Track maintenance assistance system

Abstract

The purpose of the present invention is to provide assistance so as to enable a schedule of maintenance work to be appropriately set in consideration of maintenance work conditions. This track maintenance assistance system assists in maintaining a track on which a railway vehicle travels. The system comprises: an input unit to which is inputted track state information indicating the state of the track, the track state information including any of displacement information pertaining to the track, information indicating the fineness of ballast, information indicating the state of sleepers, or fastening state information pertaining to fastening tools, and to which is inputted maintenance work condition information indicating maintenance work conditions for the track; and a processing unit that selects a plurality of proposed maintenance locations in the track on the basis of the track state information and outputs a maintenance work schedule for the plurality of selected proposed maintenance locations in accordance with a constraint condition obtained from the maintenance work condition information.

Description

Track maintenance support system

[0001] This disclosure relates to techniques for assisting in track maintenance.

[0002] Patent Literature 1 discloses a railway maintenance support method including a data analysis step of predicting the timing of maintenance work based on analytical values ​​obtained by analyzing data and set thresholds, an extraction step of extracting maintenance work plans from past maintenance work contents using predetermined extraction conditions, and a decision support step of presenting the predicted implementation times and the extracted maintenance work plans. Data related to items to be maintained, such as track displacement of rails, is exemplified as data to be analyzed.

[0003] International Publication No. 2015 / 068801

[0004] However, maintenance work is likely to be carried out in consideration of maintenance work conditions such as the time available for work, the travel time to the work site, etc. Simply presenting a predicted implementation time based on data on the items to be maintained and a maintenance work plan extracted from past examples may not allow for an appropriate maintenance work schedule.

[0005] Therefore, an object of the present disclosure is to provide support for appropriately setting a maintenance work schedule in consideration of maintenance work conditions.

[0006] The track maintenance support system supports the maintenance of a track on which railway vehicles run, and includes: an input unit that receives track condition information indicating the condition of the track, the track condition information including any one of track displacement information, information indicating ballast fineness, information indicating the condition of the sleepers, or fastener fastening state information, and maintenance work condition information indicating maintenance work conditions for the track; and a processing unit that selects a plurality of proposed maintenance locations on the track based on the track condition information, and outputs a maintenance work schedule for the selected plurality of proposed maintenance locations in accordance with constraints obtained from the maintenance work condition information.

[0007] According to this track maintenance support system, maintenance work schedules can be appropriately set in consideration of maintenance work conditions.

[0008] FIG. 1 is a block diagram showing the overall configuration of a railway system including a track maintenance support system according to an embodiment. FIG. 2 is a schematic diagram showing sensors provided on a railway vehicle. FIG. 3 is a diagram showing an example of a track path that is the target of the maintenance support system. FIG. 4 is a flowchart showing an example of processing in a processing device of the maintenance support system. FIG. 5 is a diagram showing an example of a display of proposed maintenance locations. FIG. 6 is a diagram showing an example of a maintenance work schedule display. FIG. 7 is a flowchart showing an example of processing for selecting multiple proposed maintenance locations. FIG. 8 is an explanatory diagram showing a state in which a trained model is being retrained. FIG. 9 is an explanatory diagram showing a state in which a learning model is being trained. FIG. 10 is a flowchart showing another example of processing for selecting multiple proposed maintenance locations. FIG. 11 is a flowchart showing an example of processing for generating a maintenance work schedule. FIG. 12 is an explanatory diagram showing a state in which a learning model is being trained. FIG. 13 is an explanatory diagram showing a state in which deep reinforcement learning is being performed.

[0009] 1 is a block diagram showing the overall configuration of a railway system 30 including a track maintenance support system 32. As shown in FIG.

[0010] The railway system 30 is a system for operating the railway vehicles 20, and the track maintenance support system 32 is a system for supporting the maintenance of the track 10 on which the railway vehicles 20 run.

[0011] An example of a track 10 that is a target of support by the maintenance support system 32 will be described. The track 10 is a path that guides a railway vehicle 20 along a predetermined route. In this example, the track 10 includes a first rail 12a and a second rail 12b. The two rails 12a, 12b are fixed onto ballast 16 via sleepers 13.

[0012] The ballast 16 is a track bed that supports the rails 12a, 12b. The ballast 16 includes a plurality of blocks 17 laid on the roadbed. The blocks 17 are, for example, crushed rocks or gravel. The roadbed may be the surface of the land, the lower surface of a tunnel, or the upper surface of a bridge or viaduct. Sleepers 13 are placed on the ballast 16. The sleepers 13 are rectangular parallelepiped members that are interposed between the ballast 16 and the two rails 12a, 12b and support the rails 12a, 12b. That is, a plurality of sleepers 13 are arranged on the ballast 16 in a parallel orientation with a gap between them in the extension direction of the rails 12a, 12b. The two rails 12a, 12b are arranged on the sleepers 13 with a gap between them in the extension direction of the sleepers 13 and in an orientation perpendicular to the sleepers 13. The rails 12a and 12b are fixed to the sleepers 13 by fasteners such as bucklers (see FIG. 3).

[0013] The railway vehicle 20 includes a car body 22 and a bogie 24. The bogie 24 includes a plurality of rotatably supported wheels 25W. The plurality of wheels 25W are guided by two rails 12a, 12b and run on the rails 12a, 12b. The bogie 24 supports the car body 22 from below. As the bogie 24 runs on the track 10, the railway vehicle 20, including the car body 22, runs along the track 10.

[0014] The railway vehicle 20 may be any vehicle that travels on the track 10, including an electric train, a freight train locomotive, a freight car, a passenger train locomotive, or a passenger car. The freight car or passenger car may be a trailer pulled by a locomotive, or may be a powered vehicle that has its own power. The locomotive may be an electric locomotive or an internal combustion engine such as a diesel locomotive. The railway vehicle 20 may be a commercial vehicle for transporting people or cargo, or may be a business vehicle for monitoring track conditions. The railway vehicle 20 may also be a road-rail vehicle that can travel on both tracks and roads.

[0015] As railcars 20 travel on track 10, the condition of track 10 may change.

[0016] The condition of the track 10 may be evaluated, for example, by the displacement of the track 10 including the rails 12 a, 12 b. The displacement of the track 10 may be evaluated, for example, by gauge deviation, alignment deviation, or elevation deviation. When the displacement of the track 10 progresses, maintenance work such as correcting the position or replacing the rails 12 a, 12 b is carried out in order to return the displacement to the initial state.

