Support system, support method, and program

Abstract

Output a report containing flexible advice based on the values ​​of the measurement data. [Solution] A processor of a support server (20) obtains an answer including the advice from a large-scale language model using a prompt including measurement data from each sensor (12A) provided in a processing system (11), external information for causing a large-scale language model to output advice for supporting the operation of the processing system (11) based on the measurement data, and an instruction for generating the advice based on the measurement data and the external information, and generates a report using the answer. The measurement data is assigned a first tag unique to the sensor (12A) that measured the measurement data. The external information is assigned a second tag with a tag name corresponding to the first tag assigned to the measurement data corresponding to the external information. The prompt includes the measurement data assigned a first tag corresponding to the tag to be evaluated, and the external information assigned a second tag corresponding to the tag to be evaluated.

Description

[Technical Field]

[0001] The present invention relates to a support system, a support method, and a program. [Background technology]

[0002] In a processing system installed in a chemical plant, a manufacturing line of a factory, or the like, a plurality of sensors are provided on a plurality of devices and the piping connecting them in order to grasp the operating status of the processing system. For example, a technology is known that assists in managing the operating status of a processing system by determining whether measurement data from the sensors is an abnormal value and notifying the administrator of the processing system of the determination result (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2024-36041 Summary of the Invention [Problem to be solved by the invention]

[0004] Meanwhile, a technology is being considered that automatically outputs advice to a processing system administrator to support the operation of the processing system based on measurement data. One example of such a technology is to first set rules that associate measurement data values ​​with standard data that express advice using standard phrases, symbols, etc., and then output advice based on the standard data that corresponds to the actual measurement data values ​​based on the rules. In this case, only advice that conforms to the preset standard data is output, making it impossible to output flexible advice. [Means for solving the problem]

[0005] An assistance system that solves the above problem is an assistance system that generates a report to assist in the operation of a processing system equipped with multiple sensors via a large-scale language model, wherein measurement data corresponding to the measurement value of each sensor is assigned a first tag unique to the sensor that measured the measurement data, and external information for causing the large-scale language model to output advice to assist in the operation of the processing system according to the measurement data is assigned a second tag corresponding to the first tag assigned to the measurement data corresponding to the external information, and a processor of the assistance system generates a prompt including: evaluation target measurement data assigned the first tag corresponding to the evaluation target tag; evaluation target external information assigned the second tag corresponding to the evaluation target tag; and an instruction to generate the advice based on the evaluation target measurement data and the evaluation target external information, and inputs the prompt into the large-scale language model to obtain an answer including the advice from the large-scale language model, and generates the report using the answer.

[0006] A support method for solving the above problem uses a processor included in the support system to generate a report for supporting the operation of a processing system having multiple sensors via a large-scale language model, wherein measurement data corresponding to measurement values ​​of each sensor is assigned a first tag specific to the sensor that measured the measurement data, and external information for causing the large-scale language model to output advice for supporting the operation of the processing system according to the measurement data is assigned a second tag corresponding to the first tag assigned to the measurement data corresponding to the external information, the processor generates a prompt including evaluation target measurement data assigned the first tag corresponding to an evaluation target tag, evaluation target external information assigned the second tag corresponding to the evaluation target tag, and an instruction for generating the advice based on the evaluation target measurement data and the evaluation target external information, inputs the prompt to the large-scale language model, obtains an answer including the advice from the large-scale language model, and generates the report using the answer. A program for solving the above problem causes the processor included in the support system to function as a means for executing the support method.

[0007] According to the above-mentioned assistance system, assistance method, or program, a prompt including measurement data to be evaluated and external information is input to a large-scale language model, and the large-scale language model outputs advice corresponding to the measurement data to be evaluated based on the external information. By inputting the external information into the large-scale language model along with the measurement data to be evaluated, flexible advice corresponding to the measurement data to be evaluated can be obtained from the large-scale language model without strictly setting rules that associate the values ​​of the measurement data to be evaluated with advice.

[0008] In the support system, the evaluation target measurement data may include a plurality of measurement data to which the same first tag corresponding to the evaluation target tag is assigned and which are assigned different time stamps. With the above configuration, flexible advice can be output in accordance with the change in the measurement data over time, without having to set strict rules in advance for the pattern of change in the measurement data over time and the advice corresponding to the pattern of change over time.

