Maintenance work support apparatus, method, program, and system

Abstract

To automatically generate a survival curve of equipment on the basis of inspection records described in a natural language, thereby improving efficiency of maintenance work.SOLUTION: A maintenance work support apparatus includes: an inspection record acquisition unit which acquires inspection record data described in a natural language and formed by associating dates with inspection records of equipment; a degradation event estimation unit which estimates degradation event data in which the dates are associated with degradation events indicating degradation conditions on the basis of the inspection record data; an equipment information estimation unit which estimates equipment information data in which the dates are associated with equipment information of the equipment on the basis of the inspection record data; and a survival curve generation unit which generates a survival curve indicating the transition of degradation with time regarding the equipment, on the basis of the degradation event data and the equipment information data.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] Embodiments of the present invention relate to a maintenance work support device, method, program, and system.

Background Art

[0002] Conventionally, in the maintenance of equipment and the like, maintenance personnel inspected the equipment and the like, and recorded the inspection results in natural language. Then, based on the inspection results, the deterioration state of the equipment and the like was classified and analyzed manually.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when classifying and analyzing the deterioration state manually from records described in natural language, variations due to humans may occur in the results. In addition, the cost also increases to process the records accumulated daily. For this reason, improvement in the efficiency of maintenance work has been desired.

[0005] An embodiment of the present invention aims to provide a maintenance work support device, method, program, and system capable of improving the efficiency of maintenance work using inspection records described in natural language.

Means for Solving the Problems

[0006] The maintenance work support device according to the embodiment includes: an inspection record acquisition unit that acquires inspection record data in which a date and an inspection record of a device are associated with each other in natural language; a deterioration event estimation unit that estimates deterioration event data in which the date and a deterioration event indicating the deterioration status of the device are associated with each other based on the inspection record data; a device information estimation unit that estimates device information data in which the date and the device information of the device are associated with each other based on the inspection record data; and a survival curve generation unit that generates a survival curve indicating the transition of deterioration over time for the device based on the deterioration event data and the device information data.

[0007] As a result, based on the inspection record described in natural language, the survival curve of the device can be automatically generated, so that the efficiency of the maintenance work can be improved.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

DETAILED DESCRIPTION OF THE INVENTION

[0009] [Embodiment 1] Hereinafter, Embodiment 1 will be described with reference to FIGS. 1 to 7. The maintenance work support apparatus 1a according to Embodiment 1 is an apparatus for supporting the maintenance management of equipment. The equipment to be subject to maintenance management is, for example, equipment that is delivered from a manufacturer to a customer and operates at the customer, or components that make up the equipment. An inspector from the manufacturer visits the customer regularly for the purpose of maintaining and managing the equipment and inspects the equipment. The inspector describes the inspection record of the equipment in natural language. The maintenance work support apparatus 1a is configured to be able to output the survival curve of the equipment based on the inspection record described in natural language.

[0010] The maintenance work support apparatus 1a is a computer including a CPU (not shown), a storage device, a communication I / F, an input unit, a display unit, and the like. The CPU is a processor that controls the overall operation of the maintenance work support apparatus 1a. The storage device is, for example, a ROM or the like, and stores various data and programs necessary to realize various functions in the maintenance work support apparatus 1a. The communication I / F is an interface for communicating with an external device (not shown) or the like. The input unit is an input device capable of inputting various instructions by an administrator, and is, for example, a keyboard, a mouse, or the like. The display unit is a display device such as a liquid crystal display.

[0011] FIG. 1 is a block diagram showing an example of the functional configuration of the maintenance work support device 1a according to Embodiment 1. As shown in FIG. 1, the maintenance work support device 1a includes, as functional components, an inspection record acquisition unit 101, a deterioration event estimation unit 102, a device information estimation unit 103, a storage unit 104, a management information acquisition unit 105, and a survival curve generation unit 106. Each of these functional units can be realized by a CPU (not shown) of the maintenance work support device 1a reading and executing a program stored in a storage device.

[0012] The inspection record acquisition unit 101 acquires inspection record data Rp. The inspection record data Rp is data that associates a date with an inspection record of a device. Here, the inspection record is described in natural language.

