JP6837641B2 - Installation inspection device for rail-mounted articles - Google Patents

Description

The present invention relates to a rail-mounted article mounting inspection device for inspecting rail-mounted articles such as rail bonds for mounting deterioration.

A plurality of separated rails are continuously laid on the track. The train runs on this continuously laid rail. Although the continuous rails are connected to the extent that they can run, they are physically separated and, of course, not electrically connected.

Here, when a railroad crossing is provided on the railroad track, it is preferable that the railroad crossing recognizes the proximity of the train through the rail so that the railroad crossing automatically reacts and operates. For this reason, a method is generally adopted in which a railroad crossing detects a train by passing a current or applying a voltage to a rail at a predetermined distance from the railroad crossing, and when the wheels of the train reach to cut off or energize the rail current. Here, the length of one rail is finite, and it is necessary to detect current interruption or energization due to the arrival of the train at some of the front and rear rails in the train detection section. Therefore, before and after the railroad crossing, the plurality of rails need to be electrically connected to each other.

In the electrified section, the train receives power from the trolley line and returns to the substation via the rail. That is, since the rail constitutes the return current circuit of the train current, a large train current required for driving the train flows through the rail. Therefore, a plurality of rails in the power feeding section need to be electrically connected to each other.

A rail bond is a device that electrically connects rails to each other in order to achieve such a purpose. The rail bond includes a conductive wire which is an aggregate of strands and a pair of terminals connected to the conductive wire. Each of the pair of terminals is welded or glued to each of the two adjacent rails in the traveling direction. Since the pair of terminals are connected by a conductive wire, the two rails connected by a rail bond are electrically connected to each other. By performing this on a predetermined number of rails before and after the railroad crossing, when the train arrives at a certain distance before and after the railroad crossing, the train can be detected by cutting off or energizing the current beyond the rails. ..

As a result, the train is detected by energization via the rail bond, and the processing operation based on the train detection at the railroad crossing is performed.

As such, rail bonds are an important key device in railways and the railway business. Rail bonds are often attached to the sides of the rail, such as the abdomen. In this mounting, it is mounted on the rail while realizing conductivity with the rail by welding or bonding with solder or metal wax.

Here, since the train runs on the rail, various vibrations and deformations based on the vibrations may occur on the rail. Due to this vibration and deformation, a peeled portion is generated on the welded surface (adhesive surface) of the rail bond. Alternatively, the vibration applied to the rail bond itself also causes a peeled portion on the welded surface (adhesive surface) of the rail bond with the rail. The progress of this peeled portion may cause the rail bond to fall off. This can occur not only due to vibrations and deformations applied to rails and rail bonds, but also due to climate deterioration caused by being installed in the outside world. It is difficult to completely prevent such peeling or falling off of the welded surface of the rail bond due to the nature of mounting on the rail.

If the rail bond falls off, the rails will naturally not be able to energize each other, leading to a safety problem in which trains cannot be detected. Since there will be a delay in responding after the rail bond has fallen off, it is necessary to inspect the rail bond mounting state (that is, the state of occurrence of peeling or the like that may fall off). In particular, a tapping sound inspection technique for inspecting based on the difference in timbre and the amplitude obtained by hitting a rail bond with an instrument such as a hammer has been proposed (see, for example, Patent Documents 1, 2, 3, and 4). ).

Japanese Unexamined Patent Publication No. 2011-257307 Japanese Unexamined Patent Publication No. 2000-088817 Japanese Unexamined Patent Publication No. 10-2829 Japanese Unexamined Patent Publication No. 2014-206383

In Patent Document 1, when the shockless hammer 2 hits the inspection object, the shockless hammer 2 does not bounce and is in close contact with the inspection object, and the hit is detected by the microphone 20 attached to the shockless hammer 2. By calculating the sound measurement result and grasping the state of the inspection object, it is possible to inspect the state of the inspection object regardless of the skill of the inspector. For example, the signal of the striking sound measured by the microphone 20 at the time of striking by the shockless hammer 2 is different between the portion where the laminated material of the inspection object is not peeled off and the portion where the laminated material is peeled off. We disclose a tapping sound inspection system that can easily grasp the state of an inspection object by this difference in signal.

In the tapping sound inspection system of Patent Document 1, if peeling exists in the inspection object, the peeled part becomes a new sound reflecting surface, so that the response sound signal received by the microphone is different, and this signal difference. The condition of the inspection object is judged by.

However, the rail bond is welded to the rail, and this welded surface acts as a sound reflecting surface, so it is difficult to distinguish the reflected sound from the part with peeling and the part without peeling. There is a problem. Further, in the tapping sound inspection system of Patent Document 1, the inspection is performed by continuously picking up the sound, but the rail bond also detects the vibration caused by other than the hit rail bond, and the rail bond itself. There is a problem that it is difficult to grasp the wearing state.

In Patent Document 2, the bolt 1 is hit and the pulse-shaped vibration generated in the hammer 2 is detected by the vibration detector 11. The maximum peak value GS of the vibration generated by hitting the bolt 1 that is not loose is stored in the peak reference value setter 16, and the time between the operation of the trigger detector 17 and the operation of the falling zero detector 24. The TS is measured by the counter 25 and stored in the reference value setter 10. Next, the bolt 1 to be inspected is hit, and the ratio calculator 15 obtains the ratio (GT / GS) of the maximum peak value GT of the current vibration waveform and the previous maximum peak value GS. The time TT between the operation of the trigger detector 17 and the operation of the falling zero detector 24 is measured, corrected by dividing by the ratio (GT / GS) by the corrector 8, and the reference by the comparator 9. A hitting determination device for determining whether or not the value is within a predetermined range by comparing with the value TS is disclosed.

Patent Document 2 provides airtightness and loosening of screws based on a comparison between the maximum amplitude of vibration obtained from a hammer hit and the maximum amplitude of vibration in another frequency band. inspect.

When the technique of Patent Document 2 is applied to a rail bond, in the obtained waveform after frequency conversion of the striking sound, the maximum value of the amplitude in a certain frequency domain and the maximum value of the amplitude in another frequency domain are set. Based on the comparison, the mounting state of the rail bond (wearing deterioration state such as peeling) will be inspected.

However, as will be described later, the peeling state of the welded surface of the rail bond varies, and different amplitude peaks may occur in the high frequency band caused by the peeling. In such a situation, if only the maximum value of the amplitude in the high frequency band is used as a reference, various peeling states cannot be detected, and although it should be judged as a defective state, it is judged as a good state. Problems such as That is, the technique of Patent Document 2 has a problem that it is insufficient in inspecting the mounted state of a rail-mounted article such as a rail bond.

Patent Document 3 describes a grip portion gripped by an operator, a hammer portion having a head portion fixed to the grip portion and hit by a member to be diagnosed, and a hammer portion provided on the head portion of the hammer portion to hit the member to be diagnosed. A striking device consisting of striking force detecting means for detecting force, striking force data detected when the head portion of the striking device is impacted by a member to be diagnosed, and striking sound data detected from the striking sound detecting unit are processed. A blow provided with a data processing means for detecting the processing result, a comparison means for comparing the processing result output by the data processing means with the prepared value, and a display means for displaying the comparison result of the comparison means. Disclose the diagnostic device.

When Patent Document 3 is used for the rail bond mounting state inspection, if a peak appears in the high frequency band, it is determined as a defective state. However, the peeled state of the welded surface varies, and even if a peak occurs in the high frequency band, it does not necessarily indicate a peeled state in which there is a risk of falling off. Therefore, there is a problem that a product in a good condition is determined as a defective condition, and an unnecessary replacement cost is incurred. On the contrary, the reverse problem that the rail bond in a defective state is judged as a good state may occur.

Patent Document 4 is a method for inspecting the tapping sound of a brazed article for determining peeling of the article 13 by performing frequency analysis of the tapping sound when the article 13 bonded to the substrate 12 via a brazing material is hit. , A step of collecting the tapping sound over a set period using the microphone 16 to create a striking sound type data, a step of performing frequency analysis of the striking sound type data to obtain frequency analysis data, and non-separation. The frequency analysis of the tapping sound when hitting the article is performed, and the threshold frequency is set on the higher frequency side than the maximum amplitude frequency corresponding to the maximum amplitude, and the maximum amplitude Aa of the frequency analysis data and the audible frequency region above the threshold frequency are set. The process of obtaining the existing maximum amplitude Ab, calculating the maximum amplitude ratio Ab / Aa and using it as a discrimination value is compared with the discrimination value and the set discrimination threshold, and if the discrimination value is less than the discrimination threshold, non-peeling and the discrimination value are A tapping sound inspection method including a step of determining peeling above the discrimination threshold is disclosed.

However, in Patent Document 4, the defective state of rail bond mounting is determined based on the comparison between the maximum value of the amplitude in the low frequency band and the maximum value of the amplitude in the high frequency band. At this time, if the maximum value of the amplitude in the high frequency band is small, it is judged as a good state, but depending on the peeled state of the welded surface, the maximum value of the amplitude is small, but a large number of multiple peaks may occur. .. Alternatively, multiple peaks of the same degree may occur. These vary depending on the variation of the peeled state of the welded surface.

In such a case, if the determination is made only by the maximum value of the amplitude, there is a problem that the determination becomes insufficient, such as determining the defective state as a good state or determining the good state as a defective state. That is, there is a problem that the inspection of the mounting state of the rail-mounted article (rail bond, etc.) mounted on the rail by welding or the like is insufficient only by comparing the maximum amplitude values.

