Rail surface layer quality determination device and rail surface layer quality determination method

Abstract

To provide a rail surface layer quality determination device.SOLUTION: A rail surface layer quality determination device 1 for determining the quality of a surface layer of a rail 3 comprises: a signal acquisition unit 10 for acquiring a first signal and a second signal different in phase from the first signal, as a result of detection by an eddy current probe 2 scanned along the rail 3; and a determination unit 40 configured to determine the quality of the surface layer of the rail 3 on the basis of the first and second signals.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a rail surface layer quality determination device and a rail surface layer quality determination method for determining the quality of a rail surface layer. [Background technology]

[0002] Conventionally, when repairing rails, inspections are performed to check whether there are cracks or spalling on the rails. Ultrasonic inspection is generally used to check whether there are cracks or spalling on the rails, but eddy current inspection has been proposed as a method to complement the inspection of the rail surface layer, which is a dead zone of ultrasonic inspection (for example, Patent Document 1).

[0003] Patent Document 1 describes an eddy current flaw detection method. In this eddy current flaw detection method, whether or not crack flaws and peeling flaws of railway rails, which have different causes of flaw signal generation, are mixed is determined based on the bulge of the vector waveform of a flaw signal detected as a composite signal of multiple flaw signals by an eddy current flaw detection probe. Specifically, a bulge evaluation value of the vector waveform of the flaw signal detected as a composite signal of multiple flaw signals by an eddy current flaw detection probe is calculated, and a threshold value is set for this bulge evaluation value to perform the determination. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2012-47710 A Summary of the Invention [Problem to be solved by the invention]

[0005] As described above, in the eddy current inspection method described in Patent Document 1, a bulge evaluation value based on a vector waveform of a flaw signal detected by an eddy current inspection probe is compared with a threshold value to determine whether or not a rail contains a mixture of crack flaws and peeling flaws. Meanwhile, rails include welded parts (welds) and joints. When the eddy current inspection method described in Patent Document 1 is used, flaw signals similar to those obtained when there are flaws in the welds and joints are obtained even if there are no cracks or peeling in the rail. For this reason, the eddy current inspection method described in Patent Document 1 leads to erroneous detection of cracks and peeling, and makes it impossible to properly determine whether the rail is good or bad.

[0006] Therefore, there is a need for an apparatus and method for determining the quality of a rail surface layer portion that are capable of appropriately determining the quality of the rail surface layer portion. [Means for solving the problem]

[0007] A characteristic configuration of the rail surface layer quality determination device of the present invention is that the device determines whether the rail surface layer is good or bad, and includes a signal acquisition unit that acquires a first signal and a second signal having a phase different from that of the first signal as detection results of an eddy current probe that scans along the rail, and a determination unit that determines whether the rail surface layer is good or bad based on the first signal and the second signal.

[0008] In order to ensure smooth operation of vehicles along the rail, cracks and peeling in the rail surface layer must be repaired appropriately according to their condition. The rail surface layer may contain not only cracks and peeling, but also rail welds and joints, but these welds and joints do not need to be repaired. Therefore, by detecting using not only the first signal but also both the first and second signals as in this characteristic configuration, the detection targets are expanded, and cracks, peeling, welds and joints can be properly detected (distinguished). Therefore, it becomes possible to properly determine the quality of the rail surface layer.

[0009] It is also preferable that the determination unit determines the quality of the rail surface portion based on a phase difference between the second signal and a result of combining the first signal and the second signal.

[0010] Depending on the cracks or peeling in the rail surface layer, a change occurs in the phase difference between the second signal and the combined result of the first and second signals. With this configuration, therefore, it is possible to distinguish between cracks and peeling by making a judgment based on the change in the phase difference between the second signal and the combined result of the first and second signals. This makes it possible to more accurately judge the quality of the rail surface layer.

[0011] The rail surface layer portion quality determination device further comprises a feature calculation unit that calculates feature quantities indicating features of the first signal and the second signal from the waveforms of the first signal and the second signal, respectively, and an exclusion setting unit that sets a period in which the feature quantities in the first signal and the second signal are included as an exclusion for determining the quality of the rail surface layer portion if the feature quantities satisfy a predetermined exclusion condition, and it is preferable that the determination unit individually determines the presence or absence of cracks and peeling in the rail surface layer portion based on the phase difference in the remaining periods of the first signal and the second signal other than the period set as the exclusion condition.

[0012] As described above, it is possible to detect cracks and peeling based on the first signal and the second signal from the eddy current probe. However, in addition to cracks and peeling, the rail surface layer may have parts that do not require repair but are in a state that changes the eddy current. Therefore, with the above-mentioned characteristic configuration, by setting in advance an exclusion condition that can identify parts that do not require repair but are in a state that changes the eddy current, it is possible to exclude such parts from the judgment target for the quality judgment of the rail surface layer. Therefore, by accurately judging the presence or absence of cracks and peeling in the rail surface layer, it is possible to appropriately judge the quality of the rail surface layer without erroneous judgment.

