JP2012159437A - Method for detecting periodic defect and periodic defect detecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a periodic defect detecting device being usable even when a generation period is unknown and being indifferent even when a region required for synchronous addition is a little.SOLUTION: The periodic defect detecting device includes: magnetizers 4 and 4' applying a magnetic field to a steel plate 2; a surface magnetic sensor 5 that detects a magnetic flux leakage leaked to the surface of the steel plate 2 to obtain surface measurement data; a backside magnetic sensor 5' that detects a magnetic flux leakage leaked to the backside of a test object to obtain backside measurement data; periodicity determining device 8 that determines a generation period of the periodic defect based on the surface measurement data and the backside measurement data; a data column adding section 91 that performs a synchronous addition by assembling the surface measurement data and the backside measurement data at the same position in the surface and the backside of the steel plate 2; and a defect determining section 92 that determines whether the test object has the periodic defect or not using the synchronously added measurement data.

Description

  The present invention relates to a periodic defect detection method and a periodic defect detection apparatus for detecting a periodic defect on a ferromagnetic object.

  In the manufacturing process of thin steel sheets, foreign matter attached to the rolls in the production line (or irregularities generated on the rolls themselves when the foreign matter bites into the rolls) is transferred to the steel sheets, and periodic defects called loki are formed on the steel sheets. May occur. These periodic defects, particularly when the steel sheet is being crushed by two rolls, cause irregularities on both sides of the steel sheet (if one surface is convex, the opposite surface is concave). Once these defects occur, they continue to occur until the roll is changed or the process is improved, so it is extremely important from the viewpoint of yield improvement to detect periodic defects early and take countermeasures. is there.

  Conventionally, many detection methods focusing on the periodicity of defects have been proposed as detection methods for this periodic defect. For example, first, the object is measured by a defect detection sensor, and the output signal of the sensor is synchronously added at the expected defect cycle (for example, a length corresponding to one rotation of the final rolling roll of the steel line). A method of enhancing a defect signal having a noise component from a noise component is known (see Patent Document 1).

  However, the synchronous addition can be used only when the defect cycle is known in advance. Therefore, when the defect occurrence period is not known in advance or can be changed, it is common to determine the defect period before the synchronous addition. As an example of the determination of the defect period, a plurality of defect candidates are extracted by the defect detection sensor, and the intervals of these defect candidates are compared to determine the interval as the defect period. There is something to do.

  On the other hand, in an actual production line, the degree of contact between the roll and the steel plate is not necessarily uniform. When the degree of contact between the roll and the steel sheet is weak, the defect signal may be weak, so that the defect may not be detected. In addition, since the periodic defect and noise having no periodicity are mixed in the periodic defect, there is a problem that the periodic defect and its periodicity cannot be accurately detected. Therefore, a technique for emphasizing a defect signal for determining periodicity is known (see Patent Document 2).

JP-A-6-324005 JP 2009-265087 A

  However, in the method of performing synchronous addition over a number of cycles in order to obtain a sufficiently strong signal, for example, as in the method described in Patent Document 2, a region necessary for defect inspection greatly expands in the longitudinal direction. In the manufacturing coil, the flaw detection may not end and the next manufacturing coil may be applied. That is, there are cases in which the two manufactured coils may be rejected products, resulting in a decrease in yield. Therefore, a periodic defect detection device and a periodic defect that can be used even if the generation period of the defect is unknown, and can be used even if the signal from the periodic defect is small, and the area required for synchronous addition may be small. A detection method was desired.

  The present invention has been made in view of the above, and its purpose can be used even if the generation period of periodic defects is unknown, and even if the signal from the periodic defects is minute. An object of the present invention is to provide a periodic defect detection device and a periodic defect detection method that require a small area for synchronous addition.

  In order to solve the above-described problems and achieve the object, a periodic defect detection method according to the present invention includes a magnetization step in which a magnetic field is applied to a subject using a magnetizer, and a magnet disposed on the front side of the subject. A surface measurement step for measuring surface leakage magnetic flux on the surface of the subject using a sensor to obtain surface measurement data, and measuring leakage magnetic flux on the back surface of the subject using a magnetic sensor disposed on the back side of the subject A back surface measurement step for acquiring back surface measurement data, a synchronous addition step for synchronously adding the surface measurement data and the back surface measurement data for the same position on the front surface and the back surface of the subject, and the synchronous addition measurement. And a determination step of determining whether or not the subject has a defect using data.

