Data management device, data management system, data management method, and data management program

Abstract

The data management device (1) includes a standard data collection unit (14) that collects standard data, which has been set in advance as data that can be acquired as an analysis item for data analysis, from the production equipment (5) or an operation terminal (6) associated with the production equipment (5), a non-standard data collection unit (11) that collects non-standard data, which is data that includes data related to the production equipment (5) and is not the standard data, from the production equipment (5) or the operation terminal (6), and a data placement unit that stores the non-standard data and standard data as management items associated with the non-standard data in a data storage device.

Description

[Technical Field]

[0001] The present disclosure relates to a data management device, a data management system, a data management method, and a data management program that manage data. [Background technology]

[0002] Conventionally, when troubleshooting is performed at a production site, the details of the work are handed over in language, and the details of this handover are stored as language information, for example, in an in-house network folder. However, because this language information is stored discretely in a unique form, it is difficult to search and reuse.

[0003] Furthermore, while information managed for production sites, etc., includes standardized data (hereinafter referred to as standardized data) such as design information that can be collected mechanically and is easy to reuse, this information has been managed separately from non-standardized data (hereinafter referred to as non-standardized data) such as linguistic information.

[0004] The business support system described in Patent Document 1 stores structured data and unstructured data in a database in advance, and manages forms by pasting structured data and unstructured data selected from the database into the form. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 10-31694 Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the technology of Patent Document 1, unstructured data and structured data are stored independently of each other, and there is a problem in that the unstructured data cannot be utilized efficiently.

[0007] The present disclosure has been made in view of the above, and aims to provide a data management device that can efficiently utilize unstructured data. [Means for solving the problem]

[0008] In order to solve the above-mentioned problems and achieve the objectives, the data management device of the present disclosure includes a structured data collection unit that collects structured data, which has been set in advance as data that can be acquired as analysis items for data analysis, from production equipment or an operation terminal associated with the production equipment, and an unstructured data collection unit that collects unstructured data, which is data that includes data related to the production equipment but is not the structured data, from the production equipment or the operation terminal.The data management device of the present disclosure also includes a data allocation unit that stores the unstructured data and structured data as management items associated with the unstructured data in a data storage device. [Effects of the Invention]

[0009] The data management device according to the present disclosure has the effect of enabling efficient utilization of unstructured data. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram for explaining the concept of data management executed by a data management system according to an embodiment. [Figure 2] FIG. 1 is a diagram showing a configuration of a data management system according to an embodiment; [Figure 3] FIG. 1 is a diagram showing an example of a data structure of data stored in a database by a data management device according to an embodiment; [Figure 4] FIG. 1 is a diagram for explaining an association process executed by a data management device according to an embodiment; [Figure 5] 10 is a flowchart showing a procedure of a storage process in which the data management device according to the embodiment stores data in a database; [Figure 6] 1 is a flowchart showing a procedure for a process of utilizing unstructured data in a database by a data management device according to an embodiment. [Figure 7] FIG. 1 is a diagram showing a configuration of a learning device included in a data management system according to an embodiment; [Figure 8] FIG. 1 is a diagram for explaining a neural network used by a learning device according to an embodiment. [Figure 9] 1 is a flowchart showing a procedure of a learning process executed by a learning device according to an embodiment; [Figure 10] FIG. 1 is a diagram showing a configuration of an inference device included in a data management system according to an embodiment; [Figure 11] 1 is a flowchart showing a procedure of an inference process executed by an inference device according to an embodiment; [Figure 12] FIG. 1 is a diagram illustrating a configuration example of a processing circuit when the processing circuit included in the control circuit of the data management device according to the embodiment is realized by a processor and a memory. [Figure 13] FIG. 1 is a diagram illustrating a configuration example of a processing circuit included in a control circuit of a data management device according to an embodiment, in the case where the processing circuit is configured with dedicated hardware; DETAILED DESCRIPTION OF THE INVENTION

[0011] A data management device, a data management system, a data management method, and a data management program according to embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0012] Embodiment 1 is a diagram illustrating the concept of data management performed by a data management system according to an embodiment. The data management system 100 is a system that collects and manages data handled at a production site or the like. The data management system 100 includes a utilization device 40, a data management device 1, and a production site device 20.

[0013] The data management system 100 collects, as collected data, typical data, which is typical data, and untypical data, which is untypical data, from a production site where the production site devices 20 and the like are located. The data management system 100 collects, for example, typical data and untypical data that are generated in troubleshooting performed by workers (field workers) at the production site or the like.

[0014] The data management system 100 stores and manages structured data and unstructured data in association with each other. The data management system 100 treats the structured data as an index and reads out unstructured data corresponding to the structured data based on the structured data. The data management system 100 uses the read unstructured data to analyze and learn from the unstructured data, and outputs the learning results (learning model) to the production site device 20.

[0015] The production site device 20 uses a learning model to infer unstructured data (such as messages to workers at the production site) from structured data at the production site (data indicating a trouble state, etc.). The production site device 20 outputs the inferred unstructured data to the worker, thereby providing the worker with a message according to the trouble state, etc.

[0016] In the data management system 100, a user (an operator, an administrator, etc.) of the data management system 100 sets in advance which data is structured data and which data is unstructured data.

[0017] For example, the user may set whether data is structured or unstructured depending on the data format. The user may also set the data area in which structured data and unstructured data are stored. This allows the data management system 100 to determine whether data is structured or unstructured based on the data format or the data area in which it is stored when analyzing data. For example, the data management system 100 determines that data is unstructured if the data format is "JPEG."

[0018] Of the data collected by the data management system 100, standard data is data that has been set in advance as data that can be acquired as an analysis item for data analysis, and non-standard data is data other than standard data, including data related to production equipment.

[0019] The standardized data is objective data that can be obtained as quantitative values, such as equipment operation data, which is easy to categorize and does not involve human intervention.

[0020] The standard data includes attribute information indicating the attributes of the standard data and acquired data that has actually been acquired. The attribute information is, for example, a date and time, a process (process name), a trouble code, etc. If the attribute information is a date and time, the acquired data is data indicating the actual date and time, such as "PP year QQ month RR day SS hour TT minute UU second." If the attribute information is a process, the acquired data is data identifying the type of process, such as "fourth process," and if the attribute information is a trouble code, the acquired data is data identifying the type of trouble, such as "264."

[0021] The standard data is, for example, data (sensor values, etc.) obtained from the production site device 20. The standard data is, for example, measurement values ​​such as length, weight, load, area, temperature, current value, torque, and resistance value. The production site device 20 includes production equipment (production equipment 5 described later) and an operation terminal (operation terminal 6 described later) associated with the production equipment 5.

[0022] Examples of the production equipment 5 are machine tools, coating machines, and robots. Specific examples of the standard data when the production equipment 5 is a coating machine include information on the date and time when the coating process was performed, information identifying the process in which the coating process was performed, and information on the number of coatings performed in the coating process. Examples of the operation terminal 6 are a PC (Personal Computer) and a tablet terminal. The production equipment 5 may be operated by the operation terminal 6, or may be operated by an operation panel installed in the production equipment 5.

[0023] Atypical data is subjective data that is difficult to obtain as a quantitative value. Atypical data is data that involves humans and is difficult to categorize, such as natural conversational language data (linguistic information), image data, video data, audio data, waveform data, color data, odor data, and luminosity data. Compared to typical data, atypical data is data that is highly dependent on the worker. For example, untypical personal impressions, such as color data, odor data, and luminosity data, fall under the category of atypical data. On the other hand, data that can be measured by a sensor and quantified, data represented by an electrical signal, and data classified into levels fall under the category of typical data.

[0024] Atypical data is data obtained, for example, by judgments made by workers at a production site, and the content of the data varies depending on the worker making the judgment. For this reason, atypical data is data that is difficult to predict. Specific examples of atypical data include the causes of troubles at a production site, how to respond to the troubles, and how to deal with the troubles. The causes of troubles, how to respond to the troubles, and how to deal with the troubles are obtained, for example, by an experienced worker inputting free text or voice into an operation terminal 6 or the like. In the embodiment, troubles at a production site include abnormalities (for example, equipment abnormalities, product abnormalities), malfunctions, etc. at the production site.

[0025] Furthermore, the structured data and the unstructured data are in a relationship in which the unstructured data is influenced by or dependent on the structured data. In the above-mentioned case, the unstructured data, such as natural conversational language data, image data, video data, audio data, and waveform data, is, for example, data indicating the occurrence of a malfunction. The structured data, such as data obtained from the production site equipment 20 (length, weight, load, area, temperature, current value, torque, resistance value, etc.), is data indicating the cause of a malfunction, i.e., data to be checked when a malfunction occurs. By managing the unstructured data in association with the structured data, the data management system 100 can classify and manage the unstructured data and the structured data as correlated data, enabling efficient data utilization, for example, data analysis. Note that structured data and unstructured data can also be relative. Depending on the relationship between the structured data and the unstructured data (e.g., dependency), data classified as unstructured data may become structured data, and data classified as structured data may become structured data.

[0026] The cause of the trouble becomes non-standard data when the worker determines the cause of the trouble, but becomes standard data when the control device of the production equipment 5 or the like determines the cause of the trouble. Furthermore, the trouble content becomes non-standard data when it is video data or input by the worker, and becomes standard data when it is data that is associated in advance with a trouble code.

[0027] The utilization device 40 is a device that utilizes the data managed by the data management device 1, and the data management device 1 is a device that causes the utilization device 40 to utilize the data that it manages.

[0028] The utilization device 40 provides various unstructured data (messages, etc.) to the worker using, for example, AI (Artificial Intelligence). The utilization device 40 transmits the unstructured data provided to the worker to the data management device 1, and uses a learning model generated by the data management device 1 to infer the unstructured data to be provided to the worker.

