JP2019194552A - Wire rope inspection device and method - Google Patents

Abstract

To provide a wire rope inspection device and method which can detect a disconnection point on a wire rope with a high precision and high accuracy.SOLUTION: Since a wire rope inspection device (11) according to the present invention performs an inspection of a rope (1) without contacting it, the wire rope inspection device detects the disconnection point starting from a signal with a low S/N ratio. In order to do this, the device includes: a frequency analyzer (111) for measuring the signal by a non-contact type magnetizing device (121) and a magnetic sensor (122) and deriving the frequency characteristics for each time interval for the leakage flux signal of the test object wire rope digitized by the A/D converter (13); and a disconnection determining device (112) for judging the existence or absence of a disconnection point by the frequency characteristic of the leakage flux signal of the wire rope to be tested derived by the frequency analyzer utilizing the decision model formed in advance by using the frequency characteristics for each time interval for the leakage magnetic flux signals of the learning wire rope.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a wire rope inspection apparatus and method used for inspection of a wire rope.

  Currently, wire ropes are used in various fields. This wire rope is a consumable item, and the cutting of the wire rope often causes a serious accident. For this reason, it is very important to conduct maintenance and inspection on the wire rope with high accuracy. The wire rope inspection device is a device that supports an operator so that the maintenance inspection work can be easily performed.

  Some wire rope inspection devices detect a broken wire location by using a measurement result of leakage magnetic flux leaking from the wire rope based on the magnetization of the wire rope (see, for example, Patent Document 1 and Non-Patent Document 1). This prior art utilizes the property that the leakage magnetic flux increases locally at the disconnection location.

Japanese Patent No. 5178471

Kaneda, Kawada, Hayashi, Tokui, "Damage detection of elevator wire rope by wavelet analysis", Transactions of the Society of Instrument and Control Engineers, Vol.34, No.10, 1466/1471 (1998)

  In the prior art as described in Patent Document 1, the detection of the broken portion using the measurement result of the leakage magnetic flux is performed by extracting the peak point of the voltage signal output from the sensor that detects the magnetic flux and extracting the peak point. And the changes in the surrounding signals. However, it is necessary to use a sensor that can detect magnetic flux in a non-contact manner in order to reduce the cost of maintenance and inspection work.

  When measuring leakage magnetic flux in a non-contact manner, the sensor outputs a signal having a low signal-to-noise ratio. For this reason, in the conventional method of detecting the broken point by extracting the peak point of the output signal from the sensor, it is difficult to extract the peak point itself. As a result, this conventional method cannot be expected to detect a disconnection point with high accuracy.

  Further, in the conventional method described in Non-Patent Document 1, attention is paid to the frequency characteristics of the output signal, not the output signal itself of the sensor, wavelet analysis is performed on the signal output from the sensor, The disconnection point is detected by using two straight lines on the frequency-intensity space derived from a signal in a range of only white noise not including a damaged portion as a threshold value. Note that using two straight lines as a threshold corresponds to setting a threshold for each frequency component.

  Usually, the wire rope has a structure in which a plurality of strands having a structure in which a plurality of strands are twisted are further twisted. Due to such a structure, the surface of the wire rope has irregularities due to strands of strands and strands.

  The irregularities on the surface caused by the structure of the wire rope periodically change the leakage magnetic flux. Periodic changes in leakage flux become periodic noise. Due to this periodic noise, it is actually difficult to obtain a straight line as a threshold value in the frequency-intensity space. For this reason, the conventional method using two straight lines in the frequency-intensity space as a threshold cannot be expected to detect a broken portion with high accuracy.

  The present invention has been made to solve such a problem, and an object of the present invention is to provide a wire rope inspection apparatus and method capable of detecting a disconnection point on a wire rope in a non-contact and highly accurate manner.

  The wire rope inspection device according to the present invention is premised on being used for inspection of a wire rope, has a shape that covers the entire circumference of the wire rope in a non-contact manner, a magnetizer that magnetizes the wire rope, It has a shape that covers the entire circumference of the wire rope by contact, and by sampling and quantizing the output voltage from the magnetic sensor that changes the output voltage by changing the leakage magnetic flux from the wire rope A frequency analysis is performed using an A / D converter that outputs a digital signal representing the measurement result of leakage magnetic flux and a digital signal output from the A / D converter when the wire rope to be inspected is magnetized by a magnetizer. Frequency analyzer for deriving frequency characteristics by time and frequency by time derived from the measurement results of leakage flux from magnetized learning wire rope Using a decision model generated from sexual, having a disconnection determination unit determines the presence or absence of broken point in the inspection target wire rope from the frequency characteristic of each time-frequency analyzer is derived.