[0017] The condition of the track 10 may be evaluated by the granulation of the ballast 16. The granulation of the ballast 16 may be evaluated, for example, by a parameter that depends on the size of the agglomerates 17 contained in the ballast 16. When the granulation of the ballast 16 progresses, maintenance work to replace the ballast 16 may be performed.

[0018] The condition of the track 10 may be evaluated based on the condition of the sleepers 13. As deterioration of the sleepers 13 progresses, wear, cracks, etc. may occur in the sleepers 13, so the degree of deterioration of the sleepers 13 can be evaluated by observing the external shape of the sleepers 13. When the degree of deterioration of the sleepers 13 reaches a predetermined degree of deterioration, maintenance work is carried out to replace the sleepers 13.

[0019] The condition of the track 10 may be evaluated by the tightening state of the fasteners 18. The tightening state of the fasteners may be evaluated, for example, by the number of fasteners 18 per unit distance of the track 10. As the number of loosened fasteners 18 increases, the number of fasteners 18 per unit distance of the track 10 decreases. As the number of loosened fasteners 18 increases, maintenance work is performed to tighten or replace the fasteners 18.

[0020] The railway system 30 includes a terminal device 40, a maintenance support system 32, a data server 90, and a maintenance management device 100. The track maintenance support system 32 includes a processing device 70. In addition to the processing device 70, the maintenance support system 32 may include the terminal device 40, the data server 90, or the maintenance management device 100.

[0021] The terminal device 40, the maintenance support system 32, the data server 90, and the maintenance management device 100 are connected to each other via a communication network 38 so as to be able to communicate with each other. Track status information is transmitted from the terminal device 40 to the data server 90. As a result, track status information collection data 92a is accumulated in the data server 90. The maintenance support system 32 can receive the status of each position on the track 10 from the data server 90. The maintenance management device 100 stores maintenance work condition information, which is the condition for performing maintenance. The maintenance support system 32 can receive the maintenance work condition information from the maintenance management device 100. The maintenance support system 32 selects multiple proposed maintenance locations based on the track status information. Furthermore, the maintenance support system 32 determines a maintenance work schedule for the multiple proposed maintenance locations based on the maintenance work condition information.

[0022] The communication network 38 may be a wired or wireless network, or may be a combination of these. The communication network 38 may also be a public communication network or a communication network using dedicated lines.

[0023] It is not essential that the terminal device 40, the maintenance support system 32, the data server 90, and the maintenance management device 100 are connected to each other so as to be able to communicate with each other via the communication network 38. For example, the maintenance support system 32 and the data server 90 may be connected to each other so as to be able to communicate with each other via a dedicated communication line. The maintenance support system 32 may have the functions of the data server 90 or the maintenance management device 100 built in. Furthermore, data correlating the state of the track 10 with a position on the track 10 may be transmitted directly from the terminal device 40 to the maintenance support system 32. In this case, the data server 90 may be omitted.

[0024] The configuration of each part of the railway system 30 will be described.

[0025] The terminal device 40 is a device mounted on the railway vehicle 20. The terminal device 40 is a device for detecting the state of the track 10 when the railway vehicle 20 travels on the track 10 and providing the detection results to the maintenance support system 32.

[0026] For example, the terminal device 40 includes a sensor 42 as a track condition detection unit and a communication device 46 as a transmission unit. The sensor 42 is a sensor that detects the condition of the track 10. The communication device 46 is a communication device that includes a transmission circuit.

[0027] In this embodiment, the terminal device 40 includes a running position detection unit 44. The running position detection unit 44 detects the running position of the railcar 20 on the track 10. The running position is the position of the railcar 20 in the longitudinal direction of the track 10. The running position of the railcar 20 may be a position (e.g., kilometers) based on a fixed position in the longitudinal direction of the track 10 (e.g., the starting point of the track or any station), or may be a position based on an arbitrary position in the longitudinal direction of the track 10. For example, the running position detection unit 44 may include a rotation speed detection sensor that detects the number of rotations of the wheels and output a running distance from any position based on the detection result of the rotation speed detection sensor. A sensor that detects the vehicle speed based on the number of rotations of the railcar 20 is sometimes called a tachometer generator. Because the running distance can be determined by integrating the speed, the running position detection unit 44, which includes a rotation speed detection sensor, may output the speed at regular intervals.

[0028] Furthermore, for example, the running position detection unit 44 may include a GPS (Global Positioning System) receiving unit in a GNSS (Global Navigation Satellite System), and output latitude and longitude information obtained from a signal received by the GPS receiving unit, or a position in the longitudinal direction of the track 10 based on the latitude and longitude information.

[0029] The terminal device 40 may also include a track condition information data generating device 50. The track condition information data generating device 50 is, for example, a computer including a processor configured with electric circuits. The track condition information data generating device 50 generates data that associates the running position detected by the running position detecting unit 44 with the track condition detected by the sensor 42. The data is transmitted to the outside of the railway vehicle 20 via the communication device 46.

[0030] If the maintenance support system 32 includes a terminal device 40 mounted on the railway vehicle 20, the state of the track 10 is detected when the railway vehicle 20 travels on the track 10. Therefore, the state of the track 10 is quickly updated, and it becomes easier to more appropriately select proposed maintenance locations based on the state of the track 10.

[0031] FIG. 2 is a schematic diagram showing a sensor 42 provided on the railway vehicle 20. The sensor 42 may be an image sensor supported from the railway vehicle 20 so as to capture an image of the area below. The imaging range of the sensor 42 may be a range including the rail 12a or 12b. If the imaging range includes the rail 12a or 12b, the position of the rail 12a or 12b in a planar view can be detected based on the captured image. Furthermore, by irradiating the imaging range with slit light and applying the light-section method, the three-dimensional position of the rail 12a or 12b can be detected based on the captured image. The imaging range may also be a range including the sleepers 13. In this case, the degree of deterioration of the sleepers 13 can be determined by determining the similarity between images of normal sleepers 13 and images of deteriorated sleepers 13 with respect to images of the sleepers 13 captured in the imaging range. The number of fasteners 18 per unit distance can be detected by determining the presence or absence of fasteners 18 captured in the images of the sleepers 13. The imaging range may also be a range including the ballast 16. The fineness of the ballast 16 may be detected by extracting the edges of the lumps 17 of the ballast 16 that appear in the imaging range, and determining the size of the lumps 17 from the extracted edges, the edge proportion, or the degree of edge variation.