  In order to solve the above-described problems and achieve the object, a periodic defect detection device according to the present invention detects a magnetizer that applies a magnetic field to a subject and a leakage magnetic flux that leaks to the surface of the subject. A front surface magnetic sensor for acquiring measurement data, a back surface magnetic sensor for detecting a leakage magnetic flux leaking to the back surface of the subject and acquiring back surface measurement data, and a period based on the front surface measurement data and the back surface measurement data Periodicity determining means for determining the occurrence cycle of sexual defects, synchronous adding means for synchronously adding the surface measurement data and the back surface measurement data for the same position on the front and back surfaces of the subject, and the synchronous addition Determination means for determining whether or not the subject has a periodic defect using measurement data.

  According to the periodic defect detection device and the periodic defect detection method of the present invention, the periodic defect generation period can be used even if the generation period of the periodic defect is unknown, and the signal from the periodic defect is minute. The area required for synchronous addition can be reduced.

FIG. 1 is a schematic diagram of a periodic defect detection device according to a first embodiment of the present invention. FIG. 2 is a functional block diagram of the periodicity determination device and the defect determination device. FIG. 3 is a flowchart showing processing of the periodicity determination device and the defect determination device. FIG. 4 is a diagram for explaining the correspondence between the process shown in FIG. 3 and the inspection of the steel sheet. FIG. 5 is a schematic diagram of a measurement signal when the period of the periodic defect and the interval between the small regions coincide with each other. FIG. 6 is a schematic diagram of a measurement signal when the period of the periodic defect and the interval between the small regions do not match. FIG. 7 is a schematic diagram of a measurement signal when the interval between the small regions is changed. FIG. 8 is a flowchart showing preprocessing of the periodicity determination device and the defect determination device. FIG. 9 is a schematic diagram for explaining the thinning process. FIG. 10 is a schematic diagram of a periodic defect detection device according to a second embodiment of the present invention. FIG. 11 is a schematic diagram for explaining measurement signals in the second embodiment. FIG. 12 is a graph showing an example of detecting periodic defects according to the embodiment of the present invention.

  Hereinafter, embodiments of a periodic defect detection device according to the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited by this embodiment.

[First Embodiment]
FIG. 1 is a schematic diagram showing a configuration of a leakage magnetic flux flaw detector 1 according to a first embodiment of the present invention. Note that FIG. 1 is a perspective view for easy understanding of the device arranged under the steel plate 2.

  As shown in FIG. 1, a leakage magnetic flux flaw detector 1 according to a first embodiment of the present invention detects periodic defects 3 present on the surface and surface layer of a steel plate 2 traveling in the direction of an arrow. Here, the periodic defect 3 is one in which surface defects caused by the rolling roll are periodically present in the rolling direction of the steel plate 2 (in the direction of the arrow in FIG. 1). In the present embodiment, the present invention will be described assuming that the subject is a steel plate 2. However, the subject is not limited to the steel plate 2, and a ferromagnetic material in which leakage magnetic flux is generated from a defective portion by applying a magnetic field. In general, the present invention can be appropriately implemented.

  1 includes a magnetizer 4, 4 ′, magnetic sensors 5, 5 ′, signal preprocessing devices 6, 6 ′, A / D converters 7, 7 ′, and periodicity. A determination device 8 and a defect determination device 9 are provided. The magnetizers 4 and 4 ′ are classified into a magnetizer 4 for detecting the surface of the steel plate 2 and a magnetizer 4 ′ for detecting the back surface, and the magnetic sensors 5 and 5 ′ are magnetic sensors for detecting the surface of the steel plate 2. 5 and a magnetic sensor 5 ′ for detecting the back surface.

  The pair of the surface detecting magnetizer 4 and the magnetic sensor 5 are arranged to face each other with the steel plate 2 interposed therebetween, the steel plate 2 is magnetized by the magnetizer 4 from the back side of the steel plate 2, and the leakage flux on the surface of the steel plate 2 is magnetized. Detected by sensor 5. On the other hand, a pair of magnetizer 4 ′ for detecting the back surface and magnetic sensor 5 ′ are arranged to face each other across the steel plate 2, and the steel plate 2 is magnetized from the front side of the steel plate 2 by the magnetizer 4 ′. Is detected by the magnetic sensor 5 '. In the present embodiment, as described above, the magnetizer 4 and the magnetic sensor 5 are arranged so as to face each other with the steel plate 2 interposed therebetween. However, as long as the arrangement space and magnetic interference do not occur, the magnetizer 4 and the magnetic sensor 5 are magnetized. The container 4 and the magnetic sensor 5 can also be arranged on the same side of the steel plate 2. Further, a configuration in which the front surface detection magnetizer 4 and the back surface detection magnetizer 4 ′ are made common is also possible.