[0013] It is also preferable that the feature amount includes an amplitude value, a maximum value, a minimum value, and a time difference at which the maximum value and the minimum value are obtained of the waveform.

[0014] With this configuration, the first signal and the second signal can be easily evaluated based on the feature amounts of the waveforms of the first signal and the second signal, respectively, and therefore it is possible to appropriately set areas that are not subject to the pass / fail judgment of the rail surface layer portion.

[0015] It is also preferable that the non-target setting unit determines whether or not the detection result has been obtained at a welded portion of the rail based on the amplitude value of the first signal and the maximum value, the minimum value, and the time difference of the second signal, and determines whether or not the detection result has been obtained at a joint of the rail based on the amplitude values ​​of the first signal and the second signal, and sets the detection results obtained at the welded portion and the joint as non-target.

[0016] As described above, eddy currents change at rail welds and rail joints even though rail repair is not required. With this configuration, the detection results obtained at rail welds and rail joints can be set outside the scope of the pass / fail judgment of the rail surface layer. Therefore, it is possible to prevent erroneous judgment in the judgment of whether rail repair is required or not.

[0017] It is also preferable that the judgment unit judges that the crack is present in the rail surface layer portion when the phase difference is within a predetermined first angle range, and judges that the peeling is present in the rail surface layer portion when the phase difference is within a second angle range outside the first angle range.

[0018] With this configuration, peeling and cracks can be properly detected, and the area requiring repairs can be easily identified.

[0019] It is also preferable that the non-target setting unit determines whether or not the detection result has been obtained at the welded portion based on learning data obtained by machine learning using the first signal and the second signal obtained at the welded portion of the rail as teacher data.

[0020] With this configuration, the detection results obtained at the welded portion can be appropriately detected, and therefore the detection results obtained at such welded portions can be set as excluded from the target, thereby preventing erroneous determination of the quality of the rail surface layer portion.

[0021] Furthermore, the method for determining whether the rail surface layer portion is good or bad according to the present invention includes a signal acquisition step of acquiring a first signal and a second signal having a phase different from that of the first signal as detection results of an eddy current probe scanned along the rail, and a determination step of determining whether the rail surface layer portion is good or bad based on the first signal and the second signal.

[0022] With this characteristic configuration, like the rail surface portion quality determination device described above, it is possible to accurately identify cracks, peeling, welds, and joints in the rail surface portion, thereby making it possible to appropriately determine the quality of the rail surface portion.

[0023] The rail surface layer portion quality determination method further comprises a feature calculation step of calculating feature quantities indicating features of the first signal and the second signal from the waveforms of the first signal and the second signal, respectively, and an exclusion setting step of setting a period in which the feature quantities in the first signal and the second signal are included as an exclusion for determining the quality of the rail surface layer portion if the feature quantities satisfy a predetermined exclusion condition, and it is preferable that the determination step individually determines the presence or absence of cracks and peeling in the rail surface layer portion based on a phase difference between the first signal and the second signal in a remaining period other than the period set as the exclusion condition.

[0024] With this characteristic configuration, like the rail surface portion quality determination device described above, it is possible to accurately determine the presence or absence of cracks or peeling in the rail surface portion, thereby more appropriately determining the quality of the rail surface portion. [Brief description of the drawings]

[0025] [Figure 1] 1 is a block diagram showing the configuration of a rail surface portion quality determination device; [Diagram 2] FIG. 13 is a diagram illustrating an example of an exclusion condition. [Diagram 3] FIG. 4 shows a first signal and a second signal obtained at a welded portion of a rail. [Figure 4] FIG. 2 shows a first and a second signal obtained at a rail joint. [Diagram 5] FIG. 4 is an explanatory diagram of sensitivity adjustment of an eddy current probe. [Figure 6] FIG. 4 is an explanatory diagram of a phase difference. [Figure 7] FIG. 13 is a diagram illustrating an example of a determination criterion. [Figure 8] FIG. 13 shows a signal obtained at a portion of the rail surface where a crack exists. [Figure 9] FIG. 13 is a diagram showing a signal obtained in a portion where peeling occurs in the surface layer of the rail. [Figure 10] FIG. 1 is a diagram showing the condition of a rail surface where continuous cracks exist. [Figure 11] FIG. 11 shows a first signal and a second signal obtained from the rail of FIG. 10. [Figure 12] FIG. 1 is a diagram showing the condition of a rail surface where continuous peeling exists. [Figure 13] FIG. 13 shows a first signal and a second signal obtained from the rail of FIG. 12. [Figure 14] 5 is a flowchart showing a process for determining the quality of a rail surface layer portion. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] The rail surface layer portion quality determination device and rail surface layer portion quality determination method according to the present invention are configured to be able to appropriately determine the quality of the rail surface layer portion. The rail surface layer portion quality determination device 1 and rail surface layer portion quality determination method according to the present embodiment are described below. However, the rail surface layer portion quality determination device 1 and rail surface layer portion quality determination method are not limited to the following embodiment, and various modifications are possible without departing from the gist of the invention.

