Maintenance planning system, maintenance planning method, and failure prediction device

Abstract

To efficiently perform maintenance inspection of a station facility.SOLUTION: A maintenance plane system comprises: a plurality of station facilities 1 provided at a station; an information acquisition unit that acquires information on each station facility 1; a failure prediction device 4 that acquires information from the information acquisition unit and performs failure prediction for each station facility 1 from the information; and a maintenance plan device 12 that generates a maintenance plan from a prediction result made by the failure prediction device 4, where the prediction result is the probability that each station facility 1 will fail during a predetermined first period. The maintenance plan device 12 has a maintenance day sorting unit 53 that selects a station facility 1 to be maintained on each of a plurality of maintenance days, and a movement route generation unit 54 that determines an order of movement to the station facilities 1 to be maintained on each maintenance day and a movement route on a route network.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a maintenance planning system, a maintenance planning method, and a failure prediction device for the maintenance of station facilities. [Background technology]

[0002] Conventionally, the maintenance of ticket gate equipment has been carried out by repairing it whenever a breakdown occurs, and by carrying out maintenance and inspection of each ticket gate equipment every 1 to 6 months. In addition, in order to prevent breakdowns during use of the ticket gate equipment as much as possible and to carry out maintenance and inspection efficiently, the maintenance and inspection cycle has been reviewed from time to time. For example, as shown in Patent Document 1, the operation status and failure history of each ticket gate equipment (ticket gate) are continuously referenced, and the operation status of each ticket gate equipment is manually determined to extract ticket gate equipment that is likely to break down, and the maintenance and inspection cycle is reviewed, such as by bringing forward the timing of the maintenance and inspection of the ticket gate equipment.

[0003] Maintenance inspections are performed by maintenance personnel who visit the ticket gates, and each maintenance personnel performs maintenance inspections on multiple ticket gates in a day. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2004-38747 A Summary of the Invention [Problem to be solved by the invention]

[0005] Since maintenance staff maintain multiple ticket gate devices, they are required to perform maintenance and inspections efficiently.

[0006] An object of the present invention is to efficiently perform maintenance and inspection of station facilities. [Means for solving the problem]

[0007] In order to achieve the above-mentioned object, a maintenance planning system according to one embodiment of the present invention comprises a failure prediction device that acquires information from a plurality of station equipment installed in a station and predicts failure of each of the station equipment from the information, and a maintenance planning device that generates a maintenance plan from the prediction result performed by the failure prediction device, wherein the prediction result is the probability that each of the station equipment will fail during a predetermined first period, and the maintenance planning device has a maintenance day allocation unit that selects the station equipment for which maintenance will be performed on each of a plurality of maintenance dates, and a travel route generation unit that determines the order of travel to the station equipment for which maintenance will be performed on each of the maintenance dates and the travel route on the line network.

[0008] Furthermore, a maintenance planning method according to one embodiment of the present invention has a computer execute the steps of: predicting failures of a plurality of station equipment installed in a station from information acquired for each of the station equipment, and outputting a prediction result which is the probability that each of the station equipment will fail during a predetermined first period; selecting the station equipment to be maintained on a plurality of maintenance dates based on the prediction result; and determining the order of travel to the station equipment to be maintained on each of the maintenance dates and the travel route on the railway network.

[0009] According to the above configuration, by selecting the station equipment to be maintained on each maintenance date based on the prediction results, the maintenance date for each station equipment can be allocated according to the urgency of the maintenance, which increases the likelihood that maintenance will be performed before the station equipment breaks down and prevents the station equipment from breaking down during operation.

[0010] In addition, by determining the order in which station equipment will be moved and the route on the railway network for which maintenance will be performed on each maintenance day, maintenance personnel can efficiently move to the station equipment that requires maintenance, making it possible to perform maintenance efficiently.

[0011] In addition, the failure prediction device may perform the failure prediction for each of the station equipment using a single trained model that has been machine-learned to output the prediction result when the information is input.

[0012] With this configuration, failure prediction can be performed easily and accurately using AI technology, increasing the likelihood that maintenance can be performed before station equipment breaks down, and preventing station equipment from breaking down during operation.

[0013] In addition, the multiple station facilities may include station facilities in which the information formats are different from one another, and the failure prediction device may perform prediction pre-processing to match the formats of the input information when inputting the information into the trained model.

[0014] With this configuration, even if station equipment has different information formats, the prediction preprocessing can ensure consistency of information input to the trained model. Therefore, a single trained model can predict failures of station equipment with different information formats, enabling efficient failure prediction.

[0015] In addition, it is preferable that the failure prediction device includes a learning pre-processing unit that generates learning information by performing learning pre-processing to align the format of the information acquired during a specified learning period, a prediction pre-processing unit that generates prediction information by performing prediction pre-processing to align the format of the information acquired during a specified prediction period, a model acquisition unit that acquires the trained model generated by machine learning using the learning information as input data and failure information when a failure occurs in each of the station facilities during the learning period as teacher data when the prediction information is input, so as to output the prediction result for each of the station facilities, and a failure prediction unit that inputs the prediction information into the trained model and outputs the prediction result for each of the station facilities.

[0016] With this configuration, it is possible to easily and accurately predict failures using AI technology (AI model), which increases the possibility that maintenance can be performed before the station equipment breaks down, and prevents the station equipment from breaking down during operation. Even if the station equipment has different information formats, a common trained model can be used to predict failures of station equipment with different information formats, making it possible to perform failure predictions efficiently.

