JPH023455B2 - - Google Patents

Description

[Detailed Description of the Invention] The present invention relates to a method for locating a sound source position in a pressure vessel,
In particular, the present invention relates to a device suitable for detecting the location of a foreign object by locating the location of a collision sound generated when the foreign object collides with a location within a pressure vessel.

The present inventors have previously proposed a method for locating the sound source position in a pressure vessel (Japanese Patent Application No. 54-118306
36165], the wave height data and arrival time difference data of acoustic signals from a large number of known sound sources whose positions are known in advance are stored in the memory of a computer, etc., and the unknown sound source is selected from among them. By searching for the closest match to each data,
An effective method is to locate the unknown sound source.

However, in the conventional method described above, the pattern recognition method for comparing and matching the data of the unknown sound source with the known data in the file was not sufficient, so there was a high probability of misidentification. There were flaws.

An object of the present invention is to provide a method for locating the position of a sound source in a pressure vessel, which is capable of locating the position of a sound source with high accuracy by improving a pattern recognition method for comparing and collating data of a known sound source and data of an unknown sound source. It is in.

The present invention provides for installing at least three acoustic detectors for detecting sounds generated within a pressure vessel at appropriate locations related to the pressure vessel, capturing acoustic signals from an unknown sound source output from the detectors, At least one of the peak value data and the time difference data obtained by calculating the time difference between the peak value of the signal and the signal arrival time between the detectors, and a large number of known sound source positions corresponding to the data in advance. By comparing and collating the known wave height value data stored in the memory and the known time difference data and searching for known data that is most similar to the data, it is possible to determine whether the unknown sound source position is closest to any of the known sound source positions. In determining whether there is a first dispersion value or a function of the dispersion value consisting of the difference between the logarithm of the peak value of each of the detectors and the logarithm of the known peak value of each of the detectors, A second variance value consisting of the difference between the time difference of each detector and the known time difference of each detector or a function of the variance value is calculated for each of the known sound source positions, and This method attempts to locate the sound source position with high accuracy by comparing mathematical average values, and is characterized by a method of comparing and matching data related to unknown sound sources with data related to known sound sources.

Hereinafter, the comparative verification method of the present invention will be explained in detail.

The outline of the comparative verification method of the present invention is that patterns are set from the peak value data and signal arrival time data of the acoustic signals emitted from the sound sources inside the pressure vessel, and based on these patterns, two sound sources A,
A pattern distance between sound sources A and B (for example, known sound source A and unknown sound source B) is defined, and it is determined whether the distance between sound sources A and B is close based on the magnitude of this pattern distance.

First, the peak value pattern distance D ^A _AB based on the peak value data of sound sources A and B is detected by the i-th (i = 1 to N) acoustic detector (hereinafter simply referred to as a detector) S _i The wave height values of sound sources A and B are respectively a _i and b _i
is defined as the following equation (1).

In addition, Var in formula (1) represents variance, and is a general quantity.
This is defined by the following equation (2) for X _i .

Similarly, the time difference pattern distance D ^T _AB based on the signal arrival time data between the sound sources A and B is determined by the difference between the time when the signal arrives at the detector S _i and the reference time, that is, the time difference, for the sound sources A and B, respectively. As τ _Ai and τ _Bi ,
It is defined as in the following equation (3).

D ^T _AB = {Var (τ _Ai − τ _Bi )} ^1/2 …(3) Furthermore, by combining the above pattern distances D ^A _AB and D ^T _AB , the pattern distance D ^C _AB can be calculated as shown in the following equation (4). Define. Note that α in the equation is a coefficient having a value of 0≦α≦1, and is related to the weight of the so-called weighted average value.

D ^C _AB = (1 - α) D ^T _AB + αD ^A _AB ... (4) Here, the above pattern distances D ^A _AB , D ^T _AB ,
The relationship between D ^C _AB and the geometric distance d between sound sources A and B will be explained. Assume that the positional relationship between the sound sources A and B and the detector S _i is as shown in FIG.
Let the distances between B and detector S _i be γ _Ai and γ _Bi , respectively,
A straight line connecting sound source A and sound source B or detector S _i ,
Assuming that the right angle formed by AS ₁ is θ _i , the relationship shown in the following equation (5) holds true between the distance d between the sound sources A and B and the above γ _Ai and γ _Bi .

γ ² _Bi = γ ² _Ai + d ² −2γ _Ai・dcosθ _i …(5) On the other hand, the peak value of detector Si for sound sources A and B
a _i and b _i can be expressed by the following equation (6).

a _i =K _A γ _Ai ^-J , b _i =K _B γ _Bi ^-J ...(6) In the above equation (6), K _A and K _B are the sound source A,
is a constant that depends on the intensity of the impact at B, and J
is the attenuation constant in the sound transmission medium.

Now, assuming that d≪γ _Ai , equations (5) and (6) to equation (1) become the following equation (7).

Similarly, equation (2) becomes the following equation (8).

