Inference data generation device and inference data generation method

Abstract

This invention provides a data generation device capable of creating data for accurately predicting the deterioration of railway tracks. [Solution] The estimation data creation device comprises a storage device and a processing device. The storage device acquires and stores displacement data obtained by measuring the railway track. The processing device processes the displacement data for each section of the track, which is divided into multiple sections, calculates one of the peaks showing a maximum or minimum as a representative value, and stores it as deterioration estimation data for estimating track deterioration. The processing device defines a boundary range that spans the boundary between the Nth section and the N+1th section. If a peak exists in the boundary range, the processing device determines whether or not a predetermined handover condition is met. If the processing device determines that the handover condition is met, it either hands over the peak in the Nth section of the boundary range to the N+1th section, or hands over the peak in the N+1th section of the boundary range to the Nth section, and creates deterioration estimation data.

Description

Technical Field

[0001] The present application mainly relates to an estimation data creation device that processes displacement data obtained by measuring a railway track to create deterioration estimation data for estimating track deterioration.

Background Art

[0002] Patent Document 1 relates to an inspection system for detecting irregularities in railway tracks. The inspection system detects longitudinal force using a sensor mounted on a railway vehicle. The longitudinal force refers to the longitudinal force generated in a member disposed between an axle set and a bogie. The longitudinal force is associated with track irregularities. Therefore, the inspection system performs calculation based on the detected longitudinal force to detect track irregularities.

Prior Art Literature

Patent Document

[0003]

Patent Document 1

Summary of Invention

Problem to be Solved by the Invention

[0004] In railway track inspection, track deterioration is estimated based on information obtained by measuring the track. Since a track is an extremely long object, it is common practice to divide the track into a plurality of sections in the longitudinal direction, and create and process track displacement data on a section-by-section basis. Further, time-series displacement data may be created in order to grasp temporal changes in track displacement.

[0005] Here, consider a case where a displacement exists near the boundary between the N-th section and the (N+1)-th section of the track. Since the displacement data includes slight errors and fluctuations, there is a possibility that this displacement is detected in the N-th section at one point in time and detected as being in the (N+1)-th section at another point in time. As a result, the displacement data in each section is not properly processed, and the estimation accuracy of track deterioration is lowered.

[0006] This application has been filed in view of the above circumstances, and its main purpose is to provide a prediction data generation device capable of creating data for accurately predicting the deterioration of railway tracks. [Means for solving the problem]

[0007] The problem that this application aims to solve is as described above, and next, the means for solving this problem and their effects will be explained.

[0008] According to a first aspect of this application, an inference data creation device having the following configuration is provided. That is, the inference data creation device comprises a storage device and a processing device. The storage device acquires and stores displacement data obtained by measuring a railway track. The processing device processes the displacement data for each section of the track, which is divided into a plurality of sections, calculates one of the peaks showing a maximum or minimum as a representative value, and stores it as deterioration inference data for inferring track deterioration. The processing device defines a boundary range that spans the boundary between the Nth section and the N+1th section. If the peak exists in the boundary range, the processing device determines whether predetermined delivery conditions are met. If the processing device determines that the delivery conditions are met, it delivers the peak located in the Nth section of the boundary range to the N+1th section, or delivers the peak located in the N+1th section of the boundary range to the Nth section, and creates the deterioration inference data.

[0009] According to a second aspect of this application, the following method for creating estimated data is provided: Displacement data obtained by measuring the railway track is acquired and stored. The displacement data is processed for each section of the track, which is divided into multiple sections, and one of the peaks showing a maximum or minimum value is calculated as a representative value. A boundary range is defined that spans the boundary between the Nth section and the N+1th section of the track. If the peak exists within the boundary range, it is determined whether or not predetermined delivery conditions are met. If it is determined that the delivery conditions are met, the peak located in the Nth section of the boundary range is transferred to the N+1th section, or the peak located in the N+1th section of the boundary range is transferred to the Nth section to create the estimated deterioration data. [Effects of the Invention]

[0010] According to this application, it is possible to create information that shows highly accurate inspection results for railway tracks. [Brief explanation of the drawing]