[0009] The assistance system may include a first database that stores the measurement data and the first tag in association with each other, and a second database that stores the external information and the second tag in association with each other. According to the above configuration, the processor can execute a process for generating a prompt using the measurement data stored in the first database and the external information stored in the second database. Furthermore, the processor can identify the external information corresponding to the evaluation target measurement data from the external information stored in the second database based on the tag attached to the external information. [Effects of the Invention]

[0010] According to the present invention, it is possible to output a report including flexible advice according to the values ​​of the measurement data. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic diagram showing the overall configuration of the support system. [Figure 2] FIG. 2 is a schematic diagram showing the hardware configuration of the support server. [Figure 3] FIG. 3 is a block diagram showing the configuration of the support server. [Figure 4] FIG. 4 is a flowchart of the report generation process. DETAILED DESCRIPTION OF THE INVENTION

[0012] An embodiment of the support system, support method, and program will be described below with reference to Figures 1 to 4. The support system of this embodiment is a computer system for notifying a manager of a processing system of advice for supporting the operation of the processing system provided in a chemical plant, a production line of a factory, or the like.

[0013] <Overall structure> 1, the support system of this embodiment includes a support server 20. The support server 20 outputs a report to notify the manager of a processing system 11 of advice for supporting the operation of the processing system 11 included in a plant 10. The support server 20 uses a large language model (LLM) included in an LLM server 30 to generate the report.

[0014] Plant 10 may be any facility with a processing system 11, such as a chemical plant or a factory production line. The processing system 11 is at least a part of the plant 10. For example, the processing system 11 has a configuration in which a plurality of devices 11A are connected by piping 11B. An example of the devices 11A is a processing device that performs some kind of processing on the processing target or other devices 11A. An example of the devices 11A is a tank that temporarily stores the processing target or other raw materials. The devices 11A include, but are not limited to, at least one of a pump, a compressor, a turbine, a heat exchanger, a distillation column, a reaction vessel, a dryer, a heating furnace, a fan, and a valve. The devices 11A also include a control terminal that controls the process. The control terminal includes, but is not limited to, at least one of a solenoid valve, a positioner, a motor of a motor pump, and a relay switch connected to an opening / closing control device that controls the operation of an electric heater.

[0015] The plant 10 includes a sensor group 12 provided in a treatment system 11. The sensor group 12 includes, as its components, a plurality of sensors 12A for monitoring the operating status of the treatment system 11. The plurality of sensors 12A detect physical quantities representing the state of facilities such as the equipment 11A and the piping 11B. Preferably, the sensor 12A is provided in each of the plurality of equipment 11A and each of the plurality of piping 11B, but it may be provided in any measurement target among the plurality of equipment 11A and the plurality of piping 11B. The sensor 12A may be, for example, a flow meter, a level gauge, a thermometer, a pressure gauge, a vibrometer, an analyzer, an ammeter, a voltmeter, a speed meter, or a sensor that detects the opening of a valve, but is not limited to these, and may be any sensor that measures any physical quantity.

[0016] The plant 10 includes a management device 13. The management device 13 is, for example, a data server, but may also be a computer terminal used by a manager of the plant 10. The management device 13 stores measurement data measured by each sensor 12A constituting the sensor group 12. The management device 13 accumulates measurement data from the past to the present by repeatedly storing measurement data corresponding to the measurement values ​​of each sensor 12A at predetermined time intervals.

[0017] For example, the management device 13 may store the measurement values ​​measured by each sensor 12A as the measurement data. For example, the management device 13 may store values ​​obtained by performing calculations on the measurement values ​​measured by each sensor 12A as the measurement data. For example, the measurement data may be a value representing the ratio of the measurement value to the design value (theoretical value) of the physical quantity measured by the sensor 12A. For example, the measurement data may be a value calculated using two or more measurement values, such as a pressure difference representing the difference between the measurement values ​​of two pressure gauges.

[0018] Furthermore, a first tag unique to the sensor 12A that measured the measurement data is assigned to the measurement data stored in the management device 13. The first tag is, for example, a character string composed of letters, numbers, symbols, or a combination thereof. That is, a character string unique to the sensor 12A that measured the measurement data is assigned to each measurement data as a first tag. For example, the same first tag is assigned to multiple pieces of measurement data acquired by the same sensor 12A at different times. Furthermore, a timestamp that identifies the measurement date and time is assigned to each measurement data.

[0019] The first tag may include, for example, information that can identify the plant 10 from which the measurement data was acquired, information that can identify the equipment to be measured, and information that can identify the sensor 12A that measured the measurement data. The first tag may also include, for example, information that can identify a physical quantity represented by the value of the measurement data.

