WO2009123296A1 - 周期性欠陥検出装置及びその方法 - Google Patents
周期性欠陥検出装置及びその方法 Download PDFInfo
- Publication number
- WO2009123296A1 WO2009123296A1 PCT/JP2009/056914 JP2009056914W WO2009123296A1 WO 2009123296 A1 WO2009123296 A1 WO 2009123296A1 JP 2009056914 W JP2009056914 W JP 2009056914W WO 2009123296 A1 WO2009123296 A1 WO 2009123296A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- small
- evaluation index
- defect
- periodic
- small area
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
- G01N27/72—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables
- G01N27/82—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables for investigating the presence of flaws
- G01N27/83—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables for investigating the presence of flaws by investigating stray magnetic fields
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/8914—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the material examined
- G01N2021/8918—Metal
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
- G01N21/8922—Periodic flaws
Definitions
- the present invention relates to a periodic defect detection apparatus and method for detecting a periodic defect that occurs periodically in a strip or column made of metal, plastic, or other material.
- Patent Document 1 The technique described in Patent Document 1 is one of the methods using periodicity.
- the object is first measured by a defect detection sensor, and the sensor output signal is measured at the expected period (in Patent Document 1, the length corresponding to one rotation of the final rolling roll of the steel line).
- This is a technique that uses synchronous addition and emphasizes a defective signal with a period from other noise components without a period.
- this method can only be used when the period is predicted in advance. For example, in a steel production line, the roll causing the defect may wear out and change its diameter. When the diameter changes, the defect generation cycle naturally changes, so the above method is difficult to apply. So far, several methods have been proposed to deal with the problem of changing the occurrence cycle of this defect.
- the first method is to measure a subject with a defect detection sensor, perform threshold processing on the sensor output signal, extract a plurality of spider candidates, compare a plurality of sputum capture intervals, and match them. In this case, it is determined that a periodic defect has occurred with the matching interval as the period.
- this method is actually applied, there are the following problems.
- the degree of contact between the roll and the steel sheet is not necessarily uniform, for example, the rolling reduction changes with a rolling roll.
- the defect level is small, so the defect signal is weak and may not be detected.
- signals from minor defects such as sudden defects that do not have periodicity, surface roughness of the originally harmless steel sheet, and magnetic characteristics (for magnetic detectors) are detected. May occur in combination with periodic defects. For this reason, in the method of determining periodicity by simply comparing the interval between defect candidates, the interval between defect candidates does not match due to undetected defect candidates or noise such as sudden defects or overdetection, and periodic defects and There is a problem that the periodicity cannot be accurately detected.
- a detection method using autocorrelation is known as a second method for dealing with the problem of changing the occurrence cycle of defects (for example, see Patent Document 2).
- the detection method using this self-correlation also has a problem that the accuracy of determining periodic defects decreases when the sensor output signal from the subject contains many noise components. In addition, if the detection sensitivity is lowered to suppress this overdetection, there is a problem that a minor signal from a light defect cannot be detected.
- Patent Document 1 Japanese Patent Laid-Open No. 6-3 2 400 5
- Patent Document 2 Japanese Patent Laid-Open No. Sho 5 8-1 5 6 8 4 2
- Patent Document 3 Japanese Patent Laid-Open No. 2 0 06-1 0 5 7 9 1
- the object of the present invention can be used even when the generation cycle of defects changes, and is high even in the case of periodic defects generated not only in the final rolling tool but also in a plurality of rolls having different diameters, especially when the defects are minor.
- a periodic defect detection device such as a belt-like body that can be determined with accuracy and a method thereof are provided.
- Figure 18 shows an example of the correlation calculation for a typical periodic defect measurement signal.
- Reference signal and input signal When the waveforms are completely identical or similar, the period of the periodic defect is obtained from the interval between the locations where the correlation value is large, using the fact that the correlation coefficient value is large.
- the reference signal is cut out from the input signal (that is, the sensor measurement signal in the case of defect detection).
- S ZN is low, S ZN of correlation calculation is not improved.
- the inventors have focused on the feature that once a periodic defect occurs, a plurality of defects are repeatedly generated (for example, 5 times or more), and the present invention improves S / N. I came up with it. Correlation calculation is data for one period and has the advantage of being able to calculate the period, but is not an operation that uses data for multiple periods. On the other hand, in the present invention, improvement of SZN was realized by utilizing data for a plurality of cycles.
- the present inventors have also obtained knowledge that there is a preferable range in the data area (number of data) for performing the correlation calculation in order to further improve SZN. In other words, it was found that the higher the ratio of the number of defective signal data, the better the data area for correlation calculation.
- the present invention has been made on the basis of the knowledge as described above, and has the following configuration.
- the periodic defect detection device is:
- a plurality of small areas whose area length is shorter than the above-mentioned areas are separated in the arrangement direction of the periodic defects so that all the adjacent distance intervals are equal to each other.
- An evaluation index calculating means for calculating an evaluation index of similarity between signal patterns among a plurality of signals selected by the small region selecting means
- a setting value changing means for changing the position of the small area and the distance interval, and repeating the calculation process of the small area selecting means and the evaluation index calculating means;
- the distance interval is determined as a cycle.
- a period determining means
- the small region selection means is as follows. ⁇
- a small area selection means Determine the position of one small area shorter than the area, set it as the first small area, and select a signal corresponding to the position of the first small area from the sensor output.