[0013] FIG. 2 is a diagram showing an example of the inspection record data Rp according to Embodiment 1. As shown in FIG. 2, a plurality of inspection record data Rp are registered with the inspection date and the inspection record associated therewith. The inspection record is text data that describes information about the device to be inspected in natural language as shown in FIG. 2. The inspection record is created by an inspector who has performed the inspection of the device and is transmitted from, for example, a terminal device held by the inspector to the maintenance work support device 1a. The inspection record acquisition unit 101 stores the acquired inspection record in the storage unit 104 as inspection record data Rp associated with a record number that uniquely identifies the inspection record and the inspection date.

[0014] The deterioration event estimation unit 102 estimates deterioration event data including a deterioration event of the device based on the inspection record data Rp. The deterioration event of the device indicates the state of deterioration. For example, the deterioration event includes various states related to the device, such as malfunctions of the device. The deterioration event estimation unit 102 can estimate the deterioration event using a method designated by the administrator from among morphological analysis or deep learning.

[0015] For example, the deterioration event estimation unit 102 estimates a deterioration event using morphemes obtained by performing morphological analysis on the inspection record of the inspection record data Rp. Specifically, the deterioration event estimation unit 102 prepares N types (N is an integer of 1 or more) of deterioration event classifications for the morphemes in advance. The deterioration event estimation unit 102 extracts morphemes related to the deterioration event (that is, nouns, adjectives, verbs, etc.) from the inspection record data Rp, and estimates the deterioration event corresponding to the morpheme from the occurrence frequency, occurrence times, etc. of the extracted morpheme. The deterioration event estimation unit 102 outputs the estimated deterioration event in association with the record number and inspection date of the inspection record as deterioration event data, and stores it in the storage unit 104.

[0016] FIG. 3 is a diagram showing an example of the deterioration event data estimated by the deterioration event estimation unit 102 of Embodiment 1. The deterioration event estimation unit 102 performs the above-described estimation process of the deterioration event for each of the inspection records of the inspection record data Rp acquired by the inspection record acquisition unit 101. That is, FIG. 3 corresponds one-to-one with FIG. 2.

[0017] Specifically, the deterioration event estimation unit 102 extracts the morpheme "gear" from the inspection record of the inspection record data Rp with the record number #1 in FIG. 2. The deterioration event estimation unit 102 estimates "idling of the preloaded gear" as a deterioration event based on the N types of deterioration event classifications prepared in advance for the morpheme "gear" and the occurrence frequency, occurrence times, etc. of "gear".

[0018] For example, the deterioration event estimation unit 102 can estimate a deterioration event using a learned model learned by deep learning. Specifically, the deterioration event estimation unit 102 uses a deterioration estimation model M, which is a learned model stored in the storage unit 104, for estimating the deterioration event. The deterioration estimation model M is a learned model that inputs an inspection record and outputs an estimated deterioration event. The deterioration estimation model M is constituted by, for example, a neural network and is learned and constructed by machine learning such as deep learning. The deterioration event estimation unit 102 inputs the inspection record acquired by the inspection record acquisition unit 101 into the deterioration estimation model M, acquires the deterioration event output from the deterioration estimation model M, and estimates deterioration event data by associating the acquired deterioration event with the inspection date.

[0019] The device information estimation unit 103 estimates device information data including the device information of the device based on the inspection record data Rp. The device information is information regarding the device and includes, for example, various types of information that can identify the device, such as the model number, model name, and serial number of the device. The device information estimation unit 103 estimates the device information by the same method as the method for estimating the deterioration event used in the deterioration event estimation unit 102.

[0020] For example, when a morphological analysis method is used for estimating the deterioration event, the device information estimation unit 103 estimates the device information using the morphemes obtained by performing morphological analysis on the inspection record of the inspection record data Rp. FIG. 4 is a diagram showing an example of the device information data estimated by the device information estimation unit 103 of the first embodiment.

