CN102445493A - Modulation multifrequency eddy current testing method - Google Patents

Abstract

The invention relates to a modulation multifrequency eddy current testing method, comprising the following steps: generate a frequency modulation excitation signal, amplify the frequency modulation excitation signal, drive an eddy current sensor, and respectively test a benchmark testing piece and an object to be tested; collect response signals of the eddy current sensor when testing the benchmark testing piece and the object to be tested, perform an FFT (Fast Fourier Transformation) transformation to the response signals of the eddy current sensor to obtain a frequency spectrum of the response signals of the eddy current sensor, and then analyze the frequency spectrum of the response signals of the eddy current sensor and respectively calculate total spectrogram energy, spectrogram center-of-gravity shift, spectrogram kurtosis and spectrogram skewness of each frequency spectrum of the response signals of the eddy current sensor; finally, perform threshold value judgment to the total spectrogram energy, spectrogram center-of-gravity shift, spectrogram kurtosis and spectrogram skewness of the frequency spectrum of the response signals of the eddy current sensor of the object to be tested and the frequency spectrum of the response signals of the eddy current sensor of the benchmark testing piece to judge whether the object to be tested has defects and the defect location or other parameters or not. The modulation multifrequency eddy current testing method can effectively reduce peak factors of the excitation signal and the response signal of an eddy current testing system and increase the testing speed.

Description

Modulation multifrequency Eddy detection method

Technical field

The present invention relates to the multifrequency Eddy detection method that adopt in a kind of Non-Destructive Testing field, particularly a kind of modulation multifrequency Eddy detection method is applicable to that the interference of testing process suppresses the perhaps multiparameter detection of checked object.

Background technology

Eddy detection technology is safeguarded at the nuclear power station steam pipe; The Aero-Space equipment is safeguarded; Advantages such as industries such as pipe, line, bar production have vital role, and that eddy detection technology has is contactless, be prone to realize robotization, detectability is strong and obtained application more and more widely.

Conventional eddy current detection method adopts single frequency sinusoidal signal excitation eddy current sensor, and the quantity of information of obtaining is limited, is difficult to adapt to increasingly high Non-Destructive Testing demand.1970, U.S. scientist Libby H L at first proposed the multifrequency Eddy detection method, detected in order to the interference inhibition in the realization EDDY CURRENT process and the multiparameter of checked object.The multifrequency Eddy detection method adopts several different frequency sinusoidal signal excitation eddy current sensors, utilizes under the different frequency, and parameter has the different principles that change to realize.The testing result that under different frequency, obtains is carried out analyzing and processing through certain method, extracts a plurality of desired parameters, perhaps suppresses undesired signal.

The pumping signal mode that existing multifrequency Eddy detection method adopts has two kinds, and first kind is synchronous synthesis mode, shown in Fig. 1 a and Fig. 1 b; Second kind is asynchronous synthesis mode, shown in Fig. 2 a and Fig. 2 b.

Pumping signal adopts the sinusoidal signal stack of a plurality of frequency components in the multifrequency Eddy detection method of synchronously synthetic energisation mode, and the peak value of pumping signal is higher, and peak factor is bigger, to the voltage range requirement broad of driving circuit and rear end amplifying circuit.

Pumping signal adopts the sinusoidal signal of a plurality of frequency components to drive eddy current sensor successively in the multifrequency Eddy detection method of asynchronous synthetic energisation mode, needs constantly to switch exciting signal frequency, and detection time is longer.

Summary of the invention

Technical matters to be solved by this invention is, the deficiency to existing multifrequency Eddy detection method provides a kind of modulation multifrequency Eddy detection method, and it can effectively reduce the peak factor of eddy detection system pumping signal and response signal, the raising detection speed.

