JPH05264517A - Method and device for monitoring damage of aircraft structural member - Google Patents

Abstract

PURPOSE:To discriminate AE signal for detecting a damage which is generated and can be developed in a structural member while an aircraft is flying from a noise signal. CONSTITUTION:An AE signal and a noise signal which are radiated from a structural member of an aircraft are received by a sensor 1, a low-frequency noise is eliminated from them by a high-pass filter 2, and then they are amplified by a main amplifier 4 and are converted to digital signal by a transient recorder 5, thus storing them as a time-series signal waveform data. An analysis device 5 obtains various kinds of parameters (rising time, life time, spectrum band, first order moment, and second order moment) and then judges as the AE signal when each parameter is within a specified range and as the noise signal when it is outside the specified range.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for monitoring the damage that a structural member develops and propagates while an aircraft is flying by using acoustic radiation (hereinafter referred to as AE).

[0002]

2. Description of the Related Art Researches using an AE to detect damages occurring and developing in structural members of an aircraft in flight have been conducted since the mid-1970s. However, in flight aircraft, mechanical friction and electrical noise generated by air friction, friction between structural members, and hydraulic system are generated. Therefore, when the sensor is used to detect the AE signal, these noise signals are mixed and hinder the monitoring of damage.

Therefore, a technique for discriminating between an AE signal to be detected and a noise signal has been proposed. For example, Japanese Patent Laid-Open No. 63-243753 discloses a technique for removing a noise signal by utilizing the fact that an AE signal has a peak value in a specific frequency region.

[0004]

However, a wide variety of noises are generated in a flying aircraft. Therefore, among the noise signals mixed in the AE signal, there is also one having a spectrum distribution characteristic very similar to that of the AE signal. Therefore, AE
It was not possible to sufficiently discriminate between the AE signal and the noise signal only by utilizing the fact that the signal has the peak value in the specific frequency region.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a method and an apparatus for monitoring damage to aircraft structural members, which can accurately distinguish between an AE signal and a noise signal.

[0006]

SUMMARY OF THE INVENTION An aircraft structural member damage monitoring method of the present invention eliminates low frequency noise signals contained in a signal received by a sensor for receiving acoustic radiation provided in an aircraft structural member and digitally Converting to a signal and storing as time-series signal waveform data; using the time-series signal waveform data, a rising time from when the signal exceeds a threshold to reaching a maximum amplitude, and when the signal is Determining a duration from the time when the threshold is exceeded to the time when the threshold is attenuated to a threshold or less; a step of performing a fast Fourier transform using the time-series signal waveform data to obtain a spectrum distribution; Dividing the frequency band into two frequency bands and obtaining the ratio of the area of each of the divided frequency bands to the area of the entire spectrum distribution; Determining a first-order moment of the mean value of the spectrum distribution, 2 of the variance of the spectral distribution
The step of obtaining the second moment, the obtained rise time, duration, the ratio of the area of each divided frequency band to the entire area of the spectrum distribution, the values of the first moment, and the second moment are calculated in advance. A step of discriminating a signal included in a predetermined range as an AE signal and a signal not included in the predetermined range as a noise signal.

Further, the aircraft structural member damage monitoring apparatus of the present invention includes a sensor provided on the structural member of the aircraft for receiving an acoustic signal, and a filter for removing a low frequency noise signal from the signal received by the sensor. An A / D converter for converting the signal from which the low frequency noise signal has been removed by the filter into a digital signal, a means for storing the converted digital signal as time series signal waveform data, and the time series signal waveform data And means for determining the rise time from the time the signal exceeds the threshold to the maximum amplitude and the duration from the time the signal exceeds the threshold to the time below the threshold is attenuated. , Means for obtaining a spectral distribution by performing a fast Fourier transform using the time-series signal waveform data, and the spectral distribution having at least two frequencies. A means for obtaining the ratio of the area of each of the divided frequency bands and the area of the entire spectrum distribution, a means for obtaining the first moment which is the average value of the spectrum distribution, and the dispersion of the spectrum distribution. The means for obtaining the second moment, the obtained rise time, duration, the ratio of the area of each divided frequency band to the entire area of the spectrum distribution, the values of the first moment, and the second moment are respectively It is characterized in that it is provided with means for discriminating a signal included in a predetermined range as an AE signal and a signal not included as a noise signal.