[0021] Specifically, as shown in FIG. 4, the device information estimation unit 103 extracts a morpheme "Type A ACB" as a morpheme by regular expression from the inspection record of the record number #1 of the inspection record data Rp in FIG. 1. The device information estimation unit 103 estimates "Type A ACB" as device information based on the device information classification prepared in advance for the morpheme "Type A ACB", the occurrence frequency of "Type A ACB", the number of occurrences, etc. The device information estimation unit 103 outputs the estimated device information as device information data by associating it with the record number and inspection date of the inspection record data Rp. That is, FIG. 4 corresponds one-to-one with FIG. 2.

[0022] Also, for example, the device information estimation unit 103 can also estimate device information using the learned model. At this time, the device information estimation unit 103 uses the device estimation model N for the estimation of device information. The device information estimation unit 103 inputs the inspection record data Rp acquired by the inspection record acquisition unit 101 into the device estimation model N, and acquires the device information output from the device estimation model N, thereby estimating the device information. The device estimation model N is a learned model that inputs an inspection record and outputs the device information to be estimated. The device estimation model N is constituted by, for example, a neural network and is learned and constructed by machine learning such as deep learning. The device information estimation unit 103 performs the above-described device information estimation process for each of the inspection record data Rp acquired by the inspection record acquisition unit 101. The device information estimation unit 103 associates the estimated device information with the record number and inspection date of the inspection record data Rp, and outputs it to the storage unit 104 as device information data.

[0023] In addition, when the device information cannot be estimated by the estimation method using morphological analysis or deep learning, the device information estimation unit 103 outputs information indicating that the device information cannot be estimated to the storage unit 104 in association with the record number. In the example of FIG. 4, "-" corresponds to the information indicating that the device information cannot be estimated.

[0024] The storage unit 104 stores various data such as the inspection record data Rp, the deterioration event data and device information data inferred based on the inspection record data Rp, the deterioration estimation model M, the device estimation model N, and various information such as programs.

[0025] The management information acquisition unit 105 acquires device management information in which device information and management information of the device are associated from the management information database 500. Here, the management information of the device is information regarding the installation location of the device and test information of the device. The information regarding the installation location of the device corresponds to an installation location ID, company / organization name, address, installation location, etc., and is registered in the installation location information table 502 described later. The test information of the device is various information regarding the test of the device, and for example, corresponds to a test information ID, test date, test number, implementation details, etc. The test information of the device is registered in the test information table 503 described later. The management information database 500 is stored in an external device (not shown in the figure), for example, and includes a device information table 501, an installation location information table 502, and a test information table 503.

[0026] Figure 5 is a diagram showing an example of the data structures of the device information table 501, the installation location information table 502, and the test information table 503 according to Embodiment 1. In the device information table 501, a device information ID is associated with device information such as a device model number, model name, serial number, etc. and registered. In the installation location information table 502, an installation location ID is associated with information regarding the installation location of the device such as company / organization name, address, installation location, etc. and registered. In the test information table 503, a test information ID is associated with test information of the device such as test date, test number, implementation details, etc. and registered. Further, in the test information table 503, for each test information ID, the device information ID corresponding to the test information and the installation location information ID corresponding to the installation location information are registered.

[0027] The survival curve generation unit 106 generates a survival curve regarding the device based on the deterioration event data and the device information data. Here, the survival curve is a graph showing the transition of deterioration over time. The survival curve generation unit 106 generates a survival curve using the Kaplan-Meier method. In generating the survival curve by the Kaplan-Meier method, the survival curve generation unit 106 creates a survival curve from the number of operating days of the device and the survival probability of the device with respect to the number of operating days.

[0028] Specifically, the survival curve generation unit 106 calculates the operating days of the device based on the device management information acquired by the management information acquisition unit 105. More specifically, the survival curve generation unit 106 identifies the test information corresponding to the device information estimated by the device information estimation unit 103 from the device information table 501 and the test information table 503.

[0029] For example, the survival curve generation unit 106 acquires from the test information table 503 the test information corresponding to "Type A ACB" of device information ID #1 in the device information table 501 of FIG. 5. From test information IDs #1 and #1000 corresponding to device information ID #1, it is identified that "Type A ACB" was "newly introduced" on November 1, 2000, and "periodic inspection" was carried out on November 1, 2023. The survival curve generation unit 106 calculates the operating days of "Type A ACB" based on the date and time of "newly introduced" and the date and time of "periodic inspection".

