Information processing device, prediction model generation method, information processing method and program

Abstract

To provide an information processing apparatus that predicts a deterioration state of equipment accurately, a prediction model generation method, an information processing method, and a program.SOLUTION: An information processing apparatus (10) includes deterioration state prediction means (16) which predicts a deterioration state of target equipment, using a prediction model which is selected from among a plurality of prediction models corresponding to multiple pieces of equipment, and which corresponds to the target equipment out of the multiple pieces of equipment included in a plant. The multiple prediction models may correspond to at least one of kind and location of the multiple pieces of equipment.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, a prediction model generation method, an information processing method, and a program. [Background technology]

[0002] For example, there are known techniques for maintaining and managing multiple pieces of equipment contained in a plant such as a water supply and sewerage facility. As one of the techniques for maintaining and managing equipment, a system for supporting the formulation of equipment renewal plans has been proposed. For example, Patent Document 1 discloses a system that can display risk information for equipment on a plant-by-plant basis. Also, for example, Patent Document 2 discloses a system that prevents equipment maintenance costs from concentrating in a particular fiscal year. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-173575 [Patent Document 2] Japanese Patent Application Laid-Open No. 2014-016691 Summary of the Invention [Problem to be solved by the invention]

[0004] Here, highly accurate estimation of the deterioration state of equipment enables accurate planning of equipment renewal. However, the technology of Patent Document 1 can display the registered past or present deterioration state of equipment, but does not estimate the deterioration state of the equipment. The technology of Patent Document 2 uses a theoretical deterioration formula or a statistical deterioration model, but performs processing to shift the deterioration curve based on the risk taken into account from the importance of the equipment, and does not place importance on the accuracy of the deterioration curve.

[0005] In view of the above circumstances, an object of the present disclosure is to provide an information processing device, a prediction model generation method, an information processing method, and a program that can predict the deterioration state of equipment with high accuracy. [Means for solving the problem]

[0006] An information processing device according to an embodiment of the present disclosure includes: The plant is provided with a deterioration state prediction means for predicting the deterioration state of a piece of equipment that is the target of prediction, using a prediction model selected from a plurality of prediction models corresponding to the piece of equipment included in the plant.

[0007] An information processing device according to an embodiment of the present disclosure includes: The information processing device further includes a prediction model generation means for generating the plurality of prediction models used in the deterioration state prediction means by machine learning using learning data in which indicators of performance data of equipment in other plants are used as explanatory variables and the deterioration state of the equipment in the other plants is used as a target variable.

[0008] A method for generating a prediction model according to an embodiment of the present disclosure includes: A prediction model generation method executed by an information processing device, comprising: The method includes a step of generating the plurality of prediction models used in the deterioration state prediction means of the information processing device by machine learning using learning data in which indicators of performance data of equipment of other plants are used as explanatory variables and the deterioration state of equipment of the other plants is used as a target variable.

[0009] A program according to an embodiment of the present disclosure includes: An information processing device The plurality of prediction models used in the deterioration state prediction means of the information processing device are made to function as a prediction model generation means that generates prediction models by machine learning using learning data in which indicators of performance data of equipment in other plants are used as explanatory variables and the deterioration state of equipment in the other plants is used as a target variable.

[0010] An information processing method according to an embodiment of the present disclosure includes: An information processing method executed by an information processing device, and predicting the deterioration state of the equipment that is the target of prediction using a prediction model selected from a plurality of prediction models corresponding to the equipment that is the target of prediction among a plurality of equipment included in the plant.

[0011] A program according to an embodiment of the present disclosure includes: An information processing device The deterioration state prediction means functions as a deterioration state prediction means for predicting the deterioration state of the equipment that is the target of prediction, using a prediction model selected from a plurality of prediction models corresponding to the equipment that is the target of prediction among a plurality of equipment included in the plant. [Effects of the Invention]

[0012] According to the present disclosure, it is possible to provide an information processing device, a prediction model generation method, an information processing method, and a program that predict the deterioration state of equipment with high accuracy. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of an information processing device according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a diagram for explaining generation of a prediction model. [Figure 3] FIG. 3 is a diagram illustrating an example of a facility map. [Figure 4] FIG. 4 is a diagram illustrating an example of an image indicating the degree of urgency of updating. [Figure 5] FIG. 5 is a diagram illustrating images of a plurality of scenarios. [Figure 6] FIG. 6 is a diagram illustrating an example of an image of update costs. [Figure 7] FIG. 7 is a diagram illustrating an example of a comparison image of life cycle costs. [Figure 8] FIG. 8 is a flowchart illustrating an example of processing of an information processing method according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, an information processing device, a prediction model generation method, an information processing method, and a program according to embodiments of the present disclosure will be described with reference to the drawings.

[0015] FIG. 1 is a diagram illustrating an example of the configuration of an information processing device 10 according to this embodiment. The information processing device 10 is a device that executes information processing to support the creation of an update plan for multiple pieces of equipment included in a plant. In this embodiment, the plant is a water supply and sewerage facility, but is not limited to this and may be, for example, an industrial product manufacturing factory. The equipment included in the plant may be, for example, but is not limited to, pumps, valves, reaction tank aeration devices, flow meters, etc. Furthermore, the equipment included in the plant is not limited to mechanical equipment or electrical equipment, but includes civil engineering, construction, and building equipment, etc. The information processing device 10 has a hardware configuration such as a personal computer. Here, the information processing device 10 may be configured with multiple pieces of hardware and may be referred to as an information processing system.

