JP2605516B2 - Magnetic strip flaw detection method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for magnetically detecting defects or welds on the surface and inside of a non-magnetic metal strip such as steel or alloy which runs continuously.

[0002]

2. Description of the Related Art As a known document of a method for detecting a surface and an internal flaw of a steel sheet such as a thin steel strip in a running state, for example, a "method for detecting a flaw on a steel sheet" disclosed in Japanese Patent Application No. 1-272183. (Hereinafter referred to as Document 1). 10 to 12 are drawings according to the above-mentioned Document 1, FIG. 10 is a block diagram of a magnetic steel flaw detector for a steel plate, FIG. 11 is a schematic cross-sectional view of a flaw detection unit, and FIG. It is explanatory drawing with 2 threshold values. 10 and 11, 51 is a steel plate, 52 is a hollow roll, 53 is a coil, 54 is a magnetic core, 55 is a sensor, 56 and 59 are comparison circuits, 57 is a peak value detection circuit, and 58 is an average value detection circuit. It is.

In the magnetic flaw detector having the structure shown in FIG. 10, a flaw detecting portion housed in a hollow coil 52 shown in FIG. 11 is installed below a continuously running steel plate. The traveling steel sheet 51 is magnetized by exciting the magnetic core 54 by the coil 53 in the hollow coil 52, and if a flaw exists on the surface or inside of the magnetized steel sheet, a leakage magnetic flux is generated at the flaw portion. This leakage magnetic flux is applied to the magnetized core 5
It is detected by a sensor 55 provided between the four. The detection signal obtained from the sensor 55 is compared with a preset first threshold value (shown in FIG. 12) in a comparator 56, and a detection signal exceeding the first threshold value detects a main flaw of the steel sheet. I do.

A detection signal obtained from the sensor 55 is input to a peak value detection circuit 57, and among the detection signals from the sensor 55, the peak value of the detection signal that does not reach the first threshold value is determined in the moving direction of the steel sheet. A predetermined peak interval is detected along the distance, and a desired peak value is selected as a representative of the plurality of peak values among the plurality of peak values detected in this manner. It is repeated a plurality of times along the moving direction of the steel plate. Then, the average value of the representative values selected a plurality of times is calculated by the average value calculation circuit 58, and the signal level of the fine flaw discrimination signal obtained by using the average value thus calculated as the second threshold is shown in FIG. Is shown in Then, the second threshold value is supplied to the comparison circuit 59,
A detection signal exceeding the second threshold is detected as a minute flaw signal as compared with a detection signal not reaching the first threshold. Naturally, a detection signal equal to or smaller than the second threshold value is processed as a noise signal.

As a well-known document of a noise removing device for a magnetic flaw detector for a steel plate, for example, a "noise removing device for a thin steel strip magnetic flaw detector" disclosed in Japanese Utility Model Laid-Open No. 33336/1991 (hereinafter referred to as Reference 2) There is. 13 to 17
FIG. 13 is a partially cutaway cross-sectional view of the magnetic flaw detection portion, FIG. 14 is an electric circuit diagram of FIG. 13, FIG. 15 is an output waveform diagram of FIG. 14, and FIG. FIG. 17 is a waveform diagram illustrating a detection signal of Document 2 with respect to the element spacing and the output. 13 and 14, 61 is an exciting coil, 62 is an electromagnet, X is a thin steel strip, Y
Is a flaw, and A and B are magnetically sensitive elements.

In the device shown in FIGS. 13 and 14,
In a flaw detector that generates a magnetic field in the traveling direction of the traveling thin steel strip and detects a leakage magnetic flux generated by internal or surface flaws by a group of magnetically sensitive elements and generates a flaw signal, the flaw detection device detects a magnetic flux between electromagnets 62 shown in FIG. The magnetic sensitive element group is divided into two rows A and B in the running direction, and the interval is 0.5 to
When the arrangement is made so as to be 2.0 mm and the electric circuit of FIG. 14 is used to obtain an added value obtained by adding detection values (instantaneous detection values shown in FIG. 15) obtained from the two magnetically sensitive elements, As compared with the conventional detection signal shown in FIG. 17A, the flaw signal is doubled as compared with the conventional detection signal shown in FIG. 17B, and the noise signal is reduced as compared with the conventional detection signal. Can be In this case, as shown in FIG. 16, the distance between the A and B element groups is 0.5 to 2.0 mm (the peak is about 1.0 m).
m) discloses that internal or surface flaws can be detected with high sensitivity.

[0007]

However, the method for detecting flaws on a steel sheet according to the above-mentioned document 1 has a problem that the second threshold cannot be flexibly changed according to the type of flaw on the steel sheet. That is, the flaws on the steel sheet include point flaws (having a large or small hole diameter), linear flaws (there is a running direction and a direction perpendicular thereto), planar flaws, and welds (welding before cold rolling). And welding after cold rolling), and it was not easy to set the second threshold to an appropriate value corresponding to each of these flaws.

The noise removing device in the thin steel strip magnetic flaw detector according to Document 2 is not always effective for various types of flaws on the steel sheet, and an added value obtained by adding instantaneous detection values obtained from two magnetically sensitive elements is used. When taken out, there is a problem that the flaw signal is reduced and the signal-to-noise ratio (hereinafter referred to as S / N) may be deteriorated. This is because Document 2 adds instantaneous detection values of the same polarity obtained at the same timing from two magnetically sensitive elements and outputs the sum.
If a peak signal of the same polarity cannot be obtained from the two magnetically sensitive elements at the timing of addition, the S / N is reduced instead.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem, and a metal capable of obtaining a flaw detection signal with an improved S / N corresponding to various flaws on the inside or surface of a metal strip. An object of the present invention is to provide a method and apparatus for magnetic flaw detection of a belt.