[0027] Fig. 1 is a block diagram showing a schematic configuration of a rail surface layer quality determination device 1. As shown in Fig. 1, the rail surface layer quality determination device 1 is configured with a signal acquisition unit 10, a feature amount calculation unit 20, a non-target setting unit 30, and a determination unit 40. Each functional unit is constructed with hardware or software or both, with a CPU as a core member, in order to perform processing related to quality determination of the surface layer of the rail 3.

[0028] The rail surface layer part quality determination device 1 performs flaw detection (hereinafter, "eddy current flaw detection") using changes in eddy currents generated in the rail 3 (surface of the rail 3) when a coil through which an alternating current flows is brought close to the rail 3, and determines the quality of the surface layer part of the rail 3. Such eddy current flaw detection devices 5 that perform flaw detection using eddy currents are well known, so detailed explanations will be omitted. The eddy current flaw detection device 5 can be mounted on a rail flaw detection vehicle 4 that performs flaw detection on the rail 3, and the rail flaw detection vehicle 4 can be moved to perform flaw detection. Of course, the eddy current flaw detection device 5 can also be mounted on a frame, and the frame can be moved by an operator by pushing it manually to perform flaw detection. The rail surface layer part quality determination device 1 acquires detection results from the eddy current flaw detection device 5 via, for example, a network, but it may or may not be mounted on the rail flaw detection vehicle 4.

[0029] The signal acquisition unit 10 acquires a first signal and a second signal as detection results of the eddy current probe 2 that scans along the rail 3. The eddy current probe 2 is provided in an eddy current flaw detector 5, and detects eddy currents that occur in the rail 3 when a coil (not shown) through which an AC current as described above flows is brought close to the rail 3. The first signal is a signal that changes depending on the length and depth of a flaw in the surface layer of the rail, the angle at which the flaw is formed (the angle of entry into the rail surface), and the direction. Therefore, if the acquired first signal changes from a reference signal (described later) that serves as a preset reference, it can be assumed that a flaw with a different length, depth, angle, and direction from the reference flaw detected by the reference signal has been detected. The second signal is a signal that has a different phase from the first signal and changes depending on the depth of the flaw.

[0030] The reference signal is a signal obtained by adjusting the sensitivity of the eddy current probe 2 (described later) based on a signal acquired by the eddy current probe 2 when a coil through which an AC current flows is brought close to the rail 3 to be inspected, on which a reference flaw is provided in advance, and an eddy current is generated on the detection surface. The first signal corresponds to a signal detected with the same phase as the phase of the reference signal. The second signal is a signal detected so that the phase differs from the first signal by a predetermined angle (in this embodiment, the phase differs by 90 degrees) in order to observe the change in the eddy current two-dimensionally. In other words, the difference in phase by a predetermined angle means that the detection vector direction of the second signal is different from the detection vector direction of the first signal by a predetermined angle on the XY plane. Such first and second signals can be acquired by configuring two synchronous detectors to detect in directions different from each other by a predetermined angle. The first and second signals acquired by the signal acquisition unit 10 are each transmitted to the feature amount calculation unit 20 (described later) as waveform data (hereinafter also simply referred to as "waveform"). The predetermined angle is not limited to 90 degrees, but may be 60 degrees or 30 degrees, and can be set appropriately depending on the type of rail 3, the shape of the detection target, etc.

[0031] The process of acquiring a first signal and a second signal having a different phase from the first signal as detection results of the eddy current probe 2 scanned along such a rail 3 is referred to as a signal acquisition step in the rail surface portion quality determination method for determining whether the surface portion of the rail 3 is good or bad.

[0032] The feature calculation unit 20 calculates feature amounts indicating the features of the first and second signals from the waveforms of the first and second signals. The feature amounts indicating the features of the first and second signals indicate the features included in the waveforms of the first and second signals, and in this embodiment, the feature amounts used are the amplitude values, maximum values, minimum values, and time differences at which the maximum and minimum values ​​are obtained of the waveforms of the first and second signals.

[0033] The amplitude value of the waveform is the maximum amplitude value among the amplitude values ​​of the first signal and the second signal. The maximum value is the maximum value of the first signal and the second signal. The minimum value is the minimum value of the first signal and the second signal. The time difference at which the maximum and minimum values ​​are obtained is the time difference between the time when the first signal becomes the maximum value and the time when the minimum value is reached, and the time difference between the time when the second signal becomes the maximum value and the time when the minimum value is reached. The feature calculated by the feature calculation unit 20 is transmitted to the non-target setting unit 30 described later.