[0029] The unstructured data that the utilization device 40 transmits to the data management device 1 is unstructured data to be managed. The data management device 1 manages the unstructured data received from the utilization device 40 and performs learning using the unstructured data.

[0030] 1 shows a case where the utilization device 40 is a production management device 41, a maintenance support device 42, an anomaly detection device 43, and a predictive maintenance device 44. Note that the utilization device 40 only needs to include at least one of the production management device 41, the maintenance support device 42, the anomaly detection device 43, and the predictive maintenance device 44. Furthermore, the utilization device 40 may also include devices other than these four devices.

[0031] The production management device 41 is a device that manages the production status at the production site. For example, the production management device 41 manages and adjusts the production schedule based on the progress of the production process and the production schedule. The production management device 41 notifies the workers of work instruction messages for adjusting the production schedule. For example, the production management device 41 notifies the workers of a message such as "The worker in the fourth process, please move on to the fifth process." This message (non-standard data) is sent to the data management device 1.

[0032] Furthermore, the production management device 41 infers non-standard data such as a message to be notified using the learning model received from the data management device 1. When the production management device 41 infers a message or the like, it notifies the worker of the inferred message or the like.

[0033] The maintenance support device 42 is a device that provides maintenance support to workers at the production site and the like via chat (generated AI such as chatGPT). The person with whom the worker chats using the maintenance support device 42 is AI. Note that the person with whom the worker chats using the maintenance support device 42 may also be a human (such as a production manager or technical manager). The maintenance support device 42 notifies the worker of an answer to a question from the worker. For example, the maintenance support device 42 receives a question from the worker such as, "There's been a time overrun problem with the coating machine C1 in the fourth process. Can you tell me the cause?" and notifies the worker of the answer to this question. This answer (non-standard data) is associated with the question and sent to the data management device 1. The answer notified by the maintenance support device 42 includes the cause of the problem and countermeasures for the problem.

[0034] Furthermore, the maintenance support device 42 infers non-standard data such as answers to questions using the learning model received from the data management device 1. When the maintenance support device 42 infers an answer or the like, it notifies the worker of the inferred answer or the like.

[0035] The anomaly detection device 43 is a device that detects anomalies in the production process at the production site. The anomaly detection device 43 determines and detects anomalies in the production process based on, for example, sensor values, timer values, etc. When the anomaly detection device 43 detects an anomaly, it notifies the worker with a message indicating the detected anomaly. For example, the anomaly detection device 43 notifies the worker with a message such as, "Time-over troubles are occurring frequently in the fourth process." This message (non-standard data) is sent to the data management device 1.

[0036] Furthermore, the anomaly detection device 43 infers non-standard data such as a message to be notified using the learning model received from the data management device 1. When the anomaly detection device 43 infers a message or the like, it notifies the worker of the inferred message or the like.

[0037] The predictive maintenance device 44 is a device that predicts abnormalities in the production process at the production site. The predictive maintenance device 44 predicts abnormalities in the production process based on, for example, sensor values, timer values, etc. When the predictive maintenance device 44 predicts an abnormality, it notifies the worker with a message indicating the predicted abnormality. For example, the predictive maintenance device 44 notifies the worker with a message such as, "The number of processing times for the fourth process has exceeded 1,600. If processing continues at this rate, there is a possibility of a trouble stopping the process." This message (non-standard data) is sent to the data management device 1.

[0038] Furthermore, the predictive maintenance device 44 infers non-standard data such as a message to be notified using the learning model received from the data management device 1. When the predictive maintenance device 44 infers a message or the like, it notifies the worker of the inferred message or the like.

[0039] The data management device 1 includes a database group 30, which is a data storage device. The data management device 1 collects unstructured data generated at the production site from at least one device disposed in the production site device 20 and the utilization device 40, and stores the data in the database group 30.

[0040] Furthermore, the data management device 1 collects design information 21, trouble occurrence information 22, and trouble response information 23 from the production site devices 20 and stores them in the database group 30. The design information 21 and the trouble occurrence information 22 contain structured data, and the trouble response information 23 contains structured data and unstructured data. In the embodiment, data that contains structured data and unstructured data, such as the trouble response information 23, may be referred to as unstructured data.

[0041] Trouble occurrence information 22 and trouble response information 23 are information generated at the production site and are transmitted from the production site devices 20 to the data management device 1. Design information 21 is information used at the production site and is transmitted from the production site devices 20 to the data management device 1. Note that the data management device 1 may collect the design information 21 from devices other than the production site devices 20 (such as design devices).

[0042] The design information 21 is, for example, production line design information 21. The design information 21 includes, for example, information on the external configuration of the production line, information on components (devices, equipment, etc.) of the production line, information on the operation and function of each component, information on the specifications (production capacity, etc.) of each component, information on the operation flow of the production line, etc.

[0043] Figure 1 shows a case where the design information 21 includes information indicating a process, information on the coating machine C1 used in this process, information on the robot used in this process, and information on the program used in this process.

[0044] The information indicating the process is, for example, information identifying a process such as the fourth process. The information on the coater C1 is the name of the manufacturer of the coater C1, its serial number, and the date of manufacture. The information on the robot is the name of the manufacturer of the robot, its serial number, and the date of manufacture. The information on the program is information identifying the program.

[0045] The trouble occurrence information 22 is information about troubles that have occurred at the production site. The trouble occurrence information 22 includes, for example, information about the date and time when the trouble occurred, information indicating the process in which the trouble occurred, a trouble code, details of the trouble, and the number of processes performed by the device that caused the trouble (such as the number of applications by the coating machine C1). The details of the trouble include the name of the device that caused the trouble (the coating machine C1), and information about the trouble itself, such as time over.

[0046] The trouble response information 23 is information on how workers handle trouble that occurs at the production site. The trouble response information 23 includes, for example, information on the date and time when the trouble occurred, information indicating the process in which the trouble occurred, the trouble code, the cause of the trouble, the response to the trouble, and the response time.

[0047] The cause of the trouble is information about the cause of the trouble, such as a clogged nozzle pipe. The response to the trouble is information about what response was made to the trouble, such as replacing the pipe. The response time is the time required to respond to the trouble. Of the trouble response information 23, the cause of the trouble and the response to the trouble are non-standard data, while the rest are standard data.

[0048] The non-standard data of the trouble response information 23 includes at least one of a trouble cause and a trouble response. In this case, at least one of the trouble cause and the trouble response is trouble information, which is non-standard data related to the trouble. In other words, the non-standard data of the trouble response information 23 includes trouble information, which is at least one of a trouble cause and a trouble response.

[0049] If the trouble content is non-standard data, the non-standard data of the trouble response information 23 may include the trouble information. In this case, the trouble information is at least one of the trouble content, the cause of the trouble, and the trouble response. That is, the non-standard data of the trouble response information 23 includes trouble information that is at least one of the trouble content, the cause of the trouble, and the trouble response.

[0050] By having the cause of the trouble clarified by the skilled worker and taking appropriate measures to deal with the trouble according to the cause of the trouble, the data management device 1 is able to collect the trouble response information 23 generated by the skilled worker, and can provide the non-standard data learned and inferred using the trouble response information 23 generated by the skilled worker to the inexperienced worker, the utilization device 40, etc.

[0051] The data management system 100 uses the structured data and unstructured data stored in the database group 30 by the data management device 1 for production schedule adjustment, maintenance support, anomaly detection, predictive maintenance, etc. Furthermore, the data management device 1 uses data obtained as unstructured data (such as work instructions for workers) to notify workers of how to deal with trouble (how to deal with the trouble) when an anomaly is detected.

[0052] 2 is a diagram showing the configuration of a data management system according to an embodiment. As described above, the data management system 100 includes a utilization device 40, a data management device 1, and a production site device 20.

[0053] The data management device 1 collects structured data and unstructured data from the production site devices 20. Specifically, the data management device 1 collects structured data and unstructured data from the production equipment 5 included in the production site devices 20 or the operation terminal 6 associated with the production equipment 5. The data management device 1 also collects unstructured data from the utilization device 40.

[0054] The data management device 1 includes an unstructured data collection unit 11, an unstructured data analysis unit 12, an unstructured data allocation unit 13, a structured data collection unit 14, a structured data analysis unit 15, and a structured data allocation unit 16. The data management device 1 also includes a database group 30 and a data processing unit 17.

[0055] The database group 30 includes a design DB 31, a trouble occurrence DB 32, and a trouble response DB 33. The design DB 31 stores design information 21. The trouble occurrence DB 32 stores trouble occurrence information 22. The trouble response DB 33 stores trouble response information 23.

[0056] The standard data collection unit 14 collects standard data from the production equipment 5 or the operation terminal 6 by accessing the production equipment 5 or the operation terminal 6. The standard data collection unit 14 also collects standard data by receiving standard data transmitted from the production equipment 5 or the operation terminal 6.

[0057] The routine data collection unit 14 collects design information 21 in advance, for example, before the production site device 20 starts operating. The routine data collection unit 14 also collects trouble occurrence information 22 after the production site device 20 starts operating. The routine data collection unit 14 transmits the collected routine data to the routine data analysis unit 15.

[0058] The standard data collection unit 14 collects the design information 21 from the production equipment 5 or the operation terminal 6 by accessing a memory or the like in the production equipment 5 or the operation terminal 6 that stores the design information 21. The memory address or the like of the memory to be accessed by the standard data collection unit 14 is set in advance in the standard data collection unit 14 by an administrator or the like.

[0059] Furthermore, the standard data collection unit 14 receives trouble occurrence information 22 transmitted from the production equipment 5 or the operation terminal 6, thereby collecting the trouble occurrence information 22 from the production equipment 5 or the operation terminal 6. When the production equipment 5 or the operation terminal 6 transmits standard data to the standard data collection unit 14, the administrator sets the device address of the standard data collection unit 14 in the production equipment 5 or the operation terminal 6 as the transmission destination of the standard data.

