JPH11344408A - Sound source probing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a sound source surveying device to maintain its sound source-locating function even when the frequency characteristics of used acoustic sensors are not flat by measuring in advance the phase lag of a microphone at every frequency and making phase correction at every frequency at the time of calculating the intensity of a sound source in a monitoring area. SOLUTION: A high-speed ADC 31 converts analog detecting signals into digital signals. An FFT processor 32 is a fast Fourier transformation device and has such a function that shifts detected acoustic signals S1-S8 to a frequency domain by Fourier transforming the signals S1-S8. A cross spectrum processor 33 calculates the cross spectra of all combinations of two different signals. The cross spectra can be found as the products of the Fourier-transformed signals and the complex conjugates of the Fourier-transformed signals. A sound source intensity distribution preparing section 34 with phase correcting function for acoustic sensor takes the sum of all calculated cross spectra by shifting the phases of the spectra in accordance with a virtual sound source point, but simultaneously makes phase correction in accordance with the frequency characteristics of microphones.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention calculates a sound source intensity distribution in a monitoring area using detection signals of a plurality of acoustic sensors installed around the monitoring area, and clarifies a position where a strong sound source exists. More particularly, the present invention relates to a device suitable for confirming a leak occurrence position in a plant or the like.

[0002]

2. Description of the Related Art As a conventional technique for locating leaked sound from equipment and pipes placed in a closed space in a power plant,
There is a method using an acoustic sensor array (Development of a two-dimensional leak detection system using an acoustic method: Atomic Energy Society of Japan, "1
997 Autumn Convention ", F29). By using a plurality of acoustic microphones and giving different time delays to the output signals of the microphones and superimposing them, an acoustic sensor having high sensitivity to sound from a specific angle is realized. By using a plurality of microphones, a reverberation component that causes a problem in a closed space is canceled to capture sound from an actual sound source.

[0003] In general, the frequency band of a thin microphone extends to a high frequency, but the sensitivity tends to decrease as the microphone becomes thinner. For this reason, it is difficult to obtain a microphone whose frequency band extends to a high frequency and has high sensitivity.
The sensitivity of a microphone does not become completely zero from a certain frequency, but has a sensitivity according to the frequency. For this reason,
Depending on the acoustic conditions of the monitoring area, the frequency characteristics of the microphone may not be flat and a frequency band in which the sensitivity changes may have to be used. In such a frequency domain, there is a possibility that the conventional method using only the acoustic sensor array may not provide sufficient positioning accuracy. In this regard, the prior art has not been sufficiently considered.

[0004]

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an apparatus for sound source search capable of maintaining sound source position locating performance even when the frequency characteristics of an acoustic sensor used are not flat. .

[0005]

In order to achieve the above object, a sound source detecting apparatus according to the present invention measures a phase delay for each frequency of a microphone in advance and calculates a sound source intensity in a monitoring area. For this, means for correcting the phase for each frequency is used.

That is, in a region where the sensitivity of the microphone changes with the frequency, the phase characteristic also changes according to the change in sensitivity. For this reason, when the sensitivity changes with the frequency, the sound speed apparently changes due to the phase change of the output signal with respect to the input acoustic wave. Since a difference occurs in the phase delay for each frequency, in the position location method in which the sound velocity is constant at all frequencies, the position location error also differs for each frequency. Therefore, the means for correcting the phase for each frequency acts to make the sound speed constant regardless of the frequency.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below.

FIG. 1 shows a configuration of a sound source searching apparatus according to the present invention. This is an example applied to sound source search in a closed space in a plant. Eight acoustic sensors 11 to 18 are provided, and their output signals are amplified by the amplifier 20. The sound source scanning unit 30 calculates a sound source intensity distribution in the monitoring area from detection signals of the acoustic sensors 11 to 18 arranged around the monitoring area. The sound source location and display unit 40 has a function of performing sound source location from the position of the intensity peak of the sound source intensity distribution, and displaying the result.

Next, the concept of calculating the sound source intensity distribution in the monitoring area will be described with reference to FIGS. FIG. 2 shows the relationship between the distance between the sound source and the acoustic sensor in the set monitoring area. Now, assuming that a leak sound is generated from the sound source point, the signals S1 to S8 are detected with a delay corresponding to the distance between the sound source and the installation positions of the acoustic sensors 11 to 18. FIG. 3 shows the temporal relationship between the signal generated by the sound source and the detection signals S1 to S8. A virtual sound source point is set in the monitoring area, the sound is generated from the virtual sound source point, the propagation delay time of the detection signals S1 to S8 is corrected, and the average of these signals is obtained. The result is the sound source signal SB shown in the lower part of FIG. When the virtual sound source point and the real sound source match, the phase becomes uniform due to the propagation time delay correction and the amplitude increases, and when they do not match, the phase does not match and the amplitude is small. By obtaining the sound source signal for each virtual sound source point, the sound source signal at each position in the monitoring area can be calculated.

