JP2007225303A - Apparatus and method for synchronous detection, and method and apparatus for measuring magnetism - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a synchronous detection method and apparatus for performing A/D conversion prior to synchronous detection even in the case that a primary frequency is high and performing synchronous detection at low costs and provide a magnetic measuring method and apparatus. <P>SOLUTION: In the synchronous detection method for converting signals to be detected from analog to digital form and performing synchronous detection by digital signal processing, the sampling frequency on the A/D conversion is set lower than twice the frequency of a reference signal and equal to a maximum frequency of signals after synchronous detection or higher. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a synchronous detection method and apparatus, a magnetic measurement method and apparatus, and more particularly to a synchronous detection method and apparatus for performing synchronous detection by digital signal processing, and a magnetic measurement method and apparatus.

  The eddy current measurement method, the AC leakage magnetic flux flaw detection method, and the AC magnetoresistance measurement method are measurement methods in which an AC magnetic field is applied to the measurement target from the outside and a response in the measurement target is measured. Here, since the response of the measurement target is obtained as a signal having the same frequency as the AC magnetic field applied from the outside, a method such as synchronous detection is used to detect the response of the measurement target without being affected by the noise signal from the outside. Be taken.

  Since this synchronous detection needs to perform signal processing with very high speed and high accuracy, it has been generally performed by a synchronous detection circuit constituted by an analog circuit so far.

However, when performing eddy current measurement, AC leakage magnetic flux flaw detection, or AC magnetic measurement method that requires the use of multiple sensors at the same time, the same number of synchronous detection circuits are required as the number of sensors, or proportional to the number of sensors. Therefore, there are problems such as an increase in circuit scale and cost. Patent Document 1 discloses a technique for performing synchronous detection using a digital signal.
JP 59-90044 A

  A method of performing A / D conversion on the measurement result to which the technique disclosed in Patent Literature 1 is applied before synchronous detection and processing the synchronous detection on software is conceivable. If this method is used, since it can be processed on software, synchronous detection can be realized on a PC, for example, with a simple program.

  However, in order to realize this, it is necessary to sufficiently increase the sampling frequency when the measured signal is digitized by A / D conversion. According to the sampling theorem, the maximum frequency of the measured signal (in the case of measurement with synchronous detection such as eddy current measurement, AC leakage magnetic flux flaw detection, AC magnetic measurement, the frequency of the primary signal (excitation signal) applied from the outside is measured. It is desirable to sample at a frequency greater than twice the main frequency component of the signal. Therefore, in measurements that perform synchronous detection such as eddy current measurement, AC leakage magnetic flux flaw detection, and AC magnetic measurement, particularly when the frequency of the primary signal is high, sampling at a very high frequency is required.

  As for detection resolution, the signal change (for example, signal intensity change due to a defect) in the secondary signal (detection signal detected by the detection sensor) is significantly different from the signal change of the primary signal. Since A / D conversion with high resolution is required in a small and wide measurement range, an A / D converter with a high bit number is required.

  The present inventors have conducted eddy current flaw detection measurement of defects in the steel sheet, and confirmed that the minimum number of bits required for this realization is about 16 bits, and more preferably about 20 bits or more. In order to realize such a high bit number A / D conversion, the ΣΔ conversion method is usually used, but it is difficult to increase the sampling frequency in this method. Conversely, in order to realize A / D conversion with a high sampling frequency, a so-called flash type conversion method is generally used. However, since a conversion element is required for each bit, the number of modules increases because the number of modules increases. There is a problem that is not suitable for A / D conversion. Under such circumstances, it is difficult to realize A / D conversion with a high number of bits and high speed (several hundred kHz or more).

  The present invention has been made in view of the above circumstances, and a synchronous detection method and apparatus for performing synchronous detection at low cost by performing A / D conversion before synchronous detection even when the primary frequency is high, a magnetic measurement method, and An object is to provide an apparatus.

  The invention according to claim 1 of the present invention is a synchronous detection method for performing synchronous detection by digital signal processing after A / D conversion of a signal to be detected, and a sampling frequency at the time of the A / D conversion is set as a reference signal. The synchronous detection method is characterized in that it is smaller than twice the frequency of the signal and is at least twice the maximum frequency of the signal after synchronous detection.