[0032] The sensor 42 does not have to be an image sensor. For example, if the displacement of the rails 12 a, 12 b is large, it is conceivable that the railway vehicle 20 will sway significantly. Therefore, an acceleration sensor that detects the sway of the railway vehicle 20 may be used as the sensor, and the displacement of the rails 12 a, 12 b may be determined based on the acceleration detected by the acceleration sensor.

[0033] It is not essential that the maintenance support system 32 include a terminal device 40 mounted on the railway vehicle 20. The track condition information may be based on data input as a result of visual inspection by an inspection worker.

[0034] 1 , the data server 90 is a computer equipped with a storage device 92. The data server 90 may be a cloud server. Information associating a position with a track state is transmitted from the terminal device 40 to the data server 90. The data server 90 stores collected data 92a in which the track state is associated with a position on the track 10.

[0035] The data server 90 may store data transmitted from multiple railway vehicles 20. By collecting data transmitted from multiple railway vehicles 20, the data server 90 can comprehensively collect information on the state of the track 10. When multiple data on the state of a specific track 10 are collected, it is advisable to label the data with the detection date and time and store the data. This makes it possible to understand changes over time in the track state at each position on the track 10.

[0036] The maintenance management device 100 is a computer equipped with a storage device 102 that stores maintenance work condition information 102a. The maintenance work condition information 102a is information that indicates the maintenance work conditions for the track 10.

[0037] For example, when maintenance work is performed at multiple locations, the movement route, the order of movement, or the time required for movement is determined by referring to the route of the track 10 on which the railway vehicle 20 runs. Therefore, the route information indicating the route of the track 10 on which the railway vehicle 20 runs is an example of a maintenance work condition.

[0038] Furthermore, the time required for maintenance work is determined based on the amount of work required for the maintenance work and labor amount information provided for the maintenance work. The amount of work can be determined by the type of maintenance work, the work distance, etc. The type of maintenance work is, for example, the type of work, such as moving or replacing the rails 12a and 12b, replacing the sleepers 13, or replacing the ballast 16. The work distance is, for example, the distance of the track 10 to be maintained. The labor amount information provided for the maintenance work is, for example, the number of workers provided for the maintenance work or the work time. For example, for one maintenance location, the time required for maintenance work can be determined by dividing the amount of work by the labor amount. The labor amount information is an example of a maintenance work condition.

[0039] Furthermore, the time when maintenance work can be performed is determined based on the diagram of the track 10. That is, the time period when no railway vehicles 20 are running on the track 10 is determined as the time period when maintenance work can be performed. The diagram information is also an example of a maintenance work condition.

[0040] The maintenance work conditions are registered in the maintenance management device 100 as known information, or as conditions that can be provided for maintenance work. For example, the route and diagram information of the track 10 are registered in the maintenance management device 100 as known information. The route of the track 10 may be changed due to a temporary disruption. Furthermore, the diagram information may be changed due to a temporary change in the running time of the railcars 20. Furthermore, labor amount information may be registered each time according to the number of workers available to work at each date and time.

[0041] The maintenance assistance system 32 downloads data correlating track conditions with positions on the track 10 from the data server 90 and selects multiple proposed maintenance locations. The maintenance assistance system 32 also downloads data including maintenance condition information from the maintenance management device 100 and can determine a maintenance work schedule for the multiple proposed maintenance locations based on the maintenance work condition information.

[0042] In this embodiment, the maintenance support system 32 is provided at any location other than the track 10. For example, the maintenance support system 32 is provided at a management base installed on the ground to monitor the track 10. The maintenance support system 32 may also be provided at a base for performing maintenance work. The maintenance support system 32 may be implemented in a portable terminal device and carried by a worker. If the data server 90 is a cloud server, each processing function of the maintenance support system 32 may be implemented in the cloud server.

[0043] The maintenance support system 32 includes a processing device 70. The processing device 70 is configured by a computer including a processor 72 such as a CPU, a storage device 74, a communication device 76, and the like.

[0044] The processing device 70 is communicatively connected to the data server 90 via a communication device 76. Track condition information indicating the condition of the track 10, including any one of displacement information of the track 10, information indicating the granulation of the ballast 16, information indicating the condition of the sleepers 13, and fastening status information of the fasteners 18, is input to the communication device 76. Note that the track condition information including any one of displacement information of the track 10, information indicating the granulation of the ballast 16, information indicating the condition of the sleepers 13, and fastening status information of the fasteners 18 includes not only the case where the track condition information includes only one of the plurality of pieces of information, but also the case where the track condition information includes multiple types of information from the plurality of pieces of information. Thus, for example, the track condition information may include one or more types of information selected from the group including displacement information of the track 10, information indicating the granulation of the ballast 16, information indicating the condition of the sleepers 13, and fastening status information of the fasteners 18.

[0045] The processing device 70 is communicably connected to the maintenance management device 100 via a communication device 76. Maintenance work condition information indicating the maintenance work conditions for the track 10 is input to the communication device 76.

[0046] The communication device 76 is an example of an input unit including an input circuit. The communication device 76 is an example of an input unit to which track condition information and maintenance work condition information are input. The input unit to which track condition information is input and the input unit to which maintenance work condition information is input may be configured with separate input circuits.

[0047] The processor 72 includes a circuit. The processing device 70 receives collected data 92a stored in the data server 90 via the communication device 76 and stores the data in the storage device 74. The data downloaded to the storage device 74 may be a portion of the collected data 92a in the data server 90 that belongs to the track 10 for which maintenance is to be determined. The processor 72, as a processing unit, executes processing in accordance with the program 74a stored in the storage device 74. As a result, the processor 72 executes processing to select a plurality of proposed maintenance locations on the track 10 based on the track condition information. Information 74c of the selected plurality of proposed maintenance locations is stored in the storage device 74.

[0048] Furthermore, the processing device 70 receives maintenance work condition information stored in the maintenance management device 100 via the communication device 76 and stores the maintenance work condition information 74d in the storage device 74. The processor 72, as a processing unit, executes processing in accordance with the program 74a stored in the storage device 74, thereby executing processing to output a maintenance work schedule for the selected plurality of proposed maintenance locations in accordance with the constraints obtained from the maintenance work condition information 74d. The output here may be output from the processor 72 and used for internal processing in the processing device 70, or may be output to the outside of the processing device 70. The output maintenance work schedule 74e is stored in the storage device 74.