[0008]

The signal radiated from the structural member of the aircraft is received, the low frequency noise signal is removed, the signal is converted into a digital signal and stored as time series signal waveform data. Using this time-series signal waveform data, the rise time from when the signal exceeds the threshold to when it falls below the threshold, and from the time when the signal crosses the threshold to below the threshold Duration is required. Further, fast Fourier transform is performed on the time-series signal waveform data to obtain the spectrum distribution, and the ratio of the area of each divided frequency band to the area of the entire spectrum distribution is obtained. Further, the first moment, which is the average value of the spectral distribution, and the second moment, which is the variance of the spectral distribution, are obtained. Among the obtained parameters, the AE signal is determined to be within the predetermined range, and the noise signal is determined to be outside the predetermined range.

Such discrimination processing can be performed using the damage monitoring device of the present invention.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

When damage occurs and propagates, the AE signal emitted from the structural member has a certain peculiarity in the waveform shape and the spectrum distribution shape. On the other hand, since the noise signal generated during flight has various causes, its waveform shape and spectrum distribution shape have variations.
Therefore, in the present invention, such a difference in waveform shape and spectrum shape is quantitatively expressed using various parameters,
By judging whether or not it is within the upper and lower limit thresholds, A
The E signal and the noise signal are discriminated.

FIG. 1 shows the structure of a damage monitoring device according to this embodiment. The sensor 1, the preamplifier 2, the high pass filter 3, the main amplifier 4, the transient recorder 5, and the analysis device 6 are connected in series.

The sensor 1 is attached to an airframe structural member and is emitted when damage occurs and progresses.
It receives an E signal. The preamplifier 2 receives the signal in which the AE signal output from the sensor 1 and the noise signal are mixed, and amplifies the signal. The high pass filter 3 removes a low frequency noise signal from the signal output from the preamplifier 2. The main amplifier 4 amplifies the signal output from the high pass filter 3. Transient recorder 5
Stores the signal amplified by the main amplifier 4 in the form of a digital signal by A / D conversion, and stores it as time series signal waveform data arranged in time series. The analyzer 6 receives the time-series signal waveform data and discriminates between the AE signal and the noise signal.

FIG. 2 shows a calculation procedure performed by the analyzer 6. First, at step 101, two types of parameters relating to the time domain are obtained using the time-series signal waveform data. The time-series signal waveform data represents a change in signal voltage with respect to time, as shown in FIG.
Using this time-series signal waveform data, the rise time from when the signal exceeds the threshold value to the threshold value is attenuated by the sensor 1, and when the signal exceeds the threshold value and after the threshold value is attenuated to the threshold value or less. And the duration of time required to do so are required.

Next, at step 102, a fast Fourier transform (hereinafter referred to as FFT) is performed using the time-series signal waveform data to obtain a spectrum distribution.

Then, in order to quantitatively express the distribution shape from the obtained spectral distribution, the following eight kinds of parameters in the frequency domain are obtained in step 103. First, the spectrum distribution is divided into five at arbitrary frequency bands, and spectrum bands SB1 to SB5 representing the ratio of each area to the total area are obtained.

FIG. 3B shows a spectrum distribution shape showing a distribution of signal energy with respect to frequency. The frequency bands here are 0.20 to 0.25, 0.25 to 0.
29, 0.30 to 0.35, 0.40 to 0.55, 0.
It is classified into 5 from 57 to 0.63 (MHz). And
Spectral bands SB1 to SB5 are obtained using the following equations (1) to (5) for each frequency band.

[0018]

[Equation 1] The first moment μ1 and the second moment μ2 are
It is calculated using the equations (6) and (7).

[0019]

[Equation 2]

Further, as shown in FIG. 3C, the frequency at which the energy on the vertical axis takes a value f (x2) which is 20 dB lower than the peak value f (x1) (hereinafter, -20d).
X1 is obtained.