[0010]

SUMMARY OF THE INVENTION A magnetic flaw detection method and apparatus for a metal strip according to the present invention generates a magnetic field in the running direction of a running metal strip and generates a magnetic flux leaking due to a defect inside or on the surface of the metal strip. In the metal flaw detection method and apparatus for detecting a defect signal by detecting a plurality of magnetically sensitive elements arranged in a row in a direction perpendicular to the running direction of the metal strip, a method for detecting a defect is provided. Selecting means for selecting the two magnetically sensitive elements so as to form a pair, and two detection signals respectively obtained for each pair of the magnetically sensitive elements selected by the selecting means, for each fixed traveling section of the metal strip. Peak hold means for detecting a maximum peak value, respectively, and a difference between two maximum peak values detected by the peak hold means for each pair of the magnetically sensitive elements. The difference calculating means for outputting this as a flaw signal value for each running section of each pair of magnetically sensitive elements, and the flaw signal value for each running section of each pair of magnetically sensitive elements output by the difference calculating means. Compared to a preset first threshold,
A first comparing means for detecting a main flaw of the metal strip when the first threshold value is exceeded, and a flaw signal value for each traveling section of each pair of magnetically sensitive elements output by the difference calculating means; After multiplying by a coefficient of, a predetermined bias value is added to obtain a calculated value of the corresponding section, and a moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated. A second threshold value calculating means for outputting the moving average value as a second threshold value of each in a next traveling section; and a flaw signal for each traveling section of each pair of magnetically sensitive elements outputted by the difference computing means. A second comparing means for comparing with a second threshold value calculated and output by the threshold value calculating means in the immediately preceding traveling section and detecting the second threshold value as a minute flaw of the metal band when the second threshold value is exceeded; It is provided with.

The method and apparatus for magnetically flaw detecting a metal strip according to the present invention generate a magnetic field in the running direction of the running metal strip, and reduce the leakage magnetic flux generated by a defect inside or on the surface of the metal strip. A magnetic band flaw detection method and apparatus for detecting a defect signal by detecting a plurality of magnetically sensitive elements arranged in two rows in a direction perpendicular to the direction of the metal strip.
Each of the magnetically sensitive elements in the front row of the plurality of magnetically sensitive elements arranged in a row and each of the magnetically sensitive elements in the rear row located immediately behind the magnetically sensitive elements are selected as a pair, or each magnetically sensitive element in the front row is selected. Selecting means for selecting the sensing element and each of the magnetic sensing elements in the rear row located at positions other than the position immediately behind the pair, and two detections respectively obtained for each pair of the magnetic sensing elements selected by the selecting means Peak holding means for respectively detecting the maximum peak value of the signal for each fixed traveling section of the metal belt, and two peaks detected by the peak holding means for each pair of the magnetically sensitive elements.
Difference calculating means for calculating the difference between the two maximum peak values, and outputting this as a flaw signal value for each traveling section of each pair of magnetically responsive elements; and First comparing means for comparing a flaw signal value for each traveling section with a first threshold value set in advance and detecting, when the value exceeds the first threshold value, a main flaw of the metal strip; After multiplying the flaw signal value of each pair of magnetically responsive elements output by the means for each traveling section by a predetermined coefficient, a predetermined bias value is added to obtain a calculated value of the corresponding section, and the calculated value and the past A second threshold value calculating means for calculating a moving average value with a plurality of section calculated values stored over the section and outputting the moving average value as a respective second threshold value in a next traveling section; Output of each pair of magnetically sensitive elements in each running section Is compared with its own second threshold value calculated and output by the second threshold value calculating means in the immediately preceding traveling section, and when it exceeds the second threshold value, this is detected as a minute flaw in the metal strip. 2 comparing means.

[0012]

In the present invention, a magnetic field is generated in the running direction of the running metal strip, and the leakage magnetic flux generated by a defect inside or on the surface of the metal strip is arranged in a line in a direction perpendicular to the running direction of the metal strip. In the method and the apparatus for detecting a metal band by detecting a defect signal by using a plurality of magnetically sensitive elements provided, the selecting means is configured such that two magnetically sensitive elements at a predetermined interval from the plurality of magnetically sensitive elements are paired. select. The predetermined interval refers to a case where a defect having a maximum dimension determined as a detection target is provided in a metal band, and one magnetic sensing element for detecting the defect is moved in a width direction of the metal band. If the position where the maximum value of the flaw signal is obtained from the sensitive element is the first position and the position where no flaw signal is obtained thereafter is the second position, the position between the first position and the second position is determined. Means the interval. Therefore, it is not the distance between the magnetically sensitive elements arranged in the one row. In the case of a steel strip, the predetermined interval is usually about 10 to 20 times the lateral width of the defect having the largest dimension. The peak hold means
The maximum peak value of each of the two detection signals obtained for each pair of the magnetically sensitive elements selected by the selection means is detected for each fixed traveling section of the metal band. The difference calculating means calculates the difference between the two maximum peak values detected by the peak hold means for each pair of the magnetically sensitive elements, and calculates the difference between the two maximum peak values for each pair of running sections of the magnetically sensitive elements. Output as a value. The first comparing means compares the flaw signal value of each pair of magnetically sensitive elements output from the difference calculating means for each traveling section with a first threshold value set in advance, and
If the threshold value is exceeded, this is detected as a major flaw in the metal strip. The second threshold value calculating means multiplies a flaw signal value of each pair of magnetically responsive elements output by the difference calculating means for each running section by a predetermined coefficient, and then adds a predetermined bias value to calculate the corresponding bias value. A moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated, and the moving average value is calculated for each of the second moving sections in the next traveling section.
Output as a threshold. The second comparing means calculates a flaw signal for each traveling section of each pair of magnetically responsive elements output by the difference computing means in the immediately preceding traveling section and outputs the flaw signal. If the value exceeds the second threshold as compared with the second threshold, this is detected as a minute flaw in the metal strip.