- a plurality of second small areas are arranged at equal distance intervals in the arrangement direction of the periodic defects, and from the sensor output, the plural Second subregion selection means for selecting a signal corresponding to the position of the second subregion of
- the set value changing means changes the position of the first small area and the distance interval, and repeats the arithmetic processing of the small area selecting means and the evaluation index calculating means.
- the senor is preferably a magnetic sensor that excites a subject made of a magnetic metal member to obtain a leakage magnetic flux signal.
- the length of the small region is approximately the same as the assumed maximum defect.
- the evaluation index calculation means calculates a value for evaluating similarity in each of the small regions, and obtains the evaluation index by combining these values. .
- the evaluation index calculation means calculates a value for evaluating similarity in each of the small regions, and adds the values to obtain the evaluation index. preferable.
- a value for evaluating similarity in each of the small regions is a correlation value between the small regions.
- the periodic defect detection device according to the present invention it is preferable that a value for evaluating similarity in each of the small regions is a correlation value between the small regions.
- Periodicity determining means for obtaining defect candidates for periodic defects based on the output of the sensor; and defect determining means for determining at least the presence or absence of defects based on the defect candidates and the sensor output;
- the periodicity determining means includes
- a first two-dimensional small area smaller than the two-dimensional area is selected within the two-dimensional area, and a plurality of second dimensions separated from the first two-dimensional small area by a predetermined distance in the arrangement direction of the periodic defects.
- a two-dimensional subregion is selected with the same size as the first two-dimensional subregion, and the similarity index is calculated by calculating the similarity index between the signal patterns of the sensor output corresponding to the two-dimensional subregion.
- the distance is set as a period, and a process for determining that a defect candidate exists in each two-dimensional small region is repeated with the first calculation process while changing the distance,
- the first calculation process is repeated until the distance satisfies a period range in which periodic defects can occur, the position of the first two-dimensional small area is changed in the original two-dimensional area, and the first 1 2D small area fills the specified range in the original 2D area
- the periodic defect detection method includes:
- a plurality of small areas whose area length is shorter than that of the above-mentioned areas are determined so that the adjacent distance intervals are all equal in the arrangement direction of the periodic defects, the positions are determined, and the positions of the plurality of small areas are determined.
- the small region selection step is preferably as follows.
- a position of a small region having a length shorter than the region is determined as a first small region, and a periodic defect is determined based on the position of the first small region.
- a plurality of second small areas are arranged at equal distances from each other, and the position of the first small area and the position of the plurality of second small areas are output from the sensor output.
- the setting value changing step the position and distance interval of the first small area are changed, and (b) and (c) are repeated.
- the periodic defect detection method according to the present invention includes:
- a first two-dimensional small area smaller than the two-dimensional area is selected within the two-dimensional area, and a plurality of second dimensions separated from the first two-dimensional small area by a predetermined distance in the arrangement direction of the periodic defects.
- a two-dimensional subregion is selected with the same size as the first two-dimensional subregion, and the similarity index is calculated by calculating the similarity index between the signal patterns of the sensor output corresponding to the two-dimensional subregion.
- a sensor for obtaining a signal for evaluating the property of a region having a length longer than an expected defect period on the object, a plurality of small regions having a region length shorter than the region, and a periodic defect
- a small region selection means for determining positions by separating the adjacent distance intervals in the arrangement direction to be equal, and selecting signals corresponding to the positions of the plurality of small regions from the sensor output;
- An evaluation index calculating means for calculating an evaluation index of similarity between signal patterns among a plurality of signals selected by the small area selecting means, and changing the position of the small area and the distance interval, A setting value changing unit that repeats the calculation process of the selection unit and the evaluation index calculating unit; and a period determining unit that determines the distance interval as a cycle when the evaluation index is higher than a preset value.
- a signal input step for obtaining a sensor output for evaluating a property of a region having a length longer than a predicted defect period on the object; and (b) the region length is A plurality of small areas shorter than the recording area are separated in the arrangement direction of the periodic defects so that all adjacent distance intervals are equal, the position is determined, and the signals corresponding to the positions of the plurality of small areas are A small area selection step for selecting from the sensor output; and (c) an evaluation index calculation step for calculating an evaluation index for the similarity between the signal patterns between the plurality of signals selected by the small area selection means; (D) a setting value changing step that repeats (b) and (c) by changing the position and distance interval of the small area, and the evaluation index obtained in (e) (c) is higher than a preset value In such a case, the period determination is to determine the distance interval as a period. Therefore, even if the generation cycle of the defect fluctuates, the defect can be easily detected, and the minor periodic defect generated
- FIG. 1 is a configuration diagram of the periodic defect detection device according to the first embodiment.
- FIG. 4 is a diagram showing an example of a function block of the periodicity determination device according to the first exemplary embodiment.
- FIG. 3 is a flowchart showing the processing steps of the periodicity determination device and the defect determination device of the first embodiment.
- 4A and 4B are explanatory diagrams of the arithmetic processing by the periodicity determination device.
- FIG. 5 is a flowchart showing the processing steps of the periodicity determination device and the defect determination device of the second embodiment.
- 6A to 6C are explanatory diagrams showing the relationship between the data sampling pitch and the defect detection.
- FIG. 7 is a configuration diagram of the periodic defect detection device according to the third embodiment.
- Fig. 8 is a schematic diagram of magnetization in the width direction.
- Fig. 9 shows an example of flaw detection on a roll-type micro unevenness defect that occurred during steel plate manufacturing.