[0011] [Figure 1] A diagram showing an orbital degradation estimation system including an estimation data generation device according to one embodiment of this application. [Figure 2] A graph showing that the peaks representing representative values ​​are located in different sections at different points in time. [Figure 3] A flowchart illustrating the overall process of peak transfer. [Figure 4] A graph showing waveform alignment using offset. [Figure 5] This graph explains the meaning of setting a range of +α. [Figure 6] A flowchart illustrating the transfer process for peak data transfer from the N+1th block to the Nth block. [Figure 7] A graph showing the situation in which peak transfer occurs from the N+1th block to the Nth block. [Figure 8] This graph shows a situation where the transfer of peaks from the N+1th section to the Nth section does not occur. [Figure 9]A flowchart illustrating the transfer process for peak data transfer from section N to section N+1. [Figure 10] A graph showing the situation in which peak transfer occurs from the Nth section to the N+1th section. [Modes for carrying out the invention]

[0012] Next, embodiments of this application will be described with reference to the drawings. First, the orbital degradation prediction system 1 will be described with reference to Figure 1.

[0013] The track deterioration estimation system 1 is a system that measures the displacement of a railway track and estimates the deterioration of the track based on the measurement results. The track displacements that are measured and estimated in this embodiment include, for example, elevation displacement, alignment displacement, level displacement, track gauge displacement, or planar displacement.

[0014] First, let's describe the equipment on the railway vehicle 4. The railway vehicle 4 comprises a car body 40, a bogie 41, and wheels 42. The railway vehicle 4 also comprises a control device 31, a communication device 32, a speed sensor 33, a gyro sensor 34, a track sensor 35, and a position sensor 36.

[0015] The control device 31 is a general-purpose computer such as a PC or a dedicated computer, and is capable of executing programs. The communication device 32 is a communication module or the like, and performs communication processing. The speed sensor 33 is a sensor that measures the speed of the railway vehicle 4. The gyro sensor 34 is a sensor that measures the linear acceleration and rotational angular velocity of the railway vehicle 4 in three axes. The track sensor 35 is, for example, a camera or laser sensor, and detects the shape of the track. The position sensor 36 is, for example, a GNSS sensor, and detects the position of the railway vehicle 4. The detection results of the various sensors are output to the control device 31 in association with the time. The control device 31 controls the communication device 32 to transmit the detection results of the various sensors to the estimated data creation device 10, for example, at predetermined intervals or as soon as it arrives at the storage space.

[0016] The estimation data creation device 10 is a computer such as a server device or a PC, and includes a processing device 11, a storage device 12, and a communication device 13. The processing device 11 is a CPU or the like and performs arithmetic processing. The storage device 12 is an HDD, an SSD or the like, and stores data. The communication device 13 is a communication module or the like and performs communication processing. The communication device 13 receives data from the railway vehicle 4 via the Internet 20. The processing device 11 reads and executes a program to process the data received from the railway vehicle 4. The estimation data creation device 10 may be a single piece of hardware, or may be a plurality of pieces of hardware that perform collaborative work through distributed computing.

[0017] The estimation data creation device 10 processes data detected by the equipment on the railway vehicle 4 side to create displacement data and stores the displacement data in the storage device 12. As shown in Figure 2, the displacement data is data in which the position of the track is associated with the displacement. The position is the distance along the track from the reference position. The displacement is any one of the plurality of displacements described above. With 0 as the reference, the displacement is defined as positive when displaced to one side, and negative when displaced to the other side.

[0018] Further, in the track deterioration estimation system 1, the track is measured periodically. In the present embodiment, measurement is performed at a pace of once a day to create displacement data. Note that the period is not limited to one day, and the measurement may be performed at a pace of multiple times a day, or the time interval between each measurement may be non-uniform. As described above, displacement data is created for each of a plurality of times or dates. The estimation data creation device 10 estimates the deterioration of the track based on the displacement indicated by the displacement data, the change in the displacement, and the like. Note that the processing of calculating the track displacement from various sensors provided in the railway vehicle 4 is well-known, and thus detailed description thereof will be omitted.

