JP6678567B2 - Outlier determination method of waveform data and outlier determination system using this method - Google Patents

Description

  The present invention relates to a method for judging abnormal values of time-varying waveform data and an abnormal value judging system using the method, and more particularly, to an abnormal value judging method for orbit displacement data obtained by orbital inspection and an abnormal value using this method. It relates to a value judgment system.

  Conventionally, for the purpose of railway track maintenance management, waveform data such as track displacement data is regularly acquired using a dedicated track inspection vehicle, etc., and the extreme values included in the track inspection data etc. Detection is performed to check whether any irregularities that threaten the running safety of the railway vehicle or irregularities that reduce the ride comfort of the railway vehicle have occurred on the track.

  As described in Patent Document 1, the trajectory inspection uses, for example, an optical sensor that irradiates a laser beam to a trajectory (rail) and receives reflected light from the trajectory (rail), and receives light from the irradiation. A method of calculating the amount of displacement of a trajectory from the time up to is known. By examining the track displacement data obtained in this way and the control values, we are examining the implementation of track maintenance by extracting places with large track irregularities and places where the track irregularities are advancing rapidly. In addition, track displacement data has been measured about several times a year using a dedicated track test vehicle for conventional lines. It has become possible to carry out inspections as frequently as several times a day by using an apparatus.

JP 2015-175735 A

  However, according to such a track inspection method, in rare cases, a laser beam cannot normally detect a track (rail), and an abnormal value may be recorded in track displacement data. Light jumps ". It is known that light skipping is caused by various causes. For example, when a weed grows near a track (rail), the irradiation path of the laser light is interrupted, or when the laser light is irradiated due to rainfall or the like. There is a case where the window or the light receiving window is shut off, and a case where the sunlight affects reflected light from the orbit (rail) of the laser light.

  As shown in FIG. 5A, a conventional method for dealing with light skipping is to record a numerical value without performing any processing at the time of measurement when an abnormal value occurs in the track displacement data, The processing for smoothing outliers included in the orbit displacement data is performed by the processing. Specifically, a process of smoothing an abnormal value included in the trajectory displacement data using a moving average filter or the like is performed by post-processing. However, as shown in FIG. 5B, when the smoothing is performed, the local trajectory displacement is moderated, and the trajectory may be managed with the trajectory displacement data different from the actual trajectory shape.

  Further, as another countermeasure, since normal orbital displacement data cannot be obtained if the orbital displacement data contains an abnormal value, the section in which the abnormal value has occurred must be treated as a missing measurement. After determining whether the value is an abnormal value, data processing that treats the section where the abnormal value occurs as missing is performed, but this process is very complicated and frequently performed several times a day. There is a concern that the data processing will become more complicated due to the increase in the number of data processing targets due to the inspection.

  Therefore, the present invention has been made in view of the above-described problem, and has an abnormal value determination method capable of detecting an abnormal value due to a light jump included in track displacement data and interpolating the abnormal value. And an abnormal value determination system.

An abnormality determination method for waveform data according to the present invention includes: a data acquisition step of acquiring at least waveform data among waveform data and auxiliary data; an abnormal value detection step of detecting an abnormal value included in the waveform data; An auxiliary data abnormal value detecting step of detecting an abnormal value of the auxiliary data when an abnormal value is observed in the waveform data in the detecting step , wherein the auxiliary data is data which is not affected by the detection failure of the waveform data. or wherein the Rukoto using data determined the status of the light jump of laser beam.

  In the method for determining an abnormal value of waveform data according to the present invention, an interpolation step of interpolating an abnormal value of the waveform data after the auxiliary data is determined not to include an abnormal value in the auxiliary data abnormal value detecting step. It is preferable to provide.

  In the method for determining an abnormal value of waveform data according to the present invention, it is preferable that the interpolation step performs interpolation using a straight line, a spline curve, or a sine half-wave curve using normal values before and after the waveform data.