[0016] First, the deterioration level used in the update plan that the information processing device 10 assists in creating will be described. The deterioration level indicates the degree of deterioration of each piece of equipment, and is a quantified evaluation value that varies depending on the condition of the equipment, the number of years since installation, etc. The deterioration level may be, for example, a health level or an evaluation score. The deterioration level changes over time, decreasing or increasing in value according to a predetermined calculation formula or a degree of deterioration set by the user. The change in the deterioration level over time can be calculated by the deterioration state prediction means 16, which will be described later, and can be managed so as to correspond to each of the multiple pieces of equipment and stored in the memory unit 12.

[0017] As an example, the deterioration degree may be a health degree indicating the health of multiple pieces of equipment. The health degree changes so that the value becomes smaller over time. In this embodiment, the health degree is indicated by a value from "1" to "5." A health degree of "5" indicates that the equipment is in a healthy state and there are no problems. A health degree of "4" indicates that deterioration has begun to appear. A health degree of "3" indicates that deterioration is progressing. A health degree of "2" indicates that the equipment needs to be updated because there is a possibility of a malfunction. A health degree of "1" indicates that the equipment is in a malfunction state. The health degree returns to "5" when the equipment is updated.

[0018] On the other hand, the evaluation score is an index mainly used in water supply asset management, and is a quantitative value ranging from 0 to 100 that comprehensively evaluates physical, functional, economic, social, seismic resistance, and useful life. The lower the score, the more advanced the deterioration.

[0019] The above-mentioned health levels of "5", "4", "3", "2", and "1" can be respectively associated with evaluation points of "80 to 100 points", "60 to less than 80 points", "40 to less than 60 points", "20 to less than 40 points", and "0 to less than 20 points". In the following, an embodiment will be described assuming that the deterioration level is the health level.

[0020] The information processing device 10 has the following hardware configuration: The information processing device 10 includes a communication unit 11, a storage unit 12, a control unit 13, an input unit 20, and a display unit 21.

[0021] The communication unit 11 includes a communication interface that communicates with the input unit 20 and the display unit 21. The communication interface that communicates with the input unit 20 and the display unit 21 may be, for example, USB and HDMI (registered trademark).

[0022] The communication unit 11 may include a communication interface for connecting to a network such as the Internet. The communication interface for connecting to the Internet may include, for example, a wired or wireless LAN and a router. The information processing device 10 may acquire, for example, via a network, indicators, measured values, and actual deterioration states related to the degree of deterioration of equipment in other plants. Here, the other plants refer to plants different from the target plant for which the information processing device 10 creates an update plan.

[0023] The storage unit 12 is one or more memories. The memory may be, for example, a semiconductor memory, a magnetic memory, an optical memory, or the like, but is not limited to these, and may be any memory. The storage unit 12 is, for example, built into the information processing device 10, but may also be configured to be externally connected to the information processing device 10 via any interface.

[0024] The control unit 13 is one or more processors. The processor may be, for example, a general-purpose processor or a dedicated processor specialized for a specific process, but is not limited to these and may be any processor. The control unit 13 controls the overall operation of the information processing device 10.

[0025] The information processing device 10 has the following software configuration. One or more programs used to control the operation of the information processing device 10 are stored in the memory unit 12. When the program stored in the memory unit 12 is read by the control unit 13, it causes the control unit 13 to function as an acquisition means 14, a prediction model generation means 15, a deterioration state prediction means 16, an image generation means 17, and an output means 18. The program stored in the memory unit 12 may be read by the control unit 13 when the user uses the input unit 20 to instruct the control unit 13 to start processing.

[0026] An overview of each means is as follows. The prediction model generation means 15 generates a prediction model for predicting the degree of deterioration of equipment through machine learning using learning data. The learning data may be configured to include indexes related to the degree of deterioration of equipment in other plants, measured values, and actual deterioration states. The prediction model generation means 15 stores and manages the learning data and prediction model in the memory unit 12. As will be described in detail later, the prediction model generation means 15 generates multiple prediction models corresponding to multiple pieces of equipment included in the plant and stores them in the memory unit 12.

[0027] The degradation state prediction means 16 selects a prediction model corresponding to a piece of equipment that is a prediction target among a plurality of pieces of equipment included in the plant from a plurality of prediction models stored in the storage unit 12. The degradation state prediction means 16 predicts the degradation state of the piece of equipment using the selected prediction model. The degradation state prediction means 16 can predict the degradation state of the equipment on a piece-by-piece basis or a piece-by-component basis. In a first mode in which prediction is performed on a piece-by-piece basis for a plurality of pieces of equipment included in the plant, the degradation state prediction means 16 predicts changes in the deterioration level of each piece of equipment over time so as to be able to compare and indicate the urgency of updating each piece of equipment. In a second mode in which prediction is performed on a piece-by-piece basis for a specific piece of equipment, the degradation state prediction means 16 predicts changes in the deterioration level of a piece of equipment that constitutes the specific piece of equipment using a plurality of scenarios. The degradation state prediction means 16 stores and manages the prediction results, i.e., the predicted degradation state of the equipment, in the storage unit 12. Details of the differences between the prediction content in the first mode and the prediction content in the second mode will be described later.

[0028] The degradation state prediction means 16 also manages information such as the names, types, and locations of multiple pieces of equipment included in the plant. The degradation state prediction means 16 also manages information on the parts that make up each piece of equipment included in the plant. The degradation state prediction means 16 may also manage information on the cost required to update each piece of equipment. The degradation state prediction means 16 may manage information such as the type, location, parts, and update cost of equipment as a table linked by the name of the equipment, and store the information in the storage unit 12. The table may be in, for example, a CSV (Comma-Separated Values) format.