For solving the problems of the technologies described above, the technical scheme that the present invention adopted is: a kind of modulation multifrequency Eddy detection method, it comprises the following steps:

The first step, preparation benchmark test specimen;

Second step, generation one fm exciter signal, and, drive eddy current sensor with after this fm exciter signal amplification, the benchmark test specimen is detected;

The 3rd step, the response signal of eddy current sensor during acquisition testing benchmark test specimen, and this eddy current sensor response signal is carried out the FFT conversion obtain benchmark test specimen eddy current sensor response signal frequency spectrum, promptly obtain detecting required benchmark response signal frequency spectrum;

The 4th goes on foot, benchmark response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this benchmark response signal frequency spectrum;

The 5th step, adopt aforesaid fm exciter signal, and with this fm exciter signal like the second said amplification of step after, drive aforementioned eddy current sensor, checked object is detected;

The 6th step, the response signal of eddy current sensor during the acquisition testing checked object, and this eddy current sensor response signal is carried out the FFT conversion obtain checked object eddy current sensor response signal frequency spectrum;

The 7th goes on foot, checked object eddy current sensor response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this checked object eddy current sensor response signal frequency spectrum;

The 8th step, the spectrogram gross energy according to benchmark response signal frequency spectrum, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias are carried out threshold decision to spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of above-mentioned detected checked object eddy current sensor response signal frequency spectrum: if the spectrogram gross energy of detected checked object eddy current sensor response signal frequency spectrum greater than the spectrogram gross energy of benchmark response signal frequency spectrum, then proves to have defective on the checked object; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that greater than the spectrogram degree of bias of benchmark response signal frequency spectrum defective is a surface imperfection on the checked object; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that less than the spectrogram degree of bias of benchmark response signal frequency spectrum defective is an inherent vice on the checked object.

This fm exciter signal is the linear frequency modulation pumping signal.The optimized frequency scope of this linear frequency modulation pumping signal is 1kHz～20kHz; Time span is unsuitable long, guarantees that simultaneously FFT requires preferably to count, and elects 4096us as.

The SF of gathering the eddy current sensor response signal is 5-10 a times of fm exciter signal highest frequency.

Compare with existing method; The beneficial effect that the present invention had is: the present invention adopts the pumping signal of modulation multifrequency mode as eddy detection system; The discrete spectrum of conventional multifrequency Eddy method response signal frequency domain is become continuous spectrum; Through the characteristic quantity of calculating spectrogram, and then realize disturbing and suppress or the multiparameter detection.Compare with synchronous synthesis mode multifrequency Eddy detection method; Advantage of the present invention is: adopt a FM signal; Effectively reduce the peak factor of multi-frequency excitation signal and eddy current response signal, help the design of eddy current sensor pumping signal driving circuit and response signal rear end amplifying circuit.Compare with asynchronous synthesis mode multifrequency Eddy detection method, advantage of the present invention is: FM signal is the pumping signal of a continuous change frequency in short-term, need not to switch, and has reduced detection time, thereby has improved detection defective speed.

Description of drawings

Below in conjunction with accompanying drawing and embodiment further explain is done in invention; But modulation multifrequency Eddy detection method of the present invention is not limited to embodiment.

The time domain waveform of the existing synchronous synthesis mode multi-frequency excitation signal of Fig. 1 a;

The spectrogram of the existing synchronous synthesis mode multi-frequency excitation signal of Fig. 1 b;

The time domain waveform of the existing asynchronous synthesis mode multi-frequency excitation signal of Fig. 2 a;

The spectrogram of the existing asynchronous synthesis mode multi-frequency excitation signal of Fig. 2 b;

Fig. 3 the present invention modulates the workflow diagram of multifrequency Eddy detection method;

The benchmark test specimen that adopts in Fig. 4 embodiment of the invention;

Fig. 5 a is the linear FM signal figure that the embodiment of the invention adopts;

Fig. 5 b is the linear FM signal spectrogram that the embodiment of the invention adopts;

Fig. 6 a is the eddy current sensor response signal that embodiment of the invention detection reference test specimen obtains;

Fig. 6 b is the response signal spectrogram of embodiment of the invention detection reference test specimen eddy current sensor;

The checked object of Fig. 7 for adopting in the embodiment of the invention;

Fig. 8 is the eddy current sensor response signal spectrogram when detecting the checked object surface imperfection in the embodiment of the invention;

Fig. 9 is the eddy current sensor response signal spectrogram when detecting the checked object inherent vice in the embodiment of the invention;

The spectrogram gross energy of Figure 10 for obtaining in the embodiment of the invention;

The spectrogram centre-of gravity shift of Figure 11 for obtaining in the embodiment of the invention;

The spectrogram kurtosis of Figure 12 for obtaining in the embodiment of the invention;

The spectrogram degree of bias of Figure 13 for obtaining in the embodiment of the invention.