  In the method for determining an abnormal value of waveform data according to the present invention, the abnormal value detecting means calculates a gradient of a derivative of each waveform data among waveform data acquired a plurality of times, and determines a sharp change in the gradient. Is preferably determined.

An abnormal value determination system for waveform data according to the present invention includes a data acquisition unit that acquires at least waveform data among waveform data and auxiliary data; an abnormal value detection unit that detects an abnormal value of the waveform data; Auxiliary data abnormal value detection means for detecting an abnormal value of auxiliary data when an abnormal value is observed in the waveform data by means , wherein the auxiliary data is data which is not affected by detection failure of the waveform data or characterized Rukoto using data determined the status of the light jump of laser beam.

  Further, in the abnormal value determination system for waveform data according to the present invention, after the auxiliary data abnormal value detecting means determines that the auxiliary data does not include an abnormal value, the auxiliary data interpolating means interpolates the abnormal value of the waveform data. It is preferable to provide.

  In the system for determining an abnormal value of waveform data according to the present invention, it is preferable that the interpolation means performs interpolation using a straight line, a spline curve, or a sine half-wave curve using normal data before and after the waveform data.

  Further, in the abnormal value determination system for waveform data according to the present invention, the abnormal value detecting step calculates a gradient of a derivative of each waveform data among the waveform data acquired a plurality of times, and determines a sharp change in the gradient. Is preferably determined.

  According to the abnormal value determination method and the abnormal value determination system for waveform data according to the present invention, whether a significant value of waveform data such as track displacement data is a significant value that affects running safety and riding comfort of a railway vehicle? Or abnormalities caused by light skipping, etc., and when it is determined to be an abnormal value, the abnormal value can be appropriately interpolated. Even in such a case, data processing can be performed easily and easily.

FIG. 1 is a schematic diagram of an abnormal value determination system according to the embodiment. FIG. 4 is a flowchart of data processing of the abnormal value determination system according to the embodiment. FIG. 3 is a diagram showing a specific example of an abnormality detection unit and an auxiliary data abnormal value detection unit of the abnormal value determination system according to the embodiment. FIG. 4 is a diagram showing a specific example of an interpolation unit of the abnormal value determination system according to the embodiment. FIG. 9 is a diagram for explaining a conventional data post-processing method.

  Hereinafter, a preferred embodiment for carrying out the present invention will be described with reference to the drawings. It should be noted that the following embodiments do not limit the invention according to each claim, and not all combinations of the features described in the embodiments are necessarily essential for solving the invention. .

  FIG. 1 is a schematic diagram of an abnormal value determination system according to the present embodiment, FIG. 2 is a flowchart of data processing of the abnormal value determination system according to the present embodiment, and FIG. FIG. 4 is a diagram illustrating a specific example of an abnormal value detection unit and an auxiliary data abnormal value detection unit of the abnormal value determination system, and FIG. 4 is a diagram illustrating a specific example of an interpolation unit of the abnormal value determination system according to the present embodiment. FIG. 5 is a diagram for explaining a conventional data post-processing method.

  As shown in FIG. 1, a determination system 1 according to the present embodiment processes inspection data such as track displacement detected by an inspection device 2 installed in a commercial vehicle or a dedicated track inspection vehicle. The apparatus includes a device 3 and an output device 4 for outputting a processing result of the processing device 3.

  The inspection device 2 is a device capable of continuously inspecting the shape of a track by being mounted on a commercial vehicle or a dedicated track inspection vehicle and running on the track, and specifically, the height of the track. Displacement / Left, Height displacement / Right, Street displacement / Left, Street displacement / Right, Gauge displacement, Level displacement, Planar displacement and other measurement items can be measured. In addition, since the conventionally well-known measuring device 2 can be used for the details of the measuring device 2, detailed description is omitted, but a plurality of sensors such as a laser displacement sensor and an acceleration sensor are provided, It is preferable to perform measurement using the eccentric arrow method, the inertial arrow method, or the inertial measurement method. In addition, it is also possible to carry out the inspection by using a simple hand-operated track inspection device or a continuous rail unevenness measurement device.

