JPH0334023B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an eddy current flaw detection device, and more particularly to an eddy current flaw detection device for detecting flaws on metal surfaces.

In the eddy current flaw detection method, a metal object to be inspected and a coil are placed opposite each other, and eddy currents are generated on the surface of the object by the electromagnetic action of the coil. If there is a defect on the surface of the object, the magnitude of the eddy current changes, and the impedance of the coil changes secondarily. The impedance change of this coil is detected to determine whether there is a defect or not.

Figure 1 shows a typical example of a conventional eddy current flaw detection device.The sine wave output of the oscillator 1 is
is applied to Both detection coils 2 and 3 constitute an impedance bridge circuit. Detection coil 2
3 is usually placed facing the subject 4 as shown in FIG. If there is a defect 4' on the opposite surface of the detection coil 2 of the subject 4, the impedance of the detection coil 2 changes and the output of the bridge circuit changes. This variation is amplified by an amplifier 5 and extracted by a detector 6. Hereinafter, this circuit will be referred to as a detection circuit.

By the way, the object to be inspected is much larger than the detection coil, so in order to detect the entire surface of the object, the third
A multi-parallel arrangement method in which a large number of detection coils are arranged side by side as shown in FIG. The former method of arranging multiple devices in parallel has the advantage that the drive unit is simple, but it requires a detection circuit for each coil, making the cost extremely high, and it is difficult to align the detection sensitivity of all detection circuits to the same level. There is a problem. On the other hand, although the latter rotating probe method has the problem that the drive mechanism for the detection coil is somewhat complicated, it is cheaper and does not have the problem of setting detection sensitivity, so it is more practical than the former.

However, the conventional rotating probe method has a problem in that it is difficult to accurately determine the position and direction of the flaw. When the subject 4 moves in the direction of the arrow as shown in FIG.
It changes as shown in Figure b. As can be seen from figure b, a flaw signal is generated every rotation of the rotary probe 2. Since the area ( ) intersects the trajectory of the rotating probe 2 and the flaw, the size of the flaw can be determined if the size of the area ( ) is known. However, it is difficult to distinguish between the regions (), (), and () in FIG. 4b, and it is even more difficult to determine the direction of the flaw.

In view of the drawbacks of the conventional method, the purpose of the present invention is to
An object of the present invention is to provide an eddy current flaw detection device that makes it possible to significantly improve flaw detection accuracy.

The present invention is provided with a signal processing device that converts a time series signal detected by a flaw detection circuit into a signal distribution on the surface of the object to be examined, and extracts flaw characteristics such as the flaw position, size, and direction from the signal distribution. It is characterized by

FIG. 5 shows an embodiment in which the present invention is applied to flaw detection of a continuously cast steel billet using a rotating probe type eddy current flaw detection device.

In FIG. 5, 4 is a metal steel piece, 100 is a rotating probe, 200 and 300 are detection circuits having the same function as the detection circuit in FIG. 1, 400 is a signal processing device for determining the size and position of a flaw, and 500 600 is a travel speedometer made of a metal piece; 700 is a rotation angle θ of the rotary probe 100;
This is a detection meter.

FIG. 6 shows details of the rotating probe 100. 1
01 is a rotating shaft, 102 is a motor that rotates the rotating shaft 101, 103 is a stationary side arm, 106 is a rotating side arm, 104, 105, 107, and 108 are rotating transformers for transmitting and receiving signals between the stationary side and the rotor side. , 109 are a pair of detection coils arranged parallel to the rotation direction, and 110 are a pair of detection coils arranged perpendicular to the rotation direction. The detection coil 109 is electrically connected to the fixed arm side through rotary transformers 107 and 104. Similarly, the detection coil 110 is connected to the rotating transformers 108 and 105.
It is electrically connected to the fixed side arm through. The rotational speed of the rotating shaft 101 can be controlled by the motor 102. Therefore, the rotating arm 106 fixed to the rotating shaft 101 and the detection coil 109,
The rotational speed of 110 can be controlled by motor 102.

Returning to FIG. 5, the signal detection circuit 200 extracts the impedance change of the pair of detection coils 109 in the rotary probe 100 using a bridge circuit and outputs a detected signal.
Output as na. Similarly, the signal detection circuit 300
outputs the impedance change of the pair of detection coils 110 as a detection signal n _b .