  The invention according to claim 2 of the present invention is a magnetic measurement method using the synchronous detection method according to claim 1, wherein the reference signal is an AC excitation signal, and the signal to be detected is a detection signal. This is a magnetic measurement method characterized by performing synchronous detection.

  According to a third aspect of the present invention, there is provided A / D conversion means for inputting a signal to be detected, and synchronous detection for performing synchronous detection based on a digital signal output from the A / D conversion means. In the apparatus, the sampling frequency of the A / D conversion means is set to be smaller than twice the frequency of the reference signal and at least twice the maximum frequency of the signal after synchronous detection. Device.

  Furthermore, the invention according to claim 4 of the present invention is a magnetic measurement device comprising the synchronous detection device according to claim 3, wherein the reference signal is an AC excitation signal and the signal to be detected is a detection signal. This is a magnetic measuring device.

  The “maximum frequency of the signal after synchronous detection” in the above claims is the highest value of the frequency of the intensity change signal that is calculated in advance from the line speed and the size of the detection target (for example, the size of the defect). It is. For example, the smaller the detection target, the higher the line speed, the higher the maximum frequency.

  In the present invention, even when the excitation frequency of the excitation signal as a reference signal is high, the A / D conversion is synchronously detected at a frequency lower than twice the excitation frequency and at least twice the maximum frequency of the signal after synchronous detection. Since it was performed before, it was possible to perform synchronous detection at low cost, and it was also possible to perform AC magnetic measurements such as eddy current measurement, AC leakage magnetic flux flaw detection, and AC magnetoresistance measurement.

  Hereinafter, as an example of the synchronous detection method and the measurement method using AC magnetism, an example applied to eddy current flaw detection of a steel sheet and the synchronous detection will be shown. FIG. 1 is a diagram showing an apparatus configuration example as the best mode for carrying out the present invention. In the figure, 1 is a steel plate, 2 is a defect, 3 is an AC power supply, 4 is an excitation coil, 5 is a detection coil, 6 is an iron core, 7 is an amplifier, 8a and 8b are A / D converters, and 9 is a signal processing device. , 10 represents a synchronous detection processing unit, and 11 represents a defect determination unit.

  The steel plate 1 has a defect 2. An alternating current (excitation signal) is supplied to the exciting coil 4 from the alternating current power source 3 at a frequency fa. In the exciting coil, an alternating magnetic flux is generated by an alternating current. An eddy current is generated in the steel plate 1 by an electromagnetic induction phenomenon. The magnetic flux generated by the eddy current is detected by the detection coil 5 by the electromagnetic induction phenomenon. Here, if the defect 2 exists in the steel plate 1, the defect 2 causes a change in the eddy current generated in the steel plate 1. By detecting a change in the healthy portion of the eddy current caused by the presence of the defect in the detection coil 5, the defect can be detected.

  Here, the detection signal detected by the detection coil 5 is amplified by the amplifier 7 and is less than 1/2 of the frequency fa of the excitation signal given from the AC power supply 3 to the excitation coil 4 by the A / D converter 8a. It is sampled and converted into a digital signal by fs (fs <(fa / 2)). Similarly, the excitation signal is A / D converted by the A / D converter 8b and converted into a digital signal. These two digital signals are sent to the signal processing device 9, and the synchronous detection processing unit 10 performs synchronous detection using the detection signal detected by the detection coil 5 as a secondary signal and the excitation signal as a primary signal ( The phase may be out of phase). Thus, in the obtained signal after synchronous detection, when a defect exists in the steel plate, a signal corresponding to the size of the defect is obtained. This signal is sent to the defect determiner 9, and a signal exceeding a predetermined output is determined as a defect. Here, the signal after the synchronous detection is a signal from which unnecessary high-frequency components are removed by inserting an LPF (low pass filter) after the synchronous detection.

