Production risk analysis device and production risk analysis method

Abstract

A production risk analysis device (100) comprises: a risk scenario pattern acquisition unit (112) that, on the basis of product-related information (121) including at least one of product specifications and production process information, uses a language model (see a language model server (200)) to acquire a risk scenario pattern pertaining to the production of the product; a causal model generation unit (113) that extracts causal relationships between factors and results included in the risk scenario pattern and generates a causal model (130); and a detailed risk scenario acquisition unit (114) that, on the basis of the product-related information (121) and the causal model (130), uses the language model to acquire a detailed risk scenario pertaining to the production of the product.

Description

Production Risk Analysis Device and Production Risk Analysis Method

[0001] The present invention relates to a production risk analysis device and a production risk analysis method for assisting risk analysis in product production.

[0002] In the manufacturing industry, the external environment is becoming more complex and diverse, such as the increasing geopolitical risks like disputes and epidemics, recalls due to quality standards, the progress of labor shortages, the shortening of the product life cycle, and the diversification of product varieties. Therefore, it is becoming difficult to predict the future external environment, consider the configuration of the production line, and formulate an investment plan.

[0003] As a method for risk assessment of a manufacturing system, there is a method described in Patent Document 1. In this method, a manufacturing system is modeled by a resource graph defining the physical resources of the system, a process graph defining the services executed by the system, and the mapping between these resource graph and process graph. By performing simulations of service performance with multiple sets of operation parameters according to the resource graph and process graph, the performance indicators of the manufacturing system are modeled. Based on the modeled performance indicators, risks can be identified. A manufacturing risk indicates a change in performance indicators exceeding a predetermined threshold. And a set of operation parameters corresponding to a result without manufacturing risk can be identified. Based on this finding, changes to the manufacturing system can be determined, and the manufacturing system can be optimized to avoid risks.

[0004] Japanese Patent Application Laid-Open No. 2024-068660

[0005] In the method described in Patent Document 1, risks are identified by obtaining a resource graph and a process graph including causal dependency relationships. Creating a resource graph and a process graph including causal dependency relationships requires knowledge in a wide range of fields such as economy, politics, society, technology, and environment, and incurs a great cost. The present invention has been made in view of such a background, and an object thereof is to provide a production risk analysis device and a production risk analysis method capable of improving the efficiency of risk analysis related to product production.

[0006] To solve the above-mentioned problems, the production risk analysis device according to the present invention comprises: a risk scenario pattern acquisition unit that acquires risk scenario patterns related to the production of a product using a language model based on product-related information including at least one of product specifications and production process information; a causal model generation unit that generates a causal model by extracting the causal relationships between factors and results included in the risk scenario patterns; and a detailed risk scenario acquisition unit that acquires detailed risk scenarios related to the production of a product using the language model based on the product-related information and the causal model.

[0007] According to the present invention, it is possible to provide a production risk analysis apparatus and a production risk analysis method that enable improved efficiency in risk analysis related to product production. Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

[0008] This is a functional block diagram of the production risk analysis device according to this embodiment. This diagram illustrates the overall flow of the production risk analysis process according to this embodiment. This shows an example of a risk scenario pattern acquisition prompt according to this embodiment. This is an example of a causal model according to this embodiment. This shows an example of a detailed risk scenario acquisition prompt according to this embodiment. This is a flowchart of the production risk analysis process according to this embodiment. This is a hardware configuration diagram showing an example of a computer that implements the functions of the production risk analysis device according to the above embodiment.

[0009] <<Overview of Production Risk Analysis Device>> The following describes a production risk analysis device in an embodiment for carrying out the present invention. The production risk analysis device according to this embodiment queries a language model (language model server) based on product-related information such as product specifications, configuration, and production process to obtain a risk scenario (first risk scenario) related to product production. Next, the production risk analysis device performs text analysis on the first risk scenario to generate a causal model that includes risk factors and results.