- FIG. 10 is a diagram showing the result of autocorrelation of the leakage flux signal of FIG.
- Figure 11 shows an example of the similarity evaluation index R obtained by the signal processing of Embodiment 1
- Figure 12 shows an example of similarity evaluation index R obtained by signal processing in Embodiment 1
- Figure 13 shows an example of similarity evaluation index R obtained by signal processing in Embodiment 1
- Figure 14 shows an example of the similarity evaluation index R obtained by the signal processing of Embodiment 1
- Fig. 15A and Fig. 15B show the survey results of the distribution of the size of defects generated in the continuous annealing line.
- FIG. 16A and FIG. 16B are diagrams showing examples of S ZN of similarity evaluation index R obtained by changing the window length and window width of the region for the data in FIG.
- FIG. 17 is a diagram illustrating an example of a measurement signal according to the first embodiment.
- FIG. 18 is an explanatory diagram showing a conventional method for detecting periodic defects.
- FIG. 19 is an explanatory diagram showing a conventional method for detecting periodic defects.
- FIG. 2OA to FIG. 20C are explanatory views showing the principle of the periodic defect detection method of the present invention. Explanation of symbols
- a one-dimensional data string is described for ease of understanding, but a two-dimensional data string is applicable as described later.
- Fig. 2 OA take the first small area 1 that contains the defect, and take the second small areas 2, 3, 4, 5 (here, Although there are four, the number of the second small area is not particularly limited).
- the data is taken on the data string at the same interval as (a distance interval that is the sum of the size of a small region and the distance interval between adjacent small regions). Then, if the product-sum operation is performed on each of the first subregion 1 and the second subregions 2, 3, 4, and 5, each of the four correlation values obtained has a large value. Furthermore, the value obtained by adding these four is calculated as the similarity evaluation index scale. .
- the defect position is included in the first small region as shown in FIG. If they are exactly the same, all the correlation values will be large, and the similarity evaluation index R will definitely be large.
- the first small area does not contain any defects, or the interval d is the defect period p. If they do not agree with each other, even if any of the correlation values happens to be large, all of them will not be large, and the similarity index R will not increase.
- the position q of the first small region is then changed, and the interval d is changed in the same manner as described above to determine the similarity evaluation index.
- R Repeat the calculation of R to find the period of periodic defects from the interval d when the similarity evaluation index R is larger than the preset value or when it reaches the maximum value.
- Embodiment 1 is the principle of the present invention. Details will be described in the following embodiments. Embodiment 1.
- FIG. 1 is a configuration diagram of the periodic defect detection device according to the first embodiment.
- 1 is a steel plate (shown in a perspective view for easy understanding of the equipment placed under the steel plate)
- 2 is a period 3 is a magnetizer
- 4 is a magnetic sensor
- 5 is a signal preprocessing device with a built-in amplifier and filter circuit
- 6 is an AZD conversion device
- 7 is a periodic defect detector
- 8 is a defect determination device.
- the copper plate 1 is assumed to have a plurality of roll surface defects due to the rolling roll as the periodic defect 2 in the rolling direction (the traveling direction in FIG. 1).
- a plurality of pairs of the magnetizer 3 and the magnetic sensor 4 are arranged along the width direction of the steel plate 1 (a direction perpendicular to the arrangement direction of the periodic defects) with the copper plate 1 interposed therebetween.
- the magnetizer 3 is magnetized by being supplied with a direct current from a magnetization power source (not shown), and the magnetic flux generated between the magnetic poles by the magnetizer 3 passes through the steel plate 1.
- the magnetizer 3 is installed so that the magnetic flux flows in the width direction of the steel plate 1.
- the magnetizer and magnetic sensor pair are installed facing each other with the steel plate 1 in between, but they may be installed on the same side. If the periodic defect 2 is present in the steel plate 1, the magnetic flux is hindered, and the change can be detected by the magnetic sensor 4. If the periodic defect 2 is present on the copper plate 1, the magnetic flux is blocked by this and the change can be detected by the magnetic sensor 4.
- the traveling direction shown in the figure corresponds to the direction in which the periodic defects are aligned. Therefore, the periodic defect 2 is transferred to the magnetic sensor 4 as the steel plate 1 is conveyed by the Shonobu line or the like. The signal changes as described above each time.
- the measurement signal of the magnetic sensor 4 is collected as time-series data according to the moving amount of the steel plate in the traveling direction (that is, the position of the steel plate 1), the measurement data (characteristics) of the periodic defect 2 in the alignment direction Can be obtained.
- the senor can measure the entire width direction at the same time, it can measure over the entire length in the longitudinal direction of the steel sheet, so there are no problems because it becomes more than one cycle, but in the case of periodic defect inspection, it is generally from a cost perspective
- the sensor that measures only a part of the width direction In some cases, measurement may be performed. In that case, it is only necessary to move the I ⁇ position after measuring the above set period (more than the maximum length of multiple lengths, for example, about 3 to 5 periods) at the same width position.
- the output signal of the magnetic sensor 4 thus obtained is subjected to signal amplification by an amplifier built in the signal preprocessing device 5, and noise is removed by a filter circuit built in the signal preprocessing device 5. And sent to the A / D converter 6.
- FIG. 2 is an explanatory diagram showing an example of a functional block diagram of the periodicity determining device 7.
- the periodicity determination device 7 stores AZD-converted measurement data (in this case, a plurality of magnetic sensors 4 are arranged in the width direction, so that a two-dimensional data string) is stored as raw data.