  A DC current from a magnetizing power source (not shown) is supplied to the surface detecting magnetizer 4 and the back surface detecting magnetizer 4 ′, and the DC time current of the DC power source is the surface detecting magnetizer 4. And the back surface detection magnetizer 4 'is set equal.

  The magnetic sensors 5 and 5 ′ detect leakage magnetic fluxes on the front and back surfaces of the steel plate 2 as detection signals, and the detected detection signals are subjected to signal processing by signal preprocessing devices 6 and 6 ′ incorporating amplifiers and filter circuits. Via, it is input to the A / D converters 7 and 7 '. The A / D conversion devices 7 and 7 ′ convert the input detection signals into digital measurement data, and output the converted measurement data to the periodicity determination device 8.

  The periodicity determining device 8 and the defect determining device 9 analyze the detection signals detected by the magnetic sensors 5 and 5 ′ as described above, extract the period of the periodic defect 3, and determine the defect. . Hereinafter, the periodicity determination device 8 and the defect determination device 9 will be described in detail with reference to FIG.

  FIG. 2 is a functional block diagram of the periodicity determination device 8 and the defect determination device 9. The periodicity determination device 8 shown in FIG. 2 includes a data storage area 81, a first small area selection unit 82, a first small area data setting unit 83, a second small area selection unit 84, 2 small region data setting unit 85, similarity evaluation index calculation unit 86, and periodicity determination unit 87.

  The data storage area 81 uses the detection signals A / D converted by the A / D converters 7 and 7 ′ as measurement data (in this case, a plurality of magnetic sensors 5 and 5 ′ are arranged in the width direction. This is a storage area that is stored as a dimensional data string. Furthermore, the inside of the data storage area 81 is configured to distinguish and store measurement data from the front surface detection magnetic sensor 5 and measurement data from the back surface detection magnetic sensor 5 ′.

  The first small region selection unit 82 selects measurement data for the first small region on the surface of the steel plate 2 from the measurement data stored in the data storage region 81. The first small region data setting unit 83 sets the size and position of the small region in order for the first small region selection unit 82 to select measurement data for the first small region.

  The second small area selection unit 84 selects the measurement data of the second small area in the measurement data of the front and back surfaces of the steel plate 2 from the measurement data stored in the data storage area 81. The second small area data setting unit 85 sets a distance interval between the small areas so that the second small area selection unit 84 selects the measurement data of the second small area. In the present embodiment, the measurement data on the front surface and the measurement data on the back surface are stored separately in the data storage area 81, and accordingly, the second small region selection unit 84 is used for the measurement data on the front surface. It is also possible to divide the configuration into a second small area selection unit 84 and a second small area selection unit 84 for measurement data on the back surface.

  The similarity evaluation index calculation unit 86 inputs the measurement data selected by the first small region selection unit 82 and the second small region selection unit as described above, and the measurement data of the first small region and the first small region selection unit 82 The similarity evaluation index with the measurement data of 2 small areas is calculated. The periodicity determination unit 87 determines whether or not the measurement data has periodicity based on the similarity evaluation index, and outputs the determination result to the defect determination device 9.

  As shown in FIG. 2, the defect determination device 9 includes a data string addition unit 91, a defect determination unit 92, and a determination result storage unit 93. The data string adding unit 91 synchronously adds the data strings of the measurement data of the magnetic sensor 4 and the magnetic sensor 4 ′ based on the period of the periodic defect determined by the periodicity determining unit 87 of the periodicity determining device 8. The defect determination unit 92 determines whether or not the signal is due to the periodic defect from the measurement data that is synchronously added by the data string addition unit 91. The determination result storage unit 93 is a device for storing the determination result by the defect determination unit 92.

  Hereinafter, processing performed by the periodicity determination device 8 and the defect determination device 9 will be described with reference to the flowchart shown in FIG. Further, in order to explain the correspondence between the processing performed by the periodicity judgment device 8 and the defect judgment device 9 and the measurement position in the steel plate 2, FIG.

  In the process according to the flowchart shown in FIG. 3, first, the minimum roll circumference in the production line is set to the initial value of the period candidate d (the minimum value of the defect period).

  Thereafter, the first small area selection unit 82 uses the measurement data of the steel sheet 2 stored in the data storage area 81 as a reference area with the area A having the size in the width direction h and the rolling direction k of the steel sheet 2 as a reference area. Are selected (step S1). See FIG. 4 (a).

  From the measurement data of the steel plate 2 stored in the data storage area 81, the second small area selection unit 84 has the area A at the position where the width direction is the same as the area A and at a distance d from the area A in the rolling direction. Region B is selected with the same size as (i.e., h × k) (step S2). See FIG. 4 (a).

