Car operation management system, car control system, and car operation management method

Abstract

According to the present invention, a load distribution of future load amounts of a plurality of train formation cars is determined, a current operation load prediction is changed so that the load amounts reach a load amount target while suppressing the load distribution, a post-operation-change load prediction after an operation change is calculated, and a car allocation is planned for the plurality of train formation cars on the basis of the post-operation-change load prediction.

Description

Vehicle operation management system, vehicle control system, and vehicle operation management method

[0001] The present invention relates to a vehicle operations management system, a vehicle control system, and a vehicle operations management method.

[0002] Among public transportation systems, railways excel in mass transportation and punctuality, and are used by many people every day. For example, it is said that more than half of the travel demand in Japan is handled by railways. For this reason, if a malfunction occurs in a railway vehicle and operation is suspended, not only will it result in a significant loss of transportation revenue for the railway company, but railway users will also incur a significant economic loss. Therefore, railway companies are concerned about how to prevent malfunctions in their railway vehicles.

[0003] To prevent railway vehicle malfunctions, railway companies have traditionally created and operated train operation plans to ensure that the load on train sets is even. A train set is a combination of multiple train cars; for example, six train cars can be connected together and operated as a single train. This is the type of train set we see on train station platforms today.

[0004] One method for equalizing the load on each train set is described in, for example, Patent Document 1. Patent Document 1 describes a method in which several lines included in a timetable are combined into a single job, and multiple such jobs are grouped together to create alternate positions, and a vehicle allocation planning method in which the operations of train sets are planned so that the load on each set is uniform by allocating the jobs to each set on a daily rotation basis based on the alternate positions.

[0005] Specifically, to formation A, job 1 is assigned on the first day, job 2 on the second day, job 3 on the third day, and so on. Similarly, to formation B, job 2 is assigned on the first day, job 3 on the second day, and job 4 on the third day, so as not to overlap with jobs already assigned and to even out the load.

[0006] This method of creating tasks and police boxes and assigning them to each train has basically been passed down even in recent years, and tasks and police boxes are basically created and assigned from the timetable.

[0007] Meanwhile, recent advances in IoT technology have led to the development of technologies that estimate the degree of deterioration and load accumulated in a vehicle from sensor data. For example, Patent Document 2 is mentioned.

[0008] Patent Literature 2 proposes a method for estimating the load on a target vehicle based on vehicle driving data and status information on each onboard device while the vehicle is operating. It also shows a method for predicting when future repairs will reduce costs by changing the repair threshold for the load.

[0009] JP 2001-5846 A JP 2023-157092 A

[0010] Patent Document 2 estimates the load on the target device or facility and indicates at what load level repairs should be carried out to improve risk and costs.

[0011] However, in the case of railways, multiple trains are used to operate a single line on a daily basis, so even if repairs are carried out at a time when risks and costs of individual pieces of equipment can be improved, it is difficult to say that this has resulted in improvements in risks and costs for the line as a whole.

[0012] Furthermore, when repair times are decided for individual pieces of equipment, the repair times for each train set are not taken into consideration, and there is a risk that there will be times when multiple train sets remain at high risk of failure when viewed as a line.

[0013] Thus, an operational management method is needed that will improve risks and costs when viewed as a line.

[0014] An object of the present invention is to improve risks and costs for a railway line in a vehicle operation management system.

[0015] A vehicle operations management system according to one aspect of the present invention is a vehicle operations management system that manages the operation of a plurality of railway formation vehicles, and is characterized by having: a performance-based load prediction unit that calculates a future current operational load prediction for each of the plurality of railway formation vehicles by referring to performance; a time point load target calculation unit that calculates a load target for each future point in time; and a load prediction target calculation unit that determines a load distribution of the future load of the plurality of railway formation vehicles, calculates a post-operation change load prediction after an operation change by changing the current operation load prediction so that the load reaches the load target while suppressing the load distribution, and plans vehicle allocation for the plurality of railway formation vehicles based on the post-operation change load prediction.

[0016] According to one aspect of the present invention, in a vehicle operation management system, risks and costs can be improved for a line.