  According to the sampling theorem, if the maximum frequency included in the detection signal before A / D conversion is fmax, the sampling frequency fs needs to be fs> 2 × fmax in order to store the frequency information. In the eddy current flaw detection shown in the above example, a signal having the same frequency component as the excitation signal (usually given as a sine wave) is obtained as a detection signal. FIG. 2 schematically shows the difference between the detection signal obtained by eddy current flaw detection and the sampling frequency. In FIG. 2 (A), a signal change of 0.125 kHz due to a defect is added to the frequency fa = 3 kHz of the excitation signal for eddy current flaw detection. However, the defect exists on the time axis of 3 msec in FIG. In normal eddy current flaw detection, the synchronous detection of this signal and the original excitation signal is performed by an analog circuit, and the result is filtered to emphasize only the defect signal component.

  Here, consider a case where the signal of FIG. 2 is A / D converted before synchronous detection and synchronous detection is performed by digital signal processing. In this case, according to the sampling theorem, a sampling frequency greater than twice fmax is required, so a sampling frequency that is twice or more the excitation signal frequency fa = 3 kHz is required. The one actually sampled at 6 kHz, which is twice the frequency 3 kHz of the excitation signal, which is the maximum frequency included in the signal of FIG. 2, is shown by the broken line in FIG. As can be seen from the figure, when the object is a sine wave, if the sampling is performed twice as much as the sine wave, only the same phase component is sampled and the original signal cannot be reproduced. Fig. 2 (C) shows an example (dashed line) of sampling at 12kHz, which is four times the maximum frequency (greater than twice), but it takes four sampling points for one sinusoidal wave. And a sine wave signal can be reproduced. Thus, the original signal can be reproduced at a frequency larger than twice. Usually, sampling is performed at a frequency of 4 to 5 times the maximum frequency (more preferably 10 times or more), and 4 to 5 sampling points are taken for one sinusoidal wave (10 points or more are required). More preferred).

  In the case of eddy current flaw detection, the excitation frequency, which is the frequency of the excitation signal, is determined by the penetration depth of the eddy current, but if it is desired to reduce the penetration depth, it is necessary to increase the excitation frequency. In eddy current flaw detection such as a steel plate, an excitation frequency of several hundred kHz may be used. For this reason, the excitation frequency may be increased, but if the sampling frequency is set to about 10 times the excitation frequency as described above, for example, when the excitation frequency is several hundred kHz, sampling of several MHz is performed. A frequency is required, and it is very difficult to realize such A / D conversion.

  FIG. 3 is a diagram schematically showing a detection signal of such eddy current flaw detection that performs excitation at a frequency as high as several hundred kHz. While excitation at a high frequency makes A / D conversion difficult, the excitation frequency is very high, for example, about 100 to 100,000 times compared to the frequency of the defect signal. is there. For example, in the example of FIG. 3, the period of the excitation signal is several μsec (frequency is several hundred kHz to 1 MHz), while the period of the defect signal (time range in which the signal intensity changes due to the defect) is several milliseconds (frequency is several hundreds). Hz to 1 kHz), which is a very high excitation frequency of about 1000 times.

  For example, in a steel line, the relative speed of the object to be measured and the sensor is several m / sec to 40 m / sec. Spatial spread of the defect signal (a region where the magnetic field is disturbed and changed by the defect, not necessarily the same as the defect size. Not) is about several millimeters to several tens of millimeters. In this case, the frequency of the defect signal (hereinafter referred to as the defect frequency) is about several tens of kHz to several tens of Hz. For example, when the defect excitation frequency is set to several hundreds KHz or more, the excitation frequency is higher than the defect frequency (several tens of kHz to several tens of Hz). In particular, when the frequency of the defect signal is several kHz or less and the excitation frequency is several hundred kHz or more, the excitation frequency is significantly higher than the defect frequency. In this way, when the excitation frequency is (significantly) higher than the defect frequency (when it is desired to reduce the penetration depth), the change in the defect signal is (significantly) slower than the change in the original excitation signal. Excitation signal information is redundant.

  In addition, in such an example of flaw detection, a signal (defect signal) from a measurement object is usually not periodic. The equivalent sampling technique found in digital oscilloscopes and the like is similar in that the sampling frequency is slower than the signal frequency. However, in a digital oscilloscope, a signal to be measured has a periodicity, and a sampled signal is reproduced using the periodicity. The signal to be measured handled by the present invention has no periodicity, and is different from the technique of the digital oscilloscope in that it uses the fact that the signal to be measured is slower than the signal to be synchronously detected.