[0017] In addition, multiple operators each operate multiple station facilities, the station facilities include a mixture of multiple equipment, the preprocessing for learning is performed according to a combination of the operators and the equipment, the trained model is generated for each combination of the operators and the equipment, the preprocessing for prediction is performed according to the combination of the operators and the equipment, and the prediction result is output for each combination of the operators and the equipment.

[0018] The format (data format) of the data acquired from station equipment differs for each operator. In addition, station equipment includes ticket gates, ticket vending machines, fare adjustment machines 1c, and other equipment, and each may have different specifications, data formats, and information formats. In such cases, since the data format differs depending on the operator and the equipment, it becomes difficult to generate an appropriate trained model and predict failures.

[0019] Therefore, it is preferable to perform preprocessing (preprocessing for learning and preprocessing for prediction) for each combination of operator and equipment, and generate a trained model corresponding to each combination of operator and equipment. With the above-mentioned configuration, preprocessing can be performed according to the combination of operator and equipment, a trained model can be generated, and failure prediction can be performed. This makes it possible to generate a trained model with higher accuracy and perform failure prediction with higher accuracy. As a result, the possibility of maintenance being performed before the station equipment breaks down can be improved, and the station equipment can be prevented from breaking down during operation.

[0020] The line network may be made up of one or more railway lines, the station may be a railway station, and the station equipment may be a ticket gate.

[0021] With this configuration, maintenance of ticket gate equipment installed at railway stations can be carried out efficiently.

[0022] Furthermore, a failure prediction device according to an embodiment of the present invention is a failure prediction device that acquires information for each of a plurality of station facilities during a predetermined learning period and a predetermined prediction period, and predicts a failure of each of the station facilities from the information, and includes a learning pre-processing unit that generates learning information by performing learning pre-processing to align the format of the information acquired during the learning period, a prediction pre-processing unit that generates prediction information by performing prediction pre-processing to align the format of the information acquired during the prediction period, and the prediction information is input by machine learning using the learning information as input data and failure information when a failure occurs in each of the station facilities during the learning period as teacher data. the station equipment includes a model acquisition unit that acquires a trained model generated to output a prediction result for each of the station equipment, and a failure prediction unit that inputs the prediction information into the trained model to output the prediction result for each of the station equipment, wherein a plurality of operators each operate a plurality of the station equipment, the station equipment is a mixture of a plurality of equipment, the training preprocessing is performed according to a combination of the operator and the equipment, the trained model is generated for each combination of the operator and the equipment, the prediction preprocessing is performed according to the combination of the operator and the equipment, and the prediction result is output for each combination of the operator and the equipment.

[0023] With this configuration, it is possible to easily and accurately predict failures using AI technology, which increases the likelihood that maintenance will be performed before the station equipment breaks down, and prevents the station equipment from breaking down during operation. Even if the format (data format) of data acquired from the station equipment differs for each operator, or the station equipment has different information formats, preprocessing can be performed according to the combination of operator and equipment, a trained model can be generated, and failure prediction can be performed. This makes it possible to generate a trained model with higher accuracy and perform more accurate failure prediction. As a result, it is possible to increase the likelihood that maintenance will be performed before the station equipment breaks down, and prevents the station equipment from breaking down during operation. [Brief description of the drawings]

[0024] [Figure 1] FIG. 1 is a diagram illustrating a route network. [Diagram 2] 1 is a schematic diagram illustrating a schematic configuration of a maintenance planning system for ticket gate equipment. [Diagram 3] FIG. 2 is a diagram illustrating a configuration for managing maintenance by a maintenance company. [Figure 4] FIG. 1 is a schematic diagram illustrating a detailed configuration of a maintenance planning system for ticket gate equipment. [Diagram 5] FIG. 1 is a diagram illustrating an example of an outline of the process flow for generating a trained model. [Figure 6] FIG. 1 is a diagram illustrating an example of an outline of a process flow for performing failure prediction. [Figure 7] FIG. 1 is a diagram illustrating an example of an outline of a pretreatment flow. [Figure 8] FIG. 10 is a diagram illustrating an example of an outline of a process flow for carrying out a maintenance plan. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] As shown in Figure 1, a railway network has one or more lines (railroad lines) that connect multiple stations S. A network is usually operated by a single railway operator (corresponding to "operator"), and each railway operator operates its own network. In some regions, multiple railway operators may operate their own network.

[0026] Many stations S are equipped with various station facilities. The station facilities are ticket gate equipment such as a ticket gate 1a, ticket issuing equipment such as a ticket vending machine 1b and a fare adjustment machine 1c, and various other facilities (equipment). The ticket gate equipment and ticket issuing equipment are collectively referred to as a ticket gate equipment 1. In the following, the station facilities will be described using the ticket gate equipment 1 as an example.

[0027] When a ticket gate equipment 1 breaks down, an on-call is made by a station employee or the like to notify a maintenance company or the like of the breakdown. During the on-call, information identifying the broken ticket gate equipment 1 and the details of the breakdown are communicated. A person in charge at the maintenance company or the like who receives the on-call goes to the relevant ticket gate equipment 1 and repairs it. Furthermore, in order to prevent breakdowns during operation, each ticket gate equipment 1 is periodically inspected and maintained based on a maintenance plan. Furthermore, the maintenance plan predicts the occurrence of a breakdown in the ticket gate equipment 1, and is appropriately revised depending on the results of the breakdown prediction. A breakdown prediction is made by predicting the probability that the ticket gate equipment 1 will break down in the future from the state of the ticket gate equipment 1.