[0017] 1 is a configuration diagram of a railway system and a rolling stock operations management system according to a first embodiment. FIG. 1 is a diagram explaining the concept of rolling stock operations to reduce the incidence of vehicle failures. FIG. 2 is a diagram illustrating an overall processing flow of a rolling stock operations management system of the present invention. FIG. 3 is an image diagram of the first half of the overall processing flow of FIG. 3. FIG. 4 is a diagram illustrating examples of data items for sensing data 111. FIG. 5 is a diagram illustrating examples of data items for inspection history 112. FIG. 6 is a diagram illustrating examples of data items for environmental information 114. FIG. 7 is a diagram illustrating a specific example of a data table of load forecast targets. FIG. 8 is a diagram illustrating examples of data items storing monthly load policies as load forecast targets 116. FIG. 9 is a diagram illustrating examples of data items for timetable 113. FIG. 10 is a diagram illustrating examples of data items for load amount by train 117. FIG. 11 is a processing image diagram of steps 305 and 306. FIG. 12 is a diagram illustrating a detailed flow of step 305. FIG. 13 is a diagram illustrating examples of work and shift boxes. FIG. 14 is an image diagram of creating work from a timetable. FIG. 15 is an image diagram of the results of allocation by step 306. FIG. 16 is a diagram illustrating examples of data items for a rolling stock allocation plan 119. FIG. 17 is a diagram illustrating a detailed flow of risk and cost evaluation. FIG. 18 is a diagram illustrating examples of data items for storing load amount by train set and failure rate among the evaluation results 120. 20 is a diagram showing an example of data items for storing failure rates and loss costs by line among the evaluation results 120. FIG. 21 is a diagram showing an example of an output result obtained by comparing current operations from the present time onwards with proposed changes using the load amounts by formation of FIG. 19. FIG. 22 is a diagram showing an example of an output result obtained by comparing current operations from the present time onwards with proposed changes using the failure rates by line of FIG. 20. FIG. 23 is a diagram showing a flow from the formulation of a vehicle allocation plan to actual train control in a railway using the vehicle operations management system of the present invention. FIG. 24 is a diagram showing a situation in which one or more new vehicle replacement dates are set for a period from the present time onwards in Example 2. FIG. 25 is a diagram showing one proposed operation method for ensuring that the replacement target is met by the start of the vehicle replacement period. FIG. 26 is a diagram showing one proposed operation method for ensuring that the replacement target is met by the end of the vehicle replacement period. FIG. 27 is a diagram showing a flow for determining a combination of new vehicle replacement dates and target railway formation vehicles. FIG. 28 is a configuration diagram of a railway system and a vehicle operations management system according to Example 2.

[0018] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The same reference numerals denote the same components, and the basic configurations and operations are the same.

[0019] This embodiment 1 relates to a vehicle operation management method and management system that estimates the load accumulation status of each vehicle based on sensing data sent from each vehicle and the vehicle inspection history, and presents a specific means for formulating a vehicle operation plan based on the load accumulation status to achieve a desirable future load variation.

[0020] 1 is a configuration diagram of a railway system and a rolling stock operations management system positioned therein according to Example 1. The rolling stock operations management system 101 is composed of a computer device equipped with information processing resources such as an input unit, storage devices (main storage device and auxiliary storage device), a processing unit, an output unit (e.g., a liquid crystal display device), a communication unit, and buses that interconnect these units.

[0021] The input unit is composed of, for example, a keyboard, mouse, touch panel, etc., and the maintenance manager issues instructions to execute processes, which are then used to perform load forecasting, load-based shift planning, shift allocation calculations, risk and cost evaluations, etc., as described below. The storage device is equipped with a main storage device 102 (for example, memory) and an auxiliary storage device 103 (for example, HDD).

[0022] The main memory device 102 stores programs that realize each function of the present invention (load prediction target calculation unit 104, each time point load amount target calculation unit 105, load-based shift planning unit 106, shift allocation calculation unit 107, line load prediction unit 108 (a line refers to each line that appears on the timetable), performance-based load prediction unit 109, and risk and cost evaluation aggregation unit 110).

[0023] The auxiliary storage device 103 also stores various types of information used in the present invention (sensing data 111, inspection history 112, timetable 113, environmental information 114, estimation model 115, load forecast target 116, line-specific load amount 117, station information 118, vehicle allocation plan 119, and evaluation results 120). The processing unit is made up of a processor (CPU), and functions as an arithmetic execution unit that executes the processing of the program, thereby enabling the various functions of the present invention to be exerted.

[0024] The output unit can confirm the item setting status and processing results of the present invention. The communication unit has a function of communicating with each system and vehicle described below that is external to the vehicle operations management system, receiving information sent from the above systems and vehicles and storing it in the auxiliary storage device 103. In addition, it transmits information to each system and vehicle based on the results of each program executed in the main storage device by the processing unit.

[0025] The bus connects the input unit, output unit, processing unit, communication unit, main storage unit, and auxiliary storage unit, and contributes to realizing the functions of the present invention by transferring information.

[0026] The communication unit is connected via a network to, for example, a transportation plan management system 121, an operation management system 122, and a railway vehicle formation 123. The transportation plan management system 121 manages the transportation plans of each railway vehicle formation. It also manages crew allocation plans for operating these railway vehicle formations.

[0027] The operation plan 122 is a system that actually manages and controls the operation of each railway vehicle formation based on the transportation plan created by the transportation plan management system 121. Furthermore, if some railway vehicles are unable to operate as planned, the operation plan 122 cooperates with the transportation plan management system 121 to revise the transportation plan to an appropriate one and continue operation.

[0028] The railway formation vehicle 123 performs train control based on the operation control content or control instructions of the traffic control system 122. At this time, the railway formation vehicle 123 senses each device and each apparatus of the railway formation vehicle using the sensor device 124 and transmits the results to each system via the network. The vehicle operations management system 101 receives the sensing data sent by the sensor device 124 and stores it as history in an auxiliary storage device. The number of railway formation vehicles to be managed is arbitrary, and at least one railway formation vehicle is the target.

[0029] FIG. 2 is a diagram illustrating the concept of vehicle operation for reducing the incidence of vehicle breakdowns, which is the basis of the present invention.