  According to the present invention, it is possible to detect a disconnection point on a wire rope in a non-contact and highly accurate manner.

It is a figure explaining the structural example of the wire rope inspection apparatus which concerns on Embodiment 1 of this invention. It is a figure explaining the application example of the wire rope inspection apparatus which concerns on Embodiment 1 of this invention. It is a figure explaining the example of installation of a measurement module. It is a figure explaining the time change example of the digital signal output from an A / D converter, when the leakage magnetic flux of the wire rope which has a broken part is detected using a non-contact-type measurement module. It is a figure explaining the time change example of the digital signal output from an A / D converter, when the leakage magnetic flux of the wire rope which has a broken part is detected using a contact-type measurement module. It is a figure explaining the example of the frequency characteristic obtained by wavelet transformation, when the leakage magnetic flux of a disconnection location is detected. It is a figure explaining the example of the frequency characteristic in the area | region of the two-dimensional coordinate of time and a frequency by the two-dimensional coordinate of a frequency and an amplitude. It is a figure explaining the example of the frequency characteristic in the other area | region of the two-dimensional coordinate of time and a frequency by the two-dimensional coordinate of an amplitude. It is a flowchart explaining the flow of the whole process in the case of making the wire rope inspection apparatus which concerns on Embodiment 1 of this invention determine the disconnection location of a wire rope. It is a figure explaining the example of the mask pattern used as a disconnection determination model with the wire rope inspection apparatus which concerns on Embodiment 2 of this invention.

  Hereinafter, each embodiment of the wire rope inspection apparatus and method according to the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
FIG. 1 is a diagram illustrating a configuration example of a wire rope inspection apparatus according to Embodiment 1 of the present invention.

  The wire rope inspection apparatus 11 according to the first embodiment is an apparatus corresponding to a plurality of wire ropes as shown in FIG. Although three wire ropes 1 are shown in FIG. 1, the number of wire ropes 1 that can be handled by the wire rope inspection device 11 is not particularly limited.

  The wire rope inspection apparatus 11 detects a breakage point where it is estimated that a breakage has occurred in the wire rope 1 by utilizing the characteristic that when the wire rope 1 is magnetized, the leakage magnetic flux increases at the breakage point. Hereinafter, unless otherwise specified, the disconnection location is used to indicate an estimated disconnection location.

  Measurement module 12 is used to measure the leakage flux of each wire rope 1, and A / D (Analog to Digital) conversion is used to convert the analog signal output from measurement module 12 by measuring the leakage flux to a digital signal. A vessel 13 is used.

  The measurement module 12 suppresses leakage of magnetic flux into the atmosphere, and two magnetizers 121 that are spaced apart from each other in the longitudinal direction of the wire rope 1 and a magnetic sensor group 122 that is disposed between the two magnetizers 121. Yoke 123 is provided. The two magnetizers 121 cover the entire circumference of the wire rope 1 in a non-contact manner for each wire rope 1. Thereby, the magnetizer 121 generates a uniform magnetic field over the entire circumference of the wire rope 1 and magnetizes each wire rope 1. As the magnetizer 121, a permanent magnet or an electromagnet can be used. The shape of these magnets is such that the wire rope 1 is sandwiched between two annular and semi-annular ones, the annular one is divided into a plurality of blocks, the plurality of block-like ones, and the wire rope 1 It is arranged so as to surround the entire circumference. The shape may be any shape as long as a uniform magnetic field can be generated over the entire circumference of the wire rope 1 as a result.

  Each of the magnetic sensors constituting the magnetic sensor group 122 measures the leakage magnetic flux of one wire rope 1, that is, converts the leakage magnetic flux into an analog signal and outputs the analog signal. Each magnetic sensor has a shape that covers the entire circumference of the wire rope 1 in a non-contact manner, like the magnetizer 121, so that the measurement sensitivity of the leakage magnetic flux becomes uniform along the entire circumference of the wire rope 1. As the magnetic sensor, a coil, a Hall element, various MR (Magneto Resistive) elements, or the like can be used. These magnetic sensors have two annular and semi-annular ones arranged so that the wire rope 1 is sandwiched between them, one in which the annular one is divided into a plurality of pieces, and a plurality of block-like ones in the wire rope It is arranged so as to surround the entire circumference of 1. In the first embodiment, each magnetic sensor measures the leakage magnetic flux in a non-contact manner in order to reduce the cost of maintenance and inspection work.

