JP2006058278A - Elastic wave generation position calculation device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To calculate in real time the generation position of a damage or a collision, concerning an elastic wave generation position calculation device and a method for calculating the generation position of an elastic wave (AE) generated when a damage is generated in a structure or when an object collides with the structure from the outside. <P>SOLUTION: This elastic wave generation position calculation device 10 is equipped with a plurality of AE (acoustic emission) sensors 11 installed on the surface of a thin plate 1, for detecting the AE; and a computer 13 for calculating in real time the generation position of the AE generated on the thin plate 1. The computer 13 performs continuous wavelet transformation to detection signals of each sensor 11, calculates each arrival time of the elastic wave to each sensor 11 based on calculated each wavelet coefficient, and estimates the generation position of the AE based on the sensor position and an arrival time difference. In this case, in the continuous wavelet transformation processing, the detection signal of the sensor 11 is subjected to FFT transformation, and mother wavelet is subjected to FFT transformation, and each result is multiplied together relative to each frequency, and a signal acquired by multiplication is subjected to inverse FFT transformation, to thereby calculate the wavelet coefficient. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an elastic wave generation position calculation apparatus and method for calculating a generation position of an elastic wave (acoustic emission) generated when damage occurs in a structure or an object collides with the structure from the outside. It is about.

  When damage occurs in the structure or an object collides with the structure from the outside, an elastic wave called acoustic emission (hereinafter referred to as AE) is generated. By monitoring this AE, damage and collision of the structure can be observed in real time.

  Attaching multiple AE sensors to the structure, observing the elastic waves excited by the damage generated in the structure and the collision of flying objects. From the relationship between the arrival time of the elastic waves to each sensor and the mounting position of each sensor A system for calculating the occurrence position of damage or collision is used in a plant or the like.

  However, when the structure is a thin plate, pipe, or rod, the elastic wave propagates as a guided wave having a plurality of propagation modes, and the velocity changes depending on the propagation mode, frequency, and the like. For this reason, AE observed by the AE sensor becomes a wave having dispersibility, and it is very difficult to accurately measure the arrival time from the detection signal of the AE sensor.

  Therefore, an attempt is made to perform continuous wavelet transform on the detection signal of the AE sensor, calculate the wavelet coefficient of the detection signal, and measure the arrival time of the induced wave from the wavelet coefficient (for example, Non-Patent Document 1). , 2).

J. of Acoustic Emission, Vol.18 (2000) pp.51-60 Proc. Of 44th AIAA / ASME / ASCE / AHS Structures, Structural Dynamics, and Materials Conference (2003) AIAA2003-1937

  By the way, if the continuous wavelet transform is calculated by a normal method, it is necessary to perform the convolution work at each position while shifting the mother wavelet by the number of samples of the detection signal on the time axis, and the calculation processing cost is very high.

  Therefore, it has been difficult to apply the continuous wavelet transform to a system that calculates the occurrence position of damage or collision in real time.

  An object of the present invention is to provide an elastic wave generation position calculation apparatus and method that can reduce the calculation cost and calculate the occurrence position of damage or collision in real time.

  An elastic wave generation position calculation apparatus according to the present invention is an elastic wave generation position calculation apparatus that calculates the generation position of an elastic wave (acoustic emission) generated in a thin plate, pipe-shaped or rod-shaped object, and the surface of the object A plurality of sensors that detect vibrations of the surface of the object, and a calculation unit that calculates a generation position of the elastic wave on the surface of the object based on a detection signal of each sensor. Wavelet transform means for calculating wavelet coefficients by performing continuous wavelet transform on the detection signals detected by the sensors, and arrival times of elastic waves at the sensors based on the wavelet coefficients calculated by the wavelet transform means Arrival time calculating means for calculating the position of each sensor on the surface of the object, Calculation means for calculating a position where the elastic wave is generated in the object based on a time difference between arrival times of the elastic waves to the sensor, and the wavelet transform means converts the detection signal detected by the sensor into an FFT. (Fast Fourier Transform) transform to generate frequency domain detection signal, and frequency domain mother wavelet obtained by FFT transform of mother wavelet in wavelet transform and the above frequency domain detection signal are multiplied for each frequency The wavelet coefficient is calculated by performing inverse FFT transform on the frequency domain signal obtained by the multiplication.