[0030] The survival curve generation unit 106 counts the presence or absence of the device's degradation events based on the results of the estimation processes in the degradation event estimation unit 102 and the device information estimation unit 103. Then, the survival curve generation unit 106 calculates the survival probability with respect to the operating days of the device based on the count result and the operating days of the device, and generates a survival curve.

[0031] For example, for "Device Type A ACB", the survival curve generation unit 106 counts "1" if any degradation event including "idling" has occurred (not survived), and counts "0" if no degradation event has occurred (survived). The survival curve generation unit 106 generates a survival curve by calculating the survival probability of "Type A Accumulator Gear" with respect to the operating days from the count result.

[0032] FIG. 6 is a diagram showing an example of the survival curve generated by the survival curve generation unit 106 of Embodiment 1. Specifically, FIG. 6 shows the survival curves for Device Type A ACB and Device Type B BCB. The horizontal axis represents the operating days, and the vertical axis represents the survival probability (the vertical axis in FIG. 6 shows a ratio, and "1.0" corresponds to a survival probability of 100%).

[0033] As can be seen from FIG. 6, when the number of operating days has passed three years, the survival probability of equipment A type ACB becomes 50%. This indicates that in the third year, some deterioration event has occurred in, for example, 50 out of 100 units of equipment A type ACB. On the other hand, for equipment B type BCB, the survival probability is 75% even in the third year. By referring to such a survival curve, it is possible to determine an appropriate inspection timing for each piece of equipment, such as performing an inspection on equipment A type ACB in the second year and performing an inspection on equipment B type BCB after the third year.

[0034] Note that in FIG. 6, examples of survival curves for each piece of equipment are shown, but the survival curves are not limited to the examples in FIG. 6. The survival curve generation unit 106 may generate, for example, a survival curve for each deterioration event such as "idle rotation of the preloading gear" instead of a survival curve for each model, or alternatively, a survival curve for each combination of a model and a deterioration event such as "idle rotation of the preloading gear of equipment A type ACB".

[0035] Next, the flow of the survival curve generation process will be described with reference to FIG. 7. FIG. 7 is a flowchart for explaining the flow of the survival curve generation process in the maintenance work support device 1a of Embodiment 1.

[0036] The inspection record acquisition unit 101 acquires inspection record data Rp of the equipment (step S111).

[0037] Next, when the administrator designates an estimation method for deterioration events and the like, the deterioration event estimation unit 102 determines whether an estimation method by deep learning has been selected (step S112).

[0038] When the administrator designates the estimation method using deep learning (step S112: YES), the deterioration event estimation unit 102 inputs the inspection record data Rp acquired by the inspection record acquisition unit 101 into each of the deterioration estimation model M and the device estimation model N (step S113). The deterioration event estimation unit 102 acquires the deterioration event output from the deterioration estimation model M, and the device information estimation unit 103 acquires the device information output from the device estimation model N (step S114).

[0039] On the other hand, when the administrator does not designate the estimation method using deep learning (step S112: NO), the deterioration event estimation unit 102 estimates the deterioration event data using the morphemes obtained by performing morphological analysis on the inspection record data Rp (step S115). Next, the device information estimation unit 103 estimates the device information data using the morphemes obtained by performing morphological analysis in the same manner (step S116).

[0040] The management information acquisition unit 105 acquires the device management information from the management information database 500 (step S117).

[0041] Next, the survival curve generation unit 106 calculates the operating days of the device based on the management information acquired by the management information acquisition unit 105 (step S118). The survival curve generation unit 106 calculates the survival probability of the device using the Kaplan-Meier method (step S119). The survival curve generation unit 106 generates a survival curve based on the calculated survival probability (step S120). Thus, the survival curve generation process ends.

[0042] [Summary] According to the maintenance work support device 1a of Embodiment 1, the survival curve generation unit 106 generates a survival curve regarding the device based on the deterioration event data and the device information data estimated from the inspection record data Rp described in natural language.

[0043] Thus, instead of the conventional manual analysis, the survival curve of the device can be automatically generated, eliminating the variation in manual analysis and reducing the human cost. As a result, the efficiency of the maintenance operation is improved.