Here, the rail bond is welded to the rail by a welding agent such as solder. Due to the nature of being welded by a welding agent, it is inevitable that some degree of peeling will occur on the welded surface between the rail bond and the rail. That is, it is necessary to determine whether it is a defective state in which there is a possibility of falling off or a good state in which the possibility of falling off is immediately low, depending on the rate of peeling and the state thereof.

If it is mounted with screws or bolts, the rail bond may be inspected for looseness by simple digital detection of whether the screws or bolts are loose or not. As described above, in the rail bond mounted by the welding agent, it is not possible to determine the good state or the bad state of the mounted state only by the presence or absence of peeling. The wearing state can be determined by more accurately detecting the peeling rate and state and the correlation between the actual peeling rate and the dropout rate.

For example, a judgment based only on the maximum amplitude value or a judgment based only on the ratio of the maximum amplitude in the low frequency band to the maximum amplitude in the high frequency band may not sufficiently reflect the actual peeling ratio or peeling state. If the actual peeling ratio and peeling state cannot be sufficiently reflected, there is a problem that the rail bond that is originally in a bad state is judged as a good state, or the rail bond that is originally in a good state is judged as a bad state. It can occur.

That is, there is a problem that it is insufficient to determine the rail bond mounting state only by a single algorithm such as the maximum value of the amplitude and the ratio of the amplitudes in different bands.

An object of the present invention is to provide a mounting inspection device for a rail-mounted article that determines a mounting state of a rail bond or the like welded (bonded) to a rail with solder, metal wax, or the like.

In view of the above problems, the rail-mounted article mounting inspection device of the present invention includes a storage unit that stores a waveform after time-frequency conversion of the tapping sound obtained by hitting the rail-mounted article mounted on the rail.
An amplitude determination unit that determines the mounting state of the rail-mounted article by an amplitude algorithm based on the amplitude ratio of the waveform in a predetermined range.
An area determination unit that determines the mounting state of the rail-mounted article by an area algorithm based on the area ratio of the waveform in a predetermined range.
A control unit that controls the combination of determination operations in the amplitude determination unit and the area determination unit,
A final determination unit for determining the final mounting state of the rail-mounted article based on the respective determination results of the amplitude determination unit and the area determination unit is provided.
The amplitude determination unit
An amplitude detector that detects the maximum value Ka of the amplitude in the low frequency band of the waveform in a predetermined range and the maximum value Kb of the amplitude of the high frequency band of the waveform,
An amplitude ratio calculation unit that calculates Kb / Ka, which is an amplitude ratio,
It has an amplitude algorithm determination unit that determines the mounting state of the rail-mounted article by comparing the amplitude ratio "Kb / Ka" with the amplitude ratio threshold value.
The area judgment unit is
An area calculation unit that calculates the waveform area of the waveform in a predetermined range,
In the waveform area, a frequency value calculation unit that calculates a specific frequency value that is a frequency value for dividing a waveform that has a predetermined area ratio, and
It has an area algorithm determination unit that compares a specific frequency value with an area ratio threshold value and determines the mounting state of a rail-mounted article.
The waveform is a shape based on the frequency axis,
The final determination unit determines the mounting state as a "good state" or a "bad state" based on the respective determination results of the amplitude algorithm determination unit and the area algorithm determination unit.

The mounting inspection device for rail-mounted articles of the present invention determines the mounting state by combining the determination results of a plurality of different algorithms. With this combination, it is possible to eliminate the state in which an erroneous judgment is made by a single algorithm, and it is possible to judge an accurate wearing state.

In particular, by combining the two algorithms of amplitude ratio and area ratio, it is possible to determine a more accurate mounting state while taking advantage of each. Even for a mounting state that differs depending on the variation of the peeling state, a good state or a bad state can be accurately determined by combining different algorithms.

In addition, the mounted state of the rail-mounted article, which was determined to be in a defective state at the time of one determination by a certain algorithm but is actually in a good state, can be correctly determined as a good state. As a result, the cost required for replacement can be reduced.

This is a photograph of rails with rail bonds attached. It is a side view which shows the rail bond attached to a rail from the side. It is a front view which shows the welding surface of the rail bond where the peeled part occurred. It is a front view which shows the welding surface of the rail bond where the peeled part occurred. It is a front view which shows the welding surface of the rail bond where the peeled part occurred. It is a frequency characteristic diagram of the hitting sound obtained by hitting. It is a frequency characteristic diagram of the hitting sound obtained by hitting. It is a block diagram of the mounting inspection device of a rail-mounted article according to the first embodiment of the present invention. It is an internal block diagram of the amplitude determination part. It is a schematic diagram of a waveform. It is a schematic diagram which shows the state which detects the amplitude Ka and the amplitude Kb in the waveform of a predetermined range. It is an internal block diagram of the area determination part in Embodiment 1 of this invention. It is a schematic diagram which shows the area calculation of the waveform in Embodiment 1 of this invention. It is explanatory drawing explaining the determination in the determination part in Embodiment 2 of this invention. It is a schematic diagram of the waveform including noise in Embodiment 2 of this invention. It is explanatory drawing explaining the state of determination in the determination part in Embodiment 2 of this invention. It is a block diagram of the mounting inspection apparatus in Embodiment 4 of this invention.

The mounting inspection device for a rail-mounted article according to the first aspect of the present invention includes a storage unit that stores the time-frequency-converted waveform of the tapping sound obtained by hitting the rail-mounted article mounted on the rail, and a predetermined range. Amplitude determination unit that determines the mounted state of the rail-mounted article by the amplitude algorithm based on the amplitude ratio of the waveform in, and the area that determines the mounted state of the rail-mounted article by the area algorithm based on the area ratio of the waveform in a predetermined range. Based on the determination results of the determination unit, the control unit that controls the combination of the determination operations in the amplitude determination unit and the area determination unit, and the amplitude determination unit and the area determination unit, the final mounting state of the rail-mounted article is determined. The final determination unit for determination is provided, and the amplitude determination unit includes an amplitude detection unit that detects the maximum value Ka of the amplitude in the low frequency band of the waveform in a predetermined range, the maximum value Kb of the amplitude in the high frequency band of the waveform, and the amplitude. It has an amplitude ratio calculation unit that calculates the ratio Kb / Ka, and an amplitude algorithm determination unit that compares the amplitude ratio "Kb / Ka" with the amplitude ratio threshold to determine the mounting state of the rail-mounted article. The area determination unit includes an area calculation unit that calculates the waveform area of the waveform in a predetermined range, and a frequency value calculation unit that calculates a specific frequency value that is a frequency value for dividing the waveform that has a predetermined area ratio in the waveform area. It has an area algorithm determination unit that compares a specific frequency value with an area ratio threshold to determine the mounting state of the rail-mounted article, the waveform has a shape based on the frequency axis, and the final determination unit determines the amplitude algorithm. Based on the respective determination results of the unit and the area algorithm determination unit, the mounting state is determined as a "good state" or a "bad state".

With this configuration, it is attached by welding like a rail bond, and it is possible to more accurately inspect the possibility of falling off corresponding to various peeling states of the welded surface.

In the mounting inspection device for rail-mounted articles according to the second aspect of the present invention, in addition to the first invention, the amplitude determination unit is in a "good state" if the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value. If the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value, it is judged as "defective state".
The area determination unit determines that the specific frequency value is equal to or less than the area ratio threshold value as a "good state", and that the specific frequency value is greater than the threshold value as a "bad state".
When each of the determination result in the amplitude determination unit and the determination result in the area determination unit is in a defective state, the final determination unit finally determines the mounted state of the rail-mounted article as a “defective state”.

With this configuration, it is possible to detect a more defective state by combining determination algorithms having different characteristics.

In the mounting inspection device for rail-mounted articles according to the third aspect of the present invention, in addition to the first invention, the amplitude determination unit is in a "good state" if the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value. If the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value, it is judged as "defective state".
The area determination unit determines that the specific frequency value is equal to or less than the threshold value as a "good state", and that the specific frequency value is greater than the threshold value as a "bad state".
The final determination unit is one of the amplitude determination unit and the area determination unit.
If the judgment result of the first judgment and the judgment result of the second judgment are the same, the final judgment is made based on the first judgment result.
If the judgment result of the first judgment and the judgment result of the second judgment are different, the final judgment is made in consideration of the judgments of the third and subsequent times.

With this configuration, when the determination results of the determination algorithms having different characteristics are different, the determination by the further determination algorithm can be performed to obtain the final determination. As a result, it is possible to reduce the occurrence of a determination result different from the original situation.

In the mounting inspection device for rail-mounted articles according to the fourth aspect of the present invention, in addition to the first invention, the amplitude determination unit is in a "good state" if the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value. If the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value, it is judged as "defective state".
The area determination unit determines that the specific frequency value is equal to or less than the threshold value as a "good state", and that the specific frequency value is greater than the threshold value as a "bad state".
When the number of the determination results of the "defective state" is the majority in the determination results of the amplitude determination unit and the area determination unit, the final determination unit finally determines the mounting state of the rail-mounted article as the "defective state". ..

With this configuration, when the results of the determination algorithms having different characteristics are different, a relatively high inspection result can be obtained by giving priority to the result exceeding the majority.

In the mounting inspection device for rail-mounted articles according to the fifth aspect of the present invention, in addition to the first invention, the amplitude determination unit is in a "good state" if the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value. If the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value, it is judged as "defective state".
The area determination unit determines that the specific frequency value is equal to or less than the threshold value as a "good state", and that the specific frequency value is greater than the threshold value as a "bad state".
The final determination unit weights each of the determination result of the amplitude determination unit and the determination result of the area determination unit, and when the total amount of the weighted determination results is equal to or more than a predetermined value, the rail-mounted article is attached. Is finally determined as a "defective state".