[0060] The standard data analysis unit 15 classifies each standard data into either the design information 21 or the trouble occurrence information 22 by analyzing whether each standard data is the design information 21 or the trouble occurrence information 22. An administrator sets in advance in the standard data analysis unit 15 which information is the design information 21 and which information is the trouble occurrence information 22. The standard data analysis unit 15 transmits the standard data and the classification result to the standard data allocation unit 16.

[0061] The standard data arrangement unit 16 stores each standard data in the database group 30 based on the classification result. Specifically, the standard data arrangement unit 16 stores design information 21 and information for identifying the design information 21 in the design DB 31. Furthermore, the standard data arrangement unit 16 stores trouble occurrence information 22 and information for identifying the trouble occurrence information 22 in the trouble occurrence DB 32. In this way, the standard data arrangement unit 16 stores standard data and information for identifying the standard data (hereinafter sometimes referred to as standard identification information) in the database group 30.

[0062] The unstructured data collection unit 11 is an input interface for voice, text, etc., and collects unstructured data. The unstructured data collection unit 11 collects unstructured data by receiving the unstructured data transmitted from the production equipment 5, the operation terminal 6, or the utilization device 40. The following describes the case where the unstructured data collection unit 11 mainly collects unstructured data from the production equipment 5 or the operation terminal 6.

[0063] The non-standard data collection unit 11 collects non-standard data (in the embodiment, trouble response information 23 including standard data and non-standard data) from the production equipment 5 or the operation terminal 6, and transmits it to the non-standard data analysis unit 12. The non-standard data collection unit 11 collects the trouble response information 23 after the production site device 20 starts operating.

[0064] The unstructured data collection unit 11 receives trouble response information 23 transmitted from the production equipment 5 or the operation terminal 6, and thereby collects the trouble response information 23 from the production equipment 5 or the operation terminal 6. The administrator sets the device address of the unstructured data collection unit 11 in the production equipment 5 or the operation terminal 6 as the destination of the unstructured data. Note that when the unstructured data collection unit 11 collects unstructured data from the utilization device 40, the administrator sets the device address of the unstructured data collection unit 11 in the utilization device 40 as the destination of the unstructured data.

[0065] The unstructured data collection unit 11 may collect unstructured data from the production site device 20 by accessing the production site device 20 or the utilization device 40. In other words, the unstructured data collection unit 11 may collect the trouble response information 23 from the production equipment 5 or the operation terminal 6 by accessing a device that generates and outputs the trouble response information 23 among the production equipment 5, the operation terminal 6, or the utilization device 40. In this case, the device addresses and the like of the devices that the unstructured data collection unit 11 accesses are set in advance in the unstructured data collection unit 11 by an administrator or the like.

[0066] The unstructured data analysis unit 12 analyzes which structured data each piece of unstructured data corresponds to, based on the unstructured data collected by the unstructured data collection unit 11 and the structured data in the database group 30. In other words, the unstructured data analysis unit 12 analyzes the unstructured data and structured data, and analyzes which structured data the unstructured data is linked to.

[0067] Each time the unstructured data collection unit 11 collects unstructured data, the unstructured data analysis unit 12 analyzes which of the structured data each piece of unstructured data corresponds to, and associates the unstructured data with the structured data based on the analysis results. That is, each time the unstructured data analysis unit 12 receives unstructured data from the unstructured data collection unit 11, it associates the unstructured data with the structured data for each received data unit.

[0068] The unstructured data analysis unit 12 analyzes the structured data and unstructured data, selects structured data whose degree of association with the unstructured data is higher than a specific value, and associates the analyzed unstructured data with the selected structured data. By analyzing the structured data and unstructured data, the unstructured data analysis unit 12 selects structured data corresponding to the unstructured data, and associates the structured data with the unstructured data using the selected structured data as a management item. The unstructured data analysis unit 12 associates the structured data with the unstructured data using the attribute information of the structured data as a higher-level management item and the acquired data of the structured data as a lower-level management item.

[0069] The non-standard data analysis unit 12 associates the trouble response information 23 with the trouble occurrence information 22, for example, based on the date and time information, process information, and trouble code included in the trouble response information 23 and the date and time information, process information, and trouble code included in the trouble occurrence information 22. The non-standard data analysis unit 12 also associates the trouble response information 23 with the design information 21 based on the process information included in the trouble response information 23 and the process information included in the design information 21.

[0070] For example, when a trouble occurs and a worker at the production site inputs a response to the trouble as voice (non-standard data) into the operation terminal 6, the non-standard data analysis unit 12 associates this instruction with standard data such as the content of the trouble.

[0071] Furthermore, for example, when a user makes a statement resulting from viewing structured data stored in the database group 30, the unstructured data analysis unit 12 associates the unstructured data with the viewed structured data. When the unstructured data analysis unit 12 detects that the operation terminal 6 has accessed the database group 30 and receives unstructured data corresponding to the statement from the unstructured data collection unit 11 within a specific time period after detection, the unstructured data analysis unit 12 associates the unstructured data corresponding to this statement with the accessed structured data.

[0072] For example, when a problem such as an abnormality or malfunction occurs at a production site, standard data such as a trouble code is stored in the database group 30. An operator may check (understand the current situation) the trouble code stored in the database group 30 using an operation terminal 6 such as a tablet terminal, and identify the cause of the trouble based on the trouble code. In some cases, an operator may visually check the location where the trouble occurred and identify the cause of the trouble. In this case, the operator inputs the cause of the trouble into the operation terminal 6 or the like.

[0073] When an operator inputs the cause of a trouble into the operation terminal 6, the non-standard data collection unit 11 acquires trouble response information 23 including the trouble code from the operation terminal 6. In this case, the non-standard data analysis unit 12 associates the cause of the trouble with the trouble code. As a result, the non-standard data including the cause of the trouble and the standard data including the trouble code are stored in the database group 30 in an associated manner.

[0074] Furthermore, the worker performs recovery work (actual troubleshooting) for the trouble based on the cause of the trouble. In this case, when troubleshooting, the worker may issue work instructions appropriate to the on-site situation based on the worker's experience. When the worker inputs work instructions (voice data) for the recovery work into the operation terminal 6, the non-standard data collection unit 11 collects non-standard data indicating the work instructions. In this case, the non-standard data analysis unit 12 associates the work instructions with the standard data. That is, when the non-standard data analysis unit 12 receives non-standard data indicating the work instructions from the non-standard data collection unit 11, it associates the standard data, such as the trouble code and cause of the trouble stored in the database group 30, with the non-standard data, such as the work instructions.

[0075] The unstructured data analysis unit 12 transmits unstructured data (voice data) containing the content of utterances such as work instructions and the like, and the type identification information of the type data associated with the unstructured data, to the unstructured data allocation unit 13, and the unstructured data allocation unit 13 stores the unstructured data and the type identification information in the trouble response DB 33. As a result, the unstructured data containing the content of utterances such as work instructions is temporarily stored as voice data in the trouble response DB 33. This voice data is analyzed in post-processing by the unstructured data analysis unit 12 or the data processing unit 17. The analysis in post-processing (post-event analysis) will be described later.

[0076] As another method for associating non-standard data with standard data, when a problem occurs and an operator makes a statement due to standard data being stored in the database group 30, the non-standard data analysis unit 12 may associate the stored standard data with the non-standard data such as the statement.

[0077] In this case, the unstructured data analysis unit 12 detects that a problem has occurred, and that structured data corresponding to the problem has been stored in the group of databases 30. The unstructured data analysis unit 12 detects that a problem has occurred when it receives problem occurrence information 22 from the unstructured data collection unit 11. Furthermore, when the unstructured data analysis unit 12 detects that problem occurrence information 22 has been stored in the group of databases 30, it detects that structured data corresponding to the problem has been stored in the group of databases 30. The unstructured data analysis unit 12 detects that structured data corresponding to the problem has been stored in the group of databases 30, and when it receives unstructured data corresponding to a message made within a specific time period after the detection from the unstructured data collection unit 11, it associates the unstructured data corresponding to the message with the structured data stored in the group of databases 30.

[0078] The unstructured data analysis unit 12 may later associate the unstructured data with the structured data after the unstructured data has been stored in the database group 30. In this case, the unstructured data analysis unit 12 may analyze the constituent words of the unstructured data received from the unstructured data collection unit 11 by text mining or the like, and associate the unstructured data with structured data that includes words similar to the constituent words. The unstructured data analysis unit 12 extracts structured data that includes words similar to the constituent words of the unstructured data based on a thesaurus, and associates the extracted structured data with the unstructured data.

[0079] For example, the unstructured data analysis unit 12 associates the selected type data with the unstructured data after the fact. When the unstructured data is selected mechanically, the unstructured data analysis unit 12 selects the type data at the time of trouble occurrence. When the type data is selected manually, the unstructured data analysis unit 12 selects the type data specified by the user operating the operation terminal 6.

[0080] Note that words similar to constituent words are not limited to words with similar meanings, but may also include words in the same category. That is, the unstructured data analysis unit 12 may determine words included in the same category as the constituent words as words similar to the constituent words. In this case, the unstructured data analysis unit 12 determines the category of the constituent words included in the unstructured data, for example, based on a thesaurus.

[0081] Furthermore, when the unstructured data is text, the unstructured data analysis unit 12 determines the category to which the constituent words of the unstructured data belong by processing similar to that of the generation AI. When the unstructured data is text, the unstructured data analysis unit 12 divides the text into its constituent elements and determines what the text wants to say, that is, the direction (vector). The unstructured data analysis unit 12 sets a category at the end of this direction, and the unstructured data analysis unit 12 determines this category as the category to which the constituent words of the unstructured data belong. The unstructured data analysis unit 12 transmits the unstructured data and the standard identification information of the standard data associated with this unstructured data to the unstructured data arrangement unit 13.