[0029] The image generation means 17 generates an image visually showing the deterioration levels of multiple pieces of equipment or multiple scenarios in a manner that allows for comparison. Here, "comparably" means that the changes in the deterioration levels of multiple pieces of equipment or the changes in the deterioration level of a specific piece of equipment according to multiple scenarios are displayed in a manner that allows for comparison. In this embodiment, the images generated by the image generation means 17 include an "image 211 indicating the degree of urgency of renewal (see FIG. 4)" and an "image 212 of multiple scenarios (see FIG. 5)." In the first mode, the image generation means 17 acquires the prediction results stored in the storage unit 12 by the deterioration state prediction means 16 and generates an image 211 indicating the degree of urgency of renewal that allows for comparison of changes over time in the deterioration levels of multiple pieces of equipment included in the plant. In the second mode, the image generation means 17 acquires the prediction results stored in the storage unit 12 by the deterioration state prediction means 16 and generates an image 212 of multiple scenarios that allows for comparison of changes in the deterioration levels of parts constituting a specific piece of equipment according to multiple scenarios.

[0030] The acquisition means 14 is a means for acquiring data used for generating a prediction model by the prediction model generation means 15, for predicting the deterioration state of equipment by the degradation state prediction means 16, and for generating an image by the image generation means 17, from the storage unit 12 or the like. The acquisition means 14 may acquire learning data from the storage unit 12 or the like and output it to the prediction model generation means 15. The acquisition means 14 may acquire a prediction model selected by the degradation state prediction means 16 from the storage unit 12 and output it to the degradation state prediction means 16. The acquisition means 14 may acquire information such as the type, location, parts, and renewal cost of equipment from the storage unit 12 and output it to the degradation state prediction means 16. The acquisition means 14 may acquire a prediction result of the degradation state prediction means 16 from the storage unit 12 and output it to the image generation means 17. Furthermore, the acquisition means 14 may acquire a time range for degradation state prediction set by the user, a designation of the first mode or the second mode by the user, and the like, based on the user's operation of the input unit 20, and output them to the degradation state prediction means 16.

[0031] The output means 18 outputs the image generated by the image generating means 17 to the display unit 21 .

[0032] The input unit 20 is an interface that accepts user operations on the information processing device 10. The input unit 20 may be configured with hardware such as a keyboard and a mouse.

[0033] The display unit 21 is an interface for displaying information to the user from the information processing device 10. The display unit 21 may have a hardware configuration such as a liquid crystal display or an OEL (Organic Electro-luminescence) display.

[0034] The information processing device 10 is used to support the creation of an update plan for multiple pieces of equipment included in a plant, and in creating this update plan, the deterioration level of the equipment is predicted. Specifically, predicting the deterioration level means calculating the deterioration level from the present onwards. A prediction model is used to predict the deterioration level of the equipment. Before predicting the deterioration level, the prediction model is generated by machine learning and stored in the storage unit 12.

[0035] 2 is a diagram for explaining the generation of a prediction model. The prediction model generation means 15 acquires indicators, measured values, and actual deterioration states related to the degree of deterioration of equipment in other plants via the acquisition means 14. These indicators and the like are received by the communication unit 11 periodically or at any timing and stored as performance data in the storage unit 12. The prediction model generation means 15 generates and classifies learning data using the performance data read from the storage unit 12 by the acquisition means 14. For example, indicators contained in the performance data may include the type of equipment, installation location, age, operating hours, operating frequency, inspection or investigation status, repair or renewal status, etc., as shown in FIG. 2.

[0036] The type of equipment is an indicator of the equipment's classification, such as a pump, valve, reaction tank aeration device, or flow meter. The installation location is an indicator of where the equipment is located. The installation location may include specific locations, such as sewage treatment plant A or wastewater treatment facility B, and may also include location types, such as indoors, outdoors, or in a corrosive gas environment. The age is an indicator of the number of years since the equipment was installed or updated. The operating time is an indicator of the cumulative operating time since the equipment was installed or updated. The operating frequency is an indicator of how often the equipment operates, such as 12 hours per day. The inspection or inspection status is an indicator of the equipment's inspection status, such as once a month. The repair or update status is an indicator of the equipment's recency, such as whether some parts were repaired one year ago or updated (i.e., replaced with new parts) three years ago. These indicators affect the degree of equipment deterioration. For example, the longer the age of the equipment, the greater the degree of deterioration. Furthermore, equipment with long operating times and high operating frequencies is more susceptible to deterioration than other equipment. Furthermore, for example, frequent inspections and repairs can slow down the deterioration of equipment.

[0037] Here, the prediction model generation means 15 generates multiple prediction models by dividing the prediction models according to some of the indicators included in the performance data. In this embodiment, the prediction model generation means 15 generates multiple prediction models according to the type and installation location of the equipment. That is, a different prediction model is generated for each combination of the type and installation location of the equipment. For example, even for the same type of equipment, the deterioration level over time may vary depending on the installation location, such as whether the equipment is indoors, outdoors, or in a corrosive gas environment. In this embodiment, the prediction model generation means 15 generates different prediction models according to the installation location, thereby providing a prediction model that can predict (calculate) the deterioration state of the equipment with high accuracy. Here, the prediction model generation means 15 may generate multiple prediction models according to at least one of the type and installation location of the equipment. Furthermore, the prediction model generation means 15 may add other indicators to the type and installation location of the equipment and generate different prediction models for each combination of these indicators.