With this configuration, each of the results of the determination algorithms having different characteristics is weighted, so that a result assuming an empirical rule or the like can be obtained.

In the rail-mounted article mounting inspection device according to the sixth aspect of the present invention, in addition to the first invention, the modified amplitude ratio is weighted to the amplitude ratio "Kb / Ka" calculated by the amplitude ratio calculation unit. Amplitude ratio correction part to be calculated and
A specific frequency value correction unit that weights the specific frequency value calculated by the frequency value calculation unit to calculate the correction specific frequency value is further provided.
When the total amount of the corrected amplitude ratio and the corrected specific frequency value is equal to or more than a predetermined value, the final determination unit finally determines the mounted state of the rail-mounted article as a “defective state”.

With this configuration, the final judgment can be obtained based on the calculated numerical values of the different judgment algorithms. As a result, the test can be performed with more mathematical accuracy.

In the rail-mounted article mounting inspection device according to the seventh aspect of the present invention, in addition to the first invention, the predetermined range is from the first frequency to the second frequency.
1st frequency> 2nd frequency,
The second frequency is equal to or higher than the excess component removal frequency,
The excess component removal frequency is a frequency value at which the hitting sound of the rail itself can be removed in the hitting sound.

With this configuration, it is difficult to include sounds from other than the welded surface between the rail and the mounted article in the waveform used for the determination.

In the rail-mounted article mounting inspection device according to the eighth aspect of the present invention, in addition to any one of the first to seventh inventions, the amplitude ratio threshold value and the area ratio threshold value each have a rail-mounted article mounted state. It is determined based on the measured value judged as a good condition.

With this configuration, by setting a threshold value from the measured value, the final determination result actually matches.

In the rail-mounted article mounting inspection device according to the ninth aspect of the present invention, in addition to the eighth invention, each of the amplitude ratio threshold value and the area ratio threshold value is variable or selectively has a plurality of values. ..

With this configuration, it is possible to flexibly respond to the accumulation of work history and the characteristics of the site.

In the rail-mounted article mounting inspection device according to the tenth aspect of the present invention, in addition to any one of the first to ninth inventions, the low frequency band includes the maximum amplitude in the waveform.
The low frequency band and the high frequency band are divided at a frequency value higher than the frequency value corresponding to the maximum amplitude.

With this configuration, it is possible to reliably calculate the amplitude ratio that can detect the problem of the wearing state.

In the mounting inspection device for rail-mounted articles according to the eleventh invention of the present invention, in addition to any one of the seventh to tenth inventions, the area calculation unit has a waveform from the first frequency to the second frequency. The waveform area is calculated by calculating the integrated value.

With this configuration, it is easy to calculate the area.

In the mounting inspection device for rail-mounted articles according to the twelfth invention of the present invention, in addition to any one of the seventh to eleventh inventions, the area of the waveform from the first frequency to the second frequency is set to Sa, and the first Let Sb be the area of the waveform from the frequency to the specific frequency value.
The frequency value calculation unit calculates a specific frequency value in which Sb / Sa is a predetermined area ratio.

With this configuration, the relationship between the area ratio and the specific frequency value can be reliably determined.

In the rail-mounted article mounting inspection device according to the thirteenth invention of the present invention, in addition to any one of the first to twelfth inventions, the control unit first has an amplitude determination unit and finally an area determination unit. Make a judgment.

With this configuration, determination algorithms with different characteristics can be optimally combined.

In the rail-mounted article mounting inspection device according to the fourteenth invention of the present invention, in addition to any one of the first to twelfth inventions, the control unit is determined by the amplitude determination unit a single time or a plurality of times. Have the area determination unit perform the determination once or multiple times in combination.

With this configuration, the mounting state can be determined more accurately by the determination algorithms having different characteristics.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Analysis by the inventor)
The inventor analyzed the insufficiency of determining the mounted state of the rail-mounted article using only the conventional techniques (including Patent Documents 1 to 4) described in the background technique. As explained in the analysis results below, it is possible to digitally determine the presence or absence of peeling only by the maximum amplitude value and amplitude ratio as in the prior art, but the peeling ratio and peeling state and the dropout due to this. It is difficult to accurately determine the possibility (defective state) of.

For rail-mounted articles welded with a welding agent such as solder, the mounting state cannot be determined only by the presence or absence of peeling, but the mounting state can be determined by the peeling ratio or the peeling state. That is, the peeling ratio and the peeling state have a correlation with the wearing state such as the possibility of falling off. For this reason, it may not be possible to accurately determine the mounting state if the determination is made only by the maximum value of the amplitude or the amplitude ratio of the prior art.

FIG. 1 is a photograph in which a rail bond is attached to a rail. The track is provided with rails 200, and adjacent rails 200 are connected on the same line. This is because the length of the rail 200 is finite, and by connecting the rails 200 having the finite length, it is possible to form a track with a distance. The adjacent rails 200 are not electrically connected to each other, and the rail bond 100 is attached to each of the adjacent rails 200. By this straddling mounting, the rail bond 100 can electrically connect the adjacent rails 200 to each other.

FIG. 2 is a side view showing the rail bond mounted on the rail from the side surface. The rail bond 100 is mounted on the side surface of the rail 100. The rail bond 100 includes a wire 103 that realizes an electrical connection, a terminal 101 that is welded to the rail 200, and a welding surface 102 that is a surface that contacts the terminal 101 and the rail 200.
At this time, the rail bond 100 is welded to the rail 200 by a welding material such as metal brazing or solder (it is said to be welded, but it means that it is attached by a welding material, and it can be widely attached such as welding and adhesion. The meaning is explained as welding). The welding material 110 is applied to the welding surface 102, and the welding material 110 realizes welding to the rail 200. The welding material 110 welds the welding surface 102 to the rail 200, and the rail bond 100 is attached to the rail 200.

Here, the train runs on the rail 200. As the train travels, vibration is applied to the rail 200 and the rail bond 100. Further, the rail 200 is deformed by the running of the train, and the deformation of the rail 200 may apply stress to the welding surface 102 between the rail 200 and the rail bond 100. Due to such vibration and deformation, or due to climatic factors in the outside world or natural deterioration, a peeled portion may be formed on the welding material 110 of the welding surface 102.

FIG. 3 is a front view showing a welded surface of the rail bond in which the peeled portion is generated. In the rail bond 100 of FIG. 3, a peeled portion 130 is formed at the right end of the welded surface 102, and the welded portion 120 is the rest of the peeled portion 130.

The rail bond 100 in the peeled state as shown in FIG. 3 has a risk of falling off because the peeled portion 130 is formed at the right end. That is, the mounted state is a defective state.

FIG. 4 is a front view showing a welded surface of the rail bond in which the peeled portion is generated. Unlike FIG. 3, peeled portions 130 are formed at both ends. The area of each of the peeled portions 130 at both ends is smaller than that in the case of FIG. 3, but since the peeled portions 130 are generated at both ends, the rail bond 100 of FIG. 4 may fall off. That is, the mounted state can be considered as a defective state.

FIG. 5 is a front view showing a welded surface of the rail bond in which the peeled portion is generated. In FIG. 5, unlike FIGS. 4 and 5, there are a plurality of peeled portions 130 which are not so large. Although the total area of the peeled portion 130 has a certain size, the welded portion 120 at the peripheral portion remains, and the risk of falling off is low. That is, it can be said that the wearing state is in a good state.

In the rail bond 100, the welding surface 102 is welded to the rail 200 by the welding material 110. Due to such characteristics, the occurrence of peeling varies depending on the area, location, number, form, etc. of the peeled portion 130, and the peeling cannot all be grasped by one concept. When inspecting the mounted state of the rail bond 100 by tapping sound, it is necessary to consider the wide variety of such peeled states.

FIG. 6 is a frequency characteristic diagram of the hitting sound obtained by hitting. FIG. 6 shows a waveform obtained by converting the tapping sound into time and frequency. In this waveform, if the B / A value is equal to or higher than a predetermined value in the maximum value A of the amplitude in the low frequency band and the maximum value B of the amplitude in the high frequency band, it is considered that the mounting state of the rail bond 100 is in a defective state. It is the prior art that determines.

Here, in the case of the peeled state as shown in FIG. 3, the waveform shown in FIG. 6 is obtained, and the defective state determined is likely to match the defective state in the actual mounting state, so that there are few judgment problems. ..

However, in the case of the peeled state as shown in FIG. 4, it is conceivable that the waveform is as shown in FIG. FIG. 7 is a frequency characteristic diagram of the hitting sound obtained by hitting. In the case of FIG. 7, the maximum value B of the amplitude in the high frequency band is not so large as compared with the case of FIG. Therefore, the value of B / A becomes small, and according to the prior art, it is determined that the mounted state is in a good state.

However, as described above, in the peeled state of FIG. 4, there is a risk of falling off, and such a determination is not preferable. That is, there may be a problem of determining that the condition is good even though the condition is actually defective.

On the other hand, in the case of the peeled state as shown in FIG. 5, the total area of the dispersed peeled portions 130 is large, so that the waveform can be as shown in FIG. In this case, the mounted state of the rail bond 100 shown in FIG. 5 is determined to be a defective state. However, as described above, in the case of FIG. 5, the possibility of dropping is low and the condition is actually good. As described above, it may occur that the state is actually in a good state but is determined as a bad state.

Further, the hitting of the rail bond 100 is manually performed by the operator. Therefore, there is a problem peculiar to the inspection of the mounted state of the rail bond 100 that the obtained tapping sound varies and the waveforms shown in FIGS. 6 and 7 also vary. Even with such variations in the waveform, there is a problem that it is not possible to accurately determine the wearing state from the maximum amplitude value. It is possible to automate the striking, but inspecting many rail bonds requires the operator to strik one after another, so there is still a need for the operator to manually strik. ..