- the size and position of the small area are selected when selecting the storage area 71, the first small area selector 72 for selecting the data of the first small area, and the data of the first small area.
- the first small area data setting section 73, the second small area selection section 74 for selecting the second small area data, and the second small area data when selecting the second small area data Similarity evaluation index by inputting the data selected from the second small area data setting section 75, which sets the size and distance interval of the small area, and the first small area selection section and the second small area selection section
- the similarity evaluation index calculation unit 7 6 that calculates, and whether or not there is periodicity from the similarity evaluation index
- a determination result storage unit 78 for storing the determined result and outputting the result to the defect determination unit.
- the first small area data setting section 73 changes the position of the first small area sequentially and sets it in the first small area selection section, and the second small area data setting section 75 sets the second small area data setting section 75.
- FIG. 3 is a flowchart showing the processing steps of the periodicity determining device 7 and the defect determining device 8, and FIG. 4 is an explanatory diagram of the arithmetic processing.
- the first small region selection unit 7 2 has a region 1 (reference) of the measured range h in the width direction and 1 (el) in the rolling direction.
- Select (Area) is preferably set to the value of the minimum roll circumference (minimum value of the defect period) as the initial value of g used in step S 2 and later described below.
- the second small area selector 7 4 has the same size in the width direction as viewed from the area 1 and the same size as the area 1 at a distance d in the rolling direction. Select area 2 with. Similarly, a region 3 is selected at a position 2d away from the region 1, a region 4 is selected at a position 3d away, and a region 5 is selected at a position 4d away.
- Region 1 corresponds to the first small region
- regions 2 to 5 correspond to the second small region
- the similarity evaluation index calculator 7 6 calculates the following (Equation 1) at the corresponding locations in Region 1 and Region 2, and calculates the correlation value R 1 2 between Region 1 and Region 2.
- X (i, j) is the value of the i-th point in the width direction and the j-th point in the rolling direction within the entire measured range of the digitized sensor output.
- R 45 ⁇ x (4d) * x (i, j + 5d)
- the similarity evaluation index calculator 76 calculates the following (Equation 3) and adds the correlation values R 12, R 23, R 34, and R 45 to obtain the similarity evaluation index R.
- Equation 3 an attempt is made to eliminate the influence of variation and bias for each region where the correlation values of a plurality of regions are added, and this is a unique process for detecting periodic defects.
- the periodicity determination unit 77 determines that there is a periodic defect candidate when the similarity evaluation index R is equal to or greater than a preset threshold value.
- S7 The process of S1 to S6 is repeated by shifting the position q of region 1 serving as a reference for similarity evaluation in the rolling direction.
- the amount of shift ⁇ q at this time is preferably a value smaller than 12 of the size 1 (el) in the rolling direction of the region in order to evaluate without omission.
- the lower limit of the shift amount ⁇ q is the measured rolling direction sampling interval of the digitized sensor output. However, it takes a long time to calculate, so it can be determined appropriately. Then, evaluation can be performed without omission by repeating until the position q in the region 1 reaches the maximum value of the defect period (maximum roll circumference) (until the predetermined range of S 7 in FIG. 3 is covered).
- step S1 the initial value of d is set to the minimum roll circumference.
- the initial value of d in S1 is The value of q may be set. In other words, if the value of the position q is always set from 0 (zero) to the maximum value of the defect cycle (maximum roll circumference value), the calculation is repeated in the range where q exceeds d, so the condition that q exceeds d It is preferable that the calculation is not performed because efficient calculation is possible.
- the defect determination device 8 determines whether it is a defect from numerical values such as the signal strength of the defect candidate, the length in the width direction, the length in the rolling direction, and the shape of the defect. The result is output together with the period d obtained by the sex judging device 7.
- the defect determination device 8 may perform the above determination after improving the S ZN by synchronously adding the signals of the defective portions based on the period obtained by the periodicity determination device 7. Then, when the process of S9 is completed, the process returns to step S6.
- the above flowchart is an example of a processing procedure, and the processing procedure may be changed as appropriate.
- the iterative process for changing the interval d is in the iterative process for changing the position of the area 1, but it may be reversed, and the evaluation process for the similarity evaluation index R of S5 is also similar. Although it has been described that the process is executed every time the index R is calculated, the evaluation process of the similarity evaluation index R may be performed after all the repeated processes are completed. In addition, the change of the position in the width direction of data area 1 of S8 does not have to be performed when one-dimensional data is targeted. Yes.
- Equation 4 shows the formula for calculating the synchronous addition value y (i, j). If the value of y (i, j) exceeds the specified value, it is judged as a defect. In this way, the calculation method can be simplified, and particularly the addition between small areas can be performed, so that the amount of calculation can be greatly saved.
- the noise level N is determined online based on the data determined as having no periodic defects in S5, and y (i, j) force For example, if it exceeds 3N, it is determined as a defect. (In other words, if S ZN> 3, it is a defect). N may be determined by the maximum value of the predetermined area or the mean square error.
- ⁇ i j + x + d) + x i, j + 2d) + x i, j + 3d) + x i, j + Ad)
- a defect meter generally increases the resolution of a defect signal in order to improve the accuracy of distinguishing a defect from a harmless part.
- a determination process for determining whether or not the measurement signal includes a periodic component ie, a periodic defect error. If a periodic component is included (when a periodic defect candidate is detected), the hazard level of the defect is determined using the period and location information as the second stage.