  The yoke 123 is a structure that is used to fix the two magnetizers 121 and the magnetic sensor group 122 apart from each other in the longitudinal direction of the wire rope 1 in addition to suppressing leakage of magnetic flux to the outside. The yoke 123 is manufactured by processing a member such as iron or steel into a cylindrical shape or a plate shape. In the first embodiment, this magnet 123 is used to combine the two magnetizers 121 and the magnetic sensor group 122 as the measurement module 12.

  FIG. 2 is a diagram for explaining an application example of the wire rope inspection apparatus according to the first embodiment of the present invention. Here, an application example of the wire rope inspection apparatus according to the first embodiment will be specifically described.

  FIG. 2 shows a schematic configuration example of a rope type elevator. In the rope type elevator, as shown in FIG. 2, the car 21 and the counterweight 22 are connected by the wire rope 1 to reduce the load when the car 21 is raised and lowered. In the configuration example shown in FIG. 2, the wire rope 1 is wrapped around the sheave 23 twice by hanging the wire rope 1 on the sheave 23 via the deflecting wheel 24 and further hanging on the sheave 23 via the deflecting wheel 24. The full wrap method is adopted.

  In order to specify the disconnection location, it is necessary to detect the leakage magnetic flux while changing the positional relationship between the measurement module 12 and the wire rope 1. In the rope type elevator as shown in FIG. 2, the car 21 is raised and lowered by winding and lowering the wire rope 1 via the sheave 23. At the time of winding up and down, the wire rope 1 moves in the longitudinal direction. For this reason, even if it is the measurement module 12 fixed, most of the magnetic flux leakage of the wire rope 1 can be measured.

  As the installation location of the measurement module 12, the locations 25 and 26 shown in FIG. 2, that is, between the sheave 23 and the deflector 24 can be considered. The reason is that at these locations 25 and 26, the wire rope 1 is supported by the sheave 23 and the deflecting wheel 24, and the position variation in the radial direction of the wire rope 1 becomes smaller. For this reason, both the contact between the wire rope 1 and the magnetizer 121 and the contact between the wire rope 1 and the magnetic sensor can be avoided or minimized. Thereby, the magnetizer 121 and the magnetic sensor can be used for a longer time.

  FIG. 3 is a diagram illustrating an installation example of the measurement module. FIG. 3A is an example in the direction of the arrow B shown in FIG. 2 when the measurement module 12 is installed at the location 25, and FIG. 3B is a diagram when the measurement module 12 is installed at the location 26. The example in the viewpoint of the arrow A direction shown in FIG.

  As shown in FIGS. 3A and 3B, the number of wire ropes 1 varies depending on the locations 25 and 26. The number of wire ropes 1 at the location 25 located on the upper side is twice the number of wire ropes 1 at the location 26 located on the lower side. This is because the full wrap method in which the wire rope 1 is wound around the sheave 23 twice is adopted. In the place 26, not only the number of the wire ropes 1 is small, but also the interval between the wire ropes 1 becomes wider. Therefore, the place 26 is preferable as the place where the measurement module 12 is installed. By installing in the place 26, it can be expected that the manufacturing cost of the measurement module 12 is further suppressed.

  In the elevator shown in FIG. 2, as shown in FIGS. 3A and 3B, the number of wire ropes 1 and the interval between adjacent wire ropes 1 differ depending on the installation location. Therefore, the measurement module 12 needs to be manufactured assuming an installation location. From this, the measurement module 12 shown in FIG. 1 is an example, and the elevator shown in FIG. 2 is not assumed.

  The measurement module 12 is a relatively heavy structure. Therefore, relatively large power is required to move the measurement module 12. Since the facility such as an elevator that moves the wire rope 1 itself in the longitudinal direction does not need to move the measurement module 12, it is preferable to apply the wire rope inspection apparatus according to the first embodiment. Applicable facilities and structures are not limited to elevators.

  Returning to the description of FIG. The A / D converter 13 samples an analog signal input from each magnetic sensor constituting the magnetic sensor group 122 and quantizes the sample value, thereby converting the input analog signal into a digital signal. The sampling period is set in consideration of the relative moving speed between the assumed measurement module 12 and the wire rope 1, and the quantization resolution is set in consideration of the local increase in leakage magnetic flux at the disconnection point. Has been.

  For example, when the measurement module 12 is fixed without moving, when the width in the longitudinal direction of the wire rope 1 at the minimum disconnection point to be detected is w and the winding speed or the unwinding speed of the wire rope 1 is v, the sampling period t Is to satisfy the relationship of t ≦ w / (10 × v). Further, when ΔE is a local increase in the leakage magnetic flux assumed at the disconnection location, the quantization resolution r is set to satisfy the relationship r ≦ ΔE / 10.