[0019] Furthermore, since the orbit is a very long object, the estimation data creation device 10 divides the orbit into multiple sections for processing. Specifically, the estimation data creation device 10 calculates a representative value of displacement for each section and estimates the deterioration of the orbit based on the magnitude or time change of the representative value for each section. The representative value is selected from the displacement peaks. A peak indicates a maximum or minimum. In this embodiment, the peak with the largest absolute value among the peaks is treated as the representative value. The time-series data showing the representative values ​​for each section calculated in this way is the deterioration estimation data. The estimation data creation device 10 estimates the deterioration of the orbit based on the deterioration estimation data. Note that the estimation of orbital deterioration may be performed using a device other than the estimation data creation device 10. In this case, the estimation data creation device 10 does not estimate the deterioration of the orbit.

[0020] Next, referring to Figure 2, we will explain the challenges related to determining representative values ​​for each section.

[0021] Displacement data is affected by track jitter, measurement errors, disturbance noise, or the track displacement itself. Therefore, even if the same track displacement is observed, the location of that displacement may differ depending on the measurement date and time. In particular, track displacement occurring near the boundary between two sections is affected by the above-mentioned factors and may be detected in one section of the boundary at one time and in the other section at another time.

[0022] For example, consider a situation where orbital displacement occurs near the boundary between adjacent sections N and N+1, as shown in Figure 2. At time T-1, shown in the upper part of Figure 2, the peak near the boundary is located in section N+1. Because this peak has a larger absolute value compared to other peaks, it becomes the representative value for section N+1. On the other hand, at time T, shown in the lower part of Figure 2, the peak near the boundary is located in section N. Therefore, this peak becomes the representative value for section N. When such a situation occurs, even though the orbital displacement has not changed significantly, the representative values ​​for sections N and N+1 change significantly. As a result, the accuracy of the degradation prediction data decreases, making it difficult to predict orbital degradation with high accuracy.

[0023] To address this challenge, the estimation data generation device 10 of this embodiment has a function to transfer peaks between adjacent sections. This function ensures that peaks are detected in only one of the two adjacent sections. Therefore, the accuracy of the degradation estimation data is maintained, making it possible to estimate track degradation with high accuracy.

[0024] The following explanation of the peak transfer process is based on the flowchart shown in Figure 3. The flowchart in Figure 3 represents the process executed by the inference data creation device 10, and in particular, the processing device 11.

[0025] First, the processing unit 11 performs waveform alignment of the displacement data (S101). As mentioned above, multiple displacement data exist because they are created for each time or date and time. These multiple displacement data may contain errors in position. The upper part of Figure 4 shows multiple displacement data from different times superimposed. As shown in this figure, since the peak positions of the multiple displacement data are different, it is easy for the peak position to shift to a different section. In this embodiment, the peaks of multiple displacement data are aligned by offsetting the measurement position of the displacement data. In this embodiment, the latest displacement data is used as a reference, and the other displacement data are offset to bring them closer to the latest displacement data. The amount of offset can be achieved by performing a known matching process on the waveform shown by the latest displacement data. Note that using the latest displacement data as a reference is just one example; other displacement data may be used as a reference.

[0026] Note that waveform alignment of displacement data is not limited to offset. For example, the waveform position can be changed by changing the sampling interval when creating the displacement data. For instance, if the sampling interval of the reference data is 0.25m, waveform alignment can be performed by changing the sampling interval of another data set to 0.249m. Generally, changing the sampling interval is more effective than offset in accommodating long-distance deviations.

[0027] Next, the processing unit 11 detects multiple peaks for each interval + α range (S102). + α is a predetermined value for extending the interval, for example, a value of 1 / 5 or 1 / 10 or less of the length of one interval. In other words, in this embodiment, peak detection is performed not only in the target interval but also in a part of the adjacent interval. The reason for calculating multiple peaks, not just the largest peak, is that the largest peak may not exist due to peak passing or filtering.