At step 104, the AE signal 105 or the noise signal 106 is calculated by using the ten types of parameters thus obtained.
Is determined.

The procedure of this discrimination processing will be described with reference to the flowchart of FIG. Here, the order of using the 10 types of parameters is the rise time and duration, which are the two types of parameters in the time domain, then the spectral bands SB1 to SB5 in the frequency domain, and the first moment μ.
The order is first, second moment μ2, and −20 dB frequency x1.

First, in steps 200 and 201,
The discrimination process is performed using the rise time and the duration that are two types of parameters in the time domain that can be easily analyzed and can remove many noise signals. Experience has shown that the AE signal is within a predetermined range for each time. Therefore, a signal that is not within this predetermined range can be removed as a noise signal.

Next, in steps 202 and 203, noise signals are removed using the spectral bands SB1 to SB5. By this processing, it is possible to remove a noise signal having a characteristic in a specific frequency band different from the AE signal.

FIG. 5 shows the time-series signal waveform and spectrum distribution shape of the AE signal. 6 to 10 show time-series signal waveforms and spectrum distribution shapes of noise signals detected during flight, respectively. The noise signal is generated by various mechanical factors or electrical factors as described above. The eight noise signals shown here have peak values in almost the same frequency region, but have different spectral distribution shapes.

The noise signal shown in FIG. 6 is 0.20.
The area in the frequency band of 0.25 MHz is larger than the area of the entire distribution. Therefore, the value of the spectral band SB1 obtained by using the equation (1) can be removed as a noise signal that is not within the predetermined range in which the AE signal exists.

The noise signal shown in FIG.
The area occupied by the frequency band of 0.29 MHz is large and can be removed by using the spectral band SB2.

The noise signal shown in FIG. 8 is 0.30.
The area ratio of the frequency band of 0.35 MHz is large and can be removed by the spectrum band SB3. Since the noise signal shown in FIG. 9 has a large area ratio of the frequency band of 0.40 to 0.55 MHz, it is determined by the spectrum band SB4. It can be removed. The noise signal shown in FIG.
The area ratio of the frequency band of 7 to 0.63 MHz is large. Therefore, it can be removed using the spectral band SB5.

As described above, in step 202 and step 203, by removing the noise signal in which the values of the spectral bands SB1 to SB5 exceed the predetermined range, the peak value which could not be discriminated in the past is almost the same as that of the AE signal. Noise signals can also be removed.

Next, an example of the result of discriminating the AE signal and the noise signal due to damage using the spectral band SB3 will be shown. The number of AE signals detected was about 350, and the number of noise signals was about 11000. These A
FIG. 14 shows a part of the values of the spectral band SB3 obtained for the E signal and the noise signal. here,
“Δ” indicates an AE signal, and “□” indicates a noise signal.
As is clear from FIG. 14, most of the AE signals are within the predetermined range from the upper threshold SB3a (= 0.155) to the lower threshold SB3b (= 0.060). Therefore, a signal outside this predetermined range can be removed as a noise signal.

The noise signal remaining without being removed using the above seven parameters is set to 1 in step 204.
It is removed using the parameter of the second moment μ1. Figure 1
The noise signal shown in No. 1 has less spectral distribution variance and has a lower peak frequency than the AE signal of FIG.
Therefore, the value of the first-order moment μ1 is smaller than that of the AE signal, and can be removed as a noise signal that is not within the predetermined range.

In step 205, the second moment μ2
Parameters are used. The noise signal as shown in FIG. 12 has a wide distribution. Therefore, the value of the second moment μ2 of the noise signal exceeds the predetermined range in which the AE signal exists, and the noise signal can be removed.

Further, in step 206, -20d
The discrimination processing is performed using the B frequency x1. In the noise signal shown in FIG. 13, the -20 dB frequency x1 is smaller than the minimum value that the AE signal can take. Therefore, it can be removed as a noise signal whose -20 dB frequency is not within the predetermined range.

As described above, by judging whether or not the values of the 10 kinds of parameters are within the predetermined range, A
It is possible to distinguish between the E signal and the noise signal.