[0049] A display device 78 and an input unit 79 may also be connected to the processing device 70. The display device 78 may be a liquid crystal display device, an organic EL (Electro-Luminescence) display device, or the like. A display device provided on a smartphone, a tablet terminal, or the like may also be used as the display device 78. The input unit 79 accepts various instructions from the user for the processing device 70. The input unit 79 may be a keyboard including a plurality of switches, a mouse, a touch panel, or the like.

[0050] FIG. 3 is a diagram showing an example of a route of the track 10 that is a target of the maintenance support system 32. For ease of explanation, FIG. 3 shows a simplified route. The diagram includes a first route R1, a second route R2, and a third route R3. The first route R1 is a route that passes left and right in FIG. 3. Branch positions A and B are set midway along the first route R1. The second route R2 is a route that detours midway along the first route R1. The second route R2 branches off from the first route R1 at branch positions A and B, and detours between positions A and B on the first route R1. A via position C is set midway along the second route R2. The third route R3 is a route that detours midway along the first route R1. The third route R3 branches off from the first route R1 at branch positions A and B, and detours between positions A and B on the first route R1. The second route R2 and the third route R3 detour on opposite sides of the first route R1. Route positions D and E are set along the third route R3.

[0051] Each time the railcar 20 travels along the first route R1, the second route R2, and the third route R3, the status of each position on the track 10 is detected and stored in a data server 90. The maintenance support system 32 determines whether maintenance is required for each position on the first route R1, the second route R2, and the third route R3 based on the stored status of the track 10. In this case, maintenance may be proposed for multiple scattered positions on the first route R1, the second route R2, and the third route R3. When maintenance locations are scattered along the first route R1, the second route R2, and the third route R3, it is preferable to determine a maintenance work schedule that takes into account maintenance work conditions and enables efficient movement between work locations. The present disclosure relates to a technology for proposing an appropriate maintenance work schedule.

[0052] 4 is a flowchart showing an example of processing in the processing device 70 of the maintenance support system 32. It is assumed that, in performing this processing, the track status information from the data server 90 and the maintenance work condition information from the maintenance management device 100 relating to the track to be subjected to maintenance support have been downloaded in advance.

[0053] After the process starts, in step S1, track condition information is read and maintenance proposal locations are selected based on the track condition information.

[0054] The proposed maintenance location may be specified as a one kilometer location on the track 10 or as a predetermined distance range on the track 10 .

[0055] In the next step S2, the maintenance suggested portion is displayed on the display device 78, for example.

[0056] In the next step S3, the maintenance work condition information is read in. For the plurality of proposed maintenance locations, a maintenance work schedule for the selected plurality of proposed maintenance locations is determined and output in accordance with the constraints obtained from the maintenance work conditions.

[0057] The maintenance work schedule is a schedule that is determined so that work can be carried out efficiently by moving between multiple proposed maintenance locations. For example, the maintenance work schedule is a schedule that determines the dates for work at multiple proposed maintenance locations.

[0058] In the next step S4, the maintenance work schedule is displayed on the display device 78, for example.

[0059] Fig. 5 is a diagram showing an example of displaying suggested maintenance locations. For example, suggested maintenance locations are information in which location information is associated with maintenance items. In Fig. 5, suggested maintenance locations are displayed in list format. Suggested maintenance locations may also be displayed on a route map, as in Fig. 6.

[0060] The location information may be specified by, for example, the line name and the distance in kilometers. The location information may also be specified as a location including a certain range. The maintenance item is determined according to the track condition for which maintenance is determined to be necessary. For example, if it is determined that maintenance is necessary based on the track displacement, rail displacement correction may be associated as the maintenance item. If the track displacement significantly deviates from the reference value, rail replacement may be associated as the maintenance item. Furthermore, if it is determined that maintenance is necessary based on the granularity of the ballast 16, ballast replacement may be associated as the maintenance item. If it is determined that maintenance is necessary based on the number of fasteners 18, fastening work may be associated as the maintenance item. If it is determined that maintenance is necessary based on the condition of the sleepers 13, sleeper replacement may be associated as the maintenance item.

[0061] In the example shown in FIG. 5, an identification code (ID) for identifying the proposed maintenance location is associated with each proposed maintenance location. A priority level is also associated with each proposed maintenance location. The priority level is set according to the priority level of the maintenance work. For example, a high priority level is associated with "A," and a next-lower priority level is associated with "B." Examples of criteria for determining the priority level will be described later.

[0062] It is not necessary for a maintenance proposal to have a priority level associated with it.

[0063] 6 is a diagram showing an example of a display of a maintenance work schedule. For example, the maintenance work schedule is information in which maintenance work locations on the track 10 are associated with the dates of the maintenance work.

[0064] In the example shown in FIG. 6, maintenance work locations are displayed on a route map showing routes R1, R2, and R3. Maintenance work locations may be specified by route name and kilometer distance. Maintenance work locations may also be specified as locations that include a certain range. Maintenance items may be associated with maintenance work locations. As in FIG. 5, maintenance items are maintenance work details determined according to track conditions that are determined to require maintenance. Maintenance work schedules may be determined by date or period.

[0065] 6, arrows 78a indicating the direction of movement between maintenance work locations may be added to the route map. By looking at the arrows 78a, workers can recognize the direction of movement when moving between maintenance work locations. This allows workers to move efficiently in the direction of movement assumed when generating the maintenance work schedule.

[0066] The maintenance work schedule may be displayed in a list format, similar to that shown in FIG.

[0067] An example of a process in which the processing device 70 selects a plurality of proposed maintenance locations on the track 10 based on track condition information will be described with reference to the flowchart shown in FIG.

[0068] In step S11, a determination target is selected from a plurality of target points in a maintenance target area or a maintenance target track.

[0069] In the next step S12, the track condition Ctrack is read in. The track condition Ctrack is, for example, an index value indicating the deterioration state of the track 10. The track condition Ctrack may be, for example, the amount of displacement of the rails 12a, 12b relative to a reference position, an index value indicating the fineness of the ballast 16, an index value indicating the degree of deterioration of the sleepers 13, or a value indicating the number of fasteners 18 per unit distance.

[0070] In the next step S13, the reference value Cth is read in. The reference value Cth is a value that is set in advance by the user of this system and stored in the storage device 74 as a value indicating the necessity of maintenance, for example.

[0071] In the next step S14, the track condition Ctrack is compared with a reference value Cth. For example, if the track condition Ctrack is a value indicating that the larger it is, the more severe the deterioration, then it is determined whether the track condition Ctrack exceeds the reference value Cth. The magnitude relationship for the determination may be reversed depending on the nature of the index value.