[0044] In addition, since the appropriate inspection timing of the device can be estimated from the survival curve, the maintenance operation can be performed more efficiently. Specifically, for example, when the survival probability is equal to or higher than a predetermined threshold (e.g., 80%), the inspection frequency can be reduced, and when the survival probability falls below the threshold, the inspection frequency can be increased. Thus, for example, instead of TBM (Time - Based Maintenance), operations such as CBM (Condition - Based Maintenance), RBM (Risk - Based Maintenance), and RCM (Reliability Centered Maintenance) with reduced inspection frequency become possible.

[0045] Alternatively, since the deterioration site of the device can be estimated from the survival curve of the device, inspections can be performed efficiently. Also, since the procurement timing of replacement devices and the like can be adjusted, optimal inventory management is always possible.

[0046] Conventionally, in some cases, sensors and the like are provided on the devices to be inspected, and the state of the device is inspected based on the output from the sensors and the like. In such cases, for example, deterioration and failures that are difficult to detect with sensors, such as "rust generation", "screw loosening", and "minor damage", are often not output as inspection results.

[0047] On the other hand, according to the maintenance operation support device 1a of the present embodiment, since the survival curve is generated based on the inspection record data Rp described in natural language by the inspector, it is possible to automatically output even deterioration and failures that are difficult to detect with sensors as a survival curve. Thereby, the accuracy of maintenance management is enhanced.

[0048] [Embodiment 2] Next, Embodiment 2 will be described with reference to FIGS. 8 to 10. The maintenance work support apparatus according to Embodiment 2 is different from the above-described Embodiment 1 in that it estimates device information using the management information database 500 in addition to the inspection record data Rp. In the following, when having the same configuration and functions as those of the above-described Embodiment 1, the description thereof may be omitted.

[0049] In Embodiment 2, for example, it is assumed that device information data cannot be estimated only from the inspection record data Rp. Specifically, this is the case where the inspection record data Rp does not include a character string from which device information data can be estimated.

[0050] In the maintenance work support apparatus according to Embodiment 2, the management information acquisition unit 105 acquires device management information. Specifically, the management information acquisition unit 105 acquires a device information table 501, an installation location information table 502, and a test information table 503 that are associated with each other from the management information database 500.

[0051] The device information estimation unit 103 estimates device information data based on the inspection record data Rp, the device information table 501, the installation location information table 502, and the test information table 503. Specifically, the device information estimation unit 103 extracts a character string (symbol, character string, etc.) from the inspection record of the inspection record data Rp using, for example, morphological analysis. The device information estimation unit 103 collates the extracted character string with the device information table 501, the installation location information table 502, and the test information table 503. As a result of the collation, for example, when a character string registered in either the installation location information table 502 or the test information table 503 matches the extracted character string, the device information estimation unit 103 estimates the device information corresponding to the management information including the matching character string from the device information table 501.

[0052] The survival curve generation unit 106 generates a survival curve based on the degradation event and the device information data estimated by the above-described estimation process. The survival curve generation unit 106 can generate a survival curve for each company, each installation location, for example, based on the management information corresponding to the estimated device information data.

[0053] Next, the flow of the survival curve generation process will be described with reference to FIG. 8. FIG. 8 is a flowchart for explaining the flow of the survival curve generation process in the maintenance service support apparatus 1 according to the second embodiment.

[0054] The inspection record acquisition unit 101 acquires inspection record data Rp of the device (step S131).

[0055] Next, the management information acquisition unit 105 acquires device management information from the management information database 500 (step S132).

[0056] Next, the device information estimation unit 103 estimates device information data based on the inspection record data Rp and the device management information (step S133).

[0057] Next, when the administrator designates an estimation method for the deterioration event, the deterioration event estimation unit 102 determines whether the estimation method based on deep learning has been selected (step S134).

[0058] When it is determined that the administrator has selected the estimation method based on deep learning (step S134: YES), the deterioration event estimation unit 102 inputs the inspection record data Rp acquired by the inspection record acquisition unit 101 into the deterioration estimation model M (step S135). The deterioration event estimation unit 102 acquires the deterioration event output from the deterioration estimation model M (step S136).