[0038] The prediction model generation means 15 generates training data using the indicators of the performance data as explanatory variables and the actual deterioration state of the equipment as the objective variable. In this embodiment, the equipment healthiness is used as the objective variable. Furthermore, some, but not all, of the indicators exemplified above may be used as explanatory variables. The prediction model generation means 15 uses the generated training data to generate multiple prediction models according to the type and installation location of the equipment through machine learning. The machine learning method is not particularly limited. As an example, a neural network as illustrated in FIG. 2 may be used, or a random forest may be used.

[0039] The predictive model generation means 15 also generates predictive models for the components that make up the equipment. For example, if a specific piece of equipment is composed of components such as a distribution board, a protective relay, an instrument transformer, and a vacuum circuit breaker, a predictive model is generated for each of these components. The method for generating predictive models for components is the same as the method for generating predictive models for equipment. For example, the predictive model generation means 15 generates predictive models for multiple components according to the type and installation location of the components. The same methods for generating learning data and machine learning as those for equipment can be used.

[0040] The prediction model generation means 15 stores the generated prediction models for a plurality of pieces of equipment and a plurality of parts in the storage unit 12. The prediction model generation means 15 also performs management, for example, updating of the prediction models. Before prediction is performed by the degradation state prediction means 16, the prediction model generation means 15 generates a plurality of prediction models and stores them in the storage unit 12.

[0041] The degradation state prediction means 16 selects a prediction model corresponding to a piece of equipment that is a prediction target among a plurality of pieces of equipment included in a plant from a plurality of prediction models stored in the storage unit 12, and predicts the degradation state of the equipment using the selected prediction model. In this embodiment, the degradation state prediction means 16 selects a prediction model based on the type and installation location of the equipment or part that is the prediction target. The degradation state prediction means 16 can predict the degradation state of the equipment with high accuracy by selecting a prediction model appropriate for the type and installation location of the equipment or part from a plurality of prediction models.

[0042] Here, in the first mode, the degradation state prediction means 16 predicts the degradation state for each piece of equipment included in the plant, and in the second mode, performs prediction for each component of specific equipment. The degradation state prediction means 16 acquires, for example, a user's designation of the first mode or the second mode via the acquisition means 14. In the first mode, the degradation state prediction means 16 may target all of the pieces of equipment included in the plant as prediction targets, but when receiving a user instruction via the acquisition means 14 to target only some of the equipment as prediction targets, the degradation state prediction means 16 follows the user's instruction. In the second mode, the degradation state prediction means 16 acquires the specific equipment specified by the user via the acquisition means 14 and targets the specific equipment as prediction targets. At this time, the degradation state prediction means 16 also acquires information on the types and installation locations of components constituting the specific equipment from the memory unit 12.

[0043] The degradation state prediction means 16 may receive a time range specified by the user via the acquisition means 14. The time may be in fiscal years, which are accounting year divisions. The user may input, for example, a time range of approximately 20 years from fiscal year 2021 (see FIG. 4). The user may also input the number of years elapsed from the evaluation point in time as the time range (see FIG. 5). As an example, the number of years elapsed may be set to up to 20 years. The degradation state prediction means 16 predicts the degradation state of the equipment within the specified time range. The degradation state prediction means 16 can obtain a predicted degradation level, which is the output, by, for example, advancing the number of years elapsed, which is the input to the prediction model, from the present within the time range (adding one year at a time). Here, the time range of 20 years is merely an example, and may be set longer or shorter. For example, when a setting screen is displayed on the display unit 21, the user may click on the area for setting the time range with a mouse or the like, and input or change the time range using a keyboard or the like.

[0044] FIG. 3 is a diagram illustrating an example of an equipment map of a plant. Using the plant in FIG. 3 as an example, the predictions made by the degradation state prediction means 16 and the images generated by the image generation means 17 will be described in detail below. The plant shown in FIG. 3 has areas A to E. Areas A, B, and D are indoors. Areas C and E are outdoors. Multiple pieces of equipment are installed in areas A to E. A-1, B-1, etc. indicate the names of the equipment. For example, B-1 and B-3 are indoors but in a corrosive gas environment. For example, C-3 is in an outdoor environment exposed to wind and rain. In the example of FIG. 3, equipment of the same type is colored similarly. For example, A-1 and C-3 are the same type of equipment. For example, B-1 and E-1 are the same type of equipment. For example, B-3, C-1, and D-1 are the same type of equipment.

[0045] First, the first mode, in which prediction is performed on a facility-by-facility basis for multiple pieces of equipment included in a plant, will be described. The degradation state prediction means 16 selects a prediction model from multiple prediction models stored in the storage unit 12 for each piece of equipment A-1 to E-1, depending on the type and location of the equipment, and predicts the degradation state of the equipment over a specified time range using the selected prediction model. For example, A-1 and C-3 are the same type of equipment, but are installed in different locations—indoors and outdoors. Therefore, the degradation state prediction means 16 selects different prediction models for A-1 and C-3. For example, B-1 and E-1 are the same type of equipment, but are installed in different locations—whether or not they are in a corrosive gas environment. Therefore, the degradation state prediction means 16 selects different prediction models for B-1 and E-1. For example, B-3, C-1, and D-1 are the same type of equipment, but are similarly installed in different locations. Therefore, the degradation state prediction means 16 selects different prediction models for B-3, C-1, and D-1. The degradation state predicting means 16 stores the prediction results for each of the facilities A-1 to E-1 in the storage unit 12.