[0028] As shown in FIG. 2, in order to detect the state of the ticket gate equipment 1, each part constituting the ticket gate equipment 1 is provided with an operation sensor (not shown) that acquires the operation information 6. The operation information 6 includes information (company ID) that identifies the railway operator and information (equipment ID) that identifies the type of equipment the corresponding ticket gate equipment 1 is, such as a ticket gate machine 1a, a ticket vending machine 1b, or a fare adjustment machine 1c. Furthermore, the ticket gate equipment 1 may be provided with an environment sensor (not shown) inside or around it that acquires environment information 7 such as temperature, humidity, illuminance, air pressure, and weather. The operation sensor and the environment sensor function as an information acquisition unit 49 that acquires information from the ticket gate equipment 1. A state indicating the state of the ticket gate equipment 1 is detected from the operation information 6, or information including the operation information 6 and the environment information 7.

[0029] In particular, when it is determined that there is a high probability of failure in the near future and the urgency of maintenance is high, maintenance and inspection of the ticket gate equipment 1 will be performed within a few days. At this time, multiple ticket gate equipment 1 to be subjected to maintenance and inspection on a specific maintenance date are selected taking into consideration at least one of the urgency of the maintenance and inspection and the efficiency of movement to the multiple ticket gate equipment 1 by train, and an efficient movement order and movement route to each ticket gate equipment 1 taking into consideration the line network is determined.

[0030] In this way, by selecting a ticket gate device 1 that requires maintenance and inspection urgently and performing the maintenance and inspection immediately, it is possible to perform maintenance on the ticket gate device 1 before a breakdown occurs, and it is possible to prevent breakdowns in the ticket gate device 1 that is in operation. Also, by determining an efficient movement order and movement route to the selected ticket gate device 1, movement to each ticket gate device 1 can be performed efficiently, and maintenance and inspection can be performed efficiently.

[0031] Next, an outline of the overall configuration of the maintenance planning system for the ticket gate device 1 will be described with reference to FIG. 1 and FIG.

[0032] One or more ticket gate devices 1 are connected to an Internet line 2. In addition, a trained model generation device 3, a failure prediction device 4, a server 5, and a maintenance planning device 12 are connected to the Internet line 2. For example, the trained model generation device 3, the failure prediction device 4, and the server 5 operate on the cloud.

[0033] The server 5 acquires and stores operation information 6 and environmental information 7 over time from each ticket gate device 1 via the Internet line 2. Furthermore, the server 5 acquires and stores on-call data 8 indicating the content of on-calls made at each ticket gate device 1 over time via the Internet line 2. Specifically, the on-call data 8 may be transmitted directly from the ticket gate device 1 to the server 5, or may be collected at a call center and transmitted from the call center to the server 5 or the trained model generation device 3.

[0034] The trained model generation device 3 inputs the operation information 6, the environmental information 7, and the failure information extracted from the on-call data 8 acquired from each ticket gate device 1 into an AI (artificial intelligence) 9 to perform machine learning and generate a trained model 24. Here, the AI ​​9 may be provided on the Internet line 2 or may be provided within the trained model generation device 3. The specific configuration of the trained model generation device 3 will be described in detail later.

[0035] The failure prediction device 4 inputs the operation information 6 and the environmental information 7 acquired from each ticket gate device 1 into the trained model 24, and outputs failure prediction information 36, which is a prediction result for each ticket gate device 1. The specific configuration of the failure prediction device 4 will be described in detail later.

[0036] The maintenance planning device 12 formulates a maintenance plan for each ticket gate device 1, and modifies the maintenance plan based on the failure prediction information 36 so as to increase the maintenance priority of ticket gate devices 1 that are likely to fail as necessary. Furthermore, the maintenance planning device 12 selects which ticket gate devices 1 are to be maintained on which maintenance days among several maintenance days, and further determines the order of movement to the ticket gate devices 1 to be maintained and an efficient movement route on each maintenance day. The specific configuration of the maintenance planning device 12 will be described in detail later.

[0037] As shown in Figures 2 and 3, maintenance of the ticket gate equipment 1 is performed by a maintenance company. Maintenance companies include manufacturers that both manufacture, sell, and maintain the ticket gate equipment 1. Each maintenance company maintains the ticket gate equipment 1 of one or more manufacturers. Each railway operator introduces ticket gate equipment 1 of one or more manufacturers. Therefore, each maintenance company generally performs maintenance on the ticket gate equipment 1 introduced by multiple railway operators. Note that while Figure 3 shows one maintenance company, a maintenance company that performs maintenance is determined for each manufacturer or ticket gate equipment 1, and there are also cases where multiple maintenance companies perform maintenance on the ticket gate equipment 1 equipped by one or more railway operators.

[0038] The ticket gate devices 1 installed at each railway operator and each maintenance company are connected to a maintenance HUB 50. The maintenance HUB 50 corresponds to a server 5, and creates and modifies the maintenance plans performed by each maintenance company.

[0039] The maintenance HUB 50 collectively manages multiple ticket gate devices 1 operated by multiple railway operators and maintained by multiple maintenance companies. That is, the maintenance HUB 50 collects information on these ticket gate devices 1, predicts failures, and creates maintenance plans. In this way, the maintenance HUB 50 collectively manages the ticket gate devices 1 across multiple railway operators and multiple maintenance companies, making it possible to perform maintenance of the ticket gate devices 1 extremely efficiently compared to managing the ticket gate devices 1 for each railway operator and each maintenance company.

[0040] [Trained model generation device] Next, a specific configuration of the trained model generation device 3 and an example of a trained model generation flow will be described using FIGS. 4 and 5 while also referring to FIGS. 1 to 3.