[0034] The process of calculating feature quantities indicating the characteristics of the first and second signals from the waveforms of the first and second signals is referred to as a feature quantity calculation step in the method for determining the quality of the rail surface portion.

[0035] When the feature quantity satisfies a preset exclusion condition, the exclusion setting unit 30 sets the period in which the feature quantity in the first signal and the second signal is included as an exclusion for determining whether the surface layer part of the rail 3 is good or bad. Here, the eddy current also changes in, for example, the welded part of the rail 3 and the joint of the rail 3. Such a welded part of the rail 3 and the joint of the rail 3 are not defective in determining whether the surface layer part of the rail 3 is good or bad. Here, the welded part of the rail 3 can be determined based on the amplitude value of the first signal and the maximum value, minimum value, and time difference of the second signal. Also, the joint of the rail 3 can be determined based on the amplitude values ​​of the first signal and the second signal.

[0036] Therefore, the non-target setting unit 30 determines whether or not the detection result has been obtained at a welded portion of the rail 3 based on the amplitude value of the first signal and the maximum value, minimum value, and time difference of the second signal, and determines whether or not the detection result has been obtained at a joint of the rail 3 based on the respective amplitude values ​​of the first signal and the second signal, and sets the detection results obtained at the welded portion and the joint as non-targets.

[0037] Such non-target conditions may be stored in advance in the non-target setting unit 30. An example of the non-target conditions is shown in FIG. 2. In this embodiment, the following conditions are applicable to the waveform acquired at the welded portion of the rail 3: (#1) the maximum value of the first signal is equal to or smaller than a first threshold value; (#2) the minimum value of the first signal is equal to or greater than a second threshold value; (#3) the maximum value of the second signal is greater than the first threshold value; (#4) the minimum value of the second signal is smaller than a second threshold value; (#5) the time difference between the maximum and minimum values ​​is greater than the first time; Here, the first threshold value and the second threshold value may be set by supervised learning using past performance data as training data, or different values ​​may be used depending on the type of the rail 3 to be used.

[0038] In addition, the conditions applicable to the joint of rail 3 are as follows: (#6) the amplitude value of the first signal is greater than a third threshold value; (#7) the amplitude value of the second signal is greater than a third threshold value; Here, the third threshold value may be set by supervised learning using past performance data as training data, or a different value may be used depending on the type of the rail 3 to be tested.

[0039] In the detection results of the eddy current probe 2 shown below, the first signal is a signal whose sensitivity has been adjusted based on a reference signal, and the second signal is a signal whose phase differs by 90 degrees from that of the first signal. FIG. 3 shows the detection results of the eddy current probe 2 acquired at the welded portion of the rail 3. Specifically, FIG. 3(A) shows the waveform of the first signal, and FIG. 3(B) shows the waveform of the second signal. In this embodiment, the detection results of the eddy current probe 2 are shown with the vertical axis representing the signal value and the horizontal axis representing time.

[0040] As shown in Fig. 3A, the first signal does not exceed the first threshold, so it satisfies the condition (#1) that "the maximum value of the first signal is equal to or less than the first threshold." Also, as shown in Fig. 3A, the first signal does not fall below the second threshold, so it satisfies the condition (#2) that "the minimum value of the first signal is equal to or greater than the second threshold."

[0041] Also, as shown in FIG. 3B, the second signal exceeds the first threshold, so condition (#3) that "the maximum value of the second signal is greater than the first threshold" is satisfied. Also, as shown in FIG. 3B, the second signal is below the second threshold, so condition (#4) that "the minimum value of the second signal is smaller than the second threshold" is satisfied. Furthermore, as shown in FIG. 3B, the time difference between the timing at which the maximum value of the second signal is acquired and the timing at which the minimum value of the second signal is acquired is greater than a preset first time, so condition (#5) that "the time difference between the acquisition of the maximum value and the minimum value is greater than the first time" is satisfied.

[0042] Therefore, since the detection result of the eddy current probe 2 shown in Figure 3 corresponds to a waveform obtained at the welded portion of the rail 3, the non-target setting unit 30 sets the period including the detection result of the eddy current probe 2 shown in Figure 3 as non-target.

[0043] Fig. 4 also shows the detection results of the eddy current probe 2 obtained at the joint of the rail 3. Specifically, Fig. 4(A) shows the waveform of the first signal, and Fig. 4(B) shows the waveform of the second signal.

[0044] As shown in Fig. 4A, the amplitude value of the first signal exceeds the third threshold, so that the condition (#6) that "the amplitude value of the first signal is greater than the third threshold" is satisfied. Also, as shown in Fig. 4B, the amplitude value of the second signal exceeds the third threshold, so that the condition (#7) that "the amplitude value of the second signal is greater than the third threshold" is satisfied.