The above-described embodiments are merely examples and do not limit the present invention. Of the parameters used in the discrimination processing, for example, the number of spectral bands SB is five in the embodiment, but can be set to any number. Further, in the embodiment, the -20 dB frequency is used as the parameter, but it is possible to perform the accurate determination process without using the frequency depending on the noise signal generation situation.

[0036]

As described above, the method for monitoring damage to an aircraft structural member according to the present invention uses parameters of rise time, duration, spectral band, first moment, and second moment for detected signals. Since the noise signal is obtained by removing the noise that does not fall within the predetermined range in which the AE signal exists, it is possible to accurately perform the discrimination processing between the AE signal and the noise signal. Further, such a damage monitoring method of the present invention can be performed by using the damage monitoring device of the present invention.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an aircraft structural member damage monitoring device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure of a method of monitoring damage to an aircraft structural member according to an embodiment of the present invention.

FIG. 3 is an explanatory view showing a temporal change of a signal detected in the damage monitoring method.

FIG. 4 is a flowchart showing a signal analysis procedure in the damage monitoring method.

FIG. 5 is an explanatory diagram showing a temporal change and a spectral distribution of a detected AE signal.

FIG. 6 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 7 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 8 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 9 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 10 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 11 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 12 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 13 is an explanatory diagram showing a temporal change and a spectral distribution of a detected noise signal.

FIG. 14 is a spectrum band SB3 (frequency band 0.30 to 0.35 MH) of the detected AE signal and noise signal.
FIG.

[Explanation of symbols]

 1 sensor 2 preamplifier 3 high-pass filter 4 main amplifier 5 transient recorder 6 analyzer

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hidehiko Takagi 1-6-3 Haneda Airport, Ota-ku, Tokyo Nihon Airlines Co., Ltd.

Claims (2)

[Claims] 1. A step of receiving an acoustic radiation signal generated in a structural member of an aircraft by a sensor, removing a low frequency noise signal included in the signal, converting the signal to a digital signal, and storing it as time series signal waveform data. , Using the time-series signal waveform data, the rise time from when the signal exceeds the threshold until reaching the maximum amplitude, and until the signal decays below the threshold after exceeding the threshold. And a step of obtaining a spectral distribution by performing a fast Fourier transform using the time-series signal waveform data, dividing the spectral distribution into at least two frequency bands, the area of each of the divided frequency bands And a step of obtaining a ratio of the area of the entire spectrum distribution, and a step of obtaining a first moment which is an average value of the spectrum distribution. And a step of obtaining a second moment which is the variance of the spectral distribution, the rise time, the duration, the ratio of the area of each of the divided frequency bands to the overall area of the spectral distribution, the first moment , And a value of the second moment that is included in a predetermined range, respectively, is determined as an acoustic emission signal, and a value that is not included is determined as a noise signal, and the aircraft structural member. Damage monitoring method. 2. A sensor provided on a structural member of an aircraft for receiving an acoustic radiation signal, a filter for removing a low frequency noise signal from a signal received by the sensor, and the filter for removing the low frequency noise signal. A / D converter for converting the converted signal into a digital signal, means for storing the converted digital signal as time-series signal waveform data, and using the time-series signal waveform data, the signal has a threshold value. A means for obtaining a rise time from reaching the maximum amplitude to reaching the maximum amplitude and a duration from the time when the signal exceeds the threshold to the time when the signal attenuates below the threshold, and high-speed using the time-series signal waveform data. Means for performing a Fourier transform to obtain a spectral distribution, and dividing the spectral distribution into at least two frequency bands, each of the divided frequencies A means for obtaining a ratio between the area of the region and the area of the entire spectrum distribution, a means for obtaining a first moment which is an average value of the spectrum distribution, and a means for obtaining a second moment which is a variance of the spectrum distribution were obtained. The rising time, the duration, the ratio of the area of each of the separated frequency bands to the total area of the spectrum distribution, the values of the first moment and the second moment are included in respective predetermined ranges. Is a sound emission signal, and a device that does not include a noise emission signal is determined as a noise signal.