[0041] The trained model generation device 3 includes a data communication unit 13, a model data acquisition unit 14 (corresponding to a "model acquisition unit"), a memory unit 15, a generation control unit 16, a model generation unit 17, and a learning preprocessing unit 18.

[0042] The data communication unit 13 transmits and receives data to and from the ticket gate devices 1 installed at stations S on each line of each railway operator via the Internet line 2. The model data acquisition unit 14 acquires operation information 6 and environmental information 7 acquired from each ticket gate device 1 during a predetermined learning period as model operation information 20 and model environmental information 21 linked to the ticket gate device 1 via the data communication unit 13. The model data acquisition unit 14 also acquires on-call data 8 performed during the predetermined learning period. The model data acquisition unit 14 stores the model operation information 20, the model environmental information 21, and the on-call data 8 in the storage unit 15 (step #1 in FIG. 5).

[0043] The generation control unit 16 includes a processor such as a CPU or a GPU, and controls the operation of the trained model generation device 3, including the model data acquisition unit 14 and the model generation unit 17.

[0044] The learning preprocessing unit 18 performs preprocessing on the model operation information 20 and the model environment information 21 to generate learning information 22. The preprocessing is performed by assigning different coefficients or weightings to the model operation information 20 and the model environment information 21 for each ticket gate device 1, or by calculating unique features (step #2 in FIG. 5). Note that the preprocessing may be performed in any configuration, but may also be performed based on, for example, correlations between specifications, data formats, and other information formats between each ticket gate device 1 learned by machine learning.

[0045] As described above, the ticket gate equipment 1 is a mixture of equipment manufactured by different manufacturers and different models. In addition, the format (data format) of data acquired varies from one railway operator to another. In addition, the ticket gate equipment 1 includes the ticket gate machine 1a, the ticket vending machine 1b, the fare adjustment machine 1c, and other equipment. Therefore, the data acquired from the ticket gate equipment 1 is a mixture of equipment with different information formats such as specifications, data formats, and equipment, and the information formats of the model operation information 20 and the model environment information 21 are not necessarily the same. The differences and the degree of the differences in the information formats of the respective ticket gate equipment 1 can be known in advance. Therefore, the learning preprocessing unit 18 performs preprocessing on the model operation information 20 and the model environment information 21 in order to standardize the information formats of the model operation information 20 and the model environment information 21. For example, even if the manufacturers and models are different and the information formats of the acquired data are different, there is a possibility that the differences can be absorbed by machine learning. In contrast, the differences in the equipment of the ticket gate equipment 1 have a significant impact on the accuracy of machine learning and failure prediction. Similarly, differences in the format of acquired data have a significant effect on the accuracy of machine learning and failure prediction. Therefore, the learning preprocessing unit 18 performs processing for each combination of a railway operator and a ticket gate device 1. Specifically, the learning preprocessing unit 18 performs preprocessing on the model operation information 20 and the model environment information 21 for each combination of a railway operator and a ticket gate device 1, and generates learning information 22 for each combination of a railway operator and a ticket gate device 1. The learning preprocessing unit 18 stores the generated multiple pieces of learning information 22 in the storage unit 15. As a result, appropriate learning information 22 is generated from the model operation information 20 and the model environment information 21 according to the combination of the railway operator and the ticket gate device 1.

[0046] The model generation unit 17 uses the learning information 22 generated by preprocessing the model operation information 20 and the model environment information 21 as input data, inputs the teacher data as on-call data 8 to the AI ​​9 for machine learning, and generates a trained model 24 (step #3 in FIG. 5). At this time, the model generation unit 17 generates an individual trained model 24 for each combination of the railway operator and the ticket gate device 1.

[0047] The trained model 24 is a model used for performing failure prediction to obtain the probability that the ticket gate device 1 will fail within a predetermined prediction period, for example, within seven days. The trained model 24 is, for example, a decision tree. The trained model 24 may be an ensemble model such as XGBoost, Random Forest, LightGBM, CatBoost, or AdaBoost in which multiple decision trees are associated with each other.

[0048] In this way, by performing preprocessing on the model operation information 20 and the model environmental information 21, it is possible to optimize the model operation information 20 and the model environmental information 21 for each combination of railway operator and ticket gate device 1, even for ticket gate devices 1 with different information formats, and it is possible to generate a trained model 24 that can perform failure prediction for each combination of railway operator and ticket gate device 1, even for ticket gate devices 1 with different information formats.

[0049] [Failure prediction device] Next, a specific configuration example of the failure prediction device 4 will be described using FIGS. 4 and 6 while also referring to FIGS. 1 to 3.

[0050] The failure prediction device 4 includes a data communication unit 28, a model acquisition unit 29, a prediction data acquisition unit 30, a storage unit 31, a prediction control unit 32, a failure prediction unit 33, and a prediction preprocessing unit .

[0051] The data communication unit 28 transmits and receives data to and from the ticket gate equipment 1, the trained model generation device 3, etc., via the Internet line 2. The model acquisition unit 29 acquires the trained model 24 used for failure prediction from the trained model generation device 3 via the data communication unit 28, and stores it in the storage unit 31 (step #1 in FIG. 6). The model acquisition unit 29 acquires a plurality of trained models 24 generated for each combination of the railway operator and the ticket gate equipment 1.