[0082] The unstructured data allocation unit 13 stores the unstructured data associated with the structured data in a database group 30, which is a data storage device. The unstructured data allocation unit 13 stores the unstructured data and the structured identification information of the structured data associated with the unstructured data in the database group 30. As a result, the data management device 1 stores the unstructured data and the structured data as a management item associated with the unstructured data in the database group 30. In other words, the data management device 1 stores the unstructured data and the structured data in the database group 30 in a state where they are associated with each other.

[0083] In this way, the data management device 1 selects the type of data corresponding to the untyped data, associates the selected type of data with the untyped data, and then stores the selected type of data in the database group 30 as a management item.

[0084] If the unstructured data is trouble response information 23, the unstructured data allocation unit 13 stores the unstructured data in the trouble response DB 33. Furthermore, if the unstructured data is, for example, voice collected from the utilization device 40, the unstructured data allocation unit 13 stores the unstructured data, such as voice, associated with the structured data in a DB (one of the design DB 31, the trouble occurrence DB 32, and the trouble response DB 33) in which the structured data is stored, based on the analysis results of the unstructured data analysis unit 12. Because the structured data is already stored in one of the DBs, the unstructured data allocation unit 13 can determine the DB in which to store the unstructured data, such as voice, by selecting the structured data associated with the unstructured data. In other words, the unstructured data allocation unit 13 stores the unstructured data, such as voice, in the same DB in which the structured data associated with the unstructured data is stored. For example, when the structured data associated with the unstructured data to be stored is stored in the design DB 31 , the unstructured data placement unit 13 stores the unstructured data to be stored in the design DB 31 .

[0085] If the unstructured data and the structured data can be associated with each other, the unstructured data arrangement unit 13 may store the unstructured data and the structured data in separate DBs.

[0086] This allows the data management device 1 to easily search for unstructured data corresponding to the structured data using the structured data as a search index, thereby enabling efficient utilization of unstructured data corresponding to the knowledge or tacit knowledge possessed by individual workers.

[0087] In other words, when a problem occurs at a production site, the data management system 100 can search for the non-standard data corresponding to the problem by using the standard data corresponding to the problem as a search index and searching for the non-standard data corresponding to the standard data.

[0088] For example, the data management system 100 can search for non-standard data (cause of trouble, details of trouble, etc.) corresponding to a trouble based on a trouble code, information on the process in which the trouble occurred, etc.

[0089] The data processing unit 17 organizes and analyzes the structured data and unstructured data stored in the database group 30, and learns the unstructured data corresponding to the structured data. The data processing unit 17 transmits a learning model, which is the learning result, to the utilization device 40. Details of the data processing unit 17 will be described later.

[0090] The production management device 41 has a production schedule adjustment unit 410, which is equipped with an AI support unit (learning model operation unit) 411. The production schedule adjustment unit 410 adjusts the production schedule based on the progress of the production process and the production schedule.

[0091] The AI ​​support unit 411 supports the adjustment of the production schedule using a learning model (hereinafter sometimes referred to as a schedule adjustment learning model) received from the data management device 1. The schedule adjustment learning model used by the AI ​​support unit 411 is a learning model that infers messages related to the production schedule from standard data such as trouble codes and process names. For example, the AI ​​support unit 411 uses the schedule adjustment learning model to infer work instruction messages (messages corresponding to standard data) for adjusting the production schedule. The production schedule adjustment unit 410 notifies the worker of the inferred messages.

[0092] The production management device 41 is deployed at the production site without having a schedule adjustment learning model. Thereafter, the production management device 41 receives the schedule adjustment learning model from the data management device 1 and stores it in the production schedule adjustment unit 410, which then becomes equipped with the AI ​​support unit 411. The production management device 41 may be deployed at the production site with a schedule adjustment learning model. In this case, the schedule adjustment learning model is updated at the production site as needed.

[0093] The maintenance support device 42 has a maintenance support unit 420, which is equipped with an AI support unit 421. The maintenance support unit 420 provides maintenance support to workers at the production site via chat.

[0094] The AI ​​support unit 421 provides maintenance support via chat using a learning model (hereinafter sometimes referred to as a maintenance support learning model) received from the data management device 1. The maintenance support learning model used by the AI ​​support unit 421 is a learning model that infers messages (answers to questions) related to maintenance support from standard data such as trouble codes and process names. For example, the AI ​​support unit 421 uses the maintenance support learning model to infer a message to be output when notifying the cause of a trouble. The maintenance support unit 420 notifies the worker of the inferred message.

[0095] The maintenance support device 42 is placed at the production site without having a maintenance support learning model. Thereafter, the maintenance support device 42 receives the maintenance support learning model from the data management device 1 and stores it in the maintenance support unit 420, so that the maintenance support unit 420 is equipped with an AI support unit 421. The maintenance support device 42 may be placed at the production site in a state where it has a maintenance support learning model. In this case, the maintenance support learning model is updated at the production site as needed.

[0096] The anomaly detection device 43 has an anomaly detection unit 430, which is equipped with an AI support unit 431. The anomaly detection unit 430 determines and detects an anomaly in the production process based on sensor values, timer values, etc.

[0097] The AI ​​support unit 431 detects abnormalities in the production process using a learning model (hereinafter sometimes referred to as an anomaly detection learning model) received from the data management device 1. The anomaly detection learning model used by the AI ​​support unit 431 is a learning model that infers messages related to anomaly detection from standard data such as trouble codes and process names. For example, the AI ​​support unit 431 uses the anomaly detection learning model to infer messages to be output when an abnormality is detected and messages to be output when notifying the occurrence status of the abnormality (such as the frequency of occurrence). The anomaly detection unit 430 notifies the worker of the inferred messages.

[0098] The anomaly detection device 43 is deployed at the production site without having an anomaly detection learning model. Thereafter, the anomaly detection device 43 receives the anomaly detection learning model from the data management device 1 and stores it in the anomaly detection unit 430, thereby providing the anomaly detection unit 430 with an AI support unit 431. The anomaly detection device 43 may be deployed at the production site with an anomaly detection learning model. In this case, the anomaly detection learning model is updated at the production site as needed.

[0099] The predictive maintenance device 44 has a predictive maintenance unit 440, which is equipped with an AI support unit 441. The predictive maintenance unit 440 predicts abnormalities in the production process at the production site.

[0100] The AI ​​support unit 441 predicts anomalies in the production process using a learning model (hereinafter sometimes referred to as an anomaly prediction learning model) received from the data management device 1. The anomaly prediction learning model used by the AI ​​support unit 441 is a learning model that infers messages related to anomaly prediction from standard data such as trouble codes and process names. For example, the AI ​​support unit 441 uses the anomaly prediction learning model to infer a message to be output when an anomaly is predicted. The predictive maintenance unit 440 notifies the worker of the inferred message.

[0101] The predictive maintenance device 44 is deployed at the production site without having an anomaly prediction learning model. Thereafter, the predictive maintenance device 44 receives the anomaly prediction learning model from the data management device 1 and stores it in the predictive maintenance unit 440, thereby providing the predictive maintenance unit 440 with an AI support unit 441. The predictive maintenance device 44 may be deployed at the production site with an anomaly prediction learning model. In this case, the anomaly prediction learning model is updated at the production site as needed.

[0102] 3 is a diagram illustrating an example of the data structure of data stored in a database by a data management device according to an embodiment. The data stored in the database group 30 by the data management device 1 includes structured data D1 and unstructured data d2, and the structured data D1 and the unstructured data d2 are associated with each other.

[0103] For example, when a problem occurs, the data management device 1 receives information indicating the problem (trouble situation) from the production equipment 5 or the like. In this case, the data management device 1 stores the trouble situation as structured data in the database group 30. In this case, when the trouble situation is notified to a user (operator) via a management screen of the production equipment 5 or the like, the operator inputs a message such as "The cause of the problem is a clogged nozzle pipe" into the operation terminal 6. The operation terminal 6 transmits the message input by the operator to the data management device 1. The data management device 1 stores the message (unstructured data) received from the operation terminal 6 in the database group 30 in association with the trouble situation (structured data) received from the production equipment 5. Information on the date and time when the problem occurred, information indicating the process in which the problem occurred, and the like, contained in the structured data D1, are associated with a message such as "The cause of the problem is a clogged nozzle pipe" contained in the unstructured data D1. Therefore, for example, when searching for the trouble code 264, it becomes easier to find the free-text cause of the problem.

[0104] The standard data D1 is, for example, the trouble occurrence information 22 described in Fig. 1. That is, the standard data D1 includes, for example, information on the date and time when the trouble occurred, information indicating the process in which the trouble occurred, the trouble code, the details of the trouble, the number of processes performed by the device that caused the trouble, etc. The unstandardized data d2 is, for example, unstandardized data acquired from the operation terminal 6, the production equipment 5, the utilization device 40, etc.

[0105] The data management system 100 stores the structured data D1 and the unstructured data d2 in association with each other. Therefore, if a problem occurs later, the stored unstructured data can be efficiently utilized. For example, if a problem occurs at a production site, the data management system 100 can efficiently utilize the cause of the problem and the corresponding troubleshooting based on the process where the problem occurred and the trouble code. The data management system 100 can also provide messages based on the number of applications. For example, if the predictive maintenance device 44 has an anomaly prediction learning model based on training data that combines a processing count of 1,600 or more and device shutdowns, when a problem occurs in the fourth process and the processing count exceeds 1,600, the data management system 100 can notify the worker with a message such as, "The processing count for the fourth process has exceeded 1,600. Continuing the process may result in a trouble-related shutdown." Furthermore, if the anomaly detection device 43 has an anomaly detection learning model based on learning data that combines the normal response time in the fourth step and the frequency with which that response time is exceeded, the data management system 100 can notify the worker of a message such as, "Time-over troubles are occurring frequently in the fourth step."