  In the elastic wave generation position calculation method according to the present invention, a plurality of sensors for detecting vibrations of the surface of the object are attached to the surface of the thin plate, pipe-shaped or bar-shaped object, and an elastic wave propagating through the object ( In the elastic wave generation position calculation method for calculating the generation position of acoustic emission, a step of calculating wavelet coefficients by performing continuous wavelet transform on the detection signals detected by the sensors, and calculating each wavelet coefficient Calculating the arrival time of the elastic wave to each sensor based on the position of the sensor, the position of the sensor in the object, and the time difference of the arrival time of the elastic wave to the sensor, the object of the elastic wave Calculating the occurrence position on the surface of the sensor, and in the continuous wavelet transform, the sensor The detected signal detected by the FFT is subjected to FFT (Fast Fourier Transform) transform to generate a frequency domain detection signal, and the mother wavelet in the frequency domain obtained by performing FFT transform on the mother wavelet in the wavelet transform, and detection of the above frequency domain A wavelet coefficient is calculated by multiplying a signal for each frequency and performing inverse FFT transform on the frequency domain signal obtained by the multiplication.

  According to the present invention, it is possible to reduce the calculation cost when performing continuous wavelet transform on the detected induced wave, and to calculate the occurrence position of damage or collision in real time by incorporating the calculation device into the measurement device. it can.

  Hereinafter, an AE occurrence position calculation system to which the present invention is applied will be described.

  FIG. 1 is a diagram schematically showing the configuration of an AE occurrence position calculation system 10 to which the present invention is applied.

  The AE generation position calculation system 10 is a system that detects an elastic wave called acoustic emission (AE) generated on the measurement object 1 and measures the generation position of the AE on the measurement object 1 in real time. AE occurs when the measurement object 1 is damaged or when an object collides with the measurement object 1 from the outside.

  The measurement object 1 is a structure in which when AE occurs, the AE propagates in a mode called a plate wave (Lamb wave) which is a kind of induced wave. Here, the measurement object 1 is an example of a flat plate structure. However, if the AE propagates as a guided wave, the measurement object 1 is not a flat plate structure but a rod-like structure or piping. It may be a pipe-like thin plate structure.

The system configuration AE generation position calculation system 10 receives a plurality of AE sensors 11 (11-1, 11-2,..., 11-n) arranged on the surface of the measurement object 1 and detection signals of the AE sensors 11. An analog / digital converter 12 for converting into digital data and a computer 13 for performing arithmetic processing are provided.

  Each AE sensor 11 is a sensor that detects an AE when the AE occurs on the measurement object 1. That is, each AE sensor 11 detects vibration generated at the mounting position of the measurement object 1 and converts the vibration into an electric signal. A sensor such as a pressure sensor or an acceleration sensor may be used.

  The measuring object 1 is provided with a plurality of AE sensors 11 on the surface thereof. An example of the attachment position of each AE sensor 11 with respect to the surface of the measurement object 1 is shown in FIG.

  The plurality of AE sensors 11 (11-1, 11-2,..., 11-n) each line when a virtual triangular lattice pattern is drawn on the surface of the measurement object 1, as shown in FIG. It is provided on the intersection. In other words, when an arbitrary point on the surface of the measuring object 1 is specified, if three AE sensors 11 that are close to the point are selected, the three AE sensors 11 are arranged in a triangle. Yes. For example, when the point A in FIG. 2 is specified, the three AE sensors that are close to the point A are the AE sensor 11-3, the AE sensor 11-7, and the AE sensor 11-8. It is composed.