Summary,
(1) Due to the variation of the peeled portion, there is a characteristic of the welded rail bond that the risk of falling off due to peeling cannot be accurately determined only by the maximum value of the amplitude of the hammered sound wave type.
(2) Obtaining a tapping sound The striking is performed manually by the operator. Therefore, there is a characteristic peculiar to rail bond that the waveform of the obtained tapping sound tends to vary.

It is difficult to accurately find the peeling ratio and the peeling state of the rail-mounted article only by the maximum value of the amplitude and the amplitude ratio. As a result, the inventor analyzed that it was insufficient in determining the mounting state of rail-mounted articles such as rail bonds.
Based on this analysis, the present invention has been reached in which a plurality of algorithms are combined for determination.

(Embodiment 1)

(Overview)
In the present specification, the rail bond will be described as a rail-mounted article.

FIG. 8 is a block diagram of a rail-mounted article mounting inspection device (hereinafter, abbreviated as “mounting inspection device” if necessary) according to the first embodiment of the present invention. The mounting inspection device may be grasped as a mounting inspection method. For this reason, some of the elements of the mounting inspection device are devices and hardware, and some of them are software (all may be hardware or all may be software). There is a form of.

FIG. 8 shows the basic configuration of the mounting inspection device. The mounting inspection device 1 includes a storage unit 2, an amplitude determination unit 3, an area determination unit 4, a control unit 5, and a final determination unit 6.

The storage unit 2 stores the time-frequency-converted waveform of the tapping sound obtained by striking the rail-mounted article. When a rail-mounted article such as a rail bond mounted on a rail is hit, a tapping sound is obtained. For example, the rail bond attached to the rail is hit with a hammer or the like. The hitting sound produces a hitting sound, which is collected by a microphone or the like.

The hitting sound in the collected state has a hitting sound wave shape along the time axis. When the striking sound wave shape along the time axis is time-frequency-converted, a waveform along the frequency axis can be obtained. For example, a process such as a fast Fourier transform (FFT) may be performed. For example, a waveform along the frequency axis as shown in FIG. 6 can be obtained. The storage unit 2 stores the waveform of the tapping sound along the frequency axis. That is, the waveform stored by the storage unit 2 has a shape along the frequency axis. As a result, the storage unit 2 stores the waveform along the frequency axis as shown in FIG. 6, and the waveform along the frequency axis can be output to the amplitude determination unit 3 and the area determination unit 4. ..

The amplitude determination unit 3 reads out the waveform along the frequency axis stored in the storage unit 2. At this time, one waveform is stored for each of the hitting sounds obtained by hitting each of the rail bonds, and one of the stored waveforms is read out and a determination described later is performed. If it is necessary to determine the next tapping sound, the waveform corresponding to the next tapping sound is read out and the determination is performed. This also applies to the area determination unit 4.

FIG. 9 is an internal block diagram of the amplitude determination unit. The amplitude determination unit 3 determines the mounting state of the rail-mounted article by an amplitude algorithm based on the amplitude ratio of the waveform in a predetermined range. FIG. 10 is a schematic diagram of the waveform. The waveform along the frequency axis stored in the storage unit 2 has a form as shown in FIG. In this waveform, the amplitude determination unit 3 uses a waveform in a predetermined range as shown in FIG.

That is, the amplitude determination unit 3 determines the mounting state by using the amplitude ratio of the waveform in this predetermined range.

The amplitude determination unit 3 includes an amplitude detection unit 31, an amplitude ratio calculation unit 32, and an amplitude algorithm determination unit 33 for determining the mounting state.

The amplitude detection unit 31 detects the maximum value Ka of the amplitude of the waveform in the low frequency band in the predetermined range and the maximum value Kb of the amplitude of the waveform in the high frequency band in the initial range. FIG. 11 is a schematic diagram showing a state in which the amplitude Ka and the amplitude Kb in the waveform in the predetermined range are detected. A predetermined range is set for the waveform. As an example of setting this predetermined range, the predetermined range is from the first frequency to the second frequency.

The predetermined range is divided into a low frequency band and a high frequency band, and the amplitude detection unit 31 detects the maximum value Ka of the amplitude of the waveform in the low frequency band and the maximum value Kb of the amplitude of the waveform in the high frequency band.

The amplitude ratio calculation unit 32 calculates Kb / Ka, which is an amplitude ratio. Here, the maximum value Kb of the amplitude in the high frequency band reflects an unfavorable loudness in the striking sound. That is, it reflects the sound generated by the peeling generated on the welded surface of the rail bond. On the other hand, the amplitude Ka in the low frequency band is not easily affected by the peeling state.

The amplitude algorithm determination unit 33 compares the amplitude ratio Kb / Ka with the amplitude ratio threshold value (arbitrarily set) to determine the mounting state of the rail-mounted article. When the amplitude ratio Kb / Ka is larger than the amplitude ratio threshold value, the amplitude algorithm determination unit 33 determines the mounted state of the rail-mounted article (rail bond) as a defective state, and when it is equal to or less than the amplitude ratio threshold value, it is regarded as a good state. judge.

As described above, the magnitude of the amplitude Kb in the high frequency band reflects the large peeling state of the welded surface of the rail bond, and the fact that the amplitude ratio Kb / Ka is larger than the amplitude ratio threshold is relative. This is because it is considered to indicate that the mounting state is in a defective state.

In this way, the amplitude determination unit 3 determines the mounting state of the rail bond based on the amplitude ratio Kb / Ka, and outputs the determination result to the final determination unit 6.

The amplitude determination unit 3 can determine the mounting state at high speed and clearly.

The area determination unit 4 determines the mounting state of the rail-mounted article by an area algorithm based on the area ratio of the waveform in a predetermined range. The waveform in the predetermined range is the same as that described with reference to FIG. 10, and is a waveform in the range from the first frequency to the second frequency.

FIG. 12 is an internal block diagram of the area determination unit according to the first embodiment of the present invention.

The area calculation unit 41 calculates the area of the waveform in a predetermined range. That is, the area of the waveform in the range from the first frequency to the second frequency is calculated. As shown in FIG. 10, the first frequency and the second frequency are arbitrarily set in the waveform. Here, the relationship is that the first frequency> the second frequency.

The area calculation unit 41 calculates the area of the waveform region between the first frequency and the second frequency. The calculation of the area is shown in FIG. FIG. 13 is a schematic view showing the area calculation of the waveform in the first embodiment of the present invention. The portion of the waveform in FIG. 13 with diagonal lines from the first frequency to the second frequency is the waveform area calculated by the area calculation unit 41.

In this waveform area, the frequency value calculation unit 422 calculates the frequency value of the waveform delimiter having a predetermined area ratio. The calculated frequency value of the delimiter is a specific frequency value. In FIG. 13, a delimiter in which the waveform area has a predetermined ratio is found. Since this delimiter is defined on the frequency axis, this delimiter can be represented by a frequency value. This is a specific frequency value, and the frequency value calculation unit 422 calculates this specific frequency value.

For example, in the waveform area, a specific frequency value that divides the waveform area into 9: 1 is calculated. As an example of the method for measuring the predetermined area ratio, it may be possible to see the ratio of the entire waveform area from the first frequency to the second frequency and the area of the waveform in the range from the first frequency to the specific frequency value. .. That is, the frequency value calculation unit 42 determines the ratio of the total waveform area to the area of a certain range (the area from the first frequency to the specific frequency value or the area from the second frequency to the specific frequency value). , May be measured.

Alternatively, the entire waveform area is separated by a specific frequency value, and the area of the waveform in the range from the first frequency to the specific frequency value and the area of the waveform in the range from the second frequency to the specific frequency value, You may also measure the ratio of.

In FIG. 13, the ratio of Sa and Sb is calculated with the entire waveform area from the first frequency to the second frequency as Sa and the area of the waveform from the first frequency to the specific frequency value as Sb. As an example, a frequency value having a ratio such as Sa: Sb = 9: 1 is calculated (calculated while exploring). Here, in order to explain the area ratio, a specific frequency value is used as a reference for the area ratio, but in reality, by searching for a place where the area ratio becomes a predetermined ratio, a specific frequency value is finally obtained. Is determined. That is, the use of the specific frequency value in the explanation of Sa, Sb, etc. in FIG. 13 does not mean that the area ratio is calculated after the specific frequency value is determined first, but the area ratio is determined. After that, it means that the specific frequency value is determined. For convenience of explaining which range and which range of the area ratio is the area ratio that is finally determined as the area ratio, the separation criteria of Sa and Sb are explained as specific frequency values.

The area ratio is calculated while searching, and the frequency value calculation unit 42 finds a frequency value having a predetermined area ratio. This found frequency value is a specific frequency value.

That is, in FIG. 13, the frequency value calculation unit 422 may calculate the frequency value of the delimiter in which Sb / Sa is a predetermined area ratio as a specific frequency value. Alternatively, in FIG. 13, a delimiter frequency value in which Sb / Sc is a predetermined area ratio may be calculated as a specific frequency value. Alternatively, in FIG. 13, a delimiter frequency value in which Sc / Sa has a predetermined area ratio may be calculated as a specific frequency value.

The frequency value calculation unit 422 calculates the ratio of the area of the waveform in the range on the side based on a certain frequency value, using the waveform area as one reference. A certain frequency value at which this ratio becomes a predetermined ratio is calculated as a specific frequency value.

The area algorithm determination unit 43 determines the rail bond mounting state by comparing the specific frequency value with the area ratio threshold value. Being equal to or less than the area ratio threshold means that the specific frequency value is closer to the low frequency side. At this time, it indicates that the area of the high frequency band is relatively small, and that the peeled state of the welded surface of the rail bond is small. In other words, it can be said that it is in a good condition where the possibility of dropping out is not approaching.