[0106] The data processing unit 17 performs data organization, data analysis, and data learning on the data stored in the database group 30. The data organization performed by the data processing unit 17 is a process (first process described later) of categorizing unstructured data relative to structured data into major categories, and the data analysis performed by the data processing unit 17 is a process (second process described later) of categorizing the unstructured data categorized into major categories into medium categories. Note that the data processing unit 17 may categorize the unstructured data relative to structured data into minor categories, or may categorize the unstructured data into categories smaller than the minor categories.

[0107] The data processing unit 17 stores the categorized combinations of the regular data and the unregular data in the database group 30. The data processing unit 17 may also transmit the categorized combinations of the regular data and the unregular data to an external device such as the operation terminal 6.

[0108] The data processing unit 17 uses the standard data as an index to acquire multiple pieces of unstructured data from the design DB 31, the trouble occurrence DB 32, and the trouble response DB 33, and merges the acquired unstructured data. That is, the data processing unit 17 uses the standard data as an index to search for unstructured data from the database group 30, and merges the searched unstructured data. The standard data that the data processing unit 17 uses as an index includes one or more data items (date and time, process name, equipment name, trouble code, etc.).

[0109] For example, when organizing unstructured data for each process, the data processing unit 17 searches the databases 30 for unstructured data associated with a process (e.g., the fourth process), and merges and organizes the retrieved unstructured data as unstructured data related to the process. For example, the data processing unit 17 searches the databases 30 for unstructured data using the fourth process as an index, and merges the retrieved unstructured data for the fourth process.

[0110] Furthermore, when organizing non-standard data corresponding to a trouble, the data processing unit 17 searches for the non-standard data from the databases 30 using the trouble code, date and time, and process as indexes, and merges the retrieved non-standard data. In this way, the data processing unit 17 searches for non-standard data associated with the trouble code, date and time, and process, and organizes it as non-standard data related to the trouble code, date and time, and process.

[0111] The data processing unit 17 organizes data according to the purpose of data analysis set by the user. To this end, the data processing unit 17 reads and organizes data according to the purpose of data analysis from all or part of the design DB 31, the trouble occurrence DB 32, and the trouble response DB 33.

[0112] For example, the data processing unit 17 associates non-standard data (e.g., voice) with standard data (e.g., trouble codes) that have already been stored in the database group 30 after the fact. Fields into which non-standard data are input are set in the database group 30, and the data processing unit 17 analyzes the content of the non-standard data stored in these fields and associates the non-standard data with standard data in fields into which appropriate standard data corresponding to the non-standard data has been input.

[0113] The data processing unit 17 also analyzes the organized unstructured data according to the purpose of the data analysis. For the categorized combinations of structured data and unstructured data, the data processing unit 17 categorizes the combinations of structured data and unstructured data into more detailed categories based on the structured data, and performs analysis according to the category items.

[0114] For example, when the data processing unit 17 searches for and merges unstructured data associated with the fourth step (structured data), the data processing unit 17 selects and stylizes a plurality of unstructured data that can be classified into a higher concept from the merged unstructured data. For example, the data processing unit 17 selects and stylizes unstructured data that have common management items from the merged unstructured data.

[0115] The management items are objects of management managed by the data management device 1, and are, for example, trouble items, productivity items, quality items, etc. In other words, examples of standardized data, which is data that the data processing unit 17 standardizes by conceptualizing unstandardized data, are trouble items, productivity items, quality items, etc.

[0116] A trouble item is a data item related to trouble, a productivity item is a data item related to productivity, and a quality item is a data item related to quality. Examples of trouble items include transportation trouble, processing trouble, and packaging trouble.

[0117] The data processing unit 17 associates the selected unstructured data with the structured data, which is data that has been generalized and structured (fourth step), and the original structured data (fourth step). This enables the data processing unit 17 to categorize the combinations of structured data and unstructured data categorized into major categories into medium categories.

[0118] The purpose of the data analysis corresponds to the use of the data desired by the user. The data processing unit 17 receives the purpose of the data analysis from the operation terminal 6. Note that in FIG. 2, the connection line between the data processing unit 17 and the operation terminal 6 is omitted.

[0119] When analyzing unstructured data, the data processing unit 17 analyzes the unstructured data, for example, based on the behavior of workers at the production site when a problem occurs. For example, the data processing unit 17 analyzes what work should be performed with priority when a problem occurs, and the work time for that work, as an objective index, based on the behavior of workers at the production site when a problem occurs. In this way, the data processing unit 17 analyzes what work should be performed with priority based on the worker's knowledge when a problem occurs, and what work results (work time) can be obtained.

[0120] The actions of workers at the production site when a problem occurs include, for example, instructions on the location to check first for the problem, confirmation work to identify the cause of the problem if an abnormality is found in the checked location, and effective methods (trouble response) implemented to solve the problem once the cause of the problem has been identified. The data processing unit 17 judges the actions of the worker based on the voice uttered by the worker, the information input by the worker to the production equipment 5 or the operation terminal 6, etc., and analyzes the unstructured data based on the judgment result.

[0121] For example, the data processing unit 17 organizes know-how (procedures, etc.) for dealing with troubles for each cause of trouble. That is, the data processing unit 17 categorizes the causes of troubles, which are non-standard data, and associates each cause of trouble with a corresponding trouble response.

[0122] The data processing unit 17 also organizes know-how (procedures, work points, etc.) according to the purpose, such as productivity improvement in the manufacturing industry, quality improvement in the manufacturing industry, employee development, and employee education.

[0123] For example, the data processing unit 17 categorizes the non-standard data into categories such as non-standard data related to productivity improvement, non-standard data related to quality improvement, non-standard data related to employee development, and non-standard data related to employee education.

[0124] Furthermore, the data processing unit 17 categorizes the unstructured data related to productivity improvement into categories such as unstructured data that improves the productivity of a first product, unstructured data that improves the productivity of a second product, ..., unstructured data that improves the productivity of an Nth product (N is a natural number). Then, the data processing unit 17 associates each of the categorized unstructured data with know-how for improving productivity.

[0125] Furthermore, the data processing unit 17 categorizes the non-standard data related to employee development into categories such as non-standard data for achieving a first development objective, non-standard data for achieving a second development objective, ..., non-standard data for achieving an Mth development objective (M is a natural number).The data processing unit 17 then associates each of the categorized non-standard data with know-how for achieving the development objectives.

[0126] Furthermore, the data processing unit 17 learns combinations of structured data and unstructured data that have been subjected to data organization or data analysis. The data processing unit 17 acquires the combinations of structured data and unstructured data as training data, and uses the training data to generate a training model for inferring unstructured data from structured data.

[0127] The data processing unit 17 generates a schedule adjustment learning model, a maintenance support learning model, an anomaly detection learning model, and an anomaly prediction learning model. For example, when a malfunction occurs, the data processing unit 17 learns what work instructions a worker should execute based on the worker's knowledge in response to the malfunction, and what work results (work time) can be obtained.

[0128] The data processing unit 17 transmits the schedule adjustment learning model to the production management device 41, and transmits the maintenance support learning model to the maintenance support device 42. In addition, the data processing unit 17 transmits the anomaly detection learning model to the anomaly detection device 43, and transmits the anomaly prediction learning model to the predictive maintenance device 44.

[0129] As a result, the production management device 41 uses the schedule adjustment learning model to infer a work instruction message for adjusting the production schedule from standard data of the production process (progress of the production process, production schedule, etc.), and transmits it to the operation terminal 6, etc. Then, the operation terminal 6 outputs (audio output, display, etc.) the message received from the production management device 41.

[0130] Furthermore, the maintenance support device 42 uses the maintenance support learning model to infer answers to questions input by the worker, and transmits the answers to the operation terminal 6, etc. Then, the operation terminal 6 outputs the answers received from the maintenance support device 42.

[0131] Furthermore, the anomaly detection device 43 uses the anomaly detection learning model to infer a message to be output when an anomaly is detected from the standard data when the anomaly is detected, and transmits the message to the operation terminal 6, etc. Then, the operation terminal 6 outputs the message received from the anomaly detection device 43.

[0132] Furthermore, the predictive maintenance device 44 uses the anomaly prediction learning model to infer a message to be output when an anomaly is predicted from the standard data when the anomaly is predicted, and transmits the message to the operation terminal 6, etc. Then, the operation terminal 6 outputs the message received from the anomaly detection device 43.

[0133] This allows the data management system 100 to issue work instructions in an appropriate language according to the acquired standard data (for example, the trouble occurrence information 22).

[0134] The data management device 1 does not necessarily have to include the routine data collection unit 14, the routine data analysis unit 15, and the routine data placement unit 16. In this case, the production site device 20 is provided with the functions of the routine data collection unit 14, the routine data analysis unit 15, and the routine data placement unit 16. The production site device 20 collects routine data, analyzes in which area of ​​the database group 30 each routine data should be stored, and stores the routine data in the analyzed area. The data processing unit 17 may be located outside the data management device 1. In this case, the data management device 1 is connected to the data processing unit 17, which is a data processing device. Furthermore, the data processing unit 17, which is a data processing device, is connected to the utilization device 40.

[0135] Here, the process of associating structured data with unstructured data and the process of learning the correspondence between structured data and unstructured data will be described. Fig. 4 is a diagram for explaining the association process executed by the data management device according to the embodiment. The process on the left side of Fig. 4 is the first process (data organization), and the process on the right side is the second process (data analysis). The first process is categorizing major items, and the second process is categorizing medium items.