  Further, the second small area selection unit 84 similarly selects the area C at a position 2d away from the area A, selects the area D at a position corresponding to the area A on the back surface of the steel plate 2, and d from the area D. Region E is selected at a distant position. Area A corresponds to the first small area described above, and areas B, C, D, and E correspond to the second small area described above.

ここで、ｘ（ｉ，ｊ）は、鋼板２の表面の測定データにおける幅方向ｉ番目かつ圧延方向ｊ番目の点の値であり、ｙ（ｉ，ｊ）は、鋼板２の裏面の測定データにおける幅方向ｉ番目かつ圧延方向ｊ番目の点の値である。 The similarity evaluation index calculation unit 86 performs correlation values R _AB , R _BC , R _CD , region A and region B, region B and region C, region C and region D, region D and region E according to the following formula 1. R _DE is obtained (step S3).

Here, x (i, j) is the value of the i-th point in the width direction and j-th direction in the measurement data of the surface of the steel plate 2, and y (i, j) is the measurement data of the back surface of the steel plate 2. Is the value of the i th point in the width direction and the j th point in the rolling direction.

Thereafter, the similarity evaluation index calculation unit 86 similarly calculates the similarity evaluation index R by adding the correlation value R _AB , the correlation value R _BC , the correlation value R _CD, and the correlation value R _DE according to the following formula 2. (Step S4). Here, the addition of the correlation values of a plurality of regions is intended to eliminate the influence of variations and deviations for each region, and is a unique process for detecting periodic defects.

  Next, the periodicity determination unit 87 determines that there is a periodic defect candidate when the evaluation index R is equal to or greater than a preset threshold value (step S5). When it is determined by this determination that there is a periodic defect candidate, the determination result is transmitted to the defect determination device 9 with the period d as a determination result and a data string of measurement data.

  On the other hand, if it is not determined by this determination that there is a periodic defect candidate, the period candidate d is changed to d + Δ (d = d + Δ), and the processing from step S2 to step S5 is repeated (step S6). See FIG. 4 (b).

  Here, the shift amount Δ of the period candidate is set to a constant smaller than the size k in the rolling direction of a predetermined region. In order to eliminate detection omission, Δ <k / 2 is desirable. The range of change of the period candidate shift amount d is a range of periods in which periodic defects can occur. In the line of the steel plate 2, it is desirable to cover the circumference of the roll in the production line where roll wrinkles can occur. In addition, since the change of the shift amount d of the cycle candidate may be performed within a range before and after the circumference assumed for each roll (for example, a range of about several tens of mm), the circumference of each roll in the line is greatly different. If there is a range in which no periodic defect is generated, the period candidate shift amount d need not be set in the range.

  Next, the process of steps S1 to S6 is repeated with the position of the region A serving as a reference for similarity evaluation shifted in the rolling direction (step S7). As a shift amount at this time, it is desirable to set a value smaller than ½ of the size k in the rolling direction of the region in order to eliminate detection omission. The lower limit of the shift amount is the rolling direction sampling interval measured by the digitized sensor output. However, since it takes a calculation time, it can be appropriately adjusted and set.

  Further, the process of steps S1 to S7 is repeated by shifting the position of the region A serving as the reference for similarity evaluation in the width direction (step S8). As a shift amount at this time, it is desirable to set a value smaller than ½ of the size h in the width direction of the region in order to evaluate without omission.

As a result of performing the above periodicity determination, when it is determined that there is a periodic defect candidate of the period d, the data string adding unit 91 has the same measurement position on both surfaces of the steel plate 2 based on the period d. A data string of defect candidate detection signals is synchronously added a predetermined number of times (in this embodiment, five times from region A to region E) (step S9). That is, as described above, if x (i, j) and y (i, j) are the values of the i-th point in the width direction and the j-th direction in the rolling direction in the measurement data of the front and back surfaces of the steel plate 2, synchronous addition is performed. The value A (i, j) is a value represented by the following expression 3.

  In the synchronous addition performed by the data string adding unit 91, since the position and cycle of the defect candidate are clearly known in the previous step, the synchronous addition is performed only for the range having the same size (or the same extent) as the region A. This is sufficient for defect detection. In other words, the signals of the defective portions of the small areas A, B, C on one side and the small areas D, E on the opposite side at the time when it is determined that the defect candidate is included may be most simply added synchronously.

  When the synchronous addition value A (i, j) exceeds a preset threshold value, the defect determination unit 92 determines that a periodic defect exists at this position, and the defect rolling direction, the width direction position, The length of the period and the measurement data of the surrounding area are stored in the determination result storage unit 93 (step S10).