Further, in the present invention, a magnetic field is generated in the traveling direction of the traveling metal band, and the leakage magnetic flux generated by a defect inside or on the surface of the metal band is arranged in two rows in a direction perpendicular to the traveling direction of the metal band. In the method and the apparatus for detecting a metal strip to obtain a defect signal by detecting a plurality of magnetically sensitive elements arranged in a plurality of magnetically sensitive elements, the selecting means includes a magnetic sensitive element in a front row of the plurality of magnetically sensitive elements arranged in the two rows. Either select the element and each magnetically sensitive element in the rear row immediately behind it in a pair, or pair each magnetically sensitive element in the front row with each magnetically sensitive element in the rear row other than immediately behind. And so on. The peak hold means detects the maximum peak value of each of the two detection signals obtained for each pair of the magnetically sensitive elements selected by the selection means, for each fixed traveling section of the metal band. The difference calculating means calculates the difference between the two maximum peak values detected by the peak hold means for each pair of the magnetically sensitive elements, and calculates the difference between the two maximum peak values for each pair of running sections of the magnetically sensitive elements. Output as a value. The first comparing means compares the flaw signal value of each pair of magnetically responsive elements output by the difference calculating means for each traveling section with a first threshold value set in advance, and when the signal value exceeds the first threshold value, Is detected as the main flaw of the metal strip. The second threshold value calculating means multiplies a flaw signal value of each pair of magnetically responsive elements output by the difference calculating means for each running section by a predetermined coefficient, and then adds a predetermined bias value to calculate the corresponding bias value. As a value, a moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated, and the moving average value is output as its own second threshold value in the next traveling section. The second comparing means calculates a flaw signal for each traveling section of each pair of magnetically responsive elements output by the difference computing means in the immediately preceding traveling section and outputs the flaw signal. If the value exceeds the second threshold as compared with the second threshold, this is detected as a minute flaw in the metal strip.

[0014]

1 is a system diagram of a flaw detection signal processing section for a metal strip according to the present invention. In the figure, 1-1 and 1-2 are a pair of magnetically sensitive elements (hereinafter, referred to as magnetic sensors) at relatively close positions selected from a magnetically sensitive element group (hereinafter, referred to as a magnetic sensor group) included in the magnetic flaw detection unit. ), Which are referred to as # 1 and # 2 magnetic sensors. How to select the pair of magnetic sensors will be described with reference to FIGS. The magnetic flaw detection unit will be described with reference to FIGS. 2A and 2B are waveform diagrams for explaining the operation of FIG. 1. FIG. 2A shows a section update signal, FIG. 2B shows a # 1 sensor output and its peak hold value, and FIG. 2C shows a # 2 sensor output. And (d) show the calculated value of the difference between the peak hold values of # 1 and # 2.

Reference numeral 2 in FIG. 1 denotes a section signal generator, which receives a metal band transport speed signal, generates and outputs a section update signal every time the metal band is transported by a predetermined section length. In the present invention, the # 1 and # 2 magnetic sensors 1-1 and 1-
In order to detect the output signal of No. 2 not as the instantaneous value but as the maximum peak value in the predetermined section length, the length of the predetermined section is set so as not to be too large (for example, about 10 cm in the case of a thin steel strip). Is done. Now, assuming that the transport speed in the case of a thin steel strip as a metal strip is 200 m / min and the section length is 10 cm, a section update signal may be generated at intervals of about 33.3 msec (milliseconds). The generated section update signal is supplied to two peak hold circuits 3-1 and 3
-2, which is supplied to the difference operation circuit 4 and the second threshold value operation circuit 5. FIG. 2A shows the section update signal and sections a, b, c, d, and e between the signals.

3-1 and 3-2 in FIG. 1 correspond to # 1
And # 2 peak hold circuit, and the section signal generator 2
Each time a section update signal is supplied from the controller, the peak hold value, which is the maximum peak value in the previous section, is calculated by the difference calculation circuit 4.
, The value is reset, and in this section, the # 1 and # 2 magnetic sensors 1-1 and 1 are respectively provided.
After unifying the polarities of the sensor output signals sequentially input from -2, the maximum peak value is obtained and held. In response to the input of the next section update signal, the peak hold value held so far is supplied to the difference calculation circuit 4, and the operation of the next peak hold is repeated again. Here, the unification of the polarity of the sensor output signal means that the output signal of both the positive polarity and the negative polarity is originally obtained from the magnetic sensor. Therefore, for example, only the negative signal is taken out, the amplification degree is 1, and only the polarity is inverted. This means that the signal is inverted to a positive polarity signal via an inverting amplifier to make all the sensor output signals uniform to the positive polarity signal (see FIG. 8). As a result, the peak hold circuit may obtain the maximum voltage value (here, the maximum peak value) of one polarity (for example, positive polarity).