[0136] In the first processing, the data processing unit 17 associates the structured data (one or more data items) stored in the database group 30 with the plurality of unstructured data stored in the database group 30. For example, in the first processing, the data processing unit 17 associates the unstructured data with the structured data D10 of a specific process.

[0137] 4 shows a case where the standard data D10 is the fourth step. The unstructured data corresponding to the standard data D10 of the fourth step includes, for example, trouble content, troubleshooting, and the results of the troubleshooting. 4 shows a case where the standard data D10 of the fourth step is associated with unstructured data da1 to da3 indicating trouble content, unstructured data db1 to db3 indicating troubleshooting, and unstructured data dc indicating the results of the troubleshooting.

[0138] The non-standard data da1 indicating the details of the trouble is, for example, a message indicating, "There is a possibility that trouble AA will occur." The non-standard data db1 indicating how to deal with the trouble is, for example, a message indicating, "If trouble AA occurs, it is better to take action BB."

[0139] For example, for the fourth step, the data processing unit 17 may notify the worker of the unstructured data corresponding to the fourth step. In this case, the data processing unit 17 provides the worker of the fourth step with a message such as, "There is a possibility that trouble AA may occur in the fourth step. If trouble AA occurs, it is better to take action BB." The data processing unit 17 provides the message to the worker by transmitting the message to the operation terminal 6 or the like.

[0140] 4 shows a case where the trouble content is set in unstructured data, but the trouble content may be set in structured data. The unstructured data may also include the cause of the trouble.

[0141] Furthermore, in the second process, the data processing unit 17 associates the non-standard data with the trouble item of the fourth step. The trouble item is, for example, a transport trouble. The trouble item may also be a processing trouble, a packaging trouble, etc. The trouble item is data in which the non-standard data (trouble content) is generalized into a higher-level concept.

[0142] The merged unstructured data may include various types of unstructured data, such as unstructured data corresponding to trouble, unstructured data corresponding to productivity, and unstructured data corresponding to quality. The data processing unit 17, for example, selects and stylizes unstructured data that have common item contents from the merged unstructured data. In other words, the data processing unit 17 selects and stylizes multiple unstructured data that can be classified into a higher concept from the unstructured data merged in the first process.

[0143] The data processing unit 17 selects and stylizes unstructured data that have common trouble items from the merged unstructured data, for example. Figure 4 shows a case where unstructured data da2 and da3 are unstructured data that can be used to superordinately conceptualize the trouble content. The data processing unit 17 superordinately conceptualizes and stylizes the unstructured data da2 and da3, and associates them with the original structured data D10.

[0144] In FIG. 4, when the data that has been generalized by conceptualizing the non-standard data da2 and da3 is a trouble item (such as a transportation trouble), the data processing unit 17 associates the non-standard data corresponding to the standard data D10 and the standard data D20 indicating the trouble item with the standard data D10 and D20.

[0145] The data processing unit 17 skips unstructured data da1, db2, etc. that do not correspond to the generalized and standardized standard data D20, and associates the unstructured data da2, da3, db1, db3, and dc that correspond to the standard data D20 with the standard data D20. This increases the correlation of troubleshooting to the standard data.

[0146] For example, for a trouble item in the fourth process, the data processing unit 17 may notify a worker (such as the operation terminal 6) of non-standard data corresponding to the trouble item in the fourth process. In this case, the data processing unit 17 provides the worker with a message corresponding to the trouble item in the fourth process. For example, the data processing unit 17 provides the worker in the fourth process with a message such as, "If a transport trouble occurs in the fourth process, it is better to take CC action." The data processing unit 17 provides the worker with the message by sending the message to the operation terminal 6 or the like.

[0147] Furthermore, the data processing unit 17 learns the characteristics of the fourth step for the first process, and learns the characteristics of problems in the fourth step for the second process. The data processing unit 17 learns the characteristics of the fourth step for the first process using the unstructured data da1 to da3, db1 to db3, and dc. That is, the data processing unit 17 learns the correspondence between the structured data D10 and the unstructured data da1 to da3, db1 to db3, and dc.

[0148] For example, the structured data D10 may include multiple data items such as date and time, process name, equipment name, and trouble code. In this case, the data processing unit 17 learns the correspondence between the structured data including multiple types of data items and multiple unstructured data. The data processing unit 17 generates a learning model for inferring how to handle the trouble corresponding to the fourth process and transmits it to the utilization device 40.

[0149] Furthermore, the data processing unit 17 learns the characteristics of the trouble item in the fourth step for the second process. The data processing unit 17 learns the characteristics of the trouble item in the fourth step for the second process using the unstructured data da2, da3, db1, db3, and dc. That is, the data processing unit 17 learns the correspondence between the structured data D10 and D20 and the unstructured data da2, da3, db1, db3, and dc. This allows the data processing unit 17 to learn how to deal with the trouble item in the fourth step while taking into account the characteristics of the fourth step.

[0150] The data processing unit 17 generates a learning model for inferring a troubleshooting response corresponding to the trouble item in the fourth step and transmits it to the utilization device 40. This learning model, for example, infers a troubleshooting response corresponding to the trouble item in the fourth step. The learning model generated by the data processing unit 17 for the first process is the first learning model, and the learning model generated by the data processing unit 17 for the second process is the second learning model.

[0151] When the utilization device 40 infers non-standard data using the learning model generated for the first process, it infers a message such as, for the fourth step, "Since there is a possibility that an AA problem may occur in the fourth step, if an AA problem occurs, it is better to take the BB response."

[0152] Furthermore, when the utilization device 40 infers non-standard data using the learning model generated for the second process, it infers specific instructions such as, for example, for a trouble item in the fourth step, "For trouble AA in the fourth step, take action BB."

[0153] The data processing unit 17 may perform a third process on the combination of the typical data and the untypical data associated in the second process, thereby categorizing the combination of the typical data and the untypical data into smaller categories. The data processing unit 17 may associate the untypical data with, for example, a transport trouble due to a temperature abnormality. The data processing unit 17 may also associate the untypical data with, for example, a transport trouble due to a current abnormality or a transport trouble due to a voltage abnormality. For example, similar to the second process, the data processing unit 17 generates typical data (e.g., typical data D30) that superordinates the atypical data (trouble content), and associates the atypical data corresponding to the typical data D10, D20, and D30 with the typical data D10, D20, and D30. The data processing unit 17 also generates a learning model for the third process using a process similar to the second process.

[0154] 5 is a flowchart showing the procedure of a storage process in which the data management device according to the embodiment stores data in a database. The data management device 1 collects standard data from the production site devices 20 and stores the data in the database group 30 (step S10).

[0155] After that, when a problem occurs, the data management device 1 collects unstructured data at the time of the problem from the production site device 20 or the utilization device 40 (step S20). The data management device 1 associates the structured data with the unstructured data (step S30). The data management device 1 stores the unstructured data associated with the structured data in the database group 30 (step S40). The structured data associated with the unstructured data becomes a data management item in the database group 30.

[0156] This allows the data management device 1 to efficiently utilize unstructured data. For example, the data management device 1 can search for unstructured data such as troubleshooting information by using structured data such as trouble codes as an index.

[0157] 6 is a flowchart showing the processing steps of a utilization process in which the data management device according to the embodiment utilizes unstructured data in a database. The data processing unit 17 of the data management device 1 learns the correspondence between the structured data and the unstructured data stored in the database group 30 (step S50). Specifically, the data processing unit 17 learns the correspondence between the structured data and the unstructured data, thereby generating a learning model for inferring the unstructured data from the structured data. The data processing unit 17 transmits the inferred learning model to the utilization device 40.

[0158] When a new problem occurs, the utilization device 40 outputs to the utilization device 40 unstructured data corresponding to the structured data of the newly occurred problem (step S60). Specifically, the utilization device 40 inputs the structured data of the newly occurred problem into the learning model. As a result, the learning model outputs unstructured data (inference result) corresponding to the structured data. The utilization device 40 outputs the unstructured data (instructions for dealing with the problem, etc.) that is the inference result to the operation terminal 6, etc., thereby notifying the worker of the unstructured data corresponding to the problem.

[0159] Next, we will explain the learning process and inference process executed by the data management system 100. In the data management system 100, a learning device is arranged in the data processing unit 17 of the data management device 1, and an inference device is arranged in each device within the utilization device 40.

[0160] 7 is a diagram showing the configuration of a learning device provided in a data management system according to an embodiment. The learning device 50 is connected to a learning model storage unit 55. The learning model storage unit 55 may be disposed inside the data management device 1 or may be disposed inside the utilization device 40. The learning model storage unit 55 may also be disposed outside the data management device 1 and the utilization device 40.

[0161] The learning device 50 has a data acquisition unit 51 and a model generation unit 52. The data acquisition unit 51 acquires structured data 71A from the database group 30. The data acquisition unit 51 also acquires unstructured data 72A from the database group 30. The data acquisition unit 51 is a first data acquisition unit.

[0162] The model generation unit 52 learns appropriate unstructured data 72A corresponding to the structured data 71A, based on training data created based on a combination of the structured data 71A and the unstructured data 72A sent from the data acquisition unit 51. In other words, the model generation unit 52 learns appropriate unstructured data 72A based on the training data when the structured data 71A is satisfied. That is, the model generation unit 52 generates a training model M1 that infers appropriate unstructured data 72A from the structured data 71A. Here, the training data is data in which the structured data 71A and the unstructured data 72A are associated with each other.

[0163] The training data used by the model generation unit 52 when learning how to categorize unstructured data into major categories relative to structured data is the first training data. The training data used by the model generation unit 52 when learning how to categorize unstructured data into medium categories relative to structured data is the second training data.