[0045] Therefore, since the detection result of the eddy current probe 2 shown in Fig. 4 corresponds to the waveform acquired at the joint of the rail 3, the exclusion setting unit 30 sets the period including the detection result of the eddy current probe 2 shown in Fig. 4 as an exclusion. Note that the joints of the rail 3 include a normal joint in which the rail 3 is sandwiched between a pair of iron fish plates and fastened with bolts, an insulated joint in which an insulating material is sandwiched between a pair of fish plates and fastened with bolts while being electrically insulated, and an adhesive insulated joint in which an insulating material is sandwiched between the rail 3 and the fish plate and fixed with an adhesive, and all of these joints can be set as exclusions if they meet the conditions corresponding to the joints of the rail 3 described above.

[0046] The non-target setting unit 30 can be configured to determine whether or not the detection result is obtained at the welded portion of the rail 3 based on learning data obtained by machine learning using the first signal and the second signal obtained at the welded portion of the rail 3 as teacher data. The learning data obtained by machine learning using the first signal and the second signal obtained at the welded portion of the rail 3 as teacher data is data that has been learned (machine learned) so that when the first signal and the second signal obtained at the welded portion of the rail 3 are input to the non-target setting unit 30, the non-target setting unit 30 determines that the first signal and the second signal are obtained at the welded portion of the rail 3. By using such learning data, the non-target setting unit 30 can exclude detection data obtained at the welded portion of the rail 3 from the target.

[0047] When such a feature quantity satisfies a preset exclusion condition, the process of setting the period in which the feature quantity in the first signal and the second signal is included as being excluded from the pass / fail judgment of the surface layer of the rail 3 is referred to as the exclusion setting step in the rail surface layer pass / fail judgment method.

[0048] The judgment unit 40 judges the quality of the rail surface layer portion based on the first signal and the second signal. As described above, if the first signal changes with respect to the reference signal, it can be assumed that a flaw with a different length, depth, angle, and direction from the reference flaw detected by the reference signal has been detected. In addition, since the second signal is a signal that changes according to the depth of the flaw, if the second signal changes, it can be assumed that the depth of the flaw has changed. Therefore, when acquiring the reference signal, it is preferable to detect, for example, a reference flaw that should be judged to be a flaw. This enables the judgment unit 40 to judge the quality of the rail surface layer portion based on the first signal and the second signal.

[0049] The judgment unit 40 judges the quality of the rail surface layer portion based on the phase difference between the second signal and the result of combining the first and second signals. Here, before inspecting the rail surface layer portion, the sensitivity of the eddy current probe 2 is adjusted. FIG. 5A shows the result of combining the first and second signals acquired when inspecting (inspecting) the reference flaw as described above. It is preferable that the combined signal (combined signal) of the first and second signals is a Lissajous figure as shown in FIG. 5A. In this case, the first signal indicates a change along the X-axis, and the second signal indicates a change along the Y-axis. Of course, it is not necessary to create a Lissajous figure, and the first and second signals may be directly analyzed and calculated.

[0050] Furthermore, in order to perform an appropriate judgment, the composite signal is rotated as shown in FIG. 5B, and the eddy current probe 2 is set so that the flaw signal obtained when the flaw detection is actually performed is output along the Y-axis direction. Therefore, when the flaw detection is actually performed on the rail 3 to be inspected, the judgment is made based on the second signal. Furthermore, as shown in FIG. 5C, the composite signal is amplified so that the peak value of the rotated composite signal becomes a predetermined value (Y1 in the example of FIG. 5). With the sensitivity of the eddy current probe 2 adjusted in this manner, the flaw detection is performed on the rail 3 to be actually inspected.

[0051] The result of combining the first and second signals acquired using the eddy current probe 2 with the sensitivity adjusted as described above is shown in Fig. 6. The determination unit 40 calculates the phase difference (phase angle) with respect to the X-axis of the combined signal obtained by combining the first and second signals.

[0052] Here, it is preferable that the judgment unit 40 judges the presence or absence of cracks and peeling of the surface layer of the rail 3 individually based on the phase difference between the second signal in the remaining period other than the period set as the non-target period among the first and second signals and the result of synthesis of the first and second signals. The remaining period other than the period set as the non-target period is a period other than the period set as the non-target period for the quality judgment of the surface layer of the rail 3 by the above-mentioned non-target setting unit 30. The phase difference between the second signal and the result of synthesis of the first and second signals corresponds to the phase difference (phase angle) with respect to the X-axis of the synthetic signal obtained by synthesizing the above-mentioned first and second signals. Note that, although the angle with respect to the X-axis is taken as the phase difference in FIG. 6, since the X-axis and the Y-axis are mutually orthogonal, it is also possible to calculate the phase difference with respect to the Y-axis. In this case, if the synthetic signal is in the first quadrant, the phase difference is 90 degrees minus the angle between the Y-axis and the synthetic signal, and if the synthetic signal is in the second quadrant, the phase difference is 90 degrees plus the angle between the Y-axis and the synthetic signal.