[0059] On the other hand, when it is determined that the administrator has not selected the estimation method based on deep learning (step S134: NO), the deterioration event estimation unit 102 estimates the deterioration event data using the morphemes obtained by performing morphological analysis (step S137).

[0060] Thereafter, the processes of steps S138 to S140 are the same as the processes of steps S118 to 120 in the first embodiment, and thus the description thereof is omitted. Thus, the survival curve generation process of the second embodiment ends.

[0061] [Summary] As described above, according to the maintenance service support device of Embodiment 2, even when it is impossible to estimate device information only from the inspection record data Rp, by using the inspection record data Rp and the device management information in combination, it becomes possible to estimate the device information. As a result, since the estimation accuracy of the device is improved, the generation accuracy of the survival curve is also improved. In addition, according to the maintenance service support device of Embodiment 2, the same effects as those of the above-described Embodiment 1 are achieved.

[0062] [Embodiment 3] Hereinafter, Embodiment 3 will be described with reference to FIGS. 9 to 10. The maintenance service support device 1c of Embodiment 3 is different from the above-described Embodiments 1 and 2 in that it generates a survival curve using on-site measurement information in addition to the deterioration event and device information. In the following, when having the same configuration and functions as those of the above-described Embodiment 1 or 2, the description thereof may be omitted.

[0063] FIG. 9 is a block diagram showing an example of the configuration of the maintenance service support device 1c of Embodiment 3. The maintenance service support device 1c of Embodiment 3 further includes an on-site measurement information acquisition unit 107 as a functional configuration.

[0064] As shown in FIG. 9, the on-site measurement information acquisition unit 107 acquires on-site measurement information 600. The on-site measurement information 600 is information regarding the operating status of the device. More specifically, the on-site measurement information 600 includes information such as the temperature, humidity, voltage, etc. of the location where the device is installed as information regarding the installation environment of the device, and information regarding the number of operations of the device and the operation frequency per unit time as the operation information of the device, as well as deterioration diagnosis information, environmental pollution information, etc. The on-site measurement information 600 is remotely acquired by, for example, a sensor provided in the device. Alternatively, the on-site measurement information 600 may be acquired by an inspector during a regular inspection of the device. The on-site measurement information acquisition unit 107 acquires the on-site measurement information 600 and outputs it to the storage unit 104 in association with the device information.

[0065] The survival curve generation unit 106 generates a survival curve for the device based on the degradation event, device information, and on-site measurement information 600. The survival curve generation unit 106 generates a survival curve using, for example, the COX proportional hazards model. In generating the survival curve using the COX proportional hazards model, the survival curve generation unit 106 takes into account the on-site measurement information 600 of the device in addition to the survival probability and the number of operating days of the device.

[0066] Specifically, for example, the survival curve generation unit 106 can generate a survival curve for each installation environment of the device or for each operating status of the device. Thereby, for example, comparison of the survival curves between a device placed in a high-temperature environment and other devices, or comparison of the survival curves between a device with an annual operation frequency of N times or more and other devices can be performed. As a result, for each installation environment or operating status of the device, an appropriate inspection timing of the device can be estimated, so that the maintenance work can be performed more efficiently.

[0067] In addition to the above-described COX proportional hazards model, the survival curve generation unit 106 can generate a semi-parametric survival curve using (Ridge COX regression model, Lasso COX regression model, Random Survival Forest model, etc.).

[0068] Next, with reference to FIG. 10, the flow of the survival curve generation process will be described. FIG. 10 is a flowchart for explaining the flow of the survival curve generation process in the maintenance work support device 1c of the third embodiment. Note that the processes of steps S141 to S147 in FIG. 10 are the same as the processes of steps S131 to S137 in the second embodiment, and thus the description thereof is omitted.

[0069] In step S148, the on-site measurement information acquisition unit 107 acquires on-site measurement information regarding the operating status of the device.

[0070] The survival curve generation unit 106 calculates the number of operating days of the device from the date of new incorporation to the inspection date based on the device management information acquired by the management information acquisition unit 105 (step S149).

