed by a computer, comprising: A failure mode analysis support method including a process sc for asking the interactive AI about the cause of a failure in product information or a manufacturing process and obtaining a solution to the cause of the failure.

10. 7. The failure mode analysis support method according to claim 6, further comprising a process sc for asking the interactive AI about the cause of failure obtained as a response in the process sb, and obtaining a response for a solution to the cause of failure.

11. 11. The failure mode analysis support method according to claim 10, wherein the interactive AI is a trained interactive AI that has learned using data recording correlations between failure causes and solutions to the failure causes as teaching data.

12. A process sc1 is performed to create a data table in which the product or manufacturing process information input in the input process, the failure mode obtained in the process sa, the failure cause obtained in the process sb, and the solution to the failure cause obtained in the process sc are compiled in a table format. The failure mode analysis support method according to claim 10, further comprising:

13. Further comprising a process sd for updating the interactive AI; where: The process sd for updating the interactive AI includes: A process sd1 of acquiring corrected data in which the relationship between the input information and the output information of the interactive AI has been corrected; A process sd2 of making the interactive AI learn the correction data as teaching data; The failure mode analysis support method according to claim 1 , further comprising:

14. A computer program causing a computer to carry out the method according to any one of claims 1 to 12.

15. a memory for storing the computer program of claim 14; and one or more processors capable of executing the computer program stored in the memory.

Images