  As the threshold value setting in the above determination, a method of determining a noise level N using measurement data determined as having no periodic defect and setting 3N as a threshold value using this noise level N is conceivable. (That is, if S / N> 3, it is a defect). Note that the noise level N may be determined by the maximum value of the predetermined area or the mean square error. Further, in the case of a defect having a spread in the width direction, there is an effect of improving the S / N when the same rolling direction position is integrated in the width direction, so this method can be combined.

  The processing performed by the periodicity determination device 8 and the defect determination device 9 according to the first embodiment of the present invention has been described above. However, the processing procedure is an example, and the processing procedure may be changed as appropriate. For example, in the above processing procedure, the iterative loop for changing the period candidate d is in the iterative loop for changing the position of the region A, but the reverse may be possible. Also, the evaluation process of the similarity evaluation index R in step S5 has been described to be executed every time the similarity evaluation index R is calculated in the above processing procedure. However, after all the repetition processes are completed, the similarity evaluation is performed. You may make it perform the evaluation process of the parameter | index R. FIG. Further, the change in the width direction position in the data area A in step S8 may not be performed when one-dimensional data is targeted.

Here, the correlation values R _AB , R _BC , R _CD , R _DE and the similarity evaluation index R used for the above-described periodicity determination will be described in more detail with reference to FIGS.

Conventionally, even if sensors are installed on the front and back sides of the steel plate 2, the measurement data is independently evaluated for each surface. However, in the present invention, the surface at the occurrence position of the flaw in the flaw caused by the roll, since it became clear that can be detected equivalent signals on the back both the correlation value R _AB based on this _{finding, R} BC, _{R CD,} R _DE and similarity evaluation index R are defined.

  5 to 7 are schematic diagrams of detection signals for explaining the principle of periodicity determination using the similarity evaluation index. FIG. 5 is a schematic diagram showing a case where the period p of the periodic defect and the interval d between the first and second small regions coincide with each other. FIG. FIG. 7 is a schematic diagram showing a case where the distance d between the second small regions and the second small region do not coincide with each other. FIG. 7 shows the interval p ′ between the periodic defect period p and the first and second small regions. It is the schematic diagram which showed the case where it changed to. 5 to 7, for ease of illustration, a part of a two-dimensional data string is extracted and described using a one-dimensional data string.

When the period p of the periodic defect shown in FIG. 5 and the interval d between the first and second small regions coincide with each other, if the periodic defect is included in the first small region A, the second on the surface. The small regions B and C and the second small regions D and E on the back surface also include periodic defects. Therefore, according to the above-described equation 1, when calculating the product-sum operation in the first small area A and the second small area B (that is, calculating the correlation values R _AB , R _BC , R _CD , R _DE ), The signal due to the defect is amplified. Therefore, when the similarity evaluation index R obtained by adding these four correlation values R _AB , R _BC , R _CD , and R _DE is calculated, the similarity evaluation index R is a large value reflecting the periodicity of defects.

On the other hand, when no defect is included in the first small area A as shown in FIG. 6 or the interval d does not coincide with the defect period p, among the correlation values R _AB , R _BC , R _CD , R _DE Even if any of them happens to have a large value, all of them do not become a large value, so the similarity evaluation index R does not increase. Therefore, the similarity evaluation index R defined above can be used for periodicity determination when the interval d between the small regions is set as a period candidate.

  The processing performed by the periodicity determination device 8 and the defect determination device 9 described above is based on the above-described concept. For example, as shown in FIG. 7, the interval d is slightly changed to d ′ (= d + Δd). By changing each one to cover a predetermined range (for example, up to the maximum possible length of the defect period), and determining the similarity evaluation index R for each interval d, the periodicity with the interval d as a period can be determined. Do it.

  Next, processing of the period determination device 7 and the defect determination device 8 according to the present embodiment will be described from another viewpoint with reference to the flowchart of FIG. The process described below is a process obtained by adding a pre-process to the process shown in FIG. 3. In order to avoid duplication of explanation, the description of the flowchart shown in FIG. 8 is also made by quoting the steps shown in FIG. Omitted.

  First, as an initial setting, the sampling pitch of the A / D converters 7 and 7 'is set to a value that can detect the minimum length of the defect, and measurement data of the magnetic sensors 5 and 5' are acquired (step S11). The measurement data is stored in the data storage area 81 (step S12). At this time, the measurement data of the magnetic sensor 5 for front surface measurement and the measurement data of the magnetic sensor 5 ′ for back surface measurement are stored in the data storage area 81.

  Next, a low-pass filter (for example, moving average) is applied to the rolling direction of the acquired measurement data (step S13). This is because the data is thinned out when the subsequent periodicity calculation is performed, so that the defect can be detected even if the defect position and the sampling position are shifted. That is, it is considered that there is a possibility that a defect signal does not remain when the thinning process is performed with the measurement data as it is.