[0010] Next, the production risk analysis system queries the language model based on product-related information and causal models to obtain risk scenarios related to product production (second risk scenarios). Because the language model is queried including the causal model generated based on the first risk scenario, the second risk scenarios contain more detailed information than the first risk scenarios. For this reason, below, the first risk scenarios will also be referred to as risk scenario patterns, and the second risk scenarios as detailed risk scenarios.

[0011] The production risk analysis system collects information related to the risk factors included in risk scenarios. It also performs simulations of response costs and production costs in the event of a risk occurrence, and calculates the impact on costs.

[0012] Such production risk analysis devices enable efficient risk analysis related to product production. For example, they reduce the need to analyze the causal relationships between risks and factors, and allow for risk analysis that includes complex causal relationships where causal relationships are linked in a chain. They also enable risk analysis that includes a wide range of perspectives, such as economic, political, and environmental factors. In addition, they enable quantitative risk assessment.

[0013] ≪Configuration of the Production Risk Analysis Device≫ Figure 1 is a functional block diagram of the production risk analysis device 100 according to this embodiment. The production risk analysis device 100 is a computer and comprises a control unit 110, a storage unit 120, and an input / output unit 180. User interface devices such as a display, keyboard, and mouse are connected to the input / output unit 180. The input / output unit 180 also has a communication device and is capable of sending and receiving data with the language model server 200. The language model server 200 is a server that provides a service of returning answers to inquiries called prompts using a language model. An example of the language model server 200 is an artificial intelligence chatbot.

[0014] ≪Production Risk Analysis Device: Memory Unit≫ The memory unit 120 is composed of memory devices such as ROM (Read Only Memory), RAM (Random Access Memory), and SSD (Solid State Drive). The memory unit 120 stores product-related information 121, risk scenario information 122, causal model 130, causal factor information 123, information resource information 124, evaluation information 125, and program 128. The program 128 includes a description of the processing to be executed by the functional unit provided in the control unit 110, which will be described later. The various contents of the memory unit 120 may be stored in an external storage device such as a cloud server and read as needed.

[0015] <<Production Risk Analysis Device: Control Unit>> The control unit 110 is configured to include a CPU (Central Processing Unit) and comprises a reception unit 111, a risk scenario pattern acquisition unit 112, a causal model generation unit 113, a detailed risk scenario acquisition unit 114, a causal factor information acquisition unit 115, and an evaluation unit 116. The control unit 110 may also be configured to include a GPU (Graphics Processing Unit), an NPU (Neural (network) Processing Unit), an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), etc.

[0016] Figure 2 is a diagram illustrating the overall flow of the production risk analysis process according to this embodiment. The contents stored in the memory unit 120 and the functions of the control unit 110 will be explained with reference to Figure 2.

[0017] ≪Reception Department / Product-Related Information≫ The reception department 111 receives information about the product to be manufactured, entered by the user of the production risk analysis device 100, and stores it in the product-related information 121. The product-related information 121 is information such as the product specifications, configuration, production process, and production line entered by the user. Examples of production process information include: "If the process involves assembling hard parts, both human workers and robots are candidates." and "In the case of difficult parts such as wire harnesses, human workers are candidates because handling by robots is difficult."

[0018] <<Risk Scenario Pattern Acquisition Unit / Risk Scenario Pattern (Risk Scenario Information)>> The Risk Scenario Pattern Acquisition Unit 112 generates a Risk Scenario Pattern Acquisition Prompt 510 (see Figure 3 below) and sends it to the Language Model Server 200. The Risk Scenario Pattern Acquisition Unit 112 stores the Risk Scenario Pattern 410, which is the risk scenario returned by the Language Model Server 200, in the Risk Scenario Information 122.

[0019] Figure 3 shows an example of a risk scenario pattern acquisition prompt 510 according to this embodiment. Element 511 indicates product specifications, configuration, production process, etc., which are included in product-related information 121. Examples of risk scenario patterns 410 include "Production site A has a risk of conflict and a risk of supply chain disruption," and "Production site A tends to have a labor shortage."