[0052] The prediction data acquisition unit 30 acquires the operation information 6 and the environmental information 7 acquired from each ticket gate device 1 during a predetermined prediction period as prediction operation information 34 and prediction environmental information 35 via the data communication unit 28, and stores them in the storage unit 31 (step #2 in FIG. 6). The predetermined prediction period can be set arbitrarily, and is, for example, the day before a failure prediction is made. The prediction control unit 32 includes a processor such as a CPU, and controls the operation of the failure prediction device 4, including the failure prediction unit 33.

[0053] The prediction preprocessing unit 38 performs preprocessing on the prediction operation information 34 and the prediction environment information 35 to generate prediction information 42. The preprocessing is performed by assigning different coefficients or different weights to the prediction operation information 34 and the prediction environment information 35 for each ticket gate device 1, similar to the preprocessing performed by the learning preprocessing unit 18. The prediction preprocessing unit 38 performs preprocessing on the prediction operation information 34 and the prediction environment information 35 for each combination of the railway operator and the ticket gate device 1, and generates prediction information 42 for each combination of the railway operator and the ticket gate device 1 (step #3 in FIG. 6). Note that the preprocessing may be performed in any configuration, but may be performed based on the correlation, for example, by learning the correlation between the specifications and the format of information such as data format between each ticket gate device 1 by machine learning.

[0054] The failure prediction unit 33 inputs the preprocessed prediction operation information 34 and prediction environment information 35 (prediction information 42) into the corresponding trained model 24 to perform failure prediction (step #4 in FIG. 6), stores the prediction result, failure prediction information 36, in the memory unit 31, and then outputs it to the maintenance planning device 12 (step #5 in FIG. 6). In this way, the failure prediction unit 33 can perform failure prediction using a common algorithm / AI model by using the prediction information 42 and trained model 24 generated corresponding to the combination of the railway operator and the ticket gate device 1.

[0055] Failure prediction information 36 indicates the probability that each of the ticket gate devices 1 will fail within a predetermined period (corresponding to the "first period"). If the predetermined period is seven days, the probability that the ticket gate device 1 will fail within seven days is indicated by a value between 0 and 1 for each of the ticket gate devices 1. The closer this value is to 1, the higher the probability of failure.

[0056] In this way, by performing preprocessing on the prediction operation information 34 and the prediction environment information 35, it is possible to unify the prediction operation information 34 and the prediction environment information 35 for each combination of railway operator and ticket gate device 1, even for ticket gate devices 1 with different information formats. Even for ticket gate devices 1 with different information formats, it is possible to perform failure prediction using a common algorithm / AI model and a trained model 24 corresponding to the combination of the railway operator and the ticket gate device 1.

[0057] [Pretreatment] Next, the above-mentioned preprocessing will be described with reference to FIG. 2 and FIG. 4 and FIG. 7. In the following description, the learning preprocessing unit 18 and the prediction preprocessing unit 38 are simply referred to as the preprocessing unit. In the following description, the case where maintenance is performed in a situation where a ticket gate 1a with a device ID of "00" and a ticket vending machine 1b with a device ID of "01" are installed in two railway operators with company IDs of "00" and "01" will be described as an example, but the number of railway operators may be one or more than three, and the ticket gate equipment 1 may also be one or more than three. In addition, the learning preprocessing unit 18 and the prediction preprocessing unit 38 perform similar preprocessing, but may perform different processing, for example, a configuration may be adopted in which the prediction preprocessing is not performed.

[0058] First, the preprocessing unit extracts the company ID and the equipment ID from the operation information 6 (the model operation information 20 or the prediction operation information 34) stored in the storage unit 15 or the storage unit 31.

[0059] The preprocessing unit then checks the company ID to identify the railway operator that operates the corresponding ticket gate device 1 (step #1 in FIG. 7), and checks the device ID to identify whether the corresponding ticket gate device 1 is the ticket gate device 1a, the ticket vending machine 1b, or the fare adjustment machine 1c (step #2 in FIG. 7). For example, company ID "00" corresponds to railway operator A, and company ID "01" corresponds to railway operator B. Device ID "00" corresponds to ticket gate device 1a, and device ID "01" corresponds to ticket vending machine 1b. Furthermore, if necessary, the preprocessing unit may associate device ID "10" with fare adjustment machine 1c, and device ID "11" with another device. The preprocessing unit performs subsequent processing according to the identified combination of railway operator and device.

[0060] Next, the preprocessing unit calls a preprocessing module according to the combination of the railway operator and the equipment (step #3 in FIG. 7). For example, when the company ID is "00" and the equipment ID is "00", the preprocessing unit reads out a preprocessing module specialized for the ticket gate 1a equipped by the railway operator A. When the company ID is "00" and the equipment ID is "01", the preprocessing unit reads out a preprocessing module specialized for the ticket vending machine 1b equipped by the railway operator A, when the company ID is "01" and the equipment ID is "00", the preprocessing unit reads out a preprocessing module specialized for the ticket vending machine 1a equipped by the railway operator B, when the company ID is "01" and the equipment ID is "00", and when the company ID is "01" and the equipment ID is "01", the preprocessing unit reads out a preprocessing module specialized for the ticket vending machine 1b equipped by the railway operator B.

[0061] Each preprocessing module corrects and processes the input data into a format appropriate for the combination of the railway operator and the equipment for generating a trained model and for fault prediction. Each preprocessing module has the same algorithm, and the information, coefficients, and weighting for correcting and processing the input data are optimized according to the combination of the railway operator and the equipment. Specifically, each preprocessing module has a method for calculating features specialized for each combination of the railway operator and the equipment, as well as information, coefficients, weighting, etc. for the calculation. The preprocessing module may be a function block or software, or may simply be a table that lists the information, coefficients, and weighting for correcting and processing the input data.