D ^T _AB = d√( _i ) …(8) As is clear from the above equations (7) and (8), the pattern distance
Both D ^A _AB and D ^T _AB are quantities proportional to the geometric distance d between the sound sources A and B.

Note that D ^T _AB also depends on θ _i , and D ^A _AB is further
It also depends on γ _Ai . This means that the pattern distance and the geometric distance do not necessarily have a one-to-one correspondence, but this will be discussed later.

Hereinafter, the method of the present invention will be described in detail based on an example of a sound source position locating device in a pressure vessel to which a specific comparison and verification method is applied.

A preferred embodiment is shown in FIGS. 2 and 3.

As shown in FIG. 2, _N detectors S ₁ , S ₂ . . . It has been entered. This pressure vessel internal sound source position locating device 100 is constructed as shown in FIG. In FIG. 3, the detector S ₁ ,
The signals from S ₂ ,...S _N are sent to the amplifier 10 _i (however,
i=1 to N) are input to the computer 110. The computer 110 analyzes the signals from each amplifier and calculates the signal arrival time difference and peak value. The storage device 120 is stored in advance in which the signal arrival time difference and peak value of the acoustic signals output from each detector when each part of the pressure vessel is hit with a hammer or the like are associated with the hit position, that is, the known sound source position. It is now recorded as standard pattern data. In other words, let A _k be the known sound source position in the pressure vessel (where k = 1 to K), the peak value of the output signal of the i-th detector S _i corresponding to the known sound source position A _k be A _ki , and the signal arrival time difference be A k . It is now recorded as τ _Aki .

Further, when the signal from the unknown sound source B is inputted to the computer 110, the signal arrival time difference τ _Bi and the peak value b _i of the signal from the unknown sound source B are calculated in the same manner as described above, and further, From the standard pattern data in the storage device 120, the above (1), (3), (4)
The following equations (9) to (11) are calculated based on the equations, and pattern distances D ^C _Ak,B between the sound source B and each known sound source a _k are calculated.

D ^T _Ak,B = {Var(τ _Aki −τ _Bi )} ^1/2 …(10) D ^C _Ak,B = (1−α) D ^T _Ak,B +αD ^A _Ak,B …(11) Like this The operation of the embodiment configured as shown in FIG. 4 will be explained below along with the flowchart shown in FIG.

When a device start command is given in step 111, the computer 110 is activated, and in step 112, the output signals of the amplifiers 10 ₁ to 10 _N are taken in at a predetermined sampling period, and the data is recorded in the storage device 120. Ru. The data storage area of the storage device 120 is allocated, for example, equivalent to "1" kiloword corresponding to each detector, and when the number of samplings exceeds "1" kiloword, the oldest data is discarded and the data is stored. will be operated to write new data on it. Furthermore, in step 112, at the same time as the data sampling operation, it is constantly determined whether or not abnormal noise is occurring.
This determination is based on whether the sampled data exceeds a preset value. In addition,
This determination can be made by providing a comparator outside the computer 110 to determine whether each detector signal exceeds a set level, and if it does, the computer 110 is caused to respond by an interrupt signal. is also possible. Now, when the computer 110 detects the occurrence of abnormal noise, data sampling is stopped after T seconds in step 113. This time T is 20ms to 50m when targeting a reactor pressure vessel.
A value of about s is desirable.

Next, in step 114, the signal arrival time difference γ _Bi of the unknown sound source B and the peak value b _i (where , i=1~N)
is extracted. Further, in step 115, the standard pattern data is sequentially read out from the storage device 120, and the pattern distance D ^C _{Ak from the pattern data of the unknown sound source B is determined by equations (9), (10), and (11). B}
(However, k=1 to K) is calculated. The K pattern distances D ^C _Ak,B calculated in this way are:
As mentioned above, this is correlated to the geometric distance d between the unknown sound source B and the known sound source A _k . Therefore, in step 116, the one with the minimum value of the K pattern distances D ^C _Ak,B is selected, and the position number k of the known sound source A _k corresponding to the minimum value is displayed on the display device 130. As a result, it is known that the unknown sound source B is located closest to the known sound source A _k .

The effects of this embodiment will be described below with reference to illustrated data obtained through experiments using this embodiment.

In the experiment, various parts of the pressure vessel 1 shown in FIG. hitting position (sound source A,
B) Mutual geometric distance d and the above (1), (3), (4)
Each pattern distance D ^A _AB , D ^T _AB , calculated by the formula
The relationship with D ^C _AB is shown in FIGS. 5 to 7, respectively.

As is clear from FIG. 5 or 6, there is a strong correlation between the pattern distances D ^A _AB , D ^T _AB and the sound source distance d, and as the distance d becomes smaller, the pattern distance
D ^A _AB and D ^T _AB are also becoming smaller.