  Thereafter, by using the measurement data subjected to the process of step S13 at a rate of once every several sampling pitches (for example, once every 4 times or once every 2 to 8 times). Then, periodicity evaluation data is created (step S14). Then, the created periodicity evaluation data is stored in the data storage area 81 (step S15).

  Thereafter, the processing from step S1 to S8 shown in FIG. 3 is performed (step S16). When a storage area for storing periodicity evaluation data is provided separately from the data storage area 81 in step S15, step S16 replaces the measurement data stored in the data area 81 with periodicity evaluation data. This is performed using the data stored in the data storage area.

  And when it becomes periodicity in the process of step S16 (namely, step S1-S8 of FIG. 3), the process of step S9-S10 of FIG. 3 is performed and defect determination is performed (step S17).

  Here, with reference to FIG. 9, a supplementary explanation of the thinning process in step S <b> 14 will be given.

  As can be understood from FIG. 9 (A), a series of width direction data separated at equal intervals in the width direction at the same position in the rolling direction is used (in the figure, striped measurement data having a pitch of 4p is used. ) And defect detection failure may occur depending on the timing. Therefore, in order to avoid such a situation, for example, as shown in FIGS. 9B and 9C, it is conceivable to suppress omission of defect detection by shifting the position in the rolling direction in the width direction and thinning out. .

[Second Embodiment]
FIG. 10 is a schematic diagram showing the configuration of the leakage magnetic flux flaw detector 1 according to the second embodiment of the present invention. The leakage flux testing apparatus 1 according to the first embodiment is configured to excite in the longitudinal direction of the steel plate 2, but the leakage flux testing apparatus 1 according to the second embodiment is configured to magnetize in the width direction of the steel plate 2. In this configuration, the magnetic flux in the vertical direction is arranged in the width direction to detect the magnetic flux leakage. Note that FIG. 7 is also a perspective view for easy understanding of the device arranged under the steel plate 2.

  7 includes a magnetizer 4, 4 ′, magnetic sensors 5, 5 ′, signal preprocessing devices 6, 6 ′, A / D conversion devices 7, 7 ′, and periodicity. A determination device 8 and a defect determination device 9 are provided. The magnetizers 4 and 4 ′ are classified into a magnetizer 4 for detecting the surface of the steel plate 2 and a magnetizer 4 ′ for detecting the back surface, and the magnetic sensors 5 and 5 ′ are magnetic sensors for detecting the surface of the steel plate 2. 5 and a magnetic sensor 5 ′ for detecting the back surface.

  The pair of the surface detecting magnetizer 4 and the magnetic sensor 5 are arranged to face each other with the steel plate 2 interposed therebetween, and the steel plate 2 is magnetized in the width direction by the magnetizer 4 from the back side of the steel plate 2 to leak the surface of the steel plate 2. Magnetic flux is detected by the magnetic sensor 5. On the other hand, a pair of magnetizer 4 'for detecting the back surface and magnetic sensor 5' are arranged to face each other with the steel plate 2 interposed therebetween, and the steel plate 2 is magnetized in the width direction by the magnetizer 4 'from the front side of the steel plate 2. 2 is detected by the magnetic sensor 5 '.

  Since the periodicity determination device 8 and the defect determination device 9 have the same configuration as that of the first embodiment, detailed description thereof is omitted individually. That is, the periodicity determination device 8 and the defect determination device 9 have the same configuration as that shown in FIG. 2 and can execute the processing shown in the flowcharts of FIGS. 3 and 8.

  According to the leakage magnetic flux flaw detector 1 having the above configuration, the leakage magnetic flux generated by the defect has a positive / negative distribution in the width direction (see FIG. 11). In the leakage magnetic flux flaw detector 1 according to the present embodiment, since the detection signal has such a characteristic positive and negative peak, even if data sampling (thinning processing of measurement data) as shown in FIG. 9C is performed. The probability of non-detection can be reduced, and the accuracy of periodicity evaluation can be increased.

  Further, according to the leakage magnetic flux flaw detector 1 according to the present embodiment, since detection signals relating to defects are generated across a plurality of magnetic sensors in the width direction, similarity of defect signals having two-dimensional characteristics is evaluated. By doing so, it becomes possible to evaluate more accurately in evaluating periodicity.

[Detection example]
Here, a detection example of a periodic defect according to the implementation of the present invention will be described.

  FIG. 12 is a graph showing the result of synchronous addition of periodic defect signals for 5 cycles on one side and 5 cycles on the back side. As can be understood from the graph shown in FIG. 12, a signal considered to be a defect signal is suppressed at a position of 200 mm of the steel plate 2 by suppressing the influence of ambient noise. Therefore, it can be understood that defects can be detected equally even by evaluation using defect data obtained from the front and back surfaces.