[0030] 2 is expressed as a line graph, with the horizontal axis representing time and the vertical axis representing the load on each railway vehicle formation. While the vertical axis represents the load for the sake of explanation, it may represent any measured or estimated value that may be a cause of a failure, or it may be considered a failure rate. Reference numeral 201 indicates the transition of the load on each railway vehicle formation up to the present time, and six railway vehicle formations, namely, formation A, formation B, ..., and formation F, are shown as examples. Reference numeral 202 indicates the load on each railway vehicle formation at the present time, with the vertical dotted line 203 representing the present time and the load on each railway vehicle formation being conveniently indicated by a gray circle on dotted line 203.

[0031] The dashed line 204 to the right of the current load indicates the future load forecast for each railway vehicle. 205, 206, and 207 indicate the load targets for all railway vehicles on the target line at each point in time.

[0032] Overall, the load increases with the distance traveled and the passage of time, and it is thought that as the load increases, the failure rate will also increase. Therefore, the challenge in vehicle operation is how to operate each train set so that the load does not increase. For example, train sets that have become too loaded must be operated in a way that suppresses the load so that it does not increase in the future. Generally, the higher the load on equipment, the higher the failure rate, so it is natural to suppress the load from the train sets with the highest load.

[0033] On the other hand, in the case of railways, the daily operating volume is determined in advance by a timetable, so if a certain train set is operated to reduce the load, the load on one of the other train sets must be increased accordingly. In this case, increasing the load starting with the train set with the lightest load is thought to be an effective method for reducing the failure rate across the entire line.

[0034] In light of this, in order to reduce the failure rate over the entire line, it is considered desirable to establish a vehicle operation target policy that limits the distribution of loads on each railway vehicle formation for each load target. This idea of ​​limiting the distribution of loads on each railway vehicle formation is the core of this invention.

[0035] Fig. 3 shows the overall processing flow of the rolling stock operations management system of the present invention. First, in step 301 of Fig. 3, the actual load and future load forecast for each vehicle are calculated based on the actual load. This processing will be explained using the image of the first half of the overall processing flow in Fig. 4.

[0036] 201, 204, and 205 in Figure 4 are the same as those in Figure 2. 401 shows a load forecast for the current operation. 402 shows an average load forecast target for achieving the load forecast target 205. 403 shows an example of a load forecast target for each month for achieving the load forecast target 205. The actual load value in step 301 may be a measured value of some kind, or may be an estimated load value composed of multiple factors. In the case of an estimated load value, it can be expressed, for example, as a linear sum composed of several factors. (Equation 1) is an example of a load estimation formula expressed as a linear sum. Such an estimation formula can be obtained by linear multiple regression from past history.

[0037] At this time, the sensing data 111, the inspection history 112, and the environmental information 114 are used as explanatory variables and objective variables.

[0038] 4 illustrates an image of operation based on a load forecast after an operation change. A load forecast target 205 is set in advance, and a current operation load forecast 401 is corrected to a load forecast 204 after an operation change so that it falls within the range of the load forecast target 205. A specific corrected operation method is shown as 403.

[0039] FIG. 5 shows an example of data items of sensing data. Reference numeral 501 denotes a train formation ID, which is an identifier for identifying a railway train formation vehicle. Reference numeral 502 denotes an equipment ID, which is an identifier for identifying the equipment installed on board. Reference numeral 503 denotes the type of equipment installed. Reference numeral 504 denotes the time when the sensing data was acquired. Reference numeral 505 denotes the location where the sensor data was acquired. It is desirable that the acquisition location has an information granularity that allows it to be linked to the environmental information 114 described below. In this example, the acquisition location is expressed as an array of [line name, up / down, distance from starting station], but this is not limited to these items depending on the configuration of the environmental information 114. Reference numeral 506 denotes the actual sensor value for the equipment ID 502.

[0040] FIG. 6 shows an example of data items in the inspection history 112. 601 is a train set ID, which indicates the identifier of the railway train set. 602 is an equipment ID, which is also an identifier of the equipment. 603 indicates a type, which indicates the type of equipment. 604 indicates the inspection time. 605 indicates the inspection result. 606 indicates values ​​for various inspection items measured during the inspection. In this example, a maximum of k sensor values ​​are used, but since the content and number of inspection items often differ depending on the type of equipment, data items may be stored as different items for each type of equipment. For example, 506 may be represented as a single column, and the values ​​may be stored as an array of (key, value) as [(inspection item 1, value 1), (inspection item 2, value 2), ...].

[0041] Fig. 7 shows an example of data items of the environmental information 114. Fig. 7 is a table showing the shape of the railroad tracks, and is a table of information necessary to predict the load amount of each line.

[0042] In particular, with regard to environmental information, it is desirable to model the environmental information 114 as an explanatory variable in order to determine the load on each railway vehicle for each timetable, which will be described later. 701 is a route identifier, and 702 is a type that identifies which alignment on the route. 703 indicates a point, showing the distance from the starting station on the route. 704 indicates the alignment, showing either a straight line or a curve on the left or right (and the curvature if applicable). 705 indicates the gradient. 706 is additional information, which is used to identify, for example, whether the route is on a bridge, inside a station, or inside a tunnel. 707 indicates the speed limit.

[0043] Using the sensing data 111, inspection history 112, and environmental information 114 (after quantification of the qualitative data), an estimation formula is obtained by regression or the like, treating them as explanatory variables. Using the equation (1) estimated in this way, in step 301, a load prediction for the current operation 401 is calculated. Specifically, by substituting a future elapsed time for the elapsed time t, it is possible to obtain a load prediction for the case where the current operation is continued. In addition, by storing the estimated model in the estimation model 115, it can be used for various prediction calculations described below.