  As shown in FIG. 1, the wire rope inspection device 11 includes a frequency analysis device 111, a disconnection determination device 112, and a data storage device 113 in addition to the measurement module 12 and the A / D converter 13. The frequency analysis device 111, the disconnection determination device 112, and the data storage device 113 are realized by one or more information processing devices, that is, a processing device such as a CPU (Central Processing Unit) mounted on the information processing device. This is realized by executing a dedicated program.

  The frequency analysis device 111 derives a temporal change in frequency characteristics from the digital signal output from the A / D converter 13, that is, a frequency characteristic for each time. Wavelet transform, short-time Fourier transform, multi-resolution analysis, or the like can be used for deriving the temporal change in frequency characteristics. Here, it is assumed that a temporal change in frequency characteristics is extracted using wavelet transform.

  It is assumed that the derivation of the temporal change in the frequency characteristic is performed when the car 21 is moved, that is, when the car 21 is raised or lowered. From this, the temporal change corresponds to a change in the positional relationship between the measurement module 12 and the wire rope 1.

  The moving speed of the car 21 is not always constant. The sampling period t needs to be changed according to the moving speed v. From this, the wire rope inspection apparatus 11 acquires the information regarding the moving speed v directly or indirectly from the control panel that controls the movement of the car 21, and the frequency analysis apparatus 111, the disconnection determination apparatus 112, and the data The storage device 113 is controlled. With this control, it is possible to specify the correspondence between the time at which the location determined to be disconnected has passed through the measurement module 12 and the position in the longitudinal direction of the wire rope 1 at the location determined to be disconnected.

  Here, with reference to FIGS. 4 to 8, the digital signal output from the A / D converter 13 and the wavelet transform using the digital signal, that is, the frequency characteristics extracted by time-frequency analysis will be specifically described. . The digital signal output from the A / D converter 13 is hereinafter also referred to as “magnetic flux measurement value”.

  FIG. 4 shows a time change example of a digital signal having a low S / N ratio output from the A / D converter when the leakage magnetic flux of the wire rope having a broken portion is measured by the non-contact measurement module 12. It is a figure explaining. FIG. 5 illustrates a time change example of a digital signal having a high S / N ratio output from the A / D converter when the leakage magnetic flux of a wire rope having a broken portion is measured by the contact type measurement module 12. It is a figure to do. 4 and 5, the horizontal axis represents time, and the vertical axis represents a digital voltage signal, that is, a magnetic flux measurement value.

  When the non-contact type measurement module 12 is used, the analog signal output from the magnetic sensor is a signal having a low S / N ratio. However, even with a signal having a low S / N ratio, the irregularities such as the strands constituting the wire rope 1 and the strands constituting the strands cause periodic changes in the analog signal.

  Unlike the leakage magnetic flux that appears periodically due to the unevenness on the wire rope 1, the leakage magnetic flux from the broken portion appears in an impulse shape. Therefore, as shown in FIG. 4 and FIG. 5, the magnetic flux measurement value waveform at the disconnection portion has a periodicity different from the periodic noise.

  FIG. 6 is a diagram for explaining an example of frequency characteristics obtained by wavelet transform when leakage magnetic flux at a disconnection point is detected. FIG. 6A shows an example of frequency characteristics using three-dimensional coordinates of time, frequency, and amplitude, and FIG. 6B shows frequency characteristics of an image by expressing the amplitude of FIG. 6A in shades or colors. An example is shown.

  The wavelet transform has temporal resolution in addition to frequency resolution. For this reason, the frequency characteristics obtained by the wavelet transform can represent temporal changes in the frequency characteristics as shown in FIGS. 6 (a) and 6 (b). The expression of the frequency component at the position on the time axis, that is, the amplitude value, means that the disconnection location on the wire rope 1 can be specified in position.

  In the magnetic flux measurement value changing in an ideal impulse shape, the amplitudes of all frequency components are uniform. From this, as shown in FIG. 6A and FIG. 6B, the time when the leakage magnetic flux at the disconnection point is measured, that is, the frequency characteristic in the region 62 is in a state where the amplitudes of all frequency components are almost uniform. It becomes. FIG. 7 shows an example of frequency components in this area 62 by two-dimensional coordinates of frequency and amplitude.