[0028] Next, the processing unit 11 filters the multiple peaks detected in step S102 to exclude specific peaks (S103). Specifically, the processing unit 11 filters out and excludes local maximums with a negative sign, local minimums with a positive sign, and peaks located outside the target area. Local maximums with a negative sign and local minimums with a positive sign are protruding parts where the displacement approaches 0, so the absolute value of the displacement is small. Therefore, they cannot be representative values ​​and are excluded from the peaks. Peaks located outside the area are excluded because they are geographically outside the target area.

[0029] Now, referring to Figure 5, we will explain why, after detecting peaks in the range of interval + α in step S102, peaks located outside the section are excluded in step S103. As shown in the upper part of Figure 5, consider a situation where there are two adjacent peaks, and these peaks span the (N-1)th section and the Nth section. Generally, since two adjacent peaks only have a small displacement in a portion for some reason, it is preferable to treat the two peaks together as a single peak. However, when peak detection is performed only in the target section, as shown in the upper part of Figure 5, peaks are detected in the (N-1)th section and the Nth section respectively, and are therefore treated as separate peaks.

[0030] On the other hand, when peak detection is performed in the Nth + αth section, two adjacent peaks are treated as one peak, resulting in only one peak being detected. In the example shown in the lower part of Figure 5, the peak located in section N-1 has a larger absolute value, so the peak in section N-1 is detected, while the one in section N is not detected as a peak. This allows two adjacent peaks to be treated as one peak, enabling proper peak detection.

[0031] Furthermore, if two adjacent peaks can be treated as a single peak, other processing methods besides those described above may be used.

[0032] The process from the next step S104 to step S108 focuses on two adjacent sections and performs the transfer of peaks between them. These two sections are referred to as the Nth section and the N+1th section (N=1, 2, ...).

[0033] The processing unit 11 references the peaks of the Nth section and the N+1th section (S104). The processing unit 11 determines whether the latest and largest peak of the Nth section exists within the boundary range (S105). The boundary range is the range set in the vicinity of the boundary between the Nth section and the N+1th section. The boundary range is the range from the boundary to a reference length. The reference length is preferably a value greater than 0 and less than or equal to 1 / 5 of the length of one section, or less than or equal to 1 / 10 of the length of one section. If the determination in step S105 is Yes, the processing unit 11 executes a peak transfer process from the N+1th section to the Nth section (S106). Details of this transfer process will be described later.

[0034] If the determination in step S105 is No, the processing unit 11 determines whether the latest and largest peak in the N+1th section is located within the boundary range (S107). The definition of the boundary range is as described above. If the determination in step S107 is Yes, the processing unit 11 executes the process of transferring peaks from the Nth section to the N+1th section (S108).

[0035] Thus, in this embodiment, the transfer of peaks from the N+1th section to the Nth section is determined first, and then the transfer of peaks from the Nth section to the N+1th section is determined. As a result, peaks near the boundary that satisfy the conditions are always transferred to the section with the smaller section number. Therefore, since the destination of the peak transfer is constant, the same peak is less likely to be detected in different sections. In this embodiment, the section with the smaller section number is given priority as the destination of the peak transfer, but alternatively, the section with the larger section number may be given priority as the destination of the peak transfer. The important thing is that the side that is given priority as the destination of the peak transfer is consistently unified.

[0036] After the above transfer process, the processing unit 11 determines whether the N+1th section is the last section (S109). If it is not the last section, the processing unit 11 increases N by 1 (S110) and executes steps S104 to S108 again. If the N+1th section is the last section, the processing unit 11 terminates the process.

[0037] Next, referring to Figures 6 to 8, we will explain the details of step S106, namely the process of transferring peaks from the N+1th section to the Nth section.

[0038] First, the processing unit 11 defines the position of the latest and largest peak in the Nth section as the reference position (S201). Figure 7 illustrates the reference position. Assuming that the determination in step S105 is Yes, the reference position is within the boundary range. It is preferable to use the latest displacement data for the reference position, but past displacement data may also be used.

[0039] Next, the processing unit 11 reads the displacement data for the target time (S202). Since it is clear that the peak of the latest displacement data is in the Nth section, no peak transfer occurs, so the processing unit 11 reads the displacement data from, for example, the previous time point.