Embodiment

Modulation multifrequency Eddy detection method workflow of the present invention is as shown in Figure 3.

See also Fig. 3 to Figure 11 below, with the dull and stereotyped defects detection of aluminum be categorized as example and specify the present invention and modulate multifrequency Eddy detection method workflow:

The first step, preparation benchmark test specimen, this benchmark test specimen is flawless test specimen: as shown in Figure 4 in the present embodiment, this benchmark test specimen adopts flawless aluminum flat board 1, and it is of a size of 200mm * 200mm * 2mm.

Second step, generation one fm exciter signal, and, drive eddy current sensor with after this fm exciter signal amplification; The benchmark test specimen is detected: in the present embodiment, what this fm exciter signal adopted is linear frequency modulation pumping signal 5, shown in Fig. 5 a; The horizontal ordinate express time; Unit is us, and ordinate is represented amplitude, and unit is V.This linear frequency modulation pumping signal is the linear FM signal with rectangle envelope, and its time domain expression formula does

$S\left(t\right)=A\·\mathrm{rect}(t/T)\·\cos \left[2\mathrm{\π}(f_{0}t+\frac{1}{2}{\mathrm{Kt}}^2)\right]$

In the formula: S (t) is a linear FM signal, and A is the amplitude of linear FM signal, and t is a time variable, and T is the time width of linear FM signal, f ₀Be the centre frequency of linear FM signal, K is chirped slope, and rect (t/T) is a rectangular function.

The frequency range of this linear frequency modulation pumping signal 5 is 1kHz～20kHz, and time span is 4096us.Linear frequency modulation pumping signal 5 is carried out the FFT conversion can get linear FM signal frequency spectrum 6, shown in Fig. 5 b, horizontal ordinate is represented frequency, and unit is kHz, and ordinate is represented the FFT amplitude, no unit.After these linear frequency modulation pumping signal 5 amplifications, drive eddy current sensor, flawless aluminum dull and stereotyped 1 is detected.

The 3rd the step, according to time domain sampling theorem, confirm SF.Generally speaking, SF be fm exciter signal highest frequency 5-10 doubly.The response signal of eddy current sensor during acquisition testing benchmark test specimen, and this eddy current sensor response signal is carried out the FFT conversion obtain benchmark test specimen eddy current sensor response signal frequency spectrum, promptly obtain detecting required benchmark response signal frequency spectrum.Fig. 6 a is the detection reference test specimen that collects in the present embodiment, the dull and stereotyped 1 o'clock eddy current sensor response signal 7 of promptly flawless aluminum, and its horizontal ordinate express time, unit is us, and ordinate is represented amplitude, and unit is mV.By finding out the frequency range of eddy current sensor response signal 7 identical with time span 5 among the figure with the linear frequency modulation pumping signal.This eddy current sensor response signal 7 is carried out the FFT conversion can get benchmark test specimen eddy current sensor response signal frequency spectrum 8, shown in Fig. 6 b, horizontal ordinate is represented frequency, and unit is kHz, and ordinate is represented the FFT amplitude, no unit.

The 4th goes on foot, benchmark response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this benchmark response signal frequency spectrum.Each characteristic quantity calculating formula is following:

The spectrogram gross energy does

$\mathrm{Energy}=\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}X{\left(k\right)}^2$

In the formula: Energy is a spectrogram gross energy result of calculation, and X (k) is the Fourier transform of detection signal, and k is the horizontal ordinate variable of Fourier transform, N ₁Be the corresponding Fourier transform point horizontal ordinate of linear frequency modulation initial frequency, N ₂For linear frequency modulation stops the corresponding Fourier transform point horizontal ordinate of frequency.In the present embodiment, calculate the spectrogram gross energy of this benchmark response signal frequency spectrum, thereby when obtaining the test specimen zero defect; The i.e. spectrogram gross energy 16 of this benchmark response signal frequency spectrum, shown in figure 10, horizontal ordinate is represented defective; No unit, ordinate represent the spectrogram gross energy, no unit.