  On the other hand, in the region 61 where the leakage magnetic flux at the disconnection portion is not measured, the amplitude is large only at a specific frequency and a frequency in the vicinity thereof. This is due to periodic noise generated by irregularities on the wire rope 1. FIG. 8 shows an example of frequency characteristics in this region 61 by two-dimensional coordinates of frequency and amplitude. As shown in FIG. 8, the periodic noise generated by the irregularities on the wire rope 1 forms a peak at a specific frequency.

  The frequency analysis device 111 performs wavelet transform using a group of magnetic flux measurement values within a time range that falls within a predetermined number of samplings, and the conversion result, that is, information indicating a frequency component including time domain information is a disconnection determination device. To 112. The disconnection determination device 112 uses the wavelet transform conversion result input from the frequency analysis device 111 to determine whether or not there is a disconnection point within the time range corresponding to the magnetic flux measurement value group used for the wavelet transform of the wire rope 1. Interpolation every time, that is, every sampling point on the time domain by the A / D converter 13 or when the digital signal is interpolated by the frequency analyzer 111 at a time interval shorter than the sampling time of the A / D converter 13 Do it point by point. Thereby, the disconnection determination device 112 determines the presence / absence of a disconnection location depending on the resolution in the time domain, for example, the sampling period t, every time the frequency analysis device 111 performs one time frequency analysis.

  The disconnection determination device 112 determines the presence / absence of a disconnection location using a model constructed for the determination of the presence / absence of a disconnection location. Each frequency component, that is, an amplitude value for each frequency, is used as an input of the model. The model is hereinafter referred to as “disconnection determination model”.

  This disconnection determination model is constructed for each type of wire rope, and in the case of inspection of the wire rope 1, a disconnection determination model for the wire rope 1 is used. The disconnection determination model for the wire rope 1 is constructed by using the same measurement module 12, the same A / D converter 13, and the same frequency analysis device 111 as the learning data and the wavelet transform conversion result obtained from the wire rope 1. Is. The same A / D converter 13 is an A / D converter having the same quantization resolution r and substantially the same sampling period t. The desired sampling period t varies with the relative moving speed v between the measurement module 12 and the wire rope 1.

  As the learning wire rope 1 for obtaining learning data, a plurality of first wire ropes having no disconnection portions and a plurality of second wire ropes having disconnection portions are used. It is desirable that various types of disconnection states are included in a learning wire rope having a disconnection portion. For example, to learn multiple wire ropes with different numbers of disconnections when the disconnections are concentrated locally, and multiple wire ropes with different disconnection positions such as peak / valley / outside / inside of the rope. It should be usable. It is desirable that learning wire ropes having no broken portions include those in various states. For example, it is preferable that a plurality of wire ropes having different undulations, flatnesses, and variations in rope diameter can be used for learning. A breakage determination model is constructed by machine learning using the wavelet transform conversion results obtained by measuring the leakage magnetic flux of each of these wire ropes 1, that is, the parameters used for the breakage determination model are optimized. The conversion result of the wavelet transform is stored as learning data and used for learning in another wire rope inspection apparatus 11. The disconnection determination model can be generated on the disconnection determination device 112 using deep learning, for example, by causing the disconnection determination device 112 to implement a multilayer neural network as a disconnection determination model.

  As described above, the leakage magnetic flux at the disconnection portion has an impulse waveform. The frequency analysis by Fourier transform is performed on the magnetic flux measurement values in the time range used for the frequency analysis. Therefore, in the frequency analysis by Fourier transform, the frequency component due to the leakage magnetic flux at the disconnection portion is diffused within the time range, and it is difficult to determine the presence or absence of the disconnection portion from the frequency component. For this reason, the first embodiment employs time frequency analysis using wavelet transform. Since the frequency characteristics for each time can be extracted by time-frequency analysis, diffusion of the frequency characteristics in the time domain due to the leakage magnetic flux at the disconnection location is avoided or suppressed, and the presence / absence of the disconnection location from the frequency characteristics at each time is determined. Can be performed with high accuracy.

  In the first embodiment, the result of the time frequency analysis is used as the input of the disconnection determination model on the disconnection determination device 112. As described above, the disconnection determination model is generated by machine learning using the frequency analysis result obtained by actually measuring the leakage magnetic flux of the wire rope 1. Therefore, the influence on the determination of periodic noise due to the unevenness of the wire rope 1 is eliminated or greatly suppressed. Also from this, the disconnection location on the wire rope 1 can be detected with high accuracy. Therefore, even if the S / N ratio of the analog signal output from the magnetic sensor of the measurement module 12 based on the leakage magnetic flux of the wire rope 1 is low, it is possible to detect the disconnection location on the wire rope 1 with high accuracy.