The pattern distance D ^C _AB shown in FIG. 7 can be calculated by setting the value of α in equation (4) as an unknown quantity, and using the data shown in FIGS. 5 and 6 to calculate the pattern distance D ^C _AB and the distance d. α is determined so that the correlation coefficient has the maximum value, and the calculation is performed based on the value of α. As is clear from FIG. 7, the correlation between the pattern distance D ^C _AB calculated in this way and the distance is even stronger than that between the pattern distances D ^A _AB and D ^T _AB . . Incidentally, the correlation coefficient in Figure 7 is 0.91.

FIG. 8 shows variations in the pattern distances D ^A _AB and D ^T _AB when the same point in the pressure vessel, that is, the distance d between the sound sources is zero. Originally, if the distance d is zero, the pattern distance should also be zero, but in reality it becomes a value other than zero due to noise mixed into the signal. However, compared to FIG. 5 or FIG. 6, D ^A _AB and D ^T _AB are sufficiently small values.

FIG. 9 shows the values obtained by calculating the pattern distance D ^C _AB from the data in FIG. 8 using equation (4) and integrating it in descending order of the value. From FIG. 9, it can be seen that, for example, the probability that the value of D ^C _AB will be 0.3 or less is 85% or more. Therefore, from FIG. 7, considering that there is no sound with a distance d of 1 m or more when the value of D ^C _AB is 0.3 or less, according to this embodiment, even though the sound sources are from the same point, is 1
It can be seen that the probability of erroneously identifying a distance of m or more is less than 15%. This has implications for important issues in foreign object detection within pressure vessels. In other words,
In pressure vessels such as nuclear reactors, sounds are generated when operating control rods or opening and closing valves, but it has been difficult to distinguish between these sounds and sounds caused by foreign objects. However, by applying this example,
By storing standard pattern data of sounds generated from control rods, etc. in advance, it is possible to determine whether an unknown sound source is a sound generated from a control rod, etc., or a sound caused by a foreign object.

As explained above, according to the present invention, there is an effect that the sound source position can be located with high accuracy.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of the positional relationship between the sound sources A and B and the detector S _i for explaining the present invention, and FIGS. 2 and 3 are device configuration diagrams of an embodiment to which the present invention is applied. , FIG. 4 is a flowchart for explaining the operation of the embodiment shown in FIGS. 2 and 3, and FIGS.
The figures show the relationship between the experimental values of the pattern distances D ^A _AB , D ^T _AB , and D ^C _AB and the sound source distance d, respectively. Figure 8 shows the pattern distance D ^A _AB obtained for the same sound source position and
FIG. 9 is a diagram showing the variation in experimental values of D ^T _AB , and is a diagram showing the rate of integration of experimental values of pattern distance D ^C _AB obtained for the same sound source position, starting from the smallest value. DESCRIPTION OF SYMBOLS 1... Pressure vessel, 100... Sound source position locating device in pressure vessel, 110... Computer, 120... Storage device, 130... Display device, S _i ... Acoustic detector.

Claims (1)

[Claims] 1. At least three acoustic detectors for detecting sounds generated within a pressure vessel are installed at appropriate locations related to the pressure vessel, and an acoustic signal from an unknown sound source is output from the detectors. at least one of peak value data and time difference data obtained by calculating the peak value of the signal and the time difference between the signal arrival times between the detectors, and a large number of previously known sound source positions corresponding to the data. By comparing and collating known wave height data and known time difference data stored in correspondence with each other to search for known data that is most similar to the data, it is possible to determine whether the unknown sound source position is closest to any of the known sound source positions. In the method for locating a sound source in a pressure vessel to determine whether the sound source is a and a second variance value consisting of the difference between the time difference of each of the detectors and the known time difference of each of the detectors, or a function of the variance value, and a pattern distance consisting of a mathematical average value of A method for locating a sound source position in a pressure vessel, characterized in that an unknown sound source position is located by calculating for each position and comparing each pattern distance. 2 In the invention described in claim 1,
A method for locating a sound source position in a pressure vessel, wherein the first dispersion value function is a square root of the dispersion value. 3 In the invention described in claim 1,
The first dispersion value function is the dispersion value divided by the product of the standard deviation of the logarithm of the peak value of each of the detectors and the standard deviation of the logarithm of the known peak value of the corresponding detector. A method for locating the position of a sound source inside a pressure vessel. 4 In the invention described in claim 1,
The first dispersion value function is the square root of the dispersion value divided by the product of the standard deviation of the logarithm of the peak value of each of the detectors and the standard deviation of the logarithm of the known peak value of the corresponding detector. A method for locating the position of a sound source in a pressure vessel. 5. The invention according to claims 1 to 4, wherein the mathematical average value is any one of an arithmetic average value, a weighted average value, and a geometric average value. Location method. 6 In the invention described in claim 5,
A method for locating a sound source position in a pressure vessel, wherein the weighted average value is obtained by setting one of the weights to zero. 7 In the invention described in claim 5,
A method for locating a sound source position in a pressure vessel, wherein the weighted average value is obtained by setting a weight related to a wave height value to be less than or equal to a weight related to a time difference.