  In the conventional periodic defect detection, in order to detect a minute periodic defect on a steel plate with a plate width of 1800 mm, the measurement data is repeatedly correlated with 10 cycles, and a defect with a cycle of about 2000 mm is placed at a position of 350 mm in the width direction. It was necessary to detect. In that case, even if the rolling roll was replaced to eliminate the source of the defect, the steel plate for 20 m had passed through the line so far, and this part could not be shipped as a product.

  However, according to the implementation of the present invention, it is possible to perform evaluation using data measured on both the front and back surfaces of a steel plate having the same specifications. That is, in order to detect a periodic defect having a period of about 2000 mm as described above, it is sufficient to perform correlation processing from measurement data for five periods on the front surface and five periods on the back surface. Therefore, in the past, steel sheets for 20 m were rejected, but only 10 m were rejected as products.

  As described above, according to the periodic defect detection device according to the embodiment of the present invention, the magnetizers 4 and 4 ′ for applying a magnetic field to the steel plate 2 and the leakage magnetic flux leaking to the surface of the steel plate 2 are detected to obtain surface measurement data. Generation of periodic defects based on the surface measurement data and the back surface measurement data, and the back surface magnetic sensor 5 ′ for acquiring the back surface measurement data by detecting the leakage magnetic flux leaking to the back surface of the steel plate 2. Using the periodicity determination device 8 that determines the period, the data string addition unit 91 that synchronously adds the surface measurement data about the same position on the front and back surfaces of the steel plate 2 and the back surface measurement data, and the synchronously added measurement data And the defect determination unit 92 for determining whether or not the steel sheet 2 has a periodic defect, even if the detection signal from the periodic defect is very small, the area required for the synchronous addition may be small, and Area required for the inspection is also improved yield since it is unnecessary to extend to the subject produced in the next lot.

  Furthermore, according to the periodic defect detection method according to the embodiment of the present invention, the periodicity determination device 8 includes the first small region selecting unit 82 that selects the first small region from the surface measurement data, and the first measurement from the surface measurement data. The second small region on the front surface is selected at a position separated from the one small region by the distance of the period candidate, and the second small region on the back surface is selected at a position corresponding to the first small region and the second small region on the front surface from the back surface measurement data. The second small region selection unit 84 that selects two small regions, and the similarity of period candidates from the correlation values of the measurement data in the first small region, the second small region on the front surface, and the second small region on the back surface Since the similarity evaluation index calculation unit 86 for calculating the evaluation index and the period determination unit 87 that determines the period candidate as the generation period of the periodic defect when the similarity evaluation index is equal to or greater than the threshold value, the periodic defect Can be used even if the generation period is unknown And allows the determination of the periodic defect in the synchronized addition of the measurement data of the small region only.

  As described above, according to the periodic defect detection method according to the embodiment of the present invention, the magnetization step of applying a magnetic field to the subject using the magnetizers 4 and 4 ′ and the magnetic sensor 5 arranged on the front side of the steel plate 2 are used. Measuring the leakage flux on the surface of the steel plate 2 to obtain the surface measurement data, and measuring the leakage flux on the back surface of the steel plate 2 using the magnetic sensor 5 'disposed on the back side of the steel plate 2 to measure the back surface. Defects in the steel plate 2 using the back side measurement step for acquiring data, the synchronous addition step for synchronously adding the surface measurement data and the back side measurement data for the same position on the front and back surfaces of the steel plate 2 and the synchronously added measurement data And a determination step for determining whether or not there is a region, the area required for synchronous addition may be small even if the detection signal from the periodic defect is minute, and the region required for defect inspection is Yield since it is unnecessary to extend to the subject produced in Tsu preparative also improved.

  Furthermore, according to the periodic defect detection method according to the embodiment of the present invention, the similarity evaluation index calculation for calculating the similarity evaluation index of the measurement data for the periodic candidates of the periodic defect by combining the surface measurement data and the back surface measurement data And a cycle selection step of selecting a cycle candidate as a synchronous addition cycle when the similarity evaluation index is equal to or greater than a threshold, and the synchronous addition step performs synchronous addition based on the synchronous addition cycle, so that a periodic defect occurs. Even if the period is unknown, it can be used.