  The data storage device 113 stores the determination result of the disconnection location output as data from the disconnection determination device 112. The storage location is, for example, a hard disk drive, an optical disk, a semiconductor storage device, an information processing device for data storage, and the like. Examples of the semiconductor storage device include a USB (Universal Serial Bus) memory, an SD (Secure Digital) card, and the like. Examples of the optical disk include a DVD-RAM (Digital Versatile Disk Random Access Memory) and a DVD-RW (ReWritable). Hereinafter, the data indicating the determination result of the disconnection location output from the disconnection determination device 112 is referred to as “disconnection inspection data”. Note that a communication device for transmitting data to another information processing device may be provided together with or instead of the data storage device 113.

  The disconnection inspection data is stored in the data storage device 113 together with, for example, the magnetic flux measurement value and the wavelet transform result. This is to enable the operator to visually recognize changes in the magnetic flux measurement values, changes in the frequency components over time, and the like.

  FIG. 9 is a flowchart for explaining the overall processing flow when the wire rope inspection apparatus according to the first embodiment of the present invention is to determine the broken portion of the wire rope. Next, the overall processing flow will be described in detail with reference to FIG. With this overall processing flow, the wire rope inspection method according to the first embodiment of the present invention is realized.

  First, in step S <b> 1, a disconnection determination model is generated on the disconnection determination device 112. As described above, the generation of the disconnection determination model is performed by learning using a plurality of wire ropes 1 without disconnection portions and a wavelet transform conversion result obtained from each of the plurality of wire ropes 1 having disconnection portions. Is called. The conversion result of the wavelet transform may be obtained in real time by actually magnetizing the learning wire rope 1 and measuring the leakage magnetic flux and performing frequency analysis using the measurement result. It may be obtained by saving and reading the data.

  Steps S <b> 2 to S <b> 5 are processes that are executed when the disconnection location of the wire rope 1 that is actually the inspection target is detected using the disconnection determination model generated in step S <b> 1. Here, in addition to the disconnection determination device 112, a basic flow including processing executed by each magnetic sensor, the A / D converter 13, and the frequency analysis device 111 constituting the magnetic sensor group 122 of the measurement module 12 is included. Is shown. It should be noted that the magnetizer 121 is in a state where the wire rope 1 is always magnetized while the disconnection portion of the wire rope 1 to be inspected, that is, the disconnection portion is being detected.

  In step S2, each magnetic sensor outputs an analog signal having a voltage corresponding to the detected leakage magnetic flux, thereby converting the detected leakage magnetic flux into a voltage. In step S3, the analog signal output from each magnetic sensor is sampled by the A / D converter 13 at the sampling period t, the sample value is quantized with the resolution r, and the digital signal indicating the quantization result, that is, the magnetic flux measurement value is obtained. Output.

  In step S4, the frequency analysis device 111 performs wavelet transform using a magnetic flux measurement value group within a time range in which a predetermined number of samplings are accommodated, and extracts a temporal change in frequency characteristics. In step S5, the disconnection determination device 112 determines whether or not it is a disconnection portion, that is, a disconnection portion, using the temporal change of the frequency characteristic extracted by the frequency analysis device 111, that is, the frequency characteristic for each time as an input of the disconnection determination model. To detect the disconnection. The detection of this disconnection part completes the basic processing flow.

  Conversion of the leakage magnetic flux measured by each magnetic sensor into a voltage and digitization by the A / D converter 13 are continuously performed while energized. While the power is being supplied, the magnetization of the wire rope 1 by the magnetizer 121 is continuously performed. The time-frequency analysis by the frequency analysis device 111 may be performed every time a necessary number of magnetic flux measurement values are obtained, or may be divided after the magnetic flux measurement values of the entire range to be inspected of the wire rope 1 are obtained. Also good. The determination of the presence or absence of a disconnection portion by the disconnection determination device 112 may be performed in a unit of set of frequency characteristics for each time obtained by one time frequency analysis, or the entire range of the wire rope 1 to be inspected. It may be performed in a lump after the frequency characteristics of each are obtained.

Embodiment 2. FIG.
In the first embodiment, the disconnection determination device 112 detects the presence or absence of a disconnection location, that is, the detection of the disconnection location, using a disconnection determination model that is a multilayer neural network. In contrast, the second embodiment uses a mask pattern of frequency characteristics as a disconnection determination model. In the following description, the reference numerals used in the first embodiment are used as they are, and only the portions different from the first embodiment are noted.