[0062] Next, the preprocessing unit reads input data stored in the storage unit 15 or the storage unit 31. When generating the trained model 24, the training preprocessing unit 18, which is a preprocessing unit, reads the on-call data 8, the model operation information 20, and the model environment information 21 from the storage unit 15. When predicting a failure, the prediction preprocessing unit 38, which is a preprocessing unit, reads the prediction operation information 34 and the prediction environment information 35 from the storage unit 31 (step #4 in FIG. 7 ).

[0063] Next, the preprocessing unit reads a feature list in which feature values ​​corresponding to the combination of the railway operator and the equipment are listed from a preprocessing module corresponding to the combination of the railway operator and the equipment.

[0064] Here, the feature quantity according to the combination of the railway operator and the equipment is obtained, for example, as follows. First, the preprocessing unit collects at least a part of the operation information 6 (operation information for model 20 or operation information for prediction 34) and the environment information 7 (environment information for model 21 or environment information for prediction 35) for a predetermined period in the past (for example, one to two years) for each combination of the railway operator and the equipment. Next, the preprocessing unit calculates the average, variance, increase rate, and total for each item of all the collected data to generate the feature quantity. Next, the preprocessing unit inputs the generated feature quantity into an AI model for machine learning to generate a provisional trained model. Next, the preprocessing unit uses the AI ​​model to evaluate the provisional trained model and calculate the importance of the generated input data. Then, the preprocessing unit lists a predetermined number of feature quantities with high importance as feature quantities corresponding to the combination of the railway operator and the equipment. Note that the feature quantity listing may be performed by a dedicated function block or software provided separately, instead of the preprocessing unit.

[0065] Finally, the preprocessing unit calculates the features listed for the combination of the railway operator and the equipment from the operation information 6 (operation information for model 20 or operation information for prediction 34) and the environmental information 7 (environmental information for model 21 or environmental information for prediction 35) according to the loaded feature list. This process may be executed by a preprocessing module according to the combination of the railway operator and the equipment. The calculated features become the learning information 22 or prediction information 42 according to the combination of the railway operator and the equipment.

[0066] The pre-processing module is specialized for each railway operator having a different format (data format) of acquired data and for each piece of equipment having different acquired information and specifications. Therefore, the pre-processing unit can calculate feature quantities optimized according to the combination of the railway operator and the equipment, and generate the learning information 22 or the prediction information 42. As a result, the model generation unit 17 can generate an appropriate trained model according to the combination of the railway operator and the equipment. In addition, the failure prediction unit 33 can appropriately perform failure prediction according to the combination of the railway operator and the equipment while performing failure prediction using a common AI model.

[0067] [Maintenance Planning Device] Next, a specific configuration example of the maintenance planning device 12 will be described using FIGS. 4 and 8 while referring to FIGS. 1 to 3. FIG.

[0068] The maintenance planning device 12 includes a data communication unit 51, a maintenance plan control unit 52, a maintenance day allocation unit 53, a movement route generation unit 54, a memory unit 55, and an output unit 59. The maintenance plan control unit 52 includes a processor such as a CPU, and controls the operation of the maintenance planning device 12, such as the maintenance day allocation unit 53 and the movement route generation unit 54.

[0069] The data communication unit 51 transmits and receives data to and from the failure prediction device 4 and the like via the Internet line 2. The data communication unit 51 acquires the failure prediction information 36 from the failure prediction device 4, and stores it in the storage unit 55 (step #1 in FIG. 8).

[0070] The storage unit 55 stores the failure prediction information 36 acquired from the failure prediction device 4. Furthermore, the storage unit 55 stores in advance route information 57 including the arrangement order and positional relationship of the stations S on each route, the connection relationship of each route, operation information such as a timetable, and the like.

[0071] The maintenance date allocation unit 53 selects the ticket gate devices 1 to be maintained on each maintenance date within a predetermined period, taking into consideration the ticket gate devices 1 that require urgent maintenance based on the failure prediction information 36.

[0072] Specifically, the maintenance day allocation unit 53 first extracts the failure probability of each ticket gate device 1 from the failure prediction information 36, which is the prediction result of the failure prediction, by linking it to the ticket gate device 1. The failure probability is the probability that the relevant ticket gate device 1 will fail within a predetermined period (corresponding to the "first period"), and may be, for example, the probability of failure within 7 days, and may include the probability of failure within 7 days and the probability of failure within 3 days (step #2 in FIG. 8).

[0073] Next, the maintenance day allocation unit 53 extracts ticket gate devices 1 with a failure probability equal to or higher than a predetermined value. For example, the maintenance day allocation unit 53 extracts ticket gate devices 1 with a failure probability of 0.95 (95%) or higher within seven days (step #3 in FIG. 8). The failure probability of each ticket gate device 1 indicates the urgency of maintenance, and the higher the failure probability, the more urgent the need for maintenance.

[0074] Next, the maintenance day allocation unit 53 sets each day within a predetermined period (first period) from the next day onwards as a maintenance day, and selects the ticket gate device 1 to be maintained on each maintenance day. This selection is performed for each maintenance company (step #4 in FIG. 8). Note that the maintenance day is not limited to the next day onwards, but may be set on the same day or later, or may be set two or more days later. Furthermore, the maintenance day is not limited to being set over a predetermined period (first period), but may be set within any period.