The # 1 and # 2 peak hold circuits 3-
According to 1 and 3-2, as the peak hold value obtained for each section, the maximum value of only the noise signal generated if there is no flaw to be detected in the metal band in the section,
If there is any flaw in the metal band, the maximum value of the superimposed signal of the noise signal and the flaw signal is output. FIGS. 2B and 2C show the output value of the # 1 magnetic sensor and the peak hold value V1a, V1a, V1a, V2 for each of the sections a, b, c, and d determined by the section update signal.
V1b, V1c and V1d, and the output values of the magnetic sensor of # 2 and the peak hold values V2a, V2b,
V2c and V2d are shown, respectively. Each of the peak hold values in this example indicates the maximum value of only the generated noise signal without any flaw to be detected in the metal band in each of the sections a to d.

Reference numeral 4 in FIG. 1 denotes a difference operation circuit, # 1 and #
This is a circuit for calculating the difference between the peak hold values of the previous section output from the two peak hold circuits 3-1 and 3-2 each time each section is updated. As described above, the peak hold value for each section includes a case where only the noise signal component is provided and a case where the noise signal component and the flaw signal component are superimposed. Therefore, the output signal of the difference calculation circuit 4 is a case where there is no flaw in any of the corresponding sections of the area handled by the two magnetic sensors and the peak hold values of the two sensor outputs are only the maximum peak value of the noise signal. Simply obtains the difference between the two noise signal maximum values (that is, the noise level difference signal). Since both of the two peak hold values shown in the examples of FIGS. 2B and 2C are only the maximum value of the noise signal, FIG. The difference between the two obtained maximum values of the noise signal (that is, the noise level difference signal) is calculated, and is shown as the difference operation value in the current section.

Further, there is a flaw in a certain section of the area in which one of the magnetic sensors # 1 and # 2 is in charge, and the peak hold value of the sensor output includes the maximum value of the superposition signal of the flaw signal and the noise signal. When only the maximum value of the noise signal is obtained as the peak hold value from the other sensor, when the difference between these two signals is calculated, the noise signal is almost canceled and only the maximum value of the flaw signal is obtained. Will be done. Therefore, in this case, the difference calculation circuit 4 outputs the maximum value of the flaw signal of the immediately preceding section (this is referred to as a flaw signal of the present invention) every time the traveling section is updated. As described above, the difference calculation circuit 4 in the present invention sequentially outputs the noise level difference signal or the flaw signal in the immediately preceding section every time the traveling section of the metal belt is updated.

In order to obtain the flaw signal described above, in the present invention, one of the output signals of the two magnetic sensors contains a flaw signal component but the other does not contain a flaw signal component. A pair of magnetic sensors are selected and used so that the output signal of the magnetic sensor contains a noise signal component of substantially the same level. The method of selecting the pair of magnetic sensors will be described in detail with reference to FIGS.
The outline is that the mutual distance is not too short and too far from the magnetic sensor group arranged in one or two rows in a direction perpendicular to the running direction according to the size and shape of the flaw of the metal strip. Two nearby magnetic sensors are selected. The difference calculation circuit 4 outputs the flaw signal for each section in the case where there is a flaw and the noise level difference signal in the case where there is no flaw as the # 1 and # 2 comparison circuits 6-1 and 6-2 and the second signal. Each of them is supplied to the threshold value calculation circuit 5.

Reference numeral 5 in FIG. 1 denotes a second threshold value calculation circuit for detecting a minute flaw based on a noise level difference signal provided from the difference calculation circuit 4 when there is no flaw in the metal strip having a high probability of occurrence. This is a circuit for calculating two thresholds. However, although the occurrence probability is small, a flaw signal in the case of a flaw may also be input, so that the influence of this flaw signal is removed as much as possible by a moving average process described later. Now, the noise level difference signal ΔV as shown in FIG.
i (ΔVi = V1i to V2i, i = a, b, c, d,
..) Are input, in order to calculate the second threshold, first, a coefficient α having a value of about 1 (for example, 0.7
<Α <1.5), a fixed bias voltage β is added thereto, and the operation value Vi (Vi = α · ΔVi +) in that section is added.
β) is calculated. As a specific example, when the input signal is 0.5V, if α = 1 and β = 0.5V, the operation value Vi in that section is 1V.

Next, in order to average the fluctuation of the calculated value Vi for each section, the calculated value Vi of the current section and a plurality of stored values over a plurality of past sections (that is, a predetermined length of the metal strip). , And a moving average with the calculated section values Vi-n,.
The threshold value is supplied to the # 2 comparison circuit 6-2. This moving average has a small occurrence probability as an input signal, but when a flaw signal is input or when the noise level changes suddenly, these effects are removed as much as possible so that the second threshold value does not change rapidly. It is done to make it. In this embodiment, the coefficient α and the bias voltage β are set so that the second threshold value is approximately in the vicinity of 1 V. However, the second threshold value always fluctuates according to the noise level detected by the magnetic sensor. . Then, a flaw signal exceeding the second threshold value is detected as a minute flaw.

Reference numeral 6-1 denotes a # 1 comparison circuit which compares a flaw signal input from the difference calculation circuit 4 with a first threshold (for example, 1.5 V) fixed to a level higher than the second threshold. , And those that exceed the first threshold are detected as major flaws. Reference numeral 6-2 denotes a # 2 comparison circuit which compares a similarly input flaw signal with a second threshold value supplied from the second threshold value calculation circuit 5 and detects a signal exceeding the second threshold value as a minute flaw. I do.