  FIG. 10 is a diagram illustrating an example of a mask pattern used as a disconnection determination model in the wire rope inspection apparatus according to the second embodiment of the present invention. Here, an example of the mask pattern 71 is shown on two-dimensional coordinates with the horizontal axis representing frequency and the vertical axis representing amplitude.

  As shown in FIG. 8, in the portion of the wire rope 1 where there is no breakage point, the frequency characteristic forms a peak at a specific frequency due to the influence of periodic noise due to the unevenness of the wire rope 1. In the second embodiment, paying attention to this, the mask pattern 71 is generated. Generation of such a mask pattern 71 also corresponds to parameter optimization.

  For this reason, in step S1 shown in FIG. 9, the disconnection determination device 112 is caused to generate a mask pattern 71 by learning using the conversion results of the wavelet transform respectively obtained from a plurality of wire ropes 1 without disconnection portions. . In step S <b> 5, the disconnection determination device 112 determines the presence / absence of a disconnection location based on whether or not there is a frequency component overlapping with the mask pattern 71, or the degree of overlap of the frequency component, for example, the size of the overlapping area.

  In the prior art disclosed in Non-Patent Document 1, the disconnection location is detected by using two straight lines in the frequency-intensity space as threshold values. The leakage magnetic flux at the broken portion appearing in an impulse shape is in a state where the amplitudes of all frequency components are almost uniform. This means that the amount of change of each frequency component due to the presence or absence of a broken portion is relatively small.

  The prior art disclosed in Non-Patent Document 1 is based on the premise that a rope tester including a magnetic sensor is pressed against a wire rope. Therefore, the S / N ratio of the analog signal output from the magnetic sensor is larger than the analog signal output from the magnetic sensor that detects the leakage magnetic flux in a non-contact manner. Therefore, when using a magnetic sensor that measures leakage magnetic flux in a non-contact manner, the amount of change in each frequency component due to the presence or absence of a broken portion is further reduced.

  If the straight line is used as a threshold value, the straight line is determined so as to avoid a peak appearing at a specific frequency when the frequency characteristic as shown in FIG. Therefore, there is a high possibility that the frequency characteristics of the disconnection location will be equal to or less than a straight line where the amplitudes of all frequencies are threshold values. This means that noise greatly affects the detection of the disconnection location.

  On the other hand, in the case of the mask pattern 71 as shown in FIG. 10, not only the influence of periodic noise due to the irregularities of the wire rope 1 can be avoided, but also more frequency components can be used for detecting the disconnection location. It becomes possible. For example, at a frequency with a small amplitude, the amplitude becomes larger due to the leakage magnetic flux at the broken portion. The frequency with the small amplitude also affects the detection of the broken portion. For this reason, compared to the prior art disclosed in Non-Patent Document 1, it is possible to detect a disconnection point with higher accuracy.

Embodiment 3 FIG.
In the first embodiment, the disconnection determination device 112 detects a disconnection location using a disconnection determination model that is a multilayer neural network. On the other hand, the third embodiment uses a self-encoder, which is a kind of neural network, as the disconnection determination model. In the following description, the reference numerals used in the first embodiment are used as they are, and only the portions different from the first embodiment are noted.

  The self-encoder is generated on the disconnection determination device 112. In the learning for generating the self-encoder, the conversion results of the wavelet transform respectively obtained from a plurality of wire ropes 1 having no broken portions are used. In the self-encoder thus generated, the difference between the input and the output falls within a certain range in the portion where there is no disconnection. The difference between the input and the output becomes larger at the portion where the disconnection is present. For this reason, the threshold value is determined using the wavelet transform result obtained from each of the plurality of wire ropes 1 having a broken portion.

  In the third embodiment, the above is performed in step S1. In step S5, the disconnection determination device 112 compares the input and the output, and determines that there is a disconnection portion when the absolute value of the difference is equal to or greater than a threshold value.

  Even if such a self-encoder is used as a disconnection determination model, more frequency components can be used effectively as in the second embodiment. Therefore, the disconnection location can be detected with higher accuracy than in the past.

  In the first to third embodiments, the frequency analyzer 111 directly inputs a magnetic flux measurement value that is a digital signal from the A / D converter 13, but the magnetic flux leakage value is from the A / D converter 13. It is not necessary to input directly. That is, the frequency analysis device 111 may acquire the magnetic flux leakage value from a storage medium that stores the magnetic flux leakage value, or may acquire the magnetic flux leakage value via a network. For this reason, the frequency analysis device 111 and the disconnection determination device 112 may not be used simultaneously with the measurement module 12 and the A / D converter 13.