- the defect signal is emphasized to determine the type and degree of the defect.
- step S 5 Detection of defect candidates (step S 5) by periodicity evaluation (similarity of similarity evaluation by correlation calculation: steps S 1 to S 8 above)
- step S 9 Perform defect determination (step S 9) to determine whether it is defective or harmless
- the second embodiment is the same as the processing procedure of FIG. 3 described in the first embodiment, and is a process performed in the preceding stage of S 1, and therefore, the description of the parts overlapping with the first embodiment is omitted.
- FIG. 5 is a flowchart showing the processing steps of the period determination device 7 and the defect determination device 8 according to the second embodiment.
- the signal is thinned out at a rate of once every several sampling pitches (eg, once every 4 times or once every 2 to 8 times) when the data created in S 1 3 was measured. Create periodicity evaluation data.
- step S1 to S8 is performed as in the first embodiment.
- step S15 when the periodicity evaluation data storage area is separate from the data storage area 71 to store the periodicity evaluation data and the periodicity evaluation data is stored in it. This calculation is performed using the data stored in the periodicity evaluation data storage area instead of the data area 71 data.
- the defect determination is performed by performing the process of S9 described in the first embodiment using the data in the defect harmfulness determination data storage area.
- the spatial resolution in the width direction of the measurement signal is, for example, wZ 2
- the defect signal is at least two data positions in the width direction. Is obtained. If w and 4, defect signals are obtained at four data positions in the width direction.
- defect signals can be obtained from a plurality of data positions in the width direction, and defects can be removed by decimating the rolling direction position in the width direction as shown in Fig. 6B. It is suppressed.
- the amount of shift as shown in Fig. 6C does not necessarily have to be one data position.
- Embodiment 1 in FIG. 1 In the measurement of Embodiment 1 in FIG. 1, excitation is performed in the longitudinal direction, but as shown in FIG. 7, sensors that sense a magnetic field in the vertical direction are arranged in the width direction and magnetized in the width direction. Moyo Yes. In this case, the magnetic flux generated by the defect has a positive / negative distribution in the width direction (Fig. 8).
- the similarity between two-dimensional regions is evaluated. In this way, the defect signal has a characteristic positive and negative peak due to widthwise magnetization.
- the probability of non-detection can be reduced and the accuracy of periodicity evaluation can be improved.
- periodicity can be evaluated even if thinning as described above.
- Figure 9 shows an example of flaw detection on a roll-type micro uneven defect that occurred in a steel sheet production line.
- the position of the defect is indicated by an arrow.
- the defect signal has a small difference compared to the noise signal, and automatic detection is difficult if it remains as it is.
- Figure 10 shows the result of autocorrelation calculation as an example of general signal processing applied to this signal. With the original signal, the difference between the defect signal and the noise signal is small, and even if autocorrelation is performed, the period of the defect (1670 mm) cannot be detected.
- FIGS. 11 and 12 show the results of applying the signal processing of the first embodiment.
- Figure 11 shows the size of the first small region, which is 8 mm in the width direction and 50 mm in the rolling direction. is there.
- the value of the similarity evaluation index R is larger at the position of 1 6 7 O mm (indicated by the arrow), which is the defect period, than the other periods. It can be seen that there is a periodic defect).
- Fig. 1 2 shows that the area size is the same as Fig. 6, and 4 combinations of areas are taken to calculate the correlation value.
- This is the similarity index.
- the correlation value between the four sets of regions is added to obtain the similarity evaluation index, but it can be obtained by other methods such as multiplication or addition with weights. It doesn't matter.
- the correlation value is calculated in two adjacent regions. This is important in order to minimize the effect of shifting (blurring) in the width direction when the subject is traveling. In particular, for iron and copper lines, there is a blur in the width direction called meandering when running, and this calculation method is very effective. However, in the case of a subject with a small amount of shake during running, it does not necessarily have to be two adjacent pairs. For example, a simple calculation method can be obtained by calculating the correlation between the first reference position and the second, third, fourth, and so on.
- Fig. 1 3 is for Fig. 1 1
- Fig. 1 4 is for Fig. 1 2
- the first small area is the size of 100 mm in the rolling direction and the other conditions are the same. is there.
- This measurement data Therefore, it can be seen that the force similarity evaluation value with a size of 100 mm in the rolling direction is worse, and the similarity evaluation value changes depending on the size of the region.
- the reason for this is that if the first small area is made too large compared to the defect size, the number of noise signal data increases with respect to the number of defect signal data, and the influence of the noise signal increases. Becomes worse.
- a defect having an area of 10 mm 2 or more is called a micro uneven defect, and is particularly difficult to detect.
- the inventors investigated the distribution of the size of the defects generated in the continuous annealing line, and as a result, the length in the width direction was 2 mm to 8 mm. About 3 to 5 Omm in the rolling direction was the main one.
- the force S is appropriate. Areas where correlation values are calculated for samples of various sizes (minimum size defects (length 3 mm x width 3 mm), maximum size defects (length 5 Omm x width 1 Omm))
- the similarity evaluation index R of (Equation 3) was calculated by changing the width (h) and length (1) of the window frame of the small area of 1 and the second area. Representative results are shown in FIGS. 16A and 16B.
- the vertical axis in Fig. 16A and Fig. 16B is the ratio of the evaluation index values (referred to as SZN) between the defective part and the healthy part.