[0071] Next, when the administrator designates a method for calculating the survival probability, the survival curve generation unit 106 determines whether the method for calculating the survival probability using the COX proportional hazards model has been selected (step S150).

[0072] When it is determined that the administrator has selected the method for calculating the survival probability using the COX proportional hazards model (step S150: YES), the survival curve generation unit 106 calculates the survival probability using the COX proportional hazards model (step S151).

[0073] On the other hand, when it is determined that the administrator has not selected the method for calculating the survival probability using the COX proportional hazards model (step S150: NO), the survival curve generation unit 106 calculates the survival probability using a method other than the COX proportional hazards model (step S152). The survival curve generation unit 106 generates a survival curve based on the calculated survival probability (step S153). Thus, the survival curve generation process of Embodiment 3 ends.

[0074] Note that the estimation processes of the device deterioration information and the device information corresponding to steps S142 to S147 in FIG. 10 have been described above as using the same method as in Embodiment 2, but are not limited thereto. For example, the estimation processes of the device deterioration information and the device information may be performed using the method described in Embodiment 1 (corresponding to steps S112 to S116 in FIG. 7).

[0075] [Summary] As described above, according to the maintenance work support device 1c of Embodiment 3, in addition to for each device and for each deterioration event, a survival curve for each installation environment and operating condition of the device can be generated. Thereby, it becomes possible to perform maintenance work with higher accuracy and efficiency. In addition, according to the maintenance work support device 1c of Embodiment 3, the same effects as those of Embodiments 1 and 2 described above are obtained.

[0076] [Other Embodiments] In the above-described embodiment, it has been described that the device information estimation unit 103 estimates device information in the same manner as the method for estimating the degradation event used in the degradation event estimation unit 102, but it is not limited thereto. The device information estimation unit 103 may estimate device information by a method different from the method for estimating the degradation event used in the degradation event estimation unit 102.

[0077] The maintenance operation support device in the above-described embodiment may be any device that can execute various processes related to the support of device maintenance management. For example, it may be a personal computer or a server. Further, the maintenance operation support device is not limited to a physical server and may be a virtual server by software.

[0078] In the above embodiment, hereinafter, the maintenance operation support device 1a has been described as being configured as a single device, but it is not limited thereto. For example, each of the following functional units (inspection record acquisition unit 101, degradation event estimation unit 102, device information estimation unit 103, storage unit 104, management information acquisition unit 105, and survival curve generation unit 106) of the maintenance operation support device 1a, or some of the functional units, or the functions of each unit may be distributed among a plurality of devices on the network and configured as a maintenance operation support system as a whole.

[0079] As described above, the maintenance operation support device of the embodiment includes a processor such as a CPU, a storage device such as a ROM and a RAM, an external storage device such as an HDD, an SSD, and a CD drive device, a display device such as a display device, and an input device such as a keyboard and a mouse, and has a hardware configuration using an ordinary computer.

[0080] The maintenance operation support program executed by the maintenance operation support device of the above-described embodiment is provided by being pre-embedded in a ROM or the like.

[0081] The maintenance operation support program executed by the maintenance operation support device of the above-described embodiment may be configured to be recorded on a computer-readable recording medium such as a CD-ROM, a flexible disk (FD), a CD-R, or a DVD (Digital Versatile Disk) in an installable format or an executable format file and provided.

[0082] Furthermore, the maintenance operation support program executed by the maintenance operation support device of the above-described embodiment may be configured to be stored on a computer connected to a network such as the Internet and downloaded via the network for providing. Also, the maintenance operation support program executed by the maintenance operation support device of the above-described embodiment may be configured to be provided or distributed via a network such as the Internet.

[0083] The maintenance operation support program executed by the maintenance operation support device of the above-described embodiment has a module configuration including the above-described respective functional units. As actual hardware, the CPU (processor) reads the maintenance operation support program from the above ROM and executes it, whereby the above respective functional units (inspection record acquisition unit 101, deterioration event estimation unit 102, device information estimation unit 103, storage unit 104, management information acquisition unit 105, survival curve generation unit 106, on-site measurement information acquisition unit 107) are loaded onto the main storage device, and each functional unit is generated on the main storage device.