Spectrogram centre-of gravity shift does

$\mathrm{Barycenter}=\frac{\underset{k=N_1}{\overset{{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="HDA0000028150910000053.png,HDA0000028150910000071.png"\; state="\{\{state\}\}">N</figure-callout>}}_2}{\mathrm{\&Sigma;}}}\mathrm{kX}\left(k\right)}{\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}X\left(k\right)}$

In the formula: Barycenter is a spectrogram centre-of gravity shift result of calculation, and X (k) is the Fourier transform of detection signal, and k is the horizontal ordinate variable of Fourier transform, N ₁Be the corresponding Fourier transform point horizontal ordinate of linear frequency modulation initial frequency, N ₂For linear frequency modulation stops the corresponding Fourier transform point horizontal ordinate of frequency.In the present embodiment, calculate the spectrogram centre-of gravity shift of this benchmark response signal frequency spectrum, thereby when obtaining the test specimen zero defect; The i.e. spectrogram centre-of gravity shift 19 of this benchmark response signal frequency spectrum, shown in figure 11, horizontal ordinate is represented defective; No unit, ordinate represent the spectrogram barycentre offset, no unit.

The spectrogram kurtosis does

$\mathrm{Kurtosis}=\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}{[X\left(k\right)-\stackrel{\&OverBar;}{X}]}^4/{\mathrm{<figure-callout\; id="4"\; label="SD"\; filenames="HDA0000028150910000053.png,HDA0000028150910000071.png"\; state="\{\{state\}\}">SD</figure-callout>}}^4$

Wherein

$\stackrel{\&OverBar;}{X}=\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}X\left(k\right),$ $\mathrm{SD}={\left(\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}{[X\left(k\right)-\stackrel{\&OverBar;}{X}]}^2\right)}^{\frac{1}{2}}$

In the formula: Kurtosis is a spectrogram kurtosis result of calculation, and X (k) is the Fourier transform of detection signal, and k is the horizontal ordinate variable of Fourier transform, N ₁Be the corresponding Fourier transform point horizontal ordinate of linear frequency modulation initial frequency, N ₂For linear frequency modulation stops the corresponding Fourier transform point horizontal ordinate of frequency.

is the average of X (k), and SD is the standard variance of X (k).In the present embodiment, calculate the spectrogram kurtosis of this benchmark response signal frequency spectrum, thereby when obtaining the test specimen zero defect; The i.e. spectrogram kurtosis 22 of this benchmark response signal frequency spectrum, shown in figure 12, horizontal ordinate is represented defective; No unit, ordinate represent the spectrogram kurtosis, no unit.

The spectrogram degree of bias does

$\mathrm{Skenwness}=\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}{[X\left(k\right)-\stackrel{\&OverBar;}{X}]}^3/{\mathrm{<figure-callout\; id="3"\; label="SD"\; filenames="HDA0000028150910000053.png,HDA0000028150910000072.png"\; state="\{\{state\}\}">SD</figure-callout>}}^3$

Wherein

$\stackrel{\&OverBar;}{X}=\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}X\left(k\right),$ $\mathrm{SD}={\left(\frac{1}{N_2-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="HDA0000028150910000012.png,HDA0000028150910000041.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{k=N_1}{\overset{N_2}{\mathrm{\&Sigma;}}}{[X\left(k\right)-\stackrel{\&OverBar;}{X}]}^2\right)}^{\frac{1}{2}}$

In the formula: Skewness is a spectrogram degree of bias result of calculation, and X (k) is the Fourier transform of detection signal, and k is the horizontal ordinate variable of Fourier transform, N ₁Be the corresponding Fourier transform point horizontal ordinate of linear frequency modulation initial frequency, N ₂For linear frequency modulation stops the corresponding Fourier transform point horizontal ordinate of frequency.

is the average of X (k), and SD is the standard variance of X (k).In the present embodiment, calculate the spectrogram degree of bias of this benchmark response signal frequency spectrum, thereby when obtaining the test specimen zero defect; The i.e. spectrogram degree of bias 25 of this benchmark response signal frequency spectrum, shown in figure 13, horizontal ordinate is represented defective; No unit, ordinate represent the spectrogram degree of bias, no unit.

The 5th step, adopt aforesaid fm exciter signal, and with this fm exciter signal like the second said amplification of step after, drive aforementioned eddy current sensor, checked object is detected.In the present embodiment, as shown in Figure 7, checked object still adopts aluminum dull and stereotyped 1; But in the middle of aluminum dull and stereotyped 1, be uniformly distributed with the defective 2,3,4 of three kinds of different depths, wherein, defective 2 is of a size of 10mm * 1.0mm * 0.5mm; Defective 3 is of a size of 10mm * 1.0mm * 1.0mm, and defective 4 is of a size of 10mm * 1.0mm * 1.5mm, and depth of defect increases successively; And agreement: when dull and stereotyped 1 upper surface of aluminum detected, defective was all represented surface imperfection; When dull and stereotyped 1 lower surface of aluminum detected, defective was all represented inherent vice.