[0046] In the first mode, the image generating means 17 acquires the prediction results stored in the storage unit 12 by the degradation state predicting means 16 and generates an "image 211 indicating the degree of urgency of renewal" that allows for comparison of the changes over time in the degree of deterioration of each of the facilities A-1 to E-1. FIG. 4 is a diagram illustrating an example of the image 211 indicating the degree of urgency of renewal. In the image 211 indicating the degree of urgency of renewal, the health levels of multiple facilities are displayed as a combination of time and the multiple facilities. The image 211 indicating the degree of urgency of renewal displays time on the first axis (the horizontal axis in the example of FIG. 4) and displays multiple facilities on the second axis (the vertical axis in the example of FIG. 4) perpendicular to the first axis. The display items on the vertical and horizontal axes may be reversed. That is, the image 211 indicating the degree of urgency of renewal may display time on the vertical axis and multiple facilities on the horizontal axis. The time (year) displayed in the time display portion 211A corresponds to the time range specified by the user. Furthermore, the names of the plurality of pieces of equipment (A-1 to E-1) are displayed in a display portion 211B for the plurality of pieces of equipment.

[0047] Furthermore, the health levels of each of the multiple pieces of equipment by time (by year) are displayed in different colors depending on the value. In this embodiment, a health level of "5" is displayed in white, "1" is displayed in black, and "4" to "2" are displayed in increasingly darker gray as the value decreases. As another example, a health level of "5" may be displayed in green, "4" in light green, "3" in yellow, "2" in orange, and "1" in red. Here, the display colors are not particularly limited. Furthermore, numbers, which are health level values, do not need to be displayed, but may be displayed as shown in FIG. 4.

[0048] As shown in FIG. 4, the image 211 indicating the degree of urgency of updating allows the user to intuitively understand the period until the annual health level calculated by the degradation state prediction means 16 reaches "2," which indicates the possibility of a functional outage. Furthermore, by displaying each piece of equipment A-1 to E-1 side by side, the user can compare the state of deterioration of each piece of equipment. Also, in FIG. 4, for example, when comparing equipment A-1 and C-3 of the same type, the user can intuitively understand that the health level of C-3, which is installed outdoors, is deteriorating faster than that of A-1, which is installed indoors. Also, in FIG. 4, for example, when comparing equipment B-1 and E-1 of the same type, the user can intuitively understand that the health level of B-1, which is installed indoors but used in a corrosive gas environment, is deteriorating faster than that of E-1, which is installed outdoors.

[0049] In the first mode, the image generation means 17 generates an image 211 indicating the degree of urgency of updating, for example, by the following process. The image generation means 17 acquires the prediction results stored in the memory unit 12 by the degradation state prediction means 16 via the acquisition means 14, extracts information on multiple pieces of equipment, and assigns the information to the second axis of the image 211 indicating the degree of urgency of updating. The image generation means 17 also extracts the health levels within a set time range from the prediction results and assigns colors corresponding to each value. The image generation means 17 assigns the changes in the health levels of each of the multiple pieces of equipment over time to the first axis of the image 211 indicating the degree of urgency of updating, using the assigned colors. By this process, the image generation means 17 can generate the image 211 indicating the degree of urgency of updating shown in FIG. 4.

[0050] Here, the image generating means 17 may generate an image of an equipment map as shown in Fig. 3 and display it together with the image 211 indicating the urgency of the update. In the equipment map image, the positions of multiple pieces of equipment in the plant are indicated by circles of different colors according to the type of equipment. The user can easily understand the type of equipment and the environment where the equipment is installed from the equipment map image.

[0051] The image generation means 17 generates an image of the equipment map by, for example, the following processing. The image generation means 17 acquires, via the acquisition means 14, map information of the entire plant stored in the memory unit 12, as well as information on the positions of multiple pieces of equipment in the plant and the types of equipment. The image generation means 17 maps the position information of the multiple pieces of equipment onto a map of the entire plant and places circles at those positions. The image generation means 17 extracts information on the types of multiple pieces of equipment and assigns colors corresponding to each type. The image generation means 17 colors the assigned colors to circles placed on the map corresponding to each piece of equipment. Through this processing, the image generation means 17 can generate the image of the equipment map shown in FIG. 3.

[0052] Next, a second mode in which prediction is performed for a specific piece of equipment on a part-by-part basis will be described. The degradation state prediction means 16 acquires the specific piece of equipment designated by the user via the acquisition means 14 and sets the specific piece of equipment as the prediction target. For example, if the specific piece of equipment is D-1, the degradation state prediction means 16 acquires information on the types and installation locations of the parts that make up D-1 from the storage unit 12. The degradation state prediction means 16 selects a prediction model from multiple prediction models stored in the storage unit 12 according to the type and installation location of the part, and predicts the degradation state of the equipment for a specified time range using multiple scenarios for each part using the selected prediction model.

[0053] Here, the major components that make up the equipment may include core components that cannot be replaced individually. If a core component fails, all of the major components must be replaced together to continue operating the equipment. In contrast, if a major component that is not a core component fails, it is possible to replace only the failed component.