FIG. 3 is a schematic cross-sectional view of the magnetic flaw detection section cut along a plane parallel to the running direction of the metal strip, and FIG. 4 is a schematic cross-sectional view of the magnetic flaw detection section cut along a plane perpendicular to the running direction of the metal strip. It is. In the figure, 14a and 14b are metal bands 12
And a pair of hollow rolls respectively provided on the upper side and the lower side so as to sandwich the roll. Each hollow roll 14a, 14
b is formed of a non-magnetic material. Although the outer diameters are different in size, the thickness t1 of the upper hollow roll 14a is set smaller than the thickness t2 of the lower hollow roll 14b. This is because a downward force is applied to the lower hollow roll 14b due to the weight of the metal band 12, so that the thickness t2 is increased to prevent deformation and damage of the hollow roll, and that the upper hollow roll 14a is magnetically sensitive. This is because, since the element group is included, the thickness t1 is reduced to obtain an effect of improving the magnetic detection sensitivity.

One end of each of the hollow fixed shafts 21a and 21b is penetrated through each central shaft of each of the hollow rolls 14a and 14b. The other end of the fixed shaft 21b of the lower hollow roll 14b is fixed to a frame of a building (not shown). On the other hand, the other end of the fixed shaft 21a of the upper hollow roll 14a is urged by a very weak spring (not shown) with respect to the fixed shaft 21b of the lower hollow roll 14b. Each of the fixed shafts 21a, 21b is located on a respective central axis of each of the hollow rolls 14a, 14b, so that a pair of rolling bearings 22a, 22b, 23a, 2 are respectively provided.
The hollow rolls 14a and 14b are supported on inner peripheral surfaces at both ends via 3b. Therefore, each hollow roll 14
a and 14b rotate freely with the fixed shafts 21a and 21b as the rotation center axes. In the hollow roll 14a, a magnetically sensitive element group (magnetic sensor group) 24 in which a plurality of magnetically sensitive elements (magnetic sensors) are arranged in the axial direction faces the fixed shaft 21a via the support member 25 in a posture facing downward. Fixed. The tip of the magnetically sensitive element group 24 is opposed to the inner peripheral surface of the upper hollow roll 14a with a small gap. The detection signal output from the magnetically sensitive element group 24 for each element is the signal line cable 1 via the fixed shaft 21a.
The signal is led to the signal processing unit of FIG. 1 via 8a.

On the other hand, in the lower hollow roll 14b,
Each of the free ends 26 has a magnetized core 26 having a substantially U-shaped cross section.
a is fixed to the fixed shaft 21 via the support member 27 in a posture in which a faces upward. And each free end 26 of the magnetized iron core 26
a is opposed to the inner peripheral surface of the hollow roll 14b with a small gap. A magnetized coil 28 is wound below the magnetized core 26. The magnetizing current of the magnetizing coil 28 is supplied from the magnetizing power supply device via the power supply cable 17a via the inside of the fixed shaft 21b. Therefore, the magnetically sensitive element 24 housed in the upper hollow roll 14a is
, And opposes a magnetizer including a magnetic core 26 and a magnetizing coil 28 housed in the lower hollow roll 14b. Further, a distance d1 between the magnetically sensitive element 24 and the metal strip 12;
And the distance between the magnetizer and the metal strip 12 is fixed.

In the magnetic flaw detector having such a configuration,
When the metal band 12 is run at a constant speed in the direction of the arrow shown in FIG. 3, a downward force due to the tension of the metal band 12 and a downward force due to gravity are applied to the lower hollow roll 14b. The side hollow roll 14b rotates in the direction of the arrow. The upper hollow roll 14a is also urged by the lower spring, so that it rotates in the direction of the arrow. Then, when the magnetizing power supply device is activated and an exciting current is supplied to the magnetizing coil 28, a closed magnetic path is formed by the magnetizing core 26 and the running metal band. If a defect such as a flaw or a bubble exists inside or on the surface of the metal band 12, a magnetic path in the metal band 12 is disturbed, and a leakage magnetic flux is generated. This magnetic flux is applied to the magnetic sensing element group 2
4 is detected as a defect signal for each magnetically sensitive element (magnetic sensor).

FIG. 5 is a diagram for explaining a method of selecting magnetic sensor groups arranged in one line and magnetic sensors of each pair according to the present invention. FIG. 5A schematically shows n magnetic sensors S1 to Sn arranged at equal intervals in one row in a direction perpendicular to the running direction of the metal strip. Here, the interval between each magnetic sensor is 1
When the area where flaws can be detected is determined by the two magnetic sensors,
It is set so that a gap is not generated between the detection regions by the adjacent magnetic sensors, that is, the sensor detection regions slightly overlap.

Next, a method of selecting each pair of magnetic sensors will be described. The magnetic sensor group arranged in one line in a direction perpendicular to the traveling direction of the metal strip is suitable for detecting a point-like flaw or a linear flaw in the traveling direction. The selection criteria for a pair of magnetic sensors are:
The output signal of one magnetic sensor contains a sufficient flaw signal, and the output signal of the other magnetic sensor contains little or no flaw signal. In addition, the output signals of both magnetic sensors include noise signals of substantially the same level. This is because the difference between the peak hold values of the pair of magnetic sensor output signals is calculated to obtain the flaw signal of the present invention. Therefore, specifically, the distance between the two magnetic sensors is not too close (if both are too close both include a flaw signal) and not too far (if too far, the two magnetic sensors A pair of magnetic sensors at a predetermined interval described in the section of the operation are selected because the noise level difference signal of the sensor is large.

In FIGS. 5B and 5C, two magnetic sensors arranged at a distance twice the arrangement interval from a group of magnetic sensors arranged linearly at equal intervals, that is, odd-numbered numbers shown in FIG. Competitor's magnetic sensors S1 and S3, S5 and S7, ...
(C) Even numbered magnetic sensors S2, S4, S6
, S8,... Are selected as a pair of magnetic sensors. This is because the predetermined interval is almost twice as long as the arrangement interval of the magnetic sensors.