[0072] If it is determined that the track condition Ctrack exceeds the reference value Cth, the process proceeds to step S15, and if it is determined that the track condition Ctrack does not exceed the reference value Cth, the process proceeds to step S18.

[0073] In step S15, the location to be judged is registered as a location for which maintenance is proposed.

[0074] In step S18, the location to be determined is registered as a location not requiring maintenance.

[0075] If the track condition Ctrack is equal to the reference value Cth, the process may proceed to either step. The magnitude relationship for the determination may be reversed depending on the nature of the index value. If the track condition Ctrack includes multiple index values, a location may be registered as a maintenance suggested location if it is determined that maintenance is required based on a comparison of any of the index values ​​with the reference value. When registering a maintenance suggested location, it is preferable to register the maintenance work corresponding to the index value that led to the determination that maintenance is required.

[0076] After step S15 or step S18, the process proceeds to step S16. In step S16, it is determined whether or not all target points in the maintenance target area or the maintenance target track have been determined. If it is determined that they have not been determined, the process returns to step S11 and the subsequent processes are repeated. If it is determined that they have been determined, the process proceeds to step S17, and the selection results of the proposed maintenance points are output as proposed maintenance point information 74c. Here, output refers to output from the processor 72, and includes not only output to outside the processing device 70, but also output for processing within the processing device 70. The output proposed maintenance point information 74c is stored in the storage device 74 or displayed on the display device 78.

[0077] After this, the process ends.

[0078] In the above process, the processing device 70 may determine the priority of the proposed maintenance location. For example, the priority may be determined so that the greater the difference between the track condition Ctrack and the reference value Cth, the higher the priority. The priority may be ranked according to the magnitude of the difference between the track condition Ctrack and the reference value Cth. A range of the difference between the track condition Ctrack and the reference value Cth may be associated with each of multiple priority levels, and one of the multiple priority levels may be determined based on the difference between the track condition Ctrack and the reference value Cth.

[0079] The processing device 70 may predict the future state of the track 10 based on the track condition and select multiple maintenance proposal locations on the track 10 based on the predicted future state of the track 10. For example, in FIG. 7 , step S19 may be performed to convert the track condition Ctrack into a future value. That is, if the change in the track condition over time is stored, the deterioration rate of the track condition per unit period can be estimated. The deterioration rate is multiplied by the elapsed time from the time of detection of the track condition to the time of scheduled maintenance to predict the deterioration progression state Cv. Therefore, the future state of the track 10 is predicted by adding the deterioration progression state Cv to the track condition Ctrack at the time of detection. The predicted future state of the track 10 may be compared with a reference value Cth to determine whether maintenance is required. If the scheduled maintenance time is determined as a period, the end of the scheduled maintenance period may be set as the scheduled maintenance time.

[0080] In the above example, the necessity of maintenance is determined by comparing the track condition Ctrack with the reference value Cth.

[0081] 8, the selection of maintenance locations may be performed by inputting track condition information 74b into a trained model 172 that has undergone machine learning to estimate proposed maintenance locations, and outputting the results of inferring multiple proposed maintenance locations. The trained model 172 is configured, for example, by a neural network, and reads out the track condition information 74b and performs an identification process to identify whether or not to propose the location as a maintenance location.

[0082] The processing device 70 may be configured to include a trained model 172 and a retraining program 174 and to function as a reinforcement learning device 170 .

[0083] The trained model 172 is generated, for example, by a learning device 180 configured by a computer having a training model 182 and a model generation unit 184, as shown in FIG. 9 . The model generation unit 184 learns the process of determining whether maintenance is required from track condition information based on a training dataset. The training dataset may be a plurality of datasets combining track condition information and determination results on whether maintenance is required for the track condition information. Then, the trained model 172 that infers whether maintenance is required based on the track condition information is generated from the training dataset.

[0084] Here, as described above, the track condition information included in the training data set includes any one of the displacement information of the track 10, information indicating the granulation of the ballast 16, information indicating the condition of the sleepers 13, and information on the fastening condition of the fasteners 18. If the track condition information includes multiple conditions, it is possible to make a composite inference about the need for maintenance based on multiple track conditions.

[0085] The training data set may include maintenance criteria established by a national or public institution, thereby learning criteria for determining whether maintenance is necessary depending on the circumstances of the country to which the railway system 30 belongs.

[0086] The training data set may include maintenance determination criteria determined by the operator of the railway system 30. In this way, criteria for determining whether maintenance is necessary according to the circumstances of the operator of the railway system 30 are learned.

[0087] The training data set may include regional information, and seasonal or weather information at the time of scheduled maintenance. Regional information is latitude and longitude or subdivisions of the region in which the track 10 is laid. Seasons are periods divided by weather, such as spring, summer, autumn, winter, rainy season, or cold season. Seasons may also include holiday seasons such as Christmas. This allows for learning of criteria for determining whether maintenance is necessary, taking into account regional circumstances, seasons, weather, etc.

[0088] If the trained model 172 is trained based on the above-mentioned maintenance judgment criteria, regional information, and seasonal or weather information at the time of scheduled maintenance in addition to the track condition information, the above-mentioned maintenance judgment criteria, regional information, and seasonal or weather information at the time of scheduled maintenance in addition to the track condition information may be input to the trained model 172. In this way, the necessity of maintenance is determined based on the above-mentioned maintenance judgment criteria, regional information, and seasonal or weather information at the time of scheduled maintenance in addition to the track condition information.

[0089] The reinforcement learning device 170 includes a relearning program 174. The input unit 79 of the system 32 can function as a maintenance necessity judgment input unit that can input maintenance necessity judgment information based on the track condition information input to the trained model 172.

[0090] A user of the present system 32 can examine whether maintenance is necessary by looking at the track condition information and the maintenance necessity result input to the trained model 172. The examination result can be input to the re-learning program 174 via the input unit 79. The re-learning program 174 adjusts parameters that the trained model 172 uses to determine whether maintenance is necessary, based on the track condition information and the examination result input to the trained model 172.

[0091] FIG. 10 is a flowchart showing an example of a process in which the reinforcement learning device 170 selects a plurality of proposed maintenance points on the track 10.

[0092] In step S21, the track state to be judged is read.

[0093] In the next step S22, the track condition information is input to the trained model 172, and a determination result as to whether maintenance is required is output. The determination result is displayed on the display device 78 together with the track condition information. A user of the system 32 checks the track condition information and the determination result, and carefully examines whether maintenance is required for the track 10. The user can carefully examine whether maintenance is required for the track 10 while referring to the determination result.