[0040] Next, the processing unit 11 determines whether the distance from the position of the largest peak in the Nth section to the reference position is greater than or equal to the first threshold S1 (S203). This process determines whether the latest largest peak is located in the N+1th section. That is, as shown in Figure 7, if the largest peak was in the Nth section at the latest time T, but the same peak moved to the N+1th section at the previous time T-1, then a different peak will become the largest peak in the Nth section. Therefore, the new largest peak will be located far from the reference position. The first threshold S1 is preferably determined according to the characteristics of the displacement data, etc. As an example, the first threshold S1 is preferably a value from 0 to 1 / 5 of the length of section 1.

[0041] Furthermore, if the distance from the position of the largest peak in the Nth section to the reference position is less than the first threshold S1, it is determined that the peak has not moved to the N+1th section. That is, as shown in Figure 8, in this situation, the largest peak at time T has only moved slightly (by an amount within the first threshold S1) within the Nth section, so a peak transfer is unnecessary.

[0042] If the determination in step S203 is Yes, the processing unit 11 determines whether the peak of the N+1th section is within a range of the reference position from the second threshold S2 (S204). This process determines whether the largest peak that was in the Nth section at the most recent time is now in the N+1th section. That is, as shown in Figure 7, if the peak of the Nth section at time T is located in the N+1th section at time T-1, then this peak should be located in the range of the N+1th section close to the reference position at time T-1. The second threshold S2 is the threshold that defines this range. The second threshold S2 is determined, for example, based on the degree of positional deviation that may occur in the displacement data. As an example, the second threshold S2 is preferably a value that is greater than the reference length that defines the boundary range, and is up to 1 / 5 of the length of one section.

[0043] The conditions described in steps S105, S203, and S204 correspond to the transfer conditions. If all three of these conditions are met, the transfer conditions are satisfied and the peak is transferred. Specifically, the processing unit 11 transfers the peak closest to the reference position from the N+1th section to the Nth section (S205). Transfer means deleting the data associated with one section and associating it with the other section. Therefore, for example, the deleted peak in the N+1th section is treated as a peak that does not exist in the N+1th section at the same time. As mentioned above, since multiple peaks are detected for one section, if the largest peak is transferred and deleted, another peak will be treated as the largest peak. Also, the transfer conditions are just an example, and at least one of the three conditions described above may be omitted.

[0044] If the determination in step S203 or S204 is No, the delivery conditions are not met. Therefore, the processing unit 11 determines whether there are any unprocessed time points (S206), and if there are, it identifies another time point (S207). Then, the processing unit 11 performs the processing in steps S203 to S205 on the displacement data of the identified time point. After that, if the processing unit 11 determines that there are no unprocessed time points, it terminates the peak delivery process from the N+1th section to the Nth section.

[0045] Thus, in this embodiment, since the same transfer conditions are set for the displacement data at each time point, the same peak is less likely to be detected in different sections.

[0046] Next, with reference to Figures 9 and 10, we will describe the details of step S108, namely the process of transferring peaks from the N+1th section to the Nth section.

[0047] This process is shown in Figure 9. The steps S301 to S307 shown in Figure 9 correspond to the steps S201 to S207 in Figure 6 described above. These processes are the same except that the Nth section and the N+1th section are swapped; all other aspects and the purpose of the processes are the same. Therefore, a detailed explanation of the processes is omitted.

[0048] For example, as shown in Figure 10, if at the most recent time T there is a maximum peak in the boundary range of the N+1th section (S107), and at time T-1 the distance from the reference position to the peak of the N+1th section is greater than or equal to the first threshold S1 (S303), and the peak of the Nth section is within the range of the reference position to the second threshold S2 (S304), then the peak closest to the reference position is passed from the Nth section to the N+1th section (S305).

[0049] By performing the above processing, peaks in the boundary range spanning the Nth section and the N+1th section are passed on as needed, making it less likely for the same peak to be detected in different sections. As a result, it is possible to create deterioration prediction data that allows for highly accurate estimation of railway track deterioration.