[0020] The plant 10 includes a user terminal 14. The user terminal 14 is, for example, a computer terminal managed by a user such as an employee of the business entity that manages the plant 10. The business entity may be a single company or a collection of companies that are economically and organizationally related. The user terminal 14 is, for example, a desktop or laptop personal computer, but may also be a mobile terminal such as a tablet or smartphone that supports a mobile communication system. An application program for displaying reports sent from the support server 20 and a web page browser are installed on the user terminal 14. While FIG. 1 illustrates a configuration in which the user terminal 14 is located inside the plant 10, the user terminal 14 may also be a computer terminal located outside the plant 10.

[0021] As an example, the support server 20 is a server managed by a company separate from the company that manages the processing system 11, but it may also be a server managed by the company that manages the processing system 11. The support server 20 is connected to the management device 13 of the plant 10, the user terminal 14, and the LLM server 30 via a network line so that they can communicate with each other. The support server 20 collects measurement data from the management device 13 of the plant 10. The support server 20 also causes the large-scale language model included in the LLM server 30 to generate an answer, which is advice for supporting the operation of the processing system 11 and includes advice corresponding to the measurement data. The support server 20 outputs a report using the answer obtained from the large-scale language model to the user terminal 14. Note that the report output by the support server 20 may be output to the user terminal 14 via another server or the like.

[0022] The report generated by the support server 20 may be in the form of, for example, an email containing text data. In this case, the support server 20 transmits the report in the form of an email to a predetermined mail server. The user terminal 14 obtains the report in the form of an email from the mail server. The report generated by the support server 20 may also be in the form of an image displayed by an application program or a web page browser. The image may be a still image or a video.

[0023] As an example, the LLM server 30 is a server managed by a company separate from the company that manages the assistance server 20, but it may also be a server managed by the company that manages the assistance server 20. The large-scale language model provided in the LLM server 30 is a model trained from a large-scale text corpus. The large-scale language model is used to perform natural language understanding tasks. The large-scale language model has the ability to interpret sentences given as prompts and generate appropriate responses in that context.

[0024] <Hardware configuration> Next, an example of the hardware configuration of the support server 20 will be described with reference to Fig. 2. Note that the hardware configuration of the support server 20 is not limited to the following example, and other hardware configurations may be used. Furthermore, the management device 13 and the user terminal 14 may have the same hardware configuration as the following.

[0025] As shown in FIG. 2, the support server 20 includes a processor 21, a storage unit 22, a memory 23, an input / output IF 24, and a communication IF 25. The processor 21 is hardware for executing an instruction set written in a program, and is composed of an arithmetic unit, a register, a peripheral circuit, and the like.

[0026] The memory unit 22 is a storage for saving data. The memory unit 22 is, for example, a flash memory, a hard disk drive (HDD), etc. The memory 23 temporarily stores programs, data to be processed by the programs, etc. The memory 23 is, for example, a volatile memory such as a dynamic random access memory (DRAM).

[0027] The input / output IF24 is an interface including an input device for receiving operation input from the administrator of the support server 20 and an output device for presenting information to the administrator. The input device is, for example, a pointing device such as a mouse, a keyboard, etc. The input device may be a microphone for voice input. The output device is, for example, a display panel such as a liquid crystal display panel or an organic EL panel. The input / output IF24 may be, for example, a touch panel that combines a display panel as a display unit that displays images and a touchpad as an input unit that receives operations by the user. The communication IF25 is an interface for inputting and outputting signals for communication with external devices.

[0028] <Processor 21> The processor 21 of the support server 20 executes a report generation process to output a report including advice for supporting the operation of the processing system 11 based on the measurement data.

[0029] 3, the processor 21 executes a program for report generation processing, thereby functioning as a data collection unit 21A, a prompt generation unit 21B, an advice acquisition unit 21C, a report generation unit 21D, etc. The program for report generation processing is stored in the storage unit 22 of the support server 20.

[0030] The data collection unit 21A executes a measurement data collection process to collect measurement data from the management device 13 of the plant 10. The data collection unit 21A also stores the measurement data collected from the management device 13 in the memory unit 22. The process of collecting measurement data from the management device 13 by the data collection unit 21A is performed at predetermined intervals, such as every few hours, every half day, or every day. In this case, the data collection unit 21A collects measurement data stored in the management device 13 from the last time measurement data is collected from the management device 13 until the next time measurement data is collected from the management device 13. Note that the process of collecting measurement data from the management device 13 by the data collection unit 21A may be a process of transmitting measurement data from the management device 13 to the assistance server 20 at predetermined intervals.