[0164] The learning device 50 is used to learn the unstructured data 72A corresponding to the structured data 71A, but may be a device separate from the data management device 1 and connected to the data management device 1 via a network. The learning device 50 may also be located in the utilization device 40. Furthermore, the learning device 50 may exist on a cloud server.

[0165] The model generation unit 52 can use known algorithms such as supervised learning, unsupervised learning, reinforcement learning, etc. As an example, a case where a neural network is applied to the learning algorithm used by the model generation unit 52 will be described.

[0166] The model generation unit 52 learns appropriate unstructured data 72A corresponding to the structured data 71A by so-called supervised learning, for example, according to a neural network model. Here, supervised learning refers to a technique in which data pairs (learning data) of inputs and results (labels) are provided to the learning device 50, whereby the learning device 50 learns the features contained in the learning data and infers the results from the inputs.

[0167] A neural network consists of an input layer consisting of multiple neurons, an intermediate layer (hidden layer) consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer may be one layer or two or more layers.

[0168] Fig. 8 is a diagram illustrating a neural network used by a learning device according to an embodiment. For example, in a three-layer neural network as shown in Fig. 8, when multiple inputs are input to input layers X1 to X3, the input values ​​are multiplied by weights w11 to w16 and input to intermediate layers Y1 and Y2, and the results are further multiplied by weights w21 to w26 and output from output layers Z1 to Z3. The output results vary depending on the values ​​of the weights w11 to w16 and w21 to w26.

[0169] 7 learns atypical data 72A corresponding to typical data 71A through so-called supervised learning in accordance with learning data created based on a combination of typical data 71A and atypical data 72A acquired by a data acquisition unit 51. In other words, the neural network used by the learning device 50 in Fig. 7 learns atypical data 72A corresponding to typical data 71A through so-called supervised learning in accordance with typical data 71A and atypical data 72A created based on a combination of a first input and a second input (correct answer) acquired by the data acquisition unit 51.

[0170] That is, when the neural network receives the first input, that is, the structured data 71A, as input to the input layers X1 to X3, it adjusts the weights w11 to w16 and w21 to w26 so that the results output from the output layers Z1 to Z3 approach the structured data 72A, thereby learning the structured data. The model generation unit 52 generates a learning model M1 by executing the above-described learning. The learning device 50 stores the neural network with the adjusted weights w11 to w16 and w21 to w26 as the learning model M1 in the learning model storage unit 55. The learning model storage unit 55 stores the learning model M1 output from the model generation unit 52.

[0171] Next, a processing procedure of the learning process performed by the learning device 50 will be described with reference to Fig. 9. Fig. 9 is a flowchart showing the processing procedure of the learning process performed by the learning device according to the embodiment.

[0172] The data acquiring unit 51 acquires learning data to be used for learning (step S110). Specifically, the data acquiring unit 51 acquires the structured data 71A and the unstructured data 72A corresponding to the structured data 71A.

[0173] The model generation unit 52 executes a learning process using the standard data 71A and the unstandard data 72A (step S120). Specifically, the model generation unit 52 learns the unstandard data 72A corresponding to the standard data 71A by so-called supervised learning in accordance with learning data created based on a combination of the standard data 71A and the unstandard data 72A acquired by the data acquisition unit 51, and generates a learning model M1.

[0174] After generating the learning model M1, the model generation unit 52 outputs the learning model M1 to the learning model storage unit 55 (step S130). The learning model storage unit 55 stores the learning model M1 generated by the model generation unit 52.

[0175] The inference devices arranged in each device within the utilization device 40 have the same configuration and execute the same processing. FIG. 10 is a diagram showing the configuration of the inference device provided in the data management system according to the embodiment. The inference device 60 has a data acquisition unit 61 and an inference unit 62. The inference device 60 is connected to a learning model storage unit 55.

[0176] The data acquisition unit 61 acquires the standard data 71B from outside (such as a production site) of the inference device 60. The standard data 71B is the same information as the standard data 71A.

[0177] The inference unit 62 receives the structured data 71B sent from the data acquisition unit 61. The inference unit 62 also reads out the learning model M1 from the learning model storage unit 55. The inference unit 62 uses the learning model M1 to infer unstructured data 72B corresponding to the structured data 71B. That is, the inference unit 62 inputs the structured data 71B acquired by the data acquisition unit 61 into the learning model M1, thereby being able to output appropriate unstructured data 72B inferred from the structured data 71B. The data acquisition unit 61 is a second data acquisition unit.

[0178] The inference device 60 is used to infer unstructured data 72B corresponding to structured data 71B, but may be a device separate from the utilization device 40 and connected to the utilization device 40 via a network. The inference device 60 may also be located in the data management device 1. The inference device 60 may also reside on a cloud server.

[0179] Here, we have described the case where the inference device 60 outputs unstructured data 72B using the learning model M1 learned by the model generation unit 52 of the data management device 1, but the inference device 60 may also obtain the learning model M1 from an external device such as another data management device and output the unstructured data 72B based on this learning model M1.

[0180] For example, the learning device 50 generates a learning model M1 using training data including standard data, trouble information (such as the cause of the trouble and how to deal with the trouble) corresponding to the standard data, and trouble items. In this case, the inference device 60 acquires the standard data and the trouble items corresponding to the standard data, and uses the learning model M1 to infer trouble information corresponding to the standard data and the trouble items.

[0181] Next, a processing procedure of the inference device 60 inferring the unstructured data 72B using the learning model M1 will be described with reference to Fig. 11. Fig. 11 is a flowchart showing the processing procedure of the inference process executed by the inference device according to the embodiment.

[0182] The data acquiring unit 61 acquires inference data to be used for inference of the unstructured data 72B (step S210). Specifically, the data acquiring unit 61 acquires structured data 71B from a production site or the like. The data acquiring unit 61 sends the structured data 71B, which is data for inference, to the inference unit 62. The inference unit 62 acquires the structured data 71B from the data acquiring unit 61 and acquires the learning model M1 from the learning model storage unit 55.

[0183] The inference unit 62 inputs the standard data 71B into the learning model M1 (step S220) to obtain appropriate unstandard data 72B. That is, the inference unit 62 infers the unstandard data 72B from the standard data 71B acquired by the data acquisition unit 61, using the learning model M1 for inferring the unstandard data 72B from the standard data 71B.

[0184] The inference unit 62 outputs the unstructured data 72B inferred using the learning model M1 and the structured data 71B (step S230). Specifically, the inference unit 62 outputs the unstructured data 72B obtained by the learning model M1 to an external device such as a display device (not shown).

[0185] The display device displays the unstructured data 72B corresponding to the structured data 71B (step S240), thereby enabling the worker to refer to the unstructured data 72B corresponding to the structured data 71B.

[0186] In the present embodiment, an example has been described in which the model generation unit 52 uses a supervised learning algorithm as a learning algorithm, but the learning algorithm used by the model generation unit 52 is not limited to a supervised learning algorithm. The model generation unit 52 can also apply a reinforcement learning algorithm, an unsupervised learning algorithm, a semi-supervised learning algorithm, etc., in addition to the supervised learning algorithm.

[0187] Deep learning, which learns to extract features themselves, can also be used as the learning algorithm of the model generation unit 52. The model generation unit 52 may also perform machine learning according to other known methods, such as genetic programming, functional logic programming, or support vector machines.

[0188] Furthermore, the model generation unit 52 may learn unstructured data 72A corresponding to the structured data 71A according to learning data created by a plurality of data management devices. Furthermore, the model generation unit 52 may acquire learning data from a plurality of data management devices used in the same area, or may learn unstructured data 72A corresponding to the structured data 71A using learning data collected from a plurality of data management devices operating independently in different areas.

[0189] In addition, a data management device that collects learning data may be added to or removed from the targets during the process. Furthermore, the learning device 50 that has learned the unstructured data 72A corresponding to the structured data 71A may be applied to another data management device, and the unstructured data 72A corresponding to the structured data 71A may be re-learned and updated for the other data management device.

[0190] In the embodiment, the case where the data management device 1 collects and manages standard data and unstandard data from the production site device 20 and allows the utilization device 40 to utilize the data has been described, but the data management device 1 may also manage unstandard data collected from the utilization device 40.

[0191] Specifically, the data management device 1 associates the unstructured data collected from the utilization device 40 with structured data related to this unstructured data and stores them in a DB (DataBase). For example, the data management device 1 associates at least one of the cause of the trouble and the content of the trouble included in the trouble response information 23 with the design information 21 and the trouble occurrence information 22.

[0192] Furthermore, the data management device 1 associates a message (non-standard data) such as "A time-over problem has occurred in the applicator C1 of the fourth process. Please tell us the cause." with design information 21, which is standard data corresponding to this message, and stores them in the database group 30. The design information 21 corresponding to the message "A time-over problem has occurred in the applicator C1 of the fourth process. Please tell us the cause." includes information indicating the "fourth process" and information indicating the "applicator C1." Note that non-standard data is not limited to natural conversational language data, and may be any data that does not have a standardized structure.

[0193] Next, we will explain the hardware configuration of the data management device 1. The data management device 1 is realized by a processing circuit. The processing circuit may be a processor and memory that executes a program stored in a memory, or may be dedicated hardware.