The 6th step, the response signal of eddy current sensor during the acquisition testing checked object, and this eddy current sensor response signal is carried out the FFT conversion obtain checked object eddy current sensor response signal frequency spectrum.In the present embodiment, respectively defective 2,3,4 is detected, eddy current sensor response signal when gathering each the detection respectively, and those eddy current sensor response signals are carried out the FFT conversion obtain corresponding response signal frequency spectrum.As in this enforcement, when the checked object upper surface is detected, the response signal frequency spectrum 10 of defective 2, the response signal frequency spectrum 11 of defective 3, the response signal frequency spectrum 12 of defective 4 have been obtained; As shown in Figure 8, horizontal ordinate is represented frequency, and unit is kHz; Ordinate is represented the FFT amplitude, no unit.When the checked object lower surface is detected, obtain the response signal frequency spectrum 13 of defective 2, the response signal frequency spectrum 14 of defective 3, the response signal frequency spectrum 15 of defective 4, as shown in Figure 9; Horizontal ordinate is represented frequency; Unit is kHz, and ordinate is represented the FFT amplitude, no unit.

The 7th goes on foot, checked object eddy current sensor response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this checked object eddy current sensor response signal frequency spectrum.Each characteristic quantity computing method is identical with aforementioned the 4th step.

Checked object surface imperfection and the spectrum analysis of inherent vice eddy current sensor response signal through the 6th step was obtained calculate: 1, the spectrogram gross energy 18 of the eddy current sensor response signal frequency spectrum of the spectrogram gross energy 17 of each surface imperfection response signal frequency spectrum and each inherent vice; Shown in figure 10; Horizontal ordinate is represented defective; No unit, ordinate represent the spectrogram gross energy, no unit; 2, the eddy current sensor response signal spectrogram centre-of gravity shift 21 of the eddy current sensor response signal spectrogram centre-of gravity shift 20 of each surface imperfection and inherent vice, shown in figure 11, horizontal ordinate is represented defective, no unit, ordinate represent the spectrogram barycentre offset, no unit; 3, the eddy current sensor response signal spectrogram kurtosis 23 of the eddy current sensor response signal spectrogram kurtosis 24 of each surface imperfection and inherent vice, shown in figure 12, horizontal ordinate is represented defective, no unit, ordinate represent the spectrogram kurtosis, no unit; 4, the eddy current sensor response signal spectrogram degree of bias 27 of the eddy current sensor response signal spectrogram degree of bias 26 of surface imperfection and inherent vice, shown in figure 12, horizontal ordinate is represented defective, no unit, ordinate represent the spectrogram degree of bias, no unit.

The 8th step, the spectrogram gross energy according to benchmark response signal frequency spectrum, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias are carried out threshold decision to spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of above-mentioned detected checked object eddy current sensor response signal frequency spectrum: if the spectrogram gross energy of detected checked object eddy current sensor response signal frequency spectrum greater than the spectrogram gross energy of benchmark response signal frequency spectrum, then proves to have defective on the checked object; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that greater than the spectrogram degree of bias of benchmark response signal frequency spectrum the defective on the checked object is a surface imperfection; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that less than the spectrogram degree of bias of benchmark response signal frequency spectrum the defective on the checked object is an inherent vice.As can beappreciated from fig. 10; The spectrogram gross energy 18 of the eddy current sensor response signal frequency spectrum of the spectrogram gross energy 17 of each surface imperfection response signal frequency spectrum and each inherent vice all greater than the spectrogram gross energy 16 of benchmark response signal frequency spectrum, just can judge whether to exist defective through the spectrogram gross energy that detection obtains thus.As can beappreciated from fig. 11; The spectrogram centre-of gravity shift 20 of each surface imperfection response signal frequency spectrum is positioned on the spectrogram centre-of gravity shift 19 of benchmark response signal frequency spectrum; The spectrogram centre-of gravity shift 21 of each inherent vice response signal frequency spectrum is positioned under the spectrogram centre-of gravity shift 19 of benchmark response signal frequency spectrum, and the spectrogram centre-of gravity shift that obtains through detection computations thus just can be judged the residing diverse location of defective.As can beappreciated from fig. 12; The spectrogram kurtosis 24 of each surface imperfection response signal frequency spectrum is positioned under the spectrogram kurtosis 22 of benchmark response signal frequency spectrum; The spectrogram kurtosis 23 of each inherent vice response signal frequency spectrum is positioned on the spectrogram kurtosis 22 of benchmark response signal frequency spectrum, and the spectrogram kurtosis that obtains through detection computations thus just can be judged the residing diverse location of defective.As can beappreciated from fig. 13; The spectrogram degree of bias 26 of each surface imperfection response signal frequency spectrum is positioned on the spectrogram degree of bias 25 of benchmark response signal frequency spectrum; The spectrogram degree of bias 27 of each inherent vice response signal frequency spectrum is positioned under the spectrogram degree of bias 25 of benchmark response signal frequency spectrum, and the spectrogram degree of bias that obtains through detection computations thus just can be judged the residing diverse location of defective.