[0053] The judgment unit 40 judges whether there is a crack or peeling in the surface layer of the rail 3 based on judgment criteria. The judgment unit 40 judges that there is a crack in the surface layer of the rail when the phase difference (the phase difference between the second signal and the combined result of the first signal and the second signal) is within a predetermined first angle range, and judges that there is peeling in the surface layer of the rail when the phase difference is within a second angle range outside the first angle range. An example of the judgment criteria is shown in FIG. 7. In this embodiment, the conditions applicable to a waveform acquired in a portion of the surface layer of the rail 3 where a crack exists are as follows: (#11) The phase difference is within the range of 72 degrees to 108 degrees. Therefore, in this case, the first angle range corresponds to the range of 72 degrees to 108 degrees.

[0054] In addition, the following conditions are applicable to the peeling of the surface layer of rail 3: (#12) The phase difference is smaller than 72 degrees or larger than 108 degrees. Therefore, in this case, the second angle range corresponds to a range less than 72 degrees and a range greater than 108 degrees.

[0055] Fig. 8 shows the detection results of the eddy current probe 2 obtained in a portion where peeling occurs in the surface layer of the rail 3. Specifically, Fig. 8(A) shows the waveform of the first signal, and Fig. 8(B) shows the waveform of the second signal.

[0056] Although not shown, the phase difference is within the range of 72 degrees to 108 degrees, and satisfies the condition (#11) that "the phase difference is within the range of 72 degrees to 108 degrees."

[0057] Therefore, since the detection result of the eddy current probe 2 shown in Figure 8 meets the criteria for determining a crack in the surface layer of the rail 3, the determination unit 40 determines that a crack was present in the area where the detection result shown in Figure 8 was obtained.

[0058] Fig. 9 shows the detection results of the eddy current probe 2 obtained in a portion where peeling occurs in the surface layer of the rail 3. Specifically, Fig. 9(A) shows the waveform of the first signal, and Fig. 9(B) shows the waveform of the second signal.

[0059] Although not shown, the phase difference is smaller than 72 degrees or larger than 108 degrees, satisfying the condition (#14) that "the phase difference is smaller than 72 degrees or larger than 108 degrees."

[0060] Therefore, since the detection result of the eddy current probe 2 shown in Figure 9 satisfies the criteria for determining peeling of the surface portion of the rail 3, the determination unit 40 determines that peeling has occurred in the area where the detection result shown in Figure 9 was obtained.

[0061] According to the above-mentioned criteria for cracks, it is possible to simultaneously judge the presence or absence of not only cracks but also minute shelling scratches. Also, according to the criteria for peeling, it is possible to simultaneously judge the presence or absence of not only peeling but also shelling scratches that cannot be judged according to the criteria for cracks.

[0062] Fig. 10 shows a state in which cracks are continuously present along the extension direction of the rail 3. In this case, as shown in Fig. 11(B), relatively large amplitude values ​​are continuously present in the second signal. In addition, the phase difference between the second signal shown in Fig. 11(B) and the composite result of the first signal shown in Fig. 11(A) and the second signal shown in Fig. 11(B) is within the range of 72 degrees - 108 degrees. Therefore, the determining unit 40 determines that cracks are continuously present in the surface layer of the rail 3 from the detection result of the eddy current probe 2 shown in Fig. 11.

[0063] Fig. 12 shows a state in which peeling exists continuously along the extension direction of the rail 3. In this case, as shown in Fig. 13(B), relatively large amplitude values ​​are intermittently present in the second signal. In addition, the phase difference between the second signal shown in Fig. 13(B) and the composite result of the first signal shown in Fig. 13(A) and the second signal shown in Fig. 13(B) is smaller than 72 degrees or larger than 108 degrees. Therefore, the determining unit 40 determines that peeling exists continuously on the surface layer of the rail 3 from the detection result of the eddy current probe 2 shown in Fig. 11.

[0064] In this way, the determining unit 40 can determine whether peeling or cracking occurs continuously for a predetermined period of time based on the first signal and the second signal.

[0065] If peeling or cracks are present, the determination unit 40 may determine that the condition of the rail 3 is bad and that repair is required, and if peeling and cracks are not present, the determination unit 40 may determine that the condition of the rail 3 is not bad and that repair is not required. Furthermore, since peeling and cracks in the surface layer of the rail 3 can be distinguished from each other based on the determination result of the determination unit 40, it is possible to easily determine whether the surface of the rail 3 should be resurfaced or the rail 3 should be replaced. Furthermore, it is also possible to determine whether repair should be performed promptly depending on the state of peeling and cracks in the surface layer of the rail 3. Such a determination result by the determination unit 40 may be stored in a storage unit (not shown) or may be displayed on a display device.

