JPH09274016A - Method and apparatus for detecting flaw of magnetic metal element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a minute flaw incapable of being sufficiently detected by a conventional leakage flux flaw detection method by performing operation relatively emphasizing flaw signals present in at least two signal waveforms in common with respect to a noise signal. SOLUTION: The leakage flux detection signals from the sensors 12, 13 provided on the upper and rear surfaces of a steel panel 11 pass through band- pass filters 16, 17 to be multiplied by a multiplier 18 and the presence and grade of a flaw are judged on the basis of the output of the multiplier by a judging circuit 19. The components of the signals become large as compared with a noise component by multiplication and even a minute flaw can be detected in a good S/N ratio.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for detecting minute defects such as inclusions existing inside or on the surface of a magnetic metal body by a leakage magnetic flux flaw detection method.

[0002]

2. Description of the Related Art In the case of a metallic material, if there is a defect inside or on the surface, there may be a problem in its quality such as strength. Therefore, for quality control or quality assurance, X-ray transmission method, ultrasonic flaw detection method, etc. Various non-destructive inspections have been carried out. For example, an X-ray transmission method and an ultrasonic flaw detection method are used to detect cracks in a welded portion of a steel pipe. On the other hand, regarding defects near the surface and thin strips, especially online,
An electromagnetic method, that is, a leakage magnetic flux method or an eddy current flaw detection method is often used.

In an ultra-thin steel sheet for can-making materials, defects such as inclusions may cause problems such as flange cracks during can-making. Therefore, online high-speed defects using the magnetic flux leakage method, etc. Detectors have been developed. For example, in Japanese Laid-Open Patent Publication No. 3-175352, as shown in FIG. 5, a steel plate 21 is sandwiched, a magnetizer 22 for magnetizing the steel plate is provided on one side, and a leakage caused by a defect is provided on the other side. A method has been proposed in which a magnetic sensor array 23 for detecting magnetic flux is arranged in the width direction, and the magnetic sensor array 23 is placed in each of the non-magnetic rolls 24 and 25 for flaw detection. According to such a method, the distance between the sensor 23 and the inspection object 21 can be kept small and stable, and minute defects in the steel sheet can be detected at high speed without contact.

[0004]

However, in the prior art including the method described in Japanese Patent Application Laid-Open No. 3-175352, when the internal or surface defects to be detected are small, or the internal defects of a thick metal strip are detected. In the case of detecting the defect, the defect signal becomes weak and the S / N ratio relatively deteriorates, making it difficult to detect the defect. In particular, background noise caused by the object to be inspected is often a problem as noise. The formation noise can be reduced by optimizing the measurement conditions, but as the defect signal becomes smaller, it becomes difficult to distinguish the formation noise from the defect signal, which makes it difficult to detect the defect. . For example, in the case of detecting foreign matter such as inclusions, which is a problem in ultra-thin steel plates for can making, there is a problem that the conventional technique cannot obtain sufficient detectability.

The present invention has been made to solve the above problems, and an object thereof is to enable detection of minute defects which could not be sufficiently detected by the conventional leakage magnetic flux flaw detection method.

[0006]

Means for Solving the Problems The above-mentioned problem is to apply a magnetic field to a magnetic metal body to be inspected to magnetize an inspection part, and to detect leakage magnetic flux generated from a defect of the magnetic metal body at least as a measurement condition.
A plurality of magnetic flux sensor flaw detection signals are obtained by detecting with a plurality of magnetic sensors whose items are different from each other, and the plurality of leak magnetic flux flaw detector signals are detected from the substantially same inspection site. Using a complex flaw detection signal obtained by performing signal processing including a calculation for relatively emphasizing a defect signal commonly present in at least two of the signals with respect to the noise signal, This is solved by a flaw detection method for a magnetic metal body, which is characterized by detecting defects.

In this method, (1) a magnetizer for magnetizing a magnetic metal body, and (2) a row of three magnetic poles made of an E-shaped ferromagnetic body, in which the running direction of the magnetic metal body is formed. A pair of E-type magnetic sensors arranged on both sides so as to sandwich the same portion of the magnetic metal body to be inspected, and an E-type core arranged along the center and a coil wound around the central magnetic pole of the E-type core, (3) A signal processing including an operation for emphasizing a defect signal commonly present in the two leakage magnetic flux flaw detection signals detected by the pair of E-type magnetic sensors relative to the noise signal is performed. A magnetic metal flaw detector comprising an arithmetic unit for obtaining a composite flaw detection signal obtained by performing the operation, and (4) a determination circuit for determining the presence or absence of a defect of the magnetic metal body or the grade of the defect from the composite flaw detection signal. Can be carried out.