FIGS. 5B and 5C show an example of selecting a pair of magnetic sensors instead of adjacent magnetic sensors from a group of magnetic sensors arranged linearly at equal intervals. Although shown, the present invention is not limited to this. Depending on the size of the flaw of the metal strip and the arrangement interval of the magnetic sensors, two magnetic sensors at a distance three times the arrangement interval, that is, S1 and S4 , S2 and S5, S3 and S6,... May be respectively selected as a pair of magnetic sensors.

FIG. 6 is a diagram for explaining a method of selecting magnetic sensor groups arranged in two rows and each pair of magnetic sensors according to the present invention. FIG. 6 (a) shows two directions perpendicular to the running direction of the metal band.
S11 to Sn of the front row and S2 of the rear row arranged in a row
1 to S2n schematically show a total of 2 × n magnetic sensors. Here, the distance between the magnetic sensors in the front row and the rear row (that is, the sensor distance in the running direction) is restricted at its upper limit to be within the distance between the two free ends 26a of the magnetized iron core 26 shown in FIG. The sensor interval in the traveling direction is set so as to be suitable for the type of flaw to be detected within the upper limit range, and the details will be described with reference to FIGS. 7 and 8. The intervals between magnetic sensors adjacent in the horizontal direction in each row are the same as those in FIG.

Next, a method of selecting each pair of magnetic sensors from the magnetic sensor group arranged in two rows in FIG. 6 will be described. FIG. 6B shows S11 and S21, S12 and S22,... S1n of the magnetic sensors arranged before and after in the running direction of the metal strip.
And S2n are selected as a pair of magnetic sensors. In this case, it is a selection method suitable for detecting a minute flaw or a weld. 6 (c) and (d) show the diagonally rearward magnetic sensors S11 and S22, S13 and S, except for the magnetic sensors in the front row and the back row.
, S12 and S21, S14 and S23,... Are each selected as a pair of magnetic sensors. This case is a selection method suitable for detecting a planar flaw.

FIG. 7 is a view for explaining the sensor-sensitive width in the running direction according to the type of flaw in the thin steel strip according to the present invention.
FIG. 8 is a waveform diagram of the sensor response width of FIG. The sensor interval in the traveling direction when the sensors in FIG. 6 are arranged in two rows is a width at which the sensor can detect a signal in response to the flaw signal when the target flaw passes directly below the sensor (hereinafter referred to as a sensitive width). ), The detection sensitivity (ie, S / N) in the difference calculation circuit 4 of FIG. 1 can be almost doubled. Generally, a sensor having a large flaw can obtain a large sensor output, but a small flaw has a small sensor output. Therefore, it is extremely effective to double the detection sensitivity by the difference calculation of the present invention.

FIG. 7 shows the results of experimentally determining the sensor response width in the running direction for various types of flaws starting from a small flaw of 0.2 mmφ. FIG. 8A shows a drill hole (φ = 0.
2B) shows the waveform of the magnetic sensor output and the waveform after unifying the polarities of the welded portion before cold rolling, respectively. From the magnetic sensor, a negative polarity output signal and a positive polarity output signal corresponding to one cycle of AC with respect to the traveling direction of the horizontal axis are obtained. Then, the traveling distance corresponding to this one cycle is indicated as a sensitive width. The waveform after the unification of the polarities is shown as a waveform obtained by inverting the negative polarity signal of the magnetic sensor output to the positive polarity signal and performing full-wave rectification on the alternating current. From the experimental results of FIG. 7, the sensor interval in the running direction in FIG. 6 is 3 to 15 mm for the micro flaw of the thin steel strip, 20 to 30 mm for the weld before cold rolling of the thin steel strip, and It can be seen that it is appropriate to set the width to 50 to 200 mm for the weld after cold rolling. In this way, the sensor interval in the traveling direction is set according to the type of flaw to be detected.

FIG. 9 is a diagram showing an embodiment of the selection means for selecting each pair of magnetic sensors according to the present invention. In the figure, 24 is a magnetic sensor group arranged in two rows, 30 is a wiring change board, 40 is a measuring device, and the measuring device 40 has only # 1 to # 4 pair peak hold circuits 41 to 44 on the input side. It is shown. Each pair peak hold circuit has
And two peak hold circuits identical to the # 1 peak hold circuit and # 2 peak hold circuit shown in FIG. The wiring change board 30 shown in FIG. 9 can manually change the wiring between the input terminal and the output terminal manually.
For example, the magnetic sensor paired with the magnetic sensor S11 is S2.
1 or by changing the wiring to S22 and connecting to the input side of the # 1 pair peak hold circuit 41. Further, it is also possible to input the output signals of the respective magnetic sensors to the respective multiplexers and take out the desired output signals of the two magnetic sensors as a pair using the selection control signal. The present invention thus doubles the S / N of the flaw signal obtained by calculating the difference between the peak hold values of the magnetic sensor outputs of each pair selected manually or automatically according to the type of flaw in the metal strip. Not only can the main flaws be detected using the first threshold value, but also a wide range of flaws can be detected with high sensitivity up to a minute flaw or a weld using the second threshold value calculated from the difference calculation result. it can.