[0066] The process of determining the quality of the rail surface layer portion based on such a first signal and a second signal is called the determining step in the rail surface layer portion quality determining method. As described above, in the determining step, the presence or absence of cracks and peeling in the rail surface layer portion is determined individually based on the phase difference between the second signal and a combined result of the first and second signals in the remaining period other than the period set as non-target, among the first and second signals.

[0067] As described above, the quality of the rail surface layer can be judged by first setting the period including the feature quantities calculated from the waveforms of the first and second signals as excluded from the quality judgment, and then individually judging the presence or absence of cracks and peeling for the remaining period other than the excluded period. However, it is also possible to perform the judgment according to the flowchart shown in FIG. 14, for example.

[0068] First, it is determined whether detection has been performed continuously over a predetermined period (step #51). If detection has been performed continuously (step #51: Yes), and if the phase difference between the second signal and the result of combining the first and second signals is within the range of 72 degrees to 108 degrees (step #52: Yes), and if the absolute values ​​of the first and second signals are equal to or greater than a predetermined value (step #53: Yes), the determination unit 40 determines that there is a joint (step #54) and ends the process. If the absolute values ​​of the first and second signals are not equal to or greater than a predetermined value in step #53 (step #53: No), the determination unit 40 determines that there is a continuous crack (step #55) and ends the process.

[0069] In step #52, if the phase difference between the second signal and the combined result of the first and second signals is not within the range of 72 degrees-108 degrees (step #52: No) and the absolute values ​​of the first and second signals are equal to or greater than a predetermined value (step #56: Yes), the judgment unit 40 judges that there is a seam (step #57) and ends the process. In step #56, if the absolute values ​​of the first and second signals are not equal to or greater than a predetermined value (step #56: No), the judgment unit 40 judges that there is continuous peeling (step #58) and ends the process.

[0070] In step #51, if detection has not been performed continuously over a predetermined period (step #51: No), if the phase difference between the second signal and the combined result of the first and second signals is within the range of 72 degrees-108 degrees (step #59: Yes) and the absolute values ​​of the first and second signals are equal to or greater than a predetermined value (step #60: Yes), the determination unit 40 determines that there is a seam (step #61) and ends the process. In step #60, if the absolute values ​​of the first and second signals are not equal to or greater than the predetermined value (step #60: No), the determination unit 40 determines that there is a crack (step #62) and ends the process.

[0071] In step #59, if the phase difference between the second signal and the combined result of the first and second signals is not within the range of 72 degrees-108 degrees (step #59: No) and the absolute values ​​of the first and second signals are equal to or greater than a predetermined value (step #63: Yes), the judgment unit 40 judges that there is a seam (step #64) and ends the process. In step #63, if the absolute values ​​of the first and second signals are not equal to or greater than a predetermined value (step #63: No), the judgment unit 40 judges that there is peeling (step #65) and ends the process.

[0072] In the flowchart of FIG. 14, no judgment is made regarding the welded parts of the rail 3, but it is possible to add judgment regarding the welded parts.

[0073] Other embodiments In the above embodiment, the judgment unit 40 has been described as judging the quality of the rail surface layer portion based on the phase difference between the second signal and the combined result of the first signal and the second signal. However, the judgment unit 40 can also be configured to judge the quality of the rail surface layer portion based on, for example, the phase difference between the first signal and the combined result of the first signal and the second signal.

[0074] In the above embodiment, the rail surface layer portion quality determination device 1 is described as including the feature calculation unit 20 and the non-target setting unit 30, and the determination unit 40 determines the presence or absence of cracks and peeling in the rail surface layer portion individually based on the phase difference in the remaining periods of the first and second signals other than the periods set as non-target. However, the rail surface layer portion quality determination device 1 can also be configured without including the feature calculation unit 20 and the non-target setting unit 30. In this case, it is preferable that the determination unit 40 determines the presence or absence of cracks and peeling in the rail surface layer portion individually for all periods of the first and second signals.

[0075] In the above embodiment, the feature amount has been described as including the amplitude value of the waveform, the maximum value, the minimum value, and the time difference at which the maximum value and the minimum value are obtained, but it is also possible to use other features.

[0076] In the above embodiment, the non-target setting unit 30 has been described as determining whether or not the detection result has been acquired at a welded portion of the rail 3 based on the amplitude value of the first signal and the maximum value, minimum value, and time difference of the second signal, and determining whether or not the detection result has been acquired at a joint of the rail 3 based on the amplitude values ​​of the first signal and the second signal, and setting the detection results acquired at the welded portion and the joint as non-target. However, the non-target setting unit 30 may be configured to perform either one of determining whether or not the detection result has been acquired at a welded portion of the rail 3 and determining whether or not the detection result has been acquired at a joint of the rail 3.