Since signal processing including an operation for emphasizing a defective signal commonly present in at least two signals among a plurality of signals relative to a noise signal is performed, the S / N ratio is improved. It is possible to improve, and it is possible to detect minute defects with high accuracy.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION The measurement condition refers to a condition generally considered in the detection of magnetic flux leakage, and items include, for example, a filter, a measurement surface of a metal band, a distance between a sensor and an inspection object, and a magnetization. Force, magnetization direction, sensor type, etc. are considered. In the present invention, a plurality of leakage magnetic flux flaw detection signals are obtained by a plurality of magnetic sensors, at least of which different from each other.

As a signal to be calculated, a pair of signals obtained from the front and back surfaces of the non-inspection magnetic metal body at substantially the same location can be used. The defect signal is similarly obtained on the front and back surfaces of the magnetic metal body to be inspected, whereas the noise component is affected by the surface properties of the magnetic metal body to be inspected and the like, so that different signals are produced on the front and back surfaces. Therefore, the defect signal component can be increased relative to the noise component by calculating the signals on the front and back surfaces.

As a signal to be calculated, it is possible to use a signal that has been subjected to signal processing by a plurality of filters having mutually different filter constants. That is, for example, in FIG.
As shown in, when the first bandpass filter is a lower frequency band including the defective frequency and the second bandpass filter is a higher frequency band including the defective frequency,
Although the defect signal is common to both, the noise component does not always appear at the same position because the frequency bands of both are different. Therefore, the defective signal component can be relatively increased with respect to the noise component by calculating the signals that have respectively passed through these bandpass filters.

In this case, using a plurality of sensors,
Each output may be passed through a filter having a different frequency band, or only one sensor may be used and the output from the sensor may be branched and passed through a filter having a different frequency band.

As a signal to be calculated, a signal obtained from a plurality of magnetic sensors having different distances (lift-off) between the sensor and the metal body flaw detection surface can be used.
When the lift-off changes, as shown in FIG. 4, the cover range on the metal body detected by the sensor changes, so that the noise source to be detected changes, and as a result, the noise detected by the sensor changes. Becomes Therefore, if the lift-off changes, the defect signal will be detected in the same way, whereas the noise will be different. Therefore, the defect signal component can be relatively increased with respect to the noise component by calculating the signals obtained from the sensors having different lift-offs.

Signals obtained from a plurality of different types of magnetic sensors can be used as signals to be calculated. Depending on the type of sensor, the range of coverage may differ, or it may differ in that it has a differential function such as a sensor using an E-type core and that it does not have a differential function like a general sensor. Both sensors are designed to detect defective signals, but the noise signals are different due to these types of differences. Therefore, the defect signal component can be relatively increased with respect to the noise component by calculating the signals obtained from the plurality of sensors of different types.

In addition to the items shown here, if a signal from a defect is obtained and noise is different in the output under each measurement condition, it is used as a signal to be calculated. be able to.

In this way, for signals obtained from substantially the same inspection site among a plurality of leakage magnetic flux flaw detection signals obtained under a plurality of mutually different measurement conditions, at least two of the plurality of signals are concerned. Defects are detected by using a composite flaw detection signal obtained by performing signal processing including an operation for emphasizing a defect signal commonly present in the signals with respect to a noise signal.

When the two types of leakage magnetic flux flaw detection signals are used, the calculation is generally expressed by the equation (1).

A (x) = M1 (x) * M2 (x) (1) where x is the position on the inspection body and M1 (x) is obtained under a certain measurement condition (measurement condition 1). The leakage flux flaw detection signal value at the position x, M2 (x) indicates the leakage flux flaw detection signal value at the position x obtained under another measurement condition (measurement condition 2), *
Is an operator (eg multiplication, sum of squares, M1 (x), M2
(x) Addition after taking the absolute value of both, M1 (x), M2
(x) An operator that takes the smaller one after taking the absolute value of both). A (x) is the result of the calculation at position x.

At this time, a signal change due to a defect can be detected under all of a plurality of different measurement conditions even if the magnitudes thereof are different, so that the change appears at the same position, and noise is positionally different. If the measurement conditions are selected so as to output in different ways, it is possible to improve the S / N ratio by utilizing the difference between the characteristics of the defect signal and the noise. At the position where the ground noise is large in one of the leakage magnetic flux flaw detection signals, the ground noise is small in the other leakage magnetic flux flaw detection signal. Therefore, for example, by multiplying the signal values from the same inspected portion of both, the two signals are common. The noise portion is relatively weakened as compared with the defect signal appearing at. The same effect can be expected in addition after the absolute values of both the sum of squares and M1 (x) and M2 (x) are taken. In the calculation that takes the smaller of the absolute values of both, the result of the calculation that takes the smaller one does not become so small because the defect magnetic flux flaw detection signal has a large value in the defective part, while the noise part May be small, and by taking the smaller one, the calculation result may be considerably smaller. Therefore, the defective part is relatively
To be emphasized.