  The processing device 3 includes a CPU (Central Processing Unit) for executing a processing program (for example, a method of extracting a rapid portion of waveform data such as orbital displacement data), a ROM (Read Only Memory) for storing the processing program, and a CPU. A RAM (Random Access Memory) for temporarily storing data required for processing.

  The output device 4 is preferably provided with, for example, an image display device such as a liquid crystal display device, an organic EL (Electro-Luminescence) display device, or a printing device of an ink jet printing system or a laser printing system.

  Next, the operation of the abnormal value determination system using the waveform data abnormal value determination method according to the present embodiment will be described. As shown in FIG. 2, the waveform data detected by the inspection device 2 as a data acquisition unit is input to the processing device 3 to start processing the waveform data. Note that the waveform data will be described in the case of performing data processing on inter-gauge displacement in the orbital displacement data.

  First, in the abnormal value detecting means (S101), it is confirmed whether or not there is an abnormal value included in the waveform data having a significant value in the waveform data of the gauge displacement. Specifically, as shown in FIG. 3A, when there is a portion exceeding a predetermined threshold, this is detected as an abnormal value. If the waveform data does not include an abnormal value, the data processing ends.

  Next, when an abnormal value is detected in the waveform data, in the auxiliary data abnormal value detecting means (S102), the presence or absence of an abnormal value is detected for data which is different from the gauge displacement and which is not affected by a light jump which is a detection failure. Check. Specifically, the waveform data detected by the measuring device 2 (elevation displacement / left, elevation displacement / right, street displacement / left, street displacement / right, gauge displacement, level displacement, planar displacement, and the like) Among the intermediate processing sensors of the displacement sensor for outputting the waveform data, and other measurement items, for example, a level displacement, a planar displacement, a vehicle body acceleration, and the like can be used as auxiliary data.

  At this time, as shown in FIG. 3 (b), in the auxiliary data, when there is a significant value at the same location as the significant value in the waveform data of gauge displacement, the waveform data indicates The determined extreme value is determined to be caused by an abnormality in the trajectory, and a warning is issued for an abnormality in the waveform data (S104). The warning of the abnormality of the waveform data is performed by being displayed on the output device 4.

  On the other hand, as shown in FIG. 3C, in the auxiliary data, when there is no significant value in the same place as the place where the significant value in the gauge displacement waveform data, the waveform data is shown. The extremely large value is determined to be an abnormal value caused by light skipping or the like.

  Further, as the auxiliary data, data (light jump determination data) for determining the light jump state of the laser light of the laser displacement sensor used to measure the gauge displacement can be used.

  At this time, as shown in FIG. 3D, in the auxiliary data, if there is no light jump determination information in the same location as the location where the value is significant in the gauge displacement waveform data, the waveform The large value shown in the data is determined to be caused by an abnormality in the trajectory, and a warning is issued for the abnormality in the waveform data (S104).

  On the other hand, as shown in FIG. 3E, in the auxiliary data, when there is information of light jump determination at the same place as the place where the waveform data of the gauge displacement has a significant value, the waveform data The indicated large value is determined to be an abnormal value caused by light skipping or the like.

  Further, the waveform data determined to be a light skip is subjected to interpolation of a very large value in the waveform data interpolation means (S103). More specifically, as shown in FIG. 4, after recognizing an interval between the start point and the end point, which are extremely large, as an abnormal section, the abnormal data (extraordinary value data) included in the abnormal section is deleted and the data is deleted. Measurement. Thereafter, the start point and end point of the abnormal section are assumed to be normal data, and interpolation is performed using a straight line, a spline curve, or a sine half-wavelength curve using the data of the start point and end point of the abnormal section. This presupposes that interpolation is performed within the range between the start point and the end point of the abnormal section because the trajectory to be measured does not deform beyond a certain level because it has a predetermined rigidity. .