[0044] Next, in step 302, a load forecast target is calculated. The load forecast target is indicated by 205 in FIG. 4. The load forecast target is defined as the spread of load amount for a specific period. There are several possible specific methods for calculating this load forecast target, but it may be defined as shown in (Equation 2), for example.

[0045] Here, β is the predicted decrease per unit elapsed time, and is set to be a target setting that reduces the distribution of the load on each railway vehicle, which is the core of the present invention described with reference to FIG. 2.

[0046] As explained in FIG. 2, there may be multiple load forecast targets. Calculations are performed for each load forecast target using, for example, (Equation 2). When there are multiple load forecast targets, α and β in Equation 2 above are explained as fixed values, but these may be changed depending on the load forecast target interval. For example, calculations may be performed once using the current time 203 and the first load forecast target 205 in FIG. 2, and then α and β may be redefined to be different between the load forecast target 205 and the load forecast target 206, and calculations may be performed again.

[0047] FIG. 8 shows a specific example of a data table for load forecast targets. The first line shows the data item name, and the second and subsequent lines show the actual data values, with each line representing one record of data. In the following explanation, data examples will be explained in this table format. 801 shows the ID of the load forecast target, and 802 shows the time period indicating the elapsed time of the load forecast target. The elapsed time can be calculated as the difference between the time period and the current time. 803 shows the average value of the load forecast target, and 804 shows its deviation. Furthermore, from these average values ​​and deviations, an upper limit (average value + deviation) in 805 and a lower limit (average value - deviation) in 806 may be calculated and stored.

[0048] Next, in step 303, a load forecast target for each month that satisfies the load forecast target is calculated for each railway vehicle. 402 in FIG. 4 shows an average load forecast toward the next load forecast target 205. 403 also shows the load policy for each month. The load policy for each month is assigned as an initial solution in accordance with this average load forecast 402. Here, automatic assignment may also be performed so that the load forecast target 205 is reached, but some deviation from the average load forecast 402 is allowed. For example, the load policy may be assigned using random numbers at first, and then automatically recalculated and assigned so as to finally reach 402.

[0049] 9 shows an example of data items storing the load policy for each month as the load forecast target 116. This is a table corresponding to the load forecast target 116 in FIG. 1, and this table is used to determine the current operating line 2101 and the changed operating lines 2102, 2103, and 2104 in FIG.

[0050] Reference numeral 901 indicates the train formation ID. Reference numeral 902 indicates the measurement date. Reference numeral 903 indicates the load amount on the measurement date of the relevant railway train formation. Reference numeral 904 indicates the type of stored information. Records marked "change proposal" indicate records assigned as the initial solution. Records marked "current operation" indicate the case where the current operation remains unchanged. Reference numeral 905 indicates a predicted change in the load amount. Here, the predicted change data can be expressed, for example, as an array of (time, load policy, expected load amount). By expressing it in this way, future load changes in relation to the load policy can be stored in a single record.

[0051] Next, in step 304, the load on each train set is estimated for each line in the timetable.

[0052] 10 shows an example of data items for the timetable 113. It shows a timetable table, with one record representing one line. This line is used to predict the load for each line.

[0053] Reference numeral 1001 denotes a line ID. Reference numeral 1002 denotes a line ID, which indicates an identifier for identifying each line (called a line) in the timetable. Reference numeral 1003 denotes the direction of travel of this line. Reference numeral 1004 denotes the starting point, and reference numeral 1005 denotes the terminal point. Reference numeral 1006 denotes the arrival and departure stations and times, which in this example are expressed as an array of (station name, time, arrival / departure or passing type). For each line in this timetable, the load amount is estimated for each railway vehicle. This load amount is calculated using the estimation formula shown in (Equation 1).

[0054] 11 shows an example of data items for the load amount by train 117. By checking the increased load amount, a train with an increased load amount that is close to the load at the time of alternate assignment is assigned to the railway formation.

[0055] Reference numeral 1101 indicates a route ID. Reference numeral 1102 indicates a route ID. Reference numeral 1103 indicates a train formation ID, which stores the identifier of the railway formation vehicle. Reference numeral 1104 indicates an increased load amount. This makes it possible to store the increased load amount for each railway formation vehicle for each route.

[0056] Before explaining steps 305 and 306, an image of the processing will be explained using Fig. 12. Through the processing up to step 303 described above, a load policy for each railway vehicle for each month is determined as shown in 1201. In accordance with this load policy, work (jobs) are assigned on a daily basis for each load policy from multiple shift boxes with different loads. 1202 shows an image of how jobs are assigned from shift box X 1203 with a low load to a railway vehicle for which a low-load policy has been established in the third month.

[0057] For example, low load 1 is assigned job 1 in shift X 1203. Similarly, daily jobs are assigned according to other load policies. 1204 is an example of the assignment result for a high-load railway vehicle set, which is the result of assignment from shift Z 1205.

[0058] In normal vehicle allocation, there is generally only one alternate station, but in this invention, multiple alternate stations with different loads are automatically created and allocated according to the load policy, aiming to operate in such a way that the load distribution of each railway vehicle across the entire line is minimized. How this processing is realized will be explained in steps 305 and 306.