[0050] (Feature 1) The estimation data creation device 10 of this embodiment comprises a storage device 12 and a processing device 11. The storage device 12 acquires and stores displacement data obtained by measuring the railway track. The processing device 11 processes the displacement data for each section of the track, which is divided into multiple sections, calculates one of the peaks showing a maximum or minimum as a representative value, and stores it as deterioration estimation data for estimating track deterioration. The processing device 11 defines a boundary range that spans the boundary between the Nth section and the N+1th section. If a peak exists in the boundary range, the processing device 11 determines whether or not a predetermined handover condition is met. If the processing device 11 determines that the handover condition is met, it either hands over the peak in the Nth section of the boundary range to the N+1th section, or hands over the peak in the N+1th section of the boundary range to the N section, and creates deterioration estimation data.

[0051] When a peak exists within a boundary range, it can be transferred according to a certain rule, making it easier to associate peaks with the same section in displacement data from different time points. As a result, it is possible to create deterioration prediction data that allows for highly accurate estimation of railway track deterioration.

[0052] (Feature 2) In the inferred data creation device 10 of this embodiment, it is predetermined which of the following will take priority: the transfer of peaks from the Nth section to the N+1th section, or the transfer of peaks from the N+1th section to the Nth section. The processing device 11 determines whether the transfer conditions for the priority transfer are met, and then determines whether the transfer conditions for the non-priority transfer are met.

[0053] This determines whether peaks that may exist within the boundary range are preferentially assigned to the Nth section or the N+1th section, thus making it less likely that the same peak will be associated with different sections.

[0054] (Feature 3) In the estimation data creation device 10 of this embodiment, the processing device 11 transfers peaks to displacement data at multiple time points using common transfer conditions.

[0055] By using common transfer conditions for displacement data at multiple points in time, the situation where the same peak is associated with different sections becomes less likely.

[0056] (Feature 4) In the estimation data creation device 10 of this embodiment, the processing device 11 processes displacement data for each section of the track and calculates multiple peaks. If there are peaks that have been passed to another section, the processing device 11 selects a peak from the remaining peaks and treats it as the representative value for that section.

[0057] This ensures that even if a peak is transferred to another section, the peak does not cease to exist, allowing for the creation of appropriate degradation prediction data.

[0058] (Feature 5) In the estimation data creation device 10 of this embodiment, the processing device 11 determines reference data based on displacement data at multiple time points, and either offsets the measurement position of the displacement data or changes the sampling interval of the measurement position of the displacement data so that at least one displacement data approaches the reference data.

[0059] Because the waveforms of displacement data from multiple time points tend to align more easily, it becomes less likely that the same peak will be associated with different sections.

[0060] (Feature 6) In the estimation data creation device 10 of this embodiment, the processing device 11 determines the reference position based on the maximum peak of the displacement data at any given time. The delivery conditions include the distance from the reference position to the maximum peak being greater than or equal to the first threshold S1.

[0061] This allows us to detect when the location of the largest peak has changed significantly, or in other words, when the location of the largest peak has moved to a different section.

[0062] (Feature 7) In the estimation data creation device 10 of this embodiment, the processing device 11 determines the reference position based on the maximum peak of the displacement data at any given time. The delivery conditions include that the distance from the reference position to the peak of the adjacent section is within a second threshold.

[0063] It is possible to detect peaks that may have moved to a different section.

[0064] (Feature 8) In the inferred data creation device 10 of this embodiment, the processing device 11 receives the peak that is closest to the reference position from among the peaks whose distance from the reference position to the peak of the adjacent section is within the second threshold.

[0065] This allows us to shift the most appropriate peak.

[0066] The above-mentioned features 1 to 8 can be combined, for example, as follows to realize an inferred data generation device. The same applies to the method or program. [Configuration 1] Predictive data generation device having feature 1 [Configuration 2] An inference data generation device having feature 2 in addition to configuration 1. [Configuration 3] An inference data generation device having the features 3 in addition to Configuration 1 or 2. [Configuration 4] An inference data generation device having one of Configurations 1 to 3, plus Feature 4. [Configuration 5] An inference data generation device having one of Configurations 1 to 4, plus Feature 5. [Configuration 6] An inference data generation device having one of Configurations 1 to 5, plus feature 6. [Configuration 7] An inference data generation device having one of Configurations 1 to 6, plus feature 7. [Configuration 8] An inference data generation device having one of Configurations 1 to 7, plus feature 8. [Configuration 9] An inference data generation device having one of Configurations 1 to 8, plus feature 9.