In particular, when the area ratio of the corrugated area is used as a reference, there is little variation due to various modes of peeling. This is because it is considered that the corrugated area shows a state closer to the peeling mode regardless of the peeling area and the position. The tapping sound generates various waveforms depending on the peeling mode. It is considered that the waveforms of these various shapes are determined by the peeling mode, and the waveform determined by the peeling mode more accurately reflects the peeling mode.

The area of the waveform that more accurately reflects the peeling mode is numerical information that more accurately reflects the peeling mode. Therefore, it is considered that the area ratio that more reflects the peeling mode and the specific frequency value converted from this into the judgment criteria more accurately reflect the peeling mode.

The area determination unit 4 can make a more realistic determination.

The final determination unit 6 receives the determination result from the amplitude determination unit 3 and the determination result from the area determination unit 4 having the respective characteristics. Based on the respective determination results, the final determination unit 6 finally determines the mounting state of the rail-mounted article as either a “good state” or a “bad state”. A more accurate judgment can be made by combining the judgment results of the different characteristics and making a final judgment.

That is, there is less concern that the rail bond, which is in a defective state with a risk of falling off, is determined to be in a good state, and that the rail bond, which is in a good state, is determined to be in a defective state. As a result, it is possible to ensure the safety and economy of the rail-mounted articles. Further, as will be described later, it is possible to flexibly control whether to raise the priority of ensuring safety or the priority of ensuring economic efficiency depending on the method of determination by the final determination unit 6.

The defective state indicates a state in which there is a high risk that the rail bond will fall off, which is the opposite of the good state.

As described above, the mounting inspection device 1 according to the first embodiment realizes more accurate and accurate determination of the mounting state of the rail-mounted article, and at the same time, the priority of ensuring safety and the priority of ensuring economic efficiency according to the actual situation. Flexible control such as is possible.

Next, each determination and variations of the final determination based on the determination will be described.

(Part 1: Based on the agreement of each judgment result)
The amplitude determination unit 3 determines that the rail bond mounting state is in a good state if the amplitude ratio Kb / Ka is equal to or less than the amplitude ratio threshold value, and that it is in a bad state if it is larger than the amplitude ratio threshold value. This is a determination made only by the amplitude determination unit 3.

On the other hand, the area determination unit 4 determines that the rail bond mounting state is in a good state if the specific frequency value is equal to or less than the area ratio threshold value, and that the rail bond is in a bad state if it is larger than the area ratio threshold value. This is a determination made only by the area determination unit 4. Both are as described above.

Each of the determination result by the amplitude determination unit 3 and the determination result by the area determination unit 4 is output to the final determination unit 6. Here, the control unit 5 controls the number of determinations by the amplitude determination unit 3, the number of determinations by the area determination unit 4, the determination order between the amplitude determination unit 3 and the area determination unit 4, and the like.

When the determination result of the amplitude determination unit 3 and the determination result of the area determination unit 4 are in a defective state, the final determination unit 6 sets the mounted state of the rail-mounted article as a “defective state” and makes a final determination. To do. On the contrary, when the determination result of the amplitude determination unit 3 and the determination result of the area determination unit 4 are in good condition, the final determination unit 6 sets the mounting state of the rail-mounted article as "good state". Make a final decision.

When the determination result of the amplitude determination unit 3 and the determination result of the area determination unit 4 having different characteristics match each other, the final determination unit 6 adopts the matching result. By adopting this matching result, the mounting state can be determined with higher accuracy.

(Part 2: Sequential judgment)
Each of the amplitude determination unit 3 and the area determination unit 4 determines the mounting state of the rail-mounted article by the processing as described above. Each determination result is a determination result in each algorithm.

The final determination unit 6 is used in each of the amplitude determination unit 3 and the area determination unit 4.
If the judgment result of the first judgment and the judgment result of the second judgment are the same, the final judgment is made based on the first judgment result (in other words, the second judgment result).

If the judgment result of the first judgment and the judgment result of the second judgment are different, the final judgment is made in consideration of the judgment results of the third and subsequent times.

For example
1st Amplitude Judgment Unit: Good Condition Area Judgment Unit: Good Condition 2nd Amplitude Judgment Unit: Good Condition If Area Judgment Unit: Good Condition, the final judgment unit is in good condition judged in each of the 1st and 2nd times. Is the final judgment result.

Or
1st Amplitude Judgment Unit: Good State Area Judgment Unit: Good State 2nd Amplitude Judgment Unit: Bad State Area Judgment Unit: In the case of a bad state Have them make the second judgment.

3rd Amplitude Judgment Unit: Defective State Area Judgment Unit: Defective State

If the third time is such a result, the final determination unit 6 determines the mounting state of the rail-mounted article as a defective state.

Alternatively, in the first and second times, the determination by the set of the amplitude determination unit 3 and the area determination unit 4 is not considered as one set as described above, but each of the amplitude determination unit 3 or the area determination unit 4 is considered. The judgment may be considered as the first time, the second time, and the third time.

In this case, it is determined as follows.

1st amplitude determination unit: good state 2nd area determination unit: good condition When the first judgment is made by the amplitude judgment unit 3 and the second judgment is made by the area judgment unit 4, this is the case. Such a judgment result may be obtained. In this case, the determination results of the different determination algorithms are the same for the first time and the second time, and the final determination unit 6 makes a final determination with the mounted state of the rail-mounted article as a good state.

Or
1st Amplitude Judgment Unit: Good Condition 2nd Area Judgment Unit: Bad Condition

In the case of, the control unit 5 causes the third determination to be processed. In this case, since the first time is the amplitude determination unit 3, the control unit 5 may operate the amplitude determination unit 3 at the third time.

Third amplitude determination unit: If a defective state is obtained, the final determination unit 6 determines the defective state as the final determination in consideration of the results of the second and third times (different determination algorithms).

Or
3rd Amplitude Judgment Unit: Good condition

Then, the final determination unit 6 may conclude that the good condition, which is a large number of results, is the final determination. Alternatively, the control unit 5 causes the area determination unit 4 to perform the fourth determination process.

4th area determination unit: If it is in a good state, 3 times out of 4 times are in a good state, and each of the different judgment algorithms gives a good state result, so that the final judgment unit 6 is in a good state. Is determined as the final determination.

The control unit 5 interlocks with the final determination unit 6 to determine the order of determination in each of the amplitude determination unit 3 and the area determination unit 4, and considers the order, number, ratio, variation, etc. of the individual determination results. The final determination unit 6 determines the final mounting state. The final judgment processing procedure described here is an example, and it is appropriate to decide whether to adopt the order, the number, the ratio, the variation, the first result, the last result, the intermediate result, etc., or to adopt them in combination. It should be decided.

(Part 3: Priority is given to the result of the majority)
Each of the amplitude determination unit 3 and the area determination unit 4 determines the mounting state of the rail-mounted article by the processing as described above. Each determination result is a determination result in each algorithm.

Here, in each of the determination result by the amplitude determination unit 3 and the determination result by the area determination unit 4, (the determination by the amplitude determination unit 3 may be single or multiple times, and the determination by the area determination unit 4 is performed. , It may be single or multiple times), and the final judgment is made by the one with the larger number of judgment results of good condition or bad condition.

For example
1st Amplitude Judgment Unit: Defective State 2nd Area Judgment Unit: Good State 3rd Amplitude Judgment Unit: If it is in a defective state, the final judgment is made as a defective state because the number of times the defective state is judged is large.

Or
1st Amplitude Judgment Unit: Good State 2nd Area Judgment Unit: Good State 3rd Amplitude Judgment Unit: In the case of a bad state, the final judgment unit 6 makes a final judgment as a good state.

Of course, the final judgment may be made using the judgment results of four or more times.

(Part 4: Judgment by weighting)
Each of the amplitude determination unit 3 and the area determination unit 4 determines the mounting state of the rail-mounted article by the processing as described above. Each determination result is a determination result in each algorithm.

Here, it is also preferable that the final determination unit 6 weights each of the determination result of the amplitude determination unit 3 and the determination result of the area determination unit 4, and then makes the final determination based on the total amount of the determination results.

For example
1st Amplitude Judgment Unit: Good Condition 2nd Area Judgment Unit: Bad Condition

In this case, it is assumed that the determination result of the amplitude determination unit 3 is multiplied by 0.7 as the weighting coefficient, and the determination result of the area determination unit 4 is multiplied by 0.5 as the weighting coefficient.

In this case, since 0.7 × good state and 0.5 × bad state, the total amount of good state becomes large. As a result, the final determination unit 6 determines that the condition is good.

Or
1st Amplitude Judgment Unit: Defective State 2nd Area Judgment Unit: Good Condition 3rd Amplitude Judgment Unit: Defective State 4th Area Judgment Unit: Good Condition

0.7 x bad condition, 0.5 x good condition, 0.7 x bad condition, 0.5 x good condition

Therefore, the defective state becomes large as a total amount, and the final determination unit 6 determines as a defective state.

As described above, in addition to the determination processing procedure by the control unit 5, the final determination unit 6 makes a final determination according to various processing procedures.

(Control of order by the control unit)
As described in Part 1 to Part 3, when the final determination is performed by the final determination unit 6, the control unit 5 controls the determination order of the amplitude determination unit 3 and the area determination unit 4. For example, the control unit 5 causes the amplitude determination unit 3 to make a determination first, and finally the area determination unit 4 to make a determination. As an example, the order is as follows.