[0031] The prompt generation unit 21B executes a prompt generation process to generate a prompt to be input to the large-scale language model of the LLM server 30. The prompt includes measurement data to be evaluated, external information, and an instruction to generate advice to assist the operation of the processing system 11 based on the measurement data to be evaluated and the external information.

[0032] The evaluation target measurement data included in the prompt includes one or more pieces of measurement data to which a first tag defined as an evaluation target tag is assigned. The evaluation target tag is a first tag assigned to measurement data for the large-scale language model to evaluate or determine the operating status of the processing system 11 when the large-scale language model generates advice to support the operation of the processing system 11. In other words, the large-scale language model evaluates or determines the operating status of the processing system 11 based on the measurement data to which a first tag corresponding to the evaluation target tag is assigned, and then outputs advice according to the operating status.

[0033] The prompt generation unit 21B may define only one first tag as the evaluation target tag, or may define multiple different first tags as the evaluation target tags. For example, the prompt generation unit 21B may define any predetermined first tag as the evaluation target tag. Alternatively, the prompt generation unit 21B may first perform a process to determine whether measurement data represents abnormal behavior, and then define the first tag of the measurement data determined to represent abnormal behavior as the evaluation target tag. In this case, the storage unit 22 stores, for each first tag of the measurement data, a determination condition such as a threshold value for determining whether the measurement data represents abnormal behavior. Note that the process of determining whether measurement data represents abnormal behavior is not limited to being performed by the prompt generation unit 21B, and may be performed by another computer system.

[0034] The prompt generation unit 21B may generate a prompt for each first tag corresponding to the evaluation target tag. In this case, the prompt generation unit 21B may include, in one prompt, multiple pieces of measurement data to which the same first tags corresponding to the evaluation target tags are assigned but which are different timestamps. In other words, the evaluation target measurement data may include multiple pieces of measurement data to which the same first tags corresponding to the evaluation target tags are assigned but which are different timestamps. The prompt generation unit 21B may include, in one prompt, multiple pieces of measurement data to which different first tags corresponding to the evaluation target tags are assigned. In other words, the evaluation target measurement data may include multiple pieces of measurement data to which different first tags corresponding to the evaluation target tags are assigned. Furthermore, the prompt generation unit 21B may include, in addition to the evaluation target tags, measurement data to which first tags necessary for generating a report are assigned, in the prompt.

[0035] The external information included in the prompt includes information for causing the large-scale language model to output advice for supporting the operation of the processing system 11 in accordance with the measurement data to be evaluated. The external information is, for example, text data written in any language, but may also be image data containing characters and symbols. Examples of external information are given below.

[0036] For example, the external information includes advice instructing or suggesting operation of device 11A of processing system 11 according to the value of the measurement data. As an example, the external information includes a threshold value for the measurement data and advice regarding operation of device 11A as a response when the measurement data exceeds the threshold. For example, if the first tag of the measurement data representing the pressure at an arbitrary valve is V1, an example of the external information is a sentence such as "When the measurement data assigned the first tag of V1 is less than X1 [barG], increase the flow rate of the gas feed." In this case, the large-scale language model outputs advice instructing or suggesting operation of device 11A to increase the flow rate of the gas feed when the actual measurement value of the measurement data assigned the first tag of V1 exceeds the threshold included in the external information.

[0037] An example of the external information includes an upper or lower limit that the measurement data must satisfy, or a numerical range defined by the upper and lower limits. An example of the external information is a sentence such as, "The measurement data tagged with the first tag V1 is maintained within a range of X1 [barG] or more and X2 [barG] or less." In this case, if the actual measured value of the measurement data tagged with the first tag V1 falls outside the numerical range included in the external information, the large-scale language model outputs advice instructing or suggesting an operation of device 11A to increase or decrease the value of the measurement data. As in the above example, the external information may include information that allows advice regarding the operation of device 11A to be inferred from the measurement data. In other words, the external information does not necessarily have to include advice regarding the operation of device 11A itself.

[0038] Furthermore, the advice for supporting the operation of the processing system 11 is not limited to advice regarding the operation of the device 11A, and may be, for example, advice instructing or suggesting observation of the behavior of the measurement data or the state of the device 11A. An example of the external information in this case is a sentence such as "If the measurement data assigned the first tag of V1 is X3 [barG] or more, caution is required." In this case, when the actual measurement value of the measurement data assigned the first tag of V1 exceeds a threshold included in the external information, the large-scale language model outputs advice instructing or suggesting observation of the future behavior of the measurement data or the operation of the device 11A.