- the small area is applicable to length: 1 Omm to 10 Omm, width: lmm to 30mm (S / N ⁇ 2), length: 20mm to 8 Omm, width: 2mm to 20mm (S / N ⁇ 2.5) is the preferred range, length: 25mm ⁇ 62mm, width: 7mn! -11 mm (S / N ⁇ 3) is a more preferable range.
- SZN is 2 or more, and automatic defect detection is possible. It can be seen that S ZN is 3 or more and is optimal when the width and length of the window frame of the region are approximately the same as the maximum width and length of the assumed defect.
- the measurement signal from the magnetic sensor of this example is approximately 1 in the rolling direction. Since it has a spread of about 0 mm, the difference in the preferred range does not appear as large as the relative ratio of defect sizes in Fig. 9.
- the lower limit of the preferred range may be the maximum defect size, but the upper limit of the preferred range is considered to be based on the spread of this measurement signal, so if the preferred range is labeled differently, the maximum defect size of 1 Z 4 or more, And less than or equal to 10 times the minimum spread of the signal, more preferably 2 Z 5 or more of the maximum defect size, and less than or equal to 8 times the minimum spread of the signal, and more preferably the maximum defect size. It is 12 or more in size and 6 times or less of the minimum signal spread.
- leakage magnetic flux flaw detection is shown.
- other methods may be used instead of leakage magnetic flux flaw detection, and a surface defect meter using a normal camera may be used.
- Other defect detection means such as a sensor using infrared light, a thermography, an ultrasonic sensor, and an eddy current sensor may be used.
- the technique of the above-mentioned Patent Document 3 uses optical defect inspection and utilizes the fact that the ground pattern, which is noise in the sound part of the obtained signal, has the same periodicity as the circumference of the final rolling roll. ing.
- the ground pattern which is noise in the sound part of the obtained signal
- the technique of the above-mentioned Patent Document 3 uses optical defect inspection and utilizes the fact that the ground pattern, which is noise in the sound part of the obtained signal, has the same periodicity as the circumference of the final rolling roll. ing.
- sound in the sound part is affected not only by the surface irregularities, but also by distortions caused by subtle conditions such as rolling reduction and temperature unevenness during rolling. Noise does not always have the same period as the final mill roll. For this reason, especially in the case of leakage magnetic flux inspection, the effect of evaluating the similarity for each area around the defective part is significant.
- magnetic sensors such as Hall elements, coils, magnetoresistive elements, and S Q U I D can be used as the magnetic flux leakage sensors.
- a plurality of sensors are arranged in the width direction, but a method of traversing one or a plurality of sensors may be used.
- a method for flaw detection in a partial area in the width direction may be used.
- a method may be used in which a part of the width direction is flawed for a certain length, and the flaw is detected repeatedly by changing the position in the width direction.
- the magnetizer 3 is installed so that the magnetic flux flows in the width direction of the steel plate.
- the magnetizer and magnetic sensor pair are installed opposite to each other with the steel plate 1 sandwiched therebetween, but they may be installed on the same side.
Landscapes
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Immunology (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Physics & Mathematics (AREA)
- Electrochemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Engineering & Computer Science (AREA)
- Textile Engineering (AREA)
- Investigating Or Analyzing Materials By The Use Of Magnetic Means (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020107021180A KR101288734B1 (ko) | 2008-03-31 | 2009-03-27 | 주기성 결함 검출 장치 및 그 방법 |
EP09726665.4A EP2264446B1 (en) | 2008-03-31 | 2009-03-27 | Periodic defect detecting device and method for the same |
US12/935,307 US9008975B2 (en) | 2008-03-31 | 2009-03-27 | Apparatus for detecting periodic defect and method therefor |