[0067] The functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, dedicated processors, integrated circuits, ASICs (Application Specific Integrated Circuits), 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 circuits. In this disclosure, a circuit, unit, or means is hardware that performs the enumerated functions, or hardware programmed to perform the enumerated functions. The hardware may be hardware disclosed herein, or other known hardware that is programmed or configured to perform the enumerated functions. If the hardware is a processor, which is considered a type of circuit, then the circuit, means, or unit is a combination of hardware and software, and the software is used to configure the hardware and / or the processor. [Explanation of Symbols]

[0068] 10. Estimation Data Generation Device 11 Processing Unit 12 Storage device 13. Communication equipment

Claims

1. A memory device that acquires and stores displacement data obtained by measuring railway tracks, A processing device that processes the displacement data for each section of a track divided into multiple sections, calculates one of the peaks showing a maximum or minimum value as a representative value, and stores it as deterioration estimation data for estimating track deterioration, Equipped with, The aforementioned processing apparatus is Define the boundary range that spans the boundary between the Nth section and the N+1th section. If the aforementioned peak exists within the boundary range, it is determined whether the predetermined delivery conditions are met. An estimation data creation device that, when it is determined that the transfer conditions are met, transfers the peaks located in the N sections of the boundary range to the N+1 section, or transfers the peaks located in the N+1 section of the boundary range to the N section, thereby creating the estimated degradation data.

2. An inference data generation device according to claim 1, It is predetermined which of the following will take priority: the transfer of the peak from the N section to the N+1 section, or the transfer of the peak from the N+1 section to the N section. The processing device is an inference data generation device that, after determining whether the delivery conditions for the preferred delivery are met, determines whether the delivery conditions for the less preferred delivery are met.

3. An inference data generation device according to claim 1, The processing device is an inference data creation device that transfers the peaks to the displacement data at multiple time points using common transfer conditions.

4. An inference data generation device according to claim 1, The processing device processes the displacement data for each section of the track and calculates multiple peaks. The processing device is an inferred data generation device that, if there is a peak that has been passed to another section, selects the peak from the remaining peaks and treats it as the representative value for the section in question.

5. An inference data generation device according to claim 1, The processing device determines reference data based on the displacement data at multiple time points, and either offsets the measurement position of the displacement data or changes the sampling interval of the measurement position of the displacement data so that at least one of the displacement data approaches the reference data, thereby creating estimated data.

6. An inference data generation device according to claim 1, The processing device determines a reference position based on the maximum peak of the displacement data at any given time. An inference data generation device, wherein the delivery conditions include the distance from the reference position to the maximum peak being greater than or equal to a first threshold.

7. An inference data generation device according to claim 1, The processing device determines a reference position based on the maximum peak of the displacement data at any given time. An inference data generation device, wherein the delivery conditions include the distance from the reference position to the peak in the adjacent section being within a second threshold.

8. An inference data generation device according to claim 6, The processing device is an inference data generation device that transfers the peak closest to the reference position from among the peaks whose distance from the reference position to the peak in the adjacent section is within a second threshold.

9. By measuring the railway tracks, displacement data is obtained and stored. The displacement data is processed for each section of the track, which is divided into multiple sections, and one of the peaks showing a maximum or minimum value is calculated as a representative value. Define the boundary range that spans the boundary between the Nth section and the N+1th section of the orbit, If the aforementioned peak exists within the boundary range, it is determined whether the predetermined delivery conditions are met. A method for creating estimated data, which, when it is determined that the transfer conditions are met, transfers the peaks located in the N sections of the boundary range to the N+1 section, or transfers the peaks located in the N+1 section of the boundary range to the N section, thereby creating the estimated degradation data.

Images