[0084] As described above, the embodiments of the present invention have been described. However, the embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention and are included in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0085] 1a, 1c... conservative business support device, 101... inspection record acquisition unit, 102... deterioration event estimation unit, 103... machine information estimation unit, 104... memory unit, 105... management information acquisition unit, 106... survival curve generation unit, 107... on-site measurement information acquisition unit, 500... management information database, Rp... inspection record data.

Claims

1. An inspection record acquisition unit that acquires inspection record data associating a date with an inspection record of a device, described in natural language; A deterioration event estimation unit that estimates deterioration event data associating the date with a deterioration event indicating the deterioration status of the device based on the inspection record data; A device information estimation unit that estimates device information data associating the date with device information of the device based on the inspection record data; A survival curve generation unit that generates a survival curve indicating the transition of deterioration over time for the device based on the deterioration event data and the device information data; A maintenance operation support device comprising: A maintenance operation support device.

2. A management information acquisition unit that acquires device management information in which device information of the device is associated with management information of the device; Further comprising: The device information estimation unit: Estimates the device information data based on the inspection record data and the device management information. The maintenance operation support device according to Claim 1.

3. A local measurement information acquisition unit that acquires local measurement information regarding the operating status of the device; Further comprising: The survival curve generation unit: Generates a survival curve for the device based on the deterioration event data, the device information data, and the local measurement information. The maintenance operation support device according to Claim 2.

4. The deterioration event estimation unit: Estimates the deterioration event data using morphemes obtained by performing morphological analysis on the inspection record. The maintenance operation support device according to Claim 1.

5. A storage unit that stores a deterioration estimation model, which is a learned model that inputs the inspection record data and outputs the deterioration event data; further comprising: The deterioration event estimation unit: Estimates the deterioration event data from the inspection record data using the deterioration estimation model. The maintenance operation support device according to Claim 1.

6. The survival curve generation unit: Generates the survival curve using the Kaplan-Meier method. The maintenance operation support device according to Claim 1.

7. The survival curve generation unit: Generates the survival curve using the Cox proportional hazards model. The maintenance operation support device according to Claim 3.

8. The device information data includes information regarding the model of the device. The management information includes information regarding the installation location of the device. The local measurement information includes information regarding the installation environment of the device. The survival curve generation unit: Generate at least one survival curve for each model of the device, each installation location, and each installation environment based on the deterioration event data, the device information data, and the on-site measurement information. The maintenance service support device according to claim 3.

9. A maintenance service support method executed by a maintenance service support device, comprising: An inspection record acquisition step of acquiring inspection record data in which a date and an inspection record of a device are associated with each other and described in natural language; A deterioration event estimation step of estimating deterioration event data in which the date and a deterioration event indicating the deterioration status of the device are associated with each other based on the inspection record data; A device information estimation step of estimating device information data in which the date and the device information of the device are associated with each other based on the inspection record data; A survival curve generation step of generating a survival curve showing the transition of deterioration over time for the device based on the deterioration event data and the device information data; Including Method.

10. A program for causing a computer of a maintenance service support device to execute, comprising: An inspection record acquisition step of acquiring inspection record data in which a date and an inspection record of a device are associated with each other and described in natural language; A deterioration event estimation step of estimating deterioration event data in which the date and a deterioration event indicating the deterioration status of the device are associated with each other based on the inspection record data; A device information estimation step of estimating device information data in which the date and the device information of the device are associated with each other based on the inspection record data; A survival curve generation step of generating a survival curve showing the transition of deterioration over time for the device based on the deterioration event data and the device information data; A program for causing the computer to execute.

11. A maintenance service support system, comprising: An inspection record acquisition unit that acquires inspection record data in which a date and an inspection record of a device are associated with each other and described in natural language; A deterioration event estimation unit that estimates deterioration event data in which the date and a deterioration event indicating the deterioration status of the device are associated with each other based on the inspection record data; A device information estimation unit that estimates device information data in which the date and the device information of the device are associated with each other based on the inspection record data; A survival curve generation unit that generates a survival curve showing the transition of deterioration over time for the device based on the deterioration event data and the device information data; Comprising System.

Images