[0039] The external information may include statistical values ​​of past measurement data measured while the processing system 11 was operating normally. In this case, the statistical value is, for example, at least one selected from the mean, median, standard deviation, and quartile. In this case, the large-scale language model compares the actual measurement value of the measurement data with the statistical value included in the external information, and can output advice regarding the operation of the device 11A of the processing system 11, taking into account how much the actual measurement value of the measurement data deviates from the statistical value. The external information may include design values ​​(target values) of the measurement data.

[0040] The external information may include reference information for inferring the state of the processing system 11 from the measurement data. The reference information is, for example, information that associates the value of the measurement data with the state of the processing system 11. An example of the reference information is a sentence such as, "When the measurement data assigned with the first tag V1 is X4 [barG] or less, the processing system 11 is stopped." An example of the reference information is a sentence such as, "When the measurement data assigned with the first tag V1 is X5 [barG] or more, a specific process (e.g., an adsorbent regeneration process) is being performed." For example, if the first tag of the measurement data representing the flow rate of a fluid discharged from a given pump is P1, an example of the reference information is a sentence such as, "When the measurement data assigned with the first tag P1 is Y1 [t / h] or more, the pump is operating." Another example of the reference information is a sentence such as, "When the measurement data assigned with the first tag P1 is typically in the range of Y1 [t / h] or more and Y2 [t / h] or less when the pump is operating."

[0041] The external information may include information about the first tag. The information about the first tag may include, for example, the unit of the measurement data to which the first tag is assigned. The information about the first tag may include, for example, a description of the meaning of the physical quantity represented by the measurement data to which the first tag is assigned. An example of the description of the meaning of the physical quantity represented by the measurement data is a sentence such as, "The measurement data to which the first tag of V1 is assigned represents the pressure in the synthesis process."

[0042] The external information may be an existing document on how to use the equipment included in the processing system 11, such as an operating manual for the equipment 11A. By including the contents of an existing document, such as the operating manual for the equipment 11A, in the prompt, the information contained in the document can be used to generate advice using a large-scale language model.

[0043] The advice acquisition unit 21C executes an advice acquisition process. In the advice acquisition process, the advice acquisition unit 21C inputs the prompt generated by the prompt generation unit 21B into the large-scale language model of the LLM server 30. Then, the advice acquisition unit 21C acquires an answer including advice for supporting the operation of the processing system 11 from the large-scale language model. The answer output by the large-scale language model is, for example, text data, but may also be image data, audio data, or a combination thereof.

[0044] The report generation unit 21D uses the answer obtained from the large-scale language model to generate a report including advice for supporting the operation of the processing system 11. The report may be, for example, text data, image data, or audio data including at least a portion of the answer obtained from the large-scale language model, or a combination thereof. Alternatively, the report may be, for example, data including at least a portion of an answer obtained by further processing the answer obtained from the large-scale language model using the large-scale language model.

[0045] The report generation unit 21D may output the evaluation target measurement data together with the advice in a report. The evaluation target measurement data included in the report may be in text format or in any format such as a graph or a table. The report generation unit 21D may include, together with the advice, various types of information included in external information such as a process flow diagram related to the advice or a photograph of the equipment in the report. The report generation unit 21D may include, together with the advice, the result of determining whether or not the evaluation target measurement data exhibits abnormal behavior based on external information in the report.

[0046] Furthermore, the report generating unit 21D transmits the generated report to the user terminal 14. Note that the report generating unit 21D may transmit the generated report to the user terminal 14 via another server (for example, a mail server) or the like.

[0047] <Storage section 22> As shown in FIG. 3, the storage unit 22 includes, as databases, a measurement data storage unit 22A, an external information storage unit 22B, a prompt storage unit 22C, and a report storage unit 22D.

[0048] The measurement data storage unit 22A is an example of a first database that stores the measurement data acquired by the data collection unit 21A from the management device 13 of the plant 10. For example, the measurement data storage unit 22A stores the measurement data, a first tag assigned to the measurement data, and a timestamp indicating the measurement date and time of the measurement data in association with each other.

[0049] The external information storage unit 22B is an example of a second database that stores external information for outputting advice to the large-scale language model to support the operation of the processing system 11 according to the measurement data. The external information is assigned a second tag whose tag name corresponds to the first tag assigned to the measurement data corresponding to the external information. Note that the tag name of the first tag corresponds to the tag name of the second tag, meaning that the second tag assigned to the external information related to the measurement data can be identified from the first tag assigned to the measurement data. Furthermore, the tag name refers to the character string that constitutes each of the first tag and the second tag.