1st Amplitude Judgment Unit 2nd Amplitude Judgment Unit 3rd Area Judgment Unit

Of course, the determination may be performed in an order other than this example.
Further, the control unit 5 may be made to perform a combination of the determination by the amplitude determination unit 3 at a single time or a plurality of times and the determination by the area determination unit 4 at a single time or a plurality of times. As an example, the order is as follows.

1st Amplitude Judgment Unit 2nd Area Judgment Unit 3rd Area Judgment Unit 4th Amplitude Judgment Unit

Of course, the determination may be performed in an order other than this example.

The final determination unit 6 determines the final mounting state based on the determination result of the amplitude determination unit 3 and the determination result of the area determination unit 4 obtained according to the order of the control unit 5. The determination algorithm in the final determination unit 6 is as described in Part 1 to Part 3 and the like.

The details of other elements will be described below.

(Memory)
As described above, the storage unit 2 stores the waveform of the tapping sound after time-frequency conversion. The operator hits the rail bond attached to the rail with a striking device such as a hammer. By setting a sound collecting device such as a microphone, it is possible to collect the tapping sound generated by hitting. The collected tapping sound is time-frequency axis-converted and converted into a frequency-based waveform. This conversion is performed in an external element of the storage unit 2.

The operator collects tapping sounds one after another for the plurality of rail bonds, and the storage unit 2 stores the waveforms one after another.

The storage unit 2 may be an element included in the wearing inspection device 1 or a separate element. For example, the storage device separate from the mounting inspection device 1 may be the storage unit 2, or the storage device included in the mounting inspection device 1 may be the storage unit 2. For example, when the mounting inspection device 1 is a general-purpose computer, a storage device separate from the general-purpose computer may be the storage unit 2, or the memory included in the general-purpose computer may be the storage unit 2.

Alternatively, the storage unit 2 may be a storage device on the cloud, and may have a mechanism for reading a waveform from the storage device on the cloud by connecting the wearing inspection device 1 to the network.

Since the storage unit 2 is separate from the mounting inspection device 1, in the inspection of the mounting state of the rail bond, the time and place for the operator to collect the tapping sound and to determine the mounting state based on the tapping sound. Will be able to be implemented separately. As a result, work efficiency is improved.

The storage unit 2 may store a plurality of waveforms corresponding to different rail bonds, and the area calculation unit 41 may read out these plurality of waveforms one after another and perform the next processing.

As described above, the mounting state inspection device 1 according to the first embodiment can inspect the mounting state more accurately and suitable for the actual site by combining the amplitude determination algorithm and the area determination algorithm having different characteristics. Further, the control unit 5 and the final determination unit 6 change the determination order and the final determination processing method according to the characteristics (type, position, size, years after installation, etc.) of the target rail-mounted article. Is also suitable.

(Embodiment 2)

Next, the second embodiment will be described. In the second embodiment, an additional device in the area determination unit 4 will be described. The area determination unit 4 has elements as shown in FIG. Details of each part will be described below.

(Area calculation unit)
As described above, the area calculation unit 41 calculates the waveform area of the waveform in the range of the first frequency to the second frequency. Here, the area calculation unit 41 calculates Sa, which is the area in the range from the first frequency to the second frequency. Alternatively, the area calculation unit 41 calculates Sb, which is the area of the waveform in the divided range from the first frequency to the specific frequency value. Alternatively, the area Sc of the waveform in the separated range from the second frequency to the specific frequency value is calculated. Alternatively, a plurality of these are calculated.

In any case, the area calculation unit 41 may calculate the area where the area ratio can be calculated in the entire waveform area from the first frequency to the second frequency. For example, if the area ratio is found as Sa, which is the total area, and Sb, which is the area from the first frequency to the specific frequency value, these two are calculated. Alternatively, if the area ratio is found using the area Sb and the area Sc, each of the areas of Sb and Sc is calculated.

Here, Sc is the area of the difference between Sa and Sb (the remaining portion excluding Sb from Sa), as is clear from FIG.

In this way, the area of the range required to find the area ratio is calculated.

Here, the area calculation unit 41 may calculate the waveform area by gradually calculating the integrated value of the waveform from the first frequency to the second frequency. By gradually calculating, it is possible to record each stepwise area in the middle of calculation and each of the total waveform area integrated up to the final frequency up to the second frequency.

The area calculation unit 41 calculates the waveform area while recording each of the stepwise area and the final area, so that a portion having a predetermined area ratio can be found according to the area calculation.

In this way, the area calculation unit 41 calculates the area of the entire waveform or a partial range in which the area ratio for calculating the specific frequency value can be found. In the calculation, the calculation may be performed by integration or various algorithms.

(Frequency value calculation unit)
The frequency value calculation unit 42 calculates a frequency value in which the two areas have a predetermined area ratio in the waveform area calculated by the area calculation unit 41 as a specific frequency value. The area ratio of the total waveform area from the first frequency to the second frequency and the area of a part of the waveform is calculated. Alternatively, the area ratio of the area of a part of the waveform from the first frequency to the second frequency and the area of the remaining waveform is calculated.

In FIG. 13, for example, the area ratio of Sa, which is the area of the entire waveform, and Sb, which is the area of a part of the waveform on the first frequency side, is calculated. Alternatively, the area ratio of Sb, which is the area of a part of the waveform on the first frequency side of the entire waveform, and Sc, which is the area of a part of the waveform on the second frequency side, is calculated. In either case, the area ratio is calculated in the total waveform area from the first frequency to the second frequency.

The frequency value calculation unit 42 calculates a frequency value having an area ratio that is a preset predetermined area ratio as a specific frequency value. For example, a frequency value in which the area Sa and the area Sb have a predetermined area ratio “9: 1” is calculated as a specific frequency value. A frequency value at which Sb / Sa has a predetermined area ratio is calculated as a specific frequency value. Of course, the predetermined area ratio "9: 1" is an example and may be another value.

Alternatively, a frequency value in which Sb / Sc is a predetermined area ratio may be calculated as a specific frequency value.

In FIG. 13, a certain frequency value is calculated as a specific frequency value. The frequency value calculation unit 42 outputs the calculated specific frequency value to the area algorithm determination unit 43.

Here, the area ratio is arbitrarily determined. For example, after actually measuring the area ratio appropriate for finding a specific frequency value between the actual rail bond that is judged to be in a good state and the actual rail bond that is judged to be in a bad state. Then, the area ratio may be determined. That is, the area ratio may be determined based on actual measurements and history.

Alternatively, the area ratio may be determined from a criterion that is theoretically judged to be a defective state.

It is also preferable that the area ratio is variable. The area ratio used in the frequency value calculation unit 42 is variable, and it is also preferable that the area ratio is changed according to the setting and the situation. Alternatively, the frequency value calculation unit 42, which is variable, may appropriately adopt a necessary area ratio to calculate the specific frequency value.

As described above, the frequency value calculation unit 42 calculates the frequency value having a predetermined area ratio as a specific frequency value. As described above, since the specific frequency value is based on the waveform area reflecting various peeling modes of the welded surface of the rail bond, the specific frequency value more accurately reflects the various peeling modes. ..

The frequency value calculation unit 42 outputs this specific frequency value to the determination unit, so that the area algorithm determination unit 43 can more accurately determine the mounting state of the rail bond.

(Area algorithm judgment unit)
The area algorithm determination unit 43 determines the mounting state of the rail-mounted article by comparing the specific frequency value with a predetermined threshold value. As described above, the specific frequency value is considered to indicate various peeling modes (peeling area, peeling position, number of peeling locations, spreading method, shape, etc.) of the welded surface of the rail bond.

FIG. 14 is an explanatory diagram illustrating determination by the determination unit according to the second embodiment of the present invention. The frequency value calculation unit 42 calculates a specific frequency value. This specific frequency value varies depending on the waveform. In FIG. 14, a specific frequency value A of a certain value and a specific frequency value B of another value are shown. It is calculated based on different waveforms.

The area algorithm determination unit 43 stores a predetermined threshold value in advance. In FIG. 14, a threshold value at a certain frequency is set. The area algorithm determination unit 43 compares the specific frequency value A with the threshold value. Since the specific frequency value A is larger than the threshold value, the rail bond on which the specific frequency value A is based is determined as a “defective state”. Similarly, the area algorithm determination unit 43 compares the specific frequency value B with the threshold value. Since the specific frequency value B is equal to or less than the threshold value, the rail bond that is the basis of the specific frequency value B is determined as a "good state".

The area algorithm determination unit 43 determines the mounting state of the rail bond by comparing the preset area ratio threshold value with the specific frequency value.

The area ratio threshold value may be determined based on an actually measured value in which the actual rail bond mounting state is determined as a good state (or is determined as a defective state). For example, the boundary value between the specific frequency value of the rail bond determined to be in a good state and the specific frequency value of the rail bond determined to be in a bad state may be set as the area ratio threshold value. Further, the actually measured value may be calculated by accumulating a plurality of actually measured results. That is, the area ratio threshold value may be calculated from the accumulation of specific frequency values in the actual mounting state of the plurality of rail bonds.

In determining the area ratio threshold value, it may be determined from actual measurement in advance, or it may be determined by updating during the actual mounting inspection. In the latter case, the optimum area ratio threshold value is recalculated and updated from the specific frequency values accumulated in learning.

It is also preferable that the area ratio threshold value can be changed based on the determination result of the area algorithm determination unit 43. For example, if too many defective states are detected in the determination result, but there is not much dropout in the actual rail bond, the area ratio threshold value may be too strict. In this case, the area ratio threshold is relaxed by increasing the area ratio threshold.

Alternatively, if too many good conditions are detected, but the actual rail bond falls off frequently, the area ratio threshold value may be too loose. In this case, the area ratio threshold value is made stricter, such as by lowering the area ratio threshold value, so that a more actual judgment can be realized.