  When performing synchronous detection, signal processing is performed by synchronizing the excitation signal that is the primary signal and the detection signal that is the secondary signal, but at this time, the secondary signal is converted into the excitation signal that is the primary signal, In this state, a defect signal is applied. At this time, since the frequencies of the excitation signal and the defect signal are greatly different, using this fact, only the signal component corresponding to the excitation signal is aliased, and the signal component corresponding to the defect signal is not aliased. If the secondary signal is sampled at a high frequency, the defect signal component included in the secondary signal is not affected, and the excitation signal component included in the secondary signal can be reduced in a pseudo manner. As a result, the low-frequency component excluding the high-frequency component can be obtained, and the same effect as the signal processing by thinning out the secondary signal can be obtained. This process also facilitates the synchronous detection process.

  Therefore, even if the sampling frequency in A / D conversion is less than twice the excitation frequency, if you select an appropriate value, it can be converted to a digital signal without damaging the information of the detection signal (defect signal), and the same level as before Eddy current testing can be realized. Here, in order not to impair the information that is finally obtained, a sampling frequency that is at least twice the maximum signal change frequency indicated by the target for detection and measurement is required at a minimum. In order to obtain the above, a sampling frequency of 5 times or more of the maximum frequency is desirable, and a sampling frequency of 10 times or more is more desirable.

  FIGS. 4a and 4b are examples in which sampling is performed at a low sampling rate (900 kHz, 480 kHz) with respect to a high frequency (1 MHz) signal (original sine wave signal). 4A and 4B, the original sine wave signal is connected by a solid line, sampling points are circled, and sampling points are connected by a broken line. It can be seen from the figure that the curve connected by the broken line constitutes a sine wave having a lower frequency than the original sine wave by aliasing.

  In this way, even if sampling is performed at a frequency lower than the excitation frequency, fb = fa-fs × n (n is a positive integer and fb is a positive and minimum value due to aliasing, fa is the excitation frequency, and fs is the sampling frequency. A sine wave with a frequency) is obtained. The digitized detection signal obtained by digitization in this way is synchronously detected with the reference signal shown below. The reference signal is a digital signal obtained by A / D converting the excitation signal with the same sampling period as the detection signal, or by synchronous detection with a sine wave with the same period as fb (made by a sine wave oscillator, function generator, etc.) The same eddy current flaw detection as before is possible without impairing the defect information. At this time, the frequency of the signal to be synchronously detected can be lowered, and the signal processing of the synchronous detection is easy.

Next, how to select the sampling frequency fs will be described with reference to the following equations and figures. First, the following relational expression is assumed.
fb = fa- (fs × n) (n is a positive integer and fb is positive and minimum) …… (1)
Ta = 1 / fa Ts = 1 / fs Tb = 1 / fb (2)
The phase angle α is as follows.
α = Ts / Tb * 360 (3)
n means that at the time of sampling, the next sampling point is on the nth wave in the original sine wave, and α is a sine wave formed by aliasing (in the following description, This is equivalent to what is expressed as a phase in an aliasing waveform).

  FIG. 5 is a diagram illustrating a sampling state and a sine wave formed by aliasing when the sampling frequency is changed and the phase angle α is changed. According to this figure, it can also be seen that when the sampling frequency is changed, the frequency of the sine wave formed by aliasing also changes in accordance with the above formula. If you change the sampling frequency and select a value that satisfies 0 <α ≤ 120 as shown in Fig. 5 (a) (α = 101), you can take four or more sampling points for one sine wave. Although some of the sine wave information is missing, information such as the frequency of the aliasing waveform (= 220 kHz) can be obtained, and it can also be confirmed that it is a periodic signal. As shown in Fig. 5 (b), if you select a value that satisfies 0 <α ≤ 72 (α = 72), you can take 5 or more sampling points for one sine wave. You can get more.