CN2009801121710A CN102007400B (zh) | 2008-03-31 | 2009-03-27 | 周期性缺陷检测装置及其方法 |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008-089364 | 2008-03-31 | ||
JP2008089364 | 2008-03-31 | ||
JP2009-061501 | 2009-03-13 | ||
JP2009061501A JP5453861B2 (ja) | 2008-03-31 | 2009-03-13 | 周期性欠陥検出装置及びその方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2009123296A1 true WO2009123296A1 (ja) | 2009-10-08 |
Family
ID=41135663
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/056914 WO2009123296A1 (ja) | 2008-03-31 | 2009-03-27 | 周期性欠陥検出装置及びその方法 |
Country Status (7)
Country | Link |
---|---|
US (1) | US9008975B2 (ja) |
EP (1) | EP2264446B1 (ja) |
JP (1) | JP5453861B2 (ja) |
KR (1) | KR101288734B1 (ja) |
CN (1) | CN102007400B (ja) |
TW (1) | TWI392865B (ja) |
WO (1) | WO2009123296A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104751288A (zh) * | 2015-03-30 | 2015-07-01 | 北京首钢自动化信息技术有限公司 | 一种钢卷分段多维在线质量判定系统及其方法 |
CN111314801A (zh) * | 2020-02-13 | 2020-06-19 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | 一种支持动态调度的数据采集系统及方法 |
Families Citing this family (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102011083405A1 (de) * | 2010-12-21 | 2012-06-21 | Sms Siemag Ag | Verfahren und Vorrichtung zur Oberflächeninspektion von Bandstücken |
JP5614312B2 (ja) * | 2011-02-01 | 2014-10-29 | Jfeスチール株式会社 | 周期性欠陥検出方法および周期性欠陥検出装置 |
AT511773A1 (de) * | 2011-07-15 | 2013-02-15 | Voestalpine Stahl Gmbh | Vorrichtung und verfahren zur detektion wenigstens eines periodisch auftretenden fehlers an einem gegenstand |
KR101309966B1 (ko) * | 2011-12-15 | 2013-09-17 | 주식회사 포스코 | 강판의 결함 탐상 장치 |
KR101482347B1 (ko) | 2012-12-27 | 2015-01-13 | 주식회사 포스코 | 강판의 내부 결함 검출 장치 및 방법 |
CN103630547B (zh) * | 2013-11-26 | 2016-02-03 | 明基材料有限公司 | 具有周期性结构的光学薄膜的瑕疵检测方法及其检测装置 |
JP6287379B2 (ja) * | 2014-03-11 | 2018-03-07 | 株式会社リコー | 画像検査装置、画像検査システム及び画像検査方法 |
US9966883B2 (en) * | 2014-08-08 | 2018-05-08 | Johnson Electric S.A. | Magnetic sensor and an integrated circuit |
US9954469B2 (en) * | 2014-08-08 | 2018-04-24 | Johnson Electric S.A. | Magnetic sensor and an integrated circuit |
US9716453B2 (en) * | 2014-08-08 | 2017-07-25 | Johnson Electric S.A. | Magnetic sensor and an integrated circuit |
US20160344320A1 (en) * | 2014-08-08 | 2016-11-24 | Johnson Electric S.A. | Magnetic sensor integrated circuit, motor component and application apparatus |
US9973121B2 (en) * | 2014-08-08 | 2018-05-15 | Johnson Electric S.A. | Magnetic sensor and an integrated circuit |
JP6358002B2 (ja) * | 2014-09-16 | 2018-07-18 | 旭硝子株式会社 | 不具合搬送用ロールの特定方法、およびガラスリボンにおける疵発生防止方法 |
CN105572143B (zh) * | 2015-12-17 | 2018-05-25 | 湖北第二师范学院 | 压延过程中压延材料表面周期性缺陷的检测方法 |
CN106959336B (zh) * | 2016-01-11 | 2020-07-28 | 宝山钢铁股份有限公司 | 用于漏磁检测标定的背景噪声去除装置及方法 |
JP6587988B2 (ja) * | 2016-07-07 | 2019-10-09 | 株式会社日立ビルシステム | ハンドレールの検査システムおよびハンドレールの検査方法 |
CN106053559A (zh) * | 2016-07-30 | 2016-10-26 | 董超超 | 用于金属材料表面缺陷的检测系统 |
CN106468677B (zh) * | 2016-09-06 | 2018-11-09 | 电子科技大学 | 一种脉冲涡流热成像中缺陷的自动分割方法 |
CN106918639B (zh) * | 2017-04-18 | 2019-07-02 | 清华大学 | 沿磁化方向的单元组合缺陷漏磁信号计算方法 |
CN107475509B (zh) * | 2017-08-04 | 2019-03-19 | 北京首钢股份有限公司 | 一种带钢的周期性缺陷的分析方法 |
CN113260854A (zh) | 2018-12-25 | 2021-08-13 | 杰富意钢铁株式会社 | 学习完成模型的生成方法、学习完成模型、表面缺陷检测方法、钢材的制造方法、合格与否判定方法、等级判定方法、表面缺陷判定程序、合格与否判定程序、判定系统及钢材的制造设备 |
JP7107439B2 (ja) * | 2019-06-28 | 2022-07-27 | Jfeスチール株式会社 | 丸棒の表面疵検査方法及び表面疵検査装置 |
CN110779928B (zh) * | 2019-11-19 | 2022-07-26 | 汪科道 | 缺陷检测装置及方法 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS58156842A (ja) | 1982-03-15 | 1983-09-17 | Toshiba Corp | ロ−ル疵検出装置 |
JPH06324005A (ja) | 1993-05-13 | 1994-11-25 | Nippon Steel Corp | 鋼板の圧延ロール疵検出方法 |
JPH07198627A (ja) * | 1994-01-06 | 1995-08-01 | Nippon Steel Corp | 金属表面欠陥検査装置 |
JPH08160006A (ja) * | 1994-11-30 | 1996-06-21 | Kawasaki Steel Corp | 鋼板の欠陥検出方法 |
JP2006105791A (ja) | 2004-10-05 | 2006-04-20 | Nippon Steel Corp | 帯状体や柱状体の周期性疵検出方法およびその装置 |
JP2006153614A (ja) * | 2004-11-29 | 2006-06-15 | Nippon Steel Corp | 帯状体や柱状体の周期疵検出方法およびその装置 |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CH612152A5 (ja) * | 1976-01-26 | 1979-07-13 | Rieter Ag Maschf | |
US4463425A (en) * | 1980-07-17 | 1984-07-31 | Terumo Corporation | Period measurement system |
JPH0786474B2 (ja) * | 1988-09-09 | 1995-09-20 | 富士写真フイルム株式会社 | 欠陥周期の測定方法 |