The area algorithm determination unit 43 outputs these results to the final determination unit 6.
Further, in the first frequency and the second frequency which are predetermined ranges, the first frequency> the second frequency, and the second frequency is equal to or higher than the excess component removal frequency. Here, the excess component removal frequency is a frequency value at which the hitting sound of the rail itself (not including the hitting sound of the rail-mounted article) can be removed in the hitting sound.

If the striking sound of the rail itself is included in the area calculation waveform, the waveform that is not based on the problem (peeling of the welded surface) between the rail-mounted article and the rail will be included in the area calculation, and the identification that reflects the peeling mode will be included. The calculation of the frequency value may be adversely affected. Therefore, this adverse effect is reduced by setting the frequency (which is a part of the low frequency) that is the range of the tapping sound of the rail itself as the excess component removal frequency and the second frequency being equal to or higher than this excess component removal frequency. it can.

(Noise removal in area calculation)
FIG. 15 is a schematic diagram of a waveform including noise according to the second embodiment of the present invention. The waveform of a certain tapping sound after time-frequency conversion may be the waveform shown in FIG. At this time, as shown in FIG. 15, noise 100 may be included in the waveform. Noise as a very peculiar waveform may occur in the waveform instead of the change in gain on the frequency axis.

There are various causes of such noise. For example, when hitting with a hammer, there are impurities such as sand between the hammer and the rail-mounted article, the hit is inappropriate, the microphone, etc. It is conceivable that other sounds may be mixed in when collecting the sound in. If noise due to such an inappropriate cause is included in the calculation of the waveform area, the waveform area will be different from the original waveform area of the rail-mounted article. As a result, it may not be possible to calculate the specific frequency value that should be, and it may not be possible to correctly determine the mounting state.

It is also preferable that the area calculation unit 41 removes the portion determined as the noise 100 from the calculation target of the waveform area to calculate the waveform area. By calculating the waveform area excluding the noise 100 portion, it is possible to calculate the specific frequency value that should be originally. As a result, it is possible to determine the wearing state more accurately.

Whether the noise 100 is regarded as noise or not may be determined from the relationship between the difference between the gain before and after and the bandwidth and the like. Alternatively, it may be judged from a method such as filtering.

(Comparison with multiple thresholds)
FIG. 16 is an explanatory diagram illustrating a state of determination in the determination unit according to the second embodiment of the present invention. The area algorithm determination unit 43 compares the specific frequency value with the threshold value and determines the mounting state. At this time, it is also preferable that the area algorithm determination unit 43 has a plurality of threshold values. In FIG. 16, the area algorithm determination unit 43 has three threshold values A, a threshold value B, and a threshold value C. The area algorithm determination unit 43 may make a determination by stepwise comparing a plurality of threshold values with a specific frequency value. For example, in FIG. 16, the specific frequency value is equal to or greater than the threshold value A and equal to or less than the threshold value B. In this case, it may be determined as an intermediate state between a defective state and a good state.

Alternatively, any one of a plurality of threshold values A to C may be selectively selected and compared for determination. For example, when the area algorithm determination unit 43 selects the threshold value A, the specific frequency value is larger than the threshold value A, so that the rail bond is determined to be in a defective state. On the other hand, when the area algorithm determination unit 43 selects the threshold value B or the threshold value C, the specific frequency value is equal to or less than the threshold value B (threshold value C), so that the rail bond is determined to be in a good state.

By flexibly changing the threshold value according to various factors such as the target rail bond mounting position, rail condition, and mounting timing, it is possible to determine the more actual rail bond mounting state.

As described above, in the mounting inspection device 1 for rail-mounted articles according to the second embodiment, the determination accuracy of the area determination unit 4 can be improved or the variation can be expanded by various means of the area determination unit 4. it can. As a result, it gets closer to reality.

(Embodiment 3)

Next, the third embodiment will be described. In the third embodiment, an additional device in the amplitude determination unit 3 will be described. In addition, as described in the second embodiment, in the additional device in the area determination unit 4, the device commonly used in the amplitude determination unit 3 will also be described.

(Amplitude ratio threshold)
The amplitude algorithm determination unit 33 compares the amplitude ratio Kb / Ka with the amplitude ratio threshold value to determine the mounting state of the rail-mounted article. This amplitude ratio threshold value, like the area ratio threshold value, may be determined based on an actually measured value in which the mounted state of the rail-mounted article is determined to be in a good state.

For example, the boundary value between the specific frequency value of the rail bond determined to be in a good state and the amplitude ratio Kb / Ka of the rail bond determined to be in a bad state may be set as the amplitude ratio threshold value. Further, the actually measured value may be calculated by accumulating a plurality of actually measured results. That is, the amplitude ratio threshold value may be calculated from the accumulation of the amplitude ratio Kb / Ka in the actual mounting state of the plurality of rail bonds.

In determining the amplitude ratio threshold value, it may be determined from actual measurement in advance, or it may be determined by updating during the actual mounting inspection. In the latter case, the optimum amplitude ratio threshold value is recalculated and updated from the amplitude ratio Kb / Ka accumulated in learning.

It is also preferable that the amplitude ratio threshold value can be changed based on the determination result of the amplitude algorithm determination unit 33. For example, if too many defective states are detected in the determination result, but there is not much dropout in the actual rail bond, the amplitude ratio threshold value may be too strict. In this case, relaxation such as increasing the amplitude ratio threshold value is performed.

Alternatively, if too many good conditions are detected but the actual rail bond falls off frequently, the amplitude ratio threshold value may be too loose. In this case, the amplitude ratio threshold value is made stricter, such as by lowering the amplitude ratio threshold value, so that a more actual judgment can be realized.

Further, the amplitude algorithm determination unit 33 has a plurality of amplitude ratio threshold values, and these may be selectively selected. In this case, as described with reference to FIG. 16, the amplitude algorithm determination unit 33 may selectively proceed with the comparison with the amplitude ratio threshold value.

(High frequency band and low frequency band)
The low frequency band is a band including the maximum amplitude in the waveform, and the low frequency band and the high frequency band are separated by a frequency value higher than the frequency value corresponding to the maximum amplitude. By setting the band that does not include the maximum amplitude as the high-frequency band in this way, the amplitude ratio Kb / Ka can be accurately found.

As described above, the mounting inspection device 1 according to the third embodiment has a wide range of accuracy and variations in the amplitude determination unit 3, and can expand the inspection accuracy and inspection target of the final rail-mounted article.

(Embodiment 4)

Next, the fourth embodiment will be described. In the fourth embodiment, a variation will be described in which the amplitude determination unit 3 and the area determination unit 4 directly weight the calculated amplitude ratio and the specific frequency value to perform the final determination. FIG. 17 is a block diagram of the mounting inspection device according to the fourth embodiment of the present invention.

In the amplitude determination unit 3, the amplitude ratio calculation unit 32 calculates the amplitude ratio Kb / Ka. In the area determination unit 4, the frequency value calculation unit 42 calculates a specific frequency value.

The amplitude determination unit 3 further includes an amplitude ratio correction unit 35 that weights the amplitude ratio Kb / Ka to calculate the correction amplitude ratio. Similarly, the area determination unit 4 further includes a specific frequency value correction unit 45 that weights the specific frequency value and calculates the correction specific frequency value.

For example, the amplitude ratio is multiplied by 0.7 as a weighting factor, and the specific frequency value is multiplied by 0.5 as a weighting factor.

The corrected amplitude ratio and the corrected specific frequency value are output to the final determination unit 6. The final determination unit 6 determines the mounting state of the rail-mounted article based on these values corrected by the weighting. For example, the final determination unit 6 compares the modified amplitude ratio with the amplitude ratio threshold (which may be the same as or different from that described in embodiments 1 to 3). Similarly, the final determination unit 6 compares the modified specific frequency value with the area ratio threshold (which may be the same as or different from that described in embodiments 1 to 3).

In each of these comparisons, if both are below the threshold value, it is judged as a good condition. Alternatively, if both are larger than the threshold value, it is determined as a defective state. If the comparison with each threshold value is different, another comparison result is newly obtained and the final judgment is made. These may be determined in a pattern similar to that described in the first to third embodiments of the first embodiment.

(Realization as a method and software)
As described above, the mounting inspection device 1 described in the first to fourth embodiments may have some or all of its elements realized by hardware, software, or hardware. It may be realized by a mixture of hardware and software.

When the final implementation mode of the mounting inspection device 1 is a computer or a mobile terminal, each element such as the amplitude determination unit 3 may be hardware that is a dedicated circuit included inside, or operates by a processor. It may be software that does.

When understood as software, it is realized by the configuration of the processor and the memory accessed by it. Each element is stored in the memory as a computer program in which the processor can operate.

The present invention also includes a state in which the mounting inspection device 1 is software before it is realized in a computer or a mobile terminal. Further, it may be grasped as an invention of a method in which each procedure of the mounting inspection device 1 described in the first to fourth embodiments is executed.