Here, in order to facilitate understanding of the present invention, the characteristics of the detection signals n _a and n _b for flaws parallel to and perpendicular to the traveling direction of the steel billet, which are most commonly found in continuously cast steel billets, will be described.

FIG. 7 shows changes in the detection signals n _a and n _b around the flaw parallel to the direction of steel billet travel. The signal n _a has a positive value on one side of the boundary of the flaw F and a negative value on the other side. The extreme values of the signal _{n a} extend in a ridge shape on both sides of the flaw F in parallel to the flaw F. The distance between these extreme values is approximately equal to the distance between the two coils constituting the detection coil 109. Although it is easy to see that the flaw F is between these extreme values, it is quite difficult to determine the edge of the flaw F from this power distribution. On the other hand, the signal n _b has an extreme value near the edge of the flaw F. Therefore, the edge of flaw F can be determined relatively easily from signal n _b , _but 2
It is not possible to determine whether the two extreme values are related to each other. As mentioned earlier the signal n _a
Since the extreme values of n extend along the scratch, they provide a clue to determine whether the extreme values of the signal n _b are at the edge of the same scratch. That is, the end of the flaw F, that is, the length of the flaw F can be determined from both the signals n _a and n _b .

FIG. 8 shows changes in the detection signals n _a and n _b around the flaw F perpendicular to the direction of movement of the iron piece. Both signals n _a ,
It can be seen that the characteristics of n _b are opposite to those in Figure 7. In other words, the continuity of flaw F is determined from signal n _b ,
The edge of the wound F can be determined from the signal n _a .

Note that the numerical values in FIGS. 7 and 8 indicate scratch potential values.

As a result of conducting similar studies on various cases, the following conclusions were reached.

(1) One of the signals n _a and n _b has an extreme value at the edge of the scratch, and the other extreme value represents the continuity of the scratch.

(2) In the case of a scratch that is longer than the distance between a pair of detection coils, the distance between the extreme values of signal n _a d _a and the distance between the extreme values of signal n _b
The extreme value of the signal corresponding to the longer one of d _{and b} corresponds to the wound edge.

(3) In the case of a flaw that is shorter than the distance between the pair of detection coils, the distances between extreme values d _a and d _b are both approximately equal to the distance d _p between the pair of coils.

Now, returning to FIG. 5, the detailed configuration of the signal processing device 400 will be explained with reference to FIG. 9.

In FIG. 9, 410 is a data input unit for inputting data, 420 is a signal conversion storage unit for allocating input data to an orthogonal grid on the surface of the subject, and 43
0 is a recognition unit that discriminates scratches. Among the elements constituting the data input storage sections 410 and 420, 401 is a rotation angle monitor section, 402 is a data input section, 403 is a detection coil position calculation section, 404 is a flaw potential calculation section, 405 is a flaw determination section, 406
is a flaw signal output processing section. Rotation angle monitor section 4
01 inputs the angle signal from the rotation angle detector 700, and the input angle monitor unit 401 inputs the angle signal from the rotation angle detector 700, and inputs the input signal to the data input unit 402 every time the rotation angle Δ changes by a certain angle tθ. Outputs data input commands. Upon receiving this data input command, the data input section 402 outputs the output signals n _{a and} n _b of the signal detection circuits 200 and 300 as well as the traveling speedometer 6 of the metal steel piece 4.
The speed signal V of 00 is A-D converted and input. The data input unit 402 uses the input data V to calculate the distance L from the tip of the metal steel piece 4 to the center of the rotating probe (the distance the steel piece moves) using the following equation.

L=L'+V・Δτ Where L' is the distance traveled when the data was input last time, and Δτ is the elapsed time from the previous time to this time.

The detection coil position calculation section 403 is the data input section 4
Detection coil 109 at the time of inputting data in step 02
and the position of 110 is determined. When the x and y axes are taken as shown in FIG. 10, the position (x _a , ya) of the detection coil 109 is calculated from the following equation.

x _a = RR - RRcos (θ + θ _p ) y _a = L + RRsin (θ + θ _p ) where RR is the radius of the rotating probe, θ is the rotation angle, and θ _p is the timing when the tip of the steel piece 4 passes through the center of the rotating probe is the rotation angle at .