[0094] The results of the user's review of the system 32 are input to the reinforcement learning device 170 via the input unit 79. In the next step S23, the review is read.

[0095] In the next step S24, it is determined whether maintenance is necessary in view of the inspection results. If it is determined that maintenance is necessary, the process proceeds to step S25, and if it is determined that maintenance is not necessary, the process proceeds to step S27.

[0096] In step S25, the location to be determined is registered as a location where maintenance is proposed. In step S27, the location to be determined is registered as a location where maintenance is not required.

[0097] After step S25 or step S27, the process proceeds to step S26. In step S26, the inspection results are stored in the storage device 74, and the track condition information and the inspection results are input to the re-learning program 174 to re-learn the trained model 172. As a result, the user's inspection results are reflected in the trained model 172.

[0098] This completes the process of determining whether maintenance is necessary for a given location and the re-learning process. In this case, the judgment criteria based on the customs or unique circumstances of the user or business are reflected in the trained model 172.

[0099] The priorities may be presented even when multiple maintenance suggested locations are proposed by inputting the data into the trained model 172, which has undergone machine learning to estimate the maintenance suggested locations. In this case, it is preferable to use a training data set labeled with priority levels. The priorities may also be determined according to the classification score values.

[0100] An example of processing in which the processing device 70 outputs a maintenance work schedule for the selected plurality of proposed maintenance locations in accordance with constraints obtained from the maintenance work condition information will be described with reference to the flowchart shown in FIG.

[0101] In step S31, the maintenance proposal location is read.

[0102] In the next step S32, maintenance work conditions are read in. The maintenance work conditions are conditions that can become constraints when performing maintenance work on a plurality of proposed maintenance locations.

[0103] For example, the route of the track 10 is an example of a maintenance work condition because it restricts the route of a railway vehicle for work when moving between multiple proposed maintenance locations. The route of the track 10 may not be input to the processing device 70 each time a maintenance work schedule is determined, but may be stored in the storage device 74 as a predetermined route. Since the route of the track 10 may be changed due to temporary road disruptions, it may be modified as necessary when determining a maintenance work schedule.

[0104] Furthermore, since maintenance work is performed during times when the railway vehicle 20 is not running, the diagram information of the railway vehicle 20 restricts the maintenance work time. Therefore, the diagram information is also an example of a maintenance work condition. The diagram information determined by the operator of the railway system 30 is provided as the maintenance work condition.

[0105] The time or duration required for maintenance work is determined by the amount of work required for that work and the amount of labor provided for that work. Since the amount of labor limits the time or duration required for maintenance work, the amount of labor is also an example of a maintenance work condition. The amount of labor provided as a maintenance work condition is determined by the number of workers that the business operator can provide at each date and time.

[0106] In the next step S33, a maintenance work schedule is generated based on the plurality of proposed maintenance locations and the labor work conditions. The maintenance work schedule is generated so that the plurality of proposed maintenance locations can be efficiently traveled to and the maintenance work can be efficiently performed. The efficiency may be evaluated based on travel distance or travel time.

[0107] In the next step S34, the generated maintenance work schedule is output. The output maintenance work schedule is displayed on the display device 78 and / or stored in the storage device 74 as maintenance work schedule 74e data.

[0108] The maintenance work schedule generation process can be considered as a process of determining the route order that minimizes travel costs when starting from the maintenance work base station and visiting multiple proposed maintenance locations in accordance with the constraints imposed by the route of the track 10. In other words, the maintenance work schedule generation process can be realized by applying various methods for solving the traveling salesman problem.

[0109] For example, as shown in FIG. 12 , the processing device 70 may generate a maintenance work schedule by inputting a selected plurality of proposed maintenance locations and maintenance work conditions into a trained model 272 that has undergone machine learning to estimate a maintenance work schedule. The machine learning may be an algorithm that performs deep reinforcement learning based on the proposed maintenance locations and the maintenance work conditions, or may be an algorithm that performs machine learning using the proposed maintenance locations, the maintenance work conditions, and the maintenance work schedule as ground truth data. The trained model 272 may be configured, for example, by a model suitable for optimizing a conditional combinatorial problem, such as a pointer network.

[0110] FIG. 13 is a block diagram showing an example of machine learning by the learning device 270. As shown in the figure, an agent 280 performing deep reinforcement learning includes a schedule generation policy determiner 282 and a reinforcement learner 284. Maintenance work conditions, including multiple proposed maintenance locations and route information indicating the route of the track 10 on which the railway vehicle 20 travels, are provided as an environment. The agent 280 determines a maintenance schedule for the multiple proposed maintenance locations, for example, a work order, according to the policy in the schedule generation policy determiner 282 in accordance with the provided environment. By providing the maintenance schedule to the environment, a reward for the maintenance schedule is determined. The reward is evaluated, for example, by the travel time required to move the maintenance locations according to the maintenance schedule. The reward may also be evaluated by the travel distance or travel cost. The reward is provided to the reinforcement learner 284, which then reflects the learning result in the schedule generation policy determiner 282 based on the reward. By repeating the above process, the policy by the schedule generation policy determiner 282 is learned. A trained model 272 is generated based on the parameters of the learning result.

[0111] The processing device 70 may generate a maintenance work schedule by applying an algorithm for solving the traveling salesman problem, such as a branch-and-bound method that combines branching and bounding operations, an approximation algorithm such as a hill-climbing method that searches for the best route based on randomly created routes, or a brute-force algorithm that evaluates all routes.

[0112] The environment may include region- or time-specific information, such as latitude and longitude, subdivision, season, weather, etc. In this case, if the information input to the trained model 272 includes region- or time-specific information, a maintenance work schedule appropriate for the region or time is generated.

[0113] The path of the track 10 among the maintenance work conditions can be provided as a condition for determining the order of maintenance work. On the other hand, the diagram information is a condition for limiting the time required for maintenance work. Furthermore, the labor amount is a condition for specifying the time required for maintenance work.

[0114] The processing device 70 may first determine an order for visiting the selected plurality of proposed maintenance locations based on the selected plurality of proposed maintenance locations and the route of the track 10. After this, the processing device 70 may allocate the required maintenance work time estimated from the amount of labor to the maintenance work time constrained by the diagram information in accordance with the selected order of visiting.

[0115] The diagram information may be considered as a constraint that determines the travel time, and the amount of work may be considered as a constraint that determines the stay time, and the processing device 70 may generate a maintenance work schedule by applying an algorithm that solves the constrained traveling salesman problem.