  If a value satisfying 0 <α ≦ 50 is selected as shown in FIG. 5 (c) (α = 50), more sampling points can be taken, and most of the sine wave information can be obtained. It can be seen that the aliasing waveform reproduced by sampling is also a sine wave. If a value satisfying 0 <α ≦ 36 is selected (α = 36), 10 or more sampling points can be taken for one sine wave, and more sine wave information can be obtained (Fig. 5 (d)).

  In order to represent sine wave information more accurately, it is better that θ is smaller, but in fact, it can be confirmed that it is a periodic signal by selecting a value that satisfies 0 <α ≦ 120 as seen in FIG. Eddy current measurement and leakage flux flaw detection are possible, and by selecting a value that satisfies 0 <α ≦ 72, eddy current measurement and leakage flux flaw detection can be performed with high accuracy, and by selecting a value that satisfies 0 <α ≦ 50 Eddy current measurement and magnetic flux flaw detection become possible, and by selecting a value satisfying 0 <α ≦ 50, highly accurate eddy current measurement and magnetic flux flaw detection become possible. The smaller α is, the higher the accuracy is, but the more sampling points are required. Since the amount of signal processing increases as the number of sampling points increases, it is appropriate to select a value of α at which the sampling points are about 7 to 10.

  In Fig. 5 (e), the sampling frequency at which α = 45 is selected. When the sampling frequency at which α = 360 / m (m = 1,2,3,...) Is selected in this way, The peak point of the aliasing waveform can be taken as the sampling point. The peak point of the sine wave is the maximum point (or the minimum point) and the slope is 0, and it has important information, but α = 360 / m (m = 1, 2, 3, ...) By selecting a suitable sampling frequency, it becomes possible to obtain peak point information and improve the accuracy of eddy current measurement and leakage magnetic flux flaw detection.

  Further, in eddy current flaw detection, information on the phase difference of the detection signal with respect to the excitation signal is important, and will be described below.

  FIG. 6 is an enlarged view of the waveform of FIG. 4 in the time axis direction. When ta in the figure is used, the phase angle of the original sine wave at the sampling point X is θa = ta / Ta (Ta is the period of the original sine wave). In addition, when the phase angle θb in the sine wave formed by the aliasing phenomenon at the sampling point X is expressed using tb in the figure, θb = tb / Tb (Tb is the period of the sine wave formed by the aliasing phenomenon) and Become. At this time, the phase angles in the respective waveforms coincide and θa = θb. Therefore, the phase change in the original sine wave appears as a phase change of the sine wave formed by aliasing. Therefore, if the number of sampling points is sufficiently secured and the original sine wave information is accurately left, even if sampling is performed at a frequency lower than the excitation frequency, the phase difference information of the detection signal of the eddy current flaw detection can be obtained. is there. (If the number of sampling points is too small, there is not enough sine wave information left, and phase difference information is not sufficient).

  FIG. 7 shows an example of sampling with the phase difference changed. Fig. 7 (a) shows the case where the phase difference between the digitized detection signal (shown by the broken line) and the excitation signal (shown by the solid line) is 0 degrees. Fig. 7 (b) shows the phase of the detection signal and the excitation. This is when the signal phase difference is 90 degrees. 7 (a) and 7 (b), it can be seen that the phase difference between the sine waves of the solid line and the broken line is 90 degrees. Therefore, the excitation signal is sampled at the same sampling period as the detection signal, and the same phase difference as the excitation signal and the detection signal is obtained by performing synchronous detection with the sampled waveform of the detection signal using that signal as a reference signal. Can be obtained.

  In the above, the excitation signal is sampled at the same sampling cycle as that of the detection signal to form the reference signal. However, if a periodic wave having the same cycle as the sine wave formed by sampling the detection signal is used as the reference signal, the excitation signal And the relative change in the phase difference between the detection signals can be recognized, and the same effect can be obtained. The periodic wave with the same period as the sine wave formed by sampling here is obtained by sampling the excitation signal at a different sampling period from the detection signal and then changing the frequency by signal processing such as thinning out the data However, a specified periodic wave may be generated digitally by software, or a method of generating a signal from a separately provided signal generator and performing A / D conversion may be used. In addition, a sine wave need not be used if it is a periodic wave.