Similarly, the position of the detection coil 110 (x _b , yb)
is calculated using the following formula.

x _b = RR + RRcos (θ + θ _p ) y _b = L−RRsin (θ + θ _p ) Furthermore, the detection coil position calculation unit 403
In the orthogonal grid space taken as _shown in the figure _{, the number of the grid point (I a ,}
J _a ) to the next integer value.

I _a ′=X _a /Δx, J _a ′=y _a /Δy In this case, I _a =I _a ′(ifx _a −I _a ′Δx≦Δx/2) I _a ′+1(ifx _a −I _a ′Δx >Δx/2) J _a = J _a ′ (ify _a −J _a ′Δy≦Δy/2) J _a ′+1 (ify _a −J _a ′Δy＞Δy/2) Similarly, the coordinates of the detection coil 110 ( _xb , _yb )
is also subtracted to the number (I _b , J _b ) of the neighboring grid point. The flaw potential calculation unit 404 first calculates the signal n _i
(i=a or b) is converted into a quantized signal N _i (integer).

N _i =n _i /Δn+0.5 Here, Δn represents the quantization unit of the signal,
It is a constant determined by the following formula.

Δn=(n _i ^MAX − n _i ^MIN )/m n _i ^MAX , n _i ^MIN : Maximum and minimum values of input level of signal n _i , m : Number of signal levels For example, looking at the detection number na, n _a ^MAX , n _a ^M
When ^IN is 100 and -100 and m is 5, the quantized signal N _a has the following value.

N _a =2…(60<n _a ≦100) 1…(20<n _a ≦60) 0…(−20<n _a ≦20) −1…(−60<n _a ≦−20) −2… (-100<n _a ≦ -60) Next, the flaw potential calculation unit 404 converts the quantized signals N _a and N _b to the grid points (I _a , J _a ), (I _b , J _b ).
In other words, the quantized signal N a is stored at address (I a , J _a ) of the memory area where the flaw potential data of the detection coil 109 is stored, and the quantized signal N _a is stored at the address (I _a , J a ) of the memory area where the flaw potential data of the detection coil 110 is stored. The quantized signal N _b is stored at address I _b , J _b ). However, data may be assigned to the same address several times depending on the moving speed of the steel piece and the rotational speed of the rotary probe. In such a case, the data N _a ,
To store N _b , first the data N _a ′ already stored at addresses (I _a , J _a ) and (I _b , J _b ),
Compare with N _b ′. Then, the absolute values of N _a and N _b are
Only when the absolute values are larger than the absolute values of N _a ′ and N _b ′, respectively, the contents of the corresponding address are rewritten to N _a and N _b . FIG. 12 shows the contents of the memory area after a certain flaw has passed under the rotating probe. The extreme value of the detection coil 110 appears at the edge of the wound, as shown in FIG. results have been obtained. The flaw determination unit 405 first searches for extreme values from flaw potential data regarding the detection coil 109. When the steel piece 4 moves and the Jth column of the orthogonal grid passes under the rotating probe, the data in the Jth column no longer changes. At this time, positive and negative extreme values are searched from the data in the J-th column, and data other than the extreme values in the J-th column are cleared to zero. In this way, the steel piece 4
As the data moves, only the data circled in Figure 12 remains.

On the other hand, an extreme value is searched for from the flaw potential data of the detection coil 110 as follows. The absolute value of each data in the Jth column is compared with the absolute value of each data in the (J-1)th column, and if the former is larger, the data in the (J-1)th column is cleared to zero. If these operations are repeated, the area circled in FIG. 12 will remain.

The flaw determination unit 405 sets the extreme value circled in FIG. 12 as "1" and sets the other data as zero.
Overlap these two data. As a result, the range of flaws as shown by the thick line in FIG. 13 can be determined.

Next, as shown in Fig. 13, the distance d _a between the opposing sides,
Calculate d _b . Distance a _a and d _b are both distances between coils
When d is smaller than _p , the flaw can be considered to be within the area within the broken line. On the other hand, the distance d _a is d _b and d _p
If it is larger, d _a is the length of the flaw, and the flaw width d _w is defined by the following formula.

d _w = d _b − d _p …(if d _b > d _p + Δd) Δd…(if d _a ≦d _p + Δd) Here, Δd is the minimum unit of line width, which is determined as a constant in relation to the display. It is. At this time, the wound is the 13th
It can be considered to be within the dashed line in the figure. vice versa,
If the distance d _b is greater than d _a and d _p , the length of the scar is d _b
So, the scratch width can be considered as d _a − d _p . Through such processing, data outside the broken line becomes zero, and the flaw position is clearly determined. In the flaw determination unit 405, the steel piece 4
By continuously performing such signal processing according to the movement of the steel piece, it is possible to detect the distribution of flaws on the steel piece.