  DESCRIPTION OF SYMBOLS 1 Wire rope, 11 Wire rope inspection apparatus, 12 Measurement module, 13 A / D converter, 71 Mask pattern, 111 Frequency analysis apparatus, 112 Disconnection determination apparatus, 113 Data storage apparatus, 121 Magnetizer, 122 Magnetic sensor group.

Claims (15)

A wire rope inspection device used for inspection of a wire rope,
A shape that covers the entire circumference of the wire rope in a non-contact manner, and a magnetizer that magnetizes the wire rope;
It has a shape that covers the entire circumference of the wire rope in a non-contact manner, and a magnetic sensor that changes an output voltage by a change in leakage magnetic flux from the wire rope;
An A / D converter that outputs a digital signal representing the measurement result of the leakage magnetic flux by the magnetic sensor by sampling and quantizing the output voltage from the magnetic sensor;
A frequency analysis device for performing frequency analysis using a digital signal output from the A / D converter and deriving frequency characteristics for each time when the wire rope to be inspected is magnetized by the magnetizer;
Using the determination model generated from the frequency characteristics for each time derived from the measurement result of the leakage magnetic flux from the magnetized learning wire rope, the inspection is performed from the frequency characteristics for each time derived by the frequency analyzer. A disconnection determination device for determining the presence or absence of a disconnection point in the target wire rope;
Wire rope inspection device having The disconnection determination device determines the presence or absence of the disconnection location for each position in the time domain by using the frequency characteristics for each time, and detects the disconnection location.
The wire rope inspection apparatus according to claim 1. The learning wire rope is a first wire rope without the disconnection part and a second wire rope with the disconnection part,
The determination model includes the measurement results of the leakage magnetic flux from the first wire rope and the time analysis obtained by the frequency analysis performed using the measurement results of the leakage magnetic flux from the second wire rope. Generated by learning using the frequency characteristics of
The wire rope inspection apparatus according to claim 1 or 2. The learning wire rope is a first wire rope without the disconnection portion,
The determination model is generated by learning using the frequency characteristics for each time obtained by the frequency analysis performed using the measurement result of the leakage magnetic flux from the first wire rope as learning data.
The wire rope inspection apparatus according to claim 1 or 2. The determination model is a neural network.
The wire rope inspection apparatus according to claim 3 or 4. The determination model is a self-encoder.
The wire rope inspection apparatus according to claim 4. The determination model is a mask pattern obtained by setting an amplitude threshold for each frequency.
The wire rope inspection apparatus according to claim 4. The magnetizer and the magnetic sensor are attached to the lower wire rope across the sheave and the deflector,
The wire rope inspection apparatus according to any one of claims 1 to 7. A wire rope inspection method used for wire rope inspection,
Model generation that generates a determination model for determining the disconnection location by learning using frequency characteristics for each time derived by frequency analysis using the measurement result of leakage magnetic flux from magnetized learning wire rope as learning data Steps,
Leakage magnetic flux observation step for converting the magnetic flux leakage from the magnetized wire rope to be inspected into an analog voltage signal;
An A / D conversion step of converting the analog voltage signal into a digital signal;
A frequency analysis step for deriving a frequency characteristic for each time from the digital signal;
Using the determination model, a disconnection determination step of determining the presence or absence of the disconnection location in the inspection target wire rope from the frequency characteristics for each time derived in the frequency analysis step,
A wire rope inspection method. The determination of the presence or absence of the disconnection location is performed for each sampling point on the time domain of the frequency characteristics by using the frequency characteristics for each time.
The wire rope inspection method according to claim 9. The learning wire rope is a first wire rope without the disconnection part and a second wire rope with the disconnection part,
The determination model includes the frequency for each time obtained by the frequency analysis performed using the measurement result of the leakage magnetic flux from the first wire rope and the measurement result of the leakage magnetic flux from the second wire rope. It is generated by learning using characteristics as learning data.
The wire rope inspection method according to claim 9 or 10. The learning wire rope is a first wire rope without the disconnection portion,
The determination model is generated by learning using the frequency characteristics for each time obtained by the frequency analysis performed using the measurement result of the leakage magnetic flux from the first wire rope as learning data.
The wire rope inspection method according to claim 9 or 10. The determination model is a neural network.
The wire rope inspection method according to claim 11 or 12. The determination model is a self-encoder.
The wire rope inspection method according to claim 12. The determination model is a mask pattern obtained by setting an amplitude threshold for each frequency.
The wire rope inspection method according to claim 12.

Images