By squaring this sound source signal and taking the time integration thereof, the sound source intensity for each virtual sound source point is obtained. FIG. 4 shows the result of plotting the sound source intensity on the monitoring area. The higher the intensity, the darker the black, that is, the higher the density. Thereby, it is possible to know at which position on the monitoring area a loud sound is being generated. The position of the actual sound source point can be known from the position of the intensity peak of the sound source intensity distribution. Now, since the sound source is assumed to be a point sound source, ideally only the actual sound source point position should have a high intensity, and the intensity should be 0 at other points. Some intensity is also observed at points around the sound source point. Also, the reflected sound or the like becomes a secondary sound source at various positions, and the intensity often does not become zero at positions other than the actual sound source point.

Next, the details of the sound source scanning signal processing section 30, which is a main part of the present invention, will be described. The sound source scanning signal processing unit 30 estimates the sound source intensity distribution using the detected sound signals S1 to S8. The concept of the method for estimating the sound source intensity distribution has already been described. However, as described later, the detected sound signals S1 to S1
Instead of directly adding S8 on the time axis as described in the concept, the operation is actually shifted to the frequency domain and performed. More specifically, the detected acoustic signals S1 to S8 are subjected to Fourier transform, cross spectra of all combinations of two different signals are calculated, and time correction corresponding to the virtual sound source position is given to all the calculated cross spectra. By adding, the intensity for each virtual sound source point is obtained. By arranging the intensities for each virtual sound source point as shown in FIG. 4, a sound source intensity distribution can be created.

FIG. 5 shows a specific configuration of the sound source scanning signal processing section 30. High-speed ADC 31, FFT processor 3
2, a cross spectrum processor 33, and a sound source intensity distribution creating unit 34 with a phase correction function for an acoustic sensor. The high-speed ADC 31 has a function of converting a detection signal, which is an analog signal, into a digital signal. FFT processor 32
Is a so-called fast Fourier transform device, which has a function of performing a Fourier transform on the detected acoustic signals S1 to S8 and transferring the detected acoustic signals to the frequency domain. The cross spectrum processor 33 has a function of calculating a cross spectrum of all combinations of two different signals. This cross spectrum can be obtained as the product of the Fourier transform signal and the complex conjugate of the Fourier transform signal. The sound source intensity distribution creating unit 34 with a phase correction function for the acoustic sensor adds all of the calculated cross spectra by shifting the phase according to the virtual sound source point. In the present invention, the phase correction based on the frequency characteristics of the microphone is also performed at the same time. I do it. This allows
The phase correction for each frequency can be performed for each microphone.

Next, a specific effect on the calculation of the sound source intensity distribution caused by the frequency characteristics of the microphone, which is a problem in the present invention, will be described. FIG. 6 and FIG. 7 are examples in which the frequency characteristics of the microphone are simulated by a two-pole transfer function. FIG. 6 shows the sensitivity and FIG. 7 shows the phase. The horizontal axis is the frequency normalized by the resonance frequency. The value of selectivity Q is 0.707,
Three cases of 5 and 10 are shown. When Q is small, as can be seen from FIG. 7, the phase change becomes gentle, but the change itself does not disappear. In any case, it is understood that a phase change is unavoidable near the resonance point. The phase is different for each frequency. When handling a signal near the resonance point, the phase of the sound detected by the microphone differs for each frequency, which means that the sound speed apparently differs depending on the frequency.

For example, if the resonance frequency is 10 kHz, the phase delay is 1.57 rad, so that a time delay corresponding to a distance of 8.5 mm at a sound speed of 340 m / s in air occurs. As the resonance frequency decreases and the number of poles of the transfer function increases, the time delay further increases. If the microphones have the same characteristics, the sound speed may be corrected. However, normally, it is assumed that the microphone is used in a portion where the characteristics at a low frequency than the resonance point are flat, and no effort has been made to make the characteristics near the resonance point the same. For this reason, when a frequency band near the resonance point is used, phase correction is required.

FIG. 8 and FIG. 9 show the difference in the calculation result of the sound source intensity portion distribution depending on the presence or absence of the phase correction. FIG. 8 shows a case where the phase correction is not performed, and FIG. 9 shows a case where the phase correction is performed. It can be seen that the peak is blurred unless phase correction is performed. On the other hand, it can be seen that the peak becomes clear by the phase correction, and the sound source position can be accurately obtained.