[0116] In addition, if priorities are set for the selected multiple maintenance proposal locations, the maintenance work schedule may be determined for work locations within a predetermined priority range or for work locations with the same priority level.

[0117] According to the track maintenance support system configured as described above, a plurality of proposed maintenance locations on the track 10 are selected based on the track condition information. Then, a maintenance work schedule for the selected plurality of proposed maintenance locations is output in accordance with constraints obtained from the maintenance work condition information. Therefore, the person who ultimately determines the maintenance work schedule can refer to the output maintenance work schedule and appropriately set the final maintenance work schedule in consideration of the maintenance work conditions.

[0118] Furthermore, if the track condition information includes displacement information of the track 10, information indicating the granulation of the ballast 16, information indicating the state of the sleepers 13, or information on the fastening state of the fasteners 18, it is possible to select, for example, a track 10 that has shifted significantly from its initial state, a location where the granulation of the ballast 16 has progressed, a track location where the sleepers 13 have deteriorated, or a location where a large number of fasteners 18 have fallen off, as a location where maintenance is suggested.

[0119] Furthermore, by determining the priorities of the plurality of proposed maintenance locations, the maintenance work schedule can be output taking the priorities into consideration, which makes it easier to carry out maintenance work according to the priorities.

[0120] Furthermore, by predicting the future state of the track 10 based on the track condition information and selecting multiple maintenance suggested locations on the track based on the predicted future state of the track 10, the maintenance suggested locations can be appropriately selected based on the predicted future state of the track 10. For example, maintenance can be performed first on locations with a fast rate of deterioration and later on locations with a slow rate of deterioration, rather than based only on the state at the time of detection or the current state.

[0121] Furthermore, by comparing index values ​​based on track condition information with preset reference values, it is possible to select locations for which maintenance should be proposed based on clear criteria.

[0122] Furthermore, by inputting track condition information into a trained model 172 that has undergone machine learning to estimate proposed maintenance locations, the trained model 172 outputs the results of inferring multiple proposed maintenance locations, making it possible to infer proposed maintenance locations through machine learning. For example, it is possible to easily determine whether maintenance is necessary based on a large number of evaluation values, and to make inferences that take into account circumstances that cannot be expressed numerically.

[0123] In addition, the trained model 172 is retrained based on the track condition information and maintenance necessity judgment information input into the trained model 172, so that maintenance proposal locations can be selected based on the judgment of the system operator.

[0124] Furthermore, if the maintenance work conditions include route information indicating the route of the track 10 on which the railway vehicle 20 runs, a maintenance work schedule is output based on the positions of multiple maintenance proposal locations in the route information, and therefore a more appropriate maintenance work schedule is output taking into account the travel route to the maintenance proposal locations, travel time, etc.

[0125] Furthermore, if the maintenance work condition information includes labor amount information to be provided for the maintenance work, the period required to maintain the proposed maintenance location can be estimated based on the labor amount information. A more appropriate schedule can be determined based on the estimated maintenance work period.

[0126] Furthermore, if the maintenance work condition information includes diagram information for the track 10, a more appropriate schedule can be determined based on the diagram information so that maintenance can be carried out while allocating time to the operation of the railway vehicle 20.

[0127] In addition, the selected multiple maintenance proposal locations are input into a trained model 272 that has undergone machine learning to estimate a maintenance work schedule, and a maintenance work schedule for the selected multiple maintenance proposal locations is output, thereby making it possible to propose a maintenance work schedule using machine learning.

[0128] Furthermore, if the maintenance support system 32 is equipped with a terminal device 40 including a sensor 42 as a track condition detection unit and a communication device 46 as a transmission unit, the condition of the track 10 is detected when the railway vehicle 20 travels on the track 10. This makes it easier to more appropriately select locations for which maintenance should be proposed.

[0129] Furthermore, if the maintenance work schedule is displayed on the display device 78, the maintenance work schedule can be easily grasped by visually checking the display device 78.

[0130] [Additional Notes] The present disclosure discloses the following aspects.

[0131] A first aspect is a track maintenance support system that supports the maintenance of a track on which railway vehicles run, the track maintenance support system comprising: an input unit that receives track condition information indicating the condition of the track, the track condition information including any one of track displacement information, information indicating ballast fineness, information indicating the condition of the sleepers, or fastener fastening state information, and maintenance work condition information indicating maintenance work conditions for the track; and a processing unit that selects a plurality of proposed maintenance locations on the track based on the track condition information, and outputs a maintenance work schedule for the selected plurality of proposed maintenance locations in accordance with constraints obtained from the maintenance work condition information.

[0132] This track maintenance support system outputs a maintenance work schedule for the selected multiple proposed maintenance locations based on the maintenance work condition information. The person who ultimately decides the maintenance work schedule can refer to the output maintenance work schedule and appropriately finalize the maintenance work schedule in light of the maintenance work conditions.

[0133] Furthermore, if the track condition information includes track displacement information, information indicating ballast granulation, information indicating the state of sleepers, or information on the fastening state of fasteners, it is possible to select, for example, track that has deviated significantly from its initial state, areas where ballast granulation has progressed, track areas where sleepers have deteriorated, or areas where a large number of fasteners have fallen off, as areas where maintenance should be suggested.

[0134] A second aspect is the track maintenance support system according to the first aspect, wherein the processing unit determines priorities of the plurality of proposed maintenance locations.

[0135] In this case, a maintenance work schedule can be output taking into consideration the priorities of multiple proposed maintenance locations.

[0136] A third aspect is a track maintenance support system according to the first or second aspect, wherein the processing unit predicts a future state of the track based on the track state information, and selects the plurality of proposed maintenance locations on the track based on the predicted future state of the track.

[0137] This allows appropriate selection of proposed maintenance locations based on the predicted future state of the orbit.

[0138] A fourth aspect is a track maintenance support system according to any one of the first to third aspects, wherein the processing unit selects the plurality of suggested maintenance locations on the track by comparing an index value based on the track condition information with a predetermined reference value.

[0139] In this way, by comparing the index value with the reference value, multiple maintenance proposal locations are selected, so that maintenance proposal locations can be selected based on clear criteria.

[0140] A fifth aspect is a track maintenance support system according to any one of the first to fourth aspects, wherein the processing unit inputs the track condition information into a trained model that has undergone machine learning to estimate the proposed maintenance locations, and outputs a result of inferring the plurality of proposed maintenance locations.