  In addition, according to the periodic defect detection method according to the embodiment of the present invention, the first small region selection step for selecting the first small region from the surface measurement data, the first small region and the period candidate from the surface measurement data. A second sub-region selection step for selecting the second sub-region of the front surface at a position away from the distance, and a position of the back surface at a position corresponding to the first sub-region and the second sub-region of the front surface from the back surface measurement data A second small area selection step on the back surface for selecting the second small area, and a correlation calculation step for obtaining correlation values of measurement data in the first small area, the second small area on the front surface, and the second small area on the back surface. Since the similarity evaluation index calculation step calculates the similarity evaluation index based on the correlation value, periodicity determination can be efficiently performed from the measurement data, and the measurement data for only the small area is measured. Periodicity with synchronous addition Perform the judgment of Recessed.

  As mentioned above, although embodiment of this invention was described, it should not be understood that the statement and drawing which make a part of disclosure of said embodiment limit this invention. For example, a magnetic sensor such as a Hall element, a coil, a magnetoresistive element, or a SQUID can be used as a leakage magnetic flux sensor. Further, although a plurality of sensors are arranged in the width direction, a method of traversing one or a plurality of sensors may be used. Further, although the entire width is flaw-detected, a method of flaw-detecting a partial region in the width direction may be used. In particular, in the case of periodic defects, a method may be used in which a part of the region in the width direction is flawed for a certain length, and the flaw is repeatedly detected by changing the position in the width direction.

DESCRIPTION OF SYMBOLS 1 Leakage magnetic flux inspection apparatus 2 Steel plate 3 Periodic defect 4 Magnetizer 5 Magnetic sensor 6 Signal preprocessing apparatus 7 A / D converter 8 Periodicity determination apparatus 9 Defect determination apparatus 81 Data storage area 82 1st small area selection part 83 First small region data setting unit 84 Second small region selection unit 85 Second small region data setting unit 86 Similarity evaluation index calculation unit 87 Periodicity determination unit 91 Data string addition unit 92 Defect determination unit 93 Determination result storage Part

Claims (5)

A magnetization step for applying a magnetic field to the subject;
A surface measurement step for measuring leakage magnetic flux on the surface of the subject to obtain surface measurement data;
A back surface measurement step for measuring back surface magnetic flux on the back surface of the subject and acquiring back surface measurement data;
A synchronous addition step of synchronously adding the surface measurement data and the back surface measurement data for the same position on the front and back surfaces of the subject in combination;
A determination step of determining whether the subject has a periodic defect using the synchronously added measurement data;
A periodic defect detection method comprising: A similarity evaluation index calculating step of calculating the similarity evaluation index of the measurement data for a periodic candidate of periodic defects by combining the front surface measurement data and the back surface measurement data;
A cycle selection step of selecting the cycle candidate as a synchronous addition cycle when the similarity evaluation index is greater than or equal to a threshold; and
The periodic defect detection method according to claim 1, wherein the synchronous addition step performs synchronous addition based on the synchronous addition period. A first small region selection step of selecting a first small region from the surface measurement data;
A second small area selection step for selecting a second small area of the surface at a position separated from the first small area by the distance of the period candidate from the surface measurement data;
A second small region selection step on the back surface, which selects the second small region on the back surface at a position corresponding to the first small region and the second small region on the front surface from the back surface measurement data;
A correlation calculating step for obtaining a correlation value of measurement data in the first small region, the second small region on the front surface, and the second small region on the back surface;
The periodic defect detection method according to claim 2, wherein the similarity evaluation index calculating step calculates a similarity evaluation index based on the correlation value. A magnetizer that applies a magnetic field to the subject;
A surface magnetic sensor for detecting magnetic flux leaking to the surface of the subject and acquiring surface measurement data;
A back surface magnetic sensor for detecting back surface measurement data by detecting leakage magnetic flux leaking to the back surface of the subject;
Periodicity determining means for determining the occurrence period of periodic defects based on the front surface measurement data and the back surface measurement data;
Synchronous addition means for synchronously adding the surface measurement data and the back surface measurement data for the same position on the front and back surfaces of the subject in combination,
Defect determination means for determining whether or not the subject has a periodic defect using the synchronously added measurement data;
A periodic defect detection device comprising: The periodicity determining means includes
A first small region selection step of selecting a first small region from the surface measurement data;
A second small area on the surface is selected from the surface measurement data at a position separated from the first small area by a distance of a period candidate, and the first small area and the first number of the front surface are selected from the back surface measurement data. A second small area selecting means for selecting the second small area on the back surface at a position corresponding to the two small areas;
Similarity evaluation index calculation means for calculating a similarity evaluation index of the period candidates from correlation values of measurement data in the first small area, the second small area on the front surface, and the second small area on the back surface; ,
A periodicity determining unit that determines the periodic candidate as an occurrence period of the periodic defect when the similarity evaluation index is equal to or greater than a threshold;
The periodic defect detection device according to claim 4, comprising:

Images