[0075] The failure prediction information 36 includes information specifying the station S in which the ticket gate device 1 is installed as unique information of each ticket gate device 1. Furthermore, since the failure probability is the probability of failure within seven days (a predetermined period), each day within the seven days is set as a maintenance date. Furthermore, it is determined that the higher the failure probability, the higher the probability of failure at an earlier stage.

[0076] Therefore, the maintenance day allocation unit 53 takes into consideration the failure probability of the extracted ticket gate devices 1, the positional relationship on the line of the station S where the device is installed, and line information 57, and allocates the extracted ticket gate devices 1 to each maintenance day so that the ticket gate devices 1 with a high failure probability will have the earliest possible maintenance date, and so that maintenance personnel from the maintenance company can efficiently move to each ticket gate device 1 on each maintenance day. Note that the maintenance day allocation unit 53 may allocate each ticket gate device 1 to each maintenance day without taking into consideration line information 57, but taking into consideration the positional relationship on a map, such as the latitude and longitude of the station S where each ticket gate device 1 is installed, together with the failure probability.

[0077] Next, the movement route generating unit 54 refers to the line information 57 and determines the movement order to each ticket gate device 1 and the movement route on the line for each maintenance company and each maintenance date (step #5 in FIG. 8).

[0078] The route information 57 includes the arrangement order and positional relationship of stations S on each route, the connection relationship of each route, operation information such as a timetable, etc., so that the travel route generation unit 54 can determine an efficient order and route around stations S where each ticket gate device 1 to be maintained is installed. In addition, by including fares as operation information, it is possible to determine a route taking travel expenses into consideration, and travel expenses can be optimized.

[0079] For example, the travel route generation unit 54 can determine a route by considering the following travel costs and conditions. First, the distance between stations is considered as the travel cost between each station on the line. Furthermore, between stations that are not connected on the line, if the distance between the stations is a predetermined distance (e.g., 300 m) or more, travel between the stations is not considered, and if the distance between the stations is less than the predetermined distance, travel by foot is allowed, and the walking distance is considered as the travel cost between the stations.

[0080] In addition, by receiving an input of the scheduled maintenance time for each ticket gate device 1, the movement route generation unit 54 can determine a more efficient movement order and movement route in consideration of operation information and the like.

[0081] Then, output unit 59 outputs to each maintenance company information on the ticket gate devices 1 to be maintained on each maintenance day, and the movement order and movement route of the ticket gate devices 1 to be maintained on each maintenance day (step #6 in FIG. 8). For example, output unit 59 is a display device that displays information on the ticket gate devices 1 to be maintained on each maintenance day, and the movement order and movement route of the ticket gate devices 1 to be maintained on each maintenance day. Maintenance personnel who have confirmed this information can easily perform efficient maintenance according to the information.

[0082] In this way, by allocating the ticket gate devices 1 to be maintained for each maintenance company and each maintenance date, and determining the order and route of movement between the ticket gate devices 1, it is possible to move efficiently to each ticket gate device 1 while giving priority to maintenance of the ticket gate devices 1 with a high level of urgency, thereby enabling efficient maintenance. Furthermore, the maintenance company can collectively plan maintenance for the ticket gate devices 1 of multiple railway operators without considering differences between railway operators, and can efficiently move to each ticket gate device 1 and perform efficient maintenance. As a result, it is possible to prevent the ticket gate devices 1 from breaking down during operation.

[0083] By implementing the above maintenance plan, in a situation where regular inspections are carried out seven times a year on each of the 2,000 ticket gate devices1 and malfunctions (on-call) occur twice a year, it is expected that the total number of inspections will be reduced by 25% to 35%, and the number of malfunctions will be reduced by 15% to 25%.

[0084] [Another embodiment] (1) In each of the above embodiments, the configurations of the trained model generating device 3, the failure prediction device 4, and the maintenance planning device 12 are not limited to the case of dividing them into functional blocks as shown in FIG. 4, and the functional blocks may be configured arbitrarily as long as the functions of the trained model generating device 3, the failure prediction device 4, and the maintenance planning device 12 can be realized. For example, the functional blocks of the generation control unit 16, the model generating unit 17, and the learning preprocessing unit 18 are not limited to such a configuration, and may be configured by functional blocks in which some or all of the functions of each other are integrated, or may be configured by functional blocks in which each functional block is further subdivided. In each of the above embodiments, some or all of the functions of the trained model generating device 3, the failure prediction device 4, and the maintenance planning device 12 may be configured to be integrated. For example, the failure prediction device 4 may be provided with either or both of the learning preprocessing unit 18 and the model generating unit 17.

[0085] (2) In each of the above embodiments, the trained model generation device 3 may be configured to generate the trained model 24 alone and complete the process, without transmitting the trained model 24 to the failure prediction device 4. Conversely, the failure prediction device 4 may perform failure prediction using a trained model 24 generated separately, rather than using the trained model 24 generated by the trained model generation device 3. Furthermore, the maintenance planning device 12 may perform maintenance planning using failure prediction information 36 predicted separately, rather than using failure prediction information 36 predicted by the failure prediction device 4.

[0086] (3) In each of the above embodiments, the trained model generation method realized by the trained model generation device 3, the failure prediction method realized by the failure prediction device 4, and the maintenance planning method realized by the maintenance planning device 12 may be partly or entirely realized by any configuration, not limited to the above device configuration. In addition, the trained model generation method realized by the trained model generation device 3, the failure prediction method realized by the failure prediction device 4, and the maintenance planning method realized by the maintenance planning device 12 may be partly or entirely realized by a program. The program is stored in any recording device such as the storage unit 15, the storage unit 31, or the storage unit 55, and any processor (corresponding to a computer), such as a CPU or GPU, provided in the generation control unit 16, the prediction control unit 32, or the maintenance planning control unit 52 executes the program.