The type of calculation is such that a similar defect signal is strengthened by both leakage magnetic flux signals, while both are always at the same position,
Alternatively, the calculation may be performed so that noises that do not appear as the same waveform are relatively weakened. This can be selected according to the characteristics of the noise or the defect signal, or according to the restrictions in implementing the device.

In the above example, the calculation between two types of leakage magnetic flux flaw detection signals has been described as an example, but it is not necessary to limit to two types. It is expected that the use of many signals with different measurement conditions will enhance the effect that the defect signal is emphasized and the noise is relatively less emphasized. For example,
Letting M1 (x), M2 (x), M3 (x), and M4 (x) be leakage flux flaw detection signal data taken under different conditions,
B (x) shown in the following equation (2) can be used to determine the presence or absence of a defect. B (x) = abs [M1 (x) x M2 (x) x M3 (x)] + abs [M4 (x)] (2) where abs [Y] is a function that takes the absolute value of Y. is there.

Further, here, the case where a certain one kind of calculation result is used for judging the presence or absence of a defect is described.
This is not limited to one, and it is possible to use two or more. For example, it is possible to make a final judgment by combining the judgment results of A (x) and B (x). In addition, it is possible to add a determination based on the conventional leakage magnetic flux flaw detection signal alone. In such a case, the determination circuit checks not only for one signal but also for a plurality of signals whether or not the respective thresholds are exceeded, and further, by combining the results, presence / absence of a defect and a grade. Will be the final decision.

As a result, it becomes possible to detect minute defects whose S / N is bad and cannot be detected only by the simple conventional magnetic flux leakage method.

[0024]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing an example of an apparatus according to the present invention.
In FIG. 1, 11 is a steel plate, 12 and 13 are E-type magnetic sensors, 14 is a coil, 15 is a DC magnetizing device, 16 and 17 are bandpass filters, 18 is a multiplier, and 19 is a judgment circuit.

An E-type sensor 12 is arranged on one side of the steel plate 11 and a DC magnetizing device 15 and another E-type sensor 13 are arranged on the other side. The DC magnetizing device 15 magnetizes the portion of the steel sheet 11 to be inspected to the saturation region. The steel plate 11 includes an E-type sensor 12, a magnetizer 15 and an E-type sensor 1.
The three legs (magnetic poles) 12a, 12b, 12c of the sensor core are arranged along the movement direction. Of these three coils, the coil 14 wound around the center leg 12b of the E-type sensor is used as a magnetic detection unit, and the time differential value of the magnetic flux interlinking with the coil is measured.

Due to the structural characteristic that the core is E-shaped, the magnetic flux interlinking the coil is the foot 12 at one end.
It is the difference between the magnetic flux passing through the magnetic loop including a and the middle leg 12b and the magnetic flux passing through the magnetic loop including the leg 12c at the other end and the middle leg 12b. for that reason,
Noise components common to both loops are canceled and defective signal components not common to both loops are selectively detected. The function of the E-type sensor 13 is also the same as that of the E-type sensor 1 described above.
It is similar to the action of 2.

By passing the sensor output through appropriate band pass filters 16 and 17, a leakage magnetic flux flaw detection signal (differential value)
Can be obtained.

Here, as an arithmetic operation, an example will be described in which two leakage magnetic flux flaw detection signals obtained by two sensors arranged on both sides of a steel sheet are sandwiched between the signals from the same inspection position. The sensitivity in the depth direction of the steel sheet is relatively high near the front surface side where the sensor is present and relatively low near the back surface side where the sensor is not present. Therefore, the sensor 12 and the sensor 13 output signals differently depending on the positions of the signal and the noise source in the depth direction.

In FIG. 2A, the position x on the inspection body is taken as the horizontal axis, and the vertical axis is taken as the sensor 12 after processing such as a band pass filter.
9 shows the leakage magnetic flux flaw detection signal Ma (x) due to a.
Not only the signal is large at the defective part, but also the background noise due to the inspection object is large, so the S / N ratio is about 1.8. Similarly, FIG. 2B shows the leakage magnetic flux flaw detection signal M by the sensor 13 after processing such as a band pass filter.
b (x). In this case as well, the defect signal is output at the same position as Ma (x), but the S / N ratio is still about 1.9 because of background noise caused by the inspection object. Here, multiplication is performed for each signal value from the same position. When the calculation result A (x) is expressed by an equation, the calculation of A (x) = Ma (x) × Mb (x) (3) is performed.