  Also, in the above, when the excitation signal is sampled at the same sampling cycle as the detection signal to form the reference signal, the phase information is saved if the sampling is performed in synchronization with the excitation signal sampling and the detection signal sampling. Easy and better. This shows an example in which the phase information that was originally in the sampling method of the present invention can be stored.

  FIG. 8 is a diagram showing the effect of interpolation. The waveform of the original sine wave (solid line) when the signal frequency is 1 MHz and the sampling frequency is 780 kHz (broken line) and the waveform interpolated by the following equation (4) (dotted line).

Here, x (nΔt) represents a signal obtained by sampling. The sampling indicated by the broken line in FIG. 8 is obtained by sampling a waveform having the same frequency as that shown in FIG. 5 (a) at the same sampling frequency. Although it is not sufficient, it is a waveform that can be seen to be a sine wave. Interpolation like the above equation (equivalent other interpolation equations (least square interpolation, spline interpolation, etc. may be used)) can obtain sine wave information as aliasing waveform and increase the number of sampling points The effect equivalent to that is obtained.

  In this way, interpolation increases the amount of calculation in signal processing for synchronous detection compared to when synchronous detection is performed without interpolation, but sampling (A / D conversion) is easy and accurate. An effect that enables high measurement is obtained. This shows an example in which a necessary sine wave can be reconstructed by interpolation even if the number of sampling points is small.

  In this example, the eddy current flaw detection method, which is an example of the eddy current measurement method using synchronous detection, has been described. However, the same effect can be obtained by using other eddy current measurement methods such as distance measurement. The same effect can be obtained even if the method is applied to the AC leakage magnetic flux flaw detection method and the AC magnetoresistance measurement method.

  In addition, among the eddy current flaw detection methods, a single coil method using a mutual induction type coil is used, but a self-inductive type coil may be used, and the coil method may be a self-comparison method or a standard comparison method. It goes without saying that either method may be used.

It is a figure which shows the example of an apparatus structure as the best form for implementing this invention. It is the figure which represented typically the difference in the detection signal and sampling frequency obtained by eddy current flaw detection. It is the figure which showed typically the detection signal of the eddy current flaw which performs excitation at a high frequency. It is a figure which shows the example which implemented the sampling with low sampling speed (900kHz, 480kHz) with respect to the signal of a high frequency (1MHz). It is a figure which shows the sine wave formed by aliasing at the time of changing sampling angle and changing the phase angle (alpha), and the state of sampling. 4 is an enlarged view of the waveform of FIG. 4 in the time axis direction. It is a figure which shows the example of a sampling which changed the phase difference. It is a figure showing the effect of interpolation.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Steel plate 2 Defect 3 AC power supply 4 Excitation coil 5 Detection coil 6 Iron core 7 Amplifier 8a, 8b A / D converter 9 Signal processing apparatus 10 Synchronous detection processing part 11 Defect determination part a Excitation signal b Reference signal, Primary signal c Detection signal, secondary signal, signal to be detected d digitized reference signal e digitized detection signal f signal after synchronous detection

Claims (4)

In the synchronous detection method of performing synchronous detection by digital signal processing after A / D converting the signal to be detected,
A synchronous detection method characterized in that a sampling frequency at the time of the A / D conversion is smaller than twice the frequency of the reference signal and at least twice the highest frequency of the signal after synchronous detection. A magnetic measurement method using the synchronous detection method according to claim 1,
The reference signal is an AC excitation signal, the detection target signal is a detection signal,
A magnetic measurement method characterized by performing synchronous detection. In a synchronous detection device that includes an A / D conversion unit that receives a signal to be detected, and performs synchronous detection based on a digital signal output from the A / D conversion unit,
A synchronous detection apparatus characterized in that the sampling frequency of the A / D conversion means is set to be smaller than twice the frequency of the reference signal and at least twice the maximum frequency of the signal after synchronous detection. A magnetic measurement device comprising the synchronous detection device according to claim 3,
A magnetic measuring device, wherein the reference signal is an AC excitation signal, and the signal to be detected is a detection signal.