[0087] (4) In each of the above embodiments, the ticket gate device 1, the trained model generating device 3, the failure prediction device 4, and the maintenance planning device 12 are not limited to being connected to the Internet line 2, and may be configured to be capable of transferring data between them in any configuration. For example, at least some of these devices may be connected to a network line such as an intranet other than the Internet line 2, may be configured to be capable of transmitting and receiving data via a dedicated line such as a LAN, or may be configured to transfer data via a storage medium. Also, the trained model generating device 3, the failure prediction device 4, and the maintenance planning device 12 may be provided in a maintenance HUB 50 (server 5).

[0088] (5) In each of the above embodiments, the configuration is not limited to one in which failure prediction is performed for the entire ticket gate device 1, but may be one in which the ticket gate device 1 is divided into a plurality of blocks and failure prediction is performed for each block.

[0089] (6) In each of the above embodiments, the station S is not limited to a railway station, and may be a bus station or other vehicle station.

[0090] (7) In each of the above embodiments, the ticket gate device 1 is not limited to the ticket gate 1a, but may include various station facilities (equipment) that are installed in the station S and maintained by a maintenance company. Other equipment includes, for example, platform doors, escalators, elevators, etc. [Industrial Applicability]

[0091] The present invention can be applied to a maintenance planning system, a maintenance planning method, and a failure prediction device for station facilities such as ticket gates installed in railway stations and other transportation stations. [Explanation of symbols]

[0092] 1 Ticket gate equipment (station facilities) 3. Trained model generator 4. Failure prediction device 11 Failure prediction information 12 Maintenance planning device 14 Model data acquisition section (model acquisition section) 18 Preprocessing section for learning 24 Pre-trained models 36 Failure prediction information 38 Prediction preprocessing section 53 Maintenance day distribution department 54 Movement path generation unit 57 Route Information

Claims

1. a failure prediction device that acquires information from a plurality of station facilities provided in a station and performs failure prediction for each of the station facilities based on the information; a maintenance planning device that generates a maintenance plan from a prediction result performed by the failure prediction device, the prediction result being a probability that each of the station equipment will fail during a predetermined first period; The maintenance planning device includes: a maintenance date allocation unit that selects the station equipment to be maintained on each of a plurality of maintenance dates; A maintenance planning system having a movement route generation unit that determines a movement order and a movement route on a railway network to the station facilities where maintenance is to be performed on each of the maintenance days.

2. The maintenance planning system according to claim 1, wherein the failure prediction device performs the failure prediction for each of the station facilities using a single trained model that has been machine-learned to output the prediction result when the information is input.

3. The plurality of station facilities include station facilities each having a different format of the information, The maintenance planning system according to claim 2 , wherein the failure prediction device performs a prediction pre-processing for matching the format of the information to be input when inputting the information into the trained model.

4. The failure prediction device includes: a learning preprocessing unit that performs learning preprocessing to align a format of the information acquired during a predetermined learning period and generates learning information; a prediction pre-processing unit that performs the prediction pre-processing to align a format of the information acquired during a predetermined prediction period and generates prediction information; a model acquisition unit that acquires the trained model generated by machine learning using the learning information as input data and failure information when a failure occurs in each of the station facilities during the learning period as teacher data when the prediction information is input, so as to output the prediction result for each of the station facilities; The maintenance planning system according to claim 3 , further comprising a failure prediction unit that inputs the prediction information into the learned model and outputs the prediction results for each of the station facilities.

5. A plurality of operators each operate a plurality of the station facilities, The station facilities are a mixture of multiple devices, The pre-processing for learning is performed according to a combination of the business operator and the device, The trained model is generated for each combination of the business operator and the device, The prediction pre-processing is performed according to a combination of the business operator and the device, The maintenance planning system according to claim 4 , wherein the prediction result is output for each combination of the business operator and the device.

6. The railway network comprises one or more railway lines; The maintenance planning system according to claim 1 , wherein the station is a railway station, and the station equipment is a ticket gate device.

7. a step of predicting a failure of each of a plurality of station facilities provided in a station from information acquired for each of the station facilities, and outputting a prediction result which is a probability that each of the station facilities will fail during a predetermined first period; selecting the station equipment to be maintained for each of a plurality of maintenance dates based on the prediction results; and determining a movement order and a movement route on a railway network to the station facilities to be maintained on each of the maintenance dates.

8. A failure prediction device that acquires information for each of a plurality of station facilities during a predetermined learning period and a predetermined prediction period, and performs failure prediction for each of the station facilities from the information, a learning preprocessing unit that performs learning preprocessing to align a format of the information acquired during the learning period and generates learning information; a prediction pre-processing unit that performs prediction pre-processing to align a format of the information acquired during the prediction period and generates prediction information; a model acquisition unit that acquires a trained model that is generated by machine learning using the learning information as input data and failure information when a failure occurs in each of the station facilities during the learning period as teacher data when the prediction information is input, so as to output a prediction result for each of the station facilities; a failure prediction unit that inputs the prediction information into the trained model and outputs the prediction result for each of the station facilities; A plurality of operators each operate a plurality of the station facilities, The station facilities are a mixture of multiple devices, The pre-processing for learning is performed according to a combination of the business operator and the device, The trained model is generated for each combination of the business operator and the device, The prediction pre-processing is performed according to a combination of the business operator and the device, The failure prediction device outputs the prediction result for each combination of the business operator and the equipment.

Images