The foregoing description only is used for further specifying the present invention and modulates multifrequency Eddy detection method and application; But the present invention is not limited to embodiment; Every foundation technical spirit of the present invention all falls in the protection domain of technical scheme of the present invention any simple modification, equivalent variations and modification that above embodiment did.

Claims (4)

1. a modulation multifrequency Eddy detection method is characterized in that comprising the following steps:

The first step, preparation benchmark test specimen;

Second step, generation one fm exciter signal, and, drive eddy current sensor with after this fm exciter signal amplification, the benchmark test specimen is detected;

The 3rd step, the response signal of eddy current sensor during acquisition testing benchmark test specimen, and this eddy current sensor response signal is carried out the FFT conversion obtain benchmark test specimen eddy current sensor response signal frequency spectrum, promptly obtain detecting required benchmark response signal frequency spectrum;

The 4th goes on foot, benchmark response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this benchmark response signal frequency spectrum;

The 5th step, adopt aforesaid fm exciter signal, and with this fm exciter signal like the second said amplification of step after, drive aforementioned eddy current sensor, checked object is detected;

The 6th step, the response signal of eddy current sensor during the acquisition testing checked object, and this eddy current sensor response signal is carried out the FFT conversion obtain checked object eddy current sensor response signal frequency spectrum;

The 7th goes on foot, checked object eddy current sensor response signal frequency spectrum is analyzed, and calculates spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of this checked object eddy current sensor response signal frequency spectrum;

The 8th step, the spectrogram gross energy according to benchmark response signal frequency spectrum, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias are carried out threshold decision to spectrogram gross energy, spectrogram centre-of gravity shift, spectrogram kurtosis, the spectrogram degree of bias of above-mentioned detected checked object eddy current sensor response signal frequency spectrum: if the spectrogram gross energy of detected checked object eddy current sensor response signal frequency spectrum greater than the spectrogram gross energy of benchmark response signal frequency spectrum, then proves to have defective on the checked object; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that greater than the spectrogram degree of bias of benchmark response signal frequency spectrum the defective on the checked object is a surface imperfection; If the spectrogram centre-of gravity shift of detected checked object eddy current sensor response signal frequency spectrum is less than the spectrogram centre-of gravity shift of benchmark response signal frequency spectrum; The spectrogram kurtosis of perhaps detected checked object eddy current sensor response signal frequency spectrum is greater than the spectrogram kurtosis of benchmark response signal frequency spectrum; The spectrogram degree of bias of perhaps detected checked object eddy current sensor response signal frequency spectrum proves then that less than the spectrogram degree of bias of benchmark response signal frequency spectrum the defective on the checked object is an inherent vice.

2. modulation multifrequency Eddy detection method according to claim 1 is characterized in that this fm exciter signal is the linear frequency modulation pumping signal.

3. modulation multifrequency Eddy detection method according to claim 2 is characterized in that the optimized frequency scope of this linear frequency modulation pumping signal is 1kHz～20kHz, and preferred time span is 4096us.

4. modulation multifrequency Eddy detection method according to claim 1 is characterized in that, the SF of gathering the eddy current sensor response signal is 5-10 a times of fm exciter signal highest frequency.

Images