[0050] For example, external information related to measurement data to which a specific first tag is assigned may be assigned a second tag consisting of the same character string as the specific first tag. For example, external information related to measurement data to which a specific first tag is assigned may be assigned a second tag containing a character string identical to a portion of the character string contained in the first tag. In other words, the tag names of the first tag and the second tag may be different from each other. As an example, if the first tag is the character string "AAA.BB," the second tag may be the character string "AAA-BB," and the first tag and the second tag may be associated by the character strings located before and after the "." and "-." In this case, whether the tag is the first tag or the second tag can be identified depending on whether the tag name contains a "." or a "-." As another example, if the first tag is the character string "AAABB-1," the second tag may be the character string "AAABB-2," and the first tag and the second tag may be associated by the character string located before the "-." In this case, the character or character string following the "-" can be used to identify whether the tag is a first tag or a second tag.

[0051] The external information storage unit 22B stores the external information and the second tag assigned to the external information in association with each other. For example, the external information storage unit 22B preferably stores the external information related to the measurement data assigned with the first tag corresponding to the evaluation target tag so that the external information can be searched for by the tag name of the second tag assigned to the external information.

[0052] Prompt storage unit 22C stores a template of a prompt to be input to the large-scale language model. Prompt storage unit 22C also stores text data necessary for prompt generation unit 21B to generate a prompt. Prompt storage unit 22C may also store prompts that have been generated by prompt generation unit 21B.

[0053] The report storage unit 22D stores a template of a report to be generated by the report generation unit 21D. The report storage unit 22D also stores text data necessary for the report generation unit 21D to generate a report. The report storage unit 22D may also store reports that have been generated by the report generation unit 21D so far.

[0054] <Report generation process> Next, the report generation process will be described with reference to Fig. 4. The following report generation process is repeatedly executed at predetermined intervals.

[0055] 4, in the report generation process, first, the data collection unit 21A collects measurement data stored in the management device 13 from the management device 13 of the plant 10 (step S1). The data collection unit 21A acquires the measurement data from the management device 13 and stores the acquired measurement data in the measurement data storage unit 22A. Note that the process of step S1 may be performed at any timing separate from the report generation process.

[0056] Next, the prompt generation unit 21B reads the measurement data stored in the measurement data storage unit 22A (step S2). Specifically, in step S2, the prompt generation unit 21B reads the evaluation target measurement data to which the first tag corresponding to the evaluation target tag is assigned from the measurement data storage unit 22A.

[0057] For example, in step S2, prompt generation unit 21B may first execute a process to determine whether the measurement data indicates abnormal behavior, and then define the first tag of the measurement data determined to indicate abnormal behavior as the evaluation target tag. Alternatively, in step S2, prompt generation unit 21B may define any predetermined first tag as the evaluation target tag. In step S2, prompt generation unit 21B may define only one type of first tag as the evaluation target tag, or may define two or more types of first tags as the evaluation target tag.

[0058] For example, in step S2, prompt generation unit 21B may acquire from measurement data storage unit 22A a plurality of pieces of measurement data to which the same first tags corresponding to the tags to be evaluated and which are assigned different timestamps. For example, in step S2, prompt generation unit 21B may acquire from measurement data storage unit 22A a plurality of pieces of measurement data to which different first tags corresponding to the tags to be evaluated are assigned. Furthermore, in step S2, prompt generation unit 21B may acquire measurement data to which first tags necessary for generating a report are assigned, in addition to the measurement data to which first tags corresponding to the tags to be evaluated are assigned.

[0059] Next, the prompt generation unit 21B reads the external information stored in the external information storage unit 22B (step S3). More specifically, in step S3, the prompt generation unit 21B reads the external information to which a second tag corresponding to the tag to be evaluated has been assigned from the external information storage unit 22B. At this time, the prompt generation unit 21B acquires the external information to which a second tag corresponding to the tag name of the tag to be evaluated defined in step S2 has been assigned from the external information storage unit 22B. Note that hereinafter, the external information to which a second tag corresponding to the tag to be evaluated has been assigned is referred to as the external information to be evaluated.

[0060] Next, the prompt generation unit 21B generates a prompt including the evaluation target measurement data, the evaluation target external information, and an instruction to generate advice to support the operation of the processing system 11 based on the evaluation target measurement data and the evaluation target external information (step S4). In step S4, the prompt generation unit 21B reads a prompt template including an instruction to generate advice from the prompt storage unit 22C, and generates a prompt by inserting the evaluation target measurement data and the evaluation target external information into predetermined positions in the template.