[0020] As described above, the production risk analysis device 100 includes a risk scenario pattern acquisition unit 112 that acquires a risk scenario pattern 410 related to the production of the product using a language model (see language model server 200) based on product-related information 121 which includes at least one of the product specifications and production process information.

[0021] ≪Causal Model Generation Unit and Causal Model≫ Returning to Figure 2, we will continue the explanation of the control unit 110 and the storage unit 120. The causal model generation unit 113 generates a causal model 130 by performing text analysis on the risk scenario pattern 410. Figure 4 is an example of a causal model 130 according to this embodiment. A causal model (causal relationship model) shows the relationship between a cause (factor, cause) and an effect (damage, risk). In Figure 4, the relationship is shown by an arrow with the cause as the starting point and the effect as the ending point. The causal model 130 is not limited to a graph consisting of a starting point, arrows, and an ending point, but may also be shown with text such as, for example, "There is a possibility that a risk of labor shortage will occur due to population decline as a factor." Note that although Figure 4 shows one linked graph, the results of the text analysis may also be multiple linked graphs.

[0022] As described above, the production risk analysis device 100 includes a causal model generation unit 113 that extracts the causal relationships between factors and results included in the risk scenario pattern 410 and generates a causal model 130. The causal model generation unit 113 extracts causal relationships by performing text analysis on the risk scenario pattern 410.

[0023] ≪Detailed Risk Scenario Acquisition Unit / Detailed Risk Scenario (Risk Scenario Information)≫ Returning to Figure 2, we will continue the explanation of the control unit 110 and the storage unit 120. The detailed risk scenario acquisition unit 114 generates a detailed risk scenario acquisition prompt 520 (see Figure 5 below) and sends it to the language model server 200. The detailed risk scenario acquisition unit 114 stores the detailed risk scenario 420 returned by the language model server 200 in the risk scenario information 122.

[0024] Figure 5 shows an example of the detailed risk scenario acquisition prompt 520 according to this embodiment. Element 521 shows the product specifications, configuration, production process, etc., which are included in the product-related information 121. Element 522 shows the causal model 130 in text format. An example of a detailed risk scenario 420 is as follows.

[0025] Example 1: Production site A faces the risk of conflict and supply chain disruption. As a result, the production line may be shut down for a certain period. If product production stops, production costs will increase and opportunity costs will be lost. Example 2: There is a risk of rising material costs in the production site A region. As a result, there is a risk of increased production costs. Example 3: Production site A tends to have a labor shortage. As a result, labor costs will rise, and there is a risk of increased production costs. In addition, there is a risk of decreased production yield due to an increase in early retirement of workers and a decline in skills.

[0026] As described above, the production risk analysis device 100 includes a detailed risk scenario acquisition unit 114 that acquires a detailed risk scenario 420 related to the production of the product, based on product-related information 121 and a causal model 130, using a language model (see language model server 200).

[0027] ≪Causal Factor Information Acquisition Unit and Causal Factor Information≫ Returning to Figure 2, we will continue the explanation of the control unit 110 and the storage unit 120. The causal factor information acquisition unit 115 performs a text analysis of the detailed risk scenario 420 to acquire the causes (factors) of risk occurrence. Next, the causal factor information acquisition unit 115 refers to the information resource information 124 to acquire information related to the causes and stores it in the causal factor information 123. The information resource information 124 stores reliable information resources outside the production risk analysis device 100. Examples of information resources include news provision sites, newspapers, broadcasting stations, and news agencies. The information resource information 124 may also store the reliability level of each information resource. The causal factor information acquisition unit 115, for example, accesses the server provided by the information resource, searches using the cause as a keyword, and acquires information with a high degree of relevance or recently provided information to make it the causal factor information 123.