[0054] For example, suppose a specific piece of equipment, D-1, is composed of four major components. These components are a distribution panel, a protective relay, an instrument transformer, and a vacuum circuit breaker. Of the four major components of D-1, the distribution panel is the core component. If the health rating of the distribution panel drops to "2," all four major components must be replaced to continue operating D-1. On the other hand, if the health rating of one of the protective relay, instrument transformer, or vacuum circuit breaker drops to "2," it is possible to continue operating D-1 by replacing only the component with a health rating of "2."

[0055] In this embodiment, the multiple scenarios include a first scenario and a second scenario. The first scenario is an "update scenario" in which all parts are replaced if the health level of any one part reaches "2." The second scenario is a "life extension scenario" in which, if there is a part whose health level reaches "2," only the part whose health level reaches "2" is replaced. However, even in the second scenario, if the health level of a core part reaches "2," all parts are replaced.

[0056] The degradation state prediction means 16 uses the selected prediction model to predict the degradation state of the equipment for each of the four parts in a specified time range according to a first scenario. The degradation state prediction means 16 stores the prediction results for each of the four parts in the memory unit 12 in association with the first scenario. The degradation state prediction means 16 also uses the same selected prediction model to predict the degradation state of the equipment for each of the four parts in a specified time range according to a second scenario. The degradation state prediction means 16 stores the prediction results for each of the four parts in association with the second scenario in the memory unit 12. The degradation state prediction means 16 may also calculate expenses (renewal costs) in association with the year in which at least one part was replaced. The degradation state prediction means 16 may obtain the renewal costs for each part of the equipment stored in the memory unit 12 and calculate the renewal cost for the renewal year by adding up the renewal costs corresponding to the replaced parts. The deterioration state prediction means 16 may associate the calculated renewal cost with the renewal year and store it in the storage unit 12 as part of the prediction result.

[0057] In the second mode, the image generation means 17 acquires the prediction results stored in the memory unit 12 by the degradation state prediction means 16 and generates a "multiple scenario image 212" that allows comparison of the aging deterioration of a specific piece of equipment according to different scenarios. FIG. 5 is a diagram illustrating an example of the multiple scenario image 212. In the multiple scenario image 212, the health of the specific piece of equipment is displayed as a stacked bar graph of the health of the components that make up the specific piece of equipment, showing changes over time from the time of evaluation, i.e., the present. Here, specific equipment D-1 is composed of four major components, and the health of D-1 can be expressed as the average of the health of each component. Therefore, in the example of FIG. 5, the health of each component is expressed as a value divided by 4 of the total number of components, and the health of D-1 is shown as a stacked bar graph of these values. Hereinafter, the process of adjusting and integrating the health value of each component to display the health of the specific piece of equipment is referred to as normalization.

[0058] In the first scenario, the health rating of the protective relays, voltage transformers, and vacuum circuit breakers reached "2," so all parts were replaced at the fifth year. Although significant replacement costs were incurred at the fifth year, replacement of all parts was not necessary until the facility reached 20 years of age. In the second scenario, the health rating of the protective relays, voltage transformers, and vacuum circuit breakers reached "2," so only these parts were replaced at the fifth year of age. In other words, the distribution board, a core component, was not replaced at the fifth year of age. In the second scenario, the health rating of the distribution board reached "2" at the 14th year of age, so all parts were replaced. In other words, the second scenario reduced replacement costs at the fifth year, but required significant replacement costs at the 14th year. In this way, users can compare multiple scenarios for the deterioration state of specific equipment.

[0059] In the second mode, the image generation means 17 generates images 212 of multiple scenarios by, for example, the following processing. The image generation means 17 acquires the prediction results of each component of the first scenario stored in the storage unit 12 by the degradation state prediction means 16 via the acquisition means 14. The image generation means 17 normalizes the prediction results of each component for each age. The image generation means 17 then assigns a color corresponding to each component for each age and displays the healthiness of the specific equipment as a stacked bar graph. The image generation means 17 also acquires the prediction results of each component of the second scenario stored in the storage unit 12 by the degradation state prediction means 16 via the acquisition means 14, and performs processing similar to that for the first scenario. Through this processing, the image generation means 17 can generate images 212 of multiple scenarios as shown in FIG. 5.

[0060] Here, the image generating means 17 may generate an image 213 of the renewal cost as shown in FIG. 6 and display it as part of the image 212 of the multiple scenarios. When the deterioration state predicting means 16 causes the storage unit 12 to store the calculated renewal cost in association with the renewal year, the image generating means 17 can acquire the renewal cost information. In this case, the image generating means 17 may generate the renewal cost image 213 by generating a bar graph with the number of years elapsed since the present (evaluation point) on the horizontal axis and the renewal cost value on the vertical axis, and arranging the bar graph so that the first scenario and the second scenario can be compared. The renewal cost image 213 may be displayed as part of the images 212 of the multiple scenarios, or may be displayed by switching between the images 212 of the multiple scenarios in accordance with a user instruction. For example, in the first scenario, a renewal cost of 28 million yen is incurred in the fifth year, and in the second scenario, renewal costs of 5.5 million yen and 28 million yen are incurred in the fifth and fourteenth years, respectively. By displaying the image 213 of the renewal cost, the renewal cost can be directly shown to the user, and the user can accurately understand the correspondence with the renewal budget.