[0061] Next, the advice acquisition unit 21C inputs the prompt generated by the prompt generation unit 21B into the large-scale language model of the LLM server 30, thereby acquiring an answer including advice to assist in the operation of the processing system 11 from the large-scale language model (step S5).

[0062] The advice for supporting the operation of the processing system 11 may include content instructing or suggesting operation of the device 11A of the processing system 11 according to the value of the measurement data to be evaluated. Furthermore, the advice for supporting the operation of the processing system 11 may include content instructing or suggesting observation of the behavior of the measurement data to which the first tag corresponding to the tag to be evaluated is assigned or the state of the device 11A, for example.

[0063] It is more preferable that the advice for supporting the operation of the processing system 11 is advice corresponding to the change over time of a plurality of pieces of measurement data that are assigned the same first tags corresponding to the tags to be evaluated and that are assigned different timestamps. In this case, the command input to the prompt in step S4 may be a command to generate advice corresponding to the change over time of a plurality of pieces of measurement data that are assigned different timestamps.

[0064] As an example, assume that multiple pieces of measurement data, each of which has the same first tag corresponding to the tag to be evaluated and a different timestamp, are input to the prompt as the measurement data to be evaluated. For example, if the behavior of the multiple pieces of measurement data is outside a preferred range defined by external information but shows a tendency to approach the preferred range over time, there is a possibility that the measurement data will fall within the preferred range as time passes. In such a case, the large-scale language model may provide advice suggesting that the measurement data is outside the preferred range but is showing a tendency to recover, and that the user should observe the behavior of the measurement data in the future. Furthermore, if the behavior of the multiple pieces of measurement data is outside the preferred range and shows no signs of recovery, the large-scale language model may provide advice instructing or suggesting an operation to be performed on device 11A to bring the measurement data into the preferred range.

[0065] Next, the report generation unit 21D generates a report including advice for assisting the operation of the processing system 11 using the answer including the advice acquired from the large-scale language model (step S6). For example, in step S6, the report generation unit 21D reads a report template from the report storage unit 22D and generates a report by inserting at least a part of the answer acquired from the large-scale language model into a predetermined position of the template. Then, the report generation unit 21D transmits the generated report to the user terminal 14. With the above processing, the report generation processing is completed.

[0066] <Effects of the embodiment> (1) In this embodiment, a prompt including measurement data to be evaluated and external information is input to a large-scale language model, and the large-scale language model outputs advice corresponding to the measurement data to be evaluated based on the external information. In this manner, when external information is input to a large-scale language model, flexible advice corresponding to the measurement data to be evaluated can be obtained from the large-scale language model without strictly setting rules that associate the values ​​of the measurement data to be evaluated with advice.

[0067] (2) The external information may include information for causing the large-scale language model to output advice for supporting the operation of the processing system 11 in accordance with the measurement data to be evaluated. Therefore, the external information does not need to be written in a programming language or the like, but can be written in any language. This makes it easy to create and edit the external information. For example, documents used by an administrator of the processing system 11 to manage the processing system 11 can also be used as external information. It is preferable that the external information storage unit 22B stores external information in which the contents of existing documents are converted into text data. This facilitates the creation and editing of prompts and the editing of the contents of existing documents stored as external information.

[0068] (3) When a prompt contains multiple pieces of measurement data that are assigned the same first tag corresponding to the tag to be evaluated but have different timestamps, advice can be output according to changes in the measurement data over time. If advice according to changes in the measurement data over time is to be output without using a large-scale language model, it is necessary to set rules in advance that associate expected patterns of changes in the measurement data over time with advice corresponding to each pattern. In this case, the greater the number of target measurement data, the greater the number of expected patterns of changes in the measurement data over time, making rule setting more complicated. In this regard, the present support system can output flexible advice according to changes in the measurement data over time without having to set strict rules in advance for patterns of changes in the measurement data over time and advice corresponding to those patterns.

[0069] (4) The prompt generating unit 21B can execute a process of generating a report using the measurement data stored in the measurement data storage unit 22A and the external information stored in the external information storage unit 22B.