[0028] As described above, the production risk analysis device 100 includes a causal factor information acquisition unit 115 that acquires factors included in the detailed risk scenario 420 and acquires external information (see causal factor information 123) that serves as the basis for said factors.

[0029] ≪Evaluation Unit and Evaluation Information≫ The evaluation unit 116 evaluates the risks shown in the detailed risk scenario 420. More specifically, the evaluation unit 116 extracts risks by text analysis of the detailed risk scenario 420 and stores them as evaluation information 125. Next, the evaluation unit 116 calculates the countermeasure costs, production costs, and the increase (increase rate) of production costs when the risk occurs, using production simulation, and stores them as evaluation information 125 in relation to the risks. Note that the countermeasure costs and the increase in production costs may be considered losses as they are new costs resulting from the occurrence of the risk.

[0030] Production costs include material costs and labor costs. Countermeasure costs, for example, are the costs of changing the production line. When a risk occurs, the production line is modified to minimize production costs. The evaluation unit 116 calculates the optimal product production line configuration after the risk occurs, as well as the cost of changing the production line. The following reference describes a method for changing the production line, and by using this technology, the changes and their costs can be calculated. Reference: Daiki Kajita et al., Development of an Integrated Automated Design Technology for Robot Production Lines that Can Quickly Respond to Changes in the Production Environment, Journal of the Japan Society for Precision Engineering, vol.87, no.2, pp.160-163, 2021.

[0031] When multiple risks are identified, the evaluation unit 116 calculates the cost of countermeasures and production costs for each risk. The evaluation unit 116 may also calculate the cost of countermeasures and production costs according to the degree (severity) of the risk. For example, if the risk of increased labor costs is identified, the evaluation unit 116 may calculate the cost of countermeasures and production costs by setting the degree of increase (severity), such as a 10% increase or a 20% increase.

[0032] The evaluation unit 116 performs a sensitivity analysis to analyze the degree to which a risk affects production costs. For example, the evaluation unit 116 calculates the increase in production costs when labor costs increase by 10%. The evaluation unit 116 calculates the importance of the risk according to the sensitivity and stores it as evaluation information 125 in association with the risk. The evaluation unit 116 may also calculate a higher importance of the risk if the increase (loss) in countermeasure costs and production costs is large.

[0033] The evaluation unit 116 may calculate the importance of a risk as its probability of occurrence increases. The evaluation unit 116 may calculate the probability of a risk occurring based on the likelihood of the risk being included in the causal factor information 123, the number of pieces of information related to the risk included in the causal factor information 123, the number of information resources from which the information was acquired, the probability of similar risks occurring, etc.

[0034] The evaluation unit 116 displays the risk in the evaluation information 125, the scenario in which the risk occurs (see detailed risk scenario 420), the causal factor information of the risk (causal factor information 123), the response cost when the risk occurs, the production cost, the increment of the production cost, and the importance on a display connected to the input / output unit 180.

[0035] As described above, the production risk analysis device 100 includes an evaluation unit 116 that calculates the importance of a risk based on the amount of loss (countermeasure costs, increments in production costs) that would occur if a risk included in the detailed risk scenario 420 were to occur, and the amount of external information related to the factors of the risk (see causal factor information 123).

[0036] The evaluation unit 116 calculates the loss in the event that a risk included in the detailed risk scenario 420 occurs. The evaluation unit 116 calculates the optimal product production line configuration in the event that a risk included in the detailed risk scenario 420 occurs. The evaluation unit 116 calculates the importance of a risk based on the results of the sensitivity analysis of the loss to the degree of the risk in the event that a risk included in the detailed risk scenario 420 occurs.

[0037] <Production Risk Analysis Process> Figure 6 is a flowchart of the production risk analysis process according to this embodiment. At the start of the production risk analysis process, it is assumed that the user has already entered product-related information 121.