[0061] The image generation means 17 may also generate a comparison image of life cycle costs (LCC) as shown in FIG. 7 and display it as part of the image 212 of multiple scenarios. The image generation means 17 may display the comparison image of life cycle costs instead of the image 213 of renewal costs shown in FIG. 6. The image generation means 17 may acquire renewal cost information as in FIG. 6 and generate a bar graph with the number of years elapsed since the installation year or the most recent renewal year on the horizontal axis and the numerical renewal cost values ​​on the vertical axis. The image generation means 17 executes a process of arranging the first and second scenarios for comparison, as in FIG. 6. Furthermore, the image generation means 17 extracts the time until renewal of all parts of the equipment as the number of years of use for each scenario, and defines an evaluation period from the present (evaluation point) equal to the number of years of use and displays it on the image, thereby generating a comparison image of life cycle costs. The comparison image of life cycle costs allows the user to intuitively grasp which scenario is more advantageous in terms of renewal frequency and the life cycle cost of the equipment.

[0062] In the example of FIG. 7 , in the first scenario, the service life is 15 years, and the costs incurred during the evaluation period are 28 million yen for replacement alone. In the second scenario, the service life is 24 years, and the costs incurred during the evaluation period are 33.5 million yen for life extension and replacement combined. In this example, in the first scenario, the cost per year of service life is approximately 1.87 million yen. In contrast, in the second scenario, the cost per year of service life is approximately 1.4 million yen. Therefore, in the example of FIG. 7 , it can be seen that the second scenario is advantageous from the perspective of life cycle cost. Here, the maintenance costs of the equipment may also be taken into consideration. When extracting the service life, the image generation means 17 may multiply the service life by the maintenance costs per year and add the result to the bar graph at the time of replacement. For example, if the maintenance costs per year are 100,000 yen, in the first scenario, the cost per year of service life is approximately 1.97 million yen. In contrast, in the second scenario, the cost per year of use would be approximately 1.5 million yen.

[0063] 8 is a flowchart showing an example of the process of the information processing method executed by the information processing device 10. The information processing device 10 may start the following process when, for example, the user issues an instruction to start via the input unit 20.

[0064] The degradation state prediction means 16 acquires information on equipment that is set as the prediction target among the multiple pieces of equipment included in the plant via the acquisition means 14 (step S1). In the first mode, the degradation state prediction means 16 may set all of the multiple pieces of equipment included in the plant as the prediction targets, but when a user instructs via the acquisition means 14 to set some pieces of equipment as the prediction targets, the degradation state prediction means 16 follows the user's instruction. In the second mode, the degradation state prediction means 16 acquires specific equipment specified by the user via the acquisition means 14 and sets the specific equipment as the prediction target. The information on the equipment that is set as the prediction target includes at least the type and installation location of the equipment or part, and may further include information such as renewal costs.

[0065] The degradation state predicting means 16 selects an appropriate prediction model from among a plurality of prediction models stored in the storage unit 12 based on the type and installation location of the equipment or part that is the prediction target (step S2).

[0066] When the first mode of making predictions on a facility-by-facility basis is designated by the user (Yes in step S3), the degradation state prediction means 16 makes predictions on a facility-by-facility basis for a plurality of facilities included in the plant (step S4). The degradation state prediction means 16 stores the prediction results for each facility in the storage unit 12.

[0067] The image generation means 17 acquires the prediction results stored in the memory unit 12 by the deterioration state prediction means 16 and generates an "image 211 indicating the degree of urgency of renewal" that allows for comparison of the changes over time in the degree of deterioration of multiple pieces of equipment included in the plant (step S5).

[0068] When the user designates the second mode for executing prediction for each component of a specific piece of equipment (No in step S3), the degradation state prediction means 16 predicts the degradation state for each component of a plurality of scenarios (step S6). The degradation state prediction means 16 associates the prediction results for each component with each of the plurality of scenarios and stores them in the storage unit 12.

[0069] The image generation means 17 acquires the prediction results stored in the memory unit 12 by the deterioration state prediction means 16, and generates "images 212 of multiple scenarios" that allow comparison of the changes in the degree of deterioration of a specific piece of equipment over time when different scenarios are followed (step S7).

[0070] The output means 18 outputs the image generated by the image generating means 17 to the display unit 21 (step S8), and the series of processes ends.

[0071] As described above, the information processing device 10, the information processing method, and the program according to the present embodiment can predict the state of deterioration of equipment with high accuracy by using the above-described configuration.

[0072] Although the present disclosure has been described based on the drawings and examples, it should be noted that those skilled in the art can easily make various modifications and alterations based on the present disclosure. Therefore, it should be noted that these modifications and alterations are included in the scope of the present disclosure. For example, the functions included in each means, step, etc. can be rearranged so as not to be logically inconsistent, and multiple means or steps can be combined or divided into one. [Explanation of symbols]

[0073] 10. Information processing equipment 11 Communications Department 12 Storage section 13 Control Unit 14 Acquisition method 15. Predictive model generation method 16 Deterioration state prediction method 17 Image generation means 18 Output Method 20 Input section 21 Display section 211 Image showing the urgency of the update 211A Time display section 211B Display area for multiple equipment 212 images of multiple scenarios 213 Update Cost Image

Claims

1. a degradation state prediction means for predicting a degradation state of a piece of equipment that is a prediction target among a plurality of pieces of equipment included in the plant and a part that constitutes the piece of equipment, using a prediction model selected from a plurality of prediction models according to the types and installation locations of the plurality of pieces of equipment and the types and installation locations of the parts that constitute the piece of equipment, the degradation state prediction means selects the prediction model corresponding to the prediction target from the plurality of prediction models based on the type and installation location of the equipment that is the prediction target and the types and installation locations of parts that make up the equipment; the plurality of prediction models are generated by machine learning using learning data in which indicators of performance data of equipment of other plants and parts constituting the equipment are used as explanatory variables, and deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables; The deterioration state prediction means has a first mode in which the selected prediction model is used to predict the deterioration state on a per-facility basis, and a second mode in which the deterioration state is predicted on a per-component basis, the second mode being used to predict the deterioration state ...