  Further, the position where each AE sensor 11 is arranged (coordinates with respect to the measurement object 1) is measured in advance and registered in the computer 13. That is, the positional relationship between each position of each AE sensor 11 and the measurement object 1 is registered in the computer 13. Of course, the present invention is not limited to the sensor arrangement described above, and the sensor arrangement and the number of settings can be variously changed without departing from the scope of the present invention.

  The analog / digital converter 12 receives the detection signals detected by the AE sensors 11 (11-1, 11-2,..., 11-n), amplifies the detection signals, and then performs analog-digital conversion. . Detection signals of the AE sensors 11 (11-1, 11-2,..., 11-n) digitized by the analog / digital converter 12 are input to the computer 13.

  The computer 13 calculates the generation position of the AE on the measurement object 1 based on the detection signal of each AE sensor 11 after digitization. A software program for calculating the AE occurrence position is installed in the computer 13. The computer 13 always executes the program, calculates the AE occurrence position in real time, and notifies the user. The computer 13 is also provided with a monitor 14 and an input unit 15, and displays necessary information for the user and obtains input information from the user. As for the arithmetic processing, a mode in which an arithmetic processing device is provided between the analog / digital converter 12 and the computer 13 to perform measurement is also conceivable.

Measurement Flow Next, a processing procedure for measuring the AE occurrence position by the computer 13 will be described.

  The computer 13 constantly observes AE generated on the measurement object 1. The computer 13 first executes a software program for observing the AE occurrence position. This software program 13 is always executed on the computer 13 during AE observation.

  FIG. 3 shows a processing flow of the computer 13 when a software program for observing the AE occurrence position is executed. Hereinafter, description will be made according to this processing flow.

  First, in step S1, the computer 13 monitors the detection signal of each AE sensor 11, and determines whether or not the detection signal of any one AE sensor 11 is equal to or greater than a predetermined threshold (step S1). If the threshold value is equal to or greater than the predetermined threshold value (YES in step S1), the subsequent processing from step S2 is performed, and if the output of any AE sensor 11 is less than the predetermined threshold value (NO in step S1), In step S1, the process waits. That is, the process of step S1 means that the calculation calculation of the actual AE occurrence position is started using the detection output level of the AE sensor 11 as a trigger.

  Subsequently, in step S2, the computer 13 starts capturing detection signals of all AE sensors 11, and stores the captured detection signals in the memory.

  Subsequently, in step S <b> 3, the computer 13 determines one triangular area including the AE generation source from a plurality of triangular areas surrounded by every three AE sensors 11. That is, the three AE sensors 11 surrounding the AE generation source are detected.

  As a determination method, the time when the level of the detection signal becomes equal to or higher than a predetermined threshold is obtained for all the AE sensors 11. Subsequently, the sum of the times of the three AE sensors 11 constituting the triangle is obtained for each triangle area. For example, for all 13 triangular regions shown in FIG. 2, the sum of the time points when the three AE sensors 11 constituting the triangle are equal to or greater than a predetermined threshold is obtained. Then, one triangle having the smallest sum is identified, and the identified triangle is determined as a triangle area including the AE occurrence position.

  Hereinafter, the triangular area including the AE occurrence position is referred to as a “specific triangular area”.

  Subsequently, in step S4, the computer 13 specifies the three AE sensors 11 arranged at the vertices of the specific triangular area, and performs continuous wavelet for each detection signal output from the three specified AE sensors 11. Perform conversion.

  For example, detection signals as shown in FIG. 4 are output from the three identified AE sensors 11. The detection signal output from the AE sensor 11 is a wave indicating dispersibility.

  When a continuous wavelet transform is performed on a predetermined range of detection signals output from the three AE sensors 11 (for example, a range of 500 μsec shown in FIG. 4), a waveform as shown in FIG. 5 can be obtained. it can.

  When continuous wavelet transform is performed in this way, one peak such as a point surrounded by a dotted line in FIG. 5 can be obtained.

  The details of the continuous wavelet calculation process in step S4 will be described later.