[0059] In step 305, multiple alternating stops with different loads are calculated from the timetable. This process will be described in detail using the detailed flow in Figure 13. In step 1301, the load forecast target for each railway vehicle for a certain month is obtained. Specifically, the load policy and expected load amount for each railway vehicle for the target month are obtained from the change forecast in 905 in Figure 9.

[0060] Next, in step 1302, the number of shifts with different loads to be created, the shift size, and the estimated load amount for each shift are calculated from the load forecast target. A shift is a list of the jobs to which each railway vehicle is assigned, and a job represents the operation content in which one railway vehicle is operated by the same crew and train composition.

[0061] FIG. 14 shows an example of a job and a police box. 1401 indicates the job ID, and 1402 indicates its contents. One record indicates one job. 1403 is an extracted version of the job content of 1402, and 1404 shows the data representation of 1403. In 1304, one piece of data (a, 08:50, c, 10:30) indicates that the first train departs from Station A at 08:50 and arrives at Station C at 10:30, and further, one piece of data (c, 12:10, a, 13:50) indicates that the first train departs from Station C at 12:50 and arrives at Station A at 13:50.

[0062] Since it is considered one job, it can be understood that 1403 is one job from departure from station a to return to station a. The list of jobs combining 1401 and 1402 is called a police box. The police box size is the size (number of records) of this job list. Police box information 118 is composed of and stored in two columns: the job ID 1401 and the data representation of the job content 1404.

[0063] As mentioned above, the present invention aims to reduce the distribution of loads across the entire line by creating multiple substations and assigning tasks to substations according to the load policy. To achieve this, if the load policy acquired in step 1302 above is one of three types (low load, normal load, high load), it is determined that three substations will be created. At this time, the number of train cars in each load policy becomes the desired substation size. Furthermore, for each load policy, the average expected load is calculated and used as the expected load for the substation.

[0064] Next, in step 1303, for each shift, a line that is close to the expected load is entered into the shift as a job. At this time, if there is a combination of multiple lines that results in a job that is close to the expected load, that job is entered into the shift. At this time, the shift size is also taken into consideration, and jobs are added so that each shift is close to the shift size.

[0065] An example of this task creation is shown in Figure 15. 1501 shows an example of a timetable, which in this example is a timetable from station a to station g. Each diagonal line in the figure represents a timetable line. 1502 shows an example of a task created by connecting timetable lines. There are two thick lines in the figure, each of which represents one task. For example, the line in the upper left corner is created as a task that starts at station a, terminates at station c, then turns around and returns to station a. At this time, if you want to enter the created task into a station with a different expected load, you can enter it into another station by reconnecting it with another line as shown in 1503 and creating a task.

[0066] Next, in step 1204, it is determined whether any unassigned shift positions or lines remain. If any remain, in step 1205, lines that have already been assigned as jobs are combined with unassigned lines and reassigned as new jobs. Furthermore, lines that have already been assigned as jobs are reverted to their original unassigned lines so that they can be combined with other lines.

[0067] Next, in step 1306, it is determined whether allocation has been completed for all months, and if there are any unallocated months, the target month is updated in step 1307 and the process returns to step 1301. If allocation has been completed for all months, the process ends and proceeds to step 306 in FIG.

[0068] Returning to the explanation of Figure 3, in step 306, vehicle allocation is planned in accordance with the load forecast target for each railway vehicle formation. As shown by the relationship between 1202 and 1203 and the relationship between 1204 and 1205 in Figure 12, the vehicle allocation problem was originally one of simultaneously allocating vehicles to all railway vehicles in accordance with a load policy, but it can be seen that the processing flow of the present invention in Figure 3 has been transformed into a problem of allocating work from a single shift with a single expected load of the same magnitude to a group of railway vehicles with the same load policy.

[0069] Therefore, the vehicle allocation problem can be solved by using existing methods, such as by applying a genetic algorithm (GA), as described in the aforementioned Patent Document 1.

[0070] 16 shows an image of the results of allocation in step 306. Reference numeral 1601 indicates the train set ID. Reference numeral 1602 indicates the future vehicle allocation result for the corresponding train set.

[0071] For example, for the first month in the future, train set 10001 will be in light duty, with light duty 1 assigned on the first day, light duty 2 on the second day, and light duty 3 on the third day.

[0072] FIG. 17 shows an example of data items in the vehicle allocation plan 119, and shows an example of the data representation of FIG. 16. 1701 indicates the train set ID. 1702 indicates the time period, which indicates the time period of the operation details described below. 1703 indicates the increased load amount. 1704 indicates the cumulative load amount. 1705 indicates the operation details, which in this example are expressed as an array of (operation date, work). For example, for train set 10001, the operation details are expressed as light operation 1 on the first day of 2023 / 2, light operation 2 on the second day, etc.

[0073] In step 307, it is determined whether the solution is feasible. In the present invention, the load policy may change with each month, and the assigned shifts will also change accordingly. Therefore, when the load policy changes, it is determined whether the last job of the previous month and the first job of the next month are connected. If the jobs are not connected and are not too far apart, it is possible to connect them by creating a forwarding route.

[0074] If a solution that cannot be physically connected is found, a search for a solution is started again in step 308. Here, it is necessary to avoid arriving at the same solution. In the present invention, this can be addressed by changing the load policy. For example, within the same railway vehicle, the load policy set for each month is replaced with the load policy for a certain month. This does not change the overall achievement of the load target, but by changing the load policy for that month, it becomes possible to derive a different solution.