In this case, since the size of the defect is to some extent, the defect signal appears in both signals and their positions are almost the same. Therefore, by performing multiplication, the defect corresponding portion of the signal A (x) after the operation is calculated. Grows. On the other hand, due to the difference in the sensitivity distribution in the depth direction between the two sensors, the formation noise of the formation noise does not always increase at the same position in both signals. For example, it may happen that noise from a noise source near the surface is detected by one sensor but is not detected by the other sensor. In that case, when both are multiplied, the noise corresponding part of the signal A (x) after calculation is generally not larger than that between the defective signal parts. That is, since the defect signal portion is emphasized as compared with the background noise portion, it becomes easy to determine the presence or absence of a defect by using the calculation result A (x), and the defect detectability is improved. In this example, the S / N ratio has improved to 5.2.

The determination circuit 19 compares the defect signal value with a predetermined threshold value by using the signal having a good S / N ratio obtained in this way to determine the presence / absence of a defect and the defect grade. I do.

[0032]

According to the present invention, a defect is detected by using a composite flaw detection signal obtained by computing signal waveforms obtained from substantially the same inspection portion of a plurality of eddy current flaw detection signals whose measurement conditions are different from each other. By determining the presence or absence and the degree of harmfulness, it is possible to detect a defect that is difficult to detect because the S / N ratio cannot be sufficiently obtained by the conventional method.

[Brief description of drawings]

FIG. 1 shows an embodiment of the device according to the invention.

FIG. 2 is a diagram showing a leakage magnetic flux flaw detection signal and a signal of a calculation result thereof.

FIG. 3 is a diagram showing the relationship between the frequencies of defects and noise and the band of a bandpass filter.

FIG. 4 is a diagram showing lift-off of a sensor, a detection range, and a range of a noise source.

FIG. 5 is a diagram showing an example of a conventional technique.

[Explanation of symbols]

 11 steel plate 12, 13 E-type sensor 12a, 12b, 12c E-type sensor leg 14 coil 15 DC magnetizing device 16, 17 band-pass filter 18 multiplier 19 determination circuit

Claims (8)

[Claims] 1. A magnetic flux is applied to a magnetic metal body to be inspected to magnetize an inspection portion, and a leakage magnetic flux generated from a defect of the magnetic metal body is detected by a plurality of magnetic sensors having at least one measurement condition different from each other. To obtain a plurality of leakage flux flaw detection signals, and the signals obtained from substantially the same inspection portion of the plurality of leakage flux flaw detection signals are common to at least two of the plurality of signals. Using a complex flaw detection signal obtained by performing signal processing including a calculation that relatively emphasizes the existing defect signal with respect to the noise signal,
A flaw detection method for a magnetic metal body, which comprises detecting defects in the magnetic metal body. 2. The magnetic metal body of claim 1, wherein the at least two signals include a pair of signals obtained from front and back surfaces of a non-test magnetic metal body at substantially the same location. Method of flaw detection. 3. The flaw detection of the magnetic metal body according to claim 1, wherein the at least two signals include signals processed by a plurality of filters having different filter constants. Method. 4. The at least two signals include signals obtained from a plurality of magnetic sensors having different distances between the sensor and the flaw detection surface of the metal body to be inspected. The flaw detection method for a magnetic metal body according to any one of 1. 5. The magnetic metal body according to claim 1, wherein the at least two signals include signals obtained from a plurality of magnetic sensors of different types. Flaw detection method. 6. The operation according to claim 1, wherein the operation for emphasizing a defective signal commonly present in the signals relative to the noise signal is multiplication. Item 6. A flaw detection method for a magnetic metal body according to the item. 7. A magnetizer for magnetizing a magnetic metal body, and (2) an array of three magnetic poles made of an E-shaped ferromagnetic body so that a row of three magnetic poles runs along the traveling direction of the magnetic metal body. An E-type magnetic sensor, which is composed of an E-type core arranged and a coil wound around the central magnetic pole of the E-type core, and is arranged in pairs on both sides so as to sandwich the same portion of the magnetic metal body to be inspected, (3) Performs signal processing including an operation for relatively emphasizing a defect signal commonly present in two leakage magnetic flux flaw detection signals detected by the paired E-type magnetic sensors with respect to a noise signal A magnetic metal flaw detector, comprising: an arithmetic unit for obtaining a composite flaw detection signal obtained by the above; and (4) a determination circuit for determining the presence or absence of a defect of the magnetic metal body or the grade of the defect from the composite flaw detection signal. 8. A multiplier is an arithmetic unit for obtaining a composite flaw detection signal obtained by performing an operation for relatively emphasizing a defect signal commonly present in two leakage flux flaw detection signals with respect to a noise signal. The magnetic metal flaw detector according to claim 7.