  Subsequently, in step S5, the computer 13 obtains the time when the obtained wavelet coefficient value becomes the largest, that is, the peak time from the result of continuous wavelet calculation of the detection signals of the three AE sensors 11, that is, the three AE sensors. Extract every 11th. Then, the computer 13 calculates the difference between each peak time of the three AE sensors 11 and a certain predetermined reference time, and stores these as AE arrival time difference information (time difference information t1, t2, t3).

  Subsequently, in step S6, the computer 13 determines the arrival time difference information (t1, t2, t3) of the three AE sensors 11 constituting the specific triangle, and the arrangement coordinates of these three AE sensors 11 (corresponding to the measurement object 1). The position where the AE occurs on the surface of the measuring object 1 is determined based on the coordinates of the mounting position of the AE sensor 11 to be performed.

For example, when the target structure is made of a homogeneous isotropic material, it consists of the difference in arrival time (t ₁ −t ₂ , t ₃ −t _2, etc.) between the sensor coordinates and the sensor and the velocity of the induced wave in the structure. The sound source position can be uniquely calculated by solving the simultaneous equations in the computer 13. For structures with anisotropy, such as fiber reinforced plastics (FRP) structures, the time to reach the sensor at each position in the exploration zone takes into account the sonic anisotropy. calculation Te (FIG. 6), and the resulting arrival time difference information _{(t'1 -t' 2, t'3} -t' 2 , etc.) and the time difference information obtained in the experiment _{(t 1 -t 2, t 3} - The sound source position can be determined by finding the point where t ₂ etc. is minimized.

  When the process of step S6 is completed, the computer 13 can acquire the AE occurrence position.

  The computer 13 outputs the AE occurrence position X acquired as described above as a numerical value, or displays it as a screen 20 showing the position visually as shown in FIG. 7, for example, and presents it to the user. .

CWT Next, processing contents of the continuous wavelet transform at step S4 will be described.

  The wavelet transform is a convolution operation between a signal (detection signal x (t)) having a time t as a variable to be analyzed and a bundle of signal waves Ψ localized in terms of time and frequency. Specifically, it is expressed as the following formula.

Ψ _{a, b} (t) is a special function that satisfies the admissible condition and is called a mother wavelet and includes a real part and an imaginary part. The frequency of the mother wavelet Ψ _{a, b} (t) changes when a called a scaling coefficient is changed. Therefore, arbitrary frequency information can be extracted from the detection signal x (t) by changing the scaling coefficient a. The mother wavelet moves in the time direction when b called a shift coefficient is changed. Therefore, information at an arbitrary time of the detection signal x (t) can be extracted by changing the shift coefficient b.

  First, a normal continuous wavelet transform calculation process will be described. FIG. 8 shows a calculation procedure of normal continuous wavelet transform.

First, a detection signal x (t), calculates the convolution of the initial time _{b 0} of the mother wavelet Ψ _a, b _(t), and calculates the wavelet coefficients at the initial time _{b 0} (FIG. 8 (A)).

Subsequently, a convolution of the detection signal x (t) with the mother wavelet Ψ _{a, b} (t) at time b ₁ shifted by one step time from the initial time b ₀ is calculated, and the wavelet coefficient at time b ₁ is calculated. Is calculated (FIG. 8B).

Thereafter, the time is shifted by one step, all the convolution calculations are performed, and the calculation is performed over the entire time range (time b ₀ to time b _n ) (FIG. 8C).

As a result, wavelet coefficients in the range of time b ₀ to time b _n can be obtained.

  However, the convolution calculation has a very high calculation cost. For this reason, when such a normal continuous wavelet transform calculation process is used for the measurement process of the AE occurrence position, it is almost impossible to calculate the AE occurrence position in real time.

  Therefore, the present inventor decided to use the following high-speed calculation algorithm of continuous wavelet transform for the processing.

  FIG. 9 shows a specific calculation processing procedure in step S4.

  First, a fast Fourier transform (FFT) operation is performed on the detection signal x (t) (step S11).