[0075] Within the same month: When the load policy of one railway vehicle is changed, the load policy of another railway vehicle is changed so that the total overall load does not change. For example, when the load policy of one railway vehicle is changed from normal to low load, the load policy of another railway vehicle is changed, for example, from normal to high load, so that the total overall load does not change. This is done because the amount of railway traffic does not change, as described in the vehicle operation concept in Figure 2.

[0076] If it is determined in step 307 that the solution is feasible, a risk and cost assessment is performed on the vehicle allocation plan in step 309. Fig. 18 shows a detailed flow of this risk and cost assessment. In step 1801, a certain railway vehicle is selected. Next, in step 1802, the vehicle allocation plan is referenced for that railway vehicle, and the allocated load amount for each day is obtained.

[0077] Specifically, the increased load amount 1708 in the vehicle allocation plan 119 in Fig. 17 is referenced. In step 1803, the cumulative load amount is calculated. In step 1804, the failure probability at each time point is calculated from the cumulative load amount. One method for calculating the failure probability from the load amount is the Weibull distribution. The Weibull distribution is defined in the form of (Equation 3).

[0078] From this, the failure rate is defined by (Equation 4).

[0079] Here, by inputting the load amount instead of the elapsed time, and estimating the Weibull distribution and its failure rate parameters, it is possible to obtain the failure rate from the load amount.

[0080] Next, in step 1805, it is determined whether the process has been carried out for all railway vehicle formations, and if there are any that have not been carried out, the process returns to step 1801 and processes the railway vehicle formations that have not been carried out. If not, in step 1806, the failure probability for the entire target line is calculated. The failure probability for the entire target line can be calculated using the failure rate of each railway vehicle formation using equation 5.

[0081] Next, in step 1807, the expected loss is calculated from the failure probability on the line. The expected loss can be calculated, for example, by (Equation 6).

[0082] In this way, by using the processing flow of the present invention explained in FIG. 3, it is possible to create a vehicle allocation plan that suppresses the distribution of load on the entire line, and it becomes possible to evaluate the failure rate and expected loss of the entire line based on that plan.

[0083] FIG. 19 shows an example of data items for storing the load and failure rate for each train set from the evaluation results 120. 1901 indicates the train set ID. 1902 indicates the time period, allowing the failure rate from the present point onwards to be expressed. 1903 indicates the type, indicating whether it is the failure rate under current operation or the failure rate under the proposed change. 1904 indicates the increased load of the relevant railway train set, and 1905 indicates the cumulative increase. 1906 stores the failure rate calculated in step 1804 of FIG. 18.

[0084] FIG. 20 shows an example of data items for storing failure rates and loss costs by route among the evaluation results 120. 2001 indicates the route ID, which is the route identifier. 2002 indicates the time period, and similar to 1902, this allows for understanding of failure rates and loss costs from the present time onwards. 2003 indicates the type, indicating whether it is the current operation or a proposed change. 2004 stores the failure rate of the route calculated in step 1806 of FIG. 18. 2005 similarly stores the expected loss calculated in 1807.

[0085] Figure 21 shows an example of the output results comparing the current operation and proposed changes from the present time onwards, using the load amounts for each unit shown in Figure 19. Explanations of the same parts as in Figure 2 are omitted. The dotted line in 2101 indicates the load amount if the current operation remains unchanged. Comparing the current operation and proposed changes in this example, it is easy to see that the distribution of future load amounts will be narrower. 2102 shows the average and deviation of the load amount for each unit at the present time. Continuing, 2103 shows the average and deviation of the load amount for each unit at the next load forecast target. Similarly, 2104 shows the average and deviation at the next load forecast target. By visualizing the load amount forecasts for the current operation plan and proposed changes in this way, it is possible to visually confirm that the proposed changes are expected to fall well within the load forecast target range.

[0086] Figure 22 shows an example of the output results comparing the current operation from the present point in time with the proposed change, using the failure rates by line in Figure 20. It shows the change in failure rate when viewed as a line, rather than a train set. It shows how the failure rate on a line can be reduced by operating in a way that suppresses the load variation of train sets.

[0087] Reference numeral 2201 indicates the failure rate of the line up to the present time. Reference numeral 2202 indicates the failure rate of the line under the current operation. Reference numeral 2203 indicates the failure rate of the line under the proposed change. By visualizing the change in failure rate from the present time in this way, the effect of the present invention can be quantified. Furthermore, by changing the vertical axis of FIG. 22 from the failure rate of the line to the expected loss and displaying the expected loss of 2005 in FIG. 20, it is possible to grasp the amount of loss that is likely to occur in monetary terms.

[0088] Figure 23 shows the flow from the formulation of a vehicle allocation plan to the actual train control in a railway using the vehicle operation management system of the present invention. Note that the function names and data names are basically the same as in Figure 1 and refer to the same things as in Figure 1.

[0089] In the operation target calculation 2301, starting from an execution instruction from the maintenance manager, the sensing data 111, inspection history, and environmental information are used to perform performance-based load prediction, load prediction target calculation, and time-point load calculation.