[0037]

As described above, according to the present invention, a magnetic field is generated in the running direction of the running metal band, and the leakage magnetic flux generated by a defect inside or on the surface of the metal band is perpendicular to the running direction of the metal band. In the method and apparatus for detecting a metal strip to obtain a defect signal by detecting a plurality of magnetically sensitive elements arranged in one or two rows in the direction, the metal strip may be arranged in one row corresponding to the type of flaw in the metal strip. From the plurality of magnetically sensitive elements arranged side by side, two magnetically sensitive elements at a predetermined interval are selected so as to form a pair, and from the plurality of magnetically sensitive elements arranged side by side in the two rows, Select each magnetically sensitive element and each magnetically sensitive element in the rear row immediately behind the magnetically sensitive element so that they are paired with each other, or with each magnetically sensitive element in the front row and each magnetically sensitive element in the rear row other than the position immediately behind. Each Selected to pair. The detection sensitivity of the obtained flaw signal can be doubled by calculating the difference between the peak hold values of the two detection signals obtained for each fixed traveling section for each pair of the magnetically sensitive elements selected in this way. Therefore, not only can the main flaw be detected using the first threshold, but also the second flaw calculated from the difference operation value can be detected.
By using the threshold value, an effect is obtained in which various flaws in a wide range up to a welded portion can be detected with high sensitivity.

[Brief description of the drawings]

FIG. 1 is a system diagram of a metal band flaw detection signal processing unit according to the present invention.

FIG. 2 is a waveform chart for explaining the operation of FIG.

FIG. 3 is a schematic cross-sectional view of a magnetic flaw detector cut along a plane parallel to a running direction of a metal strip.

FIG. 4 is a schematic cross-sectional view of the magnetic flaw detector cut along a plane perpendicular to the running direction of the metal strip.

FIG. 5 is a diagram for explaining a method of selecting magnetic sensor groups arranged in one row and magnetic sensors of each pair according to the present invention.

FIG. 6 is a diagram illustrating a method of selecting a magnetic sensor group arranged in two rows and a magnetic sensor of each pair according to the present invention.

FIG. 7 is a view for explaining a sensor sensitivity width in the traveling direction according to the type of flaw of the thin steel strip according to the present invention.

FIG. 8 is a waveform chart of the sensor sensitivity width of FIG. 7;

FIG. 9 is a diagram showing an embodiment of a selection means for selecting each pair of magnetic sensors according to the present invention.

FIG. 10 is a block diagram of a conventional steel sheet magnetic flaw detector.

FIG. 11 is a schematic sectional view of a conventional flaw detection unit.

FIG. 12 is an explanatory diagram of a conventional detection signal and a first threshold value and a second threshold value.

FIG. 13 is a partially cutaway sectional view of a conventional magnetic flaw detection unit.

FIG. 14 is an electric circuit diagram of FIG.

FIG. 15 is an output waveform diagram of FIG.

FIG. 16 is a characteristic diagram of a conventional magnetic sensing element interval and output.

FIG. 17 is a waveform chart illustrating a detection signal of Reference 2.

[Explanation of symbols]

 1-1 # 1 magnetic sensor 1-2 # 2 magnetic sensor 2 section signal generator 3-1 # 1 peak hold circuit 3-2 # 2 peak hold circuit 4 difference operation circuit 5 second threshold operation circuit 6-1 # 1 Comparison circuit 6-2 # 2 comparison circuit

Claims (5)