Subsequently, an FFT operation is performed on the mother wavelets _{a, b} (t) (step S12). Note that the result of the FFT operation of the mother wavelets _{a, b} (t) may be stored in advance in a memory and read out.

Subsequently, the result of the FFT operation of the detection signal x (t) and the result of the FFT operation of the mother wavelet Ψ _{a, b} (t) are multiplied for each frequency (step S13).

  Subsequently, an inverse fast Fourier transform (IFFT) operation is performed on the multiplication result in step S13 (step S14).

  Then, the signal obtained as a result of the IFFT calculation in step S14 is output as a result of the continuous wavelet calculation (wavelet coefficient).

  As described above, according to the calculation, the calculation is completed with two FFT operations, one multiplication (for each frequency), and one IFFT operation, so that the calculation cost is very low. It can be applied to the measurement process of the generation position.

It is a block block diagram of the AE generation | occurrence | production position calculation system to which this invention was applied. It is a figure which shows the attachment position of the AE sensor with respect to a measurement object. It is a processing flow of a measurement of AE generating position. It is a figure which shows the detection signal output from the AE sensor. It is a figure which shows the signal obtained by performing continuous wavelet transformation with respect to the detection signal output from the AE sensor. It is the figure which showed the method of estimating AE generation | occurrence | production position. It is a figure which shows the presentation screen to the user of a measurement result. It is the figure which showed the procedure of the normal continuous wavelet transformation process. It is the figure which showed the procedure of the simplified continuous wavelet transform used with the AE generation | occurrence | production position calculation system to which this invention was applied.

Explanation of symbols

10 AE generation position calculation system, 11 AE sensor, 12 analog / digital converter, 13 computer

Claims (2)

In an elastic wave generation position calculation device that calculates the generation position of elastic waves (acoustic emission) generated in a thin plate, pipe-shaped or rod-shaped object,
A plurality of sensors installed on the surface of the object and detecting vibrations of the surface of the object;
A calculation unit that calculates the generation position of the elastic wave on the surface of the object based on the detection signal of each sensor;
The arithmetic unit is
Wavelet transform means for calculating wavelet coefficients by performing continuous wavelet transform on the detection signals detected by each of the sensors,
Arrival time calculating means for calculating the arrival time of the elastic wave to each sensor based on each wavelet coefficient calculated by the wavelet transform means;
Calculating means for calculating a position where the elastic wave is generated in the object based on a position of each sensor on the surface of the object and a time difference between arrival times of the elastic waves at the sensors;
The wavelet transform means is
A detection signal in the frequency domain is generated by performing FFT (Fast Fourier Transform) conversion on the detection signal detected by the sensor,
The frequency domain mother wavelet obtained by FFT transforming the mother wavelet in the wavelet transform and the frequency domain detection signal are multiplied for each frequency,
A wavelet coefficient is calculated by performing inverse FFT transform on the frequency domain signal obtained by the multiplication. A plurality of sensors for detecting the vibration of the surface of a thin plate, pipe-shaped or rod-shaped object are attached, and an elastic wave for calculating the generation position of an elastic wave (acoustic emission) propagating through the object In the generation position calculation method,
Calculating wavelet coefficients by performing continuous wavelet transform on the detection signals detected by each of the sensors;
Calculating the arrival time of the elastic wave to each sensor based on each calculated wavelet coefficient;
Calculating the generation position of the elastic wave in the object based on the position of each sensor on the surface of the object and the time difference between the arrival times of the elastic waves at the sensors;
In the above continuous wavelet transform,
A detection signal in the frequency domain is generated by performing FFT (Fast Fourier Transform) conversion on the detection signal detected by the sensor,
The frequency domain mother wavelet obtained by FFT transforming the mother wavelet in the wavelet transform and the frequency domain detection signal are multiplied for each frequency,
A wavelet coefficient is calculated by performing inverse FFT transform on the frequency domain signal obtained by the multiplication.

Images