[0077] In the above embodiment, the determination unit 40 has been described as determining that there is a crack in the rail surface layer portion when the phase difference is within a predetermined first angle range, and determining that there is peeling in the rail surface layer portion when the phase difference is within a second angle range outside the first angle range. However, the determination unit 40 can also determine the presence or absence of cracks and peeling regardless of the first angle range and the second angle range, that is, without distinguishing between cracks and peeling.

[0078] In the above embodiment, the determination unit 40 is described as determining whether peeling or cracking occurs continuously for a predetermined period based on the first signal and the second signal. However, the determination unit 40 may be configured not to determine peeling or cracking that occurs continuously.

[0079] In the above embodiment, the non-target setting unit 30 is described as determining whether or not the detection result has been acquired at a welded portion of the rail 3 based on learning data obtained by machine learning the first signal and the second signal acquired at the welded portion of the rail 3 as teacher data. However, the non-target setting unit 30 can be configured without performing machine learning. Also, instead of machine learning, it is possible to determine the welded portion of the rail 3 using frequency analysis or multiple regression analysis.

[0080] In the above embodiment, an example of the exclusion condition is shown in Fig. 2, and an example of the determination criteria is shown in Fig. 7. However, the descriptions in Fig. 2 and Fig. 7 are merely examples and can be changed. [Industrial Applicability]

[0081] The present invention can be used in a rail surface layer portion quality determination device and a rail surface layer portion quality determination method for determining the quality of a rail surface layer portion. [Explanation of symbols]

[0082] 1: Rail surface quality check device 2: Overcurrent probe 3: Rail 10: Signal acquisition section 20: Feature calculation unit 30: Excluded setting section 40: Judgment section

Claims

1. A rail surface layer quality determination device for determining the quality of a rail surface layer, a signal acquiring unit that acquires a first signal and a second signal having a phase different from that of the first signal as a detection result of an eddy current probe that is scanned along a rail; a determination unit that determines whether the rail surface portion is good or bad based on the first signal and the second signal; A rail surface portion quality determination device comprising:

2. 2. The rail surface layer portion quality determination device according to claim 1, wherein the determination unit determines the quality of the rail surface layer portion based on a phase difference between the second signal and a result of combining the first signal and the second signal.

3. a feature amount calculation unit that calculates feature amounts indicating features of the first signal and the second signal from the waveforms of the first signal and the second signal; an exclusion setting unit that, when the feature quantity satisfies a preset exclusion condition, sets a period in which the feature quantity in the first signal and the second signal is included as an exclusion for determining whether the rail surface layer portion is good or bad, 3. The rail surface portion quality determination device according to claim 2, wherein the determination unit determines the presence or absence of cracks and peeling in the rail surface portion individually based on the phase difference between the first signal and the second signal in a remaining period other than the period set as the non-target period.

4. 4. The rail surface portion quality determination device according to claim 3, wherein the feature amounts include an amplitude value, a maximum value, a minimum value, and a time difference between when the maximum value and the minimum value are obtained of the waveform.

5. 5. The rail surface portion quality determination device according to claim 4, wherein the non-target setting unit determines whether the detection result has been acquired at a welded portion of the rail based on the amplitude value of the first signal and the maximum value, the minimum value, and the time difference of the second signal, and determines whether the detection result has been acquired at a joint of the rail based on the amplitude values ​​of the first signal and the second signal, and sets the detection results acquired at the welded portion and the joint as non-target.

6. 6. The rail surface layer portion quality determination device according to claim 3, wherein the determination unit determines that the crack is present in the rail surface layer portion when the phase difference is within a predetermined first angle range, and determines that the peeling is present in the rail surface layer portion when the phase difference is within a second angle range outside the first angle range.

7. 6. The rail surface portion quality determination device according to claim 5, wherein the non-target setting unit determines whether or not the detection result has been obtained at the welded portion based on learning data obtained by machine learning using the first signal and the second signal obtained at the welded portion of the rail as teacher data.

8. A method for determining whether a rail surface layer portion is good or bad, comprising the steps of: a signal acquiring step of acquiring a first signal and a second signal having a different phase from the first signal as a detection result of an eddy current probe scanned along the rail; a determining step of determining whether the rail surface portion is good or bad based on the first signal and the second signal; A method for determining whether a rail surface portion is defective or not, comprising:

9. a feature calculation step of calculating feature quantities indicating features of the first signal and the second signal from the waveforms of the first signal and the second signal, respectively; and a non-target setting step of setting a period in which the first signal and the second signal include the feature quantity as a non-target period for determining whether the rail surface layer portion is good or bad, when the feature quantity satisfies a preset non-target condition, 9. The rail surface layer portion quality determination method according to claim 8, wherein the determination step determines the presence or absence of cracks and peeling in the rail surface layer portion individually based on a phase difference between the second signal and a combination result of the first signal and the second signal, during a remaining period other than the period set as the non-target period, of the first signal and the second signal.

Images