[0038] In step S11, the risk scenario pattern acquisition unit 112 generates a risk scenario pattern acquisition prompt 510 (see Figure 3) and sends it to the language model server 200 to acquire the risk scenario pattern 410. In step S12, the causal model generation unit 113 performs text analysis on the risk scenario pattern 410 to generate a causal model 130.

[0039] In step S13, the detailed risk scenario acquisition unit 114 generates a detailed risk scenario acquisition prompt 520 (see Figure 5) and sends it to the language model server 200 to acquire the detailed risk scenario 420. In step S14, the causal factor information acquisition unit 115 performs text analysis on the detailed risk scenario 420 to acquire the causes (factors) of the risk and acquire information related to those causes.

[0040] In step S15, the evaluation unit 116 performs a text analysis of the detailed risk scenario 420 to extract risks. In step S16, the evaluation unit 116 calculates the cost of countermeasures and production costs when the risks extracted in step S15 occur.

[0041] In step S17, the evaluation unit 116 calculates the increment in production costs due to the risk, the sensitivity to losses (countermeasure costs, increment in production costs) to the risk, and the importance of the risk. In step S18, the evaluation unit 116 outputs the risk, the risk occurrence scenario, risk factor information, the response costs due to the risk occurrence, production costs, the increment in production costs, and the importance.

[0042] ≪Features of the Production Risk Analysis Device≫ The production risk analysis device 100 acquires risk scenario patterns 410 using the language model server 200 based on product-related information 121. The production risk analysis device 100 generates a causal model 130 based on the risk scenario patterns 410 and further acquires detailed risk scenarios 420 using the language model server 200.

[0043] The production risk analysis device 100 acquires information on the factors (causes) of the risks included in the detailed risk scenario 420. The production risk analysis device 100 also calculates the response cost at the time of risk occurrence, the production cost, the increment of the production cost, and the importance.

[0044] The production risk analysis device 100 includes the generated causal model 130 in the detailed risk scenario acquisition prompt 520 and queries the language model server 200 to acquire the detailed risk scenario 420. By using this detailed risk scenario 420, it is possible to reduce the analysis of the causal relationship between risks and factors required for risk analysis, or to perform risk analysis including a complex causal relationship in which causal relationships are connected in a chain. In addition, risk analysis including a wide range of perspectives including economy, politics, and environment becomes possible. Therefore, the user of the production risk analysis device 100 can efficiently analyze production risks. Furthermore, quantitative risk assessment becomes possible.

[0045] <<Modification Example>> Although several embodiments of the present invention have been described above, these embodiments are merely examples and do not limit the technical scope of the present invention. For example, the production risk analysis device 100 uses an external language model server 200. Instead of this, the production risk analysis device 100 may generate an answer (for example, the risk scenario pattern 410) to a prompt (for example, the risk scenario pattern acquisition prompt 510) using a language model.

[0046] The present invention can take various other embodiments, and furthermore, various changes such as omission and substitution can be made without departing from the gist of the present invention. These embodiments and their modifications are included in the scope and gist of the invention described in this specification and the like, and are included in the scope of the invention described in the claims and its equivalent scope.

[0047] <<Hardware Configuration>> The production risk analysis apparatus 100 according to the above-described embodiment is realized by a computer 900 having a configuration as shown in FIG. 7, for example. FIG. 7 is a hardware configuration diagram showing an example of the computer 900 that realizes the functions of the production risk analysis apparatus 100 according to the above-described embodiment. The computer 900 includes a CPU 901, a ROM 902, a RAM 903, an SSD 904, and an input / output interface 905 (described as input / output I / F (Interface) in FIG. 7). Further, the computer 900 includes a communication interface 906 (described as communication I / F in FIG. 7) and a media interface 907 (described as media I / F in FIG. 7). The computer 900 may include an HDD (Hard Disc Drive) instead of the SSD 904, or may further include an HDD in addition to the SSD 904.

[0048] The CPU 901 operates based on a program stored in the ROM 902 or the SSD 904, and performs control by the control unit 110 in FIG. 1. The ROM 902 stores a boot program executed by the CPU 901 when the computer 900 is started up, a program related to the hardware of the computer 900, and the like.