EP0927887A1 (de) * | 1997-12-17 | 1999-07-07 | Zellweger Luwa Ag | Verfahren zur Erkennung periodischer Fehler in einem längsbewegten Prüfgut |
US6266983B1 (en) * | 1998-12-09 | 2001-07-31 | Kawasaki Steel Corporation | Method and apparatus for detecting flaws in strip, method of manufacturing cold-rolled steel sheet and pickling equipment for hot-rolled steel strip |
JP2004071045A (ja) * | 2002-08-06 | 2004-03-04 | Sony Corp | デフェクト検出装置、デフェクト検出方法 |
-
2009
- 2009-03-13 JP JP2009061501A patent/JP5453861B2/ja active Active
- 2009-03-27 EP EP09726665.4A patent/EP2264446B1/en active Active
- 2009-03-27 US US12/935,307 patent/US9008975B2/en not_active Expired - Fee Related
- 2009-03-27 WO PCT/JP2009/056914 patent/WO2009123296A1/ja active Application Filing
- 2009-03-27 KR KR1020107021180A patent/KR101288734B1/ko active IP Right Grant
- 2009-03-27 CN CN2009801121710A patent/CN102007400B/zh not_active Expired - Fee Related
- 2009-03-31 TW TW098110592A patent/TWI392865B/zh not_active IP Right Cessation
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS58156842A (ja) | 1982-03-15 | 1983-09-17 | Toshiba Corp | ロ−ル疵検出装置 |
JPH06324005A (ja) | 1993-05-13 | 1994-11-25 | Nippon Steel Corp | 鋼板の圧延ロール疵検出方法 |
JPH07198627A (ja) * | 1994-01-06 | 1995-08-01 | Nippon Steel Corp | 金属表面欠陥検査装置 |
JPH08160006A (ja) * | 1994-11-30 | 1996-06-21 | Kawasaki Steel Corp | 鋼板の欠陥検出方法 |
JP2006105791A (ja) | 2004-10-05 | 2006-04-20 | Nippon Steel Corp | 帯状体や柱状体の周期性疵検出方法およびその装置 |
JP2006153614A (ja) * | 2004-11-29 | 2006-06-15 | Nippon Steel Corp | 帯状体や柱状体の周期疵検出方法およびその装置 |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104751288A (zh) * | 2015-03-30 | 2015-07-01 | 北京首钢自动化信息技术有限公司 | 一种钢卷分段多维在线质量判定系统及其方法 |
CN111314801A (zh) * | 2020-02-13 | 2020-06-19 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | 一种支持动态调度的数据采集系统及方法 |
CN111314801B (zh) * | 2020-02-13 | 2022-01-28 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | 一种支持动态调度的数据采集系统及方法 |
Also Published As
Publication number | Publication date |
---|---|
EP2264446B1 (en) | 2021-01-20 |
TW200951426A (en) | 2009-12-16 |
US9008975B2 (en) | 2015-04-14 |
CN102007400A (zh) | 2011-04-06 |
JP5453861B2 (ja) | 2014-03-26 |
TWI392865B (zh) | 2013-04-11 |
EP2264446A1 (en) | 2010-12-22 |
JP2009265087A (ja) | 2009-11-12 |
US20110040499A1 (en) | 2011-02-17 |
EP2264446A4 (en) | 2017-05-03 |
KR101288734B1 (ko) | 2013-07-23 |
KR20100116683A (ko) | 2010-11-01 |
CN102007400B (zh) | 2012-09-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2009123296A1 (ja) | 周期性欠陥検出装置及びその方法 | |
RU2549913C2 (ru) | Термографический способ контроля и контрольная установка для осуществления способа | |
RU2342653C2 (ru) | Способ неразрушающего испытания труб и устройство для его осуществления | |
JP6658711B2 (ja) | 表面欠陥検出方法及び表面欠陥検出装置 | |
JP2009265087A5 (ja) | ||
EP3344982B1 (en) | A method and system for detecting a material discontinuity in a magnetisable article | |
JP4622742B2 (ja) | 金属帯の渦流探傷方法および装置 | |
JP2012159437A (ja) | 周期性欠陥検出方法および周期性欠陥検出装置 | |
EP0021893B1 (fr) | Procédé et dispositif d'inspection de produits métalliques par courants de Foucault et application du procédé et du dispositif | |
JP5347661B2 (ja) | 帯状体の表面検査装置、表面検査方法及びプログラム | |
JP6950664B2 (ja) | 欠陥判定方法、欠陥判定装置、鋼板の製造方法、欠陥判定モデルの学習方法、及び欠陥判定モデル | |
JP2004037216A (ja) | 漏洩磁束探傷方法 | |
JP4835212B2 (ja) | 渦流探傷方法及び渦流探傷装置 | |
US10788456B2 (en) | Eddy current inspection device for nondestructive testing | |
JP2004354240A (ja) | 漏洩磁束探傷法および漏洩磁束探傷装置 | |
JP5013363B2 (ja) | 非破壊検査装置 | |
CN113671018A (zh) | 一种用于抑制钢轨漏磁检测提离干扰的滤波方法 | |
KR20100020671A (ko) | 비젼 카메라를 이용한 강재의 두께측정 및형상판별장치와,이를 이용한 강재의 두께측정방법 및형상판별방법 | |
JP3145296B2 (ja) | 欠陥検出方法 | |
JP2005157960A (ja) | 丸鋼材の本数カウント方法及び本数カウント装置 | |
KR102559191B1 (ko) | 마스크 검사 방법 및 장치 | |
KR101111261B1 (ko) | 강선의 결함 검사 장치 | |
RU2229707C1 (ru) | Способ магнитного контроля трубопроводов | |
JPH11271278A (ja) | 鋼材の欠陥検出方法 | |
JPH0481653A (ja) | 渦流探傷方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200980112171.0 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09726665 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009726665 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 20107021180 Country of ref document: KR Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12935307 Country of ref document: US |