[0090] Next, in the planning step 2302, load prediction for each route, load-based shift planning, shift allocation calculation, and risk and cost aggregation are performed, and the load for each route, shift information, vehicle allocation plan 119, and evaluation results are output.These output results can also be confirmed by the maintenance manager.The details of the processing up to this point are as explained in Figure 3.

[0091] Next, we will move on to an explanation of traffic management 2303. The transportation planning management system 121 uses the vehicle allocation plan 119 sent from the rolling stock operations management system 101 to create related plans. For example, based on the vehicle allocation plan 119, it creates plans for maintenance work and crew allocation plans. The plans created by the transportation planning management system 121 and the vehicle allocation plan 119 are then sent to the traffic management system 122 and used for traffic management. The traffic management system 122 uses the series of plans to issue operation instructions and control operations for each railway formation vehicle. The railway formation vehicle 123 receives the operation instructions and operation control information sent from the traffic management system and performs actual train control.

[0092] In this way, train control is carried out based on the vehicle allocation plan formulated by the vehicle operation system. Finally, the sensor device 124 mounted on the railway vehicle 123 measures train status information and control information, and the sensing data is stored in the sensing data 111 of the vehicle management system via a network or the like. This sensing data is used when formulating the next vehicle allocation plan.

[0093] By going through the cycle of Plan (planning) → Do (train operation) → See (sensing) in this way, it is possible to reduce the distribution of load on each train formation in the medium to long term, enabling more stable train operation with fewer breakdowns.

[0094] In addition to the first embodiment, the second embodiment relates to a vehicle operation management method and management system that presents a specific means for formulating a vehicle allocation plan that will result in an appropriate load amount by the time of replacement with a new vehicle.

[0095] FIG. 28 is a configuration diagram of a railway system according to the second embodiment and a vehicle operation management system positioned therein.

[0096] The vehicle operations management system of the second embodiment differs from the vehicle operations management system of the first embodiment shown in Fig. 1 in that a new introduction timing determination unit 240 is added. The remaining configuration is the same as that of the vehicle operations management system of the first embodiment shown in Fig. 1, and therefore a description thereof will be omitted.

[0097] Figure 24 shows a situation in which one or more new car replacement dates have been set for a period from the present time forward. Explanations of the same content as in Figure 2 will be omitted. 2401 indicates the first new car replacement date, and 2402 indicates the second new car replacement date. 2403 indicates a replacement guideline for the load. A single railway train requires several cars, and since multiple trains are required on a single line, particularly on major routes, it is rare for the entire train to be replaced at once; rather, the replacement dates are often divided into several phases. Naturally, operation must be carried out in a way that prevents breakdowns from occurring before the new car replacement date. Therefore, the challenge is to devise a car allocation plan that uses all the cars before the new car replacement date while suppressing the line's breakdown rate.

[0098] 25 shows one proposed operation method for ensuring that the load falls below the replacement guideline by the start of the rolling stock replacement period. 2501 is a reference line, which is set in parallel with the replacement guideline so that it is below the replacement guideline. For a group of railway cars due for replacement with new cars, if the group is operated so that the load reaches (or falls below) the load during the new car replacement period, appropriate replacement will be possible as long as new railway cars are prepared at least at the start of the new car replacement period.

[0099] 26 shows one proposed operating method for keeping the vehicle load below the recommended replacement level by the end of the vehicle replacement period. Reference numeral 2601 denotes a reference line, which is set so that the load remains below the recommended replacement level until the end of the new vehicle replacement period. If the operation is carried out so that this reference line is reached during the new vehicle replacement period, it is possible to operate the vehicle so that the recommended replacement level is not reached at any stage during the new vehicle replacement period, and appropriate replacement is possible.

[0100] By setting a baseline in this way and formulating a plan, it becomes possible to make a vehicle allocation plan that reduces the failure rate until the time for new vehicle replacement.The next problem is which train set of vehicles to allocate to which new vehicle replacement period.

[0101] FIG. 27 shows a flow for determining a combination of new car replacement times and target railway vehicle formations when one or more new car replacement times have been set.

[0102] In step 2701, the earliest possible replacement timing for new rolling stock is selected from among the unallocated rolling stock. In step 2702, for each unallocated railway formation, a load forecast for each month corresponding to the replacement timing in the current operation is obtained. By referring to this load forecast, it becomes easier to select railway formations that currently have a small load but will have a large load in the future.

[0103] In step 2703, the unassigned railway vehicle with the highest load is set as the replacement period. In step 2704, it is determined whether replacement has been performed at all new car replacement periods, and if not, the process proceeds to step 2705, where the next new car replacement period is selected. The process then returns to step 2702, and processing continues. If replacement has been performed at all new car replacement periods in step 2704, the process ends.

[0104] By performing the process of Figure 27, the combination of target railway vehicle formations for each new vehicle replacement period is determined. It is then clear that one combination of each (new vehicle replacement period, target railway vehicle formation) is exactly the same as the case of the first embodiment in Figure 2.

[0105] Therefore, by performing the processing of Figure 3 for the number of combinations output in Figure 27, it is possible to create a vehicle allocation plan for each new vehicle replacement period so that the load distribution among the target railway vehicle formations is small (and so that the failure rate on the route is small).