[0049] The CPU 901 controls an input device 910 such as a mouse or a keyboard and an output device 911 such as a display or a printer via the input / output interface 905. The CPU 901 acquires data from the input device 910 via the input / output interface 905, and outputs the generated data to the output device 911.

[0050] The SSD 904 stores a program executed by the CPU 901 and data used by the program. The communication interface 906 receives data from another device (not shown, such as the language model server 200) via a communication network and outputs it to the CPU 901, and transmits data generated by the CPU 901 to another device via the communication network.

[0051] The media interface 907 reads a program or data stored in the recording medium 912 and outputs it to the CPU 901 via the RAM 903. The CPU 901 loads the program from the recording medium 912 onto the RAM 903 via the media interface 907 and executes the loaded program. The recording medium 912 can be an optical recording medium such as a DVD (Digital Versatile Disk), a magneto-optical recording medium such as an MO (Magneto Optical Disk), a magnetic recording medium, a conductive memory tape medium, or a semiconductor memory.

[0052] For example, when the computer 900 functions as the production risk analysis device 100 according to the above embodiment, the CPU 901 of the computer 900 realizes the functions of the production risk analysis device 100 by executing the program 128 (see Figure 1) loaded on the RAM 903. The CPU 901 reads the program from the recording medium 912 and executes it. In addition, the CPU 901 may read the program from another device via a communication network, or it may install the program 128 from the recording medium 912 onto the SSD 904 and execute it.

[0053] 100 Production risk analysis device 111 Reception unit 112 Risk scenario pattern acquisition unit 113 Causal model generation unit 114 Detailed risk scenario acquisition unit 115 Causal factor information acquisition unit 116 Evaluation unit 121 Product-related information 122 Risk scenario information 130 Causal model 123 Causal factor information 124 Information resource information 125 Evaluation information 200 Language model server (language model) 410 Risk scenario pattern 420 Detailed risk scenario

Claims

1. A production risk analysis device comprising: a risk scenario pattern acquisition unit that acquires risk scenario patterns related to the production of a product using a language model based on product-related information including at least one of product specifications and production process information; a causal model generation unit that generates a causal model by extracting causal relationships between factors and results included in the risk scenario patterns; and a detailed risk scenario acquisition unit that acquires detailed risk scenarios related to the production of a product using the language model based on the product-related information and the causal model.

2. The production risk analysis apparatus according to claim 1, further comprising a causal factor information acquisition unit that acquires factors included in the detailed risk scenario and acquires external information that serves as the basis for said factors.

3. The production risk analysis apparatus according to claim 2, further comprising an evaluation unit that calculates the importance of a risk based on the loss incurred when a risk included in the detailed risk scenario occurs, and the amount of external information relating to the factors of the risk.

4. The production risk analysis apparatus according to claim 1, further comprising an evaluation unit for calculating losses in the event that a risk included in the detailed risk scenario occurs.

5. The production risk analysis apparatus according to claim 1, wherein the causal model generation unit performs text analysis on the risk scenario patterns to extract the causal relationships.

6. The production risk analysis apparatus according to claim 1, further comprising an evaluation unit for calculating the optimal production line configuration for the product in the event that any of the risks included in the detailed risk scenarios described above occur.

7. The production risk analysis apparatus according to claim 1, further comprising an evaluation unit that calculates the importance of a risk based on the results of a sensitivity analysis of losses to the degree of the risk when a risk included in the detailed risk scenario occurs.

8. A production risk analysis method comprising: a production risk analysis device performing the steps of: acquiring risk scenario patterns related to the production of a product using a language model based on product-related information including at least one of product specifications and production process information; generating a causal model by extracting causal relationships between factors and results included in the risk scenario patterns; and acquiring detailed risk scenarios related to the production of the product using the language model based on the product-related information and the causal model.

Images