[0141] This allows machine learning to infer suggested maintenance areas.

[0142] A sixth aspect is a track maintenance support system according to the fifth aspect, further comprising a maintenance necessity judgment input unit to which maintenance necessity judgment information based on the track condition information input to the trained model is input, and the processing unit re-trains the trained model based on the track condition information and the maintenance necessity judgment information input to the trained model.

[0143] In this case, by re-learning the trained model, maintenance proposal points can be selected based on the judgment of the operator of this system.

[0144] A seventh aspect is a maintenance support system relating to any one of the first to sixth aspects, wherein the maintenance work condition information includes route information of the track, and the processing unit outputs the maintenance work schedule based on the positions of the plurality of maintenance proposal locations in the route information.

[0145] As a result, the processing unit outputs a maintenance work schedule based on the locations of multiple maintenance proposal locations in the route information, and therefore a more appropriate maintenance work schedule is output by taking into account the travel route to the maintenance proposal locations, travel time, etc.

[0146] An eighth aspect is a track maintenance support system according to any one of the first to seventh aspects, wherein the maintenance work condition information includes labor amount information to be provided for the maintenance work, and the processing unit outputs the maintenance work schedule based on the labor amount information.

[0147] In this case, the processing unit can estimate the period required to maintain the proposed maintenance location based on the labor amount information. A more appropriate schedule can be determined based on the estimated maintenance work period.

[0148] A ninth aspect is a track maintenance support system according to any one of the first to eighth aspects, wherein the maintenance work condition information includes diagram information for the track, and the processing unit outputs the maintenance work schedule based on the diagram information.

[0149] In this case, the processing unit can determine a more appropriate schedule based on the diagram information so that maintenance can be carried out while allowing for railcar operating hours.

[0150] A tenth aspect is a track maintenance support system relating to any one of the first to ninth aspects, wherein the processing unit outputs the maintenance work schedule for the selected plurality of proposed maintenance locations by inputting the selected plurality of proposed maintenance locations into a trained model that has undergone machine learning to estimate the maintenance work schedule.

[0151] This allows maintenance work schedules to be output and proposed using machine learning.

[0152] An eleventh aspect is a track maintenance support system according to any one of the first to tenth aspects, further comprising a terminal device mounted on the railway vehicle running on the track, the terminal device including a track condition detection unit that detects the condition of the track, and a transmission unit that transmits the detection results of the track condition detection unit so that the results can be input to the input unit.

[0153] This allows the condition of the track to be detected as the train travels along the track, which makes it easier to select appropriate maintenance recommendations.

[0154] A twelfth aspect is the track maintenance support system according to any one of the first to eleventh aspects, further comprising a display device that displays the maintenance work schedule.

[0155] In this case, the maintenance work schedule can be easily grasped by visually checking the display device.

[0156] The configurations described in the above embodiment and modifications can be combined as appropriate as long as they are not mutually contradictory.

[0157] The functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, special-purpose processors, integrated circuits, application-specific integrated circuits (ASICs), conventional circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors and other circuitry. In this disclosure, a circuit, unit, or means is hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be hardware disclosed herein or other known hardware that is programmed or configured to perform the recited functions. Where the hardware is a processor, which is considered a type of circuit, the circuit, means, or unit is a combination of hardware and software, and the software is used to configure the hardware and / or processor.

[0158] The above description is illustrative in all respects and is not intended to limit the scope of the present invention. It is understood that numerous variations not illustrated can be envisaged without departing from the scope of the present invention.

Claims

1. A track maintenance support system that assists in the maintenance of the tracks on which railway vehicles run, An input unit receives track condition information, which includes either track displacement information, information indicating ballast granulation, information indicating the condition of the sleepers, or information indicating the fastening status of the fasteners, as well as maintenance work condition information indicating the maintenance work conditions for the track. A processing unit that, based on the track condition information, selects a plurality of proposed maintenance locations on the track, and outputs a maintenance work schedule for the selected plurality of proposed maintenance locations according to the constraints obtained from the maintenance work condition information, Equipped with, A track maintenance support system comprising: a processing unit that predicts the future state of the track by adding the state of deterioration progression to the track state, and selects the plurality of proposed maintenance locations on the track based on the predicted future state of the track.

2. A track maintenance support system according to claim 1, The processing unit is a track maintenance support system that determines the priority of the plurality of suggested maintenance locations.

3. A track maintenance support system according to claim 1 or claim 2, A track maintenance support system in which the processing unit selects the plurality of suggested maintenance locations in the track by comparing an index value based on the track condition information with a preset reference value.

4. A track maintenance support system according to claim 1 or claim 2, The processing unit is a track maintenance support system that outputs the result of inferring a plurality of suggested maintenance locations by inputting the track state information into a trained model that has undergone machine learning to estimate the suggested maintenance locations.

5. A track maintenance support system according to Claim 4, The system further includes a maintenance necessity determination input unit into which maintenance necessity determination information based on the orbital state information input to the trained model is input. The processing unit retrains the trained model based on the orbital state information and maintenance necessity determination information input to the trained model, and is a track maintenance support system.

6. A track maintenance support system according to claim 1 or claim 2, The aforementioned maintenance work condition information includes the track path information, The processing unit outputs the maintenance work schedule based on the locations of the plurality of suggested maintenance locations in the route information, providing a track maintenance support system.

7. A track maintenance support system according to claim 1 or claim 2, The aforementioned maintenance work condition information includes labor volume information provided for maintenance work, The processing unit outputs the maintenance work schedule based on the workload information, and is a track maintenance support system.

8. A track maintenance support system according to claim 1 or claim 2, The aforementioned maintenance work condition information includes diagram information on the track, The processing unit outputs the maintenance work schedule based on the diagram information, and is a track maintenance support system.

9. A track maintenance support system according to claim 1 or claim 2, The processing unit outputs the maintenance work schedule for the selected plurality of proposed maintenance locations by inputting the selected plurality of proposed maintenance locations into a trained model on which machine learning has been performed to estimate the maintenance work schedule, thereby providing a track maintenance support system.

10. A track maintenance support system according to claim 1 or claim 2, A track condition detection unit is mounted on the railway vehicle that travels on the aforementioned track and detects the condition of the track, A transmission unit that transmits the detection result of the orbital state detection unit to the input unit in a way that allows it to be input, A track maintenance support system further comprising terminal devices including the following.

11. A track maintenance support system according to claim 1 or claim 2, A track maintenance support system further comprising a display device for displaying the aforementioned maintenance work schedule.