[0194] FIG. 12 illustrates an exemplary configuration of a processing circuit included in a control circuit of a data management device according to an embodiment, where the processing circuit is implemented by a processor and a memory. The processing circuit 90 illustrated in FIG. 12 includes a processor 91 and a memory 92. When the processing circuit 90 includes the processor 91 and the memory 92, each function of the processing circuit 90 is implemented by software, firmware, or a combination of software and firmware. The software or firmware is written as a data management program and stored in the memory 92. The processing circuit 90 implements each function by having the processor 91 read and execute the data management program stored in the memory 92. That is, the processing circuit 90 includes the memory 92 for storing the data management program that results in the processing of the data management device 1. The data management program can also be considered a program that causes the data management device 1 to execute each function realized by the processing circuit 90. The data management program can be provided by a computer-readable storage medium storing the data management program, or by other means such as a communication medium. The data management program can also be considered a program that causes the data management device 1 to execute the processing of steps S10 to S40 of FIG. 5. The data management program may cause the data management device 1 to execute the processes of steps S50 and S60 in Fig. 6. The data management program may also cause the data management device 1 to execute the processes of steps S110 to S130 in Fig. 9.

[0195] Here, the processor 91 is, for example, a CPU (Central Processing Unit), a processing device, an arithmetic device, a microprocessor, a microcomputer, or a DSP (Digital Signal Processor), etc. Furthermore, the memory 92 is, for example, a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable ROM), or an EEPROM (registered trademark) (Electrically EPROM), a magnetic disk, a flexible disk, an optical disk, a compact disk, a minidisk, or a DVD (Digital Versatile Disc).

[0196] 13 is a diagram illustrating an example of the configuration of a processing circuit provided in a control circuit of a data management device according to an embodiment, where the processing circuit is configured with dedicated hardware. The processing circuit 93 illustrated in FIG. 13 corresponds to, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. The processing circuit 93 may be partially implemented with dedicated hardware and partially implemented with software or firmware. In this way, the processing circuit 93 can realize each of the above-described functions by dedicated hardware, software, firmware, or a combination thereof.

[0197] The hardware configurations of the learning device 50 and the inference device 60 may be the hardware configurations described in Fig. 12 or 13. The hardware configuration of the data processing unit 17 may be the hardware configurations described in Fig. 12 or 13. The hardware configurations of the production management device 41, the maintenance support device 42, the anomaly detection device 43, and the predictive maintenance device 44 may be the hardware configurations described in Fig. 12 or 13.

[0198] In this way, the data management system 100 associates and stores the structured data with the unstructured data as management items, making it possible to easily search for unstructured information by taking advantage of the ease of searching for structured information. This allows the data management system 100 to shorten the processing time for data processing (data organization, data analysis, data learning) of unstructured data.

[0199] Furthermore, the data management system 100 associates unstructured data with structured data and stores them, allowing for the aggregation of scattered unstructured data such as language data. Furthermore, the data management system 100 organizes and analyzes the unstructured data associated with structured data, allowing for the acquisition of high-quality unstructured data related to the structured data when searching for unstructured data using the structured data as an index.

[0200] Furthermore, since the data management system 100 can search for unstructured data corresponding to structured data, workers can refer to unstructured data (such as troubleshooting) for past cases (such as problems). This allows workers to quickly respond to problems and reduce time lost due to problems.

[0201] In addition, the data management system 100 learns from a combination of unstructured data and structured data as training data to generate a learning model M1, which can be used for production management, maintenance support, anomaly detection, and predictive maintenance using AI.

[0202] As described above, according to the embodiment, unstructured data and structured data as management items associated with the unstructured data are stored in the database group 30, so that the unstructured data can be utilized efficiently.

[0203] The configurations shown in the above embodiments are merely examples, and may be combined with other known technologies, and parts of the configurations may be omitted or modified without departing from the spirit of the invention. [Explanation of symbols]

[0204] 1 Data management device, 5 Production equipment, 6 Operation terminal, 11 Non-structured data collection unit, 12 Non-structured data analysis unit, 13 Non-structured data placement unit, 14 Structured data collection unit, 15 Structured data analysis unit, 16 Structured data placement unit, 17 Data processing unit, 20 Production site device, 21 Design information, 22 Trouble occurrence information, 23 Trouble response information, 30 Database group, 31 Design DB, 32 Trouble occurrence DB, 33 Trouble response DB, 40 Utilization device, 41 Production management device, 42 Maintenance support device, 43 Anomaly detection device, 44 Predictive maintenance device, 50 Learning device, 51, 61 Data acquisition unit, 52 Model generation unit, 55 Learning model memory unit, 60 Inference device, 62 Inference unit, 71A, 71B, D1, D10, D20, D30 Structured data, 72A, 72B, d2, da1 to da3, db1 to db3, dc Unstructured data, 90, 93 Processing circuit, 91 Processor, 92 Memory, 100 Data management system, 410 Production schedule adjustment unit, 411, 421, 431, 441 AI support unit, 420 Maintenance support unit, 430 Anomaly detection unit, 440 Predictive maintenance unit, M1 Learning model, X1 to X3 Input layer, Y1, Y2 Hidden layer, Z1 to Z3 Output layer, w11 to w16, w21 to w26 Weights.

Claims

1. a standard data collection unit that collects standard data, which is set in advance as data that can be acquired as analysis items for data analysis, from production equipment or an operation terminal associated with the production equipment; an unstructured data collection unit that collects unstructured data, which is data other than the structured data and includes data related to the production equipment, from the production equipment or the operation terminal; a data allocation unit that stores the unstructured data and the structured data as management items associated with the unstructured data in a data storage device; Equipped with A data management device characterized by:

2. an unstructured data analysis unit that analyzes the unstructured data to select the structured data corresponding to the unstructured data, and associates the structured data with the unstructured data using the selected structured data as a management item; 2. The data management device according to claim 1.

3. The standard data includes attribute information indicating attributes of the standard data and acquired data that has actually been acquired, the unstructured data analysis unit associates the structured data with the unstructured data, with the attribute information being a higher-level management item and the acquired data being a lower-level management item; 3. The data management device according to claim 2.

4. The unstructured data is associated with the attribute information or the acquired data.

4. The data management device according to claim 3.

5. a data processing unit that searches for unstructured data stored in the data storage device using the structured data as the management items as a search index; 2. The data management device according to claim 1.

6. the data processing unit categorizes the searched unstructured data by associating the unstructured data with each other; 6. The data management device according to claim 5.

7. the unstructured data analysis unit associates the structured data with the unstructured data when the unstructured data collection unit collects the unstructured data; 3. The data management device according to claim 2.

8. the unstructured data analysis unit analyzes the structured data and the unstructured data, selects structured data having a degree of association with the unstructured data higher than a specific value, and associates the analyzed unstructured data with the selected structured data; 3. The data management device according to claim 2.

9. a first data acquisition unit that acquires first learning data including the standard data and non-standard data corresponding to the standard data; a model generation unit that generates a first learning model for inferring unstructured data corresponding to the structured data from the structured data using the first learning data; Further comprising a learning device having 9. The data management device according to claim 1, wherein the data management device is a data management device.

10. a second data acquisition unit that acquires the standard data; an inference unit that uses the first learning model to infer and output unstructured data corresponding to the structured data acquired by the second data acquisition unit; and a reasoning device having:

10. The data management device according to claim 9.

11. The unstructured data includes trouble information, which is at least one of a trouble content, a trouble cause, and a trouble response in a production process of the production equipment.

11. The data management device according to claim 10.

12. the first data acquisition unit acquires second learning data including the standard data, the trouble information corresponding to the standard data, and trouble items standardized by superordinately conceptualizing the trouble information; the model generation unit generates a second learning model for inferring trouble information corresponding to the standard data and the trouble item from the standard data and the trouble item, using the second learning data; 12. The data management device according to claim 11.

13. the second data acquisition unit acquires the standard data and the problem items; the inference unit infers and outputs trouble information corresponding to the standard data and the trouble item acquired by the second data acquisition unit using the second learning model.

13. The data management device according to claim 12.

14. the trouble information is the trouble response indicating a work instruction including a work to be done to deal with the trouble, The inference unit infers and outputs the work instruction.

14. The data management device according to claim 13.

15. a data management device that manages standardized data that is set in advance as data that can be acquired as analysis items for data analysis, and non-standardized data that is data other than the standardized data, including data related to production equipment; a data processing device that performs data processing on the structured data and the unstructured data; and The data management device a standard data collection unit that collects the standard data from the production equipment or an operation terminal associated with the production equipment; an unstructured data collection unit that collects the unstructured data from the production equipment or the operation terminal; a data allocation unit that stores the unstructured data and the structured data as management items associated with the unstructured data in a data storage device; Equipped with the data processing device searches for unstructured data stored in the data storage device using the structured data as the management item as a search index, and performs data processing on the searched unstructured data and the structured data used as the index. A data management system comprising:

16. a utilization device that utilizes the data processing results for the structured data and the unstructured data, the data processing device uses first learning data including the standard data and standard data corresponding to the standard data to generate a learning model for inferring, from the standard data, standard data corresponding to the standard data, and outputs the learning model to the utilization device as the data processing result; The utilization device acquires the structured data, infers unstructured data corresponding to the acquired structured data using the learning model, and outputs the inferred data.

16. The data management system according to claim 15.

17. a standard data collection step in which the data management device collects standard data, which is set in advance as data that can be acquired as analysis items for data analysis, from the production equipment or an operation terminal associated with the production equipment; an unstructured data collection step in which the data management device collects unstructured data, which includes data related to the production equipment and is data other than the structured data, from the production equipment or the operation terminal; a data allocation step in which the data management device stores the unstructured data and the structured data as management items associated with the unstructured data in a data storage device; Including, A data management method comprising:

18. a standard data collection step of collecting standard data, which is set in advance as data that can be acquired as analysis items for data analysis, from a production facility or an operation terminal associated with the production facility; an unstructured data collection step of collecting unstructured data, which is data other than the structured data and includes data related to the production equipment, from the production equipment or the operation terminal; a data allocation step of storing the unstructured data and the structured data as management items associated with the unstructured data in a data storage device; A data management program that causes a computer to execute the above steps.

Images