[0106] In the above embodiment, in order to present guidelines for vehicle management that will result in risk and cost improvements when viewed as a line, the load on each vehicle is estimated from sensor information, future load changes are predicted, and a vehicle operation plan is created that reduces the load variation (variance) of each vehicle while reaching the future load target.

[0107] According to the above embodiment, by suppressing the distribution of loads across the entire railway formation while reaching future load targets, it is possible to reduce the failure probability without disproportionately distributing the load to some railway formations. Furthermore, it is possible to operate the railway formation so that it uses up the stock appropriately while suppressing the failure probability until the time for the introduction of new cars.

[0108] 101 Vehicle operation management system 102 Main memory device 103 Auxiliary memory device 104 Load forecast target calculation unit 105 Each time point load amount target calculation unit 106 Load-based shift planning unit 107 Shift allocation calculation unit 108 Line load amount prediction unit 109 Performance-based load prediction unit 110 Risk and cost evaluation aggregation unit 121 Transportation plan management system 122 Traffic management system 123 Railway train set 124 Sensor device 240 New introduction timing determination unit

Claims

1. A rolling stock operations management system that manages the operation of a plurality of railway formations, comprising: a performance-based load prediction unit that calculates a future current operational load prediction for each of the plurality of railway formations by referring to performance; a time-point load target calculation unit that calculates a load target for each future point in time; and a load prediction target calculation unit that determines a load distribution of future loads for the plurality of railway formations, calculates a post-operation change load prediction after an operation change by changing the current operation load prediction so that the load reaches the load target while suppressing the load distribution, and plans vehicle allocation for the plurality of railway formations based on the post-operation change load prediction.

2. The vehicle operations management system according to claim 1, characterized in that the performance-based load prediction unit calculates the current operational load prediction by referring to sensor information from the railway vehicle as the performance.

3. The vehicle operations management system described in claim 1, characterized in that the load forecast target calculation unit predicts future changes in the load amount to determine the load distribution, and calculates the post-operation change load forecast so as to reduce the variance, which is the variation in the load distribution.

4. The vehicle operations management system described in claim 1, characterized in that the load forecast target calculation unit determines an operation policy for the load of the railway vehicle formation for each month based on the load forecast after the operation change and plans it as the vehicle allocation.

5. The vehicle operations management system described in claim 1, characterized in that the each-point-in-time load target calculation unit calculates the load target for each of a plurality of future phases, and the load forecast target calculation unit determines the load distribution for each of a plurality of the phases, and calculates the post-operation change load forecast so as to gradually suppress the load distribution while the load reaches a plurality of the load targets.

6. The vehicle operations management system described in claim 5, characterized in that the load prediction target calculation unit calculates the load prediction after operation change so that the train load deviation of the future load amount of the multiple railway trains gradually decreases.

7. The vehicle operation management system according to claim 1, further comprising a route load prediction unit for predicting future loads for route schedules.

8. The vehicle operation management system described in claim 7, characterized in that the line load prediction unit predicts the load of the railway vehicle when it runs according to the line of the timetable by referring to the timetable and environmental information of the line.

9. A vehicle operation management system as described in claim 1, characterized in that it has a load-based shift planning unit that creates work according to the load amount and calculates the number of shifts, shift size, and expected load amount of the shift according to the load distribution of the railway vehicle formation, and a shift allocation calculation unit that allocates work within the shift including a group of railway vehicle formations divided according to a load operation policy and the expected load amount so as to suppress the load distribution.

10. The vehicle operation management system described in claim 9, characterized in that the station allocation calculation unit allocates the railway vehicle to the station that is closest to the expected load amount based on the load operation policy within a specific period, and performs the vehicle allocation for a combination of the station and the railway vehicle allocated to the station.

11. The vehicle operations management system according to claim 1, characterized in that it comprises an evaluation and aggregation unit that evaluates and aggregates risks and costs based on the post-operation change load forecast, and an output unit that displays the evaluation results evaluated by the evaluation and aggregation unit.

12. A vehicle operations management system as described in claim 1, characterized in that it has a new vehicle introduction timing determination unit that determines the railway vehicle formation to be replaced at a specified new vehicle introduction timing based on the predicted load amount.

13. The vehicle operations management system described in claim 12, characterized in that the new vehicle introduction timing determination unit plans the vehicle allocation so as to suppress the load distribution of the railway vehicle formation at the time of new vehicle introduction and reduce the failure rate of the line.

14. A vehicle control system comprising the vehicle operations management system of claim 1, a transportation planning management system, and an operation management system, wherein the transportation planning management system formulates a transportation plan based on the vehicle allocation plan for vehicle allocation sent from the vehicle operations management system, and the operation management system controls the operation of the railway vehicle formation based on the transportation plan sent from the transportation planning management system.

15. A vehicle operations management method for managing the operation of a plurality of railway formations, comprising: a performance-based load prediction step in which a performance-based load prediction unit calculates a future current operational load prediction for each of the plurality of railway formations by referring to performance; a time-point load quantity target calculation step in which a time-point load quantity target calculation unit calculates a load quantity target for each future time point; and a load prediction target calculation step in which a load prediction target calculation unit determines a load distribution of future load quantities for the plurality of railway formations, changes the current operational load prediction so that the load quantity reaches the load quantity target while suppressing the load distribution, calculates a post-operation change load prediction after the operation change, and plans vehicle allocation for the plurality of railway formations based on the post-operation change load prediction.

Images