In operation of the sound source detection device of the present invention,
It is necessary to obtain the frequency characteristics of the acoustic sensor in advance. In the present embodiment using a microphone as an acoustic detector, frequency characteristics were measured in an anechoic chamber for each microphone, and the phase correction amount for each frequency of each microphone was determined based on this data. If an acceleration sensor or the like is used as the acoustic sensor, it goes without saying that a sound source can be searched inside the container. The acceleration sensor depends on the mounting condition.
The frequency characteristics may be significantly different. In such a case, it is also possible to measure the frequency characteristics in an actual mounting state. Even if the frequency characteristic is not always an absolute value, it can be realized by using a relative frequency characteristic difference of an acoustic sensor to be used.

In this embodiment, the frequency characteristic of the acoustic sensor is corrected by correcting the phase of the cross spectrum. As another method, it is also possible to correct the detected acoustic signal itself by passing it through a filter having the opposite characteristic. The opposite characteristic is that if the acoustic sensor has a phase delay, a filter with a phase lead corresponds to it.
In the present embodiment, a method of correcting the phase of the cross spectrum is adopted from the viewpoint that the difference in the characteristics of the acoustic sensor is merely changed by the data and does not involve a filter design operation. Another characteristic is that the digital implementation realizes higher correction accuracy.

The present invention is not limited to an acoustic SG leak detection device. It can be used for the presence or absence of a sound source in the monitoring area, sound source location, and the like. For example, it can be applied to a foreign substance monitoring device that captures a partial discharge noise caused by the presence of conductive minute foreign substances inside a gas insulated switchgear and a collision noise caused by the jumping of foreign substances, and also to a sound source monitoring inside a container such as a transformer. It is possible.

As described above, the following advantages can be given as specific effects of this embodiment.

(1) By realizing the phase correction digitally, the accuracy of the correction can be increased, and as a result, a high positioning accuracy can be obtained, which has the effect of improving the performance of the sound source searching device.

(2) By realizing the phase correction digitally, it is not necessary to design a new filter for phase correction by changing the acoustic sensor or the like, so that the operation of the sound source locating apparatus is simplified, and as a result, the operation of the apparatus is simplified. It has the effect of improving economy.

(3) Digitally implementing the phase correction makes it possible to correct complicated acoustic characteristics.
This has the effect of improving the performance of the sound source detection device.

[0023]

As described above, according to the present invention, it is possible to correct the apparent change in sound speed due to the frequency characteristics of the acoustic sensor, and as a result, it is possible to improve the position localization accuracy of the sound source intensity distribution. As a result, there is an effect of improving the performance of the sound source searching device.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a sound source detection device of the present invention.

FIG. 2 is a diagram showing a positional relationship between an acoustic sensor arrangement and an actual sound source point.

FIG. 3 is an explanatory diagram of an outline of a time change of an acoustic signal and signal processing.

FIG. 4 is a diagram showing an example of a sound source intensity distribution.

FIG. 5 is an internal configuration diagram of a sound source search signal processing unit of the embodiment.

FIG. 6 is a sensitivity characteristic diagram of a microphone.

FIG. 7 is a phase characteristic diagram of a microphone.

FIG. 8 is a diagram showing a sound source intensity distribution before phase correction.

FIG. 9 is a sound source intensity distribution diagram after phase correction.

[Explanation of symbols]

Reference numerals 11 to 18: acoustic sensor, 20: amplifier, 30: sound source scanning signal processing unit, 31: high-speed ADC, 32: FFT processor, 33: cross spectrum processor, 34: sound source intensity distribution creating unit with phase correction function, 40: sound source Location and display.

Claims (2)

[Claims] 1. Sound detecting means arranged at a plurality of points near a monitoring target area; means for calculating a sound source intensity distribution in a monitoring area from a plurality of detected sound signals; A sound source location device comprising a sound source position locating means for determining a position of a sound source as a sound source position, wherein a means for correcting a phase change of an acoustic sensor for each sound detection point and each frequency is added to the sound source intensity calculating means. Sound source locator. 2. The sound source intensity calculating means according to claim 1,
Means for Fourier transforming a plurality of detected sound signals, means for calculating a cross spectrum by each combination of the plurality of detected sound signals, and means for calculating a sound source intensity from the cross spectrum for each virtual sound source point. A sound source searching apparatus characterized in that means for correcting the phase of the acoustic sensor for each frequency of each cross spectrum is added to the means.