2. the plurality of prediction models correspond to at least one of the types and installation locations of the plurality of facilities and the types and installation locations of components constituting the facilities, The information processing device according to claim 1 , wherein the installation location includes at least one of a specific location and a type of location.

3. 2. The information processing device according to claim 1, wherein the deterioration state prediction means, in the second mode, predicts the deterioration state in a first scenario in which all of the multiple parts constituting the equipment are replaced when at least one of the multiple parts constituting the equipment is deteriorated, and a second scenario in which the deteriorated parts are replaced but not the non-deteriorated parts.

4. 3. An information processing device comprising: a prediction model generation means for generating the plurality of prediction models used in the deterioration state prediction means of the information processing device according to claim 1 or 2 by machine learning using learning data in which indicators of performance data of equipment of other plants and parts constituting the equipment are used as explanatory variables and the deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables.

5. A prediction model generation method executed by an information processing device, comprising:

3. A method for generating a prediction model, comprising: a step of generating the plurality of prediction models used in the deterioration state prediction means of the information processing device according to claim 1 or 2 by machine learning using learning data in which indicators of performance data on equipment of other plants and parts constituting the equipment are used as explanatory variables, and the deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables.

6. An information processing device 3. A program that causes the plurality of prediction models used in the deterioration state prediction means of the information processing device according to claim 1 or 2 to function as a prediction model generation means that generates the prediction models by machine learning using learning data that uses indicators of performance data on equipment of other plants and parts that constitute the equipment as explanatory variables and the deterioration states of the equipment of the other plants and parts that constitute the equipment as objective variables.

7. 3. An information processing method executed by the information processing device according to claim 1, comprising: An information processing method including a step of predicting a deterioration state of a facility that is a prediction target among a plurality of facilities included in a plant and a component that constitutes the facility, using a prediction model selected from a plurality of prediction models according to the types and installation locations of the plurality of facilities and the types and installation locations of the components that constitute the facility, the prediction model corresponding to the facility and the component that constitutes the facility.

8. a degradation state prediction means for predicting a degradation state of a piece of equipment that is a prediction target among a plurality of pieces of equipment included in the plant and a part that constitutes the piece of equipment, using a prediction model selected from a plurality of prediction models according to the types and installation locations of the plurality of pieces of equipment and the types and installation locations of the parts that constitute the piece of equipment, the degradation state prediction means selects the prediction model corresponding to the prediction target from the plurality of prediction models based on the type and installation location of the equipment that is the prediction target, and the types, installation locations, and type of the installation locations of parts that constitute the equipment; the plurality of prediction models are generated by machine learning using learning data in which indicators of performance data of equipment of other plants and parts constituting the equipment are used as explanatory variables, and deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables; The deterioration state prediction means has a first mode in which the selected prediction model is used to predict the deterioration state on a per-facility basis, and a second mode in which the deterioration state is predicted on a per-component basis, the second mode being used to predict the deterioration state ...

9. a degradation state prediction means for predicting a degradation state of a piece of equipment that is a prediction target among a plurality of pieces of equipment included in the plant and a part that constitutes the piece of equipment, using a prediction model selected from a plurality of prediction models according to the types and installation locations of the plurality of pieces of equipment and the types and installation locations of the parts that constitute the piece of equipment, the degradation state prediction means selects the prediction model corresponding to the prediction target from the plurality of prediction models based on the type and installation location of the equipment that is the prediction target and the types and installation locations of parts that make up the equipment; the plurality of prediction models are generated by machine learning using learning data in which indicators of performance data of equipment of other plants and parts constituting the equipment are used as explanatory variables, and deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables; The deterioration state prediction means has a first mode in which the selected prediction model is used to predict the deterioration state on a per-facility basis, and a second mode in which the deterioration state is predicted on a per-component basis that constitutes a specific piece of equipment, and in the second mode, the information processing device predicts the renewal costs due to part replacement based on the prediction of the deterioration state on a per-component basis.

10. a degradation state prediction means for predicting a degradation state of a piece of equipment that is a prediction target among a plurality of pieces of equipment included in the plant and a part that constitutes the piece of equipment, using a prediction model selected from a plurality of prediction models according to the types and installation locations of the plurality of pieces of equipment and the types and installation locations of the parts that constitute the piece of equipment, the degradation state prediction means selects the prediction model corresponding to the prediction target from the plurality of prediction models based on the type and installation location of the equipment that is the prediction target and the types and installation locations of parts that make up the equipment; the plurality of prediction models are generated by machine learning using learning data in which indicators of performance data of equipment of other plants and parts constituting the equipment are used as explanatory variables, and deterioration states of the equipment of other plants and parts constituting the equipment are used as objective variables; The deterioration state prediction means has a first mode for predicting the deterioration state on a per-facility basis using the selected prediction model, and a second mode for predicting the deterioration state on a per-component basis constituting a specific piece of equipment, and in the second mode, the information processing device predicts the renewal costs for part replacement based on a prediction of the deterioration state of core parts that are replaced along with other parts when they deteriorate, and non-core parts that are not replaced along with other parts, among the multiple parts constituting the equipment.

11. An information processing device A program that causes the degradation state prediction means according to any one of claims 1 to 10 to function.

Images