[0070] (5) The external information storage unit 22B stores external information corresponding to measurement data so that it can be searched based on the correspondence between at least a portion of the tag name of the first tag assigned to the measurement data and at least a portion of the tag name of the second tag assigned to the external information. For example, when identifying external information using a semantic search based on the first tag of the measurement data, a database or the like is required to store search text corresponding to the first tag of the measurement data. In this regard, by assigning a second tag to the external information in advance, external information to which a second tag corresponding to the evaluation target tag is assigned can be identified as evaluation target external information based on the tag name of the second tag. Therefore, the system configuration can be simplified compared to when external information related to measurement data to which a first tag corresponding to the evaluation target tag is assigned from the external information stored in the external information storage unit 22B is identified using a semantic search based on the tag name of the measurement data.

[0071] <Example of change> This embodiment can be modified as follows: This embodiment and the following modifications can be combined and implemented within the scope of technical compatibility.

[0072] The measurement data included in the prompt may be one or more pieces of measurement data to which a first tag corresponding to the tag to be evaluated is assigned. The measurement data included in the prompt may be an evaluation target statistical value, which is a statistical value of multiple pieces of measurement data to which the same first tags corresponding to the tags to be evaluated and which are assigned different timestamps. As an example, the evaluation target statistical value refers to any one of the minimum, maximum, median, mode, and average of multiple pieces of measurement data to which the same first tags corresponding to the tags to be evaluated and which are assigned different timestamps.

[0073] In the support system, the measurement data storage unit 22A included in the storage unit 22 of the support server 20 may be omitted. In this case, the prompt generation unit 21B may be configured to acquire measurement data from the management device 13 every time a report generation process is performed.

[0074] In the support system, the external information storage unit 22B included in the storage unit 22 of the support server 20 may be omitted. In this case, the external information may be stored in another device. Alternatively, the external information may be written in the code of the program executed by the processor 21. In this case, when the processor 21 executes the program, the external information is expanded in the memory 23.

[0075] The assistance system may be realized as a single device, or may be distributed across multiple devices or subsystems that cooperate to execute a program. In the assistance system, the assistance server 20 and other devices may be realized as a single device. In the assistance system, at least part of the configuration of the assistance server 20 may be performed by a cloud server. For example, at least part of the components exemplified as databases provided in the storage unit 22 of the assistance server 20 may be replaced with databases on a cloud server. [Explanation of symbols]

[0076] 10...Plant 11...Processing system 12...Sensor group 12A...sensor 13…Management device 14...User terminal 20...Support server 21...Processor 21A...Data collection section 21B...Prompt generation unit 21C…Advice Acquisition Department 21D…Report generation section 22...Storage section 22A...Measurement data storage section 22B...External information storage section 30...LLM server

Claims

1. 1. A support system that generates a report for supporting the operation of a processing system having a plurality of sensors via a large-scale language model, comprising: a first tag unique to the sensor that measures the measurement data is assigned to the measurement data corresponding to the measurement value of each sensor; external information for causing the large-scale language model to output advice for supporting the operation of the processing system in accordance with the measurement data is assigned a second tag corresponding to the first tag assigned to the measurement data corresponding to the external information; a processor of the assistance system, generating a prompt including evaluation target measurement data to which the first tag corresponding to the evaluation target tag is assigned, evaluation target external information to which the second tag corresponding to the evaluation target tag is assigned, and an instruction to generate the advice based on the evaluation target measurement data and the evaluation target external information; inputting the prompt into the large-scale language model to obtain an answer from the large-scale language model, the answer including the advice; The answers are used to generate the report Support system.

2. The evaluation target measurement data includes a plurality of measurement data to which the same first tag corresponding to the evaluation target tag is assigned and to which different time stamps are assigned. The assistance system according to claim 1 .

3. The assistance system includes: a first database that stores the measurement data and the first tag in association with each other; a second database that stores the external information and the second tag in association with each other; The assistance system according to claim 1 or 2.

4. 1. A support method for generating a report for supporting operation of a processing system having a plurality of sensors via a large-scale language model, using a processor included in the support system, the method comprising: a first tag unique to the sensor that measures the measurement data is assigned to the measurement data corresponding to the measurement value of each sensor; external information for causing the large-scale language model to output advice for supporting the operation of the processing system in accordance with the measurement data is assigned a second tag corresponding to the first tag assigned to the measurement data corresponding to the external information; the processor: generating a prompt including evaluation target measurement data to which the first tag corresponding to the evaluation target tag is assigned, evaluation target external information to which the second tag corresponding to the evaluation target tag is assigned, and an instruction to generate the advice based on the evaluation target measurement data and the evaluation target external information; inputting the prompt into the large-scale language model to obtain an answer from the large-scale language model, the answer including the advice; The answers are used to generate the report How to help.

5. The processor included in the support system functions as a means for executing the support method according to claim 4. program.

Images