  As described above, the outlier determination system 1 using the outlier determination method according to the present embodiment determines whether the waveform data having a significant value is data indicating that a significant orbital displacement has actually occurred, or not. No significant orbital displacement has occurred, and it is possible to determine whether or not the anomaly is caused by light skipping or the like. In addition, since it is provided with an interpolating means for interpolating a very large value when it is determined that the light is skipped, a high-frequency inspection of about several times a day is performed, and a large number of data processing targets are obtained. Even in this case, the interpolation of the waveform data can be easily and reliably performed.

  Note that, in the abnormal value determination system using the abnormal value determination method according to the present embodiment, an example is described in which the abnormal value determination unit determines whether or not there is an abnormality in the waveform data by exceeding a threshold. In addition to the determination of the threshold value, an abrupt change in comparison with the preceding and following data may be calculated based on the gradient of the derivative of the waveform, and the abrupt change in the gradient may be used as a determination unit.

  Further, in the outlier determination method according to the present embodiment and the outlier determination system using the outlier determination method, the case where the waveform data is the waveform data obtained by the trajectory inspection has been described. As long as the waveform data is measured, the data can be used not only for the data of the trajectory inspection but also for the determination processing of any data. It is apparent from the description of the appended claims that embodiments with such changes or improvements can be included in the technical scope of the present invention.

1 Abnormal value judgment system 2 Inspection equipment 3 Processing equipment 4 Output equipment

Claims (8)

A data acquisition step of acquiring at least waveform data among the waveform data and the auxiliary data,
An abnormal value detection step of detecting an abnormal value included in the waveform data,
An auxiliary data abnormal value detecting step of detecting an abnormal value of auxiliary data when an abnormal value is observed in the waveform data in the abnormal value detecting step ,
Said auxiliary data, said waveform abnormal value determination method of the waveform data, characterized in Rukoto using data determined the status of the light jump detection failure of unaffected data or laser beam data. The method for determining an abnormal value of waveform data according to claim 1 ,
An abnormal value determination method for waveform data, comprising an interpolation step of interpolating an abnormal value of the waveform data after it is determined in the auxiliary data abnormal value detection step that the auxiliary data does not include an abnormal value. The method for determining an abnormal value of waveform data according to claim 2 ,
The abnormal value determination method for waveform data, wherein the interpolation step performs interpolation using a straight line, a spline curve, or a sine half-wavelength curve using normal values before and after the waveform data. The method for determining an abnormal value of waveform data according to claim 1,
The abnormal value detecting step is characterized in that, out of the waveform data acquired a plurality of times, the slope of the derivative of each waveform data is calculated, and it is determined whether or not there is a sharp change in the slope. Judgment method. Data acquisition means for acquiring at least waveform data among the waveform data and the auxiliary data,
An abnormal value detecting means for detecting an abnormal value of the waveform data,
An auxiliary data abnormal value detection unit that detects an abnormal value of auxiliary data when an abnormal value is observed in the waveform data by the abnormal value detection unit ,
The auxiliary data, the waveform data of the detected immune to bad data or abnormal value determination system of the waveform data, characterized in Rukoto using data determined the status of the light jump of laser beam. An abnormal value determination system for waveform data according to claim 5 ,
An abnormal value determination system for waveform data, comprising: an interpolation unit that interpolates an abnormal value of the waveform data after the auxiliary data abnormal value detection unit determines that the auxiliary data does not include an abnormal value. The abnormal value determination system for waveform data according to claim 6 ,
An abnormal value determination system for waveform data, wherein the interpolation means performs interpolation using a straight line, a spline curve, or a sine half-wavelength curve using normal data before and after the waveform data. An abnormal value determination system for waveform data according to claim 5 ,
The abnormal value detecting means calculates a differential slope of each of the waveform data among the waveform data acquired a plurality of times, and determines whether or not there is a sharp change in the gradient. Judgment system.

Images