The flaw signal output processing unit 406 displays the flaw distribution pattern detected by the flaw determination unit 405 on the viewer 500. Further, if an automatic scratch care device (not shown) is provided, this scratch distribution information is output to the automatic scratch care device.

As described above, according to the present invention, it is possible to accurately detect vertical cracks and horizontal scratches in a continuously cast steel billet, so maintenance work on the steel billet can be performed accurately.

In the embodiment shown in FIG. 5, a case has been described in which the outputs of the detection circuits 200 and 300 take positive and negative values. FIG. 14a shows an enlarged view of a part of the output waveform in the area () in FIG. 4b. In this signal waveform, the point where the signal changes from positive to negative approximately corresponds to the position of the flaw. Therefore, if we convert the points where the sign changes in this way into rectangular coordinates, the set of these points will be the potential values +8 and - of the detected signal na in Figure 7.
Gather between 8. In other words, it can be used as information for estimating the position and size of a flaw.

Further, depending on the type of detection circuit, the detection signal shown in FIG. 14a is integrated as shown in FIG. 14b, or is output after full-wave rectification as shown in FIG. 14c. Even in such a case, after converting the input signal into orthogonal coordinates, the characteristics of the flaw can be extracted from the signal distribution.
For example, in the case of an integrated signal as shown in Figure 14b, it is possible to use a method to find the distribution of extreme values corresponding to the peak of the integrated value, or to perform spatial differentiation of this distribution and change the sign from positive to negative. It is also possible to extract points that change and extract features as a set.

As described above, according to the present invention, since the characteristics of the flaw signal are detected from the planar distribution of the flaw signal, highly accurate flaw detection is possible.

Although the above embodiment describes the case of a rotating probe system, when a pair of detection coils arranged in parallel as shown in Fig. 3a is used, the time series of each detection circuit coupled to each coil pair Inputting a signal and allocating the data to an orthogonal grid corresponding to the position of each coil pair is much easier than using a rotating probe method. Therefore, it is clear that the present invention can be applied to a system in which a large number of probes are arranged side by side.

[Brief explanation of the drawing]

Figure 1 is a circuit configuration diagram of a typical example of an eddy current flaw detection device, Figure 2 is a layout diagram of a detection coil, Figure 3 is an explanatory diagram of a conventional flaw detection method, and Figure 4a is a circuit probe type probe and flaws. Fig. 4b is a waveform diagram of a flaw detection signal, Fig. 5 is a circuit configuration diagram showing an embodiment of the present invention, Fig. 6 is a detailed configuration diagram of the rotating probe portion, Figs. Figure 8 is a characteristic diagram of the flaw potential due to the detection coil, Figure 9 is a detailed configuration diagram of the signal processing device, Figure 10 is a diagram of the relationship between the steel piece and the x-y coordinates,
Figure 11 is a diagram of the relationship between orthogonal coordinates and data detection points, Figure 12 is an example diagram of the flaw potential of orthogonal grid points,
FIG. 13 is an example diagram showing the extent and distribution of flaws, and FIG. 14 is a waveform diagram of a detection signal. 4...Metal steel piece, 100...Rotating probe, 2
00,300...Detection circuit, 400...Signal processing device.

Claims (1)

[Scope of Claims] 1. A rotating probe that generates an eddy current in a metal object by a pair of mutually orthogonal detection coils, a detection circuit that detects an impedance change of the pair of detection coils, and a detection circuit that detects the impedance change. data input means for sampling and inputting a time series signal of values;
Signal conversion storage means for allocating each sampling data to grid points of orthogonal grid coordinates taken on the surface of the object to be detected, and determining the extreme value distribution of the signal distribution obtained by the processing of the signal conversion storage means, and determining the extreme value distribution. Identify the edge and continuity of the flaw from the distribution, and determine the location, size, and
1. An eddy current flaw detection device comprising: a flaw recognition means for determining a direction; and a signal processing device having flaw recognition means for determining a direction.