(57) [Claims] A magnetic field is generated in a traveling direction of a traveling metal band, and a leakage magnetic flux generated by a defect in the inside or a surface of the metal band is arranged in a row in a direction perpendicular to the traveling direction of the metal band. In the magnetic flaw detection method for a metal band which obtains a defect signal by detecting a magnetically sensitive element, the two magnetically sensitive elements at a predetermined interval are selected as a pair from the plurality of magnetically sensitive elements, and each of the selected ones is selected. Detecting the maximum peak value of each of the two detection signals obtained for each pair of the magnetic sensitive elements in each of the fixed running sections of the metal band, and detecting the two maximum peak values detected for each of the magnetic sensitive elements of each pair; Is calculated as a flaw signal value for each traveling section of each pair of magnetically sensitive elements, and a flaw signal value for each traveling section of each pair of magnetically sensitive elements is set to a first threshold value set in advance. Compared to If the first threshold value is exceeded, this is detected as a main flaw of the metal strip, and a predetermined bias value is added after multiplying a flaw signal value for each traveling section of each pair of the magnetic sensitive elements by a predetermined coefficient. The calculated value is used as a calculated value of the corresponding section, and a moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated. As a second threshold, comparing the flaw signal of each pair of magnetically responsive elements for each traveling section with its own second threshold calculated in the immediately preceding traveling section, and exceeding the second threshold, A magnetic flaw detection method for a metal strip, characterized by detecting this as a minute flaw in the metal strip. 2. A magnetic field generated in a traveling direction of a traveling metal band, and a leakage magnetic flux generated by a defect in the inside or a surface of the metal band is arranged in two rows in a direction perpendicular to the traveling direction of the metal band. A magnetic flaw detection method for a metal strip, which detects a defect signal by detecting the magnetically sensitive element according to (1), wherein each of the magnetically sensitive elements in the front row of the plurality of magnetically sensitive elements arranged in the two rows and the rear row just behind the magnetically sensitive elements are arranged. Each of the magnetically sensitive elements is selected as a pair, and the maximum peak value of each of the two detected signals obtained for each of the selected pairs of magnetically sensitive elements is determined for each of the constant running sections of the metal band. Calculating the difference between the two maximum peak values detected for each pair of magnetically sensitive elements, and defining this as a flaw signal value for each running section of each pair of magnetically sensitive elements; Magnetic sensitive element Comparing the flaw signal value of each traveling section with a preset first threshold value, and when the flaw signal value exceeds the first threshold value, detects this as a main flaw of the metal strip. After multiplying the flaw signal value for each traveling section by a predetermined coefficient, a predetermined bias value is added to obtain a calculated value for the corresponding section, and the calculated value and a plurality of section calculations stored over a plurality of past sections The moving average value is calculated as a second threshold value in the next traveling section, and the flaw signal of each pair of the magnetically sensitive elements in each traveling section is calculated as the preceding traveling section. A magnetic flaw detection method for a metal band, wherein the second threshold value is compared with the respective second threshold value, and when the second threshold value is exceeded, this is detected as a minute flaw in the metal band. 3. A plurality of magnetic fluxes generated by generating a magnetic field in a traveling direction of a traveling metal band and arranging a leakage magnetic flux generated by a defect inside or on the surface of the metal band in two rows in a direction perpendicular to the traveling direction of the metal band. In the magnetic band flaw detection method of detecting a defect signal by detecting the magnetic sensitive element by the magnetic sensitive element, the magnetic sensitive element in the front row of the plurality of magnetic sensitive elements arranged in the two rows and the rear row at a position other than the position immediately behind the magnetic sensitive element. Each of the magnetically sensitive elements is selected as a pair, and the maximum peak value of each of the two detected signals obtained for each of the selected pairs of magnetically sensitive elements is determined for each of the constant running sections of the metal band. Calculating the difference between the two maximum peak values detected for each pair of magnetically sensitive elements, and defining this as a flaw signal value for each running section of each pair of magnetically sensitive elements; Magnetic sensitive element Comparing the flaw signal value of each traveling section with a preset first threshold value, and when the flaw signal value exceeds the first threshold value, detects this as a main flaw of the metal strip. After multiplying the flaw signal value for each traveling section by a predetermined coefficient, a predetermined bias value is added to obtain a calculated value for the corresponding section, and the calculated value and a plurality of section calculations stored over a plurality of past sections The moving average value is calculated as a second threshold value in the next traveling section, and the flaw signal of each pair of the magnetically sensitive elements in each traveling section is calculated as the preceding traveling section. A magnetic flaw detection method for a metal band, wherein the second threshold value is compared with the respective second threshold value, and when the second threshold value is exceeded, this is detected as a minute flaw in the metal band. 4. A method in which a magnetic field is generated in a traveling direction of a traveling metal band, and a leakage magnetic flux generated by a defect in the inside or a surface of the metal band is arranged in a row in a direction perpendicular to the traveling direction of the metal band. A magnetic band flaw detection device for detecting a defect signal by detecting the magnetically sensitive element according to (a), selecting means for selecting a pair of two magnetically sensitive elements at a predetermined interval from the plurality of magnetically sensitive elements; Means for detecting the maximum peak value of each of the two detection signals obtained for each pair of magnetically sensitive elements selected by the means for each fixed traveling section of the metal band; Calculate the difference between the two maximum peak values detected for each magnetically sensitive element and output this as a flaw signal value for each running section of each pair of magnetically sensitive elements. Comparing the flaw signal value for each traveling section of each pair of magnetically sensitive elements output by the difference calculating means with a first threshold value set in advance, and determining whether the signal value exceeds the first threshold value. A first comparing means for detecting as a main flaw of the metal strip, a predetermined bias after multiplying a flaw signal value for each running section of each pair of magnetically sensitive elements outputted by the difference calculating means by a predetermined coefficient. The moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated, and the moving average value is calculated for the next traveling section. A second threshold value calculating means for outputting as its own second threshold value, and a second threshold value calculating means for outputting a flaw signal for each running section of each pair of magnetically sensitive elements output by the difference calculating means, Compare with your own second threshold calculated and output in the section It said second comparison means and the magnetic flaw detection method for a metal strip, characterized in that it comprises a detect this as a very small flaw of the metal strip when exceeding the second threshold value. 5. A magnetic field is generated in a traveling direction of a traveling metal strip, and a leakage magnetic flux generated by a defect in the inside or surface of the metal strip is arranged in two rows in a direction perpendicular to the traveling direction of the metal strip. A magnetic band flaw detection device that obtains a defect signal by detecting the magnetically sensitive element according to (1), wherein the two rows of the plurality of magnetically sensitive elements are arranged in the front row and the rear row located immediately behind the magnetically sensitive elements. Selection means for selecting each of the magnetically sensitive elements of the pair as a pair, or selecting each of the magnetically sensitive elements in the front row and the respective magnetically sensitive elements in the rear row at a position other than immediately behind as a pair. Peak holding means for detecting the maximum peak value of each of the two detection signals obtained for each pair of the magnetically sensitive elements selected by the selection means, for each fixed traveling section of the metal band; Calculating the difference between the two maximum peak values detected by the peak hold means for each pair of magnetically sensitive elements, and outputting this as a flaw signal value for each running section of each pair of magnetically sensitive elements. Comparing the flaw signal value of each pair of magnetically sensitive elements output by the difference calculation means for each traveling section with a first threshold value set in advance, and determining whether the flaw signal value exceeds the first threshold value. A first comparing means for detecting as a main flaw of the metal strip, a predetermined bias after multiplying a flaw signal value for each running section of each pair of magnetically sensitive elements outputted by the difference calculating means by a predetermined coefficient. The moving average value of the calculated value and a plurality of section calculated values stored over a plurality of past sections is calculated, and the moving average value is calculated for the next traveling section. Output as their own second threshold in A threshold calculating means, and a second threshold calculating means for calculating and outputting a flaw signal for each traveling section of each pair of magnetically sensitive elements in the preceding traveling section, which is output by the difference computing means. A second comparing means for detecting, as a threshold, a minute flaw in the metal band when the value exceeds the second threshold, the magnetic inspection apparatus for a metal band.