In this case, the process is performed by the following steps.
A storage step that stores the time-frequency-converted waveform of the tapping sound obtained by hitting the rail-mounted article mounted on the rail, and a storage step.
An amplitude determination step that determines the mounting state of the rail-mounted article by an amplitude algorithm based on the amplitude ratio of the waveform in a predetermined range, and
An area determination step for determining the mounted state of a rail-mounted article by an area algorithm based on the area ratio of the waveform in a predetermined range, and
A control step that controls a combination of determination operations in the amplitude determination step and the area determination step, and
A final determination step for determining the final mounting state of the rail-mounted article based on the respective determination results of the amplitude determination step and the area determination step is provided.
The amplitude determination step is
An amplitude detection step for detecting the maximum value Ka of the amplitude in the low frequency band of the waveform in a predetermined range and the maximum value Kb of the amplitude in the high frequency band of the waveform,
Amplitude ratio calculation step to calculate Kb / Ka, which is the amplitude ratio,
It has an amplitude algorithm determination step of comparing the amplitude ratio "Kb / Ka" with the amplitude ratio threshold value to determine the mounting state of the rail-mounted article.
The area determination step is
An area calculation step for calculating the waveform area of a waveform in a predetermined range, and
In the waveform area, a frequency value calculation step for calculating a specific frequency value which is a frequency value for dividing a waveform having a predetermined area ratio, and
It has an area algorithm determination step of comparing a specific frequency value with an area ratio threshold value and determining a mounting state of a rail-mounted article.
The waveform is a shape based on the frequency axis,
The final determination step determines the mounting state as a "good state" or a "bad state" based on the respective determination results of the amplitude algorithm determination step and the area algorithm determination step.
In this way, the present invention may be grasped as a method or software.

The rail-mounted article mounting inspection device described in the first to fourth embodiments is an example for explaining the gist of the present invention, and includes deformation and modification within a range not deviating from the gist of the present invention.

1 Mounting inspection device for rail-mounted articles 2 Storage unit 3 Amplitude determination unit 31 Amplitude detection unit 32 Amplitude ratio calculation unit 33 Amplitude algorithm 35 Amplitude ratio correction unit 4 Area determination unit 41 Area calculation unit 42 Frequency value calculation unit 43 Area algorithm judgment unit 45 Frequency value correction unit 5 Control unit 6 Final judgment unit

Claims (15)

A storage unit that stores the time-frequency-converted waveform of the tapping sound obtained by hitting the rail-mounted article mounted on the rail, and a storage unit.
An amplitude determination unit that determines the mounting state of the rail-mounted article by an amplitude algorithm based on the amplitude ratio of the waveform in a predetermined range.
An area determination unit that determines the mounting state of the rail-mounted article by an area algorithm based on the area ratio of the waveform in a predetermined range.
A control unit that controls a combination of determination operations in the amplitude determination unit and the area determination unit, and
A final determination unit for determining the final mounting state of the rail-mounted article based on the respective determination results of the amplitude determination unit and the area determination unit is provided.
The amplitude determination unit
An amplitude detection unit that detects the maximum value Ka of the amplitude of the low frequency band of the waveform in the predetermined range and the maximum value Kb of the amplitude of the high frequency band of the waveform.
An amplitude ratio calculation unit that calculates Kb / Ka, which is an amplitude ratio,
It has an amplitude algorithm determination unit that compares the amplitude ratio "Kb / Ka" with the amplitude ratio threshold value to determine the mounting state of the rail-mounted article.
The area determination unit
An area calculation unit that calculates the waveform area of the waveform in the predetermined range, and
In the waveform area, a frequency value calculation unit that calculates a specific frequency value that is a frequency value for dividing a waveform having a predetermined area ratio, and
It has an area algorithm determination unit that compares the specific frequency value with the area ratio threshold value and determines the mounting state of the rail-mounted article.
The waveform has a shape based on the frequency axis.
The final determination unit determines the mounting state as a "good state" or a "bad state" based on the respective determination results of the amplitude algorithm determination unit and the area algorithm determination unit. apparatus. The amplitude determination unit determines that the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value as "good state", and that the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value as "bad state". And
The area determination unit determines that the specific frequency value is equal to or less than the area ratio threshold value as a "good state", and that the specific frequency value is greater than the area ratio threshold value as a "bad state".
When the determination result of the amplitude determination unit and the determination result of the area determination unit are in the defective state, the final determination unit makes a final determination of the mounted state of the rail-mounted article as a “defective state”. The mounting inspection device for rail-mounted articles according to claim 1. The amplitude determination unit determines that the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value as "good state", and that the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value as "bad state". And
The area determination unit determines that the specific frequency value is equal to or less than the area ratio threshold value as a "good state", and that the specific frequency value is greater than the area ratio threshold value as a "bad state".
The final determination unit includes the amplitude determination unit and the area determination unit.
If the judgment result of the first judgment and the judgment result of the second judgment are the same, the final judgment is made based on the first judgment result.
The mounting inspection of the rail-mounted article according to claim 1, wherein if the judgment result of the first judgment and the judgment result of the second judgment are different, the final judgment is made in consideration of the third and subsequent judgments. apparatus. The amplitude determination unit determines that the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value as "good state", and that the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value as "bad state". And
The area determination unit determines that the specific frequency value is equal to or less than the area ratio threshold value as a "good state", and that the specific frequency value is greater than the area ratio threshold value as a "bad state".
When the number of the determination results of the "defective state" is the majority in the determination results of the amplitude determination unit and the area determination unit, the final determination unit sets the mounting state of the rail-mounted article to the "defective state". The mounting inspection device for rail-mounted articles according to claim 1, which is finally determined as. The amplitude determination unit determines that the amplitude ratio "Kb / Ka" is equal to or less than the amplitude ratio threshold value as "good state", and that the amplitude ratio "Kb / Ka" is larger than the amplitude ratio threshold value as "bad state". And
The area determination unit determines that the specific frequency value is equal to or less than the area ratio threshold value as a "good state", and that the specific frequency value is greater than the area ratio threshold value as a "bad state".
The final determination unit weights each of the determination result of the amplitude determination unit and the determination result of the area determination unit, and when the total amount of the weighted determination results is equal to or more than a predetermined value, the rail is mounted. The mounting inspection device for a rail-mounted article according to claim 1, wherein the mounting state of the article is finally determined as a "defective state". An amplitude ratio correction unit that calculates a correction amplitude ratio by weighting the amplitude ratio "Kb / Ka" calculated by the amplitude ratio calculation unit.
A specific frequency value correction unit that weights the specific frequency value calculated by the frequency value calculation unit to calculate a correction specific frequency value is further provided.
The final determination unit finally determines the mounted state of the rail-mounted article as a "defective state" when the total amount of the modified amplitude ratio and the modified specific frequency value is equal to or higher than a predetermined value. A mounting inspection device for the described rail-mounted article. The predetermined range is from the first frequency to the second frequency.
The first frequency> the second frequency, and
The second frequency is equal to or higher than the excess component removal frequency.
The mounting inspection device for a rail-mounted article according to any one of claims 1 to 6, wherein the excess component removing frequency is a frequency value at which the tapping sound of the rail itself can be removed in the tapping sound. The mounting inspection device for a rail-mounted article according to claim 7, wherein the area calculation unit calculates the waveform area by calculating an integrated value of the waveform from the first frequency to the second frequency. Let Sa be the area of the waveform from the first frequency to the second frequency, and Sb be the area of the waveform from the first frequency to the specific frequency value.
The mounting inspection device for a rail-mounted article according to claim 7 or 8 , wherein the frequency value calculating unit calculates the specific frequency value in which Sb / Sa is the predetermined area ratio. The rail-mounted article according to any one of claims 1 to 9 , wherein each of the amplitude ratio threshold value and the area ratio threshold value is determined based on an actually measured value in which the mounted state of the rail-mounted article is determined to be in a good state. Inspection device. The mounting inspection device for a rail-mounted article according to claim 10 , wherein each of the amplitude ratio threshold value and the area ratio threshold value is variable or selectively has a plurality of values. The low frequency band includes the maximum amplitude in the waveform.
The mounting inspection device for a rail-mounted article according to any one of claims 1 to 11 , wherein the low frequency band and the high frequency band are separated by a frequency value higher than the frequency value corresponding to the maximum amplitude. The mounting inspection device for a rail-mounted article according to any one of claims 1 to 12, wherein the control unit makes a determination in the order of the amplitude determination unit and finally the area determination unit. The rail mounting according to any one of claims 1 to 12 , wherein the control unit is made to perform a single or a plurality of determinations by the amplitude determination unit and a single or a plurality of determinations by the area determination unit. An article mounting inspection device. A storage step that stores the time-frequency-converted waveform of the tapping sound obtained by hitting the rail-mounted article mounted on the rail, and a storage step.
An amplitude determination step for determining the mounting state of the rail-mounted article by an amplitude algorithm based on the amplitude ratio of the waveform in a predetermined range, and
An area determination step for determining the mounted state of the rail-mounted article by an area algorithm based on the area ratio of the waveform in a predetermined range, and
A control step that controls a combination of the amplitude determination step and the determination operation in the area determination step, and
A final determination step for determining the final mounting state of the rail-mounted article based on the respective determination results of the amplitude determination step and the area determination step is provided.
The amplitude determination step
An amplitude detection step for detecting the maximum value Ka of the amplitude of the low frequency band of the waveform in the predetermined range and the maximum value Kb of the amplitude of the high frequency band of the waveform.
Amplitude ratio calculation step to calculate Kb / Ka, which is the amplitude ratio,
It has an amplitude algorithm determination step of comparing the amplitude ratio “Kb / Ka” with the amplitude ratio threshold value to determine the mounting state of the rail-mounted article.
The area determination step is
An area calculation step for calculating the waveform area of the waveform in the predetermined range, and
In the waveform area, a frequency value calculation step for calculating a specific frequency value which is a frequency value for dividing a waveform having a predetermined area ratio, and
It has an area algorithm determination step of comparing the specific frequency value with the area ratio threshold value to determine the mounting state of the rail-mounted article.
The waveform has a shape based on the frequency axis.
The final determination step determines the mounting state as a "good state" or a "bad state" based on the respective determination results of the amplitude algorithm determination step and the area algorithm determination step. Method.