JP2798199B2 - Noise Removal Method in Eddy Current Testing - 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 for removing noise in an eddy current flaw detection method used for detecting pitting corrosion of a thin-walled copper pipe, estimating a service life, and the like.

[0002]

2. Description of the Related Art A known eddy current flaw detection method will be described with reference to FIGS.

As shown in FIG . 9 , when an AC-flowing coil is brought close to a test object (conductor), a vortex that generates a magnetic flux in the opposite direction to the magnetic flux on the surface of the test object by the action of the magnetic flux generated by the coil. Current (eddy current) is generated. This phenomenon is called an electromagnetic induction phenomenon. If a defect exists in the test piece, the eddy current is disturbed. Therefore, by detecting this as a change in impedance with a coil, the eddy current can be applied to defect inspection. An apparatus made so that the above-described electromagnetic induction phenomenon can be used for inspection is called an eddy current flaw detector, which can excite a coil with an alternating current and display a minute impedance change generated in the coil on an oscilloscope.

FIGS. 10 to 12 show examples of eddy current flaw detection when a defect is detected using a coil. Fig. 10 and Fig. 11 show the method of arranging the inspection coil . The eddy current flaw detector is a healthy part (calibration coil).
Then, the unbalance of the impedance of the two coils installed in the inspection part (inspection coil) is displayed on an oscilloscope as shown in FIG . Generally, as shown in FIG.
The eddy current flaw detector is adjusted so that the moving direction (lift-off direction) of the bright spot when the inspection coil separates from the test object coincides with the X-axis. As a result, when the inspection coil is caused to scan on the defect, as shown in FIG. 12 , the movement of the bright spot having the component in the Y-axis direction can be obtained.

Here, the position of the bright spot of the eddy current flaw detector oscilloscope is expressed by the X-axis (adjusted in the lift-off direction = V _X ) and the Y-axis signal voltage (= V _Y ), and the distance between the origin and the bright spot is deflected. The angle between the width (= V), the line connecting the origin and the bright spot, and the X axis is called the phase angle (= θ).

At this time, as shown in FIG. 13, there is known a method for evaluating the size of the pits on the inner surface of the copper pipe by evaluating the residual inner thickness from the phase angle and then evaluating the diameter of the defect from the deflection width. However, the probe coil is excellent for detecting minute defects, but is very sensitive to the distance and positional relationship between the coil and the test object because the direction of the generated magnetic flux is orthogonal to the test object. It has been pointed out that noise may be generated due to vibration or deformation or eccentricity of the pipe, which may cause a large error in the inspection result.

For this reason, as shown in the above defect size evaluation method, when the X and Y axis signal voltages cannot be measured with high accuracy, a large error is also included in the evaluation results of the residual wall thickness and the defect diameter.

[0008]

Copper pipes are thin and small in diameter, and often have a small pitting corrosion depth and depth. Therefore, it is necessary to conduct inspections with high sensitivity. Noise is generated by various disturbance factors that occur. Therefore, measures for removing this noise are extremely important for high-precision flaw detection.

[0009] The present invention is proposed in view of the above point, and an object thereof is indispensable for performing high-accuracy flaw detection when the eddy current flaw detection method is applied to a thin-walled, small-diameter copper pipe. It is to provide a simple noise removing means (method).

[0010]

The configuration of the noise removing method in the eddy current flaw detection method according to the present invention is as follows.

A. Scan the inspection coil in the axial direction of the pipe to extract the X-axis (lift-off direction) and Y-axis signals, b. Selecting an X-axis signal in a portion of the extracted signal in which the Y-axis signal is below a certain level, c. b. A noise approximation waveform is created by approximating the X-axis signal selected in b with, for example, a quartic curve equation. d. A noise removal method in the eddy current flaw detection method, wherein only a defect indication waveform is extracted by subtracting a noise approximate waveform of c from an initial waveform based on the X-axis signal extracted in a.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the proposed noise elimination method will be described below in detail.

A. Characteristics of the noise waveform When the noise waveform as shown in FIG. 1 was examined, the following was found. Since the generated noise occurs in the same frequency band as the defect indication, an electrical so-called band-pass filter cannot be applied. Since there is no noise in the Y-axis signal component, it is sufficient to remove noise only from the X-axis signal component. When examining the waveform of the Y-axis signal component, (1) about 5 to 6 pits occur during one scan (300 mm) of the inspection coil. : F (2) Noise caused by vibration of the motor or play of the inspection coil has a few peaks during one scan. : N
1 piece (3) Noise due to deformation and eccentricity of the pipe has 5 to 8 peaks during one scan. : N2 b. 40mm (= L) so that only one or two defect signals enter
The data of the section of is extracted. : L = 300 / (F + 1)
The above a. From the data extracted from (1), a so-called X-axis signal of a sound part in which the Y-axis signal is below a certain level is selected. : A.
The selected X-axis signal is approximated by a 4 (= n) order curve. :
n = (N1 + N2) / 300 * L + 1 a.
(1) From (2) and (3) Only the defect waveform can be extracted by subtracting the noise waveform of the sound part approximated by the fourth order from the original X-axis signal.

C. Embodiment of Proposed Noise Removal Technique FIGS. 3 to 6 show the noise removal technique. Section L shown in FIG.
In the original waveform, the defect signal is buried in the noise. First, as shown in FIG. 4 , an X signal of a portion where the output of the Y-axis signal is not obtained, that is, a healthy portion is extracted. Next figure
As shown in FIG. 5 , this signal including the missing portion is approximated by an nth order curve equation. This is a pure noise signal because the originally defective portion is treated as non-defective. Finally, by subtracting this noise signal from the initial waveform, only the defect indication waveform is obtained as shown in FIG .

The section L and the order n are determined by the method shown in b.

FIG. 7 shows a waveform after noise removal, and FIG. 8 shows an oscilloscope waveform.

An example of specific data detected by the above embodiment is as follows.

[0018] MotonikuAtsu _{0.81mm V X = -1.19 (V)} V Y = 2.43 (V) θ = tan -1 (V Y N X) = 116.1 ° residual meat ratio = 44. 6% (0.36 mm remaining) In addition, since the actually measured pit depth was 0.51 mm, the remaining meat ratio =
It was 37.0% (0.30 mm remaining).

[0019]

As described above, according to the present invention, in the eddy current flaw detection method, the X-axis signal of the portion where the Y-axis signal is below a certain level is selected from the initial waveform, and this signal is approximated by the n-th order curve equation. A noise approximation waveform is created by subtracting the noise approximation waveform from the initial waveform to extract only a defect indication waveform, that is, a pitting corrosion waveform. As a result, flaw detection of a thin-walled, small-diameter copper pipe can be performed with high accuracy without being affected by noise.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of a noise waveform.

FIG. 2 is an explanatory diagram of a noise waveform.

FIG. 3 is an explanatory diagram of a noise removal method.

FIG. 4 is an explanatory diagram of a noise removal method.

FIG. 5 is an explanatory diagram of a noise removal method.

FIG. 6 is an explanatory diagram of a noise removal method.

FIG. 7 is an explanatory diagram of a waveform after noise removal.

FIG. 8 is an explanatory diagram of a waveform after noise removal.

FIG. 9 is an explanatory diagram of an electromagnetic induction phenomenon.

FIG. 10 is an explanatory diagram of an eddy current flaw detection method.

FIG. 11 is an explanatory diagram of an eddy current flaw detection method.

FIG. 12 is an explanatory diagram of an eddy current flaw detection method.

FIG. 13 is an explanatory diagram of a defect size evaluation method.

Claims (1)

(57) [Claims] 1. A method comprising: a. Scan the inspection coil in the axial direction of the pipe to extract X-axis and Y-axis signals, b. Selecting an X-axis signal in a portion of the extracted signal in which the Y-axis signal is below a certain level, c. b. Create an approximate noise waveform by approximating the X-axis signal selected in b with an n-th order curve equation. d. A noise removal method in the eddy current flaw detection method, wherein only a defect indication waveform is extracted by subtracting a noise approximate waveform